Key words: neonate, temperature measurement, warmers
Infant crib warmers are roughly diaper-sized disposable pouches of iron
powder, charcoal, water, and wood fibers which warm when exposed to the
atmosphere. They are used in hospitals to maintain an appropriate body temperature
in compromised neonates. A number of incidences of burns which resulted
when the devices were used in oxygen-rich atmospheres led to a health hazard
evaluation of one manufacturer's device. OST
was asked to perform testing on a sample of the device to provide scientific
justification for potential action against the manufacturer and to help
establish guidelines for labeling the devices.
The crib warmers were tested in a Forma Scientific model 3159 incubator, which has the capability for controlling both oxygen and carbon dioxide levels. Temperature was measured with thermistor-based instruments accurate to approximately 0.1° C. During each test, the warmer was placed atop several layers of cloth to isolate it thermally from the chamber walls. Two sensors were placed on the surface of the warmer, one on each of the two independent pockets of chemicals. In some of the tests, plastic flasks containing water were also placed on the warmer to provide thermal loads. The flasks were not intended to accurately model the thermal properties of an infant. Two volumes of water were studied, 125 cc and 60 cc. The water temperature inside each flask was monitored with a thermistor.
Warmers were tested in environments with oxygen content of 20.7% (standard air), 30%, 50%, and 75%, to span the range of oxygen content of incubators and isolettes. At each oxygen level at least three warmers were tested, except at the 75% level, where only one warmer was tested to conserve oxygen. After each warmer was opened, the temperature measurements were made every 10 minutes. As the temperature stabilized, measurements were taken every 30 minutes.
The temperature variation across the surface of the warmers was quite large - as much as 10° C when higher oxygen levels were used. This was due in part to the presence of lumps and voids in the ingredient packs, which invariably existed even after vigorous shaking of the warmers. Thus, an effort was made to place temperature sensors on these hotter areas. The water flasks, with a surface area of 80 cm², provided some averaging of these variations. The data from the tests at the different oxygen levels is summarized in figure 09. The average maximum is the mean of the maximum temperatures recorded on the individual warmers. The highest temperature represents the highest temperature recorded on the surface of any of the warmers at any time at that oxygen level. For the water load measurements, the temperature rise for the larger volume of water lagged behind the rise for the smaller volume, but the maximum temperature ultimately attained was primarily a function of positioning on the pad.
Results of OST's tests of temperature rises generated by infant crib warmers when the warmers are placed in oxygen-enriched environments. Each device tested reached unsafe temperatures when used in environments with more than 50% oxygen.
These data served to document that dangerously high temperatures can be reached by the warmers in oxygen-rich environments. The surface temperature exceeded 50° C at all of the elevated oxygen levels. Even within the water-containing flasks, the temperature exceeded 50° C at oxygen levels above 50%. Conversely, in room air, the temperature of the warmers did not meet the manufacturer's specification (38° C vs. 43° C.). These devices were recalled by the manufacturer, and all devices of this type were placed on mport detention pending review of labeling that would make clear that the devices are not to be used in oxygen-enriched environments. [Enf, PostMS]
Return to the Table of Contents.
Key words: hemolysis, test methods, blood pumps
OST scientists are investigating methods used to assess blood cell damage, so that the safety of cardiovascular devices may be appropriately evaluated. Testing for blood damage is a critical step in the approval of cardiovascular implants, such as left ventricular assist devices and of cardiopulmonary bypass systems. Some of the key aspects of this laboratory work have included the development of a suction and storage device for the atraumatic collection and handling of blood when drawn from animals, the implementation of laboratory techniques to obtain purified hemoglobin and bovine plasma for assessment tests, and the collaboration between OST and other institutions to compare testing methodologies and blood sources.
Experiments to assess hemolysis (i.e. red blood cell destruction) were performed on heart valves and on roller and centrifugal-type blood pumps. Figure 10a shows a simple loop for testing blood pumps by circulating blood in vitro. Figure 10b compares the amount of free plasma hemoglobin released from damaged red blood cells caused by two different roller-type cardiopulmonary bypass blood pumps. Hemolysis tests like these can also be used to compare how blood condition parameters, such as blood temperature, anticoagulation, animal source, pH, and lipid content, may affect the hemolysis results. The expertise developed through performance of these tests has been used to provide assistance to ASTM and ISO in their efforts to develop standards concerning hemolysis testing.
A schematic diagram of the flow loop used to evaluate the hemolytic character of a roller blood pump.
Hemolysis test results for two roller pumps. The values of free plasma hemoglobin observed are within the acceptable bounds for such pumps.
OST is also extending a study of nonroller (centrifugal) blood pumps which the Center requested from the manufacturers of these devices. Both the Center and the manufacturers are interested in the reclassification of these devices from Class III to Class II. A key component of the potential reclassification is the special control afforded by bench testing of these pumps for blood damage and the creation of a database on hemolysis testing of blood pumps. OST completed the first testing of a centrifugal pump mock loop, in which one centrifugal pump was tested for 4 hours at a flow of 4 liters per minute. Pooled blood and loop blood samples were taken every 30 minutes, and plasma-free hemoglobin and the hemolytic index were determined. The pump performed well at both room temperature and body temperature. The results from this study were used to aid ODE in making recommendations for performing hemolysis tests and interpreting tests carried out by manufacturers. [PreME, PostMS, Stds]
Return to the Table of Contents.
Key words: heart valves, cavitation, interlaboratory study, acoustic emission
Mechanical heart valves make an audible click when they close. This sound is due to the impact of the leaflets on the housing of the valve. If the "deceleration" of the leaflets is sufficiently high during impact, then a combination of fluid dynamic effects may result in transient bubble formation in very low pressure regions near the leaflets. As the closing event is completed (in about 1 millisecond), the pressures return to normal and the bubbles collapse violently. Such "collapse cavitation" has been shown to damage materials, such as the pyrolytic carbon surfaces of valve leaflets. One manufacturer's valve was withdrawn from the market several years ago after several valve failures were associated with cavitation-like damage. CDRH's guidance to manufacturers now asks for testing to evaluate any valve's potential for cavitation damage.
OST scientists planned and coordinated an interlaboratory study of an OST-proposed method to test for cavitation potential. The study participants included OST's Hydrodynamics and Acoustics Branch and four of the major manufacturers of heart valves made with pyrolytic carbon (Baxter, CarboMedics, Medtronic, and St. Jude). The protocol for the study was written by OST and accepted by the participants. OST also acted as a control laboratory, performing the measurements first and then again late in the round robin. All participants completed their measurements and submitted reports to OST. A preliminary summary was drafted which led to a meeting to discuss the results in conjunction with the meeting of the Society of Thoracic Surgeons in January. Revisions to ODE's guidance for replacement heart valves have been written and sent to ODE. The substantial changes to the guidance include 1) a newly established range of expected threshold dP/dt values; 2) a determination that a rigid mount of the test valve is unnecessary; 3) a determination that the measurement parameters for bi-leaflet valves must be more carefully defined; 4) that nucleation sites in the blood analog should be controlled; and (5) that a single valve test system is adequate and probably easier to use than a mock loop (two valve) system.
Investigations of in vitro acoustic emission techniques to detect cavitation generated by prosthetic heart valves were also completed. Acoustic techniques may make possible the detection of cavitation while the valve is still implanted in the body. OST findings indicate that the sounds associated with closing include detectable levels of noise due to collapsing bubbles. This work will be continued to determine whether the public health advantage of such a detection method can be realized. [PreME, ProA]
Return to the Table of Contents.
Key words: flow visualization, cardiovascular devices, spinal needles
The performance of many medical devices is dependent upon complex flow patterns which are difficult to assess. Physical measurements at selected points in the flow give limited information. However, a clear image of the flow might be worth a thousand individual measurements. Further, such images might point the way towards the most appropriate points at which to obtain critical data on device performance. OST has developed the capability to produce and capture such clear images using "flow visualization" techniques. These techniques will provide greater knowledge for the assessment of high-profile devices, such as heart valves, left ventricular assist devices, cardiopulmonary bypass blood pumps, and injection devices for anesthesia.
As part of the efforts to understand anesthetic injections into the spinal canal, OST scientists constructed a specialized transparent flow model to view the exit jet created by four different clinically available spinal needles (figure 11). A laser sheet approximately 1 mm in thickness was used to illuminate 20-micron-diameter fluorescent particles as they were pumped through the spinal needles at controlled rates. The particles were then viewed using magnifying video equipment. The work was conducted to corroborate the results obtained by computer simulations of ejection of anesthetic from spinal needles. The computer simulations are intended to address design issues, such as how the shape and size of the needle port hole affect anesthetic delivery. Agreement between the computational and flow visualization models is demonstrated in figure 12 for both the angle between the needle and the emerging jet and the degree of spread of the anesthetic jet. Further experiments are planned comparing the flow visualization model and the computational model to clinical data and to results obtained with OST's in vitro spinal model. [ProA, PostMS]
A schematic diagram of the flow loop used to visualize flow ejected from a spinal needle.
A video image of the pattern of fluid ejected from a spinal needle.
The pattern predicted by a computer simulation of the same ejection.
Return to the Table of Contents.
Key words: heart valves, hemolysis, regurgitation
OST is working to determine whether the spatial orientation of small heart valves affects fluid backflow under physiological pulsatile flow conditions. This work may influence the standards for bench-testing heart valves with regards to their spatial orientation during testing procedures. During FY 95, OST scientists evaluated small (17-19 mm) monoleaflet and bileaflet heart valves during backflow by three different techniques: 1) hemolysis testing; (2) in vitro volumetric backflow as a function of spatial orientation; and 3) flow visualization. No significant difference in hemolysis was noted between the valve types. Backflow through the bileaflet valve occurred mainly through the leaflet pivot indentations in the valve housing, whereas backflow through the monoleaflet valves occurred around the circumferential boundary between the leaflet on the minor orifice side of the valve and the valve housing.
Flow visualization experiments help to explain why the amount of backflow in the monoleaflet valves is dependent on the position of the valve occluder during closure relative to the direction in which gravity acts. Laxness of the occluder allows the circumferential gap between the leaflet and the housing to increase under certain spatial orientations. In contrast, backflow through the bileaflet valve was invariant under the different spatial orientations since the leaflets are more constrained and leakage occurred within the leaflet pivot points on the valve housing. A summary of the in vitro backflow tests was sent to ODE and the Office of Surveillance and Biometrics (OSB). [PostMS, ProA]
Return to the Table of Contents.
Key words: heart valves, Doppler, regurgitation, numerical model
OST is exploring methods to quantify color Doppler images of blood leaking through closed heart valves. Excessive leakage, known as regurgitation, can seriously compromise patient health. The technique involves development of software that analyzes color Doppler image frames and uses the digital information and fluid mechanics theory to compute the volume of flow (the critical clinical quantity). Unfortunately, direct experiments with color Doppler are associated with many limitations (e.g., angle dependence, aliasing, gain dependence). Therefore, a numerical model of regurgitant jets (the leaking blood) has been developed to assist in the development of the appropriate analysis software.
A useful theoretical model of jet flow, arising from boundary layer theory, predicts that fluid velocities are dependent upon 1/x, where x is the distance from a "virtual orifice." As a first step, jets were calculated using the boundary layer equations. The virtual orifice method was found to work well, as expected, because the method is based on the boundary layer solution to begin with. An alternate method for mathematically computing the regurgitant jets was needed to fully verify the virtual orifice analysis method.
The alternate method adopted was a numerical technique. Studies of fluid jets were performed using the FIDAP finite element program on the IBM RISC workstation. The computer model was designed to mimic the general tank dimensions of the regurgitation phantom, an in vitro test system built in OST. To compensate for a lack of computer resources, an axisymmetric numerical model was used instead of a three-dimensional rectangular tank. More importantly, the orifice size and shape of the numerical model were designed to match the phantom. Steady flow jets were simulated first to prove the usefulness of this line of inquiry.
Initial studies using a laminar flow numerical model produced jets that were more narrow than those seen in the phantom. Therefore, one of many possible turbulent models was used for subsequent studies. This turbulent model gave results that better matched previous in vitro data. One drawback is that the turbulent model circulation took about five times longer (about 24 hours) on the computer than the laminar model.
A total of 36 numerical studies were performed, using six nominal orifice flow velocities (100, 200, 300, 400, 500, and 600 cm/s) in six different orifice radii. A program was written to interpolate the results into a bitmapped image file comparable to color Doppler images and compatible with the virtual orifice analysis program under development for the last few years. Aliasing was simulated by another program which modified the image so that all velocities above a velocity of 250 cm/s were reduced by an integer multiple of 250 cm/s (the aliasing velocity) until they were equal to or less than 250 cm/s. This is what happens to high velocity information in an ultrasound image.
The analysis program was used to calculate the flow rates from the images. Flow rates through the six orifices varied between approximately 0.1 and 16 L/min. When aliasing was not present, linear regression of the calculated vs "measured" flow rate gave the line y = 1.081x - 0.197 (r2 = 0.999, n = 36). The overestimation may have been due to an artifact of the simulated jet.
When aliasing was introduced into the simulation, the anti-aliasing feature of the analysis program was used, which can detect where most of the aliased velocities are to be found in an iterative fashion. The resultant regression line was found to be y = 0.883x + 0.042 (r2 = 0.989, n = 24). Most likely, the underestimation was because the program does not perfectly reject all aliased velocities. The conclusion, based on these analyses of numerically simulated jets, was that the virtual orifice method for analyzing regurgitant jets works adequately well in steady, turbulent flows, even when aliasing is present. The proactive development of these test methods provides OST with the capability to participate effectively in the development of standards and to improve ODE's guidance to manufacturers of replacement heart valves. [ProA, PreME]
Return to the Table of Contents.
Key words: animal model, replacement heart valve, non-orthotopic, sheep
Mitral valve replacement in juvenile sheep has proven to be a satisfactory animal model for the preclinical evaluation of prosthetic heart valve surgical techniques, hemodynamic performance and valve-related pathology. However, due to high morbidity and mortality associated with aortic valve replacement in sheep, it is not technically feasible to evaluate unstented porcine aortic valve bioprostheses and aortic valve allografts using this animal model.
The use of a non-orthotopic site for the evaluation of unstented heart valve replacements has been suggested in the Heart Valve Guidance as an alternative model for the study of anticalcification treatments in unstented bioprostheses. The purpose of this study was to evaluate the suitability of the Hofnagel technique (implantation of a prosthetic valve in the descending thoracic aorta) with the use of an ePTFE shunt (between the aorta and the left atrium) for the study of tissue valve (xenograft; allograft) hemodynamics and valve-related pathology. The ePTFE shunt serves to increase the pulse pressure and facilitate leaflet closure, otherwise the leaflets would flutter throughout the cardiac cycle.
The use of the model for the long-term (20 weeks) assessment of tissue valve performance and valve-related pathology has been limited by variability in ePTFE shunt patency. If the ePTFE shunt becomes occluded, then adequate leaflet motion and coaptation does not occur. As a consequence of the variability associated with shunt patency, 20-week implantation studies have been discontinued. The evaluation of acute hemodynamic performance is acceptable as well as short-term studies intended to assess valve histology, hemodynamics, and leaflet cell viability. Currently, explanted allograft valves have been studied at the following time periods: fresh - 2,4,8, 10 and 14 days; cyropreserved- 2, 4, 8 and 10 days. Fibroblast cell proliferation has been demonstrated in primary tissue culture at all time periods studied (ie, fresh - 14 days; cyropreserved - 10 days). This observation is in contrast to 20-week explants in which fibroblast proliferation could not be demonstrated. Studies are being planned to extend the implantation time in 7-day increments until shunt occlusion is noted. [PreME]
Return to the Table of Contents.
Key words: spinal needles, nerve damage, in vitro model
Injection of anesthetic agents into the spinal column is usually safe and effective. Permanent nerve damage can occur, however, and this damage can be both device design- and use-related. OST scientists have been collaborating extensively with anesthesiologists at the Naval Medical Command in Bethesda to complete three experiments. The first experiment was a clinical case study combined with OST's modeling of maldistribution with pencil point needles sacrally directed with slow injection rates. Statistically significant differences were measured between needle types and injection rates. From this result, it was concluded that injections should avoid the sacral direction and should be brisk.
The second experiment was a quantitative evaluation of maldistribution and peak dye concentration as function of injection rate for two common pencil point needles. Twenty-nine trials were performed with three repeat injections in the sacral direction for five injection rates. The injections were analyzed for peak and average concentration of dye in the model. The peak concentration fell off exponentially with increasing injection rate, and a turbulence transition point was reached at between 6-8 ml/min. Above the transition to turbulence, extrapolation of the model data to the expected percent lidocaine concentration for a clinical injection did not exceed 1%, remaining clinically nontoxic.
In the third experiment, data was recorded and analyzed for 36 trials of 2 spinal needles employing simulated 2% (both isobaric and hyperbaric) and 5% (hyperbaric) anesthetic injectates. The measured maximum dye concentration was determined for the various conditions. Significantly lower concentrations were found for the 2% hyperbaric when compared to the 5% hyperbaric at injection rates of 4 and 16 ml/min. Comparison of the two 2% injectates showed that isobaric injectates should significantly lower maximum concentrations. These results suggest that diluting 5% lidocaine or using isobaric lidocaine could reduce the chance of neurotoxicity. The results of these experiments were used by OST to advise ODE and CDER scientists in regard to the labeling of anesthetic drugs packaged for use with spinal needles. [PostMS]
Return to the Table of Contents.
Key words: condoms, numerical model, virus transport
Unanswered experimental questions concerning leakage of virus and virus-sized
particles through condoms have led OST to develop a numerical model capable
of investigating physical effects on the microscopic level. New information
arising from this computational approach should have impact on the test
methods which currently provide critical information for qualifying new
materials and new devices as effective barriers. For example, the model
may resolve issues of appropriate size, shape, and charge of test viruses
used to simulate HIV and hepatitis. A first-generation model for
analyzing virus transport in cylindrical pores has been completed.
The first step was to analyze the various components of the transport equation governing virus transport through membranes. One such component is the interaction force existing between boundaries and nearby particles. The total interaction force is comprised of numerous types of forces, depending on the material composition of the particle and the boundary. The two most important forces are the van der Waals and Coulomb electrical forces. The appropriate form of the van der Waals force existing between a latex membrane and a virus was developed. Most other components in the transport equation are hydrodynamic in nature and include the hydrodynamic force and torque on the virus, as well as the virus diffusivity. All of these hydrodynamic quantities are functions of location within the membrane pore and of the virus size and shape. A finite-element approach was then developed to compute the flow around potentially odd-shaped particles in nonstandard shaped pores.
While the general virus-transport model has been under development, a simplified model was concurrently developed with the hope of obtaining some feel for the level of sophistication required to obtain reasonable agreement with experiment. The simplified model assumes spherical particles and a pore of circular cross section. The only interaction force considered is the van der Waals force. The transit time in the pore is also assumed long compared to the time to diffuse across the pore. Analysis is currently underway to determine whether any of OST's earlier experiments meet the qualifying assumptions for the simplified model. If not, it may be possible to design a new set of tests with which the simplified transport model may be compared.
A second immediate goal of the project is to translate the knowledge gained about the various components of the transport equation into computer code. Predictions of virus transport under more general conditions can then be made. [ProA]
Return to the Table of Contents.
Key words: intraocular pressure, drainage device, test methods
Implantable drainage tubes used to relieve otherwise uncontrolled high intraocular pressures are known as glaucoma shunts. Some of these devices include a pressure-sensitive valve to limit drainage, thereby preventing hypotony during the immediate post-implant time. ODE is planning to call for data on these Class III preamendments devices in order to determine whether to call for PMAs or to reclassify the devices. OST has performed bench testing on a variety of glaucoma shunts to obtain engineering information. These tests were carried out to assist ODE in its efforts to establish the appropriate regulatory path for future marketing of these devices.
Two types of tests were performed. In the first, the shunt was subjected to the gravitational pressure of 70 mmHg generated by a column of saline. The fluid drained through the shunt, reducing the column height until the pressure reduced to zero or until the valve closed. The resulting pressure vs time curve determined a flow vs pressure curve for each device. The second test used a syringe pump to simulate the production of aqueous fluid. The pressure vs time data provided some insight into the valve's ability to regulate pressure.
Experiments indicated that the flow resistance of the shunt (the slope of the pressure/flow curve) can be calculated by analyzing a drop in pressure from about 65 mmHg to a steady pressure (either the valve closing pressure or a near zero pressure). A drop from 40 mmHg to 20 mmHg will also yield sufficiently accurate information about the resistance of the shunt and do so in much shorter time. Experiments were conducted on the gravitational outflow for two different shunts and the corresponding pressure flow curves. These particular devices performed consistently, both in terms of the flow rate at any given pressure and in the closing pressure of the valve.
The advantage of this type of bench testing is that it provides a scientifically valid comparison between implants of different design and manufacture. A report to ODE concerning the testing of 20 shunts is complete. [PreME, Stds]
Continue to the Infection Control section.
Return to the Table of Contents.