2019 Volume 83 Issue 1 Pages 122-129
Background: The long-term outcomes of complete revascularization (CR) in patients with left ventricular (LV) dysfunction undergoing coronary artery bypass grafting (CABG) remain unclear.
Methods and Results: We evaluated a consecutive series of 111 patients with LV ejection fraction ≤35% who underwent isolated first-time CABG: 63 underwent CR and 48 underwent incomplete revascularization (IR). At a median follow-up of 10.1 years, the rates of death from any cause, cardiac death, and major adverse cardiac and cerebrovascular events (MACCE) were significantly greater in the IR group. After adjusting for propensity score, no significant difference was found between the CR and IR groups regarding death from any cause (hazard ratio [HR], 1.45; 95% CI: 0.75–2.81; P=0.271) and cardiac death (HR, 1.45; 95% CI: 0.68–3.10; P=0.337). In contrast, IR increased the risk of MACCE (HR, 1.92; 95% CI: 1.08–3.41; P=0.027), which was principally attributed to an increased risk of repeat revascularization (HR, 3.92; 95% CI: 1.34–11.44; P=0.013).
Conclusions: Although IR was not significantly associated with an increased risk of long-term mortality in patients with LV dysfunction who underwent CABG, CR might reduce the risks of repeat revascularization and subsequent MACCE.
Based on the current guidelines, coronary artery bypass grafting (CABG) is recommended as the preferred or reasonable therapy of choice for significant coronary artery disease (CAD) and severe left ventricular (LV) dysfunction (LV ejection fraction [LVEF] ≤0.35).1,2 The results of the Surgical Treatment for Ischemic Heart Failure trial has supported these recommendations because of the long-term survival benefit of CABG over medical therapy in patients with significant CAD and severe LV dysfunction.3,4 Therefore, CABG is considered the standard treatment for ischemic LV dysfunction; complete revascularization (CR) in such patients, however, remains controversial.
Editorial p 25
Although CR is one of the most important goals of CABG,5,6 CR is not always realistically achievable due to the variations in the complexity of coronary lesions as well as in the patient clinical status and characteristics, particularly in patients with LV dysfunction.7,8 In addition, recent guidelines do not formally address the issue of CR in detail, and very few clinical studies to date have been conducted to evaluate the impact of CR vs. incomplete revascularization (IR) in patients with LV dysfunction due to ischemic heart disease.5,9
In the present study, we investigated the long-term results after CR and IR in patients with LV dysfunction who underwent CABG.
We studied a consecutive series of 111 patients with LVEF ≤35% who underwent CABG at Mitsui Memorial Hospital between January 1994 and January 2014. The inclusion criterion was isolated first-time CABG. Exclusion criteria were previous cardiac surgery, combined procedures, recent myocardial infarction (MI), and cardiogenic shock at the time of index procedure. LVEF was preoperatively measured on echocardiography using the Simpson biplane method in all patients. The Institutional Review Board approved this retrospective study, and the need for written consent was waived.
Diagnostic coronary angiograms and operative reports of each patient were independently reviewed to assess CR. CR was defined as bypassing all lesions with >50% diameter stenosis in vessels ≥1.5 mm, as estimated on the diagnostic angiogram.10 A totally occluded vessel could also be considered as a target vessel. Left main trunk stenosis required bypass grafting to both the left anterior descending (LAD) and left circumflex (LCX) arteries to achieve CR. We also assessed the number of target distal vessels and the complexity of coronary disease, expressed as Synergy between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery (SYNTAX) score,11 which was calculated using the online calculator (available at www.syntaxscore.com). The patients were divided into 2 groups: 63 (57%) and 48 patients (43%) who underwent CR and IR, respectively. For the patients who underwent IR, we assessed the location of diseased but ungrafted vessels as well as the reason documented for not grafting an artery.
The baseline characteristics, which were retrieved from the hospital charts, were determined based on the current guidelines.12,13 The preoperative estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease formula for Japanese patients: eGFR (mL/min/1.73 m2)=194×(serum creatinine [mg/dL]−1.094×(age [years])−0.287×0.739 (if female).14
CABG TechniqueAll surgery was performed on- or off-pump with the intention of achieving CR. Median sternotomy was performed in all patients. For patients undergoing on-pump CABG, myocardial preservation during cardiopulmonary bypass involved intermittent, antegrade cold-blood cardioplegia. For patients undergoing off-pump CABG, a single deep pericardial traction suture technique was used to expose the lateral and posterior walls. A tissue stabilizer was used to reduce the amplitude of ventricular wall movement, and a soft plastic intracoronary artery shunt was used in all anastomoses.
The basic strategy for CABG was in situ grafting of the internal thoracic artery (ITA) to the LAD, which was mostly achieved using left ITA grafting. The LCX and/or diagonal branches were revascularized using the second ITA and/or saphenous vein grafts. The right coronary artery (RCA) segments were revascularized using the saphenous vein grafts or right gastroepiploic artery.
Clinical Endpoints and Follow-upThe endpoints of the present study were death from any cause, cardiac death, and major adverse cardiovascular and cerebrovascular events (MACCE) during follow-up after CABG. Death was regarded as cardiac in origin unless obvious non-cardiac causes were identified. MACCE were defined as a composite of death from any cause, non-fatal MI, non-fatal stroke, repeat revascularization, and readmission for heart failure. Non-fatal MI was defined as the presence of typical chest pain accompanied by elevated cardiac enzymes and typical changes associated with MI on 12-lead electrocardiography. Non-fatal stroke was defined as clinical evidence or a focal/global defect on computed tomography.
Follow-up angiogram was obtained for 91 patients (82.0%) at least once postoperatively. Patients who died in the same hospitalization period as the index CABG, those who refused angiography, or those with renal dysfunction were excluded from the angiographic follow-up. Graft patency was graded based on the FitzGibbon classification:15 grade A is an excellent graft, grade B is an impaired graft with stenosis >50%, and grade O is a completely occluded graft (grade A+B=patent).
Clinical follow-up was mainly obtained from the review of outpatient hospital charts. Additional follow-up information was collected through telephone interviews with the patients, relatives, or referring physicians, and mortality and/or any complications that occurred during follow-up were recorded. Follow-up examinations were completed in 92.8% of the patients with a median duration of 10.1 years (95% CI: 8.4–12.2). The follow-up ended on 31 December 2017.
Statistical AnalysisContinuous variables are expressed as mean±SD and were compared using the Student t-test. Categorical variables are expressed as proportions and were compared using the chi-squared or Fisher’s exact test. Survival curves were estimated using the Kaplan-Meier method with log-rank test to assess differences between the groups. The survival/freedom from event time started during surgery and ended at death/event or at the last follow-up.
Considering the observational nature of this study, propensity score (PS) adjustment was performed to minimize potential confounding and selection biases in the baseline patient characteristics between groups.16 The probability of undergoing IR for each patient was calculated using multivariate logistic regression analysis based on pre-specified clinically relevant covariates, such as age, New York Heart Association class III or IV, hypertension, hyperlipidemia, diabetes, extracardiac arteriopathy, eGFR, previous stroke, previous percutaneous coronary intervention, previous MI, number of target vessels, SYNTAX score, preoperative LVEF, off-pump CABG, and year of surgery. On receiver operating characteristic curve analysis, the C-statistics of the logistic model was 0.81, indicating a strong ability to differentiate between CR and IR.
PS and IR were included as covariates in adjusted linear or logistic regression models (depending on whether the outcome variable was continuous or categorical) comparing baseline characteristics between groups. The association of IR with long-term outcome was adjusted for PS by inclusion in the Cox proportional hazards model as a covariate. The results are presented as adjusted HR with 95% CI. PS-adjusted overall survival curves and freedom from MACCE were generated using the method described by Cole and Hernán.17
All reported P-values were 2-sided, and P<0.05 was considered statistically significant. We analyzed all data using JMP version 13 (SAS Institute, Cary, NC, USA) and SAS version 9.4 (SAS Institute).
Baseline patient characteristics are summarized in Table 1. A trend was observed for higher age and lower preoperative eGFR in the IR group (P=0.064 and 0.058, respectively). With respect to the angiographic features, the IR group had a significantly higher SYNTAX score (31.4±8.7 and 37.1±9.0 in the CR and IR groups, respectively; P=0.001) and higher number of target distal vessels (3.7±1.0 and 4.6±1.0 in the CR and IR groups, respectively; P<0.0001).
| Variable | CR (n=63) |
IR (n=48) |
P-value | Adjusted P-value† |
|---|---|---|---|---|
| Age (years) | 62.0±9.3 | 65.5±10.4 | 0.064 | 0.997 |
| Male | 59 (94) | 43 (90) | 0.497 | –‡ |
| NYHA class III or IV | 36 (57) | 30 (63) | 0.569 | 0.999 |
| Hypertension | 48 (76) | 36 (75) | 0.885 | 0.999 |
| Hyperlipidemia | 36 (57) | 28 (58) | 0.900 | 0.999 |
| Diabetes | 33 (52) | 26 (54) | 0.852 | 0.999 |
| Extracardiac arteriopathy | 19 (30) | 19 (40) | 0.300 | 0.994 |
| eGFR (mL/min/1.73 m2) | 54.6±28.3 | 44.6±25.8 | 0.058 | 0.997 |
| Previous stroke | 25 (40) | 22 (46) | 0.516 | 0.998 |
| Previous MI | 47 (75) | 37 (77) | 0.763 | 0.999 |
| Previous PCI | 15 (24) | 12 (25) | 0.885 | 0.999 |
| LVEF (%) | 28.2±4.5 | 27.9±5.0 | 0.808 | 0.999 |
| SYNTAX score | 31.4±8.7 | 37.1±9.0 | 0.001 | 0.993 |
| No. target distal vessels | 3.7±1.0 | 4.6±1.0 | <0.0001 | 0.988 |
| Off-pump CABG | 23 (37) | 22 (46) | 0.322 | 0.996 |
| Operation after 2004 | 34 (54) | 32 (67) | 0.177 | 0.995 |
Data given as mean±SD or n (%). †Adjusted for propensity score by inclusion as a covariate. ‡Not included in the propensity score model because of the small number of female patients. CABG, coronary artery bypass grafting; CR, complete revascularization; eGFR, estimated glomerular filtration rate; IR, incomplete revascularization; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention.
Surgical and postoperative data are listed in Table 2. No significant difference was found in the use of off-pump surgery or bilateral ITA grafting between the groups. The number of distal anastomoses was significantly lower in the IR group (3.7±1.0 and 3.2±1.0 in the CR and IR groups, respectively; P=0.004). We identified 68 ungrafted coronary arteries in the IR group. Of these, the reasons for IR were diffusely diseased and narrowed vessel (n=27; 40%); severely calcified vessel (n=10; 15%); infarcted area (n=4; 5.9%); difficult exposure and patient instability (n=4; 5.9%); technical reasons including lack of availability of a suitable conduit (n=4; 5.9%); and intramyocardial location (n=1; 1.5%). No reason was specified in the remaining 18 cases. Most of the non-grafted coronary artery territory in patients with IR was the circumflex artery territory (46%), followed by the RCA system (38%) and LAD territory (16%). Of note, the ITA was grafted to the LAD in all but 1 case in the IR group. No significant differences were found between the groups with regard to the frequency guideline-directed medication use.
| Variable | CR (n=63) |
IR (n=48) |
P-value |
|---|---|---|---|
| Use of ITA | 59 (94) | 47 (98) | 0.387 |
| Use of BITA | 15 (24) | 10 (21) | 0.710 |
| No. distal anastomoses | 3.7±1.0 | 3.2±1.0 | 0.004 |
| Postoperative angiography (n=315 grafts)† | 0.798 | ||
| FitzGibbon A | 156/174 (90) | 123/141 (87) | |
| FitzGibbon B | 13/174 (7.4) | 13/141 (9.2) | |
| FitzGibbon O | 5/174 (2.9) | 5/141 (3.5) | |
| ≥1 suboptimal graft† | 15/52 (29) | 16/39 (41) | 0.226 |
| Medications at discharge‡ | |||
| Aspirin | 60 (100) | 46 (100) | – |
| Statin | 33 (55) | 30 (65) | 0.288 |
| β-blocker | 39 (65) | 35 (76) | 0.218 |
| ACEI or ARB | 47 (78) | 35 (76) | 0.784 |
Data given as mean±SD, n (%) or n/total (%). †Overall, 91 patients underwent postoperative angiography. ‡Excluding in-hospital mortality. ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BITA, bilateral internal thoracic artery; ITA, internal thoracic artery. Other abbreviations as in Table 1.
The median duration of postoperative angiographic follow-up was 30 days (IQR, 17–85 days). Graft patency (i.e., FitzGibbon A or B) was 97.1% and 96.5% in the CR and IR groups, respectively (P=0.798; Table 2).
The in-hospital mortality rate was 4.8% (3/63) and 4.2% (2/48) for the CR and IR groups, respectively. During follow-up, death from any cause occurred in 58 of 111 patients (52%), and the 10-year overall survival rate was 49.5%±5.8% (Table 3; Figure 1A). A total of 44 patients (40%) died from cardiac causes, and the 10-year freedom from cardiac death was 63.7%±5.6% (Table 3; Figure 1B). MACCE occurred in 82 patients (74%), and the 10-year freedom from MACCE was 21.3%±5.6% (Table 3; Figure 1C).
| Variables | CR (n=63) |
IR (n=48) |
Unadjusted HR† (95% CI) |
P-value | Adjusted HR†,‡ (95% CI) |
P-value‡ |
|---|---|---|---|---|---|---|
| Death from any cause | 28 (44) | 30 (63) | 1.82 (1.07–3.11) | 0.026 | 1.45 (0.75–2.81) | 0.271 |
| Cardiac death | 21 (33) | 23 (48) | 1.92 (1.04–3.59) | 0.036 | 1.45 (0.68–3.10) | 0.337 |
| MACCE | 43 (68) | 39 (81) | 1.60 (1.02–2.51) | 0.039 | 1.92 (1.08–3.41) | 0.027 |
| Non-fatal MI | 0 (0) | 2 (4) | – | – | – | – |
| Non-fatal stroke | 10 (16) | 5 (10) | 0.67 (0.21–1.89) | 0.456 | 0.81 (0.21–2.74) | 0.745 |
| Repeat revascularization | 10 (16) | 12 (25) | 1.82 (0.78–4.33) | 0.165 | 3.92 (1.34–11.44) | 0.013 |
| Readmission for heart failure | 16 (25) | 14 (29) | 1.42 (0.68–2.94) | 0.346 | 1.35 (0.55–3.22) | 0.509 |
Data given as n (%), unless otherwise indicated. †IR:CR. ‡Propensity score adjustment. MACCE, major adverse cardiac and cerebrovascular events. Other abbreviations as in Table 1.
Kaplan-Meier curve for (A) overall survival, (B) cardiac death-free survival, and (C) major adverse cardiac and cerebrovascular event (MACCE)-free survival after coronary artery bypass grafting (CABG).
In unadjusted comparisons between the groups, the rate of death from any cause was significantly greater in the IR group (HR, 1.82; 95% CI: 1.07–3.11; P=0.026). Similarly, IR was associated with a significantly increased rate of death from cardiac causes (HR, 1.92; 95% CI: 1.04–3.59; P=0.036) and MACCE (HR, 1.60; 95% CI: 1.02–2.51; P=0.039; Table 3). On Kaplan-Meier analysis there was a statistically significant difference in the rates of death and MACCE between groups (Figure 2). Cumulative survival at 10 years was 62.1%±7.1% and 34.1%±8.5% in the CR and IR groups, respectively (log-rank P=0.024). Freedom from cardiac death at 10 years was 76.2%±7.1% in the CR group and 48.2%±9.2% in the IR group (log-rank P=0.033), and freedom from MACCE at 10 years was 31.7%±7.0% and 7.0%±4.7%, respectively (log-rank P=0.034).
Kaplan-Meier curves for (A) overall survival, (B) cardiac death-free survival, and (C) major adverse cardiac and cerebrovascular event (MACCE)-free survival for complete revascularization (CR) vs. incomplete revascularization (IR) after coronary artery bypass grafting (CABG).
After PS adjustment, no significant between-group differences were found in the incidence of death from any cause (HR, 1.45; 95% CI: 0.75–2.81; P=0.271) or cardiac death (HR, 1.45; 95% CI: 0.68–3.10; P=0.337; Table 3; Figure 3A,B). In contrast, after PS adjustment the risk of MACCE in the IR group (HR, 1.92; 95% CI: 1.08–3.41; P=0.027) was still greater than that in the CR group (Table 3; Figure 3C). This was principally attributed to an increased risk of repeat revascularization (HR, 3.92; 95% CI: 1.34–11.44; P=0.013). Of 22 patients who had repeat revascularization, stenosis or occlusion of graft was the cause of repeat revascularization in 15 patients (CR group, 7/10 patients; IR group, 8/12 patients).
Kaplan-Meier curves after propensity score adjustment for (A) overall survival, (B) cardiac death-free survival, and (C) MACCE-free survival for CR vs. IR after CABG. Abbreviations as in Figure 2.
Of 91 patients who underwent postoperative angiogram, we identified 31 with 36 failed grafts (at least 1 suboptimal graft of FitzGibbon B or O; Table 2). Completely occluded grafts (FitzGibbon O) were mostly anastomosed to RCA or LCX territories (8 of 10 grafts), while 7 of 26 impaired grafts (FitzGibbon B) were anastomosed to LAD. The relationship between postoperative graft failure and long-term outcome was assessed by including graft failure in the PS-adjusted Cox models with IR; both graft failure (HR, 3.09; 95% CI: 1.31–7.44; P=0.010) and IR (HR, 3.71; 95% CI: 1.28–11.06; P=0.016) were independent predictors of repeat revascularization. In contrast, graft failure was not significantly associated with death from any cause (HR, 1.35; 95% CI: 0.71–2.50; P=0.355), cardiac death (HR, 1.46; 95% CI: 0.70–2.93; P=0.307), or MACCE (HR, 1.59; 95% CI: 0.92–2.71; P=0.095; Table 4). On Cox modeling, the hazard of death from any cause and that of MACCE in the IR group were similar before and after inclusion of graft failure in the Cox models (HR for death from any cause, 1.45 vs. 1.45; HR for MACCE, 1.92 vs. 1.90, respectively; Tables 3,4).
| Variables | Adjusted HR† |
95% CI† | P-value† |
|---|---|---|---|
| Death from any cause | |||
| IR | 1.45 | 0.72–2.95 | 0.299 |
| Graft failure | 1.35 | 0.71–2.50 | 0.355 |
| Cardiac death | |||
| IR | 1.61 | 0.72–3.64 | 0.243 |
| Graft failure | 1.46 | 0.70–2.93 | 0.307 |
| MACCE | |||
| IR | 1.90 | 1.06–3.43 | 0.032 |
| Graft failure | 1.59 | 0.92–2.71 | 0.095 |
| Repeat revascularization | |||
| IR | 3.71 | 1.28–11.06 | 0.016 |
| Graft failure | 3.09 | 1.31–7.44 | 0.010 |
†Propensity score adjustment. Abbreviations as in Tables 1,3.
The aim of this study was to explore the effects of IR during CABG on long-term clinical outcome in patients with LV dysfunction due to ischemic heart disease. The present long-term mortality was not significantly different between the CR and IR groups during the median 10-year follow-up period, but significantly worse outcomes were found in terms of the incidence of MACCE in the IR group.
The long-term prognostic implications of CR vs. IR during CABG remain controversial. Seminal observational studies and meta-analysis have shown improved long-term survival with CR during CABG.5,6,18 In contrast, other studies showed no significant difference in the long-term outcomes between CR and IR,19 and the clinical benefit of CR was less prominent, as long as the LAD was successfully grafted, particularly using ITA.20,21 More recently, a SYNTAX post-hoc analysis showed that IR did not increase the risk of death or cardiac adverse events during 5-year follow-up after CABG.22 The discrepancy between the studies might be due to the use of different CR definitions, different patient characteristics, including anatomic coronary complexity and amount of arterial and venous grafting, and varying length of follow-up.
In the present study, CR was based on the anatomic (i.e., functional) definition of bypassing all diseased and graftable coronary segments presenting significant flow-compromising lesion. This definition has been recently advocated in other studies.10,19,20 The present IR rate of 43.6% after CABG was higher than in other studies, which ranged up to 37.8%.8 The present high IR rate might be related to the fact that we focused on patients with LV dysfunction, who have more complex CAD, with a higher prevalence of multivessel disease, diffuse and narrowed disease, and heavy calcification. The patients who underwent IR in the present study had ≥1 ungrafted diseased vessels, and most of the reasons were diffusely diseased and heavily calcified vessels. SYNTAX score, which is an angiographic method to objectively measure the complexity of CAD,11 in patients who underwent IR was higher than that in those who underwent CR (37.1±9.0 vs. 31.4±8.7; P=0.001). In addition, the IR patients tended to have a higher age and a lower renal function. These results are in line with previous studies demonstrating that coronary complexity as well as patient characteristics, such as older age and higher prevalence of severe comorbidities, frequently precluded CR.7,18
The long-term treatment efficacy of CR in patients with LV dysfunction is not well documented in the literature, to date. In the Coronary Artery Surgery Study (CASS) registry,5 wherein only 16% of patients received an ITA conduit and CR was defined as ≥1 bypass graft to each of the 3 major territories, CR improved the 6-year survival rate compared with IR (69% CR vs. 45% IR) in 114 patients with severe angina and LV dysfunction. In the poor LVEF subgroup analysis of an observational study by Yi et al, wherein only 29 patients who underwent IR were analyzed, IR was defined as the number of distal anastomoses being less than the number of diseased coronary artery segments; and CR had a better 5-year overall survival rate (80% CR vs. 54% IR) and freedom from MACCE rate (77% CR vs. 51% IR) than IR.9 Although these studies emphasized the survival advantages of CR in patients with LV dysfunction, the results must be carefully considered because many methodological differences were observed with regard to patient selection and the definition of CR. Furthermore, in the Yi et al study, coronary complexity was not considered in the analysis.9 In the present study, although CR reduced late mortality during the 10-year median follow-up period on crude analysis, this benefit disappeared after PS adjustment. In contrast, the incidence of MACCE in the IR group was significantly poorer on both crude analysis and after PS adjustment. The increased rate of MACCE in the IR group was mainly attributed to a higher rate of repeat revascularization.
The present data suggest that CR remains desirable in patients with LV dysfunction in order to reduce the rate of repeat revascularization after CABG. Not all stenoses, however, could be completely revascularized for coronary anatomic reasons, and some of these issues can be difficult or impossible to overcome. It is worth noting in this study that IR would be considered reasonable in terms of secondary prevention of mortality. In such highly complex CAD cases, the aim was to bypass the other diseased arteries, which perfused near the arteries in which peripheral anastomoses could not be performed, in addition to the revascularization of the LAD with an ITA. In addition, Melina et al showed that patients with preoperative SYNTAX score >32 tended to develop more aggressive CAD, particularly over time, and had higher mortality rates and more repeat revascularization procedures after CABG in patients with LV dysfunction.23 Guideline-recommended medications, such as β-blockers and angiotensin-converting enzyme inhibitors, after CABG might play a critical role in improving long-term outcomes in this patient population.
Study LimitationsThe main limitations of the present study were its retrospective nature and the small number of patients. Unidentified confounders may have influenced the results despite the use of PS adjustment to control for selection bias. Another limitation was the lack of data on myocardial viability, which might influence long-term outcome. Because clinical viability studies were not routinely performed preoperatively at the present institution, including these studies in the present analysis was not possible.
Although IR was not significantly associated with increased risk of mortality in patients with LV dysfunction who underwent CABG on long-term follow-up, CR might reduce the risks of repeat revascularization and subsequent MACCE. Additional studies involving a large number of patients are necessary to confirm the present results.
The authors declare no conflicts of interest.
