Widespread smell testing for COVID-19 has limited application

Having campaigned to achieve recognition that anosmia (loss of smell) is a highly prevalent symptom of COVID-19,1, 2 we were delighted that Public Health England changed the case definition on May 18, 2020.³ We agree with Cristina Menni and colleagues,⁴ that the added sensitivity attributed to adding anosmia to the case definition (less than 2%) is very likely to be a gross underestimate. Indeed, even the additional 15·9% of cases who are identified when including anosmia⁵ might still fail to capture the full benefit because access to testing in the UK has been so restricted for patients with mild disease. Data from elsewhere suggest that anosmia will have most value as a marker in mild cases that, until recently, were excluded from testing.

However, we urge caution about a call to introduce smell tests as a screening tool in some settings, such as airports and shopping centres, with the intention of denying access to those identified as having lost their sense of smell.⁴ Although new-onset and sudden-onset anosmia has a high likelihood of predicting a positive test for COVID-19 when the prevalence of disease is high, population estimates suggest that 19·1% of adults suffer from pre-existing diminished sense of smell, a figure that rises to 80% in patients older than 75·5 years. These data closely reflect the 21·7% of patients who tested negative for COVID-19 in the COVID Symptom Study who reported a loss of sense of smell.⁵ Furthermore, in patients who have developed anosmia as a result of COVID-19, chemosensory loss persists for 8 weeks in approximately 10% of cases (unpublished), but this does not reflect how infectious these individuals are to others and when they have viral clearance. The self-reported median recovery rate of 5 days, as reported by Menni and colleagues,⁴ will not be matched by the results of psychophysical smell tests. To deny access to airports or retail parks to approximately one fifth of the population on this basis risks introducing a form of discrimination and would be an intervention that goes beyond the public health benefits of reducing transmission.

We strongly advise all people who experience new-onset loss of sense of smell to self-isolate and seek confirmatory testing. However, we must not impose punitive measures on those patients who have lived without a sense of smell for many years. We encourage extreme caution in how this new finding is incorporated into policy and would suggest that clinicians and researchers working in this field be called upon to ensure that such policies are rigorously and appropriately defined.