MEETING SUMMARIES
(click on the meeting summary you wish to view)

Meeting Summaries

Joint Meeting of the Skin Diseases Interagency Coordinating Committee and the Diabetes Mellitus Interagency Coordinating Committee

National Institutes of Health
Natcher Conference Center, Conference Room G
Bethesda, Maryland
November 18, 2003

Dr. Saul Malozowski, Executive Secretary of the Diabetes Mellitus Interagency Coordinating Committee (DMICC) and Senior Advisor for Clinical Trials and Diabetes Translation, Division of Diabetes, Endocrinology, and Metabolic Diseases (DDEM), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) opened the joint meeting with the Skin Diseases Interagency Coordinating Committee (SDICC). After apologizing for the small size of the room due to the number of meetings being held at Natcher this day, Dr. Malozowski announced the topic of the joint meeting was wound healing and introduced Dr. Stephen J. Katz, Director, National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), lead agency for SDICC.

Dr. Katz welcomed the attendees and assured them that “It is not the size of the room that matters. It is the interactions that count.” He explained that the interagency committees are composed of members from other institutes and agencies of the Federal Government with a focus in a particular area, as well as voluntary organizations particularly concerned about each of these areas. Dr. Katz stated that NIAMS has an interagency working group on bone and one on lupus. Quarterly, and sometimes semiannually or annually, each of the lead National Institutes of Health (NIH) institutes, such as NIDDK or NIAMS, holds meetings. At these meetings, not only do the institutes come together but it affords them an opportunity to get together with the Centers for Disease Control and Prevention (CDC), other U.S. Department of Health and Human Services (DHHS) agencies, the Department of Defense (DoD), the Veterans Health Administration (VHA), and whoever else has a role in addressing the problem on the agenda. Thus the Federal agencies are doing things collectively, not redundantly or at odds with one another.

Dr. Katz pointed out that this seemed an opportune time to have a combined meeting with NIDDK and NIAMS because both are focused on and have an interest in wound healing, particularly in chronic wounds. He added that diabetes is the paradigm for many of the models that are used, both in humans and in animals. Dr. Katz thanked Dr. Alan Moshell and Ms. Geraldine Pollen from the NIAMS; Dr. Malozowski and Dr. Judy Fradkin, Director, DDEM, from the NIDDK; and Mr. Iain MacKenzie of the Hill Group for helping to put the joint meeting together.

According to Dr. Katz, as a consequence of a similar meeting in 1993, opportunities were identified and a request for applications (RFA) focused on wound healing was issued by NIAMS and supported by NIDDK. The RFA received a tremendous response. An important thing to note was that the RFA was sponsored by many groups, not only the NIAMS and NIDDK, but also by the National Institute of Nursing Research (NINR), National Institute on Aging (NIA), and National Institute of Child Health and Human Development (NICHD). He emphasized that there are many opportunities at NIH and many institutes that are particularly interested in this area. Over the past 10 years, NIAMS has trebled its investments in wound healing research to close to $11 million. Most of the focus is on basic research, although in the last few years, as some of those in the audience would know, there also has been a focus on clinical studies, particularly on gene therapy.

Dr. Katz stressed that the hope for a meeting like the current one was to focus on what the opportunities are, what other agencies are doing, and try to identify opportunities and gap areas for future action. For a long time, SDICC tried to develop or encourage guidelines for a standard of care against which new therapies should be compared. Dr. Katz noted this is not a simple problem, but one that, if the Federal agencies do not deal with it, whatever the responsibilities are, it will not get done. He concluded by saying that he looked forward to hearing the day’s presentations and discussions.

Before introducing Dr. Spiegel, Dr. Malozowski also thanked Dr. Moshell and stated that without his assistance, it would have been very difficult to put the agenda together for the meeting.

Dr. Spiegel reinforced Dr. Katz’s statement regarding this being an excellent opportunity for their two institutes to join with representatives of other institutes and agencies to discuss wound healing, a topic of great significance in and of itself, but particularly in relationship to diabetes. In addition to adding his thanks to those who put the meeting together, Dr. Spiegel wanted to underscore that the meeting typifies a reinvigorated spirit of collaboration among particular NIH institutes around the issues of diabetes complications. He said that diabetes is really a trans-NIH disease. In the area of complications, there are a variety of issues, such as a particular interest in wound healing in type 1 and type 2 diabetes, that bring together the various NIH institutes and centers, including, among others, the National Institute of Neurological Disorders and Stroke (NINDS); the National Heart, Lung, and Blood Institute (NHLBI); the National Eye Institute (NEI); obviously the life cycle institutes, both NIA and NICHD; and NINR. He assured the group that NIDDK intends to support research in wound healing in a collaborative way. Such support will range from basic mechanistic research in which NIDDK is spearheading new initiatives—for example, in angiogenesis, which underlies many diabetes complications and with animal models under NIDDK’s Animal Models of Diabetes Complications Consortium—through the translational area and, ultimately, to direct clinical testing in human subjects.

back to top

Dr. Spiegel added that, as many of those present knew, a special funding pool was legislated by Congress for type 1 diabetes. For fiscal years 2004 to 2008, funding is at $150 million a year for type 1 diabetes-related issues. For those who are interested, there is a special area on the NIDDK website (www.niddk.nih.gov) listing the many trans-NIH initiatives that have been created with these funds. He stated that he also was looking forward to the day’s presentations and discussions.

From the speakers’ presentations and the discussions on the epidemiology of diabetic foot ulcers, pathways to development of foot ulcers and the need for offloading, what has been learned from data on foot ulcers that would contribute to future clinical trial designs, clinical protocols for treating diabetic foot ulcers, and adjunctive therapies for foot ulcers, the following key opportunities and recommendations were identified for possible future action.

• Agree on a definition for diabetic foot ulcer.

• Identify consistent ICD codes for diabetic foot ulcers and other lower limb conditions to enhance data collection and further collaborative research.

• Issue a mandate to treat diabetic foot ulcers before complications develop.

• Establish the goal of reducing amputations related to diabetic foot ulcers to less than 10 percent.

• Establish as part of the standard of care for diabetes patients that their feet be examined at each office visit.

• Establish aggressive debridement and offloading as the standard of care for diabetic foot ulcers.

• Use a simple test to measure neuropathy (i.e., lack of sensation) in diabetic patients.

• Use the presence of neuropathy as a surrogate marker for the endpoint of ulceration in clinical trials.

• Establish surrogate markers, such as size, grade, and duration of wound, based on current FDA data, as predictors of healing to conduct phase 1 and 2 clinical trials of shorter duration and without the endpoint of complete healing.

• Design more efficient clinical trials based on epidemiologic studies and tools.

• Require offloading, preferably using an instant total contact cast, for all treatment groups in clinical trials testing new therapies for chronic wounds.

• Investigate further the refractory subset of patients who do not heal as expected and find ways to reverse the reasons for this refractory subset.

• Recognize the importance of the foot and the impressive work being done by making 2005 “the Year of the Foot” with the goal of implementing meaningful strategies in diabetic foot ulcer care.

• Organize a consensus conference to discuss diabetic foot ulcer issues such as guidelines for improved primary care and clinical practices.

Citing the late Dr. Paul Brand’s statement that “Pain is God’s greatest gift to mankind,” the group agreed that neuropathy resulting in lack of pain was seen as a major contributor to the development of foot ulcers in persons with diabetes and to the non-compliance with offloading devices that are essential to healing. Other prominent risk factors are peripheral vascular disease, past history of neuropathic foot ulcers, microvascular complications, elderly patients living alone, foot deformity, and unilateral amputees.

Topics of particular interest were photographing, measuring, and grading of the diabetic foot ulcer on a patient’s chart; the need to work with a multi-discipline team in treating the person with a diabetic foot ulcer; investigating factors contributing to chronicity in diabetic foot ulcers such as unresponsive or senescent cells present either in the wound or in the callous around the wound, a proteolytic or inflammatory environment, deficient or unavailable growth factors, or bacterial interference and therapies to address these; the use of adjunctive therapies such as platelet-derived growth factors, bilayered living skin constructs, dressings to remove or change the proteolytic environment, other anti-inflammatory agents, and removal of bacterial interference; the possibility of a biochemical assay of fluid from a wound to predict healing; recognizing and addressing the presence of stress and depression in the diabetic patient and its effect on compliance with prevention and therapies and with healing; use of sterile maggots as a debriding agent and honey or heat as healing therapies; encouraging the CDC to provide regular surveillance data from all the surveys being conducted annually, including VHA and DoD data; providing information to primary care providers and patients on NIH websites regarding standards of care, wound treatment therapies, and expectations for healing; suggestions for clinical trials and other studies; and the availability of various funding mechanisms from the NIH to further research in wound care.

The following sections detail the presentations and discussions leading to the above recommendations.

back to top

Epidemiology of Diabetic Foot Ulcers

Dr. Malozowski introduced Dr. Gayle Reiber, Veterans Administration (VA) Career Scientist and Professor, Health Services and Epidemiology, University of Washington, Seattle, who has spent 15 years of her career on the epidemiology of diabetic foot ulcers.

Dr. Reiber outlined five topics for her presentation: (1) a consensus definition for foot ulcers; (2) CDC foot ulcer history findings; (3) updated risk factors for foot ulcers; (4) current national foot ulcer data; and (5) issues for epidemiologists.

Definition of Foot Ulcers. Dr. Reiber adapted her definition from a Lazarus definition published in 1994 by inserting the italicized phrase: A foot ulcer is “a cutaneous defect extending into the dermis or subcutaneous tissue that does not undergo self-repair in a timely and orderly manner within 4 weeks of onset.” When physicians see a foot ulcer, they know it is a foot ulcer, but some of these lesions are on the periphery. They don’t heal. They are deep. Dr. Reiber stated that foot ulcers need to be better classified and offered her definition to aid the future collaborative endeavors of the group.

BRFSS History of Foot Ulcers. Dr. Reiber praised the Behavioral Risk Factor Survey (BRFSS) efforts in collecting nationwide information on risks. (See Centers for Disease Control and Prevention’s (CDC’s) online newsletter, Morbidity and Mortality Weekly Report (MMRW), November 14, 2003, at www.cdc.gov/mmwr.)With an annual budget of approximately $12.5 million dollars, it is the world’s largest telephone survey. Each year a relevant core question is, “Do you have diabetes?” More detailed diabetes information varies annually. Between the years 2000 and 2002, 44 different States and the District of Columbia put an extended module on the BRFSS and asked, “Have you ever have any sores or irritations on your feet that took over 4 weeks to heal?” There was an unusually good response rate for a telephone survey of close to 58 percent. The CDC adjusted the data for age, sex, race, and ethnic background. Significant results related to foot ulcers in those diagnosed with diabetes included the following:

• There is a history of foot ulcers in 12 percent of adults with diabetes.

• There was an increase in annual foot examinations from 56 percent to 62 percent between 1995 and 2002.

• Duration of diabetes, smoking, and insulin use each increase significantly risk.

• People at increased risk also include those who are obese and those who are not married or cohabitating, which has been shown in a number of studies.

• There was a significant protective effect with age in those over 64-years-old.

• Surprisingly, based on race and ethnicity, blacks had a protective effect and Hispanics’ odds were a little higher, but not significantly higher.

• There was no significant difference between men and women (p value = 0.2).

Dr. Reiber cited several BRFSS limitations. The survey does not cover people in institutions such as nursing homes or hospitals. Participants self-report and there is no consistent definition of foot ulcer, so persons are often confused about whether they have had a foot ulcer or a less severe lesion. Persons must have a land line telephone (not a cell phone only or no phone) to be included in the survey. Finally, possibly because of confounding for a number of variables that could not be analyzed in the data set, several risk factors did not reach statistical significance.

Another CDC finding cited by Dr. Reiber was that the number of people with diagnosed diabetes increased between the years 1980 and 2001 (National Health Survey Data from the CDC Division of Diabetes Translation). As diabetes increases, physicians can expect to see more foot ulcers. Dr. Ramsey (Ramsey, SD, Newton K, Blough D, McCulloch DK, Reiber, GE, and Wagner, EH. Incidence Outcomes and Cost of Foot Ulcers in Patients With Diabetes. Diabetes Care 1999, 33:282-387) and Dr. Moss (Moss, SE, Klein, R, and Klein, BEK. The Prevalence and Incidence of Lower Extremity Amputations in a Diabetic Population. Arch Intern Med 1992, 152:610-616) reported that the incidence or onset of new foot ulcers is about 2 to 2.5 percent in the diabetic population. Dr. Ramsey looked at the Group Health population in Seattle; Dr. Moss, the Wisconsin population. Dr. Moss reported a 10 percent prevalence.

Risk Factors for Foot Ulcers. Dr. Reiber pointed out that, whenever measured, neuropathy is uniform across studies, which makes it a highly important risk factor. Dr. Reiber referred the audience to the following reference for more information: Reiber, G.E. and Ledoux, W.R. 2002. Epidemiology of Diabetic Foot Ulcers and Amputations: Evidence for Prevention. In The Evidence Base for Diabetes Care, edited by Williams, R., Herman, W. Kinmonth, A.-L., and Wareham, N.J. John Wiley & Sons, Ltd. Other independent risk factors are ankle arm index (AAI) and transcutaneous oxygen tension (TcPO2), two different components of the circulation; prior ulcer (typically 60 percent of those presenting with foot ulcer); peripheral vascular disease (PVD); prior lower extremity amputation (LEA); and increased HbA1c.

back to top

Based on U.S. population sample data from the National Health and Nutrition Examination Survey (NHANES) between 1999 and 2000, prevalence of insensate feet is about 2-fold greater in people with diabetes and prevalence of insensate feet plus peripheral neuropathy (PN) symptoms is about 3-fold higher in people with diabetes than in those who are not. Peripheral artery disease (PAD) is 2-fold higher and if all the different lower extremity diseases are combined, they also are 2-fold higher in the population with diabetes compared to those without diabetes. Dr. Reiber remarked that it is really quite astounding the different ways that PN is measured and reported in studies. Measures that are used are signs, symptoms, nerve conduction abnormalities, and composite definitions. Depending on whether a definition like nerve conduction deficit or whether sensory examination reflex is used, there is a 73-fold difference in prevalence of neuropathy. However, despite the fact that it is reported very differently, neuropathy consistently shows up. In the type 1 Diabetes Control and Complications Trial (DCCT) baseline information, the gold standard was a 2.1 prevalence. A study done by Partanen (Partanen et al., New England Journal of Medicine, 1995) reported an 8.3 percent prevalence of neuropathy at baseline and 42 percent at 10 years in persons with type 2 diabetes.

New Outpatient and Inpatient Data. Along with standardizing a definition for foot ulcers, Dr. Reiber recommended that it would be valuable for researchers in Government agencies to use the same ICD-9 codes for foot ulcer conditions and amputations in order to share and compare data and better understand the findings. Among the many different ICD codes used for foot ulcers, the most common one is the 707 code. This is “ulcers for lower limbs except for decubitus” and that is about 55 percent of foot ulcers, with osteomyelitis, cellulitis, and abscess being among the conditions that make up the rest.

The VA, which has information on all individuals who come in for outpatient and inpatient care, does not use ICD codes for foot ulcers. They have their own list that includes varicose ulcers and lower leg conditions that are not foot ulcers. Also, since the denominator is veterans who are usually older males, the VA rates are much higher than those in the U.S. population as a whole. In the VA data, 9.4 percent of those in the youngest age group (0-44) had foot ulcers, 17.2 percent of persons in the 45-64 age group had foot ulcers, whereas those 65 and older dropped to 15 percent. Combining the VA’s 115,000 episodes in 2001 with the 184,000 discharges reported in HCUP, for a total of 299,000 episodes in persons with diabetes, rates were 2.4 percent in the 45-64 age group, 2.8 percent in the 65-74 age group, and 3.3 percent in those over 75, showing an increase with age. Comparing 1993 to 1997 data with that from 2001, Dr. Reiber noted that there was very little change. She also provided VA and HCUP 2001 data on amputations in 55,000 persons as indicative of the prevalence of foot ulcers since 84 percent of amputations are preceded by a foot ulcer.

Epidemiology Issues. In conclusion, Dr. Reiber asked the group to consider the following requests:

• Agree on a definition for foot ulcer.

• Agree on ICD codes for foot ulcers and other lower limb conditions to further collaborative research.

• Encourage the CDC to provide regular surveillance data from all the surveys being conducted annually, including VA and DoD data.

• Recognize the importance of the foot and the impressive work being done by making 2005 “the Year of the Foot” with the goal of implementing meaningful strategies in diabetic foot ulcer care.

Pathways to Foot Ulcers and the Need for Offloading

Dr. Malozowski introduced the next speaker, Dr. Andrew J.M. Boulton, Visiting Professor of Medicine, University of Miami, Professor of Medicine, University of Manchester, and Consultant Physician, Manchester Royal Infirmary, Manchester, United Kingdom

Dr. Boulton said he left Miami 10 days ago and had been in Rumania for World Diabetes Day and then Eastern Germany and England and Ireland. Given this travel and his presentation for today, he wryly assumed this made him a specialist as defined by his sponsor when he was a medical student: “A specialist is someone who comes from a long way away and brings slides.” He added that in 2005, the World Diabetes Day indeed will be focused on the foot.

Pathways to Foot Ulcers. Dr. Boulton presented a number of studies that illustrated the pathways to foot ulcers. He also put neuropathy first on his list of risk factors, noting that, in research done in combination with Dr. Reiber, this was the most important component cause, although neuropathy alone will not lead to ulceration. Other factors are peripheral vascular disease, past history of neuropathic foot ulcers (with a recurrence rate in studies of between 40 and 60 percent a year), microvascular complications, elderly patients living alone, foot deformity, and unilateral amputees, who are obviously at great risk.

Dr. Boulton showed slides illustrating that to some extent “Nothing is new in medicine.” In the 1880s, Dr. Pryce, a surgeon from Dr. Boulton’s hometown of Nottingham, looked at a series of patients and published in Lancet in 1887 the statements: “Diabetes itself may play an active part in the causation of perforating ulcers….And it is abundantly evident that the actual cause of the perforating ulcers was a peripheral nerve degeneration.” He also quoted the late Dr. Paul Brand who said “Pain [is] God’s greatest gift to mankind.” Dr. Brand, who passed away in July 2003, was the noted physician who described in leprosy the association between insensitivity and foot ulcers, rather than the cause of the ulcers being the infection itself. Dr. Boulton stressed that the importance of pain is something we forget. He stated that not having sensation really summarizes many of the problems seen in the diabetic foot in 2003, with regard both to pathways and offloading.

back to top

The United Kingdom Prospective Diabetes Study (UKPDS) study of 5,000 patients indicated that 11 percent of patients have significant neuropathy at diagnosis of type 2 diabetes. At Dr. Boulton’s center, which participated in the UKPDS, neuropathy was possibly asymptomatic in up to 50 percent of patients. He reported that in population-based community studies of 811 older patients (mean age of 65 years) with type 2 diabetes, probably over half have risk factors for foot ulceration, 42 percent have clinical evidence of significant neuropathy, and 11 percent have peripheral vascular disease (see Kumar et al., 1994; UKPDS, 1998).

When Dr. Boulton moved to Manchester in 1988 and opened the new diabetes center, there were no longitudinal studies proving that neuropathy is a major risk factor for foot ulcers. At the new center, neuropathy was assessed by a simple semi-qualitative measure—by vibration dysfunction measured by biothesiometer, a handheld vibration device that has a gradiated reading of 0 to 50, the higher the reading the more severe the sensory loss. The center’s 469 patients, who all received education, were followed for 5 years. Those with vibration perception less than 15 were considered to have no neuropathy; those over 25 had undoubted neuropathy. The risk per patient per year of developing foot ulcers went from 0.7 percent in those with no clinical neuropathy and then increased 7-fold to almost 1 in 20 in those who had neuropathy (see Young et al., Diabetes Care 1994; 17:557).

In a larger multicenter European and North American study with 1,035 patients, published in 1998 (Abbott et al. Diabetes Care 1998:21:1071), the entry criteria were a diagnosis of diabetes, no history of foot ulcers, definite risk measured by a vibration perception of more than 25 volts, and no evidence of vascular disease. The patients received standardized education and three monthly visits to the podiatrist. Yet, the annual incidence of foot ulcers in these patients was even higher than what had been reported earlier—7.2 percent. This study also showed that the foot ulcer risk increased significantly with an increment grade rise in vibration perception threshold (VPT)—5.6 percent per volt VPT.

Another Manchester study that Dr. Boulton thought was important in terms of surrogate endpoints for use in clinical trials, enrolled 169 patients and 22 controls with a spectrum of neuropathic deficits and followed them for 6 years (see Carrington et al. Diabetes Care 2002;25:2010-2015). Thirty-seven percent developed ulcers, 11 percent received amputations, and 18 percent died. The best predictor of the endpoint of neuropathy was motor nerve conduction velocity (MNCV) in the peroneal nerve. Dr. Boulton called this the missing link, because this is electrophysiology where the speed of nerve conduction can be directly measured in the patient, not a psychophysical test, and highly reproducible, with a low coefficient of variation. Dr. Boulton’s group had previously shown that MNCV correlates very well with structural abnormality in the sural nerve and is a predictor of clinical neuropathy. He noted that it is also a predictor of mortality, as is foot ulcers, which might explain the rather lower prevalence in the study Dr. Reiber cited of patients over 75, because these are the survivors. Dr. Boulton emphasized that the Manchester group’s prospective study confirmed MNCV as a predictor of the endpoint of insensate foot ulceration and indicated that electrophysiology is the best surrogate endpoint for use in clinical trials, which is important because agents are needed to teach diabetic neuropathy.

Dr. Boulton commented that researchers in the United Kingdom have the advantage in conducting population-based research. The government funded his University of Manchester group’s North West Diabetes Foot Care Study (NWDFCS), an ongoing population-based prospective study of 16,000 patients in six health-care districts to see what can be used in clinical practice as a predictor of foot ulcers. All known patients with diabetes were selected from general practitioner records. (An earlier study of just under 10,000 patients was published previously (see Abbott et al., Diabetic Med 2002;19:377)). In clinical practice, most providers have a tendon hammer, a tuning fork, a pin, and so on, so a very simple test can be done by trained research nurses and podiatrists in all the patients wherever they are seen, either at their homes, in primary care, or in secondary care. Dr. Boulton stressed that sophisticated equipment is not needed to identify the at-risk foot. There are simple tests that are applicable anywhere, such as in Romania where they often do not have sophisticated equipment.

For the NWDFCS, a simple neuropathy disability composite score was used with three sensory modalities—a vibrating 128 Hz tuning forks over the apex of the hallux, pin-prick (“can you feel the difference between sharp and blunt”) on the dorsal distal hallux, just proximal to the great toe’s nail, and similarly, hot-cold rods. If the patient cannot feel the vibrating 128 Hz tuning fork, or cannot distinguish the pin-prick, or cannot tell the difference between hot or cold, the score is one. Any hesitation, any suggestion of abnormality, the score moves towards abnormality, or a one. So if the patient has to think whether or not he/she feels a tuning fork on the hallux, then that is abnormal. Delay is abnormal. So that’s a maximum score of three abnormalities per foot or six altogether. Next ankle reflex, which is a good predictor, is tested. Normal is zero; if absent, the score is two, and if present on reinforcement, that scores a one. So the maximum score is 3 from the first test and 2 from the ankle test for a possible total of 5 per leg or 10 for both legs. Dr. Boulton repeated that this is a very simple procedure that can be done quickly in clinical practice; in fact, more quickly than it takes to describe it.

In the earlier NWDFCS, the 9,710 patients were seen in their homes, given a half-hour one-on-one education session that was appropriate to their identified level of risk, and followed for 2 years. They developed 291 ulcers, with a higher rate in males, which is a uniform finding in Western countries. The best single predictor of risk for foot ulcers was the very simple neuropathy disability score that Dr. Boulton had just described and had assured the audience was applicable in clinical practice anywhere in the world. At less than 6, the patient had a 1 percent annual risk of developing an ulcer. At 6 or more, there was a 6-fold increase in risk.

In a joint study with Dr. Reiber (Reiber, Vileikyte et al., Diabetes Care 1999;22:157-162), the two groups looked at approximately 155 incident ulcers, and using a Rothamn model, identified key components leading to the ulceration. In the pathway to ulcers, the most important component cause was neuropathy. Four out of five patients had significant neuropathy. The critical triad most commonly seen (present in 63 percent of patients with foot ulcers) was neuropathy, deformity (commonly clawing of the toes, prominence of the metatarsal heads), and trauma (often caused by inappropriate footwear). More than 80 percent of these ulcers were potentially preventable. The pathway again began with lack of pain. Asked about their feet, patients replied “I feel great” because of sensory loss. These were patients with a baseline pathology of neuropathy, along with the rest of the triad, deformity and trauma. Dr. Boulton stated that it is the job of practicing physicians, nurses, podiatrists, or other healthcare providers, to prevent these three events coming together in the same individual.

back to top

Other studies conducted over the 10-year period from the 1980s to the 1990s and referred to by Dr. Boulton also indicated the importance of foot pressure abnormalities and loads under the neuropathic foot (Boulton et al. Diabetes Care and Diabetic Medicine 1983, 1984, 1985, 1986; Veves et al. Diabetologia 1992;35:660-664; Murray et al. Diabetic Medicine1996;13:979-982) These studies showed high foot pressures are associated with first and recurrent ulcers, foot pressure abnormalities occur early in diabetic neuropathy, high foot pressures are a predictor of ulcers, and the presence of callous or of hard skin under the metatarsal heads is associated with high pressure and is a predictor of ulcers.

Need for Offloading. Dr. Boulton observed that a number of factors are known to enhance wound healing, so the job of physicians is to try to correct the underlying condition. He noted that today providers are controlling infection; providing vascular reconstruction through distal bypasses and angioplasty for patients with severely compromised peripheral circulation; and helping patients with their glycemic control, at least in the short-term, although the long-term is still a challenge. However, they have failed miserably at offloading. When asked “What dressing should I use?” Dr. Boulton would reply, “It’s not what you put on the ulcer that matters. It’s what you take off it.” He recommends aggressive debridement. Dead tissue, the callous, needs to be removed. In a clinical trial of platelet-derived growth factors (PDGFs) and a placebo, Dr. Boulton noted that where 100 percent debridement was performed at every office visit, there was more rapid healing (Steed et al., J AM Coll Surg 1996;183:61-64). Secondly, pressure must be removed.

Dr. Boulton listed a variety of devices that have been used over the years to remove pressure, including an outcast boot, a sort of crow-walker, half-shoe, special shoes, orthoses, bed rest, a wheelchair, crutches—none of which work. When asked “How can the patient be so silly? They’ve got this huge hole in their foot and they’re walking on it? Why?” he replies, “Because we don’t appreciate what it is not to have sensation.” As Dr. Paul Brand taught with regard to leprosy, pain is God’s greatest gift to mankind. Lack of pain results in a foot ulcer. If a patient’s foot does not hurt, he/she will follow the physicians’ offloading instructions to stay in bed or in a wheelchair or use crutches for a day or two, but that is all. Dr. Boulton pointed out that we live in a very strong, lay-driven, symptom-driven society. When patients go to the doctor with pain, treatment is given and they usually get better. But a patient does not go to the doctor and say, “Look, I can’t feel under my feet.” Medical schools do not teach how to take care of people who have lost sensation. So patients get out of bed and throw their crutches away.

The best tools for the foot ulcer patient, according to Dr. Boulton, are the removable cast walker and the total contact cast (TCC), which has many advantages. The TCC enforces adherence (compliance), shortens stride length, decreases cadence, reduces activity, and reduces peak pressures equal to that of a removable cast walker. He cited two major trials: Mueller et al., 1989, that compared the TCC with standard treatment, and Armstrong et al., Diabetes Care, 2001, that compared the TCC with a removable cast walker and a half shoe. At every stage, there was more healing in the total contact cast group. Even though the pressure offloading effect of the TCC and the removable cast walker is equal, the removable cast walker performed much less well.

In another study to try to understand the superior performance of the TCC over the removable cast walker, Dr. Boulton described activity patterns in patients with diabetic plantar foot ulcers. Patients were given a removable cast walker to wear, told that wearing the device was very important in all activities day and night, and told they would be monitored 24x7 with a pedometer or accelerometer, which they wore. Another accelerometer was placed inside the cast walker, without their knowledge, so their activity level with or without the cast walker could be monitored. The removable cast walker is, indeed, removable. Less than 30 percent of the footsteps taken per day by patients in the study included wearing the cast walker.

Dr. Boulton said there are, however, disadvantages to the TCC as the gold standard because, when casting an insensitive foot, it is possible to cause damage such as a new ulcer. In addition, using the TCC requires a lot of time, equipment, and new removable casts. Patients must return every week to be checked and have the cast replaced, so the cost is high. In place of the TCC, he suggested an “instant total contact cast” using the DH walker, which offloads as well as the TCC, and can be made irremovable by wrapping it in a Scotchcast. This creates a device that can be used every week. The patient can come back, have the inexpensive outer cast cut off, the HR removed, get the wound looked at, have the wound treated and debrided, and get the same HR cast put back on and rewrapped with one band of cast. The result is a significant cost-saving.

Addressing the question of why trials of removable devices are so disappointing, Dr. Boulton described an event some years ago, in the United Kingdom’s National Health Service, where patients were provided free footwear. Only 20 percent of patients reported that they wore it regularly. In a trial with the standard DH walker, it was worn for only 28 percent of daily activity. Despite all good intentions, offloading devices are not used. If they can be removed, they will be. He proceeded to describe two current key trials. One is a randomized trial in Miami testing the TCC versus the instant total contact cast based on the hypothesis that the healing rates of the two should be the same. To date, this appears to be so with 35 patients. A second trial in Tucson testing the instant TCC with a removable cast walker is based on the hypothesis that healing with the instant TCC should be faster.

back to top

Dr. Boulton strongly recommended a paradigm shift in conducting studies of new therapies for neuropathic foot ulcers. To date, most of these studies have not attended to offloading. This factor probably explains the very disappointing results of many new modalities that are being put out in the community, such as growth factors, artificial skin, and so on. He stated there is no doubt that the huge confounding variable is pressure. Patients are walking on the wounds. Hence, they are not responding in either the placebo or treatment groups. Dr. Boulton proposed that all future trials of therapies for plantar neuropathic ulcers should have standardized offloading in all treatment groups. He believes support for his proposal will be forthcoming in a paper published by Dr. Albert Piaggesi of Pisa, Italy, in November 2003 in Diabetes Care. Dr. Piaggesi’s group conducted a randomized trial of patients with chronic plantar foot ulcers in Pisa, Italy. Patients with chronic foot ulcers were randomized to a TCC for 20 days followed by an ulcerectomy or to an ulcerectomy at baseline. Comparing histological changes between the two groups, Dr. Piaggesi found that the group who had the ulcerectomy on day 1 had much more chronic inflammation with fibrosis, less angiogenesis, and more inflammation. Those who were casted for 3 weeks prior to the ulcerectomy had more evidence of angiogenesis and granulation, more like an acute wound. Dr. Boulton deemed this a very important study showing that pressure not only has a direct effect on the ulcer, but also appears to support the chronic inflammation. Prolonged repetitive pressure appears to contribute to the chronicity of diabetic foot ulcers. After pressure release, the diabetic foot ulcer in many ways resembled an acute wound.

Dr. Boulton offered the following summary conclusions, which he suggested would be discussed further during the meeting:

• Wound healing in diabetes is impaired.

• Multiple factors have shown to be impaired in diabetic wound healing.

• Cellular differences have been noted between acute and chronic wound healing.

• Failure to offload pressure from plantar neuropathic ulcers is a major contributory factor in ulcer chronicity and may explain disappointing results for potentially exciting new treatments.

To illustrate the truth of one of the aphorisms of Professor J. A. Lindsay of Belfast in Northern Ireland that Dr. Boulton thought was applicable to the diabetic foot, he quoted Dr. Brand again. Professor Lindsay had said in the early 20th century “For one mistake made for not knowing, ten mistakes are made for not looking.” When Dr. Brand was invited to address the American College of Surgeons more than 20 years ago, a surgeon asked him, “What’s the most important thing that we can do to reduce amputations in diabetes?” The audience was expecting to hear about some new ultimate device or test such as a CT scan. But Dr. Brand replied, “Take the patient’s shoes and socks off every time you see them and look at their feet.” Dr. Boulton recommended that all remember this.

Clinical Trial Design: Lessons Learned From 20,000 Foot Ulcers

In introducing the next speaker, Dr. David Margolis, Associate Professor of Dermatology and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Dr. Malozowski commented that the number of foot ulcers had probably increased since the topic was first discussed.

Dr. Margolis agreed that the number is always changing. He opened his presentation on data that has been used for the past 3 to 4 years in designing clinical trials by crediting a few of the many collaborators he had worked with at the Center for Clinical Epidemiology and Biostatistics: Drs. Jesse Berlin, Ole Hoffstad, and Lynne Taylor who had been intricately related to the studies that would be presented, and also Drs. Jill Knauss, Jonathan Kantor, Joel Gelfand, and Brian Strom. The particular data set that Dr. Margolis would be talking about was from Curative Health Services (CHS), which had funded his first exploratory look at the data set with an unrestricted grant in 1998 and 1999. CHS is a medical management system that has been in existence for about 12 or 13 years, during which time they have managed more than 150 centers in the United States, with centers entering and exiting the system. Currently CHS manages approximately 90 centers. Dr. Margolis added that he had also received funding from both NIDDK (DK 59154) and NIAMS (AR 02212 and AR 44695). The focus of his presentation would be on the not-so-positive success rates seen in clinical trials of diabetic neuropathic foot ulcers. These success rates are that only 30 to 40 percent of the patients will heal in about 20 weeks. As suggested by Dr. Boulton, these poor success rates may certainly have to do with lack of offloading. These success rates were used when products such as growth factors (of which only one has been approved) and cell-based therapies (of which two have been approved for diabetic neuropathic foot ulcers) were conducted.

Dr. Margolis remarked that today a fair number of studies are being done in 12 weeks rather than the previous 20-week standard. Again, about 25 percent of the patients will heal at 12 weeks, and about a third will heal at 20 weeks. Given these results, Dr. Margolis’ group addressed the CHS database of information from their wound care centers with the question: Are there issues with foot ulcer healing that clinically would be easy to examine and use to make predictive or prognostic estimates of who is going to do well?

Patients in the various CHS-managed centers are treated by similar algorithms, which means that Curative goes to the centers and tries to educate physicians, nurses, and other healthcare providers about how diabetic foot ulcers and other ulcers should be treated. CHS has also maintained a database during most of their 12 years on the results of basic patient information and assessments that CHS required. Initially the database was used to assess center performance and suggest ways to improve it. Because of the long-term existence of the database, there are more than 20,000 individuals with diabetic neuropathic foot ulcers within the data set.

back to top

The first thing Dr. Margolis’ group did with the CHS data set was to try to take information from the data set and find people that truly had diabetic neuropathic foot ulcers, meaning they had diabetes, had a foot ulcer, and primarily had neuropathy and did not have significant arterial disease. Different schemes were tried in order to extract the information on patients. First, they looked at the diagnostic codes used, which differ from ICD 9 codes. They then actually pulled charts to see if patients had certain attributes and created constructs and criterion that patients needed to have to confirm the diagnostic assessment that they had made from the codes.

Dr. Margolis reported that in one of the earliest studies covering patients seen between 1988 and 1997, based on conditions possessed by the patient, they could correctly code a person in the database as having a diabetic neuropathic foot ulcer about 93 percent of the time, as verified by the patient’s chart. Dr. Margolis stated that the positive predictive values derived from these early studies are very good for making clinical decisions based on large administrative databases.

One of the classical, analytic epidemiology assessments his group did initially was look at events they knew were well coded in the database, such as sex, age, prior ulcers, number of ulcers, age and size of ulcer, and grade of ulcer, and try to see whether or not these attributes were good predictors of an individual’s likelihood of healing at 20 weeks. For example, grade was predictive. Wounds were described using a grading and staging system from grade 1 to grade 6 that was similar to other grading systems in use, with the biggest difference being that the CHS system has been used with more than 20,000 individuals, not just a few hundred, and it has been used in more than 150 different centers by probably close to 500 different healthcare providers. If one goes back and looks at clinical trials that have been done, the vast majority, if not all the clinical trials, certainly the ones that have gone to the Food and Drug Administration (FDA) and been used in their clinical application, involve grade 1 or grade 2, not these higher grade, wounds. Another predictive issue, especially with venous leg ulcers, was that the older or larger a wound was, the less likely it was to heal.

From a health service perspective, an issue that was important to Dr. Margolis was that regardless of the type of regression analysis done, whether fixed effects modeling or random effects modeling or GEE-based (generalized estimating equations) models, they were unable to show any center-based effect. This means that the usual concern in designing large, multicenter clinical trials about the different centers applying the protocols somewhat differently and thus creating important center-based effects on how therapies work did not happen in this case. Whether this is due to everybody in this large system being easy to educate, all getting the message, and all doing the same thing or everybody being uneducable and doing an equally bad job is unknown. However, they were all fairly equal in these multiple centers throughout multiple different states.

Dr. Margolis’ Center then considered not only individual risk factors for healing, but groups of factors. This can be important in designing a clinical trial because of the inclusion and exclusion criteria used for patients, which one would assume might cause different rates of healing in different types of patients. Fairly complicated models or fairly simple mathematical models were used in this analysis. In all cases, the analysts had a fairly reasonable ability to discriminate. Dr. Margolis explained that what an area under the ROC curve (AROC) of 0.70 basically means is that given two patients, about 70 percent of the time it is possible, using this model, to correctly differentiate who will heal and who will not 20 weeks later. Another assessment used was the Brier score.

Dr. Margolis pointed out that using these prognostic models is still fairly complex for the average physician, because it requires adding up risk factors and making other statements, which can be confusing. However, his group generated a prediction chart with dichotomized risk factors for diabetic neuropathy foot ulcers. On this chart, they selected points they thought were important in terms of discrimination such as the area, age, and grade of the wound. A patient could either have a risk factor or not have risk factor. Again, this is important in designing a clinical trial because the designer can determine how well the control arm is going to do based on simple risk factors. Thus, a trial could be designed in which it is predictable that two-thirds or one-third or all patients will heal based on their initial risk factors and the inclusion and exclusion criteria used. This selection of criteria is very important when comparing different trials or different results.

An important factor in a clinical trial is being able to recruit patients. Because the CHS system is large and diverse, it is feasible to use its data to generalize to patients in the general population and to those who are likely to participate in a clinical trial. Using the data from the large CHS system, it is possible to create good prediction models that discriminate well and are well calibrated. Applying these models, it is possible to create cohorts of study subjects who are more or less likely to heal with standard therapy, which is critically important in a randomized clinical trial. This information enables the designer to create more honest power calculations. In the CHS group, there was minimal center effect, which is also important for power calculations. This may be more important from a health services perspective in terms of understanding who is going to do well in different systems and who is not. The one problem Dr. Margolis had with the data set was that some of the information is not digitally coded, it is free text. For example, wound location information combines free text with a computer-ready variable. Thus, it can be determined that 80 percent of the wounds are forefoot wounds, but identifying wound location in individual patients is not necessarily known. With new funding, the group hopes to resolve this problem and others by properly coding the variables.

In funding the database analysis project, Curative originally had wanted the university’s Center to evaluate a proprietary product that they had that was called platelet releasate, which they no longer sell, and determine whether or not patients who received their platelet releasate were more likely to heal. One of the different components in the platelet releasate was PDGF, which is also the active component in one FDA approved cytokine. The Center did a propensity score study. This was done about 4 or 5 years ago just as propensity score studies were being accepted. There are now multiple studies like this. What is done in a propensity score study is try to determine why someone was selected for treatment and thus model selection bias.

back to top

For the Curative patients, multiple groups were created with relative risks and the effects on healing examined. There was an approximately 14 to 60 percent increased likelihood that an individual was going to heal when they received platelet releasate, depending on which propensity category they were in. The important result of this study, according to Dr. Margolis, was that as his group was able to know the propensity categories better, they realized that these closely paralleled the severity indices that they had developed earlier for these relative risk factors. The patients in groups one and two were all patients who had wounds that were just a few centimeters, were young wounds, and were early grade wounds of two or less, which actually were the same as the patients who were in the randomized PDGF pivotal trials as well. An interesting point is that the propensity score of about 24 percent was not all that different from the 30 percent severity index in some of the PDGF pivotal trials. What to Dr. Margolis was more interesting was that patients who were not included in randomized clinical trials because their wounds were more severe may actually have benefited more from the system, but they were not studied. Again, in designing a randomized clinical trial using cytokines or other cell substitutes, based on this information, it would be important to think carefully about what patients to include and exclude.

The next thing that Dr. Margolis’ group began to do with the CHS data was to evaluate whether or not there had been a change in healing rates over time. This was of interest to them because around this time period (1998-1999) is when the new therapies came out and others would also want to know if patients were not only healing but healing faster with the new therapies. Looking at the Curative system in the early years, about a third of the patients healed. If you looked at it in the later years, about half of the patients healed. However, most of this improvement occurred well before any new therapy existed. In fact, looking at the data in greater detail shows that the percentage of patients with good prognostic signs just increased over the years. Patients were presenting with smaller, lower grade wounds of less duration. The change in healing rates and the percentage of easier to heal wounds are mirror images. The data suggests that more of these patients were being attracted to the centers and they were receiving better treatment, even before the new products became available. On the other hand, using the prognostic categories defined earlier, the evidence is that there was little change in healing rates over time for those with wounds in the most severe prognostic categories.

Dr. Margolis identified a major adverse event as another issue to be addressed when designing a clinical trial. Adverse events are of concern in clinical trials, especially in terms of how products are going to work in the future. Certainly, death is the worst adverse event that can occur in a clinical trial for a device or drug, but amputation is the worst commonly seen in a diabetes foot study. According to the CDC website (www.cdc.gov/diabetes), there has recently been a decrease in amputation rates among diabetics. This is true for all diabetics, not just those with pure neuropathy, and that may be part of the difference. However, as Dr. Reiber stated in her presentation, the CDC has also shown that there has been a rapid increase in the number of patients with diabetes, and it may well be that our definition of diabetes has changed over the years. Therefore, if one looks at the same data adjusted for the population-at-large, there has been little or no change in the amputation rate.

Dr. Margolis next presented amputation rates within the Curative system for patients with diabetic neuropathic foot ulcers who have participated in clinical trials. In general, those with significant arterial disease were excluded from trials. Within the Curative system, about 6.7 percent of the patients with diabetic neuropathic foot ulcers had an amputation within 20 weeks of care. Dr. Margolis stated that what is interesting is that almost 50 percent of these people had minor amputations such as a toe or a ray. This distinction is important because there has been much discussion in the last 5 years about the huge difference between a minor amputation and a major amputation, those that are trans-tibial or higher, in terms of rehabilitation potential and in terms of the likelihood that the patient will have additional problems. Another interesting point is that the only risk factor that is really predictive of who will have an amputation is grade. About 62 percent of those who had an amputation had a grade greater than 2.

Another factor of importance to designing a clinical trial noted by Dr. Margolis was when the amputations occur. About 45 percent of them occur within the first 4 weeks of care. This is important because many of the clinical trials that are done in wound care generally have a 2- to 4-week washout period during which patients are followed before enrolling them. Therefore, the likelihood is that it will be known that an amputation will occur before the patient is formally enrolled; however, a fair number of people will continue to have the possibility of having an amputation even years out.

Dr. Margolis also mentioned another aside that may not have to do with clinical trial management, but is interesting from a health services point of view, is that in the Curative system, the rate of amputation for a diabetic neuropathic foot ulcer really has not changed much in about 10 years. What has changed is the percentage of amputations that are minor. There were almost no minor amputations early on, whereas today the vast majority are minor amputations.

To summarize, Dr. Margolis noted that, from the large database of patients with diabetic foot ulcers, the following factors have been learned that would be useful in designing a clinical trial:

• Likelihood of finding a patient with a diabetic neuropathic foot ulcer that has various wound characteristics (e.g., size, age, and grade).

• Use of different aspects to effect inclusion/exclusion criteria and outcome (rate of healing).

• Awareness that rate of healing has changed over time, but sudden adverse events do occur, although most happen in the beginning of a trial.

Surrogate markers were the final item that Dr. Margolis wished to discuss regarding what has been learned about designing clinical trials. Information might be acquired from the CHS database to indicate how long a study needed to be. Most of the studies in the database were 20 weeks long, which is a very long period of time and resulted in very expensive studies. More recently many studies have been 12 weeks long. In determining appropriate study length, Dr. Margolis suggested that it would be very nice to have a surrogate marker, something that early on would determine if the wound was going to heal. He cited the following FDA definition for a surrogate marker: A surrogate endpoint as defined by Temple is “a laboratory measurement or a physical sign used as a substitute for a clinically meaningful that measures directly how a patient feels, functions or survives.” Standard surrogate markers today include blood pressure lowering to prevent cardiovascular disease, lipid lowering to prevent cardiovascular disease, and so forth.

back to top

To answer the question “Is there something that we can use that would allow us to do a shorter study and still be able to say that a product might actually be beneficial?”, Dr. Margolis’ group looked for potential surrogates occurring at 2, 4, and 6 weeks for a 12-week care study or at 2, 4, 6, and 8 weeks for a 20-week care study. He presented the results at 4 weeks looking at a 20-week study from a first pass through the CHS data set and looking at healing rate changes. More elaborate studies are planned for the future. In two groups of a total of about 28,000 patients, a healed group and an unhealed group, there were expected differences in terms of the duration and size of the wounds. However, the average log healing rate for those who healed in the first 4 weeks was 0.062, whereas for those who did not heal, their average log healing rate was 0.007. There is a 10-fold difference between the two numbers. According to Dr. Margolis, this indicates that it is possible to do a fairly small early study and find broad differences between the healing rates very early on that are predictive of what is going to happen later on.

For those who live in a dichotomous world and want to see how this reflects what is ultimately going to happen, Dr. Margolis said to take these rates and try to maximize differences between sensitivity and specificity using AROC curves. If the threshold for the log rate is passed, about 70 percent of the time one can actually predict at the 4th week of care who is going to heal and who is not going to heal at the 20th week, which is good for a surrogate. Dr. Margolis was not suggesting that the ultimate goal of healing be replaced in a clinical trial, but that for an earlier phase study, such as a phase 1 or phase 2, where the goal is to get information on whether or not it is worthwhile to pursue a large study, it would be valuable to be able to determine that about 70 percent of the time the trial’s premise will be correct at just 4 weeks.

Dr. Margolis summarized what has been learned so far: First, that it is fairly easy to distinguish between hard and easy to heal wounds, which will have important implications for the size of the study and for comparing studies. To compare two studies, one must first determine whether or not the patients were the same. Second, early information in the growth factor studies indicates that there might have been a more profound difference and more patients might have been helped if the focus had been on the harder to heal wounds. Third, there are very minimal center effects. This was true looking at three or four different studies. This fact has important implications for sample size calculations as well, especially if there is an important interest in center-based effects, based on some of the information Dr. Boulton discussed. For example, there could be profound differences in terms of how centers interpret offloading and how they encourage their patients to use offloading. Fourth, there is a year-based effect in the easy-to-heal wound; therefore it is necessary to be very careful about using uncontrolled clinical trial information from 2003 and comparing it against data from 1996. Fifth, amputations are unlikely adverse effects, but they do occur and they tend to be minor. That is important in terms of the consent form and in terms of how patients are counseled. It is also necessary to realize that there are going to be some amputations in almost any clinical trial. Lastly, it may be possible to use surrogates based on the change in the size of the wound, at least for the early stages of drug development.

Dr. Margolis made the following suggestions for the future of clinical trials:

• It would be possible to do cluster randomized trials with the prediction models and the surrogate models to see how they might impact practice. Predicting how patients may do with the therapy may actually influence physicians’ and other healthcare providers’ standards of care.

• It is possible to create analyses, using data and models, of the impact of subject selection strategies on clinical trial feasibility, including the impact on patient recruitment, the size of trials, how expensive the trials are going to be, and the likelihood of finding differences.

• It may be possible to go through the database, depending on funding sources, to find better surrogates than just area changes.

• It should be possible, from using this information, to design clinical trials based on how patients are going to do, with a more profound understanding of whether or not they are going to succeed or not succeed, and thus improve sample size calculations.

Brief Discussion. Dr. Kurt Stromberg of the FDA’s Division of Therapeutic Proteins, Office of Biotechnology Products, asked Dr. Margolis if he thought that the surrogate of reduction in wound size as a therapy model to shorten studies would be more meaningful if it were derived from 200 patients in randomized controlled trials versus non-randomized control data from 20,000 patients.

Dr. Margolis answered that in previous work his group had found very similar surrogates using data sets from companies that did earlier studies and looking at their control arms. He had never been given the opportunity to look at treatment arms. Part of the problem with the surrogate is that one does not know how the treatment is working through the surrogate. Certainly the treatment could be working late through the surrogate, in which case the effect will not be captured, or it could be working in an entirely different mechanism, something that emphasizes the surrogate in an odd way. All this previous work was based on status therapy. The data is not very different. What is better about the current work is that, with 20,000 people, the precision of the data is much better, In terms of doing some of the other analyses, like looking for cutpoints within that and trying to determine sensitivity and specificity, the other data sets just were not large enough to do that adequately.

Dr. Margolis added that they had done analyses involving a few thousand patients as opposed to 28,000, but that was not really a randomized trial, it was a cohort, one arm of a randomized study. To evaluate how the surrogate worked in the treatment, again it is a cohort. The two groups cannot be compared. If the question is, “Would it be worthwhile to go back and take the randomized clinical trial data that showed an effect and see whether or not the effect could have been seen at 4 weeks,” that would be a randomized situation, and that would be interesting.”

back to top

Dr. Malozowski, asked Dr. Stromberg, if, given the fact that the information is available at the FDA from many product studies that have been done and given the fact it is not important what product was used, would it be possible to access the data blindly to assess whether these surrogate endpoints are valid or not.

Dr. Stromberg replied that this is proprietary information that the FDA cannot reveal. It would have to be done through the companies themselves. Hopefully, they might be encouraged to release what is antique data because in the end this would benefit them if it enabled shorter trials.

Dr. Margolis agreed with Dr. Stromberg, but pointed out that it had been extremely difficult to obtain even the control arm data from the companies he had dealt with for his meta-analyses and there were severe restrictions on what data was available to him and on how he could use it. None of the companies were willing to give him longitudinal data. He only could obtain data on first visits and on outcomes. Dr. Margolis asked Dr. Stromberg, if he thought the companies were more likely to cooperate with Federal agencies.

Dr. Stromberg questioned where the pressure point was to encourage individual companies to mobilize information that would be useful to the country as a whole.

Dr. Malozowski, related that when he was at the FDA looking at postprandial glucose changes, although they had access to proprietary data, companies would not discuss outcomes related to glucose control measured by HbA1c, so his office looked at data from five different products and came to a conclusion that was shared with the American Diabetes Association at an appropriate time without releasing what products were involved. Internally, the data was presented as being from products a, b, c, d, and e and they were able to determine whether or not there was a relationship to glucose control. Dr. Malozowski suggested that similar exercises could be done by others internally looking at other proprietary data on wound healing.

Dr. Stromberg commented that typically, the FDA has the data since it is the agency that must request the data analyses and also the agency that must decide if shorter trials for a device or drug would be acceptable. The FDA is interested in the treatment arm, as is everyone else. The agency could mobilize a position in-house to deal with the data they have without necessarily having to publish it or release it. They could use the data to reach a conclusion that they believed was satisfactory.

Dr. Margolis repeated that he was not suggesting that the surrogate be a replacement for a pivotal study, but that it be used in earlier phase studies. He offered to work with the FDA to accomplish that goal.

back to top

Clinical Protocols for Treatment of Inpatients and Outpatients With Diabetic Foot Ulcers

Following a brief break, Dr. Malozowski introduced Dr. Harold Brem, Director, Wound Healing Program, Mount Sinai School of Medicine, New York.

Dr. Brem expressed his appreciation to Dr. Moshell and Dr. Malozowski for laying the groundwork for this meeting and creating a true interdisciplinary environment to discuss care for persons with diabetes and foot ulcers. Dr. Brem also acknowledged Dr. Hyde’s and Dr. Abraham’s generosity with their time over the last 5 years, which demonstrated to him the substantial support system the NIH had in place for clinicians and researchers. He also thanked NIDDK, the American Diabetes Association, and the United Spinal Association for their assistance in funding his research.

Need for Establishing a Standard of Care and Standard Protocol. Dr. Brem stressed that there neededs to be a focus on decreasing amputations in patients with diabetes. He believes that it is important to establish a foundation of expectation, both in and outside the hospital, on the part of the physician, the nurse, and the patient with diabetes who suffers from related wounds. Accomplishing the goal of decreasing amputations also requires the use of a standardized protocol, regardless of who is being treated or where they are being treated. Dr. Brem emphasized that the most important paradigm is that patients who do not have either ischemia or osteomyelitis should be expected to heal all of their wounds and always avoid amputation, if the protocol is strictly followed. What is not known is the timeframe in which each of these patients will heal.

Dr. Brem said that he would focus today’s presentation on such questions as:

What is the standard of care?
What is the data?
What should be the focus of future directions?

Extent of Amputation Rate in the Elderly. Dr. Brem noted that the problem of foot ulcers is particularly critical in the elderly. Dr. Brem said that aging in and of itself does not necessarily impair wound healing. What is not known at this time, is whether aging when combined with diabetes has a synergistically negative effect on healing and thereby results in the high amputation rate in the elderly (Slide 1). An alternative hypothesis is that the protocol of standard of care is not applied to the elderly immediately on first observation of the ulcer because more attention is placed on other co-morbidities.


Slide 1 (summarized from CDC Website)

Clinician’s Approach. Dr. Brem posited that clinicians who see a patient with hypertension, early breast carcinoma, or a one-millimeter melanoma expect the patient to be cured in a relatively short period of time regardless of comorbidities, based on the patient’s ability to follow strict-well-established protocols. Dr. Brem stressed that there is no reason for clinicians not to treat patients with diabetes who have a foot ulcer with the same level of care. In 2003, the data clearly suggests that in the absence of osteomyelitis or ischemia, patients who strictly adhere to a well-established protocol should be expected to heal.

Dr. Brem presented slides of a small portion of a typical day’s patient population at the Mount Sinai Wound Healing Program. He said these patients are representative of those at most wound centers. The patients were being treated for foot wounds varying in severity and underlying conditions and comorbidities. Dr. Brem stressed that the importance of the variety in clinical presentations being shown was that the protocol (which he would be describing later) was the same for each of them.

In defining a foot ulcer, Dr. Brem said it is any break in the skin of the foot of a person with diabetes; although the most common definition is of a neuropathic plantar foot ulcer, this definition should be expanded. He emphasized that the pathogenesis and sometimes the treatment would differ based on the etiology, location, duration, underlying bony abnormalities, and many other variables of the foot ulcer.

Indicating a patient with a toe ulcer and a hammer toe (Slide 2), Dr. Brem explained that the hammer toe must be corrected if the patient’s ulcer is to be expected to heal and stay healed. A patient with a foot ulcer often may have fungal nails, as shown in Slide 3. For the ulcer to heal, a combination of topical and/or systemic therapy may be necessary in addition to mechanical debridement or cutting of the nails. Significantly more investigation must be done on fungal nails.

Slide 2 Toe Ulcer Hammer Toe

Slide 3 Slide 4

Debridement. Dr. Brem emphasized that debridement must be performed sharply, as another measure to prevent amputation and accelerate healing. When the wound is debrided, the pathologist should examine all specimens. The wound should be debrided to the outer edge of the hyperkeratotic tissue (see Slide 8).


Slide 8

Debridement can sometimes be worrisome to the patient, who may fear that amputation may be inevitable, since the appearance of the wound and its concomitant infection is frightening. Dr. Brem said, that the patient should be reassured that mechanical debridement is a necessary part of treatment, because it will clean out local bacterial contamination and infection and prevent the possibility of amputation and/or osteomyelitis (See Dr. Boulton’s and Dr. Kirsner’s presentations for additional data).

back to top

Additionally, the physician treating a diabetic foot ulcer should contact the patient’s primary care physician to ascertain the patient’s current medications—in particular, ACE inhibitors—the patient’s glucose control, the ability of the patient to manage offloading, and other information related to the treatment of the wound. This contact should include sharing with the primary care physician the treatment regimen for the patient’s diabetic foot ulcer. The protocol also should include measurement of pain and quality of life.

If the wound heals less than 10 percent over a 3-week period after debridement, Dr. Brem recommended that an FDA-approved biological treatment be considered. He urged elimination of the concept of a non-healing wound. Dr. Brem insisted that if the initial treatment is appropriate and performed at onset, the term “non-healing wound” need not enter one’s vocabulary.

Dr. Brem noted that the question of clinical involvement and staffing also enters into a discussion about care for the diabetic foot ulcer patient. Included in the range of clinicians and staff for patient care that the physician treating the patient should integrate are the following: internist, radiologist, pathologist, physiatrist, physical therapist, occupational therapist, podiatrist, dermatologist, anesthesiologist (e.g., pain control), podiatrist, diabetologist, psychiatrist and nurses. There may also be a need for a neurologist, if neuropathy is a symptom. Dr. Brem stressed that all these clinicians should be aware of the possibility of a referral from the treating physician for a patient with a diabetic foot ulcer. To better assist the patient, each clinician ideally should be familiar with the protocol practiced by the others in this expanded circle of healthcare providers.Focus for Future Practices and Research. From the experience with his wound center’s patients in the program as well as from his staff’s clinical studies and experimental work, Dr. Brem offered the following suggestions for future practices and research:

• Determine a rate of healing for these patients. Based on the wound and its complications, does the wound appear to be one that will heal in 3 months? Six months? Eight months?

• Prove, by strictly practicing a standard protocol, that the amputation rate for these patients can realistically be less than 10 percent, including those with ischemia and osteomyelitis.

• Study separately the healing of wounds of patients with diabetes who do not have a diabetic foot ulcer.

• Develop angiogenesis stimulators for use in the healing of wounds.

• Develop better delivery systems for growth factors—VEGF, bFGF, EGF, GMCSF and PDGF-BB.

• Study the cell populations of patients with diabetes to observe specific changes at the cellular level with treatment by growth factors, noting changes in angiogenesis and related indicators of healing.

• Focus on outcomes with the purpose of changing the standard of care (Dr. Brem suggested this would be a particularly important area for the NIH, CDC, and the FDA.)

• Post on the NIH/NIDDK/NIAMS websites additional information regarding the expectation in healing diabetic foot ulcers, the work funded by the various institutes, and additional patient and provider information regarding available treatment modalities for this condition—specifically offloading. (Dr. Brem noted that the sites currently provide information on foot care, but very little on the care of wounds.)

• Create a Federal mandate to solve the problem of the diabetic foot ulcer, as has been done for numerous other diseases.

Regarding the last bullet, Dr. Brem emphasized that the mandate should go hand in hand with an official standard of care that would directly address the issue of why a patient under the care of one physician suffers amputation, while another with the exact same condition receives appropriate treatment and heals.

back to top

As an example of the overall goal that he proposed, Dr. Brem presented a slide of a patient who healed in 4 weeks, representing the expected standard of care from following the protocol described (see slide 9).


Slide 9

Dr. Brem concluded by saying that he looked forward to a time when amputations will have substantially decreased and perhaps have been all but eliminated for the patient with a diabetic foot ulcer. He recommended that a consensus conference to address standard of care and a standardized protocol for these conditions be organized.

Adjunctive Therapy for Diabetic Foot Ulcers: Current and Future Approaches

Dr. Malozowski introduced the final speaker prior to the discussion period, Dr. Robert Kirsner, Associate Professor, Department of Dermatology and Cutaneous Surgery, Department of Epidemiology and Public Health, University of Miami School of Medicine.

Dr. Kirsner disclosed that in discussing therapeutic approaches he would be talking about some proprietary products of companies with whom he had worked. In presenting these adjunctive therapies and looking towards the future, Dr. Kirsner said he preferred Yogi Berra’s statement that “You can observe a lot by watching” to the more pessimistic quote from Sir William Osler that “Half of us are blind, few of us feel, and we all are deaf.”

Today, Dr. Kirsner said that the group had learned, from what Dr. Margolis eloquently spoke about and what others had observed, that current practices were not very successful at treating diabetic foot ulcers. According to the outcomes from clinical trials, between 24 and 31 percent of patients are healed between 12 and 20 weeks (Diabetes Care 1999 May;22(5);692-5). Some may say, “Well, that’s a select population of difficult-to-heal patients.” Dr. Margolis also described how patients from a large database are treated in practice. Only between 30 and 50 percent of these patients’ diabetic foot ulcers healed in 32 weeks (Diabetes Care 2001;24;483-8). Dr. Boulton spoke on the importance of offloading and Dr. Brem emphasized debridement as a standard of care. Given all this information, Dr. Kirsner said that the question: remains “Why aren’t we achieving the outcomes?” Is the answer something simple or not? Is it because of poor compliance with the methodologies by patients or clinicians? Or is something else going on? Is there a refractory subset of patients that just will not respond to this modality?

In responding to the questions he had raised and the subject of the refractory subset, Dr. Kirsner first referred to Dr. Margolis’ presentation and the research literature on baseline predictors such as the size, depth, and duration of an ulcer (Margolis et al. Diabetes Care 2002;25:1835-9; Margolis et al. Arch Dem 2000;136:1531-5); healing rates or surrogate outcomes predictive of healing such as a wound being on a certain wound-healing trajectory (Robson MC, Hill DP, Woodske ME, Steed DL Arch Surg 2000;135:773-7) or decreasing in size over a certain period of time (Kantor J, Margolis DJ Br J Derm 2001;142:960-4); and the value of dichotomizing patients into healers and non-healers to look at why patients who are not going to heal do not heal.

Next, Dr. Kirsner offered proposed mechanisms—unresponsive or senescent cells present either in the wound or in the callous around the wound, a proteolytic or inflammatory environment, deficient or unavailable growth factors, or bacterial interference—as being responsible for chronicity in diabetic foot ulcers. Some of these mechanisms also are true for other chronic wounds such as abuse ulcers or pressure ulcers. He then presented current and future modalities that have been applied to address each of these mechanisms.

back to top

Cellular Senescence. Cells that are older differ from young cells in that they are less likely to undergo apoptosis, they do not grow as well, and they produce certain matrix proteins. Dr. Kirsner illustrated this difference with an example: using a platelet-derived growth factor as stimulus, his slide showed that, while fibroblasts taken from acute wounds and fibroblasts taken from a dermis responded well, those taken from chronic wounds did not and the longer the fibroblasts were in the wound bed, the less responsive they were to stimuli (Agren et al. J Invest Dermatol 1999). This information correlates with the idea of duration of ulcer being predictive of healing, along with the concept of in vitro senescence.

A number of companies have addressed the idea that cells are perhaps abnormal in a chronic wound by delivering non-senescent cells to diabetic foot ulcers to stimulate the senescent cells or provide new cells to stimulate healing. The first FDA-approved product based on this concept was Apligraf®, which is a bilayered living skin construct with neonatal fibroblasts placed in bovine type 1 collagen and then neonatal keratinocytes grown on top of it. A second product is a dermal construct, where neonatal fibroblasts are placed on an absorbable suture material, and in culture those fibroblasts produce a neodermis. These allogeneic cells, whether in Apligraf® or Dermagraft®, were applied to diabetic foot ulcers. In the pivotal trial that led to Apligraf’s® FDA approval, patients who received the product plus good wound care were 2.1 times more likely to heal than patients who received good wound care alone (Diabetes Care 2001;24:290-5). The results were similar for Dermagraft®; patients who received the product were 1.7 times more likely to heal than patients who received standard care (Diabetes Care 2003;26:1701-5). Future skin constructs include OrCel™, another bilayered cellular matrix that is different because it is cryopreserved, making it more of an off-the-shelf product that can be kept in the clinician’s office. It also is grown differently. Although bilayered, it is simultaneously cultured fibroblast and keratinocyte, so the culture time is somewhat shorter. It has been suggested that perhaps this epidermis, which is not mature, may provide greater stimulation than perhaps a mature epidermis. Looking at the concept of these tissue-engineered skins—because the products are not present for a very long time in that they are allogeneic materials—one sees that their main function is to provide growth factor stimulation to the previously non-healing wound. In OrCel™, there is a large variety of growth factors produced by the product (including IL-1b, KGF-1, M-CSF, TGF-a, and TNF-a) compared to the growth factors present in normal acute wounds treated by standard of care only. In a small 12-week pilot study done in patients with diabetic foot ulcers, greater healing was found in patients who received standard of care plus OrCel™ compared to standard of care alone (http://www.ortecinternational.com/~johncapa/technology/orcel/31).

Inflammatory Environment. The second concept presented by Dr. Kirsner of why wounds in patients with diabetes may not heal is the presence of a environment which is not conducive to healing. Unlike patients with acute wounds who go through the three phases of healing in a normal fashion, patients with non-healing diabetic foot ulcers may remain in the inflammatory phase and therefore not progress normally. There is data in other chronic wound situations indicating this chronic inflammation may be the case. Investigators have studied cytokine levels and mitogenic activity, comparing both healing ulcers and non-healing ulcers (Trangrove NJ, Bielefeldt-Ohmann H, Stacy MC Wound Rep Reg 2000). They found that there is an increased level of inflammatory cytokines in non-healing ulcers compared to healing ulcers. The importance of this is that the pro-inflammatory cytokines (such as Interleukin-1, tumor necrosis factor-alpha) stimulate the production of a proteolytic environment—specifically, some of the matrix metalloproteinases (MMPs)—and in addition, they inhibit some of the natural inhibitors of these metalloproteinases, the TIMPs (tissue-derived inhibitors of metalloproteinases) (Ito et al. Fed Euro Bio Chem Sci 1990;269:93-95; Murphy et al. Ann NY Acad Sci 1994;732:31-41). Dr. Kirsner said the problem with this scenario is that a coordinated expression is needed of both MMPs and TIMPs to result in normal healing (Plast Reconstr Surg 2000;108:236). Not only are proteinases elevated in chronic wound fluid and inhibitors lowered, but in fact the inhibitors are inactivated, resulting in an imbalance of the proteinases and inhibitors. Investigators recently studied this and found that the higher the ratio of MMP-9 (one of the proteinases) with their inhibitor (TIMP-1), the less likely this was to correlate with healing (Wound Repair Regen 2002;10:26-37). This means there exists a possible biochemical assay to predict healing. Acute wound fluid placed on cells causes them to grow. Contrarily, fluid from chronic wounds placed on fibroblasts in culture does not cause them to grow. In fact, it induces senescence. Therefore, one thing the future may hold is a way to test fluid from a wound, with a litmus-like test, and predict whether or not the wound is going to heal.

Dr. Kirsner explained that other investigators have proposed that another way to approach this chronic wound fluid that is inhibitory is to heat it (Park HY, Shon K, Phillips T: Wounds 1998;10:189-192). In the non-heated chronic wound fluid, the cell growth is low. As the temperature is raised just a few degrees, the cells grow faster. One company has come up with a product called Warm-Up that does this using a temperature control unit, an AC adaptor, a noncontact wound cover, and a warming card. They conducted a randomized trial of 20 patients and found that 70 percent of the wounds treated with the device healed after 12 weeks compared to a control group where 40 percent healed (J Foot Ankle Surg 2003;42:30-5).

back to top

Researchers at the University of Virginia thought about the proteinases in a different way. They crafted a gauze that selectively absorbs elastase, one of the destructive proteinases found in chronic wound fluid (Edwards JV, Yager DR, Cohen IK et al: Wound Repair Regen 2001;9:50-58). Another company has specifically manufactured a product that, according to them, absorbs liquid and physically binds and activates matrix metalloproteases. Each of these products is based on removing or changing the proteolytic environment. According to Dr. Kirsner, over the next few years, a number of additional anti-inflammatory agents will probably be seen. At least two are currently in clinical investigation. One is an agonist of adenosine receptors, which is an anti-inflammatory that is in early trials for the diabetic foot ulcer (Montesinos et al. J Exp Med 1997). Another anti-inflammatory agent, lactoferrin, is also in human trials for diabetic foot ulcers (Engelmayer et al. Wounds 2003).

Other investigations are underway to reverse the proteolytic environment, one of which is the topical use of doxycycline. Doxycycline is used systemically in a number of specialties (dentistry, rheumatology, vascular disease) specifically to reverse proteases. A group from the University of Florida has shown that increasing doses of doxycycline reduces TNF-alpha in a dose-dependent way and reduces the proteolytic environment. A small trial of patients that received the doxycycline topically has done quite well (Chin et al. Wounds 2003). A larger trial is ongoing.

Bacterial Interference. Dr. Kirsner stated there is a long-standing paradigm that bacterial count is paramount to non-healing of wounds, that once there is greater than perhaps 105 organisms, the wound is infected. However, he added, some people in wound healing have felt that there may be something else happening, that the bacterial count is not leading to a frank cellulitis, but perhaps just causing ulcers to be non-healing. Although the number of bacteria is an attractive concept, Dr. Kirsner suggested that if he raised the issue of beta-hemolytic streptococcus, some persons would say “That’s different.” Therefore, maybe virulence of the organism is important. If he asked, “What about an AIDS patient?” again some would say “They’re different. They’re immunosuppressed.” This raises the question “Are diabetics immunosuppressed?” They may be. Dr. Kirsner suggested that perhaps a better way to determine whether a patient is infected or has an increased bacterial burden that may lead to a poor healing wound might be to consider several critical variables affecting the wound: amount of necrotic tissue, number of organisms, bacterial virulence, and patient’s immune response. He noted that the formula would then be:

Dr. Kirsner continued by remarking that in addition to sheer numbers, perhaps quality or the type of bacteria may be causing the wound not to heal, or possibly something new that has been suggested—the formation of biofilms. With regard to quality, better agents and more agents are needed for MRSA because unfortunately patients with chronic wounds have a lot of MRSA. Dr. Kirsner cited two studies that associated the presence of MRSA with longer healing times in a large percentage of patients, 40 percent of whom had MRSA in their wounds (Kac et al. Arch Dermatol 2000;136:735-9; Tentolouris et al. Diabet Med 1999;16:787-71).

Dr. Kirsner explained that biofilms have been implicated in many chronic microbacterial diseases. The characteristics that make biofilms special are that they adhere tightly to biologic and non-biologic surfaces and they form a polysaccharide matrix that makes them resistant to conventional therapy. In special circumstances, free-floating or planktonic bacteria form a biofilm, and these biofilms have recently been found in chronic non-healing ulcers. The future holds new ways to handle these biofilms: maybe enzymes will address the polysaccharide matrix or physical modalities (ultrasound, ultraviolet light, electrical stimulation) may be used to break up biofilms and remove the bacterial interference.

Deficient and/or Unavailable Growth Factors. Dr. Kirsner said that, because of some of the factors in a chronic wound, particularly proteases (perhaps bacteria), some investigators feel platelet-derived growth factor is very useful, but it would be even better if delivered in a better way. He mentioned that Dr. Margolis had spoken of the FDA-approved recombinant platelet-derived growth factor (becaplermin) that had shown a moderate improvement in healing in its pivotal trial (Weiman TJ, Smiell JM, Su Y: Diabetes Care 1998;21:822-7). At the University of Pennsylvania, Dr. Margolis and others are planning, but have not begun, a study in gene therapy, delivering platelet-derived growth factor through a gene construct in an effort to stimulate the healing of diabetic and neuropathic foot ulcers (Margolis DJ, Cromblehome T, Herlyn M: Wound Rep Regen 2000;8:480). As recently as this year, Diabetes Care reported that epidermal growth factor had a significant healing rate of 95 percent at 12 weeks in a dose-dependent fashion when given in a randomized study of 61 patients with diabetic foot ulcers (Diabetes Care 2003;26:1856-61). Dr. Kirsner added that it is possible that other growth factors are on the horizon for diabetic foot ulcers.

Nerves and Enzymes. Dr. Kirsner reported that, in addition to the importance of sensory neuropathy and the results of a loss of protective sensation described by Dr. Boulton, other researchers have thought about diabetic neuropathy in a different way, as the loss of pro-healing cytokines. He noted that fetal wound healing is a model for the importance of nerves in healing. If wounded, a fetus does not scar. Many theories have been suggested as to why fetuses have scarless healing. Perhaps the difference is in the types of transforming growth factor beta, differences in collagen, or differences in the environment. Recently it has been suggested that neural stimulation is important for scarless fetal healing. The reason for this is that in studies involving fetal lambs, when different types of wounds were placed in the limbs, one of which was de-enervated and the other left intact, it was found that the limb that was de-enervated had slower healing and had scarring as opposed to the limb that had enervation intact, thus suggesting the importance of nerves in healing (Plast Reconstr Surg 2000 January;105(1):140-47).

back to top

Other investigators have found that fewer nerves are present in the epidermis and papillary dermis of patients with diabetes. This is also true in diabetic murine models, which had significantly fewer epidermal nerves, and the fewer nerves correlated with increased healing times in mice with diabetes (Gibran NS, Jang YC, Greenhaigh DG, Muffley LA, Underwood RA, Usui ML, Larsen J, Smith DG, Bunnett N, Ansel JC, Olerud JE J Surg Res 2002;108:122-28). This has resulted in the concept that in hyperglycemia or diabetes, there are less nerves and thus less of the factors that nerves elicit to speed healing (Spenny ML, Muangman P, Sullivan SR, Bunnett NW, Ansel JC, Olerud, JE, Gibran NS: Wound Repair & Regen 2002 Sep-Oct;10(5):295-301).

Dr. Kirsner explained that in addition to fewer nerves in the skin of those with diabetes, there are also more enzymes that break down some of the proteins that may speed healing. In the animal model for diabetes, the db/db mouse, there is greater endopeptidase activity in diabetic mice than there are in normal mice. Similarly, in patients with diabetes, there is a higher level of the protease that breaks down some of those products that nerves release compared to controls (Antezana, MA, Sullivan, SR, Usui ML, Gibran NS, Spenny ML, Larsen JA, Bunnett JC, Olerud JE J of Invest Dermatol 2002;119:1400-04).

Fewer nerves and more enzymes may lead to delayed or non-healing in diabetics. Dr. Kirsner reported that it has been suggested that, given this situation, perhaps providing something that the nerves elicit, such as substance P, might speed healing. At least in one study in animals, substance P shortened the time to closure in animal models of diabetes (Gibran NS, Jang YC, Isik FF, Greenhaigh DG, Muffley LA, Underwood RA, Usui ML, Larsen J, Smith DG, Bunnett N, Ansel JC, Olerud JE J Surg Res 2002;108:122-28).

Future of Diabetic Complications. Dr. Kirsner summarized that future research needs and opportunities were certainly to prevent, to regenerate, and to treat diabetic complications. In addition to neuropathy, there is vascular insufficiency that may cause disease directly or through neuropathy. Dr. Kirsner cited in a slide a number of efforts at gene therapy for peripheral vascular disease (PVD). The importance of the PVD studies for neuropathy is that some products such as vascular endothelial growth factor (VEGF) are now being examined for diabetic neuropathic ulcers. Thus information gained from PVD is being applied in another setting.

Returning to the refractory subset, Dr. Kirsner said he had presented several ways to reverse non-healing of chronic diabetic foot ulcers—eliminate unresponsive and/or senescent cells, alter the proteolytic/inflammatory environment, supply and protect deficient and/or unavailable growth factors, and remove bacterial interference.

Another issue Dr. Kirsner wished to address was standard of care and, regardless of whether poor compliance is by patients or clinicians, how therapeutic adjuvants to the standard of care, particularly debridement and offloading, can lead to better outcomes. As Dr. Boulton mentioned, patients are just not wearing their removable offloading devices. There are many reasons for this, but Dr. Kirsner feels it is important to understand the stress involved in patients with diabetes and their level of depression, factors that may lead to poor compliance. Dr. Kirsner therefore posed the question “If patients’ ulcers are debrided properly and healing occurs, perhaps by using a contact cast, or Dr. Boulton’s instant contact cast, is there any benefit in speeding healing in patients who may heal anyway?” He answered that he thought there was because there would also be fewer complications such as osteomyelitis and amputations, better quality of life, and less cost to society.

Dr. Kirsner’s next question was “Do any of the adjuvant therapies speed acute wound healing?” To respond he described two studies done in Miami. One was a randomized trial of platelet-derived growth factor compared to standard of care, which was bacitracin ointment. Patients had acute wounds created on their inner arms. One group received the standard of care, and the other group received the platelet-derived growth factor (rhPDGF-BB gel (Regranex®)). There was faster healing in those wounds that received platelet-derived growth factor (92.9 percent by day 22 and 100 percent by day 24) compared to the antibiotic ointment (50 percent and 57 percent, respectively) (Cohen M, Ealgstein WH J Am Acad Dermatol 2001;45:857-62). In another study of acute wounds using tissue-engineered skin, the Apligraf® product, compared to autologous skin, compared to the standard of care, which was a film dressing, the tissue-engineered skin behaved similarly to the patient’s own skin, both of which healed faster than the film dressing in a statistically significant way (Muhart M, McFalls S, Kirsner RS et al. Arch Derm 1999;135:913-18; Muhart M, McFalls S, Kirsner RS et al. Lancet 1997;350:1142).

In conclusion Dr. Kirsner stated that he feels the key adjuvants to the standard of care, which are debridement and offloading, are the mainstay of treatment. Therefore, it is critical to address issues of poor compliance. It is also highly important to investigate further the refractory subset and ways to reverse the reasons for this refractory subset.

back to top

Discussion

Dr. Malozowski opened the meeting to questions and discussions.

Dr. Steven Kravitz, Executive Director, American Professional Wound Care Association, with a membership of 1,200 U.S. clinicians, commented on Dr. Reiber’s recommendation for a consensus in use of ICD-9 codes for diabetic foot ulcers that it is probably easier to try to get physicians practicing in the VA to code in a similar manner than to get agreement among physicians participating in large clinical studies across the country. Dr. Kravitz added that one of the problems is the way billing is done in the United States. Since Medicare is interpreted by different carriers, billing is generally based on how the carriers interpret it, and thus physicians in Pennsylvania code their bills differently and those in New York or elsewhere. A national coding system is needed to eliminate these geographical biases. The question has been discussed within his association and is considered a problem.

Referring to the fact that CDC uses a particular code and other agencies use other codes, Dr. Malozowski asked “What are the barriers to implementing a universal system and how can we overcome these barriers?”

Dr. Reiber suggested that the easiest way to deal with the situation is to discuss it among one’s colleagues. For example, in presenting a set of preferred codes, one might ask others to consider these or justify use of other wound codes for reimbursement. She thought that asking the Government agencies to resolve this might become too complicated.

Dr. Brem agreed that the concept was valuable although difficult to implement. The persons doing billing, for instance, cannot simply look at a photograph and determine the correct code. Dr. Brem felt, however, it was important to work with the VA and with the Centers for Medicare & Medicaid Services (CMS) on the definitions and the modifiers. Once these are standardized by CMS, as for example”707.16 diabetic foot ulcer breaking the skin,” then there will be a financial incentive for those seeking reimbursement to follow that coding system in order to be compliant with CMS

Dr. Margolis said that he thought part of Dr. Reiber’s concern was in regard to looking at a data set and seeing the distinction, for example, between a foot ulcer and a venous leg ulcer in a person with diabetes. The codes are important not only at the healthcare provider level but also when being interpreted at the larger database level. For example, diabetes with venous leg ulcers as a 454 code is very different than a 707 code for a foot ulcer code with diabetes. Thus, relying on these codes alone can cause major differences between epidemiologic studies done by one person versus another. The best way, according to Dr. Margolis, is to pick code algorithms and then go back to the patients’ charts and see if you find what you expect to find. The new Health Insurance Portability and Accountability Act (HIPPA) regulations make this more difficult to do.

Dr. Gilman Grave, Chief, Endocrinology, Nutrition, and Growth Branch, National Institute of Child Health and Human Development (NICHD), raised a different issue. He expressed his surprise that Dr. Stanley Cohen’s work on epidermal growth factor (EGF) that NICHD has been supporting for nearly 25 years, is just now being used with diabetic ulcers. (Dr. Cohen, Professor Emeritus at Vanderbilt University, received the Nobel Prize for Physiology of Medicine for his work on EGF and its receptor, tyrosine kinase, which paved the way for all of the following receptor kinase research.) Dr. Grave thought that the use of VEGF had been an established standard of care for years to cure ulcers.

back to top

Dr. Kirsner responded that EGF was the first growth factor to be studied and shown to speed healing in humans. This study was in patients without diabetes and had their donor site wounds treated with EGF in a silver sulfadiazine base compared with the base alone and was published in the New England Journal of Medicine in 1989 by the group out of Vanderbilt. However, their work was not then reproduced. Next, studies were done based on whether there was a rationale for using epidermal growth factor for deeper wounds, for full-thickness wounds, but the pressure ulcers and venous ulcers that were studied did not heal faster in a statistically significant way. Other growth factors then became more popular, such as platelet derived growth factor, which probably has been studied the most.

Dr. Brem noted that a new trial using EGF was reported in Diabetes Care in 2003. What they did differently from the earlier trial with deep wounds was to use better delivery systems and the results were very promising.

Dr. Kirsner commented that the Annals of Internal Medicine recently reported a randomized trial of 36 non-diabetic patients who received nerve growth factor or placebo plus standard care for pressure ulcers on the feet present for less than 1month duration. Those receiving the nerve growth factor had better healing rates and a greater reduction in wound size. Dr. Reiber added that the patients, however, were only followed for 6 weeks.

Dr. Brem asked if it were possible to achieve consensus on clinical trial standards, such as length of trial, that could be established by FDA based on epidemiological information such as that provided by Dr. Margolis. Such standards would certainly be cost-effective.
Dr. Stromberg replied that current consensus on the endpoint is one of complete closure, which has not been shown to be achieved in a short-term trial.

Dr. Boulton pointed out again that the new therapies, including growth factors, have probably only shown minor results because of lack of control for offloading. Researchers he has spoken to about this are reluctant to include offloading because they believe the ulcer will heal without the offloading, which can be difficult, expensive, and time-consuming to accomplish. Dr. Boulton still considers offloading the confounding variable that, if included and controlled, would produce better success rates, and emphasized studies are needed to investigate this hypothesis. He added that another factor that needs to be addressed, as Dr. Kirsner had mentioned, is the depression that is common in patients with diabetic neuropathy. Even those who experience no pain, suffer from troubling unsteadiness, a major symptom that is rarely assessed. The impact of the psychological aspects in these patients was demonstrated by the Ohio group, headed by Dr. Ron Glaser, who showed very clearly that pro-inflammatory cytokines are much higher in patients with depression. Classic studies in dental students also found that, at the times of stress at final exams, healing rates were half what they were during vacation periods.

Dr. Stromberg agreed that data on the effect of offloading in controlling outcomes could be important information to present to sponsors of clinical trials, possibly as a requirement.

The use of honey as a wound therapy was raised as a possibility for a controlled trial. Apparently honey’s carbohydrates exert osmotic forces that break open the bacteria’s walls. Another alternate therapy, raised by Dr. Moshell, was the use of sterile maggots. Dr. Boulton said that maggots are used widely all over the world. The concept arose during the American Civil War from observing that soldiers who survived loss of their legs also had these wounds infected with maggots. There has been extensive experience in the United Kingdom in using maggots in refractory patients with neuro-ischemic ulcers who had distal vascular disease without possibility of a bypass. These living chemical factories make enzymes that are bacteriostatic. There also is uncontrolled evidence suggesting they are useful in getting rid of MRSA. They have several advantages: they are less expensive than antibiotics or surgery and they work 24x7. A controlled trial in the Ukraine showed faster healing with wounds treated with maggots, but the comparison method was not a good one. Dr. Boulton recommended that larva therapy or biosurgery—for a more acceptable term—is an area in need of controlled data to support the anecdotal experience with this therapy.

back to top

Dr. Kirsner agreed that many view it as a viable therapy but those who do not wondered if it were possible to extract the enzymes the maggots secrete and use them as a potent debriding agent. Several participants suggested this was might be one of the clinical trial areas, among other areas that are likely to be commercially non-viable, that the DMICC and SDICC might sponsor.

Dr. Judith Fradkin, Director, Division of Diabetes, Endocrinology, and Metabolic Diseases, NIDDK, said she would be talking later about such initiatives and opportunities for support, such as R01 trials, and urged those present to submit proposals. Dr. Stromberg wanted those present in the wound healing community to know that NIH supports efforts other than basic research and seconded Dr. Fradkin’s suggestion that they submit proposals for clinical trials of new therapies to enhance the field.

Dr. Fradkin noted that several researchers present were in fact being supported by NIDDK for other than basic research. She added that there also are opportunities for follow-up studies in patients who were formally enrolled in large clinical trials enrolled in clinical trials and there are certainly unanswered epidemiological questions that could be addressed in these populations. One example is the Epidemiology of Diabetes Interventions and Complications (EDIC), which is the follow-on to the Diabetes Control and Complications Trial (DCCT) and has 1,400 subjects for whom comprehensive data has been collected over the past 20 years—their blood pressure, their lipids, their inflammatory states, everything that could be measured. ACCORD (Action to Control Cardiovascular Risk in Diabetes), a clinical trial co-sponsored by NIDDK and the National Heart, Lung, and Blood Institute (NHLBI) is currently enrolling 10,000 type 2 patients being randomized to various levels of glucose, blood pressure, lipids, and controls. Dr. Fradkin asked the participants to suggest epidemiological ideas and methods that could be pursued with these populations for whom there exists or will exist extensive relevant data.

Dr. Margolis contributed that there are hundreds of questions that could be addressed via these populations. Most of what is known about the natural history of the onset of foot ulcers is really based on either small series or large databases. Some of the small series are quite large, like the VA groups; however the EDIC and proposed ACCORD groups offer relatively large, virgin populations on which there is extensive data that could provide more information than can be currently obtained from databases or past surveys. This is invaluable.

Dr. Fradkin pointed out that the data from these studies provide a potential opportunity to characterize people in a standardized way versus relying on the range of codes currently being used and their attendant problems. On the other hand, it is a smaller sample size than what is available in the huge, national databases. She added that obviously funds are limited for these studies, but the EDIC group is, in fact, developing their protocol for the next period of follow-up, and she thought they would welcome suggestions for measurements or other attributes that could be incorporated into EDIC. They have been working with Dr. Eva Feldman, Department of Neurology, University of Michigan, about a more extensive neuropathy component, possibly including nerve conduction. Dr. Fradkin emphasized that suggestions from those present would be very much welcomed.

When asked how long patients are followed, Dr. Fradkin replied that EDIC has been following them for 20 years. The next 5- to 10-year period is now under discussion. Patients are seen twice a year. EDIC has 1,400 patients for whom every conceivable measurement has been made, including coronary CT scans. Their vascular status is characterized and all sorts of markers of inflammatory mediators and so forth have been collected.

Dr. Malozowski added that the same will be true for ACCORD’s 10,000 diabetes patients in multi-centers across the United States and Canada. Over 3,000 have been recruited at the rate of about 80 per week.

Dr. Fradkin went on to say there are potential opportunities for add-on studies at relatively low additional cost, which could answer some of the remaining questions. It would also be good to figure out a standardized approach to the foot so that the two studies could be compared across each other—EDIC’s type 1 diabetes patients versus ACCORD’s type 2 patients. To apply, those interested should submit a proposal stating the nature of the examination to be conducted and what the hypothesis would be. Dr. Fradkin would be very willing to bring such a proposal to the EDIC chairs, Dr. David Nathan and Dr. Malozowski, and also to Dr. Peter Savage, Director, Division of Epidemiology and Clinical Applications, NHLBI, for ACCORD.

back to top

Dr. Brem asked if it would be possible to photograph the feet of these patients and measure their wounds, if present, with a ruler. This could resolve coding problems. All of the process data would be irrefutably standardized. Dr. Fradkin replied that she could see no reason why the feet could not be photographed; they are photographing the retina. Dr. Malozowski commented that it would be less complicated to photograph the feet.

Dr. Moshell said that, in terms of not reinventing the wheel, there is a dermatologic component of NHANES run by the National Center for Health Statistics (NCHS) based on digital photography for photographing skin. The bottom of the foot is not part of that digital photography, but the equipment and the techniques are in place. If anybody wanted to do that, Dr. Moshell offered to put them in contact with the appropriate persons at NCHS who are doing this. Dr. Kravitz added that digital photography is very inexpensive today. The American Professional Wound Care Association’s website has 10 guidelines for clinicians to use with almost any digital camera. Dr. Fradkin cautioned that the hypotheses, the questions to be addressed, and the study design need to be identified and a proposal submitted prior to asking for photographing of the patients.

Dr. Kravitz introduced the separate issues of access and cost with regard to adjuvant therapies. Medicare recently approved venous compression stockings for people with venous stasis ulcers. Dr. Kravitz strongly suggested that it would be more cost-effective, given the $40,000 noted in one report to heal a foot ulcer, for Medicare to approve the compression stockings for patients with venous stasis before an ulcer occurs. He noted there continues to be a frustrating disconnect between cost containment and practical care to decrease morbidity. Dr. Brem agreed that overall cost would be dramatically reduced by taking existing knowledge and establishing standards of care, including offloading.

In response to a question from Dr. Malozowski regarding why patients do not use the offloading device—is it a matter of aesthetics or discomfort—Dr. Boulton responded that it is because they have no pain perception and they think that as long as they are at home, they do not need it. In a previous study, it was shown with activity monitoring, that patients with foot ulcers are more active at home and less active out-of-doors, whereas patients with neuropathy and no foot ulcers are more active out-of-doors and less active at home. In foot ulcer patients with no sensation of pain, even though 75 percent of their activity is around the home, they wear the device as one would wear a shoe out in the rain. They see the dangers as being outside. There is some data to suggest that they take it off at home because they perceive the familiarity of the surroundings—the carpet, the lack of foreign objects on the floor—as safe. However, the problem with an insensate foot ulcer is that there is no safe zone. Just those few steps to the bathroom each night are the most dangerous steps. The only device that works is an irremovable cast walker. Again, it is a matter of perception. The patient thinks that if something hurts, rest it. If it does not hurt, it cannot be all that bad.

Dr. Kravitz commented that these psychological aspects are correct. A person feels just as “normal” as he/she was a couple of months earlier, and now there is the hassle of wearing this device wherever he/she goes. So, the patient takes it off for “just 5 minutes” and, unfortunately, loses the healing effects of the last 4 weeks.

Dr. Reiber observed that patients who have their first diabetic foot ulcer tend to be more compliant, more likely to offload, than they are with subsequent foot ulcers. It seems in subsequent foot ulcers, they think, “I’ve been through it before. Nothing bad will happen.” Dr. Boulton agreed with this observation that was also seen by Dr. Paul Brand in leprosy. He said the most useful time for education is at the time of the first foot ulcer. Dr. Brand once showed a cartoon of a patient with the first ulcer falling over the edge of the cliff: at this stage the patient responds well to education and can be “pulled back” from the edge of the cliff.

back to top

Dr. Malozowski commented that another perception problem is that the foot is not seen as an item of great appeal. Treating a foot ulcer is not neurosurgery. The importance of the foot ulcer in limiting activity and affecting quality of life is not recognized. He stressed that this meeting is important in raising awareness of this. Dr. Boulton agreed that people have always been more interested in treating eyes or kidneys and have the attitude that one can ignore the foot. However, this is beginning to change. In looking at the number of papers published on the diabetic foot as a proportion on all papers on diabetes from Medline, from about 0.7 percent of all papers in 1980, it is now about 4 to 5 percent, up five-fold. Part of the reason for the change is health economics.

In reply to Dr. Myrlene Staten’s (Senior Advisor, Diabetes Translational Research, Division of Diabetes, Endocrinology and Metabolic Diseases, NIDDK) query about the distribution of severity in the 20,000 patients with diabetes and foot ulcers, Dr. Margolis said that about 45 percent of the patients were grade 2 or less, with about 18 percent without major risk factors, another 18 percent having a larger wound, and another 18 percent having one of longer duration but still grade 2 or less. Dr. Staten responded that meant that more than half are at the more severe grades. They have osteomyelitis or exposed tendons. Dr. Margolis answered that there were very few patients in the highest grade however. What is being seen over time is that most of the patients are in the earlier grades and more and more of these patients are being seen. There are, however, patients with grade 3, 4, 5, and 6 ulcers.

Dr. Malozowski stated that Dr. Margolis’ slide on Prediction With Dichotomized Risk Factors for DNFU presented important, relevant information to power a study. It indicates how many patients are needed at each level to show whether or not healing rates will progress from one stage to another. Dr. Margolis said the slide would be published in December 2003 (Am J Med 2003;115:627-631).

Dr. Leonard Pogach, VA National Program Director for Diabetes, asked Dr. Margolis if there was any data on the psychosocial determinants predicting what type of people would come to the wound care centers.

Dr. Margolis replied that, although such information might be collected and might exist, it would not be in the database. He added that because he was somewhat surprised to find that the healing rates changed over time and had heard a tape of Dr. Vinicor’s speech at the ADA meeting in 2002, he had sought an explanation for the results he was seeing. He proposed that possible reasons for the increased healing rates may be that people were coming to the centers sooner or more physicians were looking at the foot than in the past. Perhaps patients are becoming more empowered through the messages issued by the foundations on the importance of the foot in diabetes and the need to see a doctor if one has a wound on the foot.

back to top


Ongoing Brief Overview of Planned Activities

Representatives of DMICC member agencies presented overviews of their current and future activities related to diabetic foot ulcers. Dr. Malozowski introduced the first presenter, Dr. Kurt Stromberg of the FDA’s Division of Therapeutic Proteins, Office of Biotechnology Products, Office of Pharmaceutical Science (OPS), Center for Drug Evaluation and Research (CDER).

Food and Drug Administration

Dr. Stromberg stated that, as discussed earlier, there are only a handful of approved products for treating diabetic foot ulcers. There are several approved non-interactive, wound healing dressings, but these are actually products for wound management, not treatment of diabetic wounds. There also are various approved debriding agents, but these are for chronic wounds and most of them are “grandfather-approved.” Given this situation, Dr. Stromberg said that it was important to ask the question “Why are there so few approved treatments?”

As emphasized by the speakers, diabetic foot ulcers are a multifactorial problem and therefore, treatment needs to be multifactorial. However, treatments tend to be single. Dr. Stromberg suggested that the FDA might help address this problem by trying to speed up clinical trials to provide more incentive for variability in types of treatments. For example, instead of a trial for one growth factor, there might be a sequential application for a number of growth factors to try to duplicate the healing process.

Dr. Stromberg explained that, whereas, complete healing and the longer trial is appropriate for the pivotal trial to actually determine effectiveness and efficacy, at the earlier stages perhaps several of the surrogate endpoints could be used to enable more trials to be carried out. Although FDA has data that might enable use of these surrogate endpoints, such as reduction in wound size as a predictor of healing, there is no vehicle or person available to analyze the data. Dr. Stromberg recommended that perhaps this is where NIH can make a major contribution by funding such analyses. The FDA would ensure that the data being looked at remains anonymous. He believed such analyses would result in developing the necessary epidemiology to support identification of surrogate markers for clinical endpoints, which he saw as an important objective resulting from today’s meeting.

In summarizing FDA’s current planned activities, Dr. Stromberg said there are no new devices, in the approval or trial stage. They are being investigated. There is a tri-center clinical focus group on wounds that developed a guideline in 1991 that will be presented to the Dermatological Diseases Advisory Committee in the spring of 2004. The guideline has been available for 2 years; a version was published in June 2001 to guide development of wound healing products. There was a meeting recently of the Anti-Infective Drugs Advisory Committee, but they did not achieve a consensus on the required outcome for diabetic foot infections. Dr. Stromberg expressed concern over the problem of the increasing obesity of the population and its association with diabetic foot ulcers and the necessity this creates for development of effective treatments.

back to top

National Institute of Diabetes and Digestive and Kidney Diseases, NIH
Dr. Malozowski presented Dr. Judith Fradkin, Director, Division of Diabetes, Endocrinology, and Metabolic Diseases (DDEM), NIDDK.

Dr. Fradkin said that when the special funding for type 1 diabetes was first received in 1998, diabetic foot disease and diabetic neuropathy were immediately identified as two understudied areas. NIDDK issued a series of requests for applications (RFA's), mostly R01s, in conjunction with the National Institute of Neurological Disorders and Stroke (NINDS) and other partners to stimulate research in this area. These include DK 98-009, Pathogenesis and Therapy of Complications of Diabetes ($6.7 million); NS 99-005, Neurologic Complications of Diabetes ($2.2 million); NS 00-002, Neurobiology of Diabetes ($4 million) and DK 01-006, Gene Therapy Approaches for Diabetes and Its Complications ($2 million). In addition, using regular NIDDK funds, the Institute issued an RFA specifically for diabetic foot disease, which resulted in funding for some of those present (DK 00-009, New Therapies for Diabetic Foot Disease, $1.5 million). Dr. Fradkin expressed pleasure in hearing today how the work supported through that RFA has come to fruition.

Dr. Fradkin noted that for several years the National Diabetes Education Program (NDEP), headed by Joanne Gallivan, and the Diabetes Clearinghouse, led by Kathy Kranzfelder, has supported an ongoing campaign called “Feet Can Last a Lifetime” that was done in partnership with voluntary organizations (American Diabetes Association, Juvenile Diabetes Foundation, American Association of Diabetes Educators, and orthopedic and podiatric societies) and other components of the U.S. Department of Health and Human Services (CDC, the Indian Health Service, HCFA now CMS, the Health Resources and Services Administration, and the VA). More than 20,000 kits have been distributed to healthcare providers, and they are now in their third production cycle. Each of the kits contains monofilaments and instructions for foot exams, stickers to put on patients’ charts to identify people who are high-risk, information about getting reimbursement for patient care, and patient education materials for people with high-risk feet. The kits also include posters to put up in the providers’ offices telling people with diabetes to take off their shoes, because, in busy practices, if people do not take off their shoes, their feet probably will not be looked at.

Dr. Fradkin agreed with Dr. Brem that NIDDK’s focus had largely been on prevention of foot ulcers, not treatment. As diabetologists, NIDDK knows about preventive care for people at risk of foot ulcers, but patients are referred to other specialists for treatment. However, she assured Dr. Brem that his recommendations had been heard by the NDEP and Clearinghouse directors. She also asked that those present contribute to helping NIDDK develop materials to specify a standard of care, such as emphasizing the importance of offloading.

Next Dr. Fradkin introduced Dr. Myrlene Staten, Senior Advisor, Diabetes Translational Research, DDEM, who is responsible for NIDDK’s bench-to-bedside translational initiatives, including a special program for those who have an idea for a new drug or new agent that might be effective but is potentially not something that would be developed by pharmaceutical companies. The program provides resources rather than financial support to help develop the product such as preclinical testing, development of agents, and so forth. At the other end of the spectrum are projects to take findings from clinical trials that have been shown to be effective in diabetes prevention and control of complications and to develop trials to show how these procedures can be more cost-effectively or more efficiently applied in clinical practice. This is research-to-practice translation. There is an ongoing program announcement that has all the advantages of a regular program announcement (PA) and an RFA in that it has three receipt dates a year, but it is reviewed by a special study section the way an RFA is reviewed. It is specifically reviewed with regard to clinical interest in improving patient care. For example, this PA could be used to develop an approach that had been shown in trials to work, such as offloading, and develop a method for applying the approach in providers’ practice in the community. Dr. Fradkin encouraged those present to look at that solicitation and to talk to Dr. Staten if interested in applying for funding under it.

There is also an RFA using special type 1 funds for a bench-to-bedside program that moves a protocol from the pilot and feasibility phase, with demonstrated milestones, into a larger clinical trial. This is available for any area of research relevant to type 1 diabetes. The next receipt date for that is February 20,, 2004. There is an innovative partner solicitation, for which the due date has passed, that Dr. Fradkin said NIDDK is considering re-issuing. This solicitation is a mechanism that allows researchers working on diabetes to further their research by bringing in a collaborator who does not currently do research on diabetes but who has some special talent such as bioengineering that would be relevant.

back to top

Finally, this year NIDDK has a special small business innovative research (SBIR) solicitation focused on developing new therapies for type 1 diabetes. This includes complications. Dr. Fradkin commented that many of those present had ideas that could be commercially viable, ranging from new products for offloading to biologicals that might be developed working with a small business. The due date for this solicitation is February 20, 2004. She added that all the initiatives she had discussed can be found on the NIDDK website (www.niddk.nih.gov) under “What’s New?”

Future plans, according to Dr. Fradkin, include working with the Juvenile Diabetes Research Foundation (JDRF) and the Department of Defense (DoD) to develop, through DoD’s Defense Advanced Research Projects Administration (DARPA) program, a project to improve wound healing focused on battlefields. (DARPA is high-risk, intensive deployment of resources program to develop a specified application driven by a pre-defined outcome. Projects usually have a lifetime of 3-5 years and a large budget of $5-10 million.) NIDDK and JDRF are talking with DoD about how their project might be directed toward applications relevant to diabetes. Also, NIDDK has been coordinating with the National Cancer Institute and JDRF to interest angiogenesis researchers in moving into the area of diabetes. A large workshop is planned for 2004 that will probably be followed by a research solicitation.

To investigate participation in any of these initiatives, most of which are trans-NIH, Dr. Fradkin urged those present to contact Dr. Staten for translational projects and Dr. Teresa Jones, Program Director for Diabetes Complications, DDEM, who is taking over administration of these grant programs from Dr. Kristin Abraham, who will now head up the Diabetes Centers Program.

A NINDS member commented that her Institute has been partnering with JDRF and NIDDK on diabetic neuropathy solicitations. She added that they may not be ready to issue another similar RFA at this time, but might consider one to address the multifactorial nature of diabetic foot ulcers, particularly the lack of sensation that occurs sub-clinically early on prior to the infection and also the age relationships.

Dr. Reiber suggested that primary care provider education is needed to give them information and guidelines about what they can do prior to referring their patient to a wound specialist. Dr. Fradkin responded that that is, in fact, the goal of the “Feet Can Last a Lifetime” campaign, to have these primary care providers look at the diabetic patient’s feet, identify who is at high-risk, and then give the patient prevention information. NDEP also has a working group of providers in various sub-specialties, including podiatrists, who have been helping develop materials with regard to the diabetic foot. Dr. Malozowski added that the group’s recommendations will be used in reviewing and evaluating a document the group is developing. In response to Dr. Reiber’s concern that the primary care provider not only know about prevention but also what questions to ask when the ulcer is present, Dr. Fradkin welcomed partnership with those present in developing appropriate materials.

back to top

National Institute of Arthritis and Musculoskeletal and Skin Diseases
Dr. Malozowski turned the podium over to Dr. Alan Moshell, Director, Skin Diseases Branch, DMICC representative from NIAMS, and cosponsor of today’s meeting.

Dr. Moshell pointed out that NIAMS, as a small institute, supports the broad area of chronic wound healing research, not diabetic foot ulcers specifically. Chronic wounds were the subject of a number of Skin Disease Interagency Coordinating Committee meetings in the early 1990s, which resulted in a workshop in January 1993, followed by an RFA and a subsequent program announcement. In the 2 years after issuance of the RFA, NIAMS’ wound healing portfolio continued to grow until in 2003 it is approximately $11 million. This included in the past a free-standing clinical trials planning grant. Dr. Margolis was successful in obtaining a NIAMS contract solicitation. NIAMS is now part of the general NIH clinical trials planning grant mechanism announced a few months ago.

Dr. Moshell noted that there is an R01 application in-house for review for a clinical trial that includes the use of sterile maggots. He asked if those present thought such a therapy, if found successful in a U.S. clinical trial, would be widely applicable in clinical use in the United States. Dr. Brem said that his colleagues’ opinions are to use what is there. In real life, however, Dr. Brem’s opinion is that the therapy would not be used if other options are available. Only centers that do not have other options available to them would use the therapy. Others present thought it would be used if found effective in a clinical trial. Dr. Stromberg said that historically, the FDA could support such a licensed product; it was done with heart patients in the 1930s. It is not inherently impossible, but it is an education problem. Dr. Margolis said that they might be adopted as standard therapy if it were demonstrated that they were successful as a permanent debriding agent. However, earlier studies indicated that diabetic foot ulcer patients had to be constantly debrided and decallousified when the maggot therapy was used, so there would be no clinical advantage then to using it for diabetic foot ulcers. It might be useful for pressure ulcers or arterial wounds. Dr. Boulton responded that they are being used in the United Kingdom and he believed they were also being used widely in the United States. Dr. David Armstrong is using them at the VA in Miami, but not for simple neuropathic foot ulcers. For those, aggressive debridement and offloading are the therapy used and most of the wounds heal. Larvae therapy is reserved and is very useful for neuroischemic wounds that can be painful to debride, in which there is not a lot of callous but a lot of slough that is very difficult to remove, even surgically. Dr. Boulton added that his experience is anecdotal, but there was a clinical trial in the Ukraine that was presented in 2000 and an abstract published. Under the right conditions, it is a useful therapy.

Dr. Moshell continued that historically, some of NIAMS manpower activities have been to stimulate epidemiology within skin disease, including chronic wounds, and again Dr. Margolis has been one of the researchers participating in this. In terms of current and future plans, the NIAMS portfolio is predominantly basic research in wound healing, but there has been some movement toward clinical and translational research. There is a Small Grants Program announcement with resources, called a PAR (a program announcement for which special referral guidelines apply), that sets aside money and has receipt deadlines three times a year. A separate committee reviews these.

NIAMS also has a high-risk RFA, which is for individuals who either have been previously funded by NIAMS and who are now looking at a different kind of work, or for individuals who are established investigators who have not been doing research in NIAMS diseases but now want to do work in these diseases. This is an annual RFA with a January deadline. Although NIAMS now participates in the NIH-wide clinical trials planning grant mechanism, they will separately review these. There is a broad range of roadmap initiatives that Dr. Zerhouni, Director, NIH, announced, all of which are essentially NIH-wide mechanisms. Several have to do with epidemiology, clinical trials and translational research.

back to top

In conclusion, Dr. Moshell brought up the subject of a consensus development conference, a type of conference that the NIH conducts several times a year. In order to hold a consensus development conference, the scientific literature must have established that there is an intervention—diagnostic, preventive, or therapeutic—that is not widely accepted in the medical community, that if it were accepted and widely applied, would improve the health of the U.S. population. Dr. Moshell asked if those present felt that the wound healing community was at the stage where such a situation existed for a standard of care for diabetic foot ulcers or other chronic ulcers and that the literature could establish this for a panel of experts outside the wound healing community. If so, he assumed the FDA would recognize this and adopt it as a standard of care for controls in wound healing clinical trials.

Dr. Reiber responded that there are two therapies well-supported by the literature that are beneficial in treating diabetic foot ulcers and associated with better outcomes—debridement and offloading—and both are highly underutilized, particularly by the primary care community. She stressed that it would be wonderful to get that message out.

Dr. Pogach expressed the opinion that, at least from the VA and DoD standpoint, standards should not be derived from consensus conferences. Standards, measures, and guidelines should be the result of a rigorous review of the scientific evidence. He concurred with Dr. Reiber’s statement that the interventions would be based on well-developed randomized clinical trials, even though there are always knowledge gaps in emerging therapies, and a consensus conference could be productive at this time. He emphasized, however, that any standard should be based on uncontroversial evidence from clinical trials, not from a consensus conference.


Dr. Moshell explained that consensus development conferences do not establish standards. They basically confirm that the literature shows that a certain practice has been established to be more effective than what is generally being used in the community and therefore should be brought to the attention of the practicing medical community and brought into greater use. Then, it is up to the FDA to decide where to go from there. Dr. Kravitz noted it is simply a matter of communicating to the general practice community that there are guidelines that have been established, published, and are well accepted and should be implemented in daily practice.

Dr. Margolis stated that although he believes debridement works, the data supporting its efficacy is incomplete. He recommended that a well-done study be conducted to definitely prove the effectiveness of debridement in healing diabetic foot ulcers. Dr. Brem said that a good reason for a consensus conference is to address such issues. He felt a consensus conference would result in agreement to substantially reduce amputations by debridement or other therapies. That would be a valuable outcome.

In response to Dr. Moshell’s question about if there was better literature to support offloading. Dr. Boulton answered that there was a recent Cochrane Update that reported controlled trial data on offloading, but there is a need for larger trials. Dr. Brem thought no one would argue the value of offloading; the discussion would be on which type should be used. Another result of a consensus conference might be agreement on outcome, which patients can be expected to heal. This would at least provide a goal.

As a result of the discussion, Dr. Moshell concluded that most of the researchers present would be willing to participate in trying to hold a consensus conference on diabetic foot ulcers.

back to top

Veterans Health Administration
Dr. Malozowski next introduced Dr. Leonard Pogach, National Program Director for Diabetes, to present a summary of the VA’s Preservation, Amputation, Care, and Treatment (PACT) Program.

Dr. Pogach noted that the VA has achieved increasing recognition as a national system of care within the United States. The agency has been improving outcomes compared to the private sector and trying to decrease disparities. He mentioned that there was actually a public law passed in 1992 (P.L. 102-405, Veterans Medical Programs Amendments) that emphasized the importance of quality amputation care. It identified veterans with amputations as a special disability group and chartered the Special Advisory Committee on Prosthetics and Special Disabilities to oversee the VA. That led to the formation of the PACT program in 1993, developed to meet the changing needs of veterans and to proactively identify and treat veterans at risk for amputations, usually as the result of neuropathy and peripheral vascular disease. Such amputations have decreased but with the increasing casualties in Iraq, now attention is focused on traumatic amputations.

The Undersecretary has issued a series of PACT directives. These are unfunded mandates, which leave medical centers to assume the costs and direction to carry them out. The first established the program and directed each medical center to establish a care system. In 1996, the directive tied the program explicitly to performance measures, and in 2001, the reissued directive called for improved performance measures and the development of high-risk registries within the VA.

Guidelines for foot exams for sensation and pulses were implemented in 1997, and in 1999, if a patient was screened by the primary care provider and considered high risk, the provider referred the patient to a podiatrist. Each quarter, every facility in the network receives foot care measure feedback on their performance from the Office of Quality and Performance. There is very little variation in ratings from facility to facility, less so than in the private sector. Although the program has been a diabetes quality improvement project measure and has been a part of the national quality improvement alliance measure since 1997, Dr. Pogach was not aware of any other agencies aside from the VA and the Indian Health Service that report, use, or publish these performance measures on a national basis. Doing so might help in terms of prevention.

Dr. Pogach commented that after patients are referred to a podiatrist, the foot care programs at individual VAs do differ widely, reflecting the lack of an evidence-based standard of care from which to impose guidelines and performance measures within the agency.

back to top

The data Dr. Pogach presented on amputation rates had not yet been reviewed and approved for distribution outside the agency. He said that hopefully it would be soon. Although the VA began collecting outpatient data in 1997, when the definition of diabetes changed, the agency relies on cross-sectional data on age-standardized amputation rates for patients from 1999 on, rather than depend on the reliability of the 1997-1998 databases. Clearly, as stated by earlier speakers, there has been a decrease in overall amputation rates, major amputations, and moderate amputation rates in persons with diabetes treated at the VA. The number of amputations is remaining relatively constant, though there is a shift from the ratio of major to moderate amputations. The amputation rates for persons without diabetes are very small in number. Eighty percent of the amputations at the VA are performed on individual veterans who have diabetes. VA data will be merged with Medicare data to determine total amputation rates. About 70 percent of VA patients with diabetes are also Medicare enrollees. If a patient comes in for diabetes medicine or a patient visit, the patient is counted as a clinical user, although the person may be receiving the majority of his/her care in other healthcare systems. This can cause denominator effects in the data.

Dr. Pogach next described the VA’s future plans. As a result of the external evaluation of the VA PACT Program in 2002, the Undersecretary felt it was time to further evaluate amputations and re-amputation rates using more rigorous methodologies. The Undersecretary also established a multi-disciplinary panel to develop data elements and performance measures, improve data accuracy and completeness, and strengthen the PACT foot care programs nationwide.

Current VA action plans are to identify unmet patient care needs, identify knowledge gaps in terms of what facilities need and what providers need, and then develop strategies to fulfill these needs. The Amputation Care Advisory Group (ACAG) is multi-disciplinary, both among disciplines and among headquarters offices. In defining how the program is to be measured, the ACAG looks at process measures (screening, risk assessment, referral), patient-centered measures (foot care behaviors, unmet needs), and outcome measures (ulcer rates in addition to amputation rates, functional status, patient satisfaction). Since the VA is a provider system, not a payer system, all the data have to be actionable. There are always competing needs; however, with better data collection and better feedback, the medical center division directors will be able to support better foot care initiatives.

Dr. Pogach remarked that he had been energized by today’s comments from the speakers and attendees. One thing the VA can do is look at their guidelines again based on the emerging better data in foot care. He assured the group that the VA would welcome being a part of the greater community in reviewing and discussing a standard of care. Fortunately, the VA is very much a learning laboratory these days and has funding available, especially for implementation strategies. In conclusion, Dr. Pogach referred the audience to the VA website (www.oqp.med.va.gov/cpg/cpg.htm) to look at some of the agency’s performance measures and guidelines.

back to top

National Institute of General Medical Sciences
Dr. Malozowski presented the final speaker, Dr. Richard A. Ikeda, Program Director, Pharmacology, Physiology, and Biological Chemistry Division, NIGMS.

Dr. Ikeda explained that the National Institute of General Medical Sciences is a little different from the organ- and disease-specific institutes at NIH, in that it is the basic science institute. Its mission is to fund research in the basic sciences of life processes, projects ranging from chemical molecules that mimic an enzyme mechanism to small-scale clinical studies with burn patients or trauma patients. In 2003, the NIGMS budget was $1.8 billion. Because NIGMS emphasizes small investigator-initiated research, it supports about 10 percent of all the research grants that are granted by the NIH. NIGMS also has a very large training component in its mission and thus supports about 45 percent of all the predoctoral trainees at the NIH and 28 percent of all trainees in general who are supported by the NIH. NIGMS also has a medical mission to support research into diseases or trauma that involves multiple organs or multiple diseases. Traditionally, this has led to research in trauma, burns, and wound healing.

The majority of NIGMS grants are R01 investigator-initiated projects, P01 program grants requiring more resources and collaborative interactions, and T32 training mechanisms. Dr. Ikeda noted that problems in wound healing are likely to be P01 grants, but encouraged those present to look at R01s as individuals, as well as the P01s, and to consider a T32 to establish a pre- or postdoctoral training program in wound healing. Because of the emphasis on investigator-initiated projects, there are fewer opportunities at NIGMS for targeted programs, set-asides, centers, and large awards. These are used sparingly to support new areas of research, emerging technologies, areas that are becoming important but are underutilized, projects that require extensive collaborations and data sharing, or research spanning disciplines that have not previously worked together.

Dr. Ikeda presented the following NIGMS current initiatives relevant to wound healing: PA 03-100, Exploratory Studies for High-Risk/High-Impact Research; PAR 02-092, Research Centers in Trauma, Burn, and Perioperative Injury; and PA 03-047, Research on Microbial Biofilms. He encouraged those present to contact him if they were interested in any of these announcements or had suggestions for other initiatives relevant to diabetic foot ulcers. He suggested that they “push the envelope” in pursuing areas of investigation, such as stem cell research.

Centers for Disease Control and Prevention
Dr. Frank Vinicor, Director of the CDC Division of Diabetes Translation, was unable to attend the meeting but forwarded a handout on the CDC’s current surveillance, epidemiology, and programmatic activities and referred the audience to the November 14, 2003, MMWR article on the CDC website. As mentioned by the meeting’s speakers, CDC is documenting preventive behaviors by professionals through NHANES and BRFSS. These surveys show that foot examination rates are slowly increasing, and amputations among persons with diabetes are not increasing, but have remained stable over the past 2 to 3 years. This data is based on discharge diagnosis and not to increased outpatient procedures. NHANES has expanded the data relevant to diabetic foot ulcers. The Healthy People 2010 objective relevant to foot disease among persons with diabetes was refined to include evidence about foot ulcers. In addition to the article cited above, MMWR, in collaboration with the Indian Health Service, published data on rates of amputations among American Indians. Manuscripts are pending regarding initial documentation of preventive care practices and rates of lower extremity amputations (LEA) within the Translating Research Into Action for Diabetes (TRIAD) project. CDC has included LEA preventive care practices as a required “intermediate indicator” of impact of its State-based Diabetes Prevention and Control Programs. This is consistent with PARTS (Performance Assessment Review Tool) and Office of Management and Budget outcome assessments. The CDC has assisted with the development and finalization of a Pharmacists, Podiatrists, Optometrists, and Dental Professionals (PPOD) Primer for NDEP partners. Along with the American Podiatric Medical Association, CDC has also collaborated with CMS to develop policies for reimbursement of diabetic foot examinations based on neuropathy, not just vasculopathy.

Dr. Malozowski closed the meeting at 1:15 p.m. with a final request to contact him with suggestions for subjects for future meetings. He announced that the next DMICC meeting would be on diabetes and obesity.

back to top



Diabetes Mellitus Interagency Coordinating Committee Meeting
Co-Sponsored by the National Institute of Diabetes and Digestive and Kidney Diseases and National Heart, Lung, and Blood Institute

Leveraging the Investment in Obesity Initiatives To Advance Diabetes Programs

National Institutes of Health
Building 31C, 6th Floor, Conference Room 6
Bethesda, Maryland
April 8, 2004

Opening Remarks

Dr. Saul Malozowski, Executive Secretary of the Diabetes Mellitus Interagency Coordinating Committee (DMICC) welcomed the attendees and expressed the Committee’s appreciation in having Dr. Jeffrey Flier and Dr. Lawrence Green present to open the scientific discussion on “Leveraging the Investment in Obesity Initiatives To Advance Diabetes Programs.” Dr. Malozowski is Senior Advisor for Clinical Trials and Diabetes Translation, Division of Diabetes, Endocrinology, and Metabolic Diseases (DDEM), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institutes of Health (NIH). Dr. Jeffrey S. Flier is Chief Academic Officer, Research and Academic Affairs, Beth Israel Deaconess Medical Center, Boston, and Dr. Lawrence W. Green is Director, Office of Science and Extramural Research, Public Health Practice Program Office, Centers for Disease Control and Prevention (CDC), Atlanta.

Dr. Malozowski explained that following Dr. Flier’s and Dr. Green’s presentations, members of NIH and other Federal agencies would present their organizations’ obesity initiatives. He then introduced Dr. Allen M. Spiegel, Director, NIDDK.

Dr. Spiegel welcomed the participants and stated that DMICC is an important group that reflects the coordinated aspect of the Federal Government’s approach to diabetes and, as illustrated by the day’s topic, its focus on obesity, given the unequivocal evidence that obesity is driving the current type 2 diabetes epidemic. As examples of this evidence, he noted that more and more is being learned about how mechanistically obesity leads to insulin resistance and that results from the Diabetes Prevention Program (DPP) showed that even relatively modest weight loss can reduce the incidence of diabetes in those at risk.

As co-chair of the NIH Obesity Task Force, along with Dr. Barbara Alving, Acting Director of the National Heart, Lung, and Blood Institute (NHLBI), Dr. Spiegel announced that the Task Force has framed a comprehensive strategic plan for obesity research based on its vision of creating a true interdisciplinary approach to this multi-dimensional problem. He observed that the group was fortunate to have for this meeting two speakers who are outstanding leaders in their respective areas of obesity research. Dr. Spiegel explained that Dr. Flier has been a leader in aspects of the study of leptin and its various actions and is co-author of two articles in the April 2, 2004, issue of Science, including “The Fat-Brain Axis Enters a New Dimension” with Dr. Joel Elmquist (Science 2004 304(5667):63-64). For the first time (albeit in mouse models), it has been shown that leptin and similar factors can control brain “wiring,” raising very important issues in terms of intrauterine environmental effects and questions about the biological underpinnings that lead to obesity. Not that these are fully deterministic in the same way that genes are deterministic, but there is evidence of a powerful biological force that may be driving some of these aspects as has been suspected.

Dr. Spiegel noted that Dr. Green has been active in smoking cessation efforts, translational research, and public health approaches and is highly knowledgeable of the importance of behavioral, lifestyle, and environmental factors in these efforts. Dr. Spiegel concluded by saying that, in keeping with the Obesity Task Force’s vision, he hoped that in the future there would be a blend of these two distinguished scientists’ expertise as a hybrid or interdisciplinary approach to research on obesity. This hopefully would break down the false dichotomy that the causes and answers are all biological or all environmental. Dr. Spiegel then introduced his co-chair and NHLBI’s Acting Director, Dr. Alving.

Dr. Alving reiterated Dr. Spiegel’s statement that when a major problem needs to be addressed, not only the NIH institutes but other Federal agencies coalesce and bring their individual initiatives forward to be shared. In addition to the NIH Obesity Task Force, there is a Food and Drug Administration (FDA) Task Force on Obesity and a Secretary’s Task Force (i.e., U.S. Department of Health and Human Services Secretary Tommy Thompson). Their efforts are interactive, rather than duplicative. She explained that the initiatives being put forth by the NIH Task Force will include bioengineering, to get a better feeling for measuring energy in/energy out; the built environment (the buildings, spaces, and products created or modified by people, such as schools, land use, transportation); prevention of childhood obesity, both in primary care settings and in daycare centers; prevention of obesity in the workplace; and new treatment options. In addition to the trans-NIH effort, individual institutes have their own commitments to obesity, such as the Women’s Health Initiative. Dr. Alving said that the latest trial results on use of estrogen-alone, which would be published in the April 14, 2004, issue of the Journal of the American Medical Association (JAMA), show that for these women, whose mean age is now 70, their average BMI is 30, indicating that the problem spans all ages.

Meeting Agenda

The meeting’s agenda provided presentations and discussions on the diabetes epidemic and its relation to obesity; the nature of obesity and the obesity epidemic; factors that tend to lead to obesity, including genetic, biological, and environmental factors; therapeutic approaches currently available; lessons learned from public health efforts in smoking cessation and other national health crises; challenges in studying obesity; and possible future directions in obesity research. Current and planned obesity research initiatives sponsored by the trans-NIH Obesity Task Force and by NIDDK, the National Cancer Institute (NCI), and NHLBI and obesity activities of the Veterans Administration (VA), FDA, CDC, and U.S. Department of Agriculture (USDA) were presented by agency representatives. Presenters’ slides can be seen at http://www.niddk.nih.gov/federal/dmicc/meetings.htm

back to top

Presentations and Discussions

Obesity Research and Relationship to Type 2 Diabetes
Jeffrey S. Flier, MD, Chief Academic Officer, Research and Academic Affairs, Beth Israel Deaconess Medical Center, Boston

Dr. Flier opened his presentation by noting that the press worldwide has published the news of the rising tide of diabetes, a disease that encompasses a group of syndromes with hyperglycemia of multiple etiologies, linked to specific complications that are heavily, but not entirely, related to vasculature. More than 17 million Americans have diabetes, of whom 90 percent have type 2 diabetes. This country now has States where 8 percent of adults have diabetes and many others where the incidence is greater than 6 percent. Whereas overall mortality for cancer seems fairly stable, and cardiovascular disease is declining, the diabetes mortality rate is alarmingly on the increase—it has become the sixth leading cause of death in the overall U.S. population and the third leading cause of death in some minority groups. Dr. Flier said that it is known that type 2 diabetes has a mixture of causes related to genes and the environment, with the genetic component being stronger in those with clinical onset at an early age. MODY (maturity-onset diabetes of the young) disorders are now quite well defined genetically, largely involving genes that impact on the beta cell and insulin secretion. Genes are still important in the late onset of the disease, but environmental factors have a correspondingly greater role.

Dr. Flier stated that the genetics of type 2 diabetes has been intensely studied, continues to be intensely studied, and can be divided into two general categories: the so-called monogenic forms, where it is easy to identify a specific genetic determinant, and the polygene category. The monogenic are the minority of cases (less than 5 percent). They involve a subgroup of mitochondrial genetic disorders (1 percent), the MODY disorders (2-3 percent), and then some rare disorders (0.5 percent) that involve either the insulin receptor or PPAR_, for example. The vast majority of the genetic influences are in the polygene category. Although there are continuing efforts in this area, in Dr. Flier’s opinion, the majority of the important polygenes that contribute to type 2 diabetes are as yet unidentified.

As an example of the critical relationship of type 2 diabetes and obesity, Dr. Flier presented information from the Nurses’ Health Study, where the age-adjusted relative risk for type 2 diabetes is highly correlated to body mass index (BMI), one of the gold standard ways of looking at body fat content and obesity. Given obesity defined clinically as a BMI of 30 and above, even in the range of body weights that formally are not called obesity, between 25 and 29.9 BMI, the Nurses’ Health Study showed a very important, powerful relationship to the risk of diabetes. Dr. Flier suggested that “If we didn’t have obesity, if everybody were lean, type 2 diabetes would virtually disappear.” He added that it would be hard to argue against the premise that treating obesity would be a powerful treatment of type 2 diabetes that would probably be unrivalled by any other kind of therapy.

While obesity is an excess of adipose tissue, Dr. Flier said there is no precise cutoff between normal and abnormal in terms of the distribution of body weight and body fat. Obesity is defined medically by a linkage to morbidity. It could be defined differently, depending on how the term is used. However, there are relationships between body fat and morbidity that are not linked just to having a BMI over 30.

The rate of obesity among U.S. adults has been steadily growing. According to the CDC’s Behavioral Risk Factor Surveillance System (BRFSS), more than 20 percent of the adult population was obese in 2000. Relatively speaking, the problem is worse in children. Dr. Flier said that, interestingly, the dramatic change in the prevalence of obesity does not imply that people are 40 pounds heavier now than they were 10 years ago. Between 1990 and 1998, the average body weight of adult males increased by 7.8 pounds; the average body weight of adult females, 8.3 pounds. The reason that this gradual increase translates into such a change in obesity is the distribution of weights in the population. The mean is close to obesity, so just a few more pounds, and a person goes from non-obese to obese. A slight shift in weight crosses the threshold from overweight to obesity.

Dr. Flier and Dr. Jeffrey Friedman (The Rockefeller University, New York) have pointed out, in talking about what causes this diabetes/obesity epidemic—what causes this shift in weight, how does one relate the genes to the environment, and so forth—it must be remembered that whatever the environment is that everyone is in, it permits a major fraction of individuals to remain lean, as well as a major fraction to become obese. Dr. Flier added that this is a worldwide event. While, for this meeting, the critical aspects of the morbidity of obesity are diabetes-related, there is no less concern regarding hypertension, dyslipidemia, atherosclerosis, cancer, and a variety of other medical syndromes.

Dr. Flier raised the question of how obesity and increased body fat have such a significant impact on health, especially diabetes, but other disorders as well. A major change in the scientific community’s thinking and a major enhancement of its understanding comes from the realization that the fat cells, the fat tissues, are an extremely active, communicative tissue, organ, cell type. Not only do these cells produce leptin, which is a critical regulator of energy balance, appetite, energy expenditure, and neuroendocrine function, but fat also produces a whole variety of circulating molecules that are either cytokines, or hormones, or in some cases, metabolites. A theme that has emerged is the clear understanding that some of these factors can actually enhance insulin action and resist coronary artery disease, and others promote insulin resistance and probably promote some of the vascular findings that are a concern, as well. This central principle—that the adipose cell, adipose tissues, and endocrine organs can be influencing systemic biology—is probably the most transforming aspect of the field in the last 10 years, and it is “picking up steam.”

back to top

Next Dr. Flier addressed the general aspects of the pathogenesis of obesity. Like diabetes, like hypertension, like everything else, there are genetic components and environmental components. The genes that are influencing obesity, just as for diabetes, are of the monogenic variety and the polygenic (susceptibility) category. It has been pointed out, and seems to be true, although still surprising, that for some people the genetic determination of body fat and obesity is as strong as the genetic determination of height. It seems counterintuitive to many people, but that is what the studies show. Just like body fat and body weight can be influenced by starvation, famine, war, and so forth, so of course, can height, and to about the same degree.

A few of these monogenic causes are known and illustrate the fact that genetic determination of body weight can exist in very powerful forms in specific cases. That has been exceptionally important in understanding how the system is wired and how it works. In the vast majority of cases, these monogenes have not been identified and almost certainly there also is a category of multiple genes creating susceptibility to particular environmental factors.

Environmental factors are also of obvious critical importance. They relate to the availability and the composition of diet, the amount and nature of physical activity, and other factors that influence energy expenditure. In Dr. Flier’s opinion, the science does not currently exist to recommend specific kinds and amounts of nutritional support for people in the community. The role of different carbohydrate and fat compositions in the diet are not yet adequately known. He added that this lack of knowledge is obvious to the public and should be viewed as a very important subject for further study.

Dr. Flier brought up the increasing portion size of food, described in a paper in JAMA in 2003, as one contributing environmental factor [Nielson SJ, JAMA, Jan 22, 2003]. For example, when the calories in a sugared soft drink are calculated and one looks at the size of the portions ingested, the effect in shifting that weight curve is obvious.

Dr. Flier stated that our genes were honed over millions of years of evolution and perhaps not prepared for the recent environment. Although, it should be pointed out that even tens of thousands of years ago, there was likely to be obesity in some subset of individuals, assuming that they were not truly starving all the time. From a physiologic point of view, obesity is a disorder of energy balance. If energy intake exceeds energy expenditure for a prolonged period of time, there is no other place for energy to be stored, in any significant amounts, beyond fat, and therefore, obesity results. Leanness is a greater threat to survival than obesity, in terms of the speed with which it can cause death; thus, it is not surprising, that there are multiple, redundant systems in place, genetically and physiologically, to resist the consequences of inadequate caloric intake. The same genetic and physiologic systems that have these consequences promote obesity in environments such as those that we live in today.

Dr. Flier explained that the first principle to understand is that the central nervous system (CNS) plays an extremely critical role in orchestrating the decision to eat, to be hungry, to go after food, and to regulate metabolism. It does this through effects, of course, on behavior, on autonomic outputs (things we are not aware of on a day-to-day, minute-to-minute basis), and neuroendocrine functions. These behaviors are integrated, in part, through very discrete centers in the brain that receive information from the periphery and bring about a variety of behaviors and physiologic changes.

Some of the metabolic feedbacks (for example, glucose) have been known for a long time. If a person’s blood glucose was lowered, the person would get hungry and various things would happen to his/her autonomic output and other functions. However, the critical physiologic regulators that provide information related to body energy stores really were not known until 10 years ago, with the discovery of the ob (obese) gene and its product, leptin. These can be categorized theoretically into acute and chronic feedback signals.

back to top

Dr. Flier cited Dr. Friedman’s observation that the gene responsible for the ob/ob mouse was a gene encoding a protein, mainly made in fat, that circulates in the blood and is a cytokine-like molecule. If the molecule is replaced in a deficient mouse, the disease essentially goes away in virtually all ways that it can be measured. This was a true landmark event in the field. Then Dr. Steve O’Rahilly and his group in Cambridge found the first family where there was an ob gene mutation—no functional leptin being made in two cousins. It was clear that the cousins had a severe dysregulation of appetite. From shortly after birth, they were always hungry. They became profoundly obese, but they could be cured with leptin therapy. Today they are essentially lean. This study shows the effects of recombinant leptin in a fully leptin-deficient child—massive obesity as an inborn error of metabolism can be cured by replacement therapy.

Dr. Flier stressed that to look at an individual who has this leptin-deficient disease in the untreated state and to say that the problem is the environment, it is the diet, is of course, absurd. To take the mother away from the child because the mother is thought not to be a good mother, as has happened on a number of occasions in cases like this, is totally unreasonable. As we learn about a pathway and a process, we have to change our understanding of the social implications as well as the medical implications.

The broader context of this, Dr. Flier pointed out, is that not only is the fat tissue (in this case, the fat cells) producing a hormone called leptin, in the absence of which a person has an overwhelming desire to eat, but it also produces a whole group of other factors. Fatty acids are more interesting than they used to be. They are not just fuel. They are also signaling molecules in a variety of ways. It has been known for years that active estrogen can be made in fat tissue, but researchers are just beginning to learn all the implications of that. Much is being learned about various complement factors, cytokines, factors that draw macrophages into fat, factors that regulate the vascular system and thrombosis, factors recently discovered that influence insulin sensitivity, and enzymes that can make local factories for the production of a classic hormone like cortisol. Dr. Flier added that he is certain that there are many more factors that have not yet been identified.

Downstream targets of products released from fat cells include sex steroid and glucocorticoid-dependent processes, inflammation and immunity, glucose and lipid metabolism, vascular tone, thrombosis, appetite (of course), autonomic nervous activity and all of its manifestations, and neuroendocrine function. For example, if the person’s cells do not make leptin, there can be no sexual maturation. Leptin replacement therapy will allow sexual maturation to take place. Dr. Flier spoke of recent studies in review now at Beth Israel showing that in a group of women with the common reproductive disorder of idiopathic hypothalamic amenorrhea (absence of regular menstrual cycles) who have low leptin levels and are lean, if they are given leptin therapy, their reproductive system reactivates. These are examples of the powerful effect of fat on areas that formerly it was not known directly responded to signals from fat. This also includes neural development, an area Dr. Flier’s group has been interested in, in which observations suggest, that in the absence of leptin, in the rodent at least, there are major abnormalities of brain development.

Dr. Flier next discussed adiponectin, the molecule identified by a number of groups that is produced very heavily by fat cells, if not uniquely by fat cells. It is a secreted molecule—actually a group of molecules—that circulates in high concentrations. They are an oligomeric species of different kinds that are just now being worked out. This group of molecules can be measured in the blood. It has been shown that with obesity (and to a significant degree, diabetes), the levels of this circulating protein or proteins are reduced. In the last year or two, there are direct demonstrations that adiponectin or one or more of its species will directly increase insulin sensitivity, increase lipid oxidization in tissues such as liver and probably muscle as well, and have major actions on what Dr. Flier called vascular protection. The exact mechanisms by which these effects occur are still being examined.

back to top

A recently published paper states that adiponectin has direct effects on targets in the brain that regulate energy expenditure, distinct from the targets regulated by leptin. This provides another part of the loop that directly regulates energy expenditure through the brain through specific receptors, a couple of which have been identified. At least one of them seems to activate AMP kinase enzyme, which is a very important enzyme for metabolic integration. In addition, drugs that increase insulin sensitivity and have other salutary effects, such as the thiazolidinediones that were observed several years ago, also induce the expression and levels of adiponectin. Dr. Flier noted that these are examples of areas of previously unknown biology that can be explored and potentially taken advantage of for therapeutic purposes.

Dr. Flier summarized several studies to illustrate key points about different regions in the hypothalamus that are important in the central circuitry for regulating energy balance. A critical area for hypothalamic regulation and integration is in the arcuate nucleus of the hypothalamus, particularly in the leptin-target neurons in this region. Leptin, by one means or another, gains access to this arcuate nucleus region. Dr. Flier focused on two classes of cells in the nucleus. One expresses two neuropeptides, AgRP (agouti-related protein) and NPY, anabolic or orexigenic products. Putting these neuropeptides into this region of the brain will induce animals to eat and to gain weight. Leptin will suppress the activity of these neurons through binding to its receptor on these neurons. Very nearby, an anatomically and genetically distinct population of cells express pro-opiomelanocortin (POMC) and something called CART; these are catabolic or anorexigenic products. When they are put into the brain, animals will eat less and will lose weight. Leptin activates these neurons.

Secondary targets of the leptin-target neurons are neurons expressing melanocortin (MC4R) receptors in other regions of the brain, which respond to mutually antagonistic ligands. One region called the PVH/VMH that has neurons that have the MC4R receptor on it receives input from these neurons in the form of a peptide called _-MSH (alpha-melanocyte-stimulating hormone) and leptin drives this part. At the same time that leptin is driving this part, it is suppressing another peptide that is an antagonist. So there is a yin-yang on this receptor, an agonist and an antagonist. Leptin drives the agonist and inhibits the antagonist; therefore this set of neurons is activated not only by leptin but also by a variety of other factors that are of interest to energy balance. Dr. Flier said that the target of these neurons is, of course, a matter of great interest. By using genetic techniques available to modulate these pathways and individual neurons, researchers are learning about what these neurons do, where they go, and how that wiring system works.

A further part of the pathway involves an area called the lateral hypothalamus. All of this is taking place within a few millimeters of physical space in the hypothalamus. The lateral hypothalamus was previously known to be an area where, if you put a lesion in it, the animal would not want to eat. Now it is known that there are several neuropeptides, but the most preeminent is melanin-concentrating hormone or MCH. When MCH is put into the brain by infusion, the animals will eat more and will get obese. If it is genetically knocked down, they will be lean. If it is genetically overexpressed, they will be prone to obesity. There are direct projections that go from the leptin-target neurons in the arcuate nucleus to MCH neurons in the lateral hypothalamus. This is highly complex, but Dr. Flier assured the group that researchers are confident that this is the underlying biology of the core system, based on results from various genetic experiments.

The current April 2, 2004, issue of Science cited earlier includes a sketch by Dr. Flier depicting three different models of how leptin affects these neurons in the arcurate nucleus. (A fourth model is under development at Dr. Flier’s laboratory.) In the first model, which is the classical approach, leptin receptors turn the neurons on or off and change electrical activity and the expression of various neuropeptides such as NPY and POMC. The second model, which came from a paper published by Dr. Friedman and others, demonstrates that, in fact, not only is leptin acting on these neurons directly through receptors, but in the ob/ob mouse that does not have any leptin, the actual synaptic connections to these neurons can be drastically altered. In some cases, there are more of these synapses; in some cases, less. Even more interesting is the fact that giving leptin to a leptin-deficient animal over a course of 6 hours will make these synapses form again. Although it is not yet understood just where these synapses are coming from and how they are regulated, they can be rapidly regulated by leptin and presumably by other factors, such as nutrition.

The third model is based on the work of Dr. Richard Simerly and his group (at the Oregon Primate Research Center in Beaverton), who have developed observations made earlier at Beth Israel. This earlier research showed that in the ob/ob mouse, its brain weight is reduced and evidences of neural development are suppressed in various regions of the brain. These could only be partially restored by giving leptin to an adult animal. There are important physical projections from these neurons to the paraventricular nucleus. Dr. Simerly very elegantly has shown that in the ob/ob mouse the paraventricular connections are drastically reduced in number and giving leptin to a fully adult ob/ob mouse only partially restores these connections. However, if leptin is given at a critical postnatal period, one at which Dr. Flier had previously shown there was a surge of leptin in the blood, possibly for developmental purposes, the connections are rapidly restored. These developments lead to a need to consider not just nerve activity or the amount of various neural peptides, but how these wiring diagrams are actually formed; the genetic, nutritional, and hormonal influences over their formation; and how they may account for some of the biological facts that are observed, including the effects of early nutrition on later biology.

Dr. Flier emphasized that thus mutations in a CNS pathway clearly can cause obesity in man, whether they are mutations in leptin, in the leptin receptor, in the POMC molecule itself that leads to _-MSH, in a processing enzyme needed to make this maturation take place, or in the melanocortin (MC4R) receptor. If these mutations exist in mice or men, the result is obesity. Most of these cases are, however, rare. On the other hand, as many as 6 percent of relatively young people with severe obesity will have a loss of function mutation at the MC4R locus. It is like the situation in the mouse, where loss of a single MC4 receptor allele will cause substantial degrees of obesity. It is a tightly regulated pathway that cannot be modified very readily without having substantial effects on body weight.

From the viewpoint of obesity in general, the problem is that most persons do not have any of these defined mutations. They have high levels of leptin proportional to their body fat and they have, not only by inference but by direct study, various degrees of leptin resistance. One of the things that researchers at Beth Israel are most convinced of and that they work on is the premise that there is a real biological fact of leptin resistance, and that leptin resistance is not total, is not global, but is highly specific for certain functions and certain locations in the body. Identifying the mechanisms for leptin resistance is currently their single greatest focus. The two molecules that have been heretofore defined as playing a role in leptin sensitivity are PTP1B, a tyrosine phosphatase, and SOCS-3, suppressor of cytokine signalling 3, a member of the cytokine signaling gene family. Dr. Ben Nneel and Dr. Barbara Kahn, as a by-product of studying PTP1B as a modulator of insulin signaling, were able to show a couple of years ago that a genetic knockdown of PTP1B will cause an animal to be more sensitive to leptin or resistant to obesity. Dr. Flier announced that a paper will hopefully be published soon that will show the role of the SOCS-3 in regulating leptin sensitivity. He is sure more information will be coming out in these areas over the next several years.

back to top

Dr. Flier said that another major area of excitement in the obesity domain that also will have great relevance to diabetes is the fact that leptin actions that researchers were initially focusing on rather selectively are seen as integrated with signals regulating meal size. This biology and physiology, in general, was known for a long time. If one expands the stomach and puts food in the stomach, elements go through nerves and inform the brain of something that should limit how much one will eat related to that individual meal. What has emerged is a flurry of work extending earlier studies with the intestinal peptide cholecystokinin (CCK) to now include PYY and ghrelin (PYY coming from the small intestine and ghrelin largely from the stomach). These molecules converge on this same limited region of the hypothalamus. Ghrelin promotes feeding and then PYY released after a meal inhibits appetite. There are still some controversies and sorting out of all the facts regarding this, but Dr. Flier thinks these will be important factors. Even though their primary focus has been on their role in making a person less hungry 10 to 15 minutes after starting eating, these same signals are converging on the pathways regulated by leptin. One thing that is known is that in the absence of leptin, these signals simply do not have a functional impact on the brain. Children, for example, who do not have leptin, will eat until their stomachs are extremely distended because there is no signal that causes them to lose their drive to eat. At the more usual level of leptin and leptin signaling, it is not known exactly how these signals interact.

With regard to the enzyme called 11 beta hydroxysteroid dehydrogenase 1 (11 beta HSD-1), the focus has been on its role in fat as a generator of local glucocorticoid cortisol. The context of this is shown by a 46-year-old male patient who has prominent abdominal obesity, hypertension, and glucose intolerance, a common profile in the obesity, diabetes, and coronary artery disease clinics. He did not have other cardinal signs of Cushing’s syndrome such as muscle weakness and striae, and tests for glucocorticoid overproduction were negative.

For years, many endocrinologists, among them Dr. Flier, wondered whether there was something still going on in this type of patient related to glucocorticoid. Most of the studies, and increasingly now, that are looking at the activity of the 11 beta HSD-1 enzyme, which activates the inactive cortisol to become active cortisol within a tissue, indicate that this activity increases when measured in the fat of people as they become more and more obese. Obese individuals have more of this cortisol-activating activity in fat, and they have more of it in visceral fat.

In transgenic mice that have an increase in the activity of this enzyme selectively, the degree of the increase is similar to what is seen in obese individuals. Dr. Flier’s laboratory created mice that had local cortisol excess in fat, the genetic-initiating event. The mice developed obesity, especially visceral obesity. Interestingly, they also ate more and developed glucose intolerance, diabetes, and when put on a high-fat diet, insulin resistance, hyper-triglyceridemia, a variety of altered adipocyte-derived circulating factors, hypertension, and fatty liver. They had the whole gamut of the metabolic syndrome—just by tuning up this enzyme about three-fold in adipose tissue. The implications of this are that local glucocorticoid reactivation in fat can cause visceral obesity and its complications. It appears that at least many obese humans have something like this. Inhibition of this enzyme might have beneficial effects on obesity and the metabolic/vascular complications of obesity. This is one of the major drug targets in the pharmaceutical field right now. The initial developments seem to be focused on application to diabetes. Dr. Flier’s viewpoint is that a really effective compound—and there is still some question as to which companies, if any, may have a really effective compound—would actually benefit the whole syndrome. This will only be known when a good compound can be tested in people.

Dr. Flier briefly mentioned several other areas of excitement within obesity research that will have larger implications for the whole metabolic field. One is the realization that, just as glucose is both a fuel and a signal, it is not surprising that lipids are a fuel, a substrate for storage, and also a signal for many mechanisms, one of which, for example, would be actions to modulate various nuclear receptors like the PPARs. Another way that they may act would be to focus on their role in the hypothalamus to regulate metabolism. All of the details of this are not known. Dr. Flier presented a slide by Dr. Luciano Rosetti, demonstrating that if long-chain fatty acids in the hypothalamic regions regulate body weight, then increasing these long-chain acetyl CoAs could have the biological consequence of inhibiting food intake and glucose production by the liver. This phenomena that Dr. Rosetti described by a pharmacologic addition of compounds is consistent with a whole variety of other manipulations, both in his lab and other laboratories, that influence these pathways either by modifying fatty acid synthase or the CPT-1 enzyme that modulates fatty acid fluxes into the mitochondria. All these pathways and others suggest that somehow these long-chain acetyl CoAs may bring about signaling changes in the hypothalamus that do modify food intake and metabolism.

Another slide from Dr. Rosetti demonstrated one part of the picture. It depicted a blow up of the arcuate nucleus, the same area that responds to leptin, and showed that there are nuclei there that respond to the addition of long-chain CoAs by a variety of mechanisms—one possible one by the action of these long-chain CoAs to influence KATP channels in various neurons by either activating or suppressing them. These second-order neurons send out signals that go to find brain areas that will influence metabolism. Some of this same kind of biology has been shown with fatty acids and also with insulin given directly into the brain. These findings modify the view that insulin has direct actions on all these peripheral organs, but it also has indirect actions in the hypothalamus through nerves to bring about some of the same actions. There is greater complexity and greater redundancy of the mechanisms.

Another frontier discussed by Dr. Flier in the pathogenesis of diabetes and obesity is the role of inflammation. It is now known that, just as in various atherosclerotic states, there are inflammatory markers in both obesity and diabetes. They are increased in obese individuals and, in some cases, even in early states in the evolution of the disease. Increasingly, these inflammatory biological factors are linked to the pathogenesis and the complications. One example is the role of cytokines in the pathogenesis of insulin resistance. These cytokines are produced in fat tissue, and very recently several groups have demonstrated that in obesity, macrophages infiltrate, occupy, and are obviously doing something within fat tissue. Just as macrophages are in the vessel wall, there are actually a lot of them in fat tissue. In fact, most of the cytokines that are in fat and are increased in fat tissue in obesity, contrary to what was originally supposed, are coming from resident macrophages. There are signals coming from fat cells to draw in those macrophages. Dr. Flier’s laboratory and those of others are now doing genetic experiments to modify factors that bring macrophages in to see what the consequences of that might be.

back to top

Dr. Jerry Shulman at Yale has been focusing on the availability of fatty acids in various peripheral tissues, liver or muscle. Using a modified slide from Dr. Shulman, Dr. Flier showed that a transport protein, FATP1, is required to get fatty acids in high concentration into the cell and then back into the fatty acetyl CoA, the same kind of molecule that seems to be operating in the brain. There are still many mechanisms being invoked, whether through a particular PKC species or through various JNK1/IKKB pathways. Whatever the exact mechanism, there is activation of serine kinase cascades that suppress insulin signaling, among other things.

SOCS-3 is definitively involved in insulin resistance. Inflammation through cytokines will induce SOCS-3 in a variety of brain and peripheral tissues through several mechanisms that will lead to a suppression of insulin signaling. Dr. Flier’s point was that all of this information about fatty acids, inflammation, leptin sensitivity, and other factors is starting to come together in a common set of pathways related to obesity and diabetes.

Other evidence about mechanisms and links between obesity and diabetes is contained in recent research by Dr. Mary Elizabeth Patti and others demonstrating that expression of mitochondrial oxidative metabolism genes is reduced in many humans who have insulin resistance in diabetes. Dr. Flier acknowledged this would include many, many people, of course, with obesity. The two most important papers to make this point recently were by the Patti group at the Joslin Diabetes Center (Patti et al., PNAS, July 8, 2003) and by Dr. Vamsi Mootha at Whitehead (Mootha et al. Nat Genetics, July 2003). This observation, that there is a common reduction in the expression of genes involved in oxidative phosphorylation in tissues of people with diabetes or even pre-diabetes, has now also led to an increasing focus on the pathway that may regulate the oxidative phosphorylation gene program, especially on the molecule PGC-1, discovered by Dr. Bruce Spiegelman. PGC-1 is a nuclear co-activator protein that, among other things, will interact with PPAR_, the receptor for the thiazolidinedione drugs. Dr. Flier stated that it appears that the biology of PGC-1 is going to play an essential role in some aspect of the altered energy expenditure and metabolic fluxes in both diabetes and obesity.

Dr. Flier remarked that there are other aspects of the pathogenesis and biology of diabetes and obesity that could be discussed before moving on to the therapeutic approaches to obesity that are available today. There is a great amount of effort, money, and concern about what to tell people about their diet, their exercise, and their behavior modification. Although not an expert in these areas regarding public health recommendations, Dr. Flier thinks that realistically not enough is yet known to comfortably be able to tell everyone what to do. The available drugs are weak, offering modest benefits along with some complications. Meridia (sibutramine), a centrally acting drug, is of relatively limited efficacy, and there is concern regarding side effects. Xenical (orlistat) reduces a mild state of a malabsorption of fat. There is a variety of over-the-counter preparations. Approximately 120,000 gastric operations are being done this year with some very major benefits but also some concern about resorting to surgery as a major therapy for obesity and diabetes.

Dr. Flier referred to the ongoing debate between the two major positions on therapeutic approaches for obesity. Position 1 says that since the recent major increase in obesity is surely caused by the environment (i.e., diet and exercise), not due to a change in genes, treat obesity by attacking the environment. People adamant about this position believe there is basically no point in studying pathways on a molecular level for the purposes of developing drugs. Just change the environmental factors.

Position 2, which is a version of Dr. Flier’s view, is that the environmental approach is good, please go ahead and try to accomplish it, but we really do not know how to do this right now; therefore, from a public health perspective, trying to develop safe and effective medications makes sense, at least until effective social, environmental, and nutritional ways can be found. The scientific community would be happier if it could avoid the drugs for treating coronary artery disease, but at present, lacking another effective approach, we use statins. Likewise, if “statins” that hit the pathway in a relevant way for obesity can be found, it would be unreasonable to not use them.

There are trials to prevent and delay the progression from impaired glucose tolerance to type 2 diabetes, some of them heavily supported by NIH and showing success through dietary and lifestyle approaches. These include the Diabetes Prevention Program, the Malmo Study, the Da Qing study, and the Finnish Diabetes Prevention Study. The Malmo Study, for example, showed a 63-percent risk reduction for type 2 diabetes after 5 years of intervention. Dr. Flier agreed that there is a definable, clear, and obvious benefit regarding diabetes from these lifestyle changes. His question is: “Are we able to press a button and have these lifestyle changes affect this epidemic?” Obviously, we must try. Meanwhile, the scientific approach of studying the problem at the pathways levels is also essential. Both scientific approaches are needed.

back to top

The pharmaceutical industry is also studying various possibilities. Dr. Flier believes 10 to 15 percent of obese persons could be successfully treated with leptin sensitivity enhancers; however, studies would need to show this. Axokine, a neurocytokine based on CNTF (ciliary neurotropic factor), which was all the way through a phase III study, was eliminated by antibodies. There are many questions about its mechanism of action. Melanocortin 4 receptor agonists are extremely logical but they are hard to develop because they are agonists and they have to work in the brain; also they have some side effects that are related to a Viagra-like action. This may or may not decrease the market for them if they work for obesity. These are still being developed. Beta 3 agonists, a long-standing effort, are mainly limited by the inability to get a really good drug that is bioavailable and selective for the beta 3 receptor in humans.

Dr. Flier briefly listed other approaches, including the MCH pathway and antagonists that published and unpublished data indicate are powerful anti-obesity drugs. It is unknown if they can successfully make it through the process required before treatment in human beings is approved. A large launch is being planned for the endocannabinoid receptor CB1 antagonist category of drugs that would, in theory, cause weight loss, reduce diabetes, and make it easier to give up smoking. More selective serotonin receptor antagonists and other interesting GPCRs (G protein coupled receptors) are being identified and may be important. The gut peptides (ghrelin antagonists, PYY agonists, GLP-1 pathway, CCK) and alternative gut approaches are leading to many therapeutic approaches. For example, the GLP pathway, which is anti-diabetic, is potentially on the pathway for regulating body weight, as well. The HSD-1 antagonists are being developed broadly by many companies. There are molecules such as DGAT 1, ACC2, and FATP1 antagonists that are directed at various aspects of fatty acid metabolism that are, in some cases, getting ready to be tested in humans in maybe a year. The JNK and JNK-related inflammatory pathways relate to obesity in diabetes and are being heavily mined. Nuclear receptors such as PPAR delta agonists seem to reduce obesity and cause fat burning and those are being studied. There is also the possibility of combination therapies.

Dr. Flier suggested he might have been a little provocative in some of his comments. He stressed, however, that there is no argument that the obesity epidemic and associated diabetes, coronary artery disease, and cardiovascular disease represent a major and increasing unmet medical need. Obesity and fat cell biology are at the core of much of the biology of the problem. New insights into the molecular basis for the regulation of energy balance and the periphery offer exciting targets for drug development. Modifications of the environment are worthy and should be pursued. It is Dr. Flier’s opinion that, unfortunately, such attempts are unlikely to reverse the epidemic soon. The goal would be to enable everyone to tolerate the effects of the environment, as many lean people do without treatment, possibly because of their genes.

Discussion

Dr. Spiegel referred to an op-ed piece in the Wall Street Journal by Dr. David Katz, who heads a CDC-funded prevention center at Yale. The article entitled “The Scarlet Burger,” basically took the position, in reference to a short story entitled “The Birthmark” by Nathanial Hawthorne, that our “birthmark” is the thrifty gene hypothesis that states humans were built to prevent starvation. Dr. Katz agreed with that hypothesis, that we will never be able to develop safe ways, or even effective ones, to block that system, because it is so intrinsic to us. In comparing an obesity-reduction drug to Dr. Flier's statin example, Dr. Spiegel said that although some people will take statins so they can eat steak all the time, very few people take statins for cosmetic reasons. On the other hand, any anti-obesity medication that turns out to be effective will be so massively overused and prescribed irrespective of any warnings put on them, that it will be the same as what happened with phen-phen. Not to say that there will be the same side effects, but individuals who want to lose a couple of pounds before their wedding will be prescribed the drug. The bar, in terms of safety, will have to be extraordinarily high.

Dr. Flier agreed that the bar should be high, but not exaggeratedly so. One cannot say that obesity is a huge public health problem that causes diabetes, hypertension, coronary artery disease, cancer, respiratory disease, and so forth, and we have to find a way to stop this growing epidemic, and at the same time say we must not use a drug because some people will take it to go from a BMI of 25 down to 21. Dr. Flier asked if there was a medication that had the same level of safety that a statin does, would it not be introduced to treat obesity because of concerns that a large number of people would want it just to be lean? The matter should be openly discussed, but as a physician, Dr. Flier would feel obligated to try to help those who could be helped by it. He also would do everything possible to ensure the drug’s safety and prevent abuse of it. Dr. Flier added that although he did not endorse this, there is some biological evidence that the population would be healthier if the curve was shifted toward more people whose BMIs were 24, 23, or 22. If a pathway could be safely tuned up so that more people had these lower BMIs, as some healthy people do, it might be a health breakthrough. Dr. Flier stated that too much fear of how a drug might be misused by some people runs the risk of underplaying the beneficial effects of having a successful medication for those who need it. He believes it is possible to find elements of the pathway that can be “tweaked” to do this.

Some of Dr. Flier’s colleagues and friends who, when they heard he was coming to speak at this meeting, urged him to make the point that they understood the interaction between genes and the environment—the biological position and the environmental position. At the same time as things are being discussed and tried to affect the environment, they are close to finding new ways of changing the pathways and hopefully finding an effective drug. Dr. Flier expressed his concern that this cannot happen if the community becomes excessively and prematurely concerned about the consequences of a drug being abused.

In response to a comment that some obese people are healthy and do not have glucose intolerance or diabetes, Dr. Flier responded that such persons may not need to be treated medically, but a question remains regarding what “health” is in an obese person. The number of people who are fully healthy at each degree of obesity is a complicated question. Granted that there are obese persons whose blood pressure and LDL are lower and their HDL higher than people who are much leaner than them, this is not surprising because the various hormones that influence pathways and result in complications have to be there in a certain form to be acted on. There are genetic variations and other kinds of variations that influence how any given obese person will respond to having low adiponectin and the level of adiponectin in some of these people will not be as low as it is in others. Dr. Flier added that we are beginning to better understand the heterogeneity between obesity and the complications of obesity.

Another comment concerned whether Dr. Flier had heard that the practice of treating obesity with thyroid hormone seems to be coming back into fashion. He replied that the problem with thyroid is that in the vast majority of instances where it has been studied and given to a non-hypothyroid individual, it does not cause much weight loss and it can cause adverse effects of hyperthyroidism. There is some interest in examining that model and trying to define agonists for thyroid hormone receptors that may be selectively able to stimulate aspects of energy expenditure, for example. That is a perfectly legitimate thing to do under clinical trial circumstances.

Dr. Judith Fradkin, Director, DDEM, NIDDK, asked what Dr. Flier thought NIH or other agencies represented at the table might do to foster development, other than continued support of the kind of basic research that he had discussed. Dr. Flier answered that he certainly would not underestimate this basic research. It is necessary to find the right balance between funding the excellent investigator-initiated studies that are looking at these hypotheses and bearing fruit and finding more effective ways to catalyze interdisciplinary research. The interdisciplinary approach, which Dr. Flier finds critical today, requires a greater degree of infrastructure than used to be required. To do first-class work in energy balance in a mouse today, a researcher cannot be a lone citizen setting up a lab and doing it on his/her own. The investigator needs the genetic models, the measurement equipment, and increasingly sophisticated ways to study behavior, and so forth.

back to top

Dr. Spiegel addressed these comments by saying that it is his perspective that NIDDK, while focusing on translational research, should not be either reinventing the wheel or doing what private industry can and should do, a lot of which is market-driven. The market here is so vast that there is no need for NIH to be heavily invested. Diabetes, cardiovascular disease, cancer, and so forth are not orphan diseases that tend to be neglected by industry. NIH investment led to aldurazyme, and even that required industry involvement. For mucopolysaccharidosis that affects a handful of people around the world, a case can be made for NIH involvement. For these other diseases in NIDDK’s areas of responsibility, Dr. Spiegel is comfortable in investing heavily in the basic research that will illuminate targets, which industry could do, but will not necessarily. In addition, he believes in using innovative ways to foster an interdisciplinary focus. For example, the NIH Director, Dr. Elias Zerhouni, added monies for the Obesity Task Force’s Fiscal Year 2005 budget and one of the initiatives from the Task Force is in the neurobiology of obesity. All the neuroscience institutes, along with NIDDK, will be invested together in this initiative. Dr. Spiegel said that clearly the convergence of neuroscientists as well as endocrinologists such as Dr. Flier is needed, and Dr. Flier agreed.

Dr. Spiegel went on to say that, however, the environmental and lifestyle approaches, by and large, are not going to be the provenance of the private sector. Nike may be interested in pushing physical activity, but the lifestyle or behavioral area is where NIH, both from basic support of economic approaches, policy approaches, and others, really needs to be, in conjunction with its sister agencies. Philosophically, that is how Dr. Spiegel would approach it.

Dr. David Acheson, Chief Medical Officer and Director of Food Safety and Security, Center for Food Safety and Applied Nutrition, FDA,, agreed with Dr. Spiegel that the incentives from industry to get involved in environmental research and change is not there because the area is not market-driven.

Dr. Spiegel also cautioned against glibness in addressing issues in this area. Food and beverages are not the same as tobacco, so pulling soda machines out of schools will not necessarily make everything fine. There are historical controls as well as parallel controls that can be compared with schools where an intervention is implemented, and the outcomes of such studies ultimately will inform policy. Rigorous science is needed, of course, and yet some things need to be implemented while the science is being developed. In regard to implementing measures for which there is not yet a rigorous evidence base, Dr. Spiegel explained that it comes down to a question of risk-benefit. If there is no risk to it, if it is perfectly reasonable, then the evidence threshold does not need to be very high to implement. He added that we need to consider such values.

Dr. Flier agreed with a comment by Dr. Philip Smith, Deputy Director, DDEM, NIDDK, about molecules and pathways involved in individual differences in energy consumption and weight gain as a good area for research. Dr. Flier referred to a paper from the Mayo Clinic a few years ago that said if one took a group of people in a GCRC type of study and paid them to eat more by agreement and then studied the extent to which they gained weight, some people gained more, some people gained less. One correlation was non-exercise-associated thermogenesis, rather than basal metabolic rate. This was related to physical movements—posture, fidgeting, and so forth. Probably there are biological determinants in part over the extent to which this kind of a pathway is being engaged. It is also known from some of the genetic models, for example the MCH pathway, that leptin or other molecules may influence not only energy intake, but also behavior and physical movement, as in the way a mouse moves around a cage in various non-purposeful but energy-burning ways. These pathways are tied in together. Ultimately, thinking along these lines, we may well have drugs that modify appetite and, at the same time, in a favorable way, cause a person to be a little less of a couch potato. The fact that we do not understand why some people are perfectly happy to not move and others need to move a lot does not mean there is not a neurobiological underpinning there.

Translational Research: Lessons From Public Health Achievements of the Late 20th Century for Emerging Health Issues of the 21st Century
Lawrence W. Green, DrPH, Director, Office of Science and Extramural Research, Public Health Practice Program Office, Centers for Disease Control and Prevention (CDC), Atlanta

Dr. Green stated that, in keeping with the spirit of the theme of the meeting, he wanted to present a convincing case for leveraging the investment in obesity, since, until there is some success in controlling obesity, there is not much else other than the investment to leverage. As Dr. Flier said, the obesity epidemic would not be turned around quickly using environmental interventions, so he trusted that a pill might be found to assist with the problem. On the other hand, there are some very persuasive cases to be made for the experience gained in turning epidemics of this kind around through a combination of educational-behavioral and environmental interventions.

In the 20th Century, and the last third of the 20th Century in particular, there have been some commendable successes: the control of birthrates in developing countries with the family planning movement that began in the 1960s; the rise and then gradual decline in the use of cigarettes; and success with hypertension control and a 58-percent reduction in stroke deaths since the early 1970s. Half of what was lost in the first half of the century to smoking, automobile-crash deaths, and cardiovascular and stroke deaths was regained in the last third of the century. Automobile injuries, especially alcohol-related injuries, have decreased through success in road design, automobile design, and seatbelt usage, a combination of educational-behavioral and environmental interventions.

Dr. Green has been particularly involved in efforts to control birthrates in Bangladesh, the tobacco epidemic, and in hypertension control. Recently, he has undertaken a critical analysis of the tobacco experience. Dr. Green agreed with Dr. Spiegel that there is no easy parallel to be drawn with the obesity epidemic insofar as the industries that must be dealt with are very different, the private sector interests are different, and the product is very different. There is virtually no redeeming value in cigarettes, whereas people cannot live without food. The strategies used for the tobacco epidemic and other initiatives do have a story to tell, however, and one that Dr. Green believes can be usefully analyzed to leverage initiatives in obesity control.

Dr. Green proposed that seven lessons be drawn from these previous initiatives, beginning with an examination of the data on the tobacco epidemic and the events that surrounded the shifts in that curve over time. Some of those shifts were associated with economics. The Great Depression was one of the first points in time when there was a very palpable reduction in tobacco consumption. The end of World War II also shifted the economics, but more to the point, shifted people’s social relationships around the product. The product access and availability was changed at the end of World War II. The first smoking cancer concern publication in the Reader’s Digest caused a gigantic leap in public awareness of the connection between tobacco and their health, and a consequent drop in consumption that was even greater than the previous two. The first Surgeon General’s report in the mid 1960’s was the point in the curve where the corner was turned from the 65-year increase in tobacco consumption to the beginning of the decrease. Probably the most salient event in recent decades was the Nonsmokers’ Rights movement. What parallel might be found in the obesity issue is a matter of conjecture and sometimes humorous speculation. Whether you can get into your airplane seat next to the person beside you may be a form of the Nonsmokers’ Rights movement in relation to obesity. Seriously, it was the galvanization of public sentiment and concern around the effect that people who smoked were having not just on themselves but on other people that really accelerated and sustained the reversal of that epidemic.

Dr. Green’s Seven Lessons
1. Surveillance and making better use of natural experiments
2. Comprehensiveness
3. Ecological Imperative
4. Threshold spending
5. Environmental influence and settings imperative
6. Education imperative
7. Evidence-based imperative and limitations

The first lesson to be drawn is the imperative of surveillance. Tobacco control has had access to good data on a population scale. Surveillance data have given us the opportunity to conduct continuous and repeated natural experiments. Surveillance has been a key to establishing baselines and trend lines that can be projected to warn against neglect of an issue and to put an issue on the public policy agenda. Surveillance also has been the key to showing change in relation to other trends, policies, and program interventions and has been the key to comparing progress in relation to objectives and programs, over time and between jurisdictions.

back to top

For example, the ongoing surveillance data in the 1990s on States’ smoke-free indoor air legislation along with corresponding data on consumption in these States allowed the drawing of some inferences about the relationship. If there had been only one-time surveys of the independent and dependent variables, it would have been impossible to determine cause and effect. Did it mean that in States in which people were smoking less, they wanted more clean air legislation? Or was it that more clean air legislation was causing people to smoke less? Having surveillance data over time and between jurisdictions has helped track the order of events or trends and has been a powerful tool in several of the public health epidemics mentioned earlier.

As an example of how data between jurisdictions and over time were used to develop best practices, Dr. Green said that surveillance data were available for 48 states at 3 different points in time (1984-88, 1990-92, and 1992-96) showing no change in their rates of tobacco consumption. Two additional States, California and Massachusetts, showed major changes. California was the first off the mark, with an increase in tobacco taxes funding major campaigns and policy initiatives in the late 1980s, so that in the early 1990s, they showed a virtual doubling in the rate of decline in smoking compared with their earlier period and compared with the other 48 States. Massachusetts got a later start, so they showed no change during that 1980-90 period, but in the subsequent period, they virtually doubled the amount of money they were investing per capita in tobacco control and achieved a near tripling of the rate of decline in tobacco consumption in that latter period. What worked? The comprehensive programs of these two States, combined with their tax increases, resulted in two- to three-times faster declines in adult smoking prevalence, a slowed rate of youth smoking prevalence compared with the rest of the Nation, and accelerated passage of local ordinances. There were similar, though later, changes in the same directions and of comparable magnitude associated with comparable program spending, in Oregon, Arizona, and in the youth population segment of Florida, where Florida conducted a similar type of program but specifically for youth.

Combined, these data gave a much better sense that something was working. What was it? The main lesson that had to be drawn when the data were examined was Lesson 2: the importance of comprehensiveness. There was no magic bullet, to be sure, in environment or in behavior. The components of tobacco control programs, when isolated, could not be shown to stand alone. There were many essential components; no one of them by itself could account for the changes that were observed in tobacco control. Also, any combination of methods was more effective than individual methods. Perhaps due to a dose effect: The more components, the better the results .The more components, the better the coverage. Dr. Green concluded that a lot can be said for comprehensiveness just from the standpoint of reaching different people with different methods, because different people will be responsive to different methods, different messages, and different environmental changes.

The Office of Smoking and Health at CDC put these lessons together from California and Massachusetts in a document entitled, “Best Practices for Comprehensive Tobacco Control Programs.” Dr. Green noted this was different from many of the best practices guidelines that come out of the systematic reviews of randomized control trials because these were not randomized controlled trials. The Office did consult the randomized controlled trial literature, but the problem was that smoking cessation studies in clinical practices had not made much progress over the decades, whereas these large-scale State programs with their combination of policy, environmental changes, and mass media efforts to change public perceptions did produce massive changes.

The components of comprehensive tobacco control programs in the “Best Practices” manual included community programs; State-wide programs, because community programs without the back-up of State-wide support could not stand alone; work with the medical care sector in maintaining a relationship to the chronic disease treatment programs for people with diseases that had been associated with their smoking; school and worksite programs; enforcement of policies that were passed because failure to enforce new laws only made a mockery of policy changes; countermarketing in relation to what the tobacco industry was doing; cessation programs; surveillance and evaluation; and administration and management. These nine things were found to be essential components of a comprehensive approach to reversing that epidemic. Dr. Green suggested looking for parallel components in a comprehensive approach to obesity and diabetes control.

back to top

The dose-response response effect seen in the analyses of these data suggested a clear relationship between how much was spent in the programs per capita relative to the effect, aside from the effect of the price increases that went with the taxes in both California and Massachusetts. Somewhere around $6 per capita, there was a point where further expenditures per capita did not show a proportionate increment in the effect in behavior change at the population level. This point of diminishing returns can be compared with the threshold level of something like $2 per capita below which very little increase in the rate of declining tobacco consumption rates could be shown. More on the threshold level follows below.

Lesson 3, the ecological imperative, was the need to address the problem at multiple levels—individual, organizational, institutional, community, State, regional, national, and international. These levels of intervention needed to be mutually supportive and complementary. There needed to be connections between the State and the community, the State and Federal initiatives, and Federal initiatives with other national and non-governmental State efforts. The MATCH model developed by Dr. Bruce Simons-Morton and colleagues has attempted to articulate how this ecological approach can play out in the planning process. The model starts at Phase 1 by defining objectives or goals in relation to health status. Phase 2 is intervention planning and has three sub-phases: (1) selecting the intervention objectives at levels of healthful behavior and policies, communities, and organizations to support the behavior through initiatives and changes in professional practices and policies; (2) identifying and selecting channels and mediators such as community leaders, community norm shapers, organizational decisionmakers, and individuals at risk; and (3) selecting the intervention approaches and target strategies to influence governments, communities, organizations, and individuals. Phases 3 and 4 are development and implementation, respectively. Phase 5 is evaluation at several levels: process, impact, and outcome evaluation. Dr. Green explained that this model was mostly built on the natural experiments provided by surveillance data.

Dr. Green said that another way to look at levels of intervention is not from the individual to organizations and communities and States and Nations but in an upstream/downstream sense of the causal chain. This premise starts with the people who are afflicted with the complications and to whom tertiary prevention initiatives are applied to prevent them from dying from their complications. Moving upstream to those who are afflicted, but without complications, secondary prevention is employed to prevent them from developing complications. Next upstream are the vulnerable populations for whom primary prevention is tried to prevent them from becoming afflicted in the first instance. This is where the obesity control efforts have their most promising contribution. Even further upstream, a strategy that might be called targeted protection is used to intervene with potentially vulnerable populations and create safer and healthier environments for those populations to prevent them from becoming vulnerable. This is where the environmental approach has its potential benefits.

General protection might be trying to improve adverse living conditions that limit people in their ability to pursue the healthful lifestyles recommended. Dr. Green presented one example that is playing out very saliently, which is the recent recognition that most people who live in relative poverty have very little access to fresh fruits and vegetables. The grocery stores to which they have access do not carry these products. So even if people are convinced that they should be eating fresh fruits and vegetables, they are having a hard time acquiring them.

The spectrum in carrying out this dynamic ranges from the work done by health professionals, particularly medical and public health policy, to the public work that needs to be done in collaboration with the educational sector, the private sector and employee programs, and activities in collaboration with other sectors of society to develop healthy public policy that will lead to more healthful environments.

Dr. Green stated that some of this is exactly what the Division of Diabetes Translation is attempting to do through its State-wide Diabetes Prevention and Control Programs (DPCPs). A study of some of these CDC model programs by Dr. Judith Ottoson and her colleagues has suggested that they could be conceptualized as follows. At the State level, the DPCP attempts to affect and strengthen the 10 public health essential functions or services. It does so initially and primarily through the environment and through its work with partners and stakeholders. In so doing, it hopes to create or to mobilize community interventions, health communications, and health system actions. These, in turn, are seen to affect the more proximal determinants of health outcomes through policy changes, system changes, and behavior changes, all of which cumulatively affect national objectives.

In addition to the point of diminishing returns and expenditure, and the dose-response effect observed in per capita spending to bring about some of these reversals of epidemics, Dr. Green said there is also an understanding that there is a threshold level of spending necessary to get any effect, which is Lesson 4. First, a critical mass of personal exposure is needed for individuals to be influenced. Second, a critical mass of population exposure is necessary to affect a detectable community response. Third, a critical distribution of exposure is necessary to reach segments of the population who are less motivated. Looking at the various States in relation to their relative success, Oregon, Arizona, and California, in a step-wise fashion, spent $2 to $4.50 to $5 per capita. Massachusetts was spending over $11 per capita, and the effects that California and Massachusetts got relative to these expenditures have a correlation of nearly one.

Lesson 5 is about environmental influence and the settings imperative. Environments provide opportunities and cues and enable choices. Social environments reinforce positive behavior and punish, or can be organized to punish, negative behavior. Legal penalties and financial incentives can be built into environments. Settings can be seen or understood as the best social definition of environments, insofar as it is within social settings that most interventions are organized and implemented. Dr. Green pointed out that worksites and schools are the best examples, because they are distinct from the medical care settings where people normally think of intervention taking place. However, worksites and schools are set up in relation to their own objectives, which usually do not include health except as an instrumental value.

back to top

Environments had the opportunity in the tobacco epidemic to intervene specifically on things like smoke-free ordinances in workplaces and restaurants and some that applied to both workplaces and restaurants. There was a gigantic leap in the number of localities that were passing such ordinances in the early 1990s. There was also the move to control the availability of vending machines that had tobacco in them. Vending machines that have high-fat/high-sugar foods in them in schools is perhaps too obvious a parallel, but they are a potential target.

For Lesson 6, the educational imperative, Dr. Green listed public awareness of risks and benefits as necessary to gain public support for change. Public interest in lifestyle options is necessary to engage people in changing their own behavior. The public needs to understand the behavioral steps to take in order to change. Public attitudes toward the options and the steps must see them as do-able. Public outrage at the conditions that have put people at risk or in danger is needed. Finally, public support of the personal and political actions that may be necessary to change environments and behavior is key, which requires a well-informed electorate supporting the actions necessary.

For Lesson 7, the evidence-based imperative, Dr. Green noted that because the evidence is so weak at this moment in time with respect to the obesity epidemic, several bridges will be necessary. Best practices indicated by research will need bridges for their application in practice in underserved areas, because the research, at least until now, has mostly been done in relatively more affluent populations. Best practices from research will also have to be bridged to appropriate adaptations for special populations. The success of individual behavior changes of the affluent will need to be bridged to the system changes needed to reach the less affluent and the less educated. Finally, a bridge is necessary between university-based, investigator-driven research to practitioner- and community-centered research to make the necessary adaptations and tailor interventions to fit the vast cultural and social differences across communities in this country.

Dr. Green offered a vision for future effectiveness- and community-based best practices that would emphasize control by practitioners, patients, clients, communities, or populations to address the many variations referred to earlier. He called for emphasizing the need for local evaluation and self-monitoring as much as Nation-wide surveillance. There is a need to synthesize research other than just randomized controlled trials, because so much of what we are trying to get a handle on in this arena does not yield easily to randomization or other experimental control. He saw research on tailoring and new informatics technologies as holding a growing potential to overcome the essential limitation of mass media, with its homogenization of messages for masses, and using new technologies to tailor the information to specific sub-populations and even down to the individual level.

Dr. Green recommended more transdisciplinary systematic study of place, setting, and culture by engaging anthropologists, social geographers, economists, and others who understand the relationship between place or setting and organization and the changes that may be necessary in the various environments to control obesity. He considers “best practice” as a process to combine and adapt packaged interventions as opposed to the tendency to apply interventions homogenously across a whole population. Packaging programs and asking communities to mount them in a public health sense at the community level simply is not working. Dr. Green recommended that packaged interventions be pulled apart and repackaged to fit the community using a population-based diagnostic planning and evaluation cycle. “Best practices” understanding has to be based not just on the evidence for the intervention having demonstrated its efficacy in one or more studies, but its effectiveness when taken to other populations. A second set of handouts on definitions of translational research and some concepts in that area were distributed to attendees in their meeting packets and went more deeply into these issues. (See From Efficacy to Effectiveness to Community and Back from Clinical Trials to Community: The Science of Translating Diabetes and Obesity Research, January 12-13, 2004, Bethesda, Maryland, http://www.niddk.nih.gov/fund/other/Diabetes-Translation/LawrenceGreen.pdf

Dr. Green proposed that a “matching, mapping, pooling, and patching” process take place. He stated that evidence will have its limitations, inevitably and forever, in relation to such moving targets as population behavior. What he is hoping to find is a way of structuring the evidence and filling the gaps in the evidence that is more systematic than what has been attempted in the past. He recommended starting by matching levels of intervention with evidence-based best practices, as illustrated by the MATCH model he described earlier. Next, more systematic ways of mapping the changes against theory is needed to identify where the empirical evidence is insufficient, since it always will be insufficient, and how theory generated from research on related behavior or problems can be applied deductively. Theory is one of the tools that social and behavioral scientists have to bring to bear in their work with populations. Third, pooling experience from model programs that have not yet been submitted to extensive evaluation will help fill the gaps in evidence and theory. Looking at model programs that seem to be working, even though their evidence is not entirely in, is a necessary part of the process of building programs at the community level. Lastly, Dr. Green cited patching the remaining gaps with local experience, indigenous wisdom about local experience in that population. Finally, there needs to be a commitment to evaluate the resulting innovations as part of the process of matching, mapping, pooling, and patching.

Summary. In summary, Dr. Green suggested that good surveillance is the initial key to leverage the initiatives in obesity, based on leveraging the successes in birth control, in injury control, in hypertension control and the reduction of the stroke epidemic, the cardiovascular disease experiences, and changes in public health epidemics in the last third of this past century. Surveillance is needed on food consumption, BMIs, and obesity, and on the policies and practices of agencies and governments. Through good surveillance, there can be better use of natural experiments.

back to top

Secondly, a comprehensive approach to this problem is required. There will be no magic bullets. Even if Dr. Flier and his colleagues are successful in developing drugs to help with obesity control, there will be problems in getting people to use them properly. All of which then brings into play the principles from the other lessons to be learned from the achievements of the late 20th century: the ecological imperative—the need to attack difficult problems at multiple levels; threshold spending—the need to get to at least a critical mass of investment before expecting to see any effect; the environmental and settings imperative—the need to adapt our strategies to different settings and to work with other sectors to do so; the educational imperative—the need to build an informed electorate and an aware public, if not an outraged public; and the evidence-based imperative and limitations that lead us to pool the data we have and to fill the gaps in those data with strategic and creative, innovative, and evaluated interventions.

Dr. Green complimented Dr. Spiegel and his colleagues for their leadership at NIH in addressing translational research issues. He congratulated NHLBI for its long history in translation research, especially the National High Blood Pressure Education Program, which he saw as a model to be emulated. He also extended his congratulations to the National Cancer Institute (NCI) for its efforts in tobacco control.

Discussion

Dr. Spiegel told the group a Senator had asked about parallels between smoking and obesity, indicating this is on congressional leaders’ minds, as well. He observed that fundamentally the situation is more complex than “you don’t have to smoke to live” versus “you have to eat to live.” With smoking, although there are pipes, cigars, smokeless tobacco, and filters, fundamentally the issue is smoke versus no smoke. With obesity, we are talking about diets of various macronutrient composition—Do you count calories or not?—and degrees of physical activity—What kind? It is much more complicated and difficult to convey any specific kind of message.

To continue the comparison between smoking cessation and obesity control, Dr. Spiegel referred to the issue of secondhand smoke and the Nonsmokers’ Rights movement, arguably a crucial inflection point, versus such issues as airline seats. This leads to the stigma issue. There is certainly far less societal hesitation in stigmatizing smoking and smokers. Given these differences in the two problems, the question is “Is there any parallel?” Dr. Spiegel suggested two worth considering. One is advertising to children, because that has some parallels in the tobacco use situation such as Joe Camel-type messages. Even here, there is tremendous resistance on the part of industry. Associated with advertising is the issue of TV watching, for which there are controlled studies, including an NHLBI one which is not yet reported but in the final stages. There appears to be efficacy to reducing children’s TV watching, but it is unclear; the larger study may illuminate whether the efficacy is attributable to decreased sedentary behavior, the effect of the ads, or eating while watching TV. The industry response is that it is strictly a matter of parental choice and responsibility.

Dr. Spiegel gave health insurance as the second parallel because an economic case can be made that obesity, like smoking, leads to increased health care costs. This brings up issues of implementation and again, stigmatization. Should obese people have to pay more for health insurance? These comparisons raise very difficult issues.

Dr. Spiegel also brought up the issues of public perception and education. Dr. Green had pointed out the need for comprehensiveness and spending to inform the public. Currently, Americans are spending some $33 billion on various forms of weight control. This clearly indicates that there is a strong perception of a problem in the public’s mind. The paradox is that it is focused on cosmetic aspects. This may be counterproductive in terms of unrealistic expectations of an ideal weight and how much weight loss is necessary to get where an individual wants to be as opposed to the more modest amount that would make an actual health difference. In this sense, in terms of the Surgeon General’s report on smoking and public perception, the parallel here may be to focus on the health imperative to balance unrealistic images of slimness portrayed by the TV and print media.

back to top

Dr. Green responded that the same reservations regarding aggressively tackling the industries supporting smoking were expressed about tobacco control very early in that experience. The peak in 1966, with the Surgeon General’s report, was preceded by a period in which nobody would touch the tobacco industry. It was followed by a period in which there was still a lot of reluctance to get tobacco use off the television screen, the radio, and the airwaves. The industry decided to take themselves off because of the high costs of their competitive advertising on mass media. They found they could be far more effective if they used billboards and more targeted media. The diffidence in being willing to potentially stigmatize obese people was the same expression of diffidence about smokers at another time in our history. There was a great reluctance to restrain their rights to smoke, to stigmatize them if they were smoking in public places. Dr. Green said that in speaking of these reservations, he was not suggesting a direction that should be taken. He was simply making the historical observation that similar concerns were expressed before.

Dr. Green added that there is a term in the social and behavioral sciences applied to health called “blaming the victim.” There is deep concern about this stigmatization problem, especially in stigmatizing people for outcomes of behavior over which they have little control. That is why it is felt that environmental conditions must aggressively be made more conducive to behavior change, rather than mercilessly issuing messages to try to change people’s behavior unilaterally.

Dr. Fradkin commented that a fundamental difference, though, between smoking and obesity is the whole secondhand smoke issue and the fact that the behavior of a person who smokes really does have a harmful effect on the people around him or her, whereas there is not a comparable issue with regard to obesity. She also asked Dr. Green to elaborate about using surveillance to make use of natural experiments going on in society. She asked what conclusions could be drawn from the surveillance that has shown such massive increases in obesity and in diabetes over the past decade in terms of the inadequacy of current public health interventions and what types of interventions should be tested in that kind of a paradigm.

Dr. Green replied that it is not yet possible to connect the changes in obesity relative to changes in practices, programs, or policies because the right practices, programs, and policies to affect obesity have not yet been identified. As practices emerge and are tried by various State governments or communities, having the surveillance in place and having the multiple baselines such surveillance provides will enable us to detect shifts in the prevalence of the problem in relation to the changes occurring down the line. For example, communities are beginning to work with the built environment. Will that make a difference? We do not have good data yet, but getting such data requires building the surveillance systems to track it.

Dr. Spiegel added that despite NHANES (National Health and Nutrition Examination Survey) and other surveillance systems, it appears that precise quantitative measures of either caloric intake or of energy expenditure do not exist. Basically, all that we have is inadequate and suspect self-reports. This fuels a debate in which the food industry can say, “It’s all lack of physical activity, of being sedentary,” and vice versa. He asked Dr. Green if this was true or an overstatement of the case. Dr. Green replied that there is an enormous effort underway in various sectors to get better measures, to validate self-reports and diary-based reports and so forth, on consumption and physical activity, but we are not there yet.

Dr. Rachel Ballard-Barbash, Associate Director, Applied Research Program, Division of Cancer Control and Population Sciences, NCI, stated that NCI’s Applied Research Program has a major focus on developing improved measures in diet, weight, and physical activity. This is a huge challenge that she considers distinctly different than the challenge faced in tobacco. As demonstrated by Dr. Green, it can only be addressed with increased focused research. With regard to surveillance, Dr. Ballard-Barbash noted that improvements are being made to NHANES such as using an accelerometer together with self-reporting data on physical activity. Even recognizing that accelerometers only give us one part, it will be possible to have a more objective measure of how the situation changes over a 4-year period. There are other techniques that can be implemented to examine data and do better surveillance of the various issues as we move forward. She agreed with Dr. Green that we need to build surveillance systems to understand what is happening in terms of interventions and policies around the country and to be able to link them to the surveillance data that we have on health behaviors. There may be a point where we will have to recognize that some of the self-reporting just does not work in this field. Hopefully, better technologies will let us find cost-effective ways to improve on that. Regarding this, Dr. Spiegel said there is an NIH bioengineering initiative to develop such technologies.

back to top

Dr. Spiegel remarked that the BRFSS data for diabetes and obesity routinely shows, State-by-State, a direct inverse correlation between education status and the severity of either obesity or diabetes. It does not track socioeconomic status, although it may be that education here is a surrogate for socioeconomic status. Dr. Adam Drewnoski claims that, at least in the state of California, there is a surveillance system for obesity and diabetes that is similar to the United States Renal Data System, which geographically across the entire country identifies by ZIP Code, county, and so forth, the extent of hemodialysis and transplantation. The California system shows that in Malibu there is very low obesity. In Alameda County, near Oakland, there is very high obesity. This again shows an inverse correlation with socioeconomic status. Dr. Green alluded to this with regard to adapting programs for the affluent versus the underserved. Dr. Spiegel recommended that there be more definitive research on the economics of this and, if it is real, ask if it reflects differences in the environment—the fact that there is a lot of Whole Foods stores in Malibu—or is it genetics. It does not have to be either/or.

Dr. Flier said there probably are some genetic elements, but probably more social and environmental elements. This situation illustrates the fuzzy border between freewill and determinism here. There are social factors that powerfully influence motivations to be lean and at the margins those operate on behavior. They are influenced by cultural norms and other factors.

Dr. Spiegel continued that if the data supports it and a case could be made that the access to fresh fruits and vegetables is a crucial determinant, then making them affordable in neighborhoods and communities where they are not, might be a very positive thing. In contrast to the tobacco situation, the food industry could be given positive incentives to ensure more nutritious, healthy food was available. Dr. Drewnoski showed a graph at a recent meeting depicting caloric density (energy density) versus cost. Items that are massively energy dense are also inexpensive, and that has been a success of the food industry. This is an area where policy and positive incentives could be effective.

Dr. Ballard-Barbash said that NCI has looked at economic costs in cancer and the variety of factors related to it and recognizes the need for similar research in the obesity area. NCI is therefore developing a program announcement for economic research in this area, with probably a number of institutes joining in this effort. Dr. Spiegel noted that it has been found that persons on food stamps simply cannot afford the 5-a-Day (now 9-a-day) program. Incentives might spur the food industry to change that.

With regard to the best practices issue, Dr. Spiegel said that NIDDK partially supports the National Weight Control Registry being managed by Dr. James Hill at the University of Colorado and Dr. Rena Wing at Brown Medical School. This is a registry of several thousands of individuals who have lost at least 10 percent body weight and maintained the weight loss for at least a year, with the average doing so for several years. What is intriguing is that these are persons who were obese (an initial requirement for participation), which may mean that genetically and within their environment, they are on the left side of the bell curve, and were not part of a group expected to lose and maintain weight loss, yet they have done so. What can be learned from these people? The registry is not rigorously built because of how people are accrued. How can the registry be leveraged? Can accrual be enhanced? Dr. Spiegel added that virtually none of these people are on a low-carb diet; however, he understands that there is not some inherent bias in the way cases are accrued that would select people who are on Atkins.

Dr. Denise Simons-Morton, Director, Clinical Applications and Prevention Program, Division of Epidemiology and Clinical Applications, NHLBI, added that a major factor for persons who are successful in losing weight and keeping it off is that they dramatically increase and maintain their physical activity levels at twice what the CDC recommends (i.e., twice the 30 minutes a day, 5 days a week, moderate intensity). The recommendation to do a higher level of physical activity than is recommended for overall health is being incorporated into guidelines for losing weight and maintaining the loss. The question is how to get people to do it or help them do it.

Dr. Ballard-Barbash stated that Dr. Simons-Morton had made a key point, and hopefully with better objective data from NHANES, there will be a clearer sense of what level of physical activity actually exists in the population, since there is a big problem with self-reported physical activity, particularly in terms of estimating actual levels.

back to top

Dr. Malozowski added that Dr. Flier had made a very strong statement that he fully agreed with, that we do not know exactly what foods to eat, in spite of diets probably being the earliest research in diabetes in the 1930s and 1940s. Unfortunately, there has not been much incentive to do research in that particular field. The issue of investment threshold in basic, clinical, and community research to move the field forward was raised in both Dr. Flier’s and Dr. Green’s presentations and needs to be carefully considered.

Dr. Gilman Grave, Chief, Endocrinology, Nutrition, and Growth Branch, Center for Research for Mothers and Children, National Institute of Child Health and Human Development, commented that exercise has to be a key element for those who maintained the weight loss, since it has been shown that people who lose weight actually lower their metabolism, so it takes even less food to keep them at a stable weight. He asked if this was for a finite period or does the change last for a lifetime.

Dr. Flier agreed that caloric restriction puts people in a starvation-type physiology. Not only do they typically get hungrier, but they become more efficient at burning energy. How long this lasts has not been studied. It would be expensive to do so for more than a matter of weeks.

Dr. Ballard-Barbash commented that there has been an extensive focus in obesity research on the fat cell and fat cell metabolism and somewhat less focus on the muscle cell and muscle cell metabolism, which must also play a large role. The genetic factors might not necessarily be the same as what is seen in obesity; there may be a whole different set of genes that govern that.

Dr. Flier answered her request for a summary on this research by saying that the muscle cell was a bit out of fashion at present. In earlier mouse knockout experiments, for example when they did a knockout of the insulin receptor for muscle, not all that much happened. The mouse did quite well. There is beginning to be an interest again because of the mitochondria research, a lot of which is focused on muscle, because muscle is quantitatively an important tissue for energy metabolism or oxidative metabolism. Studies in muscle have actually led to recent observations about a generally small, but highly reproducible, decline in many, many parts of the oxidative machinery. PGC-1 in muscle seems to be quite important in that regard. It is becoming clear that previously there was no knowledge of the degree to which tissues that were thought of as doing function “A” were actually also signaling to other tissues. The focus was on fat because there were very shocking models of lipoatrophy where it was not understood what was going on. Now this is becoming clear, and Dr. Flier believes the same thing is going to happen with muscle. For example, it is beginning to be shown that in the PPAR_ muscle-specific knockouts, muscle is sending interesting signals. What the signals are is not known yet, but they are going to the liver and other places. The capacity of muscle to be oxidative, to burn fat, and to even change the overall phenotype, from slow to fast twitch, is under genetic control, and it is potentially under pharmacologic control with the various PPARs, such as _, versus _, versus _. Dr. Flier added that a weakness in current understanding is knowledge about what the lipid-signaling molecules are that regulate the PPARs, that are so fundamental to cardiovascular and metabolic physiology. They are called fatty acid sensors, but, according to Dr. Flier, what the endogenous modulators are is not understood.

Current and Proposed Initiatives

NIDDK Obesity Initiatives
Philip F. Smith, PhD, Deputy Director, Division of Diabetes, Endocrinology, and Metabolic Diseases, NIDDK, Bethesda, Maryland

In 2003, Dr. Spiegel appointed Dr. Smith and Dr. Susan Yanovski, Director, Obesity and Eating Disorders Program, Division of Digestive Diseases and Nutrition, NIDDK, to coordinate the obesity research initiative efforts within the institute. Dr. Smith explained that to work across divisions represents a new paradigm within the institute and brings together Dr. Yanovski’s very different set of expertise and perspectives and Dr. Smith’s experience in basic science. In addition to the NIDDK obesity research initiatives that Dr. Smith would be discussing today, he referred the group to the consolidated website for the NIH obesity research initiative at http://www.obesityresearch.nih.gov. The website lists relevant open Requests for Applications (RFAs) and Program Announcements (PAs), notices of upcoming scientific meetings, and an archive of prior meetings. There is also information for the public, although that is not the primary purpose of the website.

For FY 2004 and 2005, NIDDK is offering a broad range of studies from the most basic and molecular to studies that are clinical and translational. In addition, there are several workshops planned to develop future initiatives. NIDDK’s initiatives, along with those of NHLBI and NCI, represent a broad, comprehensive approach to fulfill the vision of the NIH Obesity Task Force. Dr. Smith explained that although he will be presenting NIDDK-led initiatives, many are joint initiatives with other institutes, a practice that has been going on for many years, even prior to the formation of the NIH Obesity Task Force.

back to top

Dr. Smith stated that one of the fundamental problems in obesity research, from the basic science perspective, is that a lot is known about how to treat diabetes and obesity in the mouse, but less is known about doing so in humans. One focus over the next few years will be to learn more about the physiology and molecular basis of obesity and energy balance in general in humans. To do that, NIDDK is proposing a range of initiatives that focus on using the human as an animal model. One of these is an RFA for ancillary studies for NIDDK’s large clinical trials or clinical trial networks to study the mechanistic underpinnings of obesity in those particular trials or the mechanistic underpinnings of successful interventions in those trials. The five NIDDK initiative trials connected to this RFA are: Look AHEAD, the NASH Clinical Research Network, TODAY (which is a type 2 diabetes in adolescents trial), the follow-up of the Diabetes Prevention Program (DPPOS), and the Bariatric Surgery Clinical Research Consortium (BSCRC), which is a particularly promising model to understand the metabolic pathways involved in energy balance in humans. The National Institute of Aging (NIA) is a partner in this last initiative and one of their trials is participating. NIDDK has also provided funds for trials of pilot programs that might lead to ancillary studies. These will be on a smaller scale and designed to be very flexible and very rapid in terms of funding as opportunities arise within these trials.

Dr. Smith noted that Dr. Flier had alluded to the fact that there is lot to learn about diet composition and energy homeostasis and that there is a fair degree of controversy in the field about what we know and do not know about diet composition. NIDDK is soliciting studies in animals with well-defined diets under various exercise conditions and studies in humans under controlled dietary intake, not in terms of self-reports. The Oxygen-18 for Doubly Labeled Water for Research study will address energy expenditure and is an example of efforts to provide the scientific community with resources and reagents to do research.

To bring together a broad range of expertise in the behavioral, imaging, molecular, genetic, and physiology sciences to tackle the problem of energy balance systems within the body, Dr. Smith said NIDDK has issued a call for supplements to support the development of collaborative research teams. This RFA went out on the Internet the first week of April 2004 and quickly generated a lot of interest. Two major areas will be addressed by the RFA: (1) integrative approaches to energy balance and (2) proteomics, which will be necessary to identify biomarkers that might be good predictors of treatment efficacy or might provide clues to the linkage between obesity and comorbidities. There will be a follow-up, at least in the area of energy balance, with an RFA for consortia-type programs to bring resources in human capital together to examine energy balance in animal models, non-human primates, and primates. Another recent RFA involves the use of the zebrafish, C. Elegans, drosophilae, and the mouse as models to identify new pathways and potential new targets for drug discovery and for understanding the pathways themselves.

On the clinical side, one area of concern is the issue of translating short-term weight loss (for which we have many efficacious methods) to long-term weight maintenance. Studies recently have given us a clue that there are a number of hormones that drive us to regain weight. It is not just a matter of our metabolic set-point, but also hormones drive us to eat increased food. One investigator is looking at the possible role of leptin and leptin replacement after weight loss as a way of maintaining reduced weight and reducing the drive to eat, which shows some promise.

Following up on Dr. Flier’s presentation, Dr. Smith said that over the last 10 years, dozens of new molecules secreted by fat or at sites of fat-associated cells have been identified, and there are many more yet to be understood. One of the things that is not clear is whether these molecules differ depending on where the fat is. Clearly, fat depots are not all the same. Lipodystrophy is a good model that demonstrates the fact that people lose fat specifically in certain parts of the body and not in others. There is the notion that certain fat, such as visceral fat, may be associated more with comorbidity than other types of fat, but no one knows why that is. NIDDK is hoping to stimulate investigations to understand the key factors produced by fat at different depots, what causes fat to position in different depots, and what is the underlying basis and association with mortality.

Getting back to the human as a model organism, Dr. Smith spoke of an initiative this year on ancillary studies for clinical trials such as pilot and feasibility studies. To take advantage of the enormous opportunity offered by the breadth of the scientific community to partner basic and clinical researchers and really move discoveries in animals into humans, NIDDK issued an RFA with a receipt date in July 2004 to encourage investigations in the humanism research model.

For FY 2005, NIDDK will look at the potential effects of the maternal, neonatal environment on the development of energy balance pathways. There is a growing body of literature that suggests that gestational diabetes, for example, can affect the risk of the progeny to develop diabetes later in life. The same can be said for birthweight, both low and high, and whether an infant breastfeeds or is fed with formula. Dr. Smith asked “What is the template for those effects?” He also cited two papers co-authored by Dr. Flier in Science the week of April 4, 2004, that reflect the potential effect of such factors on the obesity epidemic. A group in Oregon demonstrated that in the leptin-deficient mouse there is a significant arborization of the connections between the sensing for leptin in the arcuate nucleus and a primary site of output, the PVH described by Dr. Flier. For example, in the wild-type animal across development (postnatal days 10 and 16 through 60), the arborization of connections can be seen from the arcuate nucleus, which contains the POMC and NPY cells, to the PVH, which contains the cells with the melanocortin 4 receptors. In the ob mouse, the development of those same pathways is heavily deficient. If the adult animals are treated, there is only a modest recovery of this connection; however, animals treated in days 10 to 16, during that critical period of development, can have the entire pattern of arborization and connections restored.

back to top

In the same issue of Science, Pinto et al. in Dr. Friedman’s lab have shown that the actual connections to the various cells in the arcuate nucleus (that is, the NPY, or food-intake driving neurons, or the POMC, or food-intake inhibition neurons) are modulated in the leptin-deficient mouse. The effects of glucose inputs are essentially reversed, going from high on the cells that drive inhibition of food to high on the cells that drive food intake. When the mouse is treated with leptin, this reverses in a very short period of time (i.e., 6 hours).

According to Dr. Smith, both of these pieces of data indicate that the energy balance system in the brain is plastic and can be affected during critical periods in development. What is not known is whether or not this is going to be a template for the kinds of effects that one would see in gestational diabetes in terms of the offspring. Clearly, this is going to require an initiative that does not just study mice. It gets at the issue of what happens in humans. Those are difficult studies to do but NIDDK is hoping to bridge into non-human primates to try to make that connection.

Dr. Smith described two groups of initiatives that are NIH-wide. One, which came from the NIH Obesity Task Force and will be led by NIDDK is the “Genetics and Genomics of Obesity.” This is a multiple-component initiative that goes from discovery in model organisms all the way up to testing of candidates in human populations and perhaps long-term prospective studies on individuals before and after development of obesity, not looking at the effects of being obese, but at the possible causes of obesity. The second one is the “Neurobiological Basis of Obesity,” which, depending on budget outcomes, will be a major trans-NIH effort supported by the Office of the Director for 2005. All of the neuroscience institutes will be keenly involved, as well many of the other institutes in the Obesity Task Force. It will focus, in the final analysis, on trying to understand the biological basis of human behavior or, in this case, human feeding behavior, a very difficult task. The approach will have to be broad and multidisciplined, requiring imaging studies in humans and studies in various animal models of behavior that relates to food intake and exercise. There will be broad participation of institutes and development of collaborative teams of investigators. NIH has a new mechanism that allows for recognition of contributions from multiple components of a team that will be implemented for the first time with this initiative. This recognition is expected to be a key component in bringing together collaborative teams.

Dr. Smith emphasized that the NIH Roadmap (http://nihroadmap.nih.gov) should not be ignored. It is a key piece of NIH’s strategy to move forward in terms of translation of basic findings to clinical intervention. There are three major areas within the Roadmap that NIDDK is heavily involved in and that might be particularly relevant to obesity, as well as other areas. One area is the “National Technology Centers for Networks and Pathways,” which will be hubs to develop proteomics technology. The long-range plan is that the hubs for technology and development would be surrounded by spokes that would be disease-relevant and supported by the institutes. NIDDK feels strongly that diabetes and obesity will be major areas within that. The NIDDK website (http://www.niddk.nih.gov) describes two current NIDDK initiatives that involve proteomics; these are not associated with the Roadmap, but will provide good partners with those technology centers.

In recognition that the bridging of disciplines is critical to moving science forward in terms of treating disease, the NIH Director, Dr. Zerhouni, has made the “Interdisciplinary Research Centers,” area the centerpiece of the Roadmap. A number of models are being tried, beginning with P20 consortium planning grants in FY 2004. This has received a very large number of applications, many of them focused on obesity. This initiative is actually unlike many of the Roadmap initiatives, since it is intended to focus on a particular disease or condition. Dr. Smith was therefore pleased that so many investigators selected obesity as a significant and complex biomedical problem. His initial review indicates that NIH has been successful in capturing what was wanted—very, very different approaches to the problem, from the built environment to behavior. NIDDK is very involved in this Roadmap initiative and the potential of its impact.

back to top

NIDDK is the lead institute for the third Roadmap area in Dr. Smith’s discussion, “Metabolomics of Steady-State, Site-Specific Tissues.” Dr. Maren Laughlin is the program contact. Dr. Smith noted that although metabolomics is probably a generation behind proteomics in terms of technology development, in the area of lipids and other signaling molecules, it will provide a key piece of the information needed to phenotype individuals to a greater degree. The focus will be on technology development with hubs supported by multiple projects, R01-based or otherwise, that would then use that technology in a particular disease. Information on these initiatives is available at the Obesity Task Force website.

NCI: Optimizing Energy Balance To Reduce the Cancer Burden
Rachel Ballard-Barbash, MD, MPH, Associate Director, Applied Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute (NCI), Bethesda, Maryland

Dr. Ballard-Barbash stated that for the first time in a bypass budget presented to Congress, for the 2005 budget the NCI Director, Dr. Andrew von Eschenbach, included a chapter on energy balance titled, “Optimizing Energy Balance To Reduce the Cancer Burden.” Surprisingly, despite probably 50 years of research in animal models and 30 or 40 years of research in humans, it is only recently that the lay public and health professionals have become aware that obesity influences cancer outcomes. Nutrition and diet have been a focus of research at NCI for many years, but only lately has there been a recognition of the fact that NCI has been funding obesity-related research, largely to understand how it influences cancer outcomes. More recently NCI has also funded research on the role of physical activity on cancer outcomes.

Dr. Ballard-Barbash explained that the term “energy balance” is used at NCI to refer to the intersects among diet, weight, and physical activity. Dr. von Eschenbach created a working group across NCI and called for the development of an energy balance initiative in July 2002, as had been done previously regarding tobacco. This group developed the statement that went into the bypass budget. The goals of this statement are three-fold.

• Understand the causes of adverse patterns of weight, physical activity, and diet in the population.

•Understand how these patterns contribute to cancer.

• Apply this knowledge to develop effective interventions for cancer prevention and control.

Four specific objectives were defined and then milestones developed under each of these objectives. Objective 1 is to discover how weight, physical activity, and diet, along with genetic and environmental factors, interact over a lifetime to influence carcinogenesis.

Dr. Green and others had discussed at this meeting the important issue that the working group formulated as Objective 2: Monitor trends and determinants of weight, physical activity, and diet and their cancer-related consequences by expanding nationwide research, particularly the surveillance infrastructure. Given that the institute has had the SEER (Surveillance, Epidemiology, and End Results) Registry for many years, Dr. Ballard-Barbash noted that NCI’s attention to population-level surveillance is much larger than that of other institutes. A large part of her program involves improving national and regional data on cancer control surveillance related to factors such as tobacco, diet, weight, screening, treatment, and cost of cancer.

Objective 3 is to develop improved measurement methods for weight and body composition, physical activity and fitness, and diet and bioactive food components, using self-reports and also advances in technology for objective reference measures. Finally, Objective 4 is to accelerate research on energy balance-related behaviors and develop interventions to improve cancer-related health outcomes, especially in high-risk populations.

Objective 1 milestones focus on issues of improved understanding of the mechanisms of obesity and carcinogenesis. The first is to discover and characterize mechanisms leading to cancer by initiating transdisciplinary research centers in the areas of energetics, physical activity, nutrition, and genetics. The Transdisciplinary Research on Energetics and Cancer (TREC) RFA, modeled after the successful transdisciplinary research on tobacco cessation, is expected to be released in June 2004. Another milestone is to improve collection of self-reported and objective measures on all of the energy balance factors within existing population studies to explore potential mechanisms by which these factors affect cancer outcomes. A third milestone involves advancing an understanding of cancer mechanisms by conducting studies in the area of energy balance within existing NCI clinical metabolic and nutrition research units. NCI has supported extensive development of animal models relevant to cancer research. One new area of exploration is examining how animal models that have been developed for the study of obesity might intersect with animal models that have been developed for the study of cancer. Additionally NCI will support basic and clinical research using proteomics and molecular technology as tools for exploring that area further.

back to top

Objective 2 milestones focus on improving surveillance activities. The objective has an extensive list of milestones, indicating NCI’s broad perspective on surveillance. Plans include expansion of nationwide surveys, most specifically within NHANES and also within NHIS (National Health Interview Survey). These efforts will improve self-report, biologic, and genetic measures within these systems that monitor health behaviors in major U.S. population groups. Because national surveys do not obtain sufficient data on smaller population subgroups, community surveillance through efforts such as the California Health Interview Survey (CHIS) are also supported. In addition, NCI supports a national survey on people’s health communication needs related to cancer that will include a component about people’s understanding of health recommendations in physical activity and nutrition. Other key aspects of health surveillance are obtaining representative data on healthcare providers’ knowledge, attitudes, and practices related to weight control and developing research resources on legislative policies related to nutrition, physical activity, and obesity. NCI is developing a PA on economic factors related to diet, physical activity, and energy balance in at-risk populations. Dr. Ballard-Barbash noted that the U.S. Department of Agriculture (USDA) has an economics PA largely focused on diet. NCI’s PA will broaden that focus to include physical activity and weight control. Finally, this objective highlights the need for training, at the national and international levels, to improve competency among future scientific leaders in this area, linking energy balance across the cancer continuum. This will include support of interdisciplinary or transdisciplinary training in basic sciences through population sciences.

Objective 3 milestones address research to improve measurement of the energy balance factors. This research includes reference biomarker research such as the OPEN (Observing Protein and Energy Nutrition) study to explore the extent of measurement error with existing self-report measures. Under this objective, NCI will expand validation research to include diet, physical activity, and fitness through the use of reference biomarkers and measures of physical fitness or activity within national and international cohort studies. NHLBI has been leading an effort promoting the development of innovative technologies, such as bioengineering and other measures to enhance accuracy of measurement; NCI has collaborated with NHLBI in this effort. Dr. Ballard-Barbash explained that one of the difficulties is that most methodologies for capturing these health behaviors have been developed in general populations. Another milestone focuses on the need to develop tools for diverse cultural populations. The last milestone under Objective 3 concerns developing better surrogate (intermediate) biomarkers as predictors of the effectiveness of diet and physical activity interventions.

Objective 4 milestones pertain to development of interventions. Given research suggesting that obesity may have an adverse effect on prognosis for breast and other cancers, NCI is supporting research on interventions that focus on weight control through diet and physical activity for cancer patients and for populations at high risk for cancer. Another milestone is to focus on the effect of sociocultural factors in the adoption of recommended behaviors and to develop approaches to improve interventions in specific populations. This milestone will examine lessons learned from the tobacco control program and the 5-a-Day program. The potential for social marketing research to enhance the effectiveness of communication is reflected in a milestone to support formative communication research in this area. The effect of food labeling to support recommendations also will be considered under this milestone. Lastly, through the TREC RFA mentioned previously, transdisciplinary research on energetics and cancer will be a focus, not just on mechanisms of how obesity influences cancer outcome, but also to identify effective population-level interventions, particularly with children and adults during critical periods of weight gain.

Dr. Ballard-Barbash concluded by noting that to-date, NCI and all of NIH has traditionally funded research that has examined micro-level factors—individual physiologic, behavioral, and genetic factors influencing a variety of health behaviors or organism behaviors and outcomes and a subsequent disease. Within NCI’s initiative on energy balance, examination of macro-level factors—contextual sociocultural, environmental, institutional, and policy factors—is also a key focus, as recommended by many of today’s speakers.

National Heart, Lung, and Blood Institute Initiatives
Denise Simons-Morton, MD, MPH, Director, Clinical Applications and Prevention Program, Division of Epidemiology and Clinical Applications, NHLBI, Bethesda, Maryland

Dr. Simons-Morton opened her presentation with a discussion on how NHLBI obesity research fits into conceptual models of translation to inform clinical and public health applications. The first phase is etiologic and determinants research, which is needed to identify potential risk factors, influences, and modifiers. This is followed by randomized controlled trials to determine the efficacy of risk factor changes, which then leads to clinical and community trials of intervention effectiveness. The distinction between phases two and three is that efficacy is basically asking what NIDDK often calls “proof of principle,” whereas effectiveness is what happens in real-life settings. Efficacy involves getting effects on health outcomes when implementing high-quality interventions, with high compliance, in an ideal population, in a research setting. Effectiveness of interventions is a function of efficacy and adherence and delivery in real-life settings. More and more the thinking is that another step is needed—dissemination and translation research, as it has been found that what has been learned is not being translated very well into public health or clinical practice. Now it is thought that we need to study approaches to achieve this final step. For example, given an effective intervention program, how can we get agencies and organizations around the country to adopt it? This means targeting not the patient population or the resident population, but the gatekeepers and organizational decisionmakers to implement such programs.

Next Dr. Simons-Morton described how the NHLBI portfolio addresses these four phases that inform clinical and public health applications. Under the etiologic and determinants research phase, there are observational studies in obesity, such as the National Growth and Health Study (NGHS) that focuses on the development of obesity in children; Framingham, which was the first study to identify obesity as an independent risk factor for cardiovascular disease and is examining additional questions today, including genetics and other mechanisms; and CARDIA (Coronary Artery Risk Development in Young Adults), an observational prospective study in young adults, 18- to 30-years-old at enrollment, that is examining obesity trends over time and relationships between diet, physical activity, and obesity on one hand and cardiovascular risk factors on the other.

back to top

NHLBI is funding a number of randomized controlled trials testing whether weight loss improves hypertension and hypercholesterolemia. The institute is working with NIDDK on the Look AHEAD (Action for Health in Diabetes) trial, which is an efficacy study testing whether intentional weight loss reduces cardiovascular disease events in people with diabetes. There are several effectiveness studies, including Pathways, which was a school-based study of American Indian schoolchildren to prevent adiposity; GEMS (Girls Health Enrichment Multi-Site Studies), which is a program of several studies testing interventions to prevent the development of obesity in African-American adolescent girls; PREMIER, a recently completed study on multiple lifestyle interventions, including weight loss in the overweight participants, to improve hypertension control; and WML (Weight Loss Maintenance), which is testing approaches to maintaining initial weight loss.

Related to the dissemination and translation area, NHLBI has funded several studies under the Environmental Interventions RFA, on which NIDDK took the lead. There is currently a Worksite Obesity RFA to find studies to test intervention approaches in worksites that include changing the environment in the worksite. Dr. Simons-Morton noted that to date studies have not tested organizational change approaches for getting interventions into real-life practice, which is a direction NHLBI would like to move toward. The closest study to doing this is the Catch On Study, a follow-up to the Child and Adolescent Trial of Cardiovascular Health (CATCH). Catch On examined the implementation of the CATCH school-based intervention program in new schools and asked what it takes to get this program implemented in schools. The answer was that it takes a “change agent” in the school.

Dr. Simons-Morton reviewed another conceptual model, the “natural history of disease and levels of prevention” model, which she approached from the opposite direction from that used by Dr. Green when he discussed this same model. NHLBI wants to do primordial prevention to prevent risk factor onset, primary prevention to reduce risk factors when present and thus prevent disease, and secondary prevention when disease is present to prevent adverse disease outcomes. There is no clear demarcation, necessarily, in the continuum, but it is a useful way to think of things and apply the model to obesity.

There are determinants of obesity, which influence diabetes, hypertension, and cholesterol, which in turn influence cardiovascular disease. Dr. Simons-Morton placed the NHLBI studies described above into the research-related phases of this natural history model. The childhood observational and intervention studies (NGHS, Pathways, and GEMS) provide information on the development of obesity and on interventions to prevent obesity development in the first place. Once obesity occurs, the focus of observational and intervention studies is to reduce obesity and prevent hypertension (PREMIER) and to identify relationships between obesity and other cardiovascular risk factors and between obesity and cardiovascular disease (Framingham and CARDIA). Look AHEAD is examining the effects on cardiovascular disease of reducing obesity. In addition to these studies, NHLBI also is conducting a variety of mechanistic and biological studies and has a number of new initiatives in the pipeline.

Dr. Simons-Morton stated that there are successful lifestyle interventions for weight loss. They consist of a number of common factors:

• They include one-on-one individual sessions with participants, often with group sessions interspersed, or maybe primarily group sessions with some individual sessions interspersed.

• They target both diet and physical activity.

• The focus is on calorie balance, not necessarily low-carb or low-fat, or the macronutrient composition, but what is considered a healthy diet, which is relatively low-fat and varied in composition.

• They use behavioral approaches. Knowledge is necessary, but not sufficient, so they use things like self-monitoring, feedback, and individualized problem solving.

• In general, the more intensive initial intervention phase is followed by a less intensive maintenance phase.

PREMIER was a three-arm, randomized trial testing the effects of multiple lifestyle interventions on blood pressure level in people with elevated blood pressure. There was an “advice-only” group and two groups—an established and established plus DASH—that had weight loss components, reduced dietary sodium, and increased physical activity. The established plus DASH group also taught people how to eat the DASH (Dietary Approaches to Stop Hypertension) diet, which is 8 to 10 servings of fruits and vegetables a day, overall low in fat, and 3 servings of low-fat dairy. The DASH diet was proven in an efficacy trial to substantially lower blood pressure. Average weight at baseline was 97 kg. At 6 months, the established and established plus DASH groups lost 5 to 6 kg on average, which was significantly different than the advice group’s average of 1 kg or less. At 18 months, there was still a large difference, but the established and established plus DASH group’s loss had crept up to an average of 4 to 5 kg, while the advice group lost from 1 to 2 kg.

Dr. Simons-Morton pointed out that the Diabetes Prevention Program results followed the same weight loss pattern as that of PREMIER. Six months is when the lifestyle arm had their major weight loss of 6 to 8 kg, compared to the metformin arm’s 2 kg and the placebo arm’s less than 1 kg loss. Then the lifestyle group’s weight gradually went up over a 4-year span to an average of a little less than 3 kg from their initial baseline. The same pattern was seen in the trials of hypertension prevention, which had a 3-year follow-up. This pattern has been seen in many studies. One could conclude that interventions only work for 6 months, but one could also conclude that interventions work as long as the intensive phase of the intervention is going on. Because what happened in all these studies is that when the study moved to the maintenance phase, people reverted somewhat back to their old behaviors. In NHLBI’s current Weight Loss Maintenance trial, once the participants achieve a certain amount of weight loss, they will be randomized into three different intervention approaches for maintenance to test what works best.

back to top

Based on these and other studies, Dr. Simons-Morton summarized what has been learned to date. We know from randomized control trials (RCTs) that obesity is a risk factor for hypertension, dyslipidemia, and other CVD risk factors. We know from randomized trials and epidemiologic evidence that lowering hypertension and dyslipidemia reduces cardiovascular disease. We know from RCTs that obesity also is a risk factor for diabetes, which is a risk factor for cardiovascular disease (CVD). The relationship between diabetes and cardiovascular disease is only known from epidemiologic evidence so far. NHLBI is sponsoring the ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial to test whether intensive control of blood glucose will reduce cardiovascular disease and provide RCT evidence.

Dr. Simons-Morton added that we also know that on the continuum of primordial to primary to secondary prevention, diet and physical activity are both important behaviors for energy balance, which influences obesity, and thus cardiovascular disease risk factors and cardiovascular disease. There is evidence for each piece of this causal pathway. We also know that educational and behavioral interventions can improve diet and physical activity, as was shown in PREMIER and DPP. Based on the evidence, NHLBI collaborated with NIH and external colleagues and developed the Obesity Education Initiative. NHLBI produced a Practical Guide that classified overweight and obesity into categories based on BMI ranges and disease risk relative to normal weight and waist circumference for men and women. The guide recommends clinical practices to address the problem. Information about the Obesity Education Initiative and guidelines for health professionals and for patients and the general public can be found at http://www.nhlbi.nih.gov.

Dr. Simons-Morton listed influences that affect the basic causal pathway that leads from diet and physical behaviors to energy imbalance to obesity to risk factors to CVD. There are intrapersonal factors (knowledge, attitudes, beliefs, and skills that influence an individual’s behavior) that have yet to be fully tapped in terms of interventions for diet and physical activity behaviors. There are environmental, community, and societal factors that affect the intrapersonal factors. These include the built environment and the availability of healthy foods, rather than the inexpensive high-fat, high-calorie foods, and places to be physically active. People are in a constant struggle against an environment that promotes a sedentary lifestyle and poor dietary intake, which we, as a human organism, have to fight. These things influence people’s knowledge and their attitudes, beliefs, and so forth, but they also directly influence behaviors. There are physiologic factors such as satiety levels and taste that affect diet and physical activity and result in energy imbalance. Finally, there are biological determinants and mechanisms such as genetics and the pathways discussed by Dr. Flier that directly affect obesity and the other risk factors for CVD. It is a complex system. NHLBI supports research in these areas and efforts to implement guidelines for what we already know.

Individuals at risk for being obese are influenced by their families, by organizations in which they play and work, by the communities in which they live, and by society’s norms and influences. Many people think that interventions to address the problem need to not only target individuals in terms of their health behaviors and adherence to recommendations, but, to be effective, interventions should also target organizations, communities, governments, and policies. Dr. Simons-Morton explained that the translational research she referred to earlier tests systems level and environmental level approaches to learn what works to influence organizations and communities to implement programs that have been found effective at the individual level.

Dr. Simons-Morton’s stated that NHLBI hosted a Think Tank on Enhancing Obesity Research at the NHLBI to consider future research directions. The Executive Summary published in January 2004 recommended (1) more basic biological research examining issues related to etiology and metabolic consequences of obesity, including research in genetics, adipose tissue biology, critical periods in obesity development, and etiologic and metabolic issues related to diet and physical activity; and (2) developing effective, practical prevention and treatment interventions, particularly related to translation into practice, including research on environmental and social determinants of diet and physical activity, influences of family environment, and interventions that could be applied in clinical practice and community settings. NHLBI is currently analyzing its portfolio with respect to specifics for future initiatives based on these recommendations.

In summary, Dr. Simons-Morton listed the following conclusions:

• Obesity is an important and proven causal risk factor for diabetes and CVD.

• Lifestyle interventions can reduce obesity, which can reduce diabetes and other CVD risk factors.

• Successful weight loss interventions include behavioral approaches for both diet and physical activity. They do not generally sustain the initial weight lost, however. Also, they are intensive, and thus of limited utility for real-world settings.

• Therefore, environmental changes and multi-level approaches are probably needed to enhance the delivery and effectiveness of weight loss programs that we know work and to promote obesity prevention.

back to top

Dr. Simons-Morton added that people in intervention programs should not have to struggle against an environment that is constantly hampering and undermining them and influencing them to do different things than what they are being taught to do. She offered two quotes. The first is known as Occam’s Razor or the law of parsimony as stated by William of Occam, a physicist who lived in the late thirteenth and early fourteenth centuries: “Entities must not be multiplied beyond what is necessary.” For example, it is simpler to describe all the planets revolving around the sun than it is the sun and all the planets revolving around the Earth. The Occam’s Razor of obesity is “Calories count, and if you want to prevent obesity, your calories in should not exceed your calories out, and if you want to reduce obesity, your calories out should exceed your calories in.” Unfortunately, things are not that simple. When it comes to behavior and society, H.L. Mencken may have provided a more accurate and relevant statement: “For every complex problem, there is a single solution that is simple, neat, and wrong.” Dr. Simons-Morton stressed “There is not a single solution to the complex problem of obesity.”

Discussion

Dr. Simons-Morton agreed with a comment from Dr. Green that although there is a relapse in lost weight for the most successful groups, the curve does not seem to ever go back to the level of lesser weight lost by the other groups. The results can be viewed as a glass half-full or a glass half-empty. The important point is that the programs do work. They just do not work as well as they could because of the environment that people live in. She added that a number of things are needed, including a comprehensive, multi-level approach; biological, mechanistic research to improve medications; individual level health promotion and behavioral skill-building; and environmental changes. Obesity is such a complex problem, we cannot say that one thing or the other is all that is required. It needs to be addressed on all fronts, including primordial, primary, and secondary prevention. One difficulty is that there are other competing needs for NIH monies.

Dr. Kelly Acton, Director, National Diabetes Programs, Indian Health Service (IHS), heartily agreed with the need for translational research. It is not that decisionmakers are necessarily resistant to moving the research to the community. IHS is trying very hard to translate the best research findings such as those from DPP. It is a question of “how to” implement these interventions in Indian communities. It is not easy to sort out what part is unique to a clinical trial and what part of the intervention is applicable to day-to-day community practice.

Dr. Simons-Morton agreed and added that is what NIH is trying to resolve through their translational research efforts. The dilemma is that there are questions to be answered at the same time as action is needed to stem the epidemic. One cannot wait for the research to come up with all of the solutions. As suggested by Dr. Spiegel, we must do what we think works now and at the same time try to obtain more information.

Brief Overview of Ongoing Activities

Veterans Adminstration (VA), Richard Harvey, PhD, Assistant Director, Preventative Behavior, VA National Center for Health Promotion/Disease Prevention, Durham, North Carolina

Following a brief introduction to the VA in general, Dr. Harvey presented information on the VA’s MOVE (Managing Overweight and Obesity for Veterans Everywhere) project, which was developed at the VA National Center for Health Promotion and Disease Prevention. The Veterans Health Administration currently treats more than 4.9 million patients a year. There are approximately 7 million enrollees in the system. The VA operates 162 medical centers and more than 1,300 community outpatient clinics. It is a very large system employing 184,000, including 15,000 MDs with an additional 25,000 affiliated MDs. The VA has strong academic affiliations throughout the country, and many of the Nation’s health professionals have trained in the VA system at one level or another.

The VA patient population is generally older (49 percent are over 65), more ill, and of a lower socioeconomic status than the general U.S. population. The largely male population is roughly 73 percent Caucasian, 15 percent African American, and 6 percent Hispanic. The female population is rapidly growing, however, with females under age 50 now making up 22 percent of outpatients. In the last 5 or 6 years, the VA has gone from largely inpatient care to outpatient care delivery systems and has been moving from the clinic into the community and into the home, using technology to achieve care within veterans’ homes.

Dr. Harvey explained that a major reason for developing the MOVE project was that limited data showed that 70 to 75 percent of veteran patients have a BMI greater than 25, BFRSS 2000 data indicated about 21 percent of those who use the VA fell into the BMI 30 or over obesity category, and the VA height and weight database collected from medical records indicated that actually 36 percent had a BMI over 30. With such a large group being overweight or obese, it was decided to develop a program for the entire VA system that would be based on the best evidence available to date, which included NIH clinical guidelines, NHLBI data and the Practical Guide developed by NHLBI for the Obesity Education Initiative, literature from the Agency for Healthcare Research and Quality’s U.S. Preventive Services Task Force (USPSTF), and other current literature.

There are currently 17 limited-scope clinical trials going on, primarily researching the feasibility of standard interventions in primary care settings. Additional full-scale trials of possibly several thousand patients at each site are being planned to prevent additional weight gain and to reduce the BMI is those with ratings over 30 back down to 25. MOVE is a comprehensive program with a public health population-based approach emphasizing lifetime rather than episodic care. Most of the research and most of the current clinical programs fall into the category of an episodic care kind of a program lasting just 12 weeks, 6 months, or whatever. One thing that behaviorists know is people do not do or keep doing uncomfortable and unpleasant things unless they have a good deal of ongoing support. As Dr. Simons-Morton mentioned, without ongoing support and fairly intensive kinds of work, there will be regression. Therefore, lifetime care was built into MOVE, since these patients are in the VA system for a lifetime, which is an advantage the VA has.

back to top

MOVE will be carried out in primary care or ambulatory care settings within the VA system without increasing staff or resources. To do this, some tasks in the primary care setting were reallocated. Dr. Harvey said the VA wanted to ensure that the program did not increase load on the physicians, who are seeing half again as many patients as they did in 1996, while the overall number of employees has decreased significantly. MOVE is designed, therefore, to be operated by non-specialized, multidisciplinary staff such as nurses, nursing assistants, dieticians, psychologists, and physical activity specialists, with just a bit of encouragement from the physician. Fully scripted patient-staff interactions were created to implement the program.

Another important feature, according to Dr. Harvey, is immediate enrollment and immediate action on a patient’s initial primary care visit. This is intended to avoid what happens with smoking cessation programs, when nearly half of the people do not show up. The program is a stepped care model, something that is imminently do-able, with an emphasis on health, not looks, again for veteran patients with a BMI of 25 or over.

At the trial sites, when a patient arrives for a routine primary care visit with their provider, their BMI is determined during the vital signs period, and those who are BMI 25 or over are advised of the scripted encounters and offered the opportunity to enroll in the MOVE program. Anyone who declines will be counseled and given a handout called, “So you’re not ready yet?” which is itself a minimal intervention, but an intervention nonetheless.

Dr. Harvey said that those who are interested complete an on-line initial assessment questionnaire, which outputs tailored reports for the patient and the staff. The patient’s report is similar to a wellness report from health-risk assessments, but is related specifically to weight and physical activity. The staff’s report details the major features of the patient for the staff member, lays out some red flags to attend to, and details some ways that the staff might be able to assist the patient with his/her efforts. In addition to the individualized profile, the patient receives a package of handouts on reduction in caloric intake, increase in physical activity, and behavior modification strategies. There are also a number of optional tailored handouts that relate to specific kinds of barriers or red flags that the patient has identified in the initial computerized assessment. These are not given out all at once because any given patient may qualify for 30 or so of them. They will be given over a period of time, as the patient is treated and is ready to deal with issue a, issue b, or issue c. Individuals who are able to walk and for whom walking may be recommended will be given a MOVE pedometer.

On this first visit, the patient receives brief counseling by the staff with regard to the handouts and the reports, agrees to one or two brief goals through a shared decision-making process, and is given a follow-up date, time, and method. The recommendations are for a follow-up a week later, and then every 2 weeks or so thereafter for as long as it takes to maintain the patient actively in the MOVE program. During follow-up visits, mostly done by telephone and about 5 minutes in length, progress is reviewed, barriers are addressed, information that the patient is ready to begin to deal with is sent to him or her, and reinforcement/encouragement is given. At some point, there will be a maintenance contact, maybe every 3 to 6 months, maybe more frequently, depending on the needs of the individual and other kinds of constraints. In any case, follow-up and support continue for the patient’s lifetime.

The above described process is the basic Level 1 part of the program. Dr. Harvey added that since many patients benefit from group sessions, the VA has scripted 50 or 60 different group sessions that can be effectively conducted by an average nurse or individual who is not usually familiar with weight control strategies. In addition to these weekly group sessions, handouts, and telephone support, individual consultations may be conducted as well, in what is loosely termed Level 2. This would be a consultation with a dietician, a psychologist, a physical activity specialist, or other specialist as indicated.

In accordance with data about weight control interventions, pharmacotherapy might be added to either of the two levels. The VA will make FDA-approved medications available to the patients who need special assistance. The VA currently has one brief intensive residential program and plans to establish standards for such a program in every region. Lastly, of course, bariatric surgical procedures would be included, as well, for individuals who may need them.

The phone counseling is primarily behaviorally based on stages of change and includes motivational interviewing and shared decision-making techniques to address identified barriers with each patient. Caloric reduction is, of course, specified, but starvation diets are not recommended. No specific diet is recommended, based on the understanding that essentially any diet will work if there is a net caloric reduction. Patients are given examples such as typical American Heart Association and/or low-carb kinds of diets. Local practitioners may outline what they wish their patients to use.

Patients, including those with various types of disabilities, are encouraged to do any kind of physical activity that increases their overall level of activity. Given the disabilities amongst the VA population, some limitations do apply. The activity should be do-able and sustainable. Behavior modification is, of course, a major part of this largely behaviorally based program. MOVE encourages gradual change, lots of stimulus control alterations, substitute behaviors, skill building, knowledge enhancement, and cognitive and behavioral changes.

back to top

Dr. Harvey emphasized that MOVE is more than a clinical program. MOVE has an executive advisory committee, of which NIDDK’s Dr. Susan Yanovski is a member. A VA steering committee is being formed to guide implementation of MOVE throughout the entire VA system. There will be a nationwide training initiative to train all primary care individuals in how to implement MOVE. A promotion campaign will address some of the environmental issues. Large posters and other items will be used to enhance awareness and motivation among staff and patients. Performance measures are being established, which, in Government, do drive change. A weight management directive will be issued, and VA/Department of Defense (DOD) clinical practice guidelines for the evaluation and treatment of overweight and obesity will be established. Currently, there are about 60 different weight management programs within the VA that vary in quality and resources. The guidelines will help standardize the program throughout the system.

The National Center for Health Promotion will continue to support the development and the evaluation of the MOVE program. The office is also developing an ongoing research agenda in obesity, physical activity, and weight control issues. One idea has been to develop a peer coaching initiative to enhance motivation and, in a sense, increase the workforce for the MOVE program.

In summary, Dr. Harvey said that through MOVE, the VA will effectively address the obesity epidemic with a robust national program, exposing every overweight VA patient to some intervention, even if it is just a handout and a very brief counseling. It is expected that veterans’ health status will improve as a result of this program, that their quality of life will improve, and that there will be long-term cost savings for the VA, given that these are long-term patients. When fully implemented, MOVE will become the largest weight management program associated with a national healthcare system in the country. The VA hopes MOVE will serve as a model for other national healthcare systems, such as that of DOD.

Discussion

Dr. Simons-Morton asked what the VA’s plans were for evaluation of MOVE. She noted that there is very little in the research literature about whether this kind of program implemented in the healthcare setting makes a difference, in terms of, for example, BMI. Was the VA planning on having some comparison non-intervention sites, or was that not possible within the VA system? It would be interesting if sites could be randomized and have BMI outcomes measured.

Dr. Harvey responded that they would be doing some feasibility and some very brief outcome work with the 17 pilot trial sites. At a much larger trial to be implemented in several VA facilities in Florida, they will be doing some evaluation research, including outcome research. The plan is to collect similar data with sites that are not participating in the MOVE program. At present, there is no imperative that sites must implement the program, or to what degree it must be implemented. Dr. Harvey suggested that he discuss programs, and possible funding opportunities, with NHLBI.

Dr. Steven Yevich, Director, VA Center for Health Promotion/Disease Prevention, interjected that the evaluation is an important component of the MOVE program, which is constantly going to be evaluated and changed as needed. The program will be tailored so as to be culturally sensitive, because the VA is in 50 states plus some territories. For example, Native American veterans in the southwest region will receive different counseling and support than veterans in inner-city New York.

Food and Drug Administration (FDA), David W.K. Acheson, MD, Chief Medical Officer and Director of Food Safety and Security, Center for Food Safety and Applied Nutrition, College Park, Maryland

Dr. Acheson summarized the initiatives of the FDA’s Obesity Working Group (OWG), which was formed in August 2003 by the Commissioner and charged to develop a clear, coherent, and effective public health message, outline a public health program, and see what could be done to enhance the food label. They were also to have a dialogue with the restaurant industry, facilitate development of more or better therapeutics, identify research gaps such as consumer behavior, and enlist the help of stakeholders to achieve their goals. The group was given a timeline of 6 months to prepare a set of recommendations to address these issues.

The FDA OWG held a series of meetings to engage stakeholders, beginning in October 2003 with a public meeting, and then in November with a workshop to explore the links between the food label and weight management. In December, they held a Health Professional Roundtable and a Consumer Roundtable. Public Docket No. 2003N-0338 was then submitted and received more than 100 comments.

Dr. Acheson said that the OWG’s conclusions from all this input basically reflected statements heard at today’s meeting—that obesity affects all segments of society. He emphasized that, whereas many things discussed at this DMICC meeting are outside of FDA’s mission, the report of the working group addressed the issue from the perspective of FDA’s mission; thus, there was a strong focus on labeling. The final report was released in March 2004 and includes both short- and long-range recommendations that are centered on the scientific fact that weight control is primarily a function of caloric balance. The full report can be found on the FDA website at http://www.fda.gov/oc/initiatives/obesity. For his summary, Dr. Acheson organized the recommendations into themes from individual working groups that independently studied the issues and then interacted with the overall team two or three times a week. These individual groups focused on labeling, enforcement, education, restaurants/industry, therapeutics, and research. Dr. Acheson then highlighted key elements in each category and referred the audience to the program handouts for further details.

back to top

The food labeling group’s recommendations included publishing an advance notice of proposed rulemaking to seek comment on how to give more prominence to calories on the food label. Suggestions were simple items such as increasing the font size, including a percent daily value column for total calories, and eliminating the listing for calories from fat. A lot of these recommendations came out of focus group testing, in terms of what people actually understand and what is helpful. Dr. Acheson said that one of the things learned about the American public through limited focus group testing was to forget math. The label must be very clear, plain, and simple. One difficulty in improving the label is that there is not a lot of space on it. A second recommendation was to seek comments authorizing health claims on certain foods that FDA would deem appropriate. An example of a health claim might be “diets low in calories may reduce the risk of obesity, which is associated with type 2 diabetes, heart disease, and certain cancers.” Among other recommendations, was one to get comments on additional columns to list quantitative amounts and percent daily values of an entire package. This was in response to the issue that, for example, a 20 oz. soda is shown as two servings, but clearly almost nobody consumes half the soda, puts it in the refrigerator, and then goes back and drinks the other half. The same is true of muffins and cookies and so forth. Although some of these would be problematic, there is a need to be more realistic of what is a serving and show the appropriate caloric content of the entire package.

Further labeling recommendations included comments on the amounts of foods that are customarily consumed and how that needs to be updated. There were several issues around carbohydrates, such as defining “low carbohydrate,” “reduced carbohydrate,” “carbohydrate free.” The labeling working group wanted to encourage manufacturers to use dietary guidance statements. An example would be, “To manage your weight, balance the calories you eat with your physical activity,” coming back again to the issue of it is input/output that is important. They also wanted to encourage manufacturers to take advantage of the flexibility of the current regulations on serving sizes, so as to label a single serving size with sodas.

Part of the enforcement working group’s goal was to make sure that the nutrition fact panel (NFP) is accurate because the panel’s accuracy is critical for consumers to monitor their intake in calories, particularly regarding serving sizes. The three recommendations of the group therefore included enforcement activities against those manufacturers who are inaccurately declaring serving sizes, highlighting in the Food Labeling Compliance Program enforcements against inaccurate declarations of serving sizes, and working with the Federal Trade Commission to target dietary supplement producers who offer false or misleading weight loss claims.

Dr. Acheson stated that just as the VA is planning a major education initiative through the MOVE program, the FDA also considers education an important part in influencing behavior. The education working group recommended that information on healthy eating choices support the bottom line FDA OWG message that “Calories Count.” Specifically, the education group recommended establishing relationships with the private and public sectors to give consumers a better understanding of the food label. Since many people do read the food label, it is probably going to help to make it clearer and make it more specific to help consumers make healthier and wiser food choices. As part of this, FDA wants to pursue relationships and partnerships with youth-oriented organizations, such as the Girl Scouts and the 4H Program, to emphasize early on the importance of caloric balance and proper diet for weight management.

In considering goals for the restaurant and food industry, Dr. Acheson pointed out that American consumers are spending approximately 46 percent of their food budget on food consumed outside the home. Obviously, it is critical to address not just what should be put on the packages that people consume in their own homes, but to identify for them what they eat outside, particularly in quick-service restaurants. Recommendations of this working group were (1) to urge the restaurant industry to launch a nationwide, voluntary, point-of-sale nutrition information campaign for consumers and (2) to encourage consumers to request nutrition information when they eat out. It was suggested that a series of options be developed for providing standardized, simple, understandable nutritional information, including caloric information at the point-of-sale. One of the strategies tested in focus groups was, when one goes to a fast-food service restaurant, which would be most helpful—to have caloric intake posted next to the menu items or just high-calorie, medium-calorie, or low-calorie indicated? Generally, the consensus was that consumers at a fast-food restaurant want to know if the item they are about to eat contains 1,500 calories. Still, it comes down to educating consumers about the concept of caloric balance, which is obviously a critical element.

The therapeutic group approached the subject in a variety of ways, but recognized that there is a subpopulation of the obese and extremely obese who require medical intervention to reduce weight and to mitigate the associated diseases. Their recommendations were to pursue things further, to convene a meeting of standing FDA advisory committees to address these challenges, and to fill in some of the knowledge gaps about existing drug therapies for obesity, similar to much of what was heard earlier at this meeting. They recommended continuing discussions with the pharmaceutical and medical device sponsors about new products and revision and reissuing for comments of the 1996 draft “Guidance for the Clinical Evaluation of Weight-Control Drugs.”

One of the mandates for the FDA OWG was to identify applied and basic research needs that include the development of healthier foods and a better understanding of consumer behavior and motivation. In this regard, the research working group recommended supporting and collaborating, as appropriate, with other obesity-related research groups, including NIH and the USDA. One suggestion was to collaborate with USDA on their national obesity prevention conference to be held in October 2004. The group also recommended the following five areas of obesity research: (1) information to facilitate consumers’ weight management decisions; (2) the relationship between overweight/obesity and food consumption patterns; (3) incentives for product reformulation; (4) the potential for FDA-regulated products unintentionally to contribute to or result in obesity; and (5) the extension of basic research findings to the regulatory environment.

back to top

The overall result of the FDA Obesity Working Group’s efforts was made clear in FDA Deputy Commissioner, now Acting Commissioner, Dr. Lester M. Crawford’s statement: “We’re going back to basics, designing a comprehensive effort to attack obesity through an aggressive, science-based, consumer-friendly program with the simple message that ‘Calories Count’.”

Centers for Disease Control and Prevention (CDC), Rodolfo Valdez, PhD, MSC, Epidemiologist, Division of Diabetes Translation, Atlanta, Georgia

Dr. Valdez stated that after 40 years of tracking public health epidemics, CDC agrees that it is time to move beyond surveillance and to establish programs to more effectively bring research findings to the community. CDC is sponsoring several such projects. The Division of Nutrition and Physical Activity has developed a comprehensive State-based program to help States maximize their efforts to prevent obesity by improving nutrition and physical activity. A PA published in the Federal Register in January 2003 resulted in 58 applications, of which 20 were selected for funding in FY 2003, 17 at the capacity-building level and 3 at the basic implementation level. These 20 programs have as their objectives to provide the population with the knowledge, skills, and motivation to modify their environments and provide opportunities for access to healthy eating and more physical activity. The interventions foster healthier behavior by mobilizing multiple levels of the social structure to achieve a balance between individual and environmental approaches for healthier lifestyles. CDC has provided a “Resource Guide for Nutrition and Physical Activity Interventions To Prevent Obesity and Other Chronic Diseases.” Topics in the Guide include caloric intake and expenditure; increased physical activity; improved nutrition, including breastfeeding and increased consumption of fruits and vegetables; and reduced television time.

A second program briefly described by Dr. Valdez was the Steps to a HealthierUS initiative proposed by Secretary Tommy Thompson of the U.S. Department of Health and Human Services (DHHS). The goal of this program is to help Americans live longer, better, and healthier. A centerpiece of the program is a 5-year cooperative agreement that provides States, cities, and Tribal entities with funds to implement chronic disease prevention efforts. The focus is on reduction of diabetes, overweight and obesity, and asthma through interventions addressing three related risk factors: physical inactivity, poor nutrition, and tobacco use. In FY 2003, approximately $14 million was distributed to fund applicants representing 15 small cities or rural communities, a Tribal consortium, and 7 large cities. These 23 communities will implement action plans to reduce health disparities and promote quality health care and prevention services in the States of Arizona, California, Colorado, Florida, Louisiana, Massachusetts, New York, Pennsylvania, Texas, and Washington, and the Intertribal Council of Michigan.

In December 2000, Congress mandated that CDC develop a media campaign to improve the health of our Nation’s youth. The resulting “VERB: It’s what you do” was launched in October 2002 as a 5-year strategic effort to promote physical activity through research, the media, partnerships, and community activities. It uses communications designed by the best youth advertisers and marketers and involves the teens themselves at all stages of planning. Dr. Valdez emphasized that the campaign is “by kids and for kids.” The campaign employs television, the Internet, and print and radio ads to reach youth 9- to 13-years-old, as well as their parents and other adults who influence youth. Messages are informational and motivational and customized to appeal to diverse populations. A recent evaluation showed that the campaign is receiving 70 percent recognition among children, so it has been very successful. Dr. Valdez stated that partnerships are critical to the success of the campaign. Through VERB, CDC’s Division of Adolescent and School Health (DASH) provides lessons and funds to 42 states, 4 territories, 15 local agencies, and 8 national organizations to initiate or expand in-school, after-school, and community programs to increase the availability and quality of physical activities for youth and reinforce the messages of the campaign.

Dr. Valdez’s group, the Division of Diabetes Translation, is studying how to implement the results of the DPP. Currently, they are asking States to assess and evaluate what programs they have now to prevent obesity and diabetes. Some pilot studies are being conducted in some States to determine just what type of efforts are needed. The Division also co-sponsors with NIDDK the National Diabetes Education Program and its “Small Steps” campaign. CDC is collaborating with the Look AHEAD project and also with environmental modifications to prevent obesity. A lot of effort is being put into screening to identify the people at risk for diabetes, and that, of course, includes identifying those who are overweight or obese. One part of the screening effort is to identify persons who have diabetes and do not know it; the other part is to identify people termed pre-diabetic, which is an important target population for implementing lifestyle changes to help them reduce weight and other risk factors and prevent the development of diabetes.

Discussion

Dr. Fradkin asked if there is any additional evaluation of VERB beyond recognition of the campaign, such as information about attitudes or behavior. Dr. Valdez answered that there is and a report has been prepared on that. In response to a question from Dr. Malozowski regarding the pilot programs, Dr. Valdez said that CDC is receiving many calls from States asking what to do. They know that a healthy diet and increased physical activity have been shown to help prevent diabetes, but they want something specific to do to implement these findings. CDC has had to hold them back a little and say, “It’s not that easy. We have to do something more.” For instance, they are told the DPP shows that to prevent one case of diabetes in a high-risk population, one has to do intervention on seven people, so that is a massive intervention effort in a community. Currently, CDC is asking the States to first assess what resources they have in order to do prevention. For example, do they have a Department of Chronic Diseases? CDC has DPCs (Diabetes Prevention Centers) in all the States and territories, but usually they are quite small, and need to evaluate the research that they have. This assessment phase is primarily what is being done now.

back to top

U.S. Department of Agriculture (USDA), Joseph T. Spence, PhD, Acting Associate Deputy Administrator, Agricultural Research Service, Beltsville, Maryland

Dr. Spence explained that USDA is organized into mission areas such as the Research, Education, and Economics (REE) mission area and the Food, Nutrition, and Consumer Service mission area, which administers the Food Assistance Program. People tend to think about farms when they think about USDA, but it is largely a food assistance organization. Unlike NIH, the extramural and intramural research entities are separate. The Cooperative States Research, Education, and Economics Service (CSREES) is the extramural arm and the Agricultural Research Service (ARS) is the intramural arm. The bulk of research at USDA is being done by ARS. The Economics Research Service (ERS) has become very interested in obesity research and risk analysis, including research on determinants of why people make the food choices that they do based on economic issues. ERS also evaluates the food assistance programs in terms of effectiveness.

The ARS Human Nutrition Research program is basically divided among six Human Nutrition Research Centers, which Dr. Spence would be describing individually. In addition, for approximately 10 years, ARS has funded the Lower Mississippi Delta/Nutrition Intervention Research Initiative. This is a partnership of seven institutions throughout Louisiana, Arkansas, and Mississippi, dealing with the rural poor in one of the most economically disadvantaged parts of the United States and developing interventions to improve individuals’ health status and diet. Fruits and vegetables are usually unavailable unless they grow these crops themselves. No amount of fat is ever wasted; it is consumed. Fast food to them is truck stops. This is a very important project for ARS, and they are beginning to see some successful interventions.

The overall mission of the ARS Nutrition Program is to define “what is a healthy diet?” Approximately 60 percent of information in the dietary guidelines process is generated by ARS; therefore, ARS is very active in doing research on dietary components. The agency has consistently had a foods-based approach to the study of nutrition. Even though centers may be doing molecular biology, the concentration is on food. Each of the centers has metabolic facilities to do between 8 and 20 in-house, long-term feeding studies, including clinical beds for in-patient studies and the ability to do large out-patient studies. The Beltsville Center in Maryland can feed about 120 people per day indefinitely. For studies, ARS typically provides all of the meals for each individual while in the study. They receive breakfast, a packed lunch, and come back for dinner, including all their weekend meals. The mission of each of the centers involves nutritional needs over the entire lifecycle, from childhood through old age. Beltsville traces its ancestry back to Dr. Wilbur Atwater and has a longstanding interest in calorimetry and body composition.

Dr. Spence stated that ARS and USDA stayed away from direct research in obesity until recently, partly because NIH’s investment in this area is so significant that they felt they would be small players. However, the compelling nature of the problem demands that USDA get involved. Another reason for their involvement is that, as the Nation’s department of food, USDA has a lot to offer regarding a food-based approach toward studying obesity.

Each of the ARS research centers has its own mission, which is given in the program handouts. The Arkansas Children’s Nutrition Center in Little Rock is relatively new. It works with women and children, emphasizing the effects of dietary factors on the prevention of atherosclerosis. This center is particularly interested in how diet affects cognitive development and immune function.

The Beltsville Center is the largest, most comprehensive, and oldest of the centers. Its mission is involved with diversity in the American population, and it has several studies intimately linked with obesity, particularly the influence of physical activity on long-term food intake and body weight and the beneficial effects of dietary fibers. As discussed today, it is very difficult for people to lose weight and keep it off. Rather than have participants go on a yo-yo dieting pattern, ARS is looking at diets, foods, and dietary consumption patterns that may not result in weight loss, but may be inherently healthier. Some of the work that the center has been doing on various plant fibers has borne this out.

The Beltsville Center is intimately involved in nutrition monitoring. The agency is working with the National Center for Health Statistics and is responsible for the dietary component of NHANES, including improving its accuracy. Because the system is not inherently accurate, Dr. Spence strongly disapproves of how people often use the data such as reporting values for calcium intake down to the nearest tenth of a milligram when the data may be only 10 to 20 percent accurate. The National Nutrient Databank is maintained at Beltsville as well. In community-based interventions, including ones related to diabetes, the center has worked with such supplements as chromium picolinate and some naturally occurring components in spices, particularly in cinnamon, which has been investigated by Dr. Richard Anderson.

The Children’s Nutrition Research Center is part of an ARS cooperative agreement with the Baylor College of Medicine, Department of Pediatrics. This center is studying obesity, particularly related to children, looking at genetic and environmental factors contributing to childhood obesity, biological influences on children’s diets, eating patterns, athletic self-concept and behavior, after-school physical activity, dietary intervention for children and families, and infant feeding patterns. They are examining how feeding patterns early in life affect children’s susceptibility to chronic diseases, including diabetes and atherosclerosis later on. Dr. Thomas Baranowski, one of the scientists there, has some innovative studies using computer games for children to pick up healthy eating habits. Since children do not prepare their own foods, a lot of work is being done with families, especially prevention work.

back to top

The Grand Forks Nutrition Center is USDA’s trace element center. Although they are currently doing little in the obesity area, they have an important role to play. As people are encouraged to consume reduced calorie diets, Dr. Spence emphasized that it is important to ensure that those diets are adequate in their content of the traditional known nutrients. Because of its location, the Grand Forks center is becoming more and more involved in Native American nutrition and Native American health. They have a mobile van that they take out to the reservations, the Northern Plains Indian Reservations, and they are becoming more and more interested in community nutrition.

The Human Nutrition Research Center on Aging, the Jean Mayer Center at Tufts University, is also a cooperative agreement. This group is conducting a number of studies related to chronic disease, obesity, and diabetes and cardiovascular disease. The major focus, in terms of obesity, is the obesity that occurs at midlife. Dr. Spence noted that people lead a healthy lifestyle and then something happens; growth hormone stops being produced and there is a propensity to gain weight. The center is studying lipoproteins and nutrition in aging (particularly Dr. Alice Lichtenstein and Dr. Ernest Schaefer), aging adipocyte and systemic metabolism, body composition and nutritional assessment in the elderly (Dr. Susan Roberts), nutritional genomics, epidemiology applied to problems of aging and nutrition, and improving cardiovascular health with diet. They are also participating in the Geisinger rural aging study, sponsored by the Geisinger Clinic in Pennsylvania. This center is very interested in the determination of energy and insulin regulation and in body composition as are some of the other centers.

The Western Human Nutrition Research Center at the University of California–Davis campus moved from the Presidio about 3 years ago. They are in the process of building a new facility. Their mission is to study nutritional interventions and the methodology associated with them. Dr. Lindsay Allen is the new center director. Dr. Allen has been involved in international nutrition and is very enthusiastic about the possibility of being involved in some nutritional interventions in the United States. The center has major ongoing projects looking at internal and external factors affecting food intake and body weight such as energy restriction, mineral homeostatis, and functional outcomes. Again, the emphasis is on maintaining a healthy intake of essential nutrients while encouraging people to lose weight.

The Nutrition Research Center Directors have been encouraged by the Secretary of Agriculture to come up with a new initiative in obesity research. They are very cognizant of the work being done at NIH and other Federal agencies. Dr. Spence remarked that it had been very helpful to hear the discussions today. He would be sharing with his colleagues at USDA what the various groups were doing and suggesting they find ways to collaborate.

Dr. Spence said that USDA’s broad outline of a plan emphasizes obesity prevention, rather than treatment of obesity. The agency is interested in helping populations at risk for gaining weight by developing effective processes and interventions. USDA is committed to a foods-based approach. They will work with plant producers and animal breeders to develop new foods, new varieties of foods, and incorporate these into diets, and then hopefully get the food producers and the food industry to adopt some of these. In fact, some of them are on the market now and are very successful. USDA will conduct research that provides the scientific basis for sound food assistance programs, which is highly important to the Department of Agriculture. USDA has been accused by some people of responsibility for the obesity epidemic because some foods are too cheap or because the agency is thought to be pushing the wrong kinds of food and so forth. It is true that one cannot produce inexpensive food that is high in calories and then be suddenly surprised that people are gaining weight. Of course, as has been said repeatedly at this meeting and elsewhere, this is a multifaceted problem, but USDA does want to have its food assistance programs based on the best available scientific evidence.

At the Secretary’s request, the Nutrition Research Center Directors put together a draft proposal that Dr. Spence outlined for the audience. The Directors recommended that there be a long-term, longitudinal study, similar to Framingham. Phase I would establish community cohorts to determine risk factors in vulnerable population groups; determine behavioral and lifestyle influences on food choices, meal patterns, purchasing, family environments, and so on; determine effects of participation in food assistance programs on choices and behaviors; and establish and refine methods to assess dietary practices and quality. Basically in Phase I, the objective is to understand why people make the food choices that they do.

Phase II is a longitudinal follow-up study of Phase I participants, who will be children, to determine how factors identified in Phase I predict weight change and body composition over time. Phase III will test intervention strategies over time. From the work USDA has done, it is known, for instance, that children can be brought into an environment where they become very interested in nutrition and, as a result, they will change their food-eating habits. Whether that continues from, say, preschool, elementary school, into high school and beyond, is another question to be answered. This phase will test various intervention strategies and follow the participants long-term to learn what works and what does not work.

Dr. Spence noted that there are similarities in what USDA is proposing and in what NIDDK and NHLBI are doing and proposing. He said he would welcome the opportunity to sit down and talk about long-term plans. The USDA can offer expertise in the food area that could be helpful to several other Federal agencies in their obesity reduction and prevention initiatives. Dr. Spence stressed that it is important to have a dialogue, because if there were ever a problem that requires thinking outside the box, obesity is that problem. Strategies need to be developed to come up with something new and fresh and make an impact on this epidemic.

Discussion

Dr. Fradkin asked Dr. Spence if ARS collects data on a State-by-State basis in terms of food purchases that could be analyzed to evaluate the impact, for example, of a CDC State-based initiative or a public health message about healthier food choices. He responded that they do collect that kind of data based on disappearance of products from shelves, which does not necessarily reflect what people are consuming. The disappearance data includes fast-food venues, but again, not what is being consumed, just what leaves the outlet. The Ag Marketing Service, which is another agency, keeps their own set of data on what is being purchased, what is being consumed, or what is being put into products on the periphery. It tracks the raw commodities. Within the data collected in NHANES, everything is broken down to raw commodities, so if somebody said they had a slice of pepperoni pizza, one can see how much wheat they consumed, how much tomato, how much meat, and so forth. The raw commodities data is compared with the disappearance data, which could provide some reasonable estimates of a change in people’s food choices, and although it would not be entirely accurate, it would be worth examining.

back to top

Dr. Malozowski commented that in the area of the main supermarkets, one knows what is being purchased. He asked if such a system is in place in closed communities as a double check. Dr. Spence answered that in Delta communities or rural communities, there is not the tracking that is available in a place like Washington. For example, in Arkansas, the disappearance data provides some data that you would not normally get because it collects how many dairy products went off of the shelves back to the warehouse, not at the individual grocery store. In rural communities, this is very difficult. They do not have the infrastructure in place.

Dr. Malozowski asked Dr. Spence if he was aware of the National Children’s Study managed by the National Institute of Child Health and Human Development (NICHD) as this could be a good match with USDA’s Phase III plans regarding following up children’s nutrition and development. Dr. Spence replied that they had learned about the project somewhat after the fact, but that Dr. Dennis Bier, the Director at Houston, has been working with NICHD’s Dr. Gilman Grave to become involved in the project.

Dr. Spiegel thanked all those present for coming to the meeting and expressed his appreciation to the speakers. He deemed this had been a very helpful session. Dr. Malozowski thanked Drs. Susan Yanovski, Brian Hoover, and Philip Smith for their assistance in organizing the meeting.

The meeting was adjourned at 4:05 p.m.

back to top


Diabetes Mellitus Interagency Coordinating Committee Meeting: Approaches Integrating Epidemiological Data on Diabetes

Rockledge 2, Room 9100-91
Bethesda, MD
June 24, 2004

Opening Remarks

Dr. Saul Malozowski, Executive Secretary, Diabetes Mellitus Interagency Coordinating Committee (DMICC), welcomed the Committee members and guests and said that this would be the final meeting of the fiscal year 2004.

Dr. Allen Spiegel, Director, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), said that the meeting’s theme would be “Approaches Integrating Epidemiological Data on Diabetes.” Although the National Institutes of Health budget will be limited during the next two years, it is important to maintain momentum in addressing the diabetes epidemic. Therefore, diabetes research resources must be leveraged by studying existing cohorts, tapping into ongoing studies, and gleaning data from completed studies that were note designed to focus specifically on diabetes.

Assembling a Raw Data Meta-Analytic Database: The Joys and Sorrows
Sue Duval, PhD, CODA Study Group, University of Minnesota

Dr. Duval discussed the use of meta-analyses and the Collaborative Study of Obesity and Diabetes in Adults (CODA), which is led by investigators at the University of Minnesota. She began by explaining that meta-analyses can prolong the lives of individual studies and seek to enhance inferences of individual studies. With individual studies, observed findings are subject to random variation that could lead to incorrect inferences. By combining data from individual studies, statistical power can be increased and consistency across studies can be assessed.
The aim of the CODA project is to use meta-analysis of individual participant data (MIPD) methods to address the following diabetes epidemiology questions:

• What simple anthropometric indices most closely predict the risk of type 2 diabetes mellitus in adults?

• Do ethnicity and other factors modify that prediction?

• Is the association of these anthropometric indices with cardiovascular disease morbidity and mortality exacerbated by their association with type 2 diabetes mellitus?

• Is it possible to predict several diabetes-related risk states using noninvasive or minimally invasive methods?

• Should screening tools differ by ethnicity?

Dr. Duval further explained that many people think of a meta-analysis as a systematic review. However, the term meta-analysis actually refers to the quantitative component, in which summary data abstracted from the published literature are synthesized to produce a powerful summary estimate that incorporates many studies. Meta-analysis of the published literature (MAL or MPL) involves an exhaustive exploration, critical evaluation, and quantitative synthesis of all unbiased evidence from published reports. In contrast, MIPD involves pooled analysis of individual, original epidemiological data, not just combining data that are available in the literature. MIPD can be conducted retrospectively or prospectively. Prospective studies can be carefully designed to help ensure that data collection is consistent across studies.

back to top

Studies using the MIPD approach are beginning to appear in the literature (e.g., meta-analyses of cancer trials) and can be considered the “gold standard” of systematic reviews. The key to success in conducting a MIPD is good international collaboration and communication with researchers who have collected the data. This type of review also requires central collection, checking, and analysis of individual participant studies, and an attempt to include all relevant studies.

In contrast to MAL, MIPD involves many researchers who become very invested in the project, and international networks of collaborating investigators develop. Other advantages of MIPD include that: more data are available; data are standardized; and it is possible to do better time-to-event analyses, to produce better adjusted/multivariate models, and to evaluate subgroup effects. In addition, heterogeneity and sampling bias in specific studies can be assessed. Disadvantages of MIPD include that: data may not be available from all published studies; there is potential for conflicts with collaborators regarding findings that are different from those published by the investigators; and substantial financial resources and infrastructure are required to get the investigators on board and to gather the data.

Dr. Duval then discussed the CODA project MIPD experience. She stated that having a large dataset allows the researchers to study broad questions with great precision. For example, questions can be asked about waist circumference versus body mass index (BMI) in fine-tuned demographic groups and in different ranges of adiposity. However, very detailed questions (e.g., about the utility of thigh circumference as a measure) may not be answerable through MIPD because only a few studies may have measured a variable of interest. Therefore, Dr. Duval emphasized the importance of prospectively identifying all variables that should be studied and developing a protocol to test important hypotheses.

One of the biggest advantages of the CODA project has been the establishment of personal relationships with the collaborators. These relationships were accomplished by phone, e-mail, and face-to-face meetings, with phone conversations and face-to-face meetings being the most productive forms of communication. In general, the collaborating researchers have been very responsive, particularly when well-known colleagues have been involved. Collaborators have also become more engaged as actual results, papers, and abstracts were developed.

Benefits of the collaborative effort have included more complete identification of studies, more balanced interpretation of the results, wider endorsement and dissemination of results, and collaboration on further research. Important factors in the CODA project’s success have been assurances to the collaborating researchers that all data sent to the data management site are secure and held in confidence, and that all published results will be in the name of the CODA Study Group. The collaborators are asked to join writing groups and have the opportunity to review drafts before publication. Other important aspects of the project are ensuring that individual studies have first rights in publishing their data and that all studies’ local review rules are followed.

Operation of the CODA MIPD has involved a commitment to obtain accurate, up-to-date data for all individuals in all relevant studies, and the greatest effort has involved establishing and maintaining collaboration, and processing the data. In addition, in merging the datasets, it has been important to determine whether the study protocols are similar and the source populations can be pooled. The least problematic task may be the data analysis.

Funding for the CODA Project began at the end of 2001, and the database now includes 37 studies, although some available databases are not yet included. Resource requirements for the project have included time, expertise (clinical, scientific, statistical, data management, computing, and administrative), money (researchers were offered up to $1,000 per study), and staff costs (which have totalled approximately 80% of the budget). The data are held at the University of Minnesota, the central project site, and a nominal steering group has been established.

Dr. Duval emphasized the importance of establishing study inclusion criteria. Inclusion criteria for the CODA project include baseline measures of age, sex, race/ethnicity, and one or more anthropometric indicators of obesity (e.g., waist circumference, BMI, or waist-to-hip ratio). Studies included in the CODA project were identified through WHO MONICA, DECODE, DECODA, and Medline searches; screening of abstracts of major international diabetes conferences; and personal communication with experts in the field. The project began by looking at follow-up studies for type 2 diabetes mellitus incidence and later expanded to include cross-sectional studies with newly diagnosed cases type 2 diabetes mellitus. Dr. Duval stressed the importance of identifying and including as many relevant studies (published and unpublished) as possible. If key studies are excluded, the results will be biased. Moreover, a large number of missing or unrepresentative trials could affect the meta-analysis results.

To establish collaboration, the CODA project team initially wrote a letter to all known investigators doing relevant research. The letter requested basic study information and protocols, but did not ask for data. It also discussed the CODA project aims and objectives, importance of the collaborative group, publication policy, collaborative group policy, and confidentiality of data. Many investigators were interested but time constraints were an issue. Ninety-nine studies received the initial letter, and 55 of the recipients gave a positive response. Of them, 52 studies were eligible, of which 37 datasets have been received and included in the CODA database. Most of the data sources are in the United States and Europe; only one dataset is in South America and only two are in Africa.

back to top

Dr. Duval listed classes of variables that would be included in an ideal dataset. These variables include:

• Class 1: Variables that can be measured using questionnaire/self-report only

• Class 2: Clinical variables that do not require drawing blood

• Class 3: Clinical variables that require drawing blood, but do not require a provocative challenge oral glucose tolerance test (OGTT )

• Class 4: Clinical variables that require OGTT

• Dependent variables—IGM (excluding diabetes), previously undiagnosed diabetes, previously diagnosed diabetes, and diabetes incidence

The CODA Project team found that data collection and transfer issues included the need to accept a variety of data formats, to assist collaborators by providing forms and payment incentives, and to accept a variety of data transfer methods. They also needed to maintain regular contact with collaborators through correspondence and meetings.

The CODA Project team has analyzed the data by stratifying the analysis by study, taking each study’s raw data and applying the same model across studies.

Discussion

The meeting participants briefly discussed the impact of the Health Insurance Portability and Accountability Act (HIPAA) on studies such as CODA. Dr. Duval commented that HIPAA has changed the face of research, but the project has not yet been impacted by the law because secondary data are being used. However, HIPAA may restrict the types of new research questions that could be asked in the future. Dr. David Jacobs of the CODA project team noted that initial study investigators may not have received patients’ consent if hospitalization records were used for studies included in the database. This issue will have to be monitored carefully.

Body Mass Index vs. Waist Circumference as Predictors of Diabetes in an International Context: An Individual Participant Data Meta-Analysis (the CODA Study)
David Jacobs, PhD, and Sue Duval, PhD, CODA Study Group, University of Minnesota

Dr. Jacobs, presenting on behalf of Dr. Duval and the CODA Study Group, described the methodology and findings of the CODA project’s meta-analysis of the association of BMI versus waist circumference (WC) as predictors of diabetes. He noted that there is overwhelming evidence that obesity is strongly associated with type 2 diabetes mellitus. The meta-analysis research questions were (slide 2):

• Which is the better predictor of type 2 diabetes, WC or BMI?
• What is the shape of the relation?
• Is the association the same in different populations?
• Is the association the same in different age groups, and for both sexes?

All projects included in the multinational collaborative CODA project (slides 3 and 4) gathered baseline glucose measurements (fasting glucose and/or oral glucose tolerance test) or incident diabetes, and baseline measurement of abdominal obesity. Analyses were restricted to studies with information on both WC and BMI, and the age range for the meta-analysis was restricted to >30 years at baseline; those under 30 will be included in future publications. Age- and sex-specific analyses used generalized linear mixed models, with random effects (slide 5). Age- and sex-adjusted risk ratios for diabetes were predicted from WC and BMI. Single-parameter models included a logistic regression model for baseline data, a proportional hazards regression model for follow-up data, and estimated absolute risk curves based on either logistic or Poisson regression. Multi-parameter models were also used. Diabetes outcomes, both prevalent and incident, included (slide 6):

• ADA definition 2003 (fasting plasma glucose ≥126 mg/dl),

• WHO definition 1999 (fasting plasma glucose ≥126 mg/dl or plasma glucose ≥200 mg/dl 2-h OGTT),

• Self-reported diabetes (medication, physician diagnosis, etc.), and

• Medication per pharmaceuticals registry.

Newly diagnosed diabetes was based on satisfying a blood glucose criterion in the absence of self-reported pre-existing diabetes.

Dr. Jacobs presented descriptive statistics for the 21 prospective studies and 13 cross-sectional studies included in the meta-analysis (slides 7-9). The prospective studies ranged in size from 658 to 52,468 subjects, with incident diabetes mellitus rates per 10,000 ranging from 16.2 to 401.7. The WC-BMI correlations ranged from 0.71 to 0.89, suggesting that the two variables are well correlated. The cross-sectional studies ranged in size from 466 to 25,902 subjects. The WC-BMI correlations ranged from 0.59 to 0.85.

back to top

The meta-analysis findings (slides 10-13) suggest that newly diagnosed diabetes, based on the latest ADA guidelines, can be predicted from BMI and WC using a logistic model, although BMI was slightly less predictive than WC. Incident diabetes was predicted from both BMI and WC using a proportional hazards model. WC appeared to be a slightly better predictor than BMI. The study also showed that when data were stratified by age group and gender, women had a slightly lower odds ratio for BMI, but not for WC, compared to men (slide 14). Dr. Jacobs noted that the cross-sectional studies produced the same kinds of results as the longitudinal studies, giving the investigators greater confidence in the poolability of the results for the cross-sectional and longitudinal data. In other words, Dr. Jacobs stated that the cross-sectional studies, with newly diagnosed diabetes as the outcome variable, and the longitudinal studies, with incident diabetes as the outcome variable, offered similar kinds of information about the association of adiposity and risk of diabetes. On the other hand, he stated that the associations of the two measures of obesity with prevalent (known) diabetes were much weaker than were those with newly identified (unknown) diabetes, perhaps because people who know they have diabetes have lost weight (data not shown). Therefore it is important to separate out the known diabetics in most studies of diabetes risk.

The above findings pertained to relative risk. To get some idea of the absolute magnitude of diabetes risk according to level of adiposity, the researchers also predicted the absolute risk of either newly diagnosed or incident diabetes for a 50-year-old individual based on the pooled data and adjusted by study (slides 15-17), and found that “the gradient is tremendous,” Dr. Jacobs reported. At the 90th or 95th percentile of obesity, the probability of newly diagnosed diabetes was much larger than at the lower percentiles of obesity. For incidence after 10 years follow-up the probabilities were also steeply graded, although they were lower than the corresponding values for newly diagnosed diabetes. Despite this high gradient of risk, however, these data suggest that even among the most obese people, 85% were not predicted to have previously unidentified diabetes or incident diabetes within the time frame examined. More cases would undoubtedly emerge with longer followup, but the data support the concept that, while obese people are at very high risk for developing diabetes, some obese people will not develop diabetes.

In addition, the investigators compared single vs. multiparameter models for both newly diagnosed and incident diabetes in relation to BMI and WC (slides 18-19). While single parameter models properly weight each study according to its precision, multiparameter models potentially introduce a peculiar kind of study bias. For example, one might use a multiparameter model to represent increasing categories of adiposity. In this case each category will be weighted optimally according to precision of the several studies contributing; however, different studies contribute differently across the range of adiposity, so the weighting of studies may vary between categories. Therefore comparison of a low adiposity category to the reference category may be greatly influenced by one set of studies, while comparison of a high adiposity category to the same reference category may be greatly influenced by another set of studies. When comparing the single-parameter vs. multi-parameter logistic and proportional hazards models, they found that the two models essentially agreed (so that no serious bias had been introduced) until the higher adiposity levels. The more flexible multiparameter models flattened compared to the less flexible single parameter models, probably reflecting difficulty in accurate measurement of weight and waist circumference in the most obese, rather than a true flattening of risk for diabetes. This finding suggests that the relationship between BMI and WC and diabetes is smooth and increasing; from the public health perspective, there appears to be no cut point for who is and who is not at risk. However, despite the graded risk, the researchers noted that the absolute increment in risk is very small for changes in BMI and WC within thin persons.

Several meta-regression analyses were presented (slide 20). These analyses use as dependent variable the study-specific slope estimating the ln(OR), in the cross-sectional studies, or ln(HRR), in the longitudinal studies, derived from each study’s individual participant data. Other ecologic characteristics of the studies were predictor variables. As with the analyses based on individuals, these ecologic analyses suggested that WC and BMI are good predictors of diabetes. Whereas the individual participant data meta-analyses suggested that the association between adiposity and risk for diabetes increased slightly with age, the meta-regression analyses did not find an age gradient (likely representing an ecologic fallacy). However, the association between adiposity and risk for diabetes was slightly weaker in the studies with low overall diabetes risk than in the studies with high overall diabetes risk. The investigators had no obvious explanation for this phenomenon, although it appears to be some sort of ceiling effect on diabetes risk.

A slide was presented that considered differences in estimated slopes according to protocol used for waist measurement (slide 21); this analysis found no significant difference in predictability of diabetes according to the 4 different waist measurement protocols, although the slope was nominally least when the measurement was taken at the narrowest point.

Dr. Jacobs noted that copious evidence in the literature suggests that visceral fat is a stronger predictor of diabetes than is subcutaneous fat. The motivation for the meta-analysis was therefore that, for screening and prediction purposes, it would be desirable to use a measure of obesity that is more specific to visceral fat. Such a measure should be easy to use and generalizable so a practitioner could readily and more accurately assess a person’s level of risk. WC is a reasonable candidate for such a measure, but perhaps it should be modified for frame size by height or hip circumference. BMI intuitively relates to fat generally, rather than specifically to visceral fat. However, the correlation between waist and BMI is about 0.8; with this high level of correlation, BMI and WC seem to offer similar information about adiposity. Therefore, it is important to ask whether WC offers an empirical improvement over BMI in prediction of diabetes, and whether other simple modifications such as incorporating hip circumference or height improve the prediction. A series of analyses (slides 22-23) was presented showing that there was little gain in predictivity of a variety of such models over use of BMI or WC alone. Some slight improvement, however, was afforded by combinations of BMI, WC, and hip circumference. Investigations of these more complex models continues.

back to top

In summary, the meta-analysis suggested that (slides 24-26):

• Waist circumference and BMI are both strongly and consistently related to diabetes risk.

• The association is largely similar whether using newly diagnosed diabetes (based on either the ADA or WHO diagnostic criteria) or incident diabetes. The association using total diabetes prevalence or known diabetes as the outcome variable is substantially weaker.

• Even modest overweight is associated with increased risk.

• There is a smooth gradient across the range of adiposity; the flexible multi-parameter models demonstrated this. They did not introduce much bias compared to the single-parameter models.

• Using several different analytic techniques, WC was consistently a slightly better predictor of risk diabetes than was BMI.

• There was further slight improvement in the prediction model when using height in the model.

• There is statistically significant heterogeneity among studies: nevertheless, all but one study showed the same qualitative increase in risk of diabetes as adiposity increased. Differences between studies suggest that adiposity is a stronger predictor in some studies than in others.

• The risk gradients were similar for men and women, but are slightly stronger at older than younger ages.

• Diabetes prevalence or incidence in the population is inversely related to the strength of the diabetes-obesity association, explaining part of the heterogeneity.

The researchers concluded that WC does not appear to offer substantial advantages over BMI and that the two are almost interchangeable in diabetes prediction. Whether there is an interaction between BMI and WC has not been investigated, although enough data are probably available through the meta-analysis to do so.

In conclusion, Dr. Jacobs said that “To prevent diabetes you have to know who will get it.” The meta-analysis showed that adiposity is important but other information indicates that it is not the only factor in diabetes risk. This study provides a good example of what can be learned by synthesizing existing data from different studies.

Discussion

During the discussion that followed Dr. Jacobs’ presentation, a participant asked whether the researchers will be able to look at duration of obesity over time (i.e., how long someone was overweight before diabetes is diagnosed). Dr. Jacobs responded that the meta-analysis would be suited to answering that type of detailed question, providing that the CODA dataset includes weight at a particular age or change in weight over time. However, only studies that asked about weight history or that developed a weight history by following participants over time could contribute information to this question. Another participant noted that subcutaneous abdominal fat could contribute to abdominal circumference, especially in some ethnic groups. Dr. Malozowski and Dr. Jacobs commented that the roles of subcutaneous versus visceral fat are of current interest. Dr. Grave suggested that it would be important to look at BMI and WC in the African-American population; a person’s descent (African versus European) may play a role in risk. However ethnicity may play a different role in one culture (e.g. African-Americans or Mexicans in the United States) than in another (e.g. native Africans or Mexicans), so that such studies may need to be restricted geographically.

Access to Data from NHLBI Population Studies
Peter J. Savage, MD, Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute

Dr. Savage discussed the availability of data from NHLBI-supported intervention and population studies. In particular, he provided an overview of participant privacy versus public access, how to protect privacy, the types of study data that are available for release by NHLBI, and the procedures for accessing the data. He noted that data from existing studies can be used for several purposes, including confirmation of results, analysis of secondary hypotheses, analysis of data for new study designs and samples, sample size estimation for future clinical trials, and teaching.

The advantages of making data available to outside investigators must not lead to violation of participants’ rights. When using data from existing studies, all investigators (original or those getting access later) have an obligation to respect the privacy rights of study participants. In recent years, there has been an evolution in understanding of informed consent and defining the boundaries of how data can be used in different kinds of studies. Investigators must consider whether informed consent forms for a given study allow specific new questions to be addressed.

NHLBI grants limited access to data from several of its large intervention and observational studies. Data requests and proposed analyses must be approved by the data recipient’s IRB. When data are provided, obvious identifiers, inadvertent identifiers, and sensitive information are removed, although such data are not completely anonymous. Individuals requesting data must sign an agreement about how the data will be used. Violation of these agreements can have serious consequences both for the outside investigator and his/her institution including restrictions on approvals of future data requests.

back to top

An alternative to using limited access data is to collaborate with investigators already involved in ongoing studies. In many cases, this is the most fruitful and efficient approach. For example, interested investigators can identify primary study investigators to serve as ancillary study sponsors. The new study and paper proposals are developed with input as needed from the primary study sponsor. Proposals and manuscripts from such collaborations must be reviewed by the study Steering Committee of the study )or its designated representatives) and NHLBI.

Dr. Savage said that advantages to working with ongoing studies include: potential access to the complete main study dataset; access to biological samples, images, recordings, etc.; access to the expertise and experience of the study investigators and, possibly, access to the cohort for new data collection and assistance with administrative requirements. The only disadvantage he listed is the additional “red tape” and delays that result from the review process.

The success of NHLBI’s promotion of data sharing is demonstrated by results to date in the Cardiovascular Health Study (CHS), a study originally established to investigate cardiovascular disease and its risk factors in the elderly. Several training grants have made use of CHS data. More than 130 approved ancillary studies have been conducted using CHS data, and collaborators have been the principal investigators for almost 70 percent of these studies. In addition, during the past three years, collaborators from outside of the original study have been primary authors for more than 60 percent of the study manuscripts.

Epidemiology datasets that are currently available from NHLBI include the Framingham Original Cohort, Framingham Offspring, Honolulu Heart Program, Atherosclerosis Risk in Communities (ARIC), and Coronary Artery Risk Development in Young Adults (CARDIA), and the CHS studies. Clinical trials datasets are available for the Asymptomatic Cardiac Ischemia Pilot (ACIP), Intermittent Positive Pressure Breathing (IPPB), Post Coronary Artery Bypass Graft Study (Post CABG), Thrombolysis in Myocardial Infarction Study (TIMI II), Lung Health Study (LHS), Digitalis Investigation Group (DIG), Beta Agonist in Mild Asthma (BAGS), Antiarrhythmics Versus Implantable Defibrillators (AVID), and Colchicine in Moderate Asthma (CIMA) studies. The NHBLI seeks to make new datasets available to others in a timely fashion. This list continues to expand.

However, NHLBI has an obligation to give the primary study investigators the opportunity to publish their findings first. For epidemiological studies, public access datasets are available five years from the close of data for examination or follow-up, and for clinical trials, datasets are available three years after publication of the primary results report. Information about how to obtain data from NHLBI can be found at: http://www.nhlbi.nih.gov/resources/deca/default.htm. The Website provides data documentation, distribution agreement forms, information for recipients’ IRBs, and policies for use of the data. Data are provided at no cost to the recipient. So far, NHLBI datasets have been requested less often than hoped, especially for studies that have not been widely publicized. From April 2000 to December 2001, the Framingham study was requested most often (35 requests), followed by the ARIC Study (13 requests) and CHS (11 requests).

Dr. Savage summarized his remarks by stating that data from a large number of NHLBI contract-funded studies are available through public access and that investigators are encouraged to access the data for new studies. Privacy is protected by reducing risk of identification, by requiring binding agreements, and by requiring IRB approval. Although not a mandate, investigators who access data are encouraged to communicate with the original study investigators to try to prevent misunderstanding about the data. Collaborations between outside investigators and investigators already involved with a study are encouraged. Overall, these policies facilitate getting the most value from the information we collect while protecting the rights of all involved.

Brief Overview of Ongoing Activities
Michael Engelgau, MD, Centers for Disease Control and Prevention (CDC)

Dr. Engelgau commented that one part of the CODA study is examining population-based strategies to detect people with pre-diabetes, which is an important CDC focus. The DETECT-2 study, a complementary study based in Copenhagen, combines approximately 27 European surveys and 23 Asian surveys in a dataset that will be used to look at strategies for detection of people with pre-diabetes. The CODA and DETECT-2 investigators are working together in this effort.

A year ago, the CDC convened 15 investigators from around the world to discuss methods for identifying people with pre-diabetes. In general, two different strategies to detect people with pre-diabetes are being studied: first, direct measurement of glucose and explicit assessment of glycemic status, and second, prediction of future diabetes risk without actually measuring a glucose value. Both strategies hold promise in various settings, Dr. Engelgau said. The recent decrease in the impaired fasting glucose (IFG) cut point from 110 to 100 changes the pool of people with undiagnosed pre-diabetes from 20 million to 40 million. Most of these individuals have isolated IFG, not impaired glucose tolerance (IGT). Scientific evidence about the IFG population is not as well-documented as the evidence for the IGT population. This has an impact on policy issues related to pre-diabetes detection and primary prevention efforts.

back to top

Furthermore, isolated IFG was not part of any of the prevention trials, so it is unclear whether that measure will have the same benefit as IGT. The group that developed the IFG criteria in 1997 recently reconvened to review four longitudinal studies from around the world. When they lowered the cut point to 100 impaired glucose tolerance, it predicted future diabetes about the same as when using previous cut points. This was an attempt to align IGT criteria with the IFG criteria.

Myrlene A. Staten, MD, National Institute of Diabetes and Digestive and Kidney Diseases

Dr. Staten said there is a critical need for a better diagnostic test for diabetes, as well as pre-diabetes and IGT. The current method requires patients to be fasted and then have a 2 hour OGTT, which must be reconfirmed on a separate occasion.. To this end, NIDDK is creating a Translational Research Program Announcement to stimulate research into a better diagnosis method. The Institute currently is supporting an application to look at a one-hour non-fasting test to determine if it is comparable to the fasting, 2 hour OGTT and if it would be more acceptable and easier for patients and providers.

Another NIDDK program announcement (PA-04-076), issued March 18, 2004, is addressing proteomic and metabolomic approaches to diagnose diabetes and pre-diabetes. It solicits applications of proteomic and metabolomic technologies for the development of novel methodologies and/or identification of new biomarkers for the diagnosis of diabetes and type 2 diabetes. The first receipt date is July 20, 2004. Additionally as part of the previously mentioned grant to study a non-fasting, one hour test for diabetes diagnosis, the investigator is collecting samples from a 2,000 people that will be put into the NIDDK sample repository for use for proteomic/metabolomic approaches.

In addition, NIDDK is interested in epidemiologic approaches to identify those at risk for diabetes.

A participant commented that a significant percentage of people with impaired glucose tolerance will not become diabetic, at least for not for a number of years. However, the window of opportunity to intervene with people who are moving toward diabetes is narrow. Having two glucose tolerance tests showing IGT is a good predictor.

Discussion

The Committee discussed the merits of hemoglobin A1C as a diagnostic indicator for type 2 diabetes. Dr. Staten said that it has been shown to a poor indicator for diagnosis of type 2 diabetes. An OGTT and single fasting glucose test have some variability but are more robust measures than is the A1C. The A1C is good way to monitor glycemia, but near the normal range, it is not a good indicator of diabetes. Nevertheless, the hemoglobin A1C was a positive advance. Other proteomic measures could also be as useful. Dr. Spiegel commented that other diagnostic research directions include epidemiology efforts and scientific technologies. In addition, work in the genetics arena includes the study of genetic polymorphisms that may be important because genetic tests could be used to assess a person’s diabetes risk.

Dr. Spiegel then introduced a brief discussion about longitudinal trends in height and the relationship of BMI to height. Increases in height have leveled in the United States and other populations’ height gain is greater, which could have implications for BMI. For example, the heights of discrete populations that move from one country to another have been shown to increase. The U.S. military has routinely gathered height data for personnel, and these data might be useful for studies in conjunction with the Department of Veterans Affairs. Data on the leveling of height in the United States may also have implications for nutrition. In addition, the Committee discussed the use of height squared versus height cubed in calculating BMI.

Closing Remarks

Dr. Malozowski stated that work to establish a Committee agenda for fiscal year 2005 will begin in July. In October, members will receive a letter requesting information for the Committee’s annual report. Finally, the participation of the attendees was acknowledged and the meeting adjourned.

back to top


Diabetes Mellitus Interagency Coordinating Committee Meeting on the Special Statutory Funding Program for Type 1 Diabetes Research

Building 31C, 6th Floor Conference Room 10
NIH Campus, Bethesda, MD
July 28, 2004

Introductory Remarks
Allen Spiegel, MD

Dr. Spiegel welcomed the DMICC members and guests, and the meeting participants introduced themselves. Dr. Spiegel thanked the representatives of the National Cancer Institute (NCI); National Heart, Lung, and Blood Institute (NHLBI), National Institute of Neurological Disorders and Stroke (NINDS); and National Eye Institute (NEI) who joined the NIDDK representatives at an April 2004 meeting on angiogenesis research in Towson, Maryland. Facilitated by the Juvenile Diabetes Research Foundation, the two-day meeting convened leaders in the field of angiogenesis research who discussed the relevance of angiogenesis research to diabetes complications and islet transplantation as it relates to vascularization. NIH was well-represented, and invited discipline-specific experts participated. There is clear movement in the cancer field to implement ways to inihibit angiogenesis, Dr. Spiegel reported. A vast amount of research information has been gained, and application of this information to diabetes complications and islet transplantation is important. A group continues to meet about how to spur research initiatives in that area. NHLBI recently issued a request for applications (RFA) on heterogeneity in the arterial, venous, and lymphatic vasculature, which is relevant to this theme, although focused on normal rather than disease states. Additional initiatives will sharpen the translational focus and address intramural initiatives.

back to top

Dr. Spiegel said that Health and Human Services Secretary Tommy Thompson has convened a series of national diabetes town hall meetings that are in part are related to the crafting of a National Diabetes Action Plan. The office of Dr. Michael O’Grady, Assistant Secretary for Planning and Evaluation, is driving this activity. The first three town hall meetings were held in Cincinatti in May and in Little Rock and Seattle in July. Secretary Thompson, other Federal health officials, and State and local health officials have participated in each of the meetings. Several hundred people have attended each meeting, which have offered opportunities for participants to make statements, ask questions, and provide input into the crafting of the National Diabetes Action Plan. The focus of the plan is not on research, although research will inform the plan. Rather, the focus is on the more distal aspects of care, which was reflected in questions fielded during the meetings. For example, diabetes education and treatment of students in school have been a concern. The school guide crafted under the joint NIDDK-Centers for Disease Control and Prevention (CDC) National Diabetes Education Program has been featured by the Secretary and has been helpful, but this area continues to be controversial and challenging. The American Federation of Teachers (AFT) issued a statement prohibiting teachers from getting involved in any aspect of care of students with diabetes at school. Additionally, the town hall meetings have addressed both type 1 and 2 diabetes and have raised topics such as stem cell research, the Department’s Diabetes Detection Initiative to reach undiagnosed persons, insurance and liability coverage issues, and obesity. The target date for presenting the National Diabetes Action Plan is November 2004. Dr. Spiegel said he is pleased to participate in the effort, but emphasized that NIH is not driving the process.

In response to a question about teachers providing care for students, Dr. Spiegel explained that concerns include students getting insulin shots at school, diabetes education, and liability issues. Other issues concern school staff other than school nurse being allowed to give students insulin shots and students testing their blood and self-administering insulin shots at school. Another participant commented that he has seen similar scenarios with school superintendents and boards of education regarding asthma care. There is a reluctance to take on such activities because they reduce teaching time, which is a reasonable concern at many schools.

Brief Historical Overview of the Special Funding Program
Judith Fradkin, MD

Dr. Fradkin presented an update about the Special Type 1 Diabetes Funding Program and led a discussion to help plan for an evaluation of the program, including planning a program advisory meeting. She then presented a brief historical overview of the program, which was first funded in fiscal year (FY) 1998 at $30 million per year for three years. The funding increased to $100 million per year for FY 2001 to FY 2003 and increased to $150 million per year for FY 2004 to FY 2008. Funding for FY 1998 to FY 2008 totals $1.14 billion. Initially, a final program evaluation was to be submitted to Congress in January 2003, but the date was extended to January 1, 2007. An interim progress report was published in April 2003 (the report can be accessed at: http://www.niddk.nih.gov/federal/planning/type1_specialfund/).

The program’s six major goals are to:

• Identify genetic and environmental causes of type 1 diabetes,
• Prevent or reverse type 1 diabetes,
• Develop cell replacement therapy,
• Prevent or reduce hypoglycemia,
• Prevent or reduce complications, and
• Attract new talent and apply new technologies to type 1 diabetes.

Dr. Fradkin proposed that the last goal be modified to include “apply new technologies”; the Committee endorsed the addition.

NIDDK has been asked to manage the program funds on behalf of HHS, with active involvement of all components of HHS, as well as the research and voluntary communities. Program principles include flexible budgeting to allow rapid response to emerging scientific opportunities and to keep type 1 diabetes funds discrete from regularly appropriated funds. Much of the funding has been used to:

• Establish large-scale collaborative, infrastructure-intensive fundamental initiatives (e.g., the Beta Cell Biology Consortium) that could not be pursued with R01 funds;

• Create major clinical trials networks;

• Promote innovative, high-risk, high-impact research that is different from typical RO1 research; and

• Promote translational research to develop new therapies.

The effort has been guided by a series of advisory group meetings, including trans-HHS meetings and meetings of outside experts. In April 2000, an Advisory Panel was convened to solicit input from experts about projects to be pursued. Twenty-two proposals were proposed and 11 were endorsed; the panel also recommended a twelfth need. The Committee also strongly endorsed the six major goals and initiatives.

back to top

In May 2002, another expert advisory panel was convened. The panel evaluated research efforts and opportunities; strongly endorsed the major goals and initiatives; and made general recommendations, including to continue support for resources and infrastructure, continue to develop and apply new type 1 diabetes technologies and to attract new research talent, and to re-issue RFAs to create ongoing research opportunities such as the Bench to Bedside and Innovative Partner RFAs.

Two meetings were held in April 2003. At one meeting, participants were asked for advice about diabetes complications research portfolio, and they recommended projects on animal models of diabetes complications and more efficient trials on complications. At the second meeting, the group encouraged coordination among the type 1 diabetes consortia and networks, and mechanisms that would attract investigators with diverse expertise.

Update on Special Funding Initiatives Since April 2003 DMICC Meeting
Judith Fradkin, MD

Dr. Fradkin updated the DMICC on the Special Funding Program activities that have taken place since the April 2003 meeting. Initiatives have included:

• In the past year, NIDDK and the National Institute of Allergy and Infectious Diseases (NIAID) jointly issued two RFAs for clinical centers and a coordinating center to develop a consortium to pursue islet transplantation clinical studies. This will be further enhanced through cooperation with the Centers for Medicare and Medicaid Services because subsequent legislation has encouraged Medicare to provide support through a demonstration project through which kidney transplant patients who receive islets will be enrolled in projects that are carried out through the Islet Transplantation Consortium. The RFA was structured with input from an advisory committee meeting held in May 2003.

• NIDDK has partnered with NCI through the T1D-RAID program to enhance translational research through a pilot program that makes RAID resources available to type 1 diabetes investigators. Two cycles of applications have been solicited and several projects have been approved for provision of resources and services through the RAID program.

• In April 2003, a meeting on proteomics and diabetes resulted in recommendations that were developed and pursued through an RFA. Applications will be reviewed in the near future for FY 2004 funding.

• Also in April 2003, NIDDK and NHLBI co-sponsored a meeting, which informed the development of an RFA on cardiovascular complications of type 1 diabetes.

• A workshop on opportunities to foster beta cell biology in the 21st century and recommendations on beta cell biology from an advisory meeting on islet transplantation led to an RFA for pilot and feasibility programs in human islet biology.

• NIDDK canvassed the Committee at a previous DMICC meeting about opportunities for research using small business partners and issued a solicitation that will be reviewed for funding this fiscal year.

• As recommended by the external advisory panel, the Bench to Bedside, Innovative Partnerships, and hypoglycemia solicitations were reissued.

• NIAID is taking the lead in developing an RFA to create a consortium for xenotransplantation studies.

• Plans are underway to develop an RFA on angiogenesis and type 1 diabetes.

A listing of all the initiatives supported by the special program can be found on the NIDDK website at: http://www.niddk.nih.gov/fund/diabetesspecialfunds/funding.htm.

Another recommendation was to enhance coordination of research through a Type 1 Diabetes Consortia Coordinating Committee. This Committee’s charge is to coordinate issues related to recruitment and enrollment; standardization of assays, phenotyping, and consents; use of clinical populations for development and validation of assays for immune and metabolic monitoring; bioinformatics; and ancillary studies. The effort is designed to develop a framework and to work with basic consortia to standardize bioinformatics efforts. Dr. Fradkin thanked Dr. Jerry Nepom and everyone present who has participated in this effort. A Website (www.niddk.nih.gov/fund/diabetesspecialfunds) is now available to provide information to investigators about resources and funding opportunities, and information for type 1diabetes patients about opportunities to participate in clinical studies.

Discussion and Planning of Expert Panel Meeting

Dr. Fradkin led a discussion about planning an Expert Panel meeting. The meeting will provide an opportunity for mid-course assessment of the status of ongoing efforts, particularly with regard to large-scale consortia, and to identify new opportunities. Issues discussed and decisions made about the Expert Panel meeting included:

Meeting timeframe

Dr. Fradkin asked the DMICC for recommendations about when the Expert Panel meeting should be held, with consideration given to implementing recommendations in 2006. If new RFAs will be issued in 2006, they would need to be published by the fall of 2005. Language in the FY 2005 House Appropriations Bill requests a report on plans for an evaluation of TrialNet, so TrialNet would be on a separate planning trajectory. Committee members felt that beginning in the fall or winter would allow more time to act on recommendations and to enable more partners to become involved.

back to top

The Committee agreed that the Expert Panel meeting should be held in the winter of 2004-05, possibly in January. Dr. Fradkin noted that members would need to be able to get needed information to her office within that timeframe.

Efforts to be discussed at the meeting

Dr. Fradkin explained that funding emphasis has shifted from research projects to consortia, so consortia or repeatedly issued RFAs will be a focus of review. She asked the Committee whether the Expert Panel should discuss large consortia (over $5 million total funding), repeated RFAs (over $10 million total funding), or continuations. She also listed 19 initiatives that the Expert Panel would be asked to evaluate. Dr. Spiegel noted that the initiatives are led by NIDDK, NIAID, NHLBI, the National Institute of Child Health and Human Development, the National Center for Research Resources, and CDC. Dr. Fradkin said the information would need to be mailed to the Expert Panel at least one month before the meeting, and material must be provided by early November.

Dr. Fradkin said that coordination will be needed for RFAs involving consortia. RFAs that are likely to be reissued will be the focus, and the lead IC for each of those RFAs will be asked to gather information from each of the participating ICs. The Committee felt that the summaries should be limited to three pages.

A suggestion was made to review the T1D-RAID program, which did not meet $5 million criterion, to look at ways to catalyze and further activate the program. Dr. Fradkin agreed, saying that the program has seen a steep increase in the number of applications, but the number could be increased further. Dr. Spiegel suggested that thought be given to a more proactive process to spur translational activities, which might be coupled with an evaluation of the RAID program. He also noted that Institutes’ capacity is a concern, and an evaluation and a substantial commitment of funds may be needed.

A question was raised about whether the panel should consider access to special funds for ongoing consortia currently supported by regularly appropriated funds (e.g., Autoimmunity Centers of Excellence). Dr. Fradkin noted that this consortium currently supports studies of multiple autoimmune diseases and questioned why type 1 diabetes studies undertaken by the consortium should not be supported regular funds along with those on the other autoimmune diseases. She pointed out that the special funds were intended to allow for initiatives that could not otherwise be pursued with regular resources rather than funding existing work.

A member said that the Diabetic Retinopathy Research Network is expanding rapidly and asked whether the review committee could discuss the possibility of additional type 1 diabetes special money for the network. Dr. Fradkin pointed out that this Network is relevant to both type 1 and type 2 diabetes and noted that special funds were not intended to replace Institutes’ funding. Dr. Spiegel added that a very compelling case would need to be made as to why type 1 diabetes funds should be used in such situations, but that flexibility in use of the funds may be needed to allocate the funds in the best possible way.

back to top

Dr. Fradkin suggested that it is important to limit the number of initiatives to be reviewed at the Expert Panel meeting to allow time for meaningful review of projects currently supported with the special funds, and allow consideration of new ideas. The focus of the panel should be evaluation of what is currently being supported with the special funds and generation of ideas for new opportunities to be pursued with the funds.

Sample template for presenting information to the Expert Panel

A sample template for presentation of information to the Expert Panel was distributed. The information, to include a statement of goals, accomplishments, future opportunities, and milestones, will be limited to three pages. Consortia should concisely identify progress toward specific goals in reference to the missions of the organizations. Reports from relevant external advisory groups, or minutes of most recent meetings of External Advisory Committees or Data Safety Monitoring Boards for clinical trials, which provide comments on progress and performance of the consortia should be appended to the summaries.

The Committee had no comments about the template. The template will be distributed to the DMICC, with summaries due November 1.

Advice to be obtained from the Expert Panel

Dr. Fradkin asked for the Committee’s advice about what the Expert Panel should be asked. The two major questions are: What is the effectiveness of the programs that wish to continue to receive type 1 diabetes funding? What are the new opportunities? The DMICC and HHS components can suggest new research opportunities, and the Expert Panel should be expected to recommend new opportunities, she said.

A suggestion was made to ask the larger scientific community to suggest possible research opportunities that could be presented to the panel at the time of the Expert Panel. This information could be gathered through a one-page form that is available online. Dr. Spiegel said that this idea would be considered, although some parameters would need to be set. A process to review suggestions would also need to be established.

The Committee had no further comments about advice to obtain from the Expert Panel.

In summary, Dr. Fradkin said that the DMICC will receive a request for completed templates (to be no longer than 3 pages; due November 1, 2004) and for Expert Panel suggestions, due within a week to 10 days. Suggestions for panel members should not be leaders of consortia but should be involved in type 1 diabetes research. A list of principal investigators who would not be eligible for the panel because of conflicts will be sent the DMICC members.

Thinking Ahead: Mandated Program Evaluation
Mary Hanlon, PhD, NIDDK Office of Scientific Program and Policy Analysis

Dr. Hanlon said that the NIDDK Science Policy Office has lead responsibility for conducting the final Special Funding Program evaluation. The due date for submitting the evaluation to Congress is January 1, 2007, although clearance will likely take more than six months. A progress report, which will be useful in conducting the final evaluation, was published in April 2003. However, the final evaluation will have a different focus than the program report.

back to top

In thinking about the Special Funding Program and what will be evaluated, her office has divided the program into three areas of focus:

• Translational research efforts (e.g., Bench to Bedside and T1D-RAID),
• Research consortia and networks (e.g., BCBC and T1DGC), and
• Investigator-initiated research (RO1 and R21).

The 2003 evaluation focused on investigator-initiated research because data were collected in 2001 and 2002, when most funding supported investigator-initiated research projects. More recently, the focus of funding has shifted to establishment of consortia. Therefore, the focus of the final evaluation and data collection will be on the consortia, although it is also important to evaluate the scientific accomplishments of investigator-initiated research projects.

Her office has identified the following questions to evaluate the overarching program effectively:

• What are the major scientific accomplishments?

• Did the program adequately identify scientific opportunities and challenges, and design initiatives to address them?

• Did the program seek the advice of external experts to identify and pursue compelling research directions?

• Did the program foster innovative and clinically oriented research?

• Did the program make a positive impact on the field of type 1 diabetes?

• What do we expect in the future and what is the potential impact of the program on type 1 diabetes?

The evaluation will be process- and output-oriented evaluation. It is not planned to be an outcome-oriented evaluation because the program is relatively new, scientific outcomes are not yet available, and many of the initiatives are multi-year.

To answer the above questions, data sources will include:

• Literature searches to identify the number of publications supported by the program and the scientific impact,

• Grantees’ progress reports to identify scientific accomplishments,

• External advisory committees and expert panels to evaluate program components such as consortia, and

• A grantee survey.

NIDDK has applied for funding for meeting logistics, graphic design, and data collection expenses.

A participant recommended that outcomes be reviewed as part of the evaluation. Doing so may be challenging, but it would be helpful to look at outcomes of the funded programs. Dr. Fradkin commented that it would be helpful for the Committee to think about a strategy to showcase what this program has meant, and how interim steps and progress are moving in the direction of achieving the program goals.

back to top

Dr. Spiegel suggested that it would be useful to look at outcomes in the context of the six program goals, some of which lend themselves more to process evaluation. If the goals are viewed as outcome measures, then progress and successes under each goal could be listed (e.g., have identified x number of genes and have used that to enhance prediction of who is at risk of type 1 diabetes). Careful language can make it clear that progress has been made toward reaching the overarching program goals.

Dr. Hanlon said she welcomes comments and questions about the evaluation. She can be reached at hanlonm@extra.niddk.nih.gov or 301-496-6623.

Closing Remarks

Dr. Fradkin thanked the DMICC members for their participation and involvement in the Expert Panel review process. Potential members of the Expert Panel will be contacted to determine a meeting date.

The meeting adjourned at 2:30 p.m.

back to top