Suparna Dutta, MD, MPH

Patients in the United States continue to experience rising out-of-pocket medical costs, with little access to the price information they desire when making decisions regarding medical care.¹ The Centers for Medicare & Medicaid Services (CMS) has taken steps toward transparency by requiring hospitals to publish price information.² In this issue of the Journal of Hospital Medicine, White and Liao³ break down the new rule, and we further discuss how this policy affects patients, hospitals, and hospitalists.

The new CMS rule requires hospitals to publish the prices of 300 “shoppable” services, including those negotiated with different payors. The rule standardizes how this information is displayed and accessed, with a daily penalty for facilities that fail to comply. Clinics and ambulatory surgical centers are currently excluded, as are facility and ancillary fees, such as those billed by pathology or anesthesiology. As White and Liao point out, a limitation for hospitalists is that this rule will only affect orders for the outpatient setting at discharge. In addition, this rule separates cost from quality. Although quality data are publicly available via CMS, price data are posted directly by hospitals, making a true value assessment difficult. To strengthen the rule, White and Liao recommend the following: increasing the financial penalty for noncompliance; aggregating data centrally to allow for comparisons; adding quality data to cost; expanding included sites and types of services; and adding common additional fees to the service price.

The larger question is whether patients will use these data in the manner intended. Previous studies have found a paradoxical relationship between patients’ expressed desire to compare prices for medical services vs documented low levels of price-shopping behavior. Mehrotra et al¹ found that lack of access to data as well as loyalty to providers were significant barriers to using price data effectively. The CMS rule increases access to the price information patients desire but cannot find. However, it is unclear whether available prices will be sufficient to change behaviors given that, aside from those with no insurance and those with high-deductible plans, most patients are fairly removed from the actual cost of service.

This rule may have a larger, unexpected impact on hospitals and access to care. Sharing price data could increase pressure on facilities to merge with larger systems in order to obtain more favorable rates via increased negotiating power. Hospitals that serve poorer communities may not be attractive merger candidates for large systems and could be left out of the push toward consolidation. Charging higher prices for the same services could lead to hospital closures or cuts in resources, potentially exacerbating health inequities for underserved populations.

On the provider end, it is unlikely that price transparency will influence resource utilization. Mummadi et al⁴ found that displaying price information in the electronic health record did not significantly influence physician ordering behavior. For hospitalists today, the emphasis on “high-value care” is already an important consideration when utilizing healthcare resources, considering the Accreditation Council for Graduate Medical Education (ACGME) requirements for residency, restrictive insurance protocols, and guidelines such as the ACR Appropriateness Criteria and the American Board of Internal Medicine’s Choosing Wisely^® campaign. Outside of extremes, separate cost data likely will not make a difference in provider ordering practices.

Although the information from this rule may not cause dramatic practice change, it will allow us to help our patients by providing those interested in price-shopping with data. This policy represents a large step toward a more transparent healthcare system, though it may have limited impact on overall healthcare costs.

Patients in the United States continue to experience rising out-of-pocket medical costs, with little access to the price information they desire when making decisions regarding medical care.¹ The Centers for Medicare & Medicaid Services (CMS) has taken steps toward transparency by requiring hospitals to publish price information.² In this issue of the Journal of Hospital Medicine, White and Liao³ break down the new rule, and we further discuss how this policy affects patients, hospitals, and hospitalists.

The new CMS rule requires hospitals to publish the prices of 300 “shoppable” services, including those negotiated with different payors. The rule standardizes how this information is displayed and accessed, with a daily penalty for facilities that fail to comply. Clinics and ambulatory surgical centers are currently excluded, as are facility and ancillary fees, such as those billed by pathology or anesthesiology. As White and Liao point out, a limitation for hospitalists is that this rule will only affect orders for the outpatient setting at discharge. In addition, this rule separates cost from quality. Although quality data are publicly available via CMS, price data are posted directly by hospitals, making a true value assessment difficult. To strengthen the rule, White and Liao recommend the following: increasing the financial penalty for noncompliance; aggregating data centrally to allow for comparisons; adding quality data to cost; expanding included sites and types of services; and adding common additional fees to the service price.

The larger question is whether patients will use these data in the manner intended. Previous studies have found a paradoxical relationship between patients’ expressed desire to compare prices for medical services vs documented low levels of price-shopping behavior. Mehrotra et al¹ found that lack of access to data as well as loyalty to providers were significant barriers to using price data effectively. The CMS rule increases access to the price information patients desire but cannot find. However, it is unclear whether available prices will be sufficient to change behaviors given that, aside from those with no insurance and those with high-deductible plans, most patients are fairly removed from the actual cost of service.

This rule may have a larger, unexpected impact on hospitals and access to care. Sharing price data could increase pressure on facilities to merge with larger systems in order to obtain more favorable rates via increased negotiating power. Hospitals that serve poorer communities may not be attractive merger candidates for large systems and could be left out of the push toward consolidation. Charging higher prices for the same services could lead to hospital closures or cuts in resources, potentially exacerbating health inequities for underserved populations.

On the provider end, it is unlikely that price transparency will influence resource utilization. Mummadi et al⁴ found that displaying price information in the electronic health record did not significantly influence physician ordering behavior. For hospitalists today, the emphasis on “high-value care” is already an important consideration when utilizing healthcare resources, considering the Accreditation Council for Graduate Medical Education (ACGME) requirements for residency, restrictive insurance protocols, and guidelines such as the ACR Appropriateness Criteria and the American Board of Internal Medicine’s Choosing Wisely^® campaign. Outside of extremes, separate cost data likely will not make a difference in provider ordering practices.

Although the information from this rule may not cause dramatic practice change, it will allow us to help our patients by providing those interested in price-shopping with data. This policy represents a large step toward a more transparent healthcare system, though it may have limited impact on overall healthcare costs.

Patients in the United States continue to experience rising out-of-pocket medical costs, with little access to the price information they desire when making decisions regarding medical care.¹ The Centers for Medicare & Medicaid Services (CMS) has taken steps toward transparency by requiring hospitals to publish price information.² In this issue of the Journal of Hospital Medicine, White and Liao³ break down the new rule, and we further discuss how this policy affects patients, hospitals, and hospitalists.

The new CMS rule requires hospitals to publish the prices of 300 “shoppable” services, including those negotiated with different payors. The rule standardizes how this information is displayed and accessed, with a daily penalty for facilities that fail to comply. Clinics and ambulatory surgical centers are currently excluded, as are facility and ancillary fees, such as those billed by pathology or anesthesiology. As White and Liao point out, a limitation for hospitalists is that this rule will only affect orders for the outpatient setting at discharge. In addition, this rule separates cost from quality. Although quality data are publicly available via CMS, price data are posted directly by hospitals, making a true value assessment difficult. To strengthen the rule, White and Liao recommend the following: increasing the financial penalty for noncompliance; aggregating data centrally to allow for comparisons; adding quality data to cost; expanding included sites and types of services; and adding common additional fees to the service price.

The larger question is whether patients will use these data in the manner intended. Previous studies have found a paradoxical relationship between patients’ expressed desire to compare prices for medical services vs documented low levels of price-shopping behavior. Mehrotra et al¹ found that lack of access to data as well as loyalty to providers were significant barriers to using price data effectively. The CMS rule increases access to the price information patients desire but cannot find. However, it is unclear whether available prices will be sufficient to change behaviors given that, aside from those with no insurance and those with high-deductible plans, most patients are fairly removed from the actual cost of service.

This rule may have a larger, unexpected impact on hospitals and access to care. Sharing price data could increase pressure on facilities to merge with larger systems in order to obtain more favorable rates via increased negotiating power. Hospitals that serve poorer communities may not be attractive merger candidates for large systems and could be left out of the push toward consolidation. Charging higher prices for the same services could lead to hospital closures or cuts in resources, potentially exacerbating health inequities for underserved populations.

On the provider end, it is unlikely that price transparency will influence resource utilization. Mummadi et al⁴ found that displaying price information in the electronic health record did not significantly influence physician ordering behavior. For hospitalists today, the emphasis on “high-value care” is already an important consideration when utilizing healthcare resources, considering the Accreditation Council for Graduate Medical Education (ACGME) requirements for residency, restrictive insurance protocols, and guidelines such as the ACR Appropriateness Criteria and the American Board of Internal Medicine’s Choosing Wisely^® campaign. Outside of extremes, separate cost data likely will not make a difference in provider ordering practices.

Although the information from this rule may not cause dramatic practice change, it will allow us to help our patients by providing those interested in price-shopping with data. This policy represents a large step toward a more transparent healthcare system, though it may have limited impact on overall healthcare costs.

Perioperative medicine is an evolving field with a rapidly growing body of literature. Because physicians and patients are often concerned about cardiac risk, we focus this review on perioperative cardiology.

The information we present here is derived from presentations at the Perioperative Medicine Summit and the annual meetings of the Society of Hospital Medicine and Society of General Internal Medicine in 2016. We surveyed perioperative literature from January 2015 through March 2016 and chose the final articles by consensus, based on relevance to clinicians who provide preoperative evaluations and postoperative care to surgical patients.

We have divided this review into four sections:

• Preoperative cardiac risk assessment
• Medical therapy to reduce postoperative cardiac complications (beta-blockers, statins, and angiotensin II receptor blockers [ARBs])
• Perioperative management of patients with a coronary stent on antiplatelet therapy
• Perioperative bridging anticoagulation.

PREOPERATIVE ASSESSMENT OF CARDIAC RISK

Functionally independent patients do better

Visnjevac O, Davari-Farid S, Lee J, et al. The effect of adding functional classification to ASA status for predicting 30-day mortality. Anesth Analg 2015; 121:110–116.

Functional capacity is an independent predictor of perioperative death and is included in the algorithm of the current joint American College of Cardiology/American Heart Association (ACC/AHA) guidelines,¹ but it is not in the Revised Cardiac Risk Index² or the American Society of Anesthesiologists (ASA) classification.³

The study. Visnjevac et al⁴ performed a retrospective, observational cohort study of 12,324 patients who underwent noncardiac surgery, stratifying rates of all-cause mortality and 30-day postoperative complications based on ASA class and functional capacity.

The ASA physical status classification is defined as:

• 1—Normal healthy patient
• 2—Patient with mild systemic disease
• 3—Patient with severe systemic disease
• 4—Patient with severe systemic disease that is a constant threat to life
• 5—Moribund patient not expected to survive without surgery.

Functional capacity was defined as the ability to perform all activities of daily living. It was prospectively assessed during the patient interview by pre-anesthesia personnel and entered into the database of the Veterans Affairs Surgical Quality Improvement Program.

Results. Within each ASA class, the mortality rate was significantly lower for functionally independent patients than for partially or fully dependent patients:

• In class 2—odds ratio (OR) 0.14 for functionally independent patients
• In class 3—OR 0.29 for functionally independent patients
• In class 4—OR 0.5 for functionally independent patients.

The mortality rate was higher for dependent patients than for independent patients who were one ASA class higher, despite the higher class having greater rates of comorbidity.

Adding functional capacity to the ASA classification improved the area under the receiver operating curve from 0.811 to 0.848 (a perfect test would have a value of 1.0), suggesting that physicians should incorporate functional capacity into their preoperative evaluation, perhaps by increasing a patient’s ASA class to the next higher class if he or she is functionally dependent.

Angina portends poor outcomes

Pandey A, Sood A, Sammon JD, et al. Effect of preoperative angina pectoris on cardiac outcomes in patients with previous myocardial infarction undergoing major noncardiac surgery (data from ACS-NSQIP). Am J Cardiol 2015; 115:1080–1084.

Coronary artery disease is a risk factor for adverse perioperative outcomes, but the risk varies depending on whether the patient has had a myocardial infarction (and how long ago) and whether he or she has anginal symptoms (and how severe they are).

The study. Pandey et al⁵ used data from the American College of Surgeons National Surgical Quality Improvement Program to evaluate the impact of stable angina in 1,568 patients who underwent noncardiac surgery after a myocardial infarction.

Results. Postoperative myocardial infarction or cardiac arrest occurred in 5.5% of patients. The incidence was significantly greater in those who had anginal symptoms before surgery than in those without symptoms (8.4% vs 5%, P = .035); reintervention rates and length of stay were also higher in this group. In multivariate analysis, preoperative angina remained a significant predictor of postoperative myocardial infarction (OR 2.49, 95% confidence interval [CI] 1.20–5.81) and re­intervention (OR 2.4, 95% CI 1.44–3.82.

The authors cautioned against relying on predictive tools such as the Revised Cardiac Risk Index that do not consider stable angina and previous myocardial infarction as separate independent risk factors.

Implications for clinical practice. While functional capacity is an integral part of the ACC/AHA guideline algorithm,¹ the findings of these two studies suggest that other current tools to calculate perioperative risk (ASA class and Revised Cardiac Risk Index) could be improved by including functional capacity and stable angina.

PERIOPERATIVE MEDICAL THERAPY

Beta-blockers help only those at high risk and may harm others

Friedell ML, Van Way CW 3rd, Freyberg RW, Almenoff PL. ß-blockade and operative mortality in noncardiac surgery: harmful or helpful? JAMA Surg 2015; 150:658–663.

Beta-blockers have been used perioperatively for nearly 2 decades to try to reduce rates of postoperative major adverse cardiovascular events. However, in view of recent trials, fewer patients are likely to benefit from this intervention than has been thought.

The study. Friedell et al⁶ retrospectively analyzed data from 343,645 patients in Veterans Affairs hospitals to determine the effect of beta-blockers on major adverse cardiac event rates after major noncardiac surgery. Beta-blockers were considered to have been used perioperatively if given any time between 8 hours before and 24 hours after surgery. The outcome studied was the mortality rate at 30 days.

The authors derived a novel risk score and used multivariate analysis to attempt to adjust for confounding factors. The risk score was based on four risk factors identified a priori:

• Serum creatinine level > 2.0 mg/dL
• Coronary artery disease
• Diabetes
• Surgery in a major body cavity (abdomen or chest).

Results. In this cohort, 43.2% of patients had received a beta-blocker. The unadjusted mortality rates by risk category for patients receiving or not receiving a beta-blocker were:

• No risk factors: 1.0% with a beta-blocker vs 0.6% without
• One or two risk factors: 1.7% vs 1.5%
• Three or four risk factors: 2.3% vs 4.5%.

After adjustment for confounding factors, the 30-day mortality rate was higher in low-risk patients and lower in high-risk patients who received beta-blockers. Odds ratios for death in beta-blocker users (entire cohort) by risk category were:

• No risk factors: 1.19
• One or two risk factors 0.97
• Three or four risk factors 0.76.

In the 3.8% of the total cohort who underwent cardiac surgery, beta-blockers had no significant effect—beneficial or harmful—in any risk group.

Jørgensen ME, Hlatky MA, Køber L, et al. ß-blocker-associated risks in patients with uncomplicated hypertension undergoing noncardiac surgery. JAMA Intern Med 2015; 175:1923–1931.

The study. Jørgensen et al⁷ investigated the association between chronic beta-blocker use for the treatment of hypertension and 30-day rates of mortality and major adverse cardiac events. Eligible patients (N = 55,320) were at least 20 years old and were undergoing any type of noncardiac surgery. The authors established that hypertension was present through use of an algorithm based on the International Classification of Diseases (10th edition). Patients with existing cardiovascular disease and renal disease were excluded. The authors used multivariate analysis to adjust for confounding factors.

Results. Twenty-six percent of the patients were on chronic beta-blocker therapy for hypertension. The mortality rate at 30 days was 1.93% in patients treated with a beta-blocker alone or in combination with other antihypertensive drugs; the rate was 1.32% for patients receiving any combination of renin-angiotensin system inhibitor, calcium antagonist, or thiazide, but no beta-blocker. Similarly, the 30-day major adverse cardiac event rates were 1.32% with beta-blockers and 0.84% without beta-blockers.

In subgroup analysis, each medication combination that included a beta-blocker was associated with higher rates of death and major adverse cardiac events than the same combination without a beta-blocker. Odds ratios for major adverse cardiac events with beta-blocker combinations ranged from 1.22 to 2.16 compared with regimens with no beta-blocker.

Implications for clinical practice. These two studies added to a growing chorus of concerns about the value and safety of beta-blockers in surgical patients. Friedell et al⁶ made an observation that was remarkably similar to one reported by Lindenauer et al⁸ in 2005: when patients were stratified by baseline risk of death, only those with the highest baseline risk benefited from beta-blocker therapy. Those in the lowest risk group actually were harmed by beta-blocker use, ie, the mortality rate was higher.

More interesting is the novel observation by Jørgensen et al⁷ that even in patients with no known cardiovascular disease who are on chronic beta-blocker therapy—presumably on stable doses and not solely for perioperative risk reduction—rates of mortality and major adverse cardiac events were higher than for patients not on chronic beta-blocker therapy.

The current studies support a cautious, selective approach to the perioperative use of beta-blockers—they should be used only in high-risk patients undergoing high-risk surgery, as has been proposed by the ACC/AHA.¹

Statins protect

Antoniou GA, Hajibandeh S, Hajibandeh S, Vallabhaneni SR, Brennan JA, Torella F. Meta-analysis of the effects of statins on perioperative outcomes in vascular and endovascular surgery. J Vasc Surg 2015; 61:519–532.

The study⁹ was a comprehensive meta-analysis of randomized controlled trials and observational studies of the effects of HMG-CoA reductase inhibitors (statins) on perioperative outcomes in patients undergoing vascular surgery (but not for intracranial or coronary artery disease). Twenty-four studies were included, 4 randomized controlled trials and 20 observational studies (including 16 cohort and 4 case-controlled studies), with a total of 22,536 patients, 8,052 receiving statins and 15,484 not receiving statins.

Results. Although there was no significant difference in cardiovascular mortality rates, patients receiving statins had significantly lower rates of all-cause mortality, myocardial infarction, stroke, and a composite of myocardial infarction, stroke, and death at 30 days postoperatively than patients not receiving statins. Additionally, there was no difference in the incidence of kidney injury between groups. The possibility of publication bias was thought to be low for all of these outcomes.

Berwanger O, Le Manach Y, Suzumura EA, et al. Association between pre-operative statin use and major cardiovascular complications among patients undergoing non-cardiac surgery: the VISION study. Eur Heart J 2016; 37:177–185.

The study. The Vascular Events in Non-cardiac Surgery Patients Cohort Evaluation (VISION) study^10,11 was an international prospective cohort study of more than 40,000 patients age 45 and older undergoing major noncardiac surgery with either general or regional anesthesia. Postoperative troponin measurements were obtained in all patients 6 to 12 hours after surgery and for the first 3 postoperative days. The authors evaluated the effect of preoperative statin use on cardiovascular outcomes at 30 days after surgery using a multivariate logistic model and propensity score analysis to correct for confounding factors. Statin use was defined as exposure within 7 days before surgery or 3 days after.

Results. In the 15,478 patients included in the analysis, statin use conferred a significant reduction in the primary outcome (composite of all-cause mortality, myocardial injury after noncardiac surgery, or stroke); the absolute risk reduction was 2.0%. Statin users also had a significantly lower risk of all-cause mortality, cardiovascular mortality, and myocardial injury after noncardiac surgery, but not of postoperative myocardial infarction or stroke. This analysis did not address the type of statin, dosing, or safety markers such as liver and muscle function.

Implications for clinical practice. With largely observational data and a few small randomized trials, these meta-analyses provide important information with respect to perioperative cardiovascular protection by statins. Starting a statin before surgery and continuing it perioperatively seems appropriate in patients at high risk (as recommended by the ACC/AHA guidelines¹). Based on other data, the benefit may be evident in as little as 5 days, as this is when statins appear to reach their plateau with regard to their vascular pleiotropic effects.¹² The incidence of adverse effects of statins, including muscle and liver injury, appears to be low in the perioperative setting.¹³

Given the inconsistent data regarding perioperative beta-blocker therapy, statins may very well be the most important perioperative medication with respect to cardiovascular risk reduction. However, a large randomized trial would help to confirm this belief.

Restart angiotensin II receptor blockers soon after surgery

Lee SM, Takemoto S, Wallace AW. Association between withholding angiotensin receptor blockers in the early postoperative period and 30-day mortality: a cohort study of the Veterans Affairs Healthcare System. Anesthesiology 2015; 123:288–306.

A concern about perioperative use of ARBs is that they impair the renin-angiotensin-aldosterone system, which maintains blood pressure under general anesthesia. ARB-induced intraoperative hypotension is particularly difficult to control, as it is often refractory to treatment with conventional adrenergic vasopressors.

The study. Lee et al¹⁴ conducted a retrospective cohort trial to evaluate the effects of continuing to withhold ARBs postoperatively. Of the 30,173 patients admitted for surgery in the Veterans Affairs system from 1999 through 2011 who were taking an ARB before surgery and who met the inclusion criteria, 10,205 (33.8%) were not restarted on their medication by postoperative day 2.

Results. The mortality rate at 30 days was higher in those whose ARBs were withheld than in those in whom it was resumed, with a multivariable-adjusted hazard ratio of 1.74 (95% CI 1.47–2.06; P < .001). The risk of withholding ARBs was more pronounced in younger patients (hazard ratio 2.52; 95% CI 1.69–3.76 in those under age 60) than in older patients (hazard ratio 1.42, 95% CI 1.09–1.85 in those over age 75).

Implications for clinical practice. While not addressing whether to continue or withhold ARBs preoperatively, this retrospective study presented evidence that delay in resuming chronic ARB therapy after surgery was common and appeared to be associated with a higher 30-day mortality rate. The ACC/AHA guidelines¹ state:

• Continuing angiotensin-converting enzyme (ACE) inhibitors or ARBs perioperatively is reasonable (class IIa recommendation, level of evidence B) (Table 1).
• If an ACE inhibitor or ARB is withheld before surgery, it is reasonable to restart it postoperatively as soon as clinically feasible (class IIa recommendation, level of evidence C).

Close attention to medication reconciliation in the postoperative period is necessary to facilitate early resumption of ARBs.

CORONARY STENTS AND ANTIPLATELET THERAPY IN NONCARDIAC SURGERY PATIENTS

Considerations in the management of noncardiac surgery patients with stents include risks of stent thrombosis, bleeding, and potentially delaying procedures to continue uninterrupted dual antiplatelet therapy. Evidence is evolving regarding the risks of perioperative complications in patients with bare-metal stents and drug-eluting stents, as well as the optimal timing before noncardiac surgery.

Bare-metal vs drug-eluting stents

Bangalore S, Silbaugh TS, Normand SL, Lovett AF, Welt FG, Resnic FS. Drug-eluting stents versus bare metal stents prior to noncardiac surgery. Catheter Cardiovasc Interv 2015; 85:533–541.

The study. Bangalore et al¹⁵ compared the safety of drug-eluting vs bare-metal stents in noncardiac surgery patients and investigated adverse events stratified by time since stent placement. This was a retrospective observational study of 8,415 patients in the Massachusetts claims database who underwent noncardiac surgery 1 year or less after percutaneous coronary intervention.

Results. There was no significant difference in the incidence of the primary outcome (composite of death, myocardial infarction, and bleeding) between the two groups.

With drug-eluting stents, patients had lower 30-day postoperative mortality rates, and their rate of the primary outcome decreased with time from percutaneous coronary intervention to surgery, being lowest beyond 90 days:

• 8.6% in days 1–30
• 7.5% in days 31–90
• 5.2% in days 91–180
• 5.8% in days 181–365 (P = .02).

With bare-metal stents, the event rate remained high over time:

• 8.2% in days 1–30
• 6.6% in days 31–90
• 8.1% in days 91–180
• 8.8% in days 181–365 (P = .60).

This study did not report information about perioperative antiplatelet management and was limited to first-generation drug-eluting stents. 

Saia F, Belotti LM, Guastaroba P, et al. Risk of adverse cardiac and bleeding events following cardiac and noncardiac surgery in patients with coronary stents: how important is the interplay between stent type and time from stenting to surgery? Circ Cardiovasc Qual Outcomes 2015; 9:39–47.

The study. Saia et al¹⁶ retrospectively examined predictors of periprocedural ischemic and bleeding events among cardiac and noncardiac surgical patients who had previously undergone percutaneous coronary intervention. They also assessed the risks associated with stent type and time from percutaneous coronary intervention to surgery.

Of 39,362 patients, 13,128 underwent procedures during the 5-year study period. The cumulative incidence of surgery was 3.6% at 30 days, 14% at 1 year, and 40% at 5 years after percutaneous coronary intervention. Almost 30% of the procedures were done urgently.

Results. The 30-day rate of postoperative cardiac death was 2.5%, nonfatal myocardial infarction 1.5%, and serious bleeding events 6.5%. Older drug-eluting stents were associated with higher risks of adverse events than newer drug-eluting stents at any time point (odds ratio 2.1 at 0–180 days, 1.9 at 6–12 months, and 1.45 after 12 months). Surgery performed 6 to 12 months after percutaneous coronary intervention had lower rates of adverse outcomes than surgery performed within 6 months. Beyond 6 months from percutaneous coronary intervention, bare-metal stents and newer drug-eluting stents did not have significantly different adverse event rates; however, newer drug-eluting stents appeared safer than bare-metal stents from 0 to 180 days.

Limitations of this study included lack of information regarding periprocedural antiplatelet management and a relatively small subset of newer drug-eluting stent patients.

Implications for clinical practice. These studies added to earlier work that demonstrated that the risk of perioperative adverse events differs by both the stent type and the time from percutaneous coronary intervention to noncardiac surgery. In patients with a drug-eluting stent, the risk levels off 90 days after percutaneous coronary intervention, suggesting that the previously recommended 12 months of uninterrupted dual antiplatelet therapy (per the 2014 ACC/AHA guidelines¹) may not be needed, particularly with newer-generation drug-eluting stents. Based on new evidence, the ACC/AHA guidelines regarding perioperative management of dual antiplatelet therapy in noncardiac surgery patients were updated,¹⁷ as noted below.

An update to the ACC/AHA guidelines on dual antiplatelet therapy

Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease. Circulation 2016 Mar 29. DOI: 10.1161/CIR.0000000000000404. [Epub ahead of print]

The 2016 update¹⁷ provides the following recommendations for patients with coronary stents who undergo noncardiac surgery:

• Delay elective surgery for 30 days after placement of a bare-metal stent (class I recommendation, level of evidence B).
• It is optimal to delay elective surgery 6 months after drug-eluting stent placement (class I recommendation, level of evidence B).
• If dual antiplatelet therapy must be discontinued, then continue aspirin if possible and restart the P2Y₁₂ inhibitor as soon as possible postoperatively (class I recommendation, level of evidence C ).
• A consensus decision among treating clinicians is useful regarding the risks of surgery and discontinuation or continuation of antiplatelet therapy (class IIa recommendation, level of evidence C).
• If dual antiplatelet therapy must be discontinued, then elective surgery should not be performed less than 30 days after bare-metal stent placement, or less than 3 months after drug-eluting stent placement (class III recommendation, level of evidence B).
• Elective surgery after drug-eluting stent placement when the P2Y₁₂ inhibitor must be discontinued may be considered 3 months after drug-eluting stent placement if the risk of surgical delay is greater than the risk of stent thrombosis (class IIb recommendation, level of evidence C).

The basic differences are the new recommendations for a minimum of 6 months of dual antiplatelet therapy as opposed to 12 months after drug-eluting stent placement before elective noncardiac surgery, and to allow surgery after 3 months (as opposed to 6 months) if the risk of delaying surgery outweighs the risk of stent thrombosis or myocardial infarction.

PERIOPERATIVE ANTICOAGULATION

The optimal perioperative management of patients with atrial fibrillation who are on warfarin is uncertain. The American College of Chest Physicians guidelines¹⁸ categorized patients with atrial fibrillation into low, moderate, and high thromboembolic risk. Based primarily on observational data, these guidelines recommended perioperative bridging anticoagulation for those at high risk but not for those at low risk. For intermediate-risk patients, there were insufficient data to make any recommendation.

Bridging may not benefit those at intermediate risk

Douketis JD, Spyropoulos AC, Kaatz S, et al; BRIDGE Investigators. Perioperative bridging anticoagulation in patients with atrial fibrillation. N Engl J Med 2015; 373:823–833.

The study. The Bridging Anticoagulation in Patients Who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery (BRIDGE) trial¹⁹ was the first randomized controlled trial to examine the effects of perioperative bridging anticoagulation in patients with atrial fibrillation without mechanical heart valves.

Results. In 1,884 patients undergoing elective surgery, the incidence of arterial thromboembolism was 0.4% in the no-bridging group and 0.3% in the bridging group (95% CI −0.6 to 0.8; P = .01 for noninferiority). Major bleeding occurred in 1.3% of patients in the no-bridging group and 3.2% in the bridging group (95% CI 0.20–0.78; P = .005 for superiority).

These results suggest that the risks of bridging therapy are greater than the benefits. Of note, the mean CHADS₂ score (1 point each for congestive heart failure, hypertension, age ≥ 75 years, and diabetes mellitus; 2 points for previous stroke or transient ischemic attack; a total score > 2 indicates significant risk of stroke) for patients enrolled in this trial was 2.3, and it may be difficult to extrapolate these results to the limited number of patients at highest risk, ie, who have a CHADS₂ score of 5 or 6. Also, this study did not address patients with arterial or venous thromboembolism.

Implications for clinical practice. Despite the limitations noted above, this study does provide guidance for management of the intermediate-risk group with atrial fibrillation as defined by the American College of Chest Physicians: a no-bridging strategy is the best option.

Perioperative medicine is an evolving field with a rapidly growing body of literature. Because physicians and patients are often concerned about cardiac risk, we focus this review on perioperative cardiology.

The information we present here is derived from presentations at the Perioperative Medicine Summit and the annual meetings of the Society of Hospital Medicine and Society of General Internal Medicine in 2016. We surveyed perioperative literature from January 2015 through March 2016 and chose the final articles by consensus, based on relevance to clinicians who provide preoperative evaluations and postoperative care to surgical patients.

We have divided this review into four sections:

• Preoperative cardiac risk assessment
• Medical therapy to reduce postoperative cardiac complications (beta-blockers, statins, and angiotensin II receptor blockers [ARBs])
• Perioperative management of patients with a coronary stent on antiplatelet therapy
• Perioperative bridging anticoagulation.

PREOPERATIVE ASSESSMENT OF CARDIAC RISK

Functionally independent patients do better

Visnjevac O, Davari-Farid S, Lee J, et al. The effect of adding functional classification to ASA status for predicting 30-day mortality. Anesth Analg 2015; 121:110–116.

Functional capacity is an independent predictor of perioperative death and is included in the algorithm of the current joint American College of Cardiology/American Heart Association (ACC/AHA) guidelines,¹ but it is not in the Revised Cardiac Risk Index² or the American Society of Anesthesiologists (ASA) classification.³

The study. Visnjevac et al⁴ performed a retrospective, observational cohort study of 12,324 patients who underwent noncardiac surgery, stratifying rates of all-cause mortality and 30-day postoperative complications based on ASA class and functional capacity.

The ASA physical status classification is defined as:

• 1—Normal healthy patient
• 2—Patient with mild systemic disease
• 3—Patient with severe systemic disease
• 4—Patient with severe systemic disease that is a constant threat to life
• 5—Moribund patient not expected to survive without surgery.

Functional capacity was defined as the ability to perform all activities of daily living. It was prospectively assessed during the patient interview by pre-anesthesia personnel and entered into the database of the Veterans Affairs Surgical Quality Improvement Program.

Results. Within each ASA class, the mortality rate was significantly lower for functionally independent patients than for partially or fully dependent patients:

• In class 2—odds ratio (OR) 0.14 for functionally independent patients
• In class 3—OR 0.29 for functionally independent patients
• In class 4—OR 0.5 for functionally independent patients.

The mortality rate was higher for dependent patients than for independent patients who were one ASA class higher, despite the higher class having greater rates of comorbidity.

Adding functional capacity to the ASA classification improved the area under the receiver operating curve from 0.811 to 0.848 (a perfect test would have a value of 1.0), suggesting that physicians should incorporate functional capacity into their preoperative evaluation, perhaps by increasing a patient’s ASA class to the next higher class if he or she is functionally dependent.

Angina portends poor outcomes

Pandey A, Sood A, Sammon JD, et al. Effect of preoperative angina pectoris on cardiac outcomes in patients with previous myocardial infarction undergoing major noncardiac surgery (data from ACS-NSQIP). Am J Cardiol 2015; 115:1080–1084.

Coronary artery disease is a risk factor for adverse perioperative outcomes, but the risk varies depending on whether the patient has had a myocardial infarction (and how long ago) and whether he or she has anginal symptoms (and how severe they are).

The study. Pandey et al⁵ used data from the American College of Surgeons National Surgical Quality Improvement Program to evaluate the impact of stable angina in 1,568 patients who underwent noncardiac surgery after a myocardial infarction.

Results. Postoperative myocardial infarction or cardiac arrest occurred in 5.5% of patients. The incidence was significantly greater in those who had anginal symptoms before surgery than in those without symptoms (8.4% vs 5%, P = .035); reintervention rates and length of stay were also higher in this group. In multivariate analysis, preoperative angina remained a significant predictor of postoperative myocardial infarction (OR 2.49, 95% confidence interval [CI] 1.20–5.81) and re­intervention (OR 2.4, 95% CI 1.44–3.82.

The authors cautioned against relying on predictive tools such as the Revised Cardiac Risk Index that do not consider stable angina and previous myocardial infarction as separate independent risk factors.

Implications for clinical practice. While functional capacity is an integral part of the ACC/AHA guideline algorithm,¹ the findings of these two studies suggest that other current tools to calculate perioperative risk (ASA class and Revised Cardiac Risk Index) could be improved by including functional capacity and stable angina.

PERIOPERATIVE MEDICAL THERAPY

Beta-blockers help only those at high risk and may harm others

Friedell ML, Van Way CW 3rd, Freyberg RW, Almenoff PL. ß-blockade and operative mortality in noncardiac surgery: harmful or helpful? JAMA Surg 2015; 150:658–663.

Beta-blockers have been used perioperatively for nearly 2 decades to try to reduce rates of postoperative major adverse cardiovascular events. However, in view of recent trials, fewer patients are likely to benefit from this intervention than has been thought.

The study. Friedell et al⁶ retrospectively analyzed data from 343,645 patients in Veterans Affairs hospitals to determine the effect of beta-blockers on major adverse cardiac event rates after major noncardiac surgery. Beta-blockers were considered to have been used perioperatively if given any time between 8 hours before and 24 hours after surgery. The outcome studied was the mortality rate at 30 days.

The authors derived a novel risk score and used multivariate analysis to attempt to adjust for confounding factors. The risk score was based on four risk factors identified a priori:

• Serum creatinine level > 2.0 mg/dL
• Coronary artery disease
• Diabetes
• Surgery in a major body cavity (abdomen or chest).

Results. In this cohort, 43.2% of patients had received a beta-blocker. The unadjusted mortality rates by risk category for patients receiving or not receiving a beta-blocker were:

• No risk factors: 1.0% with a beta-blocker vs 0.6% without
• One or two risk factors: 1.7% vs 1.5%
• Three or four risk factors: 2.3% vs 4.5%.

After adjustment for confounding factors, the 30-day mortality rate was higher in low-risk patients and lower in high-risk patients who received beta-blockers. Odds ratios for death in beta-blocker users (entire cohort) by risk category were:

• No risk factors: 1.19
• One or two risk factors 0.97
• Three or four risk factors 0.76.

In the 3.8% of the total cohort who underwent cardiac surgery, beta-blockers had no significant effect—beneficial or harmful—in any risk group.

Jørgensen ME, Hlatky MA, Køber L, et al. ß-blocker-associated risks in patients with uncomplicated hypertension undergoing noncardiac surgery. JAMA Intern Med 2015; 175:1923–1931.

The study. Jørgensen et al⁷ investigated the association between chronic beta-blocker use for the treatment of hypertension and 30-day rates of mortality and major adverse cardiac events. Eligible patients (N = 55,320) were at least 20 years old and were undergoing any type of noncardiac surgery. The authors established that hypertension was present through use of an algorithm based on the International Classification of Diseases (10th edition). Patients with existing cardiovascular disease and renal disease were excluded. The authors used multivariate analysis to adjust for confounding factors.

Results. Twenty-six percent of the patients were on chronic beta-blocker therapy for hypertension. The mortality rate at 30 days was 1.93% in patients treated with a beta-blocker alone or in combination with other antihypertensive drugs; the rate was 1.32% for patients receiving any combination of renin-angiotensin system inhibitor, calcium antagonist, or thiazide, but no beta-blocker. Similarly, the 30-day major adverse cardiac event rates were 1.32% with beta-blockers and 0.84% without beta-blockers.

In subgroup analysis, each medication combination that included a beta-blocker was associated with higher rates of death and major adverse cardiac events than the same combination without a beta-blocker. Odds ratios for major adverse cardiac events with beta-blocker combinations ranged from 1.22 to 2.16 compared with regimens with no beta-blocker.

Implications for clinical practice. These two studies added to a growing chorus of concerns about the value and safety of beta-blockers in surgical patients. Friedell et al⁶ made an observation that was remarkably similar to one reported by Lindenauer et al⁸ in 2005: when patients were stratified by baseline risk of death, only those with the highest baseline risk benefited from beta-blocker therapy. Those in the lowest risk group actually were harmed by beta-blocker use, ie, the mortality rate was higher.

More interesting is the novel observation by Jørgensen et al⁷ that even in patients with no known cardiovascular disease who are on chronic beta-blocker therapy—presumably on stable doses and not solely for perioperative risk reduction—rates of mortality and major adverse cardiac events were higher than for patients not on chronic beta-blocker therapy.

The current studies support a cautious, selective approach to the perioperative use of beta-blockers—they should be used only in high-risk patients undergoing high-risk surgery, as has been proposed by the ACC/AHA.¹

Statins protect

Antoniou GA, Hajibandeh S, Hajibandeh S, Vallabhaneni SR, Brennan JA, Torella F. Meta-analysis of the effects of statins on perioperative outcomes in vascular and endovascular surgery. J Vasc Surg 2015; 61:519–532.

The study⁹ was a comprehensive meta-analysis of randomized controlled trials and observational studies of the effects of HMG-CoA reductase inhibitors (statins) on perioperative outcomes in patients undergoing vascular surgery (but not for intracranial or coronary artery disease). Twenty-four studies were included, 4 randomized controlled trials and 20 observational studies (including 16 cohort and 4 case-controlled studies), with a total of 22,536 patients, 8,052 receiving statins and 15,484 not receiving statins.

Results. Although there was no significant difference in cardiovascular mortality rates, patients receiving statins had significantly lower rates of all-cause mortality, myocardial infarction, stroke, and a composite of myocardial infarction, stroke, and death at 30 days postoperatively than patients not receiving statins. Additionally, there was no difference in the incidence of kidney injury between groups. The possibility of publication bias was thought to be low for all of these outcomes.

Berwanger O, Le Manach Y, Suzumura EA, et al. Association between pre-operative statin use and major cardiovascular complications among patients undergoing non-cardiac surgery: the VISION study. Eur Heart J 2016; 37:177–185.

The study. The Vascular Events in Non-cardiac Surgery Patients Cohort Evaluation (VISION) study^10,11 was an international prospective cohort study of more than 40,000 patients age 45 and older undergoing major noncardiac surgery with either general or regional anesthesia. Postoperative troponin measurements were obtained in all patients 6 to 12 hours after surgery and for the first 3 postoperative days. The authors evaluated the effect of preoperative statin use on cardiovascular outcomes at 30 days after surgery using a multivariate logistic model and propensity score analysis to correct for confounding factors. Statin use was defined as exposure within 7 days before surgery or 3 days after.

Results. In the 15,478 patients included in the analysis, statin use conferred a significant reduction in the primary outcome (composite of all-cause mortality, myocardial injury after noncardiac surgery, or stroke); the absolute risk reduction was 2.0%. Statin users also had a significantly lower risk of all-cause mortality, cardiovascular mortality, and myocardial injury after noncardiac surgery, but not of postoperative myocardial infarction or stroke. This analysis did not address the type of statin, dosing, or safety markers such as liver and muscle function.

Implications for clinical practice. With largely observational data and a few small randomized trials, these meta-analyses provide important information with respect to perioperative cardiovascular protection by statins. Starting a statin before surgery and continuing it perioperatively seems appropriate in patients at high risk (as recommended by the ACC/AHA guidelines¹). Based on other data, the benefit may be evident in as little as 5 days, as this is when statins appear to reach their plateau with regard to their vascular pleiotropic effects.¹² The incidence of adverse effects of statins, including muscle and liver injury, appears to be low in the perioperative setting.¹³

Given the inconsistent data regarding perioperative beta-blocker therapy, statins may very well be the most important perioperative medication with respect to cardiovascular risk reduction. However, a large randomized trial would help to confirm this belief.

Restart angiotensin II receptor blockers soon after surgery

Lee SM, Takemoto S, Wallace AW. Association between withholding angiotensin receptor blockers in the early postoperative period and 30-day mortality: a cohort study of the Veterans Affairs Healthcare System. Anesthesiology 2015; 123:288–306.

A concern about perioperative use of ARBs is that they impair the renin-angiotensin-aldosterone system, which maintains blood pressure under general anesthesia. ARB-induced intraoperative hypotension is particularly difficult to control, as it is often refractory to treatment with conventional adrenergic vasopressors.

The study. Lee et al¹⁴ conducted a retrospective cohort trial to evaluate the effects of continuing to withhold ARBs postoperatively. Of the 30,173 patients admitted for surgery in the Veterans Affairs system from 1999 through 2011 who were taking an ARB before surgery and who met the inclusion criteria, 10,205 (33.8%) were not restarted on their medication by postoperative day 2.

Results. The mortality rate at 30 days was higher in those whose ARBs were withheld than in those in whom it was resumed, with a multivariable-adjusted hazard ratio of 1.74 (95% CI 1.47–2.06; P < .001). The risk of withholding ARBs was more pronounced in younger patients (hazard ratio 2.52; 95% CI 1.69–3.76 in those under age 60) than in older patients (hazard ratio 1.42, 95% CI 1.09–1.85 in those over age 75).

Implications for clinical practice. While not addressing whether to continue or withhold ARBs preoperatively, this retrospective study presented evidence that delay in resuming chronic ARB therapy after surgery was common and appeared to be associated with a higher 30-day mortality rate. The ACC/AHA guidelines¹ state:

• Continuing angiotensin-converting enzyme (ACE) inhibitors or ARBs perioperatively is reasonable (class IIa recommendation, level of evidence B) (Table 1).
• If an ACE inhibitor or ARB is withheld before surgery, it is reasonable to restart it postoperatively as soon as clinically feasible (class IIa recommendation, level of evidence C).

Close attention to medication reconciliation in the postoperative period is necessary to facilitate early resumption of ARBs.

CORONARY STENTS AND ANTIPLATELET THERAPY IN NONCARDIAC SURGERY PATIENTS

Considerations in the management of noncardiac surgery patients with stents include risks of stent thrombosis, bleeding, and potentially delaying procedures to continue uninterrupted dual antiplatelet therapy. Evidence is evolving regarding the risks of perioperative complications in patients with bare-metal stents and drug-eluting stents, as well as the optimal timing before noncardiac surgery.

Bare-metal vs drug-eluting stents

Bangalore S, Silbaugh TS, Normand SL, Lovett AF, Welt FG, Resnic FS. Drug-eluting stents versus bare metal stents prior to noncardiac surgery. Catheter Cardiovasc Interv 2015; 85:533–541.

The study. Bangalore et al¹⁵ compared the safety of drug-eluting vs bare-metal stents in noncardiac surgery patients and investigated adverse events stratified by time since stent placement. This was a retrospective observational study of 8,415 patients in the Massachusetts claims database who underwent noncardiac surgery 1 year or less after percutaneous coronary intervention.

Results. There was no significant difference in the incidence of the primary outcome (composite of death, myocardial infarction, and bleeding) between the two groups.

With drug-eluting stents, patients had lower 30-day postoperative mortality rates, and their rate of the primary outcome decreased with time from percutaneous coronary intervention to surgery, being lowest beyond 90 days:

• 8.6% in days 1–30
• 7.5% in days 31–90
• 5.2% in days 91–180
• 5.8% in days 181–365 (P = .02).

With bare-metal stents, the event rate remained high over time:

• 8.2% in days 1–30
• 6.6% in days 31–90
• 8.1% in days 91–180
• 8.8% in days 181–365 (P = .60).

This study did not report information about perioperative antiplatelet management and was limited to first-generation drug-eluting stents. 

Saia F, Belotti LM, Guastaroba P, et al. Risk of adverse cardiac and bleeding events following cardiac and noncardiac surgery in patients with coronary stents: how important is the interplay between stent type and time from stenting to surgery? Circ Cardiovasc Qual Outcomes 2015; 9:39–47.

The study. Saia et al¹⁶ retrospectively examined predictors of periprocedural ischemic and bleeding events among cardiac and noncardiac surgical patients who had previously undergone percutaneous coronary intervention. They also assessed the risks associated with stent type and time from percutaneous coronary intervention to surgery.

Of 39,362 patients, 13,128 underwent procedures during the 5-year study period. The cumulative incidence of surgery was 3.6% at 30 days, 14% at 1 year, and 40% at 5 years after percutaneous coronary intervention. Almost 30% of the procedures were done urgently.

Results. The 30-day rate of postoperative cardiac death was 2.5%, nonfatal myocardial infarction 1.5%, and serious bleeding events 6.5%. Older drug-eluting stents were associated with higher risks of adverse events than newer drug-eluting stents at any time point (odds ratio 2.1 at 0–180 days, 1.9 at 6–12 months, and 1.45 after 12 months). Surgery performed 6 to 12 months after percutaneous coronary intervention had lower rates of adverse outcomes than surgery performed within 6 months. Beyond 6 months from percutaneous coronary intervention, bare-metal stents and newer drug-eluting stents did not have significantly different adverse event rates; however, newer drug-eluting stents appeared safer than bare-metal stents from 0 to 180 days.

Limitations of this study included lack of information regarding periprocedural antiplatelet management and a relatively small subset of newer drug-eluting stent patients.

Implications for clinical practice. These studies added to earlier work that demonstrated that the risk of perioperative adverse events differs by both the stent type and the time from percutaneous coronary intervention to noncardiac surgery. In patients with a drug-eluting stent, the risk levels off 90 days after percutaneous coronary intervention, suggesting that the previously recommended 12 months of uninterrupted dual antiplatelet therapy (per the 2014 ACC/AHA guidelines¹) may not be needed, particularly with newer-generation drug-eluting stents. Based on new evidence, the ACC/AHA guidelines regarding perioperative management of dual antiplatelet therapy in noncardiac surgery patients were updated,¹⁷ as noted below.

An update to the ACC/AHA guidelines on dual antiplatelet therapy

Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease. Circulation 2016 Mar 29. DOI: 10.1161/CIR.0000000000000404. [Epub ahead of print]

The 2016 update¹⁷ provides the following recommendations for patients with coronary stents who undergo noncardiac surgery:

• Delay elective surgery for 30 days after placement of a bare-metal stent (class I recommendation, level of evidence B).
• It is optimal to delay elective surgery 6 months after drug-eluting stent placement (class I recommendation, level of evidence B).
• If dual antiplatelet therapy must be discontinued, then continue aspirin if possible and restart the P2Y₁₂ inhibitor as soon as possible postoperatively (class I recommendation, level of evidence C ).
• A consensus decision among treating clinicians is useful regarding the risks of surgery and discontinuation or continuation of antiplatelet therapy (class IIa recommendation, level of evidence C).
• If dual antiplatelet therapy must be discontinued, then elective surgery should not be performed less than 30 days after bare-metal stent placement, or less than 3 months after drug-eluting stent placement (class III recommendation, level of evidence B).
• Elective surgery after drug-eluting stent placement when the P2Y₁₂ inhibitor must be discontinued may be considered 3 months after drug-eluting stent placement if the risk of surgical delay is greater than the risk of stent thrombosis (class IIb recommendation, level of evidence C).

The basic differences are the new recommendations for a minimum of 6 months of dual antiplatelet therapy as opposed to 12 months after drug-eluting stent placement before elective noncardiac surgery, and to allow surgery after 3 months (as opposed to 6 months) if the risk of delaying surgery outweighs the risk of stent thrombosis or myocardial infarction.

PERIOPERATIVE ANTICOAGULATION

The optimal perioperative management of patients with atrial fibrillation who are on warfarin is uncertain. The American College of Chest Physicians guidelines¹⁸ categorized patients with atrial fibrillation into low, moderate, and high thromboembolic risk. Based primarily on observational data, these guidelines recommended perioperative bridging anticoagulation for those at high risk but not for those at low risk. For intermediate-risk patients, there were insufficient data to make any recommendation.

Bridging may not benefit those at intermediate risk

Douketis JD, Spyropoulos AC, Kaatz S, et al; BRIDGE Investigators. Perioperative bridging anticoagulation in patients with atrial fibrillation. N Engl J Med 2015; 373:823–833.

The study. The Bridging Anticoagulation in Patients Who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery (BRIDGE) trial¹⁹ was the first randomized controlled trial to examine the effects of perioperative bridging anticoagulation in patients with atrial fibrillation without mechanical heart valves.

Results. In 1,884 patients undergoing elective surgery, the incidence of arterial thromboembolism was 0.4% in the no-bridging group and 0.3% in the bridging group (95% CI −0.6 to 0.8; P = .01 for noninferiority). Major bleeding occurred in 1.3% of patients in the no-bridging group and 3.2% in the bridging group (95% CI 0.20–0.78; P = .005 for superiority).

These results suggest that the risks of bridging therapy are greater than the benefits. Of note, the mean CHADS₂ score (1 point each for congestive heart failure, hypertension, age ≥ 75 years, and diabetes mellitus; 2 points for previous stroke or transient ischemic attack; a total score > 2 indicates significant risk of stroke) for patients enrolled in this trial was 2.3, and it may be difficult to extrapolate these results to the limited number of patients at highest risk, ie, who have a CHADS₂ score of 5 or 6. Also, this study did not address patients with arterial or venous thromboembolism.

Implications for clinical practice. Despite the limitations noted above, this study does provide guidance for management of the intermediate-risk group with atrial fibrillation as defined by the American College of Chest Physicians: a no-bridging strategy is the best option.

Perioperative medicine is an evolving field with a rapidly growing body of literature. Because physicians and patients are often concerned about cardiac risk, we focus this review on perioperative cardiology.

The information we present here is derived from presentations at the Perioperative Medicine Summit and the annual meetings of the Society of Hospital Medicine and Society of General Internal Medicine in 2016. We surveyed perioperative literature from January 2015 through March 2016 and chose the final articles by consensus, based on relevance to clinicians who provide preoperative evaluations and postoperative care to surgical patients.

We have divided this review into four sections:

• Preoperative cardiac risk assessment
• Medical therapy to reduce postoperative cardiac complications (beta-blockers, statins, and angiotensin II receptor blockers [ARBs])
• Perioperative management of patients with a coronary stent on antiplatelet therapy
• Perioperative bridging anticoagulation.

PREOPERATIVE ASSESSMENT OF CARDIAC RISK

Functionally independent patients do better

Visnjevac O, Davari-Farid S, Lee J, et al. The effect of adding functional classification to ASA status for predicting 30-day mortality. Anesth Analg 2015; 121:110–116.

Functional capacity is an independent predictor of perioperative death and is included in the algorithm of the current joint American College of Cardiology/American Heart Association (ACC/AHA) guidelines,¹ but it is not in the Revised Cardiac Risk Index² or the American Society of Anesthesiologists (ASA) classification.³

The study. Visnjevac et al⁴ performed a retrospective, observational cohort study of 12,324 patients who underwent noncardiac surgery, stratifying rates of all-cause mortality and 30-day postoperative complications based on ASA class and functional capacity.

The ASA physical status classification is defined as:

• 1—Normal healthy patient
• 2—Patient with mild systemic disease
• 3—Patient with severe systemic disease
• 4—Patient with severe systemic disease that is a constant threat to life
• 5—Moribund patient not expected to survive without surgery.

Functional capacity was defined as the ability to perform all activities of daily living. It was prospectively assessed during the patient interview by pre-anesthesia personnel and entered into the database of the Veterans Affairs Surgical Quality Improvement Program.

Results. Within each ASA class, the mortality rate was significantly lower for functionally independent patients than for partially or fully dependent patients:

• In class 2—odds ratio (OR) 0.14 for functionally independent patients
• In class 3—OR 0.29 for functionally independent patients
• In class 4—OR 0.5 for functionally independent patients.

The mortality rate was higher for dependent patients than for independent patients who were one ASA class higher, despite the higher class having greater rates of comorbidity.

Adding functional capacity to the ASA classification improved the area under the receiver operating curve from 0.811 to 0.848 (a perfect test would have a value of 1.0), suggesting that physicians should incorporate functional capacity into their preoperative evaluation, perhaps by increasing a patient’s ASA class to the next higher class if he or she is functionally dependent.

Angina portends poor outcomes

Pandey A, Sood A, Sammon JD, et al. Effect of preoperative angina pectoris on cardiac outcomes in patients with previous myocardial infarction undergoing major noncardiac surgery (data from ACS-NSQIP). Am J Cardiol 2015; 115:1080–1084.

Coronary artery disease is a risk factor for adverse perioperative outcomes, but the risk varies depending on whether the patient has had a myocardial infarction (and how long ago) and whether he or she has anginal symptoms (and how severe they are).

The study. Pandey et al⁵ used data from the American College of Surgeons National Surgical Quality Improvement Program to evaluate the impact of stable angina in 1,568 patients who underwent noncardiac surgery after a myocardial infarction.

Results. Postoperative myocardial infarction or cardiac arrest occurred in 5.5% of patients. The incidence was significantly greater in those who had anginal symptoms before surgery than in those without symptoms (8.4% vs 5%, P = .035); reintervention rates and length of stay were also higher in this group. In multivariate analysis, preoperative angina remained a significant predictor of postoperative myocardial infarction (OR 2.49, 95% confidence interval [CI] 1.20–5.81) and re­intervention (OR 2.4, 95% CI 1.44–3.82.

The authors cautioned against relying on predictive tools such as the Revised Cardiac Risk Index that do not consider stable angina and previous myocardial infarction as separate independent risk factors.

Implications for clinical practice. While functional capacity is an integral part of the ACC/AHA guideline algorithm,¹ the findings of these two studies suggest that other current tools to calculate perioperative risk (ASA class and Revised Cardiac Risk Index) could be improved by including functional capacity and stable angina.

PERIOPERATIVE MEDICAL THERAPY

Beta-blockers help only those at high risk and may harm others

Friedell ML, Van Way CW 3rd, Freyberg RW, Almenoff PL. ß-blockade and operative mortality in noncardiac surgery: harmful or helpful? JAMA Surg 2015; 150:658–663.

Beta-blockers have been used perioperatively for nearly 2 decades to try to reduce rates of postoperative major adverse cardiovascular events. However, in view of recent trials, fewer patients are likely to benefit from this intervention than has been thought.

The study. Friedell et al⁶ retrospectively analyzed data from 343,645 patients in Veterans Affairs hospitals to determine the effect of beta-blockers on major adverse cardiac event rates after major noncardiac surgery. Beta-blockers were considered to have been used perioperatively if given any time between 8 hours before and 24 hours after surgery. The outcome studied was the mortality rate at 30 days.

The authors derived a novel risk score and used multivariate analysis to attempt to adjust for confounding factors. The risk score was based on four risk factors identified a priori:

• Serum creatinine level > 2.0 mg/dL
• Coronary artery disease
• Diabetes
• Surgery in a major body cavity (abdomen or chest).

Results. In this cohort, 43.2% of patients had received a beta-blocker. The unadjusted mortality rates by risk category for patients receiving or not receiving a beta-blocker were:

• No risk factors: 1.0% with a beta-blocker vs 0.6% without
• One or two risk factors: 1.7% vs 1.5%
• Three or four risk factors: 2.3% vs 4.5%.

After adjustment for confounding factors, the 30-day mortality rate was higher in low-risk patients and lower in high-risk patients who received beta-blockers. Odds ratios for death in beta-blocker users (entire cohort) by risk category were:

• No risk factors: 1.19
• One or two risk factors 0.97
• Three or four risk factors 0.76.

In the 3.8% of the total cohort who underwent cardiac surgery, beta-blockers had no significant effect—beneficial or harmful—in any risk group.

Jørgensen ME, Hlatky MA, Køber L, et al. ß-blocker-associated risks in patients with uncomplicated hypertension undergoing noncardiac surgery. JAMA Intern Med 2015; 175:1923–1931.

The study. Jørgensen et al⁷ investigated the association between chronic beta-blocker use for the treatment of hypertension and 30-day rates of mortality and major adverse cardiac events. Eligible patients (N = 55,320) were at least 20 years old and were undergoing any type of noncardiac surgery. The authors established that hypertension was present through use of an algorithm based on the International Classification of Diseases (10th edition). Patients with existing cardiovascular disease and renal disease were excluded. The authors used multivariate analysis to adjust for confounding factors.

Results. Twenty-six percent of the patients were on chronic beta-blocker therapy for hypertension. The mortality rate at 30 days was 1.93% in patients treated with a beta-blocker alone or in combination with other antihypertensive drugs; the rate was 1.32% for patients receiving any combination of renin-angiotensin system inhibitor, calcium antagonist, or thiazide, but no beta-blocker. Similarly, the 30-day major adverse cardiac event rates were 1.32% with beta-blockers and 0.84% without beta-blockers.

In subgroup analysis, each medication combination that included a beta-blocker was associated with higher rates of death and major adverse cardiac events than the same combination without a beta-blocker. Odds ratios for major adverse cardiac events with beta-blocker combinations ranged from 1.22 to 2.16 compared with regimens with no beta-blocker.

Implications for clinical practice. These two studies added to a growing chorus of concerns about the value and safety of beta-blockers in surgical patients. Friedell et al⁶ made an observation that was remarkably similar to one reported by Lindenauer et al⁸ in 2005: when patients were stratified by baseline risk of death, only those with the highest baseline risk benefited from beta-blocker therapy. Those in the lowest risk group actually were harmed by beta-blocker use, ie, the mortality rate was higher.

More interesting is the novel observation by Jørgensen et al⁷ that even in patients with no known cardiovascular disease who are on chronic beta-blocker therapy—presumably on stable doses and not solely for perioperative risk reduction—rates of mortality and major adverse cardiac events were higher than for patients not on chronic beta-blocker therapy.

The current studies support a cautious, selective approach to the perioperative use of beta-blockers—they should be used only in high-risk patients undergoing high-risk surgery, as has been proposed by the ACC/AHA.¹

Statins protect

Antoniou GA, Hajibandeh S, Hajibandeh S, Vallabhaneni SR, Brennan JA, Torella F. Meta-analysis of the effects of statins on perioperative outcomes in vascular and endovascular surgery. J Vasc Surg 2015; 61:519–532.

The study⁹ was a comprehensive meta-analysis of randomized controlled trials and observational studies of the effects of HMG-CoA reductase inhibitors (statins) on perioperative outcomes in patients undergoing vascular surgery (but not for intracranial or coronary artery disease). Twenty-four studies were included, 4 randomized controlled trials and 20 observational studies (including 16 cohort and 4 case-controlled studies), with a total of 22,536 patients, 8,052 receiving statins and 15,484 not receiving statins.

Results. Although there was no significant difference in cardiovascular mortality rates, patients receiving statins had significantly lower rates of all-cause mortality, myocardial infarction, stroke, and a composite of myocardial infarction, stroke, and death at 30 days postoperatively than patients not receiving statins. Additionally, there was no difference in the incidence of kidney injury between groups. The possibility of publication bias was thought to be low for all of these outcomes.

Berwanger O, Le Manach Y, Suzumura EA, et al. Association between pre-operative statin use and major cardiovascular complications among patients undergoing non-cardiac surgery: the VISION study. Eur Heart J 2016; 37:177–185.

The study. The Vascular Events in Non-cardiac Surgery Patients Cohort Evaluation (VISION) study^10,11 was an international prospective cohort study of more than 40,000 patients age 45 and older undergoing major noncardiac surgery with either general or regional anesthesia. Postoperative troponin measurements were obtained in all patients 6 to 12 hours after surgery and for the first 3 postoperative days. The authors evaluated the effect of preoperative statin use on cardiovascular outcomes at 30 days after surgery using a multivariate logistic model and propensity score analysis to correct for confounding factors. Statin use was defined as exposure within 7 days before surgery or 3 days after.

Results. In the 15,478 patients included in the analysis, statin use conferred a significant reduction in the primary outcome (composite of all-cause mortality, myocardial injury after noncardiac surgery, or stroke); the absolute risk reduction was 2.0%. Statin users also had a significantly lower risk of all-cause mortality, cardiovascular mortality, and myocardial injury after noncardiac surgery, but not of postoperative myocardial infarction or stroke. This analysis did not address the type of statin, dosing, or safety markers such as liver and muscle function.

Implications for clinical practice. With largely observational data and a few small randomized trials, these meta-analyses provide important information with respect to perioperative cardiovascular protection by statins. Starting a statin before surgery and continuing it perioperatively seems appropriate in patients at high risk (as recommended by the ACC/AHA guidelines¹). Based on other data, the benefit may be evident in as little as 5 days, as this is when statins appear to reach their plateau with regard to their vascular pleiotropic effects.¹² The incidence of adverse effects of statins, including muscle and liver injury, appears to be low in the perioperative setting.¹³

Given the inconsistent data regarding perioperative beta-blocker therapy, statins may very well be the most important perioperative medication with respect to cardiovascular risk reduction. However, a large randomized trial would help to confirm this belief.

Restart angiotensin II receptor blockers soon after surgery

Lee SM, Takemoto S, Wallace AW. Association between withholding angiotensin receptor blockers in the early postoperative period and 30-day mortality: a cohort study of the Veterans Affairs Healthcare System. Anesthesiology 2015; 123:288–306.

A concern about perioperative use of ARBs is that they impair the renin-angiotensin-aldosterone system, which maintains blood pressure under general anesthesia. ARB-induced intraoperative hypotension is particularly difficult to control, as it is often refractory to treatment with conventional adrenergic vasopressors.

The study. Lee et al¹⁴ conducted a retrospective cohort trial to evaluate the effects of continuing to withhold ARBs postoperatively. Of the 30,173 patients admitted for surgery in the Veterans Affairs system from 1999 through 2011 who were taking an ARB before surgery and who met the inclusion criteria, 10,205 (33.8%) were not restarted on their medication by postoperative day 2.

Results. The mortality rate at 30 days was higher in those whose ARBs were withheld than in those in whom it was resumed, with a multivariable-adjusted hazard ratio of 1.74 (95% CI 1.47–2.06; P < .001). The risk of withholding ARBs was more pronounced in younger patients (hazard ratio 2.52; 95% CI 1.69–3.76 in those under age 60) than in older patients (hazard ratio 1.42, 95% CI 1.09–1.85 in those over age 75).

Implications for clinical practice. While not addressing whether to continue or withhold ARBs preoperatively, this retrospective study presented evidence that delay in resuming chronic ARB therapy after surgery was common and appeared to be associated with a higher 30-day mortality rate. The ACC/AHA guidelines¹ state:

• Continuing angiotensin-converting enzyme (ACE) inhibitors or ARBs perioperatively is reasonable (class IIa recommendation, level of evidence B) (Table 1).
• If an ACE inhibitor or ARB is withheld before surgery, it is reasonable to restart it postoperatively as soon as clinically feasible (class IIa recommendation, level of evidence C).

Close attention to medication reconciliation in the postoperative period is necessary to facilitate early resumption of ARBs.

CORONARY STENTS AND ANTIPLATELET THERAPY IN NONCARDIAC SURGERY PATIENTS

Considerations in the management of noncardiac surgery patients with stents include risks of stent thrombosis, bleeding, and potentially delaying procedures to continue uninterrupted dual antiplatelet therapy. Evidence is evolving regarding the risks of perioperative complications in patients with bare-metal stents and drug-eluting stents, as well as the optimal timing before noncardiac surgery.

Bare-metal vs drug-eluting stents

Bangalore S, Silbaugh TS, Normand SL, Lovett AF, Welt FG, Resnic FS. Drug-eluting stents versus bare metal stents prior to noncardiac surgery. Catheter Cardiovasc Interv 2015; 85:533–541.

The study. Bangalore et al¹⁵ compared the safety of drug-eluting vs bare-metal stents in noncardiac surgery patients and investigated adverse events stratified by time since stent placement. This was a retrospective observational study of 8,415 patients in the Massachusetts claims database who underwent noncardiac surgery 1 year or less after percutaneous coronary intervention.

Results. There was no significant difference in the incidence of the primary outcome (composite of death, myocardial infarction, and bleeding) between the two groups.

With drug-eluting stents, patients had lower 30-day postoperative mortality rates, and their rate of the primary outcome decreased with time from percutaneous coronary intervention to surgery, being lowest beyond 90 days:

• 8.6% in days 1–30
• 7.5% in days 31–90
• 5.2% in days 91–180
• 5.8% in days 181–365 (P = .02).

With bare-metal stents, the event rate remained high over time:

• 8.2% in days 1–30
• 6.6% in days 31–90
• 8.1% in days 91–180
• 8.8% in days 181–365 (P = .60).

This study did not report information about perioperative antiplatelet management and was limited to first-generation drug-eluting stents. 

Saia F, Belotti LM, Guastaroba P, et al. Risk of adverse cardiac and bleeding events following cardiac and noncardiac surgery in patients with coronary stents: how important is the interplay between stent type and time from stenting to surgery? Circ Cardiovasc Qual Outcomes 2015; 9:39–47.

The study. Saia et al¹⁶ retrospectively examined predictors of periprocedural ischemic and bleeding events among cardiac and noncardiac surgical patients who had previously undergone percutaneous coronary intervention. They also assessed the risks associated with stent type and time from percutaneous coronary intervention to surgery.

Of 39,362 patients, 13,128 underwent procedures during the 5-year study period. The cumulative incidence of surgery was 3.6% at 30 days, 14% at 1 year, and 40% at 5 years after percutaneous coronary intervention. Almost 30% of the procedures were done urgently.

Results. The 30-day rate of postoperative cardiac death was 2.5%, nonfatal myocardial infarction 1.5%, and serious bleeding events 6.5%. Older drug-eluting stents were associated with higher risks of adverse events than newer drug-eluting stents at any time point (odds ratio 2.1 at 0–180 days, 1.9 at 6–12 months, and 1.45 after 12 months). Surgery performed 6 to 12 months after percutaneous coronary intervention had lower rates of adverse outcomes than surgery performed within 6 months. Beyond 6 months from percutaneous coronary intervention, bare-metal stents and newer drug-eluting stents did not have significantly different adverse event rates; however, newer drug-eluting stents appeared safer than bare-metal stents from 0 to 180 days.

Limitations of this study included lack of information regarding periprocedural antiplatelet management and a relatively small subset of newer drug-eluting stent patients.

Implications for clinical practice. These studies added to earlier work that demonstrated that the risk of perioperative adverse events differs by both the stent type and the time from percutaneous coronary intervention to noncardiac surgery. In patients with a drug-eluting stent, the risk levels off 90 days after percutaneous coronary intervention, suggesting that the previously recommended 12 months of uninterrupted dual antiplatelet therapy (per the 2014 ACC/AHA guidelines¹) may not be needed, particularly with newer-generation drug-eluting stents. Based on new evidence, the ACC/AHA guidelines regarding perioperative management of dual antiplatelet therapy in noncardiac surgery patients were updated,¹⁷ as noted below.

An update to the ACC/AHA guidelines on dual antiplatelet therapy

Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease. Circulation 2016 Mar 29. DOI: 10.1161/CIR.0000000000000404. [Epub ahead of print]

The 2016 update¹⁷ provides the following recommendations for patients with coronary stents who undergo noncardiac surgery:

• Delay elective surgery for 30 days after placement of a bare-metal stent (class I recommendation, level of evidence B).
• It is optimal to delay elective surgery 6 months after drug-eluting stent placement (class I recommendation, level of evidence B).
• If dual antiplatelet therapy must be discontinued, then continue aspirin if possible and restart the P2Y₁₂ inhibitor as soon as possible postoperatively (class I recommendation, level of evidence C ).
• A consensus decision among treating clinicians is useful regarding the risks of surgery and discontinuation or continuation of antiplatelet therapy (class IIa recommendation, level of evidence C).
• If dual antiplatelet therapy must be discontinued, then elective surgery should not be performed less than 30 days after bare-metal stent placement, or less than 3 months after drug-eluting stent placement (class III recommendation, level of evidence B).
• Elective surgery after drug-eluting stent placement when the P2Y₁₂ inhibitor must be discontinued may be considered 3 months after drug-eluting stent placement if the risk of surgical delay is greater than the risk of stent thrombosis (class IIb recommendation, level of evidence C).

The basic differences are the new recommendations for a minimum of 6 months of dual antiplatelet therapy as opposed to 12 months after drug-eluting stent placement before elective noncardiac surgery, and to allow surgery after 3 months (as opposed to 6 months) if the risk of delaying surgery outweighs the risk of stent thrombosis or myocardial infarction.

PERIOPERATIVE ANTICOAGULATION

The optimal perioperative management of patients with atrial fibrillation who are on warfarin is uncertain. The American College of Chest Physicians guidelines¹⁸ categorized patients with atrial fibrillation into low, moderate, and high thromboembolic risk. Based primarily on observational data, these guidelines recommended perioperative bridging anticoagulation for those at high risk but not for those at low risk. For intermediate-risk patients, there were insufficient data to make any recommendation.

Bridging may not benefit those at intermediate risk

Douketis JD, Spyropoulos AC, Kaatz S, et al; BRIDGE Investigators. Perioperative bridging anticoagulation in patients with atrial fibrillation. N Engl J Med 2015; 373:823–833.

The study. The Bridging Anticoagulation in Patients Who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery (BRIDGE) trial¹⁹ was the first randomized controlled trial to examine the effects of perioperative bridging anticoagulation in patients with atrial fibrillation without mechanical heart valves.

Results. In 1,884 patients undergoing elective surgery, the incidence of arterial thromboembolism was 0.4% in the no-bridging group and 0.3% in the bridging group (95% CI −0.6 to 0.8; P = .01 for noninferiority). Major bleeding occurred in 1.3% of patients in the no-bridging group and 3.2% in the bridging group (95% CI 0.20–0.78; P = .005 for superiority).

These results suggest that the risks of bridging therapy are greater than the benefits. Of note, the mean CHADS₂ score (1 point each for congestive heart failure, hypertension, age ≥ 75 years, and diabetes mellitus; 2 points for previous stroke or transient ischemic attack; a total score > 2 indicates significant risk of stroke) for patients enrolled in this trial was 2.3, and it may be difficult to extrapolate these results to the limited number of patients at highest risk, ie, who have a CHADS₂ score of 5 or 6. Also, this study did not address patients with arterial or venous thromboembolism.

Implications for clinical practice. Despite the limitations noted above, this study does provide guidance for management of the intermediate-risk group with atrial fibrillation as defined by the American College of Chest Physicians: a no-bridging strategy is the best option.

Last year, Congress passed legislation to permanently eliminate the Sustainable Growth Rate (SGR) formula, created in 1997 and designed to hold Medicare Part B or outpatient spending under control. Allowing the SGR to go into effect would have severely cut physician reimbursements in recent years, but Congress passed legislation each year to temporarily avert these cuts (also known annually as the “doc fix”). In search of a permanent solution, the passage of bipartisan legislation permanently repealing the SGR in 2015 was hailed as a way to ensure more certainty around the future of Medicare payments for physicians.

This legislation (H.R. 2, 114th Congress), sponsored by Rep. Michael C. Burgess (R-Texas) and entitled “Medicare Access and CHIP Reauthorization Act of 2015,” or MACRA, does much more than simply remove the SGR’s threat of broader Medicare payment cuts. The law changes the ways physicians are reimbursed by Medicare and continues to shift our healthcare system away from volume-based reimbursements and toward a value-based payment system.

What Is MIPS?

MACRA creates two value-based payment tracks for physicians. The first, the Merit-Based Incentive Payment System (MIPS), is closer to the old fee-for-service model of reimbursement. However, MIPS takes into account both volume and quality (i.e., payment is adjusted based on physician-quality scores). These physician-specific scores broaden the scope of quality measurement by including new measures related to resource utilization, electronic health record (EHR) use, and clinical improvement practices, along with the traditional clinical quality markers.

Under MIPS, the current Physician Quality Reporting System (PQRS), EHR Incentive Program, and Physician Value-Based Modifier all will be integrated into this single-payment adjustment.

The range of potential payment adjustments based on a physician’s MIPS score grows each year through 2022 (in 2022, adjustments can range from +9% to -9%). The program is budget neutral, which means that increases in payments to high-scoring providers will be offset by decreases in payments to low-scoring providers. For 2019 to 2024, there also will be an additional payment adjustment given to the highest MIPS performers for exceptional performance.

A benefit of MIPS is that it will streamline the various quality-reporting programs currently in place into one single program and does not ask physicians to assume any additional financial risk related to outcomes when taking care of patients. However, the particulars of how the MIPS score will be calculated are yet to be determined, and much of the utility and palatability of this score will depend on the chosen metrics. The goal of these metrics should be that they are meaningful, valid, and attributable to specific providers.

What Are APMs?

The other payment option MACRA provides for physicians allows them to opt out of MIPS and participate in the Alternative Payment Models (APMs) track. To incentivize physicians to take part in this riskier track, providers taking part in APMs will receive some extra money for their participation: a 5% annual lump sum bonus on reimbursement payments. To clarify, qualifying APMs are those where providers take on “more than nominal” financial risk, report on their quality measures, and use certified EHR technology.

To qualify as a participant in an APM (for example, the Medicare Shared Savings Program), providers must hit a threshold for percentage of total revenue received or percentage of patients from qualifying APMs. This threshold will increase over time. For example, from 2019 to 2020, providers must obtain at least 25% of their Medicare revenue or patients via APMs, whereas in 2023, 75% of their Medicare revenue or \ patients will need to come from APMs.

Providers will benefit from the increased reimbursement offered if they participate in APMs. There also is funding allocated in MACRA to help develop quality measures, with a call for physician leads to develop quality standards. This payment model, however, does come with increased financial risk for the provider contingent on patient outcomes. In addition, it may be difficult for all providers to hit the thresholds for participation.

Stick with MIPS? Or Take the Plunge with APM?

How MACRA affects you will depend a lot on the practice environment. As described above, MACRA is designed to move physicians into risk-based payment structures if possible. If possible, or otherwise, to simplify the current fee-for-service mechanism of payment by consolidating various Medicare pay-for-performance programs.

Let’s look at a few scenarios:

Hospitalist A works for a physician group that assumes risk for patients in a MACRA-approved APM and sees only those inpatients as opposed to unassigned patients. Therefore, almost all of hospitalist A’s patients are covered by risk-based contracts, and hospitalist A might be well positioned for the new APM structure.

Hospitalist B works for a group, or a university, and sees whatever patients are admitted to the hospital. Hospitalist B’s eligibility to participate in the APM will depend on the percentage of patients in alternative payment models in their market. If hospitalist B’s market has many Medicare accountable care organizations, and Medicaid and the commercial insurers compensate through a risk-sharing model, hospitalist B might reach the threshold. This is more accidental than planned, however, and hospitalist B might not be able to consistently hit this threshold year after year.

In addition, just working within the model will probably not be enough to qualify. Hospitalist B will need to also take on “more than nominal risk” as a participant in the model. In an employed academic setting, where the hospital is taking on risk as part of an APM, it is unlikely hospitalist B will qualify just by virtue of hospital employment. Hospitalist B must also meet/exceed the patient or payment thresholds under the model.

Bottom line: Given the current situation, we expect many hospitalists will likely be required to participate in MIPS and not qualify for APMs. Understanding the details and expectations now will help them be successful in the future.

Is MACRA Good for Hospitalists?

Most of organized medicine is happy to be free from the annual threat of reimbursement cuts. In addition, the new law might streamline quality reporting. But the specific upside depends on your perspective.

With APMs, a hospitalist might enjoy more upside potential, particularly for high-quality work and EHR use. However, whether it is realistic for most hospitalists to even participate in the model depends on many factors, as described previously, and SHM is advocating for the law to be implemented in ways that will more readily accommodate hospitalist practice and employment structures.

For example, the SHM Public Policy Committee has provided the Centers for Medicare & Medicaid Services (CMS) with realistic options for implementing the APM framework that would allow hospitalist B in the above example to qualify as an APM participant.

With MIPS, the benefit to hospitalists depends a fair amount on the way the law is implemented: how quality reporting happens, what metrics will count as quality improvement efforts, and how utilization of EHRs is measured.

What Issues Should Hospitalists Be Aware Of?

As MACRA is further developed, the main issue for hospitalists will be to ensure fairness in assessing quality and incentive payments. As previously encountered with quality reporting, hospitalists are not differentiated clearly from outpatient providers. As a result, they could suffer from the comparison of their quality outcomes for their sicker hospitalized patients to the patients cared for in a typical primary-care internal medicine practice. This inaccurate comparison poses problems in both models.

A potential solution would be a hospitalist-specific billing code, which would make it easier to identify hospitalists. SHM applied for and advocated for the approval of such a billing code and the request was recently approved by CMS.

In addition, as hospitalists mostly work in groups with shift-based schedules, thus sharing care of patients, individual identifiers may not be as significant as possibly looking at hospital, system, or team-based metrics. Using facility performance measures for both clinical quality and performance improvement—where hospitalists can opt to align with their hospital, which is already reporting quality outcomes—might be one way out of this conundrum. It would take into account the type of facility-level quality improvement work many hospitalists participate in. This also would decrease reporting burden for hospitalist groups.

SHM has advocated for this solution and was able to ensure this concept was included in the law; however, it is unclear when or how CMS will implement it.

To summarize, looking good in quality reporting will continue to be a challenge for hospitalists. It will be critical to keep pressure on CMS to implement solutions that account for the unique situation of our specialty.

Another issue to be aware of is the ability of hospitalists to participate in APMs. As with other facility-based providers, hospitalists have little control over whether their facility participates in an APM. Ways to ensure hospitalists can reach thresholds for participation could include allowing the various APMs that hospitalist patients are aligned with count toward an individual hospitalists’ APM participation total—a solution that SHM is advocating for Medicare to include in the APM framework.

What’s Next?

Much remains to be solidified regarding implementation of MACRA, despite the fact it goes live in a few short years (see Figure 1). CMS has asked for comments and stakeholder input regarding MIPS and APMs, and it will be releasing the first round of rules around MACRA this year.

SHM is actively working with CMS to ensure this legislation will reflect the work we are doing as hospitalists to provide high-quality clinical care for our patients and enhance the performance of our hospitals and health system. TH

Dr. Doctoroff is a hospitalist at Beth Israel Deaconess Medical Center and an instructor of medicine at Harvard Medical School in Boston. Dr. Dutta is a hospitalist at Rush University Medical Center and an assistant professor of medicine at Rush Medical College in Chicago. Both are members of the SHM Public Policy Committee.

Last year, Congress passed legislation to permanently eliminate the Sustainable Growth Rate (SGR) formula, created in 1997 and designed to hold Medicare Part B or outpatient spending under control. Allowing the SGR to go into effect would have severely cut physician reimbursements in recent years, but Congress passed legislation each year to temporarily avert these cuts (also known annually as the “doc fix”). In search of a permanent solution, the passage of bipartisan legislation permanently repealing the SGR in 2015 was hailed as a way to ensure more certainty around the future of Medicare payments for physicians.

This legislation (H.R. 2, 114th Congress), sponsored by Rep. Michael C. Burgess (R-Texas) and entitled “Medicare Access and CHIP Reauthorization Act of 2015,” or MACRA, does much more than simply remove the SGR’s threat of broader Medicare payment cuts. The law changes the ways physicians are reimbursed by Medicare and continues to shift our healthcare system away from volume-based reimbursements and toward a value-based payment system.

What Is MIPS?

MACRA creates two value-based payment tracks for physicians. The first, the Merit-Based Incentive Payment System (MIPS), is closer to the old fee-for-service model of reimbursement. However, MIPS takes into account both volume and quality (i.e., payment is adjusted based on physician-quality scores). These physician-specific scores broaden the scope of quality measurement by including new measures related to resource utilization, electronic health record (EHR) use, and clinical improvement practices, along with the traditional clinical quality markers.

Under MIPS, the current Physician Quality Reporting System (PQRS), EHR Incentive Program, and Physician Value-Based Modifier all will be integrated into this single-payment adjustment.

The range of potential payment adjustments based on a physician’s MIPS score grows each year through 2022 (in 2022, adjustments can range from +9% to -9%). The program is budget neutral, which means that increases in payments to high-scoring providers will be offset by decreases in payments to low-scoring providers. For 2019 to 2024, there also will be an additional payment adjustment given to the highest MIPS performers for exceptional performance.

A benefit of MIPS is that it will streamline the various quality-reporting programs currently in place into one single program and does not ask physicians to assume any additional financial risk related to outcomes when taking care of patients. However, the particulars of how the MIPS score will be calculated are yet to be determined, and much of the utility and palatability of this score will depend on the chosen metrics. The goal of these metrics should be that they are meaningful, valid, and attributable to specific providers.

What Are APMs?

The other payment option MACRA provides for physicians allows them to opt out of MIPS and participate in the Alternative Payment Models (APMs) track. To incentivize physicians to take part in this riskier track, providers taking part in APMs will receive some extra money for their participation: a 5% annual lump sum bonus on reimbursement payments. To clarify, qualifying APMs are those where providers take on “more than nominal” financial risk, report on their quality measures, and use certified EHR technology.

To qualify as a participant in an APM (for example, the Medicare Shared Savings Program), providers must hit a threshold for percentage of total revenue received or percentage of patients from qualifying APMs. This threshold will increase over time. For example, from 2019 to 2020, providers must obtain at least 25% of their Medicare revenue or patients via APMs, whereas in 2023, 75% of their Medicare revenue or \ patients will need to come from APMs.

Providers will benefit from the increased reimbursement offered if they participate in APMs. There also is funding allocated in MACRA to help develop quality measures, with a call for physician leads to develop quality standards. This payment model, however, does come with increased financial risk for the provider contingent on patient outcomes. In addition, it may be difficult for all providers to hit the thresholds for participation.

Stick with MIPS? Or Take the Plunge with APM?

How MACRA affects you will depend a lot on the practice environment. As described above, MACRA is designed to move physicians into risk-based payment structures if possible. If possible, or otherwise, to simplify the current fee-for-service mechanism of payment by consolidating various Medicare pay-for-performance programs.

Let’s look at a few scenarios:

Hospitalist A works for a physician group that assumes risk for patients in a MACRA-approved APM and sees only those inpatients as opposed to unassigned patients. Therefore, almost all of hospitalist A’s patients are covered by risk-based contracts, and hospitalist A might be well positioned for the new APM structure.

Hospitalist B works for a group, or a university, and sees whatever patients are admitted to the hospital. Hospitalist B’s eligibility to participate in the APM will depend on the percentage of patients in alternative payment models in their market. If hospitalist B’s market has many Medicare accountable care organizations, and Medicaid and the commercial insurers compensate through a risk-sharing model, hospitalist B might reach the threshold. This is more accidental than planned, however, and hospitalist B might not be able to consistently hit this threshold year after year.

In addition, just working within the model will probably not be enough to qualify. Hospitalist B will need to also take on “more than nominal risk” as a participant in the model. In an employed academic setting, where the hospital is taking on risk as part of an APM, it is unlikely hospitalist B will qualify just by virtue of hospital employment. Hospitalist B must also meet/exceed the patient or payment thresholds under the model.

Bottom line: Given the current situation, we expect many hospitalists will likely be required to participate in MIPS and not qualify for APMs. Understanding the details and expectations now will help them be successful in the future.

Is MACRA Good for Hospitalists?

Most of organized medicine is happy to be free from the annual threat of reimbursement cuts. In addition, the new law might streamline quality reporting. But the specific upside depends on your perspective.

With APMs, a hospitalist might enjoy more upside potential, particularly for high-quality work and EHR use. However, whether it is realistic for most hospitalists to even participate in the model depends on many factors, as described previously, and SHM is advocating for the law to be implemented in ways that will more readily accommodate hospitalist practice and employment structures.

For example, the SHM Public Policy Committee has provided the Centers for Medicare & Medicaid Services (CMS) with realistic options for implementing the APM framework that would allow hospitalist B in the above example to qualify as an APM participant.

With MIPS, the benefit to hospitalists depends a fair amount on the way the law is implemented: how quality reporting happens, what metrics will count as quality improvement efforts, and how utilization of EHRs is measured.

What Issues Should Hospitalists Be Aware Of?

As MACRA is further developed, the main issue for hospitalists will be to ensure fairness in assessing quality and incentive payments. As previously encountered with quality reporting, hospitalists are not differentiated clearly from outpatient providers. As a result, they could suffer from the comparison of their quality outcomes for their sicker hospitalized patients to the patients cared for in a typical primary-care internal medicine practice. This inaccurate comparison poses problems in both models.

A potential solution would be a hospitalist-specific billing code, which would make it easier to identify hospitalists. SHM applied for and advocated for the approval of such a billing code and the request was recently approved by CMS.

In addition, as hospitalists mostly work in groups with shift-based schedules, thus sharing care of patients, individual identifiers may not be as significant as possibly looking at hospital, system, or team-based metrics. Using facility performance measures for both clinical quality and performance improvement—where hospitalists can opt to align with their hospital, which is already reporting quality outcomes—might be one way out of this conundrum. It would take into account the type of facility-level quality improvement work many hospitalists participate in. This also would decrease reporting burden for hospitalist groups.

SHM has advocated for this solution and was able to ensure this concept was included in the law; however, it is unclear when or how CMS will implement it.

To summarize, looking good in quality reporting will continue to be a challenge for hospitalists. It will be critical to keep pressure on CMS to implement solutions that account for the unique situation of our specialty.

Another issue to be aware of is the ability of hospitalists to participate in APMs. As with other facility-based providers, hospitalists have little control over whether their facility participates in an APM. Ways to ensure hospitalists can reach thresholds for participation could include allowing the various APMs that hospitalist patients are aligned with count toward an individual hospitalists’ APM participation total—a solution that SHM is advocating for Medicare to include in the APM framework.

What’s Next?

Much remains to be solidified regarding implementation of MACRA, despite the fact it goes live in a few short years (see Figure 1). CMS has asked for comments and stakeholder input regarding MIPS and APMs, and it will be releasing the first round of rules around MACRA this year.

SHM is actively working with CMS to ensure this legislation will reflect the work we are doing as hospitalists to provide high-quality clinical care for our patients and enhance the performance of our hospitals and health system. TH

Dr. Doctoroff is a hospitalist at Beth Israel Deaconess Medical Center and an instructor of medicine at Harvard Medical School in Boston. Dr. Dutta is a hospitalist at Rush University Medical Center and an assistant professor of medicine at Rush Medical College in Chicago. Both are members of the SHM Public Policy Committee.

Last year, Congress passed legislation to permanently eliminate the Sustainable Growth Rate (SGR) formula, created in 1997 and designed to hold Medicare Part B or outpatient spending under control. Allowing the SGR to go into effect would have severely cut physician reimbursements in recent years, but Congress passed legislation each year to temporarily avert these cuts (also known annually as the “doc fix”). In search of a permanent solution, the passage of bipartisan legislation permanently repealing the SGR in 2015 was hailed as a way to ensure more certainty around the future of Medicare payments for physicians.

This legislation (H.R. 2, 114th Congress), sponsored by Rep. Michael C. Burgess (R-Texas) and entitled “Medicare Access and CHIP Reauthorization Act of 2015,” or MACRA, does much more than simply remove the SGR’s threat of broader Medicare payment cuts. The law changes the ways physicians are reimbursed by Medicare and continues to shift our healthcare system away from volume-based reimbursements and toward a value-based payment system.

What Is MIPS?

MACRA creates two value-based payment tracks for physicians. The first, the Merit-Based Incentive Payment System (MIPS), is closer to the old fee-for-service model of reimbursement. However, MIPS takes into account both volume and quality (i.e., payment is adjusted based on physician-quality scores). These physician-specific scores broaden the scope of quality measurement by including new measures related to resource utilization, electronic health record (EHR) use, and clinical improvement practices, along with the traditional clinical quality markers.

Under MIPS, the current Physician Quality Reporting System (PQRS), EHR Incentive Program, and Physician Value-Based Modifier all will be integrated into this single-payment adjustment.

The range of potential payment adjustments based on a physician’s MIPS score grows each year through 2022 (in 2022, adjustments can range from +9% to -9%). The program is budget neutral, which means that increases in payments to high-scoring providers will be offset by decreases in payments to low-scoring providers. For 2019 to 2024, there also will be an additional payment adjustment given to the highest MIPS performers for exceptional performance.

A benefit of MIPS is that it will streamline the various quality-reporting programs currently in place into one single program and does not ask physicians to assume any additional financial risk related to outcomes when taking care of patients. However, the particulars of how the MIPS score will be calculated are yet to be determined, and much of the utility and palatability of this score will depend on the chosen metrics. The goal of these metrics should be that they are meaningful, valid, and attributable to specific providers.

What Are APMs?

The other payment option MACRA provides for physicians allows them to opt out of MIPS and participate in the Alternative Payment Models (APMs) track. To incentivize physicians to take part in this riskier track, providers taking part in APMs will receive some extra money for their participation: a 5% annual lump sum bonus on reimbursement payments. To clarify, qualifying APMs are those where providers take on “more than nominal” financial risk, report on their quality measures, and use certified EHR technology.

To qualify as a participant in an APM (for example, the Medicare Shared Savings Program), providers must hit a threshold for percentage of total revenue received or percentage of patients from qualifying APMs. This threshold will increase over time. For example, from 2019 to 2020, providers must obtain at least 25% of their Medicare revenue or patients via APMs, whereas in 2023, 75% of their Medicare revenue or \ patients will need to come from APMs.

Providers will benefit from the increased reimbursement offered if they participate in APMs. There also is funding allocated in MACRA to help develop quality measures, with a call for physician leads to develop quality standards. This payment model, however, does come with increased financial risk for the provider contingent on patient outcomes. In addition, it may be difficult for all providers to hit the thresholds for participation.

Stick with MIPS? Or Take the Plunge with APM?

How MACRA affects you will depend a lot on the practice environment. As described above, MACRA is designed to move physicians into risk-based payment structures if possible. If possible, or otherwise, to simplify the current fee-for-service mechanism of payment by consolidating various Medicare pay-for-performance programs.

Let’s look at a few scenarios:

Hospitalist A works for a physician group that assumes risk for patients in a MACRA-approved APM and sees only those inpatients as opposed to unassigned patients. Therefore, almost all of hospitalist A’s patients are covered by risk-based contracts, and hospitalist A might be well positioned for the new APM structure.

Hospitalist B works for a group, or a university, and sees whatever patients are admitted to the hospital. Hospitalist B’s eligibility to participate in the APM will depend on the percentage of patients in alternative payment models in their market. If hospitalist B’s market has many Medicare accountable care organizations, and Medicaid and the commercial insurers compensate through a risk-sharing model, hospitalist B might reach the threshold. This is more accidental than planned, however, and hospitalist B might not be able to consistently hit this threshold year after year.

In addition, just working within the model will probably not be enough to qualify. Hospitalist B will need to also take on “more than nominal risk” as a participant in the model. In an employed academic setting, where the hospital is taking on risk as part of an APM, it is unlikely hospitalist B will qualify just by virtue of hospital employment. Hospitalist B must also meet/exceed the patient or payment thresholds under the model.

Bottom line: Given the current situation, we expect many hospitalists will likely be required to participate in MIPS and not qualify for APMs. Understanding the details and expectations now will help them be successful in the future.

Is MACRA Good for Hospitalists?

Most of organized medicine is happy to be free from the annual threat of reimbursement cuts. In addition, the new law might streamline quality reporting. But the specific upside depends on your perspective.

With APMs, a hospitalist might enjoy more upside potential, particularly for high-quality work and EHR use. However, whether it is realistic for most hospitalists to even participate in the model depends on many factors, as described previously, and SHM is advocating for the law to be implemented in ways that will more readily accommodate hospitalist practice and employment structures.

For example, the SHM Public Policy Committee has provided the Centers for Medicare & Medicaid Services (CMS) with realistic options for implementing the APM framework that would allow hospitalist B in the above example to qualify as an APM participant.

With MIPS, the benefit to hospitalists depends a fair amount on the way the law is implemented: how quality reporting happens, what metrics will count as quality improvement efforts, and how utilization of EHRs is measured.

What Issues Should Hospitalists Be Aware Of?

As MACRA is further developed, the main issue for hospitalists will be to ensure fairness in assessing quality and incentive payments. As previously encountered with quality reporting, hospitalists are not differentiated clearly from outpatient providers. As a result, they could suffer from the comparison of their quality outcomes for their sicker hospitalized patients to the patients cared for in a typical primary-care internal medicine practice. This inaccurate comparison poses problems in both models.

A potential solution would be a hospitalist-specific billing code, which would make it easier to identify hospitalists. SHM applied for and advocated for the approval of such a billing code and the request was recently approved by CMS.

In addition, as hospitalists mostly work in groups with shift-based schedules, thus sharing care of patients, individual identifiers may not be as significant as possibly looking at hospital, system, or team-based metrics. Using facility performance measures for both clinical quality and performance improvement—where hospitalists can opt to align with their hospital, which is already reporting quality outcomes—might be one way out of this conundrum. It would take into account the type of facility-level quality improvement work many hospitalists participate in. This also would decrease reporting burden for hospitalist groups.

SHM has advocated for this solution and was able to ensure this concept was included in the law; however, it is unclear when or how CMS will implement it.

To summarize, looking good in quality reporting will continue to be a challenge for hospitalists. It will be critical to keep pressure on CMS to implement solutions that account for the unique situation of our specialty.

Another issue to be aware of is the ability of hospitalists to participate in APMs. As with other facility-based providers, hospitalists have little control over whether their facility participates in an APM. Ways to ensure hospitalists can reach thresholds for participation could include allowing the various APMs that hospitalist patients are aligned with count toward an individual hospitalists’ APM participation total—a solution that SHM is advocating for Medicare to include in the APM framework.

What’s Next?

Much remains to be solidified regarding implementation of MACRA, despite the fact it goes live in a few short years (see Figure 1). CMS has asked for comments and stakeholder input regarding MIPS and APMs, and it will be releasing the first round of rules around MACRA this year.

SHM is actively working with CMS to ensure this legislation will reflect the work we are doing as hospitalists to provide high-quality clinical care for our patients and enhance the performance of our hospitals and health system. TH

Dr. Doctoroff is a hospitalist at Beth Israel Deaconess Medical Center and an instructor of medicine at Harvard Medical School in Boston. Dr. Dutta is a hospitalist at Rush University Medical Center and an assistant professor of medicine at Rush Medical College in Chicago. Both are members of the SHM Public Policy Committee.