Table 4.
Impact of treatment strategies on the primary outcome and its components according to the LVEF ranges and sensitivity analyses.
| AVI (within 6 months) vs. medical treatment | Univariate analysis of death or HHF HR (95% CI) and P-value | Multivariate analysis HR (95% CI) and P-value | ||
|---|---|---|---|---|
| All–cause death or HHF | All–cause death | HHF | ||
| Overall patients (n = 1,052)* | ||||
| LVEF ≤35% | 0.74 (0.28–1.95) P = 0.543 |
0.65 (0.21–1.97) P = 0.441 |
0.54 (0.10–2.86) P = 0.468 |
0.71 (0.17–2.92) P = 0.630 |
| LVEF 35–55% | 0.24 (0.10–0.60) P = 0.002 |
0.15 (0.04–0.50) P = 0.002 |
0.05 (0.01–0.44) P = 0.006 |
0.18 (0.05–0.66) P = 0.010 |
| LVEF >55% | 0.38 (0.16–0.90) P = 0.028 |
0.40 (0.14–1.15) P = 0.089 |
0.34 (0.09–1.23) P = 0.099 |
0.53 (0.12–2.31) P = 0.397 |
| Patients without CAD (n = 691) | ||||
| LVEF ≤35% | 0.51 (0.15–1.75) P = 0.284 |
0.52 (0.12–2.25) P = 0.381 |
0.65 (0.12–3.42) P = 0.612 |
0.21 (0.02–1.92) P = 0.165 |
| LVEF 35–55% | 0.19 (0.06–0.63) P = 0.007 |
0.21 (0.05–0.86) P = 0.030 |
0.08 (0.01–0.61) P = 0.015 |
0.29 (0.05–1.62) P = 0.160 |
| LVEF >55% | 0.32 (0.11–0.91) P = 0.033 |
0.34 (0.11–1.04) P = 0.060 |
0.46 (0.12–1.75) P = 0.253 |
0.21 (0.03–1.61) P = 0.132 |
| Patients without aortic disease (n = 881) | ||||
| LVEF ≤35% | 0.63 (0.22–1.86) P = 0.406 |
0.55 (0.16–1.97) P = 0.361 |
0.46 (0.08–2.64) P = 0.386 |
0.61 (0.11–3.35) P = 0.572 |
| LVEF 35–55% | 0.24 (0.08–0.66) P = 0.006 |
0.10 (0.03–0.43) P = 0.002 |
Adjusted log–rank† P = 0.007 |
0.15 (0.04–0.63) P = 0.010 |
| LVEF >55% | 0.38 (0.14–1.10) P = 0.074 |
0.41 (0.12–1.46) P = 0.169 |
0.29 (0.08–1.03) P = 0.056 |
0.63 (0.08–5.00) P = 0.658 |
| Patients without secondary MR (n = 935) | ||||
| LVEF ≤35% | 0.58 (0.16–2.03) P = 0.390 |
0.35 (0.09–1.47) P = 0.151 |
0.50 (0.10–2.69) P = 0.423 |
0.20 (0.02–1.91) P = 0.161 |
| LVEF 35–55% | 0.21 (0.07–0.58) P = 0.003 |
0.11 (0.03–0.37) P <0.001 | 0.05 (0.01–0.45) P = 0.007 |
0.12 (0.03–0.49) P = 0.003 |
| LVEF >55% | 0.45 (0.19–1.07) P = 0.070 |
0.45 (0.16–1.29) P = 0.139 |
0.30 (0.09–1.06) P = 0.062 |
0.63 (0.14–2.74) P = 0.535 |
| Patients without other cardiac or aortic surgery (n = 886)‡ | ||||
| LVEF ≤35% | 1.18 (0.28–5.03) P = 0.819 |
1.18 (0.24–5.78) P = 0.834 |
1.04 (0.13–8.37) P = 0.97 |
1.10 (0.13–9.53) P = 0.933 |
| LVEF 35–55% | 0.17 (0.04–0.71) P = 0.015 |
0.03 (0.01–0.23) P = 0.001 |
0.048 (0.01–0.54) P = 0.014 |
0.02 (0.01–0.21) P = 0.001 |
| LVEF >55% | 0.37 (0.12–1.20) P = 0.099 |
1.67 (0.30–9.31) P = 0.562 |
0.13 (0.02–1.03) P = 0.054 |
3.28 (0.55–19.48) P = 0.191 |
AVI, aortic valve intervention; CI, confidence interval; CAD, coronary artery disease; HHF, hospitalization for heart failure; HR, hazard ratio; LVEF, left ventricular ejection fraction; MR, mitral regurgitation.
To avoid immortal–time bias, the time–zero was the time of aortic valve intervention for the recipients and day 15 following the baseline echocardiography for the non-recipients. In addition, only early aortic valve interventions performed within 6 months of the baseline echocardiography were evaluated in order to reduce the impact of changes in LVEF and AR severity during follow-up on assessment.
Since no event had occurred in patients after aortic valve intervention within this LVEF range, the Cox model converged before the variable, resulting in an infinite coefficient; thus, adjusted log-rank test was adopted instead.
Patients who underwent concomitant cardiac or aortic surgery during AVI procedures and patients who received isolated cardiac or aortic surgery without AVI were excluded.