Dataset Information

A robust machine learning framework to identify signatures for frailty: a nested case-control study in four aging European cohorts.

ABSTRACT: Phenotype-specific omic expression patterns in people with frailty could provide invaluable insight into the underlying multi-systemic pathological processes and targets for intervention. Classical approaches to frailty have not considered the potential for different frailty phenotypes. We characterized associations between frailty (with/without disability) and sets of omic factors (genomic, proteomic, and metabolomic) plus markers measured in routine geriatric care. This study was a prevalent case control using stored biospecimens (urine, whole blood, cells, plasma, and serum) from 1522 individuals (identified as robust (R), pre-frail (P), or frail (F)] from the Toledo Study of Healthy Aging (R=178/P=184/F=109), 3 City Bordeaux (111/269/100), Aging Multidisciplinary Investigation (157/79/54) and InCHIANTI (106/98/77) cohorts. The analysis included over 35,000 omic and routine laboratory variables from robust and frail or pre-frail (with/without disability) individuals using a machine learning framework. We identified three protective biomarkers, vitamin D3 (OR: 0.81 [95% CI: 0.68-0.98]), lutein zeaxanthin (OR: 0.82 [95% CI: 0.70-0.97]), and miRNA125b-5p (OR: 0.73, [95% CI: 0.56-0.97]) and one risk biomarker, cardiac troponin T (OR: 1.25 [95% CI: 1.23-1.27]). Excluding individuals with a disability, one protective biomarker was identified, miR125b-5p (OR: 0.85, [95% CI: 0.81-0.88]). Three risks of frailty biomarkers were detected: pro-BNP (OR: 1.47 [95% CI: 1.27-1.7]), cardiac troponin T (OR: 1.29 [95% CI: 1.21-1.38]), and sRAGE (OR: 1.26 [95% CI: 1.01-1.57]). Three key frailty biomarkers demonstrated a statistical association with frailty (oxidative stress, vitamin D, and cardiovascular system) with relationship patterns differing depending on the presence or absence of a disability.

SUBMITTER: Gomez-Cabrero D

PROVIDER: S-EPMC8190217 | biostudies-literature | 2021 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A robust machine learning framework to identify signatures for frailty: a nested case-control study in four aging European cohorts.

Gomez-Cabrero David D Walter Stefan S Abugessaisa Imad I Miñambres-Herraiz Rebeca R Palomares Lucia Bernad LB Butcher Lee L Erusalimsky Jorge D JD Garcia-Garcia Francisco Jose FJ Carnicero José J Hardman Timothy C TC Mischak Harald H Zürbig Petra P Hackl Matthias M Grillari Johannes J Fiorillo Edoardo E Cucca Francesco F Cesari Matteo M Carrie Isabelle I Colpo Marco M Bandinelli Stefania S Feart Catherine C Peres Karine K Dartigues Jean-François JF Helmer Catherine C Viña José J Olaso Gloria G García-Palmero Irene I Martínez Jorge García JG Jansen-Dürr Pidder P Grune Tilman T Weber Daniela D Lippi Giuseppe G Bonaguri Chiara C Sinclair Alan J AJ Tegner Jesper J Rodriguez-Mañas Leocadio L

GeroScience 20210218 3

Phenotype-specific omic expression patterns in people with frailty could provide invaluable insight into the underlying multi-systemic pathological processes and targets for intervention. Classical approaches to frailty have not considered the potential for different frailty phenotypes. We characterized associations between frailty (with/without disability) and sets of omic factors (genomic, proteomic, and metabolomic) plus markers measured in routine geriatric care. This study was a prevalent c ...[more]

PMID: 33599920

Similar Datasets

Project description:BackgroundFrailty indicators can operate in dynamic amalgamations of disease conditions, clinical symptoms, biomarkers, medical signals, cognitive characteristics, and even health beliefs and practices. This study is the first to evaluate which, among these multiple frailty-related indicators, are important and differential predictors of clinical cohorts that represent progression along an Alzheimer's disease (AD) spectrum. We applied machine-learning technology to such indicators in order to identify the leading predictors of three AD spectrum cohorts; viz., subjective cognitive impairment (SCI), mild cognitive impairment (MCI), and AD. The common benchmark was a cohort of cognitively unimpaired (CU) older adults.MethodsThe four cohorts were from the cross-sectional Comprehensive Assessment of Neurodegeneration and Dementia dataset. We used random forest analysis (Python 3.7) to simultaneously test the relative importance of 83 multi-modal frailty indicators in discriminating the cohorts. We performed an explainable artificial intelligence method (Tree Shapley Additive exPlanation values) for deep interpretation of prediction effects.ResultsWe observed strong concurrent prediction results, with clusters varying across cohorts. The SCI model demonstrated excellent prediction accuracy (AUC = 0.89). Three leading predictors were poorer quality of life ([QoL]; memory), abnormal lymphocyte count, and abnormal neutrophil count. The MCI model demonstrated a similarly high AUC (0.88). Five leading predictors were poorer QoL (memory, leisure), male sex, abnormal lymphocyte count, and poorer self-rated eyesight. The AD model demonstrated outstanding prediction accuracy (AUC = 0.98). Ten leading predictors were poorer QoL (memory), reduced olfaction, male sex, increased dependence in activities of daily living (n = 6), and poorer visual contrast.ConclusionsBoth convergent and cohort-specific frailty factors discriminated the AD spectrum cohorts. Convergence was observed as all cohorts were marked by lower quality of life (memory), supporting recent research and clinical attention to subjective experiences of memory aging and their potentially broad ramifications. Diversity was displayed in that, of the 14 leading predictors extracted across models, 11 were selectively sensitive to one cohort. A morbidity intensity trend was indicated by an increasing number and diversity of predictors corresponding to clinical severity, especially in AD. Knowledge of differential deficit predictors across AD clinical cohorts may promote precision interventions.

Project description:IntroductionGlobally, schistosomes infect approximately 200 million people, with 90% of infections in sub-Saharan Africa. Schistosomiasis is hypothesized to increase HIV-1 acquisition risk, and multiple cross-sectional studies reported strong associations. We evaluated this hypothesis within four large prospective cohorts.MethodsWe conducted nested case-control analyses within three longitudinal cohorts of heterosexual HIV-1 serodiscordant couples and one female sex worker (FSW) cohort from Kenya and Uganda. The serodiscordant couples studies were conducted between 2004 and 2012 while the FSW cohort analysis included participant follow-up from 1993 to 2014. Cases HIV-1 seroconverted during prospective follow-up; three controls were selected per case. The presence of circulating anodic antigen in archived serum, collected prior to HIV-1 seroconversion, identified participants with active schistosomiasis; immunoblots determined the schistosome species. Data from serodiscordant couples cohorts were pooled, while the FSW cohort was analysed separately to permit appropriate confounder adjustment.ResultsWe included 245 HIV-1 seroconverters and 713 controls from the serodiscordant couples cohorts and 330 HIV-1 seroconverters and 962 controls from the FSW cohort. The prevalence of active schistosomiasis was 20% among serodiscordant couples and 22% among FSWs. We found no association between schistosomiasis and HIV-1 acquisition risk among males (adjusted odds ratio (aOR) = 0.99, 95% CI 0.59 to 1.67) or females (aOR = 1.21, 95% CI 0.64 to 2.30) in serodiscordant couples. Similarly, in the FSW cohort we detected no association (adjusted incidence rate ratio (aIRR) = 1.11, 95% CI 0.83 to 1.50). Exploring schistosome species-specific effects, there was no statistically significant association between HIV-1 acquisition risk and Schistosoma mansoni (serodiscordant couples: aOR = 0.90, 95% CI 0.56 to 1.44; FSW: aIRR = 0.83, 95% CI 0.53 to 1.20) or Schistosoma haematobium (serodiscordant couples: aOR = 1.06, 95% CI 0.46 to 2.40; FSW: aIRR = 1.64, 95% CI 0.93 to 2.87) infection.ConclusionsSchistosomiasis was not a strong risk factor for HIV-1 acquisition in these four prospective studies. S. mansoni was responsible for the majority of schistosomiasis in these cohorts, and our results do not support the hypothesis that S. mansoni infection is associated with increased HIV-1 acquisition risk. S. haematobium infection was associated with a point estimate of elevated HIV-1 risk in the FSW cohort that was not statistically significant, and there was no trend towards a positive association in the serodiscordant couples cohorts.

Project description:BackgroundDietary protein may slow the decline in muscle mass and function with aging, making it a sensible candidate to prevent or modulate disability progression. At present, studies providing reliable estimates of the association between protein intake and physical function, and its interaction with physical activity (PA), in community-dwelling older adults are lacking.ObjectivesWe investigated the longitudinal relation between protein intake and physical function, and the interaction with PA.MethodsWe undertook a pooled analysis of individual participant data from cohorts in the PROMISS (PRevention Of Malnutrition In Senior Subjects in the European Union) consortium (the Health Aging and Body Composition Study, Quebec Longitudinal Study on Nutrition and Successful Aging, Longitudinal Aging Study Amsterdam, and Newcastle 85+) in which 5725 community-dwelling older adults were followed up to 8.5 y. The relation between protein intake and walking speed was determined using joint models (linear mixed-effects and Cox proportional hazards models) and the relation with mobility limitation was investigated using multistate models.ResultsHigher protein intake was modestly protective of decline in walking speed in a dose-dependent manner [e.g., protein intake ≥1.2 compared with 0.8 g/kg adjusted body weight (aBW)/d: β = 0.024, 95% CI: 0.009, 0.032 SD/y], with no clear indication of interaction with PA. Participants with protein intake ≥0.8 g/kg aBW/d had also a lower likelihood of incident mobility limitation, which was observed for each level of PA. This association seemed to be dose-dependent for difficulty walking but not for difficulty climbing stairs. No associations between protein intake and other mobility limitations transitions were observed.ConclusionsHigher daily protein intake can reduce physical function decline not only in older adults with protein intake below the current RDA of 0.8 g/kg BW/d, but also in those with a protein intake that is already considered sufficient. This dose-dependent association was observed for each level of PA, suggesting no clear synergistic association between protein intake and PA in relation to physical function.

Project description:PurposeA significant hurdle in developing reliable gene expression-based prognostic models has been the limited sample size, which can cause overfitting and false discovery. Combining data from multiple studies can enhance statistical power and reduce spurious findings, but how to address the biologic heterogeneity across different datasets remains a major challenge. Better meta-survival analysis approaches are needed.Material and methodsWe presented a decentralized learning framework for meta-survival analysis without the need for data aggregation. Our method consisted of a series of proposals that together alleviated the influence of data heterogeneity and improved the performance of survival prediction. First, we transformed the gene expression profile of every sample into normalized percentile ranks to obtain platform-agnostic features. Second, we used Stouffer's meta-z approach in combination with Harrell's concordance index to prioritize and select genes to be included in the model. Third, we used survival discordance as a scale-independent model loss function. Instead of generating a merged dataset and training the model therein, we avoided comparing patients across datasets and individually evaluated the loss function on each dataset. Finally, we optimized the model by minimizing the joint loss function.ResultsThrough comprehensive evaluation on 31 public microarray datasets containing 6,724 samples of several cancer types, we demonstrated that the proposed method has outperformed (1) single prognostic genes identified using conventional meta-analysis, (2) multigene signatures trained on single datasets, (3) multigene signatures trained on merged datasets as well as by other existing meta-analysis methods, and (4) clinically applicable, established multigene signatures.ConclusionThe decentralized learning approach can be used to effectively perform meta-analysis of gene expression data and to develop robust multigene prognostic signatures.

Project description:BackgroundOur understanding of the relationship between frailty and stroke, beyond the acute phase of stroke, is limited. We aimed to estimate the prevalence of frailty in stroke survivors using differing methods of assessment and describe relationships with stroke outcomes.MethodsWe used data from three international population surveys (American Health and Retirement Survey/English Longitudinal Study of Ageing/Survey for Health and Retirement in Europe) of aging. Frailty status was assessed using the Fried frailty phenotype, a 40-item frailty index (FI) and the clinical frailty scale (CFS). We created estimates of frailty prevalence and assessed association of frailty with outcomes of mortality/hospital admission/recurrent stroke at 2 years follow-up using logistic regression models adjusted for age/sex. Additional analyses explored effects of adding cognitive measures to frailty assessments and of missing grip strength data.FindingsAcross 9617 stroke survivors, using the frailty phenotype, 23.8% (n = 2094) identified as frail; with CFS, 30.1% (n = 2906) were moderately or severely frail; using FI, 22.7% (n = 2147) had moderate frailty and 31.9% (n = 3021) had severe frailty. Frailty was associated with increased risk of mortality/hospitalization/recurrent stroke using all three measures. Adding cognitive variables to the FI produced minimal difference in prevalence of frailty. People with physical frailty (phenotype or CFS) plus cognitive impairment had a greater risk of mortality than people with an equivalent level of frailty but no cognitive impairment. Excluding people unable to provide grip strength underestimated frailty prevalence.InterpretationFrailty is common in stroke and associated with poor outcomes, regardless of measure used. Adding cognitive variables to frailty phenotype/CFS measures identified those at greater risk of poor outcomes. Physical and cognitive impairments in stroke survivors do not preclude frailty assessment.

Dataset Information

A robust machine learning framework to identify signatures for frailty: a nested case-control study in four aging European cohorts.

Publications

A robust machine learning framework to identify signatures for frailty: a nested case-control study in four aging European cohorts.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets