Introduction
Different levels of olfactory and gustatory dysfunction in patients with coronavirus disease-2019 (COVID-19) have been reported since the beginning of the pandemic. In order to prevent further spread of COVID-19, early identification and isolation are of great importance. ENT-UK advised people experiencing new onset of anosmia even without any other symptom to isolate themselves.1 The American Academy of Otolaryngology-Head and Neck Surgery proposed the use of these symptoms as a screening tool.2 Additionally, some authors believe that guidelines should include testing of patients with olfactory or gustatory dysfunctions for COVID-19.Reference Heidari, Karimi, Firouzifar, Khamushian, Ansari and Mohammadi Ardehali3 Olfactory and gustatory dysfunction are now accepted as manifestations of this novel disease,Reference Rocke, Hopkins, Philpott and Kumar4 although a thorough explanation of the relationship requires more work.
This study aimed to assess the prevalence, onset, course and severity of olfactory dysfunction and taste impairment among patients in our hospital's COVID-19 in-patient clinic. We also planned to scrutinise their accompanying nasal symptoms, possible correlation with laboratory results, pulmonary status determined by imaging and COVID-19 severity.
This study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.Reference von Elm, Altman, Egger, Pocock, Gøtzsche and Vandenbroucke5
Materials and methods
We designed a prospective observational cohort study of patients with confirmed COVID-19 who were admitted to our hospital's COVID-19 isolation units between 1 April and 1 May. This study was approved by the Turkish Ministry of Health and the local ethics committee (reference number: 2020/521). Verbal informed consent was obtained from all participants by telephone interview.
Patients who were available to be contacted by telephone within their first three days of hospital admission were included in the study. Exclusion criteria were patients who (1) had a history of permanent smell or taste loss, (2) had a history of rhinological disease or surgery, (3) had diseases causing cognitive problems, and (4) were misdiagnosed with COVID-19.
All eligible patients were reached by telephone to minimise the risk of transmission to healthcare workers and cross infection. Patients were asked to participate in an 11-item questionnaire survey. They were also informed that they would be contacted by phone two weeks after completion of their treatments for another eight-item questionnaire interview if they reported olfactory dysfunction or taste impairment in the first survey. Patients who agreed to join the study and gave consent by telephone were enrolled in the study.
Additionally, data about patient characteristics and comorbidities were reviewed from electronic medical records. Clinical features, baseline levels of serum biochemical parameters (neutrophil-to-lymphocyte ratio, C-reactive protein (CRP), ferritin, d-dimer and fibrinogen) possibly related to prognosis,6–Reference Richardson, Hirsch, Narasimhan, Crawford, McGinn and Davidson8 oxygen saturation, duration of hospitalisation, need for mechanical ventilation and admission to the intensive care unit were also noted. A thoracic radiologist examined chest computed tomography (CT) findings of all patients. Chest CT scores were assessed, and pulmonary findings were classified as: 1, mild; 2, moderate and 3, severe.Reference Ooi, Khong, Müller, Yiu, Zhou and Ho9,Reference Simpson, Kay, Abbara, Bhalla, Chung and Chung10 The patients were categorised into three groups (1, mild/moderate group; 2, severe group; and 3, critical group) according to the disease severity classification recommended by the World Health Organization (WHO).11
Given the cross-sectional observational nature of the study, sampling bias and recall bias could occur. We followed the patients’ clinic status and excluded misdiagnosed cases to overcome sampling bias. However, we did not use any randomisation method for participant selection, which may cause sampling bias. We preferred to include all eligible patients who were treated within a pre-determined time frame. By contacting the patients within the early period of their hospital stay and shortly after they were healed, we believe we decreased the risk of reporting errors to a low level.
Results
Statistical analysis
Power analysis for a bivariate correlation was performed in G*Power (Heinrich Heine University Düsseldorf, Germany) statistical power calculator to determine an adequate sample size with an alpha of 0.05, a power of 0.95 and a medium effect size (ρ = 0.3) in a two-tailed analysis.Reference Faul, Erdfelder, Buchner and Lang12 Based on the above assumptions, the required sample size was calculated as 134.
Descriptive statistics were used to define categorical variables. Associations between categorical variables were evaluated by chi-square or Fisher's exact test. Normality assumption was not satisfied by the continuous variables of the study. Therefore, non-parametric tests were used for further analysis. Spearman's rho was performed to assess the correlation between two continuous variables. Group comparisons were performed with the Mann–Whitney U test for two independent groups, with the Wilcoxon signed-rank test for more than two independent groups and Friedman's test for more than two dependent groups. Pairwise comparisons with Bonferroni correction after omnibus tests were performed. The analyses were conducted using MedCalc statistical software (version 12.7.7; Ostend, Belgium).
A total of 172 patients with COVID-19 were eligible during the study period. Among these patients, 145 responded to the first questionnaire. Five patients declared history of olfactory dysfunction. Seven patients were eventually diagnosed with other diseases mimicking COVID-19 symptoms. Finally, 133 patients were included in the study. Moreover, 46 of 49 patients who had olfactory dysfunction, taste impairment or both were able to fill out the second questionnaire (second interview could not be set with 3 patients because of intensive care unit admission).
The mean (± standard deviation (SD)) patient age was 56.03 ± 14.05 years (range, 27–93 years), and 56 patients (64.7 per cent) were male and 47 patients (35.3 per cent) were female. Demographic data, ratios of the patients who had olfactory dysfunction or taste impairment, chest CT scores and grading, and disease severity classification of the patients are presented in Table 1.
Table 1. Descriptive and clinical data
A total of 49 patients (36.8 per cent) reported having olfactory dysfunction, taste impairment or both. Moreover, 31 (23.3 per cent) and 41 (30.8 per cent) patients experienced olfactory dysfunction and taste impairment, respectively, and 23 (17.2 per cent) patients had both. Among these patients, six reported anosmia and ageusia, seven reported ageusia without any olfactory problem, and one reported ageusia with hyposmia. Olfaction and taste recovered totally in almost all patients. Only two of the patients reported partial recovery. No patient reported a permanent loss without any degree of recovery.
The average recovery (mean ± SD) time for olfactory dysfunction and taste impairment was 6.4 ± 2.12 and 6.18 ± 1.75 days, respectively. Remarkable differences were noted in the median verbal rating scale scores for olfactory function during the active phase of the disease (median, 3; range, 0–6) compared with the time prior to the disease (median, 10; range, 7–10) and 2 weeks after healing (median, 10; range, 5–10; p < 0.001). In addition, the verbal rating scale scores were significantly lower for taste sensation during the course of the disease (median, 3; range, 0–8) compared with that prior to disease (median, 10; range, 7–10) and 2 weeks after healing (median, 10; range, 5–10; p < 0.001) (Figure 1).
Fig. 1. Verbal rating scale score for olfactory dysfunction and taste impairment prior to disease, during the disease and two weeks after healing. (a) Verbal rating scale score for olfactory dysfunction and (b) verbal rating scale score for taste impairment. VRS = verbal rating scale
No significant differences in the distributions of sex, age and comorbidities were observed between the patients who suffered from olfactory dysfunction or taste impairment (Table 2). However, the median recovery time from hyposmia was 1.5 days longer in female patients (median, 7; range, 5–10) than in male patients (median, 5.5; range, 2–10; p = 0.047). Additionally, patients with hypertension reported longer recovery time from both hyposmia and dysgeusia than patients without hypertension (p = 0.033 and p = 0.032, respectively).
Table 2. Relationship between descriptive data and olfactory dysfunction or taste impairment
*n = 31; †n = 102; ‡n = 41; **n = 92
A significant association was found between olfactory dysfunction or taste impairment and nasal obstruction (p = 0.001). Although 10 (32.3 per cent) of 31 patients with olfactory dysfunction had a new-onset nasal obstruction, 12 (29.3 per cent) of 41 patients with dysgeusia reported a new-onset nasal obstruction. However, no association with other nasal symptoms was observed (Table 2).
No statistically significant association was found between olfactory dysfunction (p = 0.706) and taste impairment (p = 0.35) with either disease severity or CT grading. However, the onset of olfactory dysfunction compared with other classical symptoms of COVID-19 was the latest to occur in patients with critical disease (p = 0.04).
No statistically significant association was observed between any parameters related to olfactory dysfunction or taste impairment and duration of hospitalisation (p > 0.05). Furthermore, no correlation was noted between olfactory dysfunction or taste impairment verbal rating scale scores and laboratory parameters and CT scores (p > 0.05).
Statistically significant differences were observed in the distributions of the neutrophil count, neutrophil-to-lymphocytes ratio, CRP and fibrinogen between the patient groups when categorised based on their disease severity and CT grading. The median and p-values of the parameters are given in Tables 3 and 4. Additionally, a moderate correlation was noted between CT scores and CRP (r s = 0.396; p < 0.001) and fibrinogen levels (r s = 0.33; p < 0.001). Moreover, a weak correlation was found between CT scores and ferritin levels (r s = 0.207; p = 0.018) and peripheral capillary oxygen saturation (SpO2) (r s = −0.228; p = 0.009). However, no significant difference was observed in the initial laboratory parameters between patients who showed a CT scoring progression and the ones who did not (p > 0.05).
Table 3. Association of disease severity with clinical findings
Statistical significance was observed. †n = 88; ‡n = 36; **n = 9; §Bonferroni correction was applied. M = mild/moderate; S = severe; C = critical
Table 4. Association of chest CT grading with clinical findings
*Statistical significance was observed; †Bonferroni correction was applied. CT = computed tomography; M = mild/moderate; S = severe; C = critical; CRP = C-reactive protein
Nine (6.7 per cent) patients needed treatment in the intensive care unit after hospital admission. Of them, two patients had only hyposmia, while one had both hyposmia and dysgeusia. Four patients died among the patients transferred to the intensive care unit, and two of them reported hyposmia.
Discussion
A wide range of clinical presentations have been reported in COVID-19.Reference Richardson, Hirsch, Narasimhan, Crawford, McGinn and Davidson8,Reference Mao, Wang, Chen, He, Chang and Hong13 While the spread of this disease has expanded, physicians began to encounter more patients complaining of olfactory dysfunction and taste impairment.Reference Eliezer, Hautefort, Hamel, Verillaud, Herman and Houdart14–Reference Vaira, Salzano, Deiana and De Riu18 Personal observations and studies from several countries have been shared rapidly to call attention to the association between COVID-19 and olfactory dysfunction.Reference Eliezer, Hautefort, Hamel, Verillaud, Herman and Houdart14,Reference Vaira, Salzano, Deiana and De Riu18–Reference Marinosci, Landis and Calmy20 In this study, we evaluated olfactory dysfunction and taste impairment in patients hospitalised for COVID-19 and their relation with COVID-19 pneumonia and disease severity. Olfactory dysfunction and taste impairment were assessed by questionnaire survey conducted over the telephone, including a verbal rating scale. A subjective method was preferred as similar to other studies in the literature.Reference Beltrán-Corbellini, Chico-García, Martínez-Poles, Rodríguez-Jorge, Natera-Villalba and Gómez-Corral19,Reference Spinato, Fabbris, Polesel, Cazzador, Borsetto and Hopkins21–Reference Coelho, Kons, Costanzo and Reiter29
We found that the presence of olfactory dysfunction or taste impairment did not correlate with disease severity in patients hospitalised for COVID-19. Although two studiesReference Spinato, Fabbris, Polesel, Cazzador, Borsetto and Hopkins21,Reference Yan, Faraji, Prajapati, Ostrander and DeConde30 have suggested that olfactory dysfunction or taste impairment is a prognostic factor for the milder form of COVID-19, both studies were conducted in out-patients. Therefore, the possibility of the difference in reporting olfactory dysfunction or taste impairment between hospitalised patients and those not hospitalised should be kept in mind when interpreting these data. In this study, we assessed a certain cohort of patients (only hospitalised) to reduce this bias. Similar to our findings, studies using objective olfactory measurement methods from Italy and Iran have reported no association between chemosensory dysfunction and disease severity.Reference Moein, Hashemian, Mansourafshar, Khorram-Tousi, Tabarsi and Doty31–Reference Vaira, Hopkins, Salzano, Petrocelli, Melis and Cucurullo33 We defined disease severity according to the classification proposed by the WHO,11 which is based on reports from China,Reference Wu and McGoogan34 while in the study by Moein et al.,Reference Moein, Hashemian, Mansourafshar, Khorram-Tousi, Tabarsi and Doty31 the criteria of the Massachusetts General Hospital COVID-19 treatment guidance were used. The presence of pneumonia and the criteria proposed by Tian et al.Reference Tian, Hu, Lou, Chen, Kang and Xiang35 was used to determine disease severity in studies conducted by Vaira et al.Reference Vaira, Deiana, Fois, Pirina, Madeddu and De Vito32,Reference Vaira, Hopkins, Salzano, Petrocelli, Melis and Cucurullo33 With regard to the distribution of disease severity, our cross-sectional cohort is consistent with other large cohort studies and data stated by the WHO.11,Reference Wu and McGoogan34
Another key finding was that no statistically significant association was found between olfactory dysfunction or taste impairment and chest CT grading. Early personal observations in our clinical practice led us to probe if there was less lung involvement in patients reporting chemosensory dysfunction. The results of the study do not support our initial hypothesis. To the best of our knowledge, this is the first study investigating the association between olfactory dysfunction or taste impairment and COVID-19 pneumonia in patients hospitalised for COVID-19.
The prevalence of olfactory dysfunction or taste impairment in patients with COVID-19 has ranged from 5 per cent up to 98 per cent in the literature.Reference Mao, Wang, Chen, He, Chang and Hong13,Reference Moein, Hashemian, Mansourafshar, Khorram-Tousi, Tabarsi and Doty31 Such a wide range may be explained with the heterogeneity of the study protocols.Reference Fuccillo, Saibene, Canevini and Felisati36 Our results are consistent with those of other studies conducted with hospitalised patients.Reference Beltrán-Corbellini, Chico-García, Martínez-Poles, Rodríguez-Jorge, Natera-Villalba and Gómez-Corral19,Reference Giacomelli, Pezzati, Conti, Bernacchia, Siano and Oreni37 On the contrary, some studies have reported lower or higher olfactory dysfunction or taste impairment.Reference Lechien, Chiesa-Estomba, De Siati, Horoi, Le Bon and Rodriguez24,Reference Lechien, Cabaraux, Chiesa-Estomba, Khalife, Plzak and Hans27,Reference Coelho, Kons, Costanzo and Reiter29–Reference Vaira, Hopkins, Salzano, Petrocelli, Melis and Cucurullo33,Reference Izquierdo-Dominguez, Rojas-Lechuga, Mullol and Alobid38,Reference Vaira, Salzano and De Riu39
The study cohort had a male predominance, with a mean age of 56 years, reflecting the previously reported demographic and clinical characteristics of the disease.Reference Richardson, Hirsch, Narasimhan, Crawford, McGinn and Davidson8,Reference Fu, Wang, Yuan, Chen, Ao and Fitzpatrick40 The median recovery time from olfactory dysfunction was 1.5 days longer in female than in male patients. Although a study from Italy reports similar observations,Reference Meini, Suardi, Busoni, Roberts and Fortini25 our data show similar mean recovery time and maximum range in both female and male patients, so these should be interpreted cautiously. An unexpected finding was that patients with hypertension reported longer recovery time from both hyposmia and dysgeusia than patients without hypertension. This may be associated with the altered expression of angiotensin-converting-enzyme-2 receptor on the mucosa of the oral cavity, epithelial cells of the tongue, olfactory sustentacular cells and olfactory stem cells, especially in patients taking angiotensin-converting-enzyme-inhibitor drugs.Reference Brann, Tsukahara, Weinreb, Lipovsek, Van den Berge and Gong41 Further studies are needed to clarify this issue.
Permanent anosmia did not occur in any of the patients. After two weeks, olfaction and taste recovered totally in almost all patients. Only two patients reported a partial recovery. Together with the average recovery time for olfactory dysfunction and taste impairment, these findings support published reports suggesting an early and high rate of recovery.Reference Rocke, Hopkins, Philpott and Kumar4,Reference Beltrán-Corbellini, Chico-García, Martínez-Poles, Rodríguez-Jorge, Natera-Villalba and Gómez-Corral19,Reference Kaye, Chang, Kazahaya, Brereton and Denneny22,Reference Lechien, Chiesa-Estomba, De Siati, Horoi, Le Bon and Rodriguez24,Reference Yan, Faraji, Prajapati, Ostrander and DeConde30,Reference Vaira, Deiana, Fois, Pirina, Madeddu and De Vito32,Reference Klopfenstein, Kadiane-Oussou, Toko, Royer, Lepiller and Gendrin42 This finding distinguishes COVID-19 olfactory dysfunction from other post-viral olfactory dysfunctions, which mostly recover after longer than a year.Reference Lee, Lee, Wee and Kim43 Vaira et al.Reference Vaira, Salzano, Fois, Piombino and De Riu44 and Meini et al.Reference Meini, Suardi, Busoni, Roberts and Fortini25 suggested that olfactory dysfunction in severe acute respiratory syndrome coronavirus-2 is not related to definitive damage from the virus to the neuronal cells, considering the high rate of rapid recovery. Eliezer et al.Reference Eliezer, Hautefort, Hamel, Verillaud, Herman and Houdart14 reported normal magnetic resonance imaging findings of the olfactory bulb in COVID-19-related olfactory dysfunction and supported this hypothesis.
Although Speth et al.Reference Speth, Singer-Cornelius, Oberle, Gengler, Brockmeier and Sedaghat26 showed no correlation between olfactory dysfunction and nasal obstruction and rhinorrhoea, our study showed a significant association between olfactory dysfunction or taste impairment and nasal obstruction symptoms (p = 0.001). New-onset nasal obstruction was found in 32.3 per cent and 9.3 per cent of patients with olfactory dysfunction and taste impairment, respectively. These rates are similar to those reported from three different continents.Reference Spinato, Fabbris, Polesel, Cazzador, Borsetto and Hopkins21,Reference Kaye, Chang, Kazahaya, Brereton and Denneny22,Reference Wee, Chan, Teo, Cherng, Thien and Wong45 In a prospective study, Lechien et al.Reference Lechien, Cabaraux, Chiesa-Estomba, Khalife, Plzak and Hans27 showed an even higher rate (48.7 per cent) of concomitant nasal obstruction. Hence, nasal obstruction was not present in all patients with olfactory dysfunction, and they suggested this finding as an indicator for ‘neural’ loss of smell. We think that the coexistence of nasal obstruction in COVID-19-related olfactory dysfunction (a statistically significant association in our study) supports a ‘conductive’ loss hypothesis. Moreover, in the light of previous reports,Reference Eliezer, Hautefort, Hamel, Verillaud, Herman and Houdart14,Reference Meini, Suardi, Busoni, Roberts and Fortini25,Reference Vaira, Salzano, Fois, Piombino and De Riu44 we agree with the ‘olfactory cleft syndrome with mucosal obstruction’ hypothesis proposed by Gane et al.Reference Gane, Kelly and Hopkins15
The present study contradicts most reports showing a relatively higher rate of chemosensory dysfunction as an initial presenting symptom,Reference Rocke, Hopkins, Philpott and Kumar4,Reference Beltrán-Corbellini, Chico-García, Martínez-Poles, Rodríguez-Jorge, Natera-Villalba and Gómez-Corral19,Reference Kaye, Chang, Kazahaya, Brereton and Denneny22,Reference Ottaviano, Carecchio, Scarpa and Marchese-Ragona46 while a few studies have reported similar findings.Reference Spinato, Fabbris, Polesel, Cazzador, Borsetto and Hopkins21,Reference Lechien, Chiesa-Estomba, De Siati, Horoi, Le Bon and Rodriguez24 We found no relation between the duration of hospitalisation and the presence of olfactory dysfunction or taste impairment. Since patients with more severe disease would have had a longer hospital admission, this finding can be considered an internal verification of the study because it corroborates the previous statement that olfactory dysfunction or taste impairment are not associated with disease severity.
In this study, no correlation was noted between olfactory dysfunction or taste impairment verbal rating scale scores and laboratory parameters (p > 0.05). These results were expected as olfactory dysfunction or taste impairment, as previously mentioned, were not associated with disease severity and COVID-19 pneumonia. Obtained data about laboratory parameters are consistent with the current literature.6,Reference Richardson, Hirsch, Narasimhan, Crawford, McGinn and Davidson8,Reference Wu and McGoogan34 We believe that all these data regarding laboratory parameters show that the study group represents a well-distributed cohort with possible but minimal sampling bias and a natural disease course.
Among the strengths of the present study are: (1) the prospective design of data collection, (2) follow-up data for olfactory dysfunction or taste impairment after the first questionnaire, (3) the first study to evaluate the association between olfactory dysfunction or taste impairment and chest CT grading, and (4) grading chest CT (by an experienced thoracic radiologist) and disease severity according to validated criteria.
The major limitations of this study are: (1) lack of data concerning the period after the first questionnaire interview for patients without olfactory dysfunction or taste impairment, (2) lack of objective olfactory assessment, (3) a relatively small study population, and (4) lack of patients who were not admitted to hospital. Future prospective, controlled studies with objective olfactory assessment tools are needed to establish the association between olfactory dysfunction or taste impairment and pulmonary involvement and disease severity in COVID-19.
• Different levels of olfactory and gustatory dysfunction in patients with coronavirus disease (COVID-19) have been reported
• Olfactory dysfunction or taste impairment do not have clinical repercussions on disease severity
• Olfactory dysfunction or taste impairment do not have clinical repercussions on chest computed tomography findings
• Nasal obstruction is observed frequently as an accompanying symptom of olfactory dysfunction or taste impairment compared with other nasal symptoms
• The onset of olfactory dysfunction compared with other classical symptoms of COVID-19 was found the latest in patients with critical disease
Conclusion
Patients with COVID-19 often report olfactory dysfunction or taste impairment. Nasal obstruction may be present as an accompanying symptom. Olfactory dysfunction or taste impairment do not have clinical repercussions on disease severity and chest CT findings.
Competing interests
None declared
Open access
