Interpretation of findings

This study suggests that there is a higher prevalence of self-reported chemosensory impairment in COVID-19-positive patients compared with negative patients with early recovery of smell function in a large proportion of patients. In our series, olfactory and gustatory impairments were respectively 14 times and 11 times more likely to be associated with COVID-19 positivity. Overall, our findings suggest that loss of smell and loss of taste have a higher chance of identifying COVID-19-positive patients among those with influenza-like symptoms seen on an ambulatory basis.

The reported prevalence of olfactory dysfunction varies significantly from 5.1% in a Chinese study to 85.6% in a European study.^{Reference Spinato, Fabbris and Polesel3–Reference Moein, Hashemian and Mansourafshar5,Reference Vroegop, Eeckles and Van Rompaey7–Reference Vaira, Salzano, Deiana and De Riu9} Our results are consistent with previous reports that showed a higher prevalence of self-reported smell or taste loss among COVID-19-positive patients. Interestingly, in our study, although 31 out of 56 (55.4%) of COVID-19-positive survey participants included anosmia or hyposmia as one of their overall symptoms, only 23 (41.1%) patients maintained that they lost their sense of smell around the time of COVID-19 symptom onset (5 days prior or thereafter) with a 16% rate of “unable to recall.” This discrepancy likely represents recall bias, which is common in cross-sectional studies. Therefore, we believe that the 40% rate of smell loss around the time of COVID-19 symptom onset is a more accurate representation than the 55% overall rate.

Comparison with previous studies

Our rate of anosmia/hyposmia of 41.1% is lower than the rate of 60%–85% in recently published studies with comparable study designs or 51.5% found in another Canadian study.^{Reference Spinato, Fabbris and Polesel3,Reference Yan, Faraji, Prajapati, Boone and DeConde4,Reference Carignan, Valiquette and Grenier6,Reference Lechien, Chiesa-Estomba and De Siati8} There may be several explanations for this difference. One consideration is the cultural difference in olfactory perception.^{Reference Candau10–Reference Ferdenzi, Roberts and Schirmer12} In fact, participants in other cultures may have different thresholds for their smell loss. Another possible explanation is the presence of mutant strains with varying pathogenicity, as evidenced from genomic studies in multiple nations.^{Reference Gudbjartsson, Helgason and Jonsson13–Reference Yao, Lu and Chen16} These different strains may cause varying degrees of chemosensory impairment with regional discrepancy, although this has not been scientifically proven. Therefore, it is important to collect and examine regional data to depict a representative landscape for our population. Lastly, we do note that this difference could be a simple function of recall bias from cross-sectional surveys, as other studies do not report an “unable to recall” option, thus forcing patients to choose with an altered response rate of self-reported smell loss. Regardless, all published studies unanimously support the inclusion of smell and taste loss as important markers of COVID-19.

Clinical and research implications

A possible mechanism of olfactory dysfunction caused by SARS-CoV-2 has been demonstrated in experimental models with inoculation of coronaviruses. Coronaviruses can damage the olfactory neuroepithelium via apoptosis with a subsequent reduction of mature sensory neurons and disordering of olfactory epithelium.^{Reference Mori, Goshima and Imai17–Reference Schwob, Saha, Youngentob and Jubelt19} Another line of research suggests that the anosmia/hyposmia may result from infection of the support and perivascular cells in olfactory epithelium, via angiotensin-converting enzyme 2 receptors, which are identified as the receptors of SARS-CoV-2.^{Reference Brann, Tsukahara, Weinreb, Logan and Datta20,Reference Zhou, Yang and Wang21} Clinically, nasal inoculation of coronavirus resulted in olfactory impairment in healthy volunteers.^{Reference Åkerlund, Bende and Murphy22} Dysgeusia/ageusia is generally regarded to be secondary to loss of smell, as it may be difficult to distinguish between flavor and taste. In a large study investigating taste disorders unrelated to COVID-19, many reported loss of taste without any objective gustatory loss but found to have olfactory deficits alone.^{Reference Deems, Doty and Settle23} While it is possible that loss of taste in COVID-19 may be a distinct mechanism,^{Reference Vaira, Salzano, Fois, Piombino and De Riu24} the rate of taste loss in our patients likely represents subtle olfactory deficits. In our series, 52% of the affected individuals reported a normal sense of smell at the time of survey with many recovering within 4 weeks. This finding is consistent with previous reports and suggests that the damage to the olfactory function by COVID-19 may not be permanent.^{Reference Yan, Faraji, Prajapati, Boone and DeConde4} However, this warrants further investigation with follow-up assessments of smell function. All in all, the presence of loss of smell and taste may be a distinguishable feature of COVID-19 from other viral URTIs.

Our results have strong implications in public health measures. A recent study demonstrated 56% of residents to be COVID-19 positive but “asymptomatic” in a skilled nursing facility.^{Reference Arons, Hatfield and Reddy25} However, this study did not include the loss of smell or taste in the symptom assessment. In a previous study from Italy and our study, both demonstrate that a significant portion of the patients complained of decreased sense of smell or taste during the early phase of their disease.^{Reference Spinato, Fabbris and Polesel3} Hence, it would be essential to incorporate anosmia/hyposmia and dysgeusia/ageusia as part of future epidemiologic COVID-19 surveillance studies and tailor the screening criteria for COVID-19 accordingly. Furthermore, these symptoms can be easily assessed in the context of telemedicine and virtual care. This can, in turn, facilitate improved patient counselling for the screening of COVID-19 and help curb the spread of COVID-19 in the vulnerable population, including those in long-term care facilities and remote communities where testing may be limited. Currently, while several provincial bodies have implemented loss of smell/taste as screening symptoms, the self-assessment tool from the Canadian Federal Government still does not incorporate these symptoms. With our Ontario regional data demonstrating a high probability of COVID-19 positivity with loss of smell/taste, there is now evidence to expand public awareness of this association and ultimately improve our public health response to this pandemic.

Limitations

Interpretation of our study results is limited by recall bias, as there is a significant time lapse between diagnosis and survey administration. The recent media spotlight of COVID-19 and anosmia might have resulted in elevated rates of self-reported anosmia. We tried to mitigate this issue by asking similar, but differently worded, questions for internal validity. Future studies could be administered at the time of swab to avoid this bias. We also acknowledge non-response bias. While the study results are limited by the nature of a single-institutional-based study, we collected data over a longer period of time than previous reports to improve our sampling.^{Reference Spinato, Fabbris and Polesel3,Reference Yan, Faraji, Prajapati, Boone and DeConde4} Furthermore, the testing criteria focused on frontline workers from at-risk settings, such as healthcare or long-term care, residents in group homes, and returning travelers. This may reduce external validity. Lastly, with ethical concerns of transmission to our research team and preservation of personal protective equipment, we are presently unable to objectively measure smell in symptomatic patients during the point of assessment.

We acknowledge the need for large-scale epidemiologic studies to further investigate this association between smell and taste loss and COVID-19 positivity. Our research group currently has ongoing projects to investigate a larger regional population and to incorporate self-reported objective measures of smell using common household items. Despite the limitations, the baseline characteristics between positive and negative groups were comparable, and our data emphasize the importance of considering smell and taste as screening symptoms.