Therapeutic interventions targeting enteropathy in severe acute malnutrition modulate systemic and vascular inflammation and epithelial regeneration

Abstract

Background: Severe acute malnutrition (SAM) is the most life-threatening form of undernutrition, and children hospitalised with complications have unacceptably high mortality. Complicated SAM is a multisystem disease characterised pathophysiologically by muscle wasting, systemic inflammation, metabolic dysfunction, and malnutrition enteropathy including epithelial barrier dysfunction. There is a clear need for novel interventions to address the underlying pathogenic perturbations of complicated SAM.

Methods: In this analysis of tertiary outcomes from a phase II multi-centre trial in Zambia and Zimbabwe, multiplex biomarkers were measured in 122 children (57% male) with SAM randomised following stabilisation ('baseline') to one of four interventions for 14 days to treat malnutrition enteropathy: budesonide, N-acetylglucosamine, colostrum, or teduglutide, compared with standard-of-care. Following measurement of 35 biomarkers from day 15 plasma samples using Luminex and ELISA, the dimensionality of biomarker data was reduced using principal component analysis.

Findings: Both budesonide and colostrum reduced systemic inflammation (as measured by CD14, IL1-ra, CRP, and LBP), while children receiving colostrum had higher GLP2 and angiopoietin, and lower circulating lipopolysaccharide, suggesting better restoration of epithelial barrier function. N-acetylglucosamine, a precursor for epithelial glycosaminoglycan synthesis, increased biomarkers of epithelial regeneration (EGF, VEGF), and circulating growth factors (angiopoietin, IGFBP-3, and GCSF).

Interpretation: Interventions aimed at ameliorating malnutrition enteropathy showed plausible effects on biomarkers of inflammation and epithelial regeneration, demonstrating an interdependence of systemic inflammation and enteropathy markers seen in structural analysis. Given the interplay between inflammation and tissue restoration in malnutrition, this mechanism of action supports larger trials to determine the clinical benefits of interventions, either alone or in combination, in children with complicated SAM.

Funding: This analysis of tertiary outcomes for the TAME trial was funded by a Wellcome grant to JPS (220566/Z/20/Z). The TAME trial was funded by a grant from the Medical Research Council (UK), number MR/P024033/1. AJP is funded by Wellcome (108065/Z/15/Z). Takeda UK provided teduglutide at a discounted price.

Keywords: Randomised trial; SAM; Severe acute malnutrition; Severe malnutrition; Systemic inflammation; Vascular inflammation.

Fig. 1

CONSORT diagram for children in the TAME trial. Children were randomised to one of five treatment arms (4 IMPs, and standard of care). Three children died, and three children withdrew. Three children exited the study prior to the day 15 endpoint, so their data were not included in the final endpoint analysis. Three children exited between day 15 and day 28, so their data were included in the current analysis.

Fig. 2

Changes of the adjusted log₁₀D15 biomarker value attributable to randomised interventions, compared to the standard of care (SOC) group. Differences are shown in the log10 concentrations of each biomarker for the budesonide, teduglutide, N-acetylglucosamine (NAG), and colostrum groups, compared to the standard of care group. Results were adjusted for sex, oedema, HIV, diarrhoea, WHZ, site, and the baseline biomarker value. A p value threshold <0.10 was pre-specified as statistically significant since this is a phase II trial and marked with ∗; p < 0.05 ∗∗; p < 0.01 ∗∗∗. The full numerical results are shown in Supplementary Table S4.

Fig. 3

PCA analysis of a) day 15 systemic markers (N = 125) and b) day 15 enteropathy markers (N = 125). Biomarkers were log₁₀ transformed, standardised, and normalised before undergoing PCA. Eigenvector values, which correspond to a coefficient of orthogonal projection attributable to each biomarker following transformation for that component, are shown.

Fig. 4

PLS path modelling of the relationship between the components and the effect of the trial intervention by a) budesonide (N = 25), b)colostrum (N = 25), c) N-acetlyglucosamine (NAG) (N = 24), and d) teduglutide (N = 26) compared with standard of care (N = 25). The relationship of the PCA components (detailed in Fig. 3) are assessed for each of the trial arms, compared with the standard of care. Results displayed are the standardised path coefficients, which are shown in full in Supplementary Table S5. Paths are displayed here is the p < 0.10 for the path coefficient. In addition to the above, HIV, site, oedema, diarrhoea, and WHZ was added and connected to each of the D15 components (covariates not shown). The full model tested is shown in Supplementary Fig. S1.