Slow self-reported walking pace is associated with a higher mortality
Association of walking pace and handgrip strength with all-cause, cardiovascular, and cancer mortality: a UK Biobank observational studyLiterature - Yates T, Zaccardi F, Dhalwani NN, et al. - Eur Heart J 2017;38:3232–3240
Objectively assessed walking pace has been associated with mortality, with other studies suggesting self-reported walking pace may be a marker of CV health [1,2]. Moreover, muscle strength has been associated with all-cause and CV mortality, and handgrip strength is associated with health outcomes including frailty, falls, nutritional status and mortality [3,4].
In this analysis from the UK Biobank, the associations between self-reported walking pace and handgrip were quantified with all-cause, CV, and cancer mortality. Moreover, it was evaluated whether these associations are maintained across categories of age, BMI, and smoking status.
The UK Biobank is a large prospective cohort of middle-aged adults with the objective to improve prevention, diagnosis, and treatment of chronic disease . From the initial sample of 502 639 participants, those with prevalent cancer or CVD, as well as those without available information on all covariates, were excluded, leaving 420 727 individuals for this analysis.
The assessments of the main parameters of interest were done as follows:
- The self-reported walking pace was assessed with the question: ‘How would you describe your usual walking pace? (i) slow pace, (ii) steady/average pace, and (iii) brisk pace’.
- The handgrip strength was assessed with a hydraulic hand dynamometer while sitting.
- Cardiorespiratory fitness was assessed with a 6-min graded submaximal fitness test on a cycle ergometer, with the workload adjusted to age, height, weight, resting HR, and gender.
- Self-reported walking pace showed a strong graded association with cardiorespiratory fitness in men and women. Slow walkers achieved an average fitness level of 28.1 (95%CI: 27.7–28.6) mL x kg-1 x min-1 in women and 35.9 (95%CI: 35.3–36.5) mL x kg-1 x min-1 in men. Brisk walkers achieved levels of 34.0 (95%CI: 33.8–34.1) mL x kg-1 x min-1 in women and 42.0 (95%CI: 41.8–42.2) mL x kg-1 x min-1 in men.
- In all models, in both women and men, walking pace was associated with all-cause and CV mortality, but not with cancer mortality.
- BMI modified the associations for walking pace in men and women and handgrip strength in men. In particular, for women, the HR for all-cause mortality in slow walkers compared with fast walkers ranged from 2.16 (95%CI: 1.68–2.77) to 1.31 (95%CI: 1.08–1.60) for participants in the bottom and top BMI tertiles, respectively. Corresponding HRs for men were 2.01 (95%CI: 1.68–2.41) and 1.41 (95%CI: 1.20–1.66).
- The HRs for CV mortality for slow walkers compared with fast walkers remained above 1.7 across all categories of BMI in men and women, with a modest heterogeneity in men. A slow walking pace was also associated with cancer mortality in women and men with a low BMI.
- Handgrip strength was associated with CV mortality in men and with all-cause mortality in women and men, however, the results in women were attenuated after removing deaths occurring within the first 2 years.
- Weak handgrip strength was only associated with a higher risk of all-cause mortality in men with low BMI (HR: 1.29; 95% CI: 1.13–1.48), whereas the associations between handgrip strength and CV mortality were not affected by BMI status.
- Handgrip strength was not associated with cancer mortality in women, whereas in men, a weaker handgrip was associated with a lower risk of cancer, particularly after removing deaths occurring within the first 2 years.
In participants in the UK Biobank, a slow self-reported walking pace was associated with a higher risk of all-cause and CV mortality in women and men, particularly in those with a low BMI. These findings suggest that self-reported walking pace could be used for risk stratification, particularly for individuals with a low BMI. Handgrip strength appeared to be a less generalizable marker of risk within the general population.
In their editorial article, Grøntved and Hu comment that the protocol-based assessment of cardiorespiratory fitness in every day clinical practice for the general population is not feasible, because it is time-consuming and requires special equipment. Therefore, it is worthwhile to research alternative, cost-effective measures of fitness, as Yates et al. propose.
Yates et al. found a stronger relationship for walking pace and mortality in the tertile with the lowest BMI, which could be explained by malnutrition and sarcopenia in individuals with low BMIs. However, some participants with low BMIs may have reported slow walking pace due to subclinical disease or illness resulting in reverse causation. Another reason for the stronger relationship in the lowest BMI could be a residual confounding by smoking, which might also explain the greater cancer mortality risk in slow pace walkers. When the authors gathered data from eight cohort studies, including the one from Yates et al., an increase of 9% (95%CI: 5-14%) in all-cause mortality risk per 5 kg reduction in handgrip strength was estimated.
Alternative and feasible measures of physical fitness need to be evaluated for their prognostic value. Also, such analyses are valuable to gain further insight into the possible value of self-reported walking pace in risk prediction in the UK Biobank study and determine the validity of self-reported walking pace to discriminate between cardiorespiratory fitness levels. Finally, they state that there is a need to evaluate strategies to improve physical fitness both in healthy individuals and patients.