Association between estimated GFR and incident hypoglycaemia during hospitalization.

BACKGROUND: Kidney function is a significant factor associated with increased incidence of hypoglycaemia, especially among patients with diabetes mellitus (DM). We here quantified the association between elevated creatinine and incident hypoglycaemia among patients admitted to internal medicine departments, with and without DM. METHODS: This is a retrospective cohort analysis study. Included were all patients discharged from internal medicine units between 2010 and 2013. Patients were excluded if creatinine levels rose or dropped more than 0.3 mg/dL during hospitalization. The CKD-EPI equation was used to calculate glomerular filtration rate (eGFR). Logistic regression analysis (backward LR method) was used to study the association between eGFR and hypoglycaemia incidence. RESULTS: Included were 39 316 patients (mean age 68.0 ± 18.0 years, 49.3% males, 25.9% with DM, eGFR 69.5 ± 24.9 mL/min/1.73 m2 ). Among study participants, 6.5% had at least one hypoglycaemic event. Logistic regression modelling showed that eGFR was inversely associated with incident hypoglycaemia (OR 0.988, 95% CI 0.986-0.990, p < .001). Results were similar for patients with and without DM. Estimated GFR was negatively correlated with admission CRP levels for patients with (r = -.143, p < .001) and without DM (r = -.166, p < .001). Estimated GFR was also positively correlated with admission serum albumin levels for both patients with (r = .304, p < .001) and without DM (r = .354, p < .001). CONCLUSION: Among non-critically-ill patients hospitalized in internal medicine departments, reduced eGFR is associated with increased risk of hypoglycaemia. Glucose monitoring for all inpatients with CKD is suggested, regardless of DM status.