Dose Response of Incidental Physical Activity Against Cardiovascular Events and Mortality

Abstract

Background: Few middle-aged and older adults engage in regular leisure-time exercise. Incidental physical activity (IPA) encompasses activities of daily living outside the leisure-time domain. No dose-response study is available to guide IPA-focused interventions and guidelines. We examined the associations of device-assessed IPA intensities (vigorous [VIPA], moderate [MIPA], light [LIPA]) with major adverse cardiovascular events (MACE) and mortality, and we estimated the "health equivalence" of LIPA and MIPA against 1 minute of VIPA.

Methods: A total of 24 139 nonexercisers from the 2013 to 2015 UK Biobank accelerometry substudy (56.2% women) with a mean±SD age of 61.9±7.6 years were analyzed using a prospective cohort design. IPA energy expenditure and daily durations of VIPA, MIPA, and LIPA were calculated with a validated machine learning-based intensity classifier. MACE included incident stroke, myocardial infarction, and heart failure; CVD death; CVD mortality; and all-cause mortality.

Results: Analyses included 22 107 (MACE), 22 174 (CVD mortality), and 24 139 (all-cause mortality) participants, corresponding to 908/223/1071 events over 7.9 years of follow-up. IPA volume exhibited an L-shaped association with a nadir at ≈35 to 38 kJ·kg^-1·d^-1, corresponding to hazard ratios of 0.49 (95% CI, 0.39-0.61) for MACE, 0.33 (95% CI, 0.22-0.52) for CVD mortality, and 0.31 (95% CI, 0.25-0.38) for all-cause mortality. Any amounts of VIPA or MIPA were associated with lower risk, with a plateau of ≈14 minutes per day (VIPA) and 34 to 50 minutes per day (MIPA). The median VIPA (4.6 min/d) and MIPA (23.8 min/d) durations were associated with CVD mortality hazard ratio of 0.62 (95% CI, 0.46-0.83) and 0.50 (95% CI, 0.31-0.80), respectively. LIPA showed a subtle inverse gradient which was statistically significant only for CVD mortality at levels >130 minutes per day. One minute of VIPA was equivalent to 2.8 (MACE) to 3.4 (CVD mortality) minutes of MIPA and 34.7 (CVD mortality) to 48.5 (MACE) minutes of LIPA.

Conclusions: Any daily IPA amount of vigorous or moderate intensity was associated with lower CVD risk in a dose-response manner. LIPA had weak associations with all outcomes. One minute of vigorous or ≈3.0 to 3.5 minutes of moderate IPA was associated with a similar degree of lower CVD risk. Our findings highlight the potential cardiovascular health value of incidental physical activity, especially for people who struggle to do structured exercise.

Keywords: accelerometry; cardiovascular diseases; cohort studies; epidemiology; exercise; fitness trackers; machine learning; wearable electronic devices.

Figure 1.

Adjusted dose-response associations of total daily volume of IPA with overall MACE, CVD mortality, and all-cause mortality. A, Total major adverse cardiovascular events (MACE; n=22 107; events=908). B, Cardiovascular disease (CVD) mortality (n=22 174; events=223). C, All-cause mortality (ACM; n=24 139; events=1071). Dashed lines represent hazard ratios (HRs) and shaded areas represent their 95% CIs. Analyses were adjusted for sex, age, education, ethnicity, fruit and vegetable consumption, smoking history, alcohol consumption, sleep duration, discretionary screen time, previous cancer incidence, CVD-related medication use (insulin, blood pressure, and cholesterol), and family history of cancer and CVD. The reference point was the minimum value of the energy expenditure of physical activity (7.73 kJ·kg⁻¹·d⁻¹ PAEE). ACM analysis was additionally adjusted for previous incidence of CVD. All analyses excluded participants who had an event in the first year of follow-up and prevalent major CVD diagnosis at or before the accelerometry baseline. The circle shows the HR associated with the median PAEE value; the square shows the nadir of the dose-response curve. PAEE indicates physical activity energy expenditure.

Figure 2.

Adjusted absolute risk–based dose-response associations of daily incidental VIPA, MIPA, and LIPA with overall MACE, CVD mortality, and all-cause mortality. A, Total major adverse cardiovascular events (MACE; n=22 107; events=908). B, Cardiovascular disease (CVD) mortality (n=22 174; events=223). C, All-cause mortality (ACM; n=24 139; events=1071). Dashed lines represent hazard ratios, and shaded areas represent their 95% CIs. Analyses were adjusted for sex, age, education, ethnicity, fruit and vegetable consumption, smoking history, alcohol consumption, sleep duration, discretionary screen time, previous cancer incidence, CVD-related medication use (insulin, blood pressure, cholesterol), and family history of cancer and CVD. ACM analysis was additionally adjusted for previous incidence of CVD. Each physical activity (PA) intensity-specific spline model was mutually adjusted for PA energy expenditure volume from other intensities with established methods. All analyses excluded participants who had an event in the first year of follow-up and prevalent major CVD diagnosis at or before the accelerometry baseline. LIPA indicates light-intensity incidental physical activity; MIPA, moderate-intensity incidental physical activity; and VIPA, vigorous-intensity incidental physical activity.

Figure 3.

Adjusted dose response associations of daily incidental VIPA, MIPA, and LIPA physical activity with overall MACE, CVD mortality, and all-cause mortality. A, Total major adverse cardiovascular events (MACE; n=22 107; events=908). B, Cardiovascular disease (CVD) mortality (n=22 174; events=223). C, All-cause mortality (ACM; n=24 139; events=1071). Dashed lines represent hazard ratios (HRs); shaded areas represent their 95% CIs. Analyses were adjusted for sex, age, education, ethnicity, fruit and vegetable consumption, smoking history, alcohol consumption, sleep duration, discretionary screen time, previous cancer incidence, CVD-related medication use (insulin, blood pressure, cholesterol), and family history of cancer and CVD. All-cause mortality analysis was additionally adjusted for previous incidence of CVD. Each physical activity (PA) intensity-specific spline model was mutually adjusted for PA energy expenditure volume from other intensities with established methods. Referent data point was set to 0 for VIPA and MIPA splines and to the minimum value of LIPA (33.2 min/d) in the corresponding models. All analyses excluded participants who had an event in the first year of follow-up and prevalent major CVD diagnosis at or before the accelerometry baseline. The circle shows the HR associated with the median VLIPA value; and the square shows the nadir of the dose-response curve. LIPA indicates light-intensity incidental physical activity; MIPA, moderate-intensity incidental physical activity; and VIPA, vigorous-intensity incidental physical activity.

Figure 4.

Adjusted dose-response associations of daily incidental VIPA, MIPA, and LIPA physical activity with overall MACE, CVD mortality, and all-cause mortality, excluding frail participants (n=1924). A, Total major adverse cardiovascular events (MACE; n=19 932; events=800). B, Cardiovascular disease (CVD) mortality (n=19 989; events=197). C, All-cause mortality (ACM; n=21 785; events=937). Dashed lines represent hazard ratios (HRs), and shaded areas represent their 95% CIs. Analyses were adjusted for sex, age, education, ethnicity, fruit and vegetable consumption, smoking history, alcohol consumption, sleep duration, discretionary screen time, previous cancer incidence, CVD-related medication use (insulin, blood pressure, cholesterol), and family history of cancer and CVD. ACM analysis was additionally adjusted for previous incidence of CVD. Each physical activity (PA) intensity-specific spline model was mutually adjusted for PA energy expenditure volume from other intensities estimated with established methods. Referent data point was set to 0 for VIPA and MIPA splines and to the minimum value of LIPA in the corresponding models. All analyses excluded participants who had an event in the first year of follow-up and prevalent major CVD diagnosis at or before the accelerometry baseline. The circle shows the HR associated with the median exposure value; the square shows the nadir of the dose-response curve. LIPA indicates light-intensityincidental physical activity; MIPA, moderate-intensity incidental physical activity; and VIPA, vigorous-intensity incidental physical activity.