Abstract
This prospective study assesses the feasibility of a novel, objective olfactory test as part of an initial screening for COVID-19 in adults with unknown disease status.
Olfactory dysfunction (OD) is one of the earliest and strongest predictors of COVID-19 infection, and thus is promising as a disease screening tool.1,2 Compared with objective testing, subjective olfactory assessments significantly underreport OD.3,4 Thus, an inexpensive, quick, and sensitive method of assessing olfaction may be beneficial for the early diagnosis and spread prevention of COVID-19. In this study, we evaluate the feasibility of a novel, objective olfactory test as part of an initial screening for COVID-19 in adults with unknown disease status.
Methods
This prospective, cross-sectional study enrolled healthy adults (aged ≥ 18 years) from a single college campus COVID-19 screening site. The study was approved by the Southern Methodist University institutional review board and all participants provided written informed consent. Participants were screened for OD using a novel scent card (SAFER Diagnostics) followed immediately by polymerase chain reaction (PCR) testing for SARS-CoV-2 from nasopharyngeal swabs. The SAFER card contained a single scent in a scratch-and-sniff label that the participant identified from 8 answer options: lemon, grape, floral, blueberry, banana, mint, unsure, or no scent. Answers were processed electronically via a QR code. An incorrect choice was classified as OD. Participant demographics, medical history, COVID-19 symptoms, and subjective smell function on a binary and 10-point visual analog scale (VAS, with 0 indicating no sense of smell and 10 indicating normal sense of smell) were also collected.
Multivariable logistic regression was conducted to assess the association between SAFER and COVID-19 PCR results while controlling for variables selected a priori as common symptoms of COVID-19: fever, fatigue, and cough.5 Data analysis was conducted using SPSS version 27.0 (IBM Corp).
Results
A total of 163 participants were prospectively screened for OD using the scent card followed by SARS-CoV-2 PCR testing with characteristics summarized in Table 1. Of those who tested PCR-positive for COVID-19, 75% (12 out of 16) failed olfactory screening compared with 4.8% (7 out of 147) among those testing PCR-negative for COVID-19. The sensitivity, specificity, positive predictive value, and negative predictive value of the scent card in detecting those with COVID-19 were 75.0%, 95.2%, 63.2%, and 97.2%, respectively (Table 2). Including the symptom fatigue along with OD achieved a 93.8% sensitivity and 89.8% specificity in disease screening. The addition of fever and cough did not further increase sensitivity (Table 2).
Table 1. Univariate and Multivariate Analysis of the Correlation Between Demographic Characteristics, Medical History, COVID-19 Symptoms, and SAFER Smell Card and COVID-19 PCR Results.
| Variable | COVID-19 PCR, No. (%) | Analysis, OR (95% CI) | ||
|---|---|---|---|---|
| Positive | Negative | Univariate | Multivariate | |
| No. | 16 | 147 | ||
| Failed smell screen | 12 (75.0) | 7 (4.8) | 60 (15.36-234.38) | 80.24 (14.77-435.90) |
| Age, mean (SD) | 31.69 (14.05) | 24.31 (7.76) | NA | NA |
| Gender | ||||
| Male | 9 (56.3) | 109 (74.1) | 2.23 (0.78-6.40) | NA |
| Female | 7 (43.7) | 38 (25.9) | NA | NA |
| Race/Ethnicity | ||||
| Black | 3 (18.8) | 36 (24.5) | NA | NA |
| White | 10 (62.5) | 93 (63.3) | NA | NA |
| Asian/Pacific Islander | 1 (6.3) | 6 (4.1) | NA | NA |
| Hispanic | 2 (12.5) | 9 (6.1) | NA | NA |
| Other | 0 | 3 (2) | NA | NA |
| Tobacco status | ||||
| Smoker | 0 | 7 (4.8) | NA | NA |
| Nonsmoker | 16 (100.0) | 140 (95.2) | NA | NA |
| COVID-19 exposurea | ||||
| Yes | 13 (81.3) | 25 (17.0) | 21.15 (5.61-79.72) | NA |
| No | 3 (18.7) | 122 (83.0) | NA | NA |
| History of anosmia | ||||
| Yes | 0 | 5 (3.4) | NA | NA |
| No | 16 (100.0) | 142 (96.6) | NA | NA |
| Past medical history | ||||
| None | 15 (93.8) | 138 (93.9) | 0.98 (0.12-8.26) | NA |
| Lung disease | 0 | 0 | NA | NA |
| Heart disease | 0 | 0 | NA | NA |
| Hypertension | 0 | 5 (3.4) | NA | NA |
| Hyperlipidemia | 0 | 1 (0.7) | NA | NA |
| Liver disease | 1 (6.3) | 1 (0.7) | 9.73 (0.58-163.67) | NA |
| Diabetes | 0 | 0 | NA | NA |
| Autoimmune disease | 0 | 0 | NA | NA |
| Cancer | 1 (6.3) | 0 | NA | NA |
| Kidney disease | 0 | 0 | NA | NA |
| Neurologic disorder | 0 | 0 | NA | NA |
| Sinusitis | 0 | 3 (2.0) | NA | NA |
| Symptoms | ||||
| Cough | 5 (31.3) | 8 (5.4) | 7.90 (2.21-28.26) | 3.76 (0.48-29.51) |
| Fever | 3 (18.8) | 3 (2) | 11.08 (2.03-60.51) | 4.78 (0.37-61.06) |
| Fatigue | 6 (37.5) | 9 (6.1) | 9.20 (2.73-31.04) | 6.89 (0.70-67.59) |
| Dyspnea | 1 (6.3) | 5 (3.4) | 1.89 (0.21-17.29) | NA |
| Diarrhea | 0 | 2 (1.4) | NA | NA |
| Headache | 7 (43.8) | 13 (8.8) | 8.02 (2.56-25.07) | NA |
| Nasal congestion | 5 (31.3) | 14 (9.5) | 4.32 (1.31-14.22) | NA |
| Rhinorrhea | 5 (31.3) | 11 (7.5) | 5.62 (1.66-19.09) | NA |
| Sore throat | 6 (37.5) | 11 (7.5) | 7.42 (2.27-24.24) | NA |
| Myalgia | 2 (12.5) | 3 (2) | 6.86 (1.06-44.56) | NA |
| Nausea/vomiting | 1 (6.3) | 2 (1.4) | 4.83 (0.41-56.50) | NA |
| Anosmia | 6 (37.5) | 1 (0.7) | 87.60 (9.59-800.0) | NA |
| Loss of taste | 3 (18.8) | 0 | NA | NA |
| Olfaction on visual analog scale, mean (SD) | 5.75 (3.32) | 9.10 (1.18) | NA | NA |
Abbreviations: NA, not applicable; OR, odds ratio; PCR, polymerase chain reaction.
COVID-19 exposure was defined as the ability to identify interacting with another individual with known COVID-19 infection.
Table 2. Screening Efficacy of an Olfactory Dysfunction Screen Using a Novel Scent Card and Other COVID-19 Symptoms Compared With COVID-19 PCR Results.
| Characteristic | COVID-19 PCR, % (95% CI) | Total | |
|---|---|---|---|
| Positive | Negative | ||
| Smell screen | |||
| Fail, No. | 12 | 7 | 19 |
| Pass, No. | 4 | 140 | 144 |
| Total, No. | 16 | 147 | 163 |
| Sensitivity | 75.0 (53.1-96.9) | NA | |
| Specificity | 95.2 (91.7-98.7) | NA | |
| Predictive value | |||
| Positive | 63.2 (40.9-85.5) | NA | |
| Negative | 97.2 (94.5-99.9) | NA | |
| Smell screen + fatigue | |||
| Fail | 15 | 15 | 30 |
| Pass | 1 | 132 | 133 |
| Total | 16 | 147 | 163 |
| Sensitivity | 93.8 (81.6-100) | NA | |
| Specificity | 89.8 (84.9-94.7) | NA | |
| Predictive value | |||
| Positive | 50.0 (31.8-68.2) | NA | |
| Negative | 99.2 (97.7-100) | NA | |
| Smell screen + cough | |||
| Fail | 13 | 15 | 28 |
| Pass | 3 | 132 | 135 |
| Total | 16 | 147 | 163 |
| Sensitivity | 81.3 (61.6-100) | NA | |
| Specificity | 89.8 (84.9-94.7) | NA | |
| Predictive value | |||
| Positive | 46.4 (27.6-65.2) | NA | |
| Negative | 97.8 (95.3-100) | NA | |
| Smell screen + fever | |||
| Fail | 14 | 9 | 23 |
| Pass | 2 | 138 | 140 |
| Total | 16 | 147 | 163 |
| Sensitivity | 87.5 (70.7-100) | NA | |
| Specificity | 93.9 (90.0-97.8) | NA | |
| Predictive value | |||
| Positive | 60.9 (40.5-81.3) | NA | |
| Negative | 98.5 (96.5-100) | NA | |
| Smell screen + fatigue, cough, fever | |||
| Fail | 15 | 21 | 36 |
| Pass | 1 | 126 | 127 |
| Total | 16 | 147 | 163 |
| Sensitivity | 93.8 (81.6-100) | NA | |
| Specificity | 85.7 (80.0-91.4) | NA | |
| Predictive value | |||
| Positive | 41.7 (25.4-58.0) | NA | |
| Negative | 99.2 (97.7-100) | NA | |
Abbreviation: NA, not applicable; PCR, polymerase chain reaction.
While only 37.5% (6 out of 16) of COVID-19 test-positive participants reported subjective anosmia, 75% failed screening with the scent card. A failed scent card screen was the greatest predictor of COVID-19 positivity (odds ratio [OR], 80.24; 95% CI, 14.77-435.90) when compared with other symptoms including cough, fever, fatigue, and a history of COVID-19 exposure.
Discussion
In this study, we demonstrate that a rapid psychophysical olfaction test is feasible as a screening tool for COVID-19. Current literature suggests that individuals may fail to recognize OD that is detectable on objective testing.3,4 Furthermore, screening questionnaires may miss more than 50% of COVID-19 cases,6 owing to screening question variation, subjective interpretation of symptom severity, and intentional patient evasion. Meanwhile, antigen-based point-of-care screening remains time intensive, expensive, and unrealistic in certain environments.
In this pilot study, a novel olfactory test alone had a sensitivity of 75% and specificity of 95.2% in detecting COVID-19 using PCR testing as the gold standard. A screening test ideally prioritizes high sensitivity with potential lower specificity. Olfactory screening using a single scent may demonstrate lower test sensitivity in exchange for expediency. The optimal number of screening odorants required to achieve high sensitivity while maintaining practicality may be a topic for future studies. Furthermore, not all patients with COVID-19 experience OD. The combined screening of OD with other common COVID-19 symptoms, in this case fatigue, can help increase sensitivity. Alone, OD screening remained the greatest predictor of COVID-19 positivity compared with other symptoms.
This pilot study is limited by its small sample size, small percentage of COVID-19 PCR test-positive participants, and single-institution recruitment. Future studies with a larger sample size and a heterogeneous population may better account for other OD risk factors and optimize sensitivity using a combination of OD testing with other symptoms for screening COVID-19.
References
- 1.Yan CH, Faraji F, Prajapati DP, Boone CE, DeConde AS. Association of chemosensory dysfunction and COVID-19 in patients presenting with influenza-like symptoms. Int Forum Allergy Rhinol. 2020;10(7):806-813. doi: 10.1002/alr.22579 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Gerkin RC, Ohla K, Veldhuizen MG, et al. ; GCCR Group Author . Recent smell loss is the best predictor of COVID-19 among individuals with recent respiratory symptoms. Chem Senses. 2021;46:bjaa081. doi: 10.1093/chemse/bjaa081 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Moein ST, Hashemian SM, Mansourafshar B, Khorram-Tousi A, Tabarsi P, Doty RL. Smell dysfunction: a biomarker for COVID-19. Int Forum Allergy Rhinol. 2020;10(8):944-950. doi: 10.1002/alr.22587 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hoffman HJ, Rawal S, Li C-M, Duffy VB. New chemosensory component in the U.S. National Health and Nutrition Examination Survey (NHANES): first-year results for measured olfactory dysfunction. Rev Endocr Metab Disord. 2016;17(2):221-240. doi: 10.1007/s11154-016-9364-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhu J, Ji P, Pang J, et al. Clinical characteristics of 3062 COVID-19 patients: a meta-analysis. J Med Virol. 2020;92(10):1902-1914. doi: 10.1002/jmv.25884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gostic K, Gomez AC, Mummah RO, Kucharski AJ, Lloyd-Smith JO. Estimated effectiveness of symptom and risk screening to prevent the spread of COVID-19. Elife. 2020;9:e55570. doi: 10.7554/eLife.55570 [DOI] [PMC free article] [PubMed] [Google Scholar]