Abstract
INTRODUCTION: We evaluated the Roche Elecsys IL6 assay on the Cobas immunoassay analyser. METHOD: Serum IL6 of 144 controls were compared to 52 samples from patients with COVID-like respiratory symptoms (17 SARS-CoV-2 RT-PCR positive); 25 of these were from the intensive care unit (ICU). We compared the IL6 levels to C-reactive protein (CRP) and procalcitonin (PCT) levels in all cases. RESULTS: The IL6 assay had coefficient-of-variation (CV) of 2.3 % (34.1 pg/mL) and 2.5 % (222.5 pg/mL), a limit of quantitation <1.6 pg/mL, and was linear from 1.6 to 4948 pg/mL. There was a significant difference in IL6 values between patients with COVID-like respiratory symptoms versus controls (p < 0.001). ROC analysis showed that IL6 > 6.4 pg/mL identified symptomatic cases (AUC 0.94, sensitivity 88.2 %, specificity 97.2 %). There was a significant difference between the IL6 of symptomatic ICU/non-ICU cases (median IL6 228 vs 11 pg/mL, p < 0.0001); ROC analysis showed IL6 > 75 pg/mL (sensitivity 76.0 %, specificity 88.9 %) was superior to CRP and PCT in predicting ICU admission (AUC: IL6 0.83, CRP 0.71, PCT 0.82). CONCLUSION: The performance of Elecsys IL6 assay is in keeping with the manufacturer's claims. IL6 > 6.4 pg/mL differentiates healthy from suspected COVID-19 cases and appears to be raised earlier than the other inflammatory markers in some cases. IL6 > 75 pg/mL was a good predictor of ICU admission.