Association of cardiovascular magnetic resonance-derived circumferential strain parameters with the risk of ventricular arrhythmia and all-cause mortality in patients with prior myocardial infarction and primary prevention implantable cardioverter defibrillator

Table 5 Multivariable Cox regression model for appropriate ICD therapy

	Cox HR		Harrell’s	LR test
	(95%CI)	P value	C-statistic	χ²	P value
LVEF, scar border size			0.71	13.12	Reference
Added to null model:
Extent of impaired strain, per + 10%
Severely (<−5%)	1.0 (0.6, 1.5)	0.844	0.71	13.16	0.844
Moderately (−5, −10%)	1.5 (1.0, 2.2)	0.034	0.73	17.30	0.041
Mildly (−10, − 15%)	0.8 (0.5, 1.4)	0.487	0.71	13.62	0.482
Early diastolic strain rate, per − 0.25 1/s	1.1 (1.0, 1.1)	0.179	0.71	15.06	0.164
Late diastolic strain rate, per −0.25 1/s	1.1 (1.0, 1.2)	0.044	0.73	17.64	0.034
Mechanical dispersion, per + 25 ms	1.0 (0.7, 1.5)	0.815	0.71	13.18	0.812
LVEF, scar border size, acute revascularization			0.73	17.65	Reference
Added to null model:
Extent of impaired strain, per + 10%
Severely (<−5%)	0.9 (0.6, 1.4)	0.685	0.73	17.82	0.685
Moderately (−5, −10%)	1.5 (1.0, 2.2)	0.048	0.76	21.30	0.056
Mildly (−10, −15%)	0.8 (0.5, 1.3)	0.403	0.74	18.38	0.394
Early diastolic strain rate, per −0.25 1/s	1.0 (1.0, 1.1)	0.355	0.73	18.54	0.345
Late diastolic strain rate, per −0.25 1/s	1.1 (1.0, 1.2)	0.043	0.75	22.19	0.033
Mechanical dispersion, per + 25 ms	1.1 (0.8, 1.7)	0.490	0.74	18.12	0.495

Abbreviations as in Table 2. The incremental value of each LV strain parameter for the fit of the Cox regression model for the risk of appropriate ICD therapy as compared to the null model was assessed using the likelihood ratio (LR) chi-square statistic (χ²)

ISSN: 1532-429X