Loss of smell in COVID-19 patients: a critical review with emphasis on the use of olfactory tests

SUMMARY

Since December 2019, an outbreak of a newly isolated coronavirus (SARS-CoV-2) appeared in Wuhan, China, and then spread worldwide. Recently, it has emerged that a number of patients may present with sudden hyposmia, sometimes without other symptoms of the disease. We performed a critical review on the methods used to date to investigate the olfactory function in COVID-19 patients in order to establish which should be considered the most appropriate to use during this pandemic. Literature analysis showed that the diagnosis of hyposmia in COVID-19 patients was mainly made through subjective symptomatology collected by questionnaires and/or interview. Psychophysical tests were carried out in a few studies showing significant discrepancies between the self-reported sense of smell and test results. To date the methods used by authors to investigate smell impairment in COVID-19 patients have been very heterogeneous and predominantly based on self-reported questionnaires leading to confusing and inconclusive results. We suggest that simple validated self-administered psychophysical olfactory tests could be a valuable instrument to investigate isolated/quarantined or hospitalised COVID-19 patients referring smell impairment in order to confirm olfactory dysfunction.

Introduction

Since December 2019, a novel coronavirus SARS-CoV-2 (COVID-19) outbreak emerged in Wuhan, China, and subsequently rapidly spread to several countries ¹.

The clinical manifestations of COVID-19 range from asymptomatic infection to severe pneumonia with acute respiratory distress syndrome. Guan et al. ², in a comprehensive review of patients affected by SARS-CoV-2, described the clinical presentation of 7,736 patients who were hospitalised in China. Fever, the most frequent symptom, was present in 43.8% to 88.7% of cases. Other reported symptoms included cough (67.8%), fatigue (38.1%), sputum production (33.7%), shortness of breath (18.7%), myalgia or arthralgia (14.9%), sore throat (13.9%), headache (13.6%), chills (11.5%), nausea or vomiting (5%), diarrhoea (3.8%), nasal congestion (4.8%), haemoptysis (0.9%) and conjunctival congestion (0.8%). Alterations in taste or smell were not reported among the clinical symptoms. Mao et al. ³, studying 214 consecutive patients with laboratory-confirmed diagnosis of SARS-COV-2, reported in 36.4% both central and peripheral nervous system manifestations. Among the latter, hypogeusia and hyposmia were reported in 5.6 and 5.1% of cases, respectively. After this first report, numerous other cases of hyposmia in patients suffering from COVID-19 have been reported in the media and scientific journals, especially otolaryngology journals, so that a less serious variant of COVID-19 responsible for this symptomatology has been hypothesised ⁴. Given these reports, the British Rhinological Society (BRS), as well as the American Academy of Otolaryngology-Head and Neck Surgery (AAO-HNS), recommended that the smell loss should be considered a marker of COVID-19 infection ^5,6.

Herein, we critically reviewed reports about the methods used to date to quantify olfactory loss in COVID-19 patients, highlighting the limitations of subjective olfactory assessment and the importance of using validated olfactory tests during this pandemic.

Methodology of search strategy

We performed a full systematic review of the literature including English-language articles that were screened from several databases (PubMed, Medline, Web of Science and Google Scholar) and published from January 2020. Literature searches were performed in the beginning of May 2020. We searched articles using MeSH-terms, and/or text words such as “COVID-19”, “SARS-CoV-2”, “smell”, “smell loss”, “anosmia”, “hyposmia” and “olfactory dysfunction”. Titles were screened for relevance, followed by review of the abstract and full text. We only included peer-reviewed papers. This resulted in 24 papers that were discussed in this review. A total of 45 references were used in the full document. We did not include in this critical review articles reporting on 10 or less confirmed SARS-CoV-2 positive patients, unless a validated olfactory test was performed. Generally, we excluded articles with the lack of confirmed COVID-19 status. We divided the identified studies in two major groups, based on the methodology used to detect impairment of smell, and specifically studies using simple surveys or data extracted by medical records and studies using validated olfactory tests.

Sense of smell evaluation in confirmed COVID-19 patients by survey, questionnaires, or medical records

The majority of published studies concerning olfactory dysfunction in patients with COVID-19 infection have used survey questionnaires.

Hopkins et al. ⁵, of 2,428 subjects (most of them unconfirmed for COVID-19) who complained of smell loss, identified 74.4% reporting anosmia and 17.3% very severe smell loss. Interestingly, of the 80 patients who underwent nasopharyngeal swabs for COVID-19, only 74% tested positive for the virus.

Via telephone interview, Lee and co-workers ⁷ evaluated 3,191 laboratory confirmed COVID-19 patients complaining of acute anosmia or ageusia. Smell/taste loss was reported by 15.3% of patients (n = 488). Among these, only olfactory loss was reported by 27.7% of patients, ageusia by 20.3% and both by 52%. Similarly, Heidari et al., Clemency et al. and Gudbjartsson et al. evaluated their COVID-19 patients who complained of olfactory dysfunction through a telephone interview/verbal questionnaire ^8-10. A multicentre questionnaire study ¹¹, based on 417 mild-moderate COVID-19 patients, reported olfactory dysfunction in 85.6% of cases (n = 357). Among these, 284 (79.6%) referred anosmia, while 73 (20.4%) hyposmia. Furthermore, phantosmia and parosmia affected 12.6% and 32.4% of patients, respectively. Olfactory dysfunction appeared before (11.8%), after (65.4%) or simultaneously (22.8%) to the appearance of general or ENT symptoms. According to the authors, 25.5% of patients recovered both their sense of smell and taste during the 2 weeks following resolution of general symptoms. Among the cured patients who had residual gustatory and/or olfactory dysfunction, 53.9% complained of isolated olfactory dysfunction, while 22.5% of isolated gustatory dysfunction and 23.6% of both. Females seemed to be more affected by smell and taste dysfunctions than males ¹¹. Leichien et al. ¹² evaluated by questionnaire 1,420 mild-moderate COVID-19 positive patients. Anosmia was one of most prevalent symptoms and was reported in about 70.2% of patients. Smell loss lasted at least 7 days after the disease in 37.5% of healed patients. Benezit et al. ¹³, evaluating 259 COVID-19 positive patients through a web-based questionnaire, found that 20% of the respondents complained of hyposmia, 24% of hypogeusia and 17% of both. Other authors have reported anosmia/hyposmia through surveys in COVID-19 positive patients. Among these, Kaye et al. ¹⁴ reported that 237 COVID-19 patients, using the AAO-HNS COVID-19 Anosmia Reporting Tool for clinicians ¹⁵, complained of anosmia. In particular, 73% reported anosmia before a diagnosis of COVID-19, while 27% after. Adding smell loss to a symptom tracker phone app, Menni et al. ¹⁶ obtained surveys from 7178 subjects who reported having been tested positive for COVID-19. Of these, 65.03% complained of smell/taste loss.

Some papers reported the results of case-control studies. Beltran-Corbellini et al. ¹⁷, studying 79 SARS-CoV2 positive patients and 40 controls (patients positive for influenza) by surveys, found that smell and taste complaints were significantly more frequent in cases than in controls. Furthermore, considering only the study group, they found that the patients complaining of smell/taste alterations were significantly younger than patients without these complaints. Similarly, studying 59 COVID patients and 203 controls (COVID-19 negative patients with influenza-like symptoms) Yan et al. ¹⁸ found that the referred smell/taste impairments was independently and strongly associated with COVID-19-positivity. Finally Wee et al. ¹⁹, studying 154 patients positive to COVID-19 and 71 patients tested positive for other respiratory viruses, found olfactory/taste dysfunction being self-reported by 22.7% of the patients of the former group and by only 2.8% of the patients of the latter. The authors concluded that self-reported olfactory dysfunction had high specificity as a screening criterion for COVID-19.

Studying retrospectively smell and taste data from 128 COVID-19 patients, Yan et al. ²⁰ observed that only 26 (20.1%) required hospitalisation. Very interestingly, referred anosmia was found to be an independent factor for outpatient care. The authors concluded that smell loss in COVID-19 might be associated with a milder clinical course, as already hypothesised by Gane and coworkers ⁴ and Lee and co-workers ⁷. This could be one of the reasons why hospitalised patients usually complain less of smell/taste dysfunction. Aggarwal et al. ²¹, evaluating 43 COVID-19 positive hospitalised patients (16 of whom hospitalised for the infection), observed that on surveys smell/taste loss was reported in a very low percentage of patients (19%). The mean population age was 65.5 and 75% were males. Giacomelli et al. ²² found that 33.9% of COVID-19 hospitalised patients reported on surveys either taste or olfactory alterations, while 18.6% reported both. Similar results, obtained mainly on hospitalised patients, were reported by Klopfenstein and coworkers who observed in surveys that 47% of their 114 COVID-19 patients reported olfactory alteration (anosmia) and about 85% (46 patients) also suffered of dysgeusia ²³.

Some studies evaluated the reported sense of smell symptom together with patients reported outcome measures (PROMS), such as VAS (visual analogue scale) or SNOT 22 (sinonasal outcome test-22). Yan et al. ¹⁸ studied 59 COVID-19 patients and asked them to fill in a VAS for smell and taste. After showing that 40 patients (68%) reported smell and 42 (71%) gustatory impairment, the authors described a referred improvement of both senses with the clinical resolution of the infection in the majority of the patients. SNOT-22 has also been used in order to study the rhinological symptoms of 202 COVID-19 patients. In this study, 64% of the patients interviewed referred altered sense of smell/taste. At SNOT-22, among the patients complaining of olfactory dysfunction, 34% also complained nasal obstruction. Smell loss, as isolated symptoms, was reported by only 6 patients ²⁴.

Sense of smell evaluation in confirmed COVID-19 patients by olfactory test

Few studies in the literature have adopted validated olfactory tests to confirm the olfactory loss reported by COVID-19 patients and to assess its severity.

Eliezer et al. ²⁵ described the case of a 40-year-old woman, COVID-19 positive with acute hyposmia. A five-odour identification test confirmed the olfactory alteration, while CT scans and MRI of the nasal cavities showed bilateral inflammatory obstruction of the olfactory fissures without abnormalities of the bulbs and olfactory traits. Ottaviano et al. ²⁶ studied 6 COVID-19 patients complaining of sudden hyposmia by the “Le Nez du Vin” test, a six-odours smell identification test and in all cases found an olfactory deficit. In a study by Lechien et al. ²⁷, of 78 subjects complaining of isolated sudden hyposmia, 46 COVID-19 positive patients underwent the “Sniffin Sticks” identification sub-test (16 odours). The authors found that 52% of the patients were anosmic, 11 (24%) were hyposmic. It should be noted that 11 patients had normal olfaction (24%).

In a case-control study by Moein et al. ²⁸, 60 patients with COVID-19 (not necessarily complaining of olfactory dysfunction) carried out an olfactory study using the Persian version of the University of Pennsylvania Smell Identification Test (UPSIT). Only 35% of COVID-19 patients were aware of their olfactory deficiency, while the olfactory test showed that almost all patients (98%) had a measured olfactory dysfunction. According to UPSIT standards, 58% of patients were either severely hyposmic or anosmic, 27% had a moderate hyposmia, 13% a mild hyposmia and only one patient (2%) had normal sense of smell.

Vaira et al. ²⁹ evaluated 72 COVID-19 patients using the Connecticut Chemosensory Clinical Research Center (CCCRC) test, finding that the majority of the patients (about 83%) were affected by smell dysfunction (either hyposmia or anosmia), although only 61% of patients reported having or having had olfactory loss. Of importance, 28 patients who no longer referred olfactory dysfunction at the time of the visit were found to be hyposmic at the smell test ²⁹.

Discussion

Olfactory dysfunction can be either conductive, mainly caused by sinusitis and rhinitis due to the physical blockage of odours in reaching the receptors of olfactory neurons, or sensorineural, involving the interruption of the pathway between olfactory receptors and the olfactory cortex, mainly caused by viral infections, head injuries or neurodegenerative diseases (i.e.: Parkinson’s and Alzheimer’s diseases) ^30-32.

In general, three different types of olfactory (no radiological) testing exist: subjective (patient reported) olfactory assessment, psychophysical olfactory assessment and olfactory assessment based on electrophysiological studies ^31,33. The method used for assessing olfactory dysfunction is extremely important for accurate diagnosis, reporting outcomes and tracking olfactory changes over time ³¹. This should be considered even more important during an outbreak in which olfactory dysfunction seems to be one of the most frequently reported symptoms.

Validated questionnaires or recognised forms of evaluation, possibly quantitative and/or anchored, such as a VAS, can be used in the study of hyposmia. Nevertheless, self-assessment is not well related to the measured olfactory function ^30,31,33, as it has been well shown in the few studies where olfactory tests have been performed in COVID-19 positive patients. In particular, Lechein et al. ²⁷ found that 24% of COVID-19 patients complaining of olfactory loss were normal by the smell test, whereas Moein and colleagues ²⁸ and Vaira et al. ²⁹ observed an underestimation of the self-reported smell dysfunction. In this regard, it is known that in the general population only for anosmia is there correspondence between self-reported olfactory function and the measured one ^31,33. Due to this lack of precision, subjective evaluation of the sense of smell should always be associated with a validated olfactory test to determine the severity of the dysfunction. Furthermore, the measurement of the olfactory function with validated olfactory tests allow quantification of the extent of smell reduction and to evaluate it during clinical follow-up ^31,33.

Psychophysical and electrophysiological procedures are the most effective approaches to assess the integrity of the olfactory system in humans. Electrophysiological tests, such as olfactory event-related potentials, give an objective measure of smell, but are complex and necessitate of expensive equipment, so their clinical use is limited and are mainly reserved for research purposes ^30,32,33. Psychophysical tests, in which subjects are requested to provide a volitional response to the presentation of odourant stimuli, are much more known. Being easy to use, these tests are the most widely employed for quantifying olfactory function in clinical practice ^31,32,34.

Psychophysical olfactory tests can be divided into the threshold and supra-threshold tests. The olfactory threshold is the concentration of an odour in which 50% of stimuli are detected and 50% remain imperceptible to a subject. The supra-threshold olfactory test involves the presentation of the odour with stimuli of sufficient concentration so that they are detectable to the subject. Among commercially available psychophysical olfactory tests, the most widely known are the UPSIT and the Sniffin’ Sticks. UPSIT is a standardised microencapsulated odourant identification test where 40 odourants are presented in a scratch and sniff format with 4 response alternatives accompanying each odour, able to identify normosmia, mild, medium, severe hyposmia and anosmia. Sniffin’ Sticks uses reusable pen-like odour dispensing devices that are presented to the subject by an examiner and consists of three subtests, which allow the study of odour threshold (for n-butanol/phenyl ethyl alcohol), discrimination of 16 odourant triplets and identification of 16 odours. The latter subtest is performed using a multiple-choice task (identification of individual odours is performed from lists of four descriptors for each odour). The sum of the 3 subtest results gives a composite score, known as TDI, which can indicate normal olfactory function, hyposmia, or anosmia ^30-33,35,36. CCCRC is also a composite test, being based on threshold for n-butanol and identification of 10 odour subtests ³². Some shortened psychophysical olfactory tests are also available, mainly based on odour identification ^31-33,37. The number of items presented can range from 4 (4-Item Pocket Smell Test) to 12 odours [Screening 12 test, Cross-Cultural Smell Identification Test (CC-SIT)] ^32,33,37. When based on a small number of odours (up to 12), smell identification tests are mainly considered screening olfactory tests. Nevertheless, the use of these psychophysical instruments should be considered preferable to subjective assessment alone, as questionnaires on self-represented symptoms are not as sensitive or specific as odour identification tests, particularly for mild hyposmia ^31-33.

Viral infections of the upper respiratory tract are the most common cause of hyposmia or permanent anosmia. This loss of smell can reflect damage not only to the olfactory epithelium, but also to central olfactory structures following a viral invasion into the brain. Suzuki et al. ³⁸ identified coronavirus in nasal secretions in a patient with post-viral olfactory dysfunction. The potential route of entry into the central nervous system of SARS-CoV-2 has not yet been established, but several mechanisms including invasion of olfactory nerves and retrograde invasion of the CNS, haematogenous or lymphatic routes all seem to be possible ³⁹. The movement of the COVID-19 virus to the brain via the cribriform plate close to the olfactory bulb has been proposed as possible route for the virus to reach the brain ⁴⁰. An olfactory cleft disease has also been hypothesised as a cause of olfactory loss in patients with COVID-19 ²⁵. Interestingly, very recently, neurobiologists from Harvard Medical School analysed bulk and single-cell RNA-Seq datasets to identify cell types in the olfactory epithelium that express molecules that mediate infection by SARS-CoV-2. They found in both mouse and human datasets that olfactory sensory neurons do not express two key genes required for CoV-2 entry, ACE2 and TMPRSS2. In contrast, non-neural cells in the olfactory epithelium, olfactory epithelial support cells and stem cells of olfactory epithelium express both of these genes, suggesting that infection of these cells could directly or indirectly lead to olfactory dysfunction ⁴¹.

Conclusions

The human sense of smell can be evaluated through subjective methods and/or either psychophysical or electrophysiological olfactory tests. To date, the methods used to investigate smell impairment in COVID-19 patients have been very heterogeneous (Tab. I). Most of the publications that we identified in this review are based on olfactory self-assessment that is unreliable in COVID-19 patients compared to psychophysical tests ^27-29,42, especially in the context of a pandemic scenario and consequent restrictions in social life ⁴³. The authors probably opted for self-administered questionnaire because it can be very difficult to perform an olfactory test during SARS-CoV-2 outbreak, especially when testing recently infected patients. In fact, these patients are isolated or hospitalised, some olfactory tests require an operator who can alter the test result with the protective suit, while other tests are not disposable. For all these reasons, the results of the available studies are not conclusive and many questions remain unanswered about the actual incidence of smell impairment during COVID- 19 infection, its predictive value of a mild- moderate severity of the disease and the percentage of patients recovering.

Given to their simplicity, psychophysical tests, in which the subjects are requested to provide a volitional response to the presentation of odourant stimuli, are the most widely employed for quantifying olfactory dysfunction. Among these, odour identification tests are the most simple and fast ³². It would also be important to use disposable tests, since SARS-CoV-2 can remain viable on plastic, stainless steel and cardboard for up to 72 hours ⁴⁴. Although home-quarantined/isolated COVID-19 patients were able to perform the CCCRC in a self-administered way ⁴⁵, the UPSIT, being self-administered and readily available, should be preferred. Other shorter validated psychophysical tests, such as the CC-SIT (12 odours), the smell Diskettes Test (8 odours) and “Le Nez du Vin” (6 odours), being disposable and not requiring the supervision of a physician/nurse, could also be proposed to test patients in quarantine/isolation and possibly still contagious patients.

The involvement of taste in COVID-19 patients with chemosensory symptoms could be most probably secondary to smell loss ^30,31. Based on gustatory screening test, some authors found that 47% of patients studied with COVID-19 had taste changes ²⁹. To confirm these data, full gustatory testing, i.e. using taste strips ³¹, should be performed to confirm and quantify the taste dysfunction in COVID-19 patients.

In March 2020, the AAOHNS and BRS proposed that hyposmia and dysgeusia (in the absence of other respiratory diseases) should be added to the symptoms used to screen for CoV-2 infection ^5,6. We suggest that validated psychophysical self-administered tests could be valuable methods to investigate smell impairment in COVID-19 patients to identify as accurately as possible cases needing of closer post-infection follow-up.

More extensive studies based on validated psychophysical olfactory tests (systematically performed during active infection) may help to assess the frequency of hyposmia among COVID-19 patients, its pathogenesis, duration and potential role as a marker of disease progression or severity.

Figures and tables

Table I.

Screening and olfactory testing methods used in COVID-19 positive patients.

	Patients complaining of or with smell alterations	Verbal questionnaire	Online questionnaire	Telephone interview	Smart-phone app	UPSIT	Sniffin’ sticks	Nez du vin	Simple odourants	CCCRC	PROMS
Mao 3	11	X
Hopkins 5	2,428#		X
Lee 7	389			X
Lechien 11	357	X
Lechien 12	997	X		X##
Bénézit 13	95		X
Kaye 14	237		X
Menni 16	4,668				X*
Beltrán-Corbellini 17	25	X
Yan 18	40		X								X
Wee 19	35	X
Yan 20	75	X°	X°	X°
Giacomelli 22	31	X
Klopfenstein 23	54	X
Spinato 24	130			X							X
Eliezer 25	1								X^
Ottaviano 26	6							X§			X
Lechien 27	78	X					X^^
Moein 28	59					X**
Vaira 29	72	X***								X
Heidari 8	23	X
Clemency 9	110			X
Gudbjartsson 10	119°°			X

* COVID-19 symptom tracking. Android and IOS app

** Persian version of the 40-item University of Pennsylvania Smell Identification Test (UPSIT)

*** Not clear if questionnaire or telephone interview

^ Smell test with 5 odourants: flower rose, caramel, goat cheese, fruits, manure

^{^^} Sniffin sticks 16 odours identification sub-test only in 46

^§ Identification test (6 odours)

^# Lack of confirmed COVID-19 status in most of the cases

^## In cases of patient isolation

° Data obtained by electronic medical records. If data were not available, patients were either emailed or called

°° Loss of smell and taste were reported together.