Subclinical mastitis (SCM) is an important risk factor of postnatal HIV-1 transmission that is still poorly understood. A longitudinal sub-study of the ANRS12174 trial including 270 breastfeeding mothers in Lusaka, Zambia measured sodium (Na + ) and potassium (K + ) in archived paired breast milk samples collected at week 14, 26 and 38 postpartum to determine cumulative incidence of SCM and the effects of recurrent severe SCM on HIV-1 shedding in breast milk. A nested retrospective cohort study including 112 mothers was also done to determine longitudinal effects of SCM on four pro-inflammatory cytokines; IL6, IL8, IP10 and RANTES. The cumulative incidence for any SCM (Na + /K + ratio > 0.6) and severe SCM (Na + /K + ratio > 1) were 58.6% (95%CI: 52.7 – 64.5) and 27.8% (95%CI: 22.5 – 33.1), respectively. In majority of affected mothers (51.4%) severe SCM was recurrent. Both breasts were involved in 11.1%, 33.3% and 70% of the mothers with a single episode, 2 and 3 episodes respectively. In affected breasts, an episode of severe SCM resulted in steep upregulation of the four cytokines considered (IL8, IP10, RANTES and IL6) compared to: before and after the episode; contralateral unaffected breasts; and SCM negative control mothers. Recurrent severe SCM significantly increased the odds of shedding cell-free HIV-1 in breast milk (OR: 5.2; 95%CI: 1.7 – 15.6) whereas single episode of severe SCM did not (OR: 1.8; 95%CI: 0.8 – 4.2). A Na + /K + ratio > 1 indicative of severe SCM is an excellent indicator of breast inflammation characterized by a steep, localized and temporal upregulation of several pro-inflammatory cytokines that favor HIV-1 shedding in mature breast milk and may facilitate postnatal HIV-1 transmission through breastfeeding.