Figure 1. Adjusted Hazard Ratios and Cumulative Incidences of Long-term Outcomes for Multiple Arterial Grafting (MAG) vs Left Internal Thoracic Artery Supplemented by Saphenous Vein Grafts (LITA+SVG).

A-C, The adjusted cumulative incidences for mortality and repeated revascularization were estimated by incorporating weights that were obtained from propensity score (PS) weighting for the overall study cohort (LITA+SVG, n = 14496; MAG, n = 5580) at the end of 15-year follow-up. The adjusted cumulative incidences for myocardial infarction, stroke, heart failure, and the composite end point were estimated by incorporating weights that were obtained from PS weighting for the Discharge Abstract Database (DAD)-linked study cohort (LITA+SVG, n = 7912; MAG, n = 2633) at the end of 7-year follow-up. Cumulative incidences of mortality and the composite end point were estimated by the Kaplan-Meier method. For nonfatal long-term outcomes, cumulative incidences were estimated by the cumulative incidence function accounting for death as a competing risk. Adjusted hazard ratios were estimated from univariable Cox regression models that incorporated weights that were obtained from PS weighting. P values were calculated using the weighted Wald χ² test. B and C, The number of patients at risk was calculated by incorporating weights that were obtained from PS weighting (ie, weight of 1 for patients in the group receiving MAG and a weight equal to the odds of receiving MAG for patients in the group receiving LITA+SVG) and rounded to the nearest integer for the group receiving LITA+SVG.