Association between asymptomatic infections and linear growth in 18-24-month-old Malawian children

Abstract

Inadequate diet and frequent symptomatic infections are considered major causes of growth stunting in low-income countries, but interventions targeting these risk factors have achieved limited success. Asymptomatic infections can restrict growth, but little is known about their role in global stunting prevalence. We investigated factors related to length-for-age Z-score (LAZ) at 24 months by constructing an interconnected network of various infections, biomarkers of inflammation (as assessed by alpha-1-acid glycoprotein [AGP]), and growth (insulin-like growth factor 1 [IGF-1] and collagen X biomarker [CXM]) at 18 months, as well as other children, maternal, and household level factors. Among 604 children, there was a continuous decline in mean LAZ and increased mean length deficit from birth to 24 months. At 18 months of age, the percentage of asymptomatic children who carried each pathogen was: 84.5% enterovirus, 15.5% parechovirus, 7.7% norovirus, 4.6% rhinovirus, 0.6% rotavirus, 69.6% Campylobacter, 53.8% Giardia lamblia, 11.9% malaria parasites, 10.2% Shigella, and 2.7% Cryptosporidium. The mean plasma IGF-1 concentration was 12.5 ng/ml and 68% of the children had systemic inflammation (plasma AGP concentration >1 g/L). Shigella infection was associated with lower LAZ at 24 months through both direct and indirect pathways, whereas enterovirus, norovirus, Campylobacter, Cryptosporidium, and malaria infections were associated with lower LAZ at 24 months indirectly, predominantly through increased systemic inflammation and reduced plasma IGF-1 and CXM concentration at 18 months.

Keywords: asymptomatic infection; childhood growth faltering; insulin-like growth factor 1; structural equation modelling; stunting; systemic inflammation.

Figure 1

Mean LAZ and length deficit (cm) relative to the WHO standard (1–23 months) by completed month (n = 604). The solid line indicates participants' mean LAZ. The dashed line indicates the absolute difference between the mean length in the study sample and the 50th centile in the reference population (called length deficit here). LAZ, length‐for‐age Z‐score; WHO, World Health Organisation.

Figure 2

Pathway model for the direct and indirect associations with LAZ at 24 months of age (n = 604). The model was done with structural equation modelling. The numbers next to the lines stand for the model coefficients. The model used standardised variables, and the negative associations are indicated with a minus sign before the number. AGP, alpha‐1‐acid glycoprotein; CFI, comparative fit index; IGF‐1, insulin‐like growth factor 1; LAZ, length‐for‐age Z‐score; RMSEA, root mean square error of approximation; TLI, Tucker–Lewis index; WLZ, weight‐for‐length Z‐score.