This 3 x 3 table of 1-year probabilities of event-free survival for the 9 possibilities of concordant (white boxes) and discordant (blue boxes) interactions among the 3 treatments shows discordance to have the greatest negative effect in the population randomized to percutaneous coronary intervention (PCI) and the least effect on the population randomized to coronary artery bypass graft (CABG). Reprinted with permission from Pereira et al. (6).

Kaplan-Meier curves stratified by concordance (green lines) and discordance (black lines) for (A) all patients randomized to PCI, (B) all patients randomized to CABG, and (C) all patients randomized to medical treatment show significant differences in the PCI randomized cohort for whom the outcome of this treatment was better if it was in concordance with the preferred treatment of clinicians. Abbreviations as in Figure 1. Reprinted with permission from Pereira et al. (6).

Kaplan-Meier curves adjusted by Cox analysis for baseline clinical differences and propensity for treatment and stratified by coronary artery disease (CAD) severity were created for the medically treated cohort and a combined revascularization cohort treated by PCI or CABG. The 3 medical curves were subtracted from the 3 corresponding revascularization curves, and their differences were plotted over the 17-year follow-up interval. The absolute percent difference over time depicts the patients out of 100 who survived with revascularization who would not have survived if treated with medical therapy. Abbreviations as in Figure 1. Adapted with permission from Smith et al. (7).

These data derive from subtracting 17-year PCI from CABG Kaplan-Meier survival data after adjustment for baseline clinical differences and propensity adjustment. Survival differences expressed as the extra patients projected to survive for each 100 treated by CABG compared with PCI. Little significant difference between the 2 interventional strategies is apparent in the low-severity and intermediate-severity CAD group. However, the high-severity CAD group achieved a survival advantage of 6 patients per 100 from revascularization by CABG compared with PCI by 3 years after the procedure that remained through year 12. Abbreviations as in Figures 1 and 3. Adapted with permission from Smith et al. (7).

Additional months of survival added over 15 years of follow-up for the average patient related to choice of treatment are depicted by severity of CAD as low (black), intermediate (red), or high (white). *p < 0.05. Med = medical therapy; Revasc = revascularization; other abbreviations as in Figure 3. Reprinted with permission from Smith et al. (7).

The annual percent of isolated CABG procedures in the Society of Thoracic Surgeons’ National Cardiac Database for which a PCI had been performed within 6 h and the subset also had a myocardial infarction within 24 h of CABG declined from 1994 to 1999. Abbreviations as in Figure 1. Reprinted with permission from Haan et al. (10).

These data depict the percent of all total isolated CABG mortality that occurred in those patients who underwent surgery within 6 h of PCI (overall) and the subset sustaining a myocardial infarction (MI) <24 h before CABG. Although the incidence of urgent PCI decreased to 3 in 1,000 CABG operations, it still contributed 10% of all CABG deaths. Abbreviations as in Figure 1. Reprinted with permission from Haan et al. (10).

These institution-specific mortality rates by Risk Adjustment for Congenital Heart Surgery (RACHS-1) category showed little diversity among the 3 lowest levels of risk. Institution mortality rates seemed to vary more at higher RACHS-1 categories as the complexity of the congenital disorder increased and the numbers of patients who underwent surgery at each institution decreased. Data analysis confirmed that this diversity was not statistically significant. Reprinted with permission from Welke et al. (13).

These competing analysis outcomes show fewer patients supported by left ventricular assist device (LVAD) to have received a transplant during (A) the early era of LVAD use, 1993 to 1999, n = 39, compared with (B) the recent era of expanded LVAD use, 2000 to 2003, n = 50, when more transplantations were performed, thereby reducing the proportion of patients waiting for transplants and those dying while waiting. Reprinted with permission from Blume et al. (19).

The total death rate is compared with the mean age of patients in each of 15 series reporting mechanical (filled squares) and 23 series reporting bioprosthetic (open squares) valves in relation to the mean age of patients included in each series. The areas of squares are proportional to total patient years of follow-up. Patients in series receiving bioprosthetic valves had an older average age, but the relationship between average reported age and death rate of the reporting series seems to have a similar relationship for mechanical and bioprosthetic valves. Reprinted with permission from Lund and Bland (22).

Survival free from postoperative atrial fibrillation is stratified by left atrial volume and age over the first 7 days after cardiac surgery. Reprinted with permission from Osranek et al. (24).