Objective The objective of this study was to examine whether exposure to Intimate Partner Violence (IPV) is associated with premature termination of Exclusive Breastfeeding (EB). Per WHO recommendations, this was defined as ceasing breastfeeding or supplementing with other foods or liquids before the child was 6 months old. Method It is a prospective cohort study set in Moshi, Tanzania consisting of 1128 pregnant women with live singleton births. Women were enrolled during pregnancy and followed up with interviews during pregnancy, after birth and 2–3 years postpartum, using structured questionnaires. Emotional, physical and sexual IPV exerted by the current partner was assessed at 34 weeks gestational age with WHO questionnaires. Months of EB was assessed 2–3 years postpartum. Premature termination of EB was defined as less than 6 months of EB. Analyses were made using a logistic regression model adjusted for maternal age, education, HIV-status, alcohol use during pregnancy and parity. Confounding variables were determined using a theoretical framework approach, i.e. a Directed Acyclic Graph model to minimize bias. Results Women who were exposed to IPV had more than 50% higher odds of terminating EB before the child was 6 months old compared to women who were not exposed (aOR = 1.62, 95%CI: 1.27–2.06). Women exposed to all three types of IPV had twice the odds of early termination of EB (aOR = 1.95, 1.12; 3.37). Furthermore, the odds were tripled if exposure happened specifically during the index pregnancy (aOR = 2.93 95%CI: 1.3; 6.6). Stratified analyses showed the most severely affected groups were the mothers older than 30 and those who gave birth to girls. Conclusions The results indicated that exposure to IPV is associated with increased risk of premature termination of EB. The odds increase with multiple types of the IPV, especially when exposed during the index pregnancy.

Background Postnatal depressive symptoms measured by the Edinburgh Postnatal Depression Scale (EPDS) are reported to display measurement variance regarding factor structure and the frequency of specific depressive symptoms. However, postnatal depressive symptoms measured by EPDS have not been compared between women representing three continents. Methods A cross-sectional study including birth cohort samples from Denmark, Vietnam and Tanzania. Women were included during pregnancy at routine care sites. Depressive symptoms were selfreported 40-90 days postpartum using the EPDS. Exploratory and confirmatory factor analyses and generalized additive regression models were performed. Results A total of N=4,516 participated in the study (Denmark N=2,069, Vietnam N=1,278, Tanzania N=1,169). Factor analyses identified three factors (anhedonia, anxiety and depression) that were almost identical in the three study populations. The only variation between countries was that the item 'self-harm' loaded differently. Women from Tanzania and Denmark were more likely to have an EPDS total score above cutoff 12 (12.6% and 6.4%), compared to women from Vietnam (1.9%) (p<0.001). A low level of education was associated with significantly more depressive symptoms after adjusting for country (p<0.001). Limitations EPDS data was collected at a later time point in the Danish sample. Conclusions Postnatal depressive symptoms constitute a three-factor model across cultures including the factors anhedonia, anxiety and depression. The frequency of postnatal depressive symptoms differs between high-, medium-, and low-income countries. However, clinicians should bear in mind that low-educated women worldwide are more likely to experience postnatal depressive symptoms.

21Objective: To examine the association between postpartum depression and child growth in a 22Tanzanian birth cohort. 23Design: Prospective cohort study. 24Setting: Moshi, Tanzania. 25Population: Pregnant women over the age of 18 who sought antenatal care at two health clinics 26 in Moshi, and the children they were pregnant with, were assessed for inclusion in this study. 27Methods: The women were interviewed twice during pregnancy and three times after birth, the 28 final follow-up taking place 2-3 years postpartum. Signs of postpartum depression were assessed 29 approximately 40 days postpartum with the Edinburgh Postnatal Depression Scale (EPDS). 30Main outcome measures: Child growth was assessed with anthropometric measurements at 2-3 31 years of age, and expressed as mean z-scores. 32Results: 1128 mother-child pairs were followed throughout the duration of the study. 12.2% of 33 the mothers showed signs of postpartum depression. Adjusted mean height-for-age z-score (HAZ) 34 was significantly lower at 2-3 years follow-up for children of mothers with postpartum depression, 35 compared to children of mothers without (difference in HAZ: -0.32, 95%CI:-0.49;-0.15). Adjusted 36 mean weight-for-height z-score (WHZ) was significantly increased for the children exposed to 37 postpartum depression (difference in WHZ: 0.21, 95%CI:0.02;0.40), while there was no significant 38 difference in adjusted weight-for-age z-score (WAZ) (difference in WAZ: -0.04, 95%CI:-0.20;0.12). 39 Conclusions:We found that postpartum depressive symptoms predicted decreased linear height 40 in children at 2-3 years of age, and slightly increased weight-for-height. 41