The strategies aimed at a complete search, including descriptors (Medical Subject Headings - Mesh) and free terms, which consisted of: COVID-19, 2019-nCoV, 2019 novel coronavirus pneumonia, SARS-CoV, SARS-CoV-2, sars coronavirus, coronavirus, coronavirus infection, coronavirus COVID-19, novel coronavirus), COVID-19 clinical features, olfactory disorders, olfaction disorders and taste disorders. To identify the relevant articles for the proposed question, a search strategy was developed: (COVID-19 OR 2019-nCoV OR 2019 novel coronavirus pneumonia OR SARS-CoV OR SARS-CoV-2 OR sars coronavirus OR coronavirus OR coronavirus infection OR coronavirus COVID-19 OR novel coronavirus) AND (COVID 19 clinical features OR olfactory disorders OR olfaction disorders OR taste disorders); this strategy was adapted to the databases, and the corresponding terms were applied in Portuguese and Spanish at Lilacs database. There was no restriction on the language of publication. The complete search strategy is shown in the supplementary material (Appendix 1).

Between April 2 and 25, 2020, the following electronic databases were searched: Medline via PubMed, Cochrane Clinical Trials, ScienceDirect, Lilacs, Scopus and Google Schoolar, as well as the gray literature OpenGrey.eu, DissOnline, The New York Academy of Medicine and Reasearch Gate. To prevent citation bias, there was no manual search for the included articles and specialists in the field were not contacted.9

After the search, the references of each database were exported to the Mendeley® program (https://www.mendeley.com/) aiming to identify all articles in duplicate, promote greater selection reliability and proceed to the eligibility criteria step.

The following inclusion criteria were considered: (1) Observational studies with individuals diagnosed with coronavirus infection and (2) recording of the signs and symptoms of COVID-19, with clinical aspects related to olfactory and/or gustatory function. The following exclusion criteria were considered: (1) Studies on non-human coronavirus; (2) Review articles; (3) Experimental studies (in animals or in vitro); (4) Clinical report of OD/GD with onset prior to the SARS-CoV-2 infection.

The search for the articles was carried out by two reviewers. The titles and abstracts of the retrieved articles were independently assessed by two researchers who were not blinded to the authors or the titles of journals. The full versions of the potentially eligible articles were retrieved for further evaluation and were carried out in three stages:

• 1.

  Identification and reading of titles in different electronic databases. Articles that clearly did not meet any of the inclusion or exclusion criteria were excluded.

• 2.

  Reading of the abstracts of the studies selected in the first phase. Likewise, articles that clearly did not meet any of the inclusion or exclusion criteria were excluded.

• 3.

  All studies that were not excluded in these first two stages were read in full for the selection of the ones that would be included in this Systematic Review (SR).

The main data for each article was fully collected and inserted into a database created in the Microsoft Office Excel 2010® software. There were no disagreements between the evaluators. A standard form for data storage was created based on the model adopted by Cochrane.10 For a better presentation of the results, it was decided to consider the following variables of the selected articles: Title, author, study location, study design, sample size (N), patients’ mean age and ethnicity, both the diagnostic method and the most prevalent general signs/symptoms and the severity of the clinical picture of COVID-19, methods used in the olfactory-gustatory evaluation and the prevalence of OD and GD. The outcomes sought in the studies were both the demographic data, prevalence and clinical expressions of OD/GD, as well as possible specific treatments for such disorders in patients diagnosed with COVID-19.

The risk of bias was assessed according to the recommendations in the manual and the Newcastle-Ottawa scale adapted for cross-sectional observational studies.11 Study quality was independently assessed by two researchers and the divergences were evaluated by consensus. The maximum score to be achieved was 10 points and the items evaluated on the scale were: (1) Representativeness of the sample; (2) Sample size; (3) Characterization of the diagnosis; (4) Risk exposure assessment (risk factor); (5) Comparability, to investigate whether individuals in different groups of results are comparable, based on the study design or analysis, control of confounding factors; (6) Evaluation of results and (7) Statistical test (Table 1).

After these evaluations, the selected studies were submitted to a statistical analysis to verify the possibility of building a meta-analysis. This analysis combines and summarizes the results of several studies, thus increasing the accuracy and power of evidence of the results.

In the first phase of this SR, 1788 articles were found in six databases and none in the gray literature. After eliminating 1023 studies in duplicate, 765 were selected for the reading of titles and abstracts. A total of 749 were excluded according to the established selection criteria (62 studies on non-human coronavirus, 301 due to lack of data on clinical aspects, 375 literature reviews and 11 experimental studies in vitro) and 16 articles were selected to be read in full. After the reading, 10 articles were excluded as they did not have data on smell/taste (Table 2). Therefore, six full articles were included in the qualitative analysis (Table 3). The entire article selection process is described in Fig. 1, which shows the Prisma flow diagram for inclusion.

Figure 1.

Diagram of identification and selection flow.

(0.27MB).

The analysis of the quality of the included articles and, consequently, of the risk of bias, is shown in Table 4. All included studies are characterized as observational and cross-sectional studies. Moreover, in the final evaluation, all obtained a percentage of quality equal to or greater than 80% (8/10), and four of the six articles obtained the maximum score of 90% (9/10).

Only one of the studies carried out a census study, the other studies did not perform a random selection of patients, which compromises the representativeness of the samples, but took into account the sample size.

All studies were concerned with the participants’ selection, defining the inclusion and exclusion criteria, in order to exclude factors that could influence the results found.

The size of the studied samples was adequate to the central limit theorem for all studies, and only two studies had samples smaller than 300 subjects.

The characterization of the diagnosis and the determination of exposure were very well described in all studies. All studies controlled at least one variable, allowing an efficient comparison among all studies.

The evaluation of the results was carried out in all studies through a specific report and the methods were described and validated in all studies.

Finally, all studies showed appropriate statistical tests, with description of the statistical tests, the levels of significance and the applications used.

It was not possible to perform a meta-analysis in this SR because the articles included had very different methodologies and measures resulting from the tests.

Regarding the overall characteristics of the included articles (Table 3), the six studies were published in 2020, one in China,12 one in Europe,13 one in Iran,14 two in the United States15,16 and one in the United Kingdom.17 The Chinese study (Mao et al.)12 retrospectively evaluated 214 patients from three hospitals in Wuhan; the European study (Lechien et al.)13 cross-sectionally evaluated 417 patients in a multicentric model; the Iranian study (Moein et al.)14 studied 60 patients with COVID-19 and 60 subjects in the control group, matched for age and gender; the two North-American studies were carried out by the same group of researchers (Yan et al.)15,16; while the British study (Menni et al.)17 evaluated, through a census study, 579 COVID-19 positive and 1123 COVID-19 negative patients. The demographic data, general symptoms and the prevalence of OD and GD in patients in the six studies are depicted in Table 5.

In total, the six studies evaluated 1457 patients (Table 3). In the study by Mao et al.,12 the sample consisted only of Chinese patients; in the study by Lechien et al.,13 most of the sample consisted of European patients; in the study by Moein et al.,14 all patients were Iranians; in the two studies by Yann et al.,15,16 most patients were North-American; and in the study by Menni et al.,17 the ethnicity of the study participants was not recorded.

The exam used for diagnosing COVID-19 in the six studies was the RT-PCR (Real-Time Reverse-Transcription Polymerase Chain Reaction). Regarding the most prevalent overall signs/symptoms of COVID-19, there were differences between groups; fever was the most prevalent sign in the studies by Mao et al.12 and by Moein et al.14; fatigue was the first symptom in the first study by Yan et al.15 and of Menni et al.17; coughing was the first sign in the second study by Yan et al.16 and in the study by Lechien et al.13 (Table 5).

Mao et al.12 carried out the study at three hospitals designated for the care of patients with COVID-19, which belonged to the Union Hospital of Huazhong University of Science and Technology in Wuhan, China. Clinical data were extracted from electronic medical records and reviewed by a trained medical team. Neurological symptoms fall into three categories: skeletal muscle symptoms, symptoms or diseases of the Central Nervous System (CNS) and Peripheral Nervous System (PNS) symptoms, including OD and GD. Data on all neurological symptoms were verified by two trained neurologists. Of the 214 patients studied, 126 (58.9%) had mild-moderate disease, 127 (59.3%) were females and the mean age of the group was 52.7±15.5 years. Compared to the patients with mild-moderate disease, critically-ill patients were older (58.7±15.0 vs. 48.9±14.7 years) and showed fewer typical symptoms, such as fever in 40 (45.5%) vs. 92 (73%) and coughing in 30 (34.1%) vs. 77 (61.1%). GD was identified in 12 (5.6%), and OD in 11 (5.1%) patients; regarding these disorders, there were only references to hypogeusia and hyposmia and they were the main symptoms reported by patients with PNS alterations, with 12/19 (63%), 11/19 (57.8%); respectively. There were no records regarding the onset of olfactory/gustatory disorders, duration, resolution or specific treatments.

Lechien et al.13 carried out the multicentric European study through the COVID-19 Task Force of the Young-Otolaryngologists of the International Federations of Oto-rhino-laryngological Societies (YO-IFOS) in six Belgian, three Italian and two French hospitals. Patients with olfactory-gustatory disorders of which onset occurred prior to the pandemic, patients without laboratory diagnosis of COVID-19, patients who were in the intensive care unit at the time of the study (due to their health status) were excluded. Therefore, the sample consisted only of patients with mild to moderate disease. Clinical data were collected prospectively during the otorhinolaryngological consultation or by telephone for patients undergoing home treatment. In this study, patients filled out a questionnaire based on “The smell and taste component of the National Health and Nutrition Examination Survey” and the simplified version of the “Questionnaire of Olfactory Disorders-Negative Statements” (sQOD-NS).18 Of the 417 patients, 389 (93.3%) were European, 263 (63.1%) were females, all with mild to moderate pictures of COVID-19 and with a mean age of 36.9±11.4 years. In the evaluated sample, 85.6% and 88.8% of the patients reported OD and GD, respectively; there was a significant association between the two disorders (p<0.001). Of the 357 patients with OD, 284 (79.6%) had anosmia, 73 (20.4%) had hyposmia, 45 (12.6%) had phantosmia and 115 (32.4%) had parosmia. OD appeared before, during or after the other signs/symptoms of COVID-19 in 11.8%, 22.8% and 65.4% of cases, respectively. Among patients that had already recovered from COVID-19, 63% of them remained with OD. Among the patients who recovered their sense of smell, 33%, 39.6%, 24.2% and 3.3% did so in the periods of 1–4, 5–8, 9–14 and more than 15 days of resolution of the COVID-19 picture, respectively. In anosmia cases, olfactory function normalized over the first eight days after disease resolution in 67.8% of cases. An association was observed between fever and anosmia (p=0.014). Women were significantly more affected by OD (with both hyposmia and anosmia) and GD than men (p=0.001). Regarding the treatment for OD, saline nasal lavage, nasal corticosteroids and oral corticosteroids were used by 16.7%, 8.1% and 2.5% of the patients, respectively; GDs were treated in 1.4% of cases with l-carnitine or trace elements associated with vitamins. No results were observed regarding the effectiveness of such specific treatments for OD and GD.

Moein et al.14 carried out the study with 60 SARS-CoV-2 positive patients who were admitted to Masih Daneshvari Hospital in Tehran, Iran. At the time of the olfactory test, everyone was hospitalized during the recovery period of the disease. A control group was created with 60 healthy subjects, paired by gender and age with the 60 patients with the infection. The Persian version of the University of Pennsylvania Smell Identification Test (UPSIT), with 40 odors, was used in this study (Sensonics International, Haddon Hts., NJ, USA) with the help of a trained examiner; the test provides an absolute dysfunction index (anosmia, severe hyposmia, moderate hyposmia, mild hyposmia, normosmia, faking), as well as the relative dysfunction based on normative percentile classifications adjusted for age and gender. Of the 60 patients studied with COVID-19, 25 (42%) had mild, 29 (48%) had moderate and 6 (10%) had severe disease; 40 (66.6%) were males and the mean age of the group was 46.5±12.1 years. Of the 60 patients, OD was identified in 59 (98.3%) and GD in 14 (24%), in contrast to the control group, that showed OD in 18% of the individuals; of the 60 patients, 15 (25%) had anosmia, 20 (33%) had severe hyposmia, 16 (27%) had moderate hyposmia, 8 (13%) had mild hyposmia, and 1 (0.2%) was normal. The 59 patients with OD reported that the disorder started at the same time or immediately after the onset of the signs/symptoms of COVID-19. There were no detailed data on the type of gustatory disorder or specific treatments for OD/GD in the studied sample.

In their first study, Yan et al.15 evaluated 58 patients at the University of California San Diego Health hospital with positive laboratory tests for COVID-19 and 203 negative tests. The ODs and GDs were assessed with the help of a questionnaire using a platform (Qualtrics®). The questions were based on a 10-point continuous bar for the olfactory function, meaning: 0 (anosmia), 10 (normal smell), and a 10-point bar for the gustatory function, meaning: 0 (ageusia), 10 (normal gustatory function); the lower the score, the greater the severity of OD and GD. Of the 58 studied patients with COVID-19, 54 (93.1%) had mild disease, 29 (50%) were males and 38 (64.4%) were under 50 years of age. Loss of smell and taste was reported in 40 (68%) and 42 (71%) of COVID-19 positive individuals, respectively. In 12 (22%) patients, OD was already present at the beginning of the condition. The smell and taste function impairment was independently and strongly associated with COVID-19 positivity (Anosmia: Adjusted Odds Ratio – Aor=10.9; 95%CI: 5.08–23.5; Ageusia: Aor=10.2; 95%CI: 4.74–22.1). Compared to other symptoms of COVID-19 infection, loss of smell and taste showed the greatest magnitudes of association with COVID-19 positivity (Anosmia: OR=10.9; 95%CI: 5.6–21, 0; Ageusia: OR=11.9; 95%CI: 6.1–23.2). Of the patients who reported OD, 43 (73.6%) reported anosmia resolution at different intervals after recovery from COVID-19 (18.4% in the first week; 36.8% in the second week, and 18.4% in the third or fourth week); 15 (26.4%) of the cases did not show olfaction recovery. No results were observed regarding the resolution of the GD condition or specific treatments for both disorders.

In their second study, Yan et al.16 retrospectively evaluated 128 patients with a positive laboratory test for COVID-19 recorded in the San Diego Hospital system. Smell and taste were self-assessed in relation to the period of the disease and were compared to the levels before the disease (loss of smell/taste vs. normal function); at first, these data were extracted from patients’ medical records; when absent, the interview was performed via email or telephone. Of the 128 patients with COVID-19, 102 (79.7%) had mild disease, 67 (52.3%) were females and the mean age was 48.25±16.75 years. Of the 128 patients, 75 (58.6%) had OD and 70 (54.6%) had GD. Patients who reported loss of smell were ten times less likely to be hospitalized due to COVID-19 (OR=0.09; 95%CI: 0.01–0.74) when compared to those without loss of smell. ODs were not associated with any other measure typically related to the decision to admit the patient, suggesting that loss of smell is truly a factor that can be used as a marker for milder manifestations of COVID-19. Thus, the authors suggest that OD may be a clinical marker to help stratify disease severity at the beginning of SARS-CoV-2 infection. There was no detailed data on the type of GD present or specific treatments for any of the disorders in the studied sample.

Menni et al.17 performed a census study using the RADAR COVID-19 application with 579 patients who tested positive for SARS-CoV-2 in the UK and 1123 negative ones. Of the 579 patients, all had mild disease, 400 (69%) were females and the mean age was 40.79±11.84 years. OD and GD were present in 341 (59%) patients positive for COVID-19, compared to 202 (18%) that tested negative for the disease, resulting in an odds ratio (OR) of COVID-19 diagnosis (95%CI=6.59 [5.25; 8.27], p=1.90×10–59). The combination of loss of smell and taste, fever, persistent cough, fatigue, diarrhea, abdominal pain and loss of appetite was predictive for a positive COVID-19 test, with a sensitivity of 0.54 (0.44; 0.63); specificity of 0.86 (0.80; 0.90): ROC-AUC=0.77 (0.72; 0.82) in the set of tests and ROC-AUC cross-validation=0.75 (0.72; 0.77). Overall, without adjustments, loss of smell and taste had a positive predictive value of 61.7%. The results suggest that loss of taste/smell is a strong predictor of SARS-CoV-2 infection. There were no detailed data on the subtype of OD and GD, onset and resolution in relation to the COVID-19 severity, or specific treatments for any of the disorders.