Background: Environmental enteric dysfunction (EED), a subclinical inflammatory state of the intestinal mucosa, is thought to play a role in stunting. Stunting affects one in every fifth child under five years and is associated with impaired early development and later risk of chronic disease. EED is widespread among children living in settings where basic water, sanitation and hygiene (WASH) are lacking and is thought to result from a high pathogen burden that increases scarcely met nutrient requirements. The resultant cycle of epithelial damage and inflammation is thought to create or worsen nutrient deficiencies. Recent trials have combined WASH and smallquantity lipid-based nutrient supplement (LNS) interventions but have had little impact on growth. Small-quantity LNS may be insufficient to meet the increased nutrient demands for growth alongside chronic immune activation and intestinal disrepair. An array of biomarkers are currently being explored to explain the functional implications of EED and its role in stunting. Among them are plasma citrulline (p-cit), a marker of small intestinal enterocyte mass and faecal myeloperoxidase (f-MPO) a marker of intestinal inflammation. Objectives: The main objective of this PhD thesis was to explore environmental enteric dysfunction (EED) using selected biomarkers and investigate the interaction between nutrition intervention and EED in children with stunting. Specific objectives were to write a protocol in order to test the effects of a large-quantity LNS on growth in child stunting. To explore the correlates of p-cit and to examine the effects of LNS containing milk protein (MP) and whey permeate (WP) on p-cit and f-MPO in the study population. Finally, to explore whether the state of the gut status modifies the effects of LNS on outcomes of growth and micronutrient status. Methods: Paper I describes the MAGNUS trial protocol. A community based 2x2 factorial trial in Eastern Uganda, among 750 stunted children aged 12-59 months. Children were randomized to LNS (100 g/d for 12 weeks) containing MP and/or WP or to no supplementation. In a crosssectional study (Paper II) linear regression was used to explore baseline correlates of p-cit. The influence of covariates age, fasting and systemic inflammation were explored as well as
associations with socioeconomics, diet, micronutrient status and WASH characteristics. In Paper III, linear mixed-effects models were used to explore p-cit and f-MPO as outcomes and modifiers of the LNS intervention effect. Results: All children (N=750) were included between February and September 2020. The mean ±SD age was 32 ±11.7 months and the mean height-for-age Z-score (HAZ) was -3.02 ±0.74. In Paper II, the mean p-cit at baseline (n=730) differed according to the duration of fasting and was 20.7 ±8.9, 22.3 ±10.6 and 24.2 ±13.1 µmol/L if fasted <2, 2-5 and >5 hours, respectively. Positive correlates of p-cit were age and log10 serum insulin-like growth factor-1 (s-IGF1) and negative correlates included serum C-reactive protein, serum α1-acid glycoprotein and anaemia as well as environmental factors; food insecurity, inadequate housing materials, the wet season, an unimproved toilet and lack of soap for handwashing. Many associations attenuated after accounting for the effects of systemic inflammation. Neither MP nor WP had an effect on p-cit or f-MPO in Paper III. The LNS intervention had no effect on p-cit, though there was an 82% (95%CI: 12; 196) greater increase in f-MPO in this group compared to the unsupplemented group. In the sensitivity analysis, individuals with recent diarrhoea (n=245) were removed and the effect of LNS on f-MPO disappeared. The effect of LNS on cobalamin (B12) status was reduced in children with p-cit <20 µmol/L. The change in plasma cobalamin at 12 weeks was 20% less and the increase in plasma methylmalonic acid was 59% greater in those with low p-cit at baseline. Discussion and Conclusion: The factors found to correlate with p-cit are characteristic of environments with a high prevalence of EED. Systemic inflammation was strongly associated with p-cit and is implicated in the pathogenesis of EED and stunting. This study highlights the complex interplay that exists between p-cit and systemic inflammation. With adjustment for systemic inflammation many associations were attenuated including WASH related factors indicating that these are likely pathways of pathogen exposure and infection. There was no effect of ingredients MP and/or WP in the LNS on EED markers and no effect of LNS on enterocyte mass, although a negative effect of the LNS on intestinal inflammation appeared in those with recent diarrhoea. Finally, the beneficial effect of the LNS on cobalamin (B12) status was reducedin those with low enterocyte mass at baseline. These findings demonstrate that even large quantity LNS interventions are insufficient in themselves to improve EED. Instead, there was indication that EED may be compromising the beneficial effect of nutrition interventions, possibly through malabsorption or nutrient sequestration. To reverse EED, it seems necessary to take a combined approach of community wide WASH interventions together with improved nutrient access. A sustainable reduction in the pathogen burden would allow for intestinal repair and improve the effectiveness of nutrition interventions.