The role of albumin and the extracellular matrix on the pathophysiology of oedema formation in severe malnutrition

Abstract

Background: While fluid flows in a steady state from plasma, through interstitium, and into the lymph compartment, altered fluid distribution and oedema can result from abnormal Starling's forces, increased endothelial permeability or impaired lymphatic drainage. The mechanism of oedema formation, especially the primary role of hypoalbuminaemia, remains controversial. Here, we explored the roles of albumin and albumin-independent mechanisms in oedema formation among children with severe malnutrition (SM).

Methods: We performed secondary analysis of data obtained from two independent clinical trials in Malawi and Kenya (NCT02246296 and NCT00934492). We then used an unconventional strategy of comparing children with kwashiorkor and marasmus by matching (discovery cohort, n = 144) and normalising (validation cohort, n = 98, 2 time points) for serum albumin. Untargeted proteomics was used in the discovery cohort to determine plausible albumin-independent mechanisms associated with oedema, which was validated using enzyme-linked immunosorbent assay and multiplex assays in the validation cohort.

Findings: We demonstrated that low serum albumin is necessary but not sufficient to develop oedema in SM. We further found that markers of extracellular matrix (ECM) degradation rather than markers of EG degradation distinguished oedematous and non-oedematous children with SM.

Interpretation: Our results show that oedema formation has both albumin-dependent and independent mechanisms. ECM integrity appears to have a greater role in oedema formation than EG shedding in SM.

Funding: Research Foundation Flanders (FWO), Thrasher Foundation (15122 and 9403), VLIR-UOS-Ghent University Global Minds Fund, Bill & Melinda Gates Foundation (OPP1131320), MRC/DfID/Wellcome Trust Global Health Trials Scheme (MR/M007367/1), Canadian Institutes of Health Research (156307), Wellcome Trust (WT083579MA).

Keywords: Endothelial glycocalyx; Extracellular matrix; Oedema; Proteomics; Severe malnutrition; Starling forces.

Figure 1

(a) Association between serum albumin concentration and degree of oedema severity in the F-75 reformation clinical trial: “None” means no oedema (marasmus) whereas “+” means oedema on both feet; “++” – oedema in both feet and legs; “+++” – oedema in both feet, legs, arms, hands and face. Oedema severity was assessed by trained clinicians following World Health Organization guidelines; (b) probability of presenting with oedema based on serum albumin concentration at hospital admission in the discovery cohort: green and red dots indicate those that presented with or without oedema respectively; (c) difference in serum albumin concentration between kwashiorkor and marasmus in the validation cohort; (d) distribution of oedema status after 3 days of hospitalization in the F-75 reformulation clinical trial; (e) changes in serum albumin concentration among kwashiorkor during admission and after 3 days of hospitalization.

Figure 2

Recruitment flow diagram for the discovery and validation cohorts.

Figure 3

Association between endothelial plasma glycocalyx markers and kwashiorkor. (a,c) Association between the plasma concentrations of syndecan-1 and hyaluronic acid and kwashiorkor in the discovery cohort. (b,d) Association between the plasma concentrations of syndecan-1 and hyaluronic acid and the degree of oedema severity. “None” means no oedema (marasmus) whereas “+” means oedema on both feet; “++” – oedema in both feet and legs; “+++” – oedema in both feet, legs, arms, hands and face. Oedema severity was assessed by trained clinicians following World Health Organization guidelines; p values were estimated using ordinal logistic regression adjusted for age, sex, HIV status, and site of recruitment. Protein concentations were normalised by dividing with serum albumin concentration. (e,f) Association between the plasma concentrations of syndecan-1 and hyaluronic acid and kwashiorkor in the validation cohort in two time points: D0 is discharge from hospital, D60 is 60 days post-discharge.

Figure 4

Differential abundance of proteins between kwashiorkor and marasmus. (a) Volcano plot showing the log fold change (x-axis) and –log p value after false-discovery rate adjustment of plasma proteins (y-axis). The horizontal line signify the FDR p = 0.05. Estimates were obtained using conditional logistic regression adjusting for age, sex, HIV status and site of recruitment stratified for admission serum albumin concentration. (B) Association between the plasma concentrations of lumican and kwashiorkor in the discovery cohort using ELISA. (c) Association between the plasma concentrations of lumican and the degree of oedema severity in the discovery cohort using ELISA. “None” means no oedema (marasmus) whereas “+” means oedema on both feet; “++” – oedema in both feet and legs; “+++” – oedema in both feet, legs, arms, hands and face. Oedema severity was assessed by trained clinicians following World Health Organization guidelines; p values were estimated using ordinal logistic regression adjusted for age, sex, HIV status, and site of recruitment. Protein concentations were normalised by dividing with serum albumin concentration.

Figure 5

Association between SM phenotype and markers of ECM remodelling markers and systemic inflammatory. (a) association between MMP2 and SM phenotype, and severity of oedema in the discovery cohort. (b) associations between SM phenotype and ECM proteins and inflammatory cytokines in the validation cohort in two time points: D0 is discharge from hospital, D60 is 60 days post-discharge.