Table 5.
Univariate and multivariate Cox proportional hazard analysis in patients without therapy at baseline
| Univariate | Multivariate | Multivariate | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| Age, per 1 year increase |
1.053 (1.001–1.108) |
0.046 | 1.067 (1.012–1.125) |
0.011 | 1.060 (1.007–1.116) |
0.025 |
| NYHA | 2.405 (1.450–3.988) |
0.001 | 1.752 (1.010–3.039) |
0.056 | 1.776 (0.924–3.414) |
0.085 |
| Mayo Stage | 1.985 (1.212–3.252) |
0.006 | 1.406 (0.736–2.683) |
0.30 | 1.443 (0.779–2.672) |
0.24 |
| E/E’, per 1 unit increase |
1.073 (1.003–1.114) |
0.042 | 1.235 (0.477–3.134) |
0.28 | 1.296 (0.463–3.188) |
0.40 |
| LVEF, per 1% increase |
0.966 (0.939–0.994) |
0.016 | 0.995 (0.960–1.033) |
0.81 | 0.987 (0.954–1.021) |
0.44 |
| Septal thickness, per 1 mm increase |
1.115 (1.020–1.219) |
0.017 | 1.112 (0.988–1.251) |
0.078 | 1.086 (0.973–1.213) |
0.041 |
| ECV ≥44.0% | 4.751 (1.572–14.360) |
0.006 | 4.599 (1.493–14.165) |
0.008 | – | – |
| Global LGE | 4.041 (1.452–11.246) |
0.007 | – | – | 4.442 (1.578–12.389) |
0.015 |
All significantly prognostic factors in univariate analysis were listed. Univariate analysis was not performed for native T1 because the Kaplan-Meier curves crossed each other (Tarone-Ware, P = 0.069). Univariate analysis was not performed for log (cTnI) and log (NT-proBNP), as they were included in Mayo Stage. All clinically and statistically significant variates in univariate analysis were put into the multivariate Cox model. ECV and LGE were put in separate models because of a correlation ρ of 0.889. Backward regression was chosen. HR Hazard ratio, CI Confidence interval, NYHA New York Heart Association, LVEF Left ventricle ejection fraction, LGE Late gadolinium enhancement, ECV Extracellular volume