**Background:** Maternal postpartum depression affects an estimated 10–19% of women globally and has been linked to childhood undernutrition in low- and middle-income countries (LMICs). However, evidence from LMICs is inconsistent, and few studies have examined how specific depressive symptoms or subscales of screening tools like the Edinburgh Postnatal Depression Scale (EPDS) relate to infant growth. This study aimed to investigate associations between maternal postpartum depressive symptoms, household socioeconomic factors, infant characteristics, and infant physical growth in a low-income South African community.
**Methods:** This cross-sectional analysis used baseline data from a 6-month randomised controlled trial of daily egg supplementation in infants aged 6–9 months, conducted in Jouberton, North West Province, South Africa (February–July 2021). The analytical sample included 428 mother-infant pairs (206 boys, 222 girls) after excluding non-biological mothers, pregnant mothers, and records with missing data. Maternal postpartum depressive symptoms were assessed using the 10-item EPDS (cut-offs ≥10 and ≥13), along with three subscales: anxiety (3 items), depression (7 items), and a PHQ-2-like subscale (2 items). Infant anthropometrics (length, weight) were measured by trained fieldworkers and converted to length-for-age (LAZ) and weight-for-age (WAZ) Z-scores using WHO standards; stunting was defined as LAZ < -2 and underweight as WAZ < -2. Household demographic and socioeconomic data, infant feeding practices, and birth outcomes (birthweight, gestational age, haemoglobin) were collected via structured interviews and clinical records. Statistical methods included latent factor modelling and binomial generalised linear models with logarithmic link, adjusted for maternal age. The Benjamini–Hochberg procedure controlled the false discovery rate.
**Key Results:** The median infant age was 198 days (IQR 187–230); 25.5% were stunted and 10.5% underweight. Overall, 28.5% of mothers had EPDS total ≥10 and 15.2% had EPDS total ≥13. The total EPDS score and its subscales (anxiety, depression, PHQ-2) were not statistically significantly associated with stunting or underweight risk. However, in supplementary analyses of individual EPDS items, crying often was associated with a twofold increased risk of stunting (RR 1.97; 95% CI 1.07–3.63), and feeling scared was associated with a twofold increased risk of underweight (RR 2.03; 95% CI 1.00–4.15). Among infant characteristics, low birthweight (LBW) was strongly associated with stunting (RR 6.19; 95% CI 3.45–11.1) and underweight (RR 5.57; 95% CI 2.84–10.9). Prematurity was associated with stunting (RR 2.98; 95% CI 1.74–5.10) and underweight (RR 4.20; 95% CI 2.17–8.14). Girls had lower risk of stunting (RR 0.58; 95% CI 0.37–0.90) and underweight (RR 0.53; 95% CI 0.29–0.99) compared to boys. Maternal education below grade 10 was associated with a 90% increased risk of underweight (RR 1.90; 95% CI 1.02–3.57). Households with more than one primary school child had an 82% increased risk of stunting (RR 1.82; 95% CI 1.13–2.93). Food insecurity (hunger due to money shortage) was more prevalent among mothers of underweight infants (66.7% vs. 48.0%; P=0.0181). The latent factor model showed that household socioeconomic factors correlated with infant characteristics (anaemia, LBW, prematurity) but were not directly associated with maternal depressive symptoms or infant growth.
**Clinical Implications:** This study found no direct link between overall maternal postpartum depressive symptoms and infant stunting or underweight at 6–9 months in this South African cohort, contrasting with some prior LMIC studies. However, specific symptoms—crying often and feeling scared—were associated with poor growth, suggesting that item-level analysis of the EPDS may reveal clinically meaningful signals masked by total scores. The strong associations of LBW, prematurity, male sex, low maternal education, and household food insecurity with stunting and underweight underscore the multifactorial nature of early childhood undernutrition. These findings support routine screening for postpartum depressive symptoms—particularly in mothers of LBW or preterm infants—and integrated interventions addressing maternal mental health, nutrition, and socioeconomic support. The cross-sectional design limits causal inference, and prospective studies are needed to confirm these exploratory findings.