The present study aimed to observe the clinical features of hypoglycaemia, and identify predictors of hypoglycaemia in under-five diarrhoeal children requiring hospitalization for close observation and support. Such information could be useful to the clinicians and policy makers in developing appropriate management protocols both for identification of such children and optimizing their management. We performed a prospective study in 782 under-five children who presented with diarrhoeal illnesses. Blood glucose was determined when hypoglycaemia was suspected in 598 (62%), and 65 (11%) of them were hypoglycaemic (study group). From the other 533 non-hypoglycaemic children, 195 were randomly selected as comparison group. Bacteraemia was significantly (P = 0.026) often detected in 17 out of 260 (7%) children as opposed to 3 out of 184 (2%) children who did not have a rapid glucose test performed. Among hypoglycaemic children, 7 (11%) were bacteraemic and among non-hypoglycaemic children 10 (5%) had bacteraemia. In univariate analysis, history of shorter (<72 h) pre-admission duration of diarrhoea (75 vs. 58%, P = 0.01), documented convulsion (28 vs. 11%, P < 0.001), shorter (<72 h) hospitalization (52 vs. 33%, P = 0.01), higher case fatality rate (28 vs. 14%, P = 0.02) were associated with hypoglycaemia. In logistic regression, bacteraemic children (with clinical sepsis) were 4 times more likely to develop hypoglycaemia (OR = 4.2, 95% CI = 1.4-12.9, P = 0.012). Therefore, in a diarrhoeal disease health care service with limited resources, a rapid bedside glucose test may be considered as an inexpensive alternative in the management decisions of diagnosing bacteraemia and initiating empiric antibiotic treatment.