COVID‐19‐Associated Chemosensory Loss Continues to Decline

Abstract

Chemosensory losses have long been considered a cardinal symptom of COVID‐19 infection. Recent studies have shown changing symptom profiles with COVID‐19, including decreasing incidence of olfactory losses. We accessed the National COVID Cohort Collaborative database to identify patients with and without smell and taste loss within 2 weeks of COVID‐19 diagnosis. Peak prevalence time intervals for variants were determined from Covariants.org. Using rates of chemosensory loss during the peak time interval for “Untyped” variants as baseline (4/27/2020‐6/18/2020), odds ratios for COVID‐19‐associated smell or taste disturbance fell for each of the Alpha (0.744), Delta (0.637), Omicron K (0.139), Omicron L (0.079), Omicron C (0.061), and Omicron B (0.070) peak intervals. These data suggest that during the recent Omicron waves and potentially moving forward, the presence or absence of smell and taste disturbances may no longer have predictive value in the diagnosis of COVID‐19 infection.

Long-Term Taste and Smell Outcomes After COVID-19

“…The final participant population consisted of 340 persons with a previous history of COVID-19 (128 males [37.6%] and 212 females [62.4%]; mean (SD) age, 39.04 [14.35] years; 18 African American [5.3%], 36 Asian [10.6%], 18 Hispanic [5.3%], and 198 White [58.2%]) and 434 persons with no such history (154 males [35.5%] and 280 females [64.5%]; mean (SD) age, 39.99 [15.61] years; 53 African American [12.9%], 47 Asian [10.8%], 30 Hispanic [6.9%], and 274 White [63.1%]). Demographic information, including age and sex, is presented in Table 1 …”

“… Abbreviations: NaCl, sodium chloride; MSG, monosodium glutamate; WETT, Waterless Empirical Taste Test. a P values are based on analyses of covariance with age as the covariate and sex and SARS-CoV-2 variant common at the time of initial infection 31 as between-participant factors. …”

“… a P values are based on analyses of covariance with age as the covariate and sex and SARS-CoV-2 variant common at the time of initial infection 31 as between-participant factors. …”

“…While PCR testing is optimal for identifying exact variants, our study faced limitations in accessing PCR data, which precluded precise variant inference. As a result, the method of inferring variants based on infection timing was chosen as an alternative approach, which has been used by other investigators as well …”

“…In some analyses, the virus variant prevalent at the time of testing also served as a factor and the time between initial infection and chemosensory testing served along with age as a covariate. The virus variant most common at the time of the initial infection was determined from Reiter et al 31 Smoking behavior was initially evaluated but excluded from subsequent analyses given that it was not significant and did not interact with any other variable. Post hoc comparisons were performed using the Tukey honestly significant difference test.…”