Table 6.

Predictivity of the Multivariable Models Without and After Inclusion of Bleeding

Multivariable Model	30‐Day Mortality
	P Valuea	C‐Statistic [95% CI]
Baseline model without bleedingb	—	0.744 [0.671 to 0.818]
After inclusion of VARC‐2 LT bleeding	<0.001	0.773 [0.706 to 0.839]
After inclusion of bleeding defined as BARC ≥3	0.002	0.776 [0.694 to 0.857]
After inclusion of bleeding defined as BARC ≥2	0.011	0.762 [0.682 to 0.842]
After inclusion of TIMI major	0.001	0.768 [0.692 to 0.844]
After inclusion of TIMI major+minor bleeding	0.002	0.776 [0.701 to 0.852]
After inclusion of GUSTO severe or LT bleeding	<0.001	0.791 [0.714 to 0.869]
After inclusion of GUSTO severe or LT and moderate bleeding	0.002	0.771 [0.692 to 0.851]
After inclusion of modified VARC LT bleeding	<0.001	0.814 [0.759 to 0.870]

BARC indicates Bleeding Academic Research Consortium; BMI, body mass index; CI, confidence interval; GUSTO, Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries investigators; LT, life‐threatening; TIMI, Thrombolysis in Myocardial Infarction investigators; VARC, Valve Academic Research Consortium.

^aThe P value is obtained from a likelihood ratio test comparing the baseline model without bleeding vs multivariable model including bleeding.

^bThe baseline model includes age, sex, BMI, previous stroke, chronic obstructive pulmonary disease, and atrial fibrillation at baseline.