Diagnostic tests in COVID-19: critical appraisal of the evidence

Introduction: during the COVID-19 pandemic, many studies have been published on diagnostic tests, aimed at determining their accuracy, balance between benefits, risks, costs, fairness, and feasibility. A summary of the recommendations on the different tests available and on the different analytical or radiological characteristics that can support the diagnosis or have prognostic value is presented.

Material and methods: PICO questions were formulated and a systematic search for information was carried out. The information obtained was evaluated, synthesized, and classified following the GRADE methodology. Finally, evidence tables with recommendations were prepared.

Results and discussion: reverse transcription polymerase chain reaction is considered the gold standard, all trademarks are effective. Salivary samples have similar validity to nasopharyngeal samples and may be an option in low-prevalence outpatient settings. Antigenic tests have lower sensitivity and specificity, although they could be considered in patients with symptoms lasting less than 5 days. Serological tests are highly specific, offering better results 14 days after onset and could be useful in cases of multisystem inflammatory syndrome. In moderate and severe cases, analysis with LDH, C-reactive protein, ferritin and procalcitonin is recommended, due to the risk of coinfection. In these patients, bedside point of care-ultrasound could be an alternative in the case of trained clinicians. The quality of the evidence is, in general, low and most of the recommendations present a very low degree of evidence.