In a well-researched, recent post, Patrick O’Leary addresses the FDA’s efficacy requirements as applied to an old drug, Acthar (corticotropin), that was first approved in 1952 and granted an orphan designation in 2010 for the treatment of infantile spasms. The initial approval therefore occurred before the Drug Amendments of 1962, which instituted a “new” statutory requirement of efficacy (more on this below). O’Leary points out that Acthar’s “grandfather” status does not entirely exempt it from the FDA’s efficacy requirements, and that the drug did survive an efficacy evaluation under the DESI program. But how effective is Acthar?
Neither O’Leary nor the New York Times article on which his post is based dig very far into the clinical trial data accepted by the FDA as supporting the efficacy of the drug as a treatment for infantile spasms, and I was curious to know what the evidence says about Acthar in this regard. Clinical trial data is presented—or perhaps more accurately, “buried”—in Section 14 of a drug’s FDA-approved label; in the case of “H.P. Acthar Gel” (NDA 022432), that label can be found here. What does the clinical trial data reveal? The section is brief, just half a page, and notes that of “[t]hirteen of 15 patients (86.7%) responded to Acthar Gel as compared to 4 of 14 patients (28.6%) given prednisone (p<0.002).” Nonresponders were then given the other treatment, with the following results. “Seven of 8 patients (87.5%) responded to H.P Acthar Gel after not responding to prednisone,” while “[o]ne of the 2 patients (50%) responded to the prednisone treatment after not responding to Acthar.” As the p-value (0.002) indicates, the first figures, at least, are statistically significant. These figures were also better than I expected: 86.7% efficacy with Acthar does seem much better than 28.6% efficacy with prednisone.
One should not stop the analysis here, however. Casting some doubt on the robustness of the data just presented is the extremely small trial size of only 29 patients and fairly short duration (4 weeks). The FDA’s statistical review of Acthar expressed concern that “with such a small sample size, the study sample may not be a good representation of the intended pediatric population. Therefore the efficacy data to draw conclusions are limited.” Moreover, the statistical review noted that the primary endpoint was post-hoc rather than “prospectively defined.” It is difficult, of course, to conduct a large trial for an orphan drug indication, so its size must be understood in this context. Yet according to Questcor’s chief executive, Acthar is used to treat about 800 patients per year, and since infantile spasms usually affect those under age 1, this means several thousand patients experience the condition in just a few years. It must be wondered whether more data could have been gathered from these patients.
In fact, more data has been gathered, and that data provides another reason for physicians to be judicious when wielding their prescription pads. According to separate data from Dr. Hussain of Mattel Children’s Hospital, as reported by the New York Times, 18 of 30 (60%) babies with infantile spasm were successfully treated with prednisone. This figure, 60%, is more than twice as high as the 28.6% reported in the FDA labeling. A possible explanation, suggested by Dr. Hussain, is that the trial reported in the FDA labeling used a dose of prednisone that was too low. Hussein also reports that of the 12 babies in his study that did not respond to prednisone and were switched to Acthar, 5 were successfully treated (41%). This figure (41%) is less than half of the corresponding figure that was reported in the FDA labeling (87.5%, see above). In summary:
|FDA Labeling||Dr. Hussain|
|Prednisone is successful in __% of babies||
|Acthar helps ___% of babies who did not respond to prednisone||
The conflicting figures are, to say the least, unsettling. While there is always some uncertainty in medicine, one would hope that the statistics of an approved drug would be somewhat more robust and reproducible. Medicine is supposed to be about science, not faith, especially when it costs $125,000 per patient.
Moreover, the DESI study referred to above was less than a ringing endorsement of the drug’s efficacy overall. In an August 6, 1971 notice (36 Fed. Reg. 14509), the FDA described corticotropin (the active ingredient in Acthar) as “probably effective” in only five conditions or circumstances, none of which was “infantile spasm.” In the same notice, the FDA found the drug ineffective (“lacking substantial evidence of effectiveness”) for conditions so numerous that they will not be reproduced here. A 1977 notice (42 Fed. Reg. 11891) updated the 1971 notice, concluding that corticotropin (Acthar) was now found to “lack substantial evidence of effectiveness for the indications evaluated as probably and possibly effective in the August 6, 1971 notice.” The 1977 notice did conclude that corticotropin was effective for 12 categories of disorders, not including “infantile spasms,” but required the statement that: “[Corticotropin] has limited therapeutic value in those conditions responsive to corticosteroid therapy; in such case, corticosteroid therapy is considered to be the treatment of choice.”
More evidence will likely emerge to clarify the efficacy of Acthar in treating infantile spasm, and in the years after this evidence does emerge, doctors will gradually adjust their prescribing practices accordingly if necessary. By the time that happens, however, Acthar’s orphan drug exclusivity will have been mostly, if not completely, used up. That exclusivity expires in 2017, less than 5 years from now.
In the end, the question is whether Acthar will continue to be perceived by some as a superior drug that deserves a price 62,400% (sixty-two thousand four hundred percent) higher than prednisone, or whether it will be shown to be equal to prednisone or even not as good as it. One thing is certain: one should not take Acthar’s price as an indication of its efficacy. Many drugs are much less effective than they are promoted to be or commonly believed to be, despite stratospheric price tags. See, for example, my previous post on Tarceva. The question is whether Acthar is one of them.
Note: A final point of clarification that may not have been apparent from either O’Leary’s post or the New York Times article is that the FDA has always considered efficacy when evaluating drugs, even prior to 1962 (See, e.g., United States v. Rutherford, 442 U.S. 544, 556 (1979) (“[E]ven before the 1962 Amendments . . . the FDA considered effectiveness when reviewing the safety of drugs . . . .”); Anita Bernstein & Joseph Bernstein, An Information Prescription for Drug Regulation, 54 Buff. L. Rev. 569, 586 n.55 (2006); Note, The Drug Amendments of 1962, 38 N.Y.U. L. Rev. 1082, 1087 (1963)).