Nocturnal hypoglycaemia in type 1 diabetes--frequency and predictive factors

BACKGROUND: Nocturnal hypoglycaemia (NH) remains a problem in type 1 diabetes and spontaneous asymptomatic NH may be a risk factor for sudden death ('Dead in Bed' syndrome). AIMS: To explore whether any predictive relationship exists between the average or time-specific glycaemia and the occurrence of NH. METHODS: Twenty-five healthy patients with type 1 diabetes underwent two separate overnight periods of continuous glucose monitoring (CGM) using a MMT-7002 Medtronic MiniMed System. There was a 6-week interval before the second monitoring period. CGM glucose levels recorded between 23:00 and 08:00 h defined the nocturnal period and recorded glucose monitoring levels <3.5 mmol/l for at least 10 min during this time-defined NH. A CGM recording at 23:00 h and 08:00 h were taken as the bedtime and fasting glucose levels, respectively. RESULTS: The mean +/- SD age was 37 +/- 7 years and duration of diabetes 13 +/- 7 years; 16 (64%) were on long-acting analogue insulin. Forty-nine CGM data sets were recorded. Fourteen episodes of NH occurred in 12 patients (Group 1), 13 patients (Group 2) had no NH. Group 1 (NH) had a lower mean bedtime glucose recorded compared with Group 2 (7.7 +/- 4.3 vs. 11.4 +/- 4.0 mmol/l, P = 0.0035). Fasting glucose level was also lower in Group 1 following the occurrence of NH (P = 0.014). There was no difference in the type of insulin used between the two groups. CONCLUSION: Our data show that in normal day to day settings, NH is common and that the bedtime glucose level is a significant predictive factor.