Association Between education and walking speed: counterfactual mediation analysis in favor of a motor reserve hypothesis.

We quantified the extent to which the association between education and fast walking speed (FWS) is explained by 17 mediators (cardiovascular risk factors/diseases, comorbidities, health behaviours, socio-professional characteristics, cognition), and examined whether mediators interact with education, in favor of a reserve hypothesis. Cross-sectional analyses are based on Constances (population-based study of French adults 45-69y). 3m-FWS was measured using photoelectric cells. Education was categorized as lower/higher. After multiple imputation of missing values, we used counterfactual mediation models for multiple mediators allowing for education×mediator interactions, to estimate the total effect (TE), total indirect effect (TIE), and mediated interaction (IMD) of lower education on FWS. Analyses are based on 71,222 participants (52.6% women; mean age=57.2y; 27.2% higher education; mean FWS=180.2cm/s). In joint mediation analyses, the TE of lower education was -8.19cm/s (95% confidence interval [CI]=-8.87;-7.51), with a TIE of -5.76cm/s (95%CI=-6.10;-5.41; proportion mediated=70.3%, 95%CI=65.6;75.0). The IMD was negative (-2.52, 95%CI=-3.31;-1.72); 30.8% of the TE and 43.8% of the TIE were attributable to the IMD. Several mediators explain a large part of the association between lower education and slower FWS. The detrimental effect of mediators was more pronounced in participants with lower than in those with higher education, in agreement with a reserve hypothesis.

Genome-wide determinants of mortality and motor progression in Parkinson's disease.

There are 90 independent genome-wide significant genetic risk variants for Parkinson's disease (PD) but currently only five nominated loci for PD progression. The biology of PD progression is likely to be of central importance in defining mechanisms that can be used to develop new treatments. We studied 6766 PD patients, over 15,340 visits with a mean follow-up of between 4.2 and 15.7 years and carried out genome-wide survival studies for time to a motor progression endpoint, defined by reaching Hoehn and Yahr stage 3 or greater, and death (mortality). There was a robust effect of the APOE ε4 allele on mortality in PD. We also identified a locus within the TBXAS1 gene encoding thromboxane A synthase 1 associated with mortality in PD. We also report 4 independent loci associated with motor progression in or near MORN1, ASNS, PDE5A, and XPO1. Only the non-Gaucher disease causing GBA1 PD risk variant E326K, of the known PD risk variants, was associated with mortality in PD. Further work is needed to understand the links between these genomic variants and the underlying disease biology. However, these may represent new candidates for disease modification in PD.

Healthy lifestyle change and all-cause and cancer mortality in the European Prospective Investigation into Cancer and Nutrition cohort.

BACKGROUND: Healthy lifestyles are inversely associated with the risk of noncommunicable diseases, which are leading causes of death. However, few studies have used longitudinal data to assess the impact of changing lifestyle behaviours on all-cause and cancer mortality. METHODS: Within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, lifestyle profiles of 308,497 cancer-free adults (71% female) aged 35-70 years at recruitment across nine countries were assessed with baseline and follow-up questionnaires administered on average of 7 years apart. A healthy lifestyle index (HLI), assessed at two time points, combined information on smoking status, alcohol intake, body mass index, and physical activity, and ranged from 0 to 16 units. A change score was calculated as the difference between HLI at baseline and follow-up. Associations between HLI change and all-cause and cancer mortality were modelled with Cox regression, and the impact of changing HLI on accelerating mortality rate was estimated by rate advancement periods (RAP, in years). RESULTS: After the follow-up questionnaire, participants were followed for an average of 9.9 years, with 21,696 deaths (8407 cancer deaths) documented. Compared to participants whose HLIs remained stable (within one unit), improving HLI by more than one unit was inversely associated with all-cause and cancer mortality (hazard ratio [HR]: 0.84; 95% confidence interval [CI]: 0.81, 0.88; and HR: 0.87; 95% CI: 0.82, 0.92; respectively), while worsening HLI by more than one unit was associated with an increase in mortality (all-cause mortality HR: 1.26; 95% CI: 1.20, 1.33; cancer mortality HR: 1.19; 95% CI: 1.09, 1.29). Participants who worsened HLI by more than one advanced their risk of death by 1.62 (1.44, 1.96) years, while participants who improved HLI by the same amount delayed their risk of death by 1.19 (0.65, 2.32) years, compared to those with stable HLI. CONCLUSIONS: Making healthier lifestyle changes during adulthood was inversely associated with all-cause and cancer mortality and delayed risk of death. Conversely, making unhealthier lifestyle changes was positively associated with mortality and an accelerated risk of death.

Lifestyle changes in middle age and risk of cancer: evidence from the European Prospective Investigation into Cancer and Nutrition.

In this study, we aimed to provide novel evidence on the impact of changing lifestyle habits on cancer risk. In the EPIC cohort, 295,865 middle-aged participants returned a lifestyle questionnaire at baseline and during follow-up. At both timepoints, we calculated a healthy lifestyle index (HLI) score based on cigarette smoking, alcohol consumption, body mass index and physical activity. HLI ranged from 0 (most unfavourable) to 16 (most favourable). We estimated the association between HLI change and risk of lifestyle-related cancers-including cancer of the breast, lung, colorectum, stomach, liver, cervix, oesophagus, bladder, and others-using Cox regression models. We reported hazard ratios (HR) with 95% confidence intervals (CI). Median time between the two questionnaires was 5.7 years, median age at follow-up questionnaire was 59 years. After the follow-up questionnaire, we observed 14,933 lifestyle-related cancers over a median follow-up of 7.8 years. Each unit increase in the HLI score was associated with 4% lower risk of lifestyle-related cancers (HR 0.96; 95%CI 0.95-0.97). Among participants in the top HLI third at baseline (HLI > 11), those in the bottom third at follow-up (HLI ≤ 9) had 21% higher risk of lifestyle-related cancers (HR 1.21; 95%CI 1.07-1.37) than those remaining in the top third. Among participants in the bottom HLI third at baseline, those in the top third at follow-up had 25% lower risk of lifestyle-related cancers (HR 0.75; 95%CI 0.65-0.86) than those remaining in the bottom third. These results indicate that lifestyle changes in middle age may have a significant impact on cancer risk.

Norms for Usual and Fast Walking Speed in Adults 45-69 Years Old From the French General Population: Constances Study.

OBJECTIVES: Walking speed (WS) represents an objective measure of motor function and health. We aimed to develop usual (UWS) and fast WS (FWS) norms for the general population using a regression-based approach, while considering age, sex, height, and education. DESIGN: Cross-sectional analysis of a population-based study. SETTING AND PARTICIPANTS: French Constances study (45-69 years). METHODS: UWS/FWS were measured over 3 m (dynamic start) using photoelectric cells. We addressed selection effects (related to survey sampling and nonresponse) and missing data using a combination of inverse probability weighting (IPW) and multiple imputation (MI). Norms by sex, age, height, and education (

Associations between dietary inflammatory scores and biomarkers of inflammation in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort.

BACKGROUND: Since the first version of the dietary inflammatory index (DII®) developed in the past decade, several other versions have been developed. However, to date no study has attempted to compare these versions with respect to their associations with biomarkers of inflammation. OBJECTIVE: We aimed to investigate the relationship between four dietary inflammatory scores [DII, two energy-adjusted derivatives (E-DII and E-DIIr), and the Inflammatory Score of the Diet (ISD)], and circulating levels of several inflammatory markers and adipokines. METHODS: This study included 17 637 participants from the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort with at least one marker of inflammation measured in blood. Associations between the four scores and C-reactive protein (CRP), interleukin (IL)6, IL10, IL1RA, tumor necrosis factor-α (TNFα), soluble tumor necrosis factor receptor-1 (sTNFR1), sTNFR2, leptin, soluble leptin receptor (sLeptin R), adiponectin, and High Molecular Weight (HMW) adiponectin were evaluated using multivariable linear regressions adjusted for potential confounders. RESULTS: Positive associations were observed between the four dietary inflammatory scores and levels of CRP, IL6, sTNFR1, sTNFR2 and leptin. However, only the DII and the ISD were positively associated with IL1RA levels and only the DII and the E-DIIr were positively associated with TNFα levels. The proportion of variance of each biomarker explained by the scores was lower than 2%, which was equivalent to the variance explained by smoking status but much lower than that explained by body mass index. CONCLUSIONS: Our results suggest that the four dietary inflammatory scores were associated with some biomarkers of inflammation and could be used to assess the inflammatory potential of diet in European adults but are not sufficient to capture the inflammatory status of an individual. These findings can help to better understand the inflammatory potential of diet, but they need to be replicated in studies with repeated dietary measurements.

Association of Physical Activity and Parkinson Disease in Women: Long-term Follow-up of the E3N Cohort Study.

BACKGROUND AND OBJECTIVES: Previous cohort studies reported that a single measure of physical activity (PA) assessed at baseline was associated with lower Parkinson disease (PD) incidence, but a meta-analysis suggested that this association was restricted to men. Because of the long prodromal phase of the disease, reverse causation could not be excluded as a potential explanation. Our objective was to study the association between time-varying PA and PD in women using lagged analyses to address the potential for reverse causation and to compare PA trajectories in patients before diagnosis and matched controls. METHODS: We used data from the Etude Epidémiologique auprès de femmes de la Mutuelle Générale de l'Education Nationale (1990-2018), a cohort study of women affiliated with a national health insurance plan for persons working in education. PA was self-reported in 6 questionnaires over the follow-up. As questions changed across questionnaires, we created a time-varying latent PA (LPA) variable using latent process mixed models. PD was ascertained using a multistep validation process based on medical records or a validated algorithm based on drug claims. We set up a nested case-control study to examine differences in LPA trajectories using multivariable linear mixed models with a retrospective timescale. Cox proportional hazards models with age as the timescale and adjusted for confounders were used to estimate the association between time-varying LPA and PD incidence. Our main analysis used a 10-year lag to account for reverse causation; sensitivity analyses used 5-, 15-, and 20-year lags. RESULTS: Analyses of trajectories (1,196 cases and 23,879 controls) showed that LPA was significantly lower in cases than in controls throughout the follow-up, including 29 years before diagnosis; the difference between cases and controls started to increase ∼10 years before diagnosis (p interaction = 0.003). In our main survival analysis, of 95,354 women free of PD in 2000, 1,074 women developed PD over a mean follow-up of 17.2 years. PD incidence decreased with increasing LPA (p trend = 0.001), with 25% lower incidence in those in the highest quartile compared with the lowest (adjusted hazard ratio 0.75, 95% CI 0.63-0.89). Using longer lags yielded similar conclusions. DISCUSSION: Higher PA level is associated with lower PD incidence in women, not explained by reverse causation. These results are important for planning interventions for PD prevention.

Characteristics of reproductive history, use of exogenous hormones and walking speed among women: Data from the CONSTANCES French Cohort Study.

OBJECTIVE: To investigate the cross-sectional associations of reproductive history and use of exogenous hormones with fast walking speed (WS) in women. STUDY DESIGN: Between 2012 and 2020, 33,892 French women aged 45 years or more, recruited at health centers, underwent physical function tests and self-reported information on reproductive history and use of exogenous hormones. Linear mixed models with the center as random intercept were used to estimate the association of exposures with WS. MAIN OUTCOME MEASURES: Fast WS. RESULTS: Mean WS was 172.2 cm/s. WS increased with age at menarche (ß+1y = 0.23, 95 % confidence interval = 0.05 to 0.40), age at first birth (ß+1y = 0.20, 95 % CI = 0.13 to 0.27) and duration of breastfeeding (ßfor ≥10 vs ≤5months = 1.38; 95 % CI = 0.39 to 2.36). In addition, parity was quadratically associated with WS, with women with 3 children having the highest WS (p for U-shaped relationship < 0.01). Menopausal status had no impact on WS but age at menopause was positively associated with WS (ß+5y = 0.52, 95 % CI = 0.17 to 0.87) and partly explained the deleterious impact of artificial menopause on WS. WS increased with reproductive lifetime duration (ß+5y = 0.49, 95 % CI = 0.16 to 0.83) and decreased with time since onset of menopause (ß+5y = -0.65, 95 % CI = -0.99 to -0.31). By contrast, there was no association of WS with oral contraception and postmenopausal hormone therapy. CONCLUSION: Our findings suggest that reproductive life characteristics may be associated with WS and timing of exposure could play a role.

Statin Use and Incidence of Parkinson's Disease in Women from the French E3N Cohort Study.

BACKGROUND: Statins represent candidates for drug repurposing in Parkinson's disease (PD). Few studies examined the role of reverse causation, statin subgroups, and dose-response relations based on time-varying exposures. OBJECTIVES: We examined whether statin use is associated with PD incidence while attempting to overcome the limitations described previously, especially reverse causation. METHOD: We used data from the E3N cohort study of French women (follow-up, 2004-2018). Incident PD was ascertained using multiple sources and validated by experts. New statin users were identified through linked drug claims. We set up a nested case-control study to describe trajectories of statin prescriptions and medical consultations before diagnosis. We used time-varying multivariable Cox proportional hazards regression models to examine the statins-PD association. Exposure indexes included ever use, cumulative duration/dose, and mean daily dose and were lagged by 5 years to address reverse causation. RESULTS: The case-control study (693 cases, 13,784 controls) showed differences in case-control trajectories, with changes in the 5 years before diagnosis in cases. Of 73,925 women (aged 54-79 years), 524 developed PD and 11,552 started using statins in lagged analyses. Ever use of any statin was not associated with PD (hazard ratio [HR] = 0.87, 95% confidence interval [CI] = 0.67-1.11). Alternatively, ever use of lipophilic statins was significantly associated with lower PD incidence (HR = 0.70, 95% CI = 0.51-0.98), with a dose-response relation for the mean daily dose (P-linear trend = 0.02). There was no association for hydrophilic statins. CONCLUSION: Use of lipophilic statins at least 5 years earlier was associated with reduced PD incidence in women, with a dose-response relation for the mean daily dose. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Association between beverage consumption and risk of rheumatoid arthritis: a prospective study from the French E3N Cohort.

OBJECTIVES: To assess the relationship between consumption of largely consumed beverages (coffee, tea, alcohol and soft drinks) and the risk of RA. MATERIAL AND METHODS: The E3N Study (Étude Épidémiologique auprès des femmes de la Mutuelle Générale de l'Éducation Nationale) is a French prospective cohort including 98 995 women since 1990. Food and beverage consumption was assessed using a validated food-frequency questionnaire. Hazard ratios (HR) and their 95% CI for incident RA were estimated by Cox proportional hazards model. RESULTS: Among 62 631 women, 481 incident RA cases were identified. Consumptions of tea, alcohol and sugar-sweetened soft drinks were not associated with RA risk. We observed a linear association between coffee consumption and RA risk [≥4 cups/day vs ≤1cup/day, HR = 1.24; 95% CI (0.94, 1.64), Ptrend = 0.04], and a higher risk of RA with artificially sweetened soft drinks consumption [consumers vs not, HR = 1.66; 95% CI (1.12, 2.45)], particularly in never-smokers. Among ever-smokers, moderate liquor intake was associated with a reduced risk of RA [1-3 glasses/week vs non-consumers, HR = 0.63; 95% CI (0.43, 0.91)] and moderate wine consumption with a reduced risk of seropositive RA. CONCLUSIONS: In a large cohort of women, tea, alcohol and sugar-sweetened soft drinks consumption was not associated with RA risk, whereas consumption of coffee (especially caffeinated coffee), and artificially sweetened soft drinks was associated with higher RA risk, particularly among never-smokers. If further confirmed, these results could lead to novel mechanistic hypotheses and to simple prevention measures.

Reproductive characteristics, use of exogenous hormones and Parkinson disease in women from the E3N study.

Despite experimental studies suggesting a disease-modifying role of oestrogens, results from epidemiological studies on the relation of reproductive characteristics and hormonal exposures with Parkinson disease in women are conflicting. We used the data from the E3N cohort study including 98 068 women aged 40-65 years in 1990 followed until 2018. Parkinson disease was ascertained using a validation process based on drug claim databases and medical records. Reproductive characteristics and hormonal exposures were self-reported (11 questionnaires). Associations of exposures with Parkinson disease incidence were investigated using time-varying Cox proportional hazards regression with a 5-year exposure lag and age as the timescale adjusted for confounders. We identified 1165 incident Parkinson disease cases during a mean follow-up of 22.0 years (incidence rate = 54.7 per 100 000 person-years). Parkinson disease incidence was higher in women with early (<12 years, HR = 1.21, 95% CI = 1.04-1.40) or late age at menarche (≥14 years, HR = 1.18, 95% CI = 1.03-1.35) than in women with menarche at 12-13 years. Nulliparity was not associated with Parkinson disease, but Parkinson disease incidence increased with the number of children in parous women (P-trend = 0.009). Women with artificial (surgical, iatrogenic) menopause were at greater risk than women with natural menopause (HR = 1.28, 95% CI = 1.09-1.47), especially when artificial menopause occurred at an early age (≤45.0 years). Postmenopausal hormone therapy tended to mitigate greater risk associated with artificial or early menopause (≤45.0 years). While fertility treatments were not associated with Parkinson disease overall, ever users of clomiphene were at greater Parkinson disease risk than never users (HR = 1.81, 95% CI = 1.14-2.88). Other exposures (breastfeeding, oral contraceptives) were not associated with Parkinson disease. Our findings suggest that early and late age at menarche, higher parity, and artificial menopause, in particular at an early age, are associated with increased Parkinson disease incidence in women. In addition, there was some evidence that use of exogenous hormones may increase (fertility treatments) or decrease (postmenopausal hormone therapy) Parkinson disease incidence. These findings support the hypothesis that hormonal exposures play a role in the susceptibility to neurodegenerative diseases. If confirmed, they could help to identify subgroups at high risk for Parkinson disease.

Association between the LRP1B and APOE loci and the development of Parkinson's disease dementia.

Parkinson's disease is one of the most common age-related neurodegenerative disorders. Although predominantly a motor disorder, cognitive impairment and dementia are important features of Parkinson's disease, particularly in the later stages of the disease. However, the rate of cognitive decline varies among Parkinson's disease patients, and the genetic basis for this heterogeneity is incompletely understood. To explore the genetic factors associated with rate of progression to Parkinson's disease dementia, we performed a genome-wide survival meta-analysis of 3923 clinically diagnosed Parkinson's disease cases of European ancestry from four longitudinal cohorts. In total, 6.7% of individuals with Parkinson's disease developed dementia during study follow-up, on average 4.4 ± 2.4 years from disease diagnosis. We have identified the APOE ε4 allele as a major risk factor for the conversion to Parkinson's disease dementia [hazard ratio = 2.41 (1.94-3.00), P = 2.32 × 10-15], as well as a new locus within the ApoE and APP receptor LRP1B gene [hazard ratio = 3.23 (2.17-4.81), P = 7.07 × 10-09]. In a candidate gene analysis, GBA variants were also identified to be associated with higher risk of progression to dementia [hazard ratio = 2.02 (1.21-3.32), P = 0.007]. CSF biomarker analysis also implicated the amyloid pathway in Parkinson's disease dementia, with significantly reduced levels of amyloid ß42 (P = 0.0012) in Parkinson's disease dementia compared to Parkinson's disease without dementia. These results identify a new candidate gene associated with faster conversion to dementia in Parkinson's disease and suggest that amyloid-targeting therapy may have a role in preventing Parkinson's disease dementia.

Body Mass Index, Abdominal Adiposity, and Incidence of Parkinson Disease in French Women From the E3N Cohort Study.

BACKGROUND AND OBJECTIVES: Previous studies on the relationship between body mass index (BMI) and Parkinson disease (PD) provided inconsistent results, likely due to reverse causation explained by weight loss during the prodromal phase. We examined the association of BMI and abdominal adiposity with PD incidence using lagged analyses to address the potential for reverse causation and compared BMI trajectories in patients before diagnosis and matched controls. METHODS: We used data from the E3N cohort study of French women with a 29-year follow-up (1990-2018). BMI (kg/m2) was computed based on self-reported weight and height up to 11 times; up to 6 waist circumference (WC) and hip circumference measures were available. PD diagnoses were validated based on medical records and drug claim databases. Multivariable time-varying Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% CIs according to BMI categories (underweight <18.5 kg/m2; normal = [18.5-25.0[ kg/m2; overweight = [25.0-30.0[ kg/m2; obese ≥30.0 kg/m2). Exposures were lagged by 5 years in main analyses; we used longer lags (10 and 20 years) in sensitivity analyses. We examined trajectories of BMI categories within a nested case-control study using multivariable generalized estimating equations multinomial logistic models. RESULTS: Of 96,702 women (baseline age = 40-65 years), 1,164 developed PD. PD incidence was lower (HR = 0.76, 95% CI = 0.59-0.98, p = 0.032) among women with obesity compared with those with normal BMI. There was a similar association in analyses using longer lag times (20 years, 598 cases, HR = 0.52, 95% CI = 0.30-0.88, p = 0.016). A similar pattern was seen for WC and waist-height ratio but not waist-hip ratio. Trajectories of BMI categories (1,196 patients and 23,876 controls) showed that obesity was less frequent in patients with PD before diagnosis than in controls, with a statistically significant difference 29 years before. In addition, the frequency of obesity decreased 5-10 years before diagnosis in patients. DISCUSSION: In this large cohort of women with a long follow-up, obesity was associated with a lower hazard of PD, even when measured 20 years before diagnosis, in agreement with Mendelian randomization studies. Our analyses underscore the importance of lagged analyses to account for reverse causation. These findings warrant further investigations to understand the mechanisms underlying this inverse association.

Comparison of manual and automated measures of walking speed: Distance and pace matter.

BACKGROUND: Walking speed (WS) represents a global marker of individual health and provides a simple and objective measure of motor performances for use in clinical and research settings. WS is most often measured over relatively short distances at usual (UWS) or fast (FWS) pace, using manual (e.g., stopwatch) or automated methods (e.g., photoelectric cells). As the time needed to walk over these distances is very short, we hypothesized that measurement error related to manual compared to automated WS measures is more pronounced for shorter distances and FWS and investigated the reliability and agreement of WS in a subsample of the Constances cohort at two paces and over two distances. METHODS: We recruited 100 community-dwelling participants (50 % women) aged 45-70y (mean = 56.1y). WS was measured manually (stopwatches) and using photoelectric cells, at two paces (UWS/FWS) and over two distances (3 m/5 m). Agreement was examined using Bland and Altman plots and intraclass correlation coefficients (ICC). RESULTS: Participants were on average 169.8 cm tall, and their mean body mass index was 25.4 kg/m2. Agreement between manual stopwatches and photoelectric cells was excellent (ICCs between 0.92 and 0.97), but it was lower for smaller distances, with significantly lower ICCs over 3 m compared to 5 m both for UWS (differenceICC = -0.04) and FWS (differenceICC = -0.05). Bias of manual measures was constant for UWS and increased with increasing FWS. There were inter-rater effects, with better agreement for UWS and 5 m compared to FWS and 3 m. CONCLUSIONS: Both distance and pace have an influence on the reliability of WS measures using manual timing methods. Our findings also suggest the presence of rater effects and better agreement for 5 m and UWS. These findings are helpful for the design of studies that include manual measures of WS, especially FWS, in order to reduce measurement error and suggest that longer distances are preferable.

Machine Learning-Based Prediction of Impulse Control Disorders in Parkinson's Disease From Clinical and Genetic Data.

Goal: Impulse control disorders (ICDs) are frequent non-motor symptoms occurring during the course of Parkinson's disease (PD). The objective of this study was to estimate the predictability of the future occurrence of these disorders using longitudinal data, the first study using cross-validation and replication in an independent cohort. Methods: We used data from two longitudinal PD cohorts (training set: PPMI, Parkinson's Progression Markers Initiative; test set: DIGPD, Drug Interaction With Genes in Parkinson's Disease). We included 380 PD subjects from PPMI and 388 PD subjects from DIGPD, with at least two visits and with clinical and genetic data available, in our analyses. We trained three logistic regressions and a recurrent neural network to predict ICDs at the next visit using clinical risk factors and genetic variants previously associated with ICDs. We quantified performance using the area under the receiver operating characteristic curve (ROC AUC) and average precision. We compared these models to a trivial model predicting ICDs at the next visit with the status at the most recent visit. Results: The recurrent neural network (PPMI: 0.85 [0.80 - 0.90], DIGPD: 0.802 [0.78 - 0.83]) was the only model to be significantly better than the trivial model (PPMI: ROC AUC = 0.75 [0.69 - 0.81]; DIGPD: 0.78 [0.75 - 0.80]) on both cohorts. We showed that ICDs in PD can be predicted with better accuracy with a recurrent neural network model than a trivial model. The improvement in terms of ROC AUC was higher on PPMI than on DIGPD data, but not clinically relevant in both cohorts. Conclusions: Our results indicate that machine learning methods are potentially useful for predicting ICDs, but further works are required to reach clinical relevance.

Incidence of Parkinson's disease in French women from the E3N cohort study over 27 years of follow-up.

Parkinson's disease (PD) is an uncommon disease with a long prodromal period and higher incidence in men than women. Large cohort studies of women with a long follow-up are needed. Within the E3N French cohort study (98,995 women, 40-65 years at baseline), we identified 3,584 participants who self-reported PD or used anti-parkinsonian drugs over 27 years (1992-2018). We obtained medical records to validate PD diagnosis (definite, probable, possible, no). When medical records were not available, we used a validated algorithm based on drug claims to predict PD status. We retained a PD diagnosis for 1,294 women (medical records, 62%; algorithm, 38%). After exclusion of prevalent/possible cases, cases without age at diagnosis, and women lost to follow-up, our analyses included 98,069 women, of whom 1,200 had incident PD (mean age at diagnosis = 71.8 years; incidence rate = 0.494/1,000 person-years). Age-adjusted incidence rates increased over the six first years of follow-up, possibly due to healthy volunteer bias, and remained stable thereafter, similar to incidence rates in women from Western Europe. Forty three percent of PD cases occurred after 20 years of follow-up (2012-2018). The cumulative incidence of PD from 50 to 90 years was 2.41% (95% confidence interval [CI] = 2.27-2.65). PD incidence was lower in ever than never smokers (hazard ratio = 0.86, 95% CI = 0.76-0.96). In conclusion, we estimated PD incidence rates in French women over a 27-year follow-up, and showed stable incidence between 2002 and 2018. The long follow-up and large sample size make this study a valuable resource to improve our knowledge on PD etiology in women.

Association Between Occupational Exposure to Formaldehyde and Cognitive Impairment.

BACKGROUND AND OBJECTIVES: To our knowledge, no study has investigated the effect of exposure to formaldehyde on cognition in the general population. Our objective was to examine the association between occupational exposure to formaldehyde and cognitive impairment in middle-aged and young-old adults (≥45 years). METHODS: In the French CONSTANCES cohort, cognitive function was assessed with a standardized battery of 7 cognitive tests to evaluate global cognitive function, episodic verbal memory, language abilities, and executive functions (e.g., Digit Symbol Substitution Test [DSST]). A global cognitive score was created using principal component analysis. Cognitive impairment was assessed in reference to norms of neuropsychological battery according to age, sex, and education. Lifetime exposure to formaldehyde was assessed using a French Job Exposure Matrix created in the framework of the Matgéné project. After performing multiple imputation, separate modified Poisson regression models were used to evaluate the association between cognitive impairment (<25th percentile) and formaldehyde exposure (exposed/never exposed), exposure duration, cumulative exposure index (CEI), and combination of CEI and time of last exposure. RESULTS: Among 75,322 participants (median age 57.5 years, 53% women), 8% were exposed to formaldehyde during their professional life. These participants were at higher risk of global cognitive impairment (for global cognitive score: adjusted relative risk [aRR] 1.17; 95% confidence interval [CI] 1.11-1.23), after adjusting for confounders (age, sex, education, income, solvent exposure, Effort-Reward Imbalance, night shift, repetitive work, and noisy work). They were at higher risk of cognitive impairment for all cognitive domains explored. Longer exposure duration and high CEI were associated with cognitive impairment, with a dose-effect relationship for exposure duration. Recent exposure was associated with impairment in all cognitive domains. Time did not fully attenuate formaldehyde-associated cognitive deficits especially in highly exposed individuals (for DSST: high past exposure aRR 1.23; 95% CI 1.11-1.36; high recent exposure: aRR 1.24; 95% CI 1.13-1.35). DISCUSSION: Our findings highlight the long-term detrimental effect of formaldehyde exposure on cognitive health in a relatively young population.