Leg Strength and Nutrition Together Predict 16-Year Death Risk in Older Adults

A simple question—how strong are your legs, and how well nourished are you—may predict how long you will live with surprising accuracy. A new 16-year study of more than 1,000 Japanese adults aged 70 and older reports that combining directly measured leg-press strength with a standard nutritional index dramatically sharpens the ability to identify older adults at high risk of dying. The research, published in the Journal of Cachexia, Sarcopenia and Muscle, draws on the Tsurugaya Project, a community-based cohort study conducted in Sendai City, Miyagi Prefecture, and offers some of the strongest long-term evidence yet that muscle power and nutritional status act as complementary, not redundant, signals of longevity.

The scientific rationale behind the study rests on two well-established but rarely merged lines of evidence. Malnutrition in older adults is a recognized prognostic factor for total mortality, and in geriatric medicine it is commonly assessed using the Geriatric Nutritional Risk Index, or GNRI. This index integrates serum albumin levels—a marker of protein status and systemic health—with the ratio of actual to ideal body weight derived from height. Separately, muscle weakness has repeatedly been linked to elevated mortality risk, whether measured through handgrip, quadriceps, or respiratory muscle strength. Yet most previous studies that combined nutrition and strength focused on hospitalized patients or people with cancer, and the few community-based analyses either followed participants for only four or five years or relied on body mass index, a contested proxy for nutritional status.

The new study distinguishes itself in a second, more technical way: it measured leg-press strength directly rather than estimating it. Leg-press strength captures the simultaneous explosive extension of the hip, knee, and ankle joints in both lower extremities, recruiting the gluteal muscles, femoral muscles, and calf muscles. Because it engages most of the major leg muscle groups at once, it evaluates a far larger muscle mass than handgrip or isolated quadriceps tests and reflects the functional capacity needed for complex movements such as rising from a chair or climbing stairs. Notably, it is the only conventional strength measure that incorporates the calf muscles, a point of growing relevance since the Asian Working Group for Sarcopenia’s 2025 consensus update designated calf circumference as the key primary tool for sarcopenia case-finding.

Participants were drawn from the Tsurugaya Project, which in 2002 invited 2,730 community-dwelling residents aged 70 or older; 1,198 enrolled and 1,175 consented. After excluding individuals lacking leg-press data, Mini-Mental State Examination data, or albumin values, 1,065 participants entered the final analysis. Each participant sat on a horizontal leg extension device with feet secured on the pedal at a 90-degree knee angle and exerted maximal power as rapidly as possible across five trials, with the highest value recorded in watts. To account for body size, strength was divided by body mass index. Nutritional status was calculated from the GNRI formula: serum albumin in grams per liter multiplied by 1.489, plus the ratio of measured to optimal body weight multiplied by 41.7, with optimal weight determined by sex-specific Lorentz equations.

Rather than importing cutoffs designed for short-term outcomes in hospitalized patients, the researchers derived study-specific thresholds optimized for 16-year all-cause mortality using maximally selected rank statistics. This yielded sex-specific leg-press cutoffs of 272.63 W/kg/m² for men and 129.78 W/kg/m² for women, and a sex-neutral GNRI threshold of 105.11. Critically, the two indices proved nearly independent: the correlation coefficient between leg-press strength and GNRI was just 0.096, and the variance inflation factor of 1.009 ruled out multicollinearity. That statistical independence is precisely what makes their combination informative—each index carries distinct prognostic information that the other cannot supply. Crossing the two binary classifications produced four groups: high-strength/good-GNRI (589 people), high-strength/poor-GNRI (196), low-strength/good-GNRI (157), and low-strength/poor-GNRI (123).

Over 16 years of follow-up, during which mortality data were collected from the Sendai Municipal Authority through July 2018, 415 participants died. Kaplan–Meier survival curves revealed a stark gradient across the four groups. Five-year mortality was 5.60 percent in the high-strength/good-GNRI group but 27.64 percent in the low-strength/poor-GNRI group—nearly five times higher. Intriguingly, the high-strength/poor-GNRI group initially showed better survival than the low-strength/good-GNRI group, hinting that preserved muscle power might partially buffer the dangers of suboptimal nutrition. In fully adjusted Cox proportional hazards models accounting for sex, age, smoking history, Timed Up and Go performance, cognitive scores, and histories of hypertension, diabetes, cardiovascular disease, liver disease, and kidney disease, hazard ratios for all-cause mortality were 1.55 for high-strength/poor-GNRI, 1.69 for low-strength/good-GNRI, and 2.47 for low-strength/poor-GNRI, compared with the reference group. A stepwise increase in risk across categories remained highly significant.

Analyses treating the measures as continuous variables reinforced the pattern: each unit increase in leg-press strength was associated with a small but significant reduction in mortality risk in both men and women, and each unit increase in GNRI was similarly protective. To guard against reverse causality—the possibility that pre-existing illness at baseline had already weakened both strength and nutrition—the researchers excluded the 30 participants who died within the first two years. The stepwise survival gradient persisted, with an adjusted hazard ratio of 2.39 for the low-strength/poor-GNRI group, suggesting the findings were not merely an artifact of pre-existing disease.

The most consequential results concern predictive performance. The concordance index, which quantifies how well a model discriminates between those who die and those who survive, was 0.54 for muscle strength alone and 0.59 for GNRI alone—but rose to 0.63 for the combined MS-GNRI index, a statistically significant improvement. Net reclassification improvement, which measures how many individuals are correctly shifted into appropriate risk categories, reached 0.62 over muscle strength alone and 0.41 over GNRI alone, implying gains in predictive accuracy of 62 percent and 41 percent respectively. Time-dependent area-under-the-curve analyses told the same story at both 8 years (0.70 for the combined index versus 0.65 and 0.61 for the individual measures) and 12 years (0.67 versus 0.64 and 0.59).

The authors place these findings in the context of Japan’s leading causes of death—malignancy, cardiovascular disease, and pneumonia. Prior cohort work suggests diminished leg-press strength raises cancer mortality risk, while studies of quadriceps weakness in the very old point to possible links with pneumonia death; malnutrition, by contrast, worsens mortality across all major causes. The study has limitations the authors acknowledge candidly: without head-to-head comparison, it cannot establish whether leg-press strength outperforms handgrip or quadriceps measures, the cohort was exclusively Japanese so generalizability to other ethnic groups remains untested, and the study-specific cutoffs require validation in independent cohorts. Still, the message is clear and clinically actionable. Adequate nutrition is essential for longevity, but preserved muscle strength appears to be a critical, partially independent determinant of survival. Whether deliberately training leg-press power can actually reduce mortality risk remains the field’s next great question—and one that could reshape how clinicians screen the world’s aging populations.

Subject of Research: Combining leg-press muscle strength and nutritional status to predict long-term all-cause mortality in community-dwelling older adults

Article Title: Integrating Leg‐Press Strength and Nutritional Status Improves 16‐Year Mortality Risk Stratification in General Older Adults

Article References: Du, Y., Okazaki, T., Li, X., Obata, K., Suzuki, N., Miura, T., Miyagi, M., Nagatomi, R., Kogure, M., Nakaya, N., Hozawa, A., & Ebihara, S. (2026). Integrating Leg‐Press Strength and Nutritional Status Improves 16‐Year Mortality Risk Stratification in General Older Adults. Journal of Cachexia, Sarcopenia and Muscle, 17(5), Article e70375. https://doi.org/10.1002/jcsm.70375

Image Credits: AI Generated

DOI: 10.1002/jcsm.70375

Keywords: leg-press strength, Geriatric Nutritional Risk Index, all-cause mortality, older adults, sarcopenia, malnutrition, Tsurugaya Project, risk stratification, muscle strength, aging, cohort study, serum albumin