The abrupt discontinuation of treatment for depression with tricyclic antidepressants has been associated with a withdrawal syndrome attributed to rebound effects on neurochemicals. This withdrawal syndrome is usually milder following discontinuation of selective serotonin reuptake inhibitors (SSRIs). Nevertheless, several case reports document withdrawal symptoms such as convulsions, irritability, tremor, and abnormal crying in infants born to mothers who took SSRIs during pregnancy. Sanz and colleagues used the World Health Organization (WHO) database to study any association between SSRI use during pregnancy and a neonatal withdrawal syndrome.

The WHO database contains information on suspected cases of adverse drug reactions from 81 countries and maintains more than 3 million case records. Data are statistically analyzed every three months to identify any developing patterns linking drugs to specific adverse effects. This analysis results in a statistical indicator that adjusts for the rate of reporting to the database, combinations of drugs, and increased frequency of reporting some adverse effects. Any positive value indicates an increased reporting of associations between a specific drug and a specific adverse reaction. Indicators that become statistically significant are carefully reviewed by an expert panel.

Late in 1999, the indicator for paroxetine (Paxil) was found to be 2.58, past the predefined threshold level. This triggered an investigation into neonatal withdrawal syndrome associated with SSRI use. By 2003, researchers had identified 93 cases of neonatal withdrawal associated with SSRI use during pregnancy (see accompanying table) Original case records were analyzed to rule out alternative explanations for the adverse effect, account for combinations of medication use, and confirm the use of the materal SSRI in question. Any cases in which an adjuvant medication could have contributed were classified as doubtful. These included mothers who took antipsychotics, other antidepressants, or antianxiety medications. The strongest association of neonatal withdrawal syndrome and SSRI use was with paroxetine. In 43 of the 51 “certain” cases, paroxetine was the only drug taken by women who were pregnant. The dosage ranged from 10 to 50 mg daily. By 2003, the indicator for the entire SSRI group was calculated to be 2.68. The indicator values for individual drugs were: paroxetine, 4.07; sertraline (Zoloft), 1.20; citalopram (Celexa), 1.92; and fluoxetine (Sarafem), 1.07.

The authors conclude that the risk of neonatal withdrawal syndrome, including convulsions, appears to be increased following maternal use of an SSRI during pregnancy. The preliminary data suggest that two thirds of reported cases of suspected SSRI-induced neonatal withdrawal syndrome were associated with paroxetine. The authors discuss possible mechanisms for the increased association with paroxetine, including more potent inhibition of norepinephrine reuptake, reduced selectivity in receptor blockade, enhanced muscarinic blockade, and metabolism dependent on different cytochrome enzyme systems. The authors caution that paroxetine should be avoided if possible during pregnancy and that physicians should be vigilant for adverse neonatal effects following maternal SSRI use and report their occurrence.