Grip Strength and Healthspan: What the Cohort Data Shows

Why do simple physical tests predict death better than most lab panels?

Physical capacity measures are standardised tests of what a body can do: how hard a hand squeezes, how fast a person walks, how easily they rise from the floor, how much oxygen they can use under load. In the Prospective Urban Rural Epidemiology (PURE) study of 139,691 adults aged 35 to 70 across 17 countries, published in The Lancet in 2015, every 5 kg reduction in grip strength carried a hazard ratio of 1.16 for all-cause mortality (95% CI 1.13 to 1.20). PURE also reported that grip strength predicted all-cause and cardiovascular mortality more strongly than systolic blood pressure did.

Capacity tests perform well because each one summarises many systems at once. A grip reading reflects muscle mass, motor unit recruitment, nutritional status, chronic inflammation, joint integrity and central nervous system function in a single number, and it costs nothing beyond a dynamometer. A single blood marker reports on one pathway. Capacity tests report on the combined output of all of them.

Is grip strength a cause of longer life or a marker of it?

Grip strength is a marker of biological condition, and the available genetic evidence does not support it being a cause of survival. Willems and colleagues, writing in Nature Communications in 2017, identified 16 genetic loci for grip strength in 195,180 people and then used Mendelian randomisation to test causality. In the EPIC-Norfolk cohort of 21,043 participants with 5,699 deaths, they found no evidence for a causal relationship between muscular strength and all-cause mortality (hazard ratio per kg 0.96, 95% CI 0.91 to 1.03, P = 0.265).

A larger analysis in the same paper, using parental lifespan in UK Biobank across 102,072 paternal and 83,315 maternal deaths, returned a hazard ratio of 1.00 (95% CI 0.98 to 1.03, P = 0.739). The same methods did suggest a causal effect of higher genetically predicted grip strength on lower fracture risk (odds ratio 0.95, 95% CI 0.90 to 0.99, P = 0.02), so the null result for mortality is not simply a failure of statistical power.

The practical consequence is specific. Training a squeeze until a dynamometer reading rises does not purchase the mortality difference observed between strong and weak people in cohort studies. Capacity tests are triage instruments that place a person in a risk distribution. A low score is a reason to investigate the underlying causes, not a reason to practise the test.

What do the grip strength cohorts actually show?

The largest single-cohort estimate comes from UK Biobank. Celis-Morales and colleagues, publishing in The BMJ in 2018, followed 502,293 participants aged 40 to 69 for a mean of 7.1 years and recorded 13,322 deaths. Each 5 kg lower grip strength was associated with a hazard ratio for all-cause mortality of 1.20 in women (95% CI 1.17 to 1.23) and 1.16 in men (95% CI 1.15 to 1.17). Adding grip strength improved the predictive performance of an office-based risk score.

Normative values matter for interpreting an individual reading. Dodds and colleagues pooled 60,803 observations from 49,964 participants across 12 British population studies in PLoS ONE in 2014 and reported a peak mean grip of 51 kg for males aged 29 to 39 and 31 kg for females aged 26 to 42. By age 80, weak grip was present in 23.0% of men and 26.6% of women in that pooled sample.

How fast do you have to walk?

Usual gait speed of 0.8 m/s marks the point at which predicted survival matches the median life expectancy for a person's age and sex, according to the pooled analysis by Studenski and colleagues in JAMA in 2011. That analysis combined individual data from nine cohorts covering 34,485 community-dwelling adults aged 65 and over, followed for 6 to 21 years, with a mean age of 73.5 years, a mean gait speed of 0.92 m/s and 17,528 deaths.

The pooled hazard ratio was 0.88 per 0.1 m/s faster walking (95% CI 0.87 to 0.90). Speeds of 1.0 m/s or faster corresponded to survival longer than age and sex alone would predict, and speeds under 0.6 m/s marked elevated risk. Prediction based on age, sex and gait speed was as accurate as prediction based on age, sex, chronic conditions, smoking history, blood pressure, body mass index and prior hospitalisation.

What does the sit-to-stand test predict?

The sitting-rising test scores a person out of 10, five points for lowering to the floor and five for standing again, with one point deducted for each hand, knee, forearm or other support used. Brito and colleagues tested 2,002 adults aged 51 to 80 in Rio de Janeiro and followed them for a median of 6.3 years, reporting in the European Journal of Preventive Cardiology in 2014 that each one-point lower score was associated with a 21% higher risk of all-cause mortality.

A larger follow-up from the same group, published by Araujo and colleagues in 2025, tracked 4,282 adults aged 46 to 75 for a median of 12.3 years. Natural-cause mortality was 42.1% among those scoring 0 to 4 against 3.7% among those scoring 10 (hazard ratio 3.84, 95% CI 2.25 to 6.97). Cardiovascular mortality was 14.1% against 0.9% (hazard ratio 6.05, 95% CI 2.29 to 20.94). The test loads the knees, hips and lumbar spine, and anyone with a recent joint replacement, symptomatic osteoporosis or impaired balance should not attempt it without supervision.

How much does cardiorespiratory fitness change the picture?

Cardiorespiratory fitness, usually expressed as VO2max or as metabolic equivalents (METs) achieved on a treadmill, produces the largest effect sizes of any capacity measure. Mandsager and colleagues examined 122,007 patients who underwent exercise treadmill testing at the Cleveland Clinic between 1991 and 2014, with a median follow-up of 8.4 years, and reported in JAMA Network Open in 2018 an adjusted hazard ratio of 5.04 for the lowest fitness group against elite performers (95% CI 4.10 to 6.20).

That analysis found no upper limit of benefit. Elite performers had lower mortality than merely high performers (adjusted hazard ratio 0.77, 95% CI 0.63 to 0.95). For comparison, adjusted hazard ratios in the same cohort were 1.29 for coronary artery disease, 1.41 for smoking and 1.40 for diabetes. Kokkinos and colleagues reported comparable gradients in 750,302 US veterans aged 30 to 95 in the Journal of the American College of Cardiology in 2022, with the lowest mortality risk at roughly 14.0 METs and a hazard ratio of 4.09 for the least fit fifth against the extremely fit (95% CI 3.90 to 4.20).

The American Heart Association scientific statement led by Ross in 2016 concluded that cardiorespiratory fitness predicts cardiovascular and all-cause mortality at least as powerfully as hypertension, smoking, obesity, dyslipidaemia and type 2 diabetes, and recommended that all adults have fitness measured or estimated annually.

What are the cut-off values for each measure?

Cut-off values come from sarcopenia consensus panels and from the mortality cohorts above. Thresholds differ by region because the reference populations differ, so the panel should always be named alongside the number.

• Grip strength, EWGSOP2 (Europe, 2019): low is under 27 kg in men and under 16 kg in women.
• Grip strength, AWGS (Asia, 2019): low is under 28 kg in men and under 18 kg in women.
• Grip strength, FNIH Sarcopenia Project (2014, pooled sample of 26,625 older adults): low is under 26 kg in men and under 16 kg in women.
• Grip strength, population peak (Dodds 2014): 51 kg mean for males aged 29 to 39 and 31 kg mean for females aged 26 to 42.
• Usual gait speed: 0.8 m/s or slower is low physical performance under EWGSOP2; AWGS uses under 1.0 m/s over 6 metres; in Studenski 2011, 0.8 m/s approximated median life expectancy and 1.0 m/s or faster exceeded it.
• Five-times chair stand: over 15 seconds under EWGSOP2; 12 seconds or longer under AWGS.
• Short Physical Performance Battery: 8 points or fewer under EWGSOP2; 9 or fewer under AWGS.
• Timed Up and Go: 20 seconds or longer under EWGSOP2.
• Sitting-rising test: scored 0 to 10; scores of 0 to 4 carried a hazard ratio of 3.84 for natural-cause mortality against a score of 10 (Araujo 2025).
• Cardiorespiratory fitness: lowest observed mortality risk at approximately 14.0 METs among 750,302 veterans (Kokkinos 2022).

These thresholds identify a risk category rather than a diagnosis. None of them should be used to start, stop or change any treatment without clinical assessment.

How do you measure these at home?

Gait speed is the easiest measure to reproduce accurately. Mark a 4-metre course on level ground, start from standing, walk at usual pace as if walking down the street, and time the walk with a stopwatch. That is the protocol behind the Studenski pooled analysis, so a home reading maps onto the published thresholds.

The chair-stand and sitting-rising tests need no equipment beyond a standard dining chair and clear floor space. For the five-times chair stand, sit with arms folded across the chest and stand fully five times as quickly as is safe. For the sitting-rising test, start standing, lower to a seated position on the floor and return to standing, deducting one point per support used from a starting score of 10.

Grip strength is the one measure that cannot be improvised. PURE, UK Biobank and every sarcopenia cut-off above were generated with calibrated handheld dynamometers, a Jamar device in the case of PURE. Spring-loaded fitness grippers and grip figures from consumer hardware do not produce values comparable to those thresholds. VO2max estimates from wrist wearables are approximations rather than measurements, and the MET values in the fitness cohorts came from graded exercise testing under supervision. Consumer devices show the same gap between a displayed number and a validated measurement in sleep-stage scoring.

What improves the outcome rather than the score?

Evidence that training changes mortality is separate from evidence that markers predict it, and it is weaker. Momma and colleagues pooled 16 prospective cohort studies in the British Journal of Sports Medicine in 2022 and found muscle-strengthening activity associated with a 10% to 17% lower risk of all-cause mortality, cardiovascular disease, total cancer, diabetes and lung cancer, following a J-shaped curve with maximum risk reduction at roughly 30 to 60 minutes per week. Aerobic work carries a separate dose-response literature, set out in the weekly volume of low-intensity cardio a non-athlete needs.

Change in fitness carries the closest thing to a within-person result. Blair and colleagues re-tested 9,777 men and reported in JAMA in 1995, with a mean of 4.9 years between examinations and 5.1 years of follow-up afterwards, age-adjusted death rates per 10,000 man-years of 122.0 for men unfit at both examinations, 67.7 for men who moved from unfit to fit, and 39.6 for men fit at both. The improvers had a 44% lower mortality risk than the persistently unfit (95% CI 25% to 59%).

Both findings remain observational and cannot establish causation. The same limit applies to other long-running lifestyle cohorts, including the Finnish sauna mortality data. What they support is a narrow claim: the behaviours that raise capacity scores are also associated with lower mortality, while the score itself is a readout and not a lever. A person whose grip strength or gait speed falls below the thresholds above holds useful information and a reason to seek clinical assessment of the causes, which range from undernutrition and deconditioning to occult cardiac, neurological and malignant disease.