Abstract
The impact of maternal factors on the size of HIV-exposed-uninfected (HEU) infants and breast milk composition is poorly understood. Anthropometry, bio-electrical impedance, haemoglobin and HIV viral load data of women living with HIV (WLWH) and without HIV (WLWOH) were compared and related to their infants' anthropometric Z-scores and breast milk macronutrients 6 weeks and 6 months postnatally. At both time points, WLWH (6-week: n = 83; 6-month: n = 63) had lower reactance (measure of body cell mass) (6-week: p = 0.016; 6-month: p < 0.001), phase angle (PhA) (measure of cell health) (6-week: p = 0.001; 6-month: p = 0.002) and haemoglobin (6-week: p = 0.002; 6-month: p = 0.004) than WLWOH (6-week: n = 90; 6-month: n = 73). HEU infants had lower weight-for-age Z-scores (WAZ) (6-week: p = 0.010; 6-month: p = 0.005). Breast milk composition did not differ between groups. At 6 weeks, HEU infants had lower head circumference-for-age Z-scores (HCAZ) (p = 0.014). Bivariate regression demonstrated maternal HIV predicted lower infant WAZ (ß = -0.442; p = 0.011) and HCAZ (ß = -0.445; p = 0.014). Maternal body mass index (BMI) and mid-upper arm circumference were positively associated with breast milk protein content (ß = 0.018; p = 0.014 and ß = 0.025; p = 0.002, respectively). At 6 months (bivariate regression) maternal HIV predicted lower infant WAZ (ß = -0.609; p = 0.005) and length-for-age Z-scores (ß = -0.741; p = 0.018). Higher maternal BMI and PhA were associated with higher infant WAZ (ß = 0.622; p = 0.015 and ß = 0.055; p = 0.017, respectively). On multivariable analysis, maternal HIV remained a predictor of lower WAZ (ß = -0.568; p = 0.024). In conclusion, maternal HIV infection and phenotype predict the size of infants and breast milk composition up to 6 months postnatally.