Individuals (n) | Events (n) | IR* (per 100 PY) | Subdistribution HR (95% CI) | P value | P value for trend | ||
All-cause death† | Persistent non-exercisers | 2871 | 671 | 4.8 | 1 (reference) | <0.001 | |
Exercise dropouts | 754 | 131 | 4.6 | 0.92 (0.70 to 1.22) | 0.562 | ||
New exercisers | 1363 | 174 | 3.5 | 0.73 (0.58 to 0.91) | 0.006 | ||
Exercise maintainers | 1088 | 88 | 2.9 | 0.53 (0.38 to 0.73) | <0.001 | ||
CV death | Persistent non-exercisers | 2871 | 231 | 1.5 | 1 (reference) | <0.001 | |
Exercise dropouts | 754 | 48 | 1.6 | 1.30 (0.85 to 1.99) | 0.233 | ||
New exercisers | 1363 | 53 | 1.2 | 0.76 (0.49 to 1.17) | 0.205 | ||
Exercise maintainers | 1088 | 18 | 0.6 | 0.34 (0.17 to 0.64) | 0.001 | ||
Non-CV death | Persistent non-exercisers | 2871 | 440 | 3.2 | 1 (reference) | 0.006 | |
Exercise dropouts | 754 | 83 | 3.0 | 0.73 (0.52 to 1.04) | 0.079 | ||
New exercisers | 1363 | 121 | 2.3 | 0.73 (0.56 to 0.95) | 0.021 | ||
Exercise maintainers | 1088 | 70 | 2.3 | 0.61 (0.42 to 0.90) | 0.012 |
P values were evaluated by the likelihood ratio test.
*Weighted incidence rates were computed after inverse probability of treatment weighting.
†The Fine and Gray method was not used when calculating the risk of all-cause death.
CV, cardiovascular; IR, incidence rate; PY, person-years.