Early diagnosis of SARS-CoV-2 is essential to limiting the spread of the virus and managing infected patients during hospitalization. The sensitivity of RT-qPCR is contested by the fact that it is time-consuming, executed by trained technicians in proper environment for material extraction. Here, we evaluated the first SARS-CoV-2 antigen test recommended by the World Health Organization at September, 2020 as an alternative for immediate diagnosis of symptomatic and suspected patients at a hospital in Brazil during the epidemic peak. All patients were submitted to RT-qPCR and rapid antigen test using nasopharyngeal swabs rigorously collected at the same time. Demographics, baseline comorbidities, symptoms and outcomes were considered. Prediction analysis revealed that previous stroke, chronic obstructive pulmonary disease, desaturation and tachypnea were the most relevant determinants of the death of COVID-19 patients. Comparison between the rapid antigen test and RT-qPCR revealed an overall PPV of 97%, extended to 100% when performed between 4 and 15 days of symptoms, with an accuracy of 90-91% from days 1 to 7 and a ‘Substantial’ agreement. The rapid antigen test presented no inconclusive result. Among the discordant results and RT-qPCR inconclusives, 72% presented bilateral multifocal ground-glass opacities on imaging and other exams alterations. The median time to obtain RT-qPCR results was 83.6 hours, against 15 minutes for the rapid test, precious time for deciding on patient isolation and management. Knowledge of the risk factors and a rapid diagnosis upon patient admission is critical to reduce mortality of COVID-19 patients, hospital internal costs and in-hospital transmission.

No abstract