Validation of calculated ionised calcium for intensive care patients

A previously published formula for calculating ionised calcium (iCa) is currently used by our institution. This study assessed the validity of this formula for critically ill patients. The plasma iCa in 365 samples from 108 critically ill adults were calculated using a published formula, and compared with iCa measured using an ion selective electrode. The calculated and measured iCa ranges were 0.86-1.77 mmol/L and 0.87-1.72 mmol/L respectively. Bias between the methods was 0.004 mmol/L (Bland Altman). Linear regression demonstrated a high correlation [r2 = 0.89; calculated iCa = (0.902.measured iCa) + 0.1299]. For patients with albumin <20 g/L bias was 0.005 mmol/L [r2 = 0.95; calculated iCa = (1.0792.measured iCa) + 0.091]. Differences at medical decision points were substantially greater than the desirable bias quoted in quality specifications.This study found that a significant bias exists between calculated and measured iCa in critically ill patients at the medical decision limits. This highlights the importance of measuring rather than calculating iCa in this group.