Key result
Preoperative coronary revascularization may benefit intermediate-risk vascular surgery patients, whereas recent RCTs opposing it suffer from significant selection bias and methodological limitations.
Why the study?
Does preoperative coronary revascularization improve long-term survival and prevent postoperative myocardial infarction in high-risk patients undergoing vascular surgery?
Does preoperative coronary revascularization improve long-term survival and prevent postoperative myocardial infarction in high-risk patients undergoing vascular surgery?
Answers are grounded in this paper and cite their sources.
This editorial highlights the limitations of recent RCTs and argues that preoperative coronary revascularization should still be considered for high-risk vascular surgery patients with extensive ischemia.
Dr. Kertai has done an excellent job in the core review1 in distilling the essence from the complex and multifaceted literature on preoperative coronary revascularization. However, under the surface, a titanic debate takes place on some major issues: Does any randomized controlled trial (RCT) always provide better evidence than any observational study? Can two recent RCTs completely overturn the data accumulated by a number of previous observational studies? Was the methodology used to generate the data adequately convincing? First some basic facts that must be acknowledged in the discussion: Vascular surgery patients have high prevalence of significant coronary artery disease (CAD) of which 18% have severe triple vessel disease (TVD) and 4% left main disease (LMD),2 who are the target of any possible coronary revascularization. Patients with peripheral vascular disease have a 6–15-fold higher long-term mortality from CAD than patients without peripheral vascular disease.3 Patients with TVD or LMD have worse long-term survival than those with only single or double-vessel CAD. This is not only true for patients with unstable but also for patients with stable CAD4 and more so for patients with coexisting reduced left ventricular function. Coronary revascularization by coronary artery bypass graft (CABG) improves long-term survival and event-free survival in patients with multi-vessel (TVD and/or LMD) CAD, especially if left ventricular function is impaired.5,6,7,8 This has never been proven though for percutaneous coronary intervention (PCI) with or without stenting, although several studies have reported comparable long-term results for PCI and CABG. Recently the COURAGE study found no survival benefit with PCI over optimal medical therapy;9 however, this RCT has been criticized regarding patient selection bias (less than 10% of the patients evaluated were included), PCI methodology and incomplete revascularization in most patients (59%), weakening its conclusions.10,11 Large areas of moderate-to-severe ischemia on noninvasive radionuclide imaging of the heart predict long-term outcome even better than the number of diseased vessels on coronary angiography.12 Similar or even better results have been obtained for dobutamine stress echocardiography.13 Although no RCT was designed to compare survival with revascularization versus medical therapy in patients with ischemia on myocardial perfusion scans, a large scale observational study (10,627 consecutive patients)14 demonstrated a better survival with revascularization over medical therapy when moderate-to-severe ischemia (>20%) was present on myocardial perfusion scan, especially in high-risk patients (elderly, women, diabetics) (Fig. 1). Figure 1.: Log Hazard ratio for revascularization (Revasc) versus medical therapy (Medical Rx) as a function of % myocardium ischemic Based on final Cox proportional hazards model. (Adopted with permission from Circulation 2003;107;2900–7).LIMITATIONS OF RCTS There is an abundance of literature discussing the pros and cons of RCTs versus observational studies.15,16,17 All agree that RCT is the best study design for investigating primary treatment effects in large phase III pharmaceutical trials in compliant populations. However, observational studies are not only more feasible but may actually give more accurate results than RCTs when a clinical dilemma is seeking an answer where there is no one clear-cut solution but rather it strongly depends on various complex patients’ features and on overall clinical judgment. Although the impact of allocation biases, spurious observations, and missing covariates cannot be ignored, patients in observational studies better represent those seen in real clinical practice. A RCT provides optimal results when the clinicians providing a new drug or treatment are blinded (preferably, double-blinded) and impartial to its effect. A RCT loses its effectiveness when the physicians enrolling patients are biased as to the potential benefits or harms of the investigated treatment, simply because of their often justifiable reluctance to randomize patients who they believed may either benefit or be harmed by the treatment (or its prevention). Typical to such RCTs is that too few screened patients, those who are not likely to be affected by the randomization, are finally randomized. This obviously raises a fundamental issue of selection bias, and seriously calls into question the validity and generalizeability of such RCTs. Also, typically, these RCTs are likely to find lack of treatment effect on patient outcomes, since most patients who could have benefited or been harmed by the treatment were excluded. Two RCTs opposing preoperative coronary revascularization in candidates for vascular surgery have been recently published.18,19 However, these RCTs suffer from many of the above-mentioned limitations. LIMITATIONS OF THE CARP TRIAL17 The CARP trial aimed to challenge the long debated concept that preoperative coronary revascularization in high-risk vascular surgery patients improves long-term outcome. While this is a worthy and justifiable goal, it must be judged also by the means used to achieve it. The trial screened 5859 vascular surgery patients from 18 Veterans Adminstration hospitals in the United States, each center using different methods for screening patients (Eagle's criteria/Revised Cardiac Risk Index, with or without noninvasive testing). There was no unified method for screening patients’ preoperative risk, neither were the American Heart Association/American College of Cardiology guidelines systematically implemented which recommend noninvasive testing before coronary angiography in patients with moderate clinical predictors. Vascular surgery patients typically have low functional capacity and asymptomatic CAD, which is difficult to evaluate with noninvasive testing. Preoperative nuclear stress testing was performed in only 62% of randomized CARP patients, and only 44% of randomized patients had moderate-large ischemia on nuclear imaging. As a result, only 32% of randomized patients had TVD or 2.9% of all 5859 screened patients. This proportion is in marked contrast to the expected 18% of patients with TVD according to the data obtained from Hertzer et al.'s study using systematic coronary angiography in 1000 vascular surgery patients. Thus, the majority of patients with TVD in the CARP trial – those most likely to benefit from revascularization – were either not identified by the screening process or excluded because of the clinicians’ judgment that they should not be randomized. The other 68% of randomized patients with single or double-vessel disease were unlikely to have a survival benefit from coronary revascularization. Although the CARP study did not find differences in postoperative myocardial infarction (MI) with or without preoperative revascularization, it did find (in one of its sub-studies)20 that the more complete the preoperative coronary revascularization, particularly with CABG, the less were postoperative MIs. However, the same authors previously reported that postoperative MI after elective vascular surgery is associated with worse long-term survival.21 Thus, based on CARP's own data, complete preoperative revascularization in the right patients prevented postoperative MI, and therefore improved long-term survival. LIMITATIONS OF THE DECREASE-V TRIAL18 In theory, the DECREASE-V trial could have filled the gap that the CARP study created by aiming to randomize the very high-risk patients with significant inducible ischemia on noninvasive preoperative testing who are most likely to benefit from coronary revascularization. However, several issues in the methodology and results make this study unrealistic and cast serious doubt on its generalizeability: “Of 430 high-risk patients, 101 (23%) showed extensive ischemia and were randomly assigned to revascularization (n = 49) or no revascularization.” Apparently, all 49 patients with extensive ischemia on preoperative testing eventually underwent coronary revascularization. However, this is not compatible with real life as we know it. Up to 1/3 of vascular surgery patients with severe ischemia on noninvasive testing may not be suitable for revascularization, based on their coronary anatomy. Problems such as poor runoff, multiple small vessel CAD mainly in diabetics, chronic total obstructions not suitable for PCI in patients who are too sick for CABG, etc. may render them unsuitable for revascularization. The DECREASE-V trial “forced” all 49 patients with severe ischemia on noninvasive testing to undergo coronary revascularization, with no exclusions. The only clinical judgment applied was if it would be by PCI or CABG. This may explain the extremely high morbidity in the revascularization arm (two patients had ruptured aortic aneurysm before the vascular surgery) and the lack of postoperative benefit. Patients in the revascularization arm had extremely high 30-day mortality after vascular surgery, twice as high as the control arm (22% vs. 11%). This is in contrast to multiple studies which report reduced mortality and MI after vascular surgery in patients who had prior successful coronary revascularization, particularly with CABG.22,23 No explanation was given to the markedly high postoperative mortality despite successful and complete revascularization (in 86% of patients). Perhaps the “catastrophic” incidence of postoperative complications was because double antiplatelet therapy was not discontinued in patients who had surgery shortly after PCI,24 but this is not reported. Postoperative MI rate was also extremely high in both groups (35% and 31%) and all were Q-wave infarctions.25 This is in contrast to multiple studies in vascular surgery patients, which showed that postoperative Q-wave MI is rare whereas close to 100% are non-Q-wave infarctions. In a study that monitored hundreds of vascular surgery patients with continuous on-line 12-lead electrocardiogram perioperatively, none of the patients had Q-wave MI.26 The same group that published the DECREASE-V trial has reported that prophylactic perioperative beta-blockage does not prevent postoperative cardiac events in patients with extensive myocardial ischemia on preoperative testing, similar to those randomized in the DECREASE-V trial.27 POSTOPERATIVE MI: UNSTABLE PLAQUE OR UNSTABLE PATIENT? One of the main arguments of those opposed to preoperative coronary revascularization is that revascularization to arteries with stable coronary lesions is futile because it does not prevent postoperative MI allegedly caused by plaque rupture of relatively non-occluding (<70% stenosis) but unstable coronary plaques. We will not elaborate much on this issue here. However, after years that we have been preaching for the concept that prolonged, stress-induced ST-depression-type ischemia is a distinct and common mechanism for MI, particularly in the postoperative period,28 it has recently been accepted by the widest community of cardiologists. A new “Universal classification of MI” has been presented in the European Society of Cardiologists congress (2007) as a consensus document by the European Society of Cardiologists, American College of Cardiologists, American Heart Association and World Heart Federation (in press). This new universal classification includes 5 types of MI: Type I is spontaneous MI related to an acute coronary event (plaque rupture, fissuring or erosion); Type II is MI secondary to prolonged oxygen supply-demand imbalance (anemia, hypotension etc.); Types III to IV are sudden cardiac death, post-PCI and post CABG, respectively. The purpose of coronary revascularization in major vascular surgery patients with typically long-standing and stable CAD is to prevent type-II MI, both perioperatively and in the long-term. Although beta-blockers may help prevent some of these perioperative events, they are not yet a substitute for coronary revascularization in patients with severe CAD. PREOPERATIVE CORONARY REVASCULARIZATION AS A CLINICAL DILEMMA RCTs treat scientific questions as if they have only one possible answer – yes or no – given that all confounding factors are controlled. However, in reality, complex clinical issues may have more than one answer and a careful clinical judgment must be made for each individual patient with his/her special complex combination of predictors and demographic features. When the clinical judgment itself is evaluated, rather than the quality of a specific medication or therapy, there is probably no better alternative but to refer to observational cohort studies. Such is the case when coronary revascularization is considered in patients with complex preoperative clinical features, different results on preoperative cardiac stress testing and multiple possible combinations of coronary anatomy findings. Our recent findings29 show that intermediate-risk vascular surgery patients, those with 2–3 risk predictors of the Long-Term Survival Score (LTSS), were most likely to benefit from preoperative coronary revascularization. In contrast, low-risk patients, those with low LTSS 0–1 had good long-term survival that was not significantly affected by the revascularization. Similarly, high-risk patients (LTSS ≥4) had poor early and long-tern survival after vascular surgery that was also not significantly affected by coronary revascularization, even if severe ischemia on noninvasive testing was found and treated. This result could only be obtained by a cohort study that included all consecutive patients undergoing vascular surgery, without the unavoidable exclusions curtailing RCTs.
No takes yet. Share an insight, caveat, or question.
Landesberg et al. (2008) conducted an editorial in Coronary artery disease in vascular surgery patients. Preoperative coronary revascularization vs. No revascularization / medical therapy was evaluated. Preoperative coronary revascularization may benefit intermediate-risk vascular surgery patients, whereas recent RCTs opposing it suffer from significant selection bias and methodological limitations.
Synapse has enriched 10 related papers. Consider them for comparative context: