Evaluation of 36 formulas for calculating plasma osmolality

Abstract

Purpose: Measuring or calculating plasma osmolality is of interest in critical care medicine. Moreover, the osmolal gap (i.e. the difference between the measured and calculated osmolality) helps in the differentiation of metabolic acidosis. A variety of formulas for calculating osmolality have been published, most of them relying on sodium, urea and glucose. A novel formula developed by Zander has recently been published, which also takes into account the effects of potassium, chloride, lactate and bicarbonate on osmolality. We evaluate the previously published formulas including the novel formula by comparing calculated and measured osmolality.

Methods: Arterial or venous blood samples from 41 outpatients and 195 acutely ill inpatients (total 236 subjects) were used to compare measured osmolality with calculated osmolality as obtained from 36 published formulas including the new formula. The performance of the formulas was statistically evaluated using the method of Bland and Altman.

Results: Mean differences up to 35 mosmol/kg H(2)O were observed between measured and calculated osmolality using the previously published formulas. In contrast, the novel formula had a negligible mean difference of 0.5 mosmol/kg H(2)O. The novel formula also had the closest 95 % limits of agreement ranging from -6.5 to 7.5 mosmol/kg H(2)O.

Conclusion: Only 4 out of the 36 evaluated formulas gave mean differences between measured and calculated osmolality of less than 1 mosmol/kg H(2)O. Zander's novel formula showed excellent concordance with measured osmolality and facilitates a more precise diagnosis based on blood gas analysers. The new equation has the potential to replace separate measurements of osmolality in many cases.