Estimating Personalized Risk of Mild Cognitive Impairment in Parkinson's Disease through Comprehensive Risk Prediction Tools

Estimating Personalized Risk of Mild Cognitive Impairment in Parkinson's Disease through Comprehensive Risk Prediction Tools | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Estimating Personalized Risk of Mild Cognitive Impairment in Parkinson's Disease through Comprehensive Risk Prediction Tools Claire Pauly, Sonja R Jónsdóttir, Gabriel Martinez tirado, Anja Ophey, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9361292/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract Cognitive impairment (CI) is common in Parkinson’s disease (PD), affecting up to 30% of patients at diagnosis and increasing over time. This study evaluated and compared existing risk prediction tools for screening mild cognitive impairment (MCI) and assessing risk of subsequent cognitive decline in people with PD by validating and extending them with PD-specific and modifiable risk factors. Data from 201 PwPD and 231 controls from the Luxembourg Parkinson’s study were analyzed; 46.77% of PwPD had MCI. The Lifestyle for Brain Health (LIBRA) and preliminary disease Risk Estimator for Decline in Cognition Tool (pPREDICT) risk scores were evaluated and expanded to create a new tool called “ Modi fiable factors for cog nitive decline in PD” (ModiCogPD), and predictive performance was assessed using ROC–AUC, logistic regression, and Cox models. PD-specific scores (pPREDICT and ModiCogPD) showed better performance than LIBRA, with higher scores associated with increased risk of cognitive decline over time, while LIBRA was more informative in early disease stages. Health sciences/Diseases Health sciences/Neurology Biological sciences/Neuroscience Health sciences/Risk factors Figures Figure 1 Figure 2 Introduction Parkinson’s disease (PD) is the fastest growing neurological disorder worldwide and is estimated to reach 14.2 million cases by 2040 1 . People living with PD (PwPD) experience both motor symptoms - such as tremor, gait disorders and rigidity - and non-motor symptoms including hyposmia, depression and Parkinson’s disease–related mild cognitive impairment (PD-MCI) 2 , 3 . Given the increasing incidence of PD cases, there has been a heightened interest in understanding and addressing the disease's non-motor symptoms, particularly cognitive impairment (CI). This is attributable to the profound impact of CI on quality of life of those affected, often exacerbating caregiver load and healthcare expenses 4 , 5 . In a multicentric survey, CI has been listed as one of the top 10 research priorities for the management of PD 6 and as one of the most important unmet needs in PD 7 . PD-MCI may already be present in up to 30% of PwPD at the time of diagnosis 8 . This proportion increases to 40–50% after 5 years 9 , 10 , and up to 75% of PwPD develop dementia within 20 years of diagnosis 11 – 13 . CI and PD-MCI may be characterized by worsening of cognitive functioning in one or more cognitive domains, such as attention, memory and executive functioning. Despite the absence of disease-modifying treatments, it is imperative to identify PwPD at risk for developing PD-MCI and potentially dementia, thereby enabling early detection, diagnosis and management, and most importantly improve their quality of life as non-pharmacological interventions such as cognitive training 14 and physical activity may support cognitive functioning in early stages 15 , 16 . Several tools have been developed to estimate the risk of cognitive decline in PwPD. On the one hand, complex tools such as those based on Dopamine Transporter (DaT) imaging results, cerebrospinal fluid (CSF) and genetic data, are associated with higher costs and time constraints, leading to a reduced scalability for clinical practice. On the other hand, simple tools based on risk factors, such as lifestyle, demographics, patient reported outcomes and routinely collected clinical data can be easily implemented in the daily clinical work 17 – 19 . Given the growing evidence of the impact that lifestyle factors play in the development of CI, risk factor-based tools have gained increasing relevance in the field. Overall, comprehensive risk estimation tools with sufficient specificity for PwPD remain limited, particularly regarding the integration of modifiable risk factors. Existing approaches include PD-unspecific tools such as the Lifestyle for Brain Health (LIBRA) 20 , as well as PD-specific instruments like the Montreal Parkinson’s Risk of Dementia Scale (MoPaRDS) 21 . However, these tools do not yet provide a comprehensive framework that extensively captures both PD-specific characteristics and modifiable risk domains. Evidence from the literature indicates that addressing modifiable risk factors may prevent up to 45% of dementia cases 22 , with new updates suggesting that this figure could grow up to 65% 23 . Furthermore, incorporating PD-specific modifiable risk factors into these predictive tools could enhance their accuracy, clinical relevance, and overall utility. Estimating individualized risk of cognitive decline may help inform risk-adapted counseling and monitoring strategies, boosting early preventive interventions and inclusion in clinical trials, providing a strong rationale for developing and refining predictive tools. In line with this need for more comprehensive and PD-tailored risk stratification tools, Carlisle and colleagues (2023) developed the preliminary disease Risk Estimator for Decline in Cognition Tool (pPREDICT) 24 by including PD-specific modifiable and/or treatable risk factors for cognitive decline, such as REM Sleep Behaviour Disorder (RBD), motor severity, depression, anxiety, excessive daytime sleepiness, obesity, physical activity and several vascular risk factors. Some of these risk factors overlap with the overall 14 modifiable risk factors that account for up to 45% of worldwide dementia cases 22 , and which may prevent or delay the onset of dementia, as well as positively influence cognitive outcomes in PwPD 22 , 25 . However, several potentially relevant risk factors proposed by the Lancet Standing commission (Livingston et al ., 2024), such as diabetes, apathy, exposure to toxins and solvents, and alcohol consumption, were not included in previous prediction tools, despite evidence suggesting that they may influence cognition in PwPD 26 . This underscores the relevance of expanding the pPREDICT to develop a new prediction tool for PD-MCI, which incorporates these potentially modifiable dementia risk factors. In the present study, we developed an extended CI risk score for PwPD named ModiCogPD (“ Modi fiable factors for cog nitive decline in PD ”), including the factors mentioned previously. As few studies evaluated the existing multidimensional risk prediction tools (e.g LIBRA, pPREDICT), particularly within the same well-characterized PD cohort, we investigated whether the two established dementia risk scores (LIBRA, pPREDICT) and the newly developed ModiCogPD can predict PD-MCI in the Luxembourg Parkinson’s Study 27 , a population-based, longitudinal, monocentric study. Furthermore, we evaluated their prognostic value and their association with subsequent cognitive decline over a follow-up period of up to 8 years. Finally, we aimed to identify modifiable lifestyle factors associated with increased cognitive risk that may allow for future risk-adapted preventive strategies in PwPD. Results Descriptive statistics of the study data A total of 201 PwPD and 231 control participants from the Luxembourg Parkinson’s Study (Hipp et al ., 2018) met the eligibility criteria for analysis. Both groups present significant differences in socio-demographic and clinical characteristics at the baseline visit, except for the total years of education (Table 2 ). Further details on the individuals risk factors included in the risk scores are given in the Supplementary table 1 . Table 2 Socio-demographic, clinical characteristics and risk score of PwPD ,PD-NC, PD-MCI and control participants at the baseline visit. Numerical variables are reported as mean and SD, together with minimum and maximum values (Min–Max). Categorical and ordinal variables are described by count (n) and percentage. Group differences between PwPD and control participants were assessed using appropriate statistical tests appropriate to the distribution of each variable. Continuous variables were analyzed using either two-tailed t-tests or Mann–Whitney U tests, whereas categorical variables were assessed using Pearson’s chi-squared test. Statistical significance was set at α = 0.05. Variables PD (n = 201) Mean (std); Min-Max or n(%) PD-NC (n = 107) Mean (std); Min-Max or n(%) PD-MCI (n = 94) Mean (std); Min-Max or n(%) Control (n = 231) Mean (std); Min-Max or n (%) P -value (PD vs Control) P -value (PD-NC vs PD-MCI) Socio-demographics Age (years) 64.25 (10.38); 38–89 60.95 (9.79); 38–82 68.01 (9.79); 41–89 59.83 (11.4); 22–80 p < 0.001 p < 0.001 Sex (female) 58 (28.86%) 36 (33.65%) 22 (23.40%) 90 (48%) p < 0.001 p = 0.11 Years of Education 13.84 (3.63); 6–26 14.53 (3.55); 7–25 13.04 (3.57); 6–26 14.43 (3.61); 6–24 p = 0.105 p = 0.003 Clinical Global Cognition (MoCA: 0–30) 25.85 (2.80); 18–30 27.03 (2.18); 21–30 24.52 (2.85); 18–30 28.36 (1.31); 23–30 p < 0.001 p < 0.001 Motor Symptoms (MDS-UPDRS III: 0-109) 29.91 (13.70); 3–75 28.04 (13.07); 6–75 32.05 (14.14); 3–74 3.05 (4.26); 0–30 p < 0.001 p = 0.054 Depression (BDI-I: 0–63) 8.95 (6.35); 0–45 9.15 (6.03); 0–30 8.73 (6.73); 0–45 5.42 (5.03); 0–29 p < 0.001 p = 0.482 Apathy (SAS: 0–42) 12.99 (5.35); 1–33 12.16 (5.37); 1–32 13.94 (5.19); 2–33 9.57 (4.49); 0–27 p < 0.001 p = 0.015 Hoehn and Yahr (Disease staging) 1.0: 21 (10.45%) 1.5: 18 (8.96%) 2.0: 117 (58.21%) 2.5: 25 (12.45%) 3.0: 18 (8.96%) 4.0: 2 (1%) 1.0: 15 (14.02%)) 1.5: 15 (14.02%) 2.0: 61 (57.01%) 2.5: 8 (7.47%) 3.0: 8 (7.47%) 1.0: 6 (6.39%) 1.5: 3 (3.19%) 2.0: 56 (59.57%) 2.5: 17 (18.09%) 3.0: 10 (10.64%) 4.0: 2 (2.13%) - - p < 0.001 Disease Duration (in years) 4.58 (4.96); 0–29 3.88 (3.74);0–19 5.39 (5.99); 0–29 - - p = 0.257 Levodopa Equivalent Daily Dose (in mg) 603.50 (388.25); 25-1854 598.38 (378.84); 100–1854 609.28 (400.76);25-1850 - p = 0.021 p = 0.856 Risk Factors LIBRA 3.57 (2.327);0-11.5 3.23 (2.386) 0-11.5 3.95 (2.210) ;1–11 2.74 (1.94); -1-6 p < 0.001 0.017 pPREDICT 4.92 (2.099); 0–10 4.16 (1.953);0–10 5.78 (1.930) ;0–10 2.67 (1.54); 0–6 p < 0.001 p < 0.001 ModiCogPD 11.61 (3.291) 5–23 10.86 (3.14) ;6–23 12.47 (3.258) ;5–20 8.34 (2.30); 4–17 p < 0.001 p < 0.001 Performance of the different risk scores LIBRA presented the lowest performance in discriminating between PD-NC and PD-MCI (AUC = 0.596 ± 0.032), followed by the ModiCogPD with an AUC of 0.645 ± 0.063. The risk score with the highest performance was the original pPREDICT reaching an AUC score of 0.717 ± 0.068 (Fig. 1 ). A statistical comparison revealed no significant difference between LIBRA and the ModiCogPD score (p = 0.24); however, significant differences were observed between the pPREDICT score and both LIBRA (p < 0.001) and the ModiCogPD (p = 0.006) score. To further investigate potential improvements in the predictive performance of the ModiCogPD for cognitive status, a LR model was developed incorporating the individual subitems that compose the risk score. The LR-based ModiCogPD model demonstrated superior discrimination compared to the global ModiCogPD score, achieving an AUC of 0.723 ± 0.045 (Fig. 1 ). This performance was slightly higher than that of the pPREDICT model. . The internal validation conducted within the bootstrapping framework, showed for the ModiCogPD model that optimism correction reduced the AUC apparent from 0.831 to 0.722, suggesting the presence of small overfitting (Table 3 ). Calibration was acceptable, with a near-zero intercept and a slope close to 1 after correction, indicating that predicted probabilities were well-aligned with observed outcomes. Table 3 Performance overview of the LR-based ModiCogPD. This table presents the performance of ModiCogPD both as a global risk score and as a LR-based ModiCogPD. For the latter also calibration slope, intercept after correction, apparent AUC and optimism-corrected area under the curve (AUC) are reported. Risk scores Global Risk Score LR-based model Discrimination AUC score Discrimination AUC score Calibration slope- correct Calibration intercept Apparent AUC Optimism- corrected AUC ModiCogPD 0.645 ± 0.063 0.723 ± 0.045 0.977 0.052 0.831 0.722 Identification of the most relevant risk factors Next, we evaluated the independent associations of ModiCogPD individual risk factors with cognitive status (Table 4 ). The most relevant risk factors for the prediction of the cognitive status ( p -values < 0.05) included age, disease duration, MDS-UPDRS III, apathy, alcohol consumption and tremor/PIGD phenotype, with a more prominent PIGD phenotype being associated with increased risk. Among these, only apathy and alcohol consumption are considered modifiable traits that can be targeted to slow down the progression towards CI. Other modifiable risk factors that presented high odds without reaching statistical significance, included vigorous physical activity (OR = 1.943, p = 0.160) and vascular risk (OR = 1.703, p = 0.201). Moreover, sex showed a trend-level association with cognitive status. Table 4 Odds-ratios of individual risk factors for predicting CI status. The table reports regression coefficients, odds ratios (OR) and the respective confidence interval, and p -values for each variable included in the model. Marked in bold are those p -values that are significant. Variable Coefficient Odds Ratio 95% interval p -value Age at Assessment 0.960 2.613 1.288–5.302 0.008 Sex -0.821 0.440 0.172–1.127 0.087 Years of Education -0.545 0.580 0.298–1.126 0.108 Disease Duration 0.704 2.023 1.056–3.875 0.034 RBDSQ 0.106 1.112 0.648–1.907 0.701 MDS-UPDRS-III 0.679 1.973 1.016–3.830 0.045 BDI-I Total Score -0.329 0.720 0.332–1.561 0.405 Anxiety Disorder 0.083 1.087 0.349–3.384 0.885 BMI 0.010 1.010 0.658–1.549 0.965 Vigorous Physical Activity 0.664 1.943 0.769–4.911 0.160 Moderate Physical Activity -0.053 0.949 0.242–3.723 0.940 Vascular Risk 0.532 1.703 0.754–3.848 0.201 Starkstein Apathy Scale 0.591 1.806 1.053–3.095 0.032 MDS-UPDRS I -0.006 0.994 0.462–2.136 0.987 MDS-UPDRS II -0.520 0.595 0.269–1.312 0.198 MDS-UPDRS IV -0.407 0.666 0.427–1.037 0.072 Family History of PD 0.183 1.200 0.543–2.653 0.652 Family History of Dementia -0.198 0.820 0.330–2.035 0.669 Sniffin Sticks -0.038 0.963 0.857–1.081 0.520 Caffeine Consumption -0.849 0.428 0.118–1.553 0.197 Alcohol Consumption 1.497 0.224 0.058–0.866 0.030 Smoking -0.156 0.856 0.397–1.844 0.691 Contusion 0.019 1.019 0.434–2.394 0.965 Pesticides Exposure -0.189 0.828 0.368–1.859 0.647 Toxicants Exposure 0.178 1.195 0.550–2.599 0.653 Marital Status 0.294 1.341 0.487–3.699 0.570 Number of Children 0.445 1.561 0.754–3.232 0.230 Total Number of Spoken Languages 0.134 1.144 0.547–2.391 0.721 Diabetes - Medical History -0.933 0.393 0.109–1.421 0.155 Tremor/PIGD Phenotype -0.474 0.623 0.428–0.907 0.014 Performance of the Weighted ModiCogPD As an extension of the previous LR analyses incorporating subitems to enhance discriminatory performance, we further examined alternative weighting procedures in the development and improvement of the ModiCogPD risk score performance. Following the improved performance showed by the LR-based ModiCogPD model, we further investigated whether given clinically meaningful, integer-based weights to individual risk factors, aimed at facilitating its applicability in routine clinical practice, could enhance predictive performance, a weighted ModiCogPD score was derived using regression coefficients (see Methods subsection ‘Weighted ModiCogPD’), from the 50% training dataset. The resulting weighted score was evaluated in the remaining 50% testing subset, achieving a discrimination of AUC = 0.59, whereas the non-weighted approach yielded an AUC of 0.61 within the same testing sample (Supplementary Fig. 1). Further information on the weighting procedure is given in the supplementary material (Supplementary table 2). Predictive strength on the posterior cognitive decline trajectories The prognostic value of the three global risk scores was evaluated by examining their association with the risk and timing of cognitive decline, defined as progression to MoCA ≤ 21 for the newly diagnosed PwPD or MoCA < 26 for the global PwPD population (Fig. 2 ). The prognostic value of the risk scores in the general population showed heterogeneous associations with the hazard of cognitive decline over the course of the disease. ModiCogPD demonstrated a significant positive association, with a hazard ratio of 1.245 (95% CI: 1.084–1.430; p-value = 0.002), indicating that each one-unit increase in the score was associated with a 25% higher risk of cognitive deterioration. Similarly, pPREDICT showed a comparable effect size (HR = 1.254, 95% CI: 1.006–1.564; p-value = 0.0445), although the association was of borderline statistical significance. In contrast, LIBRA was not significantly associated with cognitive decline (HR = 1.029, 95% CI: 0.849–1.246; p-value = 0.771). By contrast, analyses performed on the newly diagnosed subgroup revealed that only the LIBRA score was significantly associated with the risk and timing of cognitive decline (HR = 1.136, 95% CI: 1.005–1.284; p-value = 0.042). Additional information is available in Supplementary table 3. Risk scores associated with cognitive domain impairment The study association between the set of risk scores and the presence of CI in a specific domain showed heterogeneous results across the risk scores. With LIBRA showing no statistical significance with the presence of impairment in any of the cognitive domains after adjusting for age, sex and years of education. However, pPREDICT and ModiCogPD showed significance in both attention (HR = 2.182, 95% interval: 1.226–3.881, p = 0.008; and HR = 1.841, 95% interval: 1.104–3.071, p = 0.019) and executive domains (HR = 3.137, 95% interval: 1.73–5.69, p = 0.0001; and HR = 1.795, 95% interval: 1.082–2.977; p = 0.024), respectively (Table 5 ). Table 5 Association between the risk scores and cognitive domain impairment. Cognitive Domains LIBRA OR [95% interval], p-value pPREDICT OR [95% interval], p-value ModiCogPD OR [95% interval], p-value Attention 1.226 [0.735–2.049] 0.435 2.182 [1.226–3.881] 0.008 1.841 [1.104–3.071] 0.019 Executive 1.112 [0.678–1.824] 0.673 3.137 [1.730–5.692] p < 0.001 1.795 [1.082–2.977] 0.024 Memory 0.906 [0.552–1.487] 0.695 1.420 [0.825–2.443] 0.207 1.417 [0.864–2.326] 0.168 Visuospatial 0.853 [0.488–1.489] 0.575 1.167 [0.645–2.115] 0.610 0.837 [0.479–1.465] 0.533 Language 0.737 [0.358–1.517] 0.408 1.641 [0.771–3.495] 0.199 1.457 [0.737–2.885] 0.280 The table shows the odds ratios (OR) with 95% confidence intervals derived from multivariable LR models examining the association between LIBRA, pPREDICT, and ModiCogPD risk scores and specific cognitive domains impairment (attention, executive, memory, visuospatial, and language). All models were adjusted for age, sex, and years of education. Discussion In this study, we evaluated and compared the performance of two established and newly developed global risk scores for CI in PwPD (pPREDICT, LIBRA and ModiCogPD), focusing on their ability to predict PD-MCI and subsequent cognitive decline. In parallel, we extended existing risk scores by incorporating additional modifiable risk factors relevant in PD (ModiCogPD) and identified the most relevant modifiable factors associated with increased CI risk in PD, with the aim of informing future risk-adapted precision preventive strategies in this population. Overall, pPREDICT showed the highest predictive performance for detecting PD-MCI at the global score level, followed closely by ModiCogPD, whereas LIBRA showed the lowest discrimination. Additionally, we investigated the contribution of individual risk factors within ModiCogPD by developing an LR-based model to assess whether differential weighting of these individual risk factors could improve the performance. This approach resulted in improved discrimination compared to the global risk score and outperformed the other models (LIBRA and pPREDICT), achieving the highest predictive accuracy. Both findings suggest that PD-specific risk scores incorporating both disease-related and modifiable factors may provide superior discrimination in PD-MCI screening compared to more general dementia risk tools such as LIBRA. Longitudinal analyses further reinforced the prognostic utility of PD-specific scores, where higher pPREDICT and ModiCogPD scores were associated with an increased hazard of cognitive decline over time in the whole study population, however, LIBRA did not significantly predict cognitive decline. This finding underscores the overall prognostic value of PD-specific multidimensional tools in estimating the risk of cognitive decline over time in a heterogeneous PD population. Interestingly, a different pattern emerged in newly diagnosed PwPD, where a higher LIBRA score was the only one associated with an increased hazard of cognitive decline over time. LIBRA was originally developed to estimate dementia risk in the general population (Schiepers et al., 2018), whereas pPREDICT and our extended version, ModiCogPD, were designed to assess cognitive decline risk specifically in PD 24 . Even though the PD-specific tools also include general risk factors, their main focus is rather on disease-related characteristics. In early PD, when motor and non-motor features may still be mild or relatively homogeneous across patients, these disease-specific variables may be less meaningful. In this context, broader risk factors, such as vascular and lifestyle-related characteristics, which are more central to LIBRA’s construction, may have a relatively stronger influence. As the disease progresses and PD-related burden becomes more pronounced, disease-specific characteristics may play a larger role in determining cognitive decline outcomes. This difference in model focus and stage-dependent expression of risk factors may help explain the variation in predictive performance across the subgroups. These findings indicate that in the early stages of PD, disease-specific variables may contribute less, highlighting the prognostic value and relevance of general risk factors in the early stages when PD characteristics are not yet fully present. Together, these results support the value of PD-tailored, multidimensional risk score tools for screening CI in pre-dementia stages, such as in PD-MCI, as well as highlighting their prognostic value during the course of the disease. The evaluation of the individual risk factors of ModiCogPD enabled us to determine which ones were the most relevant in distinguishing between PD-MCI and normal cognition, emphasizing the role of both modifiable and PD-specific variables. In our analysis, age, disease duration, motor severity (MDS-UPDRS III), apathy, alcohol consumption, and the tremor/PIGD phenotype showed a statistically significant association with CI. These findings are in line with previous studies reporting the importance of age, motor, and non-motor symptoms such as depression in cognitive decline in PD 28 . A systematic review by Guo and colleagues 29 identified advanced age, genetic variation in APOE and MAPT, gait disturbance (PIGD), motor assessments (MDS-UPDRS-III) and non-motor symptoms such as hallucinations, orthostatic hypotension and anxiety as predictors for CI in PD. Other longitudinal studies reported GBA1 risk variants as predictors for cognitive decline in PD 30,31 . Additionally, GBA1 p.E365K variant increases the risk of developing CI in PwP with normal cognition at baseline 32 . Another meta-analysis revealed several risk factors associated with PD-MCI namely older age, lower education, longer disease duration, higher levels of LEDD, more severe motor symptoms, PIGD phenotype, as well as poorer quality of life, higher levels of apathy and depression 33 . These findings reinforce the multifactorial nature of CI in PD and support the integration of multidimensional risk factors encompassing demographics, motor and non-motor symptoms, as well as potentially modifiable factors within risk prediction tools. Moreover, sex exhibited trend-level associations with cognitive status, indicating that sex-specific differences may be relevant in the application of risk score tools. Among the significant risk factors identified in our study, apathy and alcohol consumption represent potentially modifiable factors, which may constitute feasible targets for interventions aimed at slowing cognitive deterioration. Since apathy is very common in PD but often overlooked in clinical practice due to its subtle nature 34 , healthcare professionals should be aware of its association with CI and the potential pharmacological and non-pharmacological treatments. Exercise interventions, mindfulness exercises, cognitive behavioural therapy as well as repetitive transcranial magnetic stimulation are non-pharmacological treatments that could be used in PwP with apathy 35 , 36 . To our knowledge, no previous study has shown an association between alcohol consumption and CI in PD although excessive alcohol consumption has been identified as a modifiable risk factor for dementia by the report of the Lancet standing Commission, 2024)21. Existing literature reviews of alcohol use disorder link chronic excessive alcohol exposure to oxidative stress, neuroinflammation, excitotoxicity, neuronal loss, and increased risk of dementia and PD 37 . They primarily addressed PD as a neurodegenerative endpoint and did not assess alcohol-related CI within PD cohorts 37 . More studies are warranted to investigate the link between alcohol consumption and CI, and its underlying mechanism in PD with particular attention to potential differences in men and women. Other factors, such as vigorous physical activity and vascular risk factors, presented higher odds ratios without reaching statistical significance. This may be related to limited statistical power; however, these factors should be further explored in larger or independent cohorts. In particular, hypertension and diabetes type 2 should be further investigated as they are commonly associated with cognitive decline in PD and can even lead to PD with dementia 38 . In our study, we focused on vigorous physical activity to prevent CI; however, it has been shown that even mind-body exercise such as Tai Chi, Qigong, yoga and dance can improve CI in PD 39,40 . With growing evidence for the beneficial effect of physical activity, healthcare professionals and clinicians should recommend exercise as routine management for PwPD as well as specialized physiotherapy 41 , 42 . In addition, the specific role of physical activity and vascular risk factors, requires further investigation in larger cohorts to better quantify their relevance in the assessment of risk factors for PD-MCI. Taken together, the identification of potentially modifiable factors in our cohort underscores the importance of prevention-oriented strategies in PD. Accordingly, prevention efforts should focus on routine cognitive and risk assessment, standardized risk communication and education, multidomain risk-reduction interventions, and cognitive and physical training, supported by interdisciplinary collaboration across PD care networks 26 . However, public awareness of modifiable dementia risk factors remains limited. A recent survey conducted in Ireland found that although 65.6% of participants believed lifestyle changes can lower dementia risk, only 31.4% considered dementia preventable with knowledge levels varying across demographic groups 43 . These findings highlight the need for targeted and personalized prevention strategies that also take into account sociodemographic determinants of knowledge, i.e. education, sex, and age 44 . In this context, structured prevention programs, such as the “programme dementia prevention (pdp)” initiated in Luxembourg 45 , 46 , may help translate risk knowledge into actionable interventions. To contextualize these findings, it is important to consider the overall prevalence of PD-MCI within our cohort, as this reflects the clinical relevance of the observed risk associations. PD-MCI is a frequent non-motor symptom in PD and is considered as one of the most burdensome symptoms of PD. By applying the MDS-PD MCI Level II criteria with a z-score cut-off of ≤ -1.5 SD, to our PD cohort, PD-MCI was present in 46.77% of PwPD, in line with other cohorts 47 . The prevalence of PD-MCI in the literature is very heterogeneous ranging from 19 to 69.3% in PwPD 48 , due to different assessment tools applied to detect CI as well as time of assessment. Even within the same diagnostic framework, such as the MDS PD-MCI Level II criteria, prevalence estimates can differ considerably, as the recommended impairment threshold may range from − 1 to − 2 SD, leading to variability in the proportion of patients meeting the PD-MCI criteria. Given the early-moderate disease stage of the PwPD in our cohort and the criteria applied to define PD-MCI, the prevalence of CI is representative and comparable with other PD cohorts with similar age ranges and disease duration 47 . Beyond evaluating the screening potential for PD-MCI, we also explored the association between the different risk scores and the specific cognitive domains, which may provide additional insights into the early patterns of cognitive decline in PD. The association between risk scores and specific cognitive domains was heterogeneous. LIBRA showed no significant associations with any cognitive domain, whereas both pPREDICT and ModiCogPD were significantly associated with attention and executive function. No risk score was significantly associated with memory, visuospatial, or language domains. These results suggest that PD-specific scores such as pPREDICT and ModiCogPD may better capture cognitive difficulties related to attention and executive functions, which are known to occur early during the disease course 49 , 50 . Therefore, the ability of these prediction tools to identify early patterns of impairment may support earlier screening for PD-MCI and help guide focused monitoring and preventive strategies targeting the cognitive domains most at risk. Some limitations should be acknowledged in the present study and can be mainly grouped into data-related, or more methodological considerations. Regarding limitations of the underlying data, the absence of neuropsychological normative values specific to the Luxembourgish population represented a significant challenge. To address this, regression-based methods were employed to generate normative values from the control group. Although this approach has been previously described in the literature, it is less robust than using established norms derived from large, representative population samples. This limitation, together with the relatively small number of healthy controls, may have introduced additional variability in the z-score calculations and, consequently, in the classification of PwPD with CI. Further variability may have arisen from the linguistic diversity of the Luxembourgish population, which has three official languages, German, Luxembourgish and French, as well as a big community whose first language is Portuguese. Consequently, some participants completed neuropsychological testing in a non-native language. Moreover, it is not possible to completely isolate and assess a single cognitive domain, as most neuropsychological assessments require the involvement of multiple cognitive domains. As a result, neuropsychological tests may be classified differently across cognitive domains, and alternative grouping strategies may yield different domain-specific outcomes, potentially influencing the interpretation of cognitive profiles. Consistent with this, a recent scoping review shows that, although cognitive domains are central to diagnosing cognitive impairment in PD, there is limited empirical support for domain structures derived from dimensionality reduction of cognitive test data 51 . Comparisons across studies are hindered by heterogeneous statistical methods and PD sample characteristics. Finally, participation in the extended neuropsychological assessment was voluntary rather than systematically applied across the entire cohort, which may have introduced selection bias. Limitations also arise from the application of the MDS PD-MCI Level II diagnostic criteria. In particular, threshold selection for CI is a key methodological factor contributing to heterogeneity in PD-MCI classification. Although cut-offs from − 1 to -2 SD are used in the literature, the − 1.5 SD threshold was adopted, as recommended by Dalrymple-Alford et al. for the most balanced PD-MCI detection 52 . In addition, although the cognitive testing criteria was applied, it was not possible to fully implement the whole PD-MCI criteria proposed by 53 which require 1) evidence of gradual cognitive decline, due to the lack of consistent longitudinal data across the whole cohort, and 2) preserved activities of daily living (ADL), due to the absence of detailed informant-based ADL assessments across all participants and visits. Furthermore, only one neuropsychological test was available within the visuospatial domain, preventing full adherence to the MDS Level II criteria as originally described by Litvan et al. 53 . Another available risk score for PD is the Montreal Parkinson’s Risk of Dementia Scale (MoPaRDS). Similarly to the pPREDICT, it incorporates mostly PD-specific variables but does not focus on modifiable risk factors. However, we were not able to evaluate this specific prediction tool, due to specific variables, such as Bilateral disease onset not being available in our cohort so a direct comparison with it was not feasible. Similarly, minor adaptations were necessary when calculating the LIBRA score 54 , as the exact clinical variables originally used were not available; therefore, alternative assessments/questionnaires assessing comparable domains were selected. An updated version of LIBRA was published by Rosenau (2024), highlighting hearing impairment, social contact, and sleep—already included in our model—as risk factors with strong evidence for the development of dementia. Hearing impairment and social contact, particularly aspects related to loneliness and social engagement, were not included in our analysis due to the lack of available measures to assess these domains within the LuxPArk cohort. Moreover, LIBRA was originally developed to predict dementia, an outcome that could not be directly evaluated in our cohort, as no participants met the PD dementia criteria defined by 55 . In contrast, the ModiCogPD score focuses on cognitive decline defined as a worsening in cognitive performance of -1.5 SD reflecting a distinct conceptual and clinical target. Finally, the modifiable risk factor “Alcohol consumption”, included in the ModiCogPD, refers to the history of consumption and not per se to the current levels of consumption, therefore cautiousness is needed when interpreting its relevance in CI. These limitations should be taken into account when interpreting comparative performance across risk scores. Besides clinical characteristics, there are additional limitations that should be acknowledged, such as the lack of external validation for the ModiCogPD risk score, as successful replication in independent cohorts would further ensure its predictive and prognostic value and strengthen methodological confidence in the findings. The small size of the study population and the number of independent variables in the different LR models, could represent another limitation, making the models more prone to overfitting and unstable estimates; which was addressed by applying L2-regularisation. In conclusion, this study compared general and PD-specific risk scores for identifying CI and estimating the risk of future cognitive decline in PD. Our results show that PD-specific scores performed better overall in a mixed PD population, while a general lifestyle-based score (LIBRA) may be more relevant in the earliest stages of the disease. The identification of modifiable risk factors also highlights potential targets for prevention and closer clinical monitoring 26 . Overall, these findings support the use and further improvement of practical risk scores to help detect CI earlier and guide more personalized care in PD. Methods Participants 201 PwPD and 231 control participants from the Luxembourg Parkinson’s Study 27 with detailed neuropsychological assessment were included in the analyses. Participants were recruited in Luxembourg between 2015 and 2023 and provided written informed consent. The collection was approved by the National Ethics Board (CNER Ref: 201407/13) and Data Protection Committee (CNPD Ref: 446/2017). PwPD with severe CI defined as a MoCA total score < 18 at baseline were excluded from the analysis 56 . Missing values in the study population were handled differently across different variables. Participants with missing values in the neuropsychological tests were removed from the study population. While the risk factors were imputed following the most suitable methodology depending on its nature of missingness (for further details see subsection “ Risk prediction tools ”). The study population included a subset of newly diagnosed PwPD (n = 90), defined by a baseline visit within two years from diagnosis 57 . Moreover, we excluded control participants meeting criteria for PD-MCI at baseline defined as a Montreal Cognitive assessment (MoCA) total score < 26 58 , as well as participants reporting neurological comorbidities such as brain tumor, meningitis/encephalitis, multiple sclerosis, normal pressure hydrocephalus, seizures, stroke and traumatic brain injury. Neuropsychological test battery A detailed neuropsychological test battery was conducted at baseline and at yearly follow-up assessments. Based on the Movement Disorder Society (MDS) PD-MCI Level II criteria 53 , two different tests per cognitive domain were included to assess potential CI except for the visuospatial domain where only one test was available. (1) attention and working memory: Trail Making Test part A and digit span forward (2) executive functions: Frontal Assessment battery, Trail Making Test B-A (3) language: letter fluency (f), category fluency (animals) (4) memory: Word list of Consortium to Establish a Registry for Alzheimer's Disease (CERAD) Delayed Recall, Word list intermediate recall (CERAD). (5) visuospatial functions: Benton’s Judgment of Line Orientation Cognitive Impairment Definition Since population-specific normative data were not available in Luxembourg, we derived normative values from the control subgroup included in the Luxembourg Parkinson’s Study following an approach described by 59 and previously applied to the Luxembourg Parkinson’s Study 60 . PwPD were classified as having normal cognition (PD-NC) or PD-MCI according to the MDS PD-MCI Level II criteria 53 , 61 , having two impaired tests, coming from the same cognitive domain or different ones. Neuropsychological test impairment was defined as a z-score ≤ − 1.5 standard deviations (SD) relative to age-, sex-, and education-adjusted normative data. Risk Prediction tools The risk scores considered in this analysis included the LIBRA, for dementia prediction in the general population, 20,62 , and the pPREDICT, aimed to estimate cognitive decline in PD, 24 . In addition to the modifiable and non-modifiable risk factors, from Carlisle and colleagues 24 , several other risk factors were included in our proposed ModiCogPD risk score (Table 1). In order to handle missing values, the missingness mechanisms of the study variables were assessed based on observed patterns and clinical knowledge to evaluate the Missing Completely At Random (MCAR), Missing At Random (MAR), or Missing Not At Random (MNAR) assumptions. The subset of variables used to calculate the different risk scores had missingness rates between 0% and 3%. Variables exhibiting missingness patterns consistent with MNAR, such as the Sniffin’ Sticks Smell test score, were not imputed to avoid introducing systematic bias, and those participants representing a total of n = 8, were removed from the dataset. The rest of the missing values in the risk factors followed MAR or MCAR missingness mechanisms, and therefore were imputed using a mean imputation procedure. ModiCogPD For the newly designed ModiCogPD, the final raw score was obtained by summing the 31 individual risk factors after dichotomization. Dichotomization thresholds were chosen based on those presented in previous risk scores and prior literature (Table 1), resulting in a binary classification of each variable, indicating the presence (= 1) or the absence of the risk factor (= 0). Following, higher ModiCogPD scores indicate a higher risk for PD-associated cognitive decline. The final ModiCogPD score ranges from 0 to 31. Importantly, the subitems of ModiCogPD substantially overlap with the ones from pPREDICT tool, as ModiCogPD represents an extension of this framework. In addition, there is considerable overlap with the LIBRA score, given that both instruments incorporate vascular and lifestyle-related determinants of cognitive decline. The only information missing in the ModiCogPD with respect to LIBRA is hypercholesterolemia, which could not be directly assessed in our cohort (Table 1). As some of the assessments from the other prediction tools were not available in our cohort, proxies were used, when necessary. In particular, to assess depression, the Beck Depression Inventory-I (BDI-I) 63 was used instead of the Geriatric Depression Scale (GDS) 64 . Moreover, the Movement Disorder Society -Unified Parkinson's Disease Rating Scale (MDS-UPDRS) 65 part I item 4, assessing anxiety, was considered instead of the State-Trait Anxiety Inventory (STAI) 66 . To assess physical activity, we combined two items from the modified PD Risk Factor Questionnaire (PD-RFQ-U), Epi Info™ developed by Caroline Tanner addressing weekly hours of vigorous and moderate physical activity. Vascular risk factors were evaluated through questions on self-perceived orthostatic hypotension and a semi-structured interview regarding cardiovascular medical history. Table 1 : Overview of included risk factors by risk predictions tools. # new risk factors included in the ModiCogPD risk score * If “yes” in any of these questions, then 1 point was attributed Within () the variables used are indicated ADL: activities of daily living; PIGD: postural instability and gait disorder. Alcohol consumption refers to the question: In your lifetime, have you drunk 100 or more alcoholic drinks (beer, wine, liquor, spirits)? In your lifetime, have you ever regularly drunk alcohol, that is, at least one drink per week for 6 months or longer? If a participant answered “yes” to at least one of these questions, they were considered to have a history of alcohol consumption. *maybe we have to write here the question for all the RFQ derived risk : smoking, caffeine, contusion, pesticides and toxics Demographics LIBRA pPREDICT ModiCogPD Age Educational level Age Sex Years of education Age (> 65: 1 point) Sex (Male: 1 point) Years of Education (< 12 years: 1 point) Marital status#(if divorced, never married, widowed, separated: 1 point) 67 Number of children# (if no children: 1 point) Number of spoken languages# (if monolingual: 1 point) 68 Clinical MoCA score MoCA score ( 5 years: 1 point) PIGD score# ( if tremor/PIGD phenotype < = 0.90: 1 point) Family history of PD and dementia# (HPI) ( if yes: 1 point) 69 Objective hyposmia# (Sniffin Sticks 70 , Scores ≤ − 1.5 standard deviations below age- and sex-adjusted normative values: 1 point) (PD-specific) Characteristics – potentially modifiable or treatable Depression Obesity Physical inactivity Heart coronary disease Hypertension Hypercholesterolemia(missing variable, not included) Diabetes RBDSQ Depression Anxiety Obesity Low physical activity Orthostatic hypotension Hypertension Diabetes Cardiovascular disease RBDSQ (total score ≥ 5) Depression (BDI-I ≥ 16) Apathy# (Starkstein Apathy Scale ≥ 14: 1 point) Anxiety (MDS-UPDRS-I.4 and Medical History) Obesity (BMI ≥ 25kg/m2) Low physical activity* Moderate-severe motor (UPDRS-III ≥ 32 : 1 point) Moderate-severe non-motor aspects of ADL# (MDS-UPDRS-I ≥ 10: 1 point)* Moderate-severe motor aspects of ADL# (MDS-UPDRS-II ≥ 12: 1 point)* Moderate-severe motor complications# (MDS-UPDRS-IV ≥ 4: 1 point)* 71 Diabetes* (if yes: 1 point) Cardiovascular disease (if yes: 1 point) Hypertension (if yes: 1 point) Orthostatic hypotension (if yes: 1 point) Lifestyle & environmental exposures Caffeine consumption#* (RFQ) Alcohol consumption#* (RFQ) Tobacco consumption#* (RFQ) Contusion# (RFQ)* Pesticides exposure(at work and non-work)# (RFQ)* Solvents and toxicants# (RFQ)* Score range 0-15.3 0–13 0–31 Statistical Analyses Predictive Performance of the ModiCogPD To assess the general performance of the new risk score to differentiate between PD-NC and PD-MCI and enable a comparison to established risk scores, the area under the operating characteristic curve (AUC-ROC) was calculated for each risk score (LIBRA, pPREDICT, ModiCogPD). This metric measures how well a continuous score separates two different classes, in this case PD-NC and PD-MCI. A stratified five-fold cross validation was implemented ensuring that each fold preserved the same ratio of positive and negative outcomes while providing a more robust estimate of model discrimination. To enable a more granular evaluation of the predictive performance of ModiCogPD, a multivariable logistic regression (LR) model was developed to compare the use of individual risk components with the aggregated global risk score. The cognitive status of PwPD (PD-NC or PD-MCI), derived from the MDS PD-MCI Level II criteria, was used as the dependent variable, while the individual subitems of the ModiCogPD served as independent variables. Prior to model fitting, all predictors underwent a scaling step tailored to the different types of variables aiming to ensure comparable penalization across variables and to improve numerical stability. A nested cross-validation approach was used, consisting of a 5-fold CV outer loop, aiming to provide a robust estimate of the performance in unseen data, and a 3-fold CV inner loop, aimed to optimize the parameters of the model. Within each outer loop, model parameters were tuned using the inner loop, with the best model being evaluated on the held-out outer test fold. Performance metrics were then averaged across outer folds. Given that the size of the study population was limited relative to the number of independent variables (events-per-variable ratio (EPV) ~ 3), the model was considered prone to overfitting and unstable estimates. Therefore, penalised LR approaches, specifically L2 regularization, were adopted to reduce overfitting by shrinking coefficient estimates, thereby improving model stability and generalizability 72 . As penalization alters the likelihood function, standard inferential statistics such as p -values are not directly available. Therefore, unpenalized LR models were additionally fitted for descriptive and inferential purposes. In addition, internal validation was further assessed through bootstrapping to quantify potential model optimism, Calibration was evaluated by examining the agreement between predicted probabilities and observed outcomes, using the Brier score and, more specifically, the calibration intercept and slope, which indicate systematic over- or underestimation of risk and the agreement between predicted and observed probabilities across the range of predicted risk. Model optimism, reflecting potential overfitting of the LR model to the training data, was quantified using 500 bootstrap resamples. For each bootstrap sample, the model was refitted and performance metrics, including discrimination and calibration, were calculated on both the bootstrap sample (apparent performance) and the original dataset. Model optimism was estimated as the difference between the bootstrap sample and the original dataset performances. The final reported metrics were corrected by subtracting the average model optimism across all bootstrap samples 73 . Identification of the most relevant modifiable and non-modifiable risk factors Further analyses were conducted, on the LR-based risk score model with the highest discrimination, to investigate the association of modifiable and non-modifiable individual risk factors with cognitive status. A multivariate unpenalized LR model was fitted to obtain interpretable effect estimates, confidence intervals, and p -values. The model specification followed the approach described in the Methods subsection “Predictive Performance of the ModiCogPD”, with PwPD cognitive status as the outcome and individual risk score components as covariates. Weighted ModiCogPD As an extension of the previous LR analyses incorporating subitems to enhance discriminatory performance, we further examined alternative weighting procedures in the development and improvement of the ModiCogPD risk score performance. A review of the literature on weighting approaches in the CI field revealed heterogeneous approaches. While certain scores, such as LIBRA 54 , reported improved performance following the application of a weighting strategy based on expert consensus through a Delphi process, others, including MoPaRDS 21 , which employed a regression coefficient–based scoring system, did not demonstrate a clear advantage over unweighted formulations. In light of these mixed results, we aimed to formally assess whether a data-driven weighting approach, based on regression-derived coefficients, improves the discrimination and calibration of the ModiCogPD risk score compared with an unweighted approach. A regression coefficient–based scoring system was used to assign weights to the individual sub-items of the general risk score. For this approach, first the study population was split into a train and a test dataset, each of them containing 50% of the patients, followed by a scaling step tailored to the different types of variables. Once the pre-processing was finalised, a LR with all the individual risk factors was trained against the end point variable reflecting the cognitive status on the train dataset. Due to the small sample size, L2 regularization was adopted to reduce overfitting by shrinking coefficient estimates, thereby improving model stability and generalizability. Calibration and optimism were checked to ensure that the model’s apparent performance was not overestimated, especially when applied to new data. The raw coefficients were normalized by scaling to the smallest non-trivial coefficient and rounding to the closest integer, easing their usability and interpretability as a risk score 74 . Predictive strength on the posterior cognitive decline trajectories Survival analyses were used to evaluate the prognostic value for the entire set of risk scores by modeling the association between the baseline risk score and time-to-event outcomes in two study populations: the global PwPD cohort (n = 201) and a subset of newly diagnosed PwPD (n = 90). More specifically, Cox Hazard models were applied to determine the effect of baseline risk score hazard ratios on global cognitive performance trajectories, as measured by the MoCA score. In the case of the global PwPD cohort the threshold for cognitive decline was based on 55 who defined decreased global cognition as a Mini–Mental State Examination (MMSE) score < 26. As MMSE data were not available in the present study, we used a MoCA score ≤ 21 based on previously published MoCA–MMSE conversion rates 75 . Participants with cognitive performance below the specified threshold at baseline were excluded from the analyses. To account for potential confounders, age, years of education, sex and baseline MoCA score were included in the models. To account for the potential non-linear effect, particularly in the relationship between age and cognitive decline, a Cox model in which age was modeled as a restricted cubic spline with four knots was fitted 76 . The significance of the non-linear association was assessed by comparing the spline model to a linear model. As the resulting p -value indicated no significant non-linearity (for a significance level α = 0.05), age was subsequently included as a linear confounder. The presence of competing events was evaluated. Death prior to the event of interest occurred in a small subset of PwPD (n = 7) and was therefore considered a potential competing event. Given the limited number of deaths, formal competing-risk regression was not pursued. Cause-specific Cox proportional hazards models were instead fitted, with PwPD censored at the time of death. This approach estimates the instantaneous risk of the event of interest among PwPD who remain alive and event-free. Risk scores associated with cognitive domain impairment Finally, the association between the risk scores and the presence of impairment across each of the five cognitive domains was evaluated. LR models were applied to study the association between each risk score and impairment in each cognitive domain (memory, language, attention, visuospatial functions and executive functions) defined as target, adjusting for age, sex and years of education. Impairment in a cognitive domain was defined as a z-score of ≤ -1.5 SD below the age-, sex-, and years of education adjusted normative data (see section Cognitive Impairment definition for additional details ). Declarations Data availability Patient data used in the preparation of this manuscript were obtained from the National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD). NCER-PD datasets are not publicly available, as they are linked to the Luxembourg Parkinson’s Study and its internal regulations. The NCER-PD Consortium is willing to share its available data. Its access policy was devised based on the study ethics documents, including the informed consent form, as approved by National Ethics Board (CNER Ref: 201407/13) and Data Protection Committee (CNPD Ref: 446/2017). Requests to access datasets should be directed to the Data and Sample Access Committee via email: [email protected] . Code availability The underlying code for this study is available on . uniluxembourg / LCSB / Digital Medicine / Gabriel_phd_project / Collaboration With Claire And Sonja · GitLab , Acknowledgments This project was supported by the Luxembourg National Research Fund (FNR) through the FNR/PEARL/dHealthPD/14146272 and FNR/PREVENE/14781425. The National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD) was funded by the Luxembourg National Research Fund (FNR) (FNR/NCER13/BM/11264123). Data used in the preparation of this manuscript were obtained from the National Centre of Excellence in Research on Parkinson's Disease (NCER-PD). We would like to thank all participants of the Luxembourg Parkinson’s Study for their important support of our research. Furthermore, we acknowledge the joint effort of the National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD) Consortium members from the partner institutions Luxembourg Centre for Systems Biomedicine, Luxembourg Institute of Health, Centre Hospitalier de Luxembourg, and Laboratoire National de Santé generally contributing to the Luxembourg Parkinson’s Study as listed below Author Contributions : CP, SRJ, GMT conceived and designed the study. SRJ, CP, contributed to data collection. GMT contributed to the statistical analysis plan and led the statistical analyses. AO and PMC supported the statistical analysis plan with their expertise. CP, SRJ, GMT, PMC drafted the manuscript. VES, AS, DE, EK, JK and RK provided expertise related to this manuscript. All authors critically revised the manuscript for important intellectual content and approved the final version. Author’s list List of NCER-PD consortium members : Mariella GRAZIANO⁷, Alexandre BISDORFF⁵, Rene DONDELINGER⁵, Elodie THIRY³, Gelani ZELIMKHANOV³, Guy BERCHEM³, Liliana VILAS BOAS³, Linda HANSEN³, Martine GOERGEN³, Nancy DE BREMAEKER³, Nico DIEDERICH³, Romain NATI³, Roxane BATUTU³, Sylvia HERBRINK³, Jochen KLUCKEN¹,³, Rejko KRÜGER¹,²,³, Claire PAULY²,³, Lukas PAVELKA²,³, Marijus GIRAITIS²,³, Maria Fernanda NIÑO URIBE¹,³, Achilleas PEXARAS², Alexander HUNDT², Alexia MENDIBIDE², Ana Festas LOPES², Angelo FERRARI², Brian DEWITT², Carlos GAMIO², Estelle HENRY², Gaël HAMMOT², Geeta ACHARYA², Hermann THIEN², Ilsé RICHARD², Johanna TROUET², Kate SOKOLOWSKA², Katy BEAUMONT², Laura GEORGES², Lorieza CASTILLO², Lucie REMARK², Maeva MUNSCH², Margaux HENRY², Maud THERESINE², Olga KOFANOVA², Olivia ROLAND², Pauline LAMBERT², Saïda MTIMET², Wim AMMERLANN², Anne GRÜNEWALD¹, Armin RAUSCHENBERGER¹,², Dheeraj REDDY BOBBILI¹, Ekaterina SOBOLEVA¹,³, Elisa GÓMEZ DE LOPE¹, Enrico GLAAB¹, Evi WOLLSCHEID-LENGELING¹, Francoise MEISCH¹, Giuseppe ARENA¹, Ibrahim BOUSSAAD¹, Jens SCHWAMBORN¹, Kirsten ROOMP¹, ¹⁰, Michael T. HENEKA¹, Michele BASSIS¹, Muhammad ALI¹, Jade JABER¹,³, Patricia MARTINS CONDE¹, Patrick MAY¹, Paul WILMES¹, Piotr GAWRON¹, Rebecca TING JIIN LOO¹, Reinhard SCHNEIDER¹, Ruxandra SOARE¹, Sabine SCHMITZ¹, Sarah NICKELS¹, Sascha HERZINGER¹, Sinthuja PACHCHEK¹, Soumyabrata GHOSH¹, Stefano SAPIENZA¹, Valentin GROUES¹, Venkata SATAGOPAM¹, Iñigo YOLDI BERGUA¹, Gabriel MARTINEZ TIRADO¹, Jochen OHNMACHT², Anne-Marie HANFF², ¹⁰, ¹¹, Carlos VEGA², Eduardo ROSALES², Fozia NOOR², Gessica CONTESOTTO², Gloria AGUAYO², Guilherme MARQUES², Jérôme GRAAS², Joëlle FRITZ², Magali PERQUIN², Manon GANTENBEIN², Maura MINELLI², Michel VAILLANT², Myriam ALEXANDRE², Myriam MENSTER², Olena TSURKALENKO², Sibylle BÉCHET³, Jón GALES², Emna BOUHAJJA², Ulf NEHRBASS², Victoria LORENTZ², Zied LANDOULSI², Sonja JÓNSDÓTTIR², David BOUVIER⁴, Katrin FRAUENKNECHT⁴, Michel MITTELBRONN¹, ², ⁴, ¹⁰, ¹², ¹³, Roseline LENTZ⁶, Jean-Paul NICOLAY⁹, Nadine JACOBY⁸, Isabel SCHWAMINGER¹, Liyousew BORGA¹, Sijmen VAN SCHAGEN¹, Alan CASTRO MEJIA¹, Francesca TERRANOVA¹, Messaline FOMO¹, Francesca BOSCHI¹, Niloofar KHERADBIN¹, Isabelle ROLIN³ 1 Luxembourg Centre for Systems Biomedicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg 2 Luxembourg Institute of Health, Strassen, Luxembourg 3 Centre Hospitalier de Luxembourg, Strassen, Luxembourg 4 Laboratoire National de Santé, Dudelange, Luxembourg 5 Centre Hospitalier Emile Mayrisch, Esch-sur-Alzette, Luxembourg 6 Parkinson Luxembourg Association, Leudelange, Luxembourg 7 Association of Physiotherapists in Parkinsons Disease Europe, Esch-sur-Alzette, Luxembourg 8 Private practice, Ettelbruck, Luxembourg 9 Private practice, Luxembourg, Luxembourg 10 Faculty of Science, Technology and Medicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg 11 Department of Epidemiology, CAPHRI School for Public Health and Primary Care, Maastricht University Medical Centre+, Maastricht, the Netherlands 12 Luxembourg Center of Neuropathology, Dudelange, Luxembourg 13 Department of Life Sciences and Medicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg Competing interests All authors declare no financial or non-financial competing interests. References Dorsey, E. R., Sherer, T., Okun, M. S. & Bloem, B. R. The Emerging Evidence of the Parkinson Pandemic. J Parkinsons Dis 8, S3–S8 (2018). Li, X. et al. Trends and hotspots in non-motor symptoms of Parkinson’s disease: a 10-year bibliometric analysis. Front Aging Neurosci 16, 1335550 (2024). LeWitt, P. A. L. & Chaudhuri, K. R. Unmet needs in Parkinson disease: Motor and non-motor. Parkinsonism & Related Disorders 80, S7–S12 (2020). Zhang, Q., Aldridge, G. M., Narayanan, N. S., Anderson, S. W. & Uc, E. Y. Approach to Cognitive Impairment in Parkinson’s Disease. Neurotherapeutics 17, 1495–1510 (2020). Bloem, B. R., Okun, M. S. & Klein, C. Parkinson’s disease. Lancet 397, 2284–2303 (2021). Bowring, F. et al. Exploration of whether socioeconomic factors affect the results of priority setting partnerships: updating the top 10 research priorities for the management of Parkinson’s in an international setting. BMJ Open 12, e049530 (2022). Goldman, J. G. et al. Cognitive impairment in Parkinson’s disease: a report from a multidisciplinary symposium on unmet needs and future directions to maintain cognitive health. NPJ Parkinsons Dis 4, 19 (2018). Poletti, M. et al. Mild cognitive impairment and cognitive-motor relationships in newly diagnosed drug-naive patients with Parkinson’s disease. J Neurol Neurosurg Psychiatry 83, 601–606 (2012). Aarsland, D. et al. Parkinson disease-associated cognitive impairment. Nat. Rev. Dis. Primers 7, 47 (2021). Domellöf, M. E., Ekman, U., Forsgren, L. & Elgh, E. Cognitive function in the early phase of Parkinson’s disease, a five-year follow-up. Acta Neurol Scand 132, 79–88 (2015). Buter, T. C. et al. Dementia and survival in Parkinson disease: a 12-year population study. Neurology 70, 1017–1022 (2008). Hely, M. A., Reid, W. G. J., Adena, M. A., Halliday, G. M. & Morris, J. G. L. The Sydney multicenter study of Parkinson’s disease: the inevitability of dementia at 20 years. Mov Disord 23, 837–844 (2008). Gallagher, J. et al. Long-term dementia risk in Parkinson disease. Neurology 103, e209699 (2024). Gavelin, H. M. et al. Computerized cognitive training in Parkinson’s disease: A systematic review and meta-analysis. Ageing Res. Rev. 80, 101671 (2022). Folkerts, A.-K. et al. Can physical exercise be considered as a promising enhancer of global cognition in people with Parkinson’s disease? Results of a systematic review and meta-analysis. J. Parkinsons. Dis. 14, S115–S133 (2024). Wang, Z., Zhu, C., Miao, W. & Zhang, Y. Effects of resistance and balance training on motor and non-motor symptoms in patients with Parkinson’s disease: a meta-analysis. Neurol. Res. 1–14 (2025). Schrag, A., Siddiqui, U. F., Anastasiou, Z., Weintraub, D. & Schott, J. M. Clinical variables and biomarkers in prediction of cognitive impairment in patients with newly diagnosed Parkinson’s disease: a cohort study. Lancet Neurol. 16, 66–75 (2017). Yousaf, T., Pagano, G., Niccolini, F. & Politis, M. Predicting cognitive decline with non-clinical markers in Parkinson’s disease (PRECODE-2). J. Neurol. 266, 1203–1210 (2019). Gramotnev, G., Gramotnev, D. K. & Gramotnev, A. Parkinson’s disease prognostic scores for progression of cognitive decline. Sci. Rep. 9, 17485 (2019). Schiepers, O. J. G. et al. Lifestyle for Brain Health (LIBRA): a new model for dementia prevention. Int J Geriatr Psychiatry 33, 167–175 (2018). Dawson, B. K. et al. Office-Based Screening for Dementia in Parkinson Disease: The Montreal Parkinson Risk of Dementia Scale in 4 Longitudinal Cohorts. JAMA Neurol 75, 704–710 (2018). Livingston, G. et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet 404, 572–628 (2024). Mostert, C. M. et al. Broadening dementia risk models: building on the 2024 Lancet Commission report for a more inclusive global framework. EBioMedicine 120, 105950 (2025). Carlisle, T. C., Medina, L. D. & Holden, S. K. Original research: initial development of a pragmatic tool to estimate cognitive decline risk focusing on potentially modifiable factors in Parkinson’s disease. Front Neurosci 17, 1278817 (2023). Sun, C. & Armstrong, M. J. Treatment of Parkinson’s Disease with Cognitive Impairment: Current Approaches and Future Directions. Behav Sci (Basel) 11, (2021). Kalbe, E., Warnecke, T., Eggers, C., Ophey, A. & Folkerts, A.-K. Prevention of cognitive impairment and dementia in people with Parkinson’s disease: A call-to-action. J. Parkinsons. Dis. 15, 1353–1366 (2025). Hipp, G. et al. The Luxembourg Parkinson’s Study: A Comprehensive Approach for Stratification and Early Diagnosis. Front. Aging Neurosci. 10, 408277 (2018). Kwon, K.-Y., Park, S., Kim, R. O., Lee, E. J. & Lee, M. Associations of cognitive dysfunction with motor and non-motor symptoms in patients with de novo Parkinson’s disease. Sci. Rep. 12, 11461 (2022). Guo, Y. et al. Predictors of cognitive impairment in Parkinson’s disease: a systematic review and meta-analysis of prospective cohort studies. J. Neurol. 268, 2713–2722 (2021). Stoker, T. B. et al. Impact of GBA1 variants on long-term clinical progression and mortality in incident Parkinson’s disease. J. Neurol. Neurosurg. Psychiatry 91, 695–702 (2020). Davis, M. Y. et al. Association of GBA mutations and the E326K polymorphism with motor and cognitive progression in Parkinson disease. JAMA Neurol. 73, 1217 (2016). Iwaki, H. et al. Genetic risk of Parkinson disease and progression:: An analysis of 13 longitudinal cohorts. Neurol. Genet. 5, e348 (2019). Baiano, C., Barone, P., Trojano, L. & Santangelo, G. Prevalence and clinical aspects of mild cognitive impairment in Parkinson’s disease: A meta-analysis. Mov. Disord. 35, 45–54 (2020). Maher, S. et al. Treatment of apathy in Parkinson’s disease and implications for underlying pathophysiology. J. Clin. Med. 13, 2216 (2024). Plant, O. et al. A cognitive-behavioral model of apathy in Parkinson’s disease. Parkinsons Dis. 2024, 2820257 (2024). Mele, B. et al. Non-pharmacologic interventions to treat apathy in Parkinson’s disease: A realist review. Clin. Park. Relat. Disord. 4, 100096 (2021). Kamal, H. et al. Alcohol use disorder, neurodegeneration, Alzheimer’s and Parkinson's disease: Interplay between oxidative stress, neuroimmune response and excitotoxicity. Front. Cell. Neurosci. 14, 282 (2020). Aborageh, M., Hähnel, T., Martins Conde, P., Klucken, J. & Fröhlich, H. Predicting dementia in people with Parkinson’s disease. NPJ Parkinsons Dis. 11, 126 (2025). Murray, D. K., Sacheli, M. A., Eng, J. J. & Stoessl, A. J. The effects of exercise on cognition in Parkinson’s disease: a systematic review. Transl. Neurodegener. 3, 5 (2014). Zhang, X., Molsberry, S. A., Schwarzschild, M. A., Ascherio, A. & Gao, X. Association of diet and physical activity with all-cause mortality among adults with Parkinson disease. JAMA Netw. Open 5, e2227738 (2022). Ypinga, J. H. L. et al. Effects of specialised physiotherapy on mortality in Parkinson’s disease: a prospective observational study. NPJ Parkinsons Dis. 11, 214 (2025). Kalbe, E. et al. German Society of Neurology guidelines for the diagnosis and treatment of cognitive impairment and affective disorders in people with Parkinson’s disease: new spotlights on diagnostic procedures and non-pharmacological interventions. J. Neurol. 271, 7330–7357 (2024). Dukelow, T. et al. Modifiable risk factors for dementia, and awareness of brain health behaviors: Results from the Five Lives Brain Health Ireland Survey (FLBHIS). Front. Psychol. 13, 1070259 (2022). Albus, P., Folkerts, A.-K., Kessler, J., Köhler, S. & Kalbe, E. Sociodemographic differences in dementia prevention knowledge in Germany: Implications for targeted health communication. J. Prev. Alzheimers Dis. 13, 100517 (2026). Schröder, V. E. et al. Programme dementia prevention (pdp): A nationwide program for personalized prevention in Luxembourg. J. Alzheimers. Dis. 97, 791–804 (2024). Erz, D. et al. Dementia prevention through the eyes of individuals at risk: insights from a satisfaction survey within the programme for dementia prevention in Luxembourg. Front. Aging 7, 1712500 (2026). Yarnall, A. J. et al. Characterizing mild cognitive impairment in incident Parkinson disease: the ICICLE-PD study. Neurology 82, 308–316 (2014). Jellinger, K. A. Mild cognitive impairment in Parkinson’s disease: current view. Front. Cogn. 3, (2024). Johnson, D. K., Langford, Z., Garnier-Villarreal, M., Morris, J. C. & Galvin, J. E. Onset of mild cognitive impairment in Parkinson disease. Alzheimer Dis. Assoc. Disord. 30, 127–133 (2016). Karr, J. E., Graham, R. B., Hofer, S. M. & Muniz-Terrera, G. When does cognitive decline begin? A systematic review of change point studies on accelerated decline in cognitive and neurological outcomes preceding mild cognitive impairment, dementia, and death. Psychol. Aging 33, 195–218 (2018). Scharfenberg, D. et al. Understanding cognitive domains in Parkinson’s disease: A scoping review of empirical studies. Neuropsychol. Rev. (2026) doi: 10.1007/s11065-025-09691-5 . Dalrymple-Alford, J. C. et al. Characterizing mild cognitive impairment in Parkinson’s disease: MCI Criteria for Parkinson's Disease. Mov. Disord. 26, 629–636 (2011). Litvan, I. et al. Diagnostic criteria for mild cognitive impairment in Parkinson’s disease: Movement Disorder Society Task Force guidelines. Mov Disord 27, 349–356 (2012). Deckers, K. et al. Target risk factors for dementia prevention: a systematic review and Delphi consensus study on the evidence from observational studies: Major risk factors for dementia prevention. Int. J. Geriatr. Psychiatry 30, 234–246 (2015). Dubois, B. et al. Diagnostic procedures for Parkinson’s disease dementia: recommendations from the movement disorder society task force. Mov Disord 22, 2314–2324 (2007). O’Caoimh, R., Foley, M. J., Timmons, S. & Molloy, D. W. Screening for cognitive impairment in movement disorders: Comparison of the Montreal Cognitive Assessment and Quick Mild Cognitive Impairment screen in Parkinson’s disease and Lewy body dementia. J. Alzheimers Dis. Rep. 8, 971–980 (2024). Chen, J. et al. Predictors of cognitive impairment in newly diagnosed Parkinson’s disease with normal cognition at baseline: A 5-year cohort study. Front. Aging Neurosci. 15, 1142558 (2023). Nasreddine, Z. S. et al. The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society 53, 695–699 (2005). Shirk, S. D. et al. A web-based normative calculator for the uniform data set (UDS) neuropsychological test battery. Alzheimer’s Research & Therapy 3, 32 (2011). Martínez Tirado, G. et al. Data-driven clinical decision support tool for diagnosing mild cognitive impairment in Parkinson’s disease. NPJ Parkinsons Dis. 12, 15 (2026). Boel, J. A. et al. Level I PD-MCI using global cognitive tests and the risk for Parkinson’s disease dementia. Mov. Disord. Clin. Pract. 9, 479–483 (2022). Vos, S. J. B. et al. Modifiable Risk Factors for Prevention of Dementia in Midlife, Late Life and the Oldest-Old: Validation of the LIBRA Index. J Alzheimers Dis 58, 537–547 (2017). Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J. An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561–571 (1961). Yesavage, J. A. et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 17, 37–49 (1982). Goetz, C. G. et al. Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results. Movement Disorders 23, 2129–2170 (2008). Spielberger, C. D. State-trait anxiety inventory for adults. PsycTESTS Dataset American Psychological Association (APA) https://doi.org/10.1037/t06496-000 (1983). Zhang, D., Zheng, W. & Li, K. The relationship between marital status and cognitive impairment in Chinese older adults: the multiple mediating effects of social support and depression. BMC Geriatr. 24, 367 (2024). Pacifico, D. et al. Associations of multilingualism and language proficiency with cognitive functioning: epidemiological evidence from the SwissDEM study in community dwelling older adults and long-term care residents. BMC Geriatr. 23, 629 (2023). Rocca, W. A. et al. Risk of cognitive impairment or dementia in relatives of patients with Parkinson disease. Arch. Neurol. 64, 1458–1464 (2007). Hummel, T., Sekinger, B., Wolf, S. R., Pauli, E. & Kobal, G. ‘Sniffin’ sticks’: Olfactory performance assessed by the combined testing of odour identification, odor discrimination and olfactory threshold. Chem. Senses 22, 39–52 (1997). Martínez-Martín, P. et al. Parkinson’s disease severity levels and MDS-Unified Parkinson's Disease Rating Scale. Parkinsonism Relat. Disord. 21, 50–54 (2015). Cessie, S. L. & Van Houwelingen, J. C. Ridge estimators in logistic regression. J. R. Stat. Soc. Ser. C. Appl. Stat. 41, 191 (1992). Steyerberg, E. W. et al. Internal validation of predictive models: efficiency of some procedures for logistic regression analysis. J. Clin. Epidemiol. 54, 774–781 (2001). Mehta, H. B., Mehta, V., Girman, C. J., Adhikari, D. & Johnson, M. L. Regression coefficient-based scoring system should be used to assign weights to the risk index. J. Clin. Epidemiol. 79, 22–28 (2016). van Steenoven, I. et al. Conversion between mini-mental state examination, montreal cognitive assessment, and dementia rating scale-2 scores in Parkinson’s disease: Conversion of cognitive screening scales in PD. Mov. Disord. 29, 1809–1815 (2014). Molinari, N., Daurès, J. P. & Durand, J. F. Regression splines for threshold selection in survival data analysis. Stat. Med. 20, 237–247 (2001). Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterial.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 07 May, 2026 Reviewers agreed at journal 07 May, 2026 Reviewers invited by journal 05 May, 2026 Editor assigned by journal 13 Apr, 2026 Submission checks completed at journal 12 Apr, 2026 First submitted to journal 08 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9361292","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":638212201,"identity":"54574590-cb43-4b59-a1db-f63c5761cb46","order_by":0,"name":"Claire Pauly","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Claire","middleName":"","lastName":"Pauly","suffix":""},{"id":638212205,"identity":"fdb57bbf-85f3-444a-be67-176d8332c8f7","order_by":1,"name":"Sonja R Jónsdóttir","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Sonja","middleName":"R","lastName":"Jónsdóttir","suffix":""},{"id":638212208,"identity":"0d2320e2-6917-4fff-a4c5-b8784c3a651f","order_by":2,"name":"Gabriel Martinez tirado","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIie3NMYvCMBQH8FcEp4Lrc9FPcBAphANFx/saKYW6VHF0cMjUjq79KI45Ck4R14zeootI3XRRk8oNN6Rdb8gfXvJI+PEHcHH5l/EBGHy+d6ano0fAopGg2TxuSJcbQhoI/BJzElGddtD/2m3LwxLhI9udDj+b0SRQs2/dUljJQM6jnEkEKqcDHsq4RdWc1ROeBBCmmojY42FatKlKSD1ZnzV5aLI/GvL0g7yB9NG0cE1U1SKQYAMheImAbdGn6ujlYRoRlGciGJnaW9azwrutRj26j+F6T8eTTpYEZbkc2lvE+/b/PjMr0C285tPFxcXFpcoL/atUC9N7OO4AAAAASUVORK5CYII=","orcid":"","institution":"Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg","correspondingAuthor":true,"prefix":"","firstName":"Gabriel","middleName":"Martinez","lastName":"tirado","suffix":""},{"id":638212211,"identity":"09216618-418a-47bc-9fe0-8a8bc9da48aa","order_by":3,"name":"Anja Ophey","email":"","orcid":"","institution":"University Hospital Cologne and Faculty of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Anja","middleName":"","lastName":"Ophey","suffix":""},{"id":638212214,"identity":"f4db318d-a4a8-4b54-8bac-004090370699","order_by":4,"name":"Patricia Martins Conde","email":"","orcid":"","institution":"Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg","correspondingAuthor":false,"prefix":"","firstName":"Patricia","middleName":"Martins","lastName":"Conde","suffix":""},{"id":638212215,"identity":"0f340fca-3ffe-444c-97f3-f64be59a1712","order_by":5,"name":"Valerie E. Schröder","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Valerie","middleName":"E.","lastName":"Schröder","suffix":""},{"id":638212218,"identity":"b2e1f823-df3c-4680-ac14-6da46dbc5cc2","order_by":6,"name":"Amna Skrozic","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Amna","middleName":"","lastName":"Skrozic","suffix":""},{"id":638212221,"identity":"bb7d41f3-865b-4cf1-b041-230408b8510e","order_by":7,"name":"Dorothee Erz","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Dorothee","middleName":"","lastName":"Erz","suffix":""},{"id":638212224,"identity":"6613d9a2-0c60-431a-8250-cd79bc820cf9","order_by":8,"name":"National Centre of Excellence in Research on Parkinson's Disease (NCER-PD)","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"National","middleName":"Centre of Excellence in Research on Parkinson's Disease","lastName":"(NCER-PD)","suffix":""},{"id":638212225,"identity":"11eafdbd-eef5-466e-b17b-3a0814a0fae4","order_by":9,"name":"Jochen klucken","email":"","orcid":"","institution":"Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg","correspondingAuthor":false,"prefix":"","firstName":"Jochen","middleName":"","lastName":"klucken","suffix":""},{"id":638212226,"identity":"8b6cd1f5-c706-485f-91b4-b2b1bd214860","order_by":10,"name":"Elke Kalbe","email":"","orcid":"","institution":"University Hospital Cologne and Faculty of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Elke","middleName":"","lastName":"Kalbe","suffix":""},{"id":638212227,"identity":"bcc63d84-146a-407a-985d-1927468d66f9","order_by":11,"name":"Rejko Krüger","email":"","orcid":"","institution":"Transversal Translational Medicine (TTM), Luxembourg Institute of Health","correspondingAuthor":false,"prefix":"","firstName":"Rejko","middleName":"","lastName":"Krüger","suffix":""}],"badges":[],"createdAt":"2026-04-08 22:23:34","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9361292/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9361292/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109216276,"identity":"9837eaeb-0a4c-4ea8-92f1-b3fe24b8b606","added_by":"auto","created_at":"2026-05-13 18:03:44","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":82739,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of the ROC Curves for the different risk scores + LR-based ModiCogPD. \u003c/strong\u003eEach Receiver Operating Characteristic (ROC) curve represents the relationship between sensitivity and specificity across all classification thresholds for a given model, with the corresponding area under the curve (AUC) reported in the legend. The legend identifies the global risk score represented by each curve: the unweighted-ModiCogPD model (blue), LR-based ModiCogPD (Red) the pPREDICT model (orange), the LIBRA model (green). The dashed diagonal line indicates the performance of a non‑informative classifier.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9361292/v1/f7c1138db8b89751e7fe7b3b.png"},{"id":109216278,"identity":"0aa15447-0ef2-43ca-8252-6d73955683f1","added_by":"auto","created_at":"2026-05-13 18:03:44","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":22309,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredictive strength of the risk scores for cognitive decline trajectories.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eForest plot showing the hazard ratios (HR) with 95% confidence intervals derived from multivariable Cox proportional hazards models assessing the association between LIBRA, pPREDICT, and ModiCogPD scores and time to cognitive decline. Two different study populations are used 1) overall PwPD cohort and 2) a newly diagnosed subgroup, where cognitive impairment (CI) was defined as MoCA ≤ 21 and MoCA \u0026lt; 26, respectively.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9361292/v1/37a89ac1edab52c557fc2c7e.png"},{"id":109295829,"identity":"b3a41517-8a0e-4ffd-92a2-46fdaee24fa7","added_by":"auto","created_at":"2026-05-15 08:37:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":588778,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9361292/v1/a50c68d9-270c-4a56-b340-b2e62030950e.pdf"},{"id":109222266,"identity":"eb417bd2-7596-44b3-b79b-6a2bf90e1331","added_by":"auto","created_at":"2026-05-13 21:06:43","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":86783,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-9361292/v1/f61e6fae3358fc794d88f0f3.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Estimating Personalized Risk of Mild Cognitive Impairment in Parkinson's Disease through Comprehensive Risk Prediction Tools","fulltext":[{"header":"Introduction","content":"\u003cp\u003eParkinson\u0026rsquo;s disease (PD) is the fastest growing neurological disorder worldwide and is estimated to reach 14.2\u0026nbsp;million cases by 2040 \u003csup\u003e1\u003c/sup\u003e. People living with PD (PwPD) experience both motor symptoms - such as tremor, gait disorders and rigidity - and non-motor symptoms including hyposmia, depression and Parkinson\u0026rsquo;s disease\u0026ndash;related mild cognitive impairment (PD-MCI) \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Given the increasing incidence of PD cases, there has been a heightened interest in understanding and addressing the disease's non-motor symptoms, particularly cognitive impairment (CI). This is attributable to the profound impact of CI on quality of life of those affected, often exacerbating caregiver load and healthcare expenses \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. In a multicentric survey, CI has been listed as one of the top 10 research priorities for the management of PD \u003csup\u003e6\u003c/sup\u003e and as one of the most important unmet needs in PD \u003csup\u003e7\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003ePD-MCI may already be present in up to 30% of PwPD at the time of diagnosis \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. This proportion increases to 40\u0026ndash;50% after 5 years \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e, and up to 75% of PwPD develop dementia within 20 years of diagnosis \u003csup\u003e\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. CI and PD-MCI may be characterized by worsening of cognitive functioning in one or more cognitive domains, such as attention, memory and executive functioning. Despite the absence of disease-modifying treatments, it is imperative to identify PwPD at risk for developing PD-MCI and potentially dementia, thereby enabling early detection, diagnosis and management, and most importantly improve their quality of life as non-pharmacological interventions such as cognitive training \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e and physical activity may support cognitive functioning in early stages \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSeveral tools have been developed to estimate the risk of cognitive decline in PwPD. On the one hand, complex tools such as those based on Dopamine Transporter (DaT) imaging results, cerebrospinal fluid (CSF) and genetic data, are associated with higher costs and time constraints, leading to a reduced scalability for clinical practice. On the other hand, simple tools based on risk factors, such as lifestyle, demographics, patient reported outcomes and routinely collected clinical data can be easily implemented in the daily clinical work \u003csup\u003e\u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Given the growing evidence of the impact that lifestyle factors play in the development of CI, risk factor-based tools have gained increasing relevance in the field. Overall, comprehensive risk estimation tools with sufficient specificity for PwPD remain limited, particularly regarding the integration of modifiable risk factors. Existing approaches include PD-unspecific tools such as the Lifestyle for Brain Health (LIBRA) \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e, as well as PD-specific instruments like the Montreal Parkinson\u0026rsquo;s Risk of Dementia Scale (MoPaRDS) \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. However, these tools do not yet provide a comprehensive framework that extensively captures both PD-specific characteristics and modifiable risk domains. Evidence from the literature indicates that addressing modifiable risk factors may prevent up to 45% of dementia cases \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, with new updates suggesting that this figure could grow up to 65% \u003csup\u003e23\u003c/sup\u003e. Furthermore, incorporating PD-specific modifiable risk factors into these predictive tools could enhance their accuracy, clinical relevance, and overall utility. Estimating individualized risk of cognitive decline may help inform risk-adapted counseling and monitoring strategies, boosting early preventive interventions and inclusion in clinical trials, providing a strong rationale for developing and refining predictive tools.\u003c/p\u003e \u003cp\u003eIn line with this need for more comprehensive and PD-tailored risk stratification tools, Carlisle and colleagues (2023) developed the preliminary disease Risk Estimator for Decline in Cognition Tool (pPREDICT) \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e by including PD-specific modifiable and/or treatable risk factors for cognitive decline, such as REM Sleep Behaviour Disorder (RBD), motor severity, depression, anxiety, excessive daytime sleepiness, obesity, physical activity and several vascular risk factors. Some of these risk factors overlap with the overall 14 modifiable risk factors that account for up to 45% of worldwide dementia cases \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, and which may prevent or delay the onset of dementia, as well as positively influence cognitive outcomes in PwPD \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. However, several potentially relevant risk factors proposed by the Lancet Standing commission (Livingston \u003cem\u003eet al\u003c/em\u003e., 2024), such as diabetes, apathy, exposure to toxins and solvents, and alcohol consumption, were not included in previous prediction tools, despite evidence suggesting that they may influence cognition in PwPD \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. This underscores the relevance of expanding the pPREDICT to develop a new prediction tool for PD-MCI, which incorporates these potentially modifiable dementia risk factors.\u003c/p\u003e \u003cp\u003eIn the present study, we developed an extended CI risk score for PwPD named ModiCogPD \u003cem\u003e(\u0026ldquo;\u003c/em\u003e\u003cb\u003eModi\u003c/b\u003efiable factors for \u003cb\u003ecog\u003c/b\u003enitive decline in \u003cb\u003ePD\u003c/b\u003e\u0026rdquo;), including the factors mentioned previously. As few studies evaluated the existing multidimensional risk prediction tools (e.g LIBRA, pPREDICT), particularly within the same well-characterized PD cohort, we investigated whether the two established dementia risk scores (LIBRA, pPREDICT) and the newly developed ModiCogPD can predict PD-MCI in the Luxembourg Parkinson\u0026rsquo;s Study \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, a population-based, longitudinal, monocentric study. Furthermore, we evaluated their prognostic value and their association with subsequent cognitive decline over a follow-up period of up to 8 years. Finally, we aimed to identify modifiable lifestyle factors associated with increased cognitive risk that may allow for future risk-adapted preventive strategies in PwPD.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eDescriptive statistics of the study data\u003c/h2\u003e \u003cp\u003eA total of 201 PwPD and 231 control participants from the Luxembourg Parkinson\u0026rsquo;s Study (Hipp \u003cem\u003eet al\u003c/em\u003e., 2018) met the eligibility criteria for analysis. Both groups present significant differences in socio-demographic and clinical characteristics at the baseline visit, except for the total years of education (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Further details on the individuals risk factors included in the risk scores are given in the \u003cb\u003eSupplementary table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/b\u003e\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eSocio-demographic, clinical characteristics and risk score of PwPD ,PD-NC, PD-MCI and control participants at the baseline visit.\u003c/b\u003e Numerical variables are reported as mean and SD, together with minimum and maximum values (Min\u0026ndash;Max). Categorical and ordinal variables are described by count (n) and percentage. Group differences between PwPD and control participants were assessed using appropriate statistical tests appropriate to the distribution of each variable. Continuous variables were analyzed using either two-tailed t-tests or Mann\u0026ndash;Whitney U tests, whereas categorical variables were assessed using Pearson\u0026rsquo;s chi-squared test. Statistical significance was set at α\u0026thinsp;=\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePD (n\u0026thinsp;=\u0026thinsp;201)\u003c/p\u003e \u003cp\u003eMean (std); Min-Max or n(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePD-NC (n\u0026thinsp;=\u0026thinsp;107)\u003c/p\u003e \u003cp\u003eMean (std); Min-Max or n(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePD-MCI\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;94)\u003c/p\u003e \u003cp\u003eMean (std); Min-Max or n(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eControl\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;231)\u003c/p\u003e \u003cp\u003eMean (std); Min-Max or\u003c/p\u003e \u003cp\u003en (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value (PD vs Control)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value (PD-NC vs PD-MCI)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003eSocio-demographics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e64.25 (10.38); 38\u0026ndash;89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60.95 (9.79); 38\u0026ndash;82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68.01 (9.79); 41\u0026ndash;89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e59.83 (11.4); 22\u0026ndash;80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex (female)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e58 (28.86%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36 (33.65%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e22 (23.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e90 (48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYears of Education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.84 (3.63); 6\u0026ndash;26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.53 (3.55); 7\u0026ndash;25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.04 (3.57); 6\u0026ndash;26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.43 (3.61); 6\u0026ndash;24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.105\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;=\u0026thinsp;0.003\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinical\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlobal Cognition (MoCA: 0\u0026ndash;30)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e25.85 (2.80); 18\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27.03 (2.18); 21\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24.52 (2.85); 18\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e28.36 (1.31); 23\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMotor Symptoms (MDS-UPDRS III: 0-109)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e29.91 (13.70); 3\u0026ndash;75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28.04 (13.07); 6\u0026ndash;75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32.05 (14.14); 3\u0026ndash;74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.05 (4.26); 0\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.054\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepression (BDI-I: 0\u0026ndash;63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e8.95 (6.35); 0\u0026ndash;45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.15 (6.03); 0\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.73 (6.73); 0\u0026ndash;45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5.42 (5.03); 0\u0026ndash;29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.482\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eApathy (SAS: 0\u0026ndash;42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12.99 (5.35); 1\u0026ndash;33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12.16 (5.37); 1\u0026ndash;32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.94 (5.19); 2\u0026ndash;33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.57 (4.49); 0\u0026ndash;27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;=\u0026thinsp;0.015\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHoehn and Yahr (Disease staging)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.0: 21 (10.45%)\u003c/p\u003e \u003cp\u003e1.5: 18 (8.96%)\u003c/p\u003e \u003cp\u003e2.0: 117 (58.21%)\u003c/p\u003e \u003cp\u003e2.5: 25 (12.45%)\u003c/p\u003e \u003cp\u003e3.0: 18 (8.96%)\u003c/p\u003e \u003cp\u003e4.0: 2 (1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.0: 15 (14.02%))\u003c/p\u003e \u003cp\u003e1.5: 15 (14.02%)\u003c/p\u003e \u003cp\u003e2.0: 61 (57.01%)\u003c/p\u003e \u003cp\u003e2.5: 8 (7.47%)\u003c/p\u003e \u003cp\u003e3.0: 8 (7.47%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.0: 6 (6.39%)\u003c/p\u003e \u003cp\u003e1.5: 3 (3.19%)\u003c/p\u003e \u003cp\u003e2.0: 56 (59.57%)\u003c/p\u003e \u003cp\u003e2.5: 17 (18.09%)\u003c/p\u003e \u003cp\u003e3.0: 10 (10.64%)\u003c/p\u003e \u003cp\u003e4.0: 2 (2.13%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e-\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDisease Duration (in years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.58 (4.96); 0\u0026ndash;29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.88 (3.74);0\u0026ndash;19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.39 (5.99); 0\u0026ndash;29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e-\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.257\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLevodopa Equivalent Daily Dose (in mg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e603.50 (388.25); 25-1854\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e598.38 (378.84); 100\u0026ndash;1854\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e609.28 (400.76);25-1850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;=\u0026thinsp;0.021\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ep\u0026thinsp;=\u0026thinsp;0.856\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRisk Factors\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLIBRA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.57 (2.327);0-11.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.23 (2.386) 0-11.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.95 (2.210) ;1\u0026ndash;11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.74 (1.94); -1-6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.017\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epPREDICT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.92 (2.099); 0\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.16 (1.953);0\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.78 (1.930) ;0\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.67 (1.54); 0\u0026ndash;6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModiCogPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e11.61 (3.291) 5\u0026ndash;23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.86 (3.14) ;6\u0026ndash;23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.47 (3.258) ;5\u0026ndash;20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8.34 (2.30); 4\u0026ndash;17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePerformance of the different risk scores\u003c/h3\u003e\n\u003cp\u003eLIBRA presented the lowest performance in discriminating between PD-NC and PD-MCI (AUC\u0026thinsp;=\u0026thinsp;0.596\u0026thinsp;\u0026plusmn;\u0026thinsp;0.032), followed by the ModiCogPD with an AUC of 0.645\u0026thinsp;\u0026plusmn;\u0026thinsp;0.063. The risk score with the highest performance was the original pPREDICT reaching an AUC score of 0.717\u0026thinsp;\u0026plusmn;\u0026thinsp;0.068 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). A statistical comparison revealed no significant difference between LIBRA and the ModiCogPD score (p\u0026thinsp;=\u0026thinsp;0.24); however, significant differences were observed between the pPREDICT score and both LIBRA (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and the ModiCogPD (p\u0026thinsp;=\u0026thinsp;0.006) score.\u003c/p\u003e \u003cp\u003eTo further investigate potential improvements in the predictive performance of the ModiCogPD for cognitive status, a LR model was developed incorporating the individual subitems that compose the risk score. The LR-based ModiCogPD model demonstrated superior discrimination compared to the global ModiCogPD score, achieving an AUC of 0.723\u0026thinsp;\u0026plusmn;\u0026thinsp;0.045 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This performance was slightly higher than that of the pPREDICT model.\u003c/p\u003e \u003cp\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe internal validation conducted within the bootstrapping framework, showed for the ModiCogPD model that optimism correction reduced the AUC apparent from 0.831 to 0.722, suggesting the presence of small overfitting (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Calibration was acceptable, with a near-zero intercept and a slope close to 1 after correction, indicating that predicted probabilities were well-aligned with observed outcomes.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003ePerformance overview of the LR-based ModiCogPD.\u003c/b\u003e This table presents the performance of ModiCogPD both as a global risk score and as a LR-based ModiCogPD. For the latter also calibration slope, intercept after correction, apparent AUC and optimism-corrected area under the curve (AUC) are reported.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eRisk scores\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGlobal Risk Score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c7\" namest=\"c3\"\u003e \u003cp\u003eLR-based model\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eDiscrimination AUC score\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eDiscrimination\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003eAUC score\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eCalibration slope- correct\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eCalibration intercept\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eApparent AUC\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003eOptimism- corrected AUC\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModiCogPD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e \u003cp\u003e0.645\u0026thinsp;\u0026plusmn;\u0026thinsp;0.063\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e0.723\u0026thinsp;\u0026plusmn;\u0026thinsp;0.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.977\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.831\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eIdentification of the most relevant risk factors\u003c/h3\u003e\n\u003cp\u003eNext, we evaluated the independent associations of ModiCogPD individual risk factors with cognitive status (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The most relevant risk factors for the prediction of the cognitive status (\u003cem\u003ep\u003c/em\u003e-values\u0026thinsp;\u0026lt;\u0026thinsp;0.05) included age, disease duration, MDS-UPDRS III, apathy, alcohol consumption and tremor/PIGD phenotype, with a more prominent PIGD phenotype being associated with increased risk. Among these, only apathy and alcohol consumption are considered modifiable traits that can be targeted to slow down the progression towards CI. Other modifiable risk factors that presented high odds without reaching statistical significance, included vigorous physical activity (OR\u0026thinsp;=\u0026thinsp;1.943, p\u0026thinsp;=\u0026thinsp;0.160) and vascular risk (OR\u0026thinsp;=\u0026thinsp;1.703, p\u0026thinsp;=\u0026thinsp;0.201). Moreover, sex showed a trend-level association with cognitive status.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eOdds-ratios of individual risk factors for predicting CI status.\u003c/b\u003e The table reports regression coefficients, odds ratios (OR) and the respective confidence interval, and \u003cem\u003ep\u003c/em\u003e-values for each variable included in the model. Marked in bold are those \u003cem\u003ep\u003c/em\u003e-values that are significant.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCoefficient\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOdds Ratio\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e95% interval\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge at Assessment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.960\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.613\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.288\u0026ndash;5.302\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.008\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.821\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.440\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.172\u0026ndash;1.127\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.087\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYears of Education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.545\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.580\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.298\u0026ndash;1.126\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.108\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDisease Duration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.704\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.056\u0026ndash;3.875\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.034\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRBDSQ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.106\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.112\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.648\u0026ndash;1.907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMDS-UPDRS-III\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.679\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.973\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.016\u0026ndash;3.830\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.045\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBDI-I Total Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.329\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.720\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.332\u0026ndash;1.561\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.405\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety Disorder\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.083\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.087\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.349\u0026ndash;3.384\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.885\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.658\u0026ndash;1.549\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.965\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVigorous Physical Activity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.943\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.769\u0026ndash;4.911\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.160\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerate Physical Activity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.053\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.949\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.242\u0026ndash;3.723\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.940\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVascular Risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.532\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.703\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.754\u0026ndash;3.848\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.201\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStarkstein Apathy Scale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.591\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.806\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.053\u0026ndash;3.095\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.032\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMDS-UPDRS I\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.994\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.462\u0026ndash;2.136\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.987\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMDS-UPDRS II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.520\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.595\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.269\u0026ndash;1.312\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMDS-UPDRS IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.407\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.666\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.427\u0026ndash;1.037\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.072\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily History of PD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.183\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.200\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.543\u0026ndash;2.653\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily History of Dementia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.820\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.330\u0026ndash;2.035\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.669\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSniffin Sticks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.963\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.857\u0026ndash;1.081\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.520\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCaffeine Consumption\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.849\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.428\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.118\u0026ndash;1.553\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.197\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlcohol Consumption\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.497\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.224\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.058\u0026ndash;0.866\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.030\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmoking\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.856\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.397\u0026ndash;1.844\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.691\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eContusion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.434\u0026ndash;2.394\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.965\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePesticides Exposure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.189\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.368\u0026ndash;1.859\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.647\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eToxicants Exposure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.195\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.550\u0026ndash;2.599\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.653\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarital Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.294\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.341\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.487\u0026ndash;3.699\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of Children\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.445\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.561\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.754\u0026ndash;3.232\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.230\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal Number of Spoken Languages\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.144\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.547\u0026ndash;2.391\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.721\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes - Medical History\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.933\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.393\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.109\u0026ndash;1.421\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.155\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTremor/PIGD Phenotype\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.474\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.428\u0026ndash;0.907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.014\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003ePerformance of the Weighted ModiCogPD\u003c/h3\u003e\n\u003cp\u003eAs an extension of the previous LR analyses incorporating subitems to enhance discriminatory performance, we further examined alternative weighting procedures in the development and improvement of the ModiCogPD risk score performance. Following the improved performance showed by the LR-based ModiCogPD model, we further investigated whether given clinically meaningful, integer-based weights to individual risk factors, aimed at facilitating its applicability in routine clinical practice, could enhance predictive performance, a weighted ModiCogPD score was derived using regression coefficients (see Methods subsection \u0026lsquo;Weighted ModiCogPD\u0026rsquo;), from the 50% training dataset. The resulting weighted score was evaluated in the remaining 50% testing subset, achieving a discrimination of AUC\u0026thinsp;=\u0026thinsp;0.59, whereas the non-weighted approach yielded an AUC of 0.61 within the same testing sample (Supplementary Fig.\u0026nbsp;1). Further information on the weighting procedure is given in the supplementary material (Supplementary table 2).\u003c/p\u003e\n\u003ch3\u003ePredictive strength on the posterior cognitive decline trajectories\u003c/h3\u003e\n\u003cp\u003eThe prognostic value of the three global risk scores was evaluated by examining their association with the risk and timing of cognitive decline, defined as progression to MoCA\u0026thinsp;\u0026le;\u0026thinsp;21 for the newly diagnosed PwPD or MoCA\u0026thinsp;\u0026lt;\u0026thinsp;26 for the global PwPD population (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The prognostic value of the risk scores in the general population showed heterogeneous associations with the hazard of cognitive decline over the course of the disease. ModiCogPD demonstrated a significant positive association, with a hazard ratio of 1.245 (95% CI: 1.084\u0026ndash;1.430; p-value\u0026thinsp;=\u0026thinsp;0.002), indicating that each one-unit increase in the score was associated with a 25% higher risk of cognitive deterioration. Similarly, pPREDICT showed a comparable effect size (HR\u0026thinsp;=\u0026thinsp;1.254, 95% CI: 1.006\u0026ndash;1.564; p-value\u0026thinsp;=\u0026thinsp;0.0445), although the association was of borderline statistical significance. In contrast, LIBRA was not significantly associated with cognitive decline (HR\u0026thinsp;=\u0026thinsp;1.029, 95% CI: 0.849\u0026ndash;1.246; p-value\u0026thinsp;=\u0026thinsp;0.771). By contrast, analyses performed on the newly diagnosed subgroup revealed that only the LIBRA score was significantly associated with the risk and timing of cognitive decline (HR\u0026thinsp;=\u0026thinsp;1.136, 95% CI: 1.005\u0026ndash;1.284; p-value\u0026thinsp;=\u0026thinsp;0.042). Additional information is available in Supplementary table 3.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eRisk scores associated with cognitive domain impairment\u003c/h2\u003e \u003cp\u003eThe study association between the set of risk scores and the presence of CI in a specific domain showed heterogeneous results across the risk scores. With LIBRA showing no statistical significance with the presence of impairment in any of the cognitive domains after adjusting for age, sex and years of education. However, pPREDICT and ModiCogPD showed significance in both attention (HR\u0026thinsp;=\u0026thinsp;2.182, 95% interval: 1.226\u0026ndash;3.881, p\u0026thinsp;=\u0026thinsp;0.008; and HR\u0026thinsp;=\u0026thinsp;1.841, 95% interval: 1.104\u0026ndash;3.071, p\u0026thinsp;=\u0026thinsp;0.019) and executive domains (HR\u0026thinsp;=\u0026thinsp;3.137, 95% interval: 1.73\u0026ndash;5.69, p\u0026thinsp;=\u0026thinsp;0.0001; and HR\u0026thinsp;=\u0026thinsp;1.795, 95% interval: 1.082\u0026ndash;2.977; p\u0026thinsp;=\u0026thinsp;0.024), respectively (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAssociation between the risk scores and cognitive domain impairment.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCognitive Domains\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLIBRA\u003c/p\u003e \u003cp\u003eOR [95% interval], p-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003epPREDICT\u003c/p\u003e \u003cp\u003eOR [95% interval], p-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModiCogPD\u003c/p\u003e \u003cp\u003eOR [95% interval], p-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAttention\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.226 [0.735\u0026ndash;2.049] 0.435\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e2.182 [1.226\u0026ndash;3.881] 0.008\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e1.841 [1.104\u0026ndash;3.071] 0.019\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eExecutive\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.112 [0.678\u0026ndash;1.824] 0.673\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3.137 [1.730\u0026ndash;5.692] p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e1.795 [1.082\u0026ndash;2.977] 0.024\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMemory\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.906 [0.552\u0026ndash;1.487] 0.695\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.420 [0.825\u0026ndash;2.443] 0.207\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.417 [0.864\u0026ndash;2.326] 0.168\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVisuospatial\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.853 [0.488\u0026ndash;1.489] 0.575\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.167 [0.645\u0026ndash;2.115] 0.610\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.837 [0.479\u0026ndash;1.465] 0.533\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLanguage\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.737 [0.358\u0026ndash;1.517] 0.408\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.641 [0.771\u0026ndash;3.495] 0.199\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.457 [0.737\u0026ndash;2.885] 0.280\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe table shows the odds ratios (OR) with 95% confidence intervals derived from multivariable LR models examining the association between LIBRA, pPREDICT, and ModiCogPD risk scores and specific cognitive domains impairment (attention, executive, memory, visuospatial, and language). All models were adjusted for age, sex, and years of education.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we evaluated and compared the performance of two established and newly developed global risk scores for CI in PwPD (pPREDICT, LIBRA and ModiCogPD), focusing on their ability to predict PD-MCI and subsequent cognitive decline. In parallel, we extended existing risk scores by incorporating additional modifiable risk factors relevant in PD (ModiCogPD) and identified the most relevant modifiable factors associated with increased CI risk in PD, with the aim of informing future risk-adapted precision preventive strategies in this population.\u003c/p\u003e \u003cp\u003eOverall, pPREDICT showed the highest predictive performance for detecting PD-MCI at the global score level, followed closely by ModiCogPD, whereas LIBRA showed the lowest discrimination. Additionally, we investigated the contribution of individual risk factors within ModiCogPD by developing an LR-based model to assess whether differential weighting of these individual risk factors could improve the performance. This approach resulted in improved discrimination compared to the global risk score and outperformed the other models (LIBRA and pPREDICT), achieving the highest predictive accuracy. Both findings suggest that PD-specific risk scores incorporating both disease-related and modifiable factors may provide superior discrimination in PD-MCI screening compared to more general dementia risk tools such as LIBRA.\u003c/p\u003e \u003cp\u003eLongitudinal analyses further reinforced the prognostic utility of PD-specific scores, where higher pPREDICT and ModiCogPD scores were associated with an increased hazard of cognitive decline over time in the whole study population, however, LIBRA did not significantly predict cognitive decline. This finding underscores the overall prognostic value of PD-specific multidimensional tools in estimating the risk of cognitive decline over time in a heterogeneous PD population. Interestingly, a different pattern emerged in newly diagnosed PwPD, where a higher LIBRA score was the only one associated with an increased hazard of cognitive decline over time. LIBRA was originally developed to estimate dementia risk in the general population (Schiepers et al., 2018), whereas pPREDICT and our extended version, ModiCogPD, were designed to assess cognitive decline risk specifically in PD \u003csup\u003e24\u003c/sup\u003e. Even though the PD-specific tools also include general risk factors, their main focus is rather on disease-related characteristics. In early PD, when motor and non-motor features may still be mild or relatively homogeneous across patients, these disease-specific variables may be less meaningful. In this context, broader risk factors, such as vascular and lifestyle-related characteristics, which are more central to LIBRA\u0026rsquo;s construction, may have a relatively stronger influence. As the disease progresses and PD-related burden becomes more pronounced, disease-specific characteristics may play a larger role in determining cognitive decline outcomes. This difference in model focus and stage-dependent expression of risk factors may help explain the variation in predictive performance across the subgroups. These findings indicate that in the early stages of PD, disease-specific variables may contribute less, highlighting the prognostic value and relevance of general risk factors in the early stages when PD characteristics are not yet fully present. Together, these results support the value of PD-tailored, multidimensional risk score tools for screening CI in pre-dementia stages, such as in PD-MCI, as well as highlighting their prognostic value during the course of the disease.\u003c/p\u003e \u003cp\u003eThe evaluation of the individual risk factors of ModiCogPD enabled us to determine which ones were the most relevant in distinguishing between PD-MCI and normal cognition, emphasizing the role of both modifiable and PD-specific variables. In our analysis, age, disease duration, motor severity (MDS-UPDRS III), apathy, alcohol consumption, and the tremor/PIGD phenotype showed a statistically significant association with CI. These findings are in line with previous studies reporting the importance of age, motor, and non-motor symptoms such as depression in cognitive decline in PD \u003csup\u003e28\u003c/sup\u003e. A systematic review by Guo and colleagues \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e identified advanced age, genetic variation in APOE and MAPT, gait disturbance (PIGD), motor assessments (MDS-UPDRS-III) and non-motor symptoms such as hallucinations, orthostatic hypotension and anxiety as predictors for CI in PD. Other longitudinal studies reported \u003cem\u003eGBA1\u003c/em\u003e risk variants as predictors for cognitive decline in PD \u003csup\u003e30,31\u003c/sup\u003e. Additionally, \u003cem\u003eGBA1\u003c/em\u003e p.E365K variant increases the risk of developing CI in PwP with normal cognition at baseline \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Another meta-analysis revealed several risk factors associated with PD-MCI namely older age, lower education, longer disease duration, higher levels of LEDD, more severe motor symptoms, PIGD phenotype, as well as poorer quality of life, higher levels of apathy and depression \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. These findings reinforce the multifactorial nature of CI in PD and support the integration of multidimensional risk factors encompassing demographics, motor and non-motor symptoms, as well as potentially modifiable factors within risk prediction tools. Moreover, sex exhibited trend-level associations with cognitive status, indicating that sex-specific differences may be relevant in the application of risk score tools.\u003c/p\u003e \u003cp\u003eAmong the significant risk factors identified in our study, apathy and alcohol consumption represent potentially modifiable factors, which may constitute feasible targets for interventions aimed at slowing cognitive deterioration. Since apathy is very common in PD but often overlooked in clinical practice due to its subtle nature \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e, healthcare professionals should be aware of its association with CI and the potential pharmacological and non-pharmacological treatments. Exercise interventions, mindfulness exercises, cognitive behavioural therapy as well as repetitive transcranial magnetic stimulation are non-pharmacological treatments that could be used in PwP with apathy \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTo our knowledge, no previous study has shown an association between alcohol consumption and CI in PD although excessive alcohol consumption has been identified as a modifiable risk factor for dementia by the report of the \u003cem\u003eLancet\u003c/em\u003e standing Commission, 2024)21. Existing literature reviews of alcohol use disorder link chronic excessive alcohol exposure to oxidative stress, neuroinflammation, excitotoxicity, neuronal loss, and increased risk of dementia and PD \u003csup\u003e37\u003c/sup\u003e. They primarily addressed PD as a neurodegenerative endpoint and did not assess alcohol-related CI within PD cohorts \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. More studies are warranted to investigate the link between alcohol consumption and CI, and its underlying mechanism in PD with particular attention to potential differences in men and women. Other factors, such as vigorous physical activity and vascular risk factors, presented higher odds ratios without reaching statistical significance. This may be related to limited statistical power; however, these factors should be further explored in larger or independent cohorts. In particular, hypertension and diabetes type 2 should be further investigated as they are commonly associated with cognitive decline in PD and can even lead to PD with dementia \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. In our study, we focused on vigorous physical activity to prevent CI; however, it has been shown that even mind-body exercise such as Tai Chi, Qigong, yoga and dance can improve CI in PD \u003csup\u003e39,40\u003c/sup\u003e. With growing evidence for the beneficial effect of physical activity, healthcare professionals and clinicians should recommend exercise as routine management for PwPD as well as specialized physiotherapy\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. In addition, the specific role of physical activity and vascular risk factors, requires further investigation in larger cohorts to better quantify their relevance in the assessment of risk factors for PD-MCI. Taken together, the identification of potentially modifiable factors in our cohort underscores the importance of prevention-oriented strategies in PD. Accordingly, prevention efforts should focus on routine cognitive and risk assessment, standardized risk communication and education, multidomain risk-reduction interventions, and cognitive and physical training, supported by interdisciplinary collaboration across PD care networks\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. However, public awareness of modifiable dementia risk factors remains limited. A recent survey conducted in Ireland found that although 65.6% of participants believed lifestyle changes can lower dementia risk, only 31.4% considered dementia preventable with knowledge levels varying across demographic groups\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. These findings highlight the need for targeted and personalized prevention strategies that also take into account sociodemographic determinants of knowledge, i.e. education, sex, and age\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. In this context, structured prevention programs, such as the \u0026ldquo;programme dementia prevention (pdp)\u0026rdquo; initiated in Luxembourg \u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, may help translate risk knowledge into actionable interventions.\u003c/p\u003e \u003cp\u003eTo contextualize these findings, it is important to consider the overall prevalence of PD-MCI within our cohort, as this reflects the clinical relevance of the observed risk associations. PD-MCI is a frequent non-motor symptom in PD and is considered as one of the most burdensome symptoms of PD. By applying the MDS-PD MCI Level II criteria with a z-score cut-off of \u0026le; -1.5 SD, to our PD cohort, PD-MCI was present in 46.77% of PwPD, in line with other cohorts \u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. The prevalence of PD-MCI in the literature is very heterogeneous ranging from 19 to 69.3% in PwPD \u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, due to different assessment tools applied to detect CI as well as time of assessment. Even within the same diagnostic framework, such as the MDS PD-MCI Level II criteria, prevalence estimates can differ considerably, as the recommended impairment threshold may range from \u0026minus;\u0026thinsp;1 to \u0026minus;\u0026thinsp;2 SD, leading to variability in the proportion of patients meeting the PD-MCI criteria. Given the early-moderate disease stage of the PwPD in our cohort and the criteria applied to define PD-MCI, the prevalence of CI is representative and comparable with other PD cohorts with similar age ranges and disease duration\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eBeyond evaluating the screening potential for PD-MCI, we also explored the association between the different risk scores and the specific cognitive domains, which may provide additional insights into the early patterns of cognitive decline in PD. The association between risk scores and specific cognitive domains was heterogeneous. LIBRA showed no significant associations with any cognitive domain, whereas both pPREDICT and ModiCogPD were significantly associated with attention and executive function. No risk score was significantly associated with memory, visuospatial, or language domains. These results suggest that PD-specific scores such as pPREDICT and ModiCogPD may better capture cognitive difficulties related to attention and executive functions, which are known to occur early during the disease course \u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e,\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. Therefore, the ability of these prediction tools to identify early patterns of impairment may support earlier screening for PD-MCI and help guide focused monitoring and preventive strategies targeting the cognitive domains most at risk.\u003c/p\u003e \u003cp\u003eSome limitations should be acknowledged in the present study and can be mainly grouped into data-related, or more methodological considerations. Regarding limitations of the underlying data, the absence of neuropsychological normative values specific to the Luxembourgish population represented a significant challenge. To address this, regression-based methods were employed to generate normative values from the control group. Although this approach has been previously described in the literature, it is less robust than using established norms derived from large, representative population samples. This limitation, together with the relatively small number of healthy controls, may have introduced additional variability in the z-score calculations and, consequently, in the classification of PwPD with CI. Further variability may have arisen from the linguistic diversity of the Luxembourgish population, which has three official languages, German, Luxembourgish and French, as well as a big community whose first language is Portuguese. Consequently, some participants completed neuropsychological testing in a non-native language. Moreover, it is not possible to completely isolate and assess a single cognitive domain, as most neuropsychological assessments require the involvement of multiple cognitive domains. As a result, neuropsychological tests may be classified differently across cognitive domains, and alternative grouping strategies may yield different domain-specific outcomes, potentially influencing the interpretation of cognitive profiles. Consistent with this, a recent scoping review shows that, although cognitive domains are central to diagnosing cognitive impairment in PD, there is limited empirical support for domain structures derived from dimensionality reduction of cognitive test data \u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. Comparisons across studies are hindered by heterogeneous statistical methods and PD sample characteristics. Finally, participation in the extended neuropsychological assessment was voluntary rather than systematically applied across the entire cohort, which may have introduced selection bias.\u003c/p\u003e \u003cp\u003eLimitations also arise from the application of the MDS PD-MCI Level II diagnostic criteria. In particular, threshold selection for CI is a key methodological factor contributing to heterogeneity in PD-MCI classification. Although cut-offs from \u0026minus;\u0026thinsp;1 to -2 SD are used in the literature, the \u0026minus;\u0026thinsp;1.5 SD threshold was adopted, as recommended by Dalrymple-Alford et al. for the most balanced PD-MCI detection \u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e. In addition, although the cognitive testing criteria was applied, it was not possible to fully implement the whole PD-MCI criteria proposed by \u003csup\u003e53\u003c/sup\u003e which require 1) evidence of gradual cognitive decline, due to the lack of consistent longitudinal data across the whole cohort, and 2) preserved activities of daily living (ADL), due to the absence of detailed informant-based ADL assessments across all participants and visits. Furthermore, only one neuropsychological test was available within the visuospatial domain, preventing full adherence to the MDS Level II criteria as originally described by Litvan et al. \u003csup\u003e53\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAnother available risk score for PD is the Montreal Parkinson\u0026rsquo;s Risk of Dementia Scale (MoPaRDS). Similarly to the pPREDICT, it incorporates mostly PD-specific variables but does not focus on modifiable risk factors. However, we were not able to evaluate this specific prediction tool, due to specific variables, such as Bilateral disease onset not being available in our cohort so a direct comparison with it was not feasible. Similarly, minor adaptations were necessary when calculating the LIBRA score \u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e, as the exact clinical variables originally used were not available; therefore, alternative assessments/questionnaires assessing comparable domains were selected. An updated version of LIBRA was published by Rosenau (2024), highlighting hearing impairment, social contact, and sleep\u0026mdash;already included in our model\u0026mdash;as risk factors with strong evidence for the development of dementia. Hearing impairment and social contact, particularly aspects related to loneliness and social engagement, were not included in our analysis due to the lack of available measures to assess these domains within the LuxPArk cohort. Moreover, LIBRA was originally developed to predict dementia, an outcome that could not be directly evaluated in our cohort, as no participants met the PD dementia criteria defined by \u003csup\u003e55\u003c/sup\u003e. In contrast, the ModiCogPD score focuses on cognitive decline defined as a worsening in cognitive performance of -1.5 SD reflecting a distinct conceptual and clinical target. Finally, the modifiable risk factor \u0026ldquo;Alcohol consumption\u0026rdquo;, included in the ModiCogPD, refers to the history of consumption and not per se to the current levels of consumption, therefore cautiousness is needed when interpreting its relevance in CI. These limitations should be taken into account when interpreting comparative performance across risk scores.\u003c/p\u003e \u003cp\u003eBesides clinical characteristics, there are additional limitations that should be acknowledged, such as the lack of external validation for the ModiCogPD risk score, as successful replication in independent cohorts would further ensure its predictive and prognostic value and strengthen methodological confidence in the findings. The small size of the study population and the number of independent variables in the different LR models, could represent another limitation, making the models more prone to overfitting and unstable estimates; which was addressed by applying L2-regularisation.\u003c/p\u003e \u003cp\u003eIn conclusion, this study compared general and PD-specific risk scores for identifying CI and estimating the risk of future cognitive decline in PD. Our results show that PD-specific scores performed better overall in a mixed PD population, while a general lifestyle-based score (LIBRA) may be more relevant in the earliest stages of the disease. The identification of modifiable risk factors also highlights potential targets for prevention and closer clinical monitoring \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. Overall, these findings support the use and further improvement of practical risk scores to help detect CI earlier and guide more personalized care in PD.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003e201 PwPD and 231 control participants from the Luxembourg Parkinson\u0026rsquo;s Study \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e with detailed neuropsychological assessment were included in the analyses. Participants were recruited in Luxembourg between 2015 and 2023 and provided written informed consent. The collection was approved by the National Ethics Board (CNER Ref: 201407/13) and Data Protection Committee (CNPD Ref: 446/2017).\u003c/p\u003e \u003cp\u003ePwPD with severe CI defined as a MoCA total score\u0026thinsp;\u0026lt;\u0026thinsp;18 at baseline were excluded from the analysis \u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e. Missing values in the study population were handled differently across different variables. Participants with missing values in the neuropsychological tests were removed from the study population. While the risk factors were imputed following the most suitable methodology depending on its nature of missingness (for further details see subsection \u0026ldquo;\u003cem\u003eRisk prediction tools\u003c/em\u003e\u0026rdquo;). The study population included a subset of newly diagnosed PwPD (n\u0026thinsp;=\u0026thinsp;90), defined by a baseline visit within two years from diagnosis \u003csup\u003e\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMoreover, we excluded control participants meeting criteria for PD-MCI at baseline defined as a Montreal Cognitive assessment (MoCA) total score\u0026thinsp;\u0026lt;\u0026thinsp;26 \u003csup\u003e58\u003c/sup\u003e, as well as participants reporting neurological comorbidities such as brain tumor, meningitis/encephalitis, multiple sclerosis, normal pressure hydrocephalus, seizures, stroke and traumatic brain injury.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eNeuropsychological test battery\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eA detailed neuropsychological test battery was conducted at baseline and at yearly follow-up assessments. Based on the Movement Disorder Society (MDS) PD-MCI Level II criteria \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e, two different tests per cognitive domain were included to assess potential CI except for the visuospatial domain where only one test was available.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e(1) attention and working memory: Trail Making Test part A and digit span forward\u003c/p\u003e \u003cp\u003e(2) executive functions: Frontal Assessment battery, Trail Making Test B-A\u003c/p\u003e \u003cp\u003e(3) language: letter fluency (f), category fluency (animals)\u003c/p\u003e \u003cp\u003e(4) memory: Word list of Consortium to Establish a Registry for Alzheimer's Disease (CERAD) Delayed Recall, Word list intermediate recall (CERAD).\u003c/p\u003e \u003cp\u003e(5) visuospatial functions: Benton\u0026rsquo;s Judgment of Line Orientation\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eCognitive Impairment Definition\u003c/h2\u003e \u003cp\u003eSince population-specific normative data were not available in Luxembourg, we derived normative values from the control subgroup included in the Luxembourg Parkinson\u0026rsquo;s Study following an approach described by \u003csup\u003e59\u003c/sup\u003e and previously applied to the Luxembourg Parkinson\u0026rsquo;s Study \u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e. PwPD were classified as having normal cognition (PD-NC) or PD-MCI according to the MDS PD-MCI Level II criteria \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e,\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e, having two impaired tests, coming from the same cognitive domain or different ones. Neuropsychological test impairment was defined as a z-score\u0026thinsp;\u0026le;\u0026thinsp;\u0026minus;\u0026thinsp;1.5 standard deviations (SD) relative to age-, sex-, and education-adjusted normative data.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eRisk Prediction tools\u003c/h2\u003e \u003cp\u003eThe risk scores considered in this analysis included the LIBRA, for dementia prediction in the general population, \u003csup\u003e20,62\u003c/sup\u003e, and the pPREDICT, aimed to estimate cognitive decline in PD, \u003csup\u003e24\u003c/sup\u003e. In addition to the modifiable and non-modifiable risk factors, from Carlisle and colleagues \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, several other risk factors were included in our proposed ModiCogPD risk score (Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eIn order to handle missing values, the missingness mechanisms of the study variables were assessed based on observed patterns and clinical knowledge to evaluate the Missing Completely At Random (MCAR), Missing At Random (MAR), or Missing Not At Random (MNAR) assumptions. The subset of variables used to calculate the different risk scores had missingness rates between 0% and 3%. Variables exhibiting missingness patterns consistent with MNAR, such as the Sniffin\u0026rsquo; Sticks Smell test score, were not imputed to avoid introducing systematic bias, and those participants representing a total of n\u0026thinsp;=\u0026thinsp;8, were removed from the dataset. The rest of the missing values in the risk factors followed MAR or MCAR missingness mechanisms, and therefore were imputed using a mean imputation procedure.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eModiCogPD\u003c/h2\u003e \u003cp\u003eFor the newly designed ModiCogPD, the final raw score was obtained by summing the 31 individual risk factors after dichotomization. Dichotomization thresholds were chosen based on those presented in previous risk scores and prior literature (Table\u0026nbsp;1), resulting in a binary classification of each variable, indicating the presence (=\u0026thinsp;1) or the absence of the risk factor (=\u0026thinsp;0). Following, higher ModiCogPD scores indicate a higher risk for PD-associated cognitive decline. The final ModiCogPD score ranges from 0 to 31.\u003c/p\u003e \u003cp\u003eImportantly, the subitems of ModiCogPD substantially overlap with the ones from pPREDICT tool, as ModiCogPD represents an extension of this framework. In addition, there is considerable overlap with the LIBRA score, given that both instruments incorporate vascular and lifestyle-related determinants of cognitive decline. The only information missing in the ModiCogPD with respect to LIBRA is hypercholesterolemia, which could not be directly assessed in our cohort (Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eAs some of the assessments from the other prediction tools were not available in our cohort, proxies were used, when necessary. In particular, to assess depression, the Beck Depression Inventory-I (BDI-I) \u003csup\u003e63\u003c/sup\u003e was used instead of the Geriatric Depression Scale (GDS) \u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. Moreover, the Movement Disorder Society -Unified Parkinson's Disease Rating Scale (MDS-UPDRS) \u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e part I item 4, assessing anxiety, was considered instead of the State-Trait Anxiety Inventory (STAI) \u003csup\u003e\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. To assess physical activity, we combined two items from the modified PD Risk Factor Questionnaire (PD-RFQ-U), Epi Info\u0026trade; developed by Caroline Tanner addressing weekly hours of vigorous and moderate physical activity. Vascular risk factors were evaluated through questions on self-perceived orthostatic hypotension and a semi-structured interview regarding cardiovascular medical history.\u003c/p\u003e \u003cp\u003e \u003cb\u003eTable\u0026nbsp;1 : Overview of included risk factors by risk predictions tools.\u003c/b\u003e \u003c/p\u003e \u003cp\u003e# new risk factors included in the ModiCogPD risk score\u003c/p\u003e \u003cp\u003e* If \u0026ldquo;yes\u0026rdquo; in any of these questions, then 1 point was attributed\u003c/p\u003e \u003cp\u003eWithin () the variables used are indicated\u003c/p\u003e \u003cp\u003eADL: activities of daily living; PIGD: postural instability and gait disorder.\u003c/p\u003e \u003cp\u003eAlcohol consumption refers to the question: In your lifetime, have you drunk 100 or more alcoholic drinks (beer, wine, liquor, spirits)? In your lifetime, have you ever regularly drunk alcohol, that is, at least one drink per week for 6 months or longer? If a participant answered \u0026ldquo;yes\u0026rdquo; to at least one of these questions, they were considered to have a history of alcohol consumption.\u003c/p\u003e \u003cp\u003e*maybe we have to write here the question for all the RFQ derived risk : smoking, caffeine, contusion, pesticides and toxics\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eDemographics\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLIBRA\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003epPREDICT\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModiCogPD\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003cp\u003eEducational level\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003cp\u003eSex\u003c/p\u003e \u003cp\u003eYears of education\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (\u0026gt;\u0026thinsp;65: 1 point)\u003c/p\u003e \u003cp\u003eSex (Male: 1 point)\u003c/p\u003e \u003cp\u003eYears of Education (\u0026lt;\u0026thinsp;12 years: 1 point)\u003c/p\u003e \u003cp\u003eMarital status#(if divorced, never married, widowed, separated: 1 point) \u003csup\u003e\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eNumber of children# (if no children: 1 point)\u003c/p\u003e \u003cp\u003eNumber of spoken languages# (if monolingual: 1 point) \u003csup\u003e\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinical\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMoCA score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMoCA score (\u0026lt;\u0026thinsp;26: 1 point) \u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePD specific characteristics\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDisease duration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDisease duration (\u0026gt;\u0026thinsp;5 years: 1 point)\u003c/p\u003e \u003cp\u003ePIGD score# ( if tremor/PIGD phenotype\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;0.90: 1 point)\u003c/p\u003e \u003cp\u003eFamily history of PD and dementia# (HPI) ( if yes: 1 point) \u003csup\u003e\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eObjective hyposmia# (Sniffin Sticks \u003csup\u003e\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e, Scores\u0026thinsp;\u0026le;\u0026thinsp;\u0026minus;\u0026thinsp;1.5 standard deviations below age- and sex-adjusted normative values: 1 point)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e(PD-specific) Characteristics \u0026ndash; potentially modifiable or treatable\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDepression\u003c/p\u003e \u003cp\u003eObesity\u003c/p\u003e \u003cp\u003ePhysical inactivity\u003c/p\u003e \u003cp\u003eHeart coronary disease\u003c/p\u003e \u003cp\u003eHypertension\u003c/p\u003e \u003cp\u003eHypercholesterolemia(missing variable, not included)\u003c/p\u003e \u003cp\u003eDiabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRBDSQ\u003c/p\u003e \u003cp\u003eDepression\u003c/p\u003e \u003cp\u003eAnxiety\u003c/p\u003e \u003cp\u003eObesity\u003c/p\u003e \u003cp\u003eLow physical activity\u003c/p\u003e \u003cp\u003eOrthostatic hypotension\u003c/p\u003e \u003cp\u003eHypertension\u003c/p\u003e \u003cp\u003eDiabetes\u003c/p\u003e \u003cp\u003eCardiovascular disease\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRBDSQ (total score\u0026thinsp;\u0026ge;\u0026thinsp;5)\u003c/p\u003e \u003cp\u003eDepression (BDI-I\u0026thinsp;\u0026ge;\u0026thinsp;16)\u003c/p\u003e \u003cp\u003eApathy# (Starkstein Apathy Scale\u0026thinsp;\u0026ge;\u0026thinsp;14: 1 point)\u003c/p\u003e \u003cp\u003eAnxiety (MDS-UPDRS-I.4 and Medical History)\u003c/p\u003e \u003cp\u003eObesity (BMI\u0026thinsp;\u0026ge;\u0026thinsp;25kg/m2)\u003c/p\u003e \u003cp\u003eLow physical activity*\u003c/p\u003e \u003cp\u003eModerate-severe motor (UPDRS-III\u0026thinsp;\u0026ge;\u0026thinsp;32 : 1 point)\u003c/p\u003e \u003cp\u003eModerate-severe non-motor aspects of ADL# (MDS-UPDRS-I\u0026thinsp;\u0026ge;\u0026thinsp;10: 1 point)*\u003c/p\u003e \u003cp\u003eModerate-severe motor aspects of ADL# (MDS-UPDRS-II\u0026thinsp;\u0026ge;\u0026thinsp;12: 1 point)*\u003c/p\u003e \u003cp\u003eModerate-severe motor complications# (MDS-UPDRS-IV\u0026thinsp;\u0026ge;\u0026thinsp;4: 1 point)* \u003csup\u003e71\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eDiabetes* (if yes: 1 point)\u003c/p\u003e \u003cp\u003eCardiovascular disease (if yes: 1 point)\u003c/p\u003e \u003cp\u003eHypertension (if yes: 1 point)\u003c/p\u003e \u003cp\u003eOrthostatic hypotension (if yes: 1 point)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLifestyle \u0026amp; environmental exposures\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCaffeine consumption#* (RFQ)\u003c/p\u003e \u003cp\u003eAlcohol consumption#* (RFQ)\u003c/p\u003e \u003cp\u003eTobacco consumption#* (RFQ)\u003c/p\u003e \u003cp\u003eContusion# (RFQ)*\u003c/p\u003e \u003cp\u003ePesticides exposure(at work and non-work)# (RFQ)*\u003c/p\u003e \u003cp\u003eSolvents and toxicants# (RFQ)*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eScore range\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0-15.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u0026ndash;13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u0026ndash;31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analyses\u003c/h2\u003e \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e \u003ch2\u003ePredictive Performance of the ModiCogPD\u003c/h2\u003e \u003cp\u003eTo assess the general performance of the new risk score to differentiate between PD-NC and PD-MCI and enable a comparison to established risk scores, the area under the operating characteristic curve (AUC-ROC) was calculated for each risk score (LIBRA, pPREDICT, ModiCogPD). This metric measures how well a continuous score separates two different classes, in this case PD-NC and PD-MCI. A stratified five-fold cross validation was implemented ensuring that each fold preserved the same ratio of positive and negative outcomes while providing a more robust estimate of model discrimination.\u003c/p\u003e \u003cp\u003eTo enable a more granular evaluation of the predictive performance of ModiCogPD, a multivariable logistic regression (LR) model was developed to compare the use of individual risk components with the aggregated global risk score. The cognitive status of PwPD (PD-NC or PD-MCI), derived from the MDS PD-MCI Level II criteria, was used as the dependent variable, while the individual subitems of the ModiCogPD served as independent variables. Prior to model fitting, all predictors underwent a scaling step tailored to the different types of variables aiming to ensure comparable penalization across variables and to improve numerical stability. A nested cross-validation approach was used, consisting of a 5-fold CV outer loop, aiming to provide a robust estimate of the performance in unseen data, and a 3-fold CV inner loop, aimed to optimize the parameters of the model. Within each outer loop, model parameters were tuned using the inner loop, with the best model being evaluated on the held-out outer test fold. Performance metrics were then averaged across outer folds. Given that the size of the study population was limited relative to the number of independent variables (events-per-variable ratio (EPV)\u0026thinsp;~\u0026thinsp;3), the model was considered prone to overfitting and unstable estimates. Therefore, penalised LR approaches, specifically L2 regularization, were adopted to reduce overfitting by shrinking coefficient estimates, thereby improving model stability and generalizability \u003csup\u003e\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e. As penalization alters the likelihood function, standard inferential statistics such as \u003cem\u003ep\u003c/em\u003e-values are not directly available. Therefore, unpenalized LR models were additionally fitted for descriptive and inferential purposes.\u003c/p\u003e \u003cp\u003eIn addition, internal validation was further assessed through bootstrapping to quantify potential model optimism, Calibration was evaluated by examining the agreement between predicted probabilities and observed outcomes, using the Brier score and, more specifically, the calibration intercept and slope, which indicate systematic over- or underestimation of risk and the agreement between predicted and observed probabilities across the range of predicted risk. Model optimism, reflecting potential overfitting of the LR model to the training data, was quantified using 500 bootstrap resamples. For each bootstrap sample, the model was refitted and performance metrics, including discrimination and calibration, were calculated on both the bootstrap sample (apparent performance) and the original dataset. Model optimism was estimated as the difference between the bootstrap sample and the original dataset performances. The final reported metrics were corrected by subtracting the average model optimism across all bootstrap samples \u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of the most relevant modifiable and non-modifiable risk factors\u003c/h2\u003e \u003cp\u003eFurther analyses were conducted, on the LR-based risk score model with the highest discrimination, to investigate the association of modifiable and non-modifiable individual risk factors with cognitive status. A multivariate unpenalized LR model was fitted to obtain interpretable effect estimates, confidence intervals, and \u003cem\u003ep\u003c/em\u003e-values. The model specification followed the approach described in the Methods subsection \u0026ldquo;Predictive Performance of the ModiCogPD\u0026rdquo;, with PwPD cognitive status as the outcome and individual risk score components as covariates.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eWeighted ModiCogPD\u003c/h2\u003e \u003cp\u003eAs an extension of the previous LR analyses incorporating subitems to enhance discriminatory performance, we further examined alternative weighting procedures in the development and improvement of the ModiCogPD risk score performance. A review of the literature on weighting approaches in the CI field revealed heterogeneous approaches. While certain scores, such as LIBRA \u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e, reported improved performance following the application of a weighting strategy based on expert consensus through a Delphi process, others, including MoPaRDS \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, which employed a regression coefficient\u0026ndash;based scoring system, did not demonstrate a clear advantage over unweighted formulations. In light of these mixed results, we aimed to formally assess whether a data-driven weighting approach, based on regression-derived coefficients, improves the discrimination and calibration of the ModiCogPD risk score compared with an unweighted approach. A regression coefficient\u0026ndash;based scoring system was used to assign weights to the individual sub-items of the general risk score. For this approach, first the study population was split into a train and a test dataset, each of them containing 50% of the patients, followed by a scaling step tailored to the different types of variables. Once the pre-processing was finalised, a LR with all the individual risk factors was trained against the end point variable reflecting the cognitive status on the train dataset. Due to the small sample size, L2 regularization was adopted to reduce overfitting by shrinking coefficient estimates, thereby improving model stability and generalizability. Calibration and optimism were checked to ensure that the model\u0026rsquo;s apparent performance was not overestimated, especially when applied to new data. The raw coefficients were normalized by scaling to the smallest non-trivial coefficient and rounding to the closest integer, easing their usability and interpretability as a risk score \u003csup\u003e\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003ePredictive strength on the posterior cognitive decline trajectories\u003c/h2\u003e \u003cp\u003eSurvival analyses were used to evaluate the prognostic value for the entire set of risk scores by modeling the association between the baseline risk score and time-to-event outcomes in two study populations: the global PwPD cohort (n\u0026thinsp;=\u0026thinsp;201) and a subset of newly diagnosed PwPD (n\u0026thinsp;=\u0026thinsp;90). More specifically, Cox Hazard models were applied to determine the effect of baseline risk score hazard ratios on global cognitive performance trajectories, as measured by the MoCA score. In the case of the global PwPD cohort the threshold for cognitive decline was based on \u003csup\u003e55\u003c/sup\u003e who defined decreased global cognition as a Mini\u0026ndash;Mental State Examination (MMSE) score\u0026thinsp;\u0026lt;\u0026thinsp;26. As MMSE data were not available in the present study, we used a MoCA score\u0026thinsp;\u0026le;\u0026thinsp;21 based on previously published MoCA\u0026ndash;MMSE conversion rates \u003csup\u003e\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e\u003c/sup\u003e. Participants with cognitive performance below the specified threshold at baseline were excluded from the analyses. To account for potential confounders, age, years of education, sex and baseline MoCA score were included in the models. To account for the potential non-linear effect, particularly in the relationship between age and cognitive decline, a Cox model in which age was modeled as a restricted cubic spline with four knots was fitted \u003csup\u003e\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e\u003c/sup\u003e. The significance of the non-linear association was assessed by comparing the spline model to a linear model. As the resulting \u003cem\u003ep\u003c/em\u003e-value indicated no significant non-linearity (for a significance level α\u0026thinsp;=\u0026thinsp;0.05), age was subsequently included as a linear confounder. The presence of competing events was evaluated. Death prior to the event of interest occurred in a small subset of PwPD (n\u0026thinsp;=\u0026thinsp;7) and was therefore considered a potential competing event. Given the limited number of deaths, formal competing-risk regression was not pursued. Cause-specific Cox proportional hazards models were instead fitted, with PwPD censored at the time of death. This approach estimates the instantaneous risk of the event of interest among PwPD who remain alive and event-free.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eRisk scores associated with cognitive domain impairment\u003c/h2\u003e \u003cp\u003eFinally, the association between the risk scores and the presence of impairment across each of the five cognitive domains was evaluated. LR models were applied to study the association between each risk score and impairment in each cognitive domain (memory, language, attention, visuospatial functions and executive functions) defined as target, adjusting for age, sex and years of education. Impairment in a cognitive domain was defined as a z-score of \u0026le; -1.5 SD below the age-, sex-, and years of education adjusted normative data (see section \u003cem\u003eCognitive Impairment definition for additional details\u003c/em\u003e).\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eData availability\u003c/h2\u003e\n\u003cp\u003ePatient data used in the preparation of this manuscript were obtained from the National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD). NCER-PD datasets are not publicly available, as they are linked to the Luxembourg Parkinson’s Study and its internal regulations. The NCER-PD Consortium is willing to share its available data. Its access policy was devised based on the study ethics documents, including the informed consent form, as approved by National Ethics Board (CNER Ref: 201407/13) and Data Protection Committee (CNPD Ref: 446/2017). Requests to access datasets should be directed to the Data and Sample Access Committee via email: \u003cu\[email protected]\u003c/u\u003e.\u003c/p\u003e\n\u003ch2\u003eCode availability\u003c/h2\u003e\n\u003cp\u003eThe underlying code for this study is available on .\u003cu\u003euniluxembourg / LCSB / Digital Medicine / Gabriel_phd_project / Collaboration With Claire And Sonja · GitLab\u003c/u\u003e,\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eAcknowledgments\u003c/h2\u003e\n\u003cp\u003eThis project was supported by the Luxembourg National Research Fund (FNR) through the FNR/PEARL/dHealthPD/14146272 and FNR/PREVENE/14781425. The National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD) was funded by the Luxembourg National Research Fund (FNR) (FNR/NCER13/BM/11264123).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData used in the preparation of this manuscript were obtained from the National Centre of Excellence in Research on Parkinson's Disease (NCER-PD).\u003c/p\u003e\n\u003cp\u003eWe would like to thank all participants of the Luxembourg Parkinson’s Study for their important support of our research. Furthermore, we acknowledge the joint effort of the National Centre of Excellence in Research on Parkinson’s Disease (NCER-PD) Consortium members from the partner institutions Luxembourg Centre for Systems Biomedicine, Luxembourg Institute of Health, Centre Hospitalier de Luxembourg, and Laboratoire National de Santé generally contributing to the Luxembourg Parkinson’s Study as listed below\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions :\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCP, SRJ, GMT conceived and designed the study. SRJ, CP, contributed to data collection. GMT contributed to the statistical analysis plan and led the statistical analyses. AO and PMC supported the statistical analysis plan with their expertise. CP, SRJ, GMT, PMC drafted the manuscript. VES, AS, DE, EK, JK and RK provided expertise related to this manuscript. All authors critically revised the manuscript for important intellectual content and approved the final version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor’s list\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eList of NCER-PD consortium members\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eMariella GRAZIANO⁷, Alexandre BISDORFF⁵, Rene DONDELINGER⁵, Elodie THIRY³, Gelani ZELIMKHANOV³, Guy BERCHEM³, Liliana VILAS BOAS³, Linda HANSEN³, Martine GOERGEN³, Nancy DE BREMAEKER³, Nico DIEDERICH³, Romain NATI³, Roxane BATUTU³, Sylvia HERBRINK³, Jochen KLUCKEN¹,³, Rejko KRÜGER¹,²,³, Claire PAULY²,³, Lukas PAVELKA²,³, Marijus GIRAITIS²,³, Maria Fernanda NIÑO URIBE¹,³, Achilleas PEXARAS², Alexander HUNDT², Alexia MENDIBIDE², Ana Festas \u0026nbsp;LOPES², Angelo FERRARI², Brian DEWITT², Carlos GAMIO², Estelle HENRY², Gaël HAMMOT², Geeta ACHARYA², Hermann THIEN², Ilsé RICHARD², Johanna TROUET², Kate SOKOLOWSKA², Katy BEAUMONT², Laura GEORGES², Lorieza CASTILLO², Lucie REMARK², Maeva MUNSCH², Margaux HENRY², Maud THERESINE², Olga KOFANOVA², Olivia ROLAND², Pauline LAMBERT², Saïda MTIMET², Wim AMMERLANN², Anne GRÜNEWALD¹, Armin \u0026nbsp; RAUSCHENBERGER¹,², Dheeraj REDDY BOBBILI¹, Ekaterina SOBOLEVA¹,³, Elisa GÓMEZ DE LOPE¹, Enrico GLAAB¹, Evi WOLLSCHEID-LENGELING¹, Francoise MEISCH¹, Giuseppe ARENA¹, Ibrahim BOUSSAAD¹, Jens SCHWAMBORN¹, Kirsten ROOMP¹, ¹⁰, Michael T. HENEKA¹, Michele BASSIS¹, Muhammad ALI¹, Jade JABER¹,³, Patricia MARTINS CONDE¹, Patrick MAY¹, Paul WILMES¹, Piotr GAWRON¹, Rebecca TING JIIN LOO¹, Reinhard SCHNEIDER¹, Ruxandra SOARE¹, Sabine SCHMITZ¹, Sarah NICKELS¹, Sascha HERZINGER¹, Sinthuja PACHCHEK¹, Soumyabrata GHOSH¹, Stefano SAPIENZA¹, Valentin GROUES¹, Venkata SATAGOPAM¹, Iñigo YOLDI BERGUA¹, Gabriel MARTINEZ TIRADO¹, Jochen OHNMACHT², Anne-Marie HANFF², ¹⁰, ¹¹, Carlos VEGA², Eduardo ROSALES², Fozia NOOR², Gessica CONTESOTTO², Gloria AGUAYO², Guilherme MARQUES², Jérôme GRAAS², Joëlle FRITZ², Magali PERQUIN², Manon GANTENBEIN², Maura MINELLI², Michel VAILLANT², Myriam ALEXANDRE², Myriam MENSTER², Olena TSURKALENKO², Sibylle BÉCHET³, Jón GALES², Emna BOUHAJJA², Ulf NEHRBASS², Victoria LORENTZ², Zied LANDOULSI², Sonja JÓNSDÓTTIR², David BOUVIER⁴, Katrin FRAUENKNECHT⁴, Michel MITTELBRONN¹, ², ⁴, ¹⁰, ¹², ¹³, Roseline LENTZ⁶, Jean-Paul NICOLAY⁹, Nadine JACOBY⁸, Isabel SCHWAMINGER¹, Liyousew BORGA¹, Sijmen VAN SCHAGEN¹, Alan CASTRO MEJIA¹, Francesca TERRANOVA¹, Messaline FOMO¹, Francesca BOSCHI¹, Niloofar KHERADBIN¹, Isabelle ROLIN³\u003c/p\u003e\n\u003cp\u003e1 Luxembourg Centre for Systems Biomedicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg\u003c/p\u003e\n\u003cp\u003e2 Luxembourg Institute of Health, Strassen, Luxembourg\u003c/p\u003e\n\u003cp\u003e3 Centre Hospitalier de Luxembourg, Strassen, Luxembourg\u003c/p\u003e\n\u003cp\u003e4 Laboratoire National de Santé, Dudelange, Luxembourg\u003c/p\u003e\n\u003cp\u003e5 Centre Hospitalier Emile Mayrisch, Esch-sur-Alzette, Luxembourg\u003c/p\u003e\n\u003cp\u003e6 Parkinson Luxembourg Association, Leudelange, Luxembourg\u003c/p\u003e\n\u003cp\u003e7 Association of Physiotherapists in Parkinsons Disease Europe, Esch-sur-Alzette, Luxembourg\u003c/p\u003e\n\u003cp\u003e8 Private practice, Ettelbruck, Luxembourg\u003c/p\u003e\n\u003cp\u003e9 Private practice, Luxembourg, Luxembourg\u003c/p\u003e\n\u003cp\u003e10 Faculty of Science, Technology and Medicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg\u003c/p\u003e\n\u003cp\u003e11 Department of Epidemiology, CAPHRI School for Public Health and Primary Care, Maastricht University Medical Centre+, Maastricht, the Netherlands\u003c/p\u003e\n\u003cp\u003e12 Luxembourg Center of Neuropathology, Dudelange, Luxembourg\u003c/p\u003e\n\u003cp\u003e13 Department of Life Sciences and Medicine, University of Luxembourg, Esch-sur-Alzette, Luxembourg\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;Competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare no financial or non-financial competing interests.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eDorsey, E. R., Sherer, T., Okun, M. S. \u0026amp; Bloem, B. R. The Emerging Evidence of the Parkinson Pandemic. \u003cem\u003eJ Parkinsons Dis\u003c/em\u003e 8, S3\u0026ndash;S8 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, X. \u003cem\u003eet al.\u003c/em\u003e Trends and hotspots in non-motor symptoms of Parkinson\u0026rsquo;s disease: a 10-year bibliometric analysis. \u003cem\u003eFront Aging Neurosci\u003c/em\u003e 16, 1335550 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeWitt, P. A. L. \u0026amp; Chaudhuri, K. R. Unmet needs in Parkinson disease: Motor and non-motor. \u003cem\u003eParkinsonism \u0026amp; Related Disorders\u003c/em\u003e 80, S7\u0026ndash;S12 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Q., Aldridge, G. M., Narayanan, N. S., Anderson, S. W. \u0026amp; Uc, E. Y. Approach to Cognitive Impairment in Parkinson\u0026rsquo;s Disease. \u003cem\u003eNeurotherapeutics\u003c/em\u003e 17, 1495\u0026ndash;1510 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBloem, B. R., Okun, M. S. \u0026amp; Klein, C. Parkinson\u0026rsquo;s disease. \u003cem\u003eLancet\u003c/em\u003e 397, 2284\u0026ndash;2303 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBowring, F. \u003cem\u003eet al.\u003c/em\u003e Exploration of whether socioeconomic factors affect the results of priority setting partnerships: updating the top 10 research priorities for the management of Parkinson\u0026rsquo;s in an international setting. \u003cem\u003eBMJ Open\u003c/em\u003e 12, e049530 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldman, J. G. \u003cem\u003eet al.\u003c/em\u003e Cognitive impairment in Parkinson\u0026rsquo;s disease: a report from a multidisciplinary symposium on unmet needs and future directions to maintain cognitive health. \u003cem\u003eNPJ Parkinsons Dis\u003c/em\u003e 4, 19 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoletti, M. \u003cem\u003eet al.\u003c/em\u003e Mild cognitive impairment and cognitive-motor relationships in newly diagnosed drug-naive patients with Parkinson\u0026rsquo;s disease. \u003cem\u003eJ Neurol Neurosurg Psychiatry\u003c/em\u003e 83, 601\u0026ndash;606 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAarsland, D. \u003cem\u003eet al.\u003c/em\u003e Parkinson disease-associated cognitive impairment. \u003cem\u003eNat. Rev. Dis. Primers\u003c/em\u003e 7, 47 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDomell\u0026ouml;f, M. E., Ekman, U., Forsgren, L. \u0026amp; Elgh, E. Cognitive function in the early phase of Parkinson\u0026rsquo;s disease, a five-year follow-up. \u003cem\u003eActa Neurol Scand\u003c/em\u003e 132, 79\u0026ndash;88 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eButer, T. C. \u003cem\u003eet al.\u003c/em\u003e Dementia and survival in Parkinson disease: a 12-year population study. \u003cem\u003eNeurology\u003c/em\u003e 70, 1017\u0026ndash;1022 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHely, M. A., Reid, W. G. J., Adena, M. A., Halliday, G. M. \u0026amp; Morris, J. G. L. The Sydney multicenter study of Parkinson\u0026rsquo;s disease: the inevitability of dementia at 20 years. \u003cem\u003eMov Disord\u003c/em\u003e 23, 837\u0026ndash;844 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGallagher, J. \u003cem\u003eet al.\u003c/em\u003e Long-term dementia risk in Parkinson disease. \u003cem\u003eNeurology\u003c/em\u003e 103, e209699 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGavelin, H. M. \u003cem\u003eet al.\u003c/em\u003e Computerized cognitive training in Parkinson\u0026rsquo;s disease: A systematic review and meta-analysis. \u003cem\u003eAgeing Res. Rev.\u003c/em\u003e 80, 101671 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFolkerts, A.-K. \u003cem\u003eet al.\u003c/em\u003e Can physical exercise be considered as a promising enhancer of global cognition in people with Parkinson\u0026rsquo;s disease? Results of a systematic review and meta-analysis. \u003cem\u003eJ. Parkinsons. Dis.\u003c/em\u003e 14, S115\u0026ndash;S133 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Z., Zhu, C., Miao, W. \u0026amp; Zhang, Y. Effects of resistance and balance training on motor and non-motor symptoms in patients with Parkinson\u0026rsquo;s disease: a meta-analysis. \u003cem\u003eNeurol. Res.\u003c/em\u003e 1\u0026ndash;14 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchrag, A., Siddiqui, U. F., Anastasiou, Z., Weintraub, D. \u0026amp; Schott, J. M. Clinical variables and biomarkers in prediction of cognitive impairment in patients with newly diagnosed Parkinson\u0026rsquo;s disease: a cohort study. \u003cem\u003eLancet Neurol.\u003c/em\u003e 16, 66\u0026ndash;75 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYousaf, T., Pagano, G., Niccolini, F. \u0026amp; Politis, M. Predicting cognitive decline with non-clinical markers in Parkinson\u0026rsquo;s disease (PRECODE-2). \u003cem\u003eJ. Neurol.\u003c/em\u003e 266, 1203\u0026ndash;1210 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGramotnev, G., Gramotnev, D. K. \u0026amp; Gramotnev, A. Parkinson\u0026rsquo;s disease prognostic scores for progression of cognitive decline. \u003cem\u003eSci. Rep.\u003c/em\u003e 9, 17485 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchiepers, O. J. G. \u003cem\u003eet al.\u003c/em\u003e Lifestyle for Brain Health (LIBRA): a new model for dementia prevention. \u003cem\u003eInt J Geriatr Psychiatry\u003c/em\u003e 33, 167\u0026ndash;175 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDawson, B. K. \u003cem\u003eet al.\u003c/em\u003e Office-Based Screening for Dementia in Parkinson Disease: The Montreal Parkinson Risk of Dementia Scale in 4 Longitudinal Cohorts. \u003cem\u003eJAMA Neurol\u003c/em\u003e 75, 704\u0026ndash;710 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLivingston, G. \u003cem\u003eet al.\u003c/em\u003e Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. \u003cem\u003eThe Lancet\u003c/em\u003e 404, 572\u0026ndash;628 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMostert, C. M. \u003cem\u003eet al.\u003c/em\u003e Broadening dementia risk models: building on the 2024 Lancet Commission report for a more inclusive global framework. \u003cem\u003eEBioMedicine\u003c/em\u003e 120, 105950 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarlisle, T. C., Medina, L. D. \u0026amp; Holden, S. K. Original research: initial development of a pragmatic tool to estimate cognitive decline risk focusing on potentially modifiable factors in Parkinson\u0026rsquo;s disease. \u003cem\u003eFront Neurosci\u003c/em\u003e 17, 1278817 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, C. \u0026amp; Armstrong, M. J. Treatment of Parkinson\u0026rsquo;s Disease with Cognitive Impairment: Current Approaches and Future Directions. \u003cem\u003eBehav Sci (Basel)\u003c/em\u003e 11, (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalbe, E., Warnecke, T., Eggers, C., Ophey, A. \u0026amp; Folkerts, A.-K. Prevention of cognitive impairment and dementia in people with Parkinson\u0026rsquo;s disease: A call-to-action. \u003cem\u003eJ. Parkinsons. Dis.\u003c/em\u003e 15, 1353\u0026ndash;1366 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHipp, G. \u003cem\u003eet al.\u003c/em\u003e The Luxembourg Parkinson\u0026rsquo;s Study: A Comprehensive Approach for Stratification and Early Diagnosis. \u003cem\u003eFront. Aging Neurosci.\u003c/em\u003e 10, 408277 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwon, K.-Y., Park, S., Kim, R. O., Lee, E. J. \u0026amp; Lee, M. Associations of cognitive dysfunction with motor and non-motor symptoms in patients with de novo Parkinson\u0026rsquo;s disease. \u003cem\u003eSci. Rep.\u003c/em\u003e 12, 11461 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, Y. \u003cem\u003eet al.\u003c/em\u003e Predictors of cognitive impairment in Parkinson\u0026rsquo;s disease: a systematic review and meta-analysis of prospective cohort studies. \u003cem\u003eJ. Neurol.\u003c/em\u003e 268, 2713\u0026ndash;2722 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStoker, T. B. \u003cem\u003eet al.\u003c/em\u003e Impact of GBA1 variants on long-term clinical progression and mortality in incident Parkinson\u0026rsquo;s disease. \u003cem\u003eJ. Neurol. Neurosurg. Psychiatry\u003c/em\u003e 91, 695\u0026ndash;702 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis, M. Y. \u003cem\u003eet al.\u003c/em\u003e Association of \u003cem\u003eGBA\u003c/em\u003e mutations and the E326K polymorphism with motor and cognitive progression in Parkinson disease. \u003cem\u003eJAMA Neurol.\u003c/em\u003e 73, 1217 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIwaki, H. \u003cem\u003eet al.\u003c/em\u003e Genetic risk of Parkinson disease and progression:: An analysis of 13 longitudinal cohorts. \u003cem\u003eNeurol. Genet.\u003c/em\u003e 5, e348 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaiano, C., Barone, P., Trojano, L. \u0026amp; Santangelo, G. Prevalence and clinical aspects of mild cognitive impairment in Parkinson\u0026rsquo;s disease: A meta-analysis. \u003cem\u003eMov. Disord.\u003c/em\u003e 35, 45\u0026ndash;54 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaher, S. \u003cem\u003eet al.\u003c/em\u003e Treatment of apathy in Parkinson\u0026rsquo;s disease and implications for underlying pathophysiology. \u003cem\u003eJ. Clin. Med.\u003c/em\u003e 13, 2216 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePlant, O. \u003cem\u003eet al.\u003c/em\u003e A cognitive-behavioral model of apathy in Parkinson\u0026rsquo;s disease. \u003cem\u003eParkinsons Dis.\u003c/em\u003e 2024, 2820257 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMele, B. \u003cem\u003eet al.\u003c/em\u003e Non-pharmacologic interventions to treat apathy in Parkinson\u0026rsquo;s disease: A realist review. \u003cem\u003eClin. Park. Relat. Disord.\u003c/em\u003e 4, 100096 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKamal, H. \u003cem\u003eet al.\u003c/em\u003e Alcohol use disorder, neurodegeneration, Alzheimer\u0026rsquo;s and Parkinson's disease: Interplay between oxidative stress, neuroimmune response and excitotoxicity. \u003cem\u003eFront. Cell. Neurosci.\u003c/em\u003e 14, 282 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAborageh, M., H\u0026auml;hnel, T., Martins Conde, P., Klucken, J. \u0026amp; Fr\u0026ouml;hlich, H. Predicting dementia in people with Parkinson\u0026rsquo;s disease. \u003cem\u003eNPJ Parkinsons Dis.\u003c/em\u003e 11, 126 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurray, D. K., Sacheli, M. A., Eng, J. J. \u0026amp; Stoessl, A. J. The effects of exercise on cognition in Parkinson\u0026rsquo;s disease: a systematic review. \u003cem\u003eTransl. Neurodegener.\u003c/em\u003e 3, 5 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X., Molsberry, S. A., Schwarzschild, M. A., Ascherio, A. \u0026amp; Gao, X. Association of diet and physical activity with all-cause mortality among adults with Parkinson disease. \u003cem\u003eJAMA Netw. Open\u003c/em\u003e 5, e2227738 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYpinga, J. H. L. \u003cem\u003eet al.\u003c/em\u003e Effects of specialised physiotherapy on mortality in Parkinson\u0026rsquo;s disease: a prospective observational study. \u003cem\u003eNPJ Parkinsons Dis.\u003c/em\u003e 11, 214 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalbe, E. \u003cem\u003eet al.\u003c/em\u003e German Society of Neurology guidelines for the diagnosis and treatment of cognitive impairment and affective disorders in people with Parkinson\u0026rsquo;s disease: new spotlights on diagnostic procedures and non-pharmacological interventions. \u003cem\u003eJ. Neurol.\u003c/em\u003e 271, 7330\u0026ndash;7357 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDukelow, T. \u003cem\u003eet al.\u003c/em\u003e Modifiable risk factors for dementia, and awareness of brain health behaviors: Results from the Five Lives Brain Health Ireland Survey (FLBHIS). \u003cem\u003eFront. Psychol.\u003c/em\u003e 13, 1070259 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlbus, P., Folkerts, A.-K., Kessler, J., K\u0026ouml;hler, S. \u0026amp; Kalbe, E. Sociodemographic differences in dementia prevention knowledge in Germany: Implications for targeted health communication. \u003cem\u003eJ. Prev. Alzheimers Dis.\u003c/em\u003e 13, 100517 (2026).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchr\u0026ouml;der, V. E. \u003cem\u003eet al.\u003c/em\u003e Programme dementia prevention (pdp): A nationwide program for personalized prevention in Luxembourg. \u003cem\u003eJ. Alzheimers. Dis.\u003c/em\u003e 97, 791\u0026ndash;804 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErz, D. \u003cem\u003eet al.\u003c/em\u003e Dementia prevention through the eyes of individuals at risk: insights from a satisfaction survey within the programme for dementia prevention in Luxembourg. \u003cem\u003eFront. Aging\u003c/em\u003e 7, 1712500 (2026).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYarnall, A. J. \u003cem\u003eet al.\u003c/em\u003e Characterizing mild cognitive impairment in incident Parkinson disease: the ICICLE-PD study. \u003cem\u003eNeurology\u003c/em\u003e 82, 308\u0026ndash;316 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJellinger, K. A. Mild cognitive impairment in Parkinson\u0026rsquo;s disease: current view. \u003cem\u003eFront. Cogn.\u003c/em\u003e 3, (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson, D. K., Langford, Z., Garnier-Villarreal, M., Morris, J. C. \u0026amp; Galvin, J. E. Onset of mild cognitive impairment in Parkinson disease. \u003cem\u003eAlzheimer Dis. Assoc. Disord.\u003c/em\u003e 30, 127\u0026ndash;133 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarr, J. E., Graham, R. B., Hofer, S. M. \u0026amp; Muniz-Terrera, G. When does cognitive decline begin? A systematic review of change point studies on accelerated decline in cognitive and neurological outcomes preceding mild cognitive impairment, dementia, and death. \u003cem\u003ePsychol. Aging\u003c/em\u003e 33, 195\u0026ndash;218 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScharfenberg, D. \u003cem\u003eet al.\u003c/em\u003e Understanding cognitive domains in Parkinson\u0026rsquo;s disease: A scoping review of empirical studies. \u003cem\u003eNeuropsychol. Rev.\u003c/em\u003e (2026) doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11065-025-09691-5\u003c/span\u003e\u003cspan address=\"10.1007/s11065-025-09691-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDalrymple-Alford, J. C. \u003cem\u003eet al.\u003c/em\u003e Characterizing mild cognitive impairment in Parkinson\u0026rsquo;s disease: MCI Criteria for Parkinson's Disease. \u003cem\u003eMov. Disord.\u003c/em\u003e 26, 629\u0026ndash;636 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLitvan, I. \u003cem\u003eet al.\u003c/em\u003e Diagnostic criteria for mild cognitive impairment in Parkinson\u0026rsquo;s disease: Movement Disorder Society Task Force guidelines. \u003cem\u003eMov Disord\u003c/em\u003e 27, 349\u0026ndash;356 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeckers, K. \u003cem\u003eet al.\u003c/em\u003e Target risk factors for dementia prevention: a systematic review and Delphi consensus study on the evidence from observational studies: Major risk factors for dementia prevention. \u003cem\u003eInt. J. Geriatr. Psychiatry\u003c/em\u003e 30, 234\u0026ndash;246 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDubois, B. \u003cem\u003eet al.\u003c/em\u003e Diagnostic procedures for Parkinson\u0026rsquo;s disease dementia: recommendations from the movement disorder society task force. \u003cem\u003eMov Disord\u003c/em\u003e 22, 2314\u0026ndash;2324 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;Caoimh, R., Foley, M. J., Timmons, S. \u0026amp; Molloy, D. W. Screening for cognitive impairment in movement disorders: Comparison of the Montreal Cognitive Assessment and Quick Mild Cognitive Impairment screen in Parkinson\u0026rsquo;s disease and Lewy body dementia. \u003cem\u003eJ. Alzheimers Dis. Rep.\u003c/em\u003e 8, 971\u0026ndash;980 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, J. \u003cem\u003eet al.\u003c/em\u003e Predictors of cognitive impairment in newly diagnosed Parkinson\u0026rsquo;s disease with normal cognition at baseline: A 5-year cohort study. \u003cem\u003eFront. Aging Neurosci.\u003c/em\u003e 15, 1142558 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNasreddine, Z. S. \u003cem\u003eet al.\u003c/em\u003e The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool For Mild Cognitive Impairment. \u003cem\u003eJournal of the American Geriatrics Society\u003c/em\u003e 53, 695\u0026ndash;699 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShirk, S. D. \u003cem\u003eet al.\u003c/em\u003e A web-based normative calculator for the uniform data set (UDS) neuropsychological test battery. \u003cem\u003eAlzheimer\u0026rsquo;s Research \u0026amp; Therapy\u003c/em\u003e 3, 32 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez Tirado, G. \u003cem\u003eet al.\u003c/em\u003e Data-driven clinical decision support tool for diagnosing mild cognitive impairment in Parkinson\u0026rsquo;s disease. \u003cem\u003eNPJ Parkinsons Dis.\u003c/em\u003e 12, 15 (2026).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoel, J. A. \u003cem\u003eet al.\u003c/em\u003e Level I PD-MCI using global cognitive tests and the risk for Parkinson\u0026rsquo;s disease dementia. \u003cem\u003eMov. Disord. Clin. Pract.\u003c/em\u003e 9, 479\u0026ndash;483 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVos, S. J. B. \u003cem\u003eet al.\u003c/em\u003e Modifiable Risk Factors for Prevention of Dementia in Midlife, Late Life and the Oldest-Old: Validation of the LIBRA Index. \u003cem\u003eJ Alzheimers Dis\u003c/em\u003e 58, 537\u0026ndash;547 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeck, A. T., Ward, C. H., Mendelson, M., Mock, J. \u0026amp; Erbaugh, J. An inventory for measuring depression. \u003cem\u003eArch. Gen. Psychiatry\u003c/em\u003e 4, 561\u0026ndash;571 (1961).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYesavage, J. A. \u003cem\u003eet al.\u003c/em\u003e Development and validation of a geriatric depression screening scale: a preliminary report. \u003cem\u003eJ Psychiatr Res\u003c/em\u003e 17, 37\u0026ndash;49 (1982).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoetz, C. G. \u003cem\u003eet al.\u003c/em\u003e Movement Disorder Society-sponsored revision of the Unified Parkinson\u0026rsquo;s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results. \u003cem\u003eMovement Disorders\u003c/em\u003e 23, 2129\u0026ndash;2170 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpielberger, C. D. State-trait anxiety inventory for adults. \u003cem\u003ePsycTESTS Dataset\u003c/em\u003e American Psychological Association (APA) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/t06496-000\u003c/span\u003e\u003cspan address=\"10.1037/t06496-000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1983).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, D., Zheng, W. \u0026amp; Li, K. The relationship between marital status and cognitive impairment in Chinese older adults: the multiple mediating effects of social support and depression. \u003cem\u003eBMC Geriatr.\u003c/em\u003e 24, 367 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePacifico, D. \u003cem\u003eet al.\u003c/em\u003e Associations of multilingualism and language proficiency with cognitive functioning: epidemiological evidence from the SwissDEM study in community dwelling older adults and long-term care residents. \u003cem\u003eBMC Geriatr.\u003c/em\u003e 23, 629 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRocca, W. A. \u003cem\u003eet al.\u003c/em\u003e Risk of cognitive impairment or dementia in relatives of patients with Parkinson disease. \u003cem\u003eArch. Neurol.\u003c/em\u003e 64, 1458\u0026ndash;1464 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHummel, T., Sekinger, B., Wolf, S. R., Pauli, E. \u0026amp; Kobal, G. \u0026lsquo;Sniffin\u0026rsquo; sticks\u0026rsquo;: Olfactory performance assessed by the combined testing of odour identification, odor discrimination and olfactory threshold. \u003cem\u003eChem. Senses\u003c/em\u003e 22, 39\u0026ndash;52 (1997).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez-Mart\u0026iacute;n, P. \u003cem\u003eet al.\u003c/em\u003e Parkinson\u0026rsquo;s disease severity levels and MDS-Unified Parkinson's Disease Rating Scale. \u003cem\u003eParkinsonism Relat. Disord.\u003c/em\u003e 21, 50\u0026ndash;54 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCessie, S. L. \u0026amp; Van Houwelingen, J. C. Ridge estimators in logistic regression. \u003cem\u003eJ. R. Stat. Soc. Ser. C. Appl. Stat.\u003c/em\u003e 41, 191 (1992).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteyerberg, E. W. \u003cem\u003eet al.\u003c/em\u003e Internal validation of predictive models: efficiency of some procedures for logistic regression analysis. \u003cem\u003eJ. Clin. Epidemiol.\u003c/em\u003e 54, 774\u0026ndash;781 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMehta, H. B., Mehta, V., Girman, C. J., Adhikari, D. \u0026amp; Johnson, M. L. Regression coefficient-based scoring system should be used to assign weights to the risk index. \u003cem\u003eJ. Clin. Epidemiol.\u003c/em\u003e 79, 22\u0026ndash;28 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Steenoven, I. \u003cem\u003eet al.\u003c/em\u003e Conversion between mini-mental state examination, montreal cognitive assessment, and dementia rating scale-2 scores in Parkinson\u0026rsquo;s disease: Conversion of cognitive screening scales in PD. \u003cem\u003eMov. Disord.\u003c/em\u003e 29, 1809\u0026ndash;1815 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMolinari, N., Daur\u0026egrave;s, J. P. \u0026amp; Durand, J. F. Regression splines for threshold selection in survival data analysis. \u003cem\u003eStat. Med.\u003c/em\u003e 20, 237\u0026ndash;247 (2001).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-parkinsons-disease","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjparkd","sideBox":"Learn more about [npj Parkinson's Disease](http://www.nature.com/npjparkd/)","snPcode":"41531","submissionUrl":"https://submission.springernature.com/new-submission/41531/3","title":"npj Parkinson's Disease","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9361292/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9361292/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCognitive impairment (CI) is common in Parkinson\u0026rsquo;s disease (PD), affecting up to 30% of patients at diagnosis and increasing over time. This study evaluated and compared existing risk prediction tools for screening mild cognitive impairment (MCI) and assessing risk of subsequent cognitive decline in people with PD by validating and extending them with PD-specific and modifiable risk factors. Data from 201 PwPD and 231 controls from the Luxembourg Parkinson\u0026rsquo;s study were analyzed; 46.77% of PwPD had MCI. The Lifestyle for Brain Health (LIBRA) and preliminary disease Risk Estimator for Decline in Cognition Tool (pPREDICT) risk scores were evaluated and expanded to create a new tool called \u003cem\u003e\u0026ldquo;\u003c/em\u003e\u003cb\u003eModi\u003c/b\u003efiable factors for \u003cb\u003ecog\u003c/b\u003enitive decline in \u003cb\u003ePD\u0026rdquo;\u003c/b\u003e (ModiCogPD), and predictive performance was assessed using ROC\u0026ndash;AUC, logistic regression, and Cox models. PD-specific scores (pPREDICT and ModiCogPD) showed better performance than LIBRA, with higher scores associated with increased risk of cognitive decline over time, while LIBRA was more informative in early disease stages.\u003c/p\u003e","manuscriptTitle":"Estimating Personalized Risk of Mild Cognitive Impairment in Parkinson's Disease through Comprehensive Risk Prediction Tools","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-13 18:03:40","doi":"10.21203/rs.3.rs-9361292/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"243348499249154607652052260748289177666","date":"2026-05-07T08:52:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"139110128678063571702287565086006850651","date":"2026-05-07T06:05:59+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-05T08:45:30+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-13T16:24:08+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-13T03:52:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Parkinson's Disease","date":"2026-04-08T22:18:41+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-parkinsons-disease","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjparkd","sideBox":"Learn more about [npj Parkinson's Disease](http://www.nature.com/npjparkd/)","snPcode":"41531","submissionUrl":"https://submission.springernature.com/new-submission/41531/3","title":"npj Parkinson's Disease","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1575383f-d791-4205-bc17-9fb038ef6ea6","owner":[],"postedDate":"May 13th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"243348499249154607652052260748289177666","date":"2026-05-07T08:52:38+00:00","index":17,"fulltext":""},{"type":"reviewerAgreed","content":"139110128678063571702287565086006850651","date":"2026-05-07T06:05:59+00:00","index":16,"fulltext":""},{"type":"reviewersInvited","content":"7","date":"2026-05-05T08:45:30+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":67926296,"name":"Health sciences/Diseases"},{"id":67926297,"name":"Health sciences/Neurology"},{"id":67926298,"name":"Biological sciences/Neuroscience"},{"id":67926299,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2026-05-13T18:03:41+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-13 18:03:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9361292","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9361292","identity":"rs-9361292","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}