Materials and methods
Study design
We followed STROBE guidelines for this monocentric prospective comparative study. Participants were randomly assigned following simple randomization procedures (computerized random numbers) to 1 of 2 crossover groups: group 1 first performed SST with a surgical mask and then one week later performed SST without a mask, while group 2 first performed SST without a surgical mask and then one week later performed SST with a mask.
Ethical considerations
Ethical approval was granted by the local Ethics Committee (Agreement 2021-01-02). Written informed consent was obtained from all individual participants.
Study population
All the participants in this study were medical doctors and students from an academic oto-rhino-laryngology department. The inclusion criteria were, on a voluntary basis: sense of smell considered as normal without a mask (subjective self-reporting); informed consent and non-opposition were obtained. An earlier 48-hour negative Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) for the SARS-CoV-2 on nasopharyngeal swabs was needed.
The exclusion criteria were a history of chronic or acute rhinosinusitis disease, history of craniofacial trauma, history of radio or chemotherapy of the head and neck sphere, symptomatology consistent with SARS-CoV-2 one month or less before the test or recent contact case.
Data collection and SST
Before each test, patients responded to a questionnaire containing their personal and medical history, including nasosinusal and olfactory symptoms. Face masks used were surgical masks NF EN 14683.
The SST consisted in smelling pens blindly, with odorant pens placed for 3-4 seconds within 2cm of the nostrils (5). It was divided into three sub-tests: a threshold test (T), a discrimination test (D) and an identification test (I), each with a score of 16, and therefore a total score of 48 (TDI). According to TDI, patients were considered as normosmic (TDI>30.5), anosmic (TDI<16.5) or hyposmic (TDI between 16.5 and 30.5) (5). The score was calculated by a software called Olaf ®.
Statistical analysis
Statistical tests were performed using the IBM SPSS Statistics 20. Continuous values were expressed with their average number ± Standard Deviation (SD) and categorical values as numbers and percentages. The intra-group results with and without masks were compared using a paired non-parametric test. A significant difference was considered for p<0.05.
Average results for each test and each sub-test were calculated. 60% of subjects with TDI change of 5.5 or more will report clinical olfactory change, so a difference in TDI scores with and without a mask less or equal to 5 was considered as not clinically relevant (6). Differences in T, D and I scores with and without a mask were considered of clinical significance when superior to 2.5, 3, and 3, respectively (6).
The inter-group results were compared with a Mann Whitney test.