Harmonization of SARS-CoV-2 reverse transcription quantitative PCR tests to the first WHO international standard for SARS-CoV-2 RNA

Abstract

Background: Cycle threshold (Ct) values from SARS-CoV-2 reverse transcription quantitative PCR (RT-qPCR) tests are used to measure viral burden. Calibration to the First WHO International Standard for SARS-CoV-2 RNA may improve quantitative inter-assay agreement.

Methods: WHO standard was tested using four emergency use authorized RT-qPCRs to generate calibration curves and evaluate Ct value differences. Harmonization of two assays, Cepheid Xpert Xpress SARS-CoV-2 targeting E and nucleocapsid (N2) [Xpert (E) and Xpert (N2)] and a laboratory-developed test targeting E [LDT (E)], was assessed using 93 positive upper respiratory samples. Platform (target) pairs were compared via Bland-Altman analysis and Passing-Bablok regression.

Results: Ct values with the WHO standard were comparable across platforms and targets, except Xpert (N2) for which the mean difference was a median of 3.68 cycles (Interquartile Range, IQR = 3.23 to 3.76 cycles) greater than other platform (target) pairs. Using clinical samples, the mean difference of Xpert (N2) to LDT (E) was 3.64 cycles (95% Confidence Interval, CI =1.51 to 5.76). After calibration, the mean difference of Xpert (N2) to LDT (E) was 0.08 log₁₀ IU/mL (95% CI = -0.56 to 0.71) and the regression was y = 1.00x * 0.08 (95% CI slope = 0.93 to 1.07, 95% CI intercept = 0.28 to 0.42).

Conclusions: Calibration to the WHO standard resulted in the harmonization of two RT-qPCR tests, whereas analysis by Ct value alone may have led to erroneous quantitation. Harmonization to the WHO standard has the potential to improve the generalizability of clinical associations with SARS-CoV-2 RNA levels.

Fig. 1

Bland-Altman analysis comparing the cycle threshold (Ct) values of SARS-CoV-2-positive clinical upper respiratory specimens measured using the following platform (target) pairs: Xpert (E) to LDT (E) [A], Xpert (N2) to LDT (E) [B] and Xpert (N2) to Xpert (E) [C]. E, envelope; N2, nucleocapsid; Solid line, mean difference; dotted line, zero difference; dashed lines, 95% confidence intervals.

Fig. 2

Bland-Altman analysis [A-C] and Passing-Bablok regression [D-F] comparing SARS-CoV-2 RNA levels [in log₁₀ international units (IU)/mL] of clinical upper respiratory specimens measured using the following WHO-calibrated platform (target) pairs: Xpert (E) to LDT (E) [A, D], Xpert (N2) to LDT (E) [B, E] and Xpert (N2) to Xpert (E) [C, F]. E, envelope; N2, nucleocapsid; Bland-Altman [A-C]: Solid line, mean difference; dotted line, zero difference; dashed lines, 95% confidence intervals; Passing-Bablok [D-F]: Solid line, regression line; dotted line, line of identity; dashed lines, 95% confidence intervals. The mean values of the WHO standard at each concentration level are represented by the closed diamonds.