Antidepressant Study Stirs Controversy

Fran Lowry

October 22, 2012

A recent meta-analysis concluding that 2 antidepressants neither increased nor decreased suicidal thoughts and behavior in children is flawed, Harvard researchers say. However, the authors of the study stand by their article and its conclusion.

The study, led by Robert Gibbons, PhD, of the University of Chicago, in Illinois, was published in the Archives of General Psychiatry and reported by Medscape Medical News.

In a letter to the news editor of Medscape Psychiatry, Matthew J. Miller, MD, MPH, from the Harvard School of Public Health, in Boston, Massachusetts, asserts that the study's conclusion was "misleading."

Far from indicating that the antidepressant-suicide link in youth is absent, the data continue to show what the US Food and Drug Administration (FDA) data show, a 2-fold increase in suicidal thoughts and ideation among children treated with fluoxetine, Dr. Miller writes.

He and colleagues Tyler VanderWeele, PhD, and Sonja Swanson also wrote a letter to the editor of the Archives of General Psychiatry outlining their concerns about the investigators' interpretation of the antidepressant data. The letter was published as a Comment.

Selection Bias?

"Gibbons' study of highly selected placebo-controlled randomized studies of antidepressants (n = 9,185) is framed as a counterpoint to and advance over the much larger meta-analyses conducted by the FDA (n = 99,839) that found a 2-fold increased risk of suicidal thoughts and behaviors among children and young adults randomized to antidepressants, and a decreased risk in older adults," Dr. Miller notes.

But in reanalyzing these data, Dr. Gibbons and colleagues use mediation analyses, which are subject to bias, the group states.

"We know there are lots of analytic techniques that can distort the findings," Dr. Miller told Medscape Medical News in an interview. "Gibbons was very selective in the studies that he chose to evaluate."

According to Dr. Miller, the use of mediation analysis in this instance broke the randomization and introduced bias.

In spite of this, the findings still show that there is a statistically significant harmful direct effect of treatment on suicide risk in youth that is not discussed, he said.

"Even acknowledging that the technique he used could introduce bias and that he didn't attempt to use techniques that have been around, trying to mitigate that potential bias, what he actually found but did not highlight in his paper is exactly what you'd expect from a study this size — namely, that there was a harmful effect of those who were put on antidepressants compared with those put on placebo.

"The finding didn't reach statistical significance. He had a P value of 0.17, as opposed to less than 0.05, but that's because he used very few people. So the direction of the effect did not reach statistical significance but was similar in magnitude and direction to what the FDA data showed — an increased risk of suicide among younger people. It is a statistic that he does not mention or highlight in his paper whatsoever," Dr. Miller said.

No Evidence of Harm

But Dr. Gibbons stands by his interpretation of the data, acknowledging that mediation analysis can be prone to bias but insisting that his reanalysis found no evidence that antidepressants are harmful in children.

"Swanson and colleagues are correct in noting that the mediational effect is not protected completely from bias due to randomization," he told Medscape Medical News. "This protection is limited to confounders of the treatment by mediator relationship. As for potential confounders of the depressive symptom/suicide risk relationship, this study, as well as any randomized trial, cannot completely rule these out. However, we are confident that the strong mediational findings for adults and geriatric patients would not be seriously affected by an unobserved confounder since the correlation between depressive symptoms and suicide risk is nearly 0.5," he said.

"There is considerable debate in the literature regarding the appropriate way of conducting a mediational analysis, and in fact we performed the analysis several ways and obtained similar results," Dr. Gibbons added.

"We didn't find any effect for the kids. It wasn't increasing or decreasing, so you can't even do a mediation analysis. We did it for completeness and because the reviewers wanted it in the paper, but you can't even really interpret it because there's nothing to mediate. There wasn't an effect," he added.

He noted that Table 2 of the article shows the observed proportions of patients with suicidal thoughts and behavior over time, stratified by age and drug group.

"For youth, the suicidal thoughts and behavior rates are similar at baseline and week 1, lower on treatment during weeks 2 to 5, higher on treatment during weeks 6 to 8, and lower on treatment thereafter. It is hard to imagine how these data justify the statement by Miller and colleagues that 'putting aside the mediation youth the effect appears harmful.' "

"Rather, these data seem quite consistent with the null hypothesis of no effect, given that there are 4 post-baseline weeks in which the suicidal event rate for treatment was lower than placebo and 4 post-baseline weeks in which the suicidal event rate for treatment was greater than placebo," he said.

"In fact, the largest amounts of data available were for the first 6 weeks of treatment, where the predominant effect was protective," Dr. Gibbons said.

"Finally, it is important to note that while FDA did find a harmful effect for suicidal thoughts and behavior in youth that were spontaneously reported to study staff, they did not find a harmful effect for the prospective clinical ratings of suicidal thoughts and behavior," he added.

"Our study combined the prospective ratings with spontaneous reports of suicidal behavior in a more unified analysis, and like FDA's analysis of the prospective data, we also did not find an effect of treatment on suicidal thoughts and behavior in youth."

Dr. Miller has disclosed no relevant financial relationships. Dr. Gibbons reports that he has been an expert witness for the US Department of Justice and Wyeth Pharmaceuticals for 2 court cases that involved antidepressants and suicide.

Arch Gen Psychiatry. Published online February 6, 2012. Abstract