Steven L. Cohn, MD

Background

The American College of Cardiology/American Heart Association (ACC/AHA) perioperative guideline provides recommendations for the evaluation and management of cardiovascular disease in patients undergoing noncardiac surgery.

Findings

The new guideline combines the evaluation of surgery‐ and patient‐specific risk in the algorithm for preoperative cardiovascular evaluation into a single step and recommends the use of 1 of 3 tools: the Revised Cardiac Risk Index (RCRI),[1] National Surgical Quality Improvement Program (NSQIP) Surgical Risk Calculator,[2] or the NSQIP‐derived myocardial infarction and cardiac arrest calculator.[3] Estimation of risk is also simplified by stratification into only 2 groups: low risk (risk of major adverse cardiac event <1%) and elevated risk (1% risk). Coronary evaluation can be considered for patients with elevated cardiac risk and poor functional capacity, but is advised only if the results would alter perioperative management. For example, a patient with very high risk who has evidence of ischemia on stress testing may choose to forego surgery. Preoperative coronary revascularization is only indicated for patients meeting criteria in the nonsurgical setting.

For patients with previous percutaneous coronary intervention, the ACC/AHA has not changed its recommendations to optimally delay surgery for at least 30 days after bare‐metal stenting and at least 1 year after drug‐eluting stent (DES) placement. However, in patients with a DES placed 6 to 12 months previously, surgery can be performed if the risks of surgical delay outweigh the risks of DES thrombosis. After any type of coronary stenting, dual antiplatelet therapy should be continued uninterrupted through the first 4 to 6 weeks and even later whenever feasible. If not possible, aspirin therapy should be maintained through surgery unless bleeding risk is too high.

The guideline recommends perioperative continuation of ‐blockers in patients taking them chronically. Preoperative initiation of ‐blocker therapy may be considered for patients with myocardial ischemia on stress testing or 3 RCRI factors and should be started far enough in advance to allow determination of patient's tolerance prior to surgery.

Cautions

Many recommendations are based on data from nonrandomized trials or expert opinion, and the data in areas such as perioperative ‐blockade continue to evolve.

Implications

The ACC/AHA guideline continues to be a critically valuable resource for hospitalists providing perioperative care to noncardiac surgery patients.

Wijeysundera DN, Duncan D, Nkonde‐Price C, et al. Perioperative beta blockade in noncardiac surgery: a systematic review for the 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines.

J Am Coll Cardiol. 2014;64(22):24062425.

Background

Various clinical trials have reported conflicting results regarding the efficacy and safety of perioperative ‐blockers resulting in guideline committees changing their recommendations. Because of questions raised regarding the scientific integrity of the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography)‐I[4] and DECREASE‐IV[5] trials as well as the dosing of ‐blockers in POISE (PeriOperative Ischemic Evaluation) study,[6] this systematic review was performed in conjunction with the ACC/AHA guideline update[7] to evaluate the data with and without these trials.

Findings

Sixteen randomized control trials (RCTs) (n=12,043) and 1 cohort study (n=348) were included in the analysis. Perioperative ‐blockers were associated with a reduction in nonfatal myocardial infarction (MI) (relative risk [RR]: 0.69; 95% confidence interval [CI]: 0.58‐0.82; P<0.001) but an increase in bradycardia (RR: 2.61; 95% CI: 2.18‐3.12), hypotension (RR: 1.47; 95% CI: 1.34‐1.6), and nonfatal strokes (RR: 1.76; 95% CI: 1.07‐2.91; P=0.02). The POISE trial was the only one demonstrating a statistically significant increase in stroke.

The major discrepancy between the DECREASE trials and the other RCTs was related to mortalitya reduction in both cardiovascular and all‐cause death in DECREASE but an increased risk of all‐cause death in the other trials.

Cautions

Because of its size, the POISE trial heavily influences the results, particularly for mortality and stroke. Including the DECREASE trials reduces the otherwise increased risk for death to a null effect. Exclusion of the POISE and DECREASE trials leaves few data to make conclusions about safety and efficacy of perioperative ‐blockade. Several cohort studies have found metoprolol to be associated with worse outcomes than with atenolol or bisoprolol (which were preferred by the European Society of Cardiology guidelines).[8]

Implications

Perioperative ‐blockade started within 1 day of noncardiac surgery was associated with fewer nonfatal MIs but at the cost of an increase in hypotension, bradycardia, and a possible increase in stroke and death. Long‐term ‐blockade should be continued perioperatively, whereas the decision to initiate a ‐blocker should be individualized. If starting a ‐blocker perioperatively, it should be done 2 days before surgery.

Botto F, Alonso‐Coello P, Chan MT, et al.; on behalf of The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30‐day outcomes. Anesthesiology. 2014;120(3):564578.

Background

Many patients sustain myocardial injury in the perioperative period as evidenced by troponin elevations, but most do not meet diagnostic criteria for MI. Myocardial injury after noncardiac surgery (MINS) is defined as prognostically relevant myocardial injury due to ischemia that occurs within 30 days after noncardiac surgery. This international, prospective cohort study of 15,065 patients 45 years old who underwent in‐patient noncardiac surgery determined diagnostic criteria, characteristics, predictors, and 30‐day outcomes of MINS.

Findings

The diagnostic criterion for MINS was a peak troponin T level 0.03 ng/mL judged to be due to an ischemic etiology. Twelve independent predictors of MINS were identified including age 75 years, known cardiovascular disease or risk factors, and surgical factors. MINS was an independent predictor of 30‐day mortality (adjusted hazard ratio [HR]: 3.87; 95% CI: 2.96‐5.08). Age >75 years, ST elevation, or new left bundle branch block, and anterior ischemic findings were independent predictors of 30‐day mortality among patients with MINS.

Cautions

Although screening high‐risk surgical patients without signs or symptoms of ischemia with postoperative troponins will increase the frequency of diagnosing MINS, evidence for an effective treatment has not yet been established. The ACC/AHA guidelines state that routine screening is of uncertain benefit for this reason.

Implications

Because MINS is common and carries a poor 30‐day prognosis, clinical trials are needed to determine when to obtain postoperative troponins and how to prevent and treat this complication.[9] Some observational data from POISE suggest that aspirin and statins can reduce the risk of 30‐day mortality in patients with postoperative MIs.

Devereaux PJ, Mrkobrada M, Sessler DI, et al. for the POISE‐2 Investigators. Aspirin in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):14941503.

Devereaux PJ, Sessler DI, Leslie K, et al. for the POISE‐2 Investigators. Clonidine in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):15041513.

Background

Medical risk reduction with aspirin and other agents in perioperative patients remains controversial. The POISE‐2 trial is a blinded RCT examining the effects of aspirin and clonidine on outcomes in >10,000 noncardiac surgery patients at risk of cardiovascular complications. The aspirin arm of the study included the initiation group and the continuation stratum, as well as placebo. Patients in the clonidine portion of the trial received 0.2 mg of clonidine or placebo daily for the same time periods.

Findings

The primary outcome was a composite of death or nonfatal MI within 30 days of surgery. Outcomes were similar in patients initiated or continued on aspirin. No difference was seen between aspirin or placebo in the primary outcome (7.0% vs 7.1%; HR: 0.86; 95% CI: 0.86‐1.15; P=0.92). There were no differences in rates of MI, venous thromboembolism, or stroke. Major bleeding rates were higher in aspirin versus placebo‐treated patients (4.6% vs 3.8%; HR: 1.23; 95% CI: 1.01‐1.49; P=0.04).

Clonidine did not alter the composite outcome of death or nonfatal MI (7.3% vs 6.8%; HR: 1.08; 95% CI: 0.93‐1.26; P=0.29). Clinically significant hypotension, bradycardia, and nonfatal cardiac arrest were more common in clonidine‐treated patients, although no difference was detected in stroke rates.

Cautions

Although patients in the trial had cardiovascular risk factors, <24% of patients had known coronary artery disease, and <5% had coronary stents. Conclusions based on this trial regarding perioperative management of antiplatelet therapy should not include patients with coronary artery stents.

Implications

Aspirin started before surgery and continued perioperatively did not decrease the rate of death or nonfatal MI but increased the risk of major bleeding. Perioperative management of aspirin needs to be undertaken in the context of cardiac and bleeding risks. Clonidine also did not improve outcomes and increased the risk of bradycardia and hypotension. Current guidelines recommend against using alpha‐2 agonists for prevention of perioperative cardiac events⁷; however, patients already on alpha‐2 agonists should not stop them abruptly.

Background

An increasing body of literature associates obstructive sleep apnea (OSA) with an increased risk of postoperative complications. Despite evidence of risk, potential benefits of preoperative diagnosis and treatment of OSA remain unclear.

Findings

Using databases to identify patients prescribed continuous positive airway pressure (CPAP) therapy, the study compared postoperative outcomes of patients who underwent surgery any time after polysomnography (PSG) and CPAP prescription (diagnosed OSA [DOSA]) and those who had surgery during the 5 years preceding their PSG (undiagnosed OSA [UOSA]). These patients were matched with patients who underwent the same procedure for the same indication and had no insurance claims for PSG or diagnosis of sleep‐disordered breathing.

After multivariate analysis, OSA of any type was associated with increased pulmonary complications (odds ratio [OR]: 2.08; 95% CI: 1.35‐2.19). However, no significant differences in respiratory outcomes were noted between DOSA patients (N=2640) and those with UOSA (N=1571). DOSA patients did have fewer cardiovascular complications than UOSA patients (OR: 0.34; 95% CI: 0.15‐0.77). Only severe OSA (apnea‐hypopnea index >30) was associated with increased pulmonary and cardiovascular complications.

Cautions

Although this study suggests an association between preoperative diagnosis and treatment of OSA and reduced cardiovascular complications, the results are not definitive due to the inability to control for all confounding variables in a retrospective study utilizing an administrative database.

Implications

OSA is an important risk factor for postoperative complications, and this study suggests that preoperative treatment with CPAP is associated with reduced risk of cardiovascular complications, particularly in patients with severe OSA. Future controlled trials should focus on the risk‐reduction potential of preoperative diagnosis and treatment of OSA.

Mazo V, Sabat S, Canet J, et al. Prospective external validation of a predictive score for postoperative pulmonary complications. Anesthesiology. 2014;121:219231.

Background

In 2010, Canet et al. published a novel risk index, the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) index, to provide a quantitative estimate of the risk of postoperative pulmonary complications (PPCs).[10]

In the current report, Mazo and colleagues studied the ARISCAT index in a broader sample to characterize its accuracy in predicting PPC risk. The ARISCAT index is derived from clinical risk factors: (1) age, (2) preoperative oxygen saturation, (3) respiratory infection in the prior month, (4) anemia, (5) surgical site, (6) duration of surgery, and (7) emergency surgery, with varying weights based on the strength of the association in a multivariable analysis. This score can be calculated via addition of these weighted risk factors, with a score>45 equal to high risk for PPC.

Findings

Examining 5099 patients from 63 European hospitals, the authors definition of PPC included respiratory failure, pulmonary infection, pleural effusion, atelectasis, pneumothorax, bronchospasm, and aspiration pneumonitis. PPC rates were as follows: low risk (3.39%), intermediate risk (12.98%), and high risk (38.01%). The positive likelihood ratio for PPC among the highest risk group was 7.12. The C statistic for fit was 0.80. Observed PPC rates were higher than predicted for the low (3.39% vs 0.87%) and intermediate (12.98% vs 7.82%) risk groups.

Cautions

The calibration slopes were less than ideal in all subsamples, with the Western European sample performing better than the other geographic areas; suggesting that the coefficients on the ARISCAT index may benefit from recalibration to match specific populations.

Implications

This is the first major pulmonary risk index that has been externally validated. Its use of readily available clinical information, simplicity, and accuracy in estimating PPC risk make it an important addition to the toolkit during a preoperative evaluation.

Background

New‐onset atrial fibrillation (AF) is the most common perioperative arrhythmia.[11] However, little is known regarding the long‐term risks of ischemic stroke in patients who develop perioperative AF. This retrospective cohort study examined adults with no preexisting history of AF, hospitalized for surgery, and discharged free of cerebrovascular disease between 2007 and 2011 (n=1,729,360).

Findings

Of the eligible patients, 1.43% (95% CI: 1.41%‐1.45%) developed perioperative AF, and 0.81% (95% CI: 0.79%‐0.82%) had a stroke up to 1 year after discharge. Perioperative AF was associated with subsequent stroke after both cardiac (HR: 1.3; 95% CI: 1.1‐1.6) and noncardiac surgery (HR: 2; 95% CI: 1.7‐2.3). The association with stroke was stronger for perioperative AF after noncardiac versus cardiac surgery (P<0.001 for interaction).

Cautions

This is a retrospective cohort study, using claims data to identify AF and stroke. Data on duration of the perioperative AF episodes or use of antithrombotic therapies were not available.

Implications

The association found between perioperative AF and long‐term risk of ischemic stroke may suggest that perioperative AF, especially after noncardiac surgery, should be treated aggressively in terms of thromboembolic risk; however, further data will be required to validate this association.

Van Diepen S, Youngson E, Ezekowitz J, McAlister F. Which risk score best predicts perioperative outcomes in nonvalvular atrial fibrillation patients undergoing noncardiac surgery? Am Heart J. 2014;168(1):6067.

Background

Patients with nonvalvular AF (NVAF) are at increased risk for adverse perioperative outcomes after noncardiac surgery.[12] The RCRI is commonly used to predict perioperative cardiovascular events for all patients, including those with NVAF, though AF is not part of this risk assessment. The goal of this retrospective cohort study was to examine the prognostic utility of already existing NVAF risk indices, including the CHADS₂ (Congestive heart failure, Hypertension, Age 75 years, Diabetes mellitus, prior stroke or transient ischemic attack), CHA₂DS₂‐VASc (Congestive heart failure; Hypertension; Age 75 years; Diabetes mellitus; Stroke, TIA, or thromboembolism [TE]; Vascular disease; Age 65 to 74 years; Sex category [female]), and R₂CHADS₂ (Renal dysfunction, Congestive heart failure, Hypertension, Age, Diabetes, Stroke/TIA) for perioperative outcomes in patients undergoing noncardiac surgery.

Findings

A population dataset of NVAF patients (n=32,160) who underwent noncardiac surgery was examined, with outcome measures including 30‐day mortality, stroke, TIA, or systemic embolism. The incidence of the 30‐day composite outcome was 4.2% and the C indices were 0.65 for the RCRI, 0.67 for CHADS₂, 0.67 for CHA₂DS₂‐VASc, and 0.68 for R₂CHADS₂. The Net Reclassification Index (NRI), a measure evaluating the improvement in prediction performance gained by adding a marker to a set of baseline predictors, was calculated. All NVAF scores performed better than the RCRI for predicting mortality risk (NRI: 12.3%, 8.4%, and 13.3% respectively, all P<0.01).

Cautions

Patients in the highest risk category by RCRI appear to have an unadjusted higher 30‐day mortality risk (8%) than that predicted by the other 3 scores (5%, 5.6%, and 5%), indicating that these risk scores should not completely supplant the RCRI for risk stratification in this population. In addition, the overall improvement in predictive capacity of the CHADS₂, CHA₂DS₂‐VASc, and R₂CHADS₂, although superior to the RCRI, is modest.

Implications

These findings indicate that the preoperative risk stratification for patients with NVAF can be improved by utilizing the CHADS₂, CHA₂DS₂‐VASc, or R₂CHADS₂ scores when undergoing noncardiac surgery. For patients with NVAF identified as high risk for adverse outcomes, this assessment can be integrated into the preoperative discussion on the risks/benefits of surgery.

Steinberg BA, Peterson ED, Kim S, et al. Use and outcomes associated with bridging during anticoagulation interruptions in patients with atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT‐AF). Circulation. 2015;131:488494

Background

Oral anticoagulation (OAC) significantly reduces the risk of stroke in patients with AF. Many AF patients on long‐term anticoagulation undergo procedures requiring temporary interruption of OAC. Although guidelines have been published on when and how to initiate bridging therapy, they are based on observational data. Thus, it remains unclear which patients should receive bridging anticoagulation.

Findings

This is a US registry of outpatients with AF with temporary interruptions of OAC for a procedure. Of 7372 patients treated with OAC, 2803 overall interruption events occurred in 2200 patients (30%). Bridging anticoagulants were used in 24% (n=665). Bleeding events were more common in bridged than nonbridged patients (5.0% vs 1.3%; adjusted OR: 3.84; P<0.0001). The overall composite end point of myocardial infarction, stroke or systemic embolism, major bleeding, hospitalization, or death within 30 days was significantly higher in patients receiving bridging (13% vs 6.3%; adjusted OR: 1.94; P=0.0001). This statistically significant increase in the composite outcome, which includes cardiovascular events, is most likely in part secondary to inclusion of bleeding events. The recently published BRIDGE (Bridging Anticoagulation in Patients who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery) trial did not find a statistically significant difference in cardiovascular events between bridged and nonbridged patients.[13]

Cautions

Although patients who were bridged appear to have had more comorbidities and a higher mean CHADS₂ score than patients who were not bridged, it is difficult to determine which population of patients may be high risk enough to warrant bridging, as indicated by current American College of Chest Physicians guidelines, as this was not evaluated in this study

Implications

The use of bridging anticoagulation was significantly associated with higher overall bleeding and adverse event rates. The BRIDGE trial also found that forgoing bridging anticoagulation decreased the risk of major bleeding in patients with AF and was noninferior to bridging for the prevention of arterial TE.[13]

Background

The American College of Cardiology/American Heart Association (ACC/AHA) perioperative guideline provides recommendations for the evaluation and management of cardiovascular disease in patients undergoing noncardiac surgery.

Findings

The new guideline combines the evaluation of surgery‐ and patient‐specific risk in the algorithm for preoperative cardiovascular evaluation into a single step and recommends the use of 1 of 3 tools: the Revised Cardiac Risk Index (RCRI),[1] National Surgical Quality Improvement Program (NSQIP) Surgical Risk Calculator,[2] or the NSQIP‐derived myocardial infarction and cardiac arrest calculator.[3] Estimation of risk is also simplified by stratification into only 2 groups: low risk (risk of major adverse cardiac event <1%) and elevated risk (1% risk). Coronary evaluation can be considered for patients with elevated cardiac risk and poor functional capacity, but is advised only if the results would alter perioperative management. For example, a patient with very high risk who has evidence of ischemia on stress testing may choose to forego surgery. Preoperative coronary revascularization is only indicated for patients meeting criteria in the nonsurgical setting.

For patients with previous percutaneous coronary intervention, the ACC/AHA has not changed its recommendations to optimally delay surgery for at least 30 days after bare‐metal stenting and at least 1 year after drug‐eluting stent (DES) placement. However, in patients with a DES placed 6 to 12 months previously, surgery can be performed if the risks of surgical delay outweigh the risks of DES thrombosis. After any type of coronary stenting, dual antiplatelet therapy should be continued uninterrupted through the first 4 to 6 weeks and even later whenever feasible. If not possible, aspirin therapy should be maintained through surgery unless bleeding risk is too high.

The guideline recommends perioperative continuation of ‐blockers in patients taking them chronically. Preoperative initiation of ‐blocker therapy may be considered for patients with myocardial ischemia on stress testing or 3 RCRI factors and should be started far enough in advance to allow determination of patient's tolerance prior to surgery.

Cautions

Many recommendations are based on data from nonrandomized trials or expert opinion, and the data in areas such as perioperative ‐blockade continue to evolve.

Implications

The ACC/AHA guideline continues to be a critically valuable resource for hospitalists providing perioperative care to noncardiac surgery patients.

Wijeysundera DN, Duncan D, Nkonde‐Price C, et al. Perioperative beta blockade in noncardiac surgery: a systematic review for the 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines.

J Am Coll Cardiol. 2014;64(22):24062425.

Background

Various clinical trials have reported conflicting results regarding the efficacy and safety of perioperative ‐blockers resulting in guideline committees changing their recommendations. Because of questions raised regarding the scientific integrity of the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography)‐I[4] and DECREASE‐IV[5] trials as well as the dosing of ‐blockers in POISE (PeriOperative Ischemic Evaluation) study,[6] this systematic review was performed in conjunction with the ACC/AHA guideline update[7] to evaluate the data with and without these trials.

Findings

Sixteen randomized control trials (RCTs) (n=12,043) and 1 cohort study (n=348) were included in the analysis. Perioperative ‐blockers were associated with a reduction in nonfatal myocardial infarction (MI) (relative risk [RR]: 0.69; 95% confidence interval [CI]: 0.58‐0.82; P<0.001) but an increase in bradycardia (RR: 2.61; 95% CI: 2.18‐3.12), hypotension (RR: 1.47; 95% CI: 1.34‐1.6), and nonfatal strokes (RR: 1.76; 95% CI: 1.07‐2.91; P=0.02). The POISE trial was the only one demonstrating a statistically significant increase in stroke.

The major discrepancy between the DECREASE trials and the other RCTs was related to mortalitya reduction in both cardiovascular and all‐cause death in DECREASE but an increased risk of all‐cause death in the other trials.

Cautions

Because of its size, the POISE trial heavily influences the results, particularly for mortality and stroke. Including the DECREASE trials reduces the otherwise increased risk for death to a null effect. Exclusion of the POISE and DECREASE trials leaves few data to make conclusions about safety and efficacy of perioperative ‐blockade. Several cohort studies have found metoprolol to be associated with worse outcomes than with atenolol or bisoprolol (which were preferred by the European Society of Cardiology guidelines).[8]

Implications

Perioperative ‐blockade started within 1 day of noncardiac surgery was associated with fewer nonfatal MIs but at the cost of an increase in hypotension, bradycardia, and a possible increase in stroke and death. Long‐term ‐blockade should be continued perioperatively, whereas the decision to initiate a ‐blocker should be individualized. If starting a ‐blocker perioperatively, it should be done 2 days before surgery.

Botto F, Alonso‐Coello P, Chan MT, et al.; on behalf of The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30‐day outcomes. Anesthesiology. 2014;120(3):564578.

Background

Many patients sustain myocardial injury in the perioperative period as evidenced by troponin elevations, but most do not meet diagnostic criteria for MI. Myocardial injury after noncardiac surgery (MINS) is defined as prognostically relevant myocardial injury due to ischemia that occurs within 30 days after noncardiac surgery. This international, prospective cohort study of 15,065 patients 45 years old who underwent in‐patient noncardiac surgery determined diagnostic criteria, characteristics, predictors, and 30‐day outcomes of MINS.

Findings

The diagnostic criterion for MINS was a peak troponin T level 0.03 ng/mL judged to be due to an ischemic etiology. Twelve independent predictors of MINS were identified including age 75 years, known cardiovascular disease or risk factors, and surgical factors. MINS was an independent predictor of 30‐day mortality (adjusted hazard ratio [HR]: 3.87; 95% CI: 2.96‐5.08). Age >75 years, ST elevation, or new left bundle branch block, and anterior ischemic findings were independent predictors of 30‐day mortality among patients with MINS.

Cautions

Although screening high‐risk surgical patients without signs or symptoms of ischemia with postoperative troponins will increase the frequency of diagnosing MINS, evidence for an effective treatment has not yet been established. The ACC/AHA guidelines state that routine screening is of uncertain benefit for this reason.

Implications

Because MINS is common and carries a poor 30‐day prognosis, clinical trials are needed to determine when to obtain postoperative troponins and how to prevent and treat this complication.[9] Some observational data from POISE suggest that aspirin and statins can reduce the risk of 30‐day mortality in patients with postoperative MIs.

Devereaux PJ, Mrkobrada M, Sessler DI, et al. for the POISE‐2 Investigators. Aspirin in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):14941503.

Devereaux PJ, Sessler DI, Leslie K, et al. for the POISE‐2 Investigators. Clonidine in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):15041513.

Background

Medical risk reduction with aspirin and other agents in perioperative patients remains controversial. The POISE‐2 trial is a blinded RCT examining the effects of aspirin and clonidine on outcomes in >10,000 noncardiac surgery patients at risk of cardiovascular complications. The aspirin arm of the study included the initiation group and the continuation stratum, as well as placebo. Patients in the clonidine portion of the trial received 0.2 mg of clonidine or placebo daily for the same time periods.

Findings

The primary outcome was a composite of death or nonfatal MI within 30 days of surgery. Outcomes were similar in patients initiated or continued on aspirin. No difference was seen between aspirin or placebo in the primary outcome (7.0% vs 7.1%; HR: 0.86; 95% CI: 0.86‐1.15; P=0.92). There were no differences in rates of MI, venous thromboembolism, or stroke. Major bleeding rates were higher in aspirin versus placebo‐treated patients (4.6% vs 3.8%; HR: 1.23; 95% CI: 1.01‐1.49; P=0.04).

Clonidine did not alter the composite outcome of death or nonfatal MI (7.3% vs 6.8%; HR: 1.08; 95% CI: 0.93‐1.26; P=0.29). Clinically significant hypotension, bradycardia, and nonfatal cardiac arrest were more common in clonidine‐treated patients, although no difference was detected in stroke rates.

Cautions

Although patients in the trial had cardiovascular risk factors, <24% of patients had known coronary artery disease, and <5% had coronary stents. Conclusions based on this trial regarding perioperative management of antiplatelet therapy should not include patients with coronary artery stents.

Implications

Aspirin started before surgery and continued perioperatively did not decrease the rate of death or nonfatal MI but increased the risk of major bleeding. Perioperative management of aspirin needs to be undertaken in the context of cardiac and bleeding risks. Clonidine also did not improve outcomes and increased the risk of bradycardia and hypotension. Current guidelines recommend against using alpha‐2 agonists for prevention of perioperative cardiac events⁷; however, patients already on alpha‐2 agonists should not stop them abruptly.

Background

An increasing body of literature associates obstructive sleep apnea (OSA) with an increased risk of postoperative complications. Despite evidence of risk, potential benefits of preoperative diagnosis and treatment of OSA remain unclear.

Findings

Using databases to identify patients prescribed continuous positive airway pressure (CPAP) therapy, the study compared postoperative outcomes of patients who underwent surgery any time after polysomnography (PSG) and CPAP prescription (diagnosed OSA [DOSA]) and those who had surgery during the 5 years preceding their PSG (undiagnosed OSA [UOSA]). These patients were matched with patients who underwent the same procedure for the same indication and had no insurance claims for PSG or diagnosis of sleep‐disordered breathing.

After multivariate analysis, OSA of any type was associated with increased pulmonary complications (odds ratio [OR]: 2.08; 95% CI: 1.35‐2.19). However, no significant differences in respiratory outcomes were noted between DOSA patients (N=2640) and those with UOSA (N=1571). DOSA patients did have fewer cardiovascular complications than UOSA patients (OR: 0.34; 95% CI: 0.15‐0.77). Only severe OSA (apnea‐hypopnea index >30) was associated with increased pulmonary and cardiovascular complications.

Cautions

Although this study suggests an association between preoperative diagnosis and treatment of OSA and reduced cardiovascular complications, the results are not definitive due to the inability to control for all confounding variables in a retrospective study utilizing an administrative database.

Implications

OSA is an important risk factor for postoperative complications, and this study suggests that preoperative treatment with CPAP is associated with reduced risk of cardiovascular complications, particularly in patients with severe OSA. Future controlled trials should focus on the risk‐reduction potential of preoperative diagnosis and treatment of OSA.

Mazo V, Sabat S, Canet J, et al. Prospective external validation of a predictive score for postoperative pulmonary complications. Anesthesiology. 2014;121:219231.

Background

In 2010, Canet et al. published a novel risk index, the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) index, to provide a quantitative estimate of the risk of postoperative pulmonary complications (PPCs).[10]

In the current report, Mazo and colleagues studied the ARISCAT index in a broader sample to characterize its accuracy in predicting PPC risk. The ARISCAT index is derived from clinical risk factors: (1) age, (2) preoperative oxygen saturation, (3) respiratory infection in the prior month, (4) anemia, (5) surgical site, (6) duration of surgery, and (7) emergency surgery, with varying weights based on the strength of the association in a multivariable analysis. This score can be calculated via addition of these weighted risk factors, with a score>45 equal to high risk for PPC.

Findings

Examining 5099 patients from 63 European hospitals, the authors definition of PPC included respiratory failure, pulmonary infection, pleural effusion, atelectasis, pneumothorax, bronchospasm, and aspiration pneumonitis. PPC rates were as follows: low risk (3.39%), intermediate risk (12.98%), and high risk (38.01%). The positive likelihood ratio for PPC among the highest risk group was 7.12. The C statistic for fit was 0.80. Observed PPC rates were higher than predicted for the low (3.39% vs 0.87%) and intermediate (12.98% vs 7.82%) risk groups.

Cautions

The calibration slopes were less than ideal in all subsamples, with the Western European sample performing better than the other geographic areas; suggesting that the coefficients on the ARISCAT index may benefit from recalibration to match specific populations.

Implications

This is the first major pulmonary risk index that has been externally validated. Its use of readily available clinical information, simplicity, and accuracy in estimating PPC risk make it an important addition to the toolkit during a preoperative evaluation.

Background

New‐onset atrial fibrillation (AF) is the most common perioperative arrhythmia.[11] However, little is known regarding the long‐term risks of ischemic stroke in patients who develop perioperative AF. This retrospective cohort study examined adults with no preexisting history of AF, hospitalized for surgery, and discharged free of cerebrovascular disease between 2007 and 2011 (n=1,729,360).

Findings

Of the eligible patients, 1.43% (95% CI: 1.41%‐1.45%) developed perioperative AF, and 0.81% (95% CI: 0.79%‐0.82%) had a stroke up to 1 year after discharge. Perioperative AF was associated with subsequent stroke after both cardiac (HR: 1.3; 95% CI: 1.1‐1.6) and noncardiac surgery (HR: 2; 95% CI: 1.7‐2.3). The association with stroke was stronger for perioperative AF after noncardiac versus cardiac surgery (P<0.001 for interaction).

Cautions

This is a retrospective cohort study, using claims data to identify AF and stroke. Data on duration of the perioperative AF episodes or use of antithrombotic therapies were not available.

Implications

The association found between perioperative AF and long‐term risk of ischemic stroke may suggest that perioperative AF, especially after noncardiac surgery, should be treated aggressively in terms of thromboembolic risk; however, further data will be required to validate this association.

Van Diepen S, Youngson E, Ezekowitz J, McAlister F. Which risk score best predicts perioperative outcomes in nonvalvular atrial fibrillation patients undergoing noncardiac surgery? Am Heart J. 2014;168(1):6067.

Background

Patients with nonvalvular AF (NVAF) are at increased risk for adverse perioperative outcomes after noncardiac surgery.[12] The RCRI is commonly used to predict perioperative cardiovascular events for all patients, including those with NVAF, though AF is not part of this risk assessment. The goal of this retrospective cohort study was to examine the prognostic utility of already existing NVAF risk indices, including the CHADS₂ (Congestive heart failure, Hypertension, Age 75 years, Diabetes mellitus, prior stroke or transient ischemic attack), CHA₂DS₂‐VASc (Congestive heart failure; Hypertension; Age 75 years; Diabetes mellitus; Stroke, TIA, or thromboembolism [TE]; Vascular disease; Age 65 to 74 years; Sex category [female]), and R₂CHADS₂ (Renal dysfunction, Congestive heart failure, Hypertension, Age, Diabetes, Stroke/TIA) for perioperative outcomes in patients undergoing noncardiac surgery.

Findings

A population dataset of NVAF patients (n=32,160) who underwent noncardiac surgery was examined, with outcome measures including 30‐day mortality, stroke, TIA, or systemic embolism. The incidence of the 30‐day composite outcome was 4.2% and the C indices were 0.65 for the RCRI, 0.67 for CHADS₂, 0.67 for CHA₂DS₂‐VASc, and 0.68 for R₂CHADS₂. The Net Reclassification Index (NRI), a measure evaluating the improvement in prediction performance gained by adding a marker to a set of baseline predictors, was calculated. All NVAF scores performed better than the RCRI for predicting mortality risk (NRI: 12.3%, 8.4%, and 13.3% respectively, all P<0.01).

Cautions

Patients in the highest risk category by RCRI appear to have an unadjusted higher 30‐day mortality risk (8%) than that predicted by the other 3 scores (5%, 5.6%, and 5%), indicating that these risk scores should not completely supplant the RCRI for risk stratification in this population. In addition, the overall improvement in predictive capacity of the CHADS₂, CHA₂DS₂‐VASc, and R₂CHADS₂, although superior to the RCRI, is modest.

Implications

These findings indicate that the preoperative risk stratification for patients with NVAF can be improved by utilizing the CHADS₂, CHA₂DS₂‐VASc, or R₂CHADS₂ scores when undergoing noncardiac surgery. For patients with NVAF identified as high risk for adverse outcomes, this assessment can be integrated into the preoperative discussion on the risks/benefits of surgery.

Steinberg BA, Peterson ED, Kim S, et al. Use and outcomes associated with bridging during anticoagulation interruptions in patients with atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT‐AF). Circulation. 2015;131:488494

Background

Oral anticoagulation (OAC) significantly reduces the risk of stroke in patients with AF. Many AF patients on long‐term anticoagulation undergo procedures requiring temporary interruption of OAC. Although guidelines have been published on when and how to initiate bridging therapy, they are based on observational data. Thus, it remains unclear which patients should receive bridging anticoagulation.

Findings

This is a US registry of outpatients with AF with temporary interruptions of OAC for a procedure. Of 7372 patients treated with OAC, 2803 overall interruption events occurred in 2200 patients (30%). Bridging anticoagulants were used in 24% (n=665). Bleeding events were more common in bridged than nonbridged patients (5.0% vs 1.3%; adjusted OR: 3.84; P<0.0001). The overall composite end point of myocardial infarction, stroke or systemic embolism, major bleeding, hospitalization, or death within 30 days was significantly higher in patients receiving bridging (13% vs 6.3%; adjusted OR: 1.94; P=0.0001). This statistically significant increase in the composite outcome, which includes cardiovascular events, is most likely in part secondary to inclusion of bleeding events. The recently published BRIDGE (Bridging Anticoagulation in Patients who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery) trial did not find a statistically significant difference in cardiovascular events between bridged and nonbridged patients.[13]

Cautions

Although patients who were bridged appear to have had more comorbidities and a higher mean CHADS₂ score than patients who were not bridged, it is difficult to determine which population of patients may be high risk enough to warrant bridging, as indicated by current American College of Chest Physicians guidelines, as this was not evaluated in this study

Implications

The use of bridging anticoagulation was significantly associated with higher overall bleeding and adverse event rates. The BRIDGE trial also found that forgoing bridging anticoagulation decreased the risk of major bleeding in patients with AF and was noninferior to bridging for the prevention of arterial TE.[13]

Background

The American College of Cardiology/American Heart Association (ACC/AHA) perioperative guideline provides recommendations for the evaluation and management of cardiovascular disease in patients undergoing noncardiac surgery.

Findings

The new guideline combines the evaluation of surgery‐ and patient‐specific risk in the algorithm for preoperative cardiovascular evaluation into a single step and recommends the use of 1 of 3 tools: the Revised Cardiac Risk Index (RCRI),[1] National Surgical Quality Improvement Program (NSQIP) Surgical Risk Calculator,[2] or the NSQIP‐derived myocardial infarction and cardiac arrest calculator.[3] Estimation of risk is also simplified by stratification into only 2 groups: low risk (risk of major adverse cardiac event <1%) and elevated risk (1% risk). Coronary evaluation can be considered for patients with elevated cardiac risk and poor functional capacity, but is advised only if the results would alter perioperative management. For example, a patient with very high risk who has evidence of ischemia on stress testing may choose to forego surgery. Preoperative coronary revascularization is only indicated for patients meeting criteria in the nonsurgical setting.

For patients with previous percutaneous coronary intervention, the ACC/AHA has not changed its recommendations to optimally delay surgery for at least 30 days after bare‐metal stenting and at least 1 year after drug‐eluting stent (DES) placement. However, in patients with a DES placed 6 to 12 months previously, surgery can be performed if the risks of surgical delay outweigh the risks of DES thrombosis. After any type of coronary stenting, dual antiplatelet therapy should be continued uninterrupted through the first 4 to 6 weeks and even later whenever feasible. If not possible, aspirin therapy should be maintained through surgery unless bleeding risk is too high.

The guideline recommends perioperative continuation of ‐blockers in patients taking them chronically. Preoperative initiation of ‐blocker therapy may be considered for patients with myocardial ischemia on stress testing or 3 RCRI factors and should be started far enough in advance to allow determination of patient's tolerance prior to surgery.

Cautions

Many recommendations are based on data from nonrandomized trials or expert opinion, and the data in areas such as perioperative ‐blockade continue to evolve.

Implications

The ACC/AHA guideline continues to be a critically valuable resource for hospitalists providing perioperative care to noncardiac surgery patients.

Wijeysundera DN, Duncan D, Nkonde‐Price C, et al. Perioperative beta blockade in noncardiac surgery: a systematic review for the 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines.

J Am Coll Cardiol. 2014;64(22):24062425.

Background

Various clinical trials have reported conflicting results regarding the efficacy and safety of perioperative ‐blockers resulting in guideline committees changing their recommendations. Because of questions raised regarding the scientific integrity of the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography)‐I[4] and DECREASE‐IV[5] trials as well as the dosing of ‐blockers in POISE (PeriOperative Ischemic Evaluation) study,[6] this systematic review was performed in conjunction with the ACC/AHA guideline update[7] to evaluate the data with and without these trials.

Findings

Sixteen randomized control trials (RCTs) (n=12,043) and 1 cohort study (n=348) were included in the analysis. Perioperative ‐blockers were associated with a reduction in nonfatal myocardial infarction (MI) (relative risk [RR]: 0.69; 95% confidence interval [CI]: 0.58‐0.82; P<0.001) but an increase in bradycardia (RR: 2.61; 95% CI: 2.18‐3.12), hypotension (RR: 1.47; 95% CI: 1.34‐1.6), and nonfatal strokes (RR: 1.76; 95% CI: 1.07‐2.91; P=0.02). The POISE trial was the only one demonstrating a statistically significant increase in stroke.

The major discrepancy between the DECREASE trials and the other RCTs was related to mortalitya reduction in both cardiovascular and all‐cause death in DECREASE but an increased risk of all‐cause death in the other trials.

Cautions

Because of its size, the POISE trial heavily influences the results, particularly for mortality and stroke. Including the DECREASE trials reduces the otherwise increased risk for death to a null effect. Exclusion of the POISE and DECREASE trials leaves few data to make conclusions about safety and efficacy of perioperative ‐blockade. Several cohort studies have found metoprolol to be associated with worse outcomes than with atenolol or bisoprolol (which were preferred by the European Society of Cardiology guidelines).[8]

Implications

Perioperative ‐blockade started within 1 day of noncardiac surgery was associated with fewer nonfatal MIs but at the cost of an increase in hypotension, bradycardia, and a possible increase in stroke and death. Long‐term ‐blockade should be continued perioperatively, whereas the decision to initiate a ‐blocker should be individualized. If starting a ‐blocker perioperatively, it should be done 2 days before surgery.

Botto F, Alonso‐Coello P, Chan MT, et al.; on behalf of The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30‐day outcomes. Anesthesiology. 2014;120(3):564578.

Background

Many patients sustain myocardial injury in the perioperative period as evidenced by troponin elevations, but most do not meet diagnostic criteria for MI. Myocardial injury after noncardiac surgery (MINS) is defined as prognostically relevant myocardial injury due to ischemia that occurs within 30 days after noncardiac surgery. This international, prospective cohort study of 15,065 patients 45 years old who underwent in‐patient noncardiac surgery determined diagnostic criteria, characteristics, predictors, and 30‐day outcomes of MINS.

Findings

The diagnostic criterion for MINS was a peak troponin T level 0.03 ng/mL judged to be due to an ischemic etiology. Twelve independent predictors of MINS were identified including age 75 years, known cardiovascular disease or risk factors, and surgical factors. MINS was an independent predictor of 30‐day mortality (adjusted hazard ratio [HR]: 3.87; 95% CI: 2.96‐5.08). Age >75 years, ST elevation, or new left bundle branch block, and anterior ischemic findings were independent predictors of 30‐day mortality among patients with MINS.

Cautions

Although screening high‐risk surgical patients without signs or symptoms of ischemia with postoperative troponins will increase the frequency of diagnosing MINS, evidence for an effective treatment has not yet been established. The ACC/AHA guidelines state that routine screening is of uncertain benefit for this reason.

Implications

Because MINS is common and carries a poor 30‐day prognosis, clinical trials are needed to determine when to obtain postoperative troponins and how to prevent and treat this complication.[9] Some observational data from POISE suggest that aspirin and statins can reduce the risk of 30‐day mortality in patients with postoperative MIs.

Devereaux PJ, Mrkobrada M, Sessler DI, et al. for the POISE‐2 Investigators. Aspirin in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):14941503.

Devereaux PJ, Sessler DI, Leslie K, et al. for the POISE‐2 Investigators. Clonidine in patients undergoing noncardiac surgery. N Engl J Med. 2014; 370(16):15041513.

Background

Medical risk reduction with aspirin and other agents in perioperative patients remains controversial. The POISE‐2 trial is a blinded RCT examining the effects of aspirin and clonidine on outcomes in >10,000 noncardiac surgery patients at risk of cardiovascular complications. The aspirin arm of the study included the initiation group and the continuation stratum, as well as placebo. Patients in the clonidine portion of the trial received 0.2 mg of clonidine or placebo daily for the same time periods.

Findings

The primary outcome was a composite of death or nonfatal MI within 30 days of surgery. Outcomes were similar in patients initiated or continued on aspirin. No difference was seen between aspirin or placebo in the primary outcome (7.0% vs 7.1%; HR: 0.86; 95% CI: 0.86‐1.15; P=0.92). There were no differences in rates of MI, venous thromboembolism, or stroke. Major bleeding rates were higher in aspirin versus placebo‐treated patients (4.6% vs 3.8%; HR: 1.23; 95% CI: 1.01‐1.49; P=0.04).

Clonidine did not alter the composite outcome of death or nonfatal MI (7.3% vs 6.8%; HR: 1.08; 95% CI: 0.93‐1.26; P=0.29). Clinically significant hypotension, bradycardia, and nonfatal cardiac arrest were more common in clonidine‐treated patients, although no difference was detected in stroke rates.

Cautions

Although patients in the trial had cardiovascular risk factors, <24% of patients had known coronary artery disease, and <5% had coronary stents. Conclusions based on this trial regarding perioperative management of antiplatelet therapy should not include patients with coronary artery stents.

Implications

Aspirin started before surgery and continued perioperatively did not decrease the rate of death or nonfatal MI but increased the risk of major bleeding. Perioperative management of aspirin needs to be undertaken in the context of cardiac and bleeding risks. Clonidine also did not improve outcomes and increased the risk of bradycardia and hypotension. Current guidelines recommend against using alpha‐2 agonists for prevention of perioperative cardiac events⁷; however, patients already on alpha‐2 agonists should not stop them abruptly.

Background

An increasing body of literature associates obstructive sleep apnea (OSA) with an increased risk of postoperative complications. Despite evidence of risk, potential benefits of preoperative diagnosis and treatment of OSA remain unclear.

Findings

Using databases to identify patients prescribed continuous positive airway pressure (CPAP) therapy, the study compared postoperative outcomes of patients who underwent surgery any time after polysomnography (PSG) and CPAP prescription (diagnosed OSA [DOSA]) and those who had surgery during the 5 years preceding their PSG (undiagnosed OSA [UOSA]). These patients were matched with patients who underwent the same procedure for the same indication and had no insurance claims for PSG or diagnosis of sleep‐disordered breathing.

After multivariate analysis, OSA of any type was associated with increased pulmonary complications (odds ratio [OR]: 2.08; 95% CI: 1.35‐2.19). However, no significant differences in respiratory outcomes were noted between DOSA patients (N=2640) and those with UOSA (N=1571). DOSA patients did have fewer cardiovascular complications than UOSA patients (OR: 0.34; 95% CI: 0.15‐0.77). Only severe OSA (apnea‐hypopnea index >30) was associated with increased pulmonary and cardiovascular complications.

Cautions

Although this study suggests an association between preoperative diagnosis and treatment of OSA and reduced cardiovascular complications, the results are not definitive due to the inability to control for all confounding variables in a retrospective study utilizing an administrative database.

Implications

OSA is an important risk factor for postoperative complications, and this study suggests that preoperative treatment with CPAP is associated with reduced risk of cardiovascular complications, particularly in patients with severe OSA. Future controlled trials should focus on the risk‐reduction potential of preoperative diagnosis and treatment of OSA.

Mazo V, Sabat S, Canet J, et al. Prospective external validation of a predictive score for postoperative pulmonary complications. Anesthesiology. 2014;121:219231.

Background

In 2010, Canet et al. published a novel risk index, the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) index, to provide a quantitative estimate of the risk of postoperative pulmonary complications (PPCs).[10]

In the current report, Mazo and colleagues studied the ARISCAT index in a broader sample to characterize its accuracy in predicting PPC risk. The ARISCAT index is derived from clinical risk factors: (1) age, (2) preoperative oxygen saturation, (3) respiratory infection in the prior month, (4) anemia, (5) surgical site, (6) duration of surgery, and (7) emergency surgery, with varying weights based on the strength of the association in a multivariable analysis. This score can be calculated via addition of these weighted risk factors, with a score>45 equal to high risk for PPC.

Findings

Examining 5099 patients from 63 European hospitals, the authors definition of PPC included respiratory failure, pulmonary infection, pleural effusion, atelectasis, pneumothorax, bronchospasm, and aspiration pneumonitis. PPC rates were as follows: low risk (3.39%), intermediate risk (12.98%), and high risk (38.01%). The positive likelihood ratio for PPC among the highest risk group was 7.12. The C statistic for fit was 0.80. Observed PPC rates were higher than predicted for the low (3.39% vs 0.87%) and intermediate (12.98% vs 7.82%) risk groups.

Cautions

The calibration slopes were less than ideal in all subsamples, with the Western European sample performing better than the other geographic areas; suggesting that the coefficients on the ARISCAT index may benefit from recalibration to match specific populations.

Implications

This is the first major pulmonary risk index that has been externally validated. Its use of readily available clinical information, simplicity, and accuracy in estimating PPC risk make it an important addition to the toolkit during a preoperative evaluation.

Background

New‐onset atrial fibrillation (AF) is the most common perioperative arrhythmia.[11] However, little is known regarding the long‐term risks of ischemic stroke in patients who develop perioperative AF. This retrospective cohort study examined adults with no preexisting history of AF, hospitalized for surgery, and discharged free of cerebrovascular disease between 2007 and 2011 (n=1,729,360).

Findings

Of the eligible patients, 1.43% (95% CI: 1.41%‐1.45%) developed perioperative AF, and 0.81% (95% CI: 0.79%‐0.82%) had a stroke up to 1 year after discharge. Perioperative AF was associated with subsequent stroke after both cardiac (HR: 1.3; 95% CI: 1.1‐1.6) and noncardiac surgery (HR: 2; 95% CI: 1.7‐2.3). The association with stroke was stronger for perioperative AF after noncardiac versus cardiac surgery (P<0.001 for interaction).

Cautions

This is a retrospective cohort study, using claims data to identify AF and stroke. Data on duration of the perioperative AF episodes or use of antithrombotic therapies were not available.

Implications

The association found between perioperative AF and long‐term risk of ischemic stroke may suggest that perioperative AF, especially after noncardiac surgery, should be treated aggressively in terms of thromboembolic risk; however, further data will be required to validate this association.

Van Diepen S, Youngson E, Ezekowitz J, McAlister F. Which risk score best predicts perioperative outcomes in nonvalvular atrial fibrillation patients undergoing noncardiac surgery? Am Heart J. 2014;168(1):6067.

Background

Patients with nonvalvular AF (NVAF) are at increased risk for adverse perioperative outcomes after noncardiac surgery.[12] The RCRI is commonly used to predict perioperative cardiovascular events for all patients, including those with NVAF, though AF is not part of this risk assessment. The goal of this retrospective cohort study was to examine the prognostic utility of already existing NVAF risk indices, including the CHADS₂ (Congestive heart failure, Hypertension, Age 75 years, Diabetes mellitus, prior stroke or transient ischemic attack), CHA₂DS₂‐VASc (Congestive heart failure; Hypertension; Age 75 years; Diabetes mellitus; Stroke, TIA, or thromboembolism [TE]; Vascular disease; Age 65 to 74 years; Sex category [female]), and R₂CHADS₂ (Renal dysfunction, Congestive heart failure, Hypertension, Age, Diabetes, Stroke/TIA) for perioperative outcomes in patients undergoing noncardiac surgery.

Findings

A population dataset of NVAF patients (n=32,160) who underwent noncardiac surgery was examined, with outcome measures including 30‐day mortality, stroke, TIA, or systemic embolism. The incidence of the 30‐day composite outcome was 4.2% and the C indices were 0.65 for the RCRI, 0.67 for CHADS₂, 0.67 for CHA₂DS₂‐VASc, and 0.68 for R₂CHADS₂. The Net Reclassification Index (NRI), a measure evaluating the improvement in prediction performance gained by adding a marker to a set of baseline predictors, was calculated. All NVAF scores performed better than the RCRI for predicting mortality risk (NRI: 12.3%, 8.4%, and 13.3% respectively, all P<0.01).

Cautions

Patients in the highest risk category by RCRI appear to have an unadjusted higher 30‐day mortality risk (8%) than that predicted by the other 3 scores (5%, 5.6%, and 5%), indicating that these risk scores should not completely supplant the RCRI for risk stratification in this population. In addition, the overall improvement in predictive capacity of the CHADS₂, CHA₂DS₂‐VASc, and R₂CHADS₂, although superior to the RCRI, is modest.

Implications

These findings indicate that the preoperative risk stratification for patients with NVAF can be improved by utilizing the CHADS₂, CHA₂DS₂‐VASc, or R₂CHADS₂ scores when undergoing noncardiac surgery. For patients with NVAF identified as high risk for adverse outcomes, this assessment can be integrated into the preoperative discussion on the risks/benefits of surgery.

Steinberg BA, Peterson ED, Kim S, et al. Use and outcomes associated with bridging during anticoagulation interruptions in patients with atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT‐AF). Circulation. 2015;131:488494

Background

Oral anticoagulation (OAC) significantly reduces the risk of stroke in patients with AF. Many AF patients on long‐term anticoagulation undergo procedures requiring temporary interruption of OAC. Although guidelines have been published on when and how to initiate bridging therapy, they are based on observational data. Thus, it remains unclear which patients should receive bridging anticoagulation.

Findings

This is a US registry of outpatients with AF with temporary interruptions of OAC for a procedure. Of 7372 patients treated with OAC, 2803 overall interruption events occurred in 2200 patients (30%). Bridging anticoagulants were used in 24% (n=665). Bleeding events were more common in bridged than nonbridged patients (5.0% vs 1.3%; adjusted OR: 3.84; P<0.0001). The overall composite end point of myocardial infarction, stroke or systemic embolism, major bleeding, hospitalization, or death within 30 days was significantly higher in patients receiving bridging (13% vs 6.3%; adjusted OR: 1.94; P=0.0001). This statistically significant increase in the composite outcome, which includes cardiovascular events, is most likely in part secondary to inclusion of bleeding events. The recently published BRIDGE (Bridging Anticoagulation in Patients who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery) trial did not find a statistically significant difference in cardiovascular events between bridged and nonbridged patients.[13]

Cautions

Although patients who were bridged appear to have had more comorbidities and a higher mean CHADS₂ score than patients who were not bridged, it is difficult to determine which population of patients may be high risk enough to warrant bridging, as indicated by current American College of Chest Physicians guidelines, as this was not evaluated in this study

Implications

The use of bridging anticoagulation was significantly associated with higher overall bleeding and adverse event rates. The BRIDGE trial also found that forgoing bridging anticoagulation decreased the risk of major bleeding in patients with AF and was noninferior to bridging for the prevention of arterial TE.[13]

Study Design and Patients

This was a before‐and‐after study designed to assess whether implementation of a VTE prophylaxis quality improvement intervention could improve the rate of appropriate thromboprophylaxis in hospitalized medical patients at the State University of New York, Downstate Medical CenterUniversity Hospital of Brooklyn, an urban university teaching hospital of approximately 400 beds. This initiative, conducted as part of a departmental quality assurance and performance improvement program, did not require institutional review board approval. After an informal survey revealed a prophylaxis rate of approximately 50%, a more formal baseline assessment of the rate of medical patients receiving VTE prophylaxis was conducted during October 2002. This assessment was a single sampling of all medical inpatients on 2 of the medical floors on a single day. The results were consistent with those of the informal survey as well as those from an international registry.19 The results from the baseline study indicated that VTE prophylaxis was underused: only 46.9% of our medical inpatients received any form of prophylaxis. The prophylaxis rate was assessed again in 2 sampling periods beginning 12 and 18 months after implementation of the intervention. Data were collected monthly and combined into 3‐month blocks. The first postintervention sample (n = 116 patient charts) was drawn from a period 12‐14 months after implementation and the second (n = 147 patient charts) from a period 18‐20 months after implementation.

On a randomly designated day in the latter half of each month during each sampling period, all charts on 2 primary medical floors were reviewed and included in the retrospective analysis. Patients who were not on the medical service were excluded from analysis. Patients, as well as their medications, were identified using a list generated from our pharmacy database. We chose this method and schedule for several reasons. First, we sought to reduce the likelihood of including a patient more than once in a monthly sample. Second, by waiting for the latter half of the month we sought to allow house staff a chance to acquire knowledge from the educational program introduced on the first day of the month. Third, we wanted to allow house staff the time to actualize new attitudes reinforced by the audit‐and‐feedback element. The house staff included approximately 4 interns and 4 residents each month plus 10‐15 attendings or hospitalists.

Data Collection

For each sampling period we conducted a medical record (paper) review, and the Division Chief of General Internal Medicine also interviewed the medical house staff and attending physicians. Data collected included risk factors for VTE, contraindications to anticoagulant prophylaxis, type of VTE prophylaxis received, and appropriateness of the prophylaxis. Prophylaxis was considered appropriate when it was given in accordance with a risk stratification scheme (Table 1) adapted from the 2001 ACCP guideline recommendations for surgical patients20 and modified for medical inpatients, similar to the risk assessment model by Caprini et al.21 Prophylaxis was also considered appropriate when no prophylaxis was given for low‐risk patients or when full anticoagulation was given for another indication (Table 1). Questionable prophylaxis was defined as UFH given every 12 hours to a high‐risk patient. All other prophylaxis was deemed inappropriate (including no prophylaxis if prophylaxis was indicated, use of enoxaparin at incorrect prophylactic doses such as 60 or 20 mg, IPC alone for a high‐risk patient with no contraindication to pharmacological prophylaxis, and the use of warfarin if no other indication for it). The risk factors for thromboembolism and contraindications to anticoagulant prophylaxis are given in Table 2. Non‐ambulatory was defined as an order for bed rest with or without bathroom privileges or was judged based on information obtained from the medical house staff and nurses about whether the patient was ambulatory or had been observed walking outside his or her room. Data on pharmacological prophylaxis were obtained from the hospital pharmacy. Information on use of mechanical prophylaxis was obtained by house staff interviews or review of the order sheet. The house officer or attending physician of each patient was interviewed retrospectively to determine the reason for admission and the risk factors for VTE present on admission. Patients were classified as having low, moderate, high, or highest risk for VTE based on their age and any major risk factors for VTE (Table 1).19 All collected data were reported to the Department of Medicine Performance Improvement Committee for independent corroboration.

Intervention Strategies

The intervention introduced comprised 3 strategies designed to improve VTE prophylaxis: provider education, decision support, and audit‐and‐feedback.

Statistical Analysis

Differences in pre‐ and post‐intervention VTE prophylaxis and appropriate VTE prophylaxis rates were analyzed using the chi‐square test for categorical variables and the one‐way analysis of variance test for continuous variables. Differences were considered significant at the 5% level (P = .05).

Patients and Demographics

From October 2002 to August 2004 data were collected from 312 hospitalized medical patients: 49 patients in the baseline group during October 2002, and 116 and 147 at the 12‐ to 14‐month and 18‐ to 20‐month time points, respectively. Thus, approximately 40‐50 patients were randomly selected each month, representing 40% of the general medical service census. Patient demographics were similar between groups (Table 4). Overall, most patients were female (65.7%), and mean age was 61.2 years. The most common admission diagnoses were infection/sepsis (29.5%), chest pain/acute coronary syndromes/myocardial infarction (15.7%), heart failure (10.9%), and malignancy (9.6%). Overall, 7.1% (22 patients) had a contraindication to anticoagulant prophylaxis. The most common contraindication was active gastrointestinal bleeding on the current admission, which occurred in 18 of these patients.

Risk Factors for VTE

Patient risk factors for VTE in each data collection period are summarized in Table 4. Analysis of this data showed that the most prevalent risk factors for VTE in the 3 patient populations were age older than 40 years (262/312, 84.0% of the total patient population) and nonambulatory state (231/271, 85.2% of the total population). Overall, the average number of risk factors for VTE was approximately 3, with more than 60% of patients having 3 or more VTE risk factors (Fig. 1).

Figure 1

Prophylaxis Use

The types of VTE prophylaxis used and the proportion of patients treated appropriately are summarized for each data collection period in Tables 5 and 6, respectively. In all 3 populations, most patients received pharmacological rather than mechanical prophylaxis, most commonly UFH. At baseline, the prophylaxis decision was appropriate (in accordance with the recommendations of the ACCP guidelines) as often as it was inappropriate (42.9% of patients). The prophylaxis decision was questionable in the remaining 14.3% of patients.

Change in Prophylaxis Use

Twelve and 18 months after implementation of the quality improvement program, we observed an increase in the use of any prophylaxis, from 46.9% at baseline to 86.2% and 86.4%, respectively (Table 5; P < .01 in both groups versus baseline). This increase was a result almost entirely of an increase in the proportion of patients receiving pharmacological prophylaxis, which significantly increased, from 44.9% to 81.0% and 80.3%, at the 12‐ and 18‐month time points, respectively (Table 5; P < .01 for both groups versus baseline). Most meaningfully, there was a significant increase in the proportion of patients for whom an appropriate prophylaxis decision was made (from 42.9% to 68.1% and 85.0%, at the 12‐ and 18‐month time points, respectively; Table 6; P < .01 for both groups versus baseline). This represented a trend toward continuing increases in the use of appropriate prophylaxis as the study progressed (Fig. 2). This change was driven mainly by a significant increase in the prescribing of LMWH, almost all of which was prescribed in accordance with the 2001 ACCP guidelines (Table 6).

Figure 2

Study Design and Patients

This was a before‐and‐after study designed to assess whether implementation of a VTE prophylaxis quality improvement intervention could improve the rate of appropriate thromboprophylaxis in hospitalized medical patients at the State University of New York, Downstate Medical CenterUniversity Hospital of Brooklyn, an urban university teaching hospital of approximately 400 beds. This initiative, conducted as part of a departmental quality assurance and performance improvement program, did not require institutional review board approval. After an informal survey revealed a prophylaxis rate of approximately 50%, a more formal baseline assessment of the rate of medical patients receiving VTE prophylaxis was conducted during October 2002. This assessment was a single sampling of all medical inpatients on 2 of the medical floors on a single day. The results were consistent with those of the informal survey as well as those from an international registry.19 The results from the baseline study indicated that VTE prophylaxis was underused: only 46.9% of our medical inpatients received any form of prophylaxis. The prophylaxis rate was assessed again in 2 sampling periods beginning 12 and 18 months after implementation of the intervention. Data were collected monthly and combined into 3‐month blocks. The first postintervention sample (n = 116 patient charts) was drawn from a period 12‐14 months after implementation and the second (n = 147 patient charts) from a period 18‐20 months after implementation.

On a randomly designated day in the latter half of each month during each sampling period, all charts on 2 primary medical floors were reviewed and included in the retrospective analysis. Patients who were not on the medical service were excluded from analysis. Patients, as well as their medications, were identified using a list generated from our pharmacy database. We chose this method and schedule for several reasons. First, we sought to reduce the likelihood of including a patient more than once in a monthly sample. Second, by waiting for the latter half of the month we sought to allow house staff a chance to acquire knowledge from the educational program introduced on the first day of the month. Third, we wanted to allow house staff the time to actualize new attitudes reinforced by the audit‐and‐feedback element. The house staff included approximately 4 interns and 4 residents each month plus 10‐15 attendings or hospitalists.

Data Collection

For each sampling period we conducted a medical record (paper) review, and the Division Chief of General Internal Medicine also interviewed the medical house staff and attending physicians. Data collected included risk factors for VTE, contraindications to anticoagulant prophylaxis, type of VTE prophylaxis received, and appropriateness of the prophylaxis. Prophylaxis was considered appropriate when it was given in accordance with a risk stratification scheme (Table 1) adapted from the 2001 ACCP guideline recommendations for surgical patients20 and modified for medical inpatients, similar to the risk assessment model by Caprini et al.21 Prophylaxis was also considered appropriate when no prophylaxis was given for low‐risk patients or when full anticoagulation was given for another indication (Table 1). Questionable prophylaxis was defined as UFH given every 12 hours to a high‐risk patient. All other prophylaxis was deemed inappropriate (including no prophylaxis if prophylaxis was indicated, use of enoxaparin at incorrect prophylactic doses such as 60 or 20 mg, IPC alone for a high‐risk patient with no contraindication to pharmacological prophylaxis, and the use of warfarin if no other indication for it). The risk factors for thromboembolism and contraindications to anticoagulant prophylaxis are given in Table 2. Non‐ambulatory was defined as an order for bed rest with or without bathroom privileges or was judged based on information obtained from the medical house staff and nurses about whether the patient was ambulatory or had been observed walking outside his or her room. Data on pharmacological prophylaxis were obtained from the hospital pharmacy. Information on use of mechanical prophylaxis was obtained by house staff interviews or review of the order sheet. The house officer or attending physician of each patient was interviewed retrospectively to determine the reason for admission and the risk factors for VTE present on admission. Patients were classified as having low, moderate, high, or highest risk for VTE based on their age and any major risk factors for VTE (Table 1).19 All collected data were reported to the Department of Medicine Performance Improvement Committee for independent corroboration.

Intervention Strategies

The intervention introduced comprised 3 strategies designed to improve VTE prophylaxis: provider education, decision support, and audit‐and‐feedback.

Statistical Analysis

Differences in pre‐ and post‐intervention VTE prophylaxis and appropriate VTE prophylaxis rates were analyzed using the chi‐square test for categorical variables and the one‐way analysis of variance test for continuous variables. Differences were considered significant at the 5% level (P = .05).

Patients and Demographics

From October 2002 to August 2004 data were collected from 312 hospitalized medical patients: 49 patients in the baseline group during October 2002, and 116 and 147 at the 12‐ to 14‐month and 18‐ to 20‐month time points, respectively. Thus, approximately 40‐50 patients were randomly selected each month, representing 40% of the general medical service census. Patient demographics were similar between groups (Table 4). Overall, most patients were female (65.7%), and mean age was 61.2 years. The most common admission diagnoses were infection/sepsis (29.5%), chest pain/acute coronary syndromes/myocardial infarction (15.7%), heart failure (10.9%), and malignancy (9.6%). Overall, 7.1% (22 patients) had a contraindication to anticoagulant prophylaxis. The most common contraindication was active gastrointestinal bleeding on the current admission, which occurred in 18 of these patients.

Risk Factors for VTE

Patient risk factors for VTE in each data collection period are summarized in Table 4. Analysis of this data showed that the most prevalent risk factors for VTE in the 3 patient populations were age older than 40 years (262/312, 84.0% of the total patient population) and nonambulatory state (231/271, 85.2% of the total population). Overall, the average number of risk factors for VTE was approximately 3, with more than 60% of patients having 3 or more VTE risk factors (Fig. 1).

Figure 1

Prophylaxis Use

The types of VTE prophylaxis used and the proportion of patients treated appropriately are summarized for each data collection period in Tables 5 and 6, respectively. In all 3 populations, most patients received pharmacological rather than mechanical prophylaxis, most commonly UFH. At baseline, the prophylaxis decision was appropriate (in accordance with the recommendations of the ACCP guidelines) as often as it was inappropriate (42.9% of patients). The prophylaxis decision was questionable in the remaining 14.3% of patients.

Change in Prophylaxis Use

Twelve and 18 months after implementation of the quality improvement program, we observed an increase in the use of any prophylaxis, from 46.9% at baseline to 86.2% and 86.4%, respectively (Table 5; P < .01 in both groups versus baseline). This increase was a result almost entirely of an increase in the proportion of patients receiving pharmacological prophylaxis, which significantly increased, from 44.9% to 81.0% and 80.3%, at the 12‐ and 18‐month time points, respectively (Table 5; P < .01 for both groups versus baseline). Most meaningfully, there was a significant increase in the proportion of patients for whom an appropriate prophylaxis decision was made (from 42.9% to 68.1% and 85.0%, at the 12‐ and 18‐month time points, respectively; Table 6; P < .01 for both groups versus baseline). This represented a trend toward continuing increases in the use of appropriate prophylaxis as the study progressed (Fig. 2). This change was driven mainly by a significant increase in the prescribing of LMWH, almost all of which was prescribed in accordance with the 2001 ACCP guidelines (Table 6).

Figure 2

Study Design and Patients

This was a before‐and‐after study designed to assess whether implementation of a VTE prophylaxis quality improvement intervention could improve the rate of appropriate thromboprophylaxis in hospitalized medical patients at the State University of New York, Downstate Medical CenterUniversity Hospital of Brooklyn, an urban university teaching hospital of approximately 400 beds. This initiative, conducted as part of a departmental quality assurance and performance improvement program, did not require institutional review board approval. After an informal survey revealed a prophylaxis rate of approximately 50%, a more formal baseline assessment of the rate of medical patients receiving VTE prophylaxis was conducted during October 2002. This assessment was a single sampling of all medical inpatients on 2 of the medical floors on a single day. The results were consistent with those of the informal survey as well as those from an international registry.19 The results from the baseline study indicated that VTE prophylaxis was underused: only 46.9% of our medical inpatients received any form of prophylaxis. The prophylaxis rate was assessed again in 2 sampling periods beginning 12 and 18 months after implementation of the intervention. Data were collected monthly and combined into 3‐month blocks. The first postintervention sample (n = 116 patient charts) was drawn from a period 12‐14 months after implementation and the second (n = 147 patient charts) from a period 18‐20 months after implementation.

On a randomly designated day in the latter half of each month during each sampling period, all charts on 2 primary medical floors were reviewed and included in the retrospective analysis. Patients who were not on the medical service were excluded from analysis. Patients, as well as their medications, were identified using a list generated from our pharmacy database. We chose this method and schedule for several reasons. First, we sought to reduce the likelihood of including a patient more than once in a monthly sample. Second, by waiting for the latter half of the month we sought to allow house staff a chance to acquire knowledge from the educational program introduced on the first day of the month. Third, we wanted to allow house staff the time to actualize new attitudes reinforced by the audit‐and‐feedback element. The house staff included approximately 4 interns and 4 residents each month plus 10‐15 attendings or hospitalists.

Data Collection

For each sampling period we conducted a medical record (paper) review, and the Division Chief of General Internal Medicine also interviewed the medical house staff and attending physicians. Data collected included risk factors for VTE, contraindications to anticoagulant prophylaxis, type of VTE prophylaxis received, and appropriateness of the prophylaxis. Prophylaxis was considered appropriate when it was given in accordance with a risk stratification scheme (Table 1) adapted from the 2001 ACCP guideline recommendations for surgical patients20 and modified for medical inpatients, similar to the risk assessment model by Caprini et al.21 Prophylaxis was also considered appropriate when no prophylaxis was given for low‐risk patients or when full anticoagulation was given for another indication (Table 1). Questionable prophylaxis was defined as UFH given every 12 hours to a high‐risk patient. All other prophylaxis was deemed inappropriate (including no prophylaxis if prophylaxis was indicated, use of enoxaparin at incorrect prophylactic doses such as 60 or 20 mg, IPC alone for a high‐risk patient with no contraindication to pharmacological prophylaxis, and the use of warfarin if no other indication for it). The risk factors for thromboembolism and contraindications to anticoagulant prophylaxis are given in Table 2. Non‐ambulatory was defined as an order for bed rest with or without bathroom privileges or was judged based on information obtained from the medical house staff and nurses about whether the patient was ambulatory or had been observed walking outside his or her room. Data on pharmacological prophylaxis were obtained from the hospital pharmacy. Information on use of mechanical prophylaxis was obtained by house staff interviews or review of the order sheet. The house officer or attending physician of each patient was interviewed retrospectively to determine the reason for admission and the risk factors for VTE present on admission. Patients were classified as having low, moderate, high, or highest risk for VTE based on their age and any major risk factors for VTE (Table 1).19 All collected data were reported to the Department of Medicine Performance Improvement Committee for independent corroboration.

Intervention Strategies

The intervention introduced comprised 3 strategies designed to improve VTE prophylaxis: provider education, decision support, and audit‐and‐feedback.

Statistical Analysis

Differences in pre‐ and post‐intervention VTE prophylaxis and appropriate VTE prophylaxis rates were analyzed using the chi‐square test for categorical variables and the one‐way analysis of variance test for continuous variables. Differences were considered significant at the 5% level (P = .05).

Patients and Demographics

From October 2002 to August 2004 data were collected from 312 hospitalized medical patients: 49 patients in the baseline group during October 2002, and 116 and 147 at the 12‐ to 14‐month and 18‐ to 20‐month time points, respectively. Thus, approximately 40‐50 patients were randomly selected each month, representing 40% of the general medical service census. Patient demographics were similar between groups (Table 4). Overall, most patients were female (65.7%), and mean age was 61.2 years. The most common admission diagnoses were infection/sepsis (29.5%), chest pain/acute coronary syndromes/myocardial infarction (15.7%), heart failure (10.9%), and malignancy (9.6%). Overall, 7.1% (22 patients) had a contraindication to anticoagulant prophylaxis. The most common contraindication was active gastrointestinal bleeding on the current admission, which occurred in 18 of these patients.

Risk Factors for VTE

Patient risk factors for VTE in each data collection period are summarized in Table 4. Analysis of this data showed that the most prevalent risk factors for VTE in the 3 patient populations were age older than 40 years (262/312, 84.0% of the total patient population) and nonambulatory state (231/271, 85.2% of the total population). Overall, the average number of risk factors for VTE was approximately 3, with more than 60% of patients having 3 or more VTE risk factors (Fig. 1).

Figure 1

Prophylaxis Use

The types of VTE prophylaxis used and the proportion of patients treated appropriately are summarized for each data collection period in Tables 5 and 6, respectively. In all 3 populations, most patients received pharmacological rather than mechanical prophylaxis, most commonly UFH. At baseline, the prophylaxis decision was appropriate (in accordance with the recommendations of the ACCP guidelines) as often as it was inappropriate (42.9% of patients). The prophylaxis decision was questionable in the remaining 14.3% of patients.

Change in Prophylaxis Use

Twelve and 18 months after implementation of the quality improvement program, we observed an increase in the use of any prophylaxis, from 46.9% at baseline to 86.2% and 86.4%, respectively (Table 5; P < .01 in both groups versus baseline). This increase was a result almost entirely of an increase in the proportion of patients receiving pharmacological prophylaxis, which significantly increased, from 44.9% to 81.0% and 80.3%, at the 12‐ and 18‐month time points, respectively (Table 5; P < .01 for both groups versus baseline). Most meaningfully, there was a significant increase in the proportion of patients for whom an appropriate prophylaxis decision was made (from 42.9% to 68.1% and 85.0%, at the 12‐ and 18‐month time points, respectively; Table 6; P < .01 for both groups versus baseline). This represented a trend toward continuing increases in the use of appropriate prophylaxis as the study progressed (Fig. 2). This change was driven mainly by a significant increase in the prescribing of LMWH, almost all of which was prescribed in accordance with the 2001 ACCP guidelines (Table 6).

Figure 2