Table 2.
Time to first event according to atrial fibrillation (history or baseline electrocardiogram)
| No AF (n = 2834) | AF (n = 1910) | |
|---|---|---|
| Worsening HF event or cardiovascular death | ||
| n (%) | 494 (17.4) | 394 (20.6) |
| Event rate per 100 person‐years (95% CI) | 12.7 (11.6–13.8) | 14.8 (13.4–16.3) |
| HR (95% CI) a | Reference | 1.18 (1.03–1.34) |
| HR (95% CI) b | Reference | 1.03 (0.89–1.19) |
| HR (95% CI) c | Reference | 0.91 (0.79–1.05) |
| HF hospitalization or cardiovascular death | ||
| n (%) | 488 (17.2) | 389 (20.4) |
| Event rate per 100 person‐years (95% CI) | 12.5 (11.4–13.7) | 14.6 (13.2–16.1) |
| HR (95% CI) a | Reference | 1.17 (1.03–1.34) |
| HR (95% CI) b | Reference | 1.03 (0.89–1.19) |
| HR (95% CI) c | Reference | 0.91 (0.79–1.05) |
| HF hospitalization | ||
| n (%) | 287 (10.1) | 262 (13.7) |
| Event rate per 100 person‐years (95% CI) | 7.4 (6.5–8.3) | 9.8 (8.7–11.1) |
| HR (95% CI) a | Reference | 1.34 (1.13–1.59) |
| HR (95% CI) b | Reference | 1.10 (0.92–1.32) |
| HR (95% CI) c | Reference | 0.97 (0.81–1.17) |
| Cardiovascular death | ||
| n (%) | 291 (10.3) | 209 (10.9) |
| Event rate per 100 person‐years (95% CI) | 7.1 (6.3–8.0) | 7.3 (6.4–8.4) |
| HR (95% CI) a | Reference | 1.04 (0.87–1.24) |
| HR (95% CI) b | Reference | 0.95 (0.78–1.15) |
| HR (95% CI) c | Reference | 0.82 (0.68–1.00) |
| All‐cause death | ||
| n (%) | 347 (12.2) | 258 (13.5) |
| Event rate per 100 person‐years (95% CI) | 8.5 (7.6–9.4) | 9.0 (8.0–10.2) |
| HR (95% CI) a | Reference | 1.07 (0.91–1.26) |
| HR (95% CI) b | Reference | 0.95 (0.80–1.13) |
| HR (95% CI) c | Reference | 0.83 (0.69–0.98) |
| Stroke | ||
| n (%) | 54 (1.9) | 34 (1.8) |
| Event rate per 100 person‐years (95% CI) | 1.3 (1.0–1.7) | 1.2 (0.9–1.7) |
| HR (95% CI) a | Reference | 0.91 (0.59–1.40) |
| HR (95% CI) b | Reference | 0.84 (0.52–1.34) |
| HR (95% CI) c | Reference | 0.80 (0.50–1.28) |
AF, atrial fibrillation; CI, confidence interval; HF, heart failure; HR, hazard ratio.
Cause‐specific Cox regression models stratified according to diabetes mellitus status and adjusted for a history of HF hospitalization and randomized treatment allocation. No adjustment for HF hospitalization in model for all‐cause death.
Cause‐specific Cox regression models stratified according to diabetes mellitus status and adjusted for a history of HF hospitalization, randomized treatment allocation, age, sex, geographical region, heart rate, systolic blood pressure, body mass index, HF aetiology, left ventricular ejection fraction, New York Heart Association functional class, estimated glomerular filtration rate, hypertension, chronic obstructive pulmonary disease, and prior stroke/transient ischaemic attack. No adjustment for HF hospitalization in model for all‐cause death.
Fully adjusted cause‐specific Cox regression models, including adjustment for N‐terminal pro‐B‐type natriuretic peptide.