The Association Between Low Admission Peak Plasma Creatinine Concentration and In-Hospital Mortality in Patients Admitted to Intensive Care in Australia and New Zealand

Abstract

Objective: To evaluate the independent association between low peak admission plasma creatinine concentrations and in-hospital mortality in patients requiring critical care in Australia and New Zealand.

Design: Multicenter, binational, retrospective cohort study.

Setting: Data were extracted from the Australian and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation adult patient database.

Patients: All available records for the period 2000 to 2013 were utilized. The following exclusion criteria were applied: all readmission episodes (within the same hospital stay), missing in-hospital mortality, admission post kidney transplantation, chronic renal replacement therapy (hemodialysis or peritoneal dialysis), and missing peak plasma creatinine concentration. Demographic, anthropometric, admission, illness severity, laboratory, and outcome data were then extracted. Patients were categorized on the basis of their peak (maximum) plasma creatinine concentration recorded in the first 24 hours of ICU admission. Illness severity-adjusted associations with in-hospital mortality relative to a reference category of 70-79 μmol/L were then determined using multivariate logistic regression.

Interventions: Nil.

Measurements and main results: Data pertaining to 1,250,449 admissions were available for the study period. Following exclusions, 1,045,718 patients were included. Regression analysis identified that peak plasma creatinine concentrations less than 60 μmol/L measured in the first 24 hours after ICU admission imply a steadily increasing adjusted in-hospital mortality risk. In cases where this value is markedly low (< 30 μmol/L), the adjusted odds of dying in-hospital is over two-fold higher than the reference category and exceeds the risk implied with elevated (≥ 180 μmol/L) values. This finding was also independent of anthropometric data.

Conclusions: In a large heterogenous cohort of critically ill patients, low admission peak plasma creatinine concentrations are independently associated with increased risk-adjusted in-hospital mortality. Further research should now focus on the potential mechanisms underpinning this finding, such as a low skeletal muscle mass and/or fluid overload.