Predictors of 7-day mortality in critically ill patients with hyperglycemic crisis : A single center retrospective analysis.

Aim/hypothesis The main aim of the study was to identify point of care available laboratory and clinical predictors of 7‑day mortality in critically ill patients with a hyperglycemic crisis. Methods A retrospective study of 990 patients with the first hospitalization due to hyperglycemia was performed. Patients were classified as having diabetic ketoacidosis (DKA) or being in a hyperosmolar hyperglycemic state (HHS) according to the recommendations of the American Diabetes Association (ADA). Patients not fulfilling the ADA criteria for DKA or HHS were summarized in a third group (unclassifiable hyperglycemia, UCH). The primary outcome was 7‑day mortality, potentially relevant factors were analyzed as secondary outcomes. Results Overall, the 7‑day mortality was 7.5%, with no significant differences between DKA (7.8%), HHS (14.5%) and UCH (6.1%). Blood lactate levels were significantly higher in nonsurvivors than survivors in all three groups (mean level of 6.3 mmol/l vs. 3.4 mmol/l in DKA, 5.3 mmol/l vs. 3.1 mmol/l in HHS, 5 mmol/l vs. 2.5 mmol/l in UCH). Measured and calculated osmolality were significantly higher in nonsurvivors in the DKA group (measured osmolality 359 mosmol/kg vs. 338 mosmol/kg, calculated osmolality 315 mosmol/kg vs. 305 mosmol/kg) and patients with UCH (354 mosmol/kg vs. 325 mosmol/kg; 315 mosmol/kg vs. 298 mosmol/kg) but not in patients with HHS. Survival analysis for the DKA group showed no significant differences in 7‑day mortality when patients were compared by the ADA criteria of severity (severe, moderate, or mild). Patients with elevated calculated osmolality (> 320 mosmol/kg) and lactate (> 4 mmol/l) had the lowest 7‑day survival rate (66.7%). Conclusion/interpretation Our data showed that elevated lactate levels were associated with higher mortality in all types of hyperglycemic crises.