Multiple Sclerosis Research Repository

FDA Approval Package for Copaxone (NDA 20-622/S-015) (Approval Date: 7-2-2001)

1. From FDA Copaxone Label 020622s057lbl.pdf: “Table 4: Study 4 MRI Results: Medians of the Cumulative Number of T1 Gd-Enhancing Lesions: Copaxone (N=119): 11; Placebo (N=120): 17; P-Value: 0.0030
2. *What study is Study 3 in this label? -appears to have been the PRECISE trial
3. It is unsettling and disturbing that the FDA approved the labeling changes to Copaxone:
   1. “The objective raises the concern that human subjects were asked to participate in a medical experiment that was not necessary” (p 31/68)
   2. “Protocol 9003 was not based on a genuine uncertainty that the placebo group would have fewer MRI lesions than the Copaxone-treated group; it was already known that Copaxone reduced relapses, and that relapses are associated with T1 Gd-enhancing lesions” (p 31/68)
   3. “Unfortunately, the protocol shows that there was never any intention to attempt definitive answers to clinical questions, particularly the correlation between MRI and clinical parameters” (p 32/68)
   4. “Even if sponsor can overcome ethical concerns regarding the necessity of Protocol 9003, sponsor next stumbles on the problem of informed consent” (p 32/68)
   5. “A careful reading of some of the informed consent documents (they vary in significant ways among the 29 participating centers)…” (p 32/68)
   6. “But many forms eliminated all reference to the most significant Copaxone trial (9001).  In plain language, the purpose of 9003 was to determine the effect of Copaxone on the MRI; the clinical effect was already proven.  Prospective participants should have been told this.  Because they were not, the required element of informed consent, a statement of purpose,was absent” (p 32/68).  
   7. “The FDA must refuse to grant an NDA application if the applicant has not complied with its regulations concerning informed consent.  No informed consent form used in this study gave a clear explanation of the study’s true purpose…Most were inadequate in describing alternatives…This would have been especially valuable information to Canadian patients…” (p 33/68)
   8. “…I believe the necessity was marginal at best, and sponsor failed to counterbalance this weak necessity by making sure the investigators gave completed and accurate informed consents” (p 33/68)
   9. “In summary, the objective of 9003 was ethically flawed in that it recruited patients to participate in a study that was not necessary.  From a regulatory standpoint, the protocol violated the requirements of informed consent that are required for NDA approval, and must be refused pursuant to 21 CFS Statute 314.125(16), Statute 50.20, Statute 50.25, and Statute 50.27 if conducted under an IND…” (p 34/68)
   10. “Initially patients who had previously participated in an Oral Myelin Trial were excluded from the 9003, but the exclusion was later amended.  The amendment states that the reason for letting patients who participated in the Oral Myelin Trial participate in 9003 is because it ‘showed no clinical efficacy.’  This is a design flaw, in that the possibility of a synergistic effect between oral myelin and Copaxone cannot be excluded.  If any patients were admitted to 9003 who had received oral myelin, the possibility of a confounding variable was introduced.  Data was to which patients, if any, received oral myelin, was not provided in this submission” (p 34/68)
   11. Twenty-seven patients received steroids in violation of the protocol during the study…Other violations included using doses in excess of 1000 mg methyprednisoone IV for 3 days, unusual routes of administration (IM, intrathecal), for indications other than documented acute MS exacerbations” (p 35/68)
   12. “Another important protocol violation involved missed MRI scans” (p 35/68)
   13. “Because the data were not normally distributed, a parametric method such as was used here, was inappropriate…the highly skewed data” (p 36/68)
   14. “Biostatistics also finds that the use of adjusted means, which was done with every endpoint, was also inappropriate…The adjusted means in the proposed labeling were from invalid models and are therefore, not appropriate and misleading…the normal assumption was severely violated…In addition to the violation of the normal assumption, covariates other than those specified in the protocol were added to the model.  Therefore, in this reviewer’s opinion, all adjusted means should be removed from the labeling” (p 36/68)
   15. “There was no significant difference between the drug and placebo groups on this MRI endpoint” (the endpoint was: “proportion of T1 Gd-enhancing lesion-free patients, which was not an endpoint in the original protocol, but sponsor states it is the complementary value of the proportion of patients with Gd-enhancing lesions” (p 37/68)
   16. “…sponsor…used a parametric method inappropriate to the highly skewed data” (p 37/68)
   17. “The total number of new Gd-enhanced lesions in T1-weighted images was not normally distributed and therefore the parametric method used by sponsor was ot appropriate” (p 37/68)
   18. “…the latter approach is unsuited to these data, which are highly skewed” (p. 37/68)
   19. “If one uses the rank-transformed data, however, none of the secondary endpoints are significant” (p 38/68)
   20. “The results of 9003 are fruits of a study tainted by patient consent that was ot properly informed” (p 39/68)
   21. “…this reviewer recommends the sponsor be strongly reminded of the legal requirements for informed consent.  Whether the more drastic action of excluding some patients’ data (e.g., the Canadians’) from the analysis, or even barring any reference to the MRI Protocol 9003 in the labeling has been discussed internally in the Division.  The consensus is that the behavior was not egregious, pervasive, or reckless enough to warrant the more drastic remedies” (p 39/68)
   22. “In any discussion about preventing a sponsor from harvesting the fruits of an ethically-tainted protocol, one needs to keep in mind the potential that the Agency’s action may have unwanted effects on the public.  However, the medical community relies more on medical literature than on labeling as a source of new medical knowledge; labeling issues matter to a sponsor in large part because they form the basis for promotional claims.  Therefore, depriving a sponsor of the fruits of an ethically-flawed study harms the sponsor economically; the medical literature and its readers are unlikely to be harmed” (P 39/68)