2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM Guideline for Perioperative Cardiovascular Management for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice GuidelinesFree Access

Modified: formatting changes (eg, minor modifications such as PICO[TS] [patient population, intervention, comparator, outcome, time, setting] structure)

Adapted: substantive changes (eg, major adaptations, such as a change in Class of Recommendation [COR], Level of Evidence [LOE], drug or device classification).

There are several indices available for perioperative cardiovascular risk prediction. The American Society of Anesthesiologists (ASA) Physical Status Classification System classifies patients into categories according to their overall health status.¹⁰ With 6 predictors of risk (1 point assigned for each criterion), the RCRI is a simple, validated, and commonly used tool to assess perioperative risk of major cardiac complications.³ In a pooled analysis of 24 validation studies, the RCRI had modest risk discrimination for cardiac events in patients undergoing NCS, although there is discordance among the various risk-prediction tools in identifying low-risk patients, defined as having an estimated risk of MACE of <1%.^7,11 The Surgical Outcome Risk Tool estimates 30-day mortality after NCS based on the ASA Physical Status grade, urgency of surgery, surgical specialty and severity, cancer, and age ≥65 years.¹² The 21-component universal NSQIP surgical risk calculator may provide superior predictive discrimination.¹³ The AUB (American University of Beirut)-HAS2 cardiovascular risk index is easily calculated and used to assess 30-day event risk, stratifying patients undergoing NCS into low (score 0-1), intermediate (score 2-3), and high risk (score >3) based on 6 data elements.¹⁴ A simplified method using 3 traditional risk factors for hypertension, diabetes, and current smoking identified a low incidence of MI (0.10%) among patients without risk factors who underwent NCS.¹⁵

In a study of 600 patients undergoing NCS, self-reported poor functional capacity, defined as the inability to walk 4 blocks or climb 2 flights of stairs, was associated with an almost 2-fold greater risk of in-hospital cardiovascular events (9.6% versus 5.2%; P=0.04).¹ The BASEL-PMI (Incidence and Outcome of Perioperative Myocardial Injury After Noncardiac Surgery) study (n=4560) included patients at elevated cardiovascular risk (ASA class ≥3) and showed that functional capacity <2 flights of stairs was associated with a 1.63 higher rate of death, MI, acute HF, or life-threatening arrythmias at 30 days.² Furthermore, the addition of the functional capacity data to the RCRI significantly increased its predictive power. In a large analysis from NSQIP (n=211,410), functional capacity comprised 1 of 5 elements in a multivariate model predicting MI/MICA at 30 days after surgery.⁴ Functional capacity was classified into 3 categories: independent, partially dependent, and totally dependent. Using the same classification, a retrospective observational cohort study (n=12,324) from the U.S. Veterans Affairs Surgical Quality Improvement Project demonstrated that functional capacity was independently associated with mortality and added discriminatory power to traditional ASA classification.⁵ The METs study compared the power of the DASI score for predicting death or MI at 30 days after major NCS with cardiopulmonary exercise testing (CPET) and a subjective assessment of functional capacity by the anesthesiologist, classified as good (>10), moderate (4-10), or poor (<4).⁷ Subjective assessment of functional capacity by the clinician was not associated with outcomes, whereas the DASI score was associated with death or MI at 30 days after surgery (N=1401, mean age 65 years).⁷ DASI scores ≤34 were associated with increased odds of 30-day death or MI.⁸

Frailty is a risk marker for adverse outcomes after NCS and for reduced benefit after cardiac procedures.^1,3,4 In a meta-analysis of 56 studies involving 1.1 million older adults undergoing NCS, frailty was associated with an increased risk of 30-day mortality (relative risk, 3.71 [95% CI, 2.89-4.77]) and 30-day complications (relative risk, 2.39 [95% CI, 2.02-2.83]).⁴ Several validated tools are available to assess for frailty (Table 6).¹² Although most comparative studies have been neutral, 1 prospective evaluation found that the Clinical Frailty Scale outperformed the Fried phenotype and the Frailty Index for predicting death, disability, prolonged length of stay, or nonhome discharge after NCS in older adults.⁵ In a single-center observational study involving 9153 patients undergoing major NCS, incorporation of routine frailty screening into the preoperative assessment was associated with a significant reduction in 30-day mortality.² In some cases, older patients with advanced frailty, poor functional status, and reduced life expectancy may derive limited benefit from surgery; in these patients, goals of care and shared decision-making should be integrated into preoperative planning.¹³ In selected patients, prehabilitation before NCS may be associated with improved outcomes.^9-11,14

Several prospective observational studies and meta-analyses have documented the use of preoperative BNP and NT-proBNP concentrations to predict postoperative complications in selected patients undergoing NCS. Both preoperative and postoperative natriuretic peptide levels were independent predictors of the composite outcome of death or nonfatal MI at 30 days and at 180 days of follow-up.³ In a nested substudy of VISION (Vascular Events in Noncardiac Surgery Patients Cohort Evaluation), preoperative NT-proBNP concentrations >100 pg/mL were independently associated with all-cause mortality.¹ Optimal threshold values of BNP or NT-proBNP for perioperative risk prediction are not clearly established. However, in a recent cohort study evaluating 3597 patients undergoing NCS, the addition of NT-proBNP to traditional risk scores did not significantly improve risk prediction beyond that of risk scores combined with self-reported measures of functional status.⁷

Preoperative high-sensitivity cTn concentrations in patients without symptoms or signs of ischemia can identify patients with chronic myocardial injury as well as those at increased risk during and after the procedure.^4,8 There is no action predicated on this knowledge alone, although preoperative baseline troponin concentrations also inform the interpretation of postoperative troponin measurements and can help confirm a diagnosis of acute myocardial injury in the postoperative setting. There is limited evidence from heterogenous populations that preoperative cTn can predict short- or long-term adverse outcomes.⁴ The predictive performance of the RCRI to identify patients who develop perioperative MACE is improved with the addition of preoperative cTn concentrations.⁵

A preoperative 12-lead ECG is likely to be more valuable for patients planned for elevated-risk surgical procedures.^2,3 This is particularly true for patients with known coronary heart disease, arrhythmias, peripheral arterial disease, cerebrovascular disease, or other significant structural heart disease. Comparing a preoperative ECG with previous electrocardiographic tracings may be helpful whenever relevant abnormalities are identified, with the preoperative ECG serving as a baseline if postoperative complications develop.

The prognostic significance of several electrocardiographic abnormalities, including arrhythmias, significant pathologic Q-waves, LV hypertrophy, ST-segment depressions, QT_c prolongation, and bundle branch blocks has been identified in observational studies.^4,5 Most studies, however, report little to no added prognostication of ECGs beyond clinical risk assessment. The abnormalities on the ECG that should prompt the preoperative clinician to request further information, consultation, or testing are not well defined; however, notable abnormalities include significant Q-waves, LV hypertrophy, ST-segment elevation, ST depression, T-wave inversion, Mobitz type II or higher block, bundle branch blocks, AF, or QT interval prolongation.^9,12 The likelihood of abnormalities on a preoperative 12-lead ECG increases with patient age and when risk factors for heart disease are present, but a standard age or risk factor cutoff for recommending a preoperative ECG has not been defined. Likewise, the optimal time interval between obtaining a 12-lead ECG and elective NCS is unknown.

In general, an abnormal preoperative ECG may not substantially alter perioperative management, except for second-degree Mobitz type II or higher atrioventricular block,¹³ AF with rapid ventricular response or new-onset AF, or a prolonged QT interval.^4,10,14 Recognition of a prolonged QT interval may inform the selection of anesthetics, postoperative antiemetics, or antibiotic therapy. Incidental findings of Q-waves or bundle branch block on a preoperative ECG in an asymptomatic patient may indicate coronary artery disease (CAD) but should not lead to a decision to perform coronary revascularization before NCS. Another important reason to obtain a preoperative ECG in asymptomatic patients undergoing elevated- risk NCS with increased risk of MACE is to establish a baseline ECG for comparison should a postoperative ECG be abnormal.

Clinical risk assessment using validated tools is more useful to guide management and predict outcomes than are the findings of a single resting preoperative ECG. Available data suggest that in low-risk patients, a routine preoperative ECG has little effect on treatment or complication rates and should be omitted from standard preoperative evaluation.^11,13 One study assessed 30,892 patients undergoing NCS with shockwave lithotripsy for nephrolithiasis, in which a preoperative ECG triggered the cancellation of 13 (0.04%) treatments in low-risk patients (1 with new AF and 12 with ischemia or previous infarction). Of these patients, only 1 had a subsequent abnormal cardiac workup, and the remaining 11 ultimately underwent NCS without complications. The study concluded that in patients at low risk for cardiac complications, preoperative ECG triggered very few cancellations and did not predict cardiac complications after NCS.¹¹ Avoiding unnecessary testing can significantly save resources.¹⁵

Functional capacity of <4 METs (eg, patients unable to climb 1-2 flights of stairs or walk on a flat surface at ≥3 mph) is associated with increased risk for perioperative cardiac events regardless of the cause of the disability (eg, CAD, HF, or a noncardiac condition such as arthritis, chronic lung disease, or obesity).^18,19 In patients with elevated risk for perioperative cardiovascular events, as determined by a validated risk score (Section 3.1, “Cardiovascular Risk Indices”), and functional capacity <4 METs or indeterminate functional capacity, a stress test (exercise or pharmacological, with or without imaging depending on the clinical context) may be considered in select patients undergoing elevated-risk surgery, if high-risk myocardial ischemia is suspected (eg, left main disease or severe multivessel disease with reduced EF) or there is an indication for testing independent of planned surgery.^1,15,16 There is, however, limited evidence to support coronary revascularization before NCS in stable patients.^3,14,20,21 In general, an exercise stress test is preferable to a pharmacological stress test if the patient is able to exercise.^15,16 In patients unable to exercise, selection of a pharmacological stress test modality should be based on patient factors and local availability and expertise.^12,15,16

Stable patients with exercise capacity ≥4 METs are at relatively low risk for perioperative cardiac events.² In addition, although some observational studies have suggested that preoperative revascularization of patients with abnormal stress test findings may be associated with a reduction in postoperative ischemic complications, other studies have found no benefit.¹⁴ In the CARP (Coronary Artery Revascularization Prophylaxis) trial, 510 patients with documented CAD (at least 1 lesion with ≥70% stenosis) were randomly assigned to coronary revascularization or medical therapy before undergoing major vascular surgery.³ Patients in the revascularization group experienced a 36-day delay in time to vascular surgery. There was no difference in the incidence of perioperative MI at 30 days or in all-cause mortality at a median follow-up of 2.7 years.³ More contemporary large RCTs have demonstrated that routine coronary revascularization does not reduce mortality or risk for MI.^20,22,23 Because the benefits of preoperative revascularization appear to be limited, routine preoperative stress testing should not be performed in patients with adequate functional capacity. Similarly, preoperative stress testing is not recommended in low-risk patients (eg, RCRI=0; see Section 3.1, “Cardiovascular Risk Indices”) or in stable patients undergoing low-risk NCS because it is costly, may lead to a delay in surgery, and has not been shown to improve clinical outcomes.^2,14,16,24 This applies to patients with or without known or suspected CAD or cardiovascular risk factors.

In patients with elevated risk for perioperative cardiovascular events, as determined by a validated risk score (Section 3.1, “Cardiovascular Risk Indices”) and low (<4 METs or a DASI score of ≤34) or indeterminate functional capacity, CCTA may be considered in select cases if high-risk coronary anatomy is suspected and there is a guideline-concordant indication for testing independent of planned surgery. Risk assessment by CCTA offers incremental risk assessment over clinical risk scores. The Coronary CTA-VISION (Coronary CT Angiography to Predict Vascular Events in NCS Patients Cohort Evaluation) study (N=987) showed that use of CCTA marginally improved the risk estimation for predicting postoperative cardiovascular death and nonfatal MI compared with clinical risk score alone.¹ CCTA confers a high negative predictive value for excluding perioperative cardiovascular events.^1-4 These results were confirmed in a meta-analysis including 11 studies and 3480 patients who underwent preoperative CCTA.² The presence, extent, and severity of coronary atherosclerosis directly correlated with risk of perioperative MACE. On CCTA, patients with single- and multivessel disease demonstrated a 3-fold and an 8-fold increased risk of perioperative MACE, respectively.² However, as listed in Section 6.1.1 (“Coronary Revascularization”), evidence supporting routine coronary revascularization before planned surgery is lacking.

CCTA can improve risk estimation among patients who will experience perioperative MACE; however, compared with clinical risk scores, CCTA is >5 times as likely to inappropriately overestimate risk among patients who will not experience MACE.¹ This overestimation of risk may result in a delay or cancellation of surgery and unnecessary increased use of medical resources, thereby contributing to patient morbidity and mortality and increased medical expenses. Low-risk patients include those at low risk for cardiovascular events based on validated risk tools and those who are undergoing low-risk noncardiac surgical procedures (Section 3.1, “Cardiovascular Risk Indices,” Section 1.5, “Definitions of Surgical Timing and Risk”). In these patients, or those with functional capacity ≥4 METs with stable symptoms, the probability of uncovering high-risk coronary anatomy that would adversely affect surgical outcomes is small. This recommendation is applicable to patients with or without known or suspected CAD or cardiovascular risk factors.

Routine invasive angiography to assess the need for coronary revascularization is not recommended in the preoperative evaluation of NCS. The CARP trial of 510 patients assessed the benefit of preoperative coronary artery revascularization among patients with chronic CAD scheduled for elective vascular surgery and reported that coronary artery revascularization before elective vascular surgery does not significantly alter long-term outcomes.² Two small European RCTs also evaluated routine versus selective ICA before vascular surgery. In a trial enrolling 426 patients undergoing carotid endarterectomy without a previous history of CAD, routine preoperative ICA was associated with lower rates of early postoperative MI and improved long-term survival compared with surgeries without prior ICA.⁶ In a separate study of 208 patients undergoing major vascular surgery with an RCRI ≥2, in-hospital risk of MACE was not different in patients assigned to routine versus selective ICA, although routine ICA was associated with improved survival and freedom from MACE at approximately 5 years of follow-up.⁵ Unfortunately, in contrast to the more robust CARP trial, these 2 trials were small, nonblinded, and management of patients assigned to the selective ICA arms was not clearly defined; these findings have not been replicated in larger, more contemporary studies. In contrast, large trials that are not specific to perioperative care indicate that in the setting of stable CAD, invasive management with coronary revascularization of CAD in epicardial vessels other than the left main does not improve short- or long-term survival versus optimal medical therapy alone.⁷ Thus, ICA should be reserved for the highest-risk patients. A strategy of routine ICA with an intent to perform revascularization before elective NCS in patients with chronic coronary syndromes cannot be recommended. In general, indications for preoperative coronary angiography should be similar to those identified in nonoperative settings, such as ACS, accelerating angina despite maximal antianginal therapy, newly diagnosed moderate-severe ischemia on stress testing, and indicators of obstructive left main disease on noninvasive testing.

Uncontrolled hypertension is associated with increased perioperative complications. In a retrospective analysis of 251,000 adults undergoing elective NCS from the UK Clinical Practice Research Datalink, preoperative DBP >90 mm Hg was associated with increased 30-day postoperative mortality.²¹ Accordingly, ongoing treatment of chronic hypertension is recommended in the perioperative period. If hypertensive patients are unable to take oral medications, it is reasonable to use intravenous medications to control BP. Tight BP control mitigates long-term cardiovascular risk, but this strategy may not be appropriate for all patients in the perioperative period. Whereas maintaining an SBP >90 mm Hg may be an acceptable target for younger adults,²⁰ a higher target may be preferred in older adults, those with chronic hypertension, or both.²¹ The decision whether to hold or continue antihypertensive medications may be specific to drug class. Certain medications (eg, beta blockers, clonidine) may be associated with rebound hypertension if discontinued abruptly,²⁸ whereas others have been associated with increased risk of intraoperative hypotension when continued (eg, angiotensin-converting enzyme inhibitors [ACEi], angiotensin-receptor blockers [ARBs]).^29,30 Refer to Sections 7.2 through 7.4 for further guidance on specific classes of antihypertensive drugs.

In patients with untreated or uncontrolled hypertension, induction of anesthesia can trigger sympathetic activity, resulting in labile BP and heart rate.³¹ In a systematic review and meta-analysis of 30 observational studies, a preoperative diagnosis of hypertension was associated with a 35% increase in cardiovascular complications.² A single-center retrospective analysis of 58,276 patients undergoing NCS identified an association between preinduction SBP >160 mm Hg and a composite outcome (cardiac, neurological, or renal complication or in-hospital mortality) that was only significant in patients (n=10,512) with ≥1 components of the RCRI (eg, CAD, congestive HF, cerebrovascular accident, baseline serum creatinine >2.0 mg/dL, or preoperative insulin treatment).¹ However, an elevated BP on the day of surgery may represent a situational (”white coat hypertension”) response.³² Therefore, referring to patients’ baseline ambulatory BP is recommended to guide management. In the absence of RCRI components,¹ there is little evidence for increased risk of perioperative complications in patients with preoperative BP of <180/110 mm Hg² or at any preinduction BP.

Intraoperative hypotension is associated with postoperative myocardial injury, acute kidney injury, and mortality.^4-6,8,23 The harm threshold in observational analyses appears to be roughly a MAP <65 mm Hg or an SBP <90 mm Hg maintained for about 15 minutes.^4,33 However, the results of 3 key trials are challenging to interpret. The multicenter, randomized INPRESS (Intraoperative Noradrenaline to Control Arterial Pressure) trial studied 298 high-risk patients and reported a roughly 25% relative risk reduction if SBP was maintained within 10% from baseline versus >80 mm Hg.²⁵ Another single-center trial³⁴ randomized 458 high-risk patients to a MAP ≥60 mm Hg versus MAP ≥75 mm Hg, and POISE-3 (Perioperative Ischemic Evaluation-3)⁹ randomized 7490 patients to a hypotension-avoidance (target MAP ≥80 mm Hg) versus hypertension-avoidance (target MAP ≥60 mm Hg) intraoperative strategy. Neither of these reported benefit from a strategy targeting higher BP. However, interpretation of the INPRESS study is complicated by lack of details reported with respect to extent of hypotension, especially in the potentially harmful MAP range of 55 to 70 mm Hg. Thus, while the expert opinion is to maintain intraoperative BP targets above MAP ≥60 to 65 or SBP >90 mm Hg, there is currently insufficient trial evidence to support higher BP targets.

In the POISE-2 (Perioperative Ischemic Evaluation-2) substudy, there was increased risk of the composite outcome of MI and death for increasing duration of SBP <90 mm Hg through postoperative day 4 (OR, 2.83 per 10-minute increase).⁶ However, anesthetic and hypotension management (eg, fluid boluses, vasoactive drugs, and mechanical support) was at the discretion of the clinical team and not controlled or reported. Closer monitoring of postoperative patients in the intensive care setting may allow for earlier recognition of hypotension.³⁵ A systematic review and meta-analysis on the perioperative use of vasoactive drugs (including inotropic agents and vasoconstrictors) to treat hypotension concluded that their use may reduce postoperative complications and reduce the length of stay in adult patients having major abdominal surgery.⁷

Postoperative hypertension can occur as a result of a variety of stimuli, including pain, inflammation, anxiety, hypoxia, volume overload, urinary retention, or as a result of withdrawal of chronic antihypertensive medications.³⁶ Hypertension can increase the risk for myocardial ischemia/infarction, acute decompensated HF, cerebral ischemia, and dysrhythmias.³⁶ Propensity-matched retrospective Veterans Affairs cohort studies found that delaying the resumption of preoperative ACEi/ARBs was associated with increased 30-day mortality risk.^37,38 Chronically taken oral antihypertensive medications should be restarted as soon as clinically reasonable to avoid complications from postoperative hypertension. However, results from a nonrandomized propensity-matched cohort study of men ≥65 years with hypertension caution against intensification of antihypertensive therapy at hospital discharge due to an increased 30-day risk of readmission and serious adverse events without improvement in 1-year cardiovascular events.³⁹

SGLT2i have been associated with metabolic acidosis and euglycemic ketoacidosis in the perioperative period, which can lead to serious complications, prolonged hospital stay, and death.^1,2 The mechanism underlying SGLT2i-induced ketoacidosis is thought to be similar to starvation ketosis, with intensification of the normal metabolic effects of these agents by perioperative fasting. The diagnosis of euglycemic ketoacidosis may be missed because symptoms are often nonspecific, such as nausea, abdominal pain, and shortness of breath, and blood glucose levels may be normal or only mildly elevated. The diagnosis should be suspected when there is an anion gap metabolic acidosis and ketones in the blood or urine. In 2022, the U.S. Food and Drug Administration (FDA) updated the labeling for SGLT2i to recommend discontinuing these agents 3 to 4 days before surgery when feasible.³ This recommendation has been endorsed by the American Diabetes Association and other organizations.¹⁴

GDMT for HF with reduced EF (HFrEF) includes 4 medication classes that have been shown in multiple trials to reduce mortality and morbidity: (1) renin-angiotensin system inhibition with angiotensin receptor-neprilysin inhibitors, ACEi, or ARBs alone; (2) beta blockers; (3) mineralocorticoid receptor antagonists; and (4) SGLT2i.¹² There are also guideline-recommended therapies for HF with mildly reduced EF, HF with preserved EF, and HF with improved EF.¹² Similarly, patients with stage B pre-HF, including those with asymptomatic systolic or diastolic dysfunction, should be managed in accordance with guideline recommendations.¹² In patients hospitalized for HFrEF, discontinuation of GDMT in the absence of a direct contraindication has been associated with increased mortality and readmission.^4-8 Thus, continuation of medical therapy for HF is reasonable and likely to be beneficial for most patients undergoing NCS.

The postoperative development of pain, anemia, electrolyte imbalance, fluid shifts, and sepsis in NCS patients can contribute to new-onset POAF and a rapid ventricular response. Aggressive management of these underlying triggers is an essential component of POAF management. Hemodynamically stable patients with POAF may require specific therapy to achieve an optimal heart rate (<110 bpm). Medications that block atrioventricular nodal conduction, such as beta blockers or CCBs, can be used for ventricular rate control, and digoxin can be considered as an adjunct or if other agents are contraindicated.^21-24 Among patients with AF refractory to rate control with atrioventricular nodal blockers, rhythm control with synchronized electrical direct current cardioversion or pharmacological cardioversion with antiarrhythmic drugs can be considered. Exclusion of left atrial appendage thrombus before implementing rhythm control may be indicated in patients with a prolonged duration (>48 hours) of AF or in patients at high risk for thromboembolism. Synchronized direct current cardioversion is recommended for hemodynamically unstable AF with a rapid ventricular response associated with hypotension.

New-onset POAF after NCS is associated with thromboembolic risks comparable to those in nonsurgical patients with AF.^6,19 In a meta-analysis of 2,458,010 patients, POAF was associated with a 62% increased risk of early stroke and a 44% increased risk of early mortality within 30 days of surgery versus patients without POAF.⁶ POAF was also associated with a 37% increased risk of long-term stroke and a 37% increased risk of long-term mortality.⁶ In subgroup analyses, POAF was more strongly associated with stroke in patients undergoing NCS (hazard ratio [HR], 2.00 [95% CI, 1.70-2.35]) than in patients undergoing cardiac surgery (HR, 1.20 [95% CI 1.07-1.34]).⁶ It is unclear whether stroke mechanisms are the same in patients with POAF compared with those with nonsurgical AF. In patients with POAF after NCS, anticoagulation should be considered based on a patient’s thromboembolic stroke risk (eg, CHA₂DS₂-VASc score) and bleeding risk. In a Danish registry analysis of patients with AF after NCS, use of OAC initiated within 30 days postdischarge was associated with a 48% reduced risk of thromboembolic events compared with no anticoagulation therapy.⁴ The use of oral nonvitamin K anticoagulant or no anticoagulation in patients with new-onset AF in the postoperative period after NCS is being tested in ASPIRE-AF (Anticoagulation for Stroke Prevention in Patients with Recurrent Episodes of Perioperative AF after NCS), an ongoing RCT.¹⁹

In a longitudinal database of 10,723 patients with newly diagnosed AF (age 68±10 years, 41% women), 15% developed POAF after NCS. POAF after NCS was associated with a 39% risk of AF recurrence at 5 years, with an increased risk of HF and death.²⁰ Given these findings, POAF after NCS warrants timely outpatient follow-up to coordinate AF surveillance, determine the need for anticoagulation, optimize risk factors, titrate medications for rate control, and consider rhythm control. More intensive monitoring, such as with 1- to 2-week ambulatory electrocardiographic monitoring, 30-day electrocardiographic event monitoring, or implantable cardiac monitoring in selected patients, may be warranted. In cardiac surgery patients, continuous monitoring postsurgery is associated with a higher detection of AF.^7-9 This is consistent with previous studies showing higher sensitivity of AF detection with longer-term monitoring.^10,11 Although the optimal frequency, duration, and type of rhythm monitoring with postoperative AF remains unclear, outpatient follow-up within 3 to 6 months of NCS to evaluate the incidence of AF after NCS is recommended.²⁵

The type of CIED, the manufacturer, the model, primary indication for placement, pacing dependency, effects of a magnet, and proper functioning of the CIED must be determined.^1,5 Information is obtained from medical records, patient identification cards, the CIED management team, interrogation reports, or a preoperative interrogation.³ Clinicians should confirm battery status through recommended interrogations (every 12 months for a pacemaker, every 6 months for an ICD).¹ A plan for reprogramming or use of a magnet should be developed before beginning procedures, with input from the CIED professionals, anesthesiologists, and surgeons. Institutions should have a CIED protocol in place (Figures 3 and 4).^5,16 Management plans should incorporate the procedure details, patient positioning (eg, prone), and the type of anticipated EMI. Recommendations should state specific programming changes if needed, response to a magnet, patient pacemaker dependency, battery life, and history of serious arrhythmias.^1,5 Recommendations and changes need to be included in the medical record and accessible to all perioperative clinicians.^1-5

Magnet pacing rates vary by manufacturer and battery status. With ample battery supply, devices pace at 85 to 100 bpm.¹⁵ Magnet rates decrease with battery depletion. It is important to confirm asynchronous (VOO or DOO) pacing at the expected rate with magnet use. Magnet use with some pacemakers only forces asynchronous pacing for 10 beats and then reverts to programmed settings.¹⁷ These devices must be reprogrammed to avoid EMI. Magnets may not be effective in obese or prone patients, or if continuous generator contact cannot be ensured.^1,12 EMI from ESU, monitors, bone saws, or patient movement can trigger rate-responsive functions, resulting in undesirable tachycardia that may cause unwanted hemodynamic effects or may be misinterpreted.⁷ Deactivating the rate adaptive function is recommended during lithotripsy or electroconvulsive therapy. Monopolar ESU require a dispersive electrode to divert current away from the CIED to avoid damage. EMI is more common with monopolar than bipolar ESU and with coagulation versus cutting mode.^13,14 The dispersive electrode is positioned as far as possible from the generator and the heart. The current path should not cross the generator or the leads.⁹ There is a risk of EMI even with ESU below the umbilicus with underbody dispersive electrodes, and practitioners need to use a magnet or reprogram the CIED.^9,18

ICDs need to be programmed to asynchronous pacing in pacemaker-dependent patients. ICDs will misinterpret EMI as a tachyarrhythmia, thus triggering tachytherapies. Antitachyarrhythmia functions are disabled to avoid inappropriate therapies. Removal of a magnet rapidly restores tachytherapies if needed.⁸ Some, but not all, devices emit a tone with application of a magnet. Comments regarding ESU and dispersive electrodes noted previously for pacemakers apply to ICDs.

If external cardioversion or defibrillation is not an option during a procedure, transcutaneous pacing or defibrillator pads should be on the patient when CIED functions are disabled. When ICDs have been reprogrammed, the devices need to be interrogated and reprogrammed to their original or recommended settings before the patient is discharged to an unmonitored setting or to home.¹⁰ Patients with sudden death have been reported to registries after failure to reactivate tachytherapies.¹⁰ Patients with CIED can play a role in safe care and appropriate CIED functionality by sending a remote transmission upon discharge.

Leadless pacemakers are miniaturized, fully self-contained devices that are nonsurgically implanted in the RV via a catheter.^11,19 The pacemaker may or may not be magnet responsive and is manufacturer specific. Identification of the correct placement of a magnet can be difficult given that these devices are fully intracardiac.

Subcutaneous ICDs are more susceptible to ESU above the groin, likely due to their large surface area and location along the left chest wall.¹²

When compared with patients without prior stroke, an analysis from a large Danish registry found an increased risk of recurrent stroke, MI, and cardiovascular death in patients with recent stroke, particularly <3 months after the event (OR, 14.23), compared with 3 to 6 months (OR, 4.85), 6 to 11 months (OR, 3.04), or ≥12 months of the index event (OR, 2.47).² The odds of a postoperative ischemic stroke within the first 3 months of a prior stroke were high. However, over the ensuing months, this marked elevated risk waned and plateaued at 9 months but did not approach baseline risk, even at 12 months postinfarction (OR, 8.2). A more recent analysis of a Medicare database of ∼6 million patients undergoing elective noncardiac and non-neurological surgery found a far lower overall rate of secondary strokes compared with those with no prior stroke for operations performed <30 days or performed between 61 and 90 days after prior stroke (OR, 8.0 [95% CI, 6.37-10.10] and OR, 5.0 [95% CI, 4.0-6.29], respectively).¹ Importantly, there was only a very small decrement in risk between 61 and 90 days and 6 and 12 months (OR, 4.76 [95% CI, 4.26-5.26]). Based on this new evidence, delaying NCS for ≥6 months does not appear to provide additional protection against the occurrence of recurrent stroke.

A prospective multicenter study analyzed the preoperative sleep study results of 1218 patients without a prior diagnosis of OSA who were scheduled for major NCS.³ Among these patients, 67.6% were assigned a new diagnosis of previously recognized OSA, 30.5% had at least moderate OSA, and 11.2% had severe OSA. For those patients with severe OSA, there was an increased risk of the composite outcome of myocardial injury, cardiac death, HF, thromboembolism, AF, and stroke within 30 days of surgery.³ In a prospective study of 2877 adults who completed OSA risk screening questionnaires during preoperative assessment before NCS,¹¹ 23.7% were identified as high risk for OSA. Few patients identified as high risk had a prior OSA diagnosis, and among 207 who elected to undergo a home sleep study, 170 (82.1%) were found to have OSA (mild, n=97; moderate, n=40; severe, n=33).¹¹ Another prospective study enrolled 245 adults with ≥2 risk factors for OSA scheduled for NCS. All patients were screened with the STOP-Bang questionnaire, and among them 182 patients underwent a preoperative level III polysomnogram.^1,12 Seventy patients (38%) were diagnosed with OSA, including 11 patients (6%) with moderate to severe OSA. Although prospective evidence to support routine OSA screening before surgery is lacking, the increased prevalence of OSA in surgical patients with CVD and potential for benefit with OSA treatment provide a reasonable rationale for OSA screening before NCS.

Among patients receiving statin therapy who are planned for NCS, large cohorts report safety and possible reductions in cardiovascular complications associated with continued use of lipid lowering throughout the perioperative period.¹ In a large U.S. cohort (n=780,591), 9.9% of patients receiving perioperative lipid-lowering therapy had lower surgical mortality overall and after propensity matching than those who did not receive lipid-lowering therapy.¹ Although these data favor continuation of home-dose statins, the benefits of statin reloading or intensification before surgery are uncertain. One study randomly assigned 500 consecutive patients who were receiving statins and were admitted for urgent or emergent NCS to either reloading of atorvastatin or continuation of home statin.¹¹ In this study, statin reloading was associated with a 5.6% absolute risk reduction in perioperative MACE at 30 days, as well as reductions in AF and length of stay. Additional RCTs are needed to determine the benefits of statin intensification, reloading, or both in the perioperative period of NCS.

In a single-center RCT, 100 patients planned to undergo vascular surgery were randomly assigned to atorvastatin 20 mg daily for 45 days or placebo. Vascular surgery was performed an average of 30 days after statin initiation, and statins were continued long term in patients in whom LDL-cholesterol was ≥100 mg/dL. The primary endpoint, the composite of death from cardiac causes, nonfatal MI, stroke, and unstable angina at 6 months, occurred in 4 patients (8.0%) assigned to atorvastatin and 13 patients (26%) assigned to placebo (P=0.031).⁴ Furthermore, in a 2018 systematic review and meta-analysis, statin therapy appeared to reduce perioperative MACE and improve survival after vascular surgery.⁵ Long-term use of statin therapy is important to reduce primary and secondary atherosclerotic events in at-risk patients.¹² Evaluation practices outlined in the AHA/ACC guidelines may help to identify patients whose long-term cardiac outcomes would be improved by adherence to GDMT. However, given the lack of benefit in the LOAD trial, short-term use to lower perioperative risk requires further evaluation. As such, NCS should not be delayed if LDL-cholesterol is not already available.¹⁰

Intraoperative hypotension, particularly prolonged episodes, may increase postoperative myocardial injury and mortality.⁸ However, while omitting RAASi before surgery has been shown to reduce intraoperative hypotension, RCTs have failed to prove this strategy improves clinical outcomes.^1,2,9 The PREOP-ACEI (Prospective Randomized Evaluation of Perioperative Angiotensin-Converting Enzyme Inhibition) study for patients stable on ACEi for at least 6 weeks before planned major noncardiac nonvascular surgery observed that fewer episodes of intraoperative hypotension (SBP <80 mm Hg) occurred in patients randomized to omit the final preoperative ACEi dose compared with patients continuing use. Major adverse cardiovascular endpoints were not different, nor was postanesthesia unit recovery time or length of stay.¹ These findings are corroborated by the meta-analysis of earlier studies that included data from 6022 patients in whom ACEi or ARB was omitted or continued before planned NCS.² Although intraoperative hypotension was more common in patients who continued ACEi/ARB, there was no difference in MACE. Ongoing studies, including STOPorNOT (Impact of RAASi Continuation on Outcome after Major Surgery), will seek to answer questions about the clinical impact of continuation or omission of RAASi before planned surgery.^10,11 Recently published results from POISE-3 demonstrated no difference in major vascular events (MINS, vascular death, stroke, or cardiac arrest at 30 days) in nearly 7500 patients with vascular disease or risk factors randomized to hypotension-avoidance or hypertension-avoidance perioperative BP strategies.¹² In this study, antihypertensive agents were either withheld or continued based on the BP management strategy. Notably, 72% of patients in both groups were taking an ACEi/ARB at the time of randomization, with no excess adverse events in the group randomly assigned to continue their home BP regimen on the morning of surgery.¹²

The role of RAASi in HFrEF, HF with mildly reduced ejection fraction, and HF with preserved ejection fraction is supported by clinical guidelines.⁷ RAASi confer a significant reduction in mortality and even dose-dependent reduction in outcomes such as hospitalizations for HF, particularly in patients with LVEF <40%.⁷ The complexity of HF medication regimens and goals of care warrant careful consideration and ideally, minimal interruption. In our review of existing data, RCTs of perioperative omission of RAASi have largely excluded patients with moderate to severe HF or had limited inclusion.¹ Further, a meta-analysis in 2018 found only 1 study reporting the influence of RAASi interruption on HF outcomes.²

The POISE-2 trial was a blinded RCT of 10,010 patients undergoing major NCS, randomized to low-dose clonidine or placebo 2 to 4 hours before surgery, to evaluate the impact on 30-day risk of death or nonfatal MI and included patients were ≥45 years old and at high risk of cardiovascular complications (history of ASCVD event or ≥3 traditional risk factors).¹ If baseline SBP was ≥105 mm Hg and heart rate was ≥55 bpm, patients received 0.2 mg of clonidine orally and application of a transdermal 0.2 mg per day clonidine patch or placebo for 72 hours. At 30 days, no difference was seen in the primary outcome, composite of death or nonfatal MI, or key secondary endpoints. Patients treated with clonidine were significantly more likely to experience nonfatal cardiac arrest and clinically significant bradycardia or hypotension. Results of POISE-2, as well as 23 additional RCTs in NCS, were included in a 2018 meta-analysis that reported nonsignificant differences in all-cause mortality, cardiac mortality, and MI.³ There was, however, moderate-to-high quality evidence to demonstrate significant increased risk of hypotension and bradycardia. Importantly, in addition to clonidine, these studies included dexmedetomidine and mivaserol, which were never approved in the United States.³

The decision to perform NCS in a patient with CAD should involve the patient, the surgeon, the anesthesiologist, and the cardiologist managing the patient’s antiplatelet therapy and ischemic risk.^5-7,44 Antiplatelet therapy is frequently indicated for the prevention of ischemic cardiac events in patients with CAD. Temporary discontinuation of antiplatelet therapy can be safe depending on the timing of NCS relative to any prior PCI and time-independent indications for the procedure. The risk of antiplatelet therapy interruption should be individualized to balance MACE risks with optimal surgical timing.⁷ Decisions for continuation or cessation of aspirin, P2Y12 inhibitor, or both in the perioperative period of NCS should be undertaken by multidisciplinary team members and should involve the patient and their family.⁴⁴ The use of aspirin for primary prevention has decreased based on evidence that it does not reduce cardiovascular risks, and multidisciplinary team consensus may not be required for perioperative interruption of aspirin when it is prescribed for primary cardiovascular prevention.⁴⁵ Multidisciplinary input may be considered in those receiving aspirin or other antiplatelet therapy for secondary prevention or noncardiovascular indications.

Elective NCS after balloon angioplasty alone without placement of a stent can proceed after ≥14 days of uninterrupted DAPT; however, high-quality data to inform management of these patients are limited.^8,9

In patients with prior DES-PCI, the optimal timing of elective NCS requiring interruption of antiplatelet therapy requires careful consideration. Multiple studies have identified a continuum of declining perioperative MACE risk after PCI.¹² However, a large retrospective analysis of >20,000 patients corroborated smaller studies identifying that prior PCI remains a risk factor for perioperative MACE and bleeding to 1 year.^14,15 A small retrospective cohort identified that perioperative cardiac events still occur 6 months after DES-PCI.¹³ Perioperative risks of NCS are particularly high within the first year after PCI when coronary stent placement occurs for the treatment of acute MI.^11,16,17 In patients with DES-PCI performed for ACS, elective NCS should be delayed ≥12 months after PCI. A 12-month delay between PCI and NCS may also be appropriate for patients undergoing complex DES-PCI (eg, bifurcation stents, long stent lengths, multivessel PCI) or when details regarding prior DES-PCI are unavailable.¹⁷

In a matched cohort study from U.S. Veterans Affairs hospitals, perioperative MACE events were highest during the first 6 months after PCI and stabilized thereafter at 1%.¹⁶ A retrospective analysis of 221,379 hospital admissions for NCS found high rates of perioperative MI (4.7%), bleeding (32%), and mortality (4.4%) within 6 months of PCI.⁵ Stent type (BMS versus DES) was not associated with perioperative MACE at 6 months in a large cohort study.¹² Similarly, another registry showed no significant differences in perioperative MACE between those patients with BMS or DES, only identifying PCI for ACS as an independent MACE risk factor.¹¹ However, a multicenter prospective registry (approximately 40,000 patients) identified stent type as a risk factor for MACE events, with older-generation DES associated with higher risk of events at any time point compared with BMS.²¹ In a Canadian study (approximately 8000 patients), the incidence of MACE associated with major elective NCS performed >6 months after DES-PCI was 1.2%, with risk approaching that of intermediate-risk surgical patients without prior PCI.⁶ The variation seen in these analyses emphasizes that stent type, time from PCI, and indication for PCI represent important factors in shared decision-making regarding consideration of surgery at 6 months post-PCI.

In patients with prior PCI and a time-sensitive indication for NCS (eg, resection of malignancy), a risk assessment balancing the potential delay of surgery against perioperative MACE should be performed. Risks of perioperative MACE after NCS appear highest when surgery is performed within the first 3 months after PCI.¹⁷ A previous study found lower incidence of MACE when NCS was performed >3 months (2.8%) compared with <30 days (10.5%) after BMS-PCI.⁸ The incidence of perioperative MACE after NCS was also lowest if surgery was performed >3 months after DES-PCI.⁴⁶ In a prospective study evaluating perioperative MACE and bleeding, event rates were high in the first 30 days after PCI, with time from stent implantation to surgery <3 month as an independent risk factor for bleeding.²⁴ Finally, in a large pooled analysis of nonsurgical patients post-PCI, DAPT discontinuation at >3 months was not associated with excess stent thrombosis.⁴⁷ Taken together, these data suggest that, in selected patients, it may be reasonable to undergo NCS ≥3 months post-PCI if the benefit of surgery outweighs the risk of MACE.

Elective NCS should not be performed <30 days of PCI. Early case series reported a high incidence of bleeding and MACE after NCS scheduled within 30 days of PCI.⁴⁸ Larger studies subsequently confirmed excess risk of perioperative MACE within this timeframe post-PCI.^5,8,9,23,49 Catastrophic outcomes of NCS within 30 days of PCI have been reported, with risks of MI, stent thrombosis, bleeding, and mortality.^8,48,49 Surgical trauma leads to catecholamine surges, inflammatory cytokines, activation of the clotting cascade, enhanced platelet activation, and decreased fibrinolysis, all of which contribute to the thrombotic milieu, and consequently surgery should be delayed after PCI.²² A prospective study of patients post-PCI found that those undergoing NCS at <35 days from PCI had a 2-fold higher risk of complications compared with those undergoing surgery >90 days after PCI.²⁶ Another study found that NCS <1 month post-PCI had a higher incidence of MI (7.2% versus 0.5%), cardiac death (5% versus 0.4%), and all-cause mortality (9% versus 2.1%) than those undergoing surgery within the first 12 months after PCI.²³ Elective NCS after PCI with BMS can proceed after ≥30 days of uninterrupted DAPT, although BMS are rarely placed in the contemporary era.^8,9

In patients with prior PCI with coronary stent placement undergoing NCS, aspirin use was associated with lower rates of death and nonfatal MI (absolute risk reduction, 5.5%), with comparable major and life-threatening bleeding.²⁹ Although the risks of bleeding with DAPT are higher than those with aspirin alone,^25,48,50-53 in a meta-analysis of 46 studies including >30,000 patients undergoing NCS, both single and double antiplatelet therapy were associated with a modest increased risk of bleeding without additional thrombotic risk compared with placebo or interruption of antiplatelet therapy.⁵⁴ A systematic review found limited data to support either continuation or discontinuation of DAPT before NCS as a strategy to reduce thrombotic risk, bleeding, or mortality.³⁵ In patients planned for NCS that requires interruption of DAPT, aspirin monotherapy should be continued whenever possible.

The risk of stent thrombosis is highest in the first 4 to 6 weeks after PCI with BMS and in the first 3 months after DES implantation.^{6,9-12,25,48,52,55-58} In a Danish national registry from 2005 to 2012, NCS within 1 month of PCI-DES was associated with a 13-fold higher risk of cardiac death and a 4-fold higher risk of all-cause mortality.²³ When time-sensitive NCS is necessary within 30 days of BMS-PCI or within 3 months of DES-PCI, DAPT should be continued in the perioperative period, if feasible, from a surgical bleeding perspective.

In patients with prior PCI in whom oral anticoagulation monotherapy is planned to be interrupted before NCS, initiation of aspirin monotherapy is reasonable to reduce risks of stent thrombosis and ischemic complications. After NCS, aspirin may be discontinued, and oral anticoagulation monotherapy may be reinitiated as surgical bleeding risks permit.

In patients who are within 1 to 6 months of PCI and continue to need DAPT, use of intravenous antiplatelet therapy as a bridge for nondeferrable surgery has been inadequately studied.⁴ The BRIDGE (Bridging Antiplatelet Therapy With Cangrelor in Patients Undergoing Cardiac Surgery) trial studied oral P2Y12 inhibitor discontinuation and subsequent use of cangrelor versus placebo.⁵⁹ This study of patients undergoing CABG demonstrated greater platelet inhibition with cangrelor without excessive risk of major bleeding. The MONET BRIDGE (Maintenance of Antiplatelet Therapy in Patients with Coronary Stenting Undergoing Surgery) trial is currently underway for evaluating cangrelor as a bridging strategy in patients undergoing NCS within 12 months of PCI.⁶⁰ There are no established data on the use of glycoprotein IIB/IIIA inhibitors as a bridging strategy.^61,62

In observational studies of patients undergoing NCS, aspirin continuation was associated with a 1.5-fold greater risk of nonserious bleeding events.³⁴ Withdrawal of aspirin preceded up to 10% of perioperative acute cardiovascular syndromes.³⁴ In a small RCT of 220 patients undergoing NCS, perioperative aspirin was associated with a 7.2% absolute risk reduction for postoperative MACE.⁶³ In a meta-analysis of >30,000 patients with and without prior PCI undergoing NCS, antiplatelet therapy was associated with minimal bleeding risk and no increase in thrombotic complications.⁵⁴ In the prespecified stratum of POISE-2 (n=4382) trial, aspirin continuation did not reduce death or nonfatal MI compared with aspirin interruption (7.7% versus 7.8%; HR, 1.00 [95% CI, 0.81-1.23]).²⁷ Although aspirin should not be routinely continued in the perioperative period of NCS, continuation may be reasonable in selected patients after consideration of individualized thrombotic and bleeding risks.³⁵

The POISE-2 trial randomly assigned 10,010 patients planned for NCS and at risk for cardiovascular complications to perioperative aspirin versus placebo.²⁷ Administration of aspirin before surgery and for 30 days postoperatively did not reduce the composite of death or nonfatal MI (7.0% versus 7.1%; HR, 0.99 [95% CI, 0.86-1.15]; P=0.92) but was associated with a 23% increased hazard for major bleeding. Findings were consistent regardless of aspirin use before trial enrollment.²⁷ Among patients in POISE-2 undergoing vascular surgery, perioperative withdrawal of chronic aspirin therapy was not associated with increased cardiovascular events.⁶⁴

It is generally safe to perform surgeries with minimal bleeding risk without interrupting OAC therapy (Tables 13 and 14).^1-3,15-17 For NCS with greater bleeding risks, time-based interruption (“time reversal”) of OAC is advised.^5,18-20 A DOAC interruption protocol tested in 3007 patients with AF in the PAUSE (Perioperative Anticoagulation Use for Surgery Evaluation) study⁷ resulted in low rates of major bleeding or thromboembolism (Table 13). Findings were similar in the prospective, observational, EMIT-AF/VTE (Edoxaban Management in Diagnostic and Therapeutic Procedures) study.⁶

In select patients with recent thromboses and high residual thrombotic risk, delaying elective NCS may permit safer interruption of OAC. Time reversal of OAC is always preferred, but this may not be feasible for urgent or emergency procedures with moderate or high bleeding risk.²¹ The measurement of coagulation parameters, drug levels, or both may identify ongoing drug effects. In the absence of altered coagulation parameters or detectable drug levels, OAC reversal agents may not be necessary.^22-24 Otherwise, rapid reversal of OAC can be achieved with prothrombin complex concentrates,^25,26 andexanet alfa for factor Xa inhibitors (rivaroxaban, apixaban, or edoxaban), or idarucizumab for dabigatran (Table 15).^27,28

Procedures with higher bleeding risks (eg, neuraxial anesthesia) should be performed with complete interruption of OAC.²⁹ When minimal drug effect is desired, anticoagulants should be held for ≥5 half-lives (Table 13), ≥3 days for factor Xa inhibitors (rivaroxaban, apixaban, edoxaban), and ≥4 days for dabigatran (5-6 days, if creatinine clearance <50 mL/min).³⁰

In high thrombotic risk patients who are receiving VKA (Table 13), bridging with parenteral anticoagulation is a common practice; however, data supporting efficacy (ie, prevention of thromboembolism) or safety (ie, bleeding) are not available. A meta-analysis of patients with venous thromboembolism reported the incidence of recurrent thromboembolism to be low, regardless of perioperative management strategy or baseline thromboembolic risk, and that bridging increased the incidence of bleeding.⁸ A systematic review and meta-analysis of NCS perioperative bridging in patients with mechanical heart valves (35.8% of patients with mechanical mitral valves) identified 15 studies with 2453 bridging episodes and found that bridging increased overall bleeding, with near significant differences in major bleeding, and without the benefit of lowering thromboembolism; however, both noted the majority of results were based on poor-quality cohorts overall.¹⁰ The PERIOP-2 (Postoperative Low Molecular Weight Heparin Bridging Treatment for Patients at High Risk of Arterial Thromboembolism) study enrolled 1471 patients on VKA requiring NCS (79% AF, 14% mechanical heart valve, 7% with both).⁹ All patients had warfarin held 5 days before NCS and received low-molecular-weight heparin for 3 days preoperatively followed by postoperative randomization to placebo or low-molecular-weight heparin at either a prophylactic or full dose based on procedural bleeding risk. Thromboembolism was similar across all patient populations, and secondary outcomes of bleeding were increased with postoperative bridging. Although further RCTs are warranted, available data support limiting the use of bridging to very high thrombotic risk (eg, mechanical mitral valves) patients on VKA, with careful consideration of bleeding risk (eg, HAS-BLED,³¹ previous personal bleeding history, and perioperative bleeding risk) to determine an individualized strategy.

The BRIDGE trial enrolled 1844 patients with AF on VKA and randomized them to low-molecular-weight heparin bridge therapy or placebo starting 3 days preoperatively and continued 5 days postoperatively. All patients had warfarin held from preoperative day 5 through postoperative day 1. Bridging anticoagulation was noninferior to placebo for prevention of thromboembolism, but bridging increased the risk of major bleeding.¹¹ Subsequent analysis showed bridge therapy, along with history of renal disease and procedures with high bleeding risk, to be a baseline predictor of major bleeding.³² The pharmacokinetic properties of DOAC therapy allow for a more rapid onset and offset of anticoagulant effects and thus shorter perioperative interruptions. A simple DOAC interruption strategy without heparin bridging results in minimal time off DOAC, low bleeding, and low thromboembolic risk in patients with AF.⁷ Data from phase 3 AF studies support these findings. Some patients who required procedures in these studies (approximately 11% of DOAC-treated patients) received unfractionated heparin or low- molecular-weight heparin bridging, which resulted in increased risk of bleeding without reduction in thromboembolic risk.^18-20 Based on these studies, most patients with AF will not benefit from bridging anticoagulation.

When restarting VKA after interruption, it can take several days to achieve full anticoagulant effect. Therefore, once hemostasis is achieved after a low or moderate bleeding risk procedure, it is reasonable to restart VKA as early as 12 to 24 hours postoperatively. It is suggested to resume VKA at the previous therapeutic dose, bearing in mind that additional monitoring may be required because concomitant medications (eg, antibiotics, nonsteroidal anti-inflammatory drugs, acetaminophen), nutrition, and drug clearance may be altered during the perioperative period. After initiation of DOAC, in most patients, peak levels and a therapeutic anticoagulation effect are achieved in ∼2 to 3 hours. Therefore, DOACs should be resumed when full anticoagulation is clinically appropriate, which may be as early as 6 hours postoperatively if hemostasis has occurred. In the PAUSE trial, DOACs were resumed 48 to 72 hours after high bleeding risk procedures, with overall low bleeding and thrombotic events.^7,33

Early studies of beta-blockade suggesting benefit may have produced their results by withdrawing beta blockers in patients who had been on them long term.⁶ Acutely discontinuing beta blockers for long-term indications is harmful^1,2,7 and should be avoided. Clinical judgment should be used to titrate beta blockers as appropriate during the perioperative period, with a focus on ensuring the medication is continued through the hospital stay and at discharge unless clear contraindications arise.

Results from POISE indicate that initiating beta blockers on the day of surgery is harmful, particularly if the medication is started at higher doses.⁸ Contemporaneous studies focused on initiating beta-blockade weeks in advance and titrated to physiological effect showed clinical benefit, but these results have since been called into question.⁹ A large observational study of beta-blocker initiation before NCS suggested higher risks of death if beta blockers were initiated <7 days before surgery³ compared with patients who initiated beta blockers >31 days earlier. Patients who initiated beta blockers between 8 and 30 days did not have excess mortality, nor did they show any perioperative benefits. Thus, if a patient requires initiation of beta blockers before surgery, the medication should be initiated ≥7 days before the surgery.

Initial evidence for perioperative beta blockers was derived from several relatively underpowered RCTs and a large observational study suggesting benefit from perioperative beta blockers. These were followed by the large multicenter POISE study,⁸ which demonstrated mixed benefits (eg, moderate reduction in MACE) balanced by harms (eg, hypotension, stroke, with a net increase in all-cause mortality) when high-dose beta blockers were administered to beta blocker–naïve patients immediately before surgery and maintained throughout the perioperative period.^9,10 Similar to the systematic review for the “2014 ACC/AHA Guideline on Perioperative Cardiovascular Evaluation and Management of Patients Undergoing Noncardiac Surgery,”¹¹ this guideline excludes a group of perioperative beta-blocker trials from consideration because of concern for unreliable data leading to spurious results. Evidence from 3 meta-analyses was concordant regarding potential harms of perioperative beta-blocker administration; however, the results of each of these studies were heavily influenced by the large POISE-1 sample size.^4,11,12 Although large observational studies have suggested higher potential benefit of beta-blockade in patients at increased risk as defined by the RCRI, these results have not been replicated in clinical trials.^13,14

If not obtained within 3 months of NCS, it is reasonable to check the preoperative hemoglobin A1c before surgery.^5,15,19 Multiple studies have assessed hemoglobin A1c and surgical outcomes, but it remains controversial whether elevated levels are linked to poor postoperative outcomes or are just a marker of poor perioperative glucose control.^1-4 At this time, there is no evidence that deferring surgery to achieve better glycemic control improves cardiovascular outcomes. Although there are no validated hemoglobin A1c risk thresholds, it may be reasonable to postpone an elective surgery if hemoglobin A1c is higher than 8%.¹ Emergent or time-sensitive procedures should not be delayed to achieve a target hemoglobin A1c; instead, the focus should be on optimizing perioperative glucose control. A retrospective study of preoperative blood glucose levels in patients undergoing noncardiac and nonvascular surgery found glucose concentrations ≥200 mg/dL to be associated with a >2-fold higher all-cause mortality rate and a >4-fold cardiovascular mortality rate compared with patients with normal blood glucose levels.²⁰ This study also demonstrated that preoperative patients undergoing treatment for diabetes had lower all-cause and cardiovascular mortality.^20,21

SGLT2i, noninsulin glucose–lowering agents that facilitate glycemic control by inhibiting renal glucose reabsorption and thus promoting glycosuria, must be discontinued 3 to 4 days before surgery.⁵ A rare complication of these agents is euglycemic diabetic ketoacidosis, which is a serious postoperative complication defined as normoglycemia (blood glucose <250 mg/dL) in the presence of metabolic acidosis (pH <7.3), total decreased serum bicarbonate (<18 mEq/L), and elevated serum and urine ketones.^5-7,22 There are no clear guidelines regarding restarting SGLT2i after surgery. Ideally, they should not be recommended until the patient is clinically stable and has resumed a normal diet.

Prior recommendations to discontinue metformin in the perioperative period stemmed from the concern that lactic acidosis could be precipitated in the setting of physical stressors; however, more recent data suggest that metformin is not associated with lactic acidosis. A population-based cohort of >10,600 patients with type 2 diabetes identified 163 patients who had been hospitalized with lactic acidosis. When compared with sex- and age-matched controls, current use of metformin was not associated with a risk of lactic acidosis.⁹

There have been several studies demonstrating cardiovascular risk reduction in nonoperative patients taking metformin.^8,10-12 The UKPDS (United Kingdom Prospective Diabetes Study) showed that, in addition to lowering blood glucose, metformin reduced cardiovascular mortality in patients with obesity and type 2 diabetes. The prespecified subgroup analysis showed risk reductions of 32% for any diabetes-related endpoint, 42% for diabetes-related death, and 36% for all-cause mortality.²³ These findings were further supported by a 10-year follow-up that showed significant risk reduction persisted in the metformin group for any diabetes-related endpoint (21%), MI (33%), and death from any cause (27%).²⁴

Over the past few decades, several studies have indicated a possible myocardial protective benefit of volatile anesthetic agents over total intravenous anesthesia, most commonly propofol, in surgical patients undergoing cardiac surgery. In 2009, a meta-analysis including >6000 surgical patients undergoing NCS failed to demonstrate a difference in rates of MI among patients who received anesthesia with sevoflurane or propofol.¹⁰ Several subsequent studies, including 4 RCTs, failed to identify any significant advantage to inhaled versus intravenous anesthesia.^1-13,11 In 2011, 88 patients were randomized to sevoflurane versus propofol, with no difference in detectable cardiac troponin I elevation or in median peak release of cardiac troponin I.¹¹ In 2012, 385 surgical patients at cardiovascular risk undergoing major NCS were randomized to sevoflurane or propofol.² The incidence of myocardial ischemia, postoperative release of NT-proBNP, MACE, or delirium was not decreased with the use of sevoflurane. In 2013, 193 surgical patients scheduled for elective abdominal aortic surgery were randomized to either sevoflurane or propofol/remifentanil.¹ Again, sevoflurane did not decrease postoperative release of cardiac troponin T, postoperative complications, nonfatal coronary events, or mortality when compared with propofol and remifentanil. In 2017, another trial randomized 120 older patients with established CAD to sevoflurane versus propofol/remifentanil, with no difference observed in release of cardiac troponin T or BNP 8 or 24 hours after surgery.³

A large nationwide retrospective cohort study in Denmark was performed in surgical patients undergoing first-time open inguinal and infrainguinal arterial reconstruction procedures.⁴ A significant benefit in terms of reduction of mortality and cardiac morbidity (MI, HF, dysrhythmias) was observed with regional anesthesia (including neuraxial and peripheral anesthesia) when compared with patients under general anesthesia.⁴ In contrast, another retrospective analysis examined a large dataset of patients undergoing above knee amputation.⁶ After propensity matching, there was no observed difference in either 30-day mortality or any secondary outcomes, including cardiac complications.⁶ In a recent randomized trial of 1600 surgical patients receiving spinal or general anesthesia for surgical management of hip fracture,⁵ no difference was reported between groups in primary outcome, a composite of death, or an inability to walk 10 feet independently or with a walker or cane approximately 60 days after surgery. Other outcomes were examined, including MI, nonfatal cardiac arrest, and stroke, but no significant difference between groups was detected.⁵ These results are in accordance with a previous meta-analysis in patients after hip fracture that reported no 30-day mortality benefit.¹² Evidence related to neuraxial anesthesia for postoperative pain control is discussed in Section 8.2 (“Perioperative Pain Management”).

TEA remains a cornerstone for anesthetic technique in major abdominal surgery to treat pain in the perioperative period. An updated 2016 Cochrane review in patients undergoing abdominal aortic surgery concluded there was a decreased incidence of MI associated with epidural analgesia compared with intravenous systemic opioid-based pain management.¹ A recent RCT of patients undergoing abdominal surgery (gastrectomy, Whipple, distal esophagectomy) comparing TEA (n=60) with intravenous analgesia (n=60) demonstrated a lower rate of postoperative myocardial injury (8.33% versus 36.67%) and supraventricular tachyarrhythmia (11.66% versus 36.67%) in the TEA cohort.² A large, retrospective NSQIP database analysis of >8000 patients demonstrated the benefit of epidural analgesia compared with intravenous analgesia in open abdominal surgery for preventing cardiopulmonary complications (OR, 0.58). However, there is a lack of high-quality evidence from RCTs comparing TEA with intravenous analgesia within enhanced recovery after NCS protocols, making it difficult to assess primarily for perioperative MI and other MACE after major abdominal surgery.

In patients with hip fractures, lumbar epidural analgesia for pain control appears to decrease the incidence of perioperative MACE. A systematic review and meta-analysis from 2022 that included 4 RCTs published between 2000 and 2014 (n=221 patients) supported the use of epidural analgesia for the prevention of perioperative MACE (to include combined events of cardiac death, MI, unstable angina, HF, or new-onset AF).³