The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences

BackgroundAge-related muscle dysfunction is a major contributor to disability, frailty, and poor clinical outcomes in older adults. Muscle mass and strength provide limited insight into the multifactorial nature of muscular decline. Skeletal Muscle Function Deficit (SMFD) framework integrates multiple domains muscle mass, quality, strength, and power to capture a broader spectrum of age-related muscle dysfunction. ObjectiveTo develop and validate a composite SMFD score and evaluate its association with key geriatric outcomes in older adults. MethodsThis study used data from the InCHIANTI longitudinal cohort (1998-2018), including 1,035 participants and 3,196 total assessments. The SMFD score (range 0-20) was computed by assigning quintile-based values of muscle area, density, strength, and lower limb power. Associations with disability in basic and instrumental activities of daily living (BADL/IADL), frailty phenotype, poor physical performance (SPPB <7), hospitalization, falls number, and major chronic diseases were analyzed using mixed-effects models, adjusting for age, sex, fat area, and multimorbidity. ResultsThe SMFD score declined significantly over time and was independently associated with lower risk of BADL (OR 0.57), IADL (OR 0.70), frailty (OR 0.72), poor performance (OR 0.68), hospitalization (OR 0.96), and falls number (OR 0.96). Higher SMFD scores were also inversely associated with the prevalence and incidence of Parkinsons disease, stroke, and hip osteoarthritis. ConclusionsThe SMFD score is a valid, multidimensional measure that predicts adverse outcomes in older adults more effectively than traditional sarcopenia, dynapenia, and powerpenia. It holds promise for use in clinical assessment, risk stratification, and targeted interventions.

BackgroundFrailty and chronic inflammatory diseases are known risk factors for adverse outcomes and have been associated with epigenetic age acceleration (EAA), a quantitative estimation of biological age based on DNA methylation. We investigated the interrelationships between EAA, atherogenic dyslipidaemia and frailty; determined the best performing epigenetic clock for EAA estimation of atherosclerotic cardiovascular disease (ASCVD) risks; and prospectively analysed the capacity of EAA to predict future adverse outcomes. MethodsA structured cardiogeriatric evaluation including frailty and physical capacity assessment was performed in community-living older adults aged [&ge;]60 years who met predefined eligibility criteria and had no previous history of heart failure. We prioritised the selection of all available frail older adults from our bioresource and used computerised randomisation to select the more abundant robust and pre-frail individuals for relatively balanced analyses among groups. DNA methylation analysis was performed using the Infinium MethylationEPIC platform. Quantitative proton-nuclear magnetic resonance (NMR) was used for targeted metabolomic analysis of serum. Five common epigenetic clocks were compared for associations with frailty, systemic inflammation, lipid-metabolome and prediction of incident adverse outcomes. Clinical outcomes were queried using electronic medical record systems and by interview. FindingsAmong 535 older adults, GrimAge 2-estimated EAA (Grim2AA) performed best in classifying and predicting the frailty phenotype. Grim2AA was significantly associated with systemic inflammation indicated by GlycA, increased risk of ASCVD comorbidities, and an atherogenic lipid profile characterised by reduced low-density lipoprotein (LDL) particle size and abnormal triglycerides in lipoproteins. Small LDL particle size but not other lipid/lipoprotein features were also associated with frailty. Of the five epigenetic clocks analysed and benchmarked against the ACC/AHA ASCVD Risk Calculator, Grim2AA was the most strongly associated. For each 10-year increment in Grim2AA, the risk was estimated at 4{middle dot}92% (p=0{middle dot}0002). During a median follow-up of 4{middle dot}68 years, 55 all-cause deaths occurred and 101 individuals experienced cardiovascular hospitalisation. Analysis of cardiovascular hospitalisation revealed marked differences in EAA depending on the cause, and pointed to ASCVD (excluding stroke) as being most common with the highest median Grim2AA at 0{middle dot}73 years. Kaplan-Meier analysis did not show a difference in the rates of cardiovascular hospitalisation between Grim2AA [&ge;]0 and <0 years (p=0{middle dot}19). However, the rates of ASCVD hospitalisation (42 of 101) were significantly higher in older adults with the former (p=0{middle dot}0027). InterpretationAs the most comprehensive targeted analysis of blood lipid-metabolome and DNA methylation-based epigenetic clock to date, this study has linked frailty, inflammageing, atherogenic dyslipidaemia and ASCVD risks that can be collectively indicated by Grim2AA in older adults. Grim2AA is an independent predictor of future ASCVD hospitalisation, and may serve as a potential biomarker for monitoring disease trajectory and target for secondary prevention.

To understand the mechaisnm behind high respond (HighR) compared to low respond (LowR) to resistnace training (RT) and whey protein supplementation (20g/day), we analysied vastus laterails muscle biopsies from a total of 50 participants. Utilising the MRI muscle cross-sectional area (CSA) data, we defined responders as those who had hypertrophy exceeding the 1.7% method error. Quadriceps CSA in the lower responder (LowR) (n=25, mean age 69{+/-}5 years) and HighR (n=25, mean age 67{+/-}4 years) increased from 53.6 {+/-} 12.1 cm2 to 55.4 {+/-} 12.8 cm2 after 10 weeks of RET (3.3 {+/-} 1.7%, P < 0.001) and increased the absolute CSA in the higher responders (HighR) from 53.7 {+/-} 12.5 cm2 to 59.2 {+/-} 13.6 cm2 (10.3 {+/-} 2.0%, P < 0.001). Muscle biopsies were taken from the vastus lateralis before and after RT. We performed untargeted liquid chromatography-mass spectrometry metabolomics to investigate changes in muscle metabolic regulation. The partial least squares discriminant analysis (PLS-DA) yielded the best results using the polar extracts, achieving a 75% average correct classification rate for predicting HighR and LowR. There was no signifncat differences in metabolomic profile at the basline. Our findings revealed several metabolic pathways, including branched-chain amino acid catabolism, tryptophan metabolism (indole and kynurenine pathways), the TCA cycle, gut-derived metabolites, carnitine shuttle metabolism as prominent pathways disrupted in LowR. We provide new insights and has the potential to identify and enhance interventions targeting muscle metabolism, ultimately improving muscle mass and strength to reduce the risk of sarcopenia and frailty in older age.

ObjectiveExamine the association of ectopic adipose tissue (AT) with skeletal muscle (SM) mitochondrial bioenergetics in older adults. MethodsCross-sectional data from 829 older adults [&ge;]70 years was used. Total abdominal, subcutaneous, and visceral AT; and thigh muscle fat infiltration (MFI) was quantified by MRI. SM mitochondrial energetics were characterized using in vivo 31P-MRS (ATPmax) and ex vivo high-resolution respirometry (maximal oxidative phosphorylation (OXPHOS)). ActivPal was used to measure PA (step count). Linear regression models adjusted for covariates were applied, with sequential adjustment for BMI and PA. ResultsIndependent of BMI, total abdominal (standardized (Std.) {beta}=-0.21; R2=0.09) and visceral AT (Std. {beta}=-0.16; R2=0.09) were associated with ATPmax (p<0.01), but not after further adjustment for PA (p[&ge;]0.05). Visceral AT (Std. {beta}=-0.16; R2=0.25) and thigh MFI (Std. {beta}=-0.11; R2=0.24) were negatively associated with carbohydrate-supported maximal OXPHOS independent of BMI and PA (p<0.05). Total abdominal AT (Std. {beta}=-0.19; R2=0.24) and visceral AT (Std. {beta}=-0.17; R2=0.24) were associated with fatty acid-supported maximal OXPHOS independent of BMI and PA (p<0.05). ConclusionsSkeletal MFI and abdominal visceral, but not subcutaneous AT, are inversely associated with SM mitochondrial bioenergetics in older adults independent of BMI. Associations between ectopic AT and in vivo mitochondrial bioenergetics are attenuated by PA.

Background & AimThe current study examined cross-sectional and longitudinal associations between nocturia and frailty in a cohort of men and women aged 60 years and older, as evidence on this topic was lacking. MethodsBaseline and follow-up data from the Berlin Aging Study II (n=1671) assessed on average 7.1 (IQR 6.2-8.7) years apart were analyzed. Self-reported nocturia was dichotomized into </[&ge;] 2 micturitions per night, and frailty was assessed using the Fried Frailty Phenotype. Covariables were identified a priori based on a review of the existing literature. ResultsAt baseline, 70.2% of the participants were robust, 28.9% were pre-frail, and 0.9% were frail; 254 participants (23.6%) had self-reported nocturia. In longitudinal analyses, the prevalence and incidence of frailty at follow-up significantly increased when nocturia was present at baseline. Over the median follow-up of 7.1 years, there were 41 incident frailty cases (IR 5.15, 95% CI 3.79-7.00 per 1000 person-years). After adjusting for age, sex, morbidity burden, and baseline frailty status, baseline nocturia was associated with a 2.23-fold increased risk (95% CI 1.17-4.18) of frailty at follow-up. ConclusionNocturia is linked to a higher risk of developing frailty in older adults, both women and men.

BackgroundWhile studies have examined associations between changes in BMI and biological aging, the use of biological age estimates derived from omics other than DNA methylation data as well as nonlinearity and interactions in these associations are underexplored. ObjectiveWe aimed to investigate how BMI at ages 18 and [~]60, as well as changes in BMI from age 18 to [~]60, relate to downstream epigenetic and proteomic aging. We also examined nonlinearity and interactions in these associations. MethodsWe analyzed data from 401 Finnish participants with up to 9 self-reported or measured BMI values collected over 40 years. Olink proteomics and Illumina DNA methylation data were generated from blood samples taken at the last BMI measurement. We calculated 4 and 5 estimates of biological age from proteomic and epigenetic clocks, respectively. Changes in BMI over time were estimated using mixed-effects models. We applied generalized additive models to explore 1) nonlinearity in associations between BMI trajectories and biological aging while adjusting for chronological age and 2) smooth interactions between baseline BMI with changes in BMI and BMI at [~]60 years old. ResultsBMI at 18 and [~]60 years old and changes in BMI were associated with increased biological aging for most aging estimates. We found statistical evidence of nonlinearity for about one-third of the significant associations, mostly observed for proteomic clocks. We identified suggestive evidence for interactions between BMI at 18 years and BMI at [~]60 years in explaining variability in two proteomic clocks (p=0.07; p=0.09). ConclusionOur study illustrates the potential of proteomic clocks in obesity research and highlights that assuming linearity in associations between BMI trajectories and biological aging is a critical oversight. Associations between BMI and biological aging are likely modulated by past BMI, which warrants validation by other studies.

BackgroundMusculoskeletal pain frequently accompanies the development of mobility disability and falls in old age. To better understand this, we aimed to quantify the impact of different pain measures--recalled pain and movement-evoked pain--on 400-meter walk and stair climb time in older adults participating in the Study of Muscle, Mobility and Aging (SOMMA). MethodsIn SOMMA (N=879, age=76.3 {+/-} 5.0 years, 59% women, 84% Non-Hispanic White), participants completed usual pace 400m walk (avg=6.6 {+/-} 1.2 min.) and repeat stair climb tests (avg=26.6 {+/-} 7.2 sec.). Assessments of recalled pain included the Brief Pain Inventory short form (BPI-sf), total lower body pain (lower back, hips, knees, feet/ankles), stiffness (hip or knee), and Neuropathy Total Symptom Score (NTSS-6). Movement-evoked pain was assessed separately before and after the 400m walk and repeat stair climb tasks. Multivariable linear regression modeled the associations of pain with time to complete the tasks, reported as {beta}[95%CI] expressed per SD increment of pain measure or {beta}[95%CI] per pain categories, adjusted for age, sex, race, ethnicity, body mass index, prescription medications, and depressive symptoms. ResultsGreater degree of any pain measure was associated with longer physical performance time, though intercorrelations between recalled pain measures varied (r=0.13-0.57, p<0.05 for all). For each SD increment in lower body pain, participants had longer walk time (by 10.5 sec [6.1, 14.8]) and stair climb (by 0.6 sec [0.1, 1.1]). Compared to participants with no change in pain upon movement, walk time was longer in those with more pain upon movement (19.5 sec [10.3, 28.7]) (p<0.001) but not those with less pain upon movement; stair climb showed similar patterns. ConclusionsRecalled and movement-evoked pain measures were weakly correlated with one another but similarly associated with time to complete 400m walk and stair climb tests. Different pain assessments capture different functional domains of pain but have similar associations with physical performance in these older adults.

Frailty is a clinical syndrome in older adults characterized by heightened vulnerability to adverse outcomes, yet it remains under-assessed in routine practice due to time-intensive evaluation methods. Wearable accelerometers offer a promising approach for passive, continuous frailty monitoring. We analyzed data from 151 community-dwelling older adults ([&ge;]65 years) in the FACE Aging Study who wore hip accelerometers for seven days. Machine learning models were developed to classify baseline frailty status using an adapted frailty phenotype and predict 12-month frailty decline. Features derived from accelerometry data included sleep patterns, activity levels, sedentary behavior, and circadian rhythms. Multiple algorithms were compared, including deep neural networks (ResNet), gradient boosting machines (LightGBM, XGBoost, CatBoost), and multilayer perceptrons (MLPs). For baseline frailty classification, LightGBM achieved the highest performance with an F1 score of 0.898, precision of 0.815, perfect recall (1.000), and AUROC of 0.887. For predicting 12-month frailty decline, multilayer perceptron performed best with an F1 score of 0.812, precision of 0.891, and recall of 0.746. Key predictive features included low overall activity, high sedentary behavior, fragmented activity patterns, and poor sleep quality, which aligned with established frailty phenotypes. One week of free-living hip accelerometry data can accurately identify current frailty status and predict short-term frailty progression in older adults. These findings support the feasibility of developing automated, scalable frailty screening tools that could enable proactive interventions and transform geriatric care from reactive to preventive approaches. External validation in larger, diverse populations is needed to confirm generalizability.

BackgroundGait is a clinically relevant indicator of functional decline in aging populations. However, most studies classify older adults by chronological rather than functional age, which may obscure early impairments detectable through kinematic profiling. This study examined whether stratifying older adults by functional status using the Short Physical Performance Battery (SPPB) enhances sensitivity in detecting gait abnormalities and instability-related compensatory patterns. MethodsA total of 190 adults completed gait trials on a pressure-sensitive walkway. Twenty-eight spatial, temporal, and variability-based gait parameters were derived. Participants were categorized as young adults or older adults, who were further stratified into high- and low-functioning groups based on SPPB scores. Analysis of covariance (ANCOVA) was performed, adjusting for habitual walking speed to isolate functional effects. FindingsAfter adjusting for speed, the low-functioning group demonstrated longer stance and double-support durations, wider step width, and greater step-to-step variability in both spatial and temporal domains compared with both the high-functioning and young reference groups. These findings indicate a compensatory, instability-driven control strategy that challenges the assumption of a "slower but steady" gait in aging. High-functioning older adults exhibited gait patterns more closely resembling those of younger adults. InterpretationFunctional classification using the SPPB provided greater sensitivity than chronological age in detecting early mobility decline. Gait variability emerged as a salient biomarker of impaired neuromuscular control. Integrating quantitative gait profiling with validated functional assessments may improve early screening, targeted intervention, and fall prevention strategies.

OBJECTIVEEmerging evidence shows that perceived fatigability--the quantification of vulnerability to fatigue in relation to specific intensity and duration of activities--may be associated with cognitive function. We sought to quantify associations with multiple domains of cognitive function and the role of physical activity (PA). METHODSSOMMA participants completed the Pittsburgh Fatigability Scale (PFS) Physical and Mental subscales (each range 0-50; higher scores=greater fatigability) and three cognitive function assessments [Digit Symbol Substitution Test (DSST), executive function; Montreal Cognitive Assessment (MoCA), general function; and California Verbal Learning Test (CVLT), memory]. Linear regression quantified associations cross-sectionally between each PFS subscale and cognitive assessment scores adjusting for covariates. Effect modification by volume and intensity of accelerometer-measured PA was assessed. RESULTSIn 873 participants (59.2% women; age 76.3{+/-}5.0; 85% White), mean PFS Physical, Mental, and DSST scores were 15.8{+/-}8.7, 7.7{+/-}7.8, and 55.4{+/-}13.7. After adjustments, for each 4-point higher PFS Physical and 3-point higher PFS Mental, participants had nearly one fewer correct DSST items [{beta} coefficient and 95% confidence interval for PFS Physical: -0.69 (-1.09, - 0.29); PFS Mental: -0.64 (-0.97, -0.30)]. Volume and intensity of PA modified the association of PFS Mental and DSST (Pinteractions<0.01). All associations were strongest in those with the lowest volume and intensity of PA. PFS was not associated with MoCA or CVLT. DISCUSSIONGreater perceived fatigability may be associated with poorer executive function, but not memory. Individuals with greater perceived fatigability, particularly those less active, might benefit from interventions that reduce fatigability and may beneficially influence cognitive function.

Background and AimsObesity shortens life expectancy, yet its prognostic value in older adults remains unclear due to the obesity paradox and limitations of body mass index (BMI) in capturing visceral fat. We compared eight obesity indices and lifelong weight changes for mortality prediction and tested whether epigenetic age acceleration (EAA) mediates these associations in a US cohort. MethodsIn 2,222 NHANES (1999 - 2002) adults aged [&ge;]50 years, we calculated BMI, waist circumference, waist-to-height ratio, weight-adjusted-waist index (WWI), body roundness index, relative fat mass, conicity index (CI), and 10-year/long-term weight change. EAA was derived from five clocks (Horvath, Hannum, PhenoAge, GrimAge, GrimAge2). Cox regression, restricted cubic splines, and bootstrap mediation assessed hazard ratios (HRs), dose-response curves, and indirect effects. ResultsWWI and CI outperformed other indices. Highest quartiles raised all-cause mortality by 91% (HR 1.91, 95% CI 1.36-2.68) and 56% (HR 1.56, 95% CI 1.17-2.08), respectively. Each SD increase in WWI/CI was associated with higher GrimAge and GrimAge2 acceleration. Conversely, 10-year and long-term weight gain reduced mortality risk. Additionally, EAA was lowest with stable or mildly increased weight. Mediation analysis confirmed that EAA significantly mediates the association of both WWI/CI with mortality risk, as well as the protective effect of weight stability. ConclusionsNovel obesity indices (WWI, CI) are superior mortality predictors in older adults. Epigenetic ageing partly explains both the hazard of central adiposity and the survival benefit of weight homeostasis, supporting age-stratified obesity metrics and weight-stability targets. Structured Graphical Abstract Key QuestionDoes the type of obesity and historical weight fluctuations in middle-aged and elderly populations influence the acceleration of epigenetic aging and the associated risk of mortality? Additionally, could epigenetic aging acceleration serve as a potential mechanism linking obesity and weight changes to mortality? Key FindingIn middle-aged and older adults, WWI and CI, unaffected by the obesity paradox, exhibit superior predictive power for all-cause and cause-specific mortality compared to traditional obesity metrics, with EAA serving as the key mechanistic link. Furthermore, weight stability or mild weight gain sustained over 10 years or more substantially mitigates diverse mortality risks in this population by attenuating EAA progression. Take Home MessageThis study underscores the need to move beyond traditional obesity metrics such as BMI and body weight by incorporating newer indices like WWI and CI, which more accurately capture fat distribution and visceral adiposity. This integrated approach improves the identification of high-risk populations and helps reduce obesity-related mortality. Furthermore, it calls for a shift in weight management goals among older adults--from weight loss to weight stability--supporting the development of age-specific clinical guidelines. Importantly, given that EAA is a key mediator in the obesity-mortality relationship, slowing its progression may disrupt pathways leading to death, establishing EAA as a measurable biomarker and a potential target for interventions aimed at extending longevity in middle-aged and older adults. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=132 SRC="FIGDIR/small/25339825v1_ufig1.gif" ALT="Figure 1"> View larger version (72K): org.highwire.dtl.DTLVardef@15eb5aorg.highwire.dtl.DTLVardef@10e92feorg.highwire.dtl.DTLVardef@1d48209org.highwire.dtl.DTLVardef@a548e4_HPS_FORMAT_FIGEXP M_FIG Graphic Abstract C_FIG

BackgroundBody composition strongly influences clinical outcomes in older adults, yet body mass index (BMI) lacks discriminatory power, and standard tools such as bioelectrical impedance analysis (BIA), dual-energy X-ray absorptiometry are not routinely accessible. Deep learning enables scalable, opportunistic assessment of body composition from chest radiographs (CXRs), one of the most widely available imaging modalities. Methods and FindingsUsing the Inception-V3 architecture, we developed a deep-learning model using 107,568 paired CXR and BIA records (2016-2018). The model was temporally validated on a separate dataset of 77,655 records (2014-2015). Our model predicted skeletal muscle mass (SMM) and fat mass (FM) with high accuracy (SMM: Pearson r = 0.967, MAE 1.40 kg; FM: r = 0.924, MAE 1.61 kg). In a cohort of 5,932 older adults (aged [&ge;]65years), a 1-SD increase in CXR-predicted skeletal muscle index (SMI) was associated with a significant reduction in 10-year all-cause mortality (Hazard Ratio [HR] 0.65 [95% CI 0.58-0.73] for men; 0.80 [0.67-0.97] for women). In an external validation of 925 geriatric clinic patients, predicted SMI also showed comparable associations with geriatric parameters, including lower odds of sarcopenia (per 1 SD increase: 0.29 [0.22-0.38] for men; 0.25 [0.18-0.34] for women) and frailty (0.62 [0.48-0.78] for men; 1.00 [0.81-1.23] for women). These associations were more robust than those of BMI. Key limitations include the retrospective, single-center design and the use of a relatively healthy screening population. ConclusionA deep learning model applied to routine CXRs enables accurate estimation of skeletal muscle and fat mass, demonstrating prognostic and functional relevance comparable to BIA measurements. This approach may serve as a practical, low-cost tool for risk stratification and long-term care planning, particularly in older adults.

BackgroundBones and muscles are connected anatomically, and functionally. Preliminary evidence has shown the gut microbiome influences the aging process of bone and muscle in animal studies. However, such evidence in humans is still scarce. This study aimed to assess the microbiome-bone and microbiome-muscle associations in two cohorts of community-dwelling older adults. MethodsWe leveraged information from two large population-based cohorts, i.e., the Rotterdam Study (mean age 62.7 {+/-} 5.6 years; n=1,249) and the Framingham Heart Study (mean age 55.2 {+/-} 9.1 years; n=1,227). For individuals included in this study, gut microbiome 16S rRNA sequencing, musculoskeletal phenotyping derived from DXA images, lifestyle and socioeconomic data, and medication records were available. Per cohort, the 16S rRNA sequencing data, derived from stool, were processed with the DADA2 pipeline and taxonomies were assigned using the SILVA reference database. In addition, the microbiome functional potential was obtained with PICRUSt2. Further, we investigated the association between the human gut microbiome (alpha diversity, genera and predicted functional pathways) and appendicular lean mass (ALM), femoral neck bone mineral density (FN-BMD) and trabecular bone score (TBS) using multilinear regression models controlling for multiple confounders, and performed a joint analysis from both cohorts. Sex-stratified analyses were also conducted. ResultsThe gut microbiome alpha diversity was not associated with either tested phenotype after accounting for multiple-testing (P>1.67e-02). In the joint analysis, lower abundance of Oscillibacter (beta= -.51, 95%CI [-0.74, -.29]), Anaerotruncus (beta=-0.41, 95%CI [-0.61, - 0.21]), Eisenbergiella (beta=-0.39, 95%CI [-0.59, -.19]) and higher abundance of Agathobacter (beta=0.40, 95%CI [0.20, 0.60]) were associated with higher ALM (P<2.0e-04). Lower abundance of Anaerotruncus (beta=-0.32, 95%CI [-0.45, -.19]), Hungatella (beta=-0.26, 95%CI [-0.38, -.15]) and Clostridiales bacterium DTU089 (beta=-0.37, 95%CI [-0.55, -.19]) was associated with higher ALM only in females (P< 2.0e-04). Moreover, the biotin biosynthesis II pathway was positively associated with ALM (beta=0.44, 95% CI [0.24, 0.64]) (P<1.90e-04) in females while no associations were observed in males. We did not observe any robust association of bone traits with gut microbiome features. ConclusionOur results indicate that specific genera are associated with ALM in middle-aged and older adults and these associations can present in a sex-specific manner. Overall, our study suggests that the gut microbiome is linked to muscle aging in middle-aged and older adults. However, larger sample sizes are still needed to underpin the specific microbiome features involved.

ObjectiveSome evidence suggests that higher serum TSH may be a part of normal aging, but current studies are limited to 13-year follow-up. We examined TSH changes during 22 years of follow-up across the adult lifespan. DesignLongitudinal analyses of the population-based HUNT Study in Norway, with TSH measurements from 1995-97, 2006-08 and 2017-19. MethodsIn the overall population and in individuals without thyroid medication or disease, we estimated i) geometric mean serum TSH by age, integrating cross-sectional and longitudinal measurements using linear mixed models, ii) percentiles of the TSH distribution by age, and iii) within-individual TSH change during follow-up, expressed by geometric mean ratios (GMR) reflecting the fold change in geometric mean TSH. ResultsWe included 136,925 TSH measurements among 84,342 participants, of whom 40,615 had [&ge;]2 measurements and 13,613 had [~]22-year follow-up. Mean TSH was higher at older age in men, but weaker and less consistent in women. The TSH distribution widened at older age in men and women. Among individuals without thyroid medication or disease, mean TSH increased modestly by 0.13 mIU/L (GMR 1.09; 95%CI 1.08,1.11) during 22-year follow-up in men, but not in women (GMR 0.99; 95%CI 0.98,1.00). This increase was stronger at 0.5 mIU/L in men aged [&ge;]70 years at baseline (GMR 1.32; 95%CI 1.18,1.48). ConclusionsMean serum TSH increased with age in older men, but showed only modest or no age-related change in younger men and in women. The wider TSH distribution at older age supports the need for age-specific TSH reference ranges. Significance statementPrevious evidence of higher TSH concentrations at older age comes from cross-sectional studies and longitudinal studies with up to 13-year follow-up. We utilized a large population-based study to extend the evidence on within-individual TSH changes to [~]22-year follow-up across the adult lifespan. Mean TSH increased with age in men, modestly at 0.13 mIU/L overall, but stronger at 0.5 mIU/L in men followed from their 70s to their 90s. In women, mean serum TSH appeared stable during follow-up, but more frequent thyroid hormone supplementation may have skewed the TSH distribution away from higher, but still physiological levels. The TSH distribution widened at older age in both men and women, supporting the need for age-specific TSH reference ranges.

BackgroundGait adaptability, defined as the ability to adjust walking performance to environmental challenges, likely reflects complex interactions among the central nervous system (CNS) and other physiological systems, however, the drivers of lower gait adaptability in older adults are poorly understood. MethodsWe applied a Bayesian network framework to quantify multisystem interactions contributing to percent change in gait speed (%GSC) on transition from even to uneven surface in 159 older adults (63% women). Neuroimaging measures include total gray matter and white matter hyperintensities, striatal dopaminergic neurotransmission, and resting state functional connectivity. Other measures were obtained for domains important for locomotor control: health history, lifestyle, psychological well-being, cognition, and musculoskeletal and peripheral nervous systems (neurological exam). The Bayesian network estimated direct and indirect dependencies among variables, and predictive accuracy of %GSC from the Bayesian network was compared with that of multivariable linear regression using 10-fold cross-validation. ResultsParticipants exhibited slower gait on uneven compared to even surfaces (mean %GSC = -6.32%). The Bayesian network outperformed linear regression in predicting %GSC and identified four direct paths to %GSC from: BMI, muscle strength, striato-cortical sensorimotor connectivity, and purpose in life. Indirect paths to %GSC showed interrelations among CNS and non-CNS variables, including striatal dopaminergic neurotransmission, total gray matter volume, medications, proprioception, and sex. ConclusionsGait adaptability in older adults is influenced by interactions among functional connectivity, body composition, muscle strength, and psychological well-being. Strengthening both neural and physical systems through targeted interventions may mitigate declines in gait instability and preserve mobility with aging.

BackgroundIndividuals with Mild Cognitive Impairment (MCI) have more gait variability under dual-task conditions than cognitively healthy adults. However, characteristics associated with this susceptibility of gait to dual-task stress are unknown. MethodsTesting was performed at baseline in the Study of Mental And Resistance Training (SMART). Ninety-three adults with MCI (age 70{+/-}6.8 years; 66.6% female) performed a single- and dual-task walk (cognitive distractor=letter fluency), in random order. Linear and non-linear gait variability were measured using force-sensitive insoles. Cognitive performance during dual-tasking was assessed by the number of correct words vocalized. Cognitive function, brain Magnetic Resonance Imaging (MRI), muscle strength, aerobic capacity, body composition, physical and psychosocial function were also assessed as potential correlates of gait dynamics. ResultsGait dynamics worsened during dual-tasking, with decrements in both stride time variability (p<0.001) and detrended fluctuation analysis (DFA) (p=0.001). Lower aerobic capacity and thinner posterior cingulate cortex were associated with greater decrements in DFA (p<0.05). Smaller hippocampal volume, worse psychological well-being and poorer static balance were associated with greater decrements in stride time variability (p<0.05). By contrast, cognitive performance did not change under dual-task conditions compared to seated testing (p=0.13). ConclusionsUnder dual-task conditions, participants with MCI preserved their cognitive performance at the expense of gait stability. Decrements in dual-tasking gait were associated with lower aerobic fitness, balance, psychological well-being, and brain volume in cognitively-relevant areas of the posterior cingulate and hippocampus, all potentially modifiable characteristics. Trials of targeted interventions are needed to determine the potential plasticity of gait variability in high-risk cohorts.

BackgroundLimitations in activities of daily living have widespread implications for the well-being of older adults. However, the relation between performance-based physical function and self-reported functional impairment is inconsistent. MethodsThe cohort included 6,282 White women and 310 Black women aged 65 and older participating in the Study of Osteoporotic Fractures (SOF) from 1986 to 2010 who reported no limitations in any Instrumental Activities of Daily Living (IADL) at baseline. Approximately every two to six years, participants self-reported their physical limitations and trained interviewers assessed common measures of physical performance (i.e., usual gait speed, grip strength, and chair stand time). We used Cox proportional hazards models using age as the time scale to calculate hazard ratios between individual and summary measures of physical performance and incident IADL limitations. ResultsOver follow-up, 4,193 White women and 118 Black women developed IADL impairment (IR = 451.34 and 361.52 per 10,000 person-years, respectfully). Usual gait speed was associated with IADL limitations in both race cohorts (slowest gait vs. fastest gait HR: 3.83, 95% CI: 3.41 - 4.31; HR: 2.59, 95% CI: 1.42 - 4.73). For every one-point increase in summary performance score, rate of IADL limitations was lower for both White women and Black women (HR: 0.79, 95% CI: 0.78-0.80; HR: 0.87, 95% CI: 0.81 - 0.94). ConclusionIn this longitudinal study, women with poorer performance in individual and summary measures of physical function had an increased rate of incident IADL limitations over follow-up compared to women with the best performance. These findings confirm previous research using cross-sectional data.

BackgroundVestibular complaints are common in older adults and are linked to imbalance and falls. Some older adults show impaired vestibular perception despite preserved peripheral-reflex ("vestibular agnosia"). Yet it remains unclear if vestibular agnosia is independently linked to imbalance and falls in otherwise healthy older adults. We therefore investigated the prevalence of vestibular agnosia in community-dwelling older adults, and examined its association to balance and prospective falls. MethodsVestibular perceptual thresholds were measured during yaw-plane rotational chair testing. Postural sway and instrumented Timed-Up-and-Go were assessed using wearable sensors, and falls were recorded prospectively over six-month. Vestibular agnosia was identified using K-means clustering. Multivariable regressions examined associations between perceptual thresholds and balance outcomes; logistic and negative binomial regressions evaluated associations with prospective falls. ResultsAmong 166 participants (75.4 years; 81.9% female), 18.7% were classified as having vestibular agnosia. These individuals had worse cognition and somatosensation. Elevated (i.e. worse) vestibular perceptual thresholds were independently associated with greater sway velocity when standing on foam with eyes-open (adjusted {beta}=0.002, p=0.03). Associations with other balance outcomes were attenuated after adjustment. Vestibular perceptual thresholds were not associated with prospective falls (odds of [&ge;]1 fall: adjusted OR=0.99, p=0.65; fall counts: adjusted IRR=1.02, p=0.35). ConclusionsApproximately one-fifth of healthy older adults exhibit vestibular agnosia. While elevated perceptual thresholds are independently associated with poorer balance, they did not predict falls. Vestibular perceptual testing provides complementary insight into age-related balance impairment, although its utility in fall-risk prediction requires further investigation. Key PointsO_LIApproximately one-fifth of healthy older adults had vestibular agnosia (impaired vestibular perception despite intact peripheral function) C_LIO_LIOlder adults with vestibular agnosia have poorer cognition, reduced lower limb somatosensation, and higher anxiety. C_LIO_LIHigher (i.e. worse) vestibular perceptual thresholds were independently associated with greater sway velocity when standing on foam (eyes open). C_LIO_LIHigher vestibular perceptual thresholds were only associated with slower TUG performance and greater eyes-closed foam sway in unadjusted models. C_LIO_LIVestibular perceptual thresholds did not predict prospective falls over 6 months. C_LI

The New England Centenarian Study (NECS) provides a unique resource for the study of extreme human longevity (EL). To gain insight into biological pathways related to EL, chronological age and survival, we used an untargeted serum metabolomic approach (> 1,400 metabolites) in 213 NECS participants, followed by integration of our findings with metabolomic data from four additional studies. Compared to their offspring and matched controls, EL individuals exhibited a distinct metabolic profile characterized by higher levels of primary and secondary bile acids - most notably chenodeoxycholic acid (CDCA) and lithocholic acid (LCA) - higher levels of biliverdin and bilirubin, and stable levels of selected steroids. Notably, elevated levels of both bile acids and steroids were associated with lower mortality. Several metabolites associated with age and survival were inversely associated with metabolite ratios related to NAD+ production and/or levels (tryptophan/kynurenine, cortisone/cortisol), gut bacterial metabolism (ergothioneine/ trimethylamine N-oxide, aspartate/quinolinate), and oxidative stress (methionine/methionine sulfoxide), implicating these pathways in aging and/or longevity. We further developed a metabolomic clock predictive of biological age, with age deviations significantly associated with mortality risk. Key metabolites predictive of biological aging, such as taurine and citrate, were not captured by traditional age analyses, pointing to their potential role as biomarkers for healthy aging. These results highlight metabolic pathways that may be targeted to promote metabolic resilience and healthy aging.

BackgroundPregnancy involves substantial physiological, metabolic, and immunologic adaptation, which may alter trajectories of maternal biological aging. While emerging evidence suggests that pregnancy may transiently accelerate biological aging followed by partial postpartum recovery, longitudinal studies capturing these dynamics, particularly across successive pregnancies, are limited. This study examined changes in maternal biological aging over nearly four years postpartum and assessed whether subsequent pregnancies disrupted recovery trajectories. MethodsParticipants (N = 130; aged 18 to 41 years; 23% non-Hispanic Black, 61% non-Hispanic White) were followed longitudinally across pregnancy and nearly four years postpartum. Biological aging was assessed in saliva at up to three timepoints using four biomarkers: absolute telomere length via qPCR, epigenetic clocks (GrimAge2 and PhenoAge), and pace-of-aging (DunedinPACE). Generalized additive mixed models were used to estimate nonlinear change in biological aging across time, with spline terms differentiating early postpartum recovery and later changes moderated by subsequent pregnancy. ResultsIn non-interaction models, telomere length was stable in the early ([~] 9 months) postpartum period (b=0.49, SE=0.54, P=.36), while there was a trend towards deceleration in GrimAge2 (b = -2.01, SE = 1.06, P=.06) and a significant deceleration in pace of aging (DunedinPACE; b = -0.24, SE = .04, P<.001). In later postpartum ([~]43 months), telomere length declined significantly (b = -0.81, SE = .37, P=.029), while both GrimAge2 and pace of aging stabilized. A subsequent pregnancy in the later postpartum period was independently associated with shorter telomere length (b = -0.78, SE = .36, P=.032), but not with epigenetic clocks. Time-by-late subsequent pregnancy status interaction models revealed that the acceleration in aging markers during the later postpartum period was more pronounced among women who became pregnant again, particularly for DunedinPACE, where a significant interaction (b = 0.20, SE = .07, P=.006) suggested that subsequent pregnancy disrupted the slowed pace of aging observed postpartum. Interaction terms for Time-by-late subsequent pregnancy predicting telomere length and GrimAge2 were directionally consistent with this pattern of slowed recovery but did not reach statistical significance. Associations for PhenoAge were consistent in direction with the other aging indices but did not reach significance in either non-interaction or interaction models. ConclusionPregnancy may function as a biological stressor that transiently accelerates maternal aging, while the postpartum period offers a potential window for recovery. However, subsequent pregnancies may disrupt this recovery process, compounding biological aging over time. These findings underscore the importance of postpartum recovery and interpregnancy intervals in shaping maternal aging trajectories and warrant further investigation in larger, more diverse samples with additional metabolic covariates.