Association between mild and severe hypoglycemia in people with type 2 diabetes initiating insulin

Aims Primary objective: Identify risk factors associated with severe hypoglycemia (SH) and investigate the association between mild hypoglycemia and SH in people with type 2 diabetes starting insulin. Secondary objectives: Investigate the association of demographics and clinical factors with SH incidence. Methods Integrated trial database data were obtained for 3 randomized controlled trials that included insulin-naïve people with type 2 diabetes initiating basal (insulin glargine) versus biphasic (insulin lispro mixture) insulin. Standard definitions were used for SH; mild hypoglycemia was defined as all non-SH. Cox regression identified risk factors associated with SH and the correlation between SH and mild hypoglycemia. Results Data were pooled (N = 2931). During 24–48 weeks' treatment, 2127 (72.6%) participants experienced ≥ 1 mild hypoglycemic event but no SH (mean mild hypoglycemia rate = 2.33/month). 56 participants (1.9%) experienced ≥ 1 SH event plus mild hypoglycemia (mean mild hypoglycemia rate = 3.95/month); 748 participants (25.5%) had no hypoglycemia. Among factors tested, only mild hypoglycemia rate/month was associated with SH. SH risk was higher (HR = 4.24; 95%CI = 2.57–6.99;p < 0.0001) for participants experiencing multiple mild hypoglycemic events/month compared with those experiencing ≤ 1 mild hypoglycemic event/month. Conclusions Mild hypoglycemia may predict the first SH event, which is important because SH is a strong and consistent risk factor for morbidity/mortality.