Multi-omics, multi-tissue analysis reveal role of extracellular matrix remodeling and lipid transport dysfunction in edematous malnutrition (kwashiorkor)

Multi-omics, multi-tissue analysis reveal role of extracellular matrix remodeling and lipid transport dysfunction in edematous malnutrition (kwashiorkor) | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Multi-omics, multi-tissue analysis reveal role of extracellular matrix remodeling and lipid transport dysfunction in edematous malnutrition (kwashiorkor) Fabienne Nackers, Kirrily De Polnay, Ruben Almey, Simon Daled, and 23 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8320069/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Edematous malnutrition, aka kwashiorkor, is a phenotype of severe malnutrition whose pathophysiology remains poorly understood. In this case-control study, we employed plasma lipidomics, metabolomics, and proteomics, urine metabolomics and gut microbiome profiling to delineate molecular pathways specific to kwashiorkor in 60 children aged 6–59 months from Niger compared to marasmus (n = 60) and non-malnourished children (n = 60) matched by age, sex, and clinical triage score. Features were defined as kwashiorkor-specific if they also correlated with edema severity and normalized following nutritional rehabilitation. Our analyses revealed that kwashiorkor is marked by increased extracellular matrix (ECM) degradation, evidenced by elevated plasma ECM proteins, and by disrupted sphingolipid homeostasis. Neither plasma nor urine metabolomic profiles, nor gut microbiome signatures, showed unique alterations associated with kwashiorkor. These findings suggest that kwashiorkor may be a combination of nutritional deficiencies and the disruption of the ECM and sphingolipid metabolism, potentially linked with an inflammatory syndrome. Health sciences/Molecular medicine Health sciences/Medical research/Paediatric research Biological sciences/Biochemistry/Proteomics Kwashiorkor Edematous malnutrition Extracellular matrix Sphingolipid Edema Metabolomics Proteomics Lipidomics Gut microbiome Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Edematous malnutrition, also known as kwashiorkor, is a severe form of acute malnutrition characterized by bilateral edema, fatty liver, and skin changes. It is clinically and biochemically distinct from non-edematous malnutrition, also known as marasmus or severe wasting, evident as low weight-for-height (<-3 z-scores from the 2006 WHO standards), or a mid-upper arm circumference (MUAC) below 115 mm (children 6–59 months), visible muscle atrophy, and loose skin 1 . Both conditions may co-occur where a malnourished child presents with both severe wasting and edema – known as marasmic kwashiorkor. Kwashiorkor represents a critical public health issue in low-income countries although its actual prevalence is difficult to ascertain as it is often not captured in population-based surveys 2 . Despite its clinical recognition for over eight decades, the precise molecular mechanisms underlying the pathophysiology of kwashiorkor remain incompletely understood. This gap in knowledge impedes the development of targeted interventions and therapeutic strategies to effectively treat and prevent this devastating condition. Previously, kwashiorkor had been attributed to protein deficiency 3 ; however, emerging evidence suggests that its etiology is multifactorial, involving a complex interplay of nutritional, and environmental factors. Recent studies have highlighted the potential roles of oxidative stress 16 , immune dysregulation 4 , and alterations in the gut microbiome in the development of kwashiorkor 5 . Exogenous toxicants, including aflatoxins from staple foods like maize, groundnuts, sorghum, and millet often consumed in low and middle-incomes countries could also increase oxidative stress and explain the accumulation of liver fat in kwashiorkor. However, despite finding aflatoxins in biological samples of children with kwashiorkor 6 – 10 , a causal link has not be demonstrated as the possibility of reverse causality has not been ruled out 11 . Additionally, the dysregulation of metabolic pathways, particularly those related to amino acid metabolism 12 , 13 , one-carbon metabolism 14 , and the synthesis and breakdowns of extracellular matrix (ECM) proteins 15 and endothelial dysfunction 16 , have been implicated in the disease process. Yet, the comprehensive molecular mechanisms remain to be fully elucidated. To gain a deeper understanding of the molecular underpinnings of kwashiorkor, we employed a comprehensive multi-omics approach, integrating metabolomics, lipidomics, proteomics, and gut microbiome data across multiple biological samples throughout the different treatment stages to compare it to marasmus and non-malnourished children and explore its multifaceted pathophysiology. The data may lead to the understanding of its etiology and, hopefully, to the discovery of more effective treatment strategies. Results Participant characteristics A total of 180 children aged 6–59 months were enrolled in the study, with kwashiorkor (n = 60), marasmus (n = 60) and without acute malnutrition (n = 60) matched on sex, age and clinical triage score. Majority (60%) of the children were female and age categories were evenly distributed at baseline (50% 6–23 months, 50% 24–59 months old). An even distribution was accomplished between the edema severity levels on baseline, except for one child who was recruited as presenting moderate edema (++) but was secondarily reconsidered as having low severity edema (+). Of the 60 kwashiorkor participants, 12 had at least one type of cutaneous lesion, and these children with skin lesions all had moderate to severe edema (++/+++), except for one child with the lowest edema severity score (+). Height-for-age z-scores indicated a substantial burden of stunting. Most children were not critically ill (Table 1 and Supplementary table 1), with only one child per group classified as an emergency case based on study triage score at admission. NAM children were more likely to test positive for malaria using rapid diagnostic tests compared to children with kwashiorkor or marasmus and they also presented a higher Pediatric Early Warning Scores (PEWS) at baseline. Marasmic children suffered more commonly from diarrhea. Laboratory findings upon admission for the three groups are compared in the Table 1 and additional clinical characteristics and socioeconomic indicators of the participants can be found in Supplementary table 1. Two kwashiorkor participants died. Table 1 Baseline characteristics and outcome of matched study participants Kwashiorkor (n = 60) Marasmus (n = 60) No acute malnutrition (NAM) (n = 60) Age (months) mean ± SD 24.7 (7.8) 21.3 (9.0) 25.0 (12.1) n (%) 6–11 2 (3.3) 11 (18.3) 6 (10.0) 12–23 27 (45.0) 20 (33.3) 24 (40.0) 24–35 24 (40.0) 26 (43.3) 17 (28.3) 36–47 5 (8.3) 2 (3.3) 8 (13.3) 48–59 2 (3.3) 1 (1.7) 5 (8.3) Sex (n female, %) 36 (60.0) 36 (60.0) 36 (60.0) Triage score* n (%) Green (non-urgent) 30 (50.0) 30 (50.0) 30 (50.0) Yellow (priority) 29 (48.3) 29 (48.3) 29 (48.3) Red (emergency) 1 (1.7) 1 (1.7) 1 (1.7) Initial recruitment site n (%) Children hospitalized 30 (50) 30 (50) 30 (50) Children in ambulatory treatment facility n (%) 30 (50) 30 (50) 30 (50) Clinical characteristics Level of edema at baseline n (%) 0 / 60 (100.0) 60 (100.0) + 21 (35.0) / / ++ 19 (31.7) / / +++ 20 (33.3) / / Middle upper arm circumference (mm) Median [IQR] 122 [115, 130] 111 [107, 116.5] 139 [133, 145.5] Weight for height z-score Median [IQR] -2.14 [-2.84, -1.04] -3.61 [-4.23, -3.20] -1.26 [-1.60, -0.77] Weight for age z-score Median [IQR] -3.25 [-4.11, -1.94] -4.16 [-4.91, -3.58] -1.81 [-2.36, -1.03] Height for age z-score Median [IQR] -3.29 [-4.09, -2.06] -2.88 [-4.59, -1.95] -1.79 [-2.36, -0.71] PEWS at baseline n (%) 0–2 (green) 58 (96.7) 57 (95.0) 46 (76.7) 3–4 (yellow) 2 (3.3) 1 (1.7) 11 (18.3) ≥ 5 (orange/red) 0 (00.0) 2 (3.3) 3 (5.0) Temperature ≥ 38 (° C) n (%) 3 (5.0) 8 (13.6) 18 (30.0) Severe clinical anaemia n (%) 1 (1.7) 4 (6.7) 14 (23.3) Diarrhoea (≥ 3 stools) n (%) 16 (26.7) 35 (58.3) 14 (23.3) O2 saturation ≥ 93% (ambient air) n (%) 58 (98.3) 58 (98.3) 59 (98.3) Laboratory Malaria rapid test (n positive, %) 32 (53.3) 29 (48.3) 41 (73.2) Glucose mg/dL - Median [IQR] 79 [68,87] 88 [80,100] 91.5 [81,107] Haemoglobin g/dL - Median [IQR] 9.2 [7;1,10.5] 9.9 [8.8,11.6] 9.2 [6.3,10.7] Haematocrit % - Median [IQR] 27 [21,31] 30 [27,34] 27 [19,32] Creatinine mg/dL - Median [IQR] < 0.2 [< 0.2,0.2] < 0.2 [< 0.2,<0.2] < 0.2 [< 0.2,<0.2] ALT (U/L) - Median [IQR] 35 [23,58] 18.5 [14,24] 20 [14,30] CRP (mg/L) - Median [IQR] 13.1 [1.6,32.0] 12.4 [1.5,72.4] 72.9 [16.0,141] Total bilirubin (mg/dL) - Median [IQR] 0.3 [0.2,0.4] 0.3 [0.2,0.4] 0.4 [0.2,1.1] Albumin (g/L) - Median [IQR] 17.3 [13.2,21.5] 29.9 [21.4,34.1] 31.9 [26.2,36.6] Sodium mmol/L - Median [IQR] 137 [134,140] 135 [131,137] 134 [132,137] Potassium mmol/L - Median [IQR] 3.1 [2.7,3.8] 3.4 [2.7,4] 3.8 [3.3,4.2] Lactate mmol/L - Median [IQR] 2.2 [1.3,3.1] 1.8 [1.3,2.6] 1.9 [1.4,3.1] Clinical progress and outcome Stay ≥ 1 days in intensive care n (%) 2 (6.9) 3 (10.0) 4 (13.3) Final outcome (n %) Cured 51 (85.0) 50 (83.3) 60 (100) Deceased 2 (3.3) 0 0 Exit against medical advice 4 (6.7) 9 (15.0) 0 Not responding to nutrition treatment 3 (5.0) 1 (1.7) / * The study triage score was adapted from the Emergency triage Assessment and Treatment (ETAT) score applied in the MSF program in Niger at the time of the study. Details can be found in Supplementary table 2. Two nutritional criteria (visible severe wasting, oedema on both feet) were not included in the study triage score, wherein a malnourished child presenting in yellow (priority) case indicates that the child had another priority signs besides having visible severe wasting or oedema on both feet. PEWS: Pediatric Early Warning Scores; IQR: Interquartile range. Integrated plasma lipidome, metabolome, proteome and clinical biochemistry of children with kwashiorkor, marasmus and non-acutely malnourished In this study, we imposed strict criteria for the determination of kwashiorkor-associated pathways. Inferences on individual lipids, metabolites and proteins were obtained. Moreover, we also obtained inferences at the pathway level by consolidating correlated metabolites, lipid and proteins into uncorrelated latent variables (LV) using factor analysis (FA). Each omics domain was first analyzed separately. The use of factor analysis affords us to reduce the big omics datasets into explainable LVs, which can be used for further downstream analyses. Each LV represents an underlying explanatory relationship and interaction among its component features, hence these factors provide a singular value that potentially refer to an overall behavior of particular pathway(s). This allowed us to make inferences on pathway level, instead of the traditional single feature analysis, typical of omics studies. Individual plasma lipid, metabolite and protein features, and their resulting LVs, associated with kwashiorkor must pass the following criteria: Criterion 1. The feature or LV must be different in kwashiorkor compared to marasmus and NAM at recruitment. Marasmus and NAM are grouped together in subsequent analyses. Analyses are matched for age, sex and clinical triage score. Criterion 2. The feature or LV must be associated with the severity of edema (+, ++, +++) at recruitment. Criterion 3. The feature or LV must change towards normalcy as the child with kwashiorkor recovers to a non-malnourished state following nutritional therapy. These strict criteria ensure that our findings are robust despite the small sample size of this study (n = 180). However, analyses of the urine metabolome and gut microbiome only followed criteria 1 and 2 since these analyses were only performed on samples obtained at baseline. Plasma proteomics Untargeted proteomics analysis revealed 121 proteins were upregulated and 44 were downregulated (false discovery rate [FDR] p-value < 0.05) out of 308 protein features in kwashiorkor compared to both marasmus and NAM combined (Fig. 1 a). However, when tested for association with edema severity in kwashiorkor, none of these proteins were found to have passed criterion 2 (Fig. 1 b). FA was then employed to determine clusters of proteins that share a latent relationship, potentially belonging to the same pathway. FA reduced the proteomics data into 9 proteome-LVs (pLV). Component proteins in each LV are shown in Supplementary file 1. Figure 1 c shows the differences in the pLVs between kwashiorkor and non-kwashiorkor (marasmus and NAM). Out of 9 pLVs, 7 were found to be different in kwashiorkor (pLV 1, 3, 4, 5 and 7 were higher in kwashiorkor, pLV2 and 9 were lower in kwashiorkor than non-kwashiorkor cases), and hence passed criterion 1. Regression analyses of these pLVs with kwashiorkor edema severity revealed that only pLV7 was significantly associated with increasing severity of edema in kwashiorkor at baseline (Fig. 1 d). Moreover, this pLV was also higher than in non-kwashiorkor children at baseline, but this level significantly reduced during and after nutritional therapy (Fig. 1 e), passing all 3 criteria for a candidate pathway associated with kwashiorkor. Uncovering the component proteins of pLV7 revealed that the extracellular matrix proteins lumican, gelsolin and tetranectin were involved in kwashiorkor pathophysiology. The high correlation between these individual proteins and pLV7 demonstrates that pLV7 was able to capture the behavior of these three proteins as a cluster (Fig. 1 f). Using the Human Proteome Atlas 17 , these three proteins appear to be found in many tissue types all over the body, and can be found co-expressed in 41 out of 50 human organs (Fig. 1 g). However, the organs where they are most highly expressed varied – lumican in gall blader, gelsolin in heart muscle, tetranectin in adipose tissue. Because the NAM children were more likely to have a positive malaria rapid test on admission (Supplementary table 3), a first sensitivity analysis was conducted wherein children testing positive for malaria rapid diagnostic test were removed. Also, because not all children with kwashiorkor could produce urine on admission, and hence some were included without a urine strip result (Supplementary table 3), a second sensitivity analysis was conducted including only kwashiorkor children who tested negative for urine blood and protein. Both revealed that the association of pLV7 and kwashiorkor is robust, albeit with lower power due to the smaller sample size (especially for criterion 2) (Supplementary Fig. 1). Plasma lipidomics Untargeted lipidomics analysis of plasma revealed 22 lipids were upregulated and 136 were downregulated (FDR p-value < 0.05) out of 506 lipid features in kwashiorkor compared to both marasmus and NAM combined (Fig. 2 a). However, when assessed for their association with severity of edema in kwashiorkor, none of these lipids were found to have passed criterion 2 (Fig. 2 b). Hence, no individual lipid species was found to be strongly associated with kwashiorkor in this study. FA reduced the lipidomics data into 21 lipid-LVs (lLV). Component lipids in each LV is shown in Supplementary file 1. Figure 2 c shows the differences in the lLVs between kwashiorkor and non-kwashiorkor (marasmus and NAM). Out of 21 lLVs, 9 were found to be different in kwashiorkor (lLV 14 and 16 were higher in kwashiorkor, lLV 1, 2, 5, 10, 13, 15 and 20 were lower in kwashiorkor than non-kwashiorkor cases), and hence passed criterion 1. Regression analyses of these lLVs with kwashiorkor edema severity revealed that lLV5 and lLV16 were significantly associated with increasing severity of edema in kwashiorkor at baseline (Fig. 2 d). However, only lLV5 was significantly changed throughout the course of the treatment (Fig. 1 e). At baseline, this lLV was lower in kwashiorkor than non-kwashiorkor cases, and its level increased in plasma after treatment making this lLV a strong candidate pathway associated with kwashiorkor. The component lipids of lLV5 are dominated by 17 mono- to di-unsaturated sphingomyelins (SM 35:1, 37:2, 38:1, 38:2, 39:1, 40:0, 40:1, 40:2, 41:0, 41:1, 42:0, 42:1, 42:2, 43:1, 44:0, 44:1, and 44:2), along with 5 polyunsaturated cholesterol esters (16:2, 18:3, 20:4, 20:5, 22:6), ganglioside GD3 (42:1), four saturated hexo-ceramides (39:0-OH, 40:0-OH: 41:0-OH and 42:0-OH), Ether-linked lysophosphatidylcholine (24:0) and phosphatidylcholines (20:0, 22:0, 22:3, 37:1). The high correlation between these individual lipids and lLV5 demonstrate that lLV5 was able to capture the behavior of these lipids as a cluster (Fig. 2 f). Because the NAM children were more likely to have a positive malaria rapid test on admission (Supplementary table 3), a first sensitivity analysis was conducted wherein children testing positive for malaria rapid diagnostic test were removed. Also, because not all children with kwashiorkor could produce urine on admission, and hence some were included without a urine strip result (Supplementary table 3), a second sensitivity analysis was conducted including only kwashiorkor children with negative blood and protein urine test. Both revealed that the association of lLV5 and kwashiorkor is robust, albeit with lower power due to the smaller sample size (especially for criterion 2) (Supplementary Fig. 2). Plasma metabolomics Plasma untargeted (semi-polar) metabolomics yielded a total of 2,825 features for both positive and negative ionization modes combined. Conditional logistic regression to account for the matched design on age, sex and triage score revealed 60 metabolite features different in kwashiorkor (FDR p-value < 0.05) compared to both marasmus and NAM combined (Fig. 3 a). However, when tested for association with edema severity in kwashiorkor, none of these metabolite features were found to have passed criterion 2 (Fig. 3 b). Following FA, 16 metabolome-LVs (mLV) were extracted, of which only mLV5 and mLV7 passed criterion 1. Component metabolite features in each LV is shown in Supplementary file 1. Upon investigation of their association with kwashiorkor edema severity during nutritional recovery, we found no plasma mLV passing all 3 criteria, and hence no semi-polar metabolome-specific pathways were found to be associated with kwashiorkor in this study. Plasma multi-omics integration To integrate the multiple omics studies and to contextualize the clinical meaning of the individual latent variables (LV), we used partial correlation network analysis to determine patterns of association among all modules and clinical biochemistry results for all children (kwashiorkor, marasmus and NAM) at baseline (Fig. 4 a). Plasma clinical biochemistry results are presented in Supplementary table 3. Nodes correspond to individual pLV, lLV, mLV or clinical biochemistry (both laboratory and point-of-care analysis) results for baseline plasma samples. Correlations are depicted by either a green line (indicating positive partial correlation) or a red line (negative partial correlation). Linkages from the network (Fig. 4 a) are isolated in Fig. 4 b to highlight the clinical and biochemical nodes associated with pLV7 and lLV5. As shown (Fig. 4 b), pLV7 (which represents ECM proteins) was negatively associated with total plasma protein but positively associated with blood sodium concentration. On the other hand, lLV5 was positively associated with total plasma protein, total serum cholesterol, lLV10 and lLV13, and negatively associated with pLV4 and pLV5. lLV10 consists of 15 mostly phosphatidylethanolamine lipid species, whereas lLV13 is comprised of 5 unsaturated sphingomyelins. lLVs 10 and 13 were also found to be different in kwashiorkor compared to non-kwashiorkor, but their levels were not associated with kwashiorkor severity. pLV4 is composed of protein such as actin, complement component C7, protein virilizer homolog, angiotensinogen, carboxypeptidase N subunit 2, and hemoglobin subunit beta. pLV5 is composed of alpha-1-acid glycoprotein 1, complement C5, pigment epithelium-derived factor, leucine-rich alpha-2-glycoprotein, and zinc-alpha-2-glycoprotein. The composition of the different modules can be found in Supplementary file 1. lLV5 and pLV7 are not directly correlated with each other (Fig. 4 c) despite both being associated with total plasma protein. Consistent with most findings in the literature, including our previous studies, total plasma protein content is lower in kwashiorkor than in marasmus and NAM. However, there is a considerable overlap among the groups. Many children with marasmus and NAM had low total protein in their blood but did not manifest kwashiorkor syndrome. These associations remained robust in sensitivity analyses where children who tested positive for malaria in a rapid diagnostic test (Supplementary Fig. 3) were removed or that only children with kwashiorkor who tested negative for urine blood and protein using a urine test strip (Supplementary Fig. 4) were used for the analysis. Urinary metabolome and clinical biochemistry of children with kwashiorkor, marasmus and NAM Due to difficulties in obtaining urine samples from children, we were only able to obtain baseline urine samples from 88 children (24 kwashiorkor, 32 marasmus, 32 NAM; patient characteristics in Supplementary table 4. Results for urinary biochemical analyses per group can be found in Supplementary table 5). Combining the positive and negative ionization modes together yielded a total of 4570 urine metabolomic features. After FA, the data was reduced to 8 urinary metabolome LVs (uLV). Since urine was only obtained at recruitment, only criteria 1 and 2 were applied to the urine results. There was no uLV associated with kwashiorkor (Fig. 5 ). Gut microbiome composition of children with kwashiorkor, marasmus and NAM Out of 180 children at recruitment, we were only successfully able to collect fecal samples from 38 children (14 kwashiorkor, 13 marasmus, 11 NAM; patient characteristics in Supplementary table 6). Fecal samples were then subjected to 16s rRNA gene amplicon sequencing, and the resulting gene fragment amplicon sequences were quantified and taxonomically annotated. We initially visualized the variation in the data using principal coordinates analysis (PCoA) using Bray-Curtis and Jaccard distances. Using all distance measures, no clear clustering could be found neither generally nor on the basis of group (kwashiorkor vs marasmus vs NAM), age, sex and triage score (Fig. 6 a). PERMANOVA analysis concurred with this visual inspection as differences among the groups were not significant. Moreover, no differences in alpha diversity metrics (Shannon, Faith Phylogenetic, Pielou) were found among the three groups (Fig. 6 b). Finally, we used the ANCOM module in QIIME2 to identify differentially abundant features across sample groups at the taxonomic level of family, genus and strain. Both groupwise and pairwise analyses report no differentially abundant features (p > 0.05) among kwashiorkor, marasmus and NAM (Fig. 6 c). Discussion In this study, we identified extracellular matrix (ECM) degradation and sphingolipid disruption as central metabolic processes underlying kwashiorkor. Using a stringent multi-omics design, we uncovered latent variables that were distinct in kwashiorkor compared to marasmus and non-malnourished (NAM) children, correlated with oedema severity, and normalized with treatment. This approach delineates kwashiorkor-specific pathways that move beyond simple nutritional deficiency and instead reveal maladaptive tissue remodeling and lipid signaling. We presented data from children with kwashiorkor, marasmus and NAM, matched based on age, sex, and a programmatic clinical triage score. Matching especially on triage score as a proxy for clinical severity upon admission was decided for the following reasons: first, non-malnourished children without serious illness would overtly distinguish themselves from malnourished ill children only on the basis of illness severity. Second, certain diseases tend to co-occur with either form: for instance, HIV and malaria are reported to co-occur more with marasmus than kwashiorkor 18 , 19 . Third, mortality rates have been reported to be higher in kwashiorkor than marasmus 20 . Other studies have reported the opposite, especially in areas with high burden of HIV 21 . Hence, matching for triage score, albeit imperfect, allowed to somehow disentangle pathways linked with kwashiorkor other than from clinical co-morbidities. However, differences remained between groups, with NAM children notably more likely to have a positive malaria rapid test. This study further strengthens our previous findings on the role of ECM remodeling/degradation in the pathophysiology of kwashiorkor. We have previously shown that the protein lumican was higher in kwashiorkor compared to marasmus, was associated with severity of edema, and was resolving during treatment in Kenyan and Malawian children 15 . In the previous study, we matched the children with kwashiorkor and marasmus on the basis of serum albumin concentration, which led us to conclude that both hypoalbuminemia and heightened ECM degradation play important roles in kwashiorkor development. However, this previous work did not include non-malnourished children, and only focused on proteomics. In this current study, a pLV comprising lumican, gelsolin, and tetranectin met all biomarker criteria, reinforcing the role of ECM remodeling in kwashiorkor. Lumican, a small leucine-rich proteoglycan, regulates collagen fibrillogenesis and tissue integrity 22 , 23 . This protein’s potential role in kwashiorkor has been described in our previous work 15 . Gelsolin, an actin-binding protein, also affects ECM by clearing actin filaments and modulating inflammation 24 . Plasma gelsolin (pGSN) plays key roles in actin scavenging, and immune modulation 24 , 25 . pGSN levels have been reported to significantly decreased in tissue injury, secondary organ damage, and sepsis 26 , but is increased in colon cancer, pancreatic cancer and pancreatitis 26 , and HIV 27 . Tetranectin enhances plasminogen activation, important for fibrinolysis and ECM breakdown 28 . Using the human proteome atlas tool 17 , we deduced that pLV7 represents a systemic ECM breakdown, as these proteins are co-expressed in a multitude of organs, including the lymphatic system and the liver. Our previous finding that plasma lymphatic vessel endothelial hyaluronan receptor 1 (LYVE1) levels are higher in kwashiorkor than marasmus, and are reduced after 60 days of post-discharge nutritional rehabilitation 15 indicating ECM degradation in the lymphatic system in kwashiorkor. Dysfunction of the lymphatic system integrity could potentially be playing a major role in edema formation in kwashiorkor, as it is the major route for interstitial fluid drainage back to circulation, as explained in the revised Starling model. The interstitial volume is mainly controlled by the activity of lymph flow which depends on the lymphatic function 29 . Damage to the lymph, i.e. degradation of lymph ECM, may contribute to edema formation. Moreover, ECM proteins are major constituents of the vessel wall supporting endothelial cells throughout the entire vascular system 30 . Degradation of the ECM in the vessel wall will therefore result in vascular leakage and increased capillary membrane permeability, further contributing to fluid leakage to the interstitium. The ECM is also an important reservoir for sodium, enabled by the interaction between negatively charged ECMs, specifically glycosaminoglycans, allowing them to store non-osmotic Na + ions. ECM is therefore an integral and dynamic component of sodium balance 31 . Our network analysis revealed a positive association between total blood sodium (via iStat point-of-care device) and pLV7. Disruption of the ECM integrity could therefore also influence sodium balance, and consequently contribute to water retention in kwashiorkor. Our results showed that children with kwashiorkor have higher blood sodium concentration compared to marasmus and non-malnourished children, which concur with previous findings 32 . However, it is worth noting that despite this statistical difference, median blood sodium levels in all children in this study still fall within normal range. No difference in urinary sodium was found between kwashiorkor and marasmus in this study, but both conditions had lower urinary sodium compared to non-malnourished controls (Supplementary table 5). The dysregulation of ECM homeostasis has also been implicated in liver pathologies, including liver fibrosis and non-alcoholic fatty liver disease 34 . It remains unclear whether systemic ECM homeostasis dysregulation is directly involved in kwashiorkor-related liver fat infiltration; but its participation in other liver pathologies makes this plausible. We also found that lipid metabolism is strongly associated with kwashiorkor, particularly through latent lipid variable 5 (lLV5), which captures signatures of sphingomyelins, ceramides, and gangliosides. These lipids, collectively classified as sphingolipids, are key structural components of cell membranes, especially in the central nervous system, and play critical roles in cell signaling, proliferation, and apoptosis 35 . Ceramide serves as the core molecule, which can be modified to form sphingomyelins and further conjugated with glycolipids to produce gangliosides 36 . Thus, lLV5 likely reflects overall sphingolipid homeostasis. Since circulating sphingolipids are primarily synthesized by the liver and intestine, with only a small fraction derived directly from dietary sources 37 , this may suggest that disturbances occurring in these organs may underlie the development of kwashiorkor. Moreover, network analysis showed that lLV5 is linked to total serum cholesterol, lLV10 (phosphatidylethanolamine lipid species), and lLV13 (unsaturated sphingomyelins), indicating that lLV5 reflects a pathway involved in lipid transport, potentially in the liver. There are known interactions between ECM proteins and sphingolipids. Gelsolin for instance can bind sphingosine-1-phosphate (S1P) 38 , a sphingolipid mediator formed by the metabolism of sphingomyelin 39 . Insufficient lymphatic S1P signaling was found to contribute to lymphatic dysfunction in lymphedema 40 . S1P also regulates matrix metalloproteinases, which are key enzymes involved in the degradation of ECM proteins 41 . However, our results demonstrated neither direct nor indirect (partial correlation network analysis) association between pLV7 and lLV5. This could be due to lack of statistical power given our small sample size. However, both pLV7 and lLV5 are associated with total plasma protein content. This shared association could suggest that the dysfunction observed in kwashiorkor is either driven by protein deficiency and/or, alternatively, that upstream disturbances affecting ECM remodeling and sphingolipid homeostasis may contribute to a reduction in plasma protein levels. The ECM is closely linked with inflammation; its remodeling and degradation are often triggered by inflammatory signals. On the other hand, sphingolipid homeostasis is directly linked to inflammation as its metabolism results in the production of either pro- or anti-inflammatory lipids, and specific sphingolipids, especially S1P that regulates vascular inflammation and permeability 42 , potentially leading to endothelial leakage. In fact, animal models of both acute and chronic inflammation have demonstrated that plasma S1P limits the disruption of vascular endothelial monolayers and reduces edema 42 . This suggests that the balance between inflammation and sphingolipid homeostasis is critical: dysregulation in this system may lead to an edematous reaction, while systemic imbalances in sphingolipid levels could exacerbate ECM breakdown and intensify inflammatory responses. Albumin synthesis and circulating concentration declines during active inflammation, being a negative acute phase protein. Thus, the low circulating protein levels observed in kwashiorkor may not only reflect inadequate dietary intake but also a consequence of the inflammatory response itself 43 . While our findings do not provide a complete picture of kwashiorkor pathophysiology, they highlight potential pathways that may contribute to its underlying mechanisms. A proposed mechanistic framework based on these observations is presented in Fig. 7 . Further research is needed to validate and refine this proposed model. Moreover, the specific inflammatory trigger in this framework is currently unknown. Our untargeted semi-polar metabolomics results did not reveal metabolites nor metabolome-LVs associated passing all kwashiorkor biomarker criteria. However, a targeted metabolomics study comparing kwashiorkor and marasmus revealed major differences in their metabolome, where 128 of 141 (91%) metabolites quantified were lower in serum of kwashiorkor than marasmus 44 . Another quantitative metabolite profiling study found reduced levels of metabolites involved in one-carbon metabolism in serum of kwashiorkor and marasmic-kwashiorkor compared to marasmus, especially levels of methionine and choline 14 . However, our previous targeted metabolomics study comparing kwashiorkor and marasmus matched for serum albumin concentrations showed an almost complete overlap of serum amino acids concentrations in both groups, except for phenylalanine which was lower in kwashiorkor 45 . Similar to our current study, a previous untargeted metabolomics study did not find semi-polar metabolomic differences between Nigerian children with kwashiorkor and marasmus 46 . It is possible that the metabolome is more susceptible to geographic differences, leading to inconsistent results from studies in different countries. Moreover, given the increased overall fluid retention in kwashiorkor, it is also possible that small hydrophilic metabolites are dispersed both in the intravascular volume as well as the interstitial space, which makes the apparent concentration of these metabolites lower in kwashiorkor. Our results do not suggest an increase in the intravascular space in kwashiorkor. While the endothelial glycocalyx serves as a barrier for proteins from crossing the endothelial space to the interstitial space, the movement of metabolites across the endothelium remains to be deeply investigated. Nonetheless, more investigation is needed to ascertain whether specific metabolite concentrations indeed play a role in kwashiorkor pathophysiology. The role of the gut microbiome on the pathophysiology of kwashiorkor has previously been proposed. In a study of Malawian twins discordant for kwashiorkor, fecal transplantation of stool from children with kwashiorkor (n = 9) to germ-free mice led to phenotype of malnutrition where germ-free mice transplanted with stool from non-malnourished twin (n = 9) did not 5 . This study however did not include children with marasmus, and it is therefore difficult to determine whether the effect is due to kwashiorkor specifically or due to general malnutrition, or even clinical comorbidities. Also, edema in the mice models was not reported to have occurred. Similar to our findings, a previous study also did not find differences in gut microbiome between kwashiorkor and marasmus 46 . However, gut microbiome differences between children with kwashiorkor or marasmus and the NAM group cannot be excluded, given the small sample size, similar living conditions, and limited gastroenteritis cases. In this study, we matched participants using a pragmatic triage score designed to identify children in need for immediate care. However, we acknowledge that this score, and hence matching for it, may not be optimal or detailed enough. Moreover, this study included only 60 children per group, which only offered an exploratory perspective to this complex problem. We have previously described the barriers and enablers in undertaking omics research in low-resource countries 47 , such as Niger. Several logistic field operations may not be at par with quality standards observed in high income settings. Moreover, due to security reasons, we had to resort to training field staff remotely on certain aspects, such as collection of skin lesions data, which may have led to the low prevalence of detected cases. Research in other settings is needed to ensure reproducibility of the results. Nonetheless, the alignment of our current results with our previous work and that of others offers assurances on the validity of our findings. In conclusion, these observations suggest that acute malnutrition, and particularly kwashiorkor, is not simply a state of nutrient deficiency but a multifaceted physiological condition that reconfigures systemic molecular interactions. This results in increased ECM degradation and sphingolipid homeostasis dysregulation and their interaction plays significant roles in the pathophysiology of kwashiorkor, indicating that this disease could be linked to inflammation. Whether a specific inflammatory trigger is required to develop kwashiorkor or that children with kwashiorkor are predisposed to a derailed inflammatory response regardless of the stimuli remains to be investigated. Materials and methods Study Design and Participants This matched case-control study was designed to analyze and compare biological specimens, including blood, urine, and feces, collected from children with kwashiorkor, marasmus, and non-acutely malnourished (NAM). Participants were matched by age categories, sex, and triage score as a proxy of clinical severity, with children classified into groups based on edema levels (mild, moderate, and severe), which served as an indicator of kwashiorkor severity (see Table 2 ). Table 2 Recruitment groups and sample size Kwashiorkor Marasmus Non-malnourished + 20 Matching on: - Age - Sex - Severity at baseline (triage score) 20 20 ++ 20 20 20 +++ 20 20 20 Total 60 60 60 (+) feet/ankles only, (++) feet/ankles and legs, (+++)on other part of the body Biological samples were collected at various time points: (1) upon recruitment for all study participants, (2) at discharge from inpatient care, and (3) at exit from outpatient care (known in this context as an ambulatory therapeutic feeding center, ATFC). The collection frequency varied by participant group: NAM samples were collected once for children in outpatient care and at baseline and discharge from children in inpatient care, while kwashiorkor and marasmus groups underwent additional collections at the indicated discharge/exit points. Collected samples (venous blood, stools and urine) were transported in cold chain to the Epicentre Research Laboratory in Maradi, Niger, where they were stored in preparation for shipment to Belgium, following a -80°C cold chain protocol. Upon arrival, samples were distributed among collaborating laboratories for specific analyses, conducted according to the type of biological specimen and study objectives. Study setting The study was conducted from the 26th of September 2018 to the 5th of November 2019 in the Madarounfa Health District of the Maradi region, located in a remote rural area of south-central Niger, along the Nigerian border. The local livelihood mostly relies on subsistence agriculture and animal farming. Acute childhood malnutrition is endemic in the area with seasonal peaks from May until October when the food security deteriorates before the harvest. In collaboration with the Ministry of Health of Niger, Médecins Sans Frontières (MSF) is supporting the pediatric ward, the level 1 intensive care unit and the inpatient therapeutic feeding center (ITFC) of the Madarounfa District hospital as well as five ambulatory therapeutic feeding centres (ATFC) in the district. The study was approved by the Ethics Review Board of Médecins Sans Frontières (ID: 1785) and the Comité National D’Ethique pour la Recherche en Sante (CNERS) of the Ministry of Public Health in Niger (ID: 004/2018/CNERS and 026/2020/CNERS). All the study participants were treated for free according to the MSF and national guidelines. Study population This study recruited children with severe acute malnutrition aged 6 to 59 months eligible for new admission to either inpatient care in the ITFC of the Madarounfa District Hospital or enrolment in outpatient care in one of the ATFC supported by MSF in the region of Madarounfa. In addition, the study included non-acutely malnourished children from the paediatric ward of Madarounfa Health District Hospital or from the outpatient department of Madarounfa health centre. Children were classified according to their nutritional status and the severity of the oedema among children with kwashiorkor. Nutritional status was defined as follows: Kwashiorkor was defined by the presence of bilateral nutritional edema (mild to severe) - irrespective of either weight-for-height z-score (WHZ), mid-upper arm circumference (MUAC), presence of skin lesions, hair changes or other clinical features such as apathy, irritability, or fatty liver. The severity of edema was graded as mild (+) when it was present in feet/ankles only, as moderate (++) when it was present in feet/ankles and legs, and severe (+++) when it was visible on other part of the body (face, hands, arms, trunk etc) 48 . The presence of edema on feet/ankles was determined by applying moderate thumb pressure to the dorsum of both feet for at least three seconds. If bilateral pits remained after the thumbs have been removed, then the child would be considered as having nutritional edema. All children underwent a medical history assessment and a clinical examination (by two doctors for most hospitalized children). To avoid the inclusion of edema of non-nutritional origin, edematous children who could provide urine sample underwent a urine test strip and only children with negative results (i.e. negative for protein and blood) were included in the Kwashiorkor group. The presence of albumin in urine during renal disease/nephrotic syndrome contrasts with kwashiorkor where it is usually absent or present in very small amounts. Similarly, blood is found in nephritis but most often absent in kwashiorkor. Marasmus , or severe wasting, was defined as a WHZ less than − 3 49 , and/or a MUAC less than 115 mm (in children 6–59 months) and the absence of edema (nutritional or not). Non (acutely)-malnourished (NAM) status was defined by the absence of acute malnutrition, that is: absence of nutritional edema, and a WHZ ≥ -2 and a MUAC ≥ 125mm (in children 6–59 months). Inclusion criteria Children were eligible for study participation when they fulfilled the following inclusion criteria: Aged 6 to 59 months at the time of inclusion. Diagnosed with severe acute malnutrition (kwashiorkor or marasmus) or non-acutely malnourished (see above definitions). Medical condition allowing the collection of 2 mL of blood in the 12 hours following admission and before any intravenous or intramuscular antibiotic treatment. Consent signed by the child’s parent (i.e. the mother or the father) or the child’s main caretaker if the child is an orphan. Newly admitted (i.e. not having been enrolled/admitted in the nutrition programme in the last 2 months nor in the study). Direct admission at the ITFC (i.e. not transferred to the ITFC while under treatment in an ATFC) or direct admission at the ATFC (i.e. not transferred from the ITFC). Child’s parent or main caretaker intending to remain with the child in the study area until the end of the study follow-up period. Absence of any exclusion criteria. Exclusion criteria Children were excluded from participation in case of any of the following: Medical condition not allowing the collection of 2 mL of blood in the 12 hours following admission, based on the judgement of the clinician attending the child. Medical condition not allowing the collection of blood before any type of drug or fluid is administered via the intravenous (IV) catheter (when a catheter is placed upon admission), based on the judgement of the clinician attending the child. Intravenous or intramuscular antibiotics given in the 12 hours prior to the inclusion blood sampling. For artesunate injection, this was limited to 4 hours. Informed consent not provided. Previous enrolment/admission in the nutrition programme in the last 2 months (including relapse or return after default) or in the study. Current participation in another clinical research study including blood sampling and/or investigational products. Child’s parent or main caretaker not intending to remain with the child in the study area until the end of the study follow-up period. Current edema of non-nutritional origin (based on the judgement of the attending clinician or with positive urine test strip (for proteinuria ≥ 0.3g/L (1+) and/or blood red cells or hemoglobin). Not “directly admitted” (e.g. children under treatment at the ATFC who are secondarily transferred to the ITFC due to clinical or nutritional deterioration; children transferred from the ITFC to the ATFC after nutritional stabilisation). Previous history of kwashiorkor (for patients admitted in the marasmus or non-malnourished groups). Known congenital abnormalities or a known chronic disease (including known tuberculosis, sickle cell disease, congenital heart disease…). Known HIV-positive status, clinical AIDS or having a known HIV-positive mother. [Considering the low HIV prevalence in Niger (estimated at 0.2% among adults 15–49 year-old in 2012 in the Maradi Region 50 ), there was no systematic HIV screening]. Admitted into inpatient care for trauma, burns, intoxication or surgery. Clinical biochemistry analyses Urine sodium, potassium, chloride, and creatinine levels were measured using an Abbot Architect Clinical Chemistry Analyzer (Abbott, Illinois, USA). Urinary albumin was measured using the Beckman Paragon Electrophoresis system. Plasma triglycerides, cholesterol, HDL-cholesterol, total bilirubin, total protein, albumin, aspartate transferase (AST), alanine aminotransferase (ALT) and alkaline phosphatase (ALP) levels were determined using colorimetric methods and C-reactive protein (CRP) using a particle enhanced turbidimetric immunoassay (PETIA) technique all on a Siemens Dimension Xpand autoanalyzer (Siemens Healthcare, Camberley, UK) following manufacturer instructions. i-STAT point-of-care analyzer (Abbott Point of Care, Maidenhead, UK) used with the cartridge CHEM8+ (providing simultaneously glucose, kidney function [urea nitrogen, creatinine], hematology [hematocrit, hemoglobin], electrolytes [sodium, potassium, chloride, ionized calcium], anion gap) and the cartridge CG4+ (providing simultaneously lactate, blood Gases [PCO2, PO2, TCO2, sO2], pH, HCO3, base Excess). Plasma and urine untargeted metabolomics Plasma samples were thawed on ice before metabolite extraction. Ice-cold methanol was used as the extraction solvent, added at a 2:1 ratio (1 mL solvent to 500 µL plasma). After thorough vortexing, the samples were stored at − 20°C for one hour to allow protein precipitation. Following this, the samples were centrifuged at 13,000 × g for 10 minutes at 4°C. The supernatant (750 µL) was transferred to a glass tube and dried under a gentle nitrogen stream at temperatures below 10°C. The resulting dried residue was reconstituted in 10% acetonitrile containing 0.1% formic acid for subsequent liquid chromatography-mass spectrometry (LC-MS) analysis. To create a quality control (QC) sample, 10 µL aliquots from each sample were pooled together. A dilution series of the QC sample was then prepared by mixing it with 0.1% formic acid at different ratios (1:1, 1:5, 1:10, and 1:50). Chromatographic separation was performed on a Waters Acquity UPLC I-class FTN system (Waters, Manchester, UK), with dynamic mobile phase gradients and flow rates, as shown in our previously published paper 51 to optimize peak separation. The mobile phases consisted of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B), with separation achieved using an Acquity UPLC HSS T3 column (1.8 µm, 100 Å, 1 mm × 100 mm). The injection volume was set at 5 µL, with the column maintained at 50°C. Liquid chromatography (LC) flow was directed into a Waters Synapt G2Si high-resolution mass spectrometer (MS) (Waters, Manchester, UK) using electrospray ionization (ESI) in both positive and negative modes. The capillary voltage was set at 2.75 kV for positive mode and 2.20 kV for negative mode. The source temperature was maintained at 150°C, with a desolvation temperature of 500°C. Gas flow rates were 20 L/h for the cone and 600 L/h for desolvation. The mass range covered 50–1000 Da, with a scan speed of 0.1 s in MS E centroid resolution mode. MSE collision energy was ramped from 10 to 30 V for both modes. Mass calibration was achieved using sodium formate adducts, and online correction was applied with leucine-enkephalin (200 pg/µL), infused at 20 µL/min every 10 seconds. A blank (100% acetonitrile) was injected 10 times at the beginning to condition the column followed by 10 injections of the QC sample. QC samples were re-injected after every 10 experimental samples. Experimental samples were analyzed in random order. Data acquisition was monitored using MassLynx v1.7 (Waters, UK). Data processing: chromatogram peak picking, deconvolution and peak alignment were achieved using the proprietary Progenesis QI software (Nonlinear Dynamics, Newcastle, UK). Plasma lipidomics The lipid extraction was carried out using the Protein Precipitation Liquid Extraction Protocol (Jenkins, Ronis & Koulman, 2020). A 50 µL aliquot was taken from each plasma sample for the lipidomics analysis. To each aliquot, 650 µL of chloroform solution and 100 µL of LIPID Internal Standards (Lipid-IS) were added. The samples were vortexed, followed by the addition of 250 µL of methanol. After another round of vortexing, 400 µL of acetone was added. The samples were then vortexed again and centrifuged at ~ 20,000× g for 10 minutes to remove the insoluble material. The monolayer supernatant was carefully collected. The organic extracts (chloroform, methanol, and acetone) were dried using a Concentrator Plus system (Eppendorf, Stevenage, UK) for 60 minutes at 60°C. The dried samples were reconstituted in 100 µL of a 2:1:1 mixture of propan-2-ol, acetonitrile, and water, and thoroughly mixed. The reconstituted samples were transferred into vials for analysis by LC-MS, and for each sample the volume of injection was 10 µL. Chromatographic separation of lipids was performed using an HPLC system (Shimadzu UK Limited, Milton Keynes, UK). The stationary phase was a Waters Acquity UPLC® CSH C18 column (Waters, Hertfordshire, UK; 1.7 µm, I.D. 2.1 mm × 50 mm), maintained at 55°C. The mobile phase consisted of two solutions: solution A (a 6:4 mixture of acetonitrile and water) and solution B (a 9:1 mixture of isopropanol and acetonitrile), both containing 10 mM ammonium formate. The separation was carried out at a flow rate of 0.6 mL/min and the gradient used followed the method described by (Jenkins, Ronis & Koulman, 2020). MS analysis was performed on a Thermo Scientific Exactive Orbitrap instrument with a heated electrospray ionization source (Thermo Fisher Scientific, Hemel Hempstead, UK). The ion source parameters were as follows: capillary temperature: 300°C, source heater temperature: 420°C, sheath gas flow: 40 (arbitrary), auxiliary gas flow: 15 (arbitrary), spare gas: 3 (arbitrary), source voltage: 4 kV. The mass spectrometer scan rate set at 4 Hz, giving a resolution of 25,000 (at 200 m/z) with a full-scan range of m/z 100 to 1800 with continuous switching between positive and negative mode. The raw data were processed using MS-DIAL software (version 4.9), which included peak detection, alignment, and data annotation (Tsugawa et al., 2020). The following parameters were applied: Data collection: MS1 mass range: 120–1800 Da; retention time range: 0–10 min. Peak detection: minimum peak height: 10,000 amplitude units; mass window width: 0.1 Da; smoothing method: linear weighted moving average; smoothing level: 3 scans. Identification: retention time tolerance: 0.25 min; accurate mass tolerance: 0.006 Da; identification score threshold: 70%. Alignment: retention time tolerance: 0.25 min; MS1 tolerance: 0.005 Da. Annotation of lipid molecular species was based on both accurate mass data acquired by mass spectrometry (MS) and specific retention times of each lipid from liquid chromatography (LC). The lipid database used was validated using Lipid Maps ( http://www.lipidmaps.org/tools/ms ). Lipid semi-quantification was performed using deuterated internal standards (Lipid-IS) for each lipid class, based on the known concentrations of the Lipid-IS added to each sample. This analytical process resulted in matrices containing the concentrations of annotated lipid molecular species, expressed in µM. Plasma untargeted proteomics Bottom-up proteomics sample preparation was done in fully randomized fashion using a urea-based in-sol digest. To 2 µL of neat plasma, 300 µL of denaturing buffer (8 M urea (Chem-Lab Analytical, Zedelgem, Belgium) in 1 M ammonium bicarbonate (Sigma-Aldrich, Saint Louis, USA)) was added, followed by three cycles of 3 min-ON/3 min-OFF sonication and vortexing. To reduce proteins, dithiothreitol (MP Biomedical, Irvine, USA) was added to a final concentration of 5 mM followed by incubation at 37°C for 30 min. After cooling down to room temperature, alkylation was done using iodoacetamide (Sigma-Aldrich, Saint Louis, USA) added to a final concentration of 15 mM and incubation at room temperature for 30 min in the dark. Next, the tubes were diluted to 2 M urea with 1 M ammonium bicarbonate buffer and 3 µg of Trypsin/LysC (Promega, Madison, USA) was added to achieve an estimated 1:50 wt/wt ratio of enzyme to protein. This was incubated overnight at 37°C after which resulting peptides were desalted using C18 reverse phase SPE cartridges (Strata-X, Phenomenex, Torrance, USA). The cartridges were conditioned by adding 1 mL of methanol (Biosolve, Valkenswaard, The Netherlands) and equilibrated with two additions of 1 mL ultrapure water. Samples were loaded slowly onto the cartridges and washed two times with 1 mL of 5% methanol in water before eluting in 1% formic acid (Chem-Lab Analytical, Zedelgem, Belgium) in methanol buffer and drying using a vacuum concentrator. Vacuum dried peptide samples were resuspended in 0.1% formic acid in LC-MS grade water (Biosolve, Valkenswaard, The Netherlands) and centrifuged at 16 000 x g for 10 min. The clear supernatant was transferred into vials and quantified using the Lunatic UV/Vis spectrophotometer (Unchained Labs, Pleasanton, USA). The resuspension volumes were then corrected to achieve a peptide concentration of 2 µg/µL in every sample vial. Finally, a quality control sample was prepared by pooling an equal volume from all samples. Peptide samples were acquired in randomized order, alternating between sample and blank injections to prevent carry-over, and interspersed with a calibration standard (PepCal, Sciex, Massachusetts, USA) every 5 samples and quality control mixtures every 10 samples: 1 µL of sample was injected onto a Eksigent NanoLC 425 HPLC system (Sciex, Massachusetts, USA) equipped with a Triart C18 trapping column (5 mm × 0.5 mm, YMC) and a Luna Omega Polar C18 column (150 mm × 0.3 mm, particle size 3 µm, Phenomenex), operating in capillary flow mode (5 µL/min), coupled to a TripleTOF 6600 + mass spectrometer (Sciex, Massachusetts, USA) with the Optiflow TurboV ion source operating in positive mode. A 20-minute active gradient from 3 to 30% B was utilized, with mobile phases A and B consisting of 0.1% formic acid in water (Biosolve, Valkenswaard, The Netherlands) and 0.1% formic acid in acetonitrile (Biosolve, Valkenswaard, The Netherlands), respectively. Column temperature was maintained at 30°C. A SWATH 99 variable window acquisition scheme with a total cycle time of 4 sec was used. MS1 spectra spanned a mass range of 400–1200 m/z in high sensitivity mode with 250 ms accumulation time and MS2 spectra were acquired in high sensitivity mode from 100–1500 m/z in 37.5 ms per scan. Ion source parameters were set to 4.5 kV for the ion spray voltage, curtain gas at 25 psi, nebulizer gas at 10 psi, heater gas at 20 psi, and 100°C as source temperature. Instrument maintenance was required in the middle of the batch. Peptide identification and quantification was done with DIA-NN 1.8 52 . An iterative search strategy was employed to ensure maximum coverage of the proteome by first doing four separate searches starting from four different databases: (i) a predicted library from the UniProt human proteome 53 (reviewed only, retrieved on 06/01/2022) concatenated with the GPM cRAP database, (ii) a predicted library from the PeptideAtlas 54 human plasma build (2021-07), (iii) a deeply fractionated empirical DDA plasma library (ZenoTOF 7600, Sciex, Massachusetts, USA), (iv) a predicted library from an annotated secreted protein database, i.e., supplementary file 3 from Harney et al 55 . For all searches, the following settings were used: enzyme Trypsin/P allowing for 2 missed cleavages, cysteine carbamidomethylation as fixed modification and up to 1 variable modification from methionine oxidation and protein N-term acetylation, peptide length of 6–35 with charge range 1–5, precursor m/z range 400 − 122, fragment ion m/z range 100–1500, MS1 accuracy of 12, MS2 accuracy of 25, scan window of 6, match-between-runs with smart profiling enabled and double-pass mode, heuristic protein inference based with proteotypicity definition based on gene names, and robust high precision quantification with normalization and MaxLFQ disabled. The four resulting spectral libraries were merged into one spectral library, which ended up containing 3529 protein isoforms, 2820 protein groups and 16595 precursors in 13930 elution groups. Finally, one search was performed on the entire dataset using the same settings as before but starting from the merged spectral library. The resulting precursor quantity matrix at a false discovery rate of 1% was subsequently used as input for data preprocessing. First, failed injections (1 sample and 2 QC runs) were dropped from the dataset and (potential) contaminant peptides were filtered out (33 out of 9365 peptides). Peptides were then manually matched to all possible proteins of origin using the UniProt human proteome database as described previously. Peptide abundance values were log 2 transformed and normalized using the RUV-III-C 56 algorithm to correct for instrument performance degradation. RUV-III-C used a grouping based on run order except for QC samples which were assigned to one group and peptides that differed most across the run order as determined from a two-sided unpaired t-test were used as negative controls. Finally, peptides not belonging to a smallest unique group were dropped and the normalized peptide abundances were summarized to the protein level using a robust summarization algorithm 57 . Fecal microbiome analysis Total DNA was extracted from the pellet of 200 mg fecal matter by means of bead beating with a PowerLyzer (Qiagen, Venlo, the Netherlands) and phenol/chloroform extraction 58 . The beads were precipitated and removed after centrifugation at maximal speed for 5 min. The supernatant was transferred to new tubes, followed by extraction and purification with 500 µL phenol: chloroform: isoamilyc alcohol 25:24:1 (pH 7) and 700 µL chloroform. 450 µL supernatant was transferred to new tubes containing 500 µL ice-cold Isopropyl alcohol and added 45 µL 3 M sodium acetate, followed by cooling at -20°C over 1 h. Isopropyl alcohol was discarded after centrifugation at maximum speed for 30 min and then air-dried DNA was re-suspended by adding 100 µL 1× TE buffer (10 mM Tris, 1 mM EDTA). The quality control of DNA extracts was performed by electrophoresis in a 1% (w/v) agarose gel (Life Technologies, Madrid, Spain). For each sample, a sequencing library was constructed starting from 12.5 ng input gDNA according to the ‘16S Metagenomic Sequencing Library Preparation’ guide (Illumina). Briefly, 16S rRNA gene fragments were PCR-amplified from genomic DNA for 25 cycles using the 341F (5’-CCT ACG GGN GGC WGC AG -3’) and 785Rmod (5’-GAC TAC HVG GGT ATC TAA KCC-3’) 59 primer pair. The PCR amplicons were purified using Ampure XP beads (Beckman Coulter). Sequencing adapters were added using the Nextera XT Index Kit (Illumina) with 8 PCR cycles and the resulting sequencing libraries were purified on Ampure XP beads. Quality was checked on a Bioanalyzer using a DNA 1000 chip (Agilent) and quantification was done using qPCR according to the ‘Sequencing Library qPCR Quantification Guide’ from Illumina. Finally, libraries were pooled equimolarly, spiked with 20% PhiX and sequenced as paired-end 150 on a MiSeq device (Illumina). Data preparation and metagenomics analyses were done using QIIME2 (v2020.2) 60 . For each sample, read pairs were quality-trimmed, denoised and reconstructed into 16S amplicons using the DADA2 algorithm. Operational taxonomic units (OTU) and their abundance were collected. Taxonomic assignment was done using the Greengenes 16S reference collection 61 (release 13.8). These were then used to calculate alpha diversity metrics (Shannon, Faith Phylogenetic Distance, Pielou Evenness, Observed OTUs) and beta diversity metrics (Weighted and Unweighted UniFrac, Bray-Curtis, Jaccard). Alpha diversity metrics were used to detect differences between sample groups (Kwashiorkor, Marasmus, NAM) by means of Kruskal-Wallis tests. Group comparisons using beta diversity metrics were done with the PERMANOVA module in QIIME2. We used the ANCOM module to identify differentially abundant features across sample groups at the taxonomic level of family, genus and strain. We used OTU abundancies with the full pipeline of the PICRUSt2 module to perform a functional composition analysis at the taxonomic family, genus and strain levels, based on KEGG orthologs, Enzyme Classification numbers (EC) and MetaCyc pathways. Statistical analysis Univariable analyses Scales and log-transformed levels of individual metabolite, lipid and protein features were compared between kwashiorkor and combined marasmus and NAM using conditional logistic regression ( clogit ) analysis to account for the age, sex and clinical severity matching. Significance was accepted at p < 0.05 after Benjamini Hochberg False Discovery Rate (FDR) correction 62 . Significant features were then forwarded to succeeding analysis assessing their individual levels association with severity of edema using ordinal regression using the polr function of the MASS package in R. Features showing a relationship with edema severity where assessed for their change over time using linear mixed model, where the features were tested for association with timepoint, age and sex, with subject as random effect. This was done using the lmerTest 63 package in R. Factor analysis Latent variables (LV) were calculated separately for the metabolome, lipidome and proteome for the samples collected at baseline. Factor analysis was performed using the matrix-free likelihood method of extraction followed by the “varimax” rotation using the fad 64 package in R. The Bayesian information criterion (BIC) was used to determine the optimal number of latent variables to extract. We imposed a factor loading cut-off of 0.4, which means that features that have less than 0.4 loadings to the latent variable were removed - leaving behind only those features that are contributing highly to the latent variable. Factor scores were then calculated from these variables using the psych package in R, using the Bartlett method. Factor scores were calculated for the different timepoints: baseline, hospital discharge and full nutritional and clinical recovery. These factor scores were then used for subsequent analysis to represent the entire LV, either for metabolome (mLV), proteome (pLV) or lipidome (lLV). mLV, lLV and pLVs were compared between kwashiorkor and combined marasmus and NAM using conditional logistic regression ( clogit ) analysis to account for the age, sex and clinical severity matching. Significance was accepted at p < 0.05 after Benjamini Hochberg False Discovery Rate (FDR) correction 62 . Significant LVs were then forwarded to succeeding analysis assessing their individual levels association with severity of edema using ordinal regression using the polr function of the MASS package in R. LVs showing a relationship with edema severity where assessed for their change over time using linear mixed model, where the features were tested for association with timepoint, age and sex, with subject as random effect. This was done using the lmerTest 63 package in R. As sensitivity analyses, the analyses were reperformed after removing children with malaria and again after only including children with negative blood or protein in the urine measured using a urine test strip. Partial correlation network integration Integration of the different omics domains were performed at the LV level using partial correlation network analysis using the qgraph 65 package in R. mLV, lLV and pLV were combined with clinical biochemistry results to show inter-LV association with known clinical markers. Data availability Epicentre and MSF are committed to sharing and disseminating health data from its programs and research in an open, timely, and transparent manner to promote health for populations while respecting ethical and legal obligations towards patients, research participants, and their communities. For the purpose of this study, Epicentre and MSF collected identifiable participant data. Upon publication and for as long as the ethical authorisation permits it, the minimal data set underlying the findings of this study will be made available on request. If scientifically relevant, the request may be granted in accordance with legal framework set forth by MSF data sharing policy available on its website or on request. The MSF data sharing policy ensures that data will be available upon request to interested researchers while addressing all security, legal, and ethical concerns. All data access request for non-commercial and academic research can be addressed to [email protected] , [email protected] or the corresponding authors. Such request will be submitted to the medical departments of MSF and Epicentre. In case of approval of the request, the data will be shared with researchers, subject to the establishment, within a reasonable timeline, of a data sharing agreement to provide the legal framework for data sharing – including any applicable data protection laws. Such data sharing agreement may differ depending on the nature of the data to be shared – pseudonymized or anonymized – and the sensitivity of the data. Analysis codes and summaries of metabolite, lipid, and protein features, and the gut microbiome analysis used in this study are publicly available at https://github.com/HOB-IDEALS/Kwash-Multi-omics . The data may be interactively visualized using this link: https://mudiboevans.shinyapps.io/MSFKwash/ Declarations Funding The field work in Niger, including participant recruitment and international shipment of biological samples, and analysis of the metabolome, proteome, and gut microbiome were funded by the Médecins Sans Frontières International Innovation Fund. Analysis of lipidome and blood biochemistry, as well as postdoctoral fellowship of GB Gonzales were funded by the Research Foundation Flanders (FWO; 3E020617 and 31501220), Ghent University Global Minds Fund provided travel funding for GB Gonzales. Acknowledgement We express our deepest gratitude to the participants and the communities in Niger whose trust and collaboration made this study possible. We sincerely thank the laboratory technicians, physicians and study nursing team in Niger for their dedication and meticulous work throughout all phases of clinical care, sample processing and data collection. We further thank Médecins Sans Frontières–Operational Centre Paris (MSF OCP) in Niger for their operational support and collaboration. We extend special appreciation to Alexandra Ascorra, Brigite Chokote, Iris van den Boomgaard, Emmanuel Berbain, Evans Mudibo and Mario-David Barbagallo for their guidance, coordination, and assistance throughout the project. References Briend A (2014) Kwashiorkor: still an enigma - the search must go on. Alvarez J, Dent N, Browne L, Myatt M, Briend A (2016) Putting Child Kwashiorkor on the map Williams C (1935) Kwashiorkor: a nutritional disease of children associated with a maize diet. Lancet 226(5855):1151–1152 Michael H, Amimo JO, Rajashekara G, Saif LJ, Vlasova AN (2022) Mechanisms of Kwashiorkor-Associated Immune Suppression: Insights From Human, Mouse, and Pig Studies. Front Immunol ; 13 Smith MI, Yatsunenko T, Manary MJ et al (2013) Gut Microbiomes of Malawian Twin Pairs Discordant for Kwashiorkor. Science 339(6119):548–554 Coulter JBS, Hendrickse RG, Lamplugh SM et al (1986) Aflatoxins and kwashiorkor: clinical studies in Sudanese children. Trans R Soc Trop Med Hyg 80(6):945–951 de Vries HR, Maxwell SM, Hendrickse RG (1990) Aflatoxin excretion in children with kwashiorkor or marasmic kwashiorkor–a clinical investigation. Mycopathologia 110(1):1–9 McMillan A, Renaud JB, Burgess KMN et al (2018) Aflatoxin exposure in Nigerian children with severe acute malnutrition. Food Chem Toxicol 111:356–362 Oyelami OA, Maxwell SM, Adelusola KA, Aladekoma TA, Oyelese AO (1998) Aflatoxins in autopsy kidney specimens from children in Nigeria. J Toxicol Environ Health A 55(5):317–323 Soriano JM, Rubini A, Morales-Suarez-Varela M, Merino-Torres JF, Silvestre D (2020) Aflatoxins in organs and biological samples from children affected by kwashiorkor, marasmus and marasmic-kwashiorkor: A scoping review. Toxicon 185:174–183 Abdallah MF, Briend A, Gonzales GB (2021) Aflatoxins and kwashiorkor: A commentary on the association-causation conundrum. Toxicon 190:20–21 Holt L, Snyderman S, Norton P, Roitman E, Finch J (1963) The plasma aminogram in kwashiorkor. Lancet 282(7322):1343–1348 Roediger WE (1995) New views on the pathogenesis of kwashiorkor: methionine and other amino acids. J Pediatr Gastroenterol Nutr 21(2):130–136 May T, de la Haye B, Nord G et al (2022) One-carbon metabolism in children with marasmus and kwashiorkor. eBioMedicine ; 75: 103791 Gonzales GB, Njunge JM, Gichuki BM et al (2022) The role of albumin and the extracellular matrix on the pathophysiology of oedema formation in severe malnutrition. eBioMedicine 79:103991 Amadi B, Fagbemi AO, Kelly P et al (2009) Reduced production of sulfated glycosaminoglycans occurs in Zambian children with kwashiorkor but not marasmus. Am J Clin Nutr 89(2):592–600 Thul PJ, Lindskog C (2018) The human protein atlas: A spatial map of the human proteome. Protein science: publication Protein Soc 27(1):233–244 Kessler L, Daley H, Malenga G, Graham S (2001) The impact of HIV infection on the clinical presentation of severe malnutrition in children at QECH. Malawi Med journal: J Med Association Malawi 13(3):30–33 Fevang P, Havemann K, Fevang B, Høstmark AT (2018) Malaria and Malnutrition: Kwashiorkor Associated with Low Levels of Parasitaemia. Malaria Research and Treatment ; 2018: 7153173 Badaloo AV, Forrester T, Reid M, Jahoor F (2006) Lipid kinetic differences between children with kwashiorkor and those with marasmus. Am J Clin Nutr 83(6):1283–1288 Munthali T, Jacobs C, Sitali L, Dambe R, Michelo C (2015) Mortality and morbidity patterns in under-five children with severe acute malnutrition (SAM) in Zambia: a five-year retrospective review of hospital-based records (2009–2013). Archives of public health = Archives belges de sante publique ; 73(1): 23 Chakravarti S, Magnuson T, Lass JH, Jepsen KJ, LaMantia C, Carroll H (1998) Lumican regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of lumican. J Cell Biol 141(5):1277–1286 Pietraszek K, Chatron-Colliet A, Brézillon S et al (2014) Lumican: A new inhibitor of matrix metalloproteinase-14 activity. FEBS Lett 588(23):4319–4324 Bucki R, Levental I, Kulakowska A, Janmey PA (2008) Plasma gelsolin: function, prognostic value, and potential therapeutic use. Curr Protein Pept Sci 9(6):541–551 Hsieh CH, Wang YC (2022) Emerging roles of plasma gelsolin in tumorigenesis and modulating the tumor microenvironment. Kaohsiung J Med Sci 38(9):819–825 Piktel E, Levental I, Durnaś B, Janmey PA, Bucki R (2018) Plasma Gelsolin: Indicator of Inflammation and Its Potential as a Diagnostic Tool and Therapeutic Target. Int J Mol Sci ; 19(9) Asare-Werehene M, Communal L, Carmona E et al (2020) Plasma Gelsolin Inhibits CD8 + T-cell Function and Regulates Glutathione Production to Confer Chemoresistance in Ovarian Cancer. Cancer Res 80(18):3959–3971 Iram S, Rahman S, Choi I, Kim J (2024) Insight into the function of tetranectin in human diseases: A review and prospects for tetranectin-targeted disease treatment. Heliyon 10(1):e23512 Levick JR, Michel CC (2010) Microvascular fluid exchange and the revised Starling principle. Cardiovasc Res 87(2):198–210 Lin PK, Davis GE (2023) Extracellular Matrix Remodeling in Vascular Disease: Defining Its Regulators and Pathological Influence. Arterioscler Thromb Vasc Biol 43(9):1599–1616 Lloyd IJ (1990) Is the extracellular matrix an integral and dynamic component of the sodium and water homeostatic system? Med Hypotheses 33(3):197–199 Khalil M, Kabiel A, El-Khateeb S, Aref K, El Lozy M, Jahin S, Nasr F (1974) Plasma and red cell water and elements in protein-calorie malnutrition. Am J Clin Nutr 27(3):260–267 Wharton BA, Howells GR, McCance RA (1967) CARDIAC FAILURE IN KWASHIORKOR. Lancet 290(7512):384–387 Dolin CE, Arteel GE (2020) The Matrisome, Inflammation, and Liver Disease. Semin Liver Dis 40(2):180–188 Morales A, Lee H, Goñi FM, Kolesnick R, Fernandez-Checa JC (2007) Sphingolipids and cell death. Apoptosis 12(5):923–939 Albeituni S, Stiban J (2019) Roles of Ceramides and Other Sphingolipids in Immune Cell Function and Inflammation. In: Honn KV, Zeldin DC (eds) The Role of Bioactive Lipids in Cancer, Inflammation and Related Diseases. Springer International Publishing, Cham, pp 169–191 Kuo A, Hla T (2024) Regulation of cellular and systemic sphingolipid homeostasis. Nat Rev Mol Cell Biol 25(10):802–821 Bucki R, Kulakowska A, Byfield FJ et al (2010) Plasma gelsolin modulates cellular response to sphingosine 1-phosphate. Am J Physiol Cell Physiol 299(6):C1516–C1523 Mendelson K, Evans T, Hla T (2014) Sphingosine 1-phosphate signalling. Development 141(1):5–9 Kim D, Tian W, Wu TT-H et al (2023) Abnormal Lymphatic Sphingosine-1-Phosphate Signaling Aggravates Lymphatic Dysfunction and Tissue Inflammation. Circulation 148(16):1231–1249 Kim ES, Kim JS, Kim SG, Hwang S, Lee CH, Moon A (2011) Sphingosine 1-phosphate regulates matrix metalloproteinase-9 expression and breast cell invasion through S1P3-Gαq coupling. J Cell Sci 124(Pt 13):2220–2230 Maceyka M, Spiegel S (2014) Sphingolipid metabolites in inflammatory disease. Nature 510(7503):58–67 Fuhrman MP (2002) The albumin-nutrition connection: separating myth from fact. Nutr (Burbank Los Angeles Cty Calif) 18(2):199–200 Di Giovanni V, Bourdon C, Wang DX et al (2016) Metabolomic Changes in Serum of Children with Different Clinical Diagnoses of Malnutrition. J Nutr Gonzales GB, Njunge JM, Gichuki BM et al (2021) Albumin-dependent and independent mechanisms in the syndrome of kwashiorkor. medRxiv : 2021.05.31.21257914. McMillan A, Orimadegun AE, Sumarah MW et al (2017) Metabolic derangements identified through untargeted metabolomics in a cross-sectional study of Nigerian children with severe acute malnutrition. Metabolomics 13(2):13 Nacis JS, Kamande P, Toni AT et al (2024) Barriers and enablers to the effective implementation of omics research in low- and middle-income countries. Nat Biotechnol 42(6):988–991 Frontières MS (2014) Nutrition Guidelines WHO/UN Children’s Fund. WHO Child Growth Standards and the Identification of Severe Acute Malnutrition in Infants and Children (2009) (accessed November 2016 Institut National de la Statistique (INS) (2013) et ICF International. Enquête Démographique et de Santé et à Indicateurs Multiples du Niger 2012. Calverton, Maryland: INS et ICF International, Gonzales GB, De Saeger S (2018) Elastic net regularized regression for time-series analysis of plasma metabolome stability under sub-optimal freezing condition. Sci Rep 8(1):3659 Demichev V, Messner CB, Vernardis SI, Lilley KS, Ralser M (2020) DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat Methods 17(1):41–44 Consortium TU (2020) UniProt: the universal protein knowledgebase in 2021. Nucleic Acids Res 49(D1):D480–D9 Desiere F, Deutsch EW, King NL et al (2006) The PeptideAtlas project. Nucleic Acids Res 34(suppl1):D655–D8 Harney DJ, Larance M (2022) Annotated Protein Database Using Known Cleavage Sites for Rapid Detection of Secreted Proteins. J Proteome Res 21(4):965–974 Poulos RC, Hains PG, Shah R et al (2020) Strategies to enable large-scale proteomics for reproducible research. Nat Commun 11(1):3793 Sticker A, Goeminne L, Martens L, Clement L (2020) Robust Summarization and Inference in Proteome-wide Label-free Quantification. Mol Cell Proteom 19(7):1209–1219 De Paepe K, Kerckhof FM, Verspreet J, Courtin CM, Van de Wiele T (2017) Inter-individual differences determine the outcome of wheat bran colonization by the human gut microbiome. Environ Microbiol 19(8):3251–3267 Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, Glöckner FO (2013) Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res 41(1):e1 Bolyen E, Rideout JR, Dillon MR et al (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 37(8):852–857 DeSantis TZ, Hugenholtz P, Larsen N et al (2006) Greengenes, a Chimera-Checked 16S rRNA Gene Database and Workbench Compatible with ARB. Appl Environ Microbiol 72(7):5069–5072 Benjamini Y, Hochberg Y (1995) Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J Royal Stat Soc Ser B (Methodological) 57(1):289–300 Kuznetsova A, Brockhoff PB, Christensen RHB (2017) 2017 lmerTest Package: Tests in Linear Mixed Effects Models. ; 82(13): 26 Dai F, Dutta S, Maitra RA, Matrix-Free (2020) Likelihood Method for Exploratory Factor Analysis of High-Dimensional Gaussian Data. J Comput Graphical Stat 29(3):675–680 Epskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D (2012) qgraph: Network Visualizations of Relationships in Psychometric Data. J Stat Softw 48(4):1–18 Additional Declarations There is NO Competing Interest. Supplementary Files SuppFile1.Latentfactorcompositions.xlsx Latent factor composition Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8320069","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":560344274,"identity":"7747ba75-f7be-42d8-8132-b0c1adeac185","order_by":0,"name":"Fabienne Nackers","email":"","orcid":"","institution":"Epicentre","correspondingAuthor":false,"prefix":"","firstName":"Fabienne","middleName":"","lastName":"Nackers","suffix":""},{"id":560344275,"identity":"c7eae2fd-7f56-4ca8-ae7a-1a7901bd0e6b","order_by":1,"name":"Kirrily De Polnay","email":"","orcid":"","institution":"MSF Belgium","correspondingAuthor":false,"prefix":"","firstName":"Kirrily","middleName":"","lastName":"De Polnay","suffix":""},{"id":560344276,"identity":"38ea3ec2-e028-42d1-8083-0dcc938dd5f2","order_by":2,"name":"Ruben Almey","email":"","orcid":"https://orcid.org/0000-0001-6911-5904","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Ruben","middleName":"","lastName":"Almey","suffix":""},{"id":560344277,"identity":"0145f6bc-5add-428d-a7e1-15d8574a97e2","order_by":3,"name":"Simon Daled","email":"","orcid":"","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Simon","middleName":"","lastName":"Daled","suffix":""},{"id":560344278,"identity":"d816c08b-1209-4a17-b380-7bd3fac3fb49","order_by":4,"name":"Azahir Dan Djibo","email":"","orcid":"","institution":"MSF Niger","correspondingAuthor":false,"prefix":"","firstName":"Azahir","middleName":"Dan","lastName":"Djibo","suffix":""},{"id":560344279,"identity":"e76453f3-9d55-4500-8ee4-beaecaa01efe","order_by":5,"name":"Sarah DeSaeger","email":"","orcid":"","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Sarah","middleName":"","lastName":"DeSaeger","suffix":""},{"id":560344280,"identity":"da0e1b49-6227-4262-9d78-0dc406fdb935","order_by":6,"name":"Dieter Deforce","email":"","orcid":"https://orcid.org/0000-0002-0635-661X","institution":"UGent","correspondingAuthor":false,"prefix":"","firstName":"Dieter","middleName":"","lastName":"Deforce","suffix":""},{"id":560344281,"identity":"3d1747d6-f136-418b-b340-ca67763339d7","order_by":7,"name":"Maarten Dhaenens","email":"","orcid":"https://orcid.org/0000-0002-9801-3509","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Maarten","middleName":"","lastName":"Dhaenens","suffix":""},{"id":560344282,"identity":"267cbefd-4196-4391-8139-f04b32072b8e","order_by":8,"name":"Yannick Gansemans","email":"","orcid":"https://orcid.org/0000-0001-5018-044X","institution":"Laboratory of Pharmaceutical Biotechnology, Faculty of Pharmaceutical Sciences, Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Yannick","middleName":"","lastName":"Gansemans","suffix":""},{"id":560344283,"identity":"900c22fc-828e-4782-b367-ab087fe50375","order_by":9,"name":"Griet Glorieux","email":"","orcid":"https://orcid.org/0000-0002-7641-4707","institution":"Ghent University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Griet","middleName":"","lastName":"Glorieux","suffix":""},{"id":560344284,"identity":"1f1e5af5-4cc5-4cab-9482-3413a62b9def","order_by":10,"name":"Ousmane Guindo","email":"","orcid":"","institution":"MSF Niger","correspondingAuthor":false,"prefix":"","firstName":"Ousmane","middleName":"","lastName":"Guindo","suffix":""},{"id":560344285,"identity":"c100d98c-2714-4a7a-9491-4a12a775a164","order_by":11,"name":"Benjamin Jenkins","email":"","orcid":"https://orcid.org/0000-0003-0038-9709","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Benjamin","middleName":"","lastName":"Jenkins","suffix":""},{"id":560344286,"identity":"a2c48859-02cd-41bd-99c2-1c5151cdbdf9","order_by":12,"name":"Albert Koulman","email":"","orcid":"https://orcid.org/0000-0001-9998-051X","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Albert","middleName":"","lastName":"Koulman","suffix":""},{"id":560344287,"identity":"20d2b340-c8b7-4db7-a200-bddd43ebd5f4","order_by":13,"name":"Hans-Joerg Lang","email":"","orcid":"","institution":"ALIMA","correspondingAuthor":false,"prefix":"","firstName":"Hans-Joerg","middleName":"","lastName":"Lang","suffix":""},{"id":560344288,"identity":"6bd9e2f4-ea72-43f1-b52f-299c8ed85d06","order_by":14,"name":"Celine Langendorf","email":"","orcid":"","institution":"Epicentre","correspondingAuthor":false,"prefix":"","firstName":"Celine","middleName":"","lastName":"Langendorf","suffix":""},{"id":560344289,"identity":"eea50823-e9a0-4b8f-83bb-830a3ee282a4","order_by":15,"name":"Debby Laukens","email":"","orcid":"https://orcid.org/0000-0002-1984-5850","institution":"Ghent University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Debby","middleName":"","lastName":"Laukens","suffix":""},{"id":560344290,"identity":"80314459-c01e-4cef-8d78-f30c19fa9153","order_by":16,"name":"Christopher Mambula","email":"","orcid":"","institution":"Médecins Sans Frontières, Paris, France","correspondingAuthor":false,"prefix":"","firstName":"Christopher","middleName":"","lastName":"Mambula","suffix":""},{"id":560344291,"identity":"aeb8c00e-b859-4b3e-a842-b849578742bb","order_by":17,"name":"Isabelle Mukagatare","email":"","orcid":"","institution":"MSF Niger","correspondingAuthor":false,"prefix":"","firstName":"Isabelle","middleName":"","lastName":"Mukagatare","suffix":""},{"id":560344292,"identity":"9c5471e0-36a1-41be-b9de-fb985120d497","order_by":18,"name":"Damon Parkington","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Damon","middleName":"","lastName":"Parkington","suffix":""},{"id":560344293,"identity":"7f866938-e12e-4d65-8617-780a2273663e","order_by":19,"name":"Elh Ibrahim Tassiou","email":"","orcid":"","institution":"MSF Niger","correspondingAuthor":false,"prefix":"","firstName":"Elh","middleName":"Ibrahim","lastName":"Tassiou","suffix":""},{"id":560344294,"identity":"3f6b4a7f-4a7e-4107-939b-2573f1bdb026","order_by":20,"name":"Wim Van Biesen","email":"","orcid":"","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Wim","middleName":"Van","lastName":"Biesen","suffix":""},{"id":560344295,"identity":"00207f41-2467-4b11-a291-ed4a46ec88b6","order_by":21,"name":"Tom Van de Wiele","email":"","orcid":"https://orcid.org/0000-0001-6854-5689","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Tom","middleName":"Van","lastName":"de Wiele","suffix":""},{"id":560344296,"identity":"dcaa2260-2ac5-48eb-8c79-3d83d9ed82db","order_by":22,"name":"Jill Vanmassenhove","email":"","orcid":"","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Jill","middleName":"","lastName":"Vanmassenhove","suffix":""},{"id":560344297,"identity":"23db81da-62c5-496e-99d4-8593957103eb","order_by":23,"name":"Filip Van Nieuwerburgh","email":"","orcid":"https://orcid.org/0000-0001-8815-5485","institution":"Ghent University","correspondingAuthor":false,"prefix":"","firstName":"Filip","middleName":"Van","lastName":"Nieuwerburgh","suffix":""},{"id":560344298,"identity":"2214487c-8b3b-4b14-99aa-68d70f81fb08","order_by":24,"name":"Irene Adehossi","email":"","orcid":"","institution":"MSF Niger","correspondingAuthor":false,"prefix":"","firstName":"Irene","middleName":"","lastName":"Adehossi","suffix":""},{"id":560344299,"identity":"9431f279-fc9f-4a95-af5c-3f7830f813ac","order_by":25,"name":"Andre Briend","email":"","orcid":"","institution":"University of Tampere","correspondingAuthor":false,"prefix":"","firstName":"Andre","middleName":"","lastName":"Briend","suffix":""},{"id":560344273,"identity":"6ff219a0-e0e0-4adc-afd1-b01a2178aad1","order_by":26,"name":"Gerard Bryan Gonzales","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEUlEQVRIie2PsWrDMBCGzxTiDkqzFgx+BhlDS2lwX6WHIZOHTKWjJndpyOrSl1DoklFGEC8iWQMe6gfo4NKlUENqxXQpSF076BtOQqePux/A4fiH3AxHoosnAKa4ZPrKzEo09NJj7ZUZFqIXbAoVP5+HKpHDH0r0rKKPOYjwciIb0XY75P6iarz11JzlLIuDAkR8xTa0fMpr5GSLzFMz85RHQgMCFfKSUTlmNa72WcS8XJqzKBJ/ER1B+q3sui3y1zetHGzKRXBUNoRKGAnke6IVYVlsdHdNaBr37rxc5Gl8rrKoQJWa45OTl5rcJyHdVavms0vCyYOi7fs6MSoAp7Rf7/fjrUUA8Btr2+FwOBzwDfqfYZ9p1c0WAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-6614-3520","institution":"Faculty of Medicine and Health Sciences, Ghent University","correspondingAuthor":true,"prefix":"","firstName":"Gerard","middleName":"Bryan","lastName":"Gonzales","suffix":""}],"badges":[],"createdAt":"2025-12-09 16:56:04","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8320069/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8320069/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98294223,"identity":"162ceddc-53f0-47bb-8bc5-00524c0188b9","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18755430,"visible":true,"origin":"","legend":"","description":"","filename":"Multiomicskwashiorkor.docx","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/1b14e6e19622792e1cd212ca.docx"},{"id":98435609,"identity":"630b6b53-dbcc-46d4-87ed-e47954670311","added_by":"auto","created_at":"2025-12-17 16:54:07","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs8320069.json","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/dddd0d9e6a4112435091a79b.json"},{"id":98294217,"identity":"4444dddb-b84e-4621-8de4-431c4e1d293d","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":334731,"visible":true,"origin":"","legend":"","description":"","filename":"rs83200692enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/82271ffc33f18b1d638049ac.xml"},{"id":98294221,"identity":"135c31c1-8ac3-4c98-8238-50b06483d540","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7340042,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/fdd86731c98a18385f58bc06.jpeg"},{"id":98435409,"identity":"d8bf8d64-33fa-4775-a61c-67c55b0fb2e0","added_by":"auto","created_at":"2025-12-17 16:53:42","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":222114,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/13d351c18bd20e4ff3664206.png"},{"id":98433737,"identity":"7bad3a73-469d-4853-99e3-8cbb4c6f8366","added_by":"auto","created_at":"2025-12-17 16:51:05","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":900306,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/0f7f4ebec714be956ea2ad3e.png"},{"id":98436627,"identity":"a51a4935-d236-4dc0-b12e-4df83759fa61","added_by":"auto","created_at":"2025-12-17 16:56:00","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1012618,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/90469a19bb43f2543cd6ef82.jpeg"},{"id":98436418,"identity":"6dbd0c12-1704-44e1-b46f-cf95d6aff52a","added_by":"auto","created_at":"2025-12-17 16:55:39","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":591985,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/73b703dd033e601efb71e050.jpeg"},{"id":98294236,"identity":"234ca7a0-ba4b-452c-879d-a30d80ac4878","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":486889,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/1b9554b2bd8609c0a770040a.jpeg"},{"id":98294229,"identity":"505eca38-cc4c-4a9c-9acf-a9ae434f21fb","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3097550,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/54ae953f369b37927f103571.jpeg"},{"id":98294240,"identity":"19b51bde-8610-467d-892a-e3b48227b0e4","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":707416,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/bea56bd573397b2596c4a6d6.jpeg"},{"id":98294227,"identity":"a2c30a48-b128-4bf9-bf62-45970ba97033","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41200,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/5cbaa4562c133d54ec4bd4d0.png"},{"id":98294220,"identity":"6c06bdea-52c8-4538-9418-38b13ff46635","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":147797,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/a14964e6066aa5e418cd0729.png"},{"id":98436298,"identity":"ccb8767d-2281-43ab-b056-c634513a8cba","added_by":"auto","created_at":"2025-12-17 16:55:19","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":208883,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/edee6ce4bf0dd6c4d780a5a6.png"},{"id":98436107,"identity":"0ea03582-2163-4205-8bcc-4cf53b18016a","added_by":"auto","created_at":"2025-12-17 16:54:56","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":68401,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/7b53913ff123cd8e84832833.png"},{"id":98435037,"identity":"581a9060-e840-47e2-8e87-bcec3d0a5d3d","added_by":"auto","created_at":"2025-12-17 16:52:59","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43237,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/649200505cdcc279b75f0f42.png"},{"id":98294234,"identity":"8aace93b-73d9-42ab-96de-9a6b8e6178cf","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":183002,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/0a7e938dd4c03c5274b3b543.png"},{"id":98436291,"identity":"f4333a49-e60c-4f51-a85e-01033604aceb","added_by":"auto","created_at":"2025-12-17 16:55:19","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108167,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/610b58957adc06e1a13f94bf.png"},{"id":98434824,"identity":"b20db86c-228e-459c-9c37-04fd87a5e9c1","added_by":"auto","created_at":"2025-12-17 16:52:38","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97043,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/928779ac5d32a579af4ad7b2.png"},{"id":98435404,"identity":"c9cffd67-a0fe-485c-aa89-61619bc23bbe","added_by":"auto","created_at":"2025-12-17 16:53:40","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115913,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/666d8c17ab18f0739f0845cc.png"},{"id":98294239,"identity":"50b400ac-e617-49fe-aec1-42fde409b544","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24244,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/c1676fc831c981c5556a2a41.png"},{"id":98294246,"identity":"f24a9132-a727-40dd-bb81-c420d09dc8c0","added_by":"auto","created_at":"2025-12-16 09:06:01","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":142763,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/7216a9ddd288898748e38f83.png"},{"id":98294238,"identity":"e77a9a80-ae22-440c-a4e7-cf8b1b52bc56","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":38092,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/44b28da009c8b20eefa53a63.png"},{"id":98294233,"identity":"8d4bb40a-4098-4084-974b-eb0bae6c2e87","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30805,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/88b831806b00692b1bae9728.png"},{"id":98294241,"identity":"8cb3144a-f4f6-4323-bb89-abf372adb200","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43933,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/e0dc1a48ef9c522712592b80.png"},{"id":98294242,"identity":"c5a63ef2-7726-4db7-83d5-ba29b23f6e7c","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"xml","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":330112,"visible":true,"origin":"","legend":"","description":"","filename":"rs83200692structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/437efba4718447de9c751c2e.xml"},{"id":98294244,"identity":"4d779337-5e1a-4312-8507-65f9335b45a5","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"html","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":356182,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/c51c24032f7d73bfd0d84807.html"},{"id":98294210,"identity":"65a9d207-3992-41bc-9c10-57c8242882d8","added_by":"auto","created_at":"2025-12-16 09:05:59","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":504876,"visible":true,"origin":"","legend":"\u003cp\u003eProteomics analysis of plasma samples from children with kwashiorkor, marasmus and non-acutely malnourished (NAM) at baseline, discharge from hospital and treated (either end of nutritional rehabilitation for malnourished children or end of treatment for hospitalized NAM). (a) Univariable analysis of proteins at baseline (n=180). (b) Proteins showing difference at FDR p \u0026lt; 0.05 were tested for association with edema severity among children with kwashiorkor (n=60). (c) proteome-related latent variables (pLV) associated with kwashiorkor, showing 7 out of 9 pLVs with difference at FDR p \u0026lt; 0.05 in kwashiorkor compared to marasmus and NAM combined. (d) The 7 pLVs were tested for association with edema severity among children with kwashiorkor, where only pLV7 remained of interest. (e) Trend of pLV7 at baseline (n=180), discharge from hospital (n=60) and at the end of ambulatory nutritional rehabilitation for kwashiorkor and marasmus (n=90), and end of treatment for NAM (n = 30) – collectively called \u003cem\u003etreated\u003c/em\u003e. pLV7 is significantly reduced at treatment compared to baseline among children with kwashiorkor. (f) Annotation of proteins involved in pLV7 showing the correlation of the individual proteins with pLV7. (g) Tissue localization of the pLV7-component proteins using the Human Proteome Atlas (\u003ca href=\"https://www.proteinatlas.org/\"\u003ehttps://www.proteinatlas.org/\u003c/a\u003e)\u003csup\u003e17\u003c/sup\u003e. Organs labeled with ❸ are those where all three proteins (lumican, gelsolin and tetranectin) are expressed. nTPM = normalized transcripts per million\u003csup\u003e17\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/138099e19cfd208b31c01b85.png"},{"id":98294209,"identity":"9120f7ab-d74c-42aa-bef9-ae58d883ae06","added_by":"auto","created_at":"2025-12-16 09:05:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":836701,"visible":true,"origin":"","legend":"\u003cp\u003eLipidomics analysis of plasma samples from children with kwashiorkor, marasmus and non-acutely malnourished (NAM) at baseline, discharge from hospital and treated (either end of nutritional rehabilitation for malnourished children or end of treatment for NAM). (a) Univariable analysis of lipids at baseline (n = 180). (b) Lipids showing difference at FDR p \u0026lt; 0.05 were tested for association with edema severity among children with kwashiorkor (n = 60). (c) Lipidome-related latent variables (lLV) associated with kwashiorkor, showing 9 out of 21 lLVs with difference at FDR p \u0026lt; 0.05 in kwashiorkor compared to marasmus and NAM combined. (d) The 9 lLVs were tested for association with edema severity among children with kwashiorkor, where only lLV5 and lLV16 remained of interest. (e) Trend of level of lLV5 and 16 at baseline (n = 180), discharge from hospital (n = 64) and at the end of ambulatory nutritional rehabilitation for kwashiorkor and marasmus (n = 90), and end of treatment for NAM – collectively called \u003cem\u003etreated\u003c/em\u003e(n = 30). lLV5 is significantly increased at treatment compared to baseline among children with kwashiorkor. (f) Annotation of lipids involved in lLV5 showing the correlation of the individual lipids with lLV5.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/952076eb0d73c4fe4616b4bb.png"},{"id":98435717,"identity":"d81bd4a7-3f21-47d6-b152-ded74c6e65cb","added_by":"auto","created_at":"2025-12-17 16:54:16","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":756915,"visible":true,"origin":"","legend":"\u003cp\u003eUntargeted semi-polar metabolomics analysis of plasma samples from children with kwashiorkor, marasmus and non-acutely malnourished (NAM) at baseline, discharge from hospital and treated (either end of nutritional rehabilitation for malnourished children or end of treatment for NAM). (a) Univariable analysis of metabolite features at baseline (n = 180). (b) Metabolite features showing difference at FDR p \u0026lt; 0.05 were tested for association with edema severity among children with kwashiorkor (n = 60). (c) Metabolome-related latent variables (mLV) associated with kwashiorkor, showing 2 out of 16 mLVs with difference at FDR p \u0026lt; 0.05 in kwashiorkor compared to marasmus and NAM combined (n = 180). (d) The 2 lLVs were tested for association with edema severity among children with kwashiorkor, where none were found to have an association (n = 60).\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/b8c32f7a25279cc8bf063962.png"},{"id":98294213,"identity":"e9304711-30f5-4fec-b2d6-4bedd0e654cf","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":504416,"visible":true,"origin":"","legend":"\u003cp\u003eMulti-omics integration of plasma lipidome, (semi-polar) metabolome and proteome, and clinical biochemistry results. a. Partial correlation network analysis at baseline using a “spring” rendering (in \u003cem\u003epcor\u003c/em\u003e package in R). Plasma biochemical analyses: Alb – albumin; ALP – alkaline phosphatase; ALT – alanine transaminase; Blb – biliburin; Chl – cholesterol; CRP – C-reactive protein; HDL – high density lipoprotein; Prt – total protein; TAG – total triglyceride. Point-of-care analyses (iStat): Ann – anion gap; BUN: blood urea nitrogen; Ca – ionized calcium; Cl – chloride; Crt – creatinine; Glc – glucose; Hmg – hemoglobin; Hmt – hematrocrit; K – potassium; Na – sodium. b. Direct correlation among modules and clinical biochemistry parameters linked with pLV7 and lLV5. c. Direct correlation between pLV7 and lLV5. d. Plasma total protein concentration in the 3 groups (kwashiorkor, marasmus and NAM). The pink band represents children with the same total protein concentrations despite being in different groups.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/3b2189a58ce23da69bb3c5ba.png"},{"id":98436058,"identity":"616be086-4f0d-43f0-95f7-aae650b49f74","added_by":"auto","created_at":"2025-12-17 16:54:49","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":211258,"visible":true,"origin":"","legend":"\u003cp\u003eUntargeted semi-polar metabolomics analysis of urine samples from children with kwashiorkor (n = 24), marasmus (n = 32) and non-acutely malnourished (NAM) (n = 32) at baseline. Urine metabolome-related latent variables (uLV) associated with kwashiorkor, where no difference at FDR p \u0026lt; 0.05 in kwashiorkor compared to marasmus and NAM combined was found.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/c58d211ce9feffabcd1abb9d.png"},{"id":98294232,"identity":"af74c438-b980-41af-bb71-1129d180c922","added_by":"auto","created_at":"2025-12-16 09:06:00","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":921078,"visible":true,"origin":"","legend":"\u003cp\u003eGut microbiome differences between children with kwashiorkor (n = 14), marasmus (n = 13) and non-acutely malnourished children (n = 11). (a) Principal coordinates analysis using Bray-Curtis and Jaccard distances. (b) Comparison in alpha-diversity scores among the three groups. (c) Composition of gut microbiome in individuals with either kwashiorkor, marasmus or non-malnourished children.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/fbee308722dfaadcde168027.png"},{"id":98436647,"identity":"7e7e5222-7338-4c25-9630-8b8550e931d7","added_by":"auto","created_at":"2025-12-17 16:56:01","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":444640,"visible":true,"origin":"","legend":"\u003cp\u003eProposed mechanistic framework for the pathophysiology of kwashiorkor.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/45f3297d5f07c04817ba886e.png"},{"id":98445458,"identity":"954b8d64-0dc0-4528-983a-7a7550813b77","added_by":"auto","created_at":"2025-12-17 17:19:44","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5641252,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/2374314c-a14e-48f9-9999-5e4f21154a4c.pdf"},{"id":98436459,"identity":"c8c098b3-0dbd-4be0-b8be-204eeb91e9c7","added_by":"auto","created_at":"2025-12-17 16:55:41","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":59455,"visible":true,"origin":"","legend":"Latent factor composition","description":"","filename":"SuppFile1.Latentfactorcompositions.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/a27497b9dc9531f8b43259d3.xlsx"},{"id":98436479,"identity":"af9a076c-972f-4f88-8e1c-74934d3b2e51","added_by":"auto","created_at":"2025-12-17 16:55:44","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1552214,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-8320069/v1/4dcac87b6034b1cf505abbe0.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Multi-omics, multi-tissue analysis reveal role of extracellular matrix remodeling and lipid transport dysfunction in edematous malnutrition (kwashiorkor)","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEdematous malnutrition, also known as kwashiorkor, is a severe form of acute malnutrition characterized by bilateral edema, fatty liver, and skin changes. It is clinically and biochemically distinct from non-edematous malnutrition, also known as marasmus or severe wasting, evident as low weight-for-height (\u0026lt;-3 z-scores from the 2006 WHO standards), or a mid-upper arm circumference (MUAC) below 115 mm (children 6\u0026ndash;59 months), visible muscle atrophy, and loose skin\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Both conditions may co-occur where a malnourished child presents with both severe wasting and edema \u0026ndash; known as marasmic kwashiorkor. Kwashiorkor represents a critical public health issue in low-income countries although its actual prevalence is difficult to ascertain as it is often not captured in population-based surveys\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Despite its clinical recognition for over eight decades, the precise molecular mechanisms underlying the pathophysiology of kwashiorkor remain incompletely understood. This gap in knowledge impedes the development of targeted interventions and therapeutic strategies to effectively treat and prevent this devastating condition.\u003c/p\u003e \u003cp\u003ePreviously, kwashiorkor had been attributed to protein deficiency\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e; however, emerging evidence suggests that its etiology is multifactorial, involving a complex interplay of nutritional, and environmental factors. Recent studies have highlighted the potential roles of oxidative stress\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, immune dysregulation\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e, and alterations in the gut microbiome in the development of kwashiorkor\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Exogenous toxicants, including aflatoxins from staple foods like maize, groundnuts, sorghum, and millet often consumed in low and middle-incomes countries could also increase oxidative stress and explain the accumulation of liver fat in kwashiorkor. However, despite finding aflatoxins in biological samples of children with kwashiorkor\u003csup\u003e\u003cspan additionalcitationids=\"CR7 CR8 CR9\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e, a causal link has not be demonstrated as the possibility of reverse causality has not been ruled out\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Additionally, the dysregulation of metabolic pathways, particularly those related to amino acid metabolism\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e, one-carbon metabolism \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e, and the synthesis and breakdowns of extracellular matrix (ECM) proteins\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e and endothelial dysfunction\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, have been implicated in the disease process. Yet, the comprehensive molecular mechanisms remain to be fully elucidated.\u003c/p\u003e \u003cp\u003eTo gain a deeper understanding of the molecular underpinnings of kwashiorkor, we employed a comprehensive multi-omics approach, integrating metabolomics, lipidomics, proteomics, and gut microbiome data across multiple biological samples throughout the different treatment stages to compare it to marasmus and non-malnourished children and explore its multifaceted pathophysiology. The data may lead to the understanding of its etiology and, hopefully, to the discovery of more effective treatment strategies.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipant characteristics\u003c/h2\u003e \u003cp\u003eA total of 180 children aged 6\u0026ndash;59 months were enrolled in the study, with kwashiorkor (n\u0026thinsp;=\u0026thinsp;60), marasmus (n\u0026thinsp;=\u0026thinsp;60) and without acute malnutrition (n\u0026thinsp;=\u0026thinsp;60) matched on sex, age and clinical triage score. Majority (60%) of the children were female and age categories were evenly distributed at baseline (50% 6\u0026ndash;23 months, 50% 24\u0026ndash;59 months old). An even distribution was accomplished between the edema severity levels on baseline, except for one child who was recruited as presenting moderate edema (++) but was secondarily reconsidered as having low severity edema (+). Of the 60 kwashiorkor participants, 12 had at least one type of cutaneous lesion, and these children with skin lesions all had moderate to severe edema (++/+++), except for one child with the lowest edema severity score (+).\u003c/p\u003e \u003cp\u003eHeight-for-age z-scores indicated a substantial burden of stunting. Most children were not critically ill (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Supplementary table 1), with only one child per group classified as an emergency case based on study triage score at admission. NAM children were more likely to test positive for malaria using rapid diagnostic tests compared to children with kwashiorkor or marasmus and they also presented a higher Pediatric Early Warning Scores (PEWS) at baseline. Marasmic children suffered more commonly from diarrhea. Laboratory findings upon admission for the three groups are compared in the Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and additional clinical characteristics and socioeconomic indicators of the participants can be found in Supplementary table 1. Two kwashiorkor participants died.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics and outcome of matched study participants\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eKwashiorkor\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;60)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMarasmus\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;60)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo acute malnutrition (NAM) (n\u0026thinsp;=\u0026thinsp;60)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (months) mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.7 (7.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21.3 (9.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e25.0 (12.1)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003en (%) 6\u0026ndash;11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (18.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6 (10.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u0026ndash;23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27 (45.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (33.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24 (40.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e24\u0026ndash;35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24 (40.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26 (43.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (28.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e36\u0026ndash;47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (8.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8 (13.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e48\u0026ndash;59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5 (8.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex (n female, %)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e36 (60.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36 (60.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e36 (60.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriage score* n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen (non-urgent)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 (50.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 (50.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30 (50.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYellow (priority)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e29 (48.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29 (48.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29 (48.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRed (emergency)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInitial recruitment site n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChildren hospitalized\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChildren in ambulatory treatment facility n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30 (50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical characteristics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLevel of edema at baseline\u003c/p\u003e \u003cp\u003en (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60 (100.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (35.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e++\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19 (31.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e+++\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20 (33.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMiddle upper arm circumference (mm) Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e122 [115, 130]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e111 [107, 116.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e139 [133, 145.5]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWeight for height z-score Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-2.14 [-2.84, -1.04]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-3.61 [-4.23, -3.20]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-1.26 [-1.60, -0.77]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWeight for age z-score Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-3.25 [-4.11, -1.94]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-4.16 [-4.91, -3.58]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-1.81 [-2.36, -1.03]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeight for age z-score Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-3.29 [-4.09, -2.06]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-2.88 [-4.59, -1.95]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-1.79 [-2.36, -0.71]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEWS at baseline n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u0026ndash;2 (green)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e58 (96.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e57 (95.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e46 (76.7)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u0026ndash;4 (yellow)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11 (18.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;5 (orange/red)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (00.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (5.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTemperature\u0026thinsp;\u0026ge;\u0026thinsp;38 (\u0026deg; C) n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (5.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8 (13.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18 (30.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere clinical anaemia n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (6.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (23.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiarrhoea (\u0026ge;\u0026thinsp;3 stools) n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16 (26.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35 (58.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (23.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eO2 saturation\u0026thinsp;\u0026ge;\u0026thinsp;93% (ambient air) n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e58 (98.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58 (98.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e59 (98.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLaboratory\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMalaria rapid test (n positive, %)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (53.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29 (48.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e41 (73.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlucose mg/dL - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79 [68,87]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e88 [80,100]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e91.5 [81,107]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHaemoglobin g/dL - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.2 [7;1,10.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.9 [8.8,11.6]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9.2 [6.3,10.7]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHaematocrit % - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27 [21,31]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30 [27,34]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27 [19,32]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCreatinine mg/dL - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.2 [\u0026lt;\u0026thinsp;0.2,0.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.2 [\u0026lt;\u0026thinsp;0.2,\u0026lt;0.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.2 [\u0026lt;\u0026thinsp;0.2,\u0026lt;0.2]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eALT (U/L) - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e35 [23,58]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e18.5 [14,24]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20 [14,30]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCRP (mg/L) - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.1 [1.6,32.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12.4 [1.5,72.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e72.9 [16.0,141]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal bilirubin (mg/dL) - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3 [0.2,0.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.3 [0.2,0.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.4 [0.2,1.1]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlbumin (g/L) - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17.3 [13.2,21.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29.9 [21.4,34.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e31.9 [26.2,36.6]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSodium mmol/L - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e137 [134,140]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e135 [131,137]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e134 [132,137]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePotassium mmol/L - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.1 [2.7,3.8]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.4 [2.7,4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.8 [3.3,4.2]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLactate mmol/L - Median [IQR]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.2 [1.3,3.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.8 [1.3,2.6]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.9 [1.4,3.1]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical progress and outcome\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStay\u0026thinsp;\u0026ge;\u0026thinsp;1 days in intensive care n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (6.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (10.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 (13.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFinal outcome (n %)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCured\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51 (85.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50 (83.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60 (100)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeceased\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExit against medical advice\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (6.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (15.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot responding to nutrition treatment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (5.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e* The study triage score was adapted from the Emergency triage Assessment and Treatment (ETAT) score applied in the MSF program in Niger at the time of the study. Details can be found in Supplementary table 2. Two nutritional criteria (visible severe wasting, oedema on both feet) were not included in the study triage score, wherein a malnourished child presenting in yellow (priority) case indicates that the child had another priority signs besides having visible severe wasting or oedema on both feet. PEWS: Pediatric Early Warning Scores; IQR: Interquartile range.\u003c/p\u003e \u003cp\u003e \u003cb\u003eIntegrated plasma lipidome, metabolome, proteome and clinical biochemistry of children with kwashiorkor, marasmus and non-acutely malnourished\u003c/b\u003e \u003c/p\u003e \u003cp\u003eIn this study, we imposed strict criteria for the determination of kwashiorkor-associated pathways. Inferences on individual lipids, metabolites and proteins were obtained. Moreover, we also obtained inferences at the pathway level by consolidating correlated metabolites, lipid and proteins into uncorrelated latent variables (LV) using factor analysis (FA). Each omics domain was first analyzed separately. The use of factor analysis affords us to reduce the big omics datasets into explainable LVs, which can be used for further downstream analyses. Each LV represents an underlying explanatory relationship and interaction among its component features, hence these factors provide a singular value that potentially refer to an overall behavior of particular pathway(s). This allowed us to make inferences on pathway level, instead of the traditional single feature analysis, typical of omics studies.\u003c/p\u003e \u003cp\u003eIndividual plasma lipid, metabolite and protein features, and their resulting LVs, associated with kwashiorkor must pass the following criteria:\u003c/p\u003e \u003cp\u003eCriterion 1. The feature or LV must be different in kwashiorkor compared to marasmus and NAM at recruitment. Marasmus and NAM are grouped together in subsequent analyses. Analyses are matched for age, sex and clinical triage score.\u003c/p\u003e \u003cp\u003eCriterion 2. The feature or LV must be associated with the severity of edema (+, ++, +++) at recruitment.\u003c/p\u003e \u003cp\u003eCriterion 3. The feature or LV must change towards normalcy as the child with kwashiorkor recovers to a non-malnourished state following nutritional therapy.\u003c/p\u003e \u003cp\u003eThese strict criteria ensure that our findings are robust despite the small sample size of this study (n\u0026thinsp;=\u0026thinsp;180). However, analyses of the urine metabolome and gut microbiome only followed criteria 1 and 2 since these analyses were only performed on samples obtained at baseline.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePlasma proteomics\u003c/h3\u003e\n\u003cp\u003eUntargeted proteomics analysis revealed 121 proteins were upregulated and 44 were downregulated (false discovery rate [FDR] p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) out of 308 protein features in kwashiorkor compared to both marasmus and NAM combined (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). However, when tested for association with edema severity in kwashiorkor, none of these proteins were found to have passed criterion 2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003eFA was then employed to determine clusters of proteins that share a latent relationship, potentially belonging to the same pathway. FA reduced the proteomics data into 9 proteome-LVs (pLV). Component proteins in each LV are shown in Supplementary file 1. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec shows the differences in the pLVs between kwashiorkor and non-kwashiorkor (marasmus and NAM). Out of 9 pLVs, 7 were found to be different in kwashiorkor (pLV 1, 3, 4, 5 and 7 were higher in kwashiorkor, pLV2 and 9 were lower in kwashiorkor than non-kwashiorkor cases), and hence passed criterion 1. Regression analyses of these pLVs with kwashiorkor edema severity revealed that only pLV7 was significantly associated with increasing severity of edema in kwashiorkor at baseline (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed). Moreover, this pLV was also higher than in non-kwashiorkor children at baseline, but this level significantly reduced during and after nutritional therapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee), passing all 3 criteria for a candidate pathway associated with kwashiorkor.\u003c/p\u003e \u003cp\u003eUncovering the component proteins of pLV7 revealed that the extracellular matrix proteins lumican, gelsolin and tetranectin were involved in kwashiorkor pathophysiology. The high correlation between these individual proteins and pLV7 demonstrates that pLV7 was able to capture the behavior of these three proteins as a cluster (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ef). Using the Human Proteome Atlas\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, these three proteins appear to be found in many tissue types all over the body, and can be found co-expressed in 41 out of 50 human organs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eg). However, the organs where they are most highly expressed varied \u0026ndash; lumican in gall blader, gelsolin in heart muscle, tetranectin in adipose tissue.\u003c/p\u003e \u003cp\u003eBecause the NAM children were more likely to have a positive malaria rapid test on admission (Supplementary table 3), a first sensitivity analysis was conducted wherein children testing positive for malaria rapid diagnostic test were removed. Also, because not all children with kwashiorkor could produce urine on admission, and hence some were included without a urine strip result (Supplementary table 3), a second sensitivity analysis was conducted including only kwashiorkor children who tested negative for urine blood and protein. Both revealed that the association of pLV7 and kwashiorkor is robust, albeit with lower power due to the smaller sample size (especially for criterion 2) (Supplementary Fig.\u0026nbsp;1).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003ePlasma lipidomics\u003c/h3\u003e\n\u003cp\u003eUntargeted lipidomics analysis of plasma revealed 22 lipids were upregulated and 136 were downregulated (FDR p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) out of 506 lipid features in kwashiorkor compared to both marasmus and NAM combined (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). However, when assessed for their association with severity of edema in kwashiorkor, none of these lipids were found to have passed criterion 2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Hence, no individual lipid species was found to be strongly associated with kwashiorkor in this study.\u003c/p\u003e \u003cp\u003eFA reduced the lipidomics data into 21 lipid-LVs (lLV). Component lipids in each LV is shown in Supplementary file 1. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec shows the differences in the lLVs between kwashiorkor and non-kwashiorkor (marasmus and NAM). Out of 21 lLVs, 9 were found to be different in kwashiorkor (lLV 14 and 16 were higher in kwashiorkor, lLV 1, 2, 5, 10, 13, 15 and 20 were lower in kwashiorkor than non-kwashiorkor cases), and hence passed criterion 1. Regression analyses of these lLVs with kwashiorkor edema severity revealed that lLV5 and lLV16 were significantly associated with increasing severity of edema in kwashiorkor at baseline (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). However, only lLV5 was significantly changed throughout the course of the treatment (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee). At baseline, this lLV was lower in kwashiorkor than non-kwashiorkor cases, and its level increased in plasma after treatment making this lLV a strong candidate pathway associated with kwashiorkor.\u003c/p\u003e \u003cp\u003eThe component lipids of lLV5 are dominated by 17 mono- to di-unsaturated sphingomyelins (SM 35:1, 37:2, 38:1, 38:2, 39:1, 40:0, 40:1, 40:2, 41:0, 41:1, 42:0, 42:1, 42:2, 43:1, 44:0, 44:1, and 44:2), along with 5 polyunsaturated cholesterol esters (16:2, 18:3, 20:4, 20:5, 22:6), ganglioside GD3 (42:1), four saturated hexo-ceramides (39:0-OH, 40:0-OH: 41:0-OH and 42:0-OH), Ether-linked lysophosphatidylcholine (24:0) and phosphatidylcholines (20:0, 22:0, 22:3, 37:1). The high correlation between these individual lipids and lLV5 demonstrate that lLV5 was able to capture the behavior of these lipids as a cluster (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBecause the NAM children were more likely to have a positive malaria rapid test on admission (Supplementary table 3), a first sensitivity analysis was conducted wherein children testing positive for malaria rapid diagnostic test were removed. Also, because not all children with kwashiorkor could produce urine on admission, and hence some were included without a urine strip result (Supplementary table 3), a second sensitivity analysis was conducted including only kwashiorkor children with negative blood and protein urine test. Both revealed that the association of lLV5 and kwashiorkor is robust, albeit with lower power due to the smaller sample size (especially for criterion 2) (Supplementary Fig.\u0026nbsp;2).\u003c/p\u003e\n\u003ch3\u003ePlasma metabolomics\u003c/h3\u003e\n\u003cp\u003ePlasma untargeted (semi-polar) metabolomics yielded a total of 2,825 features for both positive and negative ionization modes combined. Conditional logistic regression to account for the matched design on age, sex and triage score revealed 60 metabolite features different in kwashiorkor (FDR p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) compared to both marasmus and NAM combined (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). However, when tested for association with edema severity in kwashiorkor, none of these metabolite features were found to have passed criterion 2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003eFollowing FA, 16 metabolome-LVs (mLV) were extracted, of which only mLV5 and mLV7 passed criterion 1. Component metabolite features in each LV is shown in Supplementary file 1. Upon investigation of their association with kwashiorkor edema severity during nutritional recovery, we found no plasma mLV passing all 3 criteria, and hence no semi-polar metabolome-specific pathways were found to be associated with kwashiorkor in this study.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003ePlasma multi-omics integration\u003c/h3\u003e\n\u003cp\u003eTo integrate the multiple omics studies and to contextualize the clinical meaning of the individual latent variables (LV), we used partial correlation network analysis to determine patterns of association among all modules and clinical biochemistry results for all children (kwashiorkor, marasmus and NAM) at baseline (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Plasma clinical biochemistry results are presented in Supplementary table 3. Nodes correspond to individual pLV, lLV, mLV or clinical biochemistry (both laboratory and point-of-care analysis) results for baseline plasma samples. Correlations are depicted by either a green line (indicating positive partial correlation) or a red line (negative partial correlation).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eLinkages from the network (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea) are isolated in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb to highlight the clinical and biochemical nodes associated with pLV7 and lLV5. As shown (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), pLV7 (which represents ECM proteins) was negatively associated with total plasma protein but positively associated with blood sodium concentration. On the other hand, lLV5 was positively associated with total plasma protein, total serum cholesterol, lLV10 and lLV13, and negatively associated with pLV4 and pLV5. lLV10 consists of 15 mostly phosphatidylethanolamine lipid species, whereas lLV13 is comprised of 5 unsaturated sphingomyelins. lLVs 10 and 13 were also found to be different in kwashiorkor compared to non-kwashiorkor, but their levels were not associated with kwashiorkor severity. pLV4 is composed of protein such as actin, complement component C7, protein virilizer homolog, angiotensinogen, carboxypeptidase N subunit 2, and hemoglobin subunit beta. pLV5 is composed of alpha-1-acid glycoprotein 1, complement C5, pigment epithelium-derived factor, leucine-rich alpha-2-glycoprotein, and zinc-alpha-2-glycoprotein. The composition of the different modules can be found in Supplementary file 1.\u003c/p\u003e \u003cp\u003elLV5 and pLV7 are not directly correlated with each other (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec) despite both being associated with total plasma protein. Consistent with most findings in the literature, including our previous studies, total plasma protein content is lower in kwashiorkor than in marasmus and NAM. However, there is a considerable overlap among the groups. Many children with marasmus and NAM had low total protein in their blood but did not manifest kwashiorkor syndrome. These associations remained robust in sensitivity analyses where children who tested positive for malaria in a rapid diagnostic test (Supplementary Fig.\u0026nbsp;3) were removed or that only children with kwashiorkor who tested negative for urine blood and protein using a urine test strip (Supplementary Fig.\u0026nbsp;4) were used for the analysis.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eUrinary metabolome and clinical biochemistry of children with kwashiorkor, marasmus and NAM\u003c/h2\u003e \u003cp\u003eDue to difficulties in obtaining urine samples from children, we were only able to obtain baseline urine samples from 88 children (24 kwashiorkor, 32 marasmus, 32 NAM; patient characteristics in Supplementary table 4. Results for urinary biochemical analyses per group can be found in Supplementary table 5). Combining the positive and negative ionization modes together yielded a total of 4570 urine metabolomic features.\u003c/p\u003e \u003cp\u003eAfter FA, the data was reduced to 8 urinary metabolome LVs (uLV). Since urine was only obtained at recruitment, only criteria 1 and 2 were applied to the urine results. There was no uLV associated with kwashiorkor (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eGut microbiome composition of children with kwashiorkor, marasmus and NAM\u003c/h3\u003e\n\u003cp\u003eOut of 180 children at recruitment, we were only successfully able to collect fecal samples from 38 children (14 kwashiorkor, 13 marasmus, 11 NAM; patient characteristics in Supplementary table 6). Fecal samples were then subjected to 16s rRNA gene amplicon sequencing, and the resulting gene fragment amplicon sequences were quantified and taxonomically annotated. We initially visualized the variation in the data using principal coordinates analysis (PCoA) using Bray-Curtis and Jaccard distances. Using all distance measures, no clear clustering could be found neither generally nor on the basis of group (kwashiorkor vs marasmus vs NAM), age, sex and triage score (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). PERMANOVA analysis concurred with this visual inspection as differences among the groups were not significant. Moreover, no differences in alpha diversity metrics (Shannon, Faith Phylogenetic, Pielou) were found among the three groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb). Finally, we used the ANCOM module in QIIME2 to identify differentially abundant features across sample groups at the taxonomic level of family, genus and strain. Both groupwise and pairwise analyses report no differentially abundant features (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05) among kwashiorkor, marasmus and NAM (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we identified extracellular matrix (ECM) degradation and sphingolipid disruption as central metabolic processes underlying kwashiorkor. Using a stringent multi-omics design, we uncovered latent variables that were distinct in kwashiorkor compared to marasmus and non-malnourished (NAM) children, correlated with oedema severity, and normalized with treatment. This approach delineates kwashiorkor-specific pathways that move beyond simple nutritional deficiency and instead reveal maladaptive tissue remodeling and lipid signaling.\u003c/p\u003e \u003cp\u003eWe presented data from children with kwashiorkor, marasmus and NAM, matched based on age, sex, and a programmatic clinical triage score. Matching especially on triage score as a proxy for clinical severity upon admission was decided for the following reasons: first, non-malnourished children without serious illness would overtly distinguish themselves from malnourished ill children only on the basis of illness severity. Second, certain diseases tend to co-occur with either form: for instance, HIV and malaria are reported to co-occur more with marasmus than kwashiorkor\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Third, mortality rates have been reported to be higher in kwashiorkor than marasmus\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Other studies have reported the opposite, especially in areas with high burden of HIV\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Hence, matching for triage score, albeit imperfect, allowed to somehow disentangle pathways linked with kwashiorkor other than from clinical co-morbidities. However, differences remained between groups, with NAM children notably more likely to have a positive malaria rapid test.\u003c/p\u003e \u003cp\u003eThis study further strengthens our previous findings on the role of ECM remodeling/degradation in the pathophysiology of kwashiorkor. We have previously shown that the protein lumican was higher in kwashiorkor compared to marasmus, was associated with severity of edema, and was resolving during treatment in Kenyan and Malawian children\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. In the previous study, we matched the children with kwashiorkor and marasmus on the basis of serum albumin concentration, which led us to conclude that both hypoalbuminemia and heightened ECM degradation play important roles in kwashiorkor development. However, this previous work did not include non-malnourished children, and only focused on proteomics.\u003c/p\u003e \u003cp\u003eIn this current study, a pLV comprising lumican, gelsolin, and tetranectin met all biomarker criteria, reinforcing the role of ECM remodeling in kwashiorkor. Lumican, a small leucine-rich proteoglycan, regulates collagen fibrillogenesis and tissue integrity\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. This protein\u0026rsquo;s potential role in kwashiorkor has been described in our previous work\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Gelsolin, an actin-binding protein, also affects ECM by clearing actin filaments and modulating inflammation\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Plasma gelsolin (pGSN) plays key roles in actin scavenging, and immune modulation\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. pGSN levels have been reported to significantly decreased in tissue injury, secondary organ damage, and sepsis\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, but is increased in colon cancer, pancreatic cancer and pancreatitis\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, and HIV\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Tetranectin enhances plasminogen activation, important for fibrinolysis and ECM breakdown\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eUsing the human proteome atlas tool\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, we deduced that pLV7 represents a systemic ECM breakdown, as these proteins are co-expressed in a multitude of organs, including the lymphatic system and the liver. Our previous finding that plasma lymphatic vessel endothelial hyaluronan receptor 1 (LYVE1) levels are higher in kwashiorkor than marasmus, and are reduced after 60 days of post-discharge nutritional rehabilitation\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e indicating ECM degradation in the lymphatic system in kwashiorkor. Dysfunction of the lymphatic system integrity could potentially be playing a major role in edema formation in kwashiorkor, as it is the major route for interstitial fluid drainage back to circulation, as explained in the revised Starling model. The interstitial volume is mainly controlled by the activity of lymph flow which depends on the lymphatic function\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Damage to the lymph, i.e. degradation of lymph ECM, may contribute to edema formation. Moreover, ECM proteins are major constituents of the vessel wall supporting endothelial cells throughout the entire vascular system\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. Degradation of the ECM in the vessel wall will therefore result in vascular leakage and increased capillary membrane permeability, further contributing to fluid leakage to the interstitium.\u003c/p\u003e \u003cp\u003eThe ECM is also an important reservoir for sodium, enabled by the interaction between negatively charged ECMs, specifically glycosaminoglycans, allowing them to store non-osmotic Na\u003csup\u003e+\u003c/sup\u003e ions. ECM is therefore an integral and dynamic component of sodium balance\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. Our network analysis revealed a positive association between total blood sodium (via iStat point-of-care device) and pLV7. Disruption of the ECM integrity could therefore also influence sodium balance, and consequently contribute to water retention in kwashiorkor. Our results showed that children with kwashiorkor have higher blood sodium concentration compared to marasmus and non-malnourished children, which concur with previous findings\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. However, it is worth noting that despite this statistical difference, median blood sodium levels in all children in this study still fall within normal range. No difference in urinary sodium was found between kwashiorkor and marasmus in this study, but both conditions had lower urinary sodium compared to non-malnourished controls (Supplementary table 5).\u003c/p\u003e \u003cp\u003eThe dysregulation of ECM homeostasis has also been implicated in liver pathologies, including liver fibrosis and non-alcoholic fatty liver disease\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. It remains unclear whether systemic ECM homeostasis dysregulation is directly involved in kwashiorkor-related liver fat infiltration; but its participation in other liver pathologies makes this plausible.\u003c/p\u003e \u003cp\u003eWe also found that lipid metabolism is strongly associated with kwashiorkor, particularly through latent lipid variable 5 (lLV5), which captures signatures of sphingomyelins, ceramides, and gangliosides. These lipids, collectively classified as sphingolipids, are key structural components of cell membranes, especially in the central nervous system, and play critical roles in cell signaling, proliferation, and apoptosis\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Ceramide serves as the core molecule, which can be modified to form sphingomyelins and further conjugated with glycolipids to produce gangliosides\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Thus, lLV5 likely reflects overall sphingolipid homeostasis. Since circulating sphingolipids are primarily synthesized by the liver and intestine, with only a small fraction derived directly from dietary sources\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e, this may suggest that disturbances occurring in these organs may underlie the development of kwashiorkor. Moreover, network analysis showed that lLV5 is linked to total serum cholesterol, lLV10 (phosphatidylethanolamine lipid species), and lLV13 (unsaturated sphingomyelins), indicating that lLV5 reflects a pathway involved in lipid transport, potentially in the liver.\u003c/p\u003e \u003cp\u003eThere are known interactions between ECM proteins and sphingolipids. Gelsolin for instance can bind sphingosine-1-phosphate (S1P)\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e, a sphingolipid mediator formed by the metabolism of sphingomyelin\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Insufficient lymphatic S1P signaling was found to contribute to lymphatic dysfunction in lymphedema\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. S1P also regulates matrix metalloproteinases, which are key enzymes involved in the degradation of ECM proteins\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. However, our results demonstrated neither direct nor indirect (partial correlation network analysis) association between pLV7 and lLV5. This could be due to lack of statistical power given our small sample size. However, both pLV7 and lLV5 are associated with total plasma protein content. This shared association could suggest that the dysfunction observed in kwashiorkor is either driven by protein deficiency and/or, alternatively, that upstream disturbances affecting ECM remodeling and sphingolipid homeostasis may contribute to a reduction in plasma protein levels.\u003c/p\u003e \u003cp\u003eThe ECM is closely linked with inflammation; its remodeling and degradation are often triggered by inflammatory signals. On the other hand, sphingolipid homeostasis is directly linked to inflammation as its metabolism results in the production of either pro- or anti-inflammatory lipids, and specific sphingolipids, especially S1P that regulates vascular inflammation and permeability\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e, potentially leading to endothelial leakage. In fact, animal models of both acute and chronic inflammation have demonstrated that plasma S1P limits the disruption of vascular endothelial monolayers and reduces edema\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. This suggests that the balance between inflammation and sphingolipid homeostasis is critical: dysregulation in this system may lead to an edematous reaction, while systemic imbalances in sphingolipid levels could exacerbate ECM breakdown and intensify inflammatory responses. Albumin synthesis and circulating concentration declines during active inflammation, being a negative acute phase protein. Thus, the low circulating protein levels observed in kwashiorkor may not only reflect inadequate dietary intake but also a consequence of the inflammatory response itself\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWhile our findings do not provide a complete picture of kwashiorkor pathophysiology, they highlight potential pathways that may contribute to its underlying mechanisms. A proposed mechanistic framework based on these observations is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. Further research is needed to validate and refine this proposed model. Moreover, the specific inflammatory trigger in this framework is currently unknown.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eOur untargeted semi-polar metabolomics results did not reveal metabolites nor metabolome-LVs associated passing all kwashiorkor biomarker criteria. However, a targeted metabolomics study comparing kwashiorkor and marasmus revealed major differences in their metabolome, where 128 of 141 (91%) metabolites quantified were lower in serum of kwashiorkor than marasmus\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. Another quantitative metabolite profiling study found reduced levels of metabolites involved in one-carbon metabolism in serum of kwashiorkor and marasmic-kwashiorkor compared to marasmus, especially levels of methionine and choline\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. However, our previous targeted metabolomics study comparing kwashiorkor and marasmus matched for serum albumin concentrations showed an almost complete overlap of serum amino acids concentrations in both groups, except for phenylalanine which was lower in kwashiorkor\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Similar to our current study, a previous untargeted metabolomics study did not find semi-polar metabolomic differences between Nigerian children with kwashiorkor and marasmus\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. It is possible that the metabolome is more susceptible to geographic differences, leading to inconsistent results from studies in different countries. Moreover, given the increased overall fluid retention in kwashiorkor, it is also possible that small hydrophilic metabolites are dispersed both in the intravascular volume as well as the interstitial space, which makes the apparent concentration of these metabolites lower in kwashiorkor. Our results do not suggest an increase in the intravascular space in kwashiorkor. While the endothelial glycocalyx serves as a barrier for proteins from crossing the endothelial space to the interstitial space, the movement of metabolites across the endothelium remains to be deeply investigated. Nonetheless, more investigation is needed to ascertain whether specific metabolite concentrations indeed play a role in kwashiorkor pathophysiology.\u003c/p\u003e \u003cp\u003eThe role of the gut microbiome on the pathophysiology of kwashiorkor has previously been proposed. In a study of Malawian twins discordant for kwashiorkor, fecal transplantation of stool from children with kwashiorkor (n\u0026thinsp;=\u0026thinsp;9) to germ-free mice led to phenotype of malnutrition where germ-free mice transplanted with stool from non-malnourished twin (n\u0026thinsp;=\u0026thinsp;9) did not\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. This study however did not include children with marasmus, and it is therefore difficult to determine whether the effect is due to kwashiorkor specifically or due to general malnutrition, or even clinical comorbidities. Also, edema in the mice models was not reported to have occurred. Similar to our findings, a previous study also did not find differences in gut microbiome between kwashiorkor and marasmus\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. However, gut microbiome differences between children with kwashiorkor or marasmus and the NAM group cannot be excluded, given the small sample size, similar living conditions, and limited gastroenteritis cases.\u003c/p\u003e \u003cp\u003eIn this study, we matched participants using a pragmatic triage score designed to identify children in need for immediate care. However, we acknowledge that this score, and hence matching for it, may not be optimal or detailed enough. Moreover, this study included only 60 children per group, which only offered an exploratory perspective to this complex problem. We have previously described the barriers and enablers in undertaking omics research in low-resource countries\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e, such as Niger. Several logistic field operations may not be at par with quality standards observed in high income settings. Moreover, due to security reasons, we had to resort to training field staff remotely on certain aspects, such as collection of skin lesions data, which may have led to the low prevalence of detected cases. Research in other settings is needed to ensure reproducibility of the results. Nonetheless, the alignment of our current results with our previous work and that of others offers assurances on the validity of our findings.\u003c/p\u003e \u003cp\u003eIn conclusion, these observations suggest that acute malnutrition, and particularly kwashiorkor, is not simply a state of nutrient deficiency but a multifaceted physiological condition that reconfigures systemic molecular interactions. This results in increased ECM degradation and sphingolipid homeostasis dysregulation and their interaction plays significant roles in the pathophysiology of kwashiorkor, indicating that this disease could be linked to inflammation. Whether a specific inflammatory trigger is required to develop kwashiorkor or that children with kwashiorkor are predisposed to a derailed inflammatory response regardless of the stimuli remains to be investigated.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Participants\u003c/h2\u003e \u003cp\u003eThis matched case-control study was designed to analyze and compare biological specimens, including blood, urine, and feces, collected from children with kwashiorkor, marasmus, and non-acutely malnourished (NAM). Participants were matched by age categories, sex, and triage score as a proxy of clinical severity, with children classified into groups based on edema levels (mild, moderate, and severe), which served as an indicator of kwashiorkor severity (see Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eRecruitment groups and sample size\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eKwashiorkor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMarasmus\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNon-malnourished\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eMatching on:\u003c/p\u003e \u003cp\u003e- Age\u003c/p\u003e \u003cp\u003e- Sex\u003c/p\u003e \u003cp\u003e- Severity at baseline\u003c/p\u003e \u003cp\u003e\u003cem\u003e(triage score)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e++\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e+++\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e60\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e(+) feet/ankles only, (++) feet/ankles and legs, (+++)on other part of the body\u003c/h2\u003e \u003cp\u003eBiological samples were collected at various time points: (1) upon recruitment for all study participants, (2) at discharge from inpatient care, and (3) at exit from outpatient care (known in this context as an ambulatory therapeutic feeding center, ATFC). The collection frequency varied by participant group: NAM samples were collected once for children in outpatient care and at baseline and discharge from children in inpatient care, while kwashiorkor and marasmus groups underwent additional collections at the indicated discharge/exit points.\u003c/p\u003e \u003cp\u003eCollected samples (venous blood, stools and urine) were transported in cold chain to the Epicentre Research Laboratory in Maradi, Niger, where they were stored in preparation for shipment to Belgium, following a -80\u0026deg;C cold chain protocol. Upon arrival, samples were distributed among collaborating laboratories for specific analyses, conducted according to the type of biological specimen and study objectives.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eStudy setting\u003c/h2\u003e \u003cp\u003eThe study was conducted from the 26th of September 2018 to the 5th of November 2019 in the Madarounfa Health District of the Maradi region, located in a remote rural area of south-central Niger, along the Nigerian border. The local livelihood mostly relies on subsistence agriculture and animal farming. Acute childhood malnutrition is endemic in the area with seasonal peaks from May until October when the food security deteriorates before the harvest. In collaboration with the Ministry of Health of Niger, M\u0026eacute;decins Sans Fronti\u0026egrave;res (MSF) is supporting the pediatric ward, the level 1 intensive care unit and the inpatient therapeutic feeding center (ITFC) of the Madarounfa District hospital as well as five ambulatory therapeutic feeding centres (ATFC) in the district.\u003c/p\u003e \u003cp\u003e The study was approved by the Ethics Review Board of M\u0026eacute;decins Sans Fronti\u0026egrave;res (ID: 1785) and the Comit\u0026eacute; National D\u0026rsquo;Ethique pour la Recherche en Sante (CNERS) of the Ministry of Public Health in Niger (ID: 004/2018/CNERS and 026/2020/CNERS).\u003c/p\u003e \u003cp\u003e All the study participants were treated for free according to the MSF and national guidelines.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eStudy population\u003c/h2\u003e \u003cp\u003e This study recruited children with severe acute malnutrition aged 6 to 59 months eligible for new admission to either inpatient care in the ITFC of the Madarounfa District Hospital or enrolment in outpatient care in one of the ATFC supported by MSF in the region of Madarounfa. In addition, the study included non-acutely malnourished children from the paediatric ward of Madarounfa Health District Hospital or from the outpatient department of Madarounfa health centre. Children were classified according to their nutritional status and the severity of the oedema among children with kwashiorkor.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eNutritional status was defined as follows:\u003c/h2\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eKwashiorkor\u003c/b\u003e was defined by the presence of bilateral nutritional edema (mild to severe) - irrespective of either weight-for-height z-score (WHZ), mid-upper arm circumference (MUAC), presence of skin lesions, hair changes or other clinical features such as apathy, irritability, or fatty liver. The severity of edema was graded as mild (+) when it was present in feet/ankles only, as moderate (++) when it was present in feet/ankles and legs, and severe (+++) when it was visible on other part of the body (face, hands, arms, trunk etc)\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. The presence of edema on feet/ankles was determined by applying moderate thumb pressure to the dorsum of both feet for at least three seconds. If bilateral pits remained after the thumbs have been removed, then the child would be considered as having nutritional edema. All children underwent a medical history assessment and a clinical examination (by two doctors for most hospitalized children). To avoid the inclusion of edema of non-nutritional origin, edematous children who could provide urine sample underwent a urine test strip and only children with negative results (i.e. negative for protein and blood) were included in the Kwashiorkor group. The presence of albumin in urine during renal disease/nephrotic syndrome contrasts with kwashiorkor where it is usually absent or present in very small amounts. Similarly, blood is found in nephritis but most often absent in kwashiorkor.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eMarasmus\u003c/b\u003e, or severe wasting, was defined as a WHZ less than \u0026minus;\u0026thinsp;3\u003csup\u003e49\u003c/sup\u003e, and/or a MUAC less than 115 mm (in children 6\u0026ndash;59 months) \u003cb\u003eand the absence of edema\u003c/b\u003e (nutritional or not).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eNon (acutely)-malnourished (NAM) status\u003c/b\u003e was defined by the absence of acute malnutrition, that is: absence of nutritional edema, and a WHZ \u0026ge; -2 and a MUAC\u0026thinsp;\u0026ge;\u0026thinsp;125mm (in children 6\u0026ndash;59 months).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eInclusion criteria\u003c/h2\u003e \u003cp\u003eChildren were eligible for study participation when they fulfilled the following inclusion criteria:\u003c/p\u003e \u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eAged 6 to 59 months at the time of inclusion.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eDiagnosed with severe acute malnutrition (kwashiorkor or marasmus) or non-acutely malnourished (see above definitions).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eMedical condition allowing the collection of 2 mL of blood in the 12 hours following admission and before any intravenous or intramuscular antibiotic treatment.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eConsent signed by the child\u0026rsquo;s parent (i.e. the mother or the father) or the child\u0026rsquo;s main caretaker if the child is an orphan.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eNewly admitted (i.e. not having been enrolled/admitted in the nutrition programme in the last 2 months nor in the study).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eDirect admission at the ITFC (i.e. not transferred to the ITFC while under treatment in an ATFC) or direct admission at the ATFC (i.e. not transferred from the ITFC).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e Child\u0026rsquo;s parent or main caretaker intending to remain with the child in the study area until the end of the study follow-up period.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e Absence of any exclusion criteria.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eExclusion criteria\u003c/h2\u003e \u003cp\u003eChildren were excluded from participation in case of any of the following:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eMedical condition not allowing the collection of 2 mL of blood in the 12 hours following admission, based on the judgement of the clinician attending the child.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMedical condition not allowing the collection of blood before any type of drug or fluid is administered via the intravenous (IV) catheter (when a catheter is placed upon admission), based on the judgement of the clinician attending the child.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eIntravenous or intramuscular antibiotics given in the 12 hours prior to the inclusion blood sampling. For artesunate injection, this was limited to 4 hours.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cul\u003e\n \u003cli\u003eInformed consent not provided.\u0026nbsp;\u003c/li\u003e\n\u003c/ul\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003ePrevious enrolment/admission in the nutrition programme in the last 2 months (including relapse or return after default) or in the study.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eCurrent participation in another clinical research study including blood sampling and/or investigational products.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e Child\u0026rsquo;s parent or main caretaker not intending to remain with the child in the study area until the end of the study follow-up period.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eCurrent edema of non-nutritional origin (based on the judgement of the attending clinician or with positive urine test strip (for proteinuria\u0026thinsp;\u0026ge;\u0026thinsp;0.3g/L (1+) and/or blood red cells or hemoglobin).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eNot \u0026ldquo;directly admitted\u0026rdquo; (e.g. children under treatment at the ATFC who are secondarily transferred to the ITFC due to clinical or nutritional deterioration; children transferred from the ITFC to the ATFC after nutritional stabilisation).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003ePrevious history of kwashiorkor (for patients admitted in the marasmus or non-malnourished groups).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eKnown congenital abnormalities or a known chronic disease (including known tuberculosis, sickle cell disease, congenital heart disease\u0026hellip;).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eKnown HIV-positive status, clinical AIDS or having a known HIV-positive mother. [Considering the low HIV prevalence in Niger (estimated at 0.2% among adults 15\u0026ndash;49 year-old in 2012 in the Maradi Region\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e), there was no systematic HIV screening].\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eAdmitted into inpatient care for trauma, burns, intoxication or surgery.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eClinical biochemistry analyses\u003c/h2\u003e \u003cp\u003eUrine sodium, potassium, chloride, and creatinine levels were measured using an Abbot Architect Clinical Chemistry Analyzer (Abbott, Illinois, USA). Urinary albumin was measured using the Beckman Paragon Electrophoresis system. Plasma triglycerides, cholesterol, HDL-cholesterol, total bilirubin, total protein, albumin, aspartate transferase (AST), alanine aminotransferase (ALT) and alkaline phosphatase (ALP) levels were determined using colorimetric methods and C-reactive protein (CRP) using a particle enhanced turbidimetric immunoassay (PETIA) technique all on a Siemens Dimension Xpand autoanalyzer (Siemens Healthcare, Camberley, UK) following manufacturer instructions. i-STAT point-of-care analyzer (Abbott Point of Care, Maidenhead, UK) used with the cartridge CHEM8+ (providing simultaneously glucose, kidney function [urea nitrogen, creatinine], hematology [hematocrit, hemoglobin], electrolytes [sodium, potassium, chloride, ionized calcium], anion gap) and the cartridge CG4+ (providing simultaneously lactate, blood Gases [PCO2, PO2, TCO2, sO2], pH, HCO3, base Excess).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003ePlasma and urine untargeted metabolomics\u003c/h2\u003e \u003cp\u003ePlasma samples were thawed on ice before metabolite extraction. Ice-cold methanol was used as the extraction solvent, added at a 2:1 ratio (1 mL solvent to 500 \u0026micro;L plasma). After thorough vortexing, the samples were stored at \u0026minus;\u0026thinsp;20\u0026deg;C for one hour to allow protein precipitation. Following this, the samples were centrifuged at 13,000 \u0026times; g for 10 minutes at 4\u0026deg;C. The supernatant (750 \u0026micro;L) was transferred to a glass tube and dried under a gentle nitrogen stream at temperatures below 10\u0026deg;C. The resulting dried residue was reconstituted in 10% acetonitrile containing 0.1% formic acid for subsequent liquid chromatography-mass spectrometry (LC-MS) analysis.\u003c/p\u003e \u003cp\u003eTo create a quality control (QC) sample, 10 \u0026micro;L aliquots from each sample were pooled together. A dilution series of the QC sample was then prepared by mixing it with 0.1% formic acid at different ratios (1:1, 1:5, 1:10, and 1:50).\u003c/p\u003e \u003cp\u003eChromatographic separation was performed on a Waters Acquity UPLC I-class FTN system (Waters, Manchester, UK), with dynamic mobile phase gradients and flow rates, as shown in our previously published paper\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e to optimize peak separation. The mobile phases consisted of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B), with separation achieved using an Acquity UPLC HSS T3 column (1.8 \u0026micro;m, 100 \u0026Aring;, 1 mm \u0026times; 100 mm). The injection volume was set at 5 \u0026micro;L, with the column maintained at 50\u0026deg;C.\u003c/p\u003e \u003cp\u003eLiquid chromatography (LC) flow was directed into a Waters Synapt G2Si high-resolution mass spectrometer (MS) (Waters, Manchester, UK) using electrospray ionization (ESI) in both positive and negative modes. The capillary voltage was set at 2.75 kV for positive mode and 2.20 kV for negative mode. The source temperature was maintained at 150\u0026deg;C, with a desolvation temperature of 500\u0026deg;C. Gas flow rates were 20 L/h for the cone and 600 L/h for desolvation. The mass range covered 50\u0026ndash;1000 Da, with a scan speed of 0.1 s in MS\u003csup\u003eE\u003c/sup\u003e centroid resolution mode. MSE collision energy was ramped from 10 to 30 V for both modes. Mass calibration was achieved using sodium formate adducts, and online correction was applied with leucine-enkephalin (200 pg/\u0026micro;L), infused at 20 \u0026micro;L/min every 10 seconds.\u003c/p\u003e \u003cp\u003eA blank (100% acetonitrile) was injected 10 times at the beginning to condition the column followed by 10 injections of the QC sample. QC samples were re-injected after every 10 experimental samples. Experimental samples were analyzed in random order.\u003c/p\u003e \u003cp\u003eData acquisition was monitored using MassLynx v1.7 (Waters, UK). Data processing: chromatogram peak picking, deconvolution and peak alignment were achieved using the proprietary Progenesis QI software (Nonlinear Dynamics, Newcastle, UK).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003ePlasma lipidomics\u003c/h2\u003e \u003cp\u003eThe lipid extraction was carried out using the Protein Precipitation Liquid Extraction Protocol (Jenkins, Ronis \u0026amp; Koulman, 2020). A 50 \u0026micro;L aliquot was taken from each plasma sample for the lipidomics analysis. To each aliquot, 650 \u0026micro;L of chloroform solution and 100 \u0026micro;L of LIPID Internal Standards (Lipid-IS) were added. The samples were vortexed, followed by the addition of 250 \u0026micro;L of methanol. After another round of vortexing, 400 \u0026micro;L of acetone was added. The samples were then vortexed again and centrifuged at ~\u0026thinsp;20,000\u0026times; g for 10 minutes to remove the insoluble material. The monolayer supernatant was carefully collected. The organic extracts (chloroform, methanol, and acetone) were dried using a Concentrator Plus system (Eppendorf, Stevenage, UK) for 60 minutes at 60\u0026deg;C. The dried samples were reconstituted in 100 \u0026micro;L of a 2:1:1 mixture of propan-2-ol, acetonitrile, and water, and thoroughly mixed. The reconstituted samples were transferred into vials for analysis by LC-MS, and for each sample the volume of injection was 10 \u0026micro;L.\u003c/p\u003e \u003cp\u003eChromatographic separation of lipids was performed using an HPLC system (Shimadzu UK Limited, Milton Keynes, UK). The stationary phase was a Waters Acquity UPLC\u0026reg; CSH C18 column (Waters, Hertfordshire, UK; 1.7 \u0026micro;m, I.D. 2.1 mm \u0026times; 50 mm), maintained at 55\u0026deg;C. The mobile phase consisted of two solutions: solution A (a 6:4 mixture of acetonitrile and water) and solution B (a 9:1 mixture of isopropanol and acetonitrile), both containing 10 mM ammonium formate. The separation was carried out at a flow rate of 0.6 mL/min and the gradient used followed the method described by (Jenkins, Ronis \u0026amp; Koulman, 2020). MS analysis was performed on a Thermo Scientific Exactive Orbitrap instrument with a heated electrospray ionization source (Thermo Fisher Scientific, Hemel Hempstead, UK). The ion source parameters were as follows: capillary temperature: 300\u0026deg;C, source heater temperature: 420\u0026deg;C, sheath gas flow: 40 (arbitrary), auxiliary gas flow: 15 (arbitrary), spare gas: 3 (arbitrary), source voltage: 4 kV. The mass spectrometer scan rate set at 4 Hz, giving a resolution of 25,000 (at 200 m/z) with a full-scan range of m/z 100 to 1800 with continuous switching between positive and negative mode.\u003c/p\u003e \u003cp\u003eThe raw data were processed using MS-DIAL software (version 4.9), which included peak detection, alignment, and data annotation (Tsugawa et al., 2020). The following parameters were applied: Data collection: MS1 mass range: 120\u0026ndash;1800 Da; retention time range: 0\u0026ndash;10 min. Peak detection: minimum peak height: 10,000 amplitude units; mass window width: 0.1 Da; smoothing method: linear weighted moving average; smoothing level: 3 scans. Identification: retention time tolerance: 0.25 min; accurate mass tolerance: 0.006 Da; identification score threshold: 70%. Alignment: retention time tolerance: 0.25 min; MS1 tolerance: 0.005 Da.\u003c/p\u003e \u003cp\u003eAnnotation of lipid molecular species was based on both accurate mass data acquired by mass spectrometry (MS) and specific retention times of each lipid from liquid chromatography (LC). The lipid database used was validated using Lipid Maps (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.lipidmaps.org/tools/ms\u003c/span\u003e\u003cspan address=\"http://www.lipidmaps.org/tools/ms\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Lipid semi-quantification was performed using deuterated internal standards (Lipid-IS) for each lipid class, based on the known concentrations of the Lipid-IS added to each sample. This analytical process resulted in matrices containing the concentrations of annotated lipid molecular species, expressed in \u0026micro;M.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003ePlasma untargeted proteomics\u003c/h2\u003e \u003cp\u003eBottom-up proteomics sample preparation was done in fully randomized fashion using a urea-based in-sol digest. To 2 \u0026micro;L of neat plasma, 300 \u0026micro;L of denaturing buffer (8 M urea (Chem-Lab Analytical, Zedelgem, Belgium) in 1 M ammonium bicarbonate (Sigma-Aldrich, Saint Louis, USA)) was added, followed by three cycles of 3 min-ON/3 min-OFF sonication and vortexing. To reduce proteins, dithiothreitol (MP Biomedical, Irvine, USA) was added to a final concentration of 5 mM followed by incubation at 37\u0026deg;C for 30 min. After cooling down to room temperature, alkylation was done using iodoacetamide (Sigma-Aldrich, Saint Louis, USA) added to a final concentration of 15 mM and incubation at room temperature for 30 min in the dark. Next, the tubes were diluted to 2 M urea with 1 M ammonium bicarbonate buffer and 3 \u0026micro;g of Trypsin/LysC (Promega, Madison, USA) was added to achieve an estimated 1:50 wt/wt ratio of enzyme to protein. This was incubated overnight at 37\u0026deg;C after which resulting peptides were desalted using C18 reverse phase SPE cartridges (Strata-X, Phenomenex, Torrance, USA). The cartridges were conditioned by adding 1 mL of methanol (Biosolve, Valkenswaard, The Netherlands) and equilibrated with two additions of 1 mL ultrapure water. Samples were loaded slowly onto the cartridges and washed two times with 1 mL of 5% methanol in water before eluting in 1% formic acid (Chem-Lab Analytical, Zedelgem, Belgium) in methanol buffer and drying using a vacuum concentrator. Vacuum dried peptide samples were resuspended in 0.1% formic acid in LC-MS grade water (Biosolve, Valkenswaard, The Netherlands) and centrifuged at 16 000 x \u003cem\u003eg\u003c/em\u003e for 10 min. The clear supernatant was transferred into vials and quantified using the Lunatic UV/Vis spectrophotometer (Unchained Labs, Pleasanton, USA). The resuspension volumes were then corrected to achieve a peptide concentration of 2 \u0026micro;g/\u0026micro;L in every sample vial. Finally, a quality control sample was prepared by pooling an equal volume from all samples.\u003c/p\u003e \u003cp\u003ePeptide samples were acquired in randomized order, alternating between sample and blank injections to prevent carry-over, and interspersed with a calibration standard (PepCal, Sciex, Massachusetts, USA) every 5 samples and quality control mixtures every 10 samples: 1 \u0026micro;L of sample was injected onto a Eksigent NanoLC 425 HPLC system (Sciex, Massachusetts, USA) equipped with a Triart C18 trapping column (5 mm \u0026times; 0.5 mm, YMC) and a Luna Omega Polar C18 column (150 mm \u0026times; 0.3 mm, particle size 3 \u0026micro;m, Phenomenex), operating in capillary flow mode (5 \u0026micro;L/min), coupled to a TripleTOF 6600\u0026thinsp;+\u0026thinsp;mass spectrometer (Sciex, Massachusetts, USA) with the Optiflow TurboV ion source operating in positive mode. A 20-minute active gradient from 3 to 30% B was utilized, with mobile phases A and B consisting of 0.1% formic acid in water (Biosolve, Valkenswaard, The Netherlands) and 0.1% formic acid in acetonitrile (Biosolve, Valkenswaard, The Netherlands), respectively. Column temperature was maintained at 30\u0026deg;C. A SWATH 99 variable window acquisition scheme with a total cycle time of 4 sec was used. MS1 spectra spanned a mass range of 400\u0026ndash;1200 m/z in high sensitivity mode with 250 ms accumulation time and MS2 spectra were acquired in high sensitivity mode from 100\u0026ndash;1500 m/z in 37.5 ms per scan. Ion source parameters were set to 4.5 kV for the ion spray voltage, curtain gas at 25 psi, nebulizer gas at 10 psi, heater gas at 20 psi, and 100\u0026deg;C as source temperature. Instrument maintenance was required in the middle of the batch.\u003c/p\u003e \u003cp\u003ePeptide identification and quantification was done with DIA-NN 1.8\u003csup\u003e52\u003c/sup\u003e. An iterative search strategy was employed to ensure maximum coverage of the proteome by first doing four separate searches starting from four different databases: (i) a predicted library from the UniProt human proteome\u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e (reviewed only, retrieved on 06/01/2022) concatenated with the GPM cRAP database, (ii) a predicted library from the PeptideAtlas\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e human plasma build (2021-07), (iii) a deeply fractionated empirical DDA plasma library (ZenoTOF 7600, Sciex, Massachusetts, USA), (iv) a predicted library from an annotated secreted protein database, i.e., supplementary file 3 from Harney et al\u003csup\u003e55\u003c/sup\u003e. For all searches, the following settings were used: enzyme Trypsin/P allowing for 2 missed cleavages, cysteine carbamidomethylation as fixed modification and up to 1 variable modification from methionine oxidation and protein N-term acetylation, peptide length of 6\u0026ndash;35 with charge range 1\u0026ndash;5, precursor m/z range 400\u0026thinsp;\u0026minus;\u0026thinsp;122, fragment ion m/z range 100\u0026ndash;1500, MS1 accuracy of 12, MS2 accuracy of 25, scan window of 6, match-between-runs with smart profiling enabled and double-pass mode, heuristic protein inference based with proteotypicity definition based on gene names, and robust high precision quantification with normalization and MaxLFQ disabled. The four resulting spectral libraries were merged into one spectral library, which ended up containing 3529 protein isoforms, 2820 protein groups and 16595 precursors in 13930 elution groups. Finally, one search was performed on the entire dataset using the same settings as before but starting from the merged spectral library. The resulting precursor quantity matrix at a false discovery rate of 1% was subsequently used as input for data preprocessing.\u003c/p\u003e \u003cp\u003eFirst, failed injections (1 sample and 2 QC runs) were dropped from the dataset and (potential) contaminant peptides were filtered out (33 out of 9365 peptides). Peptides were then manually matched to all possible proteins of origin using the UniProt human proteome database as described previously. Peptide abundance values were log\u003csub\u003e2\u003c/sub\u003e transformed and normalized using the RUV-III-C\u003csup\u003e56\u003c/sup\u003e algorithm to correct for instrument performance degradation. RUV-III-C used a grouping based on run order except for QC samples which were assigned to one group and peptides that differed most across the run order as determined from a two-sided unpaired t-test were used as negative controls. Finally, peptides not belonging to a smallest unique group were dropped and the normalized peptide abundances were summarized to the protein level using a robust summarization algorithm\u003csup\u003e\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eFecal microbiome analysis\u003c/h2\u003e \u003cp\u003eTotal DNA was extracted from the pellet of 200 mg fecal matter by means of bead beating with a PowerLyzer (Qiagen, Venlo, the Netherlands) and phenol/chloroform extraction\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. The beads were precipitated and removed after centrifugation at maximal speed for 5 min. The supernatant was transferred to new tubes, followed by extraction and purification with 500 \u0026micro;L phenol: chloroform: isoamilyc alcohol 25:24:1 (pH 7) and 700 \u0026micro;L chloroform. 450 \u0026micro;L supernatant was transferred to new tubes containing 500 \u0026micro;L ice-cold Isopropyl alcohol and added 45 \u0026micro;L 3 M sodium acetate, followed by cooling at -20\u0026deg;C over 1 h. Isopropyl alcohol was discarded after centrifugation at maximum speed for 30 min and then air-dried DNA was re-suspended by adding 100 \u0026micro;L 1\u0026times; TE buffer (10 mM Tris, 1 mM EDTA). The quality control of DNA extracts was performed by electrophoresis in a 1% (w/v) agarose gel (Life Technologies, Madrid, Spain).\u003c/p\u003e \u003cp\u003eFor each sample, a sequencing library was constructed starting from 12.5 ng input gDNA according to the \u0026lsquo;16S Metagenomic Sequencing Library Preparation\u0026rsquo; guide (Illumina). Briefly, 16S rRNA gene fragments were PCR-amplified from genomic DNA for 25 cycles using the 341F (5\u0026rsquo;-CCT ACG GGN GGC WGC AG -3\u0026rsquo;) and 785Rmod (5\u0026rsquo;-GAC TAC HVG GGT ATC TAA KCC-3\u0026rsquo;)\u003csup\u003e59\u003c/sup\u003e primer pair. The PCR amplicons were purified using Ampure XP beads (Beckman Coulter). Sequencing adapters were added using the Nextera XT Index Kit (Illumina) with 8 PCR cycles and the resulting sequencing libraries were purified on Ampure XP beads. Quality was checked on a Bioanalyzer using a DNA 1000 chip (Agilent) and quantification was done using qPCR according to the \u0026lsquo;Sequencing Library qPCR Quantification Guide\u0026rsquo; from Illumina. Finally, libraries were pooled equimolarly, spiked with 20% PhiX and sequenced as paired-end 150 on a MiSeq device (Illumina).\u003c/p\u003e \u003cp\u003eData preparation and metagenomics analyses were done using QIIME2 (v2020.2)\u003csup\u003e60\u003c/sup\u003e. For each sample, read pairs were quality-trimmed, denoised and reconstructed into 16S amplicons using the DADA2 algorithm. Operational taxonomic units (OTU) and their abundance were collected. Taxonomic assignment was done using the Greengenes 16S reference collection\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e (release 13.8). These were then used to calculate alpha diversity metrics (Shannon, Faith Phylogenetic Distance, Pielou Evenness, Observed OTUs) and beta diversity metrics (Weighted and Unweighted UniFrac, Bray-Curtis, Jaccard). Alpha diversity metrics were used to detect differences between sample groups (Kwashiorkor, Marasmus, NAM) by means of Kruskal-Wallis tests. Group comparisons using beta diversity metrics were done with the PERMANOVA module in QIIME2. We used the ANCOM module to identify differentially abundant features across sample groups at the taxonomic level of family, genus and strain. We used OTU abundancies with the full pipeline of the PICRUSt2 module to perform a functional composition analysis at the taxonomic family, genus and strain levels, based on KEGG orthologs, Enzyme Classification numbers (EC) and MetaCyc pathways.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eUnivariable analyses\u003c/h2\u003e \u003cp\u003eScales and log-transformed levels of individual metabolite, lipid and protein features were compared between kwashiorkor and combined marasmus and NAM using conditional logistic regression (\u003cem\u003eclogit\u003c/em\u003e) analysis to account for the age, sex and clinical severity matching. Significance was accepted at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 after Benjamini Hochberg False Discovery Rate (FDR) correction\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e. Significant features were then forwarded to succeeding analysis assessing their individual levels association with severity of edema using ordinal regression using the \u003cem\u003epolr\u003c/em\u003e function of the MASS package in R. Features showing a relationship with edema severity where assessed for their change over time using linear mixed model, where the features were tested for association with timepoint, age and sex, with subject as random effect. This was done using the \u003cem\u003elmerTest\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e package in R.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eFactor analysis\u003c/h2\u003e \u003cp\u003eLatent variables (LV) were calculated separately for the metabolome, lipidome and proteome for the samples collected at baseline. Factor analysis was performed using the matrix-free likelihood method of extraction followed by the \u0026ldquo;varimax\u0026rdquo; rotation using the \u003cem\u003efad\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e package in R. The Bayesian information criterion (BIC) was used to determine the optimal number of latent variables to extract. We imposed a factor loading cut-off of 0.4, which means that features that have less than 0.4 loadings to the latent variable were removed - leaving behind only those features that are contributing highly to the latent variable. Factor scores were then calculated from these variables using the \u003cem\u003epsych\u003c/em\u003e package in R, using the Bartlett method. Factor scores were calculated for the different timepoints: baseline, hospital discharge and full nutritional and clinical recovery. These factor scores were then used for subsequent analysis to represent the entire LV, either for metabolome (mLV), proteome (pLV) or lipidome (lLV).\u003c/p\u003e \u003cp\u003emLV, lLV and pLVs were compared between kwashiorkor and combined marasmus and NAM using conditional logistic regression (\u003cem\u003eclogit\u003c/em\u003e) analysis to account for the age, sex and clinical severity matching. Significance was accepted at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 after Benjamini Hochberg False Discovery Rate (FDR) correction\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e. Significant LVs were then forwarded to succeeding analysis assessing their individual levels association with severity of edema using ordinal regression using the \u003cem\u003epolr\u003c/em\u003e function of the MASS package in R. LVs showing a relationship with edema severity where assessed for their change over time using linear mixed model, where the features were tested for association with timepoint, age and sex, with subject as random effect. This was done using the \u003cem\u003elmerTest\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e package in R.\u003c/p\u003e \u003cp\u003eAs sensitivity analyses, the analyses were reperformed after removing children with malaria and again after only including children with negative blood or protein in the urine measured using a urine test strip.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003ePartial correlation network integration\u003c/h2\u003e \u003cp\u003eIntegration of the different omics domains were performed at the LV level using partial correlation network analysis using the \u003cem\u003eqgraph\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e package in R. mLV, lLV and pLV were combined with clinical biochemistry results to show inter-LV association with known clinical markers.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eEpicentre and MSF are committed to sharing and disseminating health data from its programs and research in an open, timely, and transparent manner to promote health for populations while respecting ethical and legal obligations towards patients, research participants, and their communities.\u003c/p\u003e \u003cp\u003eFor the purpose of this study, Epicentre and MSF collected identifiable participant data. Upon publication and for as long as the ethical authorisation permits it, the minimal data set underlying the findings of this study will be made available on request. If scientifically relevant, the request may be granted in accordance with legal framework set forth by MSF data sharing policy available on its website or on request.\u003c/p\u003e \u003cp\u003eThe MSF data sharing policy ensures that data will be available upon request to interested researchers while addressing all security, legal, and ethical concerns. All data access request for non-commercial and academic research can be addressed to [email protected], [email protected] or the corresponding authors.\u003c/p\u003e \u003cp\u003eSuch request will be submitted to the medical departments of MSF and Epicentre. In case of approval of the request, the data will be shared with researchers, subject to the establishment, within a reasonable timeline, of a data sharing agreement to provide the legal framework for data sharing \u0026ndash; including any applicable data protection laws. Such data sharing agreement may differ depending on the nature of the data to be shared \u0026ndash; pseudonymized or anonymized \u0026ndash; and the sensitivity of the data.\u003c/p\u003e \u003cp\u003eAnalysis codes and summaries of metabolite, lipid, and protein features, and the gut microbiome analysis used in this study are publicly available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/HOB-IDEALS/Kwash-Multi-omics\u003c/span\u003e\u003cspan address=\"https://github.com/HOB-IDEALS/Kwash-Multi-omics\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The data may be interactively visualized using this link: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://mudiboevans.shinyapps.io/MSFKwash/\u003c/span\u003e\u003cspan address=\"https://mudiboevans.shinyapps.io/MSFKwash/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe field work in Niger, including participant recruitment and international shipment of biological samples, and analysis of the metabolome, proteome, and gut microbiome were funded by the M\u0026eacute;decins Sans Fronti\u0026egrave;res International Innovation Fund. Analysis of lipidome and blood biochemistry, as well as postdoctoral fellowship of GB Gonzales were funded by the Research Foundation Flanders (FWO; 3E020617 and 31501220), Ghent University Global Minds Fund provided travel funding for GB Gonzales.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe express our deepest gratitude to the participants and the communities in Niger whose trust and collaboration made this study possible. We sincerely thank the laboratory technicians, physicians and study nursing team in Niger for their dedication and meticulous work throughout all phases of clinical care, sample processing and data collection. We further thank M\u0026eacute;decins Sans Fronti\u0026egrave;res\u0026ndash;Operational Centre Paris (MSF OCP) in Niger for their operational support and collaboration. We extend special appreciation to Alexandra Ascorra, Brigite Chokote, Iris van den Boomgaard, Emmanuel Berbain, Evans Mudibo and Mario-David Barbagallo for their guidance, coordination, and assistance throughout the project.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBriend A (2014) Kwashiorkor: still an enigma - the search must go on. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003c/span\u003e\u003cspan address=\"http://www.ennonline.net/kwashiorkorstillanenigma\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlvarez J, Dent N, Browne L, Myatt M, Briend A (2016) Putting Child Kwashiorkor on the map\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilliams C (1935) Kwashiorkor: a nutritional disease of children associated with a maize diet. Lancet 226(5855):1151\u0026ndash;1152\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMichael H, Amimo JO, Rajashekara G, Saif LJ, Vlasova AN (2022) Mechanisms of Kwashiorkor-Associated Immune Suppression: Insights From Human, Mouse, and Pig Studies. Front Immunol ; 13\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith MI, Yatsunenko T, Manary MJ et al (2013) Gut Microbiomes of Malawian Twin Pairs Discordant for Kwashiorkor. Science 339(6119):548\u0026ndash;554\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoulter JBS, Hendrickse RG, Lamplugh SM et al (1986) Aflatoxins and kwashiorkor: clinical studies in Sudanese children. Trans R Soc Trop Med Hyg 80(6):945\u0026ndash;951\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Vries HR, Maxwell SM, Hendrickse RG (1990) Aflatoxin excretion in children with kwashiorkor or marasmic kwashiorkor\u0026ndash;a clinical investigation. Mycopathologia 110(1):1\u0026ndash;9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcMillan A, Renaud JB, Burgess KMN et al (2018) Aflatoxin exposure in Nigerian children with severe acute malnutrition. Food Chem Toxicol 111:356\u0026ndash;362\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOyelami OA, Maxwell SM, Adelusola KA, Aladekoma TA, Oyelese AO (1998) Aflatoxins in autopsy kidney specimens from children in Nigeria. J Toxicol Environ Health A 55(5):317\u0026ndash;323\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoriano JM, Rubini A, Morales-Suarez-Varela M, Merino-Torres JF, Silvestre D (2020) Aflatoxins in organs and biological samples from children affected by kwashiorkor, marasmus and marasmic-kwashiorkor: A scoping review. Toxicon 185:174\u0026ndash;183\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbdallah MF, Briend A, Gonzales GB (2021) Aflatoxins and kwashiorkor: A commentary on the association-causation conundrum. Toxicon 190:20\u0026ndash;21\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolt L, Snyderman S, Norton P, Roitman E, Finch J (1963) The plasma aminogram in kwashiorkor. Lancet 282(7322):1343\u0026ndash;1348\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoediger WE (1995) New views on the pathogenesis of kwashiorkor: methionine and other amino acids. J Pediatr Gastroenterol Nutr 21(2):130\u0026ndash;136\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMay T, de la Haye B, Nord G et al (2022) One-carbon metabolism in children with marasmus and kwashiorkor. \u003cem\u003eeBioMedicine\u003c/em\u003e ; 75: 103791\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonzales GB, Njunge JM, Gichuki BM et al (2022) The role of albumin and the extracellular matrix on the pathophysiology of oedema formation in severe malnutrition. eBioMedicine 79:103991\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmadi B, Fagbemi AO, Kelly P et al (2009) Reduced production of sulfated glycosaminoglycans occurs in Zambian children with kwashiorkor but not marasmus. Am J Clin Nutr 89(2):592\u0026ndash;600\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThul PJ, Lindskog C (2018) The human protein atlas: A spatial map of the human proteome. Protein science: publication Protein Soc 27(1):233\u0026ndash;244\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKessler L, Daley H, Malenga G, Graham S (2001) The impact of HIV infection on the clinical presentation of severe malnutrition in children at QECH. Malawi Med journal: J Med Association Malawi 13(3):30\u0026ndash;33\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFevang P, Havemann K, Fevang B, H\u0026oslash;stmark AT (2018) Malaria and Malnutrition: Kwashiorkor Associated with Low Levels of Parasitaemia. \u003cem\u003eMalaria Research and Treatment\u003c/em\u003e ; 2018: 7153173\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBadaloo AV, Forrester T, Reid M, Jahoor F (2006) Lipid kinetic differences between children with kwashiorkor and those with marasmus. Am J Clin Nutr 83(6):1283\u0026ndash;1288\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMunthali T, Jacobs C, Sitali L, Dambe R, Michelo C (2015) Mortality and morbidity patterns in under-five children with severe acute malnutrition (SAM) in Zambia: a five-year retrospective review of hospital-based records (2009\u0026ndash;2013). \u003cem\u003eArchives of public health\u0026thinsp;=\u0026thinsp;Archives belges de sante publique\u003c/em\u003e ; 73(1): 23\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChakravarti S, Magnuson T, Lass JH, Jepsen KJ, LaMantia C, Carroll H (1998) Lumican regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of lumican. J Cell Biol 141(5):1277\u0026ndash;1286\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePietraszek K, Chatron-Colliet A, Br\u0026eacute;zillon S et al (2014) Lumican: A new inhibitor of matrix metalloproteinase-14 activity. FEBS Lett 588(23):4319\u0026ndash;4324\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBucki R, Levental I, Kulakowska A, Janmey PA (2008) Plasma gelsolin: function, prognostic value, and potential therapeutic use. Curr Protein Pept Sci 9(6):541\u0026ndash;551\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHsieh CH, Wang YC (2022) Emerging roles of plasma gelsolin in tumorigenesis and modulating the tumor microenvironment. Kaohsiung J Med Sci 38(9):819\u0026ndash;825\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiktel E, Levental I, Durnaś B, Janmey PA, Bucki R (2018) Plasma Gelsolin: Indicator of Inflammation and Its Potential as a Diagnostic Tool and Therapeutic Target. Int J Mol Sci ; 19(9)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsare-Werehene M, Communal L, Carmona E et al (2020) Plasma Gelsolin Inhibits CD8\u0026thinsp;+\u0026thinsp;T-cell Function and Regulates Glutathione Production to Confer Chemoresistance in Ovarian Cancer. Cancer Res 80(18):3959\u0026ndash;3971\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIram S, Rahman S, Choi I, Kim J (2024) Insight into the function of tetranectin in human diseases: A review and prospects for tetranectin-targeted disease treatment. Heliyon 10(1):e23512\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLevick JR, Michel CC (2010) Microvascular fluid exchange and the revised Starling principle. Cardiovasc Res 87(2):198\u0026ndash;210\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin PK, Davis GE (2023) Extracellular Matrix Remodeling in Vascular Disease: Defining Its Regulators and Pathological Influence. Arterioscler Thromb Vasc Biol 43(9):1599\u0026ndash;1616\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLloyd IJ (1990) Is the extracellular matrix an integral and dynamic component of the sodium and water homeostatic system? Med Hypotheses 33(3):197\u0026ndash;199\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhalil M, Kabiel A, El-Khateeb S, Aref K, El Lozy M, Jahin S, Nasr F (1974) Plasma and red cell water and elements in protein-calorie malnutrition. Am J Clin Nutr 27(3):260\u0026ndash;267\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWharton BA, Howells GR, McCance RA (1967) CARDIAC FAILURE IN KWASHIORKOR. Lancet 290(7512):384\u0026ndash;387\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDolin CE, Arteel GE (2020) The Matrisome, Inflammation, and Liver Disease. Semin Liver Dis 40(2):180\u0026ndash;188\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorales A, Lee H, Go\u0026ntilde;i FM, Kolesnick R, Fernandez-Checa JC (2007) Sphingolipids and cell death. Apoptosis 12(5):923\u0026ndash;939\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlbeituni S, Stiban J (2019) Roles of Ceramides and Other Sphingolipids in Immune Cell Function and Inflammation. In: Honn KV, Zeldin DC (eds) The Role of Bioactive Lipids in Cancer, Inflammation and Related Diseases. Springer International Publishing, Cham, pp 169\u0026ndash;191\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuo A, Hla T (2024) Regulation of cellular and systemic sphingolipid homeostasis. Nat Rev Mol Cell Biol 25(10):802\u0026ndash;821\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBucki R, Kulakowska A, Byfield FJ et al (2010) Plasma gelsolin modulates cellular response to sphingosine 1-phosphate. Am J Physiol Cell Physiol 299(6):C1516\u0026ndash;C1523\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMendelson K, Evans T, Hla T (2014) Sphingosine 1-phosphate signalling. Development 141(1):5\u0026ndash;9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim D, Tian W, Wu TT-H et al (2023) Abnormal Lymphatic Sphingosine-1-Phosphate Signaling Aggravates Lymphatic Dysfunction and Tissue Inflammation. Circulation 148(16):1231\u0026ndash;1249\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim ES, Kim JS, Kim SG, Hwang S, Lee CH, Moon A (2011) Sphingosine 1-phosphate regulates matrix metalloproteinase-9 expression and breast cell invasion through S1P3-Gαq coupling. J Cell Sci 124(Pt 13):2220\u0026ndash;2230\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaceyka M, Spiegel S (2014) Sphingolipid metabolites in inflammatory disease. Nature 510(7503):58\u0026ndash;67\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFuhrman MP (2002) The albumin-nutrition connection: separating myth from fact. Nutr (Burbank Los Angeles Cty Calif) 18(2):199\u0026ndash;200\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Giovanni V, Bourdon C, Wang DX et al (2016) Metabolomic Changes in Serum of Children with Different Clinical Diagnoses of Malnutrition. J Nutr\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonzales GB, Njunge JM, Gichuki BM et al (2021) Albumin-dependent and independent mechanisms in the syndrome of kwashiorkor. medRxiv : 2021.05.31.21257914.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcMillan A, Orimadegun AE, Sumarah MW et al (2017) Metabolic derangements identified through untargeted metabolomics in a cross-sectional study of Nigerian children with severe acute malnutrition. Metabolomics 13(2):13\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNacis JS, Kamande P, Toni AT et al (2024) Barriers and enablers to the effective implementation of omics research in low- and middle-income countries. Nat Biotechnol 42(6):988\u0026ndash;991\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFronti\u0026egrave;res MS (2014) Nutrition Guidelines\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWHO/UN Children\u0026rsquo;s Fund. WHO Child Growth Standards and the Identification of Severe Acute Malnutrition in Infants and Children (2009) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003c/span\u003e\u003cspan address=\"http://www.who.int/nutrition/publications/severemalnutrition/9789241598163/en/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (accessed November 2016\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInstitut National de la Statistique (INS) (2013) et ICF International. Enqu\u0026ecirc;te D\u0026eacute;mographique et de Sant\u0026eacute; et \u0026agrave; Indicateurs Multiples du Niger 2012. Calverton, Maryland: INS et ICF International,\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonzales GB, De Saeger S (2018) Elastic net regularized regression for time-series analysis of plasma metabolome stability under sub-optimal freezing condition. Sci Rep 8(1):3659\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemichev V, Messner CB, Vernardis SI, Lilley KS, Ralser M (2020) DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat Methods 17(1):41\u0026ndash;44\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConsortium TU (2020) UniProt: the universal protein knowledgebase in 2021. Nucleic Acids Res 49(D1):D480\u0026ndash;D9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDesiere F, Deutsch EW, King NL et al (2006) The PeptideAtlas project. Nucleic Acids Res 34(suppl1):D655\u0026ndash;D8\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarney DJ, Larance M (2022) Annotated Protein Database Using Known Cleavage Sites for Rapid Detection of Secreted Proteins. J Proteome Res 21(4):965\u0026ndash;974\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulos RC, Hains PG, Shah R et al (2020) Strategies to enable large-scale proteomics for reproducible research. Nat Commun 11(1):3793\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSticker A, Goeminne L, Martens L, Clement L (2020) Robust Summarization and Inference in Proteome-wide Label-free Quantification. Mol Cell Proteom 19(7):1209\u0026ndash;1219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Paepe K, Kerckhof FM, Verspreet J, Courtin CM, Van de Wiele T (2017) Inter-individual differences determine the outcome of wheat bran colonization by the human gut microbiome. Environ Microbiol 19(8):3251\u0026ndash;3267\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, Gl\u0026ouml;ckner FO (2013) Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res 41(1):e1\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBolyen E, Rideout JR, Dillon MR et al (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 37(8):852\u0026ndash;857\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeSantis TZ, Hugenholtz P, Larsen N et al (2006) Greengenes, a Chimera-Checked 16S rRNA Gene Database and Workbench Compatible with ARB. Appl Environ Microbiol 72(7):5069\u0026ndash;5072\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamini Y, Hochberg Y (1995) Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J Royal Stat Soc Ser B (Methodological) 57(1):289\u0026ndash;300\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuznetsova A, Brockhoff PB, Christensen RHB (2017) 2017 lmerTest Package: Tests in Linear Mixed Effects Models. ; 82(13): 26\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDai F, Dutta S, Maitra RA, Matrix-Free (2020) Likelihood Method for Exploratory Factor Analysis of High-Dimensional Gaussian Data. J Comput Graphical Stat 29(3):675\u0026ndash;680\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEpskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D (2012) qgraph: Network Visualizations of Relationships in Psychometric Data. J Stat Softw 48(4):1\u0026ndash;18\u003c/span\u003e\u003c/li\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Kwashiorkor, Edematous malnutrition, Extracellular matrix, Sphingolipid, Edema, Metabolomics, Proteomics, Lipidomics, Gut microbiome","lastPublishedDoi":"10.21203/rs.3.rs-8320069/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8320069/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eEdematous malnutrition, aka kwashiorkor, is a phenotype of severe malnutrition whose pathophysiology remains poorly understood. In this case-control study, we employed plasma lipidomics, metabolomics, and proteomics, urine metabolomics and gut microbiome profiling to delineate molecular pathways specific to kwashiorkor in 60 children aged 6\u0026ndash;59 months from Niger compared to marasmus (n\u0026thinsp;=\u0026thinsp;60) and non-malnourished children (n\u0026thinsp;=\u0026thinsp;60) matched by age, sex, and clinical triage score. Features were defined as kwashiorkor-specific if they also correlated with edema severity and normalized following nutritional rehabilitation. Our analyses revealed that kwashiorkor is marked by increased extracellular matrix (ECM) degradation, evidenced by elevated plasma ECM proteins, and by disrupted sphingolipid homeostasis. Neither plasma nor urine metabolomic profiles, nor gut microbiome signatures, showed unique alterations associated with kwashiorkor. These findings suggest that kwashiorkor may be a combination of nutritional deficiencies and the disruption of the ECM and sphingolipid metabolism, potentially linked with an inflammatory syndrome.\u003c/p\u003e","manuscriptTitle":"Multi-omics, multi-tissue analysis reveal role of extracellular matrix remodeling and lipid transport dysfunction in edematous malnutrition (kwashiorkor)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-16 09:05:50","doi":"10.21203/rs.3.rs-8320069/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f044cdca-053d-4dc1-a7e0-d8986183c820","owner":[],"postedDate":"December 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":59645057,"name":"Health sciences/Molecular medicine"},{"id":59645058,"name":"Health sciences/Medical research/Paediatric research"},{"id":59645059,"name":"Biological sciences/Biochemistry/Proteomics"}],"tags":[],"updatedAt":"2026-01-10T10:50:31+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-16 09:05:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8320069","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8320069","identity":"rs-8320069","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}