Table 3

Comparison of Predictive Value of Different Variables for VT/VF

ModelsWald p ValueHR95% CILog Likelihood RatioChi-Square DifferenceLikelihood Ratio Test
p Value
Models with LVEDV
 LVEDV indexed by BSA, per 10 ml/m2<0.0011.091.05–1.123,405.2118.46<0.001
 LVEDV indexed by BSA, per 10 ml/m20.0961.040.99–1.093,397.4126.26<0.001
 RWT, per 0.01-U decrement0.0051.081.02–1.150.005
Models with LVESV
 LVESV indexed by BSA, per 10 ml/m2<0.0011.101.06–1.153,405.8717.80<0.001
 LVESV indexed by BSA, per 10 ml/m20.1341.050.99–1.113,397.9325.73<0.001
 RWT, per 0.01-U decrement0.0051.091.03–1.150.005
Models with LVEF
 LVEF, %, per unit percentage0.0210.960.93–0.993,418.425.240.022
 LVEF, %, per unit percentage0.9791.000.96–1.043,400.1123.55<0.001
 RWT, per 0.01-U decrement<0.0011.121.06–1.17<0.001
Models with LAV
 LAV indexed by BSA, per 10 ml/m2<0.0011.231.10–1.383,411.0312.63<0.001
 LAV indexed by BSA, per 10 ml/m20.1531.100.97–1.253,398.1125.55<0.001
 RWT, per 0.01-U decrement<0.0011.101.04–1.15<0.001
Models with LV mass
 LV mass indexed by BSA, per 10 g/m20.1251.050.99–1.123,421.382.290.130
 LV mass indexed by BSA, per 10 g/m20.431.030.96–1.093,399.5024.16<0.001
 RWT, per 0.01-U decrement<0.0011.111.06–1.16<0.001

Abbreviations as in Tables 1 and 2.

  • Values added to the baseline multivariate model alone and together with RWT.

  • Findings regarding risk associated with each echocardiographic parameter were obtained from separate models, each adjusted for the following covariates: treatment arm, female sex, glomerular filtration rate, previous myocardial infarction, New York Heart Association functional class, QRS duration ≥150 ms, SBP, previous VA, and age at enrollment.

  • The chi-square difference was calculated by deducting the chi-square of each model from the chi-square of the baseline model (83.30).