Improved long-term survival with coronary artery bypass graft surgery compared to percutaneous coronary intervention in diabetics with multivessel disease

Background

While randomized clinical trials have demonstrated the superiority of coronary artery bypass grafting (CABG) over percutaneous coronary intervention (PCI) in patients with diabetes and multivessel coronary artery disease (CAD), there remains a paucity of observational evidence comparing these two modalities.

Methods

Clinical and administrative databases for Canada's most populous province, Ontario, were linked to obtain records of all patients with angiographic evidence of multivessel CAD (defined as: 2-vessel and 3-vessel disease) treated with either isolated CABG or PCI from October 2008 to March 2017. Left main disease was excluded in the primary analysis. Baseline characteristics of patients undergoing CABG and PCI were compared and 1:1 propensity score matching was performed to account for baseline differences. 30-day mortality was compared in the matched groups. Late mortality and the composite of major cardiovascular and cerebrovascular events (MACCE, consisting of stroke, myocardial infarction (MI), repeat revascularization, and death) were compared between the matched groups using a stratified log rank test and Cox-proportional hazards model. The individual non-fatal components of MACCE were compared using the Fine-Gray model that accounted for death as a competing risk. A secondary analysis that included patients with left main disease was also performed for the outcome of late mortality. A sensitivity analysis that excluded patients with acute coronary syndrome was also conducted for late mortality.

Results

A total of 9,395 and 4,016 patients underwent CABG and PCI respectively. Prior to matching, CABG patients were younger (65.7 vs 68.5 years, p<0.001), more likely male (78% vs 73%, p<0.001) and with more severe CAD. Propensity score matching based on 24 baseline covariates yielded 3,782 well-balanced pairs. There was no difference in early mortality between CABG and PCI (2.3% vs 2.5%, p=0.65). The rate of all-cause mortality over 8-years was significantly higher with PCI compared to CABG (Figure- HR: 1.35, 95% CI: 1.23–1.50). The cumulative incidence of MI (HR 1.91, 95% CI: 1.66–2.20) and need for repeat revascularization (HR: 4.06, 95% CI: 3.54–4.66) were significantly higher with PCI over 8 years. There was no difference in late stroke between PCI and CABG (stroke (HR: 0.85, 95% CI: 0.68–1.07). Overall MACCE was higher in PCI compared to CABG (HR: 1.94, 95% CI: 1.80–2.09). In our secondary analysis that included patients with left main disease, findings were robust and late mortality remained higher with PCI compared to CABG (HR: 1.42, 95% CI: 1.30–1.54). In a sensitivity analysis where patients with acute coronary syndrome at the time of presentation were excluded, late mortality remained higher with PCI (HR: 1.30, 95% CI: 1.12–1.49) in 2,028 matched pairs.

Conclusions

In patients with multivessel CAD and diabetes we observed improved long-term mortality and freedom from MACCE at 8-years with CABG compared to PCI.

Figure 1

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): Canadian Institutes of Health Research

91Transcatheter valve-in-valve versus redo surgical aortic valve replacement for the management of failed biological prosthesis: early and late outcomes

Background

While the gold standard for the management of failed previous biological prosthesis was redo surgical aortic valve replacement (SAVR), valve-in-valve (ViV) transcatheter AVR (TAVR) has emerged as a less invasive option. Published studies comparing the two techniques have been small and limited to early outcomes. Herein, we compare early mortality, morbidity and late mortality between ViV TAVR and redo SAVR.

Methods

Clinical and administrative databases for Canada's most populous province, Ontario (>13 million patients), were linked to identify patients undergoing ViV and redo SAVR for a failed biological prosthesis. Baseline characteristics were compared and 1:1 propensity score matching (PSM) was performed to account for baseline differences. Standardized mean differences (SMD) were used to assess adequacy of PSM, whereby a SMD<0.10 indicated a good match. Early outcomes were compared in the matched groups using McNemar's test. In accordance to government privacy legislation, outcomes with <6 events, were presented as absolute risk difference (ARD) between ViV and Redo SAVR to prevent patient re-identification. Late mortality was compared between the matched groups using Kaplan-Meier survival curves and a Cox-proportional hazard model.

Results

Records of 558 patients undergoing intervention for a failed biological prosthesis between March 2008 to September 2017 in 11 Ontario institutions were reviewed (ViV = 214, redo SAVR = 344). Patients who underwent ViV were older (78.2±8.2 vs 69.1±11.4, p<0.001, SMD=0.92) and had more hypertension, diabetes, ischemic heart disease, atrial fibrillation, congestive heart failure compared to redo SAVR before PSM (SMD>0.20). Propensity matching on 24 variables yielded similar groups for comparison (n=133 pairs). Ages were similar between ViV and Redo SAVR (76.0±6.2 vs 76.0±8.7, SMD=0.003) along with all other comorbidities (SMD<0.1). 30-day mortality was significantly lower with ViV compared to Redo SAVR (ARD: −7.4%, 95% confidence interval (95% CI): −12.4%, −2.3%). The rate of permanent pacemaker implantation (ARD: −8.1%, 95% CI: −14.2%, −2.1%), blood transfusions (ARD: −62.2%, 95% CI: −75.2%, −49.1%) and length of stay (LOS) (mean difference: −7.8 days, 95% CI: −11.0, −4.6 days) were also lower with ViV. All-cause mortality at 5 years was similar between ViV and redo SAVR (Figure, p=0.19).

Figure 1

Conclusion

ViV TAVR was associated with lower early mortality, risk of permanent pacemaker implantation, any blood transfusion, and hospital LOS compared to redo SAVR in the largest PSM study to date. While there was no difference in late mortality at 5 years, additional studies with more subjects and longer follow-up are necessary. ViV TAVR may be the preferred approach for the treatment of failed biological prosthesis.