Self-Rated Health and Frailty in Older Adults from the Population-Based Three-City Bordeaux Cohort.

INTRODUCTION: This study aimed to investigate whether self-rated health (SRH) predict frailty and its components among community dwellers aged 75 years and older. METHODS: We ran a cross-sectional and prospective analysis from 643 and 379 participants of the Bordeaux Center (France) of the Three-City Study, respectively. We assessed SRH using a single question with 5 response options. We defined frailty as having at least 3 out of the following 5 criteria: weight loss, exhaustion, slowness, weakness, and low energy expenditure. We used multivariate logistic regression and Cox proportional hazard models. RESULTS: At baseline, poor SRH was significantly associated with frailty (odds ratio = 5.2; 95% confidence interval [CI]: 2.9-9.5) and its components except for weakness. In the prospective analysis on nonfrail participants, poor SRH was associated with the 4-year risk of slowness (hazard ratio [HR] = 1.7; 95% CI: 1.1-2.6) but not with that of frailty (HR = 1.6; 95% CI: 0.9-2.9) or the other components. CONCLUSIONS: In a French cohort of community dwellers aged 75 years or older, poorer SRH was associated with concomitant frailty and 70% higher risk of slowness over 4 years.

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.

Benefits of dimension reduction in penalized regression methods for high-dimensional grouped data: a case study in low sample size.

MOTIVATION: In some prediction analyses, predictors have a natural grouping structure and selecting predictors accounting for this additional information could be more effective for predicting the outcome accurately. Moreover, in a high dimension low sample size framework, obtaining a good predictive model becomes very challenging. The objective of this work was to investigate the benefits of dimension reduction in penalized regression methods, in terms of prediction performance and variable selection consistency, in high dimension low sample size data. Using two real datasets, we compared the performances of lasso, elastic net, group lasso, sparse group lasso, sparse partial least squares (PLS), group PLS and sparse group PLS. RESULTS: Considering dimension reduction in penalized regression methods improved the prediction accuracy. The sparse group PLS reached the lowest prediction error while consistently selecting a few predictors from a single group. AVAILABILITY AND IMPLEMENTATION: R codes for the prediction methods are freely available at https://github.com/SoufianeAjana/Blisar. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Integrating Metabolomics, Genomics, and Disease Pathways in Age-Related Macular Degeneration: The EYE-RISK Consortium.

PURPOSE: The current study aimed to identify metabolites associated with age-related macular degeneration (AMD) by performing the largest metabolome association analysis in AMD to date, as well as aiming to determine the effect of AMD-associated genetic variants on metabolite levels and investigate associations between the identified metabolites and activity of the complement system, one of the main AMD-associated disease pathways. DESIGN: Case-control association analysis of metabolomics data. PARTICIPANTS: Five European cohorts consisting of 2267 AMD patients and 4266 control participants. METHODS: Metabolomics was performed using a high-throughput proton nuclear magnetic resonance metabolomics platform, which allows quantification of 146 metabolite measurements and 79 derivative values. Metabolome-AMD associations were studied using univariate logistic regression analyses. The effect of 52 AMD-associated genetic variants on the identified metabolites was investigated using linear regression. In addition, associations between the identified metabolites and activity of the complement pathway (defined by the C3d-to-C3 ratio) were investigated using linear regression. MAIN OUTCOME MEASURES: Metabolites associated with AMD. RESULTS: We identified 60 metabolites that were associated significantly with AMD, including increased levels of large and extra-large high-density lipoprotein (HDL) subclasses and decreased levels of very low-density lipoprotein (VLDL), amino acids, and citrate. Of 52 AMD-associated genetic variants, 7 variants were associated significantly with 34 of the identified metabolites. The strongest associations were identified for genetic variants located in or near genes involved in lipid metabolism (ABCA1, CETP, APOE, and LIPC) with metabolites belonging to the large and extra-large HDL subclasses. Also, 57 of 60 metabolites were associated significantly with complement activation levels, independent of AMD status. Increased large and extra-large HDL levels and decreased VLDL and amino acid levels were associated with increased complement activation. CONCLUSIONS: Lipoprotein levels were associated with AMD-associated genetic variants, whereas decreased essential amino acids may point to nutritional deficiencies in AMD. We observed strong associations between the vast majority of the AMD-associated metabolites and systemic complement activation levels, independent of AMD status. This may indicate biological interactions between the main AMD disease pathways and suggests that multiple pathways may need to be targeted simultaneously for successful treatment of AMD.

Autoantibodies to human citrullinated fibrinogen and their subfamilies to the α36-50Cit and ß60-74Cit fibrin peptides similarly predict radiographic damages: a prospective study in the French ESPOIR cohort of very early arthritides.

OBJECTIVE: To investigate whether subfamilies of the RA-specific autoantibodies to human citrullinated fibrinogen (AhFibA) differentially associate with the RA risk factors, HLA-DRB1 shared epitope containing alleles (SE alleles) and cigarette smoking, and thus help to predict the disease outcome. METHODS: AhFibA and their anti-α36-50Cit and anti-ß60-74Cit subfamilies were assayed by ELISA, at baseline, in the French ESPOIR (Etude et Suivi des Polyarthrites Indifférenciées Récentes) cohort composed of undifferentiated arthritides and RA patients of < 6 months' duration. Cigarette smoking, SE alleles' presence, DAS28, HAQ and modified Sharp-van der Heijde Score data were obtained at baseline, and after follow-up. RESULTS: After 3 years, 701 patients were classified as having RA according to the ACR/EULAR 2010 criteria. Among them, 349 (50%), 203 (29%) and 257 (37%) were AhFibA-, anti-α36-50Cit- and anti-ß60-74Cit-positive, respectively. The presence and titres of AhFibA and their subfamilies similarly associated with SE alleles, irrespective of their fine specificity, without significant effect of smoking. Neither their presence nor their titre was associated with DAS28 or HAQ. The presence of at least one subfamily was associated with a faster Sharp/van der Heijde score progression, albeit without correlation with the titre. CONCLUSION: AhFibA and their main subfamilies are similarly associated with SE alleles without additional effect of smoking. Whatever their fine specificity was, their presence (but not their titre) similarly constituted a marker of faster joint destruction.

DeepAlienorNet: A deep learning model to extract clinical features from colour fundus photography in age-related macular degeneration.

OBJECTIVE: This study aimed to develop a deep learning (DL) model, named 'DeepAlienorNet', to automatically extract clinical signs of age-related macular degeneration (AMD) from colour fundus photography (CFP). METHODS AND ANALYSIS: The ALIENOR Study is a cohort of French individuals 77 years of age or older. A multi-label DL model was developed to grade the presence of 7 clinical signs: large soft drusen (>125 µm), intermediate soft (63-125 µm), large area of soft drusen (total area >500 µm), presence of central soft drusen (large or intermediate), hyperpigmentation, hypopigmentation, and advanced AMD (defined as neovascular or atrophic AMD). Prediction performances were evaluated using cross-validation and the expert human interpretation of the clinical signs as the ground truth. RESULTS: A total of 1178 images were included in the study. Averaging the 7 clinical signs' detection performances, DeepAlienorNet achieved an overall sensitivity, specificity, and AUROC of 0.77, 0.83, and 0.87, respectively. The model demonstrated particularly strong performance in predicting advanced AMD and large areas of soft drusen. It can also generate heatmaps, highlighting the relevant image areas for interpretation. CONCLUSION: DeepAlienorNet demonstrates promising performance in automatically identifying clinical signs of AMD from CFP, offering several notable advantages. Its high interpretability reduces the black box effect, addressing ethical concerns. Additionally, the model can be easily integrated to automate well-established and validated AMD progression scores, and the user-friendly interface further enhances its usability. The main value of DeepAlienorNet lies in its ability to assist in precise severity scoring for further adapted AMD management, all while preserving interpretability.

Metabolomics in serum of patients with non-advanced age-related macular degeneration reveals aberrations in the glutamine pathway.

Age-related macular degeneration (AMD) is a common, progressive multifactorial vision-threatening disease and many genetic and environmental risk factors have been identified. The risk of AMD is influenced by lifestyle and diet, which may be reflected by an altered metabolic profile. Therefore, measurements of metabolites could identify biomarkers for AMD, and could aid in identifying high-risk individuals. Hypothesis-free technologies such as metabolomics have a great potential to uncover biomarkers or pathways that contribute to disease pathophysiology. To date, only a limited number of metabolomic studies have been performed in AMD. Here, we aim to contribute to the discovery of novel biomarkers and metabolic pathways for AMD using a targeted metabolomics approach of 188 metabolites. This study focuses on non-advanced AMD, since there is a need for biomarkers for the early stages of disease before severe visual loss has occurred. Targeted metabolomics was performed in 72 patients with early or intermediate AMD and 72 control individuals, and metabolites predictive for AMD were identified by a sparse partial least squares discriminant analysis. In our cohort, we identified four metabolite variables that were most predictive for early and intermediate stages of AMD. Increased glutamine and phosphatidylcholine diacyl C28:1 levels were detected in non-advanced AMD cases compared to controls, while the rate of glutaminolysis and the glutamine to glutamate ratio were reduced in non-advanced AMD. The association of glutamine with non-advanced AMD corroborates a recent report demonstrating an elevated glutamine level in early AMD using a different metabolomics technique. In conclusion, this study indicates that metabolomics is a suitable method for the discovery of biomarker candidates for AMD. In the future, larger metabolomics studies could add to the discovery of novel biomarkers in yet unknown AMD pathways and expand our insights in AMD pathophysiology.

High adherence to a Mediterranean diet and lower risk of frailty among French older adults community-dwellers: Results from the Three-City-Bordeaux Study.

BACKGROUND & AIMS: Mediterranean diet (MeDi) is considered as a key component for healthy aging, including prevention of age-related disability, while its association with frailty, independent of disability has never been assessed. Our objective was to investigate the relation between MeDi adherence and frailty incidence among persons aged ≥75 years participating at the prospective population-based French Three-City Study. METHODS: The study sample consisted of 560 initially non-frail participants of the Three-City-Bordeaux center, seen at the 2009-2010 follow-up, and re-examined two years later. Adherence to MeDi was computed from a food frequency questionnaire (scored as 0-9). Frailty was defined as having at least three out of the following five slightly modified Fried frailty criteria: involuntary weight loss, exhaustion, slowness, weakness and low physical activity. Logistic regression models adjusted for sociodemographic and clinical covariates, including cognitive performance and depressive symptomatology, were used to assess the association between MeDi score and subsequent frailty risk. RESULTS: Over the 2-year follow-up, 79 participants (14%) became frail. Older adults with the highest MeDi adherence (score 6-9) had a significantly 68% frailty risk reduction (95% CI: 28-86%, p = 0.006) compared to those in the lowest MeDi category (score 0-3). Regarding the frailty criterion separately, the highest MeDi adherence was associated with a significantly reduced risk of incident slowness (OR = 0.45; 95% CI: 0.20-0.99, p = 0.04), poor muscle strength (OR = 0.44; 95% CI: 0.20-0.98, p = 0.04) and low physical activity (OR = 0.39; 95% CI: 0.18-0.82, p = 0.01), compared to the lowest MeDi adherence. CONCLUSION: In addition to its well-documented beneficial effects on health, adherence to MeDi might contribute to prevent the onset of frailty, even at late stages of life.

Plasma Concentrations of Lutein and Zeaxanthin, Macular Pigment Optical Density, and Their Associations With Cognitive Performances Among Older Adults.

Purpose: We investigated the cross-sectional associations between macular pigment optical density (MPOD), plasma lutein (L), and zeaxanthin (Z) concentrations and cognitive function in 184 older adults of the 3-City-Bordeaux cohort. Methods: MPOD was measured using the two-wavelength autofluorescence method with a modified scanning laser ophthalmoscope. Plasma L and Z (L+Z) concentrations were determined by high-performance liquid chromatography and were considered either crude or expressed as a ratio of the concentration of plasma lipids (total cholesterol [TC] + triglycerides [TG]). Cognitive performances were assessed using the following four separate neuropsychological tests: the Mini-Mental State Examination (MMSE), the Isaacs Set Test (IST), the Benton Visual Retention Test (BVRT), and the sum of the three free recalls of the Free and Cued Selective Reminding Test (FCSRT). These test results were summarized by a composite global cognitive z-score. Results: Higher MPOD at 0.5° was significantly associated with a higher composite z-score (ß = 0.15, 95% confidence interval [CI] 0.04-0.26), higher BVRT (ß = 0.39, 95%CI 0.08-0.70), and higher IST (ß = 1.16, 95%CI 0.11-2.22) performances. Higher plasma L+Z concentrations were significantly associated with higher IST scores (ß = 0.97, 95%CI 0.01-1.94). Furthermore, a higher L+Z/TC+TG ratio was associated with a higher composite z-score (ß = 0.12, 95%CI 0.01-0.23), along with higher IST (ß = 1.02, 95%CI 0.002-2.04) and FCSRT (ß = 1.55, 95%CI 0.41-2.69) performances. Conclusions: This analysis suggested that both higher MPOD and L+Z concentrations were significantly associated with higher cognitive performances. However, MPOD measurements have the advantage of being a fast and representative measure of long-term carotenoid intake.

Adapted Surgical Procedure for Argus II Retinal Implantation: Feasibility, Safety, Efficiency, and Postoperative Anatomic Findings.

PURPOSE: To evaluate the feasibility, safety, and efficiency of an adapted surgical procedure used for postmarket Argus II implantations, so as to lower risks of postoperative hypotony or conjunctivoscleral erosion, and to describe the observed anatomic characteristics of the positioning of the implanted array. DESIGN: Single-arm prospective multicenter clinical trial. PARTICIPANTS: Eighteen consecutive patients with end-stage retinitis pigmentosa. METHODS: To protect the site of insertion of the cable of the device, a scleral flap was systematically added to the standardized implantation procedure. It was associated with temporalis fascia autograft, so as to cover the episcleral-fixed electronics case. Intraoperative and postoperative data at day 1, weeks 1 and 2, and months 1, 3, and 6 were collected. Postoperative distance between electrode-array and retina was measured on spectral-domain optical coherence tomography images. Position of the array was evaluated on fundus images between months 1 and 6. MAIN OUTCOME MEASURES: Feasibility of the modified surgical technique (time constraints, intraoperative complications), variations of intraocular pressure over time, postoperative ocular findings and adverse events, postoperative distance between the array and the retina, and rotation of the array between months 1 and 6 after implantation. RESULTS: The adapted surgical technique was performed easily without associated specific complications. No cases of chronic hypotony or conjunctivoscleral erosion were reported. One serious device/procedure-related adverse event was recorded (sterile posterior uveitis), which resolved after vitrectomy. Postoperative distance between array and retina was variable: full apposition was achieved in 4 patients (22.22%), partial apposition observed in 9 patients (50.00%), and absence of strict apposition noted in 5 patients (27.78%, 4 of whom had posterior staphyloma). A statistically significant slight rotation of the array was observed between months 1 and 6 (P < 0.0001), occurring downwardly in 68.75% of cases. CONCLUSIONS: The combined use of scleral flap and temporalis fascia autograft was easily achieved and effective in preventing hypotony and conjunctival erosion in our study. Postoperative distance between semirigid array and retinal surface was variable, and increased in the case of preoperative staphyloma. A slight rotation of the device occurred over time. Further studies based on larger samples are needed to confirm our findings and determine their functional consequences.

Obesity in French Inmates: Gender Differences and Relationship with Mood, Eating Behavior and Physical Activity.

CONTEXT: Inmates, notably women, are at greater risk for obesity and metabolic complications than the general population according to several studies from high income countries. Data regarding French correctional institutions are lacking so far. To fill this gap, we have assessed in a sample from a French prison (33 females and 18 males) the gender-specific effect of incarceration on weight and body mass index (BMI) and examined their current metabolic status. Furthermore, to reveal the possible determinants of increased obesity, we analyzed emotional vulnerability, eating behavior and physical activity using self-reported questionnaires. RESULTS: In this sample, obesity (BMI≥30 kg/m2) was already frequent in women (18.2%) but rather scarce for men (11%) at prison entry. Incarceration worsened the rate of obesity in both genders (21.2% and 16.7% respectively). At the time of study, abdominal obesity estimated through waist circumference was particularly prevalent in women (69.7%) versus men (27.8%) and metabolic syndrome was detected in 33% of female against none in male inmates. Abdominal obesity was associated with female sex (p<0.03), low physical activity (p<0.05) and eating disorder (p = 0.07) in univariate analyses. Low physical activity remained significant as an explanatory factor of higher abdominal obesity in multivariate analysis. A marked difference between genders was found for practice of physical activity with a higher proportion of women compared to men being inactive (37.9% vs. 11.8%) and fewer women being very active (17.2% vs. 41.2%). CONCLUSION: This study revealed that a significant proportion of women of this correctional institution combined established obesity, a metabolic syndrome and very little practice of physical activity which put them at high risk of cardiovascular disease. Thus, obesity should be better surveyed and treated in prison, especially for female inmates. Increased physical activity, adapted to obese women, would be the first mean to decrease obesity and gender differences.

Dietary Patterns and 12-Year Risk of Frailty: Results From the Three-City Bordeaux Study.

OBJECTIVE: To analyze the association between dietary patterns and the 12-year risk of frailty and its components in community-dwelling elderly French adults. DESIGN: A prospective cohort study. SETTING: The Bordeaux sample of the Three-City Study. PARTICIPANTS: A total of 972 initially nonfrail nondemented participants (336 men and 636 women) aged 73 years on average, re-examined at least once over 12 years. MEASUREMENTS: Five sex-specific dietary clusters were previously derived at baseline. Frailty incident to the baseline visit was defined as having at least three out of the following 5 criteria: unintentional weight loss, exhaustion, low energy expenditure, slowness, and muscle weakness. Multivariate Cox proportional hazard models were used to assess the association between dietary clusters and the risk of frailty and its components. RESULTS: In total, 78 men for 3719 person-years and 221 women for 7027 person-years became frail over the follow-up. In multivariate analyses, men in the "pasta" pattern and women in the "biscuits and snacking" pattern had a significantly higher risk of frailty compared with those in the "healthy" pattern [hazard ratio (HR) 2.2; 95% confidence interval (CI) 1.1-4.4 and HR 1.8; 95% CI 1.2-2.8, respectively; P = .09 and P = .13 for the global test of significance of risk difference across clusters, respectively]. In men, "biscuits and snacking" and "pasta" patterns were significantly associated with higher risk for muscle weakness (HR 3.3; 95% CI 1.6-7.0 and HR 2.1; 95% CI 1.2-3.7, respectively; P = .003 for global test). CONCLUSIONS: This 12-year prospective population-based study suggests that some particular unhealthy dietary patterns may increase the risk of frailty in older adults.

Comparison of variable selection methods for high-dimensional survival data with competing events.

BACKGROUND: In the era of personalized medicine, it's primordial to identify gene signatures for each event type in the context of competing risks in order to improve risk stratification and treatment strategy. Until recently, little attention was paid to the performance of high-dimensional selection in deriving molecular signatures in this context. In this paper, we investigate the performance of two selection methods developed in the framework of high-dimensional data and competing risks: Random survival forest and a boosting approach for fitting proportional subdistribution hazards models. METHODS: Using data from bladder cancer patients (GSE5479) and simulated datasets, stability and prognosis performance of the two methods were evaluated using a resampling strategy. For each sample, the data set was split into 100 training and validation sets. Molecular signatures were developed in the training sets by the two selection methods and then applied on the corresponding validation sets. RESULTS: Random survival forest and boosting approach have comparable performance for the prediction of survival data, with few selected genes in common. Nevertheless, many different sets of genes are identified by the resampling approach, with a very small frequency of genes occurrence among the signatures. Also, the smaller the training sample size, the lower is the stability of the signatures. CONCLUSION: Random survival forest and boosting approach give good predictive performance but gene signatures are very unstable. Further works are needed to propose adequate strategies for the analysis of high-dimensional data in the context of competing risks.