Antidepressants and pregnancy: an embarrassment for JAMA
In retrospect, the issue of antidepressant use in pregnancy is so rife with controversy that the Journal of the American Medical Association (JAMA) should have seen this one coming a mile away.

In February 2006, JAMA published a study on depressive relapse during pregnancy. Conducted and written by major luminaries in the field of perinatal psychiatry, this naturalistic multisite study tracked depressive symptoms and antidepressant use in 201 pregnant women, all of whom had histories of depression but who were euthymic and taking antidepressants at the beginning of the observation period. The results were striking: 68% of women who stop ped taking antidepressants relapsed during the pregnancy, whereas only 26% of those who continued their medication did.⁷

While the article did not ignore the potential perils of prenatal exposure to antidepressants, the authors concluded the abstract with a statement endorsing their use: "Women with histories of depression who are euthymic in the context of ongoing antidepressant therapy should be aware of the association of depressive relapse during pregnancy with antidepressant discontinuation."⁷

At the end of the article, the funding source was reported as the National Institute of Mental Health and 2 of the 13 authors—Dr Newport and Dr Stowe—reported financial ties to drug companies. The following was associated with their names: "None of the other authors reported disclosures."⁷

On February 8, Dr Adam Urato, an obstetrician/gynecologist and faculty member at Tufts University, read the article with interest and was struck with the disclosure. Since he frequently advises patients on antidepressant use in pregnancy, he had followed the research closely and recalled that some of these authors had disclosed industry relationships in other articles. In a letter to JAMA published in the July 12 issue, he catalogued the industry relationships of the authors and pointed out that 9 had been paid by antidepressant drug makers, while only 2 reported disclosures.⁸

By this time, the story was already out—the Wall Street Journal published a front-page story entitled "Financial Ties to Industry Cloud Major Depression Study."⁹ That article not only reported the missing disclosures, but detailed the many controversies about antidepressant use in pregnancy, including recent FDA warnings linking SSRI use with pulmonary hypertension, seizure risk, and—in the case of paroxetine—cardiac malformations. The implication was that antidepressant drug makers had influenced the most prominent researchers in the field to gloss over the risks of using antidepressants in pregnancy and to instead highlight the benefits. The fact that they did not disclose their relationships made them appear even guiltier of commercial bias.

But did conflict of interest really exist?
In a letter of response to JAMA, 4 of the authors explained why they did not submit disclosures. The research, they pointed out, was not funded by industry.¹⁰ Yes, the individual researchers receive money from drug companies, but the published study was not a clinical trial of any particular drug, so no particular company stood to benefit from the results. For these reasons, said the authors, "we concluded that there was no inherent conflict of interest with respect to any of our financial associations with industry."¹⁰

How credible is this response? The core question here is whether the authors’ relationships with industry might have influenced the way they designed, conducted, or interpreted their research. Did the researchers have a financial incentive to modify their research in any way? In order to answer this question, we would have to know whether their research results are likely to lead directly to future payments from the relevant companies.

In fact, there are several possible scenarios in which the authors could financially benefit from the results:

1. They may receive research funding from a range of drug companies wanting to demonstrate the value of their own products in preventing relapse during pregnancy.
2. They may be invited to serve as paid consultants. The role of "consultant" varies, but a major role is to advise companies about what sort of research studies might yield results beneficial to their future marketing efforts.
3. They may be invited by companies to conduct lucrative continuing education activities—such as giving lectures, participating in conferences, and authoring sponsored articles—in order to disseminate their research findings.

In fact, there is evidence that some of the authors of the JAMA article are already deriving additional income in this way. An industry-funded medical education enterprise, the Massachusetts General Hospital Psychiatry Academy, now offers a symposium titled "Psychotropic Drug Use During Pregnancy," which is conducted by some of the JAMA authors. The faculty discuss the results of their study and others and endorse the use of antidepressants in pregnancy (the videotaped presentation can be viewed online11). Thus, it seems clear that in the case of the JAMA article, the following elements of potential COI were met:

• The researchers receive payments from companies that manufacture antidepressants.
• The research findings endorse the value of antidepressant therapy in pregnancy.
• The findings are likely to attract more money from industry to the researchers in the future.

Since the researchers stood to benefit financially from positive results, their judgments may have been, in the words of Thompson, "unduly influenced by a secondary interest (such as financial gain)."¹

Looked at in this way, there appears to be no question that the potential for COI existed and that full financial disclosures should have been made. Does this mean that the authors actually altered the research design to make sure the results favored antidepressants? Of course not. These are all highly ethical scientists who conduct clinical studies to answer important ques tions about psychiatric care. How ever, requiring full financial disclosure would not have invalidated the results of their research. It simply would have provided readers with information that might be relevant in their own evaluation of the research design and conclusions.