To find a correlation between the Visual Analogue Scale (VAS) and olfactory test in screening of smell disorders.
Observational study, cross-sectional, prospective for a period of 6 months. Evaluation of smell was conducted in 2 phases: self-evaluation of smell using VAS and evaluation of smell by the investigator using the olfactory test. This test was performed with 5 odors contained in 8 capsules.
The study sample included 1152 patients with 614 females (53.30%) and 538 males (46.70%), the sex ratio was 0.8. The study on 60 subjects, reported anosmic to VAS, 70% were actually anosmic after the olfactory test against 28.3% of hyposmic. Olfactory disorders determined by VAS were 32.5% while the actual prevalence determined by the olfactory test was 20.3%. There was a statistical significant connection between the olfactory perception and VAS (p <0.0001); the concordance between VAS and olfactory test is 94.5%, that indicates a correlation between the VAS and the olfactory test. There was also a statistical significant association between the odor discrimination and VAS (p <0.0001). VAS performance was determined by a sensitivity of 82.05% and a specificity of 80.17%.
There is a correlation between VAS and olfactory test; VAS can be used for screening the olfactory disorders.
Keywords: Correlation; Olfactory Disorders; Visual Analogue Scale; Olfactory Test
Olfactory sense (smell) has a chemical sense, based on this, a progress of series events follows a contact with certain known
odorant molecules with specific receptors localized in the olfactory neuroepithelium [1,2].
Smell is essential in both animals and humans life. In fact, animals have the power to recognize and discriminate a large number of chemical signals in their environment. This plays a major role in their behavior, their ability to survive in a hostile environment, but also in managing their emotions, reproduction, social life (recognition of the clan, the family) and some physiological regulations .
In humans, the inability to detect household odors significant ly limits their social activity .
Clinically, the study of smell disorders is fraught with difficulties due to the lack of physical parameter of the stimulus evolving continuously. In addition, screening and clinical exploration of smell in clinical practice requires the use of a rarely available olfactory test, particularly in Ivory Coast. Indeed, in our practice, only one box is available for clinical exploration of smell. Due to the unavailability of the olfactory test, we asked ourselves whether the visual analog scale (VAS) could not be an effective alternative and valid method in clinical evaluation of smell. The correlation research between self-assessment and measurement of olfactory function by different methods has been the subject of studies [5-9]. However, to our knowledge, it has never been the subject of study in smell disorders screening. Hence, we have undertaken this study whose objectives were to look for a correlation between VAS and olfactory test, and evaluate the VAS performance in smell disorders detecting.
Patients and Methods
Type of study
This was an observational study, cross-sectional, prospective held in schools and religious structures of Abidjan district.
We included in the study at least 15-years old subjects without any neuropsychiatric disorder. Patients were informed and gave their consent for participation in the study.
Were excluded from the study, subjects that could not carry out the smell assessment (patients with a stoma or tracheotomy).
Conduct of inquiry
The survey was conducted in several sessions by appointment in structures or institutions with strong influence (church, school). It was carried out by a team of investigators trained for this purpose after host structure agreement.
The survey data (interview and evaluation of smell) were collected on an anonymous questionnaire by a team of 3 to 5 investigators per session.
Evaluation of smell was conducted in two successive stages: self-evaluation of smell and evaluation of smell through the olfactory test.
Self-evaluation of smell by the patient
Self-evaluation of olfactory status was performed by the subject, detected by visual analogue scale (VAS) and graded from 0 to 10 according to the criteria of EP3OS.
olfactory test procedure
The smell test was performed by the investigator to assess perception, discrimination, and recognition of different odors.
This test comprises of a box (Figure 1) containing 8 capsules (GlaxoSmithKline). Five (5) different odors were used: banana, vanilla, mint, lavender and lemon. The capsules numbered 1, 2, 3, and 4 each contained the banana smell in increasing concentrations (0.1% for the capsule 1, 1% for the capsule 2, 10% for the capsule 3 and 20% for the capsule 4. Capsules numbered 11, 12, 13, and 14 respectively contain the vanilla, mint, lavender and lemon odors with single concentration. The capsules were opened and presented to a patient one by one and a time of olfactory rest of about 1 minute was allowed after each presentation. The capsules numbered from 1 to 4 were first presented to the patient who indicated whether or not he felt an odor (perception) and if so, which description best met his perception from the list that was presented to him. The same procedure was followed for capsules 11 to 14. He indicated if or not felt a smell (perception) and if so, he had to clarify whether these odors were identical or different (discrimination). The last step of the test was to ask the patient to name or identify odors of capsules numbered from 11 to 14 (recognition).
The VAS was divided by three to get an average score. Thus, a VAS score of 0 corresponded to anosmia; a VAS score between 0 and 7 was hyposmia while a VAS score greater than 7 determined normosmia.
Olfaction was considered clinically normal when the patient perceived the smell of the capsule having the lowest concentration (capsule1). The patient was found hyposmic when he perceived the odor of capsules having the highest concentrations (capsules 2, 3 and 4). When he perceived no smells of any capsules, he was considered anosmic.
Odor’s discrimination was made between 4 capsules numbered from 11 to 14. It was classified normal when the patient could differentiate at least 2 of the 4 different odors. The absence of discrimination was judged by inability to differentiate between any of the 4 odors presented.
The recognition of odors was appreciated from these same capsules numbered 11 to 14. It was classified good if the patient could name or identify a smell of one of these capsules. The lack of recognition was judged by inability to identify any of the four odors presented. VAS performance was evaluated by calculating the sensitivity and specificity, the olfactory test being the reference diagnostic test.
Statistical analyses were performed by means of SAS version 9.1 software (SAS Institute, Cary, NC). Quantitative variables were expressed as mean ± standard deviation. Variables were also expressed as frequencies and percentages. Relationships between self-assessment of smell disorders by VAS and olfactory perception by olfactory test were analyzed using logistic regression. An appropriateness test of the logistic model has been performed by the Hosmer and Lemeshow method to assess the logistic model validity. A p < 0.05 was considered statistically significant for all tests.
Epidemiological characteristics of the population
During the study period, 1152 subjects were the subject of exploration of smell. This population consisted of 614 female patients (53.30%) and 538 male (46.70%), with a sex ratio of 0.8. The mean age of patients was 35.18 ± 14.67 years; extremes were 15 years and 77 years.
Self-assessment of smell by VAS
The average score of the VAS was 7 with a standard deviation of 2.8 and a median score of 8. In our study population, 60 patients (5.2%) reported being anosmic (VAS score of 0) and 314 subjects (27.3%) considered themselves hyposmic (score of VAS between 0 and 7). Those with olfactory disorders using the VAS represented 32.5% of subjects tested. A proportion of 67.5% of subjects found their olfaction normal (VAS score between 7 and 10) (Table I)
Table 1. Distribution of subjects by visual analog scale (VAS) score.
Evaluation of smell through the olfactory test
The prevalence of olfactory disorders was 20.3% of the test subjects. The proportion of anosmic patients was 7.5% against 12.8% for hyposmic subjects (Table II).
Table II. Distribution of subjects by odors perception.
Approximately 92% of subjects were classified as normal for discrimination of odors (Figure 2A). The recognition of odors was good in 86.5% of subjects (Figure 2B).
Figure 1. Box of 8 capsules containing 5 odors and used for the olfactory test.
Figure 2. Discrimination and recognition of odors
A: Odor discrimination
B: Odor recognition
Relationship between VAS and olfactory test
Using the VAS, 70% (42 of 60 subjects) of subjects that considered they were anosmic were really anosmic after completing the olfactory test, while 28.3% were hyposmic. In addition, 94.6% of the subjects considered as normosmic according to the VAS assessment were really normosmic after the olfactory test (Table III).
Table III. Distribution of olfactory perception by VAS score.
The percentage of concordance was 94.5% and the discordance percentage was 4.9%. There was a statistically significant relation between the olfactory perception determined by the olfactory test and the VAS score (p <0.0001). Thus, a correlation existed between the VAS and the olfactory test.
There was also a statistically significant association between odor discrimination determined by the olfactory test and VAS (p <0.0001). Indeed, 70% of subjects said anosmic with VAS did not discriminate odors and 99.6% of normosmic subjects with VAS had normal olfaction according to olfactory test (Table IV).
Table IV. Distribution of the odors discrimination by VAS score
VAS performance evaluation
True positives smell disorders (anosmic and hyposmic) subjects after the olfactory test were 192.
False negatives were 42; false positives were 182 and true negatives were 736.
Thus, the sensitivity of the VAS was 82.05%; the specificity was 80.17%.
In our study, on the 60 subjects that reported being anosmic by VAS evaluation (score 0), 70% were really anosmic after the
olfactory test against 28.3% of hyposmic subjects and 1.7% of normosmic subjects. Thus, overall margin of error was 30%. More, the likelihood to have normal olfaction after the olfactory test when the subject was declared anosmic by the VAS was 1.7%.
The olfactory disorders evaluated by VAS represented 374 cases out of a total of 1152 subjects, whether 32.5%, against an effective strength of 234 cases after the olfactory test evaluation, thus a prevalence of 20.3% in our study. Thus, olfactory disorders by VAS were overestimated. These data reflect an underestimation by patients of their olfactory performance. On the contrary, in the study of Hosemann et al , olfactory disorders are underestimated in the self-assessment. In fact, only 63% of patients had declared dysosmic with the self-assessment while the proportion was to 65% after the olfactory test. The high proportion of olfactory disorders in this study is due to the study population. Indeed, this population is related to patients with disabling nasal polyposis unlike our study relates to olfactory disorders detection in the general population.
In our study, there was a correlation between the VAS and the olfactory test. This correlation was determined by percentage of concordance (94.5%) and statistically significant relation between the olfactory perception and VAS (p <0.0001). The study of Hox et al.  confirms that good correlation. In their study related to a population of patients with nasal polyposis, there was a good correlation between the evaluation by the VAS and the measurement of olfactory function through the olfactory test. Furthermore, the relationship between smell self-evaluation and olfactory function measurement through the olfactory test in dysosmic patients aware of their olfactory deficit was well correlated.
On the contrary, the relationship between olfactory function assessment and nasal obstruction sensation was low in normal subjects  and very low in the patients’ population of nasal polyposis . In addition, there was no correlation between subjective methods represented by the smell self-assessment and objective evaluation methods . The difference between our results and those of Philpott et al.  is at the methodological level.
For Nguyen et al. , visual analog scale and the olfactory test are two complementary methods of smell evaluation, especially in patients with smell fluctuations. Thus, both methods should be involved in screening or evaluation of dysosmia whenever the olfactory test is available.
Sensitivity and specificity of VAS
In our study, the sensitivity of the VAS was 82.05%. It is the ability of the VAS to detect smell disorders. As for the specificity of the VAS, which is the ability to identify normosmic subjects, it was 80.17%. In this study, where the diagnostic test (VAS) was evaluated in the olfactory disorders screening, sensitivity is from our point of view, the most relevant evaluation criterion of performance. Thus, the VAS performance is acceptable and even good enough in the olfactory disorders detection.
There is a correlation between the visual analogue scale and the olfactory test in the olfactory disorders detection. In addition, the visual analog scale performance is good, permitting olfactory disorders detection by this method, especially in case of olfactory test absence. Actually, olfactory test remains the reference clinical test in routine practice.
The authors wish to thank Professor Jean-Michel Klossek, in the Department of Otorhinolaryngology, University of Poitiers (France) for the gift of olfactory test box.