REF-1574 · Araujo et al., 2022 · British Journal of Sports Medicine · Paper · PubMed

Should I worry if I cannot stand on one leg for 10 s?

Failing went with higher death risk (HR 1.84, 1.23-2.78) in 1702 adults aged 51-75. It is a marker in a cohort, not a diagnosis; no proof balance work extends life.

Main caveat: Abstract only. No full text is held.
limited evidenceMeasured on non-athletes and recreational people

Participants were tested between 2008 and 2020 and classed YES or NO on completing the stance. Survival was worse in the NO group (log-rank 85.6, p<0.001). Adding the test to established risk factors improved prediction by -2 log likelihood and integrated discrimination improvement. The authors see potential in adding it to routine examination of middle-aged and older adults.

Our reading: the test is quick and may flag frailty or illness that other measures miss, which is the only thing this design speaks to.

Protocol

Design: one baseline test; survival curves and Cox models.
Adjustment: age, sex, BMI, comorbidities. Recent falls and physical activity were not controlled, by the authors' own statement.
Not in the abstract: recruitment source, attempts allowed, stance rules, the list of comorbidities, sex-specific results.
Funding and interests: Brazilian research councils CNPq and FAPERJ per the Europe PMC record; interests not in what was read.

Key Results

MeasureYES (could hold)NO (could not)
Share of sample79.6% (our arithmetic)20.4%
Died over a median 7 years4.6%17.5%
All-cause mortality, adjustedreferenceHR 1.84 (95% CI 1.23-2.78)
Overall deaths: 7.2%, about 123 people (our arithmetic); the abstract does not list the comorbidities, so events per covariate cannot be judged.
Reporting. The HR is printed with p<0.001, but an interval of 1.23-2.78 corresponds to about p = 0.003 (our arithmetic). Treat the interval as the result.
The crude gap shrinks. A 17.5% vs 4.6% crude split is far larger than the adjusted HR; age accounts for part of it (inference).
Not read: the size of the prediction gain, which the abstract calls significant without printing a figure.

Mechanisms

Measured: none besides the test.
Authors' interpretation: balance falls quickly after the mid-50s, raising the risk of falls and other adverse outcomes.
Inference: failing the stance may reflect neurological, circulatory or musculoskeletal illness, or low activity, rather than balance as such. Reverse causation is the first reading of any such test.

Practical Application

Should I worry if I cannot stand on one leg for 10 s? Failing went with higher death risk (HR 1.84, 1.23-2.78) in 1702 adults aged 51-75. It is a marker in a cohort, not a diagnosis; no proof balance work extends life.
Good for: a free self-check for someone aged 51-75 training for health, though what was read gives no stance rules or safety procedure. A failure is a prompt to mention it at the next medical visit (Inference: the authors propose it for routine examinations, not self-testing).
Not good for: telling anyone that a balance programme will extend life. The direct trial evidence on balance work concerns falls (REF-1575), not death.
Not good for: athletes who pass, or people under 51.
Not good for: diagnosis. Nothing here says what a failure means for one person.

Limitations

Abstract only. No full text is held.
Uncontrolled variables named by the authors: recent falls and physical activity. Smoking and medication are not named in the adjustment.
Baseline-only test with a binary result, in a sample whose source the abstract does not give.
Few deaths, one cohort, no external validation described.
Printed P value does not match the interval; see Key Results.
No intervention. No claim about improving the test can be drawn.
How this entry was checked and changed
Why this level

(from the abstract) — n ≥ 200 would earn the observational B+; one step down because only the abstract was read, the sample's source and the test's conditions are missing, the authors themselves list falls and activity as uncontrolled, and the printed P value does not match its interval