New Predictive Equations for Serum Ionized Calcium in Hospitalized Patients

OBJECTIVE: To study a new and easy way to calculate equations to predict ionized calcium (Ca(2s+)) for adult hospitalized patients with the usual laboratory and clinical parameters. SUBJECTS AND METHODS: This retrospective observational study was conducted in a third-level university hospital. An initial learning cohort (cohort L: 269 patients) was selected to derive the new equations. These equations were tested in a validation of another cohort (cohort V: 146 patients). Patients selected were hospitalized adults who had simultaneous determinations of Ca(2s+) and serum total calcium (CaTot). They were classified using their estimated glomerular filtration rate (GFRe) into normal function, moderate and severe kidney dysfunction. Demographic and biochemical parameters, in addition to comorbidities, were collected from hospital databases. Nine published equations to predict Ca(2s+) and 2 widely used equations to predict corrected CaTot were also selected to be compared to newer equations for accuracy in detecting serum calcium alterations. New equations were derived by a multiple linear-regression analysis from patients in cohort L. RESULTS: Three equations were derived containing the CaTot square root as the main independent variable. Equation 1: Ca(2s+) = 0.815 × CaTot(0.5). Equation 2: Ca(2s+) = 0.826 × CaTot(0.5) − 0.023 × renal function. Equation 3: Ca(2s+) = 0.813 × CaTot(0.5) − 0.006 × albumin(0.75) s+ 0.079. These equations performed better than published equations to predict Ca(2s+) when their error measures were analyzed in cohort V, even in special populations such as critically ill and very old patients. CONCLUSIONS: Three new equations predicting Ca(2s+) were derived requiring easily available clinical and laboratory parameters. They could be valuable in predicting hypocalcemia but are of limited use in hypercalcemia.