If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Radboud University Medical Centre; Donders Institute for Brain, Cognition and Behaviour; Department of Neurology; Center of Expertise for Parkinson & Movement Disorders; Nijmegen, the Netherlands
Radboud University Medical Centre; Donders Institute for Brain, Cognition and Behaviour; Department of Neurology; Center of Expertise for Parkinson & Movement Disorders; Nijmegen, the Netherlands
Radboud University Medical Centre; Donders Institute for Brain, Cognition and Behaviour; Department of Neurology; Center of Expertise for Parkinson & Movement Disorders; Nijmegen, the Netherlands
Corresponding author. Radboud University Medical Centre; Donders Institute for Brain, Cognition and Behaviour; Department of Neurology; Center of Expertise for Parkinson & Movement Disorders; Nijmegen, the Netherlands.
Affiliations
Radboud University Medical Centre; Donders Institute for Brain, Cognition and Behaviour; Department of Neurology; Center of Expertise for Parkinson & Movement Disorders; Nijmegen, the NetherlandsDepartment of Epidemiology; Erasmus MC University Medical Centre; the Netherlands
Olfactory dysfunction is associated with a significantly increased risk of Parkinson's disease (PD).
•
Risk estimates were higher for studies with higher % women, increased mean study age and REM-sleep behavior disorder cohorts.
•
PD risk decreased with longer latency after olfactory dysfunction onset.
•
Personalized risk estimates are needed for accurate PD diagnosis.
Abstract
Background
Interest has risen in identifying individuals at high risk of incident Parkinson's disease (PD) to facilitate inclusion in neuroprotective treatment trials. Current risk estimates of prodromal markers are based on aggregated data of an entire population, but this approach disregards differences in risk estimates by subgroups of a population. In this proof of concept, we determine subgroup-specific risk estimates of olfactory dysfunction for incident PD.
Methods
PubMed, EMBASE and Cochrane were searched for prospective studies investigating the association between olfactory dysfunction and incident PD. Random-effects meta-analysis, subgroup analyses and meta-regression were performed to investigate general and subgroup risk estimates.
Results
Individuals with odor identification dysfunction seemed to be at greater risk of incident PD compared to controls without olfactory dysfunction (OR = 4.18; 95%CI [2.47–7.07]). Risk estimates were higher in studies that included higher percentages of women (regression slope β = 0.053 increase in log odds ratio per 1% increase 1%, p = 0.0006), increased with mean study age (β = 0.21 per one year increase; p = 0.005) and in REM-sleep behavior disorder cohorts (β = 1.95; p = 0.03). Furthermore, the association between olfactory dysfunction and incident PD was most distinct in studies with shorter follow-up duration (ß = −0.56; p = 0.0047).
Conclusion
The presence of olfactory dysfunction conveys a considerably elevated risk of incident PD, likely more in studies with a higher proportion of women, older individuals or short follow-up duration. Individual patient data are warranted to confirm these findings and to yield subgroup-specific risk estimates of other common markers to refine prodromal PD criteria.
1. Background
Parkinson's disease (PD) is a progressive neurological disorder that is often preceded by a prodromal phase that largely consists of non-motor features. Olfactory deficits are a common feature of both the prodromal and clinically manifest phase of PD [
]. Various hypotheses have been proposed to explain the occurrence of early loss of smell in PD patients, for example that the olfactory bulb is affected early on as α-synucleinopathy ascends within the nervous system [
As olfactory dysfunction is present early in the disease course and appears to be common, it is regarded as a promising biomarker to timely diagnose PD, which may be incorporated in recruitment algorithms for neuroprotective treatment trials [
]. Recently, interest has risen in developing quantitative methods to estimate risk of incident PD based on presence of prodromal features, including olfactory dysfunction, REM sleep behavior disorder (RBD) and other factors [
]. Note that the current Movement Disorder Society (MDS) criteria for (prodromal) PD only provide overall estimates of the risk of incident PD based on these features [
]. However, this approach may disregard differences in risk by subgroup, as the specificity of a putative prodromal feature for PD pathology could vary depending on the characteristics of a patient subgroup. For example, a clinician would be much more aware of the risk of clinical PD if olfactory dysfunction occurred in an individual with RBD than in an individual without RBD. To date, however, empirical data assessing risk estimates of olfactory dysfunction by subgroup are scarce.
To address this key gap in knowledge, we aimed to unravel the association between olfactory dysfunction and incident PD by meta-analyzing all published longitudinal studies, with a particular focus on subgroup-specific estimates based on age, sex, presence of RBD, and latency until diagnosis of PD after detection of olfactory dysfunction.
2. Methods
2.1 Search strategy
PRISMA guidelines for conducting and reporting systematic research were used. PubMed, EMBASE and Cochrane databases were searched for eligible studies up-to-date until July 4, 2020. The search strategy is detailed in the Supplementary material. Reference lists of the included articles were hand-searched for additional eligible studies.
2.2 Selection criteria
Two reviewers (JJD, AT) independently evaluated the eligibility of studies. We only included prospective, longitudinal studies while we excluded cross-sectional studies in this meta-analysis, because of two reasons: prospective studies are less susceptible to selection bias, provided that there is no selective attrition during follow-up, as well as to differential misclassification of olfactory function during the (matched) prediagnostic phase of at-risk individuals (i.e., information bias). Disagreements were resolved during a consensus meeting with a third independent reviewer (SKLD). Studies were included if they fulfilled all predefined selection criteria (Supplementary Table 1). No restrictions were applied to study population selection criteria to ensure that both population-based as well as high-risk trait cohorts (e.g., selection based on presence of RBD) were included. All measures of olfactory dysfunction (such as identification and threshold) were considered. If there was more than one article in which a specific cohort was reported and analyzed (e.g., if updated estimates from the same cohort were published), the article with the most incident cases and if identical the study with the longest follow-up was included. If a study pooled multiple comparable cohorts, it was assessed if all individual cohorts met the inclusion criteria for separate inclusion. If not, the total pooled results from the original study were adopted in the current analysis if these met the inclusion criteria. With these measures, potential inflation of risk estimates due to insufficient cases per study is reduced. This applied to one study included in the present meta-analysis. [
In a standardized extraction form, the following variables from the included articles were collected: first author name, year of publication, manuscript title, cohort name, country of cohort, study design, study selection criteria (e.g., RBD), number of persons at risk, mean or median age at baseline, the proportion of women among those at risk, definition of olfactory dysfunction, olfactory testing instrument, mean or median duration of follow-up (after detection of olfactory dysfunction, i.e. set study length or until diagnosis), diagnostic criteria used for PD, number of individuals stratified by olfactory dysfunction and incident PD status (clinical diagnosis), age- and sex-adjusted odds ratio (OR) or hazard ratio (HR) and 95% confidence interval, number of cases of PD, and covariates used in regression model(s). If risk estimates were unavailable, the corresponding author of included articles was contacted to provide them to us. A study was interpreted as ‘high-risk’ if it included patients with any signs or symptoms other than olfactory dysfunction that are suggestive of PD, such as RBD or mild parkinsonian signs.
Quality assessment was performed for each study using a six-point modified Newcastle–Ottawa scale for cohort studies. Compared with the original scale, the following two items were removed because they were used as selection criteria for an article to be included in the meta-analysis [
] exposure was measured as part of the study protocol. Similarly, as we intended to include both studies with and without age and sex adjustment to analyze its modifying effect, we did not intend lack of adjustment to be part of the exclusion criteria. Given these modifications, a sum score was not calculated. A funnel plot and Egger's test were conducted to test for potential publication bias. Egger's test is a regression analysis of the study effect sizes on study precision, essentially returning a numerical value of the asymmetry in the funnel plot [
Unadjusted and age- and sex-adjusted ORs of olfactory dysfunction for incident PD were compiled from each included study. If no unadjusted OR was provided, an unadjusted OR was calculated based on the number of individuals stratified by olfactory dysfunction and incident PD status. Only in the meta-analysis regarding age- and sex-adjusted odds ratios, the study by Chen et al., which reported a hazard ratio, was combined with odds ratios from other studies [
]. Heterogeneity across studies was assessed using the I2 statistic. Random effects generic inverse variance models were used to meta-analyze unadjusted and, separately, age- and sex-adjusted ORs of olfactory dysfunction for incident PD. If statistical separation occurred (e.g., no normosmic individuals with subsequent incident PD), one individual was added to every stratum to avoid inflation of the OR. Across all analyses, p < 0.05 was considered to be statistically significant.
To explore whether the association of olfactory dysfunction with incident PD varied by study-level covariates, unadjusted ORs were analyzed with a univariate random-effects analysis by analyzing the standard error of the difference in the pooled unadjusted meta-analysis [
]. Study-level covariates were subdivided based on mean study age (≥66.8 or <66.8 years, [this cut-off reflects the median age across the included studies]), study percentage of females (≥47.1% or <47.1%, median percentage of females across included studies), mean study follow-up duration (≥6.2 or <6.2 years after olfactory testing, median follow-up across included studies).
To determine multivariate study-level covariates, a random-effects meta-regression analysis was conducted using the unadjusted odds ratios and the following covariates: age, sex and follow-up duration at the study level and the olfactory testing instrument. As a sensitivity analysis, a Knapp-Hartung modification was added to additionally account for heterogeneity [
]. In a separate sensitivity analysis, we added the study-level incidence rate of PD as a covariate to the meta-regression, to assess whether observed ORs were affected by the number of incident PD cases per year in a study. Analyses were performed using R packages meta, dmetar and ggplot2 [
From PubMed, the Cochrane Library and EMBASE, 2304 deduplicated articles were retrieved. After title and abstract screening, the full text of 57 articles was evaluated, after which 8 reports were included in the final meta-analysis (Supplementary Figure 1). Screening of referenced articles did not result in additional inclusions.
3.2 Characteristics and quality of included studies
Eight articles comprising twelve cohorts presented complete cohort analyses, while the study of Jennings et al. presented a nested case-control analysis. One report concerned five pooled cohorts (Depression_PD, EPI-PARK, PRIPS-Tübingen, PRIPS-Homburg and TREND) that were pooled accordingly in this study as not all individual studies met the criteria for separate inclusion [
]. In total, 8441 at-risk individuals were included in this meta-analysis. The average number of patients per study was 1055, and there were three studies with over 1000 individuals. Five studies were conducted in North-America, one in Germany, one in Italy and one in Austria. Study-level mean and median age at baseline were 68.6 and 66.8 years, respectively (range 60.0–79.7). Three studies included less than a third of women and three studies had female participation of 50% or more (range 0–53.8%, median percentage 47.1%). Five studies had a sample that was largely representative of the general population. Three studies reported high-risk cohorts: the study by Montgomery et al. only included individuals with mild parkinsonian signs, and the studies by Fereshtehnejad et al. and Mahlknecht et al. (2016) only included individuals with RBD. Follow-up ranged from 3 to 10 years, and six studies reported a follow-up duration longer than 5 years (range 5.7–10 years, median 6.2 years). Three studies also reported risk estimates for the follow-up period after the first 4 or 5 years of follow-up (e.g. in the article by Chen et al., for the first 5 years as well as the 5–9.8 year interval. The following olfactory testing instruments were used: B-SIT (Brief Smell Identification Test, n = 2), UPSIT (University of Pennsylvania Smell Identification Test, n = 3) and SS-12 or SS-16 (Sniffin’ Sticks Identification test, n = 3). The definition of olfactory dysfunction varied by study, although all studies used a customary measure of odor identification (Table 1). Each study provided an OR of the association of olfactory dysfunction with incident PD or crude numbers which facilitated the calculation of that OR. In addition, half of the studies provided a risk estimate (odds ratio [OR] or hazard ratio [HR]) for the association between olfactory dysfunction and PD, which was adjusted for age and sex. In Table 2 of the Supplementary File, an overview of the modified Newcastle-Ottawa scale is included, applied to the included studies. Representativeness of studies differed, as some studies included patients with RBD or mild parkinsonian signs. (Table 1). Notably, all studies reported complete follow-up or complete registration of all at-risk individuals lost to follow up. The funnel plot of risk estimates showed two estimates from high-risk cohorts on the far-right side and no estimates on the bottom left. (Supplementary Figure 2; Egger's test p = 0.22).
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Abbreviations: TREND (Tübingen evaluation of Risk factors for Early detection of NeuroDegeneration), PRIPS (Prospective evaluation of Risk factors for Idiopathic Parkinson's Syndrome), RBD (REM-sleep behavior disorder), HealthABC (Health, Aging and Body Composition), PARS (Parkinson Associated Risk Syndrome), SNH (substantia nigra hyperechogenicity). MPS: mild parkinsonian signs. EPIPARK: epidemiological non-motor parkinsonism symptoms study. DAT = dopamine transporter 1 Median age was used if mean age was not available.
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
Cut-offs of the subgroups are defined by the median follow-up, age and female percentage, respectively. CI = confidence interval, RBD = REM-sleep behavior disorder. 1 versus non-RBD * Odds ratio is pooled per subgroup for mean study follow-up, mean study age and study female percentage, and for RBD cohorts, high-risk cohorts and community-dwelling cohorts.
3.3 Association of olfactory dysfunction and incident PD
Significant heterogeneity in the meta-analysis of unadjusted ORs (I2 statistic 70.7%, p < 0.01) was found. The unadjusted pooled OR of olfactory dysfunction for incident PD was 4.18 [CI 95% 2.47-7.07]. Pooling the three high-risk cohorts resulted in an unadjusted OR of 9.39 [CI 95% 2.42-36.49] compared to the population-based studies with an OR of 2.75 [CI 95% 2.11-3.58] with p=0.08. The calculated age- and sex-adjusted pooled OR, based on the four studies with available data, was 3.12 [CI 95% 2.26-3.98]; by comparison, the pooled unadjusted OR in these four studies was 2.73 [CI 95% 2.09-3.57]. (see Fig. 1).
Fig. 1Hyposmia and incident Parkinson's disease: unadjusted (A) and age- and sex-adjusted (B) odds ratios across studies Unadjusted odds ratios (left) and age- and sex-adjusted odds ratios (right). The articles by Fereshtehnejad, Jennings, Mahlknecht (2016) and Montgomery did not contain age- and sex-adjusted odds ratios. Pooled unadjusted OR of all studies that reported an adjusted OR (2.73 [CI 95% 2.09–3.57]) was lower than the pooled adjusted OR (B).
Subgroup analyses of the association between olfactory dysfunction and PD are outlined in Table 2. The pooled unadjusted OR was higher in RBD (sub)cohorts (OR=5.75) than in other studies and even higher in all high-risk cohorts, which includes patients with mild parkinsonian signs (OR=9.39). The pooled unadjusted OR for a shorter-than-median follow-up duration (<6,2 years after diagnosis of olfactory dysfunction) was higher than that of studies with a longer follow-up (OR = 7.96 versus 2.73). OR was higher in studies with higher-than-median mean age of 66.8 years compared to studies with a lower-than-median mean age (OR = 4.70 versus 2.91). Pooled effects were larger for studies with a higher-than-median (≥47.1%) percentage of women (OR = 6.01 versus 2.69). These subgroup (univariate) differences were not statistically significant, although a trend for the association of olfactory dysfunction and follow-up duration was observed (p = 0.07), as well as for RBD studies versus community-dwelling cohorts.-
A multivariate random effects meta-regression demonstrated multiple statistically significant study-level covariates, namely study-level female percentage (regression slope parameter [β] = +0.053 [increase in logOR per unit increase in a study's female percentage]; p = 0.0006), shorter mean study follow-up duration (β = −0.56 for each increasing year of study follow-up; p = 0.0016), increasing mean study age (β = 0.2112 for every unit increase in mean study age; p = 0.0047) and RBD (β = 1.95 for exclusively RBD studies; p = 0.034). In Fig. 2, regression plots for sex and follow-up are shown. The type of olfaction testing instrument used did not significantly influence the risk estimates for incident PD in these studies, although a trend of lower OR for Sniffin’ Sticks studies was observed (β of [UPSIT or BSIT] vs. [Sniffin Sticks] = +1.17; p = 0.058). A sensitivity analysis using additional Knapp-Hartung modification resulted in an increase of p-values for mean study age (p = 0.096), mean study follow-up duration (p = 0.080) and mean study sex (p = 0.068). In a separate sensitivity analysis, study incidence of PD did not show a significant modifying effect on the observed OR (p = 0.46). An additional sensitivity analysis that adjusted for high-risk cohorts showed that this was an independent predictor of the association between olfactory dysfunction and incident PD (β = 1.49, p = 0.026). Although the covariates study age (p = 0.01) and gender (0.0023) were still significant, this was not true for study follow-up duration (p = 0.068).
Fig. 2Olfactory dysfunction and incident Parkinson's disease: Visualization of study-level covariates sex and follow-up duration.
In this study, we observed that having olfactory dysfunction is associated with an increased risk of incident PD. For individuals with olfactory dysfunction in the total pooled population, the unadjusted odds ratio of PD diagnosis was 4.18 compared to individuals without olfactory dysfunction. Importantly, however, the risk seems to vary across subgroups. In particular, in cohorts comprising a high percentage of women or containing exclusively individuals with confirmed RBD, olfactory dysfunction seems to convey an elevated risk of incident PD. In addition, this risk was higher in cohorts with a shorter follow-up duration after the initial establishment of olfactory dysfunction. In the next paragraphs, we carefully interpret these findings in a broader context.
There are three possible explanations for the observed higher risk estimate of olfactory dysfunction in studies with a higher percentage of women. First, olfactory dysfunction is more common in men than in women and its causes may be more heterogeneous, certainly in age groups at high risk of PD (>60 years) [
]. In particular, olfactory dysfunction may in some individuals reflect the prodromal phase of another neurodegenerative disease than PD, such as Alzheimer's disease [
]. Intriguingly, olfactory dysfunction may have higher predictive value for mild cognitive impairment in men than in women, indicating that the predictive value of olfactory dysfunction in men and women may differ between neurodegenerative diseases [
]. Large, prospective studies with robust monitoring and clinical assessment of incident PD and Alzheimer's disease are warranted to determine whether there are true sex differences in the association between olfactory dysfunction and incident neurodegenerative diseases. Second, the routes of spread of synucleinopathy may differ between men and women as a result of hormonal effects. However, previous research suggests that estrogen is likely to protect instead of harm olfactory bulb tissue [
]. Lastly, there is some debatable evidence that olfactory identification in women is superior to men's in neurodegenerative as well as physiological conditions, lowering the risk of olfactory identification for Parkinson's disease specifically [
], and our observation that the risk of PD associated with a prominent prodromal PD marker varies by the distribution of sex in a population highlights the need to consider sex-specific estimates in future prodromal PD criteria updates. Furthermore, we note that the sex-related meta-regression analyses were significantly influenced by the male-only Honolulu Study [
]. In studies that examined effect modification at an individual rather than a group level, effect estimates of olfactory dysfunction were actually higher among men, which further adds to our cautious interpretation of sex-specific differences in risk estimates [
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
The association between olfactory dysfunction and risk of PD after a shorter study follow-up period might indicate that olfactory dysfunction is indeed an imminent marker for PD. Interestingly, we observed that the risk may decrease with time if individuals remain free of clinical PD beyond the first years of follow-up. In those individuals, it seems more likely that olfactory dysfunction is caused by a different pathology. The observation of a time-varying association is congruent with the olfactory vector hypothesis as well as the Braak staging system, which suggest an ascending fashion of spread of α-synuclein in the human nervous system for at least a proportion of the PD population [
]. Another explanation for this finding should also be noted: the included studies with shorter follow-up durations more commonly consisted of cohorts at higher risk of PD. Indeed, a recent long-term study in patients with idiopathic olfactory dysfunction found that PD diagnosis was established 10.9 years on average after olfactory dysfunction onset [
]. Still, this observation highlights the need to consider time-varying effects in the prediction of incident PD in individuals with newly established olfactory dysfunction. Future studies could study this by including olfactory dysfunction as a time-dependent covariate in a survival analysis with PD diagnosis as primary outcome. In studies which assessed effect modification by disease duration at individual-level, there was no clear difference in effect estimates of olfactory dysfunction for incident PD between short- and long-term intervals. However, we also note that those studies were not specifically powered to address effect modification by disease duration [
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
We observed a substantially increased risk estimate of olfactory dysfunction for incident PD in studies comprising exclusively individuals with RBD, although the difference by RBD composition was statistically non-significant, likely due to a limited sample size. RBD is regarded as a sign of presence of neurodegeneration and a clear prodromal marker of PD, mostly in males because of high sex-specific prevalence of RBD. Therefore, its presence together with olfactory dysfunction should raise considerable suspicion for a prodromal state of PD [
]. Postuma et al. recently published a multi-center RBD cohort study in which olfactory dysfunction was the second best predictor for a combined outcome of dementia and parkinsonism after objective motor testing [
]. We hypothesize that in the case of RBD, an underlying synucleinopathy is the most likely cause of olfactory dysfunction, whereas in community-dwelling populations, a broader spectrum of causes may underlie this feature.
We hypothesized that the covariate study-level age was negatively associated with PD incidence in studies, as it is conceivable that young-onset olfactory dysfunction is likely to be more often pathological than older-onset olfactory dysfunction [
]. Instead, we observed a positive association between age and olfactory dysfunction. This finding might be explained by the fact that young-onset PD is more often caused by pathogenic mutations and earlier studies reported better olfaction in patients with Parkin and LRRK2 mutations compared to patients with idiopathic PD [
]. As previous research suggests that nonmotor symptoms in monogenic forms of PD might be less common, a less diffuse α-synuclein deposition might result in more isolated motor function-related symptoms [
Of note, three studies which are cited in the MDS research criteria for prodromal PD could not be included in this meta-analysis, either because the study's estimates compared the worst quartile of olfactory performance to only the best quartile of the population [
]. A key limitation of this study is that summary statistics were pooled from studies that were heterogeneous in several key aspects. As a consequence, our data provide insight on how the odds ratio of olfactory dysfunction for PD varies by the population mean of a key characteristic but do not yield characteristic-specific estimates. For instance, we show that the odds ratio is higher in studies with a higher percentage of women, but we could not calculate an odds ratio specific to women. In addition, studies were heterogeneous in definition of olfactory dysfunction and the choice of the olfactory function testing instrument. However, each of the instruments used assesses odor identification, and scores between UPSIT, B-SIT and SS-12 or SS-16 can reliably be transformed. [
] Therefore, we considered it admissible to meta-analyze data based on different instruments in this study. We also added olfactory testing instrument as a covariate in the meta-regression analyses and observed that it did not modify the observed results independently. We also note that analytical methods varied substantially across the included studies and some studies had a very low number of incident PD cases. To avoid inflation of risk estimates due to this, one study that pooled study results was included as the individual cohorts did not meet the inclusion criteria [
]. This does not take into account potential heterogeneity between these cohorts. Furthermore, summary statistics were used for (sub)group risk estimation, which meant that we had to make several statistical assumptions in pooling the data. Our pooled risk and subgroup estimates from high-risk cohorts, not based on individual patient data, should therefore be interpreted with caution. We consider these estimates to be an indication rather than robust quantification of true (biological) differences. An additional Knapp-Hartung sensitivity analysis, which further increases robustness in random effects meta-regressions with high heterogeneity compared to the DerSimonian-Laird estimator, increased p-values of age, sex and follow-up duration to non-significant values. Our observations should therefore encourage the initiation of collaborative efforts based on individual patient data to robustly quantify subgroup differences to allow for more personalized risk estimates [
]. Still, our observation of subgroup differences provides important novel insight on the prodromal phase of PD and set our study apart from an earlier meta-analysis on the association of olfactory dysfunction and PD [
This study suggests that olfactory dysfunction may signal imminent onset of clinical PD, especially in populations with a high proportion of women, higher age or individuals with RBD. Furthermore, the risk of clinical PD may subside over time if it does not occur in the first years of follow-up. Individual-patient data are needed to confirm these findings and to yield reliable, subgroup-specific risk estimates that could be used to refine criteria for prodromal PD and ultimately allow for earlier diagnosis and neuroprotective treatment.
Funding disclosures
The Center of Expertise for Parkinson & Movement Disorders was supported by a center of excellence grant by the Parkinson Foundation.
Bastiaan R. Bloem has received honoraria from serving on the scientific advisory board for Abbvie, Biogen, UCB, and Walk with Path, has received fees for speaking at conferences from AbbVie, Zambon, Roche, GE Healthcare, and Bial, and has received research support from The Netherlands Organisation for Scientific Research, the Michael J. Fox Foundation, UCB, Abbvie, the Stichting Parkinson Fonds, the Hersenstichting Nederland, the Parkinson Foundation, Verily Life Sciences, Horizon 2020, the Topsector Life Sciences and Health, and the Parkinson Vereniging.
Jules M. Janssen Daalen, Anouk Tosserams, Philipp Mahlknecht, Klaus Seppi and Sirwan K.L. Darweesh have no competing interests to declare.
Acknowledgements
The authors thank Sebastian Heinzel, PhD (Dept. Of Neurology, Christian-Albrechts University, Kiel, Germany) for his contributions to the completion of the dataset and the additional analyses performed.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Midbrain hyperechogenicity, hyposmia, mild parkinsonian signs and risk for incident Parkinson's disease over 10 years: a prospective population-based study.
00020-1&id=gr1.jpg){kind=link}
00020-1&id=gr2.jpg){kind=link}