key: cord-0742581-6fnv43y1
authors: Hannum, Mackenzie E; Ramirez, Vicente A; Lipson, Sarah J; Herriman, Riley D; Toskala, Aurora K; Lin, Cailu; Joseph, Paule V; Reed, Danielle R
title: Objective sensory testing methods reveal a higher prevalence of olfactory loss in COVID-19–positive patients compared to subjective methods: A systematic review and meta-analysis
date: 2020-09-29
journal: Chem Senses
DOI: 10.1093/chemse/bjaa064
sha: 619efedaaef6152d7d319aac0d19264bf29c14a0
doc_id: 742581
cord_uid: 6fnv43y1

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes coronavirus disease 2019 (COVID-19), has currently infected over 6.5 million people worldwide. In response to the pandemic, numerous studies have tried to identify causes and symptoms of the disease. Emerging evidence supports recently acquired anosmia (complete loss of smell) and hyposmia (partial loss of smell) as symptoms of COVID-19, but studies of olfactory dysfunction show a wide range of prevalence, from 5% to 98%. We undertook a search of Pubmed/Medline and Google Scholar with the keywords “COVID-19,” “smell,” and/or “olfaction.” We included any study that quantified smell loss (anosmia and hyposmia) as a symptom of COVID-19. Studies were grouped and compared based on the type of method used to measure smell loss—subjective measures such as self-reported smell loss versus objective measures using rated stimuli—to determine if prevalence differed by method type. For each study, 95% confidence intervals (CIs) were calculated from point estimates of olfactory disturbances. We identified 34 articles quantifying anosmia as a symptom of COVID-19 (6 objective, 28 subjective), collected from cases identified from January 16 to April 30, 2020. The pooled prevalence estimate of smell loss was 77% when assessed through objective measurements (95% CI of 61.4-89.2%) and 44% with subjective measurements (95% CI of 32.2-57.0%). Objective measures are a more sensitive method to identify smell loss as a result of infection with SARS-CoV-2; the use of subjective measures, while expedient during the early stages of the pandemic, underestimates the true prevalence of smell loss.

In December 2019, an outbreak of a novel coronavirus disease, coronavirus disease 2019 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that originated in Wuhan, China, rapidly spread to almost every country worldwide. As of June 5, 2020, over 6.5 million cases have been identified and over 387,155 deaths have been attributed to the virus (1) . The most common symptoms of infection include fever, dry cough, and fatigue (1) . Other accepted symptoms include difficulty breathing, sore throat, headache, nasal congestion, diarrhea, skin rash, and body aches and pains (1, 2) . However, as knowledge about the virus increased with more confirmed cases, reports of loss of smell and/or taste started to arise. Other the past few months, COVID-19 research has investigated olfactory and taste disturbances as potential symptoms of COVID-19 (3) (4) (5) . Many of these disturbances include the immediate onset of a complete loss of smell (anosmia) and/or taste (ageusia); other studies report hyposmia, a reduction in perceived odor intensity. Therefore, the Centers for Disease Control and Prevention and the World Health Organization officially included losses of smell and taste as symptoms of COVID-19, though less prevalent than some other symptoms (1) .

While olfactory loss is a common symptom of numerous viral respiratory infections (6) , recent reports suggest its prevalence might be higher with SARs-CoV-2 infection (7) . However, there is a wide reported range of olfactory disturbance prevalence, from 5% (8) to 98% (9) . Thus, there is a need to better quantify smell loss during the COVID-19 pandemic (10) . Differences in the reported values may be attributed to different recruiting and sampling methodologies, the range of symptom severity across patients, and the amount of information about COVID-19 available at the time of data collection (e.g., symptom recognition).

However, different data collection techniques used by researchers and health care professionals might also account for the different prevalence estimates reported. There are two general types of methods to measure smell loss: objective and subjective. Objective measures of smell encompass A c c e p t e d M a n u s c r i p t 4 psychophysical testing designed to measure and quantify human responses to physical stimuli. Though sparsely used in COVID-19 research to date, current psychophysical techniques encompass odor threshold tests to determine the lowest concentration of an odor that can be detected, odor discrimination tests to measure the ability to differentiate between odors, and odor identification tests, assessing the ability to correctly name odor qualities. When possible, these tests are performed repeatedly over several days to measure changes in a patient's smell abilities over time. Historically, these objective tests are often executed in a laboratory setting, under surveillance of a researcher or health care professional, to ensure proper completion. Examples of odor threshold tests in a COVID-19 population involve the use of butanol or phenylethyl alcohol at different concentrations (11) (12) (13) . The Sniffin' Sticks test, an odor discrimination and threshold test, is another method to quantify human olfactory performance (14) used now in COVID-19 patients (15, 16) . However, due to the global presence of stay-at-home orders, many researchers have adapted these objective methods to enable testing at home, by patients themselves, with common household odorants (5, 11) .

A more common technique employed to quantify smell loss in the COVID-19 population uses subjective methods, self-report through patient questionnaires or interview or the extraction of symptomatic information from a patient's electronic health records (8, 17, 18) . However, collecting information from records can be prone to underestimation of smell loss due to an initial lack of awareness that it is a symptom of COVID-19. Other subjective methods directly ask patients about their own perceived sense of smell through an online questionnaire (7, 19) , over the phone (18) , or in person with a doctor (20, 21) . However, retrospective assessments through self-report measures are often prone to recall bias (22) . The present review provides a comprehensive assessment of methodologies currently employed to quantify smell loss in COVID-19-positive patients and examines whether method type affects reported prevalence of smell loss in COVID-19 patients. Another recent systematic review examined the prevalence of olfactory loss as a symptom in COVID-19; however, it contained data A c c e p t e d M a n u s c r i p t 5 collected up until April 19, 2020, encompassed different inclusion criteria, included only 10 papers in its analysis, and additionally examined gustatory dysfunction (23) . Building on that prior meta-analysis, we sought to compare differences in prevalence estimates of smell loss collected via objective versus subjective methods. We included any study that quantified smell loss as a symptom of COVID-19, summarizing reports with publication dates up until June 5, 2020.

Article Selection: This systematic review and meta-analysis followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (24) . Figure 1 outlines the steps taken to select articles for inclusion in the meta-analysis. First, Pubmed/Medline and Google Scholar were used to retrieve literature with the keyword "COVID-19" plus "smell" and/or "olfaction" on May 15, 2020, and manual search of relevant articles via Google Scholar was also performed on June 5, 2020, yielding a total of 78 articles.

Titles and abstracts were then screened for their relevance to the topic. Thirty-two articles were initially excluded during the screening test if they were not about smell loss and COVID19, did not report cases or percentage of patients with smell loss, or if they were not written in the English language. If an abstract referenced a measure of prevalence of olfactory dysfunction in COVID-19-positive patients, it was included in the pool of articles (n = 43). Full texts were then screened to confirm positive identification of COVID-19 in patients via a nasopharyngeal swab, throat swab, RT-PCR-confirmed laboratory test, or a clinical assessment by medical professional. Five articles were excluded because patients had not tested positive for COVID-19 by one of these methods (25) (26) (27) (28) (29) . Three articles were excluded due to population bias, with patients recruited for a specific symptom alone (e.g., olfactory disorders) (30) (31) (32) . The exact data needed for our analysis were not reported in three articles, which were thus excluded (15, 33, 34) . Lastly, one article was excluded (4) because of potential data overlap with another report by the same author (35) . The meta-analysis included a total of 34 papers. surveyed. An exception was made for articles that reported taste and/or smell dysfunction when anosmia or hyposmia were not specifically reported. Articles were labeled as using either objective or subjective methods to measure smell loss based on the method the study employed. Objective methods consisted of studies having patients smell a substance, including both household items being selfadministered in their own home and smelling items in a laboratory setting (6 studies in total). Subjective measures included all other methods, for example, self-reports of overall smell loss (28 studies).

Differences in data collection across the studies required further inclusion restrictions regarding how smell loss was reported. When smell loss was reported in tandem with taste loss (e.g., "loss of taste or smell"), this value was extracted. If articles reported smell and taste loss together as well as separately, both values of positive cases with smell loss symptoms were summed to represent all patients presenting smell loss; these values did not include overlapping patients. If articles reported smell and taste loss separately, smell-loss-only values were included.

Three authors (RDH, AKT, and VAR) performed the initial screen and data extraction, and two additional authors (MEH and SJL) validated and resolved disagreements in the data extracted from the articles.

Quality of the articles selected was analyzed with a risk-of-bias assessment checklist adapted from Hoy et al (36) that contains 9 questions which are referred to here as "criteria". Two authors (RDH and SJL) completed the risk-of-bias assessment using the assessment tool outlined by Hoy et al. (36) , as described and adapted by Tong et al (23) . Two additional authors (VAR and AKT) resolved any differences. Supplementary Table S1 details the nine criteria used. Specific questions were scored as 0 (No) or 1 (Yes) for each item, with summary scores of low (0-3), moderate (4) (5) (6) , and A c c e p t e d M a n u s c r i p t 7 high (7-9) risk of bias for the entire study. Supplementary Table S1 contains the full risk-of-bias assessment for each article.

Statistical Analysis. All statistical analyses were performed using R 3.6.0 (37) and RStudio 1.2.1564 (38) . Meta analysis was conducted using the meta package in R (40) . Point estimates of the prevalence of olfactory loss were made by dividing the number of cases of olfactory loss by the total number of subjects included in the study. The Freeman-Tukey double arcsine method was used to transform proportions. A 95% CI was calculated using the Wilson score estimate of the confidence interval, as it a robust method that is reliable across small and large sample sizes (39) . An inverse variance weighting scheme was employed. These parameters were specified in the metaprop function in the R meta package. Heterogeneity was assessed using Cochran's Q and I 2 . Tests for heterogeneity were cut off at Cochran Q-values that were significant (p < 0.05) and I 2 > 50%, because an I 2 of 30-50% was suggested as a cutoff for moderate heterogeneity by Higgins and Thompson (41) . Tau 2 , was reported as a measure of between-study heterogeneity and was estimated by the Dersimmion-Laird method.

Pooled prevalence estimates were computed and reported for both a fixed-effect model and a random-effect model with parameters described. In the fixed-effect model, we assume that there is one true effect size that underlies all the studies in this analysis and that all differences in observed effects are due to sampling error. In the random-effects model, we allow that the true effect size might differ among studies. An overall pooled prevalence estimate was computed for all 34 studies to determine overall prevalence of smell loss in COVID-19 patients.

Subgroup analysis was performed with groupings for objective (N=6 studies) and subjective methods (N=28 studies) for assessing olfactory dysfunction in COVID-19 positive individuals. The R scripts and compiled data used for this analysis are available without restriction on GitHub A total of 28 studies used subjective methods (questionnaire, interview, etc.), comprising 16538 subjects, with 8,166 cases of smell loss. The reported prevalence of olfactory loss ranged from 5% to 88% per study, a larger range than for studies classified as using objective methods. The pooled estimates of the prevalence were 48.6% and 44.4% under the fixed-and random-effect models, respectively. Similar to the objective subgroup, Cochran's Q was significant (Q=5585.37, df=27, p<0.001), and the I 2 value was 99.5%, confirming the heterogeneity of the meta-analysis across the subjective studies. To account for the observed heterogeneity, our pooled prevalence estimate for olfactory loss among COVID-19 positive patients was 44.4% with a 95% CI of 32.2-57.0% under the random-effects model. (Figure 2 ).Discussion

This systematic review and meta-analysis revealed that olfactory dysfunction is a prominent symptom of COVID-19. Meta-analysis using the random-effects model computed an overall prevalence estimate of 50.2% (95% CI: 38.9-61.5%), which is very similar to, although slightly lower than, the A c c e p t e d M a n u s c r i p t 10 previously reported value of 53% in a meta-analysis of olfactory dysfunction in 10 studies (23) . Both meta-analyses confirmed that olfactory dysfunction, regardless of the measurement methodology, is identified in about half of the patients infected with SARS-CoV-2.

Methodological differences in smell loss measurement tools impact reported prevalence estimate of olfactory loss in COVID-19 patients. We further examined if the type of method used to gather information on olfactory loss caused differences in prevalence estimation. Most studies (28 of 34) included in this meta-analysis used subjective methods (self-report) to quantify the prevalence of smell loss; 6 studies used objective methods (e.g., odor threshold tests). The reliance on self-report was expedient due to the pandemic conditions and global stay-at-home orders. However, our analysis revealed a stark difference in prevalence between the two subgroups; studies using objective methods reported around 77% prevalence overall, whereas those using subjective methods reported around 44%.

There are inherent pros and cons regarding each type of methodology. Objective methods quantify smell loss and can limit any confounds because they are often conducted in a controlled environment with standardized procedures. Objective methods rely on true perception of a stimuli when presented, diminishing response and measurement bias. In contrast, subjective methods naturally encompass more variability due to a lack of standardization in how and what questions were asked.

Additionally, subjective methods are often prone to recall bias. However, they are an easy and costefficient way to collect information quickly from the intended population, as demonstrated by the numerous studies in our meta-analysis that used this type of method. Objective methods have higher time and cost requirements than do subjective methods.

The higher overall reported prevalence of olfactory loss in the studies using objective methods (77%) compared to those using subjective methods (44%) suggests that subjective methodologies may neglect to capture crucial information and consistently underestimate true smell loss in A c c e p t e d M a n u s c r i p t 11 patients. One major difference between the two types of methods is the number of variables assessed in each study. Often in the studies using subjective methods the researchers were interested in numerous aspects of COVID-19 symptoms, not just smell loss alone, whereas the studies using objective methods focused solely on sensory loss, using numerous stimuli to get a sensitive measurement of the patient's smell loss. Our findings align with a prior meta-analysis by Tong et al. (23) that found that nonstandardized methods (though all subjective methods by our criteria) severely underestimated olfactory prevalence (estimated at ~37%) compared to standardized methods (which included both subjective and objective methods), which indicated around 87% prevalence. Standardized methods outlined in the Tong et al. study consisted of both objective and subjective measures-because a method is subjective does not mean that the method is not standardized or validated. Validated subjective methods for collecting information on olfactory dysfunction in our pool of studies include a version of the Questionnaire of Olfactory Dysfunction, and the Sino-Nasal Outcome Test (42, 43) .

Overall, however, objective methods have been found to be more sensitive in detecting anosmia and hyposmia than subjective self-reports in a COVID-19 patient population (16) . Additionally, among patients who initially self-reported no smell loss, objective analysis showed mild hyposmia in 30%, again pointing to underreporting by subjective methods (13) . On the other hand, in a COVID-19 patient population recruited due to suspected olfactory loss, 38% of patients with self-reported olfactory dysfunction had normal olfactory performance using the Sniffin' Sticks test, an objective method (15) .

This over-reporting could be due to the biased and specific recruitment (patients with suspected olfactory loss) compared to the general COVID-19 patient population recruited for studies in our metaanalysis.

The higher reported prevalence of olfactory loss when using objective as compared to subjective methods to measure olfactory loss calls for further examination of the consequences of the methodologies employed. Researchers might be missing a critical symptom of COVID-19 through the use A c c e p t e d M a n u s c r i p t 12 of unstandardized, subjective methods to measure smell loss, as demonstrated by the lower prevalence estimate we found in studies classified as using subjective methods. However, objective methods are costly and time-consuming to conduct in standardized laboratory settings. Additionally, more casecontrol studies are needed to validate the reported prevalence rates and compare the findings to a SARS-Cov-2 negative population to account for pre-existing smell disorders.

Many researchers have adapted objective methods to evaluate smell loss to enable use in a home setting. Varia et al found no significant difference in patient smell loss ratings when conducted with an objective method during hospitalization (standardized setting) versus during home quarantine (13) . Irvani et al. used an app to track smell loss in COVID-19 patients over time, which revealed moderate test-retest reliability across sessions among users showing no symptoms, and significant reduction in olfactory function for those who tested positive for COVID-19 compared to those who tested negative (5) . The accessibility and adaptability of these objective approaches make them a resource-efficient strategy to obtain an accurate measure of olfactory loss in COVID-19 patients.

Sudden onset of anosmia is frequently reported to be one of the first presenting symptoms of COVID-19 (23, 44, 45) . In a cohort specifically of patients complaining of smell loss, researchers found that 83% of people reported anosmia as their first symptom of COVID-19 (32) . The actual mechanism by which SARS-CoV-2 may inhibit and disrupt smell perception is currently unknown; however, many potential theories about the cause of smell loss have been proposed. Reports suggest that, different from other coronaviruses, such as those that cause the common cold, SARS-CoV-2 can cause smell loss even in the absence of symptoms such as blockage of the nose, postnasal drip, or a runny nose, which are typical co-occurring manifestations of smell loss from other respiratory viruses (46) . The lack of nasal blockage suggests COVID-19 might be a neurotropic and neuroinvasive disease (47) . Furthermore, it is now commonly known that SARS-CoV-2 coronavirus binds to ACE2 receptors, allowing the virus to enter and infect cells. ACE2 receptors are expressed in nasal epithelium cells (48) , specifically the structures A c c e p t e d M a n u s c r i p t 13 that support olfactory neurons, leading to a theory that infection of these supporting cells might cause additional damage to the olfactory epithelium, resulting anosmia or hyposmia (49) .

Though olfactory dysfunction occurs in high prevalence in patients positive for COVID-19, time to recovery varies across the studies. Several studies report significant improvements quickly after symptom onset, e.g., (7) . Other studies reported that many patients still had not returned to normal sense of smell more than 2 weeks after initial onset of smell loss (16) . Altogether, our meta-analysis demonstrates a prevalence of identified smell loss in about half of the COVID-19 patients, supporting the need to understand the mechanism of infection, onset of symptoms, and recovery from olfactory loss due to a SARS-CoV-2 infection. Furthermore, there is lack of awareness regarding chemosensory function in subjects-many researchers combined the "loss of taste or smell" in their symptomatic findings, even though they are two completely different perceptions that may be impacted differently by SARS-CoV-2. In addition, there remains a lack of comprehensive testing of chemesthetic sensations (e.g., burn from capsaicin or cooling from menthol compounds) (50) . Given the interrelationship between smell and taste, during clinical assessments patients may report changes in taste rather than changes in smell. For patients who report changes in smell and taste function during screening questionnaires, full testing should be performed using objective standardized chemosensory assessment tools. Considering that psychophysical testing may not be possible for all patients and the current social distancing regulations, regular olfactory and gustatory self-assessment at home may be an initial recommendation. Although regular self-assessment may give information about chemosensory function during the trajectory of the disease, the results should be interpreted with caution. In addition, longitudinal assessments of chemosensory function may help identify those patients with continued impairment who may need further treatment and non-pharmacological interventions (e.g., olfactory training).

More research is needed to better establish procedures to estimate prevalence of sensory loss.

Our meta-analysis results reveal underestimates when using subjective techniques, supporting the value of adapting objective methods to estimate smell loss. As information regarding COVID-19 is constantly evolving and is being crowd-sourced, more than ever researchers need to come together on methods to best assess smell loss. To that end, we have created an on-line portal of published studies of COVID-19 and smell loss which is updated frequently (51) . M a n u s c r i p t 16 M a n u s c r i p t M a n u s c r i p t . Forest plot meta-analysis of the prevalence of olfactory dysfunction in COVID-19 patients across studies classified as using objective (top) or subjective (bottom) methodologies. "Events" indicates cases of olfactory loss; "Total" indicates total number of COVID-19-positive patients. Both fixed-effects and randomeffects models are presented. Individual study estimates are represented as "+" on the continuous horizontal line, which represents the 95% CI.

Isolated sudden onset anosmia in COVID-19 infection. A novel syndrome? Rhinology

Loss of smell and taste in combination with other symptoms is a strong predictor of COVID-19 infection

Relationship between odor intensity estimates and COVID-19 population prediction in a Swedish

Olfaction and anosmia in rhinosinusitis

Association of chemosensory dysfunction and Covid-19 in patients presenting with influenza-like symptoms

Neurological Manifestations of Hospitalized Patients with COVID-19 in Wuhan, China: a retrospective case series study

Smell dysfunction: a biomarker for COVID-19. International forum of allergy & rhinology 2020

Corona Viruses and the Chemical Senses: Past, Present, and Future

Objective evaluation of anosmia and ageusia in COVID-19 patients: Single-center experience on 72 cases

Validation of a self-administered olfactory and gustatory test for the remotely evaluation of COVID-19 patients in home quarantine

Olfactory and gustatory function impairment in COVID-19 patients: Italian objective multicenter-study

Sniffin' Sticks': Olfactory Performance Assessed by the Combined Testing of Odor Identification, Odor Discrimination and Olfactory Threshold

Objective olfactory evaluation of selfreported loss of smell in a case series of 86 COVID-19 patients

Anosmia in COVID-19 patients

Alterations in smell or taste -Classic COVID-19?

Self-reported olfactory loss associates with outpatient clinical course in COVID-19

Smell and taste symptom-based predictive model for COVID-19 diagnosis

Prevalence and Duration of Acute Loss of Smell or Taste in COVID-19 Patients

Smell and taste alterations in Covid-19: a crosssectional analysis of different cohorts

Recall bias can be a threat to retrospective and prospective research designs

The Prevalence of Olfactory and Gustatory Dysfunction in COVID-19 Patients: A Systematic Review and Meta-analysis

Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement

Self-reported symptoms of covid-19 including symptoms most predictive of SARS-CoV-2 infection, are heritable

Subjective Changes in Smell and Taste During the COVID-19 Pandemic: A National Survey-Preliminary Results

Early recovery following new onset anosmia during the COVID-19 pandemic -an observational cohort study

Utility of hyposmia and hypogeusia for the diagnosis of COVID-19. The Lancet Infectious Diseases

COVID-19 Anosmia Reporting Tool: Initial Findings

Sudden hyposmia as a prevalent symptom of COVID-19 infection

The close relationship between sudden loss of smell and COVID-19

Anosmia as a prominent symptom of COVID-19 infection

Coincidence of COVID-19 epidemic and olfactory dysfunction outbreak

Coincidence of COVID-19 Infection and Smell: Taste Perception Disorders

Real-time tracking of selfreported symptoms to predict potential COVID-19

Assessing risk of bias in prevalence studies: modification of an existing tool and evidence of interrater agreement

R: A language and environment for statistical computing

RStudio: Integrated Development for R. RStudio

Approximate Is Better than "Exact" for Interval Estimation of Binomial Proportions

How to perform a meta-analysis with R: a practical tutorial

Quantifying heterogeneity in a meta-analysis

Olfaction-associated quality of life in chronic rhinosinusitis: adaptation and validation of an olfaction-specific questionnaire

Psychometric and clinimetric validity of the 20-Item Sino-Nasal Outcome Test (SNOT-20)

Acute-onset smell and taste disorders in the context of COVID-19: a pilot multicentre polymerase chain reaction based case-control study

Anosmia and dysgeusia in patients with mild SARS-CoV-2 infection

Role of Olfaction in Human Health: A Focus on Coronaviruses

Smell and taste dysfunction in patients with COVID-19. The Lancet Infectious Diseases

SARS-CoV-2 entry factors are highly expressed in nasal epithelial cells together with innate immune genes

Possible link between anosmia and COVID-19: sniffing out the truth

More than smell. COVID-19 is associated with severe impairment of smell, taste, and chemesthesis

COVID-19 outbreak in Iraqi Kurdistan: The first report characterizing epidemiological, clinical, laboratory, and radiological findings of the disease

Predictive Value of Sudden Olfactory Loss in the Diagnosis of COVID-19

Olfactory Dysfunction and Sinonasal Symptomatology in COVID-19: Prevalence, Severity, Timing, and Associated Characteristics

High prevalence of olfactory and taste disorder during SARS-CoV-2 infection in outpatients

Olfactory and Gustatory Dysfunction in Coronavirus Disease 19 (COVID-19)

New onset of loss of smell or taste in household contacts of home-isolated SARS-CoV-2-positive subjects

Analysis of Imported Cases of COVID-19 in Taiwan: A Nationwide Study

Olfactory and gustatory dysfunctions as a clinical presentation of mild-to-moderate forms of the coronavirus disease (COVID-19): a multicenter European study

Smell and taste dysfunction during the COVID-19 outbreak: a preliminary report

Self-reported olfactory and taste disorders in SARS-CoV-2 patients: a cross-sectional study

Who should we test for COVID-19? A triage model built from national symptom surveys

Alterations in Smell or Taste in Mildly Symptomatic Outpatients With SARS-CoV-2 Infection

Features of anosmia in COVID-19

Spread of SARS-CoV-2 in the Icelandic Population

The role of self-reported olfactory and gustatory dysfunction as a screening criterion for suspected COVID-19

Loss of Taste and Smell as Distinguishing Symptoms of

Asymptomatic infection and atypical manifestations of COVID-19: Comparison of viral shedding duration

Predictive value of sudden olfactory loss in the diagnosis of

We would like to acknowledge Michael G. Tordoff for his assistance with the review of the literature.