ABSTRACT
<p><b>Background</b>Measuring total serum calcium is important for the diagnosis of diseases. Currently, results from commercial kits for calcium measurement are variable. Generally, the performance of serum calcium measurements is monitored by external quality assessment (EQA) or proficiency testing schemes. However, the commutability of the EQA samples and calibrators is often unknown, which limits the effectiveness of EQA schemes. The aim of this study was to evaluate the bias of serum calcium measurements and the commutability of processed materials.</p><p><b>Methods</b>Inductively coupled plasma mass spectrometry was applied as a comparative method, and 14 routine methods were chosen as test methods. Forty-eight serum samples from individual patients and 25 processed materials were quantified. A scatter plot was generated from patient samples, and 95% prediction intervals were calculated to evaluate the commutability of the processed materials and measurement bias at three concentration levels was used to determine the accuracy of routine assays.</p><p><b>Results</b>All assays showed high precision (total coefficient of variation [CV] <2.26%) and correlation coefficients (r > 0.99). For all assays, the mean bias for the 48 patient samples ranged from -0.13 mmol/L to 0.00 mmol/L (-5.61-0.01%), and the ranges for the three concentrations were -0.10-0.04 mmol/L (-5.71-2.35%), -0.14--0.01 mmol/L (-5.80--0.30%), and -0.19-0.04 mmol/L (-6.24-1.22%). The EQA samples, calibrators, and animal sera exhibited matrix effects in some assays; human serum pools were commutable in all assays; certificate reference materials were commutable in most assays, and only GBW09152 exhibited a matrix effect in one assay; and aqueous reference materials exhibited matrix effects in most assays.</p><p><b>Conclusions</b>Biases for most assays were within the acceptable range, although the accuracy of some assays needs improvement. Human serum pools prepared from patient samples were commutable, and the other tested materials exhibited a matrix effect.</p>