Conclusions Inconsistent With Results With Amphetamines and Divalproex

To the Editor: In their article, Russell E. Scheffer, M.D., et al. (1) reported in their conclusions and elsewhere positive summary statements that included the following: “Pediatric patients with bipolar disorder and concurrent ADHD [attention deficit hyperactivity disorder] can be safely and effectively treated with mixed amphetamine salts after their manic symptoms are stabilized with divalproex sodium” (p. 58).

These ambitious claims were made by the authors after noting what they suggested to be these limitations of their brief trial: 1) ineffectively low doses of mixed amphetamine salts, 2) a failure to increase the divalproex doses to assess greater possible response, 3) the small group size, 4) a study protocol limited to a single academic center, and 5) a failure to address long-term outcomes and safety.

The authors’ statements regarding the efficacy/tolerability of mixed amphetamine salts/divalproex might be true, but their repetitively positive published conclusions are not consistent with their evidence. The announced conclusions, likewise, that appeared in the official publication of APA that reiterated this positive news (2) failed to disclose serious research limitations.

The expressed concerns of Dr. Scheffer and colleagues regarding the limitations of their study, while justified, did not address the serious problems in their research design and reporting:

Published positive conclusions of this research effort, funded in part by a grant from the Stanley Medical Research Institute to Dr. Rush, are misleading. The Journal and Psychiatric News must be cautious about favorable generalizations from brief trials whose data from partially unblinded and open-label design do not include results from the research itself that demonstrate serious injuries (e.g., rehospitalizations and induction of mania) as a likely byproduct of the protocol. Representations of preliminary results should not suggest “safety and efficacy” when the data are limited and inconclusive.