AgingResearch.News

Sources and notes Association of Serum 25‐Hydroxyvitamin D With Frailty Across Three Prospective Cohorts: Effect Modification by Physical Activity. Food Science & Nutrition. Published paper. DOI: 10.1002/fsn3.72589 agingresearch.news/s/u6t9rk Prospective observational cohorts: China, 2038 adults aged ≥65, from 2012 to 2018; England, 6061 aged ≥50, 2012–2013 through 2016–2019; US, 5065 aged ≥50, 2016 through 2020–2022. Low/normal group sizes are not given. Low vitamin D: ≤50 nmol/L; normal: >50 nmol/L. Frailty index (FI): accumulated health deficits; new frailty: FI ≥0.25 among initially nonfrail participants. Adjustments: age, sex, education, income, marital status, smoking, drinking, diabetes, hypertension, body mass index, activity; season in China/England. China/England/US order below; parentheses give 95% confidence intervals. Low versus normal vitamin D, new-frailty hazard ratios: 1.76 (1.19–2.60), 1.27 (1.08–1.50), 1.05 (0.86–1.27). Per 10 nmol/L higher vitamin D: 0.86 (0.77–0.95), 0.95 (0.92–0.99), 0.97 (0.95–0.99). Baseline FI increases with low levels: 0.013 (0.005–0.022), 0.010 (0.004–0.015), 0.020 (0.011–0.029). FI reductions per 10 nmol/L: 0.006 (−0.008 to −0.004), 0.003 (−0.004 to −0.001), 0.002 (−0.003 to −0.001); intervals describe signed changes. Low levels also accompanied higher prevalent-frailty odds. Baseline relationships were nonlinear: overall/nonlinear p < 0.001 in China; p < 0.01 in England/US. FI plateaued in China and turned slightly upward elsewhere. New-frailty relationships showed no clear nonlinearity. Activity interactions appeared for continuous vitamin D across cohorts, with stronger new-frailty relationships among active participants; England's categorical baseline analysis found no interaction. Sensitivity hazard ratios per 10 nmol/L: - Complete cases: 0.88 (0.80–0.98), 0.95 (0.92–0.99), 0.98 (0.95–1.00). - Restricting England/US to ages ≥65, requiring repeated follow-up, and excluding early frailty over 2 years in England/US left results broadly unchanged. - Accounting for death: China 0.90 (0.82–0.99); US 0.98 (0.96–1.00). - Discrete-time models: 0.86 (0.78–0.95), 0.93 (0.89–0.97), 0.97 (0.94–0.99). - FI ≥0.21: 0.94 (0.87–1.02), 0.92 (0.89–0.95), 0.98 (0.95–1.00). Low-versus-normal analyses remained consistent. Authors' limits: - Poorer health and other participant differences could explain the relationships. - Vitamin D measures and laboratory methods differed. - Seasonal influences remained; US collection dates were unavailable. - Vitamin D was measured only at enrollment. - Frailty indicators differed across countries. - Findings exclude younger age groups.

Sources and notes Sugar rationing during the first 1000 days of life and biological aging and age-related diseases. Nature Communications; not a preprint. DOI: 10.1038/s41467-026-79639-7 agingresearch.news/s/gv2s2x The natural experiment included 54,201 UK Biobank participants and used the abrupt end of postwar sugar rationing to examine early-life exposure and later health. Comparisons concerned people exposed and unexposed to rationing, with attention to exposure duration during the first 1,000 days of life. Exposure was associated with lower biological age acceleration: a lower estimated biological age relative to age in years. The association was stronger with longer exposure and more pronounced in females. It remained consistent across groups with different genetic susceptibility. Exposure was also associated with lower risks of age-related cataract and chronic kidney disease, but not several other age-related diseases examined. The paper does not give numerical effect sizes, confidence intervals or p-values for these findings. - Other early-life differences between people exposed and unexposed to rationing might have influenced their later health.

Sources and notes Effects of Long-Term Walnut Consumption on Body Composition in the Walnuts and Healthy Aging (WAHA) Randomized Clinical Trial. Nutrients. DOI: 10.3390/nu18223779. agingresearch.news/s/jt2h5u This pre-specified secondary analysis covered Barcelona participants in a randomized trial. Of 352 randomized, 326 completed: 163 per group, including 220 women and 106 men, aged 63–79. Participants added 30–60 grams of walnuts daily, approximately 15% of energy needs, or avoided walnuts for 2 years. No food substitutions or calorie restriction were prescribed. Whole-body X-ray scans measured fat and fat-free tissue. Analyses adjusted for baseline values and changes in physical activity. Mean body composition changes were <3% in women and <5% in men. The trial detected no between-group differences in weight, BMI (weight relative to height), waist circumference, total fat or fat-free mass. Fat increased and fat-free mass decreased in both groups. Women’s abdominal-to-hip-and-thigh fat percentage ratio increased 1.8% with walnuts and decreased 0.7% with control (p = 0.009). Leptin and adiponectin, blood markers related to body fat, showed no detected differences in change, nor did their ratio. Exploratory analyses tested FTO, an obesity-linked gene, at variant rs1861868. Men with its CC version had increased weight and BMI compared with T-version carriers after walnuts; weight, BMI and waist circumference also increased compared with controls carrying CC. Their BMI rose 1.9% and waist-to-height ratio rose 1.7%, with opposite changes on control. Women carrying CC had a decreased lean-to-fat ratio and increased leptin compared with T carriers. The BMI interaction p-value was 0.023. Other subgroup comparisons found no differences in BMI or fat-percentage responses by sex, age, baseline BMI, education, diabetes, physical activity or a blood marker of walnut intake. APOE, another gene tested for differences in response, showed no major differences in body composition or fat-related blood markers. - Only healthy older adults at Barcelona were included; younger people or different regional diets could give different results. - Genetic subgroup findings were exploratory and based on small samples. - The tested FTO variant marks genetic variation; the mechanism behind its response remains unresolved. - Detailed regional fat measurements, particularly fat around internal organs, were lacking.

Sources and notes Depressive symptom dimensions, physical activity, and incident dementia: a longitudinal analysis of three international cohorts. European Archives of Psychiatry and Clinical Neuroscience. Journal article, not a preprint. DOI: 10.1007/s00406-026-02415-7 agingresearch.news/s/wnqym6 The longitudinal analysis included 27,417 participants from the China Health and Retirement Longitudinal Study, English Longitudinal Study of Ageing, and Health and Retirement Study. Depression questionnaire items were grouped into physical symptoms, depressed mood and positive feelings. Dementia required concurrent memory impairment, using education-specific thresholds, and dependence in daily activities. Statistical models adjusted for starting memory performance, and results were pooled across studies using random-effects meta-analysis. Per 1-point higher score, the pooled odds ratio was 1.43 (95% CI 1.32–1.55) for physical symptoms and 1.31 (95% CI 1.22–1.40) for depressed mood. These compare the odds of dementia at higher versus lower symptom scores. When all symptom groups were modeled together, the physical-symptom association remained in all three studies; the depressed-mood and positive-feeling associations weakened substantially. Dose–response and sensitivity analyses supported the primary findings. The associations did not clearly differ by physical activity. Statistical mediation analyses estimated the share of the association passing through physical activity at 6.7–11.0% in the English study and 5.4–8.8% in the Health and Retirement Study. Estimates in the China study were imprecise. - Dementia was defined through memory and daily-function measures, so results may differ from those based on clinical diagnoses.

Sources and notes The aging human brain exhibits reduced hemodynamic and cerebrospinal fluid flow waves during sleep. Nature Aging. Published paper, not a preprint. DOI: 10.1038/s43587-026-01270-4 agingresearch.news/s/g24p7t Age-group comparison: nighttime brain scans and electrical recordings in 31 young and 41 older adults, with one to three 25-minute runs. Daytime experiments included 31 young and 55 older adults; 20 young and 32 older participants completed both. Visual electrical-response analyses included 28 young and 54 older adults; breath-hold analyses included 21 and 39, respectively. Models included age, sleep state and sex; checks added scan site and anatomy. Fluid-wave power increased from waking to sleep by 3.82 dB in younger versus 1.15 dB in older adults (P = 0.002); dB means decibels, a measure used here for signal power. The older group’s increase was not distinguishable from chance. Flow was lower in older adults during waking and sleep, with a greater age difference during sleep (interaction P = 0.025), across every sleep stage without rapid eye movements. Blood-volume wave power increased by 3.67 dB versus 2.15 dB (P = 0.038); its age difference was also greater during sleep (interaction P = 0.021). Older adults had weaker slow brain waves and weaker coupling between brain activity, blood-volume changes and fluid flow. Some older adults had young-like flow. Visual stimulation produced similar electrical responses but smaller blood-volume and fluid responses in older adults. Breath-holding also produced smaller responses. Daytime fluid responses to visual stimulation predicted nighttime flow (P = 0.005), as did blood-vessel reactivity (P = 0.014); within-group relationships held only in older adults. Less frontal gray matter accompanied lower flow in older adults; other lobes showed no clear relationship. Anatomy and systemic rhythms did not explain lower flow; motion did not explain weaker coupling. Site adjustment did not alter the models. Anatomical adjustments preserved daytime–nighttime relationships and mediation results. Slow brain waves mediated the age–flow relationship independently and through blood-volume waves; blood-volume waves did not independently mediate it after accounting for slow brain waves. An age effect remained. - Age groups were compared rather than individuals followed through aging. - Waste clearance was not measured. - Blood volume, blood pressure, hydration, sleep deprivation history and genetic risk factors were not measured. - Cognitive impairment and neurological or sleep disorders were excluded. - Whether reduced flow precedes or follows frontal tissue loss remains unresolved.

Sources and notes Chronic kidney disease and long-term risk of cerebrovascular vulnerability. European Heart Journal Open. DOI: 10.1093/ehjopen/oeag205 agingresearch.news/s/gkjgft This population-based cohort study included 381,886 participants with kidney function, urine protein and follow-up data. It compared people with and without chronic kidney disease (CKD). Models of time until stroke examined CKD status, disease stage and kidney risk categories. CKD was associated with higher overall stroke risk: hazard ratio, a measure of relative risk, 1.61 (95% CI 1.50-1.73). For early-onset stroke, the hazard ratio was 1.89 (95% CI 1.62-2.22). Risk increased with more severe CKD, lower estimated kidney filtration, higher urine albumin relative to creatinine, and higher kidney risk categories. Kidney disease present at the start was also associated with more white matter lesions—abnormal areas in tissue connecting brain regions—smaller brain and gray matter volumes, and lower brain blood flow before a clinically apparent stroke. Blood protein analyses identified five candidate proteins associated with stroke among participants with CKD. WFDC2, a circulating protein, gave the highest AUCs, measures of how well predictions distinguish people who develop stroke from those who do not, for 3- and 5-year stroke prediction when added individually to traditional risk factors. Adding WFDC2 and HAVCR1, another blood protein, to traditional risk factors plus kidney filtration and urine protein measures increased the 3- and 5-year AUCs from 0.823 to 0.843 and from 0.792 to 0.816, respectively. Other health differences between the groups may have contributed to the stroke and brain-imaging findings.

Sources and notes Single-cell, multi-region profiling of the macaque brain across the lifespan. Cell. DOI: 10.1016/j.cell.2026.09.014. agingresearch.news/s/i92gu9 The study profiled gene activity in 5.3 million individual cells from 582 tissue samples spanning 11 brain regions of 55 rhesus macaques aged 0.5–21 years. It compared age-associated patterns across brain regions and cell types, including differences by sex. Some neuronal populations—groups of cells that send nerve signals—were less abundant at older ages. Glia, cells that support neurons, showed subtle age-associated differences in their proportions, detectable only when examining individual cells. Age-associated patterns of gene activity included both shared and differing patterns across regions and cell types. Some patterns differed by sex. The aging patterns overlapped with those seen in Alzheimer’s disease. The authors proposed shared molecular roots, with windows of vulnerability differing by age, brain region and sex. - Gene-activity profiles alone cannot show whether the overlapping patterns contribute to Alzheimer’s disease. - The atlas covered rhesus macaques, not people.

Sources and notes Food-level dietary factors reveal cardiometabolic risk hidden by diet quality scores. medRxiv preprint, not yet peer reviewed. DOI: 10.64898/2026.11.10.26367678 agingresearch.news/s/irxphh This observational analysis included 87,162 UK Biobank participants reporting intake of 148 foods and beverages, with a median 9.4 years of follow-up. A statistical method identified eight reproducible food-intake patterns. Models of time to diagnosis tested their relationships with coronary artery disease, type 2 diabetes and chronic kidney disease. Prediction models assessed what food patterns added beyond diet quality scores, and vice versa. Blood analyses covered 249 circulating metabolites, small molecules reflecting metabolism. Hazard ratios compare disease rates over follow-up; the ranges below are 95% confidence intervals. For sugar and processed foods, the hazard ratio was 1.19 for type 2 diabetes (1.11-1.27) and 1.15 for chronic kidney disease (1.07-1.23). For plant foods and fish, it was 0.90 for kidney disease (0.84-0.96); for cheese and whole grains, 0.94 for coronary artery disease (0.90-0.98). Food patterns improved disease discrimination, the ability to distinguish who develops disease, with ΔC up to +0.0049. Net reclassification improvement, a measure of improved risk classification, was 8-16%. Adding either diet score to models already containing food patterns produced negligible further improvement. The patterns were stable across geographically distinct samples. Blood profiles showed lower omega-3 fats with the sugar and processed foods pattern, and more polyunsaturated fats with the plant-based and fish pattern. - Other health or lifestyle differences between people may have contributed to their disease risk. - The analysis has not yet been peer reviewed.

Sources and notes All-cause mortality and associated factors according to age at diagnosis of type 2 diabetes in Finland. BMJ Open Diabetes Research & Care. Peer reviewed. DOI: 10.1136/bmjdrc-2026-006329 agingresearch.news/s/8drekj Retrospective dynamic registry cohort: 485,713 people, 2018 to 2023; median follow-up six years; 82,431 deaths. Diagnosis groups: ≤40 years, 28 939; 41–79 years, 422 578; ≥80 years, 34 196. Comparator: Finns without diabetes. Death ratios adjusted for age and sex. Risk-factor models also included diabetes duration, blood sugar, cholesterol, kidney risk, cardiovascular and other complications, and medication use. Death ratios (100 = expected mortality; 95% confidence intervals): overall 92 (91 to 92); diagnosis ≤40: 288 (273 to 304); ≥80: 56 (55 to 57). Adjusted mortality rate ratios (95% confidence intervals): heart failure 2.07 (2.01 to 2.13); high/very high kidney risk 1.65 (1.61 to 1.70); peripheral artery disease 1.37 (1.32 to 1.42); female sex 0.69 (0.67 to 0.70); diabetes duration 0.97 (0.97 to 0.97); baseline age 1.06 (1.06 to 1.06); diabetic eye damage 0.73 (0.68 to 0.79); elevated average blood sugar 1.05 (1.02 to 1.08). Cholesterol showed no clear mortality difference overall. Medication associations: insulin alone 3.37 (3.10 to 3.67); insulin combinations 1.63 (1.52 to 1.75); non-insulin glucose-lowering drugs 0.92 (0.86 to 0.98); lipid-lowering drugs 0.57 (0.56 to 0.59); kidney-protective drugs 0.73 (0.71 to 0.75). These were prescribing comparisons, not treatment experiments. Sensitivity checks compared patient characteristics, without revised death ratios. Mortality with missing kidney measurements was 30% versus 10% with measurements. People with diabetic eye damage had more illnesses and treatment. Blood sugar predicted mortality in older diagnosis groups, but not the youngest group. - Other illnesses were unaccounted for in the population comparison. - Older diagnosis groups may include healthier survivors and milder cases found through screening. - Earlier diagnosis dates and disease duration could be inaccurate. - People with existing diabetes had to survive until 2018. - Illnesses recorded near death received the same weight as earlier diagnoses. - Prescriptions near death could obscure medication benefits; insulin prescribing reflected greater illness severity. - Smoking, body mass index and blood pressure were unavailable. - Missing laboratory measurements differed by health and care.

Sources and notes Health-Related Quality of Life Across the Spectrum of Cardiovascular-Kidney-Metabolic Syndrome: Insights from Participant-Level Pooled Analysis of 3 Global Trials. AHA Scientific Sessions 2026 Abstract. Conference abstract, not a preprint. DOI: 10.1161/circ.154.suppl_1.5952 agingresearch.news/s/yesj7e Prespecified participant-level analysis of three trials that randomized adults to finerenone or matching placebo. Two enrolled people with chronic kidney disease and type 2 diabetes; one enrolled people with heart failure. Of 18,991 participants, 18,694 (98%) had starting quality-of-life scores: 5,531 and 7,185 from the kidney disease/diabetes trials and 5,978 from the heart-failure trial. Mean age was 67±10 years; 35% were women. The paper does not give treatment-arm counts. Quality of life was assessed at baseline and 12 months using EQ-5D-5L, a health-status questionnaire scored from below zero to 1, with higher scores indicating better health. Multivariable survival models assessed the relationship with cardiovascular death or heart failure hospitalization; the paper does not list their adjustment factors. Models of score changes adjusted for baseline values, trial, geographic region and randomized treatment. Mean starting score was 0.75±0.21. Lower scores went with older age, being a woman, higher body mass index, poorer kidney function and more coexisting illnesses (all P<0.001). Each 0.1-point lower score was associated with a higher risk of cardiovascular death or heart failure hospitalization: adjusted hazard ratio 1.10 (95% CI, 1.08–1.12; P<0.001). Finerenone's effects on that combined outcome were consistent across starting scores (interaction P=0.80). Similar findings were reported for major cardiovascular events, combined kidney outcomes and death from any cause; numerical estimates were not given. At 12 months, the quality-of-life difference favoring finerenone over placebo was +0.006 (95% CI, +0.001 to +0.011; P=0.017). Improvements were greater among people with poorer starting scores (interaction P=0.005). - Other health differences could have contributed to the relationship between poorer quality of life and later heart outcomes. - These findings concern people with chronic kidney disease and type 2 diabetes, or heart failure.