Prevalence and predictive value of ionized hypocalcemia among critically ill patients

Background: Ionized hypocalcemia is common among critically ill patients, and it has been shown to correlate with increased mortality. The purpose of this study was to examine the performance and independence of ionized calcium (Ca2+) in prediction of all-cause day-30 mortality among critically ill adult patients. Methods: Of 993 critically ill patients treated in the Helsinki University Hospital during a 24-month period, the study comprised 941 patients without calcium supplementation. Patient and laboratory data were obtained retrospectively from an intensive care database. The discriminative powers of admission and lowest Ca2+ values regarding day-30 mortality were evaluated by producing receiver operating curves (ROC). Hazard ratios for death of severe and mild hypocalcemia were calculated by Cox regression model. Results: The prevalence of ionized hypocalcemia (Ca2+ <1.16 mmol l(-1)) was 85%. Of 941 patients, 45 (4.7%) had ionized calcium >1.3 mmol l(-1) and were excluded from mortality analysis. Univariate Cox regression model revealed hazard ratios of 5.1 (95% confidence interval, CI 2.9-9.0) for severe (<0.90 mmol l(-1)) and 1.8 (95% CI 1.3-2.4) for mild ionized hypocalcemia (0.90-1.15 mmol l(-1)) on admission, but hypocalcemia was not shown to be independently associated with mortality by multivariate Cox regression model. In prediction of day-30 mortality admission and lowest Ca2+, levels had areas under curves of 0.636 and 0.671, respectively. Conclusions: Ionized hypocalcemia is common among critically ill adults and it is associated with increased mortality. Although non-survivors and survivors differ significantly in admission Ca2+, hypocalcemia is not independently associated with day-30 mortality.