Objectives: To examine the effect of diabetes per se and nocturnal hypoglycemia on sleep in children with insulin-dependent diabetes mellitus (IDDM). Design: Overnight polysomnography was performed on 3 occasions in 29 children with IDDM - twice during metabolic profiling. Sleep data were analyzed from 14 children (median [range], 8.7 [5.9 to 12.9] years) with a night of hypoglycemia and a nonhypoglycemic night. Seven children in the control group (9 [5.6 to 11.4] years) underwent 2 nights of polysomnography, once during metabolic profiling (to assess the effects of metabolic profiling), and 15 members of the control group had polysomnography only (to assess the effects of diabetes per se and to compare with the index group). Results: Children with IDDM had disrupted sleep compared with the control group (short wakes, percentage of sleep time, 0.8 [0.5 to 1.9] vs 0.0 [0.0 to 0.3], median [interquartile range], IDDM vs control group, respectively, P = .001; long wakes, 1.2 [0.7 to 3.0] vs 0.0 [0.0 to 0.0], P = .033; total number of wakes, 6 [3.5 to 11.5] vs 1 [0 to 2], P = .002). Blood sampling disrupted sleep with increased long wakes as percentage of sleep time (0.3 [0.0 to 3.8] vs 4.3 [3.1 to 16.9], nonsampling vs sampling night, respectively, P = .003). Episodes of hypoglycemia were profound, with a glucose nadir of 2.0 [1.4 to 3.3] mmol/L (35.0 [24.5 to 57.8] mg/dL) and prolonged, 308 [30 to 630] minutes, with no effect on sleep. Conclusions: Children with diabetes had disrupted sleep compared with a control group, but there was no effect of profound nocturnal hypoglycemia on sleep quality. J Pediatr 2000;137:233-8)