OBJECTIVES: The aim of this study is to investigate the impact of glucose dysregulation ((hyperglycemia and/or hypoglycemia) on ICU and in-hospital mortality.

METHODS: Patients from the Medical Information Mart for Intensive Care IV (MIMIC IV) database were used for the analysis. The first 24 hours glucose levels at ICU stays were used to divide patients into hypoglycemia (>180) , hyperglycemia, hypoglycemia & hyperglycemia, and normoglycemia groups. Propensity score matching was based on age, gender, race, Charlson Comorbidity Index (CCI), and Acute Physiology Score III (APSIII). In addition, the same analyses were performed for patients with a diabetes diagnosis and without a diabetes diagnosis respectively.

RESULTS: There were 38,290 patients in the analysis: hypoglycemia (n=805), hyperglycemia(23,196), hypoglycemia & hyperglycemia (n=1,636), and normoglycemia (n=12,653). After propensity score matching, compared with normoglycemia, hypoglycemia was not associated with ICU mortality (OR:1.01; p=1) and in-hospital mortality (OR:1.03; p=0.848); hyperglycemia was associated with higher ICU mortality (OR:1.40; p<0.001) and in-hospital mortality (OR:1.25;p<0.001); hypoglycemia & hyperglycemia showed no significant difference in ICU (OR:0.91; p=0.397) and in-hospital mortality (OR:0.89; p=0.254). In diabetic patients, hypoglycemia was not associated with ICU mortality (OR:1.18; p=0.603) and in-hospital mortality (OR:1.16; p=0.625); hypoglycemia & hyperglycemia was not associated with ICU (OR:0.94; p=0.79) and in-hospital mortality (OR:0.83; p=0.264). In nondiabetic patients; the hyperglycemia was highly associated with ICU mortality (OR:1.54; p<0.001) and in-hospital mortality (OR:1.36; p<0.001).

CONCLUSIONS: Hyperglycemia is associated with higher mortality in overall and nondiabetic patients in critical care . Hypoglycemia and hypoglycemia & hyperglycemia are not statistically significant with mortality.