BACKGROUND: Selective serotonin reuptake inhibitors (SSRIs) have been suspected of cardiac teratogenicity, but reports have been inconsistent. Our aim was to investigate the rate of nonsyndromic congenital heart defects in newborns exposed in utero to SSRIs compared with unexposed controls. METHODS: This prospective study of women who gave birth at our tertiary center from 2000 to 2007 yielded 235 women who reported first-trimester SSRI use during pregnancy. All newborns born during the study period and found to have a persistent cardiac murmur on day 2 or 3 of life were referred for examination by a pediatric cardiologist and by echocardiography. The findings were compared between the newborns who were exposed to SSRIs and those who were not. RESULTS: Nonsyndromic congenital heart defects were identified by echocardiography in 8 of 235 (3.40%) newborns exposed in utero to SSRIs and in 1083 of 67,636 (1.60%) non-exposed newborns. The difference in prevalence between the two groups was significant (relative risk, 2.17; 95% confidence interval, 1.07–4.39). The prevalence rates for paroxetine and fluoxetine exposure were 4.3% and 3.0%, respectively. All cardiac defects in the study group were mild: ventricular septal defect (6), bicuspid aortic valve (1) and right superior vena cava to coronary sinus (1). CONCLUSIONS: Newborns exposed in utero to SSRIs, have a twofold higher risk of mild nonsyndromic heart defects than unexposed infants. The data suggest that women who require SSRI treatment during pregnancy can be reassured that the fetal risk is low and possible cardiac malformations will probably be mild. Late-targeted ultrasound and fetal echocardiography at 22 to 23 weeks' gestation are recommended in this patient group. Birth Defects Research (Part A), 2009. © 2009 Wiley-Liss, Inc.