Table 3.

Summary of studies examining effects of documented severe malnutrition in childhood on non-communicable disease (NCD) outcomes

Study	Setting and population	Type/timing of severe malnutrition (SM) exposure	Outcomes	Key findings*	Risk-of-bias score (IV/EV)†
Case-control studies
Chege15	Cases: patients with diabetes 61.8±10.9 years, Kenya (n=45) Controls: age and sex-matched non-diabetics from same area attending outpatient clinics (n=45)	Self-reported episode of SM in childhood Exposure age not specified	T2D risk factors	↑ Childhood SM among diabetics	– / –
Fekadu et al16	Cases: insulin-requiring diabetics 18–40 years, Ethiopia (n=107) Controls: age and sex-matched patients attending other hospital clinics (n=110)	Self-reported episode of childhood SM Exposure age not specified	Insulin-requiring diabetes risk factors	↑ Childhood SM in diabetics	·/ –
Prospective cohort studies
Benefice et al65	Ex-malnourished: children 5.5±0.5 years, Senegal (n=52) Chronic controls: chronically undernourished children (n=54) Well-nourished controls (WN): age-matched, well-nourished children (n=33)	Marasmus Median age: 14 months	Motor fitness, anthropometry	↓ Handgrip in post-SM versus chronic controls ↓ Height/weight for age versus WN controls ↓ Distance throw, jump, agility/shuttle run versus WN controls ↔ Endurance run	·/ –
Boulé et al52	Ex-malnourished: young men 22.0±3.6 years, Mexico (n=26) Controls: young men 26.5±2.1 years with no history of SM (n=27)	Marasmus, kwashiorkor Age at admission: ≤1 years	Insulin sensitivity, abdominal obesity	↓ Insulin sensitivity w/ high abdominal fat versus fat-matched controls	· / –
Bourdon et al77	Ex-malnourished: children 9.6±1.6 years, Malawi (n=69) Sibling controls: closest in age to case child with no history of SM (n=44) Community controls: age and sex-matched (n=37)	Marasmus and kwashiorkor Median age at admission: 21.5 months	Cardiometabolic disease markers	↔ Metabolites	· / –
Cook23	Ex-malnourished: children 6.7–14.9 years, Uganda (n=31) Controls: children raised in similar environment as cases with no history of SM (n=21)	Kwashiorkor Mean age at admission: 1.9 years	Carbohydrate tolerance	↓ Glucose clearance ↑ Blood glucose 2 hours post oral glucose tolerance test (OGTT)	· / –
Francis-Emmanuel et al53	Ex-malnourished: adult marasmus survivors (MS) (n=42) and kwashiorkor survivors (KS) (n=38) 17–50 years, Jamaica Community controls: age, sex, BMI-matched (n=70) Birthweight-matched controls: age-matched (n=40)	Marasmus and kwashiorkor Age at admission: 6–18 months	Glucose metabolism	↔ Fasting plasma glucose ↑ Glucose intolerance (MS versus KS) ↓ Insulin sensitivity (MS versus KS) ↔ Insulin sensitivity (MS versus controls) ↓ Insulinogenic and oral disposition indices (MS vs all groups)	·· / –
Gonzalez-Barranco et al42	Ex-malnourished: young men 20.2±3.6 years, Mexico (n=52) Controls: young men with no history of SM (n=50)	Marasmus, kwashiorkor Mean age at admission: 4.5 months	Glucose metabolism, lipid profile, blood pressure (BP)	↑ Areas under the curves of glucose and insulin ↓ Insulin sensitivity, BP ↔ Fasting blood glucose, lipid profile	·· / –
Idohou-Dossou et al20	Ex-malnourished: children 6–8 years, Senegal (n=24) Sibling controls (SC): closest in age to case child with no history of SM (n=24) Well-nourished controls (WN): age-matched healthy children from wealthier area (n=19)	Marasmus Age at admission: 1–3 years	Biochemical nutritional indicators, growth factors, anthropometry	↓ Apolipoprotein A↔ versus WN controls, no difference between post-SM and SC ↓ Lean mass in post-SM and SC associated with low IGF-↔	· / –
Kajubi24	Ex-malnourished: adolescents 11–19 years, Uganda (n=15) Controls: adolescents with no history of SM (n=11)	Kwashiorkor Age at admission: 1.5–3 years	Pancreatic function	↔ Blood glucose post-OGTT ↓ Fasting plasma insulin	·/–
Lelijveld et al9	Ex-malnourished: children 9.6±1.6 years, Malawi (n=320) Sibling controls (SC): closest in age to case child with no history of SM (n=217) Community controls (CC): age and sex-matched with no history of SM (n=184)	Marasmus, kwashiorkor Median age at admission: 24 months	Blood markers of NCDs, physical capacity, anthropometry	↔ Glucose tolerance, glycosylated haemoglobin, blood lipids, salivary cortisol ↑ Diastolic BP in post-SM versus SC ↓ Handgrip strength versus CC/SC ↓ Lean mass versus CC but similar to SC	··/–
Moore et al25	Rural adults (mean age 35.8 years), Gambia (n=145)	Low weight-for-age z-score (WAZ) WAZ measured at 18 months	Cardiovascular disease (CVD) risk factors	↓ Fasting plasma insulin in lower WAZ quartiles ↔ Fasting blood glucose, blood glucose or insulin post-OGTT, cortisol, BP	··/–
Sheppard et al84	Ex-malnourished: adult survivors of kwashiorkor (KS) 29.82±9.03 years (n=21) or marasmus (MS) 25.02±5.69 years (n=23), Jamaica	Marasmus, kwashiorkor Mean age at admission: 11 months	Epigenetic profile in muscle tissue	↕ Differences in DNA methylation of 63 genes related to, body size/composition, glucose metabolism, musculoskeletal growth, cardiovascular pathways between MS and KS	·/–
Tennant et al43	Ex-malnourished: adult survivors of childhood kwashiorkor (n=62) 27.2±7.8 years and marasmus (n=54) 29.2±8.4 years, Jamaica Community controls: age and sex matched with no history of SM (n=45)	Marasmus, kwashiorkor Mean age at admission: 12 months	Cardiovascular structure/function	↓ Left ventricular outflow tract parameter, stroke volume, cardiac output, pulse wave velocity ↑ Diastolic BP ↔ Systolic BP ↑ Systemic vascular resistance ↑ Heart rate in MS versus KS ↔ Large vessel, cardiac remodelling	··/–

Acceptable IV and EV . Poor IV or EV . Poor IV and EV .

*Symbols for effect direction: ↑ increased; ↓ decreased; ↕ mixed (indicate statistically significant results were reported, defined as p<0.05); ↔ none (indicates no statistically significant result was reported). If no age group is indicated beside the finding, then all age groups were affected.

†The scoring system used in the risk-of-bias assessment is described in the Methods section.

T2D, type 2 diabetes.