Incidence and Prognostic Impact of Incomplete Revascularization Documented by Coronary Angiography 1 Year After Coronary Artery Bypass Grafting

Complete revascularization (CR) at the time of coronary artery bypass graft (CABG) surgery improves long-term cardiac outcomes. No studies have previously reported angiographically confirmed CR rates post-CABG. This study's aim was to assess the impact upon long-term outcomes of CR versus incomplete revascularization (IR), confirmed by coronary angiography 1 year after CABG. Randomized On/Off Bypass Study patients who returned for protocol-specified 1-year post-CABG coronary angiograms were included. Patients with a widely patent graft supplying the major diseased artery within each diseased coronary territory were considered to have CR. Outcomes were all-cause mortality and major adverse cardiovascular events (MACE; all-cause mortality, nonfatal myocardial infarction, repeat revascularization) over the 4 years after angiography. Of the 1,276 patients, 756 (59%) had CR and 520 (41%) had IR. MACE was 13% CR versus 26% IR, p <0.001. This difference was driven by fewer repeat revascularizations (5% CR vs 18% IR; p <0.001). There were no differences in mortality (7.1% CR vs 8.1% IR, p = 0.13) or myocardial infarction (4% in both). Adjusted multivariable models confirmed CR was associated with reduced MACE (odds ratio 0.44, 95% confidence interval 0.33 to 0.58, p <0.01), but had no impact on mortality. In conclusion, CR confirmed by post-CABG angiography was associated with improved MACE but not mortality. Repeat revascularization of patients with IR, driven by knowledge of the research angiography results, may have ameliorated potential mortality differences. Complete revascularization (CR) at the time of coronary artery bypass graft (CABG) surgery improves long-term cardiac outcomes. No studies have previously reported angiographically confirmed CR rates post-CABG. This study's aim was to assess the impact upon long-term outcomes of CR versus incomplete revascularization (IR), confirmed by coronary angiography 1 year after CABG. Randomized On/Off Bypass Study patients who returned for protocol-specified 1-year post-CABG coronary angiograms were included. Patients with a widely patent graft supplying the major diseased artery within each diseased coronary territory were considered to have CR. Outcomes were all-cause mortality and major adverse cardiovascular events (MACE; all-cause mortality, nonfatal myocardial infarction, repeat revascularization) over the 4 years after angiography. Of the 1,276 patients, 756 (59%) had CR and 520 (41%) had IR. MACE was 13% CR versus 26% IR, p <0.001. This difference was driven by fewer repeat revascularizations (5% CR vs 18% IR; p <0.001). There were no differences in mortality (7.1% CR vs 8.1% IR, p = 0.13) or myocardial infarction (4% in both). Adjusted multivariable models confirmed CR was associated with reduced MACE (odds ratio 0.44, 95% confidence interval 0.33 to 0.58, p <0.01), but had no impact on mortality. In conclusion, CR confirmed by post-CABG angiography was associated with improved MACE but not mortality. Repeat revascularization of patients with IR, driven by knowledge of the research angiography results, may have ameliorated potential mortality differences. The Department of Veterans Affairs (VA) “Randomized On/Off Bypass (ROOBY) Trial” was designed to compare the relative efficacy of off-pump versus on-pump coronary artery bypass grafting (CABG).1Shroyer AL Grover FL Hattler B Collins JF McDonald GO Kozora E Lucke JC Baltz JH Novitzky D On-pump versus off-pump coronary-artery bypass surgery.New Engl J Med. 2009; 361: 1827-1837Crossref PubMed Scopus (887) Google Scholar The results based upon the randomized allocation of the trial have been reported, including that patients receiving on-pump CABG had better graft patency rates than patients in the off-pump arm. Overall graft patency was 85% and about 33% of patients had at least one occluded graft.1Shroyer AL Grover FL Hattler B Collins JF McDonald GO Kozora E Lucke JC Baltz JH Novitzky D On-pump versus off-pump coronary-artery bypass surgery.New Engl J Med. 2009; 361: 1827-1837Crossref PubMed Scopus (887) Google Scholar Using the subjects who had protocol-driven 1-year post-CABG coronary angiograms, the objective of this substudy was to determine the impact of angiographically apparent incomplete revascularization (IR) versus complete revascularization (CR) upon clinical outcomes over the 4 years after the post-CABG coronary angiogram. The ROOBY trial was a single-blinded, controlled, multicenter study conducted at 18 VA medical centers from February 2002 to May 2008 with prospective patient follow-up to 1 year after CABG.1Shroyer AL Grover FL Hattler B Collins JF McDonald GO Kozora E Lucke JC Baltz JH Novitzky D On-pump versus off-pump coronary-artery bypass surgery.New Engl J Med. 2009; 361: 1827-1837Crossref PubMed Scopus (887) Google Scholar Each VA medical center's Institutional Review Board, the VA Cooperative Studies Program Human Rights Committee, and VA Clinical Merit Review Board provided ROOBY trial approvals. A ROOBY follow-up study (ROOBY-FS) was separately funded by the VA Cooperative Studies Program and received Institutional Review Board approvals from the Northport VA Medical Center, the Rocky Mountain Regional VA, and the Palo Alto Health Economics Resource Center.2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar The ROOBY-FS assessed 5-year outcomes of ROOBY patients. This study was a prespecified secondary objective of ROOBY-FS and was approved by the ROOBY-FS Executive Committee during the planning phase. The full trial protocol (Clinical Trial Registration: NCT01924442) has been published.2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar As this report was a prospective cohort analysis, the Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines for cohort studies were followed. Detailed methods and results of both ROOBY and ROOBY-FS have been published.1Shroyer AL Grover FL Hattler B Collins JF McDonald GO Kozora E Lucke JC Baltz JH Novitzky D On-pump versus off-pump coronary-artery bypass surgery.New Engl J Med. 2009; 361: 1827-1837Crossref PubMed Scopus (887) Google Scholar,2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar Clinical outcome assessments were completed for all randomized patients.2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar Using VA and non-VA database merges, the Palo Alto VA Health Economics Resource Center team documented all the post-CABG primary and secondary clinical endpoints over the 4 years after the post-CABG coronary angiogram. National nurse reviewers conducted in-depth medical chart reviews centrally. An endpoints committee adjudicated all major adverse cardiovascular event (MACE) endpoints, and any differences between medical chart review and database abstracts were reconciled. The study population consisted of ROOBY patients who returned for the study's 1-year post-CABG coronary angiogram. Patients who had an interim repeat revascularization procedure prior to 1-year coronary angiography were excluded due to a higher risk of subsequent repeat revascularizations. For purposes of the generalizability of the results, patients were asked to self-report their race. Five-year post-CABG vital status was complete for all patients. There were 3 patients who were not documented to have been seen clinically within 180 days of the 5-year post-CABG time point, and these patients were thus considered lost to follow-up for the MACE endpoint.2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar Details of the methods of the angiographic core lab assessments have been published.3Hattler B Messenger JC Shroyer AL Collins JF Haugen SJ Garcia JA Baltz JH Cleveland JC Novitzky D Grover FL Off-pump coronary artery bypass surgery is associated with worse arterial and saphenous vein graft patency and less effective revascularization: results from the Veterans Affairs Randomized On/Off Bypass (ROOBY) Trial.Circulation. 2012; 125: 2827-2835Crossref PubMed Scopus (171) Google Scholar Briefly, all study patients were asked to return for a 1-year post-CABG follow-up coronary artery angiogram unless contraindicated. All baseline and 1-year coronary angiograms were scored by a core lab, which was blinded to treatment assignment. Percent stenosis was calculated by comparing the narrowest diameter of a lesion in any view to the most normal proximal or distal reference segment in the same frame. An artery with a reference diameter ≥1.5 mm and a ≥50% diameter narrowing of a diseased segment was considered to have a clinically significant stenosis. At 1-year coronary angiography, bypass grafts were evaluated for patency. The 3 primary coronary regions, consisting of the left anterior descending (LAD), circumflex, and right coronary arteries, were evaluated using prespecified criteria. Each coronary region was assessed for CR versus IR. A coronary region was defined as having CR when a patent graft with no significant lesions was supplying the major diseased artery within the region, the graft was appropriately anastomosed distal to the primary stenosis, and there were no new significant distal coronary lesions within 1 cm of the graft anastomosis. If all 3 coronary regions had CR, then the patient had CR. Otherwise the patient was considered to have IR, which included patients with graft occlusions to significantly diseased arteries and those with diseased coronary regions which were never bypassed. As this study required patients to survive and return for the 1-year post-CABG coronary angiogram, the focus was placed upon events over the 4 years after the angiogram (i.e., the 1- to 5-year post-CABG events). The primary outcomes were all-cause death and MACE (composite of all-cause death, repeat revascularization, or nonfatal myocardial infarction). Secondary endpoints included the individual MACE subcomponents and cardiac death. Study site investigators were advised to refrain from interventions during the study-driven angiograms. Subsequent decisions regarding the need for revascularization were left to the local treating physicians, and the specific clinical indications for revascularization were not captured. Baseline variables were compared between CR and IR groups using chi-square tests and Fisher exact tests (categorical endpoints) or Wilcoxon Rank Sum and t-test analyses (continuous variables) depending on normality. For time-to-event endpoints, Kaplan-Meier curves with likelihood ratio tests compared CR versus IR associations. Two multivariable adjusted logistic regression analyses were performed to evaluate which demographic, clinical, operative, and 1-year angiographic variables were independently associated with mortality and with MACE. Variables included in both models were age, ethnicity, education level, diabetes, peripheral vascular disease (PVD), chronic obstructive pulmonary disease, atrial fibrillation, stroke, serum creatinine >1.5 mg/dl (to convert to μmol/l multiply by 88.4), left ventricular ejection fraction <35%, 3-vessel coronary artery disease, left main disease ≥50%, off-pump CABG, endoscopic vein harvest, use of internal mammary artery conduit, urgent procedure, perioperative complications, and CR. Per the ROOBY-FS protocol, primary hypotheses were evaluated for statistical significance using a p value of ≤0.01 to avoid type I errors arising from multiple comparisons; however, all p values have been reported permitting independent interpretation and assessment for potential type II errors.2Shroyer AL Hattler B Wagner TH Collins JF Baltz JH Quin JA Almassi GH Kozora E Bakaeen F Cleveland JC Bishawi M Grover FL Five-year outcomes after on-pump and off-pump coronary-artery bypass.New Engl J Med. 2017; 377: 623-632Crossref PubMed Scopus (76) Google Scholar Adjusting for baseline characteristics, multivariable logistic regression analyses used a pre-established p ≤0.05 to validate bivariate study-related findings. Both the Perry Point VA Coordinating Center and the Palo Alto VA Health Economics Resource Center teams coordinated major components of the database programming and analyses to support this ROOBY-FS 1- to 5-year follow-up analysis. SAS version 9.4 (SAS Institute, Cary, NC) was used for the analyses performed. Of the 2,203 patients randomized in the ROOBY study, 1,276 returned for 1-year (range 10 to 14 months) post-CABG coronary artery angiography and were included in this substudy. The primary reason for not having a post-CABG coronary angiogram was that the patient declined the procedure (n = 711). Other reasons included patient death prior to 1 year, repeat revascularization prior to 1 year, or a contraindication to angiography such as an elevated serum creatinine >2.0 mg/dl. Compared with the study group, excluded patients had more comorbidities including previous stroke (9.6%), PVD (19%), diabetes (47%), and hypertension (89%). The vital status was known for all 1,276 study patients. Only 3 patients (1 CR and 2 IR) were not seen for a clinical visit within 180 days of the 5-year post-CABG date and were considered lost to follow-up for the MACE and MACE subcomponent endpoints. Of the 1,276 study patients, 520 (41%) had IR and 756 (59%) had CR documented by coronary angiography. Within the IR group, approximately 47% had IR of the circumflex territory, 53% had IR of the RCA territory, and 22% had IR of the LAD territory (a patient could have 1 or more territories with IR). Significant graft stenoses and/or occlusions were found in 413 (79%) of the IR group and 79 (15%) had less grafts completed than planned during the initial CABG. Others had progression of disease or postanastomotic lesions. Comparison of baseline variables between the 2 groups found the IR group was more likely to have 3-vessel coronary artery disease and to have had an off-pump procedure (Table 1). The rates of aspirin, beta-blocker, and lipid lowering medication use were ≥90% for each medication at discharge and the 1-year visit, with no difference between groups.Table 1Baseline demographic, clinical, and operative variablesVariableCR (n = 756)IR (n = 520)p valueAge (years, mean ± SD)61.9 ± 8.262.4 ± 8.20.26Men752 (99%)517 (99%)0.91White656 (87%)445 (86%)0.54Serum creatinine >1.5 mg/dl31 (4%)28 (5%)0.28COPD146 (19%)107 (21%)0.58Peripheral vascular disease99 (13%)73 (14%)0.63Previous stroke48 (6%)36 (7%)0.68Diabetes mellitus317 (42%)214 (41%)0.78Hypertension640 (85%)434 (83%)0.57Current smoker237 (31%)173 (33%)0.503-vessel disease477 (63%)367 (71%)0.006Left main disease ≥50%176 (23%)126 (24%)0.69Ejection fraction <55%238/711 (34%)167/463 (36%)0.38SAQ-AF score (mean ± SD)63.7 ± 2764.2 ± 270.72Internal mammary artery conduit729 (96%)498 (96%)0.55Radial artery conduit53 (7%)51 (10%)0.07Endoscopic vein harvesting175/509 (34%)138/324 (43%)0.02Off-pump CABG334 (44%)290 (56%)<0.001CABG = coronary artery bypass graft surgery; COPD = chronic obstructive pulmonary disease; CR = complete revascularization; IR = incomplete revascularization; SAQ-AF = Seattle Angina Questionnaire Angina Frequency; SD = standard deviation.Percentages are rounded to nearest whole number. Open table in a new tab CABG = coronary artery bypass graft surgery; COPD = chronic obstructive pulmonary disease; CR = complete revascularization; IR = incomplete revascularization; SAQ-AF = Seattle Angina Questionnaire Angina Frequency; SD = standard deviation. Percentages are rounded to nearest whole number. There was no difference between groups in 1- to 5-year all-cause mortality (7% CR, 8% IR; Table 2). The MACE rates were 26% IR and 13% CR, p <0.001. The difference in the MACE composite outcome was primarily driven by more repeat revascularizations in the IR group, and there were no differences in mortality, nonfatal myocardial infarction, or cardiac mortality. Comparison of patients with LAD IR (n = 115) versus LAD CR (n = 1,153) found similar outcomes with a difference in MACE (30% LAD IR vs 17% LAD CR, p = 0.002), but not mortality (5% LAD IR vs 8% LAD CR, p = 0.46). Also, when a lesion severity of ≥70% (rather than 50%) was used to define IR, the mortality rate was still 8%.Table 2Primary and secondary clinical outcomes in the 4 years after post-CABG coronary angiographyVariableCR (n = 756)IR (n = 520)p valueDeath54 (7%)42 (8%)0.53MACE100 (13%)134 (26%)<0.001Nonfatal MI30 (4%)21 (4%)0.95Repeat revascularization41 (5%)93 (18%)<0.001Repeat PCI41 (5%)86 (17%)<0.001Repeat CABG07 (1%)0.001Cardiac death15 (2%)16 (3%)0.21CABG = coronary artery bypass graft surgery; CR = complete revascularization; IR = incomplete revascularization; MACE = major adverse cardiovascular event; MI = myocardial infarction; PCI = coronary mortality endpoints were complete for all 1,276 patients. For the endpoints, = for CR and = for IR. Percentages are rounded to nearest whole number. Open table in a new tab CABG = coronary artery bypass graft surgery; CR = complete revascularization; IR = incomplete revascularization; MACE = major adverse cardiovascular event; MI = myocardial infarction; PCI = coronary The mortality endpoints were complete for all 1,276 patients. For the endpoints, = for CR and = for IR. Percentages are rounded to nearest whole number. was used to assess IR was independently associated with 5-year mortality or MACE. Variables included in the were off-pump versus on-pump CABG, age, ethnicity, education level, diabetes, chronic obstructive pulmonary disease, atrial fibrillation, stroke, serum creatinine >1.5 left ventricular ejection fraction <35%, 3-vessel coronary artery disease, left main disease endoscopic vein use of internal mammary artery conduit, urgent procedure, and perioperative In the mortality IR was not independently associated with death = The variables significantly associated with 5-year mortality were (odds ratio 95% confidence interval to PVD to and left ventricular ejection fraction to Table there was no with mortality, CR was significantly associated with from MACE 0.33 to 0.58, p <0.001). Variables independently associated with worse MACE were PVD to left main coronary artery disease to and major perioperative within days of the initial CABG to Table The were for the mortality and for the MACE Kaplan-Meier from MACE the IR group had significantly more MACE events in the after the coronary angiogram, and the curves to over time associated with and MACE in the 4 years after post-CABG ratio vascular vascular main stenosis perioperative = coronary artery bypass graft surgery; = 95% confidence = left ventricular ejection MACE = major adverse cardiovascular Open table in a new tab CABG = coronary artery bypass graft surgery; = 95% confidence = left ventricular ejection MACE = major adverse cardiovascular This study found IR was in of patients who prespecified coronary angiography 1 year after CABG. In to patients with CR, patients with IR had more MACE over the 4 years after angiography, primarily due to repeat revascularizations. There was no difference in mortality between This is the study to date with protocol-specified coronary angiography after CABG to report subsequent long-term outcomes based upon the angiographic of IR versus CR. of the of graft which included a protocol-specified angiogram to 18 after CABG, reported of patients had of at least one JH vein graft after coronary artery bypass from PubMed Scopus Google Scholar reported outcomes based upon the of graft and found that patients with vein graft had more repeat revascularization, but no difference in myocardial infarction or JH between vein graft and subsequent clinical outcomes after coronary artery bypass 2012; 125: PubMed Scopus Google Scholar to study, found that left internal mammary artery graft was associated with more repeat but not with worse JH B D and of internal mammary artery graft and subsequent clinical from the of Trial.Circulation. PubMed Scopus Google Scholar of a of patients who clinically as to coronary angiograms between 1 and 18 after CABG found that long-term MACE outcomes were worse in patients with significantly diseased or occluded vein JH of saphenous vein graft to outcomes coronary artery bypass J PubMed Scopus Google Scholar to of saphenous vein grafts within 1 year after bypass graft and patient angiographic follow-up of grafts to and in patients during PubMed Google Fisher M J of bypass the Study and Google JH A M D and of an for of vein graft coronary artery bypass graft a randomized PubMed Scopus Google Scholar and/or of the or the coronary arising during graft or to secondary JH vein graft after coronary artery bypass from PubMed Scopus Google vein graft review of the and PubMed Scopus Google D The impact of medication on clinical outcomes of coronary artery a J 2017; PubMed Scopus Google Scholar As the graft rate a of to patients and their during their CABG or is not that most are to in the of such are more of an than a of clinical JH vein graft and clinical a in patients coronary artery bypass PubMed Scopus Google Scholar The and impact of CR versus IR to of in and cardiac studies a with CR at the time of defined as of 1 graft to each diseased territory, however, the of subsequent graft occlusions less F outcomes chronic of of PubMed Scopus Google M A M J D revascularization and long-term after coronary artery bypass J PubMed Scopus Google The of of revascularization during CABG with multiple arterial PubMed Scopus Google The of of revascularization during long-term follow-up after coronary artery PubMed Scopus Google Scholar study found that IR documented by 1-year post-CABG angiography was associated with subsequent but not mortality. the reason mortality was not is not and study, one that of lesions on a more of subsequent interventions and an on medication any of which may have improved mortality in these D The impact of medication on clinical outcomes of coronary artery a J 2017; PubMed Scopus Google Scholar The study has Patients were with the results may not to or all randomized patients the 1-year post-CABG study angiogram, and there was a patients for angiography. there was no difference in Seattle Angina Questionnaire Angina between patients who and not return for angiography were and p = that patients were not more likely to return for B Messenger J J Bishawi M Quin JA Almassi GH Collins JF Kozora E Grover FL Shroyer Clinical and angiographic of 1 year after coronary artery bypass graft PubMed Scopus Google Scholar The study's site investigators were from during the 1-year post-CABG study angiogram. All subsequent decisions over the 4 years after the angiogram were left to the of local As by the Kaplan-Meier from MACE repeat revascularizations in the IR group were more in the year after the study angiogram, the angiographic may have clinical In conclusion, this study provided an to evaluate long-term outcomes based upon 1-year post-CABG angiographic findings. angiographic IR was and was associated with higher was not to determine post-CABG assessments for IR could long-term MACE. review and review and review and review and review and review and Almassi review and review and The that have no known or that could have to the reported in this The for and with the and and for statistical programming