Background: New-onset hyperglycaemia is more strongly associated with mortality in critically ill patients in the Intensive Care Unit (ICU) than chronic hyperglycaemia. However, it has not been clear how best to identify new-onset, or relative, hyperglycaemia. We recently proposed a novel metric for relative hyperglycaemia, the stress hyperglycaemia ratio (SHR) (1). We investigated whether SHR independently predicts mortality in ICU patients.
Methods: We prospectively studied 1,193 consecutive adult patients admitted to ICU, Flinders Medical Centre, Adelaide between January 2016 and February 2017. The primary endpoint was in-hospital mortality. The variables of interest were admission glucose and SHR (calculated as glucose at ICU admission divided by the patient’s estimated average glucose over the prior 3 months which was derived from glycosylated haemoglobin). Survival data and APACHE III, a score derived from a comprehensive set of clinical and laboratory data that is used to predict mortality risk in ICU, were retrieved from the ICU registry.
Results: In univariate analyses, mortality was associated with admission glucose (p<0.001), SHR (p<0.001) and APACHE III score (p<0.001), but not with diabetes (p=0.498), or sex (p=0.877). A 0.1 increase in SHR was associated with a 5.3% increment in mortality. When APACHE III score was included as a covariate, SHR remained a significant predictor of mortality (p=0.002), while glucose was not (p=0.169). In subgroup analyses, SHR was associated with mortality in patients without (p<0.001) and with (p=0.004) diabetes. In contrast, glucose was significantly associated with mortality in patients without (p<0.001), but not with (p=0.083), diabetes.
Conclusion: Relative hyperglycaemia, as defined by SHR, predicts mortality in ICU patients across the glycaemic spectrum independent of other prognostic clinical variables, while absolute glucose does not. Future studies should explore whether calculating individualized therapeutic targets for glucose-lowering therapy based on relative, rather than absolute, hyperglycaemia reduces mortality in critically ill patients.