OBJECTIVE: The aim of this study was to investigate the effect of caffeine (in doses equivalent to normal daily ingestion) on rates and severity of hypoglycemia in patients with long-standing type 1 diabetes to determine the relationship between caffeine, autonomic function, and hypoglycemia. RESEARCH DESIGN AND METHODS: Using a double-blinded randomized study, we investigated the effect of caffeine versus placebo in 19 patients with long-standing type 1 diabetes using continuous glucose sensing technology and simultaneous assessment of autonomic function using Holter monitoring. RESULTS: Caffeine reduced the duration of nocturnal hypoglycemia with a mean duration of 49 minutes (range 0-235) versus 132 (0-468) minutes (P = 0.035). The reduction in duration of nighttime hypoglycemia was due to a decline in the number of episodes of moderate hypoglycemia at the expense of mild hypoglycemic episodes (P = 0.04). There was no overall correlation between reduced heart rate variability (a marker of autonomic dysfunction) and hypoglycemic events (r(s) = 0.12, P = 0.62). CONCLUSIONS: Our results suggest that caffeine is associated with a significant reduction in nocturnal hypoglycemia. The reduction in nocturnal hypoglycemia was not linked to the concomitant rise in parasympathetic activity associated with caffeine.