key: cord-1027451-1xxq8r26
authors: Agyeman, Akosua Adom; Lee Chin, Ken; Landersdorfer, Cornelia B.; Liew, Danny; Ofori-Asenso, Richard
title: Smell and Taste Dysfunction in Patients With COVID-19: A Systematic Review and Meta-analysis
date: 2020-06-06
journal: Mayo Clin Proc
DOI: 10.1016/j.mayocp.2020.05.030
sha: ceea87d130408355cd0379229d0cebc3f2826f6d
doc_id: 1027451
cord_uid: 1xxq8r26

Abstract Objective To estimate the prevalence of olfactory and gustatory dysfunctions (OGDs) among patients infected with novel coronavirus disease 2019 (COVID-19). Patients and Methods A systematic review was conducted by searching MEDLINE, EMBASE, and the preprint server MedRxiv until 11 May 2020 using the terms ‘anosmia’ or ‘hyposmia’ or ‘dysosmia’ or ‘olfactory dysfunction’ or ‘olfaction disorder’ or ‘smell dysfunction’ or ‘ageusia’ or ‘hypogeusia’ or ‘dysgeusia’ or ‘taste dysfunction’ or ‘gustatory dysfunction’ or ‘neurological’ and ‘COVID-19’ or ‘2019 novel coronavirus’ or ‘2019-nCoV’ or ‘SARS-CoV-2’. References of included studies were also manually screened. Only studies involving diagnostic-confirmed patients with COVID-19 were included. Random-effects meta-analysis was performed. Results A total of twenty-four studies with data from 8438 test-confirmed COVID-19 patients from thirteen countries were included. The pooled proportion of patients presenting with olfactory and gustatory dysfunctions was 41.0% (95% confidence interval [CI] 28.5% to 53.9) and 38.2% (95% CI 24.0 to 53.6%), respectively. Increasing mean age correlated with lower prevalence of olfactory (coefficient = -0.076; p=.02) and gustatory (coefficient = -0.073; p=.03) dysfunctions. There was a higher prevalence of olfactory dysfunctions with the use of objective measurements compared to self-reports (coefficient = 2.33; p=.01). No significant moderation of the prevalence of OGDs by gender was observed. Conclusion There is a high prevalence of OGDs among patients infected with COVID-19. Routine screening for these conditions could contribute to improved case detection in the ongoing COVID-19 pandemic. However, to better inform population screening measures, further studies are needed to establish causality.

this review (3 May 2020) , no systematic review and meta-analysis on the topic had been published.

This systematic review follows the recommendations outlined in the PRISMA statement 10 and the Cochrane collaboration handbook 11 (Supplementary Table S1 ). Given the urgent need for information to inform clinical decision making in the context of the COVID-19 pandemic, prior registration of the review protocol was not feasible.

To identify appropriate studies for the review, searches were performed in MEDLINE, EMBASE and the preprint server MedRxiv using the terms 'anosmia' or 'hyposmia' or 'dysosmia' or 'olfactory dysfunction' or 'olfaction disorder' or "smell dysfunction' or 'ageusia' or 'hypogeusia' or 'dysgeusia' or 'taste dysfunction' or 'gustatory dysfunction' or 'neurological' and 'COVID-19' or '2019 novel coronavirus' or '2019-nCoV' or 'SARS-CoV-2'

(Supplementary Table S2 ). The search was initially performed on 3 May 2020, and last updated on 11 May 2020. No language restrictions were applied. The reference lists of included studies were also hand-searched for additional articles.

Only studies with data on patients with test-confirmed COVID-19 status were eligible for inclusion. Moreover, as our objective was to estimate the prevalence of smell and taste dysfunctions separately, studies that reported a composite of smell and taste dysfunction without presenting individual data on each outcome were excluded. When studies from the same center recruiting patients over the same period were present, we selected the one with the larger sample size or more detailed information. Furthermore, we excluded case series involving <10 patients, as well as commentaries, editorials, and reviews.

The methodological quality of each study was assessed using a tool developed by Murad et al. 12 This tool involves eight items under four domains: selection, ascertainment, causality, and reporting (Supplementary Table S3 ). Only items applicable to the studies included in the review were evaluated. Numeric scoring was not performed and an overall judgement about methodological quality of the included studies was made as per the questions deemed most critical in the specific clinical scenario. 12

Data were extracted independently by two reviewers (AAA and RO), and any disagreements were resolved via consensus. For each study, first author name, country, participant mean age, proportion of females, and percentage with smell or taste disorders were extracted. If a study assessed OGDs based on patients' self-report and objective measurements, we prioritized the data from the objective assessments. Analyses were conducted using Stata SE software version 16 (StataCorp, TX, USA). Meta-analysis of prevalence was conducted via the STATA Metaprop command, 13 using the Freeman-Tukey double arcsine transformed (FTT) proportions to account for variance instability. 11 Owing to the anticipated between-study heterogeneity, a randomeffects model was used. Between-study heterogeneity was quantified using the I 2 statistic. 11 An I 2 >50% was considered to represent substantial heterogeneity. Leave-one-out sensitivity analyses were performed to assess the stability of pooled estimates. Using the STATA Metareg command, 14 meta-regression was conducted to determine the influence of study mean age, proportion of females, assessment method (subjective versus objective) and region of study (Europe versus other) on the pooled prevalence. The syntax for the meta-regression was "Metareg logitEventRate age [proportion of females] region [assessment method], wsse(logitEventSE)". Where logitEventRate=log(p/(1-p)) and logitEventSE=[sqrt(1/(p*Total) + 1/((1-p)*Total))], and p=proportion and Total=sample size. A two-tailed p-value of <.05 was considered significant. As our study relied on published literature, institutional ethical review was not required.

The searches in MEDLINE and EMBASE retrieved 341 citations. Following removal of duplicates and screening of titles and abstracts, 52 articles were selected for full text evaluation.

Fifteen articles were retained after full-text assessment. Nine additional studies were identified from the preprint server, resulting in 24 studies being included in the review ( Figure 1 ). The included studies were from China (n=1), 24 the Netherlands (n=1), 15 the UK (n=1), 19 Iran (n=1), 23 Israel (n=1), 25 South Korea (n=1), 30 US (n=2), 21,29 France (n=4), 17, 18, 26, 37 Germany (n=4), 20, 31, 32, 36 Italy (n=4), 16, 27, 35, 38 Spain (n=2), 22, 34 Belgium (n=1), 33 and one multi-country study across Europe (France, Spain, Italy, Belgium, and Switzerland). 28 The studies involved a total of 8438 patients. The mean age ranged from 34.0 to 77.0 years, and 9.1% to 67.9% of the patients were females (Table 1) . The methodological assessment of the individual studies is presented in Supplementary Table   S4 . As our review included only COVID-19 patients with diagnostic confirmation, bias from non-ascertainment of exposure was minimized. However, very few studies had reported use of objective assessment of OGDs with the majority relying on unvalidated questionnaires. Thus, studies may suffer from ascertainment bias. Most studies (20/24; 83%) did not explore other potential causes that may explain the outcome, nor could they establish with certainty that OGDs were absent prior to COVID-19 and therefore causality cannot be implied.

Of the 24 studies that reported prevalence of olfactory dysfunction, 21% (5/24) used objective assessments, whereas the remaining (19/24; 79%) mainly relied on self-reports. The reported prevalence of olfactory dysfunction ranged from 3.2 to 98.3%, and the pooled prevalence was 41.0% (95% confidence interval [CI] 28.5 to 53.9%; I 2 =99.1%) (Figure 2 The variables included in the meta-regression altogether explained 51.7% of the between-study variance relative to the prevalence of olfactory dysfunctions. There was no strong evidence of publication bias (Funnel plot, Figure S1 ; Eggers test, p=.306).

Fifteen studies involving 5649 patients, reported the prevalence of gustatory dysfunctions.

Among these, 13% (2/15) used objective assessments, whereas the remaining (13/15; 87%) relied mainly on self-reports. The reported prevalence of gustatory dysfunctions ranged from 5.6 to 62.7%, and the pooled prevalence was 38.2% (95% CI 24.0 to 53.6%: I 2 =98.8) (Figure 3) included in the meta-regression altogether explained 47.9% of the between-study variance relative to the prevalence of gustatory dysfunctions. There was no strong evidence of publication bias (Funnel plot, Figure S2 ; Eggers test, p=.604).

In this systematic review and meta-analysis, we found that 41% and 38% of diagnosticconfirmed patients with COVID-19 presented with olfactory or gustatory dysfunctions, respectively. Increasing age correlated with lower prevalence of OGDs, whereas the use of objective assessment methods correlated with higher prevalence of olfactory dysfunction. No significant moderation of the prevalence of OGDs by gender was noted. The exact mechanisms underlying OGDs among patients with COVID-19 remain unclear.

However, olfactory impairment after upper respiratory tract infection is a common occurrence in clinical settings. In particular, post-viral olfactory dysfunction has been implicated in 40% of cases of anosmia in adults, 39 with coronaviruses accounting for 10 to 15% of cases. 5 Olfactory dysfunction in COVID-19 could be related to the involvement of the olfactory bulb or to peripheral damage of the olfactory receptor cells in the nasal neuroepithelium. 40 This assertion is based on the potential neurotrophic features of SARSCoV-2. In particular, it has been demonstrated in transgenic mice that following intranasal administration of SARS-CoV (which shares similarities with SARS-CoV-2) the virus could penetrate into the brain through the olfactory bulb, leading to rapid transneuronal spread. 41 It is also well recognized that alterations in the volume and composition of saliva can disturb taste sensation. 42 Previously, epithelial cells lining salivary gland ducts were found to be early target cells of SARS coronavirus infection in the upper respiratory tracts of rhesus macaques. 43 Phylogenetic similarities between SARS-CoV and SARS-CoV-2, means the latter could also alter gustatory sensation in affected patients.

The few studies that have evaluated the clinical utility of OGDs in COVID-19 diagnosis have suggested their low sensitivity (23-43%) and high specificity (93-99%). 17, 44 Regardless, in one study, the sensitivity and specificity of OGDs were reported to be comparable to the sensitivity and specificity of a history of close contact with a confirmed COVID-19 case. 44 An analysis of 237 entries from the AAO-HNS COVID-19 Anosmia Reporting Tool suggested that anosmia was noted in 73% of patients prior to COVID-19 diagnosis and was the initial symptom in 26.6%. 45 Yan et al, 21 have also suggested that OGDs may be associated with a milder course of COVID-19 infection. This may also potentially explain the lower prevalence of OGDs with increasing mean age, as older people are more likely to experience severe COVID-19 infection compared to younger individuals. 46 Consequently, the potential higher burden of OGDs in patients with milder COVID-19 disease is concerning, as such patients may be less likely to be tested but could continue to spread the virus. Thus, public education about symptoms of OGDs may be necessary, and patients experiencing such symptoms should be advised to self-isolate pending confirmatory testing. 9 It is not yet clear whether the COVID-19-related OGDs are transient or permanent. However, among 23 COVID-19 patients with anosmia in Iran, 75% reported significant improvement over a period of two weeks. 32 Moreover, among 237 patients with anosmia on the AAO-HNS COVID-19 Anosmia Reporting platform, 27% reported improvement in symptoms, with the mean time to improvement being 7.2 days. 45 Owing to the high occurrence of anosmia in COVID-19-postive individuals, the indiscriminate use of corticosteroids, particularly in the absence of known head trauma or allergic symptoms should be discouraged, as corticosteroids may escalate COVID-19 infection. 47 Moreover, as more evidence evolves around COVID-19, further studies addressing therapeutic approaches to OGDs among infected patients will be needed.

Our study has some key strengths. Although a recent meta-analysis by Tong et al. reported the pooled prevalence of olfactory and gustatory dysfunction among patients with COVID-19 as 52.73% (95% CI, 29.64%-75.23%) and 43.93% (95% CI, 20.46%-68.95%), respectively, their analysis was based on 10 studies and involved <1700 patients. 48 Our study has provided pooled prevalence estimates of OGDs based on data from 24 studies involving more than 8400 diagnostic-confirmed COVID-19 patients from 13 countries. Furthermore, our analysis provides new insights into the potential role of individual-level characteristics such as age and gender in the presentation of OGDs among COVID-19 patients which were not explored in the study by Tong et al. 48 Despite the strengths of our study, the findings should be interpreted in light of some limitations.

There was high statistical heterogeneity, which is largely attributable to the observational nature of the studies. Further investigations revealed that the heterogeneity was not entirely explained by differences in patients age, proportion of females, region or OGDs' assessment methods.

From a methodological perspective, the design of most of the included studies preclude the confirmation of causality between COVID-19 and OGDs. Thus, to better inform population screening measures, further well-designed prospective studies using validated or objective measurement techniques are needed to establish causality. Most studies also recruited patients in European settings, which may affect the generalizability of our findings. Moreover, few studies utilized objective assessment methods for establishing the presence of OGDs, whereas the majority relied on self-reports. These may lead to bias in the ascertainment of OGDs. For example, it is possible for patients to confuse taste function and aroma sense perception, 49 33 Lastly, as the meta-regression relied on study-level characteristics such as mean age and proportion of females, ecological fallacy cannot be entirely ruled out. 11, 50 Thus, future studies adopting methods such as the use of individual participant data (IPD) meta-analysis, 11,50 may be essential towards confirming our findings.

This study found a high prevalence of OGDs among patients infected with COVID-19. Hence, the consideration of OGDs as part of the screening and diagnostic approaches for COVID-19 could help improve case detection and further curtail the spread of the virus. However, to better inform population screening measures, further well-designed studies are needed to establish causality between COVID-19 and the occurrence of OGDs. Table 1 : Characteristics of included studies 

COVID-19 Coronavirus pandemic

Clinical Characteristics of Coronavirus Disease 2019 in China

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Sniffing out the evidence; It's now time for public health bodies recognize the link between COVID-19 and smell and taste disturbance

Presentation of new onset anosmia during the COVID-19 pandemic

Coincidence of COVID-19 epidemic and olfactory dysfunction outbreak

The use of google trends to investigate the loss of smell related searches during COVID-19 outbreak

Loss of Sense of Smell as a marker of COVID-19 infection

American Academy of Otolaryngology-head and neck surgery (AAO-HNS). Anosmia, Hyposmia, and Dysgeusia Symptoms of Coronavirus Disease

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

Cochrane Handbook for Systematic Reviews of Interventions version 6

Methodological quality and synthesis of case series and case reports

Metaprop: a Stata command to perform meta-analysis of binomial data

Meta-regression in Stata

Strong associations and moderate predictive value of early symptoms for SARS-CoV-2 test positivity among healthcare workers, the Netherlands

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

Utility of hyposmia and hypogeusia for the diagnosis of COVID-19

Features of anosmia in COVID-19

Clinical features of 95 sequential hospitalised patients with novel coronavirus 2019 disease (COVID-19), the first UK cohort

Clinical infectious diseases : an official publication of the Infectious Diseases Society of America

Self-reported olfactory loss associates with outpatient clinical course in Covid-19. Int Forum Allergy Rhinol

Acute-onset smell and taste disorders in the context of Covid-19: a pilot multicenter PCR-based case-control study

Smell dysfunction: a biomarker for COVID-19. Int Forum Allergy Rhinol

Neurologic Manifestations of Hospitalized Patients With Coronavirus Disease

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

Cluster of COVID-19 in northern France: A retrospective closed cohort study

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

Clinical and Epidemiological Characteristics of 1,420 European Patients with mild-to-moderate Coronavirus Disease 2019

Clinical features, laboratory characteristics, and outcomes of patients hospitalized with coronavirus disease 2019 (COVID-19): Early report from the United States

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

Distinguishing between COVID-19 and the common cold in a primary care setting -comparison of patients with positive and negative SARS-CoV-2 PCR results

Psychophysical Olfactory Findings of Mild-to-moderate COVID-19 Patients: Preliminary

A cohort of patients with COVID-19 in a major teaching hospital in Europe

Clinical presentation and evolution of COVID-19 in immunosuppressed patients. Preliminary evaluation in a North Italian cohort on calcineurin-inhibitors based therapy

COVID-19 in people living with human immunodeficiency virus: A case series of 33 patients

Multivariable prediction model of intensive care unit transfer and death: a French prospective cohort study of COVID-19 patients

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

Olfactory disorders following upper respiratory tract infections

Defining the burden of olfactory dysfunction in COVID-19 patients

Middle East Respiratory Syndrome Coronavirus Causes Multiple Organ Damage and Lethal Disease in Mice Transgenic for Human Dipeptidyl Peptidase 4

Role of saliva in the maintenance of taste sensitivity

Epithelial cells lining salivary gland ducts are early target cells of severe acute respiratory syndrome coronavirus infection in the upper respiratory tracts of rhesus macaques

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

Anosmia Reporting Tool: Initial Findings. Otolaryngol Head Neck Surg

Clinical features of COVID-19 in elderly patients: A comparison with young and middle-aged patients

Anosmia, hyposmia, and dysgeusia as indicators for positive SARS-CoV-2 infection

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

Confusing tastes and smells: how odours can influence the perception of sweet and sour tastes

Avoiding Methodological Biases in Meta-Analysis

Belgium (Mons) Objective measurement (Olfactory dysfunction: sniffin sticks tests)

Self-report (Gustatory dysfunction, validated NHANES questionnaire