The SPRINTT trial, funded by the European Union’s Horizon 2020 program, was one of the largest randomized controlled trials ever conducted in community-dwelling frail older Europeans. It enrolled more than 1,500 men and women aged seventy and above across multiple European countries, all of whom met criteria for physical frailty and sarcopenia — a combination characterized by slowness, weakness, low physical activity and reduced muscle mass. Participants were randomized to either a moderate-intensity, multicomponent physical activity program, centered on walking and adapted to individual capacity, or to a control group receiving structured health education. The primary goal was to prevent major mobility disability, defined as the inability to walk 400 meters, a threshold with proven clinical relevance because it captures the capacity to function independently in everyday life. Earlier reports from the trial indicated that the intervention produced statistically significant but modest benefits, leaving open the crucial question of whether particular subgroups of patients benefited more than others.

The new study addresses that question through the lens of multimorbidity patterns rather than simple disease counts. Most previous analyses treated multimorbidity as a number — two diseases, three diseases, four or more — or as a crude index of cumulative burden. But clinicians have long observed that a patient with diabetes, peripheral vascular disease and osteoarthritis is not clinically equivalent to a patient with chronic obstructive pulmonary disease, heart failure and depression, even if both carry the same number of diagnoses. The researchers therefore applied statistical clustering techniques to the participants’ disease profiles, identifying groups of conditions that tended to occur together and grouping individuals according to the characteristic pattern of their chronic illnesses. This approach, sometimes described as person-centered rather than disease-centered, allowed the team to ask whether the protective effect of exercise against mobility loss was homogeneous across these clinically distinct constellations of disease.

The technical rationale for such heterogeneity is compelling. Different disease clusters disable the body through different physiological pathways. Cardiometabolic patterns, typically combining type 2 diabetes, hypertension and obesity, limit mobility largely through vascular damage, impaired muscle perfusion and peripheral neuropathy, all of which erode the capacity for sustained aerobic effort. Osteoarticular patterns, dominated by osteoarthritis and chronic pain, restrict movement through mechanical limitation and pain-avoidance behavior that accelerates deconditioning. Cardiorespiratory and neuropsychiatric patterns act through reduced aerobic reserve, breathlessness and fatigue, and through the motivational deficits associated with depression. A walking-based intervention that increases daily physical activity might, in principle, counteract deconditioning in all of these groups, but the magnitude of the achievable gain could differ substantially depending on which bottleneck — vascular, mechanical, respiratory or psychological — dominates the individual patient’s trajectory.

The analysis confirmed that suspicion in a clinically meaningful way. The benefits of the multicomponent physical activity program on the risk of developing major mobility disability were not uniform across multimorbidity patterns. For some clusters of conditions, the intervention produced clear and robust protection: frail older adults whose chronic disease burden followed certain patterns experienced a significantly lower hazard of losing the ability to walk 400 meters when they exercised regularly compared with their counterparts receiving health education alone. For other patterns, the observed benefit was attenuated and statistically uncertain, suggesting that standard exercise prescriptions may need to be adapted — or supplemented with other treatments — to reach patients whose disabling pathway is driven by disease mechanisms that physical activity alone cannot fully offset. The precise estimates, hazard ratios and interaction terms reported in the article quantify this differential effect, providing effect-size benchmarks that guideline developers and trial designers can build upon.

Methodologically, the study is a careful exercise in post-hoc subgroup science, a field notorious for false positives. The authors handled the inherent risks by defining multimorbidity patterns using prespecified statistical procedures, applying clustering algorithms to baseline disease data, testing interactions between intervention assignment and pattern membership, and adjusting for the covariates that typically confound mobility outcomes, including age, sex, baseline physical performance, body mass index and gait speed. Sensitivity analyses examined whether the results were robust to alternative cluster definitions and to variations in the handling of missing data. Although the analysis was not powered a priori for pattern-specific comparisons — a limitation the authors acknowledge — the coherence of the findings across analytic choices lends credibility to the central conclusion: multimorbidity is not a monolith, and its internal structure matters for prevention.

The implications reach well beyond the walls of geriatric medicine. Health systems across Europe, North America and Asia are confronting a demographic transition in which the majority of people over sixty-five live with at least two chronic conditions. Current guidelines for these patients are typically assembled disease by disease — a cardiology recommendation, a diabetes recommendation, an orthopedic recommendation — with little attention to how the combination of conditions changes what prevention can achieve. The SPRINTT analysis provides an empirical basis for a different model, one in which the pattern of multimorbidity becomes a stratification variable in clinical decision-making. A physical activity prescription for a frail older patient with a cardiometabolic cluster may be among the most effective interventions available; for a patient whose mobility is limited by a different cluster, the same prescription may need reinforcement with pain management, nutritional support, depression treatment or assistive technology to translate into preserved walking ability.

The findings also intersect with a growing body of research on physical resilience — the capacity of an organism to resist and recover from health stressors. Frailty, understood as a state of diminished physiological reserve, has often been treated as a single construct measured with composite scores such as the frailty phenotype or the frailty index. The SPRINTT results suggest that the trajectory of frail older adults is better predicted by a vector than by a scalar: the direction of their disease burden, not merely its magnitude, determines how plastic their mobility remains. This resonates with mechanistic work on the biology of aging, in which distinct molecular hallmarks — chronic inflammation, mitochondrial dysfunction, cellular senescence, neuromuscular junction degeneration — are differentially engaged by different chronic diseases and may respond differently to exercise as a systemic intervention. Physical activity is one of the few therapies known to act simultaneously on most of these pathways, which helps explain why it remains effective, if unevenly so, across clinically diverse populations.

Questions inevitably remain. The trial population consisted of community-dwelling Europeans selected for physical frailty and sarcopenia, so the generalizability of the pattern-specific results to frailer institutionalized populations, to non-European cohorts or to younger adults with early multimorbidity has not been demonstrated. The clustering solution chosen by the investigators is one of several statistically defensible partitions of the disease space, and alternative algorithms might yield patterns with different boundaries. And although major mobility disability is a validated and consequential endpoint, future work should examine whether multimorbidity patterns also moderate the effects of exercise on other outcomes, including falls, hospitalization, cognitive decline and mortality. Longer follow-up, larger samples and replication in independent cohorts will be needed before pattern-stratified exercise prescriptions become standard clinical practice.

Even with those caveats, the study marks a turning point in how the prevention of disability in old age can be conceptualized. It moves the field away from the blunt arithmetic of disease counting and toward a nosology of combinations, in which the specific constellation of conditions a person carries is treated as clinically actionable information. It also delivers a pragmatic message of hope with nuance: exercise remains one of the most powerful tools available for keeping frail older adults on their feet, but its power is conditional, and understanding those conditions is the key to unlocking it for everyone. As populations age and multimorbidity becomes the norm rather than the exception, the lesson from SPRINTT is that precision geriatrics — matching preventive interventions to the pattern of disease, not just to its burden — is no longer an aspiration but an evidence-backed necessity.

The need for such predictive tools is particularly urgent in Japan, one of the world’s most rapidly aging societies. Nearly 60 percent of Japanese adults aged 85 and older already receive some form of support through the country’s Long-Term Care Insurance system. Yet disability often develops gradually, through a combination of declining muscle strength, reduced physical resilience, chronic disease, inflammation, and impaired organ function. By the time a person has difficulty walking, bathing, or managing daily activities, opportunities for prevention may already be limited. A reliable biological warning signal could allow healthcare professionals to intervene while older adults still retain enough physical capacity to benefit from exercise, nutritional counseling, rehabilitation, and other supportive measures.

The research team, led by Associate Professor Yusuke Osawa of Keio University’s Graduate School of Health Management, analyzed plasma samples from participants in the Kawasaki Aging Well-being Project, or KAWP. The discovery analysis included 230 adults who were free of disability when their blood was collected and who were followed for approximately 4.5 years. Rather than selecting biomarkers based only on pre-existing theories, the investigators used a data-driven screening strategy to examine 29 circulating proteins. Machine-learning methods and multivariable statistical models were then used to test whether protein concentrations were linked to the later development of disability or death.

Two proteins stood out consistently. Participants with higher levels of B2M and cystatin C were more likely to become disabled during follow-up. In statistical terms, each increase in B2M was associated with a hazard ratio of 1.35 for future disability, while the corresponding hazard ratio for cystatin C was 1.42. A hazard ratio above one indicates an elevated likelihood of experiencing the outcome over time, although it does not mean that every person with a high value will become disabled. Importantly, the associations remained after the researchers adjusted for age, sex, kidney function, lifestyle characteristics, and other potential confounding factors. This suggests that the proteins may capture aspects of biological vulnerability not fully reflected by conventional clinical measurements.

Cystatin C is a small protein produced by most nucleated cells and filtered from the blood by the kidneys. Because its concentration is less dependent on muscle mass than creatinine, the conventional marker used to estimate kidney function, cystatin C can be particularly informative in very old adults, whose muscle mass may be substantially reduced. B2M is also connected to kidney clearance, but it has an additional biological significance. It forms part of major histocompatibility complex class I molecules found on the surface of nearly all nucleated cells and can rise in conditions involving immune activation, inflammation, and altered filtration. In this sense, elevated B2M may reflect both impaired renal handling and persistent immune-system stimulation.

The researchers next asked whether the discovery could be reproduced outside Japan. They turned to the Invecchiare in Chianti, or InCHIANTI, study, a long-running investigation of aging in Italy that followed participants for as long as 15 years. In that independent population, higher concentrations of B2M and cystatin C were again associated with an increased risk of disability, particularly among adults aged 80 and older. Replication across geographically and culturally different cohorts is a crucial test for any proposed biomarker because it helps determine whether an association reflects a broadly shared biological process rather than a feature unique to one study population, healthcare system, or lifestyle pattern.

Not every signal from the Japanese analysis survived this external test. Epidermal growth factor and interferon gamma-induced protein 10 were initially associated with mortality, but their relationships were not consistently reproduced in the Italian cohort. The contrast strengthens the significance of the B2M and cystatin C findings: the two proteins were not merely part of a large list of statistically interesting molecules, but were among the markers showing the most durable connection with disability across populations. Still, the study is observational. It demonstrates association, not proof that raising or lowering either protein directly causes functional decline. The biomarkers may be indicators of several underlying processes rather than independent drivers of disability.

The biological picture emerging from the study is consistent with the concept of “inflammaging,” the gradual increase in low-grade immune activation that accompanies aging. Unlike an acute infection, inflammaging may persist quietly for years, contributing to vascular damage, muscle wasting, impaired repair, and reduced physiological reserve. Kidney dysfunction can add another layer of risk by allowing metabolic waste products and signaling molecules to accumulate in the circulation. In an 85-year-old, even modest changes in renal function or inflammatory activity may have consequences because the body’s ability to compensate for stress is already diminished. B2M and cystatin C could therefore act as accessible readouts of overlapping processes that precede visible disability.

“Our findings suggest that preserving healthy aging requires attention not only to diseases but also to the biological processes that precede disability,” Professor Yasumichi Arai of Keio University said. Earlier identification of high-risk individuals could create a window for interventions before decline becomes irreversible. Dr. Osawa noted that both proteins can already be measured using standard clinical assays, making them more practical than experimental markers that require specialized laboratory platforms. A future screening approach might combine B2M and cystatin C with information about gait speed, grip strength, nutrition, medication use, cognition, and social circumstances. Such a tool would not predict an individual’s fate, but could help direct limited preventive resources toward people most likely to benefit.

The study’s results point toward a shift in how healthcare systems think about longevity. Extending life without preserving mobility and self-sufficiency can leave older adults and families facing years of escalating care needs. Blood-based risk assessment could eventually support a more proactive model, in which clinicians monitor biological changes before a person requires long-term assistance. The next steps will include determining whether repeated measurements improve prediction, identifying clinically meaningful threshold values, and testing whether targeted interventions can alter outcomes in people with elevated B2M or cystatin C. For now, the findings offer a scientifically grounded lead: in the oldest-old, two familiar blood proteins may provide an early glimpse of the path toward disability—and a chance to change it.

Subject of Research: People

Article Title: Plasma proteins associated with disability and mortality risks in Japanese community-dwelling octogenarians

News Publication Date: 27-Jun-2026

Web References: https://doi.org/10.1007/s11357-026-02377-7; https://www.k-ris.keio.ac.jp/html/100007234_en.html

References: GeroScience. “Plasma proteins associated with disability and mortality risks in Japanese community-dwelling octogenarians.” DOI: 10.1007/s11357-026-02377-7.

Image Credits: Dr. Yusuke Osawa, Keio University, Japan

Keywords: aging, disability risk, biomarkers, beta-2-microglobulin, B2M, cystatin C, kidney function, inflammaging, healthy aging, long-term care, gerontology, blood proteins, preventive medicine, older adults, GeroScience