Fast and flexible inference for joint models of multivariate longitudinal and survival data using integrated nested Laplace approximations

Modeling longitudinal and survival data jointly offers many advantages such as addressing measurement error and missing data in the longitudinal processes, understanding and quantifying the association between the longitudinal markers and the survival events, and predicting the risk of events based on the longitudinal markers. A joint model involves multiple submodels (one for each longitudinal/survival outcome) usually linked together through correlated or shared random effects. Their estimation is computationally expensive (particularly due to a multidimensional integration of the likelihood over the random effects distribution) so that inference methods become rapidly intractable, and restricts applications of joint models to a small number of longitudinal markers and/or random effects. We introduce a Bayesian approximation based on the integrated nested Laplace approximation algorithm implemented in the R package R-INLA to alleviate the computational burden and allow the estimation of multivariate joint models with fewer restrictions. Our simulation studies show that R-INLA substantially reduces the computation time and the variability of the parameter estimates compared with alternative estimation strategies. We further apply the methodology to analyze five longitudinal markers (3 continuous, 1 count, 1 binary, and 16 random effects) and competing risks of death and transplantation in a clinical trial on primary biliary cholangitis. R-INLA provides a fast and reliable inference technique for applying joint models to the complex multivariate data encountered in health research.

Recent Advances in Clinical Trials in Multiple System Atrophy

PURPOSE OF REVIEW

This review summarizes previous and ongoing neuroprotection trials in multiple system atrophy (MSA), a rare and fatal neurodegenerative disease characterized by parkinsonism, cerebellar, and autonomic dysfunction. It also describes the preclinical therapeutic pipeline and provides some considerations relevant to successfully conducting clinical trials in MSA, i.e., diagnosis, endpoints, and trial design.

RECENT FINDINGS

Over 30 compounds have been tested in clinical trials in MSA. While this illustrates a strong treatment pipeline, only two have reached their primary endpoint. Ongoing clinical trials primarily focus on targeting α-synuclein, the neuropathological hallmark of MSA being α-synuclein-bearing glial cytoplasmic inclusions. The mostly negative trial outcomes highlight the importance of better understanding underlying disease mechanisms and improving preclinical models. Together with efforts to refine clinical measurement tools, innovative statistical methods, and developments in biomarker research, this will enhance the design of future neuroprotection trials in MSA and the likelihood of positive outcomes.

Haemoglobin trajectories in chronic kidney disease and risk of major adverse cardiovascular events

BACKGROUND

The trajectories of haemoglobin in patients with chronic kidney disease (CKD) have been poorly described. In such patients, we aimed to identify typical haemoglobin trajectory profiles and estimate their risks of major adverse cardiovascular events (MACE).

METHODS

We used 5-year longitudinal data from the CKD-REIN cohort patients with moderate to severe CKD enrolled from 40 nationally representative nephrology clinics in France. A joint latent class model was used to estimate, in different classes of haemoglobin trajectory, the competing risks of (i) MACE + defined as the first event among cardiovascular death, non-fatal myocardial infarction, stroke or hospitalization for acute heart failure, (ii) initiation of kidney replacement therapy (KRT) and (iii) non-cardiovascular death.

RESULTS

During the follow-up, we gathered 33 874 haemoglobin measurements from 3011 subjects (median, 10 per patient). We identified five distinct haemoglobin trajectory profiles. The predominant profile (n = 1885, 62.6%) showed an overall stable trajectory and low risks of events. The four other profiles had nonlinear declining trajectories: early strong decline (n = 257, 8.5%), late strong decline (n = 75, 2.5%), early moderate decline (n = 356, 11.8%) and late moderate decline (n = 438, 14.6%). The four profiles had different risks of MACE, while the risks of KRT and non-cardiovascular death consistently increased from the haemoglobin decline.

CONCLUSION

In this study, we observed that two-thirds of patients had a stable haemoglobin trajectory and low risks of adverse events. The other third had a nonlinear trajectory declining at different rates, with increased risks of events. Better attention should be paid to dynamic changes of haemoglobin in CKD.

Statistical classification of treatment responses in mouse clinical trials for stratified medicine in oncology drug discovery

Translational oncology research strives to explore a new aspect: identifying subgroups that exhibit treatment response even during pre-clinical phases. In this study, we focus on PDX models and their implementation in mouse clinical trials (MCT). Our primary objective was to identify subgroups with different treatment responses using Latent Class Mixed Model (LCMM).We used a public dataset and focused on one treatment, encorafenib, and two indications, melanoma and colorectal cancer, for which efficacy depends on a specific mutation BRAF V600E. One LCMM per indication was implemented to classify treatment responses at the PDX level, analyzing the growth kinetics of treated tumors and matched controls within the PDX models. A simulation study was carried out to explore the performance of LCMM in this context. For both applications, LCMM identified classes for which the higher the proportion of mutated BRAF V600E PDX models the greater the treatment effect, which is aligned with encorafenib use recommendations. The simulation study showed that LCMM could identify classes with large differences in treatment effects. LCMM is a suitable tool for MCT to explore treatment response subgroups of PDX. Once these subgroups are defined, characterization of their phenotypes/genotypes could be performed to explore treatment response predictors.

Addressing unmeasured confounders in cohort studies: Instrumental variable method for a time-fixed exposure on an outcome trajectory

Instrumental variable methods, which handle unmeasured confounding by targeting the part of the exposure explained by an exogenous variable not subject to confounding, have gained much interest in observational studies. We consider the very frequent setting of estimating the unconfounded effect of an exposure measured at baseline on the subsequent trajectory of an outcome repeatedly measured over time. We didactically explain how to apply the instrumental variable method in such setting by adapting the two-stage classical methodology with (1) the prediction of the exposure according to the instrumental variable, (2) its inclusion into a mixed model to quantify the exposure association with the subsequent outcome trajectory, and (3) the computation of the estimated total variance. A simulation study illustrates the consequences of unmeasured confounding in classical analyses and the usefulness of the instrumental variable approach. The methodology is then applied to 6224 participants of the 3C cohort to estimate the association of type-2 diabetes with subsequent cognitive trajectory, using 42 genetic polymorphisms as instrumental variables. This contribution shows how to handle endogeneity when interested in repeated outcomes, along with a R implementation. However, it should still be used with caution as it relies on instrumental variable assumptions hardly testable in practice.

Random survival forests with multivariate longitudinal endogenous covariates

Predicting the individual risk of clinical events using the complete patient history is a major challenge in personalized medicine. Analytical methods have to account for a possibly large number of time-dependent predictors, which are often characterized by irregular and error-prone measurements, and are truncated early by the event. In this work, we extended the competing-risk random survival forests to handle such endogenous longitudinal predictors when predicting event probabilities. The method, implemented in the R package DynForest, internally transforms the time-dependent predictors at each node of each tree into time-fixed features (using mixed models) that can then be used as splitting candidates. The final individual event probability is computed as the average of leaf-specific Aalen-Johansen estimators over the trees. Using simulations, we compared the performances of DynForest to accurately predict an event with (i) a joint modeling alternative when considering two longitudinal predictors only, and with (ii) a regression calibration method that ignores the informative truncation by the event when dealing with a large number of longitudinal predictors. Through an application in dementia research, we also illustrated how DynForest can be used to develop a dynamic prediction tool for dementia from multimodal repeated markers, and quantify the importance of each marker.

Describing complex disease progression using joint latent class models for multivariate longitudinal markers and clinical endpoints

Neurodegenerative diseases are characterized by numerous markers of progression and clinical endpoints. For instance, multiple system atrophy (MSA), a rare neurodegenerative synucleinopathy, is characterized by various combinations of progressive autonomic failure and motor dysfunction, and a very poor prognosis. Describing the progression of such complex and multi-dimensional diseases is particularly difficult. One has to simultaneously account for the assessment of multivariate markers over time, the occurrence of clinical endpoints, and a highly suspected heterogeneity between patients. Yet, such description is crucial for understanding the natural history of the disease, staging patients diagnosed with the disease, unravelling subphenotypes, and predicting the prognosis. Through the example of MSA progression, we show how a latent class approach modeling multiple repeated markers and clinical endpoints can help describe complex disease progression and identify subphenotypes for exploring new pathological hypotheses. The proposed joint latent class model includes class-specific multivariate mixed models to handle multivariate repeated biomarkers possibly summarized into latent dimensions and class-and-cause-specific proportional hazard models to handle time-to-event data. Maximum likelihood estimation procedure, validated through simulations is available in the lcmm R package. In the French MSA cohort comprising data of 598 patients during up to 13 years, five subphenotypes of MSA were identified that differ by the sequence and shape of biomarkers degradation, and the associated risk of death. In posterior analyses, the five subphenotypes were used to explore the association between clinical progression and external imaging and fluid biomarkers, while properly accounting for the uncertainty in the subphenotypes membership.

Disease progression model anchored around clinical diagnosis in longitudinal cohorts: example of Alzheimer's disease and related dementia

BACKGROUND

Alzheimer's disease and related dementia (ADRD) are characterized by multiple and progressive anatomo-clinical changes including accumulation of abnormal proteins in the brain, brain atrophy and severe cognitive impairment. Understanding the sequence and timing of these changes is of primary importance to gain insight into the disease natural history and ultimately allow earlier diagnosis. Yet, modeling changes over disease course from cohort data is challenging as the usual timescales (time since inclusion, chronological age) are inappropriate and time-to-clinical diagnosis is available on small subsamples of participants with short follow-up durations prior to diagnosis. One solution to circumvent this challenge is to define the disease time as a latent variable.

METHODS

We developed a multivariate mixed model approach that realigns individual trajectories into the latent disease time to describe disease progression. In contrast with the existing literature, our methodology exploits the clinical diagnosis information as a partially observed and approximate reference to guide the estimation of the latent disease time. The model estimation was carried out in the Bayesian Framework using Stan. We applied the methodology to the MEMENTO study, a French multicentric clinic-based cohort of 2186 participants with 5-year intensive follow-up. Repeated measures of 12 ADRD markers stemmed from cerebrospinal fluid (CSF), brain imaging and cognitive tests were analyzed.

RESULTS

The estimated latent disease time spanned over twenty years before the clinical diagnosis. Considering the profile of a woman aged 70 with a high level of education and APOE4 carrier (the main genetic risk factor for ADRD), CSF markers of tau proteins accumulation preceded markers of brain atrophy by 5 years and cognitive decline by 10 years. However we observed that individual characteristics could substantially modify the sequence and timing of these changes, in particular for CSF level of A[Formula: see text].

CONCLUSION

By leveraging the available clinical diagnosis timing information, our disease progression model does not only realign trajectories into the most homogeneous way. It accounts for the inherent residual inter-individual variability in dementia progression to describe the long-term anatomo-clinical degradations according to the years preceding clinical diagnosis, and to provide clinically meaningful information on the sequence of events.

TRIAL REGISTRATION

clinicaltrials.gov, NCT01926249. Registered on 16 August 2013.

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.

Analysis of the 24-h activity cycle: An illustration examining the association with cognitive function in the Adult Changes in Thought study

The 24-h activity cycle (24HAC) is a new paradigm for studying activity behaviors in relation to health outcomes. This approach inherently captures the interrelatedness of the daily time spent in physical activity (PA), sedentary behavior (SB), and sleep. We describe three popular approaches for modeling outcome associations with the 24HAC exposure. We apply these approaches to assess an association with a cognitive outcome in a cohort of older adults, discuss statistical challenges, and provide guidance on interpretation and selecting an appropriate approach. We compare the use of the isotemporal substitution model (ISM), compositional data analysis (CoDA), and latent profile analysis (LPA) to analyze 24HAC. We illustrate each method by exploring cross-sectional associations with cognition in 1,034 older adults (Mean age = 77; Age range = 65-100; 55.8% female; 90% White) who were part of the Adult Changes in Thought (ACT) Activity Monitoring (ACT-AM) sub-study. PA and SB were assessed with thigh-worn activPAL accelerometers for 7-days. For each method, we fit a multivariable regression model to examine the cross-sectional association between the 24HAC and Cognitive Abilities Screening Instrument item response theory (CASI-IRT) score, adjusting for baseline characteristics. We highlight differences in assumptions and the scientific questions addressable by each approach. ISM is easiest to apply and interpret; however, the typical ISM assumes a linear association. CoDA uses an isometric log-ratio transformation to directly model the compositional exposure but can be more challenging to apply and interpret. LPA can serve as an exploratory analysis tool to classify individuals into groups with similar time-use patterns. Inference on associations of latent profiles with health outcomes need to account for the uncertainty of the LPA classifications, which is often ignored. Analyses using the three methods did not suggest that less time spent on SB and more in PA was associated with better cognitive function. The three standard analytical approaches for 24HAC each have advantages and limitations, and selection of the most appropriate method should be guided by the scientific questions of interest and applicability of each model's assumptions. Further research is needed into the health implications of the distinct 24HAC patterns identified in this cohort.

CA-125 Early Dynamics to Predict Overall Survival in Women with Newly Diagnosed Advanced Ovarian Cancer Based on Meta-Analysis Data

(1) Background: Cancer antigen 125 (CA-125) is a protein produced by ovarian cancer cells that is used for patients' monitoring. However, the best ways to analyze its decline and prognostic role are poorly quantified. (2) Methods: We leveraged individual patient data from the Gynecologic Cancer Intergroup (GCIG) meta-analysis (N = 5573) to compare different approaches summarizing the early trajectory of CA-125 before the prediction time (called the landmark time) at 3 or 6 months after treatment initiation in order to predict overall survival. These summaries included observed and estimated measures obtained by a linear mixed model (LMM). Their performances were evaluated by 10-fold cross-validation with the Brier score and the area under the ROC (AUC). (3) Results: The estimated value and the last observed value at 3 months were the best measures used to predict overall survival, with an AUC of 0.75 CI 95% [0.70; 0.80] at 24 and 36 months and 0.74 [0.69; 0.80] and 0.75 [0.69; 0.80] at 48 months, respectively, considering that CA-125 over 6 months did not improve the AUC, with 0.74 [0.68; 0.78] at 24 months and 0.71 [0.65; 0.76] at 36 and 48 months. (4) Conclusions: A 3-month surveillance provided reliable individual information on overall survival until 48 months for patients receiving first-line chemotherapy.

Explaining the association between social and lifestyle factors and cognitive functions: a pathway analysis in the Memento cohort

Background

This work aimed to investigate the potential pathways involved in the association between social and lifestyle factors, biomarkers of Alzheimer’s disease and related dementia (ADRD), and cognition.

Methods

The authors studied 2323 participants from the Memento study, a French nationwide clinical cohort. Social and lifestyle factors were education level, current household incomes, physical activity, leisure activities, and social network from which two continuous latent variables were computed: an early to midlife (EML) and a latelife (LL) indicator. Brain magnetic resonance imaging (MRI), lumbar puncture, and amyloid-positron emission tomography (PET) were used to define three latent variables: neurodegeneration, small vessel disease (SVD), and AD pathology. Cognitive function was defined as the underlying factor of a latent variable with four cognitive tests. Structural equation models were used to evaluate cross-sectional pathways between social and lifestyle factors and cognition.

Results

Participants’ mean age was 70.9 years old, 62% were women, 28% were apolipoprotein-ε4 carriers, and 59% had a Clinical Dementia Rating (CDR) score of 0.5. Higher early to midlife social indicator was only directly associated with better cognitive function (direct β = 0.364 (0.322; 0.405), with no indirect pathway through ADRD biomarkers (total β = 0.392 (0.351; 0.429)). In addition to a direct effect on cognition (direct β = 0.076 (0.033; 0.118)), the association between latelife lifestyle indicator and cognition was also mostly mediated by an indirect effect through lower neurodegeneration (indirect β = 0.066 (0.042; 0.090) and direct β =  − 0.116 (− 0.153; − 0.079)), but not through AD pathology nor SVD.

Conclusions

Early to midlife social factors are directly associated with higher cognitive functions. Latelife lifestyle factors may help preserve cognitive functions through lower neurodegeneration.

Relationship Between Brain Tissue Oxygen and Near-Infrared Spectroscopy in Patients with Nontraumatic Subarachnoid Hemorrhage

BACKGROUND

Continuous monitoring of cerebral oxygenation is one of the diagnostic tools used in patients with brain injury. Direct and invasive measurement of cerebral oxygenation with a partial brain oxygen pressure (PbtO₂) probe is promising but invasive. Noninvasive assessment of regional transcranial oxygen saturation using near-infrared spectroscopy (NIRS) may be feasible. The aim of this study was to evaluate the interchangeability between PbtO₂ and NIRS over time in patients with nontraumatic subarachnoid hemorrhage.

METHODS

This retrospective study was performed in a neurocritical care unit. Study participants underwent hourly PbtO₂ and NIRS measurements over 72 h. Temporal agreement between markers was described by their pointwise correlation. A secondary analysis assessed the structure of covariation between marker trajectories using a bivariate linear mixed model.

RESULTS

Fifty-one patients with subarachnoid hemorrhage were included. A total of 3362 simultaneous NIRS and PbtO₂ measurements were obtained. The correlation at each measurement time ranged from - 0.25 to 0.25. The global correlation over time was - 0.026 (p = 0.130). The bivariate linear mixed model confirmed the lack of significant correlation between the PbtO₂ and NIRS measurements at follow-up. NIRS was unable to detect PbtO₂ values below 20 mm Hg (area under the receiver operating characteristic curve 0.539 [95% confidence interval 0.536-0.542]; p = 0.928), and percentage changes in NIRS were unable to detect a decrease in PbtO₂ ≥ 10% (area under the receiver operating characteristic curve 0.615 [95% confidence interval 0.614-0.616]; p < 0.001).

CONCLUSIONS

PbtO₂ and NIRS measurements were not correlated. There is no evidence that NIRS could be a substitute for PbtO₂ monitoring in patients with nontraumatic subarachnoid hemorrhage.

Individual dynamic prediction of clinical endpoint from large dimensional longitudinal biomarker history: a landmark approach

Background

The individual data collected throughout patient follow-up constitute crucial information for assessing the risk of a clinical event, and eventually for adapting a therapeutic strategy. Joint models and landmark models have been proposed to compute individual dynamic predictions from repeated measures to one or two markers. However, they hardly extend to the case where the patient history includes much more repeated markers. Our objective was thus to propose a solution for the dynamic prediction of a health event that may exploit repeated measures of a possibly large number of markers.

Methods

We combined a landmark approach extended to endogenous markers history with machine learning methods adapted to survival data. Each marker trajectory is modeled using the information collected up to the landmark time, and summary variables that best capture the individual trajectories are derived. These summaries and additional covariates are then included in different prediction methods adapted to survival data, namely regularized regressions and random survival forests, to predict the event from the landmark time. We also show how predictive tools can be combined into a superlearner. The performances are evaluated by cross-validation using estimators of Brier Score and the area under the Receiver Operating Characteristic curve adapted to censored data.

Results

We demonstrate in a simulation study the benefits of machine learning survival methods over standard survival models, especially in the case of numerous and/or nonlinear relationships between the predictors and the event. We then applied the methodology in two prediction contexts: a clinical context with the prediction of death in primary biliary cholangitis, and a public health context with age-specific prediction of death in the general elderly population.

Conclusions

Our methodology, implemented in R, enables the prediction of an event using the entire longitudinal patient history, even when the number of repeated markers is large. Although introduced with mixed models for the repeated markers and methods for a single right censored time-to-event, the technique can be used with any other appropriate modeling technique for the markers and can be easily extended to competing risks setting.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-022-01660-3).

Monitoring of SARS-CoV-2 in wastewater: what normalisation for improved understanding of epidemic trends?

SARS-CoV-2 RNA quantification in wastewater has emerged as a relevant additional means to monitor the COVID-19 pandemic. However, the concentration can be affected by black water dilution factors or movements of the sewer shed population, leading to misinterpretation of measurement results. The aim of this study was to evaluate the performance of different indicators to accurately interpret SARS-CoV-2 in wastewater. Weekly/bi-weekly measurements from three cities in France were analysed from February to September 2021. The concentrations of SARS-CoV-2 gene copies were normalised to the faecal-contributing population using simple sewage component indicators. To reduce the measurement error, a composite index was created to combine simultaneously the information carried by the simple indicators. The results showed that the regularity (mean absolute difference between observation and the smoothed curve) of the simple indicators substantially varied across sampling points. The composite index consistently showed better regularity compared to the other indicators and was associated to the lowest variation in correlation coefficient across sampling points. These findings suggest the recommendation for the use of a composite index in wastewater-based epidemiology to compensate for variability in measurement results.

How Selection Over Time Contributes to the Inconsistency of the Association Between Sex/Gender and Cognitive Decline Across Cognitive Aging Cohorts

The association between sex/gender and aging-related cognitive decline remains poorly understood because of inconsistencies in findings. Such heterogeneity could be attributable to the cognitive functions studied and study population characteristics, but also to differential selection by dropout and death between men and women. We aimed to evaluate the impact of selection by dropout and death on the association between sex/gender and cognitive decline. We first compared the statistical methods most frequently used for longitudinal data, targeting either population estimands (marginal models fitted by generalized estimating equations) or subject-specific estimands (mixed/joint models fitted by likelihood maximization) in 8 studies of aging: 6 population-based studies (the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) Study (1996-2009), Personnes Âgées QUID (PAQUID; 1988-2014), the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study (2003-2016), the Three-City Study (Bordeaux only; 1999-2016), the Washington Heights-Inwood Community Aging Project (WHICAP; 1992-2017), and the Whitehall II Study (2007-2016)) and 2 clinic-based studies (the Alzheimer's Disease Neuroimaging Initiative (ADNI; 2004-2017) and a nationwide French cohort study, MEMENTO (2011-2016)). We illustrate differences in the estimands of the association between sex/gender and cognitive decline in selected examples and highlight the critical role of differential selection by dropout and death. Using the same estimand, we then contrast the sex/gender-cognitive decline associations across cohorts and cognitive measures suggesting a residual differential sex/gender association depending on the targeted cognitive measure (memory or animal fluency) and the initial cohort selection. We recommend focusing on subject-specific estimands in the living population for assessing sex/gender differences while handling differential selection over time.

Modeling repeated self-reported outcome data: a continuous-time longitudinal Item Response Theory model

Item Response Theory (IRT) models have received growing interest in health science for analyzing latent constructs such as depression, anxiety, quality of life or cognitive functioning from the information provided by each individual's items responses. However, in the presence of repeated item measures, IRT methods usually assume that the measurement occasions are made at the exact same time for all patients. In this paper, we show how the IRT methodology can be combined with the mixed model theory to provide a longitudinal IRT model which exploits the information of a measurement scale provided at the item level while simultaneously handling observation times that may vary across individuals and items. The latent construct is a latent process defined in continuous time that is linked to the observed item responses through a measurement model at each individual- and occasion-specific observation time; we focus here on a Graded Response Model for binary and ordinal items. The Maximum Likelihood Estimation procedure of the model is available in the R package lcmm. The proposed approach is contextualized in a clinical example in end-stage renal disease, the PREDIALA study. The objective is to study the trajectories of depressive symptomatology (as measured by 7 items of the Hospital Anxiety and Depression scale) according to the time from registration on the renal transplant waiting list and the renal replacement therapy. We also illustrate how the method can be used to assess Differential Item Functioning and lack of measurement invariance over time.

Determinants of Distinct Trajectories of Fatigue in Patients Undergoing Chemotherapy for a Metastatic Colorectal Cancer: 6-Month Follow-up Using Growth Mixture Modeling

OBJECTIVES

This longitudinal prospective and observational study was designed to identify fatigue trajectories during a 6-month period of chemotherapy in patients with metastatic colorectal cancer, and examine the psychosocial factors predicting these trajectories. Associations between fatigue and survival were also investigated.

METHODS

A total of 169 patients (M_age = 64.36 years, SD = 10.5) reported their fatigue levels every 2 weeks for 6 months. Psychological variables (anxiety, depression, internal control, and coping) were assessed at baseline. A Growth Mixture Model was used to identify latent trajectories of fatigue, and a multinomial logistic regression tested covariate predictors of patients' trajectories.

RESULTS

Four clinically distinct fatigue trajectories were identified: intense fatigue (6.51%), moderate fatigue (48.52%), no fatigue (33%), and increasing fatigue (11.83%). Fatigue severity was directly associated with overall survival. High depression levels were associated with fatigue severity over time for intense (OR = 1.80 [1.32-2.47]) and for moderate (OR = 1.58 [1.25-2.00]) fatigue, compared to patients reporting no fatigue. Patients who did not report fatigue were better adjusted, and had more resources, such as better internal control over the disease and less emotion-focused coping (guilt and avoidance), than those who reported intense (OR_control = 0.77 [0.65-0.92]) or moderate (OR_control = 0.89 [0.79-0.99] and OR_coping = 1.13 [1.02-1.24]) fatigue.

CONCLUSION

Fatigue trajectories differed considerably across patients with metastatic colorectal cancer. This first longitudinal study on colorectal cancer patients involving transactional variables suggests that psychosocial interventions should target these specific outcomes, in order to help patients manage their fatigue.

Co-Occurrence of Geriatric Syndromes and Diseases in the General Population: Assessment of the Dimensions of Aging

OBJECTIVES

The co-occurrence of multiple medical or psycho-social conditions (geriatric syndromes (GS) and age-related diseases) is a growing concern in older people. Given the diversity of these conditions and their complex interactions, our aim was to determine whether they could be structured into synthetic dimensions in order to facilitate the management of multimorbidity.

DESIGN

The underlying structure of 10 GSs and 8 age-related diseases was identified using a multiple correspondence analysis (MCA), and confronted to subjective and objective health outcomes.

SETTING

community residents from Bordeaux City (France) older than 75 years in 2010.

PARTICIPANTS

630 adults aged 75+ years who lived in Bordeaux and participated in the 10-year follow-up of the Three-City study.

MEASUREMENTS

GSs included physical frailty, cognitive impairment and dementia, dependency, depressive symptoms, polymedication, thinness, falls, sensory deficit, social isolation, incontinence. Age-related diseases were cancer, cardiac diseases, peripheral vascular diseases, diabetes, hypertension, pulmonary diseases, osteoporosis, other chronic diseases. Association of the MCA-derived independent dimensions was assessed with 10-year visit subjective health and well-being, and with incident death and entry into institution during the remaining cohort follow-up.

RESULTS

Most of the participants (82%) had at least two age-related syndromes or diseases. The MCA structured the 18 conditions into three major dimensions: Degradation (D) driven by GS, Vascular (V), and Psychosocial (P) representing 68.7%, 7.4%, and 5.7% of the total variance, respectively. Dimension D was a strong predictor of future death and institutionalization. Dimensions D and P were strongly associated with current well-being.

CONCLUSIONS

This work confirmed that multimorbidity is very common among older adults, and demonstrated the essential role of GS as manifestations of aging, even more than age-related diseases.

Impaired hippocampal neurogenesis in vitro is modulated by dietary-related endogenous factors and associated with depression in a longitudinal ageing cohort study

Environmental factors like diet have been linked to depression and/or relapse risk in later life. This could be partially driven by the food metabolome, which communicates with the brain via the circulatory system and interacts with hippocampal neurogenesis (HN), a form of brain plasticity implicated in depression aetiology. Despite the associations between HN, diet and depression, human data further substantiating this hypothesis are largely missing. Here, we used an in vitro model of HN to test the effects of serum samples from a longitudinal ageing cohort of 373 participants, with or without depressive symptomology. 1% participant serum was applied to human fetal hippocampal progenitor cells, and changes in HN markers were related to the occurrence of depressive symptoms across a 12-year period. Key nutritional, metabolomic and lipidomic biomarkers (extracted from participant plasma and serum) were subsequently tested for their ability to modulate HN. In our assay, we found that reduced cell death and increased neuronal differentiation were associated with later life depressive symptomatology. Additionally, we found impairments in neuronal cell morphology in cells treated with serum from participants experiencing recurrent depressive symptoms across the 12-year period. Interestingly, we found that increased neuronal differentiation was modulated by increased serum levels of metabolite butyrylcarnitine and decreased glycerophospholipid, PC35:1(16:0/19:1), levels - both of which are closely linked to diet - all in the context of depressive symptomology. These findings potentially suggest that diet and altered HN could subsequently shape the trajectory of late-life depressive symptomology.

An Item Response Theory analysis of the Unified Multiple System Atrophy Rating Scale

INTRODUCTION

The Unified Multiple System Atrophy Rating Scale (UMSARS) has four subscales that have been specifically designed for the clinical assessment of MSA patients. UMSARS I (activities of daily living) and II (motor examination) subscales are regularly used as primary endpoints in treatment trials. The main objective of this study was to identify UMSARS I and II subscale items that best describe progression over time.

METHODS

All MSA patients seen at the French Reference Centre for MSA from 2007 to 2020 were included in a prospective cohort with an annual follow-up assessment including UMSARS. The repeated measures of the 26 UMSARS I and II items were analyzed using a longitudinal Item Response Theory model to identify the most informative items for each of the five UMSARS IV disease stages. Sample size estimates were further calculated for the most informative items as a group.

RESULTS

A total of 557 MSA patients were included with a mean follow-up of 2.3 years. The majority of items progressed with disease duration or across the different UMSARS IV disability stages, with the exception of those related to dysautonomia. Roughly 70% of the scale information was carried by only 11/26 items, many reflecting the patient perspective. These yielded similar sample size estimates than UMSARS I + II items.

CONCLUSION

This study provides important information about the progression of UMSARS I and II subscale items. Improvements seem particularly necessary regarding those assessing dysautonomia. A shortened scale may be useful as outcome for future clinical trials.

Time-varying associations between an exposure history and a subsequent health outcome: a landmark approach to identify critical windows

Background

Long-term behavioral and health risk factors constitute a primary focus of research on the etiology of chronic diseases. Yet, identifying critical time-windows during which risk factors have the strongest impact on disease risk is challenging. To assess the trajectory of association of an exposure history with an outcome, the weighted cumulative exposure index (WCIE) has been proposed, with weights reflecting the relative importance of exposures at different times. However, WCIE is restricted to a complete observed error-free exposure whereas exposures are often measured with intermittent missingness and error. Moreover, it rarely explores exposure history that is very distant from the outcome as usually sought in life-course epidemiology.

Methods

We extend the WCIE methodology to (i) exposures that are intermittently measured with error, and (ii) contexts where the exposure time-window precedes the outcome time-window using a landmark approach. First, the individual exposure history up to the landmark time is estimated using a mixed model that handles missing data and error in exposure measurement, and the predicted complete error-free exposure history is derived. Then the WCIE methodology is applied to assess the trajectory of association between the predicted exposure history and the health outcome collected after the landmark time. In our context, the health outcome is a longitudinal marker analyzed using a mixed model.

Results

A simulation study first demonstrates the correct inference obtained with this approach. Then, applied to the Nurses’ Health Study (19,415 women) to investigate the association between body mass index history (collected from midlife) and subsequent cognitive decline (evaluated after age 70), the method identified two major critical windows of association: long before the first cognitive evaluation (roughly 24 to 12 years), higher levels of BMI were associated with poorer cognition. In contrast, adjusted for the whole history, higher levels of BMI became associated with better cognition in the last years prior to the first cognitive interview, thus reflecting reverse causation (changes in exposure due to underlying disease).

Conclusions

This approach, easy to implement, provides a flexible tool for studying complex dynamic relationships and identifying critical time windows while accounting for exposure measurement errors.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-021-01403-w).

Dynamic reciprocal relationships between cognitive and functional declines along the Alzheimer's disease continuum in the prospective COGICARE study

BACKGROUND

Thoroughly understanding the temporal associations between cognitive and functional dimensions along the dementia process is fundamental to define preventive measures likely to delay the disease's onset. This work aimed to finely describe the trajectories of cognitive and functional declines, and assess their dynamic bidirectional relationships among subjects at different stages of the dementia process.

METHODS

We leveraged extensive repeated data of cognition and functional dependency from the French prospective COGICARE study, designed to better characterize the natural history of cognitive and functional declines around dementia diagnosis. Cognition was measured by the Mini-Mental State Examination, the Isaacs Set Test for verbal fluency, the Benton Visual Retention Test for visuo-spatial memory, and Trail Making Test Part B for executive functioning. Functional dependency was measured by basic and instrumental activities of daily living. The study included 102 cognitively normal, 123 mildly cognitively impaired, and 72 dementia cases with a median of 5 repeated visits over up to 57 months. We used a dynamic causal model which addresses the two essential issues in temporal associations assessment: focusing on intra-individual change and accounting for time.

RESULTS

Better cognitive abilities were associated with lower subsequent decline of the functional level among the three clinical stages with an intensification over time but no reciprocity of the association whatever the clinical status.

CONCLUSION

This work confirms that the progressive functional dependency could be induced by cognitive impairment. Subjects identified as early as possible with clinically significant cognitive impairments could benefit from preventive measures before the deterioration of activities of daily living and the appearance of dementia clinical signs.

The serum metabolome mediates the concert of diet, exercise, and neurogenesis, determining the risk for cognitive decline and dementia

INTRODUCTION

Diet and exercise influence the risk of cognitive decline (CD) and dementia through the food metabolome and exercise-triggered endogenous factors, which use the blood as a vehicle to communicate with the brain. These factors might act in concert with hippocampal neurogenesis (HN) to shape CD and dementia.

METHODS

Using an in vitro neurogenesis assay, we examined the effects of serum samples from a longitudinal cohort (n = 418) on proxy HN readouts and their association with future CD and dementia across a 12-year period.

RESULTS

Altered apoptosis and reduced hippocampal progenitor cell integrity were associated with exercise and diet and predicted subsequent CD and dementia. The effects of exercise and diet on CD specifically were mediated by apoptosis.

DISCUSSION

Diet and exercise might influence neurogenesis long before the onset of CD and dementia. Alterations in HN could signify the start of the pathological process and potentially represent biomarkers for CD and dementia.

Heterogeneous Long-Term Trajectories of Dependency in Older Adults: The PAQUID Cohort, a Population-Based Study over 22 years

BACKGROUND

A critical step toward successful aging is to identify opportunities for prevention of functional decline. Our aim was to describe the heterogeneity in trajectories of dependency preceding death in elders and to identify factors associated with this heterogeneity.

METHODS

The study relied on 3,238 participants of the prospective population-based PAQUID cohort aged 65+ at baseline in 1988. Dependency was defined from an 11-item scale of basic and instrumental activities of daily living (ADL: bathing, dressing, toileting, continence, eating, and transferring; instrumental activities of daily living (IADL): telephoning, shopping, using transport, handling medication, and managing finances) collected over 22 years. Heterogeneous trajectories were estimated using a longitudinal item response theory model including latent classes.

RESULTS

Five distinct profiles of functional dependency were identified over the two last decades of life: persistently high (12%), moderate (26%), persistently low (40%), and accelerated high dependency (15%), and no dependency (8%). Main factors associated with heterogeneity included age at death, sex, education, initial cognition (Mini-Mental State Examination [MMSE] score and dementia), initial disability, and poly-medication.

CONCLUSIONS

In the two last decades of life, more than 9 elders in 10 were characterized as functional decliners. On average, around half of the elders died with no or mild dependency, while 27% live several years with a high level of limitations and would need assistance in activities of daily living, at least for 2-4 years preceding death. The identified factors associated with these trajectories are important to understand functional heterogeneity in elders and to propose interventions to postpone or prevent "chronic" disability.

Plasma carotenoids and medial temporal lobe atrophy in older adults

BACKGROUND & AIMS

Carotenoids are vegetable pigments with neuroprotective properties. Clinical studies found efficacy of specific carotenoids on improving brain perfusion and functioning with aging. However, evidence of an effect on neurodegeneration, which may require longer follow-up period to observe, is more limited. Leveraging biomarkers from a large population-based cohort study of older adults, we investigated whether blood carotenoids were associated with atrophy of the medial temporal lobe (a biomarker of neurodegeneration in aging) over 10 years.

METHODS

This study included 461 dementia-free participants from the Three-City Bordeaux study (aged ≥65) who had plasma carotenoids measured at baseline and up to three repeated brain imaging exams in the subsequent 10 years.

RESULTS

In adjusted linear mixed models, each increase of 1 SD in plasma level of total carotenoids and of β-carotene was associated with 0.02 cm³ (95% CI, 0.001-0.04; P = 0.04) and 0.02 cm³ (95% CI, 0.01-0.04; P = 0.008) smaller medial temporal lobe volume loss per year, respectively.

CONCLUSIONS

Our results based on a unique long-term prospective evaluation of a neuroimaging biomarker suggest a beneficial role of carotenoids for the prevention of age-related neurodegeneration.

Early signature in the blood lipidome associated with subsequent cognitive decline in the elderly: A case-control analysis nested within the Three-City cohort study

BACKGROUND

Brain lipid metabolism appears critical for cognitive aging, but whether alterations in the lipidome relate to cognitive decline remains unclear at the system level.

METHODS

We studied participants from the Three-City study, a multicentric cohort of older persons, free of dementia at time of blood sampling, and who provided repeated measures of cognition over 12 subsequent years. We measured 189 serum lipids from 13 lipid classes using shotgun lipidomics in a case-control sample on cognitive decline (matched on age, sex and level of education) nested within the Bordeaux study center (discovery, n = 418). Associations with cognitive decline were investigated using bootstrapped penalized regression, and tested for validation in the Dijon study center (validation, n = 314).

FINDINGS

Among 17 lipids identified in the discovery stage, lower levels of the triglyceride TAG50:5, and of four membrane lipids (sphingomyelin SM40:2,2, phosphatidylethanolamine PE38:5(18:1/20:4), ether-phosphatidylethanolamine PEO34:3(16:1/18:2), and ether-phosphatidylcholine PCO34:1(16:1/18:0)), and higher levels of PCO32:0(16:0/16:0), were associated with greater odds of cognitive decline, and replicated in our validation sample.

INTERPRETATION

These findings indicate that in the blood lipidome of non-demented older persons, a specific profile of lipids involved in membrane fluidity, myelination, and lipid rafts, is associated with subsequent cognitive decline.

FUNDING

The complete list of funders is available at the end of the manuscript, in the Acknowledgement section.

Trajectory of extracellular fluid volume over time and subsequent risks of end-stage kidney disease and mortality in chronic kidney disease: a prospective cohort study

BACKGROUND

Extracellular fluid volume (ECF) is independently associated with chronic kidney disease (CKD) progression and mortality in patients with CKD, but the prognostic value of the trajectory of ECF over time beyond that of baseline value is unknown.

OBJECTIVES

To characterize ECF trajectory and evaluate its association with the risks of end-stage kidney disease (ESKD) and mortality.

METHODS

From the prospective tricentric NephroTest cohort, we included 1588 patients with baseline measured glomerular filtration rate (mGFR) ≥15 mL min^-1 /1.73 m² and ECF measurement. ECF and GFR were measured repeatedly using the distribution volume and clearance of ⁵¹ Cr-EDTA, respectively. ESKD and mortality were traced through record linkage with the national registries. Adjusted shared random-effect joint models were used to analyse the association between the trajectory of ECF over time and the two competing outcomes.

RESULTS

Patients were mean age 58.7 years, 66.7% men, mean mGFR of 43.6 ± 18.6 mL min^-1 /1.73 m² and mean ECF of 16.1 ± 3.6 L. Over a median follow-up of 5.3 [IQR: 3.0;7.4] years, ECF increased by 136 [95%CI 106;167] mL per year on average, whilst diuretic prescription and 24-hour urinary sodium excretion remained stable. ESKD occurred in 324 (20.4%) patients, and 185 (11.6%) patients died before ESKD. A higher current value of ECF was associated with increased hazards of ESKD (adjusted hazard ratio [aHR]: 1.12 [95%CI 1.06;1.18]; P < 0.001 per 1 L increase in ECF), and death before ESKD (aHR: 1.10 [95%CI 1.04;1.17]; P = 0.002).

CONCLUSIONS

The current value of ECF was associated with the risks of ESKD and mortality, independent of multiple potential confounders, including kidney function decline. This highlights the need for a close monitoring and adjustment of treatment to avoid fluid overload in CKD patients.

Blood polyunsaturated omega-3 fatty acids, brain atrophy, cognitive decline, and dementia risk

INTRODUCTION

We searched for consistent associations of an omega-3 index in plasma (sum of eicosapentaenoic acid [EPA] and docosahexaenoic acid [DHA]) with several dementia-related outcomes in a large cohort of older adults.

METHODS

We included 1279 participants from the Three-City study, non-demented at the time of blood measurements at baseline, with face-to-face neuropsychological assessment and systematic detection of incident dementia over a 17-year follow-up. An ancillary study included 467 participants with up to three repeated brain imaging exams over 10 years.

RESULTS

In multivariable models, higher levels of plasma EPA+DHA were consistently associated with a lower risk of dementia (hazard ratio for 1 standard deviation = 0.87 [95% confidence interval, 0.76-0.98]), and a lower decline in global cognition (P = .04 for change over time), memory (P = .06), and medial temporal lobe volume (P = .02).

DISCUSSION

This prospective study provides compelling evidence for a relationship between long-chain omega-3 fatty acids levels and lower risks for dementia and related outcomes.

Long-Term Trajectories of Body Weight, Diet, and Physical Activity From Midlife Through Late Life and Subsequent Cognitive Decline in Women

Healthy lifestyles are promising targets for prevention of cognitive aging, yet the optimal time windows for interventions remain unclear. We selected a case-control sample nested within the Nurses' Health Study (starting year 1976, mean age = 51 years), including 14,956 women aged ≥70 years who were free of both stroke and cognitive impairment at enrollment in a cognitive substudy (1995-2001). Cases (n = 1,496) were women with the 10% worst slopes of cognitive decline, and controls (n = 7,478) were those with slopes better than the median. We compared the trajectories of body mass index (weight (kg)/height (m)2), alternate Mediterranean diet (A-MeDi) score, and physical activity between groups, from midlife through 1 year preceding the cognitive substudy. In midlife, cases had higher body mass index than controls (mean difference (MD) = 0.59 units, 95% confidence interval (CI): 0.39, 0.80), lower physical activity (MD = -1.41 metabolic equivalent of task-hours/week, 95% CI: -2.07, -0.71), and worse A-MeDi scores (MD = -0.16 points, 95% CI: -0.26, -0.06). From midlife through later life, compared with controls, cases had consistently lower A-MeDi scores but a deceleration of weight gain and a faster decrease in physical activity. In conclusion, maintaining a healthy lifestyle since midlife may help reduce cognitive decline in aging. At older ages, both deceleration of weight gain and a decrease in physical activity may reflect early signs of cognitive impairment.

Disease progression and prognostic factors in multiple system atrophy: A prospective cohort study

Multiple system atrophy (MSA) is a rare neurodegenerative disease, with limited understanding of disease progression and prognostic factors. We leveraged the data of a large prospective cohort of MSA to study both clinical progression and survival and assess their determinants. All consecutive patients seen at the French Reference Centre for MSA since 2007 were included in a prospective cohort with an annual follow-up including the Unified MSA Rating Scale (UMSARS). We used joint models to evaluate the risk of death, the mean trajectory of each UMSARS subscale and to determine the potential factors. Investigated factors included gender, age at baseline, MSA subtype, diagnosis certainty, type of first symptoms and the duration between symptom onset and the first visit. Among the 261 MSA patients included in our cohort, the median duration of clinical follow-up was 2.1 years (up to 10.3 years) and the median survival was 4.0 years since the first visit. Main factors for poor survival were the progression over time of UMSARS score (I + II and IV) and the severity of orthostatic hypotension. MSA subtype had no effect on progression or survival. The UMSARS I + II score progressed faster over time in subjects with autonomic dysfunction as the initial feature and in women. Despite a faster progression, women and men had similar survival. From this large MSA cohort, we confirm the rapid progression and poor prognosis of MSA. We provide additional evidence for a negative impact of early autonomic dysfunction and the severity of orthostatic hypotension on both disease progression and survival.

Is Low Psychomotor Speed a Marker of Brain Vulnerability in Late Life? Digit Symbol Substitution Test in the Prediction of Alzheimer, Parkinson, Stroke, Disability, and Depression

BACKGROUND

Dementia, stroke, depression, and disability are frequent in late life and are major causes of quality of life disruption and family burden. Even though each of these disorders relies on specific pathogenic processes, a common clinical manifestation is psychomotor slowing.

OBJECTIVE

We assessed the relevance of a simple marker of low psychomotor speed in predicting several brain outcomes: dementia, Alzheimer's disease (AD), Parkinson's disease (PD), stroke, depressive symptoms, and disability in activities of daily living (ADL) and instrumental ADL (IADL).

METHODS

PAQUID is a population-based study involving 3,777 individuals aged 65 or older prospectively followed-up with repeated clinical evaluations. After 10 years, 437 participants developed dementia, 333 developed AD, 71 developed PD, 207 reported incident stroke, 404 developed disability in ADL, 994 in IADL, and 494 developed depressive symptomology. Psychomotor speed was measured with the digit symbol substitution test (DSST). Cox proportional hazards models controlled for several confounders assessed the risk of incident outcomes.

RESULTS

Participants with low DSST performance had increased risk of incident all-type dementia (hazard ratio [HR] 3.41, p < 0.0001) and AD-type dementia (HR 3.18, p < 0.0001). Higher risk for PD (HR 2.98, p = 0.04), IADL (HR 1.82, p < 0.0001), ADL disability (HR 1.95, p = 0.001), depressive symptoms (HR 1.53, p = 0.03), and a statistical trend for stroke (HR 1.88, p = 0.09) was also found.

CONCLUSION

Low psychomotor speed is associated with an increased risk of developing various brain outcomes: dementia, AD, PD, disability, depressive symptoms, and marginally stroke. Low psychomotor speed may be the consequence of a number of discrete cerebral abnormalities and could be considered as a marker of brain vulnerability. In clinical practice, a low score in DSST should be seen as a warning sign of possible negative evolution.

Dynamic modeling of multivariate dimensions and their temporal relationships using latent processes: Application to Alzheimer's disease

Alzheimer's disease gradually affects several components including the cerebral dimension with brain atrophies, the cognitive dimension with a decline in various functions, and the functional dimension with impairment in the daily living activities. Understanding how such dimensions interconnect is crucial for Alzheimer's disease research. However, it requires to simultaneously capture the dynamic and multidimensional aspects and to explore temporal relationships between dimensions. We propose an original dynamic structural model that accounts for all these features. The model defines dimensions as latent processes and combines a multivariate linear mixed model and a system of difference equations to model trajectories and temporal relationships between latent processes in finely discrete time. Dimensions are simultaneously related to their observed (possibly multivariate) markers through nonlinear equations of observation. Parameters are estimated in the maximum likelihood framework enjoying a closed form for the likelihood. We demonstrate in a simulation study that this dynamic model in discrete time benefits the same causal interpretation of temporal relationships as models defined in continuous time as long as the discretization step remains small. The model is then applied to the data of the Alzheimer's Disease Neuroimaging Initiative. Three longitudinal dimensions (cerebral anatomy, cognitive ability, and functional autonomy) measured by six markers are analyzed, and their temporal structure is contrasted between different clinical stages of Alzheimer's disease.

A joint model for multiple dynamic processes and clinical endpoints: Application to Alzheimer's disease

As other neurodegenerative diseases, Alzheimer's disease, the most frequent dementia in the elderly, is characterized by multiple progressive impairments in the brain structure and in clinical functions such as cognitive functioning and functional disability. Until recently, these components were mostly studied independently because no joint model for multivariate longitudinal data and time to event was available in the statistical community. Yet, these components are fundamentally interrelated in the degradation process toward dementia and should be analyzed together. We thus propose a joint model to simultaneously describe the dynamics of multiple correlated components. Each component, defined as a latent process, is measured by one or several continuous markers (not necessarily Gaussian). Rather than considering the associated time to diagnosis as in standard joint models, we assume diagnosis corresponds to the passing above a covariate-specific threshold (to be estimated) of a pathological process that is modeled as a combination of the component-specific latent processes. This definition captures the clinical complexity of diagnoses such as dementia diagnosis but also benefits from simplifications for the computation of maximum likelihood estimates. We show that the model and estimation procedure can also handle competing clinical endpoints. The estimation procedure, implemented in an R package, is validated by simulations and the method is illustrated on a large French population-based cohort of cerebral aging in which we focused on the dynamics of three clinical manifestations and the associated risk of dementia and death before dementia.

Diet-Related Metabolites Associated with Cognitive Decline Revealed by Untargeted Metabolomics in a Prospective Cohort

Scope

Untargeted metabolomics may reveal preventive targets in cognitive aging, including within the food metabolome.

Methods and results

A case‐control study nested in the prospective Three‐City study includes participants aged ≥65 years and initially free of dementia. A total of 209 cases of cognitive decline and 209 controls (matched for age, gender, education) with slower cognitive decline over up to 12 years are contrasted. Using untargeted metabolomics and bootstrap‐enhanced penalized regression, a baseline serum signature of 22 metabolites associated with subsequent cognitive decline is identified. The signature includes three coffee metabolites, a biomarker of citrus intake, a cocoa metabolite, two metabolites putatively derived from fish and wine, three medium‐chain acylcarnitines, glycodeoxycholic acid, lysoPC(18:3), trimethyllysine, glucose, cortisol, creatinine, and arginine. Adding the 22 metabolites to a reference predictive model for cognitive decline (conditioned on age, gender, education and including ApoE‐ε4, diabetes, BMI, and number of medications) substantially increases the predictive performance: cross‐validated Area Under the Receiver Operating Curve = 75% [95% CI 70–80%] compared to 62% [95% CI 56–67%].

Conclusions

The untargeted metabolomics study supports a protective role of specific foods (e.g., coffee, cocoa, fish) and various alterations in the endogenous metabolism responsive to diet in cognitive aging.

Individual dynamic predictions using landmarking and joint modelling: Validation of estimators and robustness assessment

After the diagnosis of a disease, one major objective is to predict cumulative probabilities of events such as clinical relapse or death from the individual information collected up to a prediction time, usually including biomarker repeated measurements. Several competing estimators have been proposed, mainly from two approaches: joint modelling and landmarking. These approaches differ by the information used, the model assumptions and the complexity of the computational procedures. This paper aims to review the two approaches, precisely define the derived estimators of dynamic predictions and compare their performances notably in case of misspecification. The ultimate goal is to provide key elements for the use of individual dynamic predictions in clinical practice. Prediction of two competing causes of prostate cancer progression from the history of prostate-specific antigen is used as a motivated example. We formally define the quantity to estimate and its estimators, propose techniques to assess the uncertainty around predictions and validate them. We then conduct an in-depth simulation study compare the estimators in terms of prediction error, discriminatory power, efficiency and robustness to model assumptions. We show that prediction tools should be handled with care, in particular by properly specifying models and estimators.

Evaluation of the Concurrent Trajectories of Cardiometabolic Risk Factors in the 14 Years Before Dementia

IMPORTANCE

Cardiometabolic risk factors have been associated with an increased risk of dementia; yet, the optimal targets and time window for the management of cardiometabolic health to prevent dementia remain unknown.

OBJECTIVES

To model concurrently and compare the trajectories of cardiometabolic risk factors up to 14 years preceding diagnosis in individuals with dementia and matched controls free of dementia.

DESIGN, SETTING, AND PARTICIPANTS

A case-control study nested within the Three-City study, a French population-based cohort of older persons (≥65 years), included 6 home visits with neuropsychological testing between 1999 and 2014. Data analysis was performed in September 2017. A total of 785 incident dementia cases and 3140 controls matched by sex, age, educational level, and cohort center at the time of diagnosis were evaluated.

EXPOSURES

Repeated measures of body mass index (BMI) and systolic (SBP) and diastolic (DBP) blood pressure, high-density lipoprotein (HDL-C) and low-density lipoprotein cholesterol (LDL-C), triglycerides, and glycemia levels between 1999 and 2014.

MAIN OUTCOMES AND MEASURES

Incidence of dementia based on systematic detection and validated diagnosis.

RESULTS

A total of 785 cases and 3140 controls (2530 [65%] women; mean [SD] age, 76 [5] years) were included in the study. Cases presented a faster decline in BMI, slower increase of SBP and constantly lower DBP. Mean values (95% CI) 14 years before diagnosis (-14 years) and at diagnosis (year 0) for the most common profile were BMI, 26.1 (25.6-26.5) and 24.8 (24.5-25.1) for a case, and 25.7 (25.4-26.1) and 25.3 (25.0-25.5) for a control; for SBP, 135.2 (131.8-138.7) and 142.1 (140.3-143.9) mm Hg for a case, and 135.8 (132.9-138.6) and 144.9 (143.7-146.1) mm Hg for a control; for DBP, 76.5 (74.7-78.5) and 74.0 (73.1-74.9) mm Hg for a case, and 76.7 (75.1-78.3) and 75.0 (74.2-75.7) mm Hg for a control. In contrast, glycemia was higher among cases (mean fasting glucose values [95% CI] at -14 years and year 0: 89.4 [86.9-92.1] and 96.4 [93.7-99.3] mg/dL for a case, and 87.1 [85.1-89.2] and 95.3 [93.5-97.1] mg/dL for a control), with a significant case-control difference from -1.6 to -14 years prior to diagnosis. There were no significant case-control differences in trajectories of blood lipid levels (mean values [95% CI] at -14 years and year 0: for HDL-C, 70.6 [67.6-73.9] and 61.3 [58.9-63.8] mg/dL for a case, and 70.4 [67.5-73.3] and 62.3 [60.2-64.3] mg/dL for a control; for LDL-C: 147.2 [140.5-154.5] and 141.6 [136.6-146.7] mg/dL for a case, and 144.3 [138.7-150.4] and 141.2 [137.5-145.2] mg/dL for a control; for triglycerides: 115.5 [103.6-149.1] and 112.6 [104.8-120.9] mg/dL for a case, and 112.5 [103.8-144.4] and 109.7 [105.0-114.8] mg/dL for a control).

CONCLUSIONS AND RELEVANCE

In this large cohort of older persons, BMI declined in prodromal dementia, possibly reflecting early preclinical changes. Lower BP prior to dementia may reflect both a consequence and a contributing factor for the disease, whereas higher blood glucose levels may constitute a risk factor for dementia in the older age range. Overall, these findings suggest that elevated glycemia, low BP, and weight loss may be primary targets for the management of cardiometabolic health for primary and secondary prevention of dementia in the older age range.

Temporal Trends in the Level and Decline of Cognition and Disability in an Elderly Population: The PAQUID Study

In line with declining trends in dementia incidence, we compared the cognitive and functional evolution of 2 "generations" of elderly individuals aged 78-88 years, who were included 10 years apart in the French Personnes Agées Quid cohort (n = 612 in 1991-1992 and n = 628 in 2001-2002) and followed-up for 12 years with assessments of cognition and disability. The impact of specific risk factors on this evolution was evaluated. Differences between the generations in baseline levels and decline over time were estimated using a joint model to account for differential attrition. Compared with the first generation, the second generation had higher performances at baseline on 4 cognitive tests (from P < 0.005). Differences in global cognition, verbal fluency, and processing speed, but not in working memory, were mostly explained by improvement in educational level. The second generation also exhibited less cognitive decline in verbal fluency and working memory. Progression of disability was less over the follow-up period for the second generation than for the first. The cognitive state of this elderly population improved, partially due to improvements in educational level.

Association of Cardiovascular Health Level in Older Age With Cognitive Decline and Incident Dementia

IMPORTANCE

Evidence is limited regarding the relation between cardiovascular health level and dementia risk.

OBJECTIVE

To investigate the association between cardiovascular health level, defined using the 7-item tool from the American Heart Association (AHA), and risk of dementia and cognitive decline in older persons.

DESIGN, SETTING, AND PARTICIPANTS

Population-based cohort study of persons aged 65 years or older from Bordeaux, Dijon, and Montpellier, France, without history of cardiovascular diseases or dementia at baseline who underwent repeated in-person neuropsychological testing (January 1999-July 2016) and systematic detection of incident dementia (date of final follow-up, July 26, 2016).

EXPOSURES

The number of the AHA's Life's Simple 7 metrics at recommended optimal level (nonsmoking, body mass index <25, regular physical activity, eating fish twice a week or more and fruits and vegetables at least 3 times a day, cholesterol <200 mg/dL [untreated], fasting glucose <100 mg/dL [untreated], and blood pressure <120/80 mm Hg [untreated]; score range, 0-7) and a global cardiovascular health score (range, 0-14; poor, intermediate, and optimal levels of each metric assigned a value of 0, 1, and 2, respectively).

MAIN OUTCOMES AND MEASURES

Incident dementia validated by an expert committee and change in a composite score of global cognition (in standard units, with values indicating distance from population means, 0 equal to the mean, and +1 and -1 equal to 1 SD above and below the mean).

RESULTS

Among 6626 participants (mean age, 73.7 years; 4200 women [63.4%]), 2412 (36.5%), 3781 (57.1%), and 433 (6.5%) had 0 to 2, 3 to 4, and 5 to 7 health metrics at optimal levels, respectively, at baseline. Over a mean follow-up duration of 8.5 (range, 0.6-16.6) years, 745 participants had incident adjudicated dementia. Compared with the incidence rate of dementia of 1.76 (95% CI, 1.38-2.15) per 100 person-years among those with 0 or 1 health metrics at optimal levels, the absolute differences in incident dementia rates for 2, 3, 4, 5, and 6 to 7 metrics were, respectively, -0.26 (95% CI, -0.48 to -0.04), -0.59 (95% CI, -0.80 to -0.38), -0.43 (95% CI, -0.65 to -0.21), -0.93 (95% CI, -1.18 to -0.68), and -0.96 (95% CI, -1.37 to -0.56) per 100 person-years. In multivariable models, the hazard ratios for dementia were 0.90 (95% CI, 0.84-0.97) per additional optimal metric and 0.92 (95% CI, 0.89-0.96) per additional point on the global score. Furthermore, the gain in global cognition associated with each additional optimal metric at baseline was 0.031 (95% CI, 0.009-0.053) standard units at inclusion, 0.068 (95% CI, 0.045-0.092) units at year 6, and 0.072 (95% CI, 0.042-0.102) units at year 12.

CONCLUSIONS AND RELEVANCE

In this cohort of older adults, increased numbers of optimal cardiovascular health metrics and a higher cardiovascular health score were associated with a lower risk of dementia and lower rates of cognitive decline. These findings may support the promotion of cardiovascular health to prevent risk factors associated with cognitive decline and dementia.