Mortality Among Hospitalized Patients With Hypoglycemia: Insulin Related and Noninsulin Related

Abstract

Context: Hypoglycemia is common among hospitalized patients with and without diabetes mellitus. Objective: Investigate the association between spontaneous or insulin-related hypoglycemia and mortality in hospitalized patients. Design: Hypoglycemia was defined as blood glucose <70 mg/dl (3.9 mmol/l), including moderate (40 to 70 mg/dl, 2.2 to 3.9 mmol/l) and severe hypoglycemia (<40 mg/dl, 2.2 mmol/l). Use of insulin during hospitalization defined insulin-related hypoglycemia, thus patients were classified into 6 groups: non-insulin treated (NITC) and insulin-treated controls (ITC), insulin-related hypoglycemia (IH) or severe hypoglycemia (ISH), and non insulin-related hypoglycemia (NIH) and severe hypoglycemia (NISH). Setting and Patients: Historical prospectively data of patients ≥ 18 years of age, hospitalized in medical wards for any cause between January 2011 and December 2013. Main Outcome Measure: All-cause mortality at the end of follow-up. Results: The cohort included 33,675 patients, including 2605 with moderate hypoglycemia (IH, 1011; NIH, 1594) and 342 with severe hypoglycemia (ISH, 201; NISH,141). Overall end-of-follow-up mortality was 31.9% (NITC, 28.0%; ITC, 42.9%; NIH, 50.7%; IH, 55.3%; NISH, 70.9%; ISH, 69.1%). Compared with NITC, unadjusted hazard ratios (95% confidence intervals) for mortality were as follows: ITC, 1.7 (1.6 to 1.8), NIH, 2.2 (2.0 to 2.4), IH, 2.5 (2.2 to 2.7), NISH, 4.2 (3.5 to 5.2), and ISH, 3.8 (3.2 to 4.5); with P < 0.001. Following multivariate analysis, respective hazard ratios were 1.8, 2.1, 2.4, 3.2, and 3.6 (P < 0.001). Cause of admission did not affect the association. Conclusions: In hospitalized patients, hypoglycemia, either with insulin use or spontaneous, is associated with increased short- and long-term mortality.