Dose-response relationship between physical activity and mortality in adults with noncommunicable diseases: a systematic review and meta-analysis of prospective observational studies

Abstract

Background: This study aims to investigate the relationship between post-diagnosis physical activity and mortality in patients with selected noncommunicable diseases, including breast cancer, lung cancer, type 2 diabetes, ischemic heart disease, stroke, chronic obstructive pulmonary disease (COPD), osteoarthritis, low back pain and major depressive disorder.

Methods: A systematic search was conducted of PubMed, Scopus and the Web of Science from their inception to August 2018. Additionally, the search was updated in August 2019. Eligibility criteria included prospective observational studies examining the relationship between at least three physical activity categories (e.g. low, moderate, high) and all-cause mortality as the primary outcome.

Results: In total, 28 studies were included in the meta-analysis: 12 for breast cancer, 6 for type 2 diabetes, 8 for ischemic heart disease and 2 for COPD. The linear meta-analysis revealed that each 10 metabolic equivalent task hours increase of physical activity per week was associated with a 22% lower mortality rate in breast cancer patients (Summary Hazard Ratio [HR], 0.78; 95% CI: 0.71, 0.86; I²: 90.1%), 12% in ischemic heart disease patients (HR, 0.88; 95% CI: 0.83, 0.93; I²: 86.5%), 30% in COPD patients (HR, 0.70; 95% CI: 0.45, 1.09; I²: 94%) and 4% in type 2 diabetes patients (HR, 0.96; 95% CI: 0.93, 0.99; I²: 71.8%). There was indication of a non-linear association with mortality risk reductions even for low levels of activity, as well as a flattening of the curve at higher levels of activity. The certainty of evidence was low for breast cancer, type 2 diabetes and ischemic heart disease but only very low for COPD.

Conclusion: Higher levels of post-diagnosis physical activity are associated with lower mortality rates in breast cancer, type 2 diabetes, ischemic heart disease and COPD patients, with indication of a no-threshold and non-linear dose-response pattern.

Keywords: Health promotion; Longevity; Non-communicable disease; Physical activity; Public health.

Fig. 1

PRISMA flow chart. Explanations. a. Studies with ineligible exposure did not directly measure the physical activity but another facet of movement behaviour mostly sedentary time; b. For studies excluded based on the unability to quantify physical activity, it was impossible to quantify physical activity because data presented in the paper do not allow for the transformation of the activity categories in MET-h/week (e.g. when the question was “compared to other men of your age do you intend to walk slower, faster or about the same?”)

Fig. 2

Risk of bias graph

Fig. 3

Linear dose–response meta-analysis for the association between post-diagnosis physical activity and all-cause mortality

Fig. 4

Non-linear dose–response meta-analysis for the association between post-diagnosis physical activity and all-cause mortality. a) breast cancer (n = 12); b) type 2 diabetes (n = 6); c) ischemic heart disease (n = 8); and d) COPD (n = 2). The figure includes values up to 100 MET-h/week