Prevalence of alpha-synuclein seeding activity in the Olfactory Mucosa following SARS-CoV-2 infection: results from a pilot study

Study design and participants

This observational pilot study (ClinicalTrials.gov: NCT05401773) included four groups: patients with Parkinson’s disease (PD) as positive controls, individuals with prior SARS-CoV-2 infection and persistent olfactory dysfunction (COVID+OD), individuals with prior SARS-CoV-2 infection without olfactory dysfunction (COVID−OD), and healthy controls (HC).

Potentially eligible participants were identified through a broad screening approach, including consecutive referrals from the specialised post-COVID outpatient clinic as well as participants from a previous observational studies conducted at our centre [28, 31], and personal contacts or word of mouth among hospital staff and their relatives. PD participants were recruited from the movement disorders outpatient clinic.

All participants were aged ≥18 years and able to understand the study procedures. Both COVID groups required documented, PCR-confirmed SARS-CoV-2 infection ≥ 3 months before inclusion. COVID+OD required olfactory dysfunction persisting for ≥ 3 months after infection, scores < 13/16 on both the Sniffin’ Sticks Identification (SSI) and Discrimination (SSD) subtests, and no structural nasal pathology potentially explaining the dysfunction. OD was attributed to SARS-CoV-2 based on its onset in temporal association with infection. COVID−OD required no history of or current OD and scores ≥ 13/16 on both subtests. Both COVID groups excluded participants with OD predating infection or relevant pre-existing neurological disorders. HC had no history of COVID-19, OD, or relevant neurological disorders, subjectively normal olfaction, and SSI and SSD scores ≥ 13/16. PD participants met established clinical diagnostic criteria [27], had no history of COVID-19, and served as positive controls for the RT-QuIC assay. All groups required a negative SARS-CoV-2 antigen test at inclusion; HC and PD participants additionally required a negative SARS-CoV-2 antibody test unless vaccinated (ClinicalTrials.gov: NCT05401773).

The study was approved by the local ethics committee in March 2021, and recruitment continued until January 2024. All participants provided written informed consent before participation. Post-COVID-19 participants were followed for at least one year, as described previously [28, 31].

Clinical assessments

All participants underwent a standardized neurological examination, including an assessment of non-motor symptoms of PD, performed by study investigators (BH, KS, NDC, FJ, CC) under the supervision of trained movement-disorder specialists (BH, KS). Motor and non-motor symptoms were assessed using the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS, Parts I–III). In addition, cognitive function was screened using the Montreal Cognitive Assessment (MoCA) [22], and the Innsbruck REM Sleep Behavior Disorder Inventory (RBD-I) was used to identify symptoms suggestive of RBD, with a normalized score of ≥ 0.25, corresponding to at least two positive responses, being defined as a positive screen [11]. Olfactory function was assessed using the Identification (SSI) and Discrimination (SSD) subtests of Sniffin’ Sticks test battery (Burghart Medizintechnik, Wedel, Germany). To minimize cross-infection risk during the COVID-19 pandemic, odors were presented using a single-use paper sheet to which the tip of the reusable odor pen had been applied, in accordance with the recommendations of the German Society of Otorhinolaryngology, Head and Neck Surgery; consequently, the threshold subtest was omitted because it could not be safely administered under these conditions [15].

OM sampling and RT-QuIC

Olfactory mucosa (OM) samples were collected by ENT specialists medial of the middle nasal turbinate in the direction of the superior nasal turbinate using FLOQ-Brushes (Copan Italia) [4]. Following the OM collection, swabs were stored at 4 °C, and shipped on dry ice to the University of Verona for processing and aSyn-SAA analysis, as previously published [5, 36].

In particular, recombinant full-length human α-synuclein (aSyn; amino acids 1–140) was prepared in-house. aSyn cDNA was cloned into pET-28a and expressed in E. coli BL21 (DE3). After induction with 0.1 mM IPTG, the periplasmic fraction was isolated, subjected to ammonium sulfate precipitation, and further purified via anion-exchange chromatography (Q-Sepharose, GE Healthcare). Protein purity was verified by SDS-PAGE, and aliquots were dialyzed into 10 mM sodium phosphate buffer (pH 7.4) and stored at − 80 °C.

For aSyn-SAA, 2 µL of diluted OM homogenate was incubated in 98 µL reaction buffer (100 mM phosphate buffer, pH 8.2; 10 µM Thioflavin T; 0.1 mg/mL recombinant aSyn; 0.004% SDS) containing ~ 37 mg of 0.5 mm glass beads. Reactions were run in triplicate at 30 °C in a FLUOstar® Omega plate reader under cycles of shaking (1 min at 200 rpm, double orbital) and rest (14 min), with fluorescence recorded every 45 min (excitation 450 ± 10 nm, emission 480 ± 10 nm).

A reaction was considered positive if fluorescence exceeded 10% of the plate’s maximum signal. OM samples were classified as positive if ≥ 2 of 3 replicates crossed this threshold within 50 h, negative if none did, and inconclusive if one of three was positive. Inconclusive samples were reanalyzed and, if results persisted, classified as negative [5].

Statistics

All statistical analyses were performed using IBM SPSS Statistics (version 30.0; IBM Corp., Armonk, NY, USA). Age, sex and OM-SAA results were compared among people with PD, individuals with post-COVID-19, and HC. Group comparisons were performed using chi-square tests with Benjamini–Hochberg FDR-corrected post-hoc analyses, and a univariate ANOVA followed by Šidák-corrected post-hoc tests. As the PD cohort was included as a positive control group for the OM-SAA, the PD cohort was not included for further group comparisons.

Chi-square tests were applied to assess differences in categorical variables (e.g., sex, persistent hyposmia, OM-SAA positivity) among COVID+OD, COVID-OD, and HC. Group differences in OM-SAA positivity were analyzed using contingency-table statistics. Counts of OM-SAA–positive and –negative individuals were cross-tabulated across three groups (COVID+OD, COVID-OD, and HC). The overall association between group and OM-SAA status was assessed using a two-sided chi-square test of independence (2×3 design), with post-hoc comparisons corrected using the Benjamini–Hochberg FDR procedure; expected cell counts satisfied assumptions for χ² approximation. Effect size was reported as Cramer’s V. To complement these analyses, column-proportion tests with z-statistic annotation were applied for descriptive pairwise comparisons, where groups sharing the same subset label did not differ significantly at p < 0.05. To account for age differences across groups, we conducted a multivariable logistic regression predicting OM-SAA positivity, entering age and sex as covariates. Model significance was assessed using the omnibus χ² statistic, and effects are expressed as odds ratios (95% CI). Additionally, OM-SAA positivity was modelled using multivariable logistic regression for those contrasts identified as significant in the chi-square analysis, with age and sex included as covariates and p values adjusted via Benjamini–Hochberg FDR correction.

Continuous outcomes (e.g., MoCA and MDS-UPDRS scores) were analyzed using analysis of covariance (ANCOVA) models controlling for age and sex, with post-hoc comparisons corrected using the Šidák procedure. Within the post-COVID cohort, differences between participants with and without positive OM-SAA results, were assessed for key demographic and clinical measures. To account for potential confounding by age and sex, analysis of covariance (ANCOVA) model was applied to these continuous outcomes.

Age and time since COVID-19 are reported as mean ± standard deviation (SD). All remaining continuous variables are summarized as mean ± standard deviation (median; 95% confidence interval of the mean), while categorical variables are presented as absolute and relative frequencies.

Due to the lack of existing prevalence data regarding OM-SAA of COVID-19 patients and significant uncertainty regarding their prevalence in HC, a formal power calculation was not conducted. The study was designed as an exploratory pilot to inform future research. The significance threshold was set at p < 0.05 (two-sided, unless stated otherwise). Post hoc pairwise comparisons were performed and reported only when the corresponding overall test was statistically significant. For multiple comparisons, Benjamini–Hochberg FDR correction was applied to chi-square and multivariable logistic regression analyses, and the Šidák procedure was used for ANOVA and ANCOVA, with all p values referring to the adjusted level unless otherwise stated. p values are reported to two decimal places, except for values  < 0.05, which are reported to three decimal places or as p < 0.001.

Because participants were recruited consecutively, age was not fully balanced across the HC and post-COVID-19 cohorts. Therefore, an age- and sex-balanced sensitivity analysis was performed. Equal-sized subsets were selected from each study group without replacement using a reproducible optimization procedure implemented in Python (version 3.12; Python Software Foundation, Wilmington, DE, USA), following the general principles of covariate balance in observational studies [37]. The optimization prioritized minimizing between-group differences in mean age, followed by balancing sex distribution, while maximizing the number of retained participants. The final optimized cohort comprised 33 participants per group. Analyses of the clinical outcome variables, including the MoCA and MDS-UPDRS scores, were repeated in this cohort to assess the robustness of the findings.

As an additional assessment of the robustness of the selection procedure, propensity scores for the probability of belonging to the COVID+OD group were estimated from age and sex using the Propensity Score Matching procedure implemented in IBM SPSS Statistics (version 30.0; IBM Corp., Armonk, NY, USA). The COVID+OD group was selected as the reference because it was the only group that differed significantly in age from both the HC and COVID−OD groups. Participants were ranked according to their estimated propensity scores, and a propensity score-based trimming procedure was applied to obtain equally sized, age-balanced groups. The resulting subset was highly consistent with the optimized sensitivity cohort, differing by only two participants in the COVID−OD group.

OM-SAA outcomes in healthy controls and post-COVID-19 participants (Table 2)

Table 2 summarizes demographic data as well as OM-SAA results between HC and the post-COVID-19 groups. Age was comparable between the OM-SAA positive and OM-SAA negative groups (mean ± SD: 51.4 ± 17.7 vs. 46.4 ± 15.8 years, p = 0.15).

Across the 137 individuals from the HC and post-COVID-19 cohorts, OM-SAA positivity was detected in 20.4% of cases (28/137). OM-SAA positivity rates differed across groups, with 11.9% in HC, 35.7% in COVID+OD, and 15.1% in COVID-OD (Table 2). A chi-square test of independence demonstrated a significant overall association between diagnostic group and OM-SAA status (χ²(2)  = 8.84, p = 0.012), corresponding to a small-to-medium effect (Cramer’s V = 0.25). Pairwise comparisons with Benjamini–Hochberg FDR adjustment showed that OM-SAA positivity was significantly higher in the COVID+OD group than in both the HC group (FDR-adjusted p = 0.030) and the COVID-OD group (FDR-adjusted p = 0.030), whereas no significant difference was observed between the COVID-OD and HC group (FDR-adjusted p = 0.65). Post-hoc column-proportion tests yielded the same pattern, with significantly higher OM-SAA positivity exclusively in the COVID+OD group, while HC and COVID-OD did not differ.

Moreover, in the overall age- and sex-adjusted logistic regression model, diagnostic group remained significantly associated with OM-SAA positivity after adjustment for age and sex (omnibus χ²(4)  = 10.14, p = 0.038, Nagelkerke R² = 0.11; Table 2). Group allocation was the only significant predictor, whereas neither age nor sex was associated with OM-SAA positivity.

Within this model, the COVID+OD group showed significantly higher odds of OM-SAA positivity than both healthy controls and the COVID−OD group, whereas no significant difference was observed between the COVID−OD and HC groups (Table 2). Additional age- and sex-adjusted pairwise logistic regression analyses confirmed significantly higher odds of OM-SAA positivity in the COVID+OD group compared with both healthy controls and the COVID−OD group (Table 2). These findings are consistent with the categorical analyses, indicating that elevated OM-SAA positivity is specific to participants with post-COVID-19 olfactory dysfunction.

Clinical outcomes in healthy controls and post-COVID-19 participants (Table 3)

To assess group differences in continuous clinical variables between healthy controls and the post-COVID groups, univariate ANCOVAs adjusted for age and sex were performed (Table 3). MoCA scores differed significantly between groups (F = 3.50, p = 0.033, partial η² = 0.05). Age was strongly associated with MoCA performance (F = 42.14, p < 0.001, partial η² = 0.24), whereas sex showed no significant effect (F = 3.71, p = 0.06). Post-hoc comparisons indicated that this effect was driven by lower MoCA scores in the COVID+OD compared to the HC group (p = 0.038). Similarly, MDS-UPDRS total scores also differed significantly between groups (F = 10.27, p < 0.001, partial η² = 0.13). Age was significantly associated with MDS-UPDRS total scores (F = 29.21, p < 0.001, partial η² = 0.18), whereas sex showed no significant effect (F = 0.63, p = 0.43). Post-hoc comparisons revealed higher MDS-UPDRS-total scores in the COVID+OD than in both healthy controls (p < 0.001) and the COVID-OD group (p = 0.015). Significant group effects were also observed for MDS-UPDRS Part I (F = 5.01, p = 0.008, partial η² = 0.07), Part II (F = 4.05, p = 0.020, partial η² = 0.06), and Part III (F = 6.63, p = 0.002, partial η² = 0.09). Compared with healthy controls, the COVID+OD group had higher scores in Parts I (p = 0.006), II (p = 0.021), and III (p = 0.002). MDS-UPDRS Part III scores were also higher in the COVID+OD group than in the COVID−OD group (p = 0.025).

To assess the robustness of the clinical findings despite the age differences between the HC and post-COVID groups, an additional age- and sex-balanced sensitivity analysis was performed including 33 participants per group (see Table 3). The results closely mirrored the primary analyses, confirming significantly lower MoCA scores and higher MDS-UPDRS total scores in the COVID+OD group compared with healthy controls. Similarly, significantly higher MDS-UPDRS Parts I and III scores were observed, whereas the group effect for MDS-UPDRS Part II did not reach statistical significance.

Clinical outcomes by OM-SAA positivity in post-COVID participants (Table 4)

Age was comparable between the OM-SAA positive and OM-SAA negative post-COVID groups (Table 4). To test whether OM-SAA positivity was associated with olfactory function, we conducted a univariate ANCOVA with age and sex as covariates. After adjustment, OM-SAA positive participants had significantly lower SSD scores (F = 12.51, p < 0.001, partial η² = 0.12) and SSI scores (F = 7.61, p = 0.007, partial η² = 0.08). Age was strongly associated with both SSD and SSI (both p < 0.001; partial η² = 0.13 and 0.26, respectively), whereas sex showed no significant effect on either performance. As summarized in Table 4, no significant associations were observed between OM-SAA positivity and MDS-UPDRS total or subscores, MoCA, or RBD-I scores.

Ethics approval and consent to participate

Ethical approval for the trial was obtained from the Ethics Commission of the Medical University of Innsbruck (MUI), Austria. Written informed consent was obtained.

Trial registration number

ClinicalTrials.gov: NCT05401773.

Competing interests

The authors declare no competing interests related to this work.