AIMS:The aims of the present study were to examine the association between late pregnancy exposure to serotonin reuptake inhibitor (SRI) antidepressants and difficulties in achieving an adequate breast milk supply in women who have given birth to preterm infants, while accounting for the potential impacts of underlying maternal psychiatric illness. METHODS:A retrospective cohort study was carried out of 3024 women delivering liveborn preterm infants (<37 weeks' gestation) between January 2004 and December 2008. The primary outcome was postnatal domperidone use, considered to be a valid proxy for the presence and pharmacological management of low milk supply. Relative risks adjusted for maternal sociodemographic characteristics and comorbidities (aRRs) were calculated for low milk supply, comparing women with late pregnancy exposure to SRI antidepressants (n = 86), women with a psychiatric illness but no antidepressant use (n = 126) and women with neither antenatal exposures (n = 2812). RESULTS:Compared with non-exposed women, nonmedicated psychiatric illness [aRR 1.64; 95% confidence interval (CI) 1.16, 2.30] but not late pregnancy SRI use (aRR 1.00; 95% CI 0.59, 1.70) was associated with an increased risk of domperidone use, indicative of low milk supply. CONCLUSIONS:These findings do not support the previously observed negative impacts of antidepressant use on breastfeeding, instead suggesting that women with an underlying psychiatric illness appear at greatest risk of experiencing low milk supply and could benefit from additional breastfeeding education and support.