Association between glycemic variability and short-term mortality in patients with acute kidney injury: a retrospective cohort study of the MIMIC-IV database

Abstract

Acute kidney injury (AKI) represents a significant challenge to global public health problem and is associated with poor outcomes. There is still considerable debate about the effect of mean blood glucose (MBG) and coefficient of variation (CV) of blood glucose on the short-term mortality of AKI patients. This retrospective cohort study aimed to explore the association between glycemic variability and short-term mortality in patients with AKI. Data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were analyzed, including 6,777 adult AKI patients. MBG and CV on the first day of ICU admission were calculated to represent the overall glycemic status and variability during the ICU stay in AKI patients. The primary outcome indicator was ICU 30-day mortality of AKI patients. Multivariate Cox regression analysis and smoothed curve fitting were used to assess the relationship between blood glucose levels and mortality. Eventually, the ICU 30-day mortality rate of AKI patients was 23.5%. The increased MBG and CV were significantly correlated with ICU 30-day mortality (hazards ratio (HR) = 1.20, 95% confidence interval (CI) 1.14-1.27; HR = 1.08, 95% CI 1.03-1.13). The smoothed curve fitting showed a U-shaped relationship between MBG on the first day of ICU admission and ICU 30-day mortality (inflection point = 111.3 mg/dl), while CV had a linear relationship with 30-day ICU mortality. Thus, we conclude that MBG and CV were significantly associated with short-term mortality in intensive care patients with AKI. Tighter glycemic control may be an effective measure to improve the prognosis of patients with AKI.

Figure 1

Flowchart of participant selection.

Figure 2

The relationship of MBG and CV associated with ICU 30-day mortality. (A) The non-linear relationship between MBG and ICU 30-day mortality. (B) The linear relationship between CV and ICU 30-day mortality. Hazard ratios (HRs) were adjusted for age, sex, heart rate, SBP, DBP, SpO2, platelets, WBC, Cr, BUN, myocardial infarct, cerebrovascular disease, liver disease, kidney disease, malignant cancer, sepsis, hypertension, diabetes, AKI stage, RRT use, mechanical ventilation, vasoactive drug use, hypoglycemic drug use, SOFA score, and comorbidity index. MBG mean blood glucose, CV coefficient of variation, SBP systolic blood pressure, DBP diastolic blood pressure, SpO2 blood oxygen saturation, WBC white blood cell, Cr creatinine, BUN blood urea nitrogen, RRT renal replacement therapy, SOFA sequential organ failure assessment.

Figure 3

Kaplan–Meier survival curves for ICU 30-day mortality. (A) MBG (mg/dl) with ICU 30-day mortality; (B) CV (%) with ICU 30-day mortality. MBG mean blood glucose, CV coefficient of variation.

Figure 4

Subgroup analyses of the MBG and CV associated with ICU 30-day mortality. Hazard ratios (HRs) were adjusted for age, sex, heart rate, SBP, DBP, SpO2, platelets, WBC, Cr, BUN, myocardial infarct, cerebrovascular disease, liver disease, kidney disease, malignant cancer, sepsis, hypertension, diabetes, AKI stage, RRT use, mechanical ventilation, vasoactive drug use, hypoglycemic drug use, SOFA score, and comorbidity index. MBG mean blood glucose, SBP systolic blood pressure, DBP diastolic blood pressure, SpO2 blood oxygen saturation, WBC white blood cell, Cr creatinine, BUN blood urea nitrogen, RRT renal replacement therapy, SOFA sequential organ failure assessment.