The Ethics of Randomized Placebo-Controlled Trials of Antidepressants With Pregnant Women

Abstract

To determine whether placebo-controlled randomized clinical trials of antidepressant medications with pregnant participants are ethically justified. We searched PUBMED, PsychINFO, and Cochrane databases for articles related to the ethical justification of randomized placebo-controlled trials involving pregnant women and for all research articles for the treatment of depression during pregnancy, using the keys words depression, antidepressive agents, pregnancy, clinical trials, placebo, and ethics. We based our ethical framework on current literature on the ethics of maternal-fetal medicine and research and the requirements of argument-based ethics. We selected studies in which maternal, fetal, and neonatal health outcomes associated with antidepressant exposure were reported. We integrated evidence-based reasoning and argument-based ethics to address four questions to determine whether what is known as normative equipoise (evidence-based reasoning requires uncertainty about which clinical management is superior) exists: 1) Is there evidence of efficacy of antidepressant use during pregnancy? 2) Is there documented causality of serious, far-reaching, and irreversible clinical harm to the fetal or neonatal patient? 3) Is there documented causality of serious, far-reaching, and irreversible harm to depressed pregnant women in a placebo arm? 4) Are there no or only rare documented occurrences of less serious injury to the fetal or neonatal patient? Evidence-based answers to these questions support the judgment that normative equipoise exists with respect to randomized placebo-controlled trials of antidepressants with pregnant participants. Randomized placebo-controlled trials of antidepressant medications during pregnancy are ethically justified. Well-conducted and ethically justified trials of antidepressants should improve the quality of care provided to depressed pregnant women and thus address a major public health problem.