Taste alteration in COVID-19: Significant geographical differences exist in the prevalence of the symptom

Abstract

Early detection of COVID-19 is important for reduction in the spread of the disease and gustatory disturbances (GD) are known to have a strong predictive value. In the present study, we aimed to map the geographical differences in the prevalence of GD in individuals infected with SARS-CoV-2 during the first wave of COVID-19 in order to improve case identification and to facilitate prioritization. We undertook a rapid scoping review of articles published in the repository of the National Library of Medicine (MEDLINE/PubMed) and medRxiv from their inception until 3^rd September, 2020. The minimum requirements for completing a restricted systematic review were fulfilled. Of the 431 articles retrieved, 61 studies (28,374 cases confirmed with COVID-19) from 20 countries were included in the analysis. GD were most prevalent in the Americas [66.78%, 95% CI 54.77–78.79%] compared to Europe [57.18%, 95% CI 52.35–62.01%], the Middle East [38.83%, 95% CI 27.47–50.19%] and East Asia [13.1%, 95% CI 0.14–26.06%]. No differences of GD prevalence were evident between February and August 2020. The data demonstrate that there is a marked geographical distribution of GD in COVID-19 patients which, possibly, might be explained by differences in diagnostic criteria for COVID-19 case definition during the early phase of the pandemic.

Introduction

The early identification and confirmation of suspected cases of Coronavirus Disease 2019 (COVID-19) is fundamental to limiting the spread of the disease but is particularly challenging for asymptomatic or pauci-symptomatic patients. The sudden loss of taste (ageusia), with or without loss of smell (anosmia), have been cited as independent signs of the disease although, more frequently, these symptoms occur in association with the most common manifestations of the disease, namely, fever, cough and fatigue [1]. Unlike early studies [2,3], recent systematic reviews have indicated that gustatory dysfunction (GD), including ageusia, hypogeusia and dysgeusia, is common in COVID-19 patients [4]. Furthermore, loss of taste and smell have been reported to be distinguishing symptoms of COVID-19 which have a high predictive value [5]. The European Centre for Disease Prevention and Control (EDCD) was one of the first public health bodies to include sudden onset anosmia, ageusia or dysgeusia as clinical criteria to identify possible COVID-19 cases [6]. Unfortunately, these symptoms have not been used unanimously for case identification and for testing prioritization. On 5^th August 2020, however, the Centers for Disease Control and Prevention in the USA updated their COVID-19 case definition to include GD as an important clinical criterion for diagnosis [7]. Soon afterwards (7^th August, 2020), the World Health Organization updated its COVID-19 case definition to include recent onset ageusia, in the absence of any other identifiable cause, as suggestive of COVID-19 infection [8].

In the present study, we hypothesized that differences in the criteria used for COVID-19 identification by national and/or local public health bodies may reflect, at least in part, the changes of known prevalence rates of these symptoms over time and in a geographically specific manner. Specifically, we undertook a rapid systematic review of the prevalence of GD in COVID-19 cases and examined worldwide data from East Asia, the Middle East (including Turkey), Europe (including Britain) and the Americas.

Results

We identified 431 studies; 91 relevant articles met the inclusion criteria. 61 of the 91 studies were included for data analysis (Fig. 1 ). The studies were from 20 different countries; 5 studies were derived from multi-national collaborations. The majority of cohorts were from Europe (n = 40), followed by the Middle East (n = 8), North and South America (n = 6 and n = 2, respectively), East Asia (n = 6) and Africa (n = 1). Two articles [9,10] pooled multinational data within Europe and 3 studies [[11], [12], [13]] included cases from two main geographical areas. The study populations and prevalence range are depicted graphically in Fig. 2 .

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of study selection process.

Fig. 2.

Worldwide prevalence of gustatory disorders (GD) in COVID-19 positive cases. Circle size is proportional to study population. Color-coded bars indicate percentage range of GD prevalence. The map depicted in the picture was freely downloaded and modified from https://www.freeworldmaps.net/powerpoint/.

Worldwide, 14,486 of 28,374 confirmed COVID-19 cases (51.05%) reported subjective and/or objective GD (Table 1 ). Strikingly, there were significant differences of prevalence between subgroups (ANOVA, p = 0.000106; Kruskal–Wallis, p = 0.00071). Further, when each geographical region was compared, there were significant differences between all of the subgroups except Europe vs America (Suppl Table 1). Studies from East Asia reported the lowest prevalence of GD (13.1%, 95% CI 0.14–26.06%), followed by the Middle East (38.83%, 95% CI 27.47–50.19%), Europe (57.18%, 95% CI 52.35–62.01%) and the Americas (66.78%, 95% CI 54.77–78.79). We could not highlight trends of increased GD prevalence in COVID-19 patients over time, except in East Asia (Supplementary Fig. 1).

Table 1.

Studies included for the evaluation of gustatory dysfunction (GD).

Reference	GD	COVID+	Prevalence	Area
Romero-Sanchez et al. [33]	52	841	6.2%	Spain
Paderno et al. [34]	320	508	63%	Italy
Zayet et al. [35]	62	95	65%	France
Vaira et al. [36]	39	72	54.2%	Italy
Lapostolle et al. [37]	757	1487	50.9%	France
Martin-Sanz et al. [38]	114	215	53%	Spain
Lechien et al. [39]	770	1420	54.2%	Europe (multiple sites)
Villarreal et al. [40]	161	230	70%	Spain
Petrocelli et al. [41]	184	300	61.3%	Italy
Qiu et al. [13]	50,20	116,39	45.16%	France, Germany (and China)
Abalo-Lojo et al. [42]	74	131	56.5%	Spain
Zayet et al. [43]	34	70	48.6%	France
Tudrej et al. [44]	92	198	46.5%	France
Poncet-Megemont et al. [45]	81	139	58.3%	France
Patel et al. [46]	89	141	63.1%	UK
Chary et al. [47]	64	115	55.65%	France
Sierpinski et al. [48]	923	1942	47.5%	Poland
Izquierdo-Dominguez et al. [49]	442	846	52.2%	Spain
Rojas-Lechuga et al. [50]	128	197	65%	Spain
Giacomelli et al. [51]	17	59	28.8%	Italy
Menni et al., 2020 [12]	4178	6452	64.76%	UK (and USA)
Vaira et al. [52]	76	106	71.7%	Italy
Mercante et al. [53]	113	204	55.4%	Italy
Lechien et al. [10]	1136	2013	56%	Europe (multiple sites)
Klopfenstein et al. [54]	34	70	48%	France
Liguori et al. [55]	48	103	46.6%	Italy
Luers et al. [56]	50	72	69.4%	Germany
Dell'Era et al. [57]	232	355	65.4%	Italy
Magnavita et al. [58]	31	82	37.8%	Italy
Meini et al. [59]	69	100	69%	Italy
Lechien et al. [9]	342	385	88.8%	Belgium
Beltran-Corbe llini et al. [60]	28	79	35.44%	Spain
Vaira et al. [61]	234	340	67.8%	Italy
Gelardi et al. [62]	52	72	72.2%	Italy
Vacchiano et al. [63]	66	108	61%	Italy
De Maria et al. [64]	48	95	50.5%	Italy
Fistera et al. [65]	6	43	14%	Germany
Vaira et al. [66]	91	138	65.9%	Italy
Hintschich et al. [67]	18	41	44%	Germany
Chiesa-Estomba et al. [11]	718	1043	68.8%	Europe (multiple sites)
	12,043	21,062	57.18%	TOTAL EUROPE
Yan et al. [68]	42	59	71%	USA
Menni et al. [12]	490	726	67.49%	USA (and UK)
Carignan et al. [69]	85	134	63.4%	Canada
Pinna et al. [70]	10	so	10%	USA
Chiesa-Estomba et al. [71]	333	542	61.4%	South America (Multiple sites)
Lee et al. [72]	32	56	57.1%	Canada
Kempker et al. [73]	27	51	52.9%	USA
Brandao Neto et al. [74]	499	655	76.2%	Brazil
	1518	2273	66.78%	TOTAL AMERICA
Sayin et al. [75]	46	64	71.9%	Turkey
Biadsee et al. [76]	67	128	52%	Israel
Altin et al. [77]	22	81	27.2%	Turkey
Salepci et al. [78]	77	223	34.5%	Turkey
AI-Ani and Acharya [79]	28	141	19.86%	Qatar
SakaIii et al. [80]	81	172	47.1%	Turkey
Levinson et al. [81]	14	42	33.3%	Israel
Çalica Utku et al. [82]	51	143	35.7%	Turkey
	386	994	38.83%	TOTAL MIDDLE EAST
Lee et al. [83]	353	3191	11%	Korea
Mao et al. [84]	12	214	5.6%	China
Kim et al. [85]	58	172	33.7%	Korea
Qiu et al. [13]	30	239	12.55%	China (and France, Germany)
Cho et al. [86]	36	83	43.4%	Hong Kong
Liang et al. [87]	33	86	38.4%	China
	522	3985	13.1%	TOTAL EAST ASIA
Farah et al. [88]	17	60	28.3%	Somalia
	14,486	28,374	50.05%	WORLD

Discussion

An awareness of the association between taste alterations and COVID-19 infection is important for diagnosing the disease, particularly in dental and oral health settings [14]. In the present study, we followed a streamlined approach to synthesizing evidence (the rapid review) which is typically used for informing emergent decisions faced by decision-makers in health care settings. The results of this rapid systematic review show that there are distinct geographical patterns of GD in patients with established SARS-CoV-2 infection.

The first systematic assessments of the evidence available up to March 2020 failed to identify associations between anosmia/ageusia and COVID-19 [2,3]. For example, in an early systematic review involving a total of 1556 patients, olfactory or gustative dysfunctions were not reported [2]. In another early study which examined evidence of anosmia in COVID-19 patients, researchers found the symptom to be of “limited and inconclusive” value [3]. The first study reporting a 5.1 and 5.6% prevalence of hyposmia and hypogeusia, respectively, was a pre-print (non-peer-reviewed) case series of a Chinese population [15]. In sharp contrast, the most recent meta-analysis analyzing smell and taste alterations not only reported that almost half of COVID-19 patients had these symptoms but also, that 15% of patients had olfactory and gustatory abnormalities as their initial presenting symptoms [16].

Recent systematic reviews have assessed chemosensory alterations in COVID-19 patients [4,[16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30]]. Only one review, however, has focused specifically on taste changes [31]. In their pooled analysis, Aziz et al. [31] found that almost half of patients (49.8%) with COVID-19 had altered taste sensation. Similarly, when the data was pooled in the reviews assessing chemosensory alterations, the prevalence of olfactory and gustatory alterations occurred in approximately half of COVID-19 patients but when these symptoms were considered individually, there was a marked range of prevalence (3.2–100% olfactory symptoms; 0–92.6% gustatory disturbances), The results of the present study are in agreement with the existing literature on the worldwide prevalence of GD associated with COVID-19 disease; of 28,374 confirmed COVID-19 cases, 14,486 (51.05%) reported GD. Taken together, and in association with other findings [32], there is strong evidence to suggest that gustatory alterations are cardinal symptoms of COVID-19.

A major limitation of the present study is that it includes research with diverse study designs and patients with different disease severity, e.g. severe, mild, or asymptomatic COVID-19 cases. Further, the majority of the studies that were analysed were cross-sectional, retrospective and observational, hence recollection bias may have been present. Most studies were similar to those previously graded as “moderate risk of bias” [29]. Importantly, the presence of taste alterations may not have been reported in the presence of other more severe symptoms such as dyspnea, fever and productive cough which could explain the lack of association between GD and COVID-19 in the first studies published in February and March 2020. For these reasons, the true prevalence of ageusia, hypogeusia and dysgeusia might be significantly higher than reported [31].

Disturbance of taste sensation can occur as a result of local and systemic conditions, including oral, nasal, and sinus disease, metabolic (obesity, diabetes, poor nutrition, hypertension), neurological (epilepsy, Alzheimer’s and Parkinson’s disease, schizophrenia, multiple sclerosis), tumours and radiation associated with cancer treatment, drugs (allopurinol, anti-hypertensives, atatins, lithium), head trauma, and certain habits (smoking, cocaine snorting). Significantly, in the published literature, the measures for assessing GD have not been validated and the definition of dysgeusia has not been unanimously accepted. Further, the more patients and doctors are aware of the possibility of GD with COVID-19, the more cases would have been reported and investigated over time [12] although in the present study, the data failed to show a trend of increased GD prevalence over time except in East Asia. By contrast to other studies [2], our analysis focused on geographical location rather than ethnicity. Whilst this approach may have failed to identify individual genetic/ethnic determinants of infection, we believe that our methodology was better suited to study the clinical manifestations of the disease and to inform the decisions of public health surveillance bodies.

The data that were examined in the present study involved the subjective (self-reported) and objective (testing with the four basic tastes of sweet, sour, salty and bitter modalities sprayed onto the tongue in a supra-threshold doses) interpretation of GD. Previous findings, however, have shown that there are no significant differences between the subjective and objective interpretation of gustatory function [18] and, therefore, we suggest that self-reported taste alterations can be considered a reliable parameter for GD in COVID-19 patients.

In conclusion, we show that GD in COVID-19 exhibits distinct geographical patterns of prevalence. Given the potential usefulness of taste assessment in the diagnosis of mildly and pauci-symptomatic patients, we believe that it is imperative to recognize ageusia/hypogeusia/dysgeusia as a potential clinical manifestation of COVID-19, particularly in Europe and America. Dentists, therefore, may be the first healthcare providers to diagnose taste disturbances and are likely to play an important role in case identification and early diagnosis of COVID-19 cases in the future.

Methods

Study design and literature search

This study was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines and used a rapid review approach due to time constraints [89]. The study complied with the minimum requirements for completing a restricted systematic review [90]. Accordingly, the search was performed by one investigator (N.C.) and verification of a random sample of full texts for accuracy of title/abstract screening and data extraction was undertaken by the same reviewer. Key terms used for the search were (SARS-CoV-2 or COVID or COVID-19) in association with taste or ageusia or hypogeusia or dysgeusia or gustatory. The search was conducted in PubMed/MEDLINE as well as medRxiv using the advanced search (title and abstract) tool.

Study selection and data extraction

The exclusion criteria were as follows: articles not in English, duplicate publications, irrelevant articles, studies where the infection status was not clearly confirmed, studies that did not evaluate gustatory outcomes individually, simple case reports, and review or systematic review articles. Studies using telephone surveys or Apps were only included where the respondents had a confirmed COVID-19 diagnosis. For studies reporting cases from two or more geographical areas (e.g. East Asia and Europe), the data for subgroup analysis were extracted only when information from individual countries was available. Where the date of patient recruitment was not provided, the date of the article submission was used as a surrogate source of information. The primary outcome was to assess the prevalence of gustatory alterations (ageusia, hypogeusia, dysgeusia) in confirmed COVID-19 cases worldwide and in distinct geographical areas; the secondary outcome was to establish a spatio-temporal pattern of GD in published cases. No constraints were placed on the size of the cohorts to ensure a comprehensive search and to identify the maximum number of potential articles.

Statistical analysis

Subgroup analyses were based on the country of origin of the studies by pooling the actual data reported in each individual study. Differences in prevalence (% and category) among subgroups were assessed with chi-square statistics and one-way ANOVA, as appropriate. Tukey’s post-hoc test or Student’s t tests were used for comparison between group pairs. By making a further assumption that the dependent variable may not be normally distributed, the Kruskal-Wallis test was also used to compare overall differences in prevalence. Where appropriate, Pearson’s coefficient was used to assess the correlation between time and prevalence. A level of p < 0.05 was chosen to determine statistical significance.

Ethics approval and consent to participate

Not applicable.

Consent for publication

N.C. approved the final version of the manuscript.

Availability of data and material

The datasets used and/or analyzed during the current study can be made available by the corresponding author on a reasonable request.

Competing interests

The author declares no conflict of interest.

Funding

The author received no specific funding for this study.

Authors’ contributions

N.C. conceived the study, undertook the search and wrote the manuscript.

Appendix A. Supplementary data

The following is Supplementary data to this article: