Author + information
- ↵∗Address for correspondence:
Dr. Anelechi Anyanwu, Department of Cardiovascular Surgery, Mount Sinai Medical Center, 1190 Fifth Avenue, New York, New York 10029.
In this issue of the Journal, Royse et al. (1) report long-term survival in patients who had coronary artery bypass graft surgery (CABG) with total arterial revascularization, noting superior survival after 7 years compared with CABG with single internal thoracic artery (ITA) and vein grafts. Lytle et al. (2) had demonstrated similar excellent arterial conduit outcomes 2 decades ago with bilateral ITA grafts, concluding that “it has been the position of some coronary artery surgeons that the consideration of multiple arterial grafting could be ignored because no clear evidence existed that outcomes were improved for any patient subsets. That position is no longer tenable.” In the United States, however, multiarterial CABG is still largely ignored, with only 10% of CABG done using ≥2 arterial grafts, and <1% with 3 arterial grafts (3). In contrast, in Australia approximately 50% of patients receive ≥2 arterial grafts (3). Why have cardiologists and surgeons in the United States not embraced multiarterial CABG? The root of the poor uptake of arterial grafting, and the confusion around the evidence, can be understood by a careful study of current data and practice.
If Longevity Is Part of Selection Criteria, Long-Term Survival Will Be Better
Surgeons will typically only perform multiple arterial grafts if patients are expected to enjoy long survival after CABG. Propensity matching cannot adjust for this, as the critical clinical judgment that surgeons often use in decision making (“the eye ball test”) is not considered. For example, a 45-year-old unemployed male who is an alcoholic, homeless, and a heavy smoker would most probably have 1 ITA and vein grafts. Such a patient has a high probability of experiencing major adverse events, but propensity matching would pair him with a 45-year-old banking executive receiving 4 arterial grafts, and count differential adverse events as being associated with the treatment (vein grafting); in reality, negative outcomes would more likely be due to (unbalanced) patient factors. The net effect is that for therapies where benefit is conditional on longevity, one often sees separation of survival curves earlier than can be explained by the treatment, as surgeons have selected out a patient group with better survival. For example, the unmatched survival curves in Royse et al. (1) show immediate survival advantage of arterial grafts from 1 year after surgery. This treatment selection bias makes it harder to ascertain the degree to which long-term benefits are due to treatment effect as opposed to unbalanced patient factors.
If the Surgeon Has Specialized Surgical Skills, Then Long-Term Survival Will Be Superior
Arterial grafting is a technically complex procedure requiring a high level of surgical commitment and expertise. It is more difficult and time-consuming compared to a single ITA with vein grafts. The operation reported by Royse et al. (1) is intricate and involves anastomosis of the radial artery to the left ITA and then multiple sequential anastomoses to feed the coronary territories. An error in judging the length or lie of the conduit between anastomoses can have catastrophic adverse consequences. Invariably, in multisurgeon datasets, when one compares multiple arterial grafting with venous grafting, one is also comparing surgeons who frequently do a more technically challenging operation to those who do not. Comparative demonstrations of better long-term survival with arterial graft patients, therefore, partly reflect a higher level of expertise of surgeons doing the arterial grafting. Propensity matching cannot separate this effect of the operator in a skill-dependent operation—a form of bias that Blackstone referred to as “inextricable confounding” (4). The (skilled) surgeon is an integral and inseparable part of the complex treatment. Statistical comparison will be inherently biased and frequently shows exaggerated benefit with the complex operation, as the less complex procedures in the control arm are typically done by any surgeon, including low-volume and less-skilled surgeons. Where surgical skill is balanced, such as in randomized trials in which the surgeons who do the multiarterial grafts also do the vein graft patients, it has been harder to demonstrate clinical differences.
Modern Vein Procurement Strategies May Predispose to Vein Graft Failure
Ideally, veins should be procured with an atraumatic no-touch technique to avoid endothelial injury (a precursor to vein graft disease ). In the current era, however, vein procurement is typically delegated to surgeons’ assistants or junior residents. Some procurement techniques necessarily invoke direct trauma to the vein endothelium. Indeed, assimilation of available evidence suggests the widely used endoscopic vein harvest may be associated with lower 1-year graft patency (compared with open vein harvest) (6), likely related to inadvertent direct trauma, traction injury, and surgical experience of operators doing endoscopic approach. In contrast, arterial grafts are typically procured in atraumatic fashion by a skilled surgeon or surgeon’s assistant. Therefore, in terms of endothelial integrity, patients in the current era are more likely to experience a better-quality conduit, and hence one with better durability, with any arterial conduit. This differential handling of conduits has implications for clinical practice and interpretation of contemporary research studies.
Do We Need Further Studies on Arterial Revascularization?
More propensity-matched comparisons will invariably show that arterial revascularization is superior to vein grafting because of the combined effect of the conduit, the reasons described in the above text, and publication bias. Analysis of large databases is unlikely to provide robust comparative data due to unbalanced selection bias, varied surgical skills and techniques, volume effect, and limited follow-up. A recent analysis of the Society of Thoracic Surgeons database found higher operative mortality with multiple arterial revascularization in centers who used multiple arterial grafts in <5% of CABG procedures (7). Databases may, therefore, not capture the true treatment effect due to dilution by low-volume centers. Randomized trials are generally too small, underpowered, and deficient in long-term follow-up to provide definitive data. The only large randomized trial, ART (Arterial Revascularization Trial), randomized >3,000 patients to bilateral or single ITA grafting (8), and could not demonstrate 5-year outcome differences. A confounding factor in the ART trial was that surgeons were allowed to use the radial artery as a second or third graft, so some patients in the single ITA group received radial artery grafts, which conferred a survival advantage compared with those who received only veins for supplemental grafts (9). Further large multicenter trials are probably impractical and unlikely to occur. The 10-year time frame required to conduct and report these trials and the scarcity of funding support for nonindustry trials are significant obstacles to future randomized studies. In all probability, we have reached an impasse regarding data on multiple arterial grafting—one that may not change substantially with future studies.
Is There Sufficient Information to Say Arterial Grafts Are Superior to Venous Grafts?
Current evidence shows that, at least in expert hands, multiple arterial grafts can be performed with similar safety to a single ITA plus vein grafts (although the former requires more effort, skill, and cost). If multiple arterial grafts can be done safely, then rationalizing their use becomes easier. There is a persistent misperception by many surgeons that use of bilateral ITA grafts substantially increases risk of CABG, particularly that of deep sternal wound infection. In the ART study, the wound complication rate with skeletonized bilateral internal thoracic arteries was the same as that of a pedicled single ITA (10) (pedicled single ITA procurement is the predominant approach used during CABG). Several studies have shown that skeletonized bilateral ITA grafts can be safely used in diabetic individuals, the elderly, and the obese with minimal incremental risk. In a meta-analysis of studies on diabetic patients, the incidence of deep wound infection was 3.2% in bilateral compared with 2% in single ITA grafting; but, when the conduits were skeletonized, there was no difference in infection rate with bilateral ITA use (11). Radial artery grafting is not associated with specific incremental risks, provided grafted vessels have high-grade stenoses. In the majority of patients undergoing CABG, therefore, a second or third arterial graft can be added without noticeable incremental risk—this has been observed in Australia, where over one-half of patients receive ≥2 arterial grafts, and is also our own experience (Figure 1).
Assuming multiple arterial grafts can be performed safely, then (excluding technical, patient, or logistic factors) the principal barrier to wider adoption could be uncertainty regarding the superiority of arterial conduits. The recognized superiority of arterial over venous grafts is, however, well demonstrated by the universal acceptance of preferential use of an ITA for the left anterior descending artery—few would argue against the superiority of artery over vein in this context. The next question then arises as to whether we have enough data to say arterial grafts are preferable to veins for other suitable coronary targets. To answer this seems simple. Vein graft disease exists and will invariably lead to failure of up to 50% of vein grafts within 10 years of surgery (5), often causing recurrence of angina, heart failure, myocardial infarction, reintervention, and sometimes, death. On a daily basis, patients undergo stenting of diseased vein grafts in catheterization laboratories worldwide. For patients who receive arterial grafts, there is no equivalent “arterial graft disease.” Arterial grafts may fail within the first 12 months of CABG—mainly secondary to technical factors, competitive flow, and “string” phenomena—but after the first year, attrition rates are low, in contradistinction to the progressive attrition of vein grafts that continues throughout the long-term. In a pooled analysis of radial artery versus saphenous vein randomized trials (12), 50% of vein grafts had failed at 9 years compared with <25% of radial arteries—this difference occurred despite a similar graft failure rate at 3 years. The known incidence, natural history, and clinical consequences of vein graft disease and the absence of progressive late attrition of arterial grafts is enough evidence to support use of multiple arterial grafts if the procedure can be done safely.
Should Coronary Revascularization Be a Subspecialty?
In our opinion, the answer to this question is an unequivocal yes. We need acceptance that some coronary operations require an advanced level of expertise and skill that can likely only be systematically provided on a subspecialty level by surgeons practiced in those advanced techniques and working in centers of excellence. A center of excellence does not imply a volume threshold (indeed, many high-volume CABG centers perform very few multiarterial CABGs), but rather a commitment to investing in and promoting arterial grafting and other advanced revascularization techniques. Institutions doing CABG should strive to have at least 1 surgeon who subspecializes in systematic multiarterial revascularization. These subspecialists should work closely with interventional cardiologists, other surgeons in their group, and referring cardiologists to evolve local strategies to maximize arterial grafting.
Re-Examining the Economics of Arterial Grafting
Arterial grafting is more expensive than venous grafting because of longer operative time and additional surgical equipment utilized. Inherent on longer operative times is a marginal increase in post-operative morbidity and costs. In the ART study, which was conducted mainly in Europe, bilateral ITA grafting was associated with a 9% increase in 1-year costs compared with single ITA (13); such incremental costs will likely be substantially higher in the United States, where an extra hour in the operating room costs >$2,000 (14). There are also opportunity costs, as the staff and room could have been used for a subsequent case. There is no incremental reimbursement to compensate hospitals or surgeons commensurate with the additional time and effort taken to perform total arterial revascularization. A surgeon who spends 5 h doing an all-arterial CABG will effectively get paid the same as a surgeon who does a 3-h CABG with single ITA and saphenous vein for a young patient. The latter surgeon may go on to another case, yielding additional income. U.S. surgeons, arguably, get penalized financially if they routinely perform total arterial revascularization.
In recognition of the additional effort required and the long-term health benefits, health care purchasers and providers should devise means to incentivize multiple arterial grafting, such as use of a second arterial graft as a quality metric and appropriately weighted reimbursements to centers performing multiple arterial revascularization (15). Economics from a societal perspective are also affected, because the systematic use of vein grafts likely results in higher downstream health care costs due to reinterventions, medications, hospitalizations, and health disability associated with vein graft disease. It may be more logical for health care purchasers to invest more into maximizing arterial grafting, as this could ultimately reduce long-term costs.
Should We Abandon the Vein Graft?
Vein grafts and arterial grafts are complementary, and both continue to have an important role to play in surgical revascularization. There will remain scenarios where vein grafts are preferable to arterial grafts. For example, patients with poor distal targets or those expected not to survive beyond few years would have minimal incremental benefit with arterial grafting. Some have anatomy that may benefit from the higher initial flow of vein grafts. Some patients will benefit from a simpler or more expedient operation utilizing vein grafts. Sometimes arterial grafts fail technically, requiring vein grafts for bailout. Arterial conduits, before or after procurement, may be found to be inadequate, and so on. These scenarios are, however, unlikely to be operational in many of the 90% of patients who receive only 1 arterial graft in the United States today.
We have abundant biological and clinical evidence as to why veins are less desirable than arteries as CABG conduits. If multiple arterial grafting were a pill, surgeons would most likely prescribe it every day to most patients who undergo CABG based on the currently available data—its current low uptake likely reflects not lack of evidence, but rather the technical and time demands of the surgery. We do not need more studies to tell us that multiple arterial grafting is the gold standard—that was well defined by Lytle et al. 2 decades ago (2). What we need is dedicated subspecialists and designated centers of excellence for myocardial revascularization, which can carry arterial grafting forward in the United States, as Royse et al. (1) and others have done in Australia. Without question, there is a significant population of patients who should be considered for a second arterial graft, and there are also many patients who would likely benefit from total arterial revascularization. We have done enough talking and publishing—now we need to get to action.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
The Icahn School of Medicine at Mount Sinai receives royalty payments from Edwards Lifesciences and Medtronic for intellectual property related to Dr. Adams’ involvement in the development of 2 mitral valve repair rings and 1 tricuspid valve repair ring. Dr. Anyanwu has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2018 American College of Cardiology Foundation
- Royse A.G.,
- Brennan A.P.,
- Ou-Young J.
- Schwann T.A.,
- Tatoulis J.,
- Puskas J.,
- et al.
- Gosavi S.,
- Mukherjee D.
- Harky A.,
- Balmforth D.,
- Shipolini A.,
- Uppal R.
- Schwann T.A.,
- Habib R.H.,
- Wallace A.,
- et al.
- Taggart D.P.,
- Altman D.G.,
- Gray A.M.,
- et al.
- Taggart D.P.,
- Altman D.G.,
- Flather M.,
- et al.
- Benedetto U.,
- Altman D.G.,
- Gerry S.,
- et al.
- Gaudino M.,
- Benedetto U.,
- Fremes S.,
- et al.
- Gray A.M.,
- Murphy J.,
- Altman D.G.,
- et al.
- Childers C.P.,
- Maggard-Gibbons M.
- Puskas J.D.,
- Yanagawa B.,
- Taggart D.P.
- Corresponding Author
- If Longevity Is Part of Selection Criteria, Long-Term Survival Will Be Better
- If the Surgeon Has Specialized Surgical Skills, Then Long-Term Survival Will Be Superior
- Modern Vein Procurement Strategies May Predispose to Vein Graft Failure
- Do We Need Further Studies on Arterial Revascularization?
- Is There Sufficient Information to Say Arterial Grafts Are Superior to Venous Grafts?
- Should Coronary Revascularization Be a Subspecialty?
- Re-Examining the Economics of Arterial Grafting
- Should We Abandon the Vein Graft?