Preoperative cardiac evaluation and management of patients undergoing elective non-cardiac surgery

Risk stratification starts with simple bedside evaluation that integrates clinical risk factors, functional capacity and type of surgery. Patients at low risk could be offered early surgery after assessment by their general practitioner, while complex patients may need more detailed assessment by a perioperative physician or cardiologist, in liaison with anaesthetists, surgeons and GPs. This approach facilitates more efficient professional decision making, better communication with primary care-based teams, more rapid optimisation of a patient’s medical fitness for surgery, and more targeted postoperative management. It also allows patients to be better informed of both the potential benefits and risks of surgery when giving consent.

Investigations should only be performed if: a) the results are expected to accurately and significantly change clinical estimates of risk; b) these altered risk estimates consistently lead to changed management decisions; and c) the resultant management changes have been shown in clinical trials to improve clinical outcomes. As situations that satisfy all three of these criteria are rare in perioperative medicine, the value of investigations, apart from a routine 12-lead electrocardiogram (ECG), is limited in preoperative cardiac management. The most useful applications may be in reclassifying intermediate-risk patients to either low-risk (surgery can safely proceed without further intervention) or high-risk (needing more detailed evaluation and use of prophylaxis), or in determining unacceptable surgical risk in high-risk patients undergoing high-risk surgery (Box 3).8

β-Blockers are potentially useful in lowering cardiac risk by antagonising the effects of adrenaline and other stress hormones and exerting negative chronotropic and inotropic actions. However, results of randomised trials and meta-analyses suggest mixed effects, with further uncertainty resulting from the recent disclosure of several potentially fraudulent or negligent Dutch trials.26 The large Perioperative Ischemic Evaluation Study (POISE) showed a 31% reduction in the risk of non-fatal MI with β-blockers, at the expense of a 34% increased risk of all-cause mortality and 89% increased risk of non-fatal stroke.27 A recent meta-analysis of nine well conducted “secure” trials (including POISE, and excluding the “non-secure” Dutch trials) found initiation of β-blockers before surgery caused a 27% increase in 30-day all-cause mortality and a 73% increase in non-fatal stroke, while decreasing risk of non-fatal MI by 27%.28 The updated 2009 American College of Cardiology and American Heart Association guidelines give a Class 1 recommendation only for continuing β-blockers in patients with a pre-existing cardiac condition for which there is a strong indication.29

However, two large retrospective observational studies using propensity-based risk adjustment suggest that β-blockers reduce all-cause inhospital deaths proportionally to increasing cardiac risk, as measured by an RCRI score ≥ 2, while increasing deaths in those with an RCRI score < 2 (Box 5).30,31 In the former patients, one of the studies showed that, while β-blockers reduced risk of non-fatal MI and cardiac arrest, stratified analyses indicated these benefits were limited to patients undergoing non-vascular surgery.31

It thus remains unclear which patients benefit from β-blockers. If β-blockers are to be initiated, observational data suggest they be restricted to high-cardiac-risk patients,30,31 and should be commenced some weeks before surgery and haemodynamically titrated to a tolerable dose that lowers resting heart rate to 70 beats/min.32 Longer-acting agents such as atenolol appear to be safer than short-acting agents such as metoprolol.33

Statins improve endothelial function, reduce vascular inflammation and stabilise atherosclerotic plaque. Evidence of benefit of perioperative treatment in statin-naive patients is of limited quality and is dominated by observational studies34 or trials in cardiac surgery.35 The few secure randomised trials involving patients undergoing non-cardiac vascular surgery are underpowered and inconclusive.36 Patients already prescribed statins as chronic therapy should continue treatment in the perioperative period.

Angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs) have not been shown to improve outcomes in the absence of left ventricular systolic dysfunction. Indeed, observational studies suggest they predispose to severe intraoperative hypotension (generally responsive to fluid loading and vasopressors), especially if combined with β-blockers or diuretics, and may increase 30-day mortality in patients undergoing major vascular surgery.37 There is debate about whether these agents should be withheld one half-life before anaesthesia induction if their indication is purely for hypertension (unless blood pressure is uncontrolled) or, given the preponderance of day-of-surgery admissions, to recommend continuation with adequate hydration.38 A prospective randomised trial is required to clarify the safety of perioperative use of these agents.

Aspirin interacts with the cyclo-oxygenase enzyme system and irreversibly inhibits platelet aggregation, theoretically lessening risk of coronary thrombosis but increasing risk of perioperative bleeding. No adequately powered trial has assessed benefits of aspirin prophylaxis in aspirin-naive patients. In patients with known CAD, excluding those with recent coronary artery stent insertion (discussed below), risk of subsequent death or MI is increased two- to threefold if aspirin is ceased before surgery.39 While the risk of major postoperative bleeding may offset this cardiac risk for certain procedures, such as extensive skin grafting, a recent meta-analysis of 41 studies involving 49 590 surgical patients shows that, overall, the cardiac risk exceeds bleeding risk for most surgical patients with known CAD whose aspirin is withheld.40

Percutaneous coronary intervention (PCI) or coronary artery bypass grafting is only indicated before non-cardiac surgery in clinically unstable patients (those with unstable angina, recent MI or ventricular arrhythmias) with significant left main or three-vessel (or two-vessel if this includes the proximal left anterior descending artery) CAD. A large trial failed to show any perioperative or long-term benefit of prophylactic revascularisation, compared with optimal medical treatment alone, in stable patients undergoing high-risk surgery.41

Large observational studies show that symptomatic congestive heart failure (CHF) increases the absolute risk of perioperative death to 8% — more than twice the risk seen in established CAD without CHF.42 Other studies suggest stable, well controlled CHF does not necessarily increase risk.43 Current guidelines are uncertain about when left ventricular function should be reassessed using echocardiography in clinically stable patients with known CHF.7,18 The ability of BNP and N-terminal proBNP to discriminate cardiac risk among patients with CHF, who may have chronically elevated levels, has yet to be examined.18 It is also unknown whether optimising CHF management before surgery — including using a BNP-guided strategy to titrate therapy, correcting coexisting anaemia and strictly controlling ventricular rate in patients with atrial fibrillation — improves postoperative outcomes.44 What is agreed is the need to defer surgery in patients with decompensated or severe chronic CHF (worsening or new-onset CHF; New York Heart Association Class IV symptoms) until they are medically optimised and euvolaemic. In patients with newly diagnosed CHF, elective surgery should be delayed 3 months or more to allow adequate time for antifailure therapies to improve left ventricular function and remodelling.44 β-Blockers with proven mortality benefit (bisoprolol, carvedilol or metoprolol succinate) and ACE inhibitors or ARBs should be continued during the perioperative period unless precluded by hypotension or symptomatic bradycardia.

In all patients with clinical features consistent with severe valvular heart disease, preoperative echocardiography and a 12-lead ECG are mandatory in assessing valve and left ventricular dysfunction. These are also important to screen for conduction system defects caused by perivalvular fibrosis that may predispose to bradyarrhythmias requiring perioperative pacing. Ideally, patients eligible for valve reconstruction or replacement, or transcutaneous valvuloplasty or valve implantation, should undergo these procedures before elective surgery. This is particularly pertinent in patients with symptomatic or critical calcific aortic stenosis with a valve area < 0.8 cm2, which carries a 10%–28% risk of perioperative sudden cardiac death.45 Patients with severe mitral regurgitation should be medically optimised before surgery, including rate control of chronic atrial fibrillation. ACE inhibitors or ARBs prescribed for afterload reduction should be continued despite a risk of anaesthesia-induced hypotension. In the absence of prior history of infective endocarditis, mechanical prosthetic heart valves, congenital heart defects, cardiac transplantation with valvulopathy, or rheumatic heart disease in Indigenous Australians, prophylactic antibiotics are usually not required. Exceptions to this are operations associated with a high risk of bacteraemia (dental procedures and periodontal disease; genitourinary procedures; surgery involving the oropharynx, respiratory tract, sinuses, nose or ear; incision and drainage of local abscesses; or surgery through infected skin).8,46

Australian guidelines from 2009 recommend that elective surgery requiring cessation of dual antiplatelet therapy should be postponed for at least 6 weeks after insertion of bare-metal stents and 12 months after insertion of drug-eluting stents (DES).47 However, more recent American guidelines reflecting additional new evidence and experience with later-generation DES suggest a minimum period of 6 months after insertion of DES.48 A recent retrospective cohort study of more than 28 000 patients who underwent non-cardiac surgery within 2 years after stent insertion showed that major adverse cardiac events at 30 days were associated with emergency surgery, history of MI in the 6 months before surgery and an RCRI score greater than 2, but not with stent type or timing of surgery beyond 6 months after stent insertion.49 Ceasing dual therapy earlier than stipulated above carries a very high risk of stent thrombosis, with mortality rates up to 20%.50 Continuation of dual therapy confers little risk of major bleeding in most minor surgery (Box 6). In patients requiring urgent surgery associated with high bleeding risk within the recommended minimum time frames, aspirin should be continued and clopidogrel (or prasugrel or ticagrelor) withdrawn at least 5 to 7 days before surgery, depending on the agent.8,47 This should be coupled with consideration of bridging anticoagulation (heparin–tirofiban or heparin–eptifibatide) in selected highest-risk patients (although there are limited data in support of such treatments).47 Any discussion about postponing surgery or continuing, modifying or discontinuing antiplatelet therapy must involve close liaison between a patient’s GP, interventional cardiologist, anaesthetist, surgeon and haematologist to balance the risk and benefit of such decisions. Patients scheduled for PCI and requiring non-cardiac surgery in the foreseeable future should preferably receive bare-metal stents.

Whether and when to withhold anticoagulants depends on the balance between risk of thromboembolic events if interrupted and risk of major bleeding if continued (Box 6). Patients at low thromboembolic risk can cease taking anticoagulants with no need for bridging heparin, while those undergoing minor procedures with low bleeding risk do not require their cessation.51 In high-risk patients, bridging heparin is required after oral anticoagulants are ceased 5 days (for warfarin)51 or between 24 hours and 4 days (for the newer oral agents dabigatran, rivaroxaban and apixaban, as per manufacturer’s product information for each) before surgery. Bridging anticoagulation with subcutaneous low-molecular-weight heparin, if there are no contraindications, obviates the need for hospitalisation to administer intravenous unfractionated heparin. Bleeding risk with the newer anticoagulant agents is of concern, given the lack of both an antidote and reliable assays of anticoagulation effects. Early, effective and ongoing communication between GPs and specialists, combined with reference to detailed, up-to-date protocols, is required to maximise patient safety during perioperative transitions of anticoagulation.52

Obstructive sleep apnoea (OSA) affects up to 25% of adult general surgical patients and up to 77% of those undergoing bariatric surgery.53 As many as 70% of cases are undiagnosed before patients present for preoperative evaluation. In a recent meta-analysis of case–controlled and cohort studies of patients diagnosed with OSA and undergoing elective surgery, postoperative cardiorespiratory events were twice those seen in patients without OSA (3.8% v 1.7%).53 Various screening questionnaires with equivalent predictive value in identifying patients with moderate to severe OSA are easy to administer.54 In cases where known OSA is mild or screening risk is low, surgery is low risk and there are no associated comorbidities, surgery can proceed without further intervention. In all other cases, formal evaluation by a sleep physician, initiation or titration of continuous positive airway pressure (CPAP) therapy where indicated, and close liaison with anaesthetists should be undertaken. The optimal duration of CPAP therapy in newly diagnosed patients awaiting surgery and how patients with known OSA who are non-compliant with CPAP therapy should be treated remain uncertain.

Chronic obstructive pulmonary disease (COPD) frequently coexists in patients with CAD or CHF who are, or have been, smokers. COPD is an independent risk factor for major cardiopulmonary complications and can complicate assessment of functional capacity and administration of prophylactic β-blockers. Clinical history and simple bedside spirometry are sufficient to gauge disease severity in otherwise stable patients. Routine chest x-rays and formal lung function tests add little value. In the absence of moderate to severe bronchospasm, a meta-analysis supports the safety of cardioselective β-blockers in most patients with stable COPD.55 Patients with combined bronchospastic disease and CAD who are undergoing high-risk surgery might derive cardioprotective benefit from α-2 adrenergic agonists (such as clonidine).56 Before surgery, patients with unstable COPD or asthma should receive oral steroids, which do not compromise wound healing, and all patients with COPD should totally abstain from smoking for at least 6 weeks.

For patients with these devices, especially implantable cardioverter defibrillators (ICDs), GPs or anaesthetists should ideally contact the relevant cardiologist to ascertain the type of device, its indications, current settings and mode of magnetic inactivation (if applicable). Such information allows appropriate safeguards to be organised, if required, before surgery.57 Surgical diathermy, particularly in chest, head or neck surgery, can cause electrical interference that may inhibit pacemakers or trigger shocks from ICDs.