This study characterises the extent of desvenlafaxine transfer into milk and provides data on infant exposure to desvenlafaxine via breast milk in ten women with postnatal depression and their breastfed infants. Desvenlafaxine concentration in milk and plasma was measured chromatographically in milk and in maternal and infant plasma collected at steady state. Theoretic and relative infant doses via milk were estimated and the per cent drug in infant versus mother's plasma was calculated. Theoretic infant dose via milk was 85 (53-117) μg kg(-1) day(-1) (mean and 95% confidence interval) and relative infant dose was 6.8% (5.5-8.1%). The ratio of drug in infant/maternal plasma also gave an infant exposure estimate of 4.8% (3.5-6.2%) for all ten infants and 5.3% (4.2-5.7%) in the eight infants who were exclusively breastfed. No adverse effects were seen in the infants. The relative infant dose was similar to that for previous studies using venlafaxine and was supported by a separate exposure measure using the ratio of drug in the infant's plasma relative to that in the mother's plasma. The theoretic infant dose of desvenlafaxine was 41-45% of that for venlafaxine and its metabolite desvenlafaxine in previous studies, reflecting the lower recommended maternal dose for desvenlafaxine. Although our data for desvenlafaxine use in lactation are encouraging and there are supporting data from venlafaxine studies, more patients and their infants need to be studied before the safety of desvenlafaxine as a single therapeutic agent can be fully assessed.