Rationale: Physical activity (PA) has been associated with mortality in patients with COPD. We aimed to investigate 1) this association across the disease severity and 2) whether the previously reported protective effect of PA is due to its quantity and/or intensity. Methods: We pooled data about PA (Sensewear armband), clincial characteristics and mortality from patients with COPD tested in Belgium (n=179), Spain (n= 173), Germany (n=168), United Kingdom (n=101) and Brazil (n=137) for these analyses. The quantity (daily step count) and intensity (mean METs during periods of at least 1.5 METs) of PA were included as exposures. Analyses (Cox proportional hazard regression) were adjusted for age, gender, severity of airflow limitation and BMI. Results: 758 patients (74% male, 67±8 years, FEV 1 56±22 % pred , BMI 27±5 kg.m −2 , 5881±4007 steps/day, mean PA intensity 3.18±0.52 METs) were included and 149 (20%) died during the 4-year follow up. Being more active was related to lower mortality risk (Hazard ratio, 95%CI 0.85 [0.79 to 0.91] per 1000 steps more, p<0.001). The association between PA quantity and mortality was similar across GOLD stages (interaction effect p=0.39, Figure 1A) and across intensity levels (interaction effect p=0.67, Figure 1B).
Conclusion: The strength of the association between PA quantity and mortality is comparable across GOLD stages and across intensity levels.