Background
The biological mechanisms related to increased risk of stunting in infants/children who are HIV-exposed and uninfected (CHEU) are not fully clear. We assessed whether insulin-like growth factors (IGFs), which are associated with infant growth, are determinants for stunting observed among CHEU.
Methods
A random sample of 213 out of 268 infants from the antepartum part of the IMPAACT P1084s substudy with at least one serum sample collected at the birth visit, week 26 or week 74 visits in Uganda, South Africa and Malawi was selected. Stored serum was analysed for IGF-1, IGF Binding Protein (IGFBP)-1 and IGFBP-3 by Quest Diagnostics using in-house assays validated in accordance with CLIA regulations. IGF-1 was measured using liquid chromatography/mass spectrometry, IGFBP-1 using radioimmunoassay and IGFBP-3 using immunoassay. Risk ratios (RRs) for stunting (length-for-age z (LAZ) -score<-2 standard deviations (SDs)) at birth, weeks 26 and 74 were estimated using log-binomial regression models with log-transformed growth factors at birth as predictors. Cross-sectional associations between LAZ-scores and log-transformed growth factors at birth, weeks 26 and 74 were evaluated using adjusted linear regression models.
Results
213 infant participants (47% male) were included. The median (Q1, Q3) weight within 5 days of birth was 2.9 kg (2.6, 3.2) and length was 48 cm (46, 50), representing a median LAZ-score of -1 SDs (-2, 0). Although higher IGF-1 at birth was significantly associated with reduced risk of stunting at week 26 (RR: 0.4 per log10 increase; 95% confidence interval (CI): 0.19, 0.86), IGFBP-1 or IGFBP-3 at birth were not associated with future stunting (Table 1, Analysis 1). Table 1, Analysis 2 shows that, at 74 weeks, LAZ-scores increased per log10 increase in IGF-1 (1.60; 95% CI: 1.15, 2.04), decreased per log10 increase in IGFBP-1 (-0.62; 95% CI: -1.21, -0.02), and increased per log10 increase in IGFBP-3 (2.08; 95% CI: 1.26, 2.91). Similar observations were made at week 26.
Conclusions
Growth factor levels were significantly associated with LAZ-scores at 26 weeks and 74 weeks of age. However, only modest associations between growth factor levels at birth and future stunting were shown in this population. Therefore, variations in IGF levels may not be the main determinant for stunting in CHEU. Other mechanisms independent of the growth hormone/IGF axis should be explored.