AIMS/HYPOTHESIS: In the Western world, non-autoimmune diabetes in the young is believed to be driven by overweight/obesity and insulin resistance. However, it is increasingly being reported in undernourished people in low- and middle-income countries, including India. We hypothesised that these patients would show markers of chronic undernutrition and a 'thin-fat' phenotype and be predominantly beta cell-deficient.

METHODS

We studied young patients (clinically diagnosed with type 2 diabetes at <40 years) who attended the outpatient department of Assam Medical College and Hospital, Dibrugarh (in North-East India). We measured weight, height, waist and hip circumference, haemoglobin, fasting glucose, HbA, lipid, GADA and C-peptide levels, and body fat percentage (adiposity, assessed using dual-energy x-ray absorptiometry), and calculated BMI (kg/m), body roundness index and HOMA indices. Volunteers from similar socioeconomic background with normal glucose tolerance (measured by 75 g OGTT) were assessed as control participants. We also compared the anthropometric characteristics and body composition of our participants with those of non-Hispanic white Americans from the NHANES study.

RESULTS

The study included 252 control participants (136 male participants, median age 30 years, BMI 23.0 kg/m) and 240 GADA-negative young patients with diabetes (155 male participants, age 36 years, BMI 23.0 kg/m). The majority of study participants came from a relatively impoverished population of tea garden workers ('tribal' workers). Of the patients with diabetes, 28% had stunted growth (male <161.2 cm, female <149.8 cm), 27% were anaemic, 68% were lean (BMI <25 kg/m, including 14% who were underweight [BMI <18.5 kg/m]) and 32% were overweight/obese (BMI ≥25 kg/m). When assessed using dual-energy x-ray absorptiometry, 61% of control participants and 53% of patients had adiposity (body fat percentage >25% in male participants or >35% in female participants). Compared with a contemporary non-Hispanic white American population, Assamese control participants and diabetic patients had higher WHR, body roundness index, and total and truncal adiposity (assessed using dual-energy x-ray absorptiometry) across the range of BMI, thus conforming to the description of the 'thin-fat' phenotype. The diabetic patients were severely beta cell-deficient (median HOMA-B 25.7) and only moderately insulin-resistant (median HOMA-S 103) with higher triacylglycerol and lower HDL-cholesterol concentrations than control participants. Underweight patients (<18.5 kg/m) were the most hyperglycaemic (based on fasting plasma glucose and HbA), and were severely beta cell-deficient but insulin-sensitive. As previously reported, two-thirds of these patients belonged to the severely insulin-deficient diabetes (SIDD) cluster according to the Swedish diabetes subgroup classification.

CONCLUSIONS/INTERPRETATION: Diabetes in the young people of this impoverished population is heterogeneous, but the majority of patients are not overweight/obese or insulin-resistant. Overall, these participants conform to the thin-fat phenotype, and their diabetes is predominantly driven by beta cell deficiency. The sociodemographic history and physical characteristics of this population suggest a role for multigenerational undernutrition in the aetiology of non-autoimmune diabetes in these young patients from Assam.