Abstract
Objectives:
To examine the longitudinal prevalence and recovery of olfactory, gustatory, and oral chemesthetic deficits in a sizable cohort of SARS-CoV-2 infected persons using quantitative testing. To determine whether demographic and clinical factors, mainly the medications used after the COVID-19 diagnosis, influence the test measures.
Methods:
Prospective cohort in a hospital with primary, secondary, tertiary, and quaternary care. Patients with confirmed COVID-19 were tested during the acute infection phase (within 15 days of initial symptom, n = 187) and one (n = 113) and 3 months later (n = 73). The University of Pennsylvania Smell Identification Test, the Global Gustatory Test, and a novel test for chemesthesis were administered at all visits.
Results:
During the acute phase, 93% were anosmic or microsmic and 29.4% were hypogeusic. No one was ageusic. A deficit in oral chemesthesis was present in 13.4%. By 3 months, taste and chemesthesis had largely recovered, however, some degree of olfactory dysfunction remained in 54.8%. Remarkably, patients who had been treated with anticoagulants tended to have more olfactory improvement. Recovery was greater in men than in women, but was unrelated to disease severity, smoking behavior, or the use of various medications prior to, or during, COVID-19 infection.
Conclusions:
When using quantitative testing, olfactory disturbances were found in nearly all SARS-CoV-2 infected patients during the acute infection phase. Taste or chemesthetic deficits were low. Olfactory impairment persisted to some degree in over half of the patients at the 3-month follow-up evaluation, being more common in women and less common in those who had been treated earlier with anticoagulants.
Level of Evidence:
3
Introduction
Persons with chemosensory deficits due to SARS-CoV-2 frequently turn to physicians for prognosis and treatment. Unfortunately, informative data on these points are limited. Prevalence of dysfunction based on extant psychophysical tests range, over the course of a month since symptom onset, from 50.6% to 96% for olfaction,1 -4 61.3% to 67.5% for taste.3,5 The chemesthesis is responsible for the burning, cooling, or tingling sensations of food and other stimuli and is reportedly distorted in some patients, 6 and no studies have addressed the empirical frequency of these chemesthetic deficits in COVID-19.
A number of studies suggest that most patients recover their smell or taste capacities within 10 days from the onset of the COVID-19 infection, although nearly 40% does not achieve complete recovery. 1 If the chemosensory dysfunction remains after 20 days, spontaneous resolution is believed to be less likely to occur. 3 Although treatment options for these conditions are being explored, none have shown strong efficacy.7,8 It is unknown whether the drugs used for treatment during the acute phase, or prior to the infection, are related to the timing or degree to which recovery occurs.
In this study, we examined the prevalence of smell, taste, and chemesthetic disorders in patients with confirmed COVID-19 during its acute infection phase and 1 and 3 months later. We examined whether the degree of dysfunction noted at these time periods was related to age, sex, race, education, disease severity, and smoking behavior, as well as anticoagulant, antihypertensive, and diabetic medications administered before or after the COVID-19 diagnosis.
Methods
The initial study group was comprised of 187 patients with PCR-confirmed COVID-19 enrolled between September 2020 and October 2021. Of these, 113 were retested at the 1-month follow-up period and 73 at the 3-month follow-up period. The longitudinal attrition was caused by 1 death, 72 patients who could not be re-contacted or who did not wish to further participate, and 43 patients who were unable to complete a full test session. The severity of the COVID-19 was classified as either mild (ie, nonpneumonia and mild pneumonia) or severe/critical (dyspnea, respiratory frequency ≥30/min, blood oxygen saturation ≤93%, partial pressure of arterial oxygen to fraction of inspired oxygen ratio <300, and/or lung infiltrates >50% within 24 to 48 hours, respiratory failure, septic shock, and/or multiple organ dysfunction or failure). 9 Demographic and clinical factors and all the medications used previously and after the COVID-19 diagnosis were collected.
Olfaction was tested using the Portuguese language version of the University of Pennsylvania Smell Identification Test (UPSIT®). The UPSIT® is a widely employed standardized 40-item 4-alternative forced-choice olfactory test known to correlate strongly with olfactory threshold test scores. 10 Normal olfactory function was based on normative data from the Brazilian adaptation and validation study of this test. 11
Taste was measured using the Global Gustatory Test (GGT). 12 This test is a whole-mouth modified version of the regional taste test described by Soter et al. 13 It employed aqueous solutions of 0.31 M NaCl (salty), 0.015 M citric acid (sour), 0.49 M sucrose (sweet), and 0.04 M caffeine (bitter). The stimuli were presented as oral spays, with 4 replications each presented in a pre-determined random order, that is, a total of 16 trials. Patients with more than 12 correct answers out of 16 were considered normogeusic, those with 4 to 12 were considered hypogeusic, and those with less than 4 were considered ageusic.12,13
Oral chemesthesis was tested using, for the first time, a novel carbonated water test. Carbonated water is known to activate lingual nociceptors as a result of the conversion of CO2 to carbonic acid which, in turn, activates oral trigeminal neurons (Dessirier et al., 2000). In this test, the subject sipped 50 ml of carbonated water (Água com Gás Crystal, The Coca Cola Company, São Paulo, Brasil) for 10 seconds and then either swallowed or expectorated the water into a waste container. The water bottle was opened immediately before presentation to the subject. The subject rated his or her chemesthetic sensation on a modified 10-point category scale, with 0 reflecting a completely distorted tingling sensation and 10 a completely normal tingling sensation. A score of 7 or above was classified as normal, a score between 3 and 7 as moderately distorted, and a score of 3 or less as severely altered. This test was also administered to 15 control individuals (age range: 22-61 years) without a history of COVID-19 infection. Fourteen reported a completely normal tingling sensation (score of 10) and only 1 scored 5 on the scale.
In addition to the empirical testing, subjects also rated their self-perceived taste and smell abilities on a 10-point category scale, with 0 being total lack of sensation and 10 normal sensation. Subjects were excluded from the study if they had a history of head trauma, a neuropsychiatric disorder, cerebrovascular disease, chemotherapy, radiotherapy, or self-reported loss of smell or taste that preceded COVID-19. Data from subjects who were unable to complete the testing were omitted from analysis. The Universidade Estadual de Londrina’s Ethics Committee approved the study, and all patients signed the informed consent.
Statistical Analyses
Prevalence is depicted in percentages and continuous data are reported as means and standard deviations or as indicated. Percentages of normosmics versus microsmics/anosmics among groups were compared by Fisher’s exact test. Logistic regression models were used to estimate the association between the variables and the rates of normosmia, full recovery criterium. Demographic and clinical data were assessed using a stepwise procedure based on the Akaike information criterion (AIC) for achieving the best model to predict the smell recovery. 14 P values at or below .05 were considered significant. When multiple comparisons were made, the Bonferroni correction was employed. The statistical software used was STATA 17 (StataCorp LP, College Station, TX).
Results
During the acute phase of COVID-19, 174 out of the 187 patients (93%) were found to be anosmic or microsmic and 55 (29.4%) hypogeusic. None were ageusic. Chemesthesis mouth perceptions were partially or severely altered in 25 volunteers (13.4%, Figure 1). The scores on these tests can be seen in Figure 2. At this time, 36 (19.4%) subjects reported experiencing parosmia and 29 (15.6%) phantosmia. By the 1- and 3-month follow-up periods, taste and chemesthesis functions were normal in 67 (91.8%) and 72 (98.6%) patients. However, fewer than half of the patients exhibited normal olfactory function at that time. At the 1-month follow-up, the parosmia and phantosmia values of the 113 subjects decreased to 4.8% and 6.9%, respectively. At the 3-month follow-up, similar rates of parosmia and phantosmia were noted in the 73 patients that were evaluated (7.5% and 6.7%, respectively). The most frequent qualitative sensation was described as smoke-like.
Flowchart of the study with major results.
UPSIT (A), Global Gustatory Test (B), and oral chemesthesis (C) scores for the acute phase (n = 187), 1 month (n = 113), and 3 month (n = 73) follow-up periods. The distribution of the patients’ scores in each group is depicted in a violin plot. Horizontal black lines: medians and interquartile ranges. Abbreviations: GGT, global gustatory test; UPSIT, University of Pennsylvania smell identification test.
The univariate unadjusted analyses comparing data between subjects with total olfactory recovery and subjects with partial or no recovery at the 3-month evaluation are shown in Table 1. Women had a worse total normosmia rate than men (28.2% vs 64.7%, P = .002). No differences between the 2 groups were found for COVID-19 severity, age, race, education, number of comorbidities, smoking behavior, or previous use of drugs for hypertension, diabetes, or depression. None of the patients who were normosmics during the acute phase developed olfactory dysfunction at the 3-month follow-up.
Demographic and Clinical Characteristics of Patients Normosmic or Microsmic/Anosmic in the University of Pennsylvania Smell Identification Test after 3 Months SARS-CoV-2 Infection.
An analysis of the drugs administered during the acute phase found that anticoagulant use was associated with a higher percentage of subjects recovering their smell function by the 3-month time period (70.6% vs 37.5% of those who did not use, P = .02, Table 2). No such association was found for any of the other drugs used at this time, namely corticosteroids (P = .42), azithromycin (P = .14), ivermectin (P = 1.00), albendazole (P = − .32), or hydroxychloroquine (P = .12), although very few subjects were taking the latter 2 drugs, limiting statistical power to assess such an effect.
Percentages of Tested Normal Smell Function and Unadjusted Comparisons Between Patients that Used or Did Not Use Different Medications After 3 Months of SARS-CoV-2 Infection.
In the stepwise logistic regression model that including all clinical and demographic variables, the aforementioned effect of anticoagulant therapy failed to reach statistical significance (P = .23), only being a woman was associated with a worse smell outcome (OR: 0.32, 95% CI: 0.10-0.96, P = .04). As in the previous analysis, none of the medications used after SARS-CoV-2 infection was associated with a better prognosis on the olfactory outcome (ps > 0.10). Although a few patients failed to recover completely taste or chemesthesis function, the numbers were too small to provide adequate statistical power to determine whether such variables were meaningfully associated with lack of recovery.
Figure 3 shows the subjective assessments of olfactory and taste functions. An analysis of the self-reported olfactory function during the acute phase found that only 6 of the 41 patients (14.6%) who reported having a normal sense of smell were normal upon psychophysical testing. Of those tested 1 month later, even a higher number (23 out of 34, or 67.6%) exhibited this misperception. In the third-month follow-up, this percentage was 54.5% (12 out of 22 patients).
Self-reported smell (A) and taste (B) abilities on a 10-point category scale, with 0 being total lack of sensation and 10 normal sensation for the acute phase (n = 187), 1 month (n = 113), and 3 month (n = 73) follow-up periods.
Discussion
In this study, 93% of 187 Brazilian patients tested within the 15-day acute period of the SARS-CoV-12 infection exhibited olfactory dysfunction, as measured by the UPSIT®. The prevalence of the dysfunction decreased to 77.9% and 54.8% at the 1-month and 3-month follow-up periods, respectively. These findings are in general accord with earlier studies employing the UPSIT® in COVID-19 cases, although some differences are apparent. Moein et al 15 reported a 98% of 60 Iranian patients exhibited some degree of smell dysfunction on the UPSIT® during the acute period. By extending the sample size to 100, a 96% prevalence rate was noted.1,15 The latter study retested 86 of the patients and found that the prevalence rate decreased to 39% 6 to 8 weeks after symptom onset, a value somewhat lower than the 3-month prevalence rate of the present study. González et al 16 noted a 73% prevalence rate for 100 COVID-19 Chilean patients tested with the UPSIT® during the acute phase of the disorder which decreased to 41% for those tested a month later, the latter being similar to the 6 to 8 weeks prevalence observed by Moein et al. 1 In a study of 56 Australian COVID-19 patients 6 months after their diagnosis, Leedman et al 17 found 35.6% exhibited some degree of smell loss. Boscolo-Rizzo et al 18 noted, using a modified version of the UPSIT®, that 60% of 87 Italian patients tested 6 months after the COVID-19 diagnosis still had some degree of smell dysfunction. Despite some variation, all of these studies are in agreement that the vast majority of persons, when empirically tested with a 40-odorant test, experience some measurable smell loss during the acute phase of the SARS-CoV-2 infection that continues in some patients for months after the onset of the infection.
While 19.4% of the patients reported experiencing parosmia and 15.6% phantosmia during the acute phase, the parosmia and phantosmia values of the 113 subjects decreased to 4.8% and 6.9%, respectively, at the 1-month follow-up. Similar rates of parosmia and phantosmia were noted in the 73 patients evaluated at the 3-month follow-up (7.5% and 6.7%). Although few studies are available in the literature concerning the prevalence of either parosmia or phantosmia in post-COVID-19 patients, Ercoli et al 19 reported phantosmia in 11.8%, and parosmia in 23.5%, in their 17 patients who had their acute infection several months earlier.
Our study, like most others, demonstrates the disconnect between a patient’s self-report of olfactory function and measured function. On average, studies that have used psychophysical olfactory test measures have reported a higher prevalence of smell disorders than those who have relied on self-report.4,20 Such findings stress the importance of olfactory testing and making patients and physicians aware of the limitations of self-report. Early accurate identification of smell loss could lead to earlier treatment and better olfactory outcomes. 21
In accord with most other quantitative studies, we found that only a minority of persons infected with SARS-CoV-2 exhibited abnormal taste test scores (13.4%). In a study of 111 COVID-19 patients, Niklassen et al 22 found 26% to have demonstrable taste dysfunction during the infectious period, as measured by tests in which the tastants were either sprayed into the mouth or presented to the lingual surface on strips of paper embedded with tastants. After this period was over, only 6.5% continued to exhibit taste loss. In contrast, 90% of the patients initially exhibited some degree of smell loss which declined to 27% thereafter. Cao et al 23 identified 11 of 250 healthcare workers who had experienced SARS-Cov-2 infection. All had been tested with the 27-item version of the Waterless Empirical Taste Test (WETT®), a test that also uses taste strips for stimulus presentation. While scores on a 12-item smell identification test were significantly lower in those who had a recent SARS-CoV-2 infection, this was not the case for the taste test scores that fell, on average, at the 50th percentile of those from a normative group. 23 While the smell test scores were related to the time since the onset of chemosensory systems, this was not the case for the taste test scores. Somewhat higher taste dysfunction prevalence rates were noted by Vaira et al. 24 These authors used a single-trial 4-solution taste test and reported that 47% of 72 patients exhibited either hypogeusia (n = 33) or ageusia (n = 1) early in the disease process. Petrocelli et al 25 , also using a homemade 1-trial 4-solution taste test, concluded that 38% of their sample of 300 patients exhibited ageusia.
An important observation of our study is that olfaction appeared to recover earlier in patients who had been treated with anticoagulants during the acute period of infection, although this effect was not strong and was not statistically significant at the .05 alpha level when the full model was employed. It is well established that anticoagulants can attenuate inflammation and that inflammation initially leads to coagulation. The presence of coagulant factors augment the inflammatory response, although subsequently the anticoagulant response diminishes inflammatory mediators. 26 One way that COVID-19 affects the nervous system is its impact on cerebral microvasculature as manifested by vasculitis, microangiopathy, coagulopathy, and in-situ circulatory microthrombi.27 -29 Hence, anticoagulants may lower vascular thrombi within the olfactory mucosa, notably the lamina propria, olfactory bulb regions, and possibly cerebral regions involved in olfactory transmission. 29 Although other research has demonstrated that anticoagulation has benefits for COVID-19 patients independent of olfaction,30,31 the present study is the first to provide evidence that such therapy may be of value in sparing COVID-19-related olfactory dysfunction, providing one more potential therapeutic option for the treatment of such dysfunction. More research is clearly needed on this point.
Aside from the possible beneficial effects of anticoagulant therapy, our study found no other benefit from the medications that were evaluated, including corticosteroids. This is at variance with reports that such steroids may improve olfaction in COVID-19 patients.32,33 The reason for this discrepancy is not known. Our negative results for azithromycin, chloroquine, and hydroxychloroquine are consistent with the literature.34 -37 However, the evidence of the use of these drugs on the olfactory function in COVID-19 patients is sparse in the literature and more studies are needed to investigate that relationship.
Confirming previous studies,38,39 the percentage of women having a complete recovery of their sense of smell at the third-month follow-up was lower than that of the men. Being a woman was associated with a 68% lower chance of total olfactory recovery. Interestingly, 1 study showed that women also have a more extended recovery period than men. 40
As with the taste tests, the test of oral chemesthesis we employed found smaller prevalences of these symptoms compared to olfaction (13.4% vs 93%, 8% vs 76%, and 1.4% vs 54.8%, at the acute phase, 1 and 3-month follow up, respectively). Moreover, just 1 patient had altered chemesthetic perception at the third-month follow-up. Chemesthesis is well explored in the food industry. 41 However, there are very few studies that have investigated chemesthetic distortion in clinical practice, with most simply asking about the symptoms via questionnaires. 6 The present study is the first to use a novel carbonated oral water test for COVID-19 subjects. Another COVID-19 study has measured nasal chemesthesis. They employed a 70% acetic acid solution and rated the stinging sensation inside the nose after sniffing the vapor. 42 Even after 1 year of COVID-19 infection, patients reported a reduced sensitivity at the nasal cavities with this substance compared to controls. The difference between their study and ours suggests that nasal chemesthesis may be more affected than oral chemesthesis during SARS-CoV-2 convalescence.
This study has both strengths and weaknesses. Strengths include (a) the longitudinal quantitative testing of smell, taste, and chemesthesis in a sizable number of COVID-19 patients during and 2 times following the acute phase and (b) the exploration of factors, including medications, that may be associated with faster recovery over time. Among the limitations of the study are (a) the lack of a concurrent control group and (b) the drop out of participants in the longitudinal arm of the study due to logistic and personal reasons. The latter not only decreases statistical power, but theoretically could also influence the conclusions of the study, as these patients may have declined to participate or could not be contacted because their sense of smell was normal. An additional potential limitation of the study is the use of a novel test of chemesthesis for which control data are limited. Although this test is modeled on tests found useful in the food industry, its clinical application and sensitivity to COVID-19 is limited so more validation of the test would seem useful.
Conclusions
We have verified that more than half of the COVID-19 patients persists with olfactory dysfunction after 3 months of the disease onset, this sequela being more prevalent in women. Taste and chemesthetic function returned to normal in almost all patients. No medications used during this period were associated with a better outcome on the smelling capacity, even though patients who used anticoagulants tended to have a high rate of total olfactory recovery.
Footnotes
Contributor’s Statements
Marco Aurélio Fornazieri: Prof. Fornazieri conceptualized and designed the study, conducted the data collection, conducted the analyses, drafted the initial manuscript, and revised the manuscript. José Lucas Barbosa da Silva, Juliana Gameiro, Henrique Ochoa Scussiato: coordinated data collection, conducted the analyses, drafted the initial manuscript, and revised the manuscript. Rafael Antônio Matias Ribeiro Ramos, Bruno Machado Cunha, Alan Felipe Figueiredo, Eduardo Hideki Takahashi, Gabrielli Algazal Marin, Igor Ruan de Araújo Caetano, Tainara Kawane Meli, Diego Issamu Higuchi, Rafael Rodrigues Pinheiro dos Santos, Ana Carla Mondek Rampazzo: Conducted the data collection, drafted the initial manuscript, and critically reviewed the manuscript. Fábio de Rezende Pinna and Richard Louis Voegels: Prof. Voegels and Prof. Pinna conceptualized and designed the study, and critically reviewed the manuscript. Richard L. Doty: Professor Doty provided statistical guidance and critically revised and reviewed the manuscript. All authors approved the final manuscript as submitted.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: RLD is a consultant to Eisai Co, Ltd, Merck Pharmaceuticals, the Michael J. Fox Foundation for Parkinson’s Research, and Johnson & Johnson. He receives royalties from Cambridge University Press, Johns Hopkins University Press, and John Wiley & Sons, Inc. He is president of, and a major shareholder in, Sensonics International, a manufacturer and distributor of smell and taste tests, including the test used in this study. No other authors have disclosures.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Programa de Pesquisa para o Sistema Único de Saúde (PPSUS)
