Investigation of a UK biobank cohort reveals causal associations of self-reported walking pace with telomere length

Abstract

Walking pace is a simple and functional form of movement and a strong predictor of health status, but the nature of its association with leucocyte telomere length (LTL) is unclear. Here we investigate whether walking pace is associated with LTL, which is causally associated with several chronic diseases and has been proposed as a marker of biological age. Analyses were conducted in 405,981 UK Biobank participants. We show that steady/average and brisk walkers had significantly longer LTL compared with slow walkers, with accelerometer-assessed measures of physical activity further supporting this through an association between LTL and habitual activity intensity, but not with total amount of activity. Bi-directional mendelian randomisation analyses suggest a causal link between walking pace and LTL, but not the other way around. A faster walking pace may be causally associated with longer LTL, which could help explain some of the beneficial effects of brisk walking on health status. Given its simple measurement and low heritability, self-reported walking pace may be a pragmatic target for interventions.

Fig. 1. Associations between self-reported walking pace and LTL.

Data presented as β-coefficient (95% CI) for “average” (n = 212,032) and “brisk” (n = 166,641) walking pace relative to “slow” (26,804) walking pace (reference). Model 1: adjusted for age, sex, ethnicity, white blood cell count. Model 2: model 1 additionally adjusted for education level, employment status, Townsend index of deprivation, fruit and vegetable intake, processed and red meat intake, oily fish intake, regularity of adding salt to food, alcohol intake, smoking status, average sleep duration, blood pressure or cholesterol medication use, diabetes diagnosis or insulin prescription, mobility limitation, and prevalent cardiovascular disease and prevalent cancer. Model 3: model 2 additionally adjusted for total physical activity volume (MET-min/week). Model 4: model 3 additionally adjusted for body mass index.

Fig. 2. Baseline exposure distribution and predicted β-coefficient (95% CI), relative to the 10th percentile reference of each exposure.

Data presented for the association of a total physical activity and b intensity gradient with LTL (n = 86,002). Models were fitted with the use of restricted cubic splines (3 evenly-spaced knots). Predicted LTL β-coefficients and histogram data shown for values between the 1st or 99th percentiles of each exposure distribution. The reference values chosen for the exposure are at the 10th percentile (total physical activity = 18.82 mg; intensity gradient −2.77), with 25th, 50th, 75th and 90th percentiles also denoted. A higher (less negative) intensity gradient (intensity distribution of physical activity) indicates more time is habitually spent in higher intensity activities (e.g., brisk walking) over a day. Model 1: is adjusted for age, sex, ethnicity, white blood cell count. Model 2: model 1 additionally adjusted for education level, employment status, Townsend index of deprivation, season of accelerometer wear, fruit and vegetable intake, processed and red meat intake, oily fish intake, regularity of adding salt to food, alcohol intake, smoking status, average sleep duration, blood pressure or cholesterol medication use, diabetes diagnosis or insulin prescription, mobility limitation, and prevalent cardiovascular disease and prevalent cancer. Model 3: model 2 additionally adjusted for either intensity gradient or total physical activity (i.e. mutual adjustment). Model 4: model 3 additionally adjusted for body mass index.