Antidepressants and pregnancy: complexities of producing evidence-based information

In her commentary, Einarson (1) discusses a study by Nakhai-Pour and colleagues that found an increased rate of spontaneous abortion following antidepressant use in pregnancy (2). Interestingly, this nested case- control study found an increased rate of spontaneous abortion that was very similar in magnitude to that reported by Einarson and colleagues in a 2009 prospective observational study of a large convenience sample (n=1874, half exposed) of women who contacted teratology information centres (3).

Despite the replication of her own study findings, Einarson casts doubt on the association between SSRI antidepressant use and spontaneous abortion, concluding that, “this study cannot make any definitive conclusions as to whether antidepressants increase the risk of spontaneous abortion”, and that “if there is a true increased risk of spontaneous abortion caused by gestational use of antidepressants, it is very small.” (1)

Einarson raises several weaknesses and fails to mention a single strength of Nakhai-Pour’s study, such as the use of a large population- based sample and prospective recording of information on filled prescriptions for cases and controls. The latter eliminates the major problem of recall bias linked to outcome that is inherent in case-control studies. Nakhai-Pour et al. also controlled for a range of confounders likely to affect outcome, such as socio-economic status and co- morbidities, as well as indicators of depression severity and previous depression diagnoses. Pharmaco-epidemiological studies that use administrative data on filled prescriptions cannot perfectly estimate the proportion of filled prescriptions that are taken, but methods such as measuring refill rates can provide important clues. Additionally, the observed dose-response relationship in this study strongly suggests that fill data are reliable.

Einarson raises concerns that observed associations may be due to an expected 5% false positive rate, and claims that over 100 analyses were conducted. The stated aim of the study was to examine the risks of spontaneous abortion associated with antidepressant use in pregnancy, and this can therefore be presumed to be the primary outcome measure. Table 2 reports 1 outcome (crude and adjusted OR), Table 3 another 12 outcomes (again crude and adjusted OR), and Table 4 reports two evaluated dose- response relationships, for paroxetine and venlafaxine. Of these reported 15 comparisons, 13 crude and 7 adjusted ORs were significant at p<.05. Thus even with a 5% type I error the association between antidepressant use and increased risks of spontaneous abortion appears robust.

No study is perfect; nor does any study stand on its own. It is important to examine all research within the body of evidence preceding it: does it add to existing scientific knowledge? Bradford-Hill’s epidemiological criteria for causation provide a useful guide: is the relationship biologically plausible, does it follow the right temporal sequence (cause before effect), has it been replicated in different settings, is a dose-response relationship observed, etc.? The relationship between SSRI exposure and an increased risk of spontaneous abortion meets many of these criteria.

Of greater concern, however, are the implications for the advice provided to pregnant women and the clinicians who care for them. We disagree with Einarson’s characterization of the absolute risk increase of 2.8% (unadjusted) and relative increase of 1.68 (95% CI 1.38-2.06) as “very small” and unconfirmed. Moreover, there is evidence linking SSRIs to cardiac defects, persistent pulmonary hypertension, and withdrawal syndrome in the newborn (4). For women facing depression in pregnancy and considering what treatment to use, it is also important to have accurate and unbiased information on the efficacy of SSRIs. For mild to moderate major depression in particular, the efficacy of SSRIs relative to placebo has been brought into question (5;6). As such, for most depression in pregnancy, we would advocate for psychotherapy and other non- pharmacological alternatives.

Barbara Mintzes, PhD Assistant Professor, Dept Anesthesiology, Pharmacology & Therapeutics Therapeutics Initiative University of British Columbia

Elia Abi-Jaoude, MD Research Fellow Department of Psychiatry University of Toronto

Anne Rochon Ford, Co-ordinator Women and Health Protection

Reference List (1) Einarson A. Antidepressants and pregnancy: complexities of producing evidence-based information. CMAJ 2010 Jul 13;182(10):1017-8.

(2) Nakhai-Pour HR, Broy P, Berard A. Use of antidepressants during pregnancy and the risk of spontaneous abortion. CMAJ 2010 Jul 13;182(10):1031-7.

(3) Einarson A, Choi J, Einarson TR, Koren G. Rates of spontaneous and therapeutic abortions following use of antidepressants in pregnancy: results from a large prospective database. J Obstet Gynaecol Can 2009 May;31(5):452-6

(4) Tuccori M, Testi A, Antonioli L, Fornai M, Montagnani S, Ghisu N, et al. Safety concerns associated with the use of serotonin reuptake inhibitors and other serotonergic/noradrenergic antidepressants during pregnancy: a review. Clin Ther 2009 Jun;31 Pt 1:1426-53

(5) Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 2008 Feb;5(2):e45.

(6) Fournier JC, DeRubeis RJ, Hollon SD, Dimidjian S, Amsterdam JD, Shelton RC, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA 2010 Jan 6;303(1):47-53.

Conflict of Interest:

None declared