Parosmia in patients with post-infectious olfactory dysfunction in the era of COVID-19-associated olfactory impairment

Objectives A large number of patients with olfactory impairment are affected by parosmia or phantosmia. This study aimed to examine the demographic and clinical characteristics of parosmia. Methods We performed a retrospective data analysis of patients consulting at our Smell and Taste Outpatient Clinic. A total of 297 patients were included (203 women, mean age 44.4 ± 13.7 years). Olfactory function was quantified using the “Sniffin’ Sticks” composite TDI (odor threshold, determination, and identification) score. The presence of qualitative olfactory impairment was assessed trough medical history and a parosmia questionnaire. Results Most of the patients showed olfactory impairment after an infection with SARS-CoV‑2 (84%) and were diagnosed with parosmia (49%). Patients with parosmia (PAR) (n = 201) were significantly younger compared to the group without parosmia (noPAR; n = 92) (PAR 43.2 ± 13 years vs. noPAR 47 ± 15.1 years, p = 0.03) and had a slightly shorter duration of disease, without reaching statistical significance (PAR 10.3 ± 4.9 months, noPAR 13.6 ± 37.6 months, p = 0.23). They also had higher TDI scores (PAR 24.3 ± 7 points, noPAR 21.4 ± 8.2 points, p = 0.003). Conclusions Patients affected by parosmia were younger and had a better olfactory function compared to patients without parosmia.


Introduction
Olfactory disorders are common with approximately one fifth of the population being affected [4,65].Due to the coronavirus pandemic, olfaction came to the attention of the general population because an infection with SARS-CoV-2 is often associated with quantitative as well as qualitative olfactory impairment, such as parosmia and phantosmia [53].Even before the pandemic, olfactory dysfunction (OD) following infections of the upper respiratory tract was one of the most common causes of olfactory loss [14,63].
Qualitative OD is often, but not necessarily, accompanied by quantitative olfactory dysfunction [25].Parosmia, meaning a distorted sensation of smell, is often associated with infections or head trauma [2,9,33,66,68].Affected people describe the distorted smells mostly as unpleasant and disgusting, which affects their daily life.They report feeling distressed, anxious, and worried about their future [52], leading to mental health problems and in the most severe cases to clinical depression [12,29].Furthermore, they suffer from loss of appetite and reject certain foods because they are unable to tolerate many odors and flavors, resulting in re- HNO 9 • 2024 649 duced pleasure in food intake, changes in dietary patterns, social behaviors related to dining, and changes in weight [11,16].Olfactory impairment in general often results in reduced enjoyment of food [44].
Parosmia is thought to be a sign of recovery.However, the literature on this topic varies [24,39,56,59].Moreover, the pathogenesis of parosmia is widely debated.In the past, two different hypotheses have often been proposed.In "peripheral" theory, itis assumed thata loss of intact olfactory neurons results in the inability to form a complete "picture" of the odorant [17,20,37]; in "central" theory, the integration or interpretation in the brain is impaired and a distorted odor is formed [36].The results of a recent study support the peripheral theory by demon-strating a decreased olfactory threshold in COVID-19 associated with OD [47].
Patients with SARS-CoV-2 rarely complain of additional rhinitis symptoms, and pathophysiological considerations suggest that the OD is due to damage to the olfactory mucosa [5,64] rather than nasal obstruction/congestion [6].SARS-CoV-2 enters the cells via the surface receptor ACE2 [46], which is mainly expressed by the supporting cells in the olfactory mucosa.This is the most likely explanation for the rapid improvement in olfactory function in many cases, since the olfactory receptor neurons themselves are not directly affected [6].A possible entry of the virus via Neuropilin 1 [7] and transport via the olfactory and/or trigeminal nerve into the brain is discussed [28].However, the rapid recovery of symptoms in most cases contradicts this hypothesis [6].
In post-infectious OD (PIOD), improvement of olfactory function is most frequently compared with other causes of OD [66] and often occurs spontaneously in one third of the patients within 2-3 years [21,56].The recovery rate lies between 35% and 46% after 1.9 years [8] but it has been reported to be higher [19,31].In COVID-19-associated OD, the recovery rate is at 43% after an average of 3.5 months [47].Recent studies have shown that the prevalence of long-term olfactory impairment lies between 2.9% and 8.3% 2 years after COVID-19 infection [3,34].
The aim of this study was to investigate the demographic and clinical characteristics in a large group of patients with parosmia.

Study design
The present study is a retrospective data analysis of patient data, which were collected during consultations at the Smell and Taste Outpatient Clinic of the Department of Otorhinolaryngology at the University Hospital Dresden from February 2021 to July 2022.
The investigation was approved by the ethics committee of the medical faculty of theTechnical Universityof Dresden(BO-EK-254062022).The study was designed and conducted in accordance with the ethical principles defined in the current revised version of the 2013 Declaration of Helsinki.

Participants
Data were collected from 297 patients (203 female patients, 91 male patients, 3 not specified).Although there were isolated missing data, no patient was excluded as each variable was analyzed on a case-bycase basis.The patients were divided into two groups according to their medical history: parosmia present and parosmia not present.Only patients with PIOD (non-COVID), including COVID-19, were enrolled in the study.Therefore, a further distinction was made between COVID and non-COVID in the parosmia group.The differen-tiationbetweenCOVID and non-COVID was made on the basis of medical history and partly by presenting a polymerase chain reaction (PCR) test result.Other causes of OD, such as chronic rhinosinusitis, neurodegenerative diseases, or head trauma, were excluded.

Clinical examination and parosmia questionnaire
As part of the routine presentation in our smell and taste consultation, each patient received a detailed medical history, especially regarding the presence of parosmia and phantosmia symptoms.According to the frequency (daily, not daily), intensity (very intense, less intense), and consequences (weight loss, change of daily activities) of the symptoms, the parosmia was graded from 0 to III degrees [23].In addition, participants underwent nasal endoscopy by a certified otorhinolaryngologist to identify pathologies, such as massive septal deviations, nasal polyps, and signs of inflammation.To further evaluate the qualitative olfactory dysfunction some of the patients (n = 115) were asked to complete the parosmia questionnaire (PQ; . Table 1), as used in a previous study by Landis and colleagues [33].Each response was scored, and the summated score was the PQ score (min 4, max 16 points).A low PQ value is associated with a higher probability of occurrence of parosmia.
Additionally, depending on the degree of parosmia, a subgroup analysis was "Sniffin' sticks" and TDI score Quantitative olfactory function was assessed using the "Sniffin' Sticks" test battery (Burghart, Holm, Germany).Odor threshold (T), odor discrimination (D), and odor identification (I) scores were assessed as described previously [51].The sum of the results yields the TDI score, on the basis of which the olfactory function can be divided into anosmia (TDI ≤ 16 points), hyposmia(16 < TDI ≤ 30.5 points), and normosmia (TDI ≥ 30.75 points).

Subjective olfactory function
The patients were asked either to mark their olfactory function on a visual analogue scale (VAS) of 10 cm length, the left end indicating no olfactory function and the right end perfect olfactory function, or to write down their olfactory function in percent.The selection of the VAS was converted into a score ranging from 0 to 100.
Regarding the mean duration of the OD, there was no statistically significant difference between the two groups (PAR 10.3 ± 4.9 months, noPAR 13.6 ± 37.6 months, p = 0.21, 1-344 months).
The severity of the qualitative disorder was graded from 0 to III degrees in 204 patients.In total, 23, 58, 100, and 23 patients were diagnosed with parosmia of degrees 0, I, II, and III, respectively.The subgroup analysis for parosmia degree < II and ≥ II showed no significant differences between the two groups with respect to age, gender, or duration of disease.
We also observed a negative correlation between the degree of parosmia and the PQ score in the group of patients withparosmia(Pearsonr = -0.59,p < 0.001).Furthermore, the subgroup analysis showed a significant difference in the PQ score between parosmia degrees < II and ≥ II.However, there was no significant correlation between the PQ score and the parosmia degree when analyzing the two groups separately (PAR < II Pearson r = -0.38,p = 0.07; PAR ≥ II Pearson r = -0.18,p = 0.20).

Discussion
Patients affected by parosmia were younger than patients without parosmia and had a better olfactory function.These findings also match the results of an online survey conducted by Pellegrino and colleagues [54] but not those of Menzel et al. because we could not show a difference in gender and duration of disease when parosmia is present or not [47].
Several studies suggest that there is a discrepancy between the self-rated olfactory function and olfactory test results [32,43,61,62].While self-ratings and self-descriptions of olfactory function are of the highest significance in clinical olfaction, a structured history in a personal interview and the clinical examination are equally important in acquiring a less biased view of the complaints.Accordingly, olfactory tests are inevitable for assessing olfactory function.Still, there was a correlation between self-rated olfactory function and the TDI, T, and I score in our study population.
Compared to PIOD from the pre-COVID era, the proportion of individuals with parosmia appeared to have increased since COVID-19.In our study population, more than two thirds of the patients complained of parosmia (PAR 69% vs. noPAR 31%).In other studies, the prevalence was up to 76% [47,50,60].However, in the pre-COVID-19 pandemic period it was between 18% and 56% depending on the study population [30,49,56,57,67].For the last 2 years, mainly patients with post-COVID OD presented for consultation, which explains the unbalanced ratio of non-COVID and COVID-19 patients in our study cohort, as well as the longer duration of disease for the non-COVID group.In addition, only outpatients were included in our study, and therefore no general conclusion can be drawn regarding overall olfactory dysfunction in COVID-19.Furthermore, it should be noted that within our observation period there were different SARS-CoV-2 variants (peak of alpha: April-May 2021, peak of delta: September-November 2021, peak of omicron: December 2021-February 2022; [15]).
To date, there is no validated test available for the diagnosis of parosmia; only structured medical history and the parosmia questionnaire of Landis et al. are used [33].Several clinical findings were associated with parosmia such as a reduced volume of the olfactory bulb [48,57,58] and specific patterns in central processing a b Fig. 2 9 Box plot of odor threshold (T), discrimination (D), identification (I), and TDI scores (TDI) in patients with parosmia (PAR) and without (noPAR).*Statistically significant differences between the groups; ◻ median value, ♦ outliers of olfactory stimuli in functional magnetic resonance imaging (fMRI; [26]).However, none is suited for clinical diagnostics at an individual level.With our study, we confirmed that the PQ is suitable for distinguishing between patients with parosmia and patients without parosmia.We also observed a negative correlation between the degree of parosmia and the PQ score in the PAR group.Furthermore, the subgroup analysis showed a significant difference in the PQ score between parosmia degrees < II and ≥ II.However, there was no significant correlation between the PQ score and the parosmia degree in these groups, implying that the PQ is suitable for screening for the presence of parosmia but not for grading.Furthermore, the degree of parosmia should be evaluated depending on the symptoms [23].Liu et al. proposed the "SSParoT" based on the hedonic rating of odors as a tool to assess qualitative OD in healthy and affected individuals [41].However, the discrimination between parosmia and non-parosmia seems to be problematic [60].
Treatment options for parosmia are limited.Recently Altundag et al. demonstrated a positive effect of modified olfactory training in patients with parosmia following a SARS-CoV-2 infection [1].Another study revealed that the presence of parosmia is associated with a relevant recovery in olfactory function in PIOD patients after olfactory training [40].Therefore, olfactory training seems to be an effective therapy for parosmia.Garcia and colleagues treated 12 post-COVID patients with gabapentin in addition to olfactory training and topical corticosteroids [18].After 3 weeks of treatment a significant improvement was noticed in 67% of the patients.So far, gabapentin has also been shown to have a positive benefit for phantosmia in isolated cases [10].Only few medical treatment options for parosmia have been proposed in the past.Zilstorff applied cocaine hydrochloride to the olfactory mucosa that led to temporary blockage of olfactory distortion due to an anesthetic effect by blocking of the sodium channel and interfering with action potential [69].Oral substitution of alpha-lipoic acid per day for an average of 4.5 months resulted in a reduction in the percentage of parosmia in patients with post-infectious olfactory disorders (48% before vs. 22% after treatment; [22]).Although a beneficial effect of the treatment with alphalipoic acid was reported, no further investigations have been published since then.Intranasal application of sodium citrate led to a reduction in the proportion of patients with parosmia after treatment (38% before vs. 25% after treatment) but without reaching statistical significance (p = 0.17) [67].The surgical resection of olfactory mucosa is a more invasive but possibly beneficial approach in severe qualitative OD [27,38,45].To date there is one study proposing a successful treatment of unilateral peripheral parosmia by olfactory cleft blockage [42].
Parosmia is considered a positive prognostic predictor for olfactory recovery, especially in patients with PIOD [24,40,47,56].However, this was not confirmed in other studies [55].Still, it seems to be of high importance to inform the patients about this positive aspect, because parosmia is occurring more frequently since COVID-19, and some patients do not know how to cope with this symptom.It gives hope to the patients with parosmia, which is important, because of the association between parosmia and depression and/or anxiety [13,35].

Conclusion
Patients affected by parosmia were younger and had a better olfactory function compared to patients without parosmia.Qualitative olfactory dysfunction can have a major impact on the patients' quality of life, and therefore identifying these patients is mandatory in order to provide appropriate therapy.This requires further research due to the lack of specific therapy options for parosmia.Einen ausführlichen Bericht finden Sie auf der Kongresshomepage: https://kongress-hno. de

Fig. 1 8 a
Fig. 1 8 a Gender ratio in the groups with parosmia present and not present.b Proportion of parosmia in the COVID-19 and non-COVID groups.c Ratio of duration of disease in the groups with parosmia present and not present subdivided into ≤ 3 months, 4-12 months, 13-24 months, and > 24 months.d Distribution of parosmia degree in the study population ). Regarding quantitative olfactory function (T, D, I, and TDI HNO 9 • 2024 651

Table 1
Parosmia questionnaire including four questions in order to elicit the presence or absence of parosmia

Table 2
Demographic information and results of the olfactory testing a a Results are shown as mean (± standard deviation) if not stated differently *Statistically significant