Stunting is characterized by chronic inflammation in Zimbabwean infants

Abstract

Background: Stunting affects one-third of children in developing countries, but the causes remain unclear. We hypothesized that enteropathy leads to low-grade inflammation, which suppresses the growth hormone-IGF axis and mediates stunting.

Methods: We conducted a case-control study of 202 HIV-unexposed Zimbabwean infants who were stunted (height-for-age Z-score (HAZ) <-2; cases) or non-stunted (HAZ >-0.5; controls) at 18 months. We measured biomarkers of intestinal damage (I-FABP), inflammation (CRP, AGP, IL-6) and growth hormone-IGF axis (IGF-1, IGFBP3) in infant plasma at 6 weeks and 3, 6, 12 and 18 months, and in paired maternal-infant plasma at birth. Adjusted mean differences between biomarkers were estimated using regression models. Multivariate odds ratios of stunting were estimated by logistic regression.

Results: At birth, cases were shorter (median (IQR) HAZ -1.00 (-1.53, -0.08) vs 0.03 (-0.57, 0.62,); P<0.001) than controls and their mothers had lower levels of IGF-1 (adjusted mean difference (95%CI) -21.4 (-39.8, -3.1) ng/mL). From 6 weeks to 12 months of age, levels of CRP and AGP were consistently higher and IGF-1 and IGFBP3 lower in cases versus controls; IGF-1 correlated inversely with inflammatory markers at all time-points. I-FABP increased between 3-12 months, indicating extensive intestinal damage during infancy, which was similar in cases and controls. In multivariate analysis, higher log10 levels of CRP (aOR 3.06 (95%CI 1.34, 6.99); P = 0.008) and AGP (aOR 7.87 (95%CI 0.74, 83.74); P = 0.087) during infancy were associated with stunting. There were no associations between levels of I-FABP, IL-6, sCD14 or EndoCAb and stunting.

Conclusions: Stunting began in utero and was associated with low maternal IGF-1 levels at birth. Inflammatory markers were higher in cases than controls from 6 weeks of age and were associated with lower levels of IGF-1 throughout infancy. Higher levels of CRP and AGP during infancy were associated with stunting. These findings suggest that an extensive enteropathy occurs during infancy and that low-grade chronic inflammation may impair infant growth.

Figure 1. Selection of cases and controls.

14110 women were recruited within 96–2000. Eligible infants for this current study were born to women who were HIV-negative at baseline and remained uninfected through 18 months, who had anthropometry data available at 18 mo and stored plasma samples of ≥0.2 mL volume from at least 4 study time-points between 6 weeks and 18 months of age. Stunted infants were selected based on height-for-age Z-score <−2.0 at 18 mo; non-stunted infants were selected based on height-for-age Z-score >−0.5 at 18 mo.

Figure 2. Growth in cases and controls.

A: Median height-for-age Z-score (HAZ) in non-stunted (dashed line) and stunted (solid line) infants between birth and 18 months of age, with interquartile range. B: Median growth velocity in non-stunted (dashed line) and stunted (solid line) infants between birth and 18 months of age, with interquartile range. Growth velocity was calculated as height change per day, by comparing height (in centimeters) at consecutive visits and dividing by the number of days between visits.

Figure 3. Changes in biomarkers over the first 18 months of life.

Levels of A: Insulin-like growth factor 1 (IGF-1), B: IGF-binding protein 3 (IGFBP3), C: Intestinal fatty acid binding protein (I-FABP), D: Soluble CD14 (sCD14), E: IgG endotoxin core antibodies (EndoCAb), F: Interleukin-6 (IL-6), G: C-reactive protein (CRP) and H: Alpha-1 acid glycoprotein (AGP), in non-stunted (blue line) and stunted (red line) infants between birth and 18 months of age. Data shown are means with standard errors, except for IL-6, which shows medians with interquartile range (solid error bars for non-stunted infants and dashed bars for stunted infants). Mean levels, standard deviations, unadjusted and adjusted differences between cases and controls are shown in Table S1.

Figure 4. Relationships between pro-inflammatory markers and IGF-1.

Associations between levels of IGF-1 and (Panel A) AGP, (Panel B) CRP, (Panel C) soluble CD14 and (Panel D) IL-6, at 3 months of age. Data for cases and controls are combined. Spearman correlations are shown.

Figure 5. Influence of IGF-1 on birth weight and relationship between maternal and infant inflammation at birth.

A: Relationship between infant levels of IGF-1, measured within 96 hours of birth, and birth weight. Spearman correlation is shown. B: Relationship between maternal and infant AGP levels measured in paired plasma samples collected within 96 hours of delivery. Spearman correlation is shown.