Albumin adjustment of total calcium does not improve the estimation of calcium status

There is a longstanding controversy as to whether plasma measurements of total calcium should be adjusted for albumin concentration, and if so which formulas are the most appropriate.

Ionised calcium, total calcium and albumin results, analysed at the same time at Uppsala University Hospital Laboratory between February 2005 and June 2013, were retrieved from a laboratory information system. The dataset included results from 20,003 patients. Total calcium was albumin-modified by a locally derived formula, based on 3106 patients from the dataset, and formulas from the literature. The agreement between the reference method ionised calcium and unadjusted total calcium and the seven different albumin-modifying calcium formulas, respectively, were compared with intra-class correlation coefficients (ICC).

Total calcium showed substantial agreement to ionised calcium, ICC 0.85 (95% CI 0.84-0.86) for the whole validation cohort. Albumin-modified calcium by different formulas showed significantly less or equal agreement, however the locally determined formula performed better than formulas taken from the literature. Also, total calcium classified the patient as hypo-normo- or hypercalcemic right in 82% of the patients. The albumin-modified calcium did not classify patients significantly better except in the subgroup hypoalbuminemia (<30 g/L) where the local formula classified the patients slightly better than total calcium.

Albumin modification of total calcium determinations is unlikely to add valuable information, and this practice should be abandoned. Ionised calcium should be used more frequently when aberrant results for total calcium are followed up, or in patients with known hypoalbuminemia.