Severe Hypoglycemia, Risk of Recurrent Events and Weight Gain on Insulin Therapy in Patients with Type 1 Diabetes Mellitus.

摘要

Background. Few studies provide epidemiologic evidence to demonstrate acute effects of severe hypoglycemia (SH) on subsequent SH in patients with type 1 diabetes mellitus (T1DM) under a clinically relevant time range. Although it is well accepted that hypoglycemia could associate with subsequent weight gain in T1DM, there is limited direct evidence to support this hypothesis.;Methods. We conducted a secondary data analysis using data from the Diabetes Control and Complications Trial (DCCT) which randomized T1DM patients to either intensive (IT) or conventional therapy (CT) and followed them quarterly. We estimated relative risks (RR) for subsequent SH in three consecutive quarterly time windows following occurrence of SH (index SH). We estimated the effects of SH on subsequent weight change/weight (kg) in various observation periods. We used generalized estimating equations to account for the dependence of multiple-observations within a person and to adjust for confounding. Hazard ratios (HRs) of SH on substantial weight gain, overweight and obesity were estimated using Cox and marginal structural models.;Results. in both treatment arms, the greatest absolute risks and RRs for subsequent SH after index SH were observed in the first time window in three consecutive quarterly windows following index SH. In IT, the estimated effect of SH on weight change/weight was close to null effect. In CT, a weight loss was observed during the 1st 3-month following SH [adjusted difference of the means for weight change (MDWC) was -0.29 (-0.51, -0.08)] comparing to those without SH. The HRs of SH on substantial weight gain, becoming overweight or obese were also found around the null effect (HR=1) in both treatment arms.;Conclusions. This study provides direct evidence to support an acute effect of recent SH on subsequent SH episodes, and the results indicate that the immediate time periods after occurrence of SH are crucial in clinical management of T1DM to prevent subsequent SH. We did not find evidence to support an association between occurrence of SH and subsequent weight gain in patients with T1DM. Because the DCCT is a clinical trial design, one should be cautious in extrapolating our findings to all patients with T1DM.