Although treadmill exercise testing can provide an assessment of cardiorespiratory fitness, which serves as an independent prognostic indicator, numerous studies now suggest that usual gait speed, time, or distance covered during walk performance tests and weekly walking distance/time are powerful predictors of mortality and future cardiovascular events in selected patients. This review summarizes the relation between these variables and their association with cardiovascular and all-cause mortality, with specific reference to potential underlying mechanisms and implications for the clinician. Contemporary health care providers have escalating opportunities to promote lifestyle physical activity using pedometers, accelerometers, and smartphone-based health and wellness applications. In conclusion, fitness and/or ambulatory indexes should be considered a “vital sign” in middle-aged and older adults.
Peak or symptom-limited treadmill testing is widely used to screen for or evaluate signs or symptoms of ischemic heart disease. The attained treadmill speed, grade, and duration also allow an assessment of cardiorespiratory fitness (CRF), expressed as peak metabolic equivalents (METs; 1 MET = 3.5 ml O 2 /kg/min), that provides an independent risk factor for all-cause and cardiovascular (CV) mortality. In a seminal report, Blair et al reported an inverse relation between peak METs and CV and all-cause mortality in a large cohort of healthy middle-aged subjects (n = 13,344) but suggested an “asymptote of gain” beyond which further improvements in CRF conferred no additional survival benefit. This asymptote was estimated to be ∼9 and 10 METs for women and men, respectively, approximate cut points that have been substantiated by others. The investigators concluded that the peak MET values associated with the lowest death rates are attainable by most men and women who engage in regular, moderate-intensity exercise. It was further suggested that a brisk walk of 30 to 60 minutes each day would be sufficient to produce the fitness standard, that is, ∼9 and 10 METs for women and men, respectively. Conversely, an exercise capacity <5 METs generally identifies a population with heightened mortality.
To further clarify the fitness-mortality association, Blair et al reported on the relations between changes in CRF and the risk of death in men. The highest death rate (122.0/10,000 man-years) occurred in men who were unfit at both initial and follow-up clinical examinations (mean interval between examinations 4.9 years); the lowest death rate (39.6/10,000 man-years) was in men who were physically fit at both examinations. Men who improved from the unfit to the fit category between the first and second examinations had an intermediate death rate (67.7/10,000 man-years) after adjusting for age, health status, and other risk factors. For each minute increase in treadmill time between examinations, there was a 7.9% decrease in risk of mortality (p = 0.001). These findings support a cause-and-effect relation between improved CRF and reduced mortality, rather than merely an association between these variables.
More recently, a landmark investigation of 2,172 patients admitted with acute coronary syndromes evaluated the effect of preadmission physical activity status on inhospital mortality and CV health outcomes 1 month after discharge. Patients who were physically active demonstrated 0.56 lower odds of inhospital mortality and 0.80 lower odds of recurrent CV events within the first 30 days after discharge. In a cohort of 16,533 middle-aged and older men and women, with varied body habitus and risk factor profiles, the presence of low CRF was associated with an approximately twofold greater 30-year risk for CV mortality.
Although peak or symptom-limited exercise testing can provide diagnostic and prognostic information in older adults with and without coronary heart disease (CHD), numerous reports have shown that simple walking metrics, including usual gait speed, time, or distance covered during walk performance tests and weekly walking distance/time, are powerful predictors of mortality and CV events ( Table 1 ). Recently, researchers reported on a simple functional assessment tool, the ability to sit and rise from the floor, to predict all-cause mortality in a cohort of 2002 adults aged 51 to 80 years (68% men). Over an average follow-up of 6.3 years, there were 159 deaths (7.9%). Multivariate analysis adjusting for age, gender, and body mass index revealed that subjects in the lowest functional category (e.g., poor flexibility and/or muscular fitness) demonstrated a fivefold to sixfold higher all-cause mortality. Thus, simple physical performance tests may be used to screen for mobility limitations, often an early sign of functional decrease, and for risk stratifying selected population subsets.
| First Author (year) | Population | Age (yrs) | Follow- Up (yrs) | Fitness Measurement | Mortality or Cardiovascular Events |
|---|---|---|---|---|---|
| Hakim et al (1998) | 707 retired men | 61–81 | 12 | Distance walked per day | Mortality rate for men who walk <1 mile/day was ∼ twice that among those who walk >2 miles/day |
| Manson et al (1999) | 72,488 women, free of CVD or cancer | 40–65 | 8 | Time walked per week | Women who walked ≥3 hr/wk at a brisk pace, had a RR of 0.65, as compared with sedentary controls |
| Manson et al (2002) | 73,743 women, free of CVD or cancer | 50–79 | 3.2 | Average walking pace | Women who walked at 2 to 3 mph, 3 to 4 mph, and >4 mph had RRs of CVD of 0.86, 0.76, and 0.58, respectively |
| Cesari et al (2005) | 3047 older adults | 74.2 | 4.9 | Time to walk over a 6-m course | Older persons who walk <1 m/s had a higher death rate (RR=1.64) than those who walk ≥1 m/s |
| Newman et al (2006) | 3075 adults | 70–79 | 4.9 | Corridor walk (400 m) | Each additional minute of performance time was associated with an adjusted HR of 1.29 for mortality |
| Rosano et al (2008) | 3156 older adults (43% men, 29% black) | 70.4 | 8.4 | Time to walk a 15-foot course | Slower walking speed (<1.0 m/s) was associated with a greater risk of mortality and incident disability |
| Kavanagh et al (2008) | 6956 men with CVD; 12 mo walk training program | 55.1 | 9.0 | Walking distance | Each 1-mile improvement was associated with a 20% reduction in cardiac death |
| Stanaway et al (2011) | 1705 men | ≥70 | 4.9 | Time to walk 6 m | A walking speed of ≤0.82 m/s was most predictive of mortality. No men who walked ≥1.36 m/s (3 mph) died |
| Studenski et al (2011) | 34,485 older adults | 73.5 | 6–21 | Gait speed | The pooled HR per each 0.1 m/s faster gait speed was 0.88 |
| Beatty et al (2012) | 556 outpatients with stable CHD | 67.6 | 8.0 | 6MWT | Each SD decrease in distance (104 m) was associated with a 30% ∗ to 55% † higher rate of cardiovascular events |
| Yazdanyar et al (2014) | 1665 Cardiovascular Health study participants | 77 | 8.0 | 6MWT | Performance in the shortest (i.e., least distance covered) quintile was associated with increased mortality: HR = 2.1 |
| Jefferis et al (2014) | 3435 men free from CVD and HF | 68.3 | 10.9 | Time walked per week | Compared with men walking 0 to 3 hr/wk, men walking 4 to 7, 8 to 14, and 15 to 21 hr/wk had HRs for stroke of 0.89, 0.63 and 0.68 |
| Lo et al (2015) | 1119 participants with HF | ≥65 | 10 | Gait speed | Impairment in gait speed (<0.8 m/s), measured ≤1 year before HF diagnosis, was independently associated with mortality (HR, 1.37) |
∗ Adjusted for potential confounders.
Gait Speed, Survival, and CV Outcomes
Until recently, researchers had no idea of the ambulatory speed that may be synonymous with being “caught” by the Grim Reaper, that is, the walking pace that may be associated with a heightened mortality. Data from the Concord Health and Aging in Men Project, a cohort study of 1,705 men aged ≥70 years living in several suburbs in Sydney, Australia, were used to answer this question. At baseline, walking speed was carefully measured at usual pace, documenting the fastest time from 2 trials. After an average follow-up of nearly 6 years, investigators determined the particular walking speeds that were most predictive of increased mortality and survival. A natural walking speed of 0.82 m/s (∼2 mph) was most predictive of early mortality, whereas older men who walked faster than 0.82 m/s were 1.23 times less likely to die (95% confidence interval 1.10 to 1.37) than those who walked at slower paces (p = 0.0003). In contrast, no men who initially walked at speeds of ≥1.36 m/s (∼3 mph) were among the 266 deaths reported. Interestingly, the walking speed associated with the highest mortality was virtually identical to the gait speed (0.80 m/s) signifying the median life expectancy in a pooled analysis of 9 cohort studies using individual data from diverse populations (n = 34,485 community-dwelling older adults, ≥65 years; 59.6% women) with baseline gait speed data. The pooled hazard ratio per each 0.1 m/s faster gait speed was 0.88.
In outpatients with stable CHD, the 6-minute walk test has been shown to provide independent and additive information beyond traditional risk factors and a discrimination ability similar to aerobic capacity (peak METs) for predicting CV events, including heart failure (HF), myocardial infarction, and death. In adults ≥65 years with HF, impairment in gait speed (<0.8 vs ≥0.8 m/s), measured within 1 year before the diagnosis of incident HF, was independently associated with mortality (hazard ratio 1.37), after adjusting for sociodemographic and potential clinical confounders. Collectively, these data suggest that in coronary patients with and without HF, efforts aimed at improving outcomes and reducing mortality should include measures of mobility and function, such as gait speed, to identify subjects who are at risk for adverse outcomes and who may benefit from interventions aimed at preserving or enhancing mobility and functional capacity. Nevertheless, clinical trials are needed before we can confidently state that increasing walking speed will improve outcomes in the elderly population.
Potential Underlying Mechanisms
In a comprehensive meta-analysis of 43 studies of the relation between physical activity and CHD incidence, the relative risk of CHD in relation to physical inactivity ranged from 1.5 to 2.4, with a median value of 1.9. Moreover, the relative risk of a sedentary lifestyle appeared to be similar in magnitude to that associated with other major CHD risk factors. According to 2 scientific statements from the American Heart Association, existing published data satisfy the criteria required to infer a causal relation from epidemiologic research and, thus, for physical inactivity to be designated as a major CHD risk factor. These epidemiologic studies, when combined with relevant experimental and clinical investigations providing biologic plausibility ( Figure 1 ), strongly support the recommendation that regular physical activity reduces the incidence of CHD.
