Nearly one-third of the approximately 30 million people undergoing non-cardiac surgery in the US each year have known coronary artery disease (CAD) or CAD risk factors.1,2 An estimated one million suffer peri-operative cardiac complications accounting for roughly US$20 billion in annual medical costs.3 Peri-operative myocardial infarction (MI) may affect as many as 34% of high-risk individuals and result in significant long-term morbidity or cardiac death.4,5 Given these facts, a number of predictive models have been devised to help identify patients at high risk of peri-operative cardiovascular complications and to guide further diagnostic testing.6–8 Unfortunately, treatments aimed at lowering this risk have been met with questionable and limited success.
Pathophysiology of Myocardial Ischaemia
The surgical patient is subjected to a number of factors that increase the risk of peri-operative MI. Surgery can induce a physiologically stressful state due to catecholamine surge resulting in haemodynamic alterations, vasospasm, vascular inflammation from cytokine release, platelet activation and development of a hypercoaguable state with reduced fibrinolytic activity.3,9,10 The potential addition of large fluid shifts to these physiological insults may create an imbalance in myocardial oxygen, accounting for one mechanism of peri-operative MI. However, autopsy studies indicate that these stressors are as equally likely to result in the rupture or disruption of vulnerable plaques, thereby causing peri-operative MI similar to that seen in the acute coronary syndromes affecting the non-surgical population.11,12
Clinical Risk Stratification
Many risk-prediction models have been developed over the years, beginning with Goldman et al. in 1977, who developed a multifactorial risk index focused specifically on peri-operative cardiac complications.6 This risk index combined many patient-specific variables such as age, evidence of ischaemia or heart failure, valvular abnormalities, arrhythmias and overall medical condition with the need for high-risk surgery. Subsequent modification of the Goldman risk index utilising a Bayesian approach was introduced by Destky et al. in 1986, resulting in improved predictive power.7
The most recent and commonly used risk assessment model was introduced in 1999 by Lee et al.8 The Revised Cardiac Risk Index (RCRI) identifies six independent predictors of cardiac complication during non-cardiac surgery: high-risk surgery (defined as procedures carrying a >5% risk of cardiac complications), renal insufficiency (serum creatinine >2mg/dl), insulin-dependent diabetes and a history of ischaemic heart disease, congestive heart failure or cerebrovascular disease. These clinical risk factors have been incorporated into the American College of Cardiology/American Heart Association (ACC/AHA) guidelines for peri-operative cardiovascular evaluation and management.13 The guidelines endorse a five-step algorithmic approach to patients anticipating non-cardiac surgery and aim to limit further cardiac testing unless results are expected to affect additional management (see Figure 1). The algorithm first considers whether or not surgery is deemed emergent. Emergent surgery obviates the requirement/ability for further investigation beyond basic data gathering and physical exam findings, such as heart rate and blood pressure measurements, volume status, electrolytes, blood count, chest X-ray and electrocardiogram (ECG). Next, patients are evaluated for any active cardiac conditions that would dramatically increase peri-operative morbidity and mortality (see Table 1). Evaluation and management of these conditions should delay surgery to allow for treatment according to the current standard of care. Risk stratification inherent to the surgical procedure itself is then assessed (see Table 2). In the absence of active cardiac conditions, low-risk surgeries are typically allowed to proceed without delay or further testing. Conversely, intermediate- and high-risk procedures may require additional investigation. The next critical step involves assessing the patient’s functional capacity. Patients identified as having poor exercise tolerance – defined as an inability to walk four blocks or climb two flights of stairs (approximately 4 metabolic equivalents [METS]) without symptoms – have been shown to have significantly more peri-operative complications than those with good exercise tolerance.14 By contrast, patients with good functional capacity who are otherwise asymptomatic are unlikely to have their management altered by further cardiac testing and can generally proceed with surgery. In the absence of any of the above clinical variables (need for emergent surgery, active cardiac symptomatology, high-risk procedure and poor functional capacity), further risk stratification with the help of the RCRI can be used to help guide decisions in terms of the need for additional functional testing.
Peri-operative Management
Medical Therapy
Beta-adrenergic antagonists (beta-blockers) have been one of the most extensively studied medical therapeutic agents aimed at reducing peri-operative cardiovascular complications. The results of these studies have been varied and reaching a consensus hampered by heterogeneity in study design, agents used and patient populations. One of the earliest randomised studies suggesting benefit with beta-blockers utilised atenolol in patients with known or suspected CAD undergoing non-cardiac surgery.15 The treatment group experienced no difference in peri-operative cardiac events, but was noted to have less electrocardiographic ischaemia and reduced mortality at six months. Further supportive data come from the Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography (DECREASE) trial, in which high-risk patients undergoing vascular surgery receiving titrated doses of bisoprolol over an average of 37 days before surgery had a significant reduction in peri-operative cardiac death and MI compared with placebo.5 However, other trials have been unable to confirm this cardioprotective benefit and the largest randomised trial to date has suggested potential harm.16–18 Over 8,000 patients with or at risk of CAD in the PeriOperative ISchemic Evaluation (POISE) trial were randomised to fixed-dose extended-release metoprolol versus placebo on the day of surgery and were found to have a reduction in peri-operative MI at the cost of increased stroke and overall mortality casting suspicion on the beneficial effects of acute high-dose prophylactic beta-blocker use.19 By contrast, intermediate-risk patients in the recent DECREASE-IV trial who were given titrated doses of bisoprolol and either fluvastatin or placebo in a randomised unblinded fashion demonstrated a reduction in non-fatal MI and cardiac death at 30 days with no apparent increase in stroke.20 Based on cumulative data thus far, it is recommended that beta-blockers be continued in those patients already prescribed them for other indications and be considered as treatment for intermediate- and high-risk patients undergoing intermediate- or high-risk procedures.21 Beta-blockers are no longer recommended as routine treatment in low-risk patients who do not otherwise have an indication for them; if initiated, therapy should begin ideally well before the planned procedure with doses carefully titrated to optimise the resting heart rate while avoiding significant hypotension.
Given their effectiveness in the medical treatment of CAD, 3-hydroxy-3- methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) have also been investigated as a means of lowering peri-operative cardiovascular complications. The pleiotropic effects of statins are well documented and it is these effects (plaque stabilisation, reduced vascular inflammation, improved endothelial function) that would seem to make statins the ideal agents for vascular protection from the physiological stress of surgery.22 Along with observational data, the strongest evidence to date for the beneficial effects of statins stems from the DECREASE-III trial in which patients undergoing vascular surgery randomised to fluvastatin experienced significantly less myocardial ischaemia and the combined secondary end-point of cardiac mortality and MI.23,24 Statins are recommended for patients who meet National Cholesterol Education Program criteria for treatment and are felt to be beneficial in all patients undergoing vascular non-cardiac surgery.25 Statins are also considered potentially helpful in those with at least one clinical risk factor undergoing intermediate-risk procedures.13
Coronary Revascularisation
Among the indications for coronary revascularisation in the general population with stable CAD is left main or severe three-vessel CAD, symptomatic two-vessel CAD involving the proximal left anterior descending (LAD) artery and angina refractory to medical therapy.26 To date, there have been no randomised trials supporting the role of prophylactic coronary revascularisation to reduce cardiovascular complications in non-cardiac surgery. However, a number of observational studies have suggested improved outcomes in patients with CAD and successful prior revascularisation. A review of patients in the Coronary Artery Surgical Study (CASS) database with previous coronary artery bypass grafting (CABG) who subsequently underwent high-risk non-cardiac surgery revealed a significant reduction in 30-day mortality and non-fatal MI compared with patients receiving medical therapy alone.27 A similar review of Medicare recipients also reported significant improvements in both short- and long-term survival in patients undergoing CABG prior to high-risk aortic surgery.28 Likewise, observations of reduced cardiac events have been noted with percutaneous coronary intervention (PCI).29 Patients previously treated with angioplasty in the Bypass Angioplasty Revascularisation Investigation (BARI) trial later proceeding with non-cardiac surgery had low peri-operative cardiovascular complication rates that were similar to those with CABG revascularisation.30
The largest randomised trial to date designed to evaluate the strategy of prophylactic coronary revascularisation in high-risk patients undergoing vascular surgery is the Coronary Artery Revascularisation Prophylaxis (CARP) trial.31 In the CARP trial, 510 patients scheduled to undergo elective major vascular surgery deemed at high risk of cardiovascular complications were randomised to revascularisation or no revascularisation after obstructive CAD was identified by coronary angiography. Half of the patients enrolled had two or more RCRI risk factors and three-quarters had either multiple clinical risk factors or moderate to large reversible defects on non-invasive stress imaging. The majority of patients had normal left ventricular ejection fraction (EF) with single- or two-vessel disease while three-vessel disease was seen in one-third of the total population. Excluded were patients with EF <20%, severe aortic stenosis or significant left main (LM) disease (>50% angiographic stenosis). Patients randomised to receive revascularisation underwent either CABG or PCI at the discretion of the operator and experienced no reduction in peri-operative MI or death at 30 days and saw no improvement in long-term mortality after six years.
Subsequently, the recent DECREASE-V pilot study randomised 101 patients with three or more clinical risk factors and extensive ischaemia planning major vascular surgery to either revascularisation or optimal medical treatment with antiplatelet therapy and beta-blockade titrated to a resting heart rate of less than 65 beats per minute.32 The DECREASE-V patient population had more three-vessel CAD (67%) and a higher percentage of left ventricular systolic dysfunction (EF <35% in 43% of patients) than patients in the CARP trial, yet similarly observed no difference in 30-day mortality or MI with revascularisation. While small and insufficiently powered to make definitive conclusions in terms of the optimal peri-operative treatment strategy, the results from DECREASE-V are consistent with CARP trial data in patients suffering from less advanced CAD. Taken together, these trials cast doubt on the benefit of pre-operative revascularisation in an effort to reduce cardiovascular peri-operative risk.
Surgical Coronary Revascularisation
Despite the lack of definitive data supporting prophylactic coronary revascularisation for the reduction of peri-operative cardiovascular complications, additional evidence suggests a potential benefit in certain population subgroups. Subsequent analysis of the CARP trial including registry patients excluded from the original study discovered that subjects with significant LM disease were the only group to experience a significant mortality benefit with revascularisation.33
Additionally, some evidence suggests that surgical revascularisation and PCI may not be equivalent. A meta-analysis of both the CARP and DECREASE-V trials that specifically looked at the method of revascularisation performed found that PCI-treated patients had significantly increased rates of 30-day MI and composite MI plus death, while CABG patients experienced less long-term composite death and non-fatal MI.34 Further secondary analysis of the CARP trial revealed that patients undergoing CABG had a statistically significant increase in the completeness of revascularisation over PCI and that peri-operative MI was inversely related to the completeness of revascularisation.35
The decision to proceed with surgical revascularisation must take into account not only the combined risk of CABG and non-cardiac surgery together, but must also be weighed against the risk inherent in delaying a potentially necessary non-cardiac surgery. With these considerations in mind, the guidelines currently support a strategy of CABG and subsequent delay of non-cardiac surgery only in cases where patients would otherwise meet independent criteria for surgical revascularisation.13,36
Percutaneous Coronary Intervention
The accumulation of data thus far indicates that the role of PCI in prophylactic revascularisation should be largely reserved for patients suffering from acute or unstable coronary syndromes who would otherwise meet criteria for urgent revascularisation.37,38 Interestingly, previous retrospective analysis of patients who underwent PCI without stent placement (angioplasty alone) <90 days from the time of surgery were at similar risk of cardiovascular complications to those with non-revascularised CAD, yet those treated with similar PCI >90 days from surgery had fewer cardiac events than unrevascularised matched controls while still experiencing more than controls without significant CAD.39 For the vast majority of patients, PCI in the current era involves placement of either a bare-metal or a drug-eluting stent. PCI with stent placement presents a unique challenge in the peri-operative management of patients given the need for antiplatelet therapy and the time needed for adequate vessel healing and endothelialisation. The denudation of endothelium after PCI combined with the thrombogenicity of a newly placed stent is of particular concern when facing the physiological stress of surgery.
Observational data suggest an increased risk of adverse cardiac events in patients undergoing surgery shortly after PCI. Two retrospective studies showed a high number of peri-operative MIs and deaths in patients who underwent surgery within two weeks of bare-metal stent placement largely as a consequence of stent thrombosis.40,41
Drug-eluting stents are designed to limit neointimal hyperplasia and subsequently reduce target vessel revascularisation compared with bare-metal stents. The locally delivered drug aimed at reducing in-stent restenosis consequently delays the endothelialisation process and ultimately lengthens the duration of dual antiplatelet therapy needed to limit thrombogenicity. Thus, the type of stent implanted may directly affect the timing of non-cardiac surgery by balancing the likelihood of possible in-stent restenosis with bare-metal stents within six months against the improved outcomes and longer duration of antiplatelet therapy required with the use of drug-eluting stents. Despite the improvements drug-eluting stents have provided in the treatment of CAD, recent registry data suggest no difference in peri-operative cardiac events when comparing the outcomes at two years among patients treated with either bare-metal or drug-eluting stents.42 Similarly, the randomised DECREASE-V trial utilised drug-eluting stents for PCI and again showed no reduction in peri-operative cardiac complications.32
The timing of antiplatelet therapy discontinuation is of critical importance to the safety of patients with previous PCI undergoing non-cardiac surgery. The combination of aspirin and a thienopyridine has been shown to significantly reduce the risk of stent thrombosis until adequate endothelialisation occurs.43 Currently, the guidelines suggest delaying elective surgery until 14 days after balloon angioplasty, 30–45 days after bare-metal stent placement and one year after treatment with a drug-eluting stent (see Figure 2).13 In a small, single-centre registry, patients with previous PCI who had early discontinuation of their antiplatelet therapy at the time of surgery suffered significantly higher adverse cardiac event rates regardless of the type of PCI performed; however, no adverse events were seen in patients who had completed their full treatment course before surgery.44
Conclusion
As the population ages an increasing number of patients with CAD or CAD risk factors are proceeding with non-cardiac surgery. Cardiovascular complications are common and are associated with significant morbidity and economic burden. Risk-prediction models have aided in identifying patients at high risk of peri-operative cardiac complications, but not those in which revascularisation may benefit.45 Currently, pharmacotherapy with beta blockade and possibly statins has provided the best hope for risk reduction, while prophylactic revascularisation solely for the purpose of improved peri-operative outcomes has been disappointing. The knowledge that mild coronary stenoses are as likely to be responsible for coronary thrombosis and peri-operative MI as severe stenoses are may explain this difference. Continued efforts at risk reduction and identifying active cardiac conditions to treat is paramount in the evaluation of the pre-operative patient, as is constant vigilance and the timely treatment of those who suffer from a peri-operative cardiac complication.