Dose–response of intravenous calcium in the surgical intensive care unit

Abstract

Background Hypocalcemia is common in patients admitted to the surgical intensive care unit and is associated with increased morbidity and mortality. Current dosing strategies do not always achieve ionized calcium (iCa) normalization, especially in patients with severe hypocalcemia. Objective The purpose of this study was to explore the association between intravenous (IV) calcium dose and change in ionized calcium. Setting Patients admitted to the surgical intensive care unit with concomitant hypocalcemia at a large academic hospital in the United States. Method This single center, retrospective cohort study evaluated the association between IV calcium dose and subsequent change in ionized calcium level in adult surgical intensive care unit patients with hypocalcemia. The primary outcome of this study was to develop a model exploring the association between IV calcium dose and change in iCa levels. Secondary outcomes included describing the average IV calcium dose required to normalize iCa levels, average time to normalization of iCa levels, and assessing the safety of IV calcium replacement. Main outcome measure Change in iCa. Results One hundred and ninety-four patients met study criteria. In the final model initial iCa level, total calcium dose, the interaction between initial iCa level and total calcium dose, age, and pancreatitis remained. The model (R² = 0.625) is expressed by the following equation: Change in iCa level = 0.462 − 0.011 × [Ca dose] − 0.0007 × [Age] − 0.259 × [Initial iCa] + 0.076 × [initial iCa × Ca dose] − 0.076 × [Pancreatitis]. Removing two patients that received > 10 grams of total calcium improved the R² to 0.769. Lastly, a simplified model removing age and pancreatitis found a similar R² of 0.756. Conclusion We observed that change in iCa level after initial calcium dose depended on the baseline iCa. Our full and simplified model excluding two outliers predicted 76.9% and 75.6% of the variation in iCa response, respectively. If validated in other settings this model could be utilized to provide more accurate calcium dosing.