A multicenter real‐life study to determine the efficacy of corticosteroids and olfactory training in improving persistent COVID‐19‐related olfactory dysfunction

Abstract Background No definitive treatment exists to effectively restore function in patients with persistent post‐infectious olfactory dysfunction (OD). Corticosteroids have been considered as a therapeutic option in post‐infectious OD but their benefit in COVID‐19‐related OD remains unexplored. We aim to determine the role of the combination of corticosteroids plus olfactory training (OT) in improving persistent COVID‐19‐related OD. Methods A multicenter real‐life cohort study was conducted between December 2020 and April 2022 on patients with reported COVID‐19‐related OD. Only patients with confirmed OD at Sniffin' Sticks (S'S) and those who attended their 6‐month follow‐up were included. Patients were started on a combined treatment of corticosteroids and OT. Patients refusing corticosteroids or not doing any treatment formed the control groups. Visual analogue scale (VAS) for sense of smell and SNOT‐22 were used to assess patients reported symptoms. Results Sixty‐seven subjects with reported COVID‐19‐related OD were initially seen. Normosmic patients at S'S (n = 14) and those not attending their follow‐up (n = 9) were excluded. Of the 44 patients included in the analysis, 19 patients had the combined treatment (group A), 16 patients refused to take corticosteroids and did the OT alone (group B) whereas 9 patients did not do any treatment (group C). An improvement of threshold + discrimination + identification (TDI) score (p = .01) and VAS for smell (p = .01) was found in group A whereas only the TDI score improved in group B (p = .04). Presence of comorbidities, age, sex (male), and length of OD negatively influenced olfactory recovery. Conclusions Our study confirms the importance of OT in long‐term OD suggesting that the addition of corticosteroids may give a benefit in terms of patient's perceived olfaction. Level of Evidence 2b


| INTRODUCTION
Olfactory dysfunction (OD) represents a highly prevalent symptom in patients infected by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) with up to 85% of mild-to-moderate coronavirus disease 2019 (COVID-19) cases developing loss of sense of smell. 1,2 Spontaneous recovery rate of olfaction is very high within the first month following infection (recovery rate 94.6%) and it becomes 85.7% at 6 months 3 and 93% at 12 months. 4 Persistent post-infectious OD (PIOD) has been recognized as a "long-COVID" symptom, defined as a persistent symptom in individuals who recovered from COVID-19 5 and, unfortunately, no definitive treatments exist to effectively restore function. European guidelines recommend olfactory training (OT) for a minimum of 3 months to maximize the chance of smell improvement. 6 Nonetheless, OT remains ineffective in 50%-85% of subjects [7][8][9] with up to 29% of PIOD cases not improving even after long-term OT (14 months). 10 Topical and systemic corticosteroids have been considered as a therapeutic option in PIOD but their benefits for non-sinonasalrelated OD remain controversial. A systematic review published in 2019 11 suggested that systemic corticosteroids could improve olfactory loss in PIOD (Level 4), whereas a more recent one 12 concluded that systemic or topical corticosteroids remain "optional" due to the lack of high-quality studies. The rationale behind the use of corticosteroids to treat PIOD relies on its capacity to reduce a subclinical inflammation which may persist in the nose after an otherwise resolved upper respiratory tract infection. On the other hand, corticosteroids could play a role in the regeneration of the olfactory epithelium of PIOD patients, as already shown in animal models. 13,14 Studies focusing on corticosteroids as treatment of PIOD did not clarify which formulation, dose and route of administration is better in improving sense of smell and if this is more effective if combined with OT. Another question remains on whether there is a time limit from OD onset at which treatment should be started in order to observe a benefit. Ultimately, in the lack of clear evidence-based guidelines the choice is left to doctor's preferences. To date, most of the authors seem to agree that corticosteroids may have a role when started close to OD onset 15 ; however, whether this could have a role in persistent OD remains partially unexplored.
In this study, we aim to investigate the role of the combination of corticosteroids plus OT in improving persistent COVID-19-related OD in a cohort of subjects with a history of smell loss longer than 7 months.
Patients refusing to take corticosteroids and doing OT alone and those not doing any treatment were used as internal controls.

| Study design
A multicenter real-life cohort study was conducted to assess the efficacy and safety of corticosteroids in combination with OT in the treatment of persistent OD in patients with a history of mild-to-moderate

| Participants' characteristics
Patients with a reported OD that occurred following a laboratoryconfirmed SARS-CoV-2 infection referred to our smell clinics at the University College London Hospitals (London, United Kingdom) and the University Hospital of Padua (Padua, Italy) were selected. All participants provided full informed consent prior to their inclusion in the study. Data were collected on demographics, subjective characteristics of OD at onset, smoking status, comorbidities, and medications taken ( Table 1). Patients with a chronic or recent short-term oral steroid use, pregnancy, pre-existing history of OD, non-COVID-19-related OD, or other pathologies known to affect olfaction (i.e., head and neck tumors, chronic rhinosinusitis [CRS], head trauma, radio/chemotherapy of the craniofacial region, psychiatric or neurological disease) were not included in the study.

| First assessment (T 0 ) and evaluation of olfactory function
On the first visit, a fully detailed medical history was obtained. Participants were asked to report any medications they used. Factors such as duration of olfactory loss and presence of parosmia, described as the occurrence of distorted olfaction when smelling odor, were also explored. All patients underwent nasal endoscopy to exclude signs of CRS-nasal polyps, nasal discharge, and signs of rhinitis-or an obstruction/inflammation of the olfactory clefts. An MRI of the head was arranged for all patients to study the olfactory system and exclude any central causes of OD. Olfaction was evaluated using Sniffin' Sticks (S'S)-extended set (Burghart, Medisense) to obtain the odor threshold (T), discrimination (D), and identification (I) scores. Normosmia was attributed where TDI score (the sum of T, D, and I individual scores) was ≥30.75, hyposmia where TDI was >16, but <30.75, and functional anosmia if TDI ≤ 16. 16 Self-assessment of olfaction was performed using a visual analogue scale (VAS-0 represents "sense of smell absent" and 10 "sense of smell not affected") 17 whereas sinonasal symptoms were evaluated using the Sino-Nasal Outcomes Test-22 (SNOT-22). 18

| Treatment and further follow-up (T1)
Patients with no OD at S'S (TDI ≥ 30.75) were discharged back to their general practitioner (GP). Conversely, patients with a confirmed OD (TDI < 30.75) were offered a steroid treatment consisting of a 2-week course of oral corticosteroids (Prednisolone 40 mg/daily for 5 days, then tapered down over 9 days) followed by intranasal corticosteroids drops for 2 weeks (Betamethasone 0.1%, 2 drops/nostril bidaily) administered in the Kaiteki position. 19 Specific consent to start the previously mentioned treatment was sought from all patients before giving any related prescription. They were also asked to start OT, as previously described, 6 until further follow-up irrespective of whether they had done or not it before. Patients with contraindications to corticosteroids 20 or refusing to take them were asked to start OT. A further follow-up at 6 months was arranged for all patients and patient-reported outcome measures (PROMs) and S'S were repeated on that occasion. Treatment adherence was checked at follow-up by requesting specific questions about treatment (i.e., modalities of topical steroid drops administration, length of time allowed for OT, and strict adherence to instructions provided).
At follow-up, patients who did not do any treatment during the study period were kept in the analysis and formed an additional control group.

| Statistical analysis
Quantitative variables were presented as median and interquartile range whereas qualitative variables were expressed as number of observations and percentage. Considering the Wilcoxon test, to T A B L E 1 General characteristics of the whole population of dysosmic patients and according to type of treatment Interval for smell loss onset, median [P25-P75], days obtain an increase in the TDI score of 5.5 points, which corresponds to the minimal clinically important difference (MCID), 21  OT-group A), 16 patients refused to take corticosteroids and did the OT alone (group B) and 9 patients did not do any treatment despite medical recommendations (group C). Figure 1 shows the flow chart for the study population.   Table 2. Apart from a significant lower number of parosmics observed in group B at baseline (p = .01), no other significant differences were observed in the measurements either at baseline or at follow-up in the three groups (Table 2).

| Effects of the therapy on olfaction
A statistically significant improvement in the TDI score was demonstrated at follow-up in patients receiving the combined treatment  observed when comparing the number of patients reaching the MCID improvement in the three groups (p = .78).

| Influence of available variables on smell improvement
Presence of comorbidities negatively influenced the TDI and identification scores in group A (p = .04 and p = .03 respectively) and the discrimination and identification scores in group B (p < .001 and p = .007 respectively). Age and sex (male) negatively influenced identification score in group B only (p < .001 for both) whereas the length of OD negatively influenced threshold and discrimination scores in group A (p = .02 and p = .01 respectively) and the discrimination and identification scores in group B (p < .001 and p = .004 respectively) ( Table 4). All the other variables were found to not influence smell recovery.

| DISCUSSION
Corticosteroids have been considered as a therapeutic option for PIOD with many studies showing promising results. [23][24][25][26][27] It has been hypothesized that some patients with persistent PIOD may have an undetectable (not macroscopically evident) ongoing inflammation in the olfactory neuroepithelium [28][29][30] which could explain why some people could respond better than others to steroidal treatment. 30  concluded that OT with a topical nasal steroid (not better described) was more effective than OT alone, especially in the subgroup of patients with PIOD. It must be stated that, in addition to the way of administration, corticosteroid molecules differ in terms of their antiinflammatory potencies and duration of action 33 which could eventually influence their potential effect to improve sense of smell.
However, to our knowledge, the best corticosteroid molecule to use in COVID-19-related OD, or broadly in post-viral OD, has not yet been identified.
Today, most of the authors agree that, considering the systemic side effects of taking oral corticosteroids, it is not recommended to use them more than 2 weeks for the treatment of COVID-19-related OD. 34 As an option, giving a short course of oral steroids for 3-4 days has been suggested as a diagnostic tool, 31 38 has been reported to be helpful. Given the potential benefits of intranasal steroid drops, we offered a combined treatment of oral and topical steroids for a total length of treatment of 4 weeks.
Finally, our data also highlight the role of OT in persistent PIOD, as demonstrated by the fact that no statistically significant improvement was observed in those who did not do it (group C).

| Strengths and limitations
This study is the first one looking at the role of corticosteroids in patients with a persistent COVID-19-related OD. Also, all patients considered in the study had no signs of paranasal inflammation, as demonstrated by a clear MRI head. This allowed us to be more confident that any smell improvement observed in the steroid group would have not been confounded by treating an underlying sinonasal disease. The main limitation of the study is its non-randomized non-blinded design as treatments suggested were not randomly assigned. However, this represents a real-life study and it was not initially designed as a prospective controlled trial. Group C did not reach the minimal sample size; therefore, we cannot exclude that the results observed regarding this group were affected by a casual effect. Even though it could be considered a controlled study for the presence of two different control groups, their inclusion was not part of the initial study design but was a consequence of patients' own choice to take or not the treatment suggested. As an additional consequence of that, the patients reported outcomes (i.e., VAS and SNOT-22) might have been biased whereas those receiving the combined treatment were more prone to believe they could have achieved an improvement at the end of the treatment.
Also, by giving a combination of oral and topical steroid drops to patients in group A, we were not able to conclude whether the observed smell improvement was due to a particular formulation of corticosteroids or to the combination of both.

| CONCLUSIONS
Our study confirms the importance of OT in the treatment of persistent COVID-19-related OD suggesting that the addition of corticosteroids may also give a benefit in terms of patient's perceived olfaction. Topical steroid drops administered in the Kaiteki position may contribute to oral corticosteroids effect by targeting directly the olfactory neuroepithelium. However, benefits of corticosteroids must be considered against their systemic side effects and randomized controlled studies on bigger populations are strongly encouraged to better clarify their role in the treatment of persistent PIOD.

CONFLICT OF INTEREST
The authors declare that they have no conflict of interest.