Continuing Conundrum of Multiple Arterial Conduits for Coronary Artery Bypass Grafting
The idea of using multiple arterial grafts (AGs) for coronary artery bypass graft surgery (CABG) dates back more than 4 decades, and the potential clinical benefits of this strategy represent one of the most extensively studied aspects of coronary surgery. A recent MEDLINE search for studies comparing the results of CABG using bilateral versus single internal thoracic artery (BITA versus SITA) resulted in 3950 articles. Over the years, the BITA versus SITA comparison has been summarized in 6 meta-analyses, with the most recent pooling data from 89 399 patients (20 949 with BITA and 68 450 with SITA). All of them reported significantly longer survival associated with the use of BITA, with the hazard ratio for mortality varying from 0.78 to 0.81.1
The findings are similar when the second AG is the radial artery. Multiple observational studies have reported significantly longer survival associated with the use of the radial artery in addition to the SITA.2 This has been observed even in high-risk categories such as female sex, obesity, diabetes mellitus, and reoperations.2
It is important to emphasize, however, that the great majority of the reported evidence supporting multiple AGs is observational.
To date, only 3 randomized controlled trials (RCTs) of BITA versus SITA have been published. Two had a limited sample size and were clearly underpowered to detect even moderate differences in clinical outcome. The third, ART (Arterial Revascularization Trial), was started in 2004 with the aim of confirming that the use of BITA during CABG improves long-term survival. The primary outcome of ART is 10-year all-cause mortality, and the study is powered to detect a 20% relative risk reduction. Last year, the ART Investigators published the results of the planned 5-year interim analysis.3 No differences in clinical outcomes were found between patients receiving BITA and those receiving SITA, and the survival curves of the 2 groups were almost perfectly overlapping (hazard ratio, 1.04). The final results of ART will likely be published in late 2018, but because of the very low event rate in both arms and the almost identical survival curves of the 2 groups, it is the opinion of most of the experts in the field that no significant differences in the primary outcome will be found between patients with BITA and those with SITA.
So why does there seem to be such an apparent contradiction between the consistent findings of a large number of observational studies and those of the RCTs (Figure)? There are several possible reasons.
Most obviously, it is possible that the RCTs were underpowered. ART, by far the largest, used sample size calculations based on studies from the prestatin era and probably overestimated the event rate in the control group given the short follow-up time (especially because ART patients were highly compliant with optimal medical therapy). Additional possible explanations include cointervention in that 23% of patients assigned to the SITA group also had a radial artery (ie, in effect a second AG), the high percentage (16%) of crossover in the BITA arm (possibly indicative of less experience with BITA grafting technique of ART surgeons), and patient age given that almost a quarter of patients were >70 years and the treatment/age interaction at 5 years trended to significance (P=0.08). In addition, although the choice of the primary outcome (death) is indisputable, failure of grafts to vessels other than the left anterior descending may be more likely to result in cardiac events other than death.
However, it is increasingly important to acknowledge that a possible alternative explanation is that the longer survival associated with additional AGs seen in observational studies is actually determined by confounders and selection bias. This hypothesis seems particularly attractive because it is likely that surgeons reserved the BITA operation, considered more invasive and with possibly better long-term results, for patients judged healthier and with longer life expectancy. This eyeballing by surgeons, on the basis of experience and clinical judgment, takes into account variables that are not measured by our databases and is impossible to neutralize with our matching algorithms. In terms of measured covariables, the BITA cohort in the observational series was generally younger and had a significantly lower proportion of female patients and individuals with diabetes mellitus and pulmonary or peripheral vascular disease and more limited coronary disease.1 The longer survival associated with BITA in the observational studies could then be explained by measured and unmeasured covariables and by allocation bias, not by the superiority of the treatment.
If the clinical consequences of the use of multiple AGs remain to be determined, there is, however, good evidence that their patency rate is better than that of saphenous vein grafts, especially in the mid and long term. A network meta-analysis pooling data from 9 angiographic RCTs involving 2780 patients found that the use of the saphenous vein was associated with a 4-fold and 3-fold increased risk of late (≥4 years) functional graft occlusion compared with the right internal thoracic artery or the radial artery, respectively.4
Recent data also suggest that vascular mediators produced by the endothelium of the AGs exert a protective effect on the downstream coronary circulation; that is, the use of AGs can slow the rate of native vessel disease progression after CABG.5
To conclude, after >4 decades of clinical research, it is still uncertain whether the use of multiple AGs for CABG leads to superior clinical outcomes. The 5-year interim analysis of ART raised more questions than it answered and was a healthy wake-up call to the surgical community for the need of further investigations.
While the 10-year results of ART are awaited and new RCTs are on their way, what is the role of multiple AGs for CABG? Because the patency rate of AGs is better than that of saphenous veins and new data even suggest a protective effect on the native circulation, they should be routinely used in patients with reasonably long life expectancy, as long as no operative risk is added. Operator experience and center experience are key to achieving this goal.
Sources of Funding
Dr Fremes is supported in part by the Bernard S. Goldman Chair in Cardiovascular Surgery.
- © 2018 American Heart Association, Inc.
- Buttar SN,
- Yan TD,
- Taggart DP,
- Tian DH
- Gaudino M,
- Taggart D,
- Suma H,
- Puskas JD,
- Crea F,
- Massetti M
- Taggart DP,
- Altman DG,
- Gray AM,
- Lees B,
- Gerry S,
- Benedetto U,
- Flather M
- Benedetto U,
- Raja SG,
- Albanese A,
- Amrani M,
- Biondi-Zoccai G,
- Frati G