Prediction of dementia risk from multimodal repeated measures: The added value of brain MRI biomarkers.

Abstract: The utility of brain magnetic resonance imaging (MRI) for predicting dementia is debated. We evaluated the added value of repeated brain MRI, including atrophy and cerebral small vessel disease markers, for dementia prediction. We conducted a landmark competing risk analysis in 1716 participants of the French population-based Three-City Study to predict the 5-year risk of dementia using repeated measures of 41 predictors till year 4 of follow-up. Brain MRI markers improved significantly the individual prediction of dementia after accounting for demographics, health measures, and repeated measures of cognition and functional dependency (area under the ROC curve [95% CI] improved from 0.80 [0.79 to 0.82] to 0.83 [0.81 to 0.84]). Nonetheless, accounting for the change over time through repeated MRIs had little impact on predictive abilities. These results highlight the importance of multimodal analysis to evaluate the added predictive abilities of repeated brain MRI for dementia and offer new insights into the predictive performances of various MRI markers. Highlights: We evaluated whether repeated brain volumes and cSVD markers improve dementia prediction.The 5-year prediction of dementia is slightly improved when considering brain MRI markers.Measures of hippocampus volume are the main MRI predictors of dementia.Adjusted on cognition, repeated MRI has poor added value over single MRI for dementia prediction.We utilized a longitudinal analysis that considers error-and-missing-prone predictors, and competing death.

Genetic Complexities of Cerebral Small Vessel Disease, Blood Pressure, and Dementia.

Importance: Vascular disease is a treatable contributor to dementia risk, but the role of specific markers remains unclear, making prevention strategies uncertain. Objective: To investigate the causal association between white matter hyperintensity (WMH) burden, clinical stroke, blood pressure (BP), and dementia risk, while accounting for potential epidemiologic biases. Design, Setting, and Participants: This study first examined the association of genetically determined WMH burden, stroke, and BP levels with Alzheimer disease (AD) in a 2-sample mendelian randomization (2SMR) framework. Second, using population-based studies (1979-2018) with prospective dementia surveillance, the genetic association of WMH, stroke, and BP with incident all-cause dementia was examined. Data analysis was performed from July 26, 2020, through July 24, 2022. Exposures: Genetically determined WMH burden and BP levels, as well as genetic liability to stroke derived from genome-wide association studies (GWASs) in European ancestry populations. Main Outcomes and Measures: The association of genetic instruments for WMH, stroke, and BP with dementia was studied using GWASs of AD (defined clinically and additionally meta-analyzed including both clinically diagnosed AD and AD defined based on parental history [AD-meta]) for 2SMR and incident all-cause dementia for longitudinal analyses. Results: In 2SMR (summary statistics-based) analyses using AD GWASs with up to 75 024 AD cases (mean [SD] age at AD onset, 75.5 [4.4] years; 56.9% women), larger WMH burden showed evidence for a causal association with increased risk of AD (odds ratio [OR], 1.43; 95% CI, 1.10-1.86; P = .007, per unit increase in WMH risk alleles) and AD-meta (OR, 1.19; 95% CI, 1.06-1.34; P = .008), after accounting for pulse pressure for the former. Blood pressure traits showed evidence for a protective association with AD, with evidence for confounding by shared genetic instruments. In the longitudinal (individual-level data) analyses involving 10 699 incident all-cause dementia cases (mean [SD] age at dementia diagnosis, 74.4 [9.1] years; 55.4% women), no significant association was observed between larger WMH burden and incident all-cause dementia (hazard ratio [HR], 1.02; 95% CI, 1.00-1.04; P = .07). Although all exposures were associated with mortality, with the strongest association observed for systolic BP (HR, 1.04; 95% CI, 1.03-1.06; P = 1.9 × 10-14), there was no evidence for selective survival bias during follow-up using illness-death models. In secondary analyses using polygenic scores, the association of genetic liability to stroke, but not genetically determined WMH, with dementia outcomes was attenuated after adjusting for interim stroke. Conclusions: These findings suggest that WMH is a primary vascular factor associated with dementia risk, emphasizing its significance in preventive strategies for dementia. Future studies are warranted to examine whether this finding can be generalized to non-European populations.

Complexities of cerebral small vessel disease, blood pressure, and dementia relationship: new insights from genetics.

Importance: There is increasing recognition that vascular disease, which can be treated, is a key contributor to dementia risk. However, the contribution of specific markers of vascular disease is unclear and, as a consequence, optimal prevention strategies remain unclear. Objective: To disentangle the causal relation of several key vascular traits to dementia risk: (i) white matter hyperintensity (WMH) burden, a highly prevalent imaging marker of covert cerebral small vessel disease (cSVD); (ii) clinical stroke; and (iii) blood pressure (BP), the leading risk factor for cSVD and stroke, for which efficient therapies exist. To account for potential epidemiological biases inherent to late-onset conditions like dementia. Design Setting and Participants: This study first explored the association of genetically determined WMH, BP levels and stroke risk with AD using summary-level data from large genome-wide association studies (GWASs) in a two-sample Mendelian randomization (MR) framework. Second, leveraging individual-level data from large longitudinal population-based cohorts and biobanks with prospective dementia surveillance, the association of weighted genetic risk scores (wGRSs) for WMH, BP, and stroke with incident all-cause-dementia was explored using Cox-proportional hazard and multi-state models. The data analysis was performed from July 26, 2020, through July 24, 2022. Exposures: Genetically determined levels of WMH volume and BP (systolic, diastolic and pulse blood pressures) and genetic liability to stroke. Main outcomes and measures: The summary-level MR analyses focused on the outcomes from GWAS of clinically diagnosed AD (n-cases=21,982) and GWAS additionally including self-reported parental history of dementia as a proxy for AD diagnosis (ADmeta, n-cases=53,042). For the longitudinal analyses, individual-level data of 157,698 participants with 10,699 incident all-cause-dementia were studied, exploring AD, vascular or mixed dementia in secondary analyses. Results: In the two-sample MR analyses, WMH showed strong evidence for a causal association with increased risk of ADmeta (OR, 1.16; 95%CI:1.05-1.28; P=.003) and AD (OR, 1.28; 95%CI:1.07-1.53; P=.008), after accounting for genetically determined pulse pressure for the latter. Genetically predicted BP traits showed evidence for a protective association with both clinically defined AD and ADmeta, with evidence for confounding by shared genetic instruments. In longitudinal analyses the wGRSs for WMH, but not BP or stroke, showed suggestive association with incident all-cause-dementia (HR, 1.02; 95%CI:1.00-1.04; P=.06). BP and stroke wGRSs were strongly associated with mortality but there was no evidence for selective survival bias during follow-up. In secondary analyses, polygenic scores with more liberal instrument definition showed association of both WMH and stroke with all-cause-dementia, AD, and vascular or mixed dementia; associations of stroke, but not WMH, with dementia outcomes were markedly attenuated after adjusting for interim stroke. Conclusion: These findings provide converging evidence that WMH is a leading vascular contributor to dementia risk, which may better capture the brain damage caused by BP (and other etiologies) than BP itself and should be targeted in priority for dementia prevention in the population.

Genome-wide association study of a semicontinuous trait: illustration of the impact of the modeling strategy through the study of Neutrophil Extracellular Traps levels.

Over the last years, there has been a considerable expansion of genome-wide association studies (GWAS) for discovering biological pathways underlying pathological conditions or disease biomarkers. These GWAS are often limited to binary or quantitative traits analyzed through linear or logistic models, respectively. In some situations, the distribution of the outcome may require more complex modeling, such as when the outcome exhibits a semicontinuous distribution characterized by an excess of zero values followed by a non-negative and right-skewed distribution. We here investigate three different modeling for semicontinuous data: Tobit, Negative Binomial and Compound Poisson-Gamma. Using both simulated data and a real GWAS on Neutrophil Extracellular Traps (NETs), an emerging biomarker in immuno-thrombosis, we demonstrate that Compound Poisson-Gamma was the most robust model with respect to low allele frequencies and outliers. This model further identified the MIR155HG locus as significantly (P = 1.4 × 10-8) associated with NETs plasma levels in a sample of 657 participants, a locus recently highlighted to be involved in NETs formation in mice. This work highlights the importance of the modeling strategy for GWAS of a semicontinuous outcome and suggests Compound Poisson-Gamma as an elegant but neglected alternative to Negative Binomial for modeling semicontinuous outcome in the context of genomic investigations.

Association of ABO blood groups with venous thrombosis recurrence in middle-aged patients: insights from a weighted Cox analysis dedicated to ambispective design.

BACKGROUND: In studies of time-to-events, it is common to collect information about events that occurred before the inclusion in a prospective cohort. When the studied risk factors are independent of time, including both pre- and post-inclusion events in the analyses, generally referred to as relying on an ambispective design, increases the statistical power but may lead to a selection bias. In the field of venous thromboembolism (VT), ABO blood groups have been the subject of extensive research due to their substantial effect on VT risk. However, few studies have investigated their effect on the risk of VT recurrence. Motivated by the study of the association of genetically determined ABO blood groups with VT recurrence, we propose a methodology to include pre-inclusion events in the analysis of ambispective studies while avoiding the selection bias due to mortality. METHODS: This work relies on two independent cohorts of VT patients, the French MARTHA study built on an ambispective design and the Dutch MEGA study built on a standard prospective design. For the analysis of the MARTHA study, a weighted Cox model was developed where weights were defined by the inverse of the survival probability at the time of data collection about the events. Thanks to the collection of information on the vital status of patients, we could estimate the survival probabilities using a delayed-entry Cox model on the death risk. Finally, results obtained in both studies were then meta-analysed. RESULTS: In the combined sample totalling 2,752 patients including 993 recurrences, the A1 blood group has an increased risk (Hazard Ratio (HR) of 1.18, p = 4.2 × 10-3) compared with the O1 group, homogeneously in MARTHA and in MEGA. The same trend (HR = 1.19, p = 0.06) was observed for the less frequent A2 group. CONCLUSION: The proposed methodology increases the power of studies relying on an ambispective design which is frequent in epidemiologic studies about recurrent events. This approach allowed to clarify the association of ABO blood groups with the risk of VT recurrence. Besides, this methodology has an immediate field of application in the context of genome wide association studies.

A multistate approach for the study of interventions on an intermediate time-to-event in health disparities research.

We propose a novel methodology to quantify the effect of stochastic interventions for a non-terminal intermediate time-to-event on a terminal time-to-event outcome. Investigating these effects is particularly important in health disparities research when we seek to quantify inequities in the timely delivery of treatment and its impact on patients' survival time. Current approaches fail to account for time-to-event intermediates and semi-competing risks arising in this setting. Under the potential outcome framework, we define causal contrasts relevant in health disparities research and provide identifiability conditions when stochastic interventions on an intermediate non-terminal time-to-event are of interest. Causal contrasts are estimated in continuous time within a multistate modeling framework and analytic formulae for the estimators of the causal contrasts are developed. We show via simulations that ignoring censoring in intermediate and/or terminal time-to-event processes or ignoring semi-competing risks may give misleading results. This work demonstrates that a rigorous definition of the causal effects and joint estimation of the terminal outcome and intermediate non-terminal time-to-event distributions are crucial for valid investigation of interventions and mechanisms in continuous time. We employ this novel methodology to investigate the role of delaying treatment uptake in explaining racial disparities in cancer survival in a cohort study of colon cancer patients.

Impact of interventions scenarios targeting three main vascular risk factors on the future burden of dementia in France.

The epidemiological and societal burden of dementia is expected to increase in the coming decades due to the world population aging. In this context, the evaluation of the potential impact of intervention scenarios aiming at reducing the prevalence of dementia risk factors is an active area of research. However, such studies must account for the associated changes in mortality and the dependence between the risk factors. Using micro-simulations, this study aims to estimate the changes in dementia burden in France in 2040 according to intervention scenarios targeting the prevention or treatment of hypertension, diabetes and physical inactivity. Accounting for their communality and their effects on mortality, the results show that the disappearance of hypertension, diabetes and physical inactivity in France in 2020 could decrease dementia prevalence by 33% among men and 26% among women in 2040 and increase the life expectancy without dementia at age 65 by 3.4 years (men) and 2.6 years (women). Among the three factors, the prevention of hypertension would be the most efficient. These projections rely on current estimates of the risk of dementia and death associated with risk factors. Thanks to the R package developed they could be refined for different countries or different interventions and updated with new estimates.

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.

Five-Year Dynamic Prediction of Dementia Using Repeated Measures of Cognitive Tests and a Dependency Scale.

The progression of dementia prevalence over the years and the lack of efficient treatments to stop or reverse the cognitive decline make dementia a major public health challenge in the developed world. Identifying people at high risk of developing dementia could improve the treatment of these patients and help select the target population for preventive clinical trials. We used joint modeling to build a dynamic prediction tool of dementia based on the change over time of 2 neurocognitive tests (the Mini-Mental State Examination and the Isaacs Set Tests) as well as an autonomy scale (the Instrumental Activities of Daily Living). The model was estimated with data from the French cohort Personnes Agées QUID (1988-2015) and validated both by cross-validation and externally with data from the French Three City cohort (1999-2018). We evaluated its predictive abilities through area under the receiver operating characteristics curve and Brier score, accounting for right censoring and competing risk of death, and obtained an average area under the curve value of 0.95 for the risk of dementia in the next 5 or 10 years. This tool is able to discriminate a high-risk group of people from the rest of the population. This could be of help in clinical practice and research.

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.

Incidence and Risk Factors in Concussion Events: A 5-Season Study in the French Top 14 Rugby Union Championship.

BACKGROUND: Concussions are a source of major concern in rugby, and a limited number of studies have attempted to identify risk factors for these injuries. PURPOSE: To investigate the incidence of match concussion and associated risk factors, including individual workload, anthropometric variables, playing position, and season phase, in elite rugby union players. STUDY DESIGN: Case-control study; Level of evidence 3. METHODS: All concussions and explanatory variables were collected for every match over 5 consecutive seasons (2014-2018) in 1334 professional players participating in the French Top 14 championship. Concussion risk was estimated using mixed effects Poisson regression. RESULTS: Mean match concussion incidence equated to 10.4 (95% CI, 9.3-11.5) concussions for 1000 hours of game exposure. A peak was reached in the 2016-2017 season (13.7; 95% CI, 11.0-16.5). A greater risk was observed in the playoffs as compared with the first phase of the season (incidence rate ratio, 3.96; 95% CI, 2.10-7.35). In comparison with other positions, half-backs incurred the highest rate of concussion events (incidence, 16.1; 95% CI, 11.8-20.3). Irrespective of playing position, those with greater height and lower body mass reported a higher risk of concussions (P = .02), especially during tackling actions for lighter players (P = .01) and during other match events for taller players (P = .03). When adjusted for season phase, players who had accumulated a higher amount of playing time since the beginning of the season demonstrated a lower risk of concussion (P = .005). CONCLUSION: Inter- and intraseasonal variations in concussion rates were observed. Within positional groups, lighter and taller players were more at risk, with the highest incidence generally observed in half-backs. Workload was measured by the number of matches played before a concussion event, and it appeared to have a protective rather than deleterious effect on concussion risk.

Predicting Progression to Advanced Age-Related Macular Degeneration from Clinical, Genetic, and Lifestyle Factors Using Machine Learning.

PURPOSE: Current prediction models for advanced age-related macular degeneration (AMD) are based on a restrictive set of risk factors. The objective of this study was to develop a comprehensive prediction model applying a machine learning algorithm allowing selection of the most predictive risk factors automatically. DESIGN: Two population-based cohort studies. PARTICIPANTS: The Rotterdam Study I (RS-I; training set) included 3838 participants 55 years of age or older, with a median follow-up period of 10.8 years, and 108 incident cases of advanced AMD. The Antioxydants, Lipids Essentiels, Nutrition et Maladies Oculaires (ALIENOR) study (test set) included 362 participants 73 years of age or older, with a median follow-up period of 6.5 years, and 33 incident cases of advanced AMD. METHODS: The prediction model used the bootstrap least absolute shrinkage and selection operator (LASSO) method for survival analysis to select the best predictors of incident advanced AMD in the training set. Predictive performance of the model was assessed using the area under the receiver operating characteristic curve (AUC). MAIN OUTCOME MEASURES: Incident advanced AMD (atrophic, neovascular, or both), based on standardized interpretation of retinal photographs. RESULTS: The prediction model retained (1) age, (2) a combination of phenotypic predictors (based on the presence of intermediate drusen, hyperpigmentation in one or both eyes, and Age-Related Eye Disease Study simplified score), (3) a summary genetic risk score based on 49 single nucleotide polymorphisms, (4) smoking, (5) diet quality, (6) education, and (7) pulse pressure. The cross-validated AUC estimation in RS-I was 0.92 (95% confidence interval [CI], 0.88-0.97) at 5 years, 0.92 (95% CI, 0.90-0.95) at 10 years, and 0.91 (95% CI, 0.88-0.94) at 15 years. In ALIENOR, the AUC reached 0.92 at 5 years (95% CI, 0.87-0.98). In terms of calibration, the model tended to underestimate the cumulative incidence of advanced AMD for the high-risk groups, especially in ALIENOR. CONCLUSIONS: This prediction model reached high discrimination abilities, paving the way toward making precision medicine for AMD patients a reality in the near future.

Patterns of Benzodiazepine Use and Excess Risk of All-Cause Mortality in the Elderly: A Nationwide Cohort Study.

INTRODUCTION: Despite the risks associated with their use, benzodiazepines remain used more widely than wisely. In this context, a better understanding of how their patterns of use can be associated with an increased risk of death appears essential. Indeed, the studies that investigated this association so far are inconsistent and question the influence of potential biases. OBJECTIVE: The objective of this study was to investigate the association of various patterns of benzodiazepine use with all-cause mortality. METHODS: A nationwide cohort of non-prevalent benzodiazepine users aged ≥ 65 years was identified using French healthcare insurance system claims databases. Exposure to benzodiazepines considered short-term, chronic (defined as a cumulated ≥ 6-month period over the previous 12 months), ongoing, and discontinued use. Using a Cox model, adjusted hazard ratios for all-cause mortality were estimated according to benzodiazepine patterns of use; exposure and confounders were treated as time-dependent variables. RESULTS: In the cohort of 54,958 individuals aged ≥ 65 years, adjusted hazard ratios for all-cause mortality and benzodiazepines were 2.26 (95% confidence interval 1.96-2.61) for short-term use, 3.86 (3.04-4.90) for chronic use-discontinued, and 3.05 (2.17-4.29) for chronic use-ongoing. At age 80 years, these were 1.62 (1.48-1.79), 2.00 (1.82-2.19) and 1.13 (1.02-1.26), respectively. Adjusted hazard ratios show similar decreases with age for all patterns of benzodiazepine use. CONCLUSIONS: These findings confirm the existence of an excess risk of mortality associated with benzodiazepine use and provide pattern- and age-specific estimates. Higher risks were observed for patients aged < 80 years, short-term use, or chronic use recently interrupted. If the two latter can relate to an indication bias, the associations found for ongoing chronic use and short-term use conversely support a potential causal hypothesis.

Impact of benzodiazepine consumption reduction on future burden of dementia.

Dementia is a major public health issue worldwide and chronic use of benzodiazepine, which is very frequent in northern countries, was found to be a risk factor of dementia. This work aims at evaluating the impact of a reduction in chronic use of benzodiazepine on the future burden of dementia in France. Using estimations of dementia incidence and of benzodiazepine use and nation-wide projections of mortality and population sizes, a Monte Carlo approach based on an illness-death model provided projections of several indicators of dementia burden. With no change in benzodiazepine consumption, the prevalence of dementia between age 65 and 99 in France in 2040 was estimated at 2.16 millions (95% confidence interval (CI) 1.93-2.38), with a life expectancy without dementia at 65 years equal to 25.0 years (24.7-25.3) for women and 23.8 years (23.5-24.2) for men. Assuming a disappearance of chronic use of benzodiazepine in 2020, the prevalence would be reduced by about 6.6% in 2040 and the life expectancy without dementia would increase by 0.99 (0.93-1.06) year among women and 0.56 (0.50-0.62) among men. To conclude, a modest but significant reduction in future dementia burden could be obtained by applying current recommendation for duration of benzodiazepine use.

Cognitive and functional changes in prediagnostic phase of Parkinson disease: A population-based study.

INTRODUCTION: Prodromal non-motor symptoms precede, often by decades, motor signs and diagnosis of Parkinson's disease. It is however still uncertain if cognitive changes belong to the spectrum of non-motor prodromal Parkinson's disease. Thanks to the very long-term follow-up of the PAQUID population-based cohort, we assessed trajectories of cognitive complaints and functioning over a 13-year period before the diagnosis of late onset Parkinson's disease. METHODS: This study relies on a matched nested case-control sample selected from the cohort. Of the 3777 initial subjects of the cohort, 43 developed incident Parkinson's disease over the follow-up. The mean age at diagnosis was 78.0 (standard deviation = 5.8) years and 46.5% were men. These cases were matched to 86 elderly control subjects. Scores of different cognitive domains, daily function, and depressive symptoms were described throughout the follow-up using mixed-effects models. RESULTS: No significant global cognitive decline preceded the diagnosis of late onset Parkinson's disease. However, psychomotor speed appeared significantly slower 2 years before the diagnosis and depressive symptoms 12 years before. Global score of instrumental activities of daily living became altered 2-3 years preceding the diagnosis of late onset Parkinson's disease, including the use of public transportation that was altered ten years before the diagnosis. CONCLUSION: In late onset Parkinson's disease, while global cognitive functions seem preserved, psychomotor speed starts to decline 2 years before the diagnosis and activities of daily living are also impacted. Depressive symptoms appear very early in the prediagnosic phase.

Excess cumulative incidence estimation for matched cohort survival studies.

We suggest a regression approach to estimate the excess cumulative incidence function (CIF) when matched data are available. In a competing risk setting, we define the excess risk as the difference between the CIF in the exposed group and the background CIF observed in the unexposed group. We show that the excess risk can be estimated through an extended binomial regression model that actively uses the matched structure of the data, avoiding further estimation of both the exposed and the unexposed CIFs. The method naturally deals with two time scales, age and time since exposure and simplifies how to deal with the left truncation on the age time-scale. The model makes it easy to predict individual excess risk scenarios and allows for a direct interpretation of the covariate effects on the cumulative incidence scale. After introducing the model and some theory to justify the approach, we show via simulations that our model works well in practice. We conclude by applying the excess risk model to data from the ALiCCS study to investigate the excess risk of late events in childhood cancer survivors.

Long-term risk of hip or forearm fractures in older occasional users of benzodiazepines.

AIMS: This article sought to study the association between patterns of benzodiazepine (BZD) use and the risk of hip and forearm fractures in people aged 50 and 75 years or more. METHODS: In a representative cohort of the French National Health Insurance Fund of individuals aged 50 years or older (n = 106 437), we followed up BZD dispensing (reflecting their patterns of use) and the most frequent fall-related fractures (hip and forearm) for 8 years. We used joint latent class models to simultaneously identify BZD dispensing trajectories and the risk of fractures in the entire cohort and in those 75 years or older). We used a survival model to estimate the adjusted hazard ratios (aHRs) between these trajectories and the risk of fractures. RESULTS: In the entire cohort, we identified 5 BZD trajectories: non-users (76.7% of the cohort); occasional users (15.2%); decreasing users (2.6%); late increasing users (3.0%); and early increasing users (2.4%). Compared with non-users, fracture risk was not increased in either occasional users (aHR = 0.99, 95% confidence interval [CI] 0.99-1.00) or in decreasing users (aHR = 0.90, 95% CI 0.74-1.08). It was significantly higher in early increasing users (aHR = 1.86, 95% CI 1.62-2.14) and in late increasing users (aHR = 1.39, 95% CI 1.15-1.60). We observed similar trajectories and risk levels in the people older than 75 years. CONCLUSION: Occasional BZD use, which is compatible with current recommendations, was not associated with an excess risk of the most frequent fall-related fractures in people older than 50 or 75 years.

Quantile regression for incomplete longitudinal data with selection by death.

Quantile regressions are increasingly used to provide population norms for quantitative variables. Indeed, they do not require any Gaussian assumption for the response and allow to characterize its entire distribution through different quantiles. Quantile regressions are especially useful to provide norms of cognitive scores in the elderly that may help general practitioners to identify subjects with unexpectedly low cognitive level in routine examinations. These norms may be estimated from cohorts of elderly using quantile regression for longitudinal data, but this requires to properly account for selection by death, dropout and intermittent missing data. In this work, we extend the weighted estimating equation approach to estimate conditional quantiles in the population currently alive from mortal cohorts with dropout and intermittent missing data. Suitable weight estimation procedures are provided for both monotone and intermittent missing data and under two missing-at-random assumptions, when the observation probability given that the subject is alive depends on the survival time (p-MAR assumption) or not (u-MAR assumption). Inference is performed through subject-level bootstrap. The method is validated in a simulation study and applied to the French cohort Paquid to estimate quantiles of a cognitive test in the elderly population currently alive. On one hand, the simulations show that the u-MAR analysis is quite robust when the true missingness mechanism is p-MAR. This is a useful result because computation of suitable weights for intermittent missing data under the p-MAR assumption is untractable. On the other hand, the simulations highlight, along with the real data analysis, the usefulness of suitable weights for intermittent missing data. This method is implemented in the R package weightQuant.