The importance of external quality assessment data in evaluating SARS-CoV-2 virus genome detection assays.

We read with interest the comprehensive review by Wing Ying Au and Peter Pak Hang Cheung on the diagnostic performances of common nucleic acid tests for SARS-CoV-2. Although there are numerous publications in this regard, no reference was made to performance data that were collected in SARS-CoV-2-related external quality assessment (EQA) schemes. Performance studies as cited in the review provide essential data on sensitivity and specificity in evaluation settings, but such studies have so far not been published for all routinely used assays. However, EQA schemes provide data on the diagnostic and analytic performance not only of selected individual assays, but of the entire range of assays used by laboratories in their routine work and thus real-life data. It should be noted that in the latest round of the national Austrian EQA scheme for SARS-CoV-2 virus genome detection, a total of no less than 53 different assays were registered. By using the same panel of samples, these schemes show differences in the performances of individual systems, but also the intra-type variation of results—ie, the variation in results obtained from different devices of the same type. In addition, EQA is an ongoing process that monitors the performance of assays during the whole time they are approved and supplied. We therefore consider EQA data to be no less important for assessing the analytical performance of assays than performance study and evaluation data. Particularly noteworthy are the following results, which could only be found through data on SARS-CoV-2-related EQA schemes. Firstly, monitoring of dozens of different assays that are used in one or more of the hundreds of participating laboratories in one EQA round is feasible, even across national borders.2, 3 Secondly, samples with a virus load of 100 000 copies per mL or more (ie, a cycle threshold [Ct] value of less than 30) are quite reliably detected as positive, but the detection probability decreases with decreasing virus load. Thirdly, the false positive rate is very low. Fourthly, assay-specific Ct values prevent comparability of results obtained by different test systems, whereas the intra-assay and the intra-type variation of Ct values is low.4, 5 Lastly, the accuracy of the test systems, as challenged by samples from a dilution series and by two identical samples, is good for most assays, but not all.2, 4

In conclusion, we repeat that the performance of assays can only be fully assessed if both data from evaluation studies and EQA data are included.

We declare no competing interests. AG is President of ÖQUASTA, the Austrian Association for Quality Assurance and Standardization of Medical and Diagnostic Tests. CB is Chairman of the Executive Board of EQALM, the European Organisation for External Quality Assurance Providers in Laboratory Medicine.