Supervisory Review of NDA 20-785

Janet Woodcock, MD

July 7, 1998

I. INTRODUCTION

This is a supervisory medical review of NDA 20-785, a marketing application submitted by Celgene Corporation for thalidomide (trade name Thalomid) for the treatment of erythema nodosum leprosum (ENL), a designated orphan indication. The application was originally submitted on 12/24/96.

The drug thalidomide has a long FDA history. Thalidomide was approved in Europe in 1957. A US marketing application was reviewed by the Agency in 1960 and was not approved because of concerns about neuropathy associated with use of the drug. While the Agency was awaiting answers to these concerns, the link between thalidomide use and an epidemic of congenital malformations (phocomelia and other organ defects) occurring in Europe was recognized and the drug was withdrawn from marketing. The tragedy played a part in the debate around the 1962 amendments to the Federal Food, Drug, and Cosmetic Act that resulted in specific effectiveness requirements for drugs.

Celgene is seeking approval to market thalidomide for this indication.

II. PURPOSE OF SUPERVISORY REVIEW

III. EVALUATION OF THE EFFECTIVENESS DATA IN ENL

A. Data Sources

The Celgene NDA is primarily a literature submission for the purposes of effectiveness data. It contains additional primary information for one of the published trials (Hastings, et al) that was obtained during an attempt to reconstruct the records for the trial from the patients' hospital notes from many decades ago (see III.B.3.b). In addition, effectiveness data were obtained from medical records of treatment made available under IND 11,359 held by the Public Health Service and submitted under the NDA (see III.C).

B. Published Literature

1. Two double-blind controlled trials evaluated the effectiveness of 100 mg thalidomide four times daily in controlling the cutaneous manifestations of ENL. They are discussed in the following:

a. Iyer GCS, Languillon J, Ramanujam K, Tarabini-Castellani G, Terencio De Las Aguas J, Bechelli LM, Uemura K, Dominguez V and Sundaresan T. "WHO Coordinated Short-Term Double-Blind Trial with Thalidomide in the Treatment of Acute Lepra Reactions in Male Lepromatous Patients" Bull World Health Org 1971, 45, 719.

Primary endpoint: Although the protocol was reported to have a formal statistical design, the primary endpoint was not clearly reported in the publication. The presence of fever, cutaneous lesions, and other manifestations of the lepra reaction were evaluated at the beginning of treatment, 48 and 96 hours, and on day 8.

Other evaluations: Limited laboratory and clinical safety evaluations were reported.

Results:

92 patients were randomized initially, 57 had second courses of treatment, 39 had third courses, and 26 completed a fourth course for a total of 214 evaluations.

All patients were over age 14, a few were over age 55, the majority were young to middle aged males.

Patients: Patient characteristics were reported for the initial evaluation for each of the 214 courses of treatment. Not surprisingly, on account of the crossover design, patients were quite comparable with respect to lepra symptoms, number and duration of reactions prior to the trial, and previous therapy. Although skin lesions were required on entry, 5% of thalidomide treated patients and 4% of aspirin treated patients did not have skin lesions at the initial evaluation: presumably these were initial evaluations from courses of treatment subsequent to initial enrollment. Most patients were fairly ill, with fever, skin lesions, pain, anorexia and nerve lesions reported on initial evaluation.

Effect on temperature: Thalidomide was observed to have a marked effect on fever compared to aspirin. For example, in patients with temperatures greater than 38.5 ⁰C, 44 were initially assigned to thalidomide and 24 to aspirin. At 48 hours 5 thalidomide treated patients and 19 aspirin treated patients still had temperatures greater than 38.5 ⁰C, and at day 8 the numbers were 2 and 12 respectively.

Effect on skin lesions: At initial evaluation, 95% of thalidomide and 96% of aspirin patients were reported to have skin lesions. At subsequent evaluations, the following results were observed (table adapted from Table 7, p. 724 of publication):

Compared to baseline (if courses in patients without skin lesions at the start [6 thalidomide and 4 Aspirin]are not counted), 75% of thalidomide courses and 25% of aspirin courses were reported to result in a complete skin response as assessed at day 8. Including improvement, responses were reported to be 90% for thalidomide and 51% for aspirin.

Effect on other lepra manifestations: There were not significant differences in outcomes at 8 days for nerve, eye, testis, or other lesions although there were trends favoring thalidomide in some areas.

Effect on relapse: Throughout the three courses of therapy where relapse could be evaluated, patients treated with thalidomide had a lower probability of subsequent relapse than aspirin-treated patients. For example, after the first course of treatment 48% of thalidomide-treated patients relapsed during the trial, while 69% of aspirin-treated patients relapsed.

Safety evaluation: This short-term trial could only evaluate safety issues related to brief administration of thalidomide. No serious adverse events were reported.

Drops in diastolic blood pressure and pulse rate over that related to aspirin therapy were observed.

Drops in total white blood cell and neutrophil counts were observed.

No significant effect on liver enzymes, renal function tests, urinalysis or sedimentation rate was observed.

No difference in reported drowsiness was noted between the two groups.

Commentary: This was a rigorously planned study conducted at multiple sites around the world under the auspices of the World Health Organization. It evaluated the effect of thalidomide in patients with fairly severe manifestations of ENL, but ENL of an acute rather than chronic nature. The dose of the aspirin active control used (1600 mg/day) would be expected to have some analgesic and antipyretic effects in adults, but probably not maximal, and certainly would not be considered an adequate anti-inflammatory dose. Therefore, the control can be considered to approximate a placebo, and this trial does not represent a comparison to an adequate dose of an antiinflammatory agent. The 22% complete skin response at 8 days on aspirin may represent the spontaneous improvement rate in this population. These patients were untreated with other medications during the duration of their evaluations, so that confounding by other therapies should not have been an issue.

Entry criteria for the trial were not rigidly specified, except that patients should have cutaneous ENL manifestations and not be severely ill. Entry criteria in trials are often designed more narrowly to decrease variability and improve the chances of detecting a drug effect; however, broad entry criteria increase the generalizabilty of the results to a wider patient population and do not necessarily imply a poor design.

Crossover designs can pose problems due to the inability to separate effects of the prior drug on the subsequent course. However, the treatment effect is large and was observed on the first course of thalidomide vs aspirin as well as subsequent courses. Additionally, the dose (and effect) of aspirin was small. Therefore, any effects from the crossover would favor aspirin and diminish the apparent effect of thalidomide.

At the 8-day assessment, little effect was found on systemic manifestations of ENL other than fever. This is not at all surprising given the short course of therapy.

b. Sheskin J and Convit J, "Results of a Double-blind Study of the Influence of Thalidomide on the Lepra Reaction." Int J Lep 1969, 37,(2), 135.

This double-blind, placebo-controlled trial evaluated the effects of a 7-day course of 100 mg thalidomide four times a day in primarily hospitalized patients with chronic (longer than 3 months)ENL.

This was a single-center study conducted in Caracas, Venezuela.

Other evaluations: Clinical evaluation of side effects was performed. Skin biopsies were performed for bacteriologic and histologic evaluation. Laboratory tests were done, but not clearly reported in the publication.

Results:

The results of this trial provide substantiating evidence that thalidomide is effective in the treatment of ENL.

 

2. A single publication reports the results of two double-blind placebo-controlled trials of 100 mg thalidomide three times daily in patients with ENL.

Waters MFR. "An Internally-Controlled Double-blind Trial of Thalidomide in Severe Erythema Nodosum Leprosum" Lepr Rev 1971, 42,26-42.

This paper reports the results of two consecutive, double-blind, placebo-controlled crossover studies in hospitalized patients with severe, chronic, steroid-dependent ENL conducted in Malaysia.

Inclusion criteria: Patients with lepromatous leprosy with severe, histologically proven ENL. Only patients requiring at least 15 mg. prednisolone or 18 international units ACTH daily for control of ENL symptoms were eligible. All patients were taking dapsone.

Study medications: Allocation to study medications was randomized. Dispensing of tablets from prepared envelopes was handled by hospital personnel separate from the assessors. Reportedly nursing staff were not aware that 2 different tablets were being used.

Dosage regimen: Placebo or 100 mg thalidomide three times daily

Concomitant medications: Patients were given 100 mg dapsone twice weekly. Aspirin, paracetamol, or other mild analgesics were given PRN.

Schema:

First trial: This was a 16 week trial with four 4-week treatment periods. All patients were hospitalized at the start of the trial and underwent a 4-week observation period. During the second 4 weeks, patients were randomized to thalidomide or placebo. At the beginning of the third 4-week period, patients were crossed over to the other study medication. During the last 4-week period, patients were observed off study medication.

Second trial: This trial was of identical design, except that the four periods were each 6 weeks long, for a total duration of 24 weeks.

Primary endpoint: The primary endpoint was a comparison of total weekly steroid dosage. Patients were seen 6 days a week, and their steroid dosage titrated to control fever and cutaneous ENL symptoms. Upon worsening, prednisolone dose was usually raised, but this could be supplemented by ACTH if deemed necessary. ACTH units were converted to prednisolone equivalents via a formula. The state of the ENL was also formally assessed at the end of each week using a scoring system.

Other evaluations: Daily temperatures, weekly weight, blood counts every 2 weeks, urinalysis every 4 weeks, complete leprosy evaluation at the end of every 4 or 6 week trial phase.

Results:

16 week trial:

Nine adult male patients were enrolled, ranging from 21-56 yrs of age. All had moderately severe or severe chronic ENL and were receiving a mean dose of 28 mg prednisolone (1 pt was receiving ACTH alone). The duration of ENL was from 9 months to 3^1/2 years with continuous steroid treatment averaging 12 months. All ENL was biopsy proven.

Effect on steroid dosage. During the 4-week run-in period, 6 of 9 patients had relatively stable steroid requirements, 2 had increases of 20-30% and 1 had a doubling of dose.

Five patients were randomized to thalidomide in the second 4-week period. Four of the 5 had major reductions in steroid use by the end of 4 weeks: 1 patient was tapered off steroids, 3 others were reduced from 105 to 6, 140 to 43, and 140 to 48 mg prednisolone per week respectively. At the same time, the clinical condition of the ENL for these patients was scored as improved. One patient did not respond and had a markedly increased steroid dose by the end of the period.

Four patients were randomized to placebo for the same 4 weeks. Three of the 4 required increases in steroid dosage, the fourth remained stable.

Upon crossover in the third 4-week period, the 4 apparent thalidomide responders slowly relapsed and again required increased steroids. Three of the 4 crossed-over placebo patients had major reductions in steroid dosage on thalidomide accompanied by improvements in their ENL scores.

In the final 4 weeks, the 3 responders from the third period slowly relapsed toward their previous steroid requirements with ENL scores deteriorating.

Overall, the 7 patients responding to thalidomide had between 61-100% reductions in steroid requirements between the week preceding thalidomide and the 4 week of therapy.

24 week trial:

Eight patients were studied in this trial, 7 of the 8 were also in the prior trial.

Effect on steroid dosage: In the initial 6-week observation period, no patient had greater than a 25% fluctuation in steroid dosage.

In the second 6-week period, 3 patients were randomized to receive thalidomide. Two of 3 were able to stop steroids, the other had a 2/3 reduction in steroid dose. No major changes occurred for 4 of the 5 patients on placebo, 1 placebo patient had a significant (approx 50%)reduction in dosage.

In the third 6-week period, patients were crossed over. One of the thalidomide-treated patients relapsed, while the 2 who had stopped steroids remained stable. All 5 thalidomide-treated patients (crossed over from placebo) had major reductions in steroid dosage (greater than 60% reduction); 1 patient was able to stop steroids.

In the fourth 6-week period, 4 of the 5 thalidomide responders from the third period relapsed.

Other observations: In thalidomide-treated patients, decreases in white blood counts were observed (most patients had leukocytosis initially), 4 patients had rash and eosinophilia which was treated with antihistamines in 2 cases and recurred on rechallenge in these 2 cases but did not result in stopping drug. Two patients reported sleepiness in the 24 week trial only.

3. Other Controlled Trials in the Literature.

a. Pearson JMH, Vedagiri M. "Treatment of Moderately Severe Erythema Nodosum Leprosum with Thalidomide--A Double-blind Controlled Trial" Lepr Rev 1969, 40, 111-116.

This double-blind, placebo-controlled trial was conducted in Malaysia. Twelve patients with longstanding ENL were randomly allocated to placebo or thalidomide, then crossed over after 6 weeks to 6 weeks of the other treatment. The primary endpoint was a scoring system devised by the authors that took into account ENL symptoms, fever, WBC, and stibophen requirements. Interpretation of this score is not possible due to the many variables. In addition, prednisolone dose was allowed to vary in patients taking it and patients were allowed paracetamol.

The results show that patients taking thalidomide had a better clinical score than patients on placebo, despite taking less stibophen, paracetamol, and prednisolone. However, the degree of improvement of the patients is not discernable from the published report.

b. Hastings RC, Trautman JR, Enna CD, Jacobson RR. "Thalidomide in the treatment of erythema nodosum leprosum. With a note on selected laboratory abnormalities in erythema nodosum leprosum." Clin Pharm Ther 1970, 11,481-487.

This report summarizes the results of therapeutic 4-day trials of thalidomide 100 mg four times daily or placebo in hospitalized patients with chronic ENL at the US Public Health Service Hospital in Carville, Louisiana. Forty-four trials were conducted in 22 patients; 21 of the trials were not double-blinded. Thirty-one of the trials were administered in the context of acute steroid withdrawal as the patients were admitted to the infirmary, withdrawn from steroids, observed for 4 days, and enrolled if they had active disease. An attempt was made to reconstruct the records for this trial from the hospital notes since all study records including the randomization schedule had been lost. The FDA reviewers attempting to reconstruct this trial had difficulty doing so, and the context (short-term treatment after steroid withdrawal in the majority of cases) is less germane than that from other studies.

4. Unblinded Studies Reported in the Literature

The sponsor submitted additional unblinded studies from the literature in support of thalidomide for the treatment of ENL. Twenty-eight of these studies evaluated, in an open label fashion, the effectiveness of thalidomide in treating ENL. Many of the studies enrolled patients with severe steroid-dependent disease who would not be expected to have a high rate of spontaneous remission. The literature reports were written by approximately 26 different groups in 14 different countries. Investigators reported findings of prompt relief of cutaneous symptoms and fever, steroid sparing in relevant patients, and prevention of relapse, with efficacy rates very similar to those reported in the controlled studies.

Although these experiences were not controlled, they generally report a very robust treatment effect occurring in a substantial majority of patients in a short time after initiation of therapy, a combination of events not likely in the course of the natural history of the disease. These independent reports provide additional evidence of the effectiveness of thalidomide in the treatment of ENL.

C. Data from Retrospective Chart Review

Patient records from 102 patients treated under IND 11,359 were evaluated. The overall interpretation of these records is difficult, due to the spontaneous fluctuation of disease, possible confounding effects of other treatments, and the non-prospective nature of the data. However, of the 46 patients not recorded to be on prednisone, high-dose aspirin, or clofazamine, 24 patients had at least 1 "positive" challenge-dechallenge result, with remission of ENL lesions on thalidomide and recurrence upon thalidomide withdrawal. Some patients had as many as 7 challenge-dechallenge episodes in which ENL skin lesions were controlled on thalidomide and recurred after withdrawal. (See Office Director's Review dated Sept. 19, 1997, and Dr. Gao's review of Aug. 7, 1997.) Although these data are retrospective and not concurrently controlled, the multiple challenge-dechallenge-rechallenge observations are highly supportive of both the effectiveness of thalidomide and of its ability to prevent relapse, because of the unlikelihood of such repeated occurrences being observed by chance alone. The observed response rate to thalidomide in this series appeared lower than that reported in the literature; however, the doses used were much lower. Control of ENL symptoms was observed to occur at a dose of 100 mg daily in many of the episodes.

D. Ongoing Study

The sponsor is conducting a dose-comparison trial in the Philippines. This trial evaluates the response of ENL to 100 or 300 mg thalidomide for 7 days, and may provide additional information on the most effective starting dose.

E. Discussion of Efficacy Data

The majority of the effectiveness data supporting this orphan indication for thalidomide are derived from the published medical literature, primarily from studies conducted over 25 years ago. These data are supplemented by the retrospective information from IND 11,359. As recently outlined in FDA's guidance for industry, Providing Clinical Evidence of Effectiveness for Human Drug and Biological Products (May 1998), FDA can, in certain circumstances, rely on published reports alone to support approval of a new product.

Four double-blind, placebo-controlled trials of thalidomide in ENL were discussed above. The trials, conducted by 3 different investigators in 3 different countries, and evaluating different patient populations with respect to disease severity and duration, provided similar conclusions about the effectiveness of thalidomide. The drug was found to have a rapid onset of action, to favorably affect the majority of patients treated, and to have a much larger effect than placebo or active control. While each of these studies provides independent evidence for thalidomide's effectiveness, the consistency of results across studies is also striking. The conclusions are also supported by the results of a large number of open-label trials reported in the literature in which a similar rapid onset and robust effect were observed, often in patients with unremitting disease. Although the reported endpoints (skin lesions, steroid sparing) are more subjective than endpoints such as mortality, the reported effect size is also much larger than that demonstrated for most interventions. As a rheumatologist experienced in the management of patients with severe inflammatory and vasculitic disorders, I find these endpoints appropriate for the disease being treated and adequately objective when large differences in outcomes are being evaluated (e.g., a 60% reduction in steroid dose or an absence of skin lesions). Therefore, I find that the published literature may be relied upon to support the approval of thalidomide for ENL.

Although many studies have reported beneficial effects of thalidomide on fever and other systemic manifestations of ENL, the effectiveness of the drug for systemic symptoms is not well supported, and the indication should be limited to the treatment of cutaneous manifestations.

Data from the retrospective chart review, along with literature reports also support the use of thalidomide to prevent relapse in ENL. Treating physicians should attempt to reduce and then discontinue medication in patients whose disease has remitted.

A number of questions have been raised about the availability of effective alternative treatments for ENL. Corticosteroids have been shown in clinical practice to be useful in controlling ENL. However, the very serious adverse effects of long-term corticosteroid administration are well-documented. Extended high- or moderate-dose steroid therapy is not a desirable alternative. Clofazimine is an FDA approved anti-leprosy drug that is additionally approved for the treatment of ENL. Clofazimine was approved for the ENL indication based on the results of a retrospective, historically controlled study of patients with ENL treated at the National Hansen's Disease Center in Carville, Louisiana, supported by controlled trials reported in the literature. (Medical Officer's Review, April 16, 1985.) Particularly since it has a slow onset of effect in ENL, estimating the magnitude of clofazamine's treatment effect from results of an historically controlled trial is difficult. In addition, effective blinding of trials using clofazamine is highly challenging, since non-dark-skinned patients develop red skin pigmentation within about 2 weeks of starting therapy. Because of these factors, while it is clear that clofazamine is effective in ENL, the magnitude and speed of onset of the effect is not precisely known. The side effect of skin pigmentation is unacceptable to many ambulatory patients (suicides have been reported), and the drug also causes an enteritis that can be fatal and was reported in a number of ENL patients in trials. Therefore, clofazamine, while an effective alternative, also has drawbacks and cannot be considered an ideal alternative for ENL. Currently available treatments for ENL may entail, as does thalidomide, serious or unacceptable side effects for some patients.

IV. BRIDGING DATA FROM EFFICACY TRIALS AND BASIS FOR DOSE RECOMMENDATIONS

An ongoing study in the Philippines compares the initial response of ENL to 100 or 300 mg thalidomide. This should provide additional information on whether there are any benefits in starting at the higher end of the dosing recommendations.

The adequacy of safety information is considered separately in the following section.

V. SAFETY EVALUATION

The overall adverse reaction profile of thalidomide is well understood from decades of clinical use. More frequent or severe adverse events include the following:

A. Acute effects

Drowsiness is an expected event in a proportion of patients taking this drug with known sedative properties. This can be somewhat mitigated by dosing at bedtime. Possibly related and fairly frequent effects include tremor, lowered blood pressure, pulse rate, and dizziness, or orthostatic hypotension.

B. Teratogenesis

This effect is well known, has been reported to occur with a very low single dose (50 mg), and occurs at a time during gestation when many women are not aware of pregnancy. This effect raises the greatest concerns about availability of thalidomide.

C. Neuropathy

Peripheral sensory neuropathy is a known complication of thalidomide therapy, and probably the most significant risk to patients consuming the drug. Few reports have occurred in ENL patients compared to those with other conditions. Whether this is due to failure to manifest symptoms, failure to recognize the complication, or lack of susceptibility of these patients is not known. The fact that this complication may be related to cumulative dose makes the recommendations for tapering and discontinuation quite important. The potential for neuropathy and recommendations for early detection are laid out in the warnings section of the label and also in the patient information.

D. Allergic drug reactions

Rashes, eosinophilia, urticaria, and related reactions occur with some frequency in thalidomide-treated patients. These problems resolve with discontinuation, but must be distinguished from the underlying disease being treated.

E. Hematologic effects

Lowered white blood counts, including neutropenia, have been reported in patients taking thalidomide, particularly patients with underlying disorders that may affect the hematologic system (e.g., HIV infection). Severe hematologic reactions are not a serious concern in ENL patients who do not have other predisposing diseases; therefore, no specific monitoring schedule is recommended in the label.

F. Viral load in HIV-infected individuals

In the course of studying the effects of thalidomide on HIV-related illnesses, it was observed that viral load increased during drug administration. Further study is needed to evaluate this observation, and caution should be used in any HIV-positive individual.

The adverse event profile discussed above was generated from clinical experience over decades using a wide variety of manufacturers' thalidomide formulations. One issue that has arisen is how this safety information may be extrapolated to recommendations about Celgene's thalidomide. Since the recommended use of thalidomide in ENL involves titration to a desired effectiveness result, the principal question is whether the Celgene product has a different acute safety profile compared to previously used products. The Celgene product has been administered to patients in both controlled trials and open label studies, generally at doses from 100-400 mg per day. The ongoing Philippine trial in ENL has enrolled 19 patients. This study involves 7-day administration of 100 or 300 mg thalidomide/day, followed by a taper over several weeks. The reported adverse event profile to date from this study is what would be expected from previous clinical experience with the drug (e.g., primarily somnolence, dizziness, rash). In addition, a controlled clinical trial of Celgene thalidomide was conducted in patients with HIV infection and wasting. Doses of 100 or 200 mg were compared to placebo. The incidence of somnolence reported was 36% in the 100 mg dose, 38% in the 200 mg dose, and 11% on placebo. The approximately 25% excess of somnolence in the treated arms is consistent with other reports of drowsiness after thalidomide administration.

Other than teratogenesis, the most serious side effect of thalidomide administration in this population is peripheral neuropathy. Since this has been reported infrequently in ENL patients using any formulation of the drug, the dose-dependency, time course, and rate of progression is not well described with any formulation. The recommendations in the label, which involve patient and physician vigilance towards the development of early symptoms of sensory neuropathy, are appropriate.

The overall safety experience with thalidomide is exceedingly large in proportion to the intended ENL population (estimated at approximately 100 new cases per annum in the US). This situation is quite unusual for a new molecular entity, where the target population frequently may be 100- or 1000- fold larger than the number of individuals studied premarketing, a circumstance that introduces additional risk of the occurrence of rare, not-yet-detected serious adverse reactions. While not all thalidomide patients were studied in clinical trials, the prescribing circumstances closely resemble the intended use conditions of the product, a real-life scenario that is likely to elicit the fullest range of adverse experiences that will be encountered during marketing.

The actual safety experience with Celgene's thalidomide, although quite small, represents a much larger ratio of the number of premarket patients studied to the number of postmarket patients than is expected for most drugs, due to the fact that only a very small number of US patients have this orphan-designated condition.

Concern has been raised about the safety issues involved with off-label use of any approved thalidomide. Physicians have been using investigational thalidomide for a variety of very serious or life-threatening diseases under the "single-patient" and "open protocol" INDs set up by the FDA in cooperation with various manufacturers for this purpose. Any investigational use or off label use of a drug represents greater risk, because less information about safety (and effectiveness) for the indication is available. Approval of thalidomide for ENL would not mean that adequate safety information is available for other indications. Safety information in indications other than those for which marketing approval is sought is not a requirement for product approval. Due to the paucity of information on other uses, the known teratogenic effects of thalidomide, and the potential for peripheral neuropathy, the decision to use thalidomide for an off label use should not be taken lightly by physician or patient, and prudence dictates that such use occur only in very serious or life-threatening illnesses without satisfactory alternatives. The extraordinary restrictions on distribution of this drug, combined with the extensive educational efforts, should accomplish this objective.

VI. RESTRICTED DISTRIBUTION UNDER 21 CFR 314.520

A. Restriction to prescribers and pharmacies registered with the System for Thalidomide Education and Prescribing Safety (S.T.E.P.S.) program (this program includes educational materials).

B. Requirement for patient informed consent, and completion of educational materials including video.

C. Mandatory 100% patient registry and monthly survey completion.

D. Agreement by patient to comply with provisions of program, including not sharing medications, or donating blood or sperm.

E. For women with childbearing potential, agreement to use two methods of birth control, and mandatory monthly or biweekly pregnancy testing.

F. For men taking thalidomide, agreement to use barrier contraception when sexually active with a woman of childbearing potential (because the woman could be pregnant.)

G. Restriction on prescribing or dispensing more than a 28 day supply of drug.

All elements of this program have been extensively reviewed by a variety of FDA experts and consultants. The input of staff of the Centers for Disease Control and Prevention has been very helpful in the evaluation.

The current restrictions strike a balance between the need to prevent fetal exposure to the drug and the need to make the drug available without extraordinary burdens on patients and prescribers. The restrictions create a sufficient burden to deter any prescribing that is not carefully considered; however, they should not pose insuperable obstacles to health professionals and patients. The 100% registry and opportunity for follow-up allow a second tier of safeguards beyond that created by the restrictions at the prescribing level.

This represents the first time that the restricted distribution provisions under 21 CFR 314.520 have been invoked. These regulations were specifically directed to products with safety issues that could not be addressed under the conditions of ordinary approval.

The restricted distribution program for thalidomide is specifically designed to ensure that no fetal exposure to the drug occurs. Any suspected fetal exposure will be considered a serious, unexpected adverse reaction that should be reported in an expedited fashion to the FDA.

A response team at FDA has been formed to oversee postmarketing adverse event reporting and rapidly investigate any report of suspected fetal exposure. Any such verified event would represent a failure of the system and would mandate additional actions or restrictions on the drug depending on the circumstances of the failure.

This novel approach of restriction on distribution will require careful oversight to ensure that it is functioning as designed. FDA will provide a third tier of safeguards by directly inspecting the various elements of the program within the first quarter of operations, and periodically thereafter.

The provisions of 21 CFR 314.530 provide for withdrawal of any drug approved under 314.520 if the restrictions on distribution prove inadequate to ensure the safe use of the product. This represents the final tier of safety assurance under these regulations.

VII. CONCLUSIONS

The Celgene application meets the standard for demonstrating effectiveness of the drug for the intended use. The restricted distribution program under 21 CFR 314.520 enables the product to be used safely for the proposed indication of ENL. Continuing vigilance will be required by both the sponsor and the FDA to ensure that this product is used safely when it is commercially available in the US.

Janet Woodcock, M.D.

Director

Center for Drug Evaluation and Research

cc:

Orig NDA 20-785

HFD-001/Woodcock

HFD-002/Lumpkin

HFD-540/Div File

HFD-105/Walling