Paul: You are right that the results did indeed indicate that combining imipramine with CBT produced a long-term disadvantage, not merely no advantage over CBT alone. On the other hand, the study authors (and audience at the conference) were explicitly cautious not to overinterpret these findings. Also, one cannot necessarily say that any effects of adding imipramine to CBT may be accounted for by placebo effects, because combining placebo with CBT produced quite different outcomes than combining imipramine with CBT. Previous studies have rarely allowed such direct comparisons because they typically did not include both of these conditions. In this study, however, we can directly compare and evaluate the relative effects of placebo vs. imiprimane in combination with CBT. And here we have truly fascinating results: Combining CBT with placebo did not lead to the same disadvantage at follow-up than combining CBT with imipramine (there was absolutely no decline in treatment gains in the combined CBT+placebo group).
One can assume that imipramine does have some unique and real physiological effects (e.g., on mood) whereas placebo obviously does not. When imipramine was eventually dropped, these patients had more problems than patients who had been on placebo (maybe because the CBT+imipramine patients had learned to rely too much on the positive effects of the drug). This is indeed a fascinating finding, and I fully agree that we need to re-evaluate the significance and practical implications of other CBT/drug combination studies. In general, I am now even more skeptical about such combinations as I was in the past (although there are undoubtedly individual cases where such a combination may be warranted or even life-saving).