Is diabetes mellitus related to undernutrition in rural Tanzania?

OBJECTIVE To investigate the relation between undernutrition and diabetes. DESIGN Survey of glucose tolerance in rural Tanzania. SETTING Eight villages in three widely separated regions of Tanzania. SUBJECTS 8581 people aged 15 and above: 3705 men and 4876 women. MAIN OUTCOME MEASURES Oral glucose tolerance, body mass index, height, and low haemoglobin and cholesterol concentrations. RESULTS In the eight villages 42.7-56.9% of all men and 30.0-45.2% of all women had a body mass index below 20 kg/m2; the lowest quintile was 18.2 kg/m2 in men and 18.6 kg/m2 in women. The prevalence of diabetes did not change significantly from the lowest to the highest fifths of body mass index in men (lowest 1.6% (95% confidence interval 0.8% to 2.9%) v highest 1.3% (0.7% to 2.5%)) or women (1.1% (0.6% to 2.1%) v 0.5% (0.2% to 1.2%)). In men and in women prevalence of impaired glucose tolerance was greater in the lowest fifths of height (8.2% (6.3% to 10.6%), and 11.1% (9.2% to 13.3%)) respectively and body mass index (9.6% (7.5% to 12.1%), and 8.4% (6.7% to 10.5%)) than in the highest fifths (impaired glucose tolerance 4.7% (3.4% to 6.5%); and 5.1% (3.9% to 6.7%); body mass index 5.1% (3.7% to 7.0%), and 7.7% (6.2% to 9.6%). CONCLUSION Rates of diabetes were not significantly associated with low body mass index or height, but overall rates were much lower than those in well nourished Western populations. Increased impaired glucose tolerance in the most malnourished people may reflect the larger glucose load per kilogram weight. The role of undernutrition in the aetiology of diabetes must be questioned.