Risk Factors for Hypoglycemia-Related Hospitalization in Patients With Type 2 Diabetes: A Nested Case–Control Study

Abstract

Background Hypoglycemia requiring hospitalization remains a serious and costly limitation to treatment of type 2 diabetes with antidiabetic medications. Objective We identified risk factors for hypoglycemia hospitalization in patients with type 2 diabetes treated with oral antidiabetic drugs (OADs). Methods In the 2004 to 2008 MarketScan database, we identified patients with type 2 diabetes taking OADs with >12 months of enrollment. We conducted a nested case–control study, selecting cases with an inpatient admission for hypoglycemia (first event). Using the index date of the cases, we conducted incidence density sampling to identify controls (10:1 matching) with continued eligibility during that month, further matching on date of cohort entry (±1 month). The final sample was 1339 cases and 13,390 controls. We assessed patterns of OAD availability (creating 3 groups: continuous, intermittent, and nonavailability), other medication availability, previous visits for hypoglycemia, complications of diabetes, and other comorbidities in the previous 180 days. A conditional logistic regression model identified predictors of hypoglycemia hospitalization. Results Mean (SD) age of cases was 56.4 (7.0) years compared with 54.6 (7.8) years in the controls. Overall, cases had more comorbidities than controls. In multivariable modeling, previous emergency department hypoglycemia visits (odds ratio [OR] = 9.48; 95% CI, 4.95–18.15) and previous outpatient hypoglycemia visits (OR = 7.88; 95% CI, 5.68–10.93) were strongly predictive of inpatient hypoglycemia admission. Continuous metformin availability had a 38% lower rate of inpatient hypoglycemia admission (OR = 0.62; 95% CI, 0.53–0.73) and intermittent metformin availability a 24% lower rate (OR = 0.76; 95% CI, 0.64–0.92) than nonavailability of metformin. Relative to nonavailability, continuous (OR = 2.25; 95% CI, 1.93–2.63) and intermittent sulfonylurea availability (OR = 2.28; 95% CI, 1.90–2.74) had increased rates of hypoglycemia hospitalization. Intermittent thiazolidinedione availability had a slightly increased rate of hypoglycemia hospitalization (OR = 1.22; 95% CI, 1.01–1.47). Continuous availability of thiazolidinediones and continuous or intermittent use of other OADs were not predictive of hypoglycemia admission. Conclusions Previous outpatient or emergency department visits for hypoglycemia and continuous or intermittent sulfonylurea availability were found to be predictive of costly inpatient hypoglycemia admissions. Although this observational study may not be generalizable to all patients with type 2 diabetes and assessed medication availability rather than actual consumption, previous outpatient visits and prescription for OADs should serve as points of intervention and patient education.