To comprehensively calculate the cumulative prevalence of anosmia and dysgeusia in SARS-CoV-2 infection worldwide, a systematic review was conducted following guideline of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) ²⁵ . The protocol of this systematic review has been registered at PROSPERO ( CRD42020223204).

All articles reporting anosmia and dysgeusia as the symptom of COVID-19 were included. COVID-19 case was defined by a positive RT-PCR for SARS-CoV-2 from either nasopharyngeal swab, oropharyngeal swab, bronchoalveolar lavage, or cerebrospinal fluid. All cross-sectional, retrospective, and prospective studies that randomly sampled COVID-19 cases from community or hospitals were considered eligible; whereas case reports and case series, including all editorials, reviews, and commentaries, were excluded. Studies targeting specific groups such as pregnant females, children, and other groups, were excluded. Only articles written in English during 2019-2020 were included.

Three bibliographical databases (PubMed, Scopus, and Web of Science) and three preprint databases (MedRxiv, BioRxiv, and Researchsquare) were used to identify the potential articles (as of November 10th, 2020). The search criteria were as follows. PubMed ([Title] “SARS-CoV-2” OR “COVID-19” OR “Wuhan coronavirus” OR “Wuhan virus” OR “novel coronavirus” OR “nCoV” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus disease 2019 virus” OR “2019-nCoV” OR “2019 novel coronavirus” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus” OR “coronaviruses” OR “SARS 2” OR “2019-nCoV acute respiratory disease” OR “novel coronavirus pneumonia” OR “COVID”) AND ([All] “Anosmia” OR “smell loss” OR “smell dysfunction” OR “smell impairment” OR “hyposmia” OR “dysosmia” OR “olfactory dysfunction” OR “olfactory disorder”) AND (“dysgeusia” OR “taste loss” OR “taste dysfunction” OR “taste impairment” OR “gustatory dysfunction” OR “gustatory disorder” OR “hypogeusia” OR “ageusia”). Scopus ([Title] “SARS-CoV-2” OR “COVID-19” OR “Wuhan coronavirus” OR “Wuhan virus” OR “novel coronavirus” OR “nCoV” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus disease 2019 virus” OR “2019-nCoV” OR “2019 novel coronavirus” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus” OR “coronaviruses” OR “SARS 2” OR “2019-nCoV acute respiratory disease” OR “novel coronavirus pneumonia” OR “COVID”) AND ([All] “Anosmia” OR “smell loss” OR “smell dysfunction” OR “smell impairment” OR “hyposmia” OR “dysosmia” OR “olfactory dysfunction” OR “olfactory disorder”) AND (“dysgeusia” OR “taste loss” OR “taste dysfunction” OR “taste impairment” OR “gustatory dysfunction” OR “gustatory disorder” OR “hypogeusia” OR “ageusia”). Web of Science ([Title] “SARS-CoV-2” OR “COVID-19” OR “Wuhan coronavirus” OR “Wuhan virus” OR “novel coronavirus” OR “nCoV” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus disease 2019 virus” OR “2019-nCoV” OR “2019 novel coronavirus” OR “severe acute respiratory syndrome coronavirus 2” OR “coronavirus” OR “coronaviruses” OR “SARS 2” OR “2019-nCoV acute respiratory disease” OR “novel coronavirus pneumonia” OR “COVID”) AND ([All] “Anosmia” OR “smell loss” OR “smell dysfunction” OR “smell impairment” OR “hyposmia” OR “dysosmia” OR “olfactory dysfunction” OR “olfactory disorder”) AND ([All] “dysgeusia” OR “taste loss” OR “taste dysfunction” OR “taste impairment” OR “gustatory dysfunction” OR “gustatory disorder” OR “hypogeusia” OR “ageusia”).

Moreover, we searched the preprint servers MedRxiv, BioRxiv, and Researchsquare for non-peer-reviewed articles. Data were extracted from the articles as well as supplementary materials. Reference lists from the eligible articles were retrieved for further relevant studies.

The information of identified articles was imported into EndNote X9 (Thompson Reuters, Philadelphia, PA, USA). Duplicates between databases were removed. To identify eligible studies, the retrieved articles were screened based on title and abstract. The potentially eligible studies were then fully reviewed by two authors (MF and JKF). After reviewing the full texts, the eligibility of each study was decided.

Information of study characteristics, study site, study design, number of patients with anosmia, number of patients with dysgeusia, and COVID-19 characteristics such as number of patients, severity, and outcome were collected.

The primary outcomes were: (a) the global incidence of anosmia in COVID-19 patients; (b) the global incidence of dysgeusia in COVID-19 patients; (c) the association of anosmia with the severity of COVID-19; (d) the association of dysgeusia with the severity of COVID-19; (e) the association of anosmia with mortality of COVID-19; and (f) the association of dysgeusia with mortality of COVID-19. Moreover, this review was conducted to provide the possible pathogenesis of anosmia and dysgeusia in SARS-CoV-2 infection.

The cumulative prevalence rate of anosmia and dysgeusia was calculated for COVID-19 cases by dividing the number of COVID-19 cases with anosmia by the total number of COVID-19 cases with and without anosmia, and was expressed as a percentage (%) with 95% confidence intervals (95% CI). Pooled odds ratios (OR) and 95% CI were calculated to assess the association of anosmia and the occurrence of SARS-CoV-2 compared to non-SARS-CoV-2 respiratory infections. The same method was used for dysgeusia. The pooled OR and 95% CI were presented in a forest plot.

Critical assessment was conducted for the study setting and diagnosis of SARS-CoV-2 to reduce the bias. The Newcastle-Ottawa scale (NOS) ²⁶ was used as critical appraisals to assess the quality of eligible studies. Prior to analysis, gathered data from studies were evaluated for heterogeneity and potential publication bias.

To assess the association between anosmia or dysgeusia and the presence of SARS-CoV-2, Z test was performed ( p < 0.05 was considered statistically significant). Q test was used to evaluate the heterogeneity among studies, and the data with heterogeneity was analyzed using a random effect model. The reporting and publication bias were assessed using Egger’s test and a funnel plot ( p < 0.05 was considered having potential for publication bias). The data were analyzed using Review Manager version 5.3 ²⁷ .

In total, 691 articles (660 reviewed articles and 31 preprint articles) were identified through the databases; of these, 182 articles were removed as duplicates. An additional 287 articles were excluded following a screening process of the titles and abstracts due to irrelevant studies, leaving 222 references ( Figure 1 ). Full-texts of the remaining 222 references were retrieved and screened for eligibility, and this process excluded an additional 115 references as the inclusion criteria was not met. This exclusion included articles with no access ^{28, 29} , RT-PCR not clearly stated in the text ^{30- 45} , case reports ^{46- 97} , case seriess ^{98- 113} , repeated datasetss ^{114- 118} , and studies in specific groups ^{119- 130} . A complete assessment was conducted for 107 references.

The meta-analysis included 107 studies to calculate the prevalence of anosmia in COVID-19 patients. Additionally, 6 studies were excluded while calculating the prevalence of dysgeusia in COVID-19, thus leaving 101 eligible studies. In total, 20 and 16 studies were included to assess the association of anosmia and dysgeusia with the COVID-19 occurrence, respectively.

To calculate the prevalence of anosmia in COVID-19 cases, 107 studies were included comprising 32,142 COVID-19 patients, and anosmia was reported in 12,038 patients with a global pooled prevalence of 38.2% (95% CI: 36.5%, 47.2%). The list of the studies and the prevalence of anosmia in each study are presented in Table 1 .

In total, 30,901 COVID-19 patients from 101 studies were included to calculate the prevalence of dysgeusia in COVID-19. Dysgeusia was identified in 11,337 out of 30,901 COVID-19 patients resulting in a cumulative prevalence of 36.6% (95% CI: 35.2%, 45.2%). The individual studies and the prevalence of dysgeusia from each study are listed in Table 2 .

In total, 20 studies comprising 1,213 COVID-19 cases with anosmia and 2,735 non-COVID-19 patients (mostly COVID-19-like symptoms with negative RT-PCR for SARS-CoV-2) were analyzed to investigate the association between anosmia and the occurrence of COVID-19. Data suggested that anosmia was 10.2-fold more prevalent in patients with COVID-19 compared to those with COVID-19 like illness, OR 10.21 (95% CI: 6.53, 15.96) with p < 0.001 ( Figure 2 ).

In total, 16 studies comprising 1,342 COVID-19 cases with dysgeusia and 1,990 patients with other respiratory illness (COVID-19 like illness with negative RT-PCR for SARS-CoV-2) were included to assess the association between dysgeusia and the occurrence of COVID-19. Data suggested that dysgeusia was 8.6-fold more prevalent in patients with COVID-19 compared to those with other respiratory illness, with OR 8.61 (95% CI: 5.26, 14.11) and p<0.001 ( Figure 3 ).

Limited studies have assessed the association between anosmia and dysgeusia and the severity and mortality of COVID-19 cases. One study linked anosmia with a lower fatality rate and a lower ICU admission ²⁴⁰ .

The pooled prevalence of anosmia in our systematic review was 38.2% of 32,142 COVID-19 cases. This result was almost thrice the initial prevalence reported from Wuhan, China ^{174, 208} . This suggests that anosmia is a potential indicator of SARS-CoV-2 infection, and may be useful for screening and early identification of COIVD-19 patients, particularly asymptomatics ²⁴¹ . Some countries, such as the UK and US have used anosmia as an indicator for preventive measure, wherein COVID-19 patient with anosmia should commence self-isolation ^{242- 244} .

Anosmia is not only present in COVID-19 patients, but also in patients with other respiratory diseases such as influenza, parainfluenza, Eipstein Barr virus, picornavirus, and rhinovirus ^{245- 248} . However, our study demonstrated that the prevalence of anosmia was 10.2-fold higher in COVID-19 patient than that in non-COVID-19 patient. During the previous pandemics, such as severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS), anosmia was rarely reported ²⁴⁹ . Only one study reported persistent anosmia after 2 years of recovery from SARS ²⁵⁰ . Another study reported that anosmia in COVID-19 patients varied based on ethnicity; anosmia in Caucasian is three times more prevalent than in Asian population ²⁵¹ .

Dysgeusia was initially reported in 11.7% of patients who were discharged from Wuhan hospital, which persisted for at least four weeks. This result was lower than ours (36.6% out of 30,901 COVID-19 cases), which might be attributable to either lower dysgeusia prevalence in China or underestimation of this symptom itself ²⁰⁸ . Moreover, the prevalence of dysgeusia in COVID-19 patients was 8.6-fold higher than that in non-COVID-19 like illness. Herpes zoster and HIV have also been linked to gustatory dysfunction ^{252, 253} . Furthermore, another study reported that anosmia and dysgeusia have 82.5% predictive value for positive SARS-CoV-2 RT-PCR ²⁵⁴ .

Several mechanisms have been proposed to explain the emergence of anosmia in COVID-19 patients.

a. Obstruction in the nasal airway

As several viral infections in the respiratory system display blockage of nasal airway or nasal congestion, this hypothesis was initially proposed. According to this mechanism, the interaction between the odorants and olfactory receptors is inhibited by certain obstructions, thereby impairing the subsequent smelling processes ²⁵⁵ . This condition results in anosmia. The obstruction could be caused by nasal discharge or by inflammation occurring in the nasal cavity ²⁵⁶ ; however, this hypothesis can be presumably ruled out. Moreover, several studies reported that anosmia is more prevalent than nasal congestion in COVID-19 patients ^{188, 256- 259} . Interestingly, the incidence of rhinorrhea and nasal obstruction in SARS-CoV-2 infection is lower than other coronaviruses such as SARS-CoV and MERS-CoV ²⁶⁰ .

Furthermore, presumably, nasal obstruction is a secondary mechanism by which anosmia is induced in COVID-19 patients as the obstruction in viral infection typically occurs as a subsequent event after damage in the mucociliary system, thereby inhibiting the nasal discharge and leading to nasal obstruction. In certain viral infections, the mucociliary system operated by ciliated cells is impaired. A previous study reported that human coronavirus (HCoV) disrupted the nasal ciliated respiratory epithelium leading to impaired mucociliary escalator system ²⁶¹ .

b. Damage in olfactory sensory neurons

Smelling processes commence when the odorants bind to the olfactory sensory neurons (OSNs) in the olfactory epithelium located in the nasal cavity, which subsequently transmits this information through their axons to the olfactory bulb in the brain ²⁶² . According to this concept, a viral attack on the receptor neurons eventually creates disturbances in the sense of smell; however, this hypothesis remains under debate as several recent studies reported the absence of angiotensin-converting enzyme 2 (ACE2) and transmembrane protease serine 2 (TMPRSS), the key factors for the virus to enter the cell ²⁶³ , in the OSNs ^{264- 267} . These findings are supported by another study carried out by Bryche et al., who demonstrated that SARS-CoV-2 was not detected in the OSNs of hamsters ²⁶⁶ .

Moreover, after comparing the duration between anosmia incidence in COVID-19 patients and the normal cellular regeneration process, this proposed mechanism should be reconsidered. Several studies reported that COVID-19-related anosmia disappeared within 1-2 weeks, whereas regeneration of dead OSNs requires more than 2 week time period ^{188, 206, 255, 262, 268} . This discrepancy results in a temporary conclusion that COVID-19-related anosmia is not directly associated with the impairment of the OSNs.

c. Olfactory center damage in the brain

The aforementioned dysfunction of OSNs and the mechanism by which SARS-CoV-2 directly affects the olfactory center via axonal transport of the neuron remains unclear, as the OSN lacks ACE2 and TMPRSS2 which hinders viral entry into the cell ^{264- 267} . Nevertheless, the possibility of olfactory center disruption caused by SARS-CoV-2 should not be overlooked as the cause of anosmia, since a previous study concluded that human ACE2 (hACE2)-transgenic mice suffered from brain infection after intranasal inoculation with SARS-CoV ²⁶⁹ . The study found that the brain infection commenced from the olfactory bulb, which is the axonal trajectory pathway of the OSNs ²⁶⁹ . This finding suggests that SARS-CoV-2 might also first utilize another structure in the nasal cavity before it is transported into the OSNs.

d. Olfactory supporting cells dysfunction

As OSN does not express ACE2 and TMPRSS2, the virus should use another pathway to infect the olfactory system. Numerous studies have established the expression of these SARS-CoV-2 entry proteins in several supporting cells in olfactory epithelium, that is, Bowman’s gland cells, horizontal basal cells, olfactory bulb pericytes, mitral cells, sustentacular cells, and microvillar cells ^{264- 267} . Of these supporting cells, the sustentacular cells have gained immense attention as the initial site of SARS-CoV-2 infection in the olfactory epithelium. In addition to their higher expression of ACE2 and TMPRSS2 than the others, sustentacular cells are located on the surface of the nasal cavity making them vulnerable to exposure to the external environment ^{264, 267} .

Notably, sustentacular cells act as supporting cells and promote olfactory neuron in the olfactory system. These cells detoxify harmful odorants, promote odorant-receptor binding, and provide nutritional substances to support the action of olfactory receptor neurons ^{255, 264} . Considerably, it is plausible to suggest that any damage occurring in sustentacular cells will in turn affect the olfactory epithelium and produce anosmia.

The corresponding regeneration time to the recovery of anosmia also supports the notion that sustentacular cell damage relates to anosmia caused by SARS-CoV-2. As the replenishment of dead OSNs does not correspond to the duration of COVID-19-related anosmia within 1-2 weeks, the regeneration of sustentacular cells seems to be in line with that time frame ^{264, 266, 268} .

Furthermore, this hypothesis is supported by a recent study conducted by Bryche et al., who reported that SARS-CoV-2 was accumulated in sustentacular cells but not in the OSNs ²⁶⁶ . The olfactory epithelial damage and sustentacular cell loss occurred 2 days after instilling SARS-CoV-2 intranasally in golden Syrian hamsters ²⁶⁶ .

e. Inflammation-related olfactory epithelium dysfunction

It is worth noting that the cytokine storm in COVID-19 is strongly associated with organ dysfunctions, including OSNs ²³² . The dysfunction in this structure can lead to disturbance in the sense of smell ²⁷⁰ . Torabi et al. suggested that proinflammatory cytokines, particularly tumor necrosis factor a (TNF-α), may lead to COVID-19-induced anosmia ²⁷¹ . Another proinflammatory cytokine, interleukin-6 (IL-6), increased in cases presenting with anosmia ^{232 , 272} .

The mechanism used by these cytokines, in particular IL-6, to produce anosmia is not fully understood. Cazzolla et al. suggested that this effect can be caused by either peripheral or central action of the cytokines ²³² . In the periphery, IL-6 may induce apoptosis of ciliary neuronal cells in the olfactory epithelium ²⁷² , whereas in its central action, the olfactory center in the brain is attacked by the cytokine as a result of virus infection ²³² .

Although gustatory impairment is always displayed concomitantly with olfactory dysfunction, this symptom has a relatively different mechanism and is often distantly linked to the latter symptom. Several hypotheses have been proposed to explain the mechanism behind the emergence of dysgeusia in COVID-19 patients.

a. The subsequent effect of cranial nerves dysfunction

Considering the close relationship between the olfactory and gustatory system both peripherally and centrally, smell and taste dysfunction in COVID-19 often occurs concomitantly ^{256, 273} . This hypothesis describes dysgeusia as a secondary event of olfactory dysfunction ²⁷⁴ ; however, several studies revealed that the percentage of dysgeusia in COVID-19 patients is higher than symptoms related to olfactory dysfunction ^{188, 275} . Based on this finding, another mechanism may be involved in inducing SARS-CoV-2-related dysgeusia. Furthermore, COVID-19-induced dysgeusia could also occur when there is certain damage in the cranial nerves responsible for gustatory transmission (cranial nerve VII, IX, and X) ²⁷⁶ . Among these nerves, SARS-CoV-2 exposure to cranial nerve VII has gained immense attention. Based on this hypothesis, the virus initially colonizes the nasopharynx structure, then moves to the Eustachian tube, and eventually reaches the middle ear where the virus gets access to chorda tympani and causes dysgeusia ²⁷⁶ .

b. Zinc deficiency

Another interesting hypothesis underlying dysgeusia in COVID-19 is related to zinc deficiency ²⁷⁶ . This hypothesis was developed as zinc is an important mineral in carbonic anhydrase, which is pivotal in maintaining taste sensation ²⁷⁷ . Interestingly, one study reported that zinc level in patients with SARS-CoV-2 infection was significantly lower compared to that in the healthy control groups ²⁷⁸ . Alterations in the sense of taste after being treated with certain treatments, such as irradiation in cancer patients ^{279, 280} , could be prevented by zinc supplementation. Moreover, dengue fever virus and human immunodeficiency virus replication could be inhibited by zinc chelation ^{281, 282} . Furthermore, pharmacological agents influencing ACE2 activity are associated with taste disturbances ^{283, 284} .

Nevertheless, this effect does not relate to zinc deficiency as these drugs do not influence both serum and salivary zinc concentrations ²⁸⁴ . Further investigation needs to be carried out to reveal the role of zinc in dysgeusia associated with COVID-19.

c. SARS-CoV-2-bound sialic acid

SARS-CoV-2 may produce dysgeusia via interaction with sialic acid receptors ^{232, 274, 285} . Sialic acid plays a pivotal role in the taste processing pathway as it is a component of the normal salivary composition ²⁸⁶ . Moreover, reduced amount of sialic acid impairs the ability to taste ²⁸⁷ . An in silico study revealed that SARS-CoV-2 could interact with the sialic acid receptor through its spike protein ²⁸⁸ . Previously, MERS-CoV was also reported to interact with this receptor ²⁸⁹ . Following this occupancy, the gustatory threshold increases, while gustatory particles degrade at a higher rate ^{274, 287} .

d. Direct attack on several oral sites

A previous study investigated the expression of ACE2 in various tissues in the oral cavity and found that the tongue had higher ACE2 expression in comparison to other tissues, such as buccal and gingival tissues ²⁹⁰ . This finding raised a hypothesis that SARS-CoV-2 could directly attack the taste buds in the tongue, initiating inflammatory responses, and would eventually alter the sense of taste ²⁷⁶ . It is proposed that the Toll-like receptor-mediated cascade and apoptosis are the subsequent events that could lead to taste dysfunction ^{276, 291} .

A previous study investigating SARS-CoV infection in rhesus macaques revealed that, initially, the salivary gland was attacked by the virus ²⁹² . As the human salivary gland expresses a high level of ACE2 ²⁹³ , it is reasonable to pay more attention to the vulnerability of this gland against SARS-CoV-2 exposure. Disruption in the activity of the salivary gland would produce either imbalance in salivary composition or impairment of salivary flow, which could ultimately result in dysgeusia ²⁷⁶ .

All data underlying the results are available as part of the article and no additional source data are required.

Figshare: PRISMA checklist for ‘Anosmia and dysgeusia in SARS-CoV-2 infection: Incidence, effects on COVID-19 severity and mortality, and the possible pathobiology mechanisms - A systematic review and meta-analysis’, https://doi.org/10.6084/m9.figshare.13323080.v1 ²⁹⁴ .

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).