Sensitivity and specifity of prediction models based on gustatory disorders in diagnosing COVID-19 patients: a case-control study.

Adamczyk, K.; Herman, M.; Fraczek, J.; Piec, R.; Szykula-Piec, B.; Zaczynski, A.; Wojtowicz, R.; Bojanowski, K.; Rusyan, E.; Krol, Z.; Wierzba, W.; Franek, E.

Abstract Objective: to quantitatively assess disturbances of sweet, sour and salty and bitter tastes in a group of young, asymptomatic or oligosymptomatic COVID-19 patients; establish a reliable, sensitive and specific test that can diagnose COVID-19 on the basis of taste disorders and publish the results according to STARD 2015 statement (Standards for Reporting Diagnostic Accuracy). Design: case-control study Setting: isolated rooms in the Central Clinical Hospital of the Ministry of the Interior and Administration in Warsaw (which had been transformed into an infectious hospital) and a dormitory in one of Warsaw universities. Participants: 52 SARS-CoV-2 positive (51 men, mean age 21.7 years) and 36 negative students (34 men, mean age 20.8 years). Main outcome measures: a gustatory function assessment (sweet, salty, sour and bitter taste), with flavour concentrations established previously in healthy subjects was conducted for all subjects. Each participant received one tasteless reference and nine flavour tablets with sucrose concentrations of 40, 80 and 106.4 mg/ml; NaCl at 13.5, 17 and 27 mg/ml; ascorbic acid at 6.25 and 12.5 mg/ml and grapefruit extract at 40 mg/ml. Results: the only taste that was impaired significantly more frequently in COVID-19 patients was the sweet taste at the lowest flavour concentration (40 mg/ml, p = 0.002). Different screening and diagnostic models were constructed using the examined variables. The highest accuracy screening test consisted of the positive result of a three-question questionnaire (self-reported loss of taste, self-reported loss of smell, or fever within the last month (positive if at least one present) and/or ageusia of sweet taste at a sucrose concentration of 40 mg/ml. The sensitivity of the model was 94% with a specificity of 55%. The highest accuracy diagnostic test consisted of ageusia of sweet taste at a sucrose concentration of 106.4 mg/ml or/and ageusia of salty taste at an NaCl concentration of 13.5 or 17 mg/ml. The specificity of the test was found to be 100%, and the sensitivity was 34%. Conclusion: as the most effective way of controlling the present pandemic involves testing the wider population for symptomatic, oligosymptomatic and asymptomatic carriers of SARS-CoV-2 and isolating or hospitalising infected subjects, in the present study, an inexpensive, simple, fast and sensitive (94%) screening test that can be used for such a purpose is proposed. In addition, a specific (100%) diagnostic test that could be used to refer patients to quarantine in the case of limited availability of genetic or serological tests is proposed.