Covid-19 affects taste independently of smell: results from a combined chemosensory home test and online survey from a global cohort (N=10,953)

Abstract

People often confuse smell loss with taste loss, so it is unclear how much gustatory function is reduced in patients self-reporting taste loss. Our pre-registered cross-sectional study design included an online survey in 12 languages with instructions for self-administering chemosensory tests with ten household items. Between June 2020 and March 2021, 10,953 individuals participated. Of these, 3,356 self-reported a positive and 602 a negative COVID-19 diagnosis (COVID+ and COVID-, respectively); 1,267 were awaiting test results (COVID?). The rest reported no respiratory illness and were grouped by symptoms: sudden smell/taste changes (STC, N=4,445), other symptoms excluding smell or taste loss (OthS, N=832), and no symptoms (NoS, N=416). Taste, smell, and oral irritation intensities and self-assessed abilities were rated on visual analog scales. Compared to the NoS group, COVID+ was associated with a 21% reduction in taste (95% Confidence Interval (CI): 15-28%), 47% in smell (95%-CI: 37-56%), and 17% in oral irritation (95%-CI: 10-25%) intensity. In all groups, perceived intensity of smell (r=0.84), taste (r=0.68), and oral irritation (r=0.37) was correlated. Our findings suggest most reports of taste dysfunction with COVID-19 were genuine and not due to misinterpreting smell loss as taste loss (i.e., a classical taste-flavor confusion). Assessing smell and taste intensity of household items is a promising, cost-effective screening tool that complements self-reports and helps to disentangle taste loss from smell loss. However, it does not replace standardized validated psychophysical tests.

Figure 1.

Flow diagram showing the participants included in the study. Exclusions were based on implausible data. Participants were split based on their response (yes/no) to the question asking whether they have a respiratory illness. Only those who responded with “yes” to the respiratory illness question were asked whether they had COVID-19. They were further split according to their report of a positive (COVID+) or negative (COVID−) diagnosis or if they were waiting for a test result and were suspected to have COVID-19 (COVID?). Those who responded “no” to the respiratory illness question were split by the symptoms reported: only smell and/or taste-related changes (STC), any OtherSymptom but smell/taste changes (OthS), or NoSymptoms at all (NoS). Age ± SD in years (y); Gender is reported as women (w) and men (m), remaining participants identified as “other” or did not share.

Figure 2.. Perceived intensity of taste, smell, and oral irritation when sampling food or household items for six groups of participants.

Participants are grouped according to COVID diagnosis or symptoms (from left to right) into COVID positive (COVID+; N=3,275), unknown COVID status (COVID?; N=1,224), and COVID negative (COVID−; N=579), those who reported sudden smell/taste changes (STC; N=4,271), those with other symptoms excluding smell or taste loss (OthS; N=802), and those with no symptoms (NoS; N=396). They rated perceived intensity of smell, taste, and oral irritating stimuli using a visual analog scale (0–100). Points represent individual subject data (jittered horizontally), the center horizontal bars depict the median, the shapes reflect the density of the distribution, and the colored areas show interquartile ranges. For a similar presentation of data for self-reported chemosensory ability, see Supplementary Figure 3.

Figure 3.. Multiple Factor Analysis (MFA) on self-reported chemosensory abilities and perceived intensities.

(A) Correlation circle including all ratings. (B) Map of the six groups (COVID+, N=3,275; COVID−, N=579; COVID?, N=1,224; STC, N=4,271; OthS, N=802; NoS, N=396) with 95% confidence ellipses.

Figure 4.. Differences in self-reported abilities and perceived intensities for smell, taste, and oral irritation between three clusters obtained by the Agglomerative Hierarchical Clustering (AHC) on perceived intensities irrespective of reported diagnosis.

(A) 3D plot on smell, taste, and oral irritation intensities. Dots represent individual subject data, clusters are color-coded. (B) Self-reported abilities and perceived intensities of foods and items of the three chemosensory modalities of smell, taste, and oral irritation for the three clusters. Points represent individual subject data (jittered horizontally), the center horizontal bars depict the medians, the shapes reflect the density of the distribution, and the colored areas show the interquartile range. Cluster one (green) was minimally impaired, while cluster 2 (orange) was severely impaired for all three chemical senses; cluster 3 (blue) showed severe loss of smell but not of taste or oral irritation.