Background: The effect of leprosy on the sense of smell is not yet fully established. Studies that have relied only on patients’ perceptions may have under- or over-estimated the change in smell perception. A validated and psychophysical method is necessary to avoid these errors in assessment.

Objectives: This study aimed to validate the existence of olfactory involvement in leprosy patients.

Methods: A cross-sectional, controlled study was conducted, in which individuals with leprosy (exposed individuals) and individuals without leprosy (control patients) were recruited. For each exposed individual, we selected two control patients. A total of 108 patients (72 control patients and 36 exposed individuals) with no history of infection with the new coronavirus (COVID-19) took the University of Pennsylvania Smell Identification Test(UPSIT).

Results: Most exposed individuals had olfactory dysfunction [n = 33, 91.7% (CI 95%: 77.5%–98.3%)] when compared with the control patients [n = 28, 38.9% (CI 95%: 27.6%-51.1%)], but only two (5.6%) had olfactory complaints. The olfactory function was significantly worse among exposed individuals [UPSIT leprosy = 25.2 (CI 95%: 23.1–27.3) when compared with the UPSIT control patients = 34.1 (CI 95%: 33.0–35.3); p<0.001]. The risk of olfactory loss was higher among the exposed individuals [OR: 19.5 (CI 95%: 5.18–105.70; p< 0.001)].

Conclusions: Olfactory dysfunction was highly prevalent among exposed individuals, although they had little or no self-knowledge of the disorder. The results show that it is important to assess the sense of smell in exposed individuals.

A cross-sectional, controlled study was conducted, in which individuals with leprosy (exposed individuals) and individuals without leprosy (control patients) were recruited. For each exposed individual (new cases sampled at the time of diagnosis and existing cases undergoing MDT), we selected two control patients. The study was conducted through non-probabilistic convenience sampling, which means that all cases of leprosy in the city of Londrina-PR, from  a period of 13 months betwen 2021-2022, were included in the study, provided that they met the inclusion/exclusion criteria and that the patients involved agreed to participate. A signed consent form was obtained from all patients before participation. The study was approved by the Ethics Committee in Research Involving Human Beings at the State University of Londrina, CAAE: 42159420.0.0000.523. We excluded patients aged under 18 years, those who had olfactory disease caused by traumatic brain injury, radiation therapy to the face, and post-airway infection dysfunctions (including COVID-19), and those who did not understand the procedures involved in the test (16 individuals in all). For data collection, patients were asked to fill out a form and medical information was retrieved from their electronic medical records. A test with 40 odors, namely the University of Pennsylvania Smell Identification Test (UPSIT), was conducted to assess the olfactory function. Each participant was given 4 booklets with 10 substances, which they had to scrape and sniff, before identifying them by marking an option from among four alternatives. They had to choose one (see supplementary material). The UPSIT value ranged from 0 to 40, based on the number of correct answers, classifying the olfactory function of the individual as normosmia, hyposmia, anosmia, or possible simulator. This categorization differed when applied to men and women, wherein the latter presented greater olfactory acuity. We performed Fisher's Exact Test to compare categorical variables. Continuous variables were analyzed using the Wilcoxon-Mann-Whitney Test. After verifying the normality of the quantitative variables using the Shapiro-Wilk test, the nonparametric Mann-Whitney and Kruskal-Wallis tests were used to compare the results obtained. Spearman's nonparametric correlation test was used to verify the relationship between discrete quantitative variables. A comparison of the risk of olfactory loss between the exposed and control groups was carried out using logistic regression. We considered the values of statistical significance (p-value) under 0.05. The 95% confidence interval was calculated. All data were compiled in an Excel spreadsheet for further statistical analysis.

Stata (version 13.0, Statacorp Texas), Jamovi, and R studio were used for statistical analysis.