Evaluation of analytical performance of homocysteine LC-MS/MS assay and design of internal quality control strategy

Objectives: Liquid chromatography-tandem mass spectrometry (LC-MS/MS) has become a common technique in clinical laboratories in recent years. Because most methods are laboratory-developed tests (LDTs), their reproducibility and quality control (QC) have been controversial. In this study, Westgard Sigma Rules were used to evaluate the analytical performance and establish an individualised internal QC (IQC) strategy for these LDTs.

Methods: Taking the LC-MS/MS LDT method for homocysteine (Hcy) as an example, the 'desirable specifications' from the Biological Variation Database were used as quality goals. Based on the external quality assessment (EQA) samples, bias was calculated and the coefficient of variation (CV) was also calculated by IQC measurements for six consecutive months. The analytical performance was evaluated by calculated sigma metrics and an IQC strategy was designed using the Westgard Sigma Rules with run size.

Results: Over 116 days within 6 months, a total of 850 data points were collected for each of IQC 1 and IQC 2. The monthly coefficient of variation CV% was 2.57-4.01%, which was non-significant (p-value: 0.75). The absolute bias% for IQC1 and IQC2 was 1.23 and 1.87%, respectively. The allowable total error (TEa) was selected as 15.5%, Sigma metrics were 4.02 and 4.30, and the analytical performance was 'Good'. The 1_3s/2_2s/R_4s/4_1s multi rules (n=4, r=1) with a run size of 200 samples were suggested for the Hcy IQC scheme. The quality goal index (QGI) values were over 1.2, indicating that trueness needed to be improved.

Conclusions: The analytical performance of the Hcy LC-MS/MS LDT conformed to the Six Sigma rating level, achieving 'good' (four Sigma). Clinical practice indicated that calibration bias was the primary factor affecting trueness.