Many (> 40%) women discontinue antidepressants during pregnancy because of concerns about effects on the foetus, based on information from inadequately-controlled studies. The sibling-control study design provides the best control for confounding factors, notably maternal depression. The purpose of this review was to investigate the evidence from sibling-control analyses for adverse outcomes in offspring associated with antidepressant exposure during pregnancy. Fourteen sibling-control studies were identified through searches of PubMed and Embase. Outcomes included preterm birth, small for gestational age, neonatal size, birth defects, autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), behavioural problems, neurodevelopmental deficits, and scholastic attainment. For the majority of these outcomes, no statistically significant associations were found when comparing exposed and unexposed siblings. Single studies reported associations with preterm birth, reduced gestational age, ADHD, anxiety at 36 months, and lower mathematics test scores, which persisted in the sibling-control analyses. However, differences were small and possibly not clinically significant. Moreover, effects of residual confounding could not be excluded. These findings provide evidence that many of the previously reported associations between prenatal antidepressant exposure and adverse outcomes in offspring are no longer statistically significant when exposed offspring are compared with unexposed siblings. The few statistically significant differences in sibling-control analyses were generally small with doubtful clinical significance. Decisions on antidepressant treatment during pregnancy should be made individually, based on evidence from properly controlled studies, not on misleading information based on studies that have not controlled adequately for confounding factors.