Objectively measured physical capability levels and mortality: systematic review and meta-analysis

Abstract

Objective: To do a quantitative systematic review, including published and unpublished data, examining the associations between individual objective measures of physical capability (grip strength, walking speed, chair rising, and standing balance times) and mortality in community dwelling populations.

Design: Systematic review and meta-analysis.

Data sources: Relevant studies published by May 2009 identified through literature searches using Embase (from 1980) and Medline (from 1950) and manual searching of reference lists; unpublished results were obtained from study investigators.

Study selection: Eligible observational studies were those done in community dwelling people of any age that examined the association of at least one of the specified measures of physical capability (grip strength, walking speed, chair rises, or standing balance) with mortality.

Data synthesis: Effect estimates obtained were pooled by using random effects meta-analysis models with heterogeneity between studies investigated.

Results: Although heterogeneity was detected, consistent evidence was found of associations between all four measures of physical capability and mortality; those people who performed less well in these tests were found to be at higher risk of all cause mortality. For example, the summary hazard ratio for mortality comparing the weakest with the strongest quarter of grip strength (14 studies, 53 476 participants) was 1.67 (95% confidence interval 1.45 to 1.93) after adjustment for age, sex, and body size (I(2)=84.0%, 95% confidence interval 74% to 90%; P from Q statistic <0.001). The summary hazard ratio for mortality comparing the slowest with the fastest quarter of walking speed (five studies, 14 692 participants) was 2.87 (2.22 to 3.72) (I(2)=25.2%, 0% to 70%; P=0.25) after similar adjustments. Whereas studies of the associations of walking speed, chair rising, and standing balance with mortality have only been done in older populations (average age over 70 years), the association of grip strength with mortality was also found in younger populations (five studies had an average age under 60 years).

Conclusions: Objective measures of physical capability are predictors of all cause mortality in older community dwelling populations. Such measures may therefore provide useful tools for identifying older people at higher risk of death.

Fig 1 Flow diagram for identification of published studies. Includes identification of studies for additional review of other health outcomes, reported elsewhere

Fig 2 Flow diagram showing contact with authors and ascertainment of results for inclusion in review

Fig 3 Hazard ratios of mortality per 1 kg increase in grip strength with adjustment for age, sex (where appropriate), and body size. B=both sexes; F=women only; M=men only; MrOS=Osteoporotic Fractures in Men Study; SOF=Study of Osteoporotic Fractures. *Estimates adjusted for multiple factors as results from models adjusted for age, sex, and body size were not available. Adjustments were as follows: Cesari (2008)—age, sex, body mass index (BMI), cognitive performance, number of clinical conditions, albumin, total cholesterol; Newman—age, sex, race, height, smoking, physical activity, number of chronic conditions, education, interleukin-6, Center for Epidemiologic Studies Depression scale (CES-D), DXA body composition; Shibata—blood pressure, cholesterol, albumin, visual retention, education, BMI, history of chronic diseases, alcohol, smoking, activities of daily living, electrocardiographic changes; Takata—sex, smoking, BMI, systolic blood pressure, marital status, total cholesterol, glucose, complications from prevalent disease

Fig 4 Hazard ratios of mortality comparing weakest with strongest quarter of grip strength with adjustment for age, sex (where appropriate), and body size. B=both sexes; F=women only; M=men only; MrOS=Osteoporotic Fractures in Men Study; SOF=Study of Osteoporotic Fractures

Fig 5 Summary hazard ratios of mortality from meta-analyses comparing each quarter of grip strength, walking speed, and chair rise time with highest quarter, including results adjusted for age, sex (where appropriate), and body size (n=number of data points included in meta-analysis)

Fig 6 Hazard ratios of mortality comparing slowest with fastest quarter of walking speed with adjustment for age, sex (where appropriate), and body size. B=both sexes; F=women only; M=men only. *When results from analyses of H-EPESE by Markides et al and Ostir et al were included in this meta-analysis in place of Al Snih’s results, the findings were unchanged

Fig 7 Hazard ratios of mortality comparing slowest with fastest quarter of chair rise time with adjustment for age, sex (where appropriate), and body size. B=both sexes; F=women only; M=men only; MrOS=Osteoporotic Fractures in Men Study; SOF=Study of Osteoporotic Fractures