Secondary Logo

Journal Logo

Cardiovascular and Thoracic Anesthesiology: Original Clinical Research Report

Is Compliance With Surgical Care Improvement Project Cardiac (SCIP-Card-2) Measures for Perioperative β-Blockers Associated With Reduced Incidence of Mortality and Cardiovascular-Related Critical Quality Indicators After Noncardiac Surgery?

Kertai, Miklos D. MD, PhD*; Cooter, Mary MSc*; Pollard, Richard J. MD; Buhrman, William MD; Aronson, Solomon MD, MBA*; Mathew, Joseph P. MD, MHSc*; Stafford-Smith, Mark MD, FRCPC*

Author Information
doi: 10.1213/ANE.0000000000002577

Abstract

KEY POINTS

  • Question: The relationship between β-blocker withdrawal and mortality and cardiovascular-related critical quality indicators has not been studied in a contemporary cohort of patients undergoing noncardiac surgery.
  • Findings: Perioperative withdrawal of β-blockers was associated with increased risk for mortality within 48 hours after noncardiac surgery and with decreased risk for need of vasopressor during the early postoperative period and a shorter stay in the postanesthesia care unit.
  • Meaning: Perioperative continuation of β-blocker therapy is an advisable quality improvement target to prevent mortality, but could be associated with higher risk for perioperative hypotension and increased use of immediate postoperative health care resources.

Despite significant advances in perioperative care of noncardiac surgery patients, cardiovascular complications remain the major cause of perioperative morbidity and mortality, with myocardial infarction occurring in 1%–5% of unselected patients undergoing noncardiac surgery.1 These findings correlate with the increasing prevalence of advanced age, preexisting coronary artery disease, congestive heart failure, peripheral vascular disease, diabetes mellitus, and renal insufficiency in patients presenting for noncardiac surgery.1 Previous studies have shown that a set of these risk factors, as well as noninvasive testing, may help to identify patients who are at an increased risk for perioperative cardiovascular complications and thus could benefit from perioperative β-blocker (BB) therapy.2–4 BBs have been shown to reduce perioperative and long-term nonfatal myocardial infarction and cardiac death.4,5 However, starting BB therapy before noncardiac surgery could increase the risk for perioperative stroke and mortality.6

In patients who have been receiving BBs chronically, on the other hand, the risk for immediate postoperative and 1-year mortality and cardiovascular complications increases significantly when BB therapy is discontinued before noncardiac surgery.7,8 Thus, for secondary prevention of cardiovascular morbidity and mortality, the American College of Cardiology/American Heart Association (ACC/AHA) Task Force recommends that BB therapy continue perioperatively in patients who are receiving BBs longitudinally.1 This recommendation is consistent with the Surgical Care Improvement Project Cardiac (SCIP-Card-2) measures, which mandate that, as a quality improvement metric, BB therapy be continued from 24 hours before surgical incision through discharge from postanesthesia/recovery area in all surgery patients receiving BB therapy before admission.9 However, the ACC/AHA guideline and SCIP-Card-2 measures are based on expert opinion and studies with relatively small sample sizes conducted in a single practice setting and not on association studies that explored the relationship between BB withdrawal and immediate postoperative mortality and cardiovascular complications nor studies that focused on patients undergoing vascular or intermediate-risk elective noncardiac surgical procedures.1,4,7–9 Thus, in patients who receive BB therapy chronically, the association between perioperative withdrawal of BBs and risk for perioperative mortality or cardiovascular complications is unclear. Therefore, we assessed the association between perioperative BB withdrawal and mortality and cardiovascular complications in surgical cases of patients from a large community-based anesthesiology group practice who received BB therapy chronically and underwent emergent or elective noncardiac surgery.

METHODS

Data Source

The Institutional Review Board for Clinical Investigations at Duke University Medical Center in Durham, North Carolina, approved this study and waived the requirement for informed consent. The Department of Anesthesiology at the Duke University Medical Center served as a statistical analysis center for this study. A data set of unique encounters with patients who underwent elective or emergent noncardiac procedures between January 1, 2009, and December 31, 2014, was curated from the quality assurance database at American Anesthesiology, Inc, a large community-based anesthesiology group practice and a division of MEDNAX National Medical Group. The procedures were performed at 22 inpatient and outpatient facilities, pain clinics, endoscopy suites, intensive care units, and obstetric suites across 49 treatment locations. The inpatient and outpatient facilities of the MEDNAX National Medical Group vary in number of beds (<25 to >200), settings (urban and rural), and payer-mix. American Anesthesiology, Inc, uses the Quantum Clinical Navigation System (MEDNAX, Inc, Sunrise, FL), an internally designed quality improvement program that requires the anesthesiology care team (anesthesiologist, nurse anesthetists, and postoperative care unit nurses) to complete a standardized quality assurance sheet for each patient undergoing a surgical procedure with anesthetic care. The accuracy and completeness of every sheet is verified by a dedicated quality control nurse at each location within 48 hours and electronically transferred to a centralized database.

Study Population

On the basis of the quality assurance database at American Anesthesiology, Inc, we identified a set of 732,704 records of patients who underwent elective or emergent surgery during the study period. We excluded 106,120 cases that met any of the following criteria: younger than 18 years, unknown sex, and unknown American Society of Anesthesiologists (ASA) physical status classification system scores; 215,360 cases that involved a cardiac surgical procedure or an outpatient clinic procedure; and 936 cases with an unknown status for preoperative or perioperative BB use or unknown type of anesthetic procedure. Of the remaining 410,288 cases with known BB therapy status, based on the SCIP-Card-2 definition of preoperative/perioperative BB therapy, 343,533 cases were not receiving BB therapy before admission; 3829 cases were on BB therapy before admission but did not receive BB therapy perioperatively (withdrawn from BB therapy group); and 62,926 cases were on BB therapy before admission and also received BB therapy perioperatively (continued on BB therapy group). Given the SCIP-Card-2 definition of preoperative/perioperative BB therapy, no specific data were recorded on the quality assurance sheet regarding the number of the 343,533 cases without BB therapy before admission, who were then administered BB therapy during the perioperative period. Thus, 66,755 cases were eligible for inclusion in the current analysis.

Data Extraction

We extracted all available demographic and clinical data that were recorded on the quality assurance sheet (Supplemental Digital Content, Figure 1, http://links.lww.com/AA/C88). These included demographic data, anesthetizing location, type of providers involved, anesthetic and analgesic techniques, procedure-related factors, comorbidities, positive initiatives such as preoperative and perioperative BB use (according to the SCIP-Card-2 definition), efficiency indicators, anesthesia practitioner performance indicators, and several important clinical outcomes up to 48 hours after induction of anesthesia or start of monitored anesthesia care.

Classification of Outcomes

The primary outcome chosen for the current study was all-cause mortality, and the secondary outcomes of the study were cardiovascular-related critical quality indicators (CQIs) that had been reported up to 48 hours after induction of anesthesia or start of monitored anesthesia care, ie, those collected on the Quantum Clinical Navigation System Quality Assurance Sheet and those previously described as outcome indicators associated with perioperative initiation or discontinuation of BBs4–8,10: blood pressure support requiring vasopressor initiation, electrocardiographic changes requiring treatment, unplanned admission to intensive care unit, postanesthesia care unit stay >2 hours, and a combination of cardiac arrest and myocardial infarction. These cardiovascular-related CQIs are described in detail in Supplemental Digital Content, Table 1, http://links.lww.com/AA/C88.

Statistical Analysis

Summary statistics are presented as means (± standard deviation) for continuous variables and as group frequencies and percentages for categorical variables. T-tests or χ2 tests were used for group comparisons of those with and without continuation of perioperative BB therapy. Due to the size of the data set, we supplemented the group comparisons with standardized differences to identify clinically meaningful differences on a standardized scale.11–13 For numeric variables, the standardized difference (d) was defined as

where and are the sample mean of the variable in each group, and and are the respective sample variances. For binary categorical variables, d was defined as:

where and are the proportion of patients with the characteristic in each group. We employed the Yang and Dalton formulation for categorical variables with 3 or more levels, which treats the variable as multiple nonredundant dichotomous variables and uses the Mahalanobis distance to define the standardized difference.13 We defined negligible covariate imbalance between groups as d < 0.10, which occurs when >90% of data overlaps in the 2 groups.12

Comparisons of demographic and clinical characteristics of cases withdrawn from BB therapy versus cases continued on BB therapy revealed several notable differences; therefore, we conducted nearest neighbor propensity-score matching to reduce potential for confounding in the analysis. The propensity score model was based on demographic and clinical patient characteristics that were considered potential confounders based on findings of prior studies and availability of these characteristics in the current data set. Subsequently, cases were matched based on the derived propensity scores in a 1:4 ratio to balance our interest in retaining as many mortality events and cardiovascular-related CQIs as possible and also reducing residual confounding of differences in demographic and clinical characteristics.14 The resulting propensity score distributions were evaluated for balance, and demographic and clinical characteristics of the matched cohort were compared via standardized differences. If imbalance remained after the matching, we explored model reformulation or used a caliper as necessary to improve the match.

The number of perioperative events in the current study was relatively limited. Therefore, to avoid overfitting and to enable assessment of the interaction between demographic and clinical risk factors and patterns of BB use for all-cause mortality and cardiovascular complications, we used the Revised Cardiac Risk Index suggested by Lee et al,15 including age >65 years, high-risk surgery type (aortic and major vascular surgery, duodenopancreatic surgery, bile duct surgery, major lung surgery) as defined by Kristensen et al,16 coronary artery disease, diabetes mellitus, and renal insufficiency (creatinine ≥2 mg/dL on admission).

Univariable and multivariable conditional logistic regression models were applied to evaluate the association between patterns of BB use and each of all-cause mortality and cardiovascular-related CQIs, as well as the interaction between BB use and the modified Revised Cardiac Risk Index on each of these outcomes. Conditional logistic regression was used to account for the analysis cohort derivation via propensity score matching, specifying an exchangeable correlation structure among clustered propensity-matched patients. The discriminatory ability of the multivariable models was quantified by the c-index, which is identical to the area under the receiver operating characteristics curve. The c-index ranges from 0.5 (not predictive at all) to 1.0 (optimal calibration).17 The fit of the multivariable regression models to the data was further evaluated using the Hosmer–Lemeshow goodness-of-fit test.18 Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) are reported. In addition, we derived the absolute risk difference, number needed to treat (NNT), and number needed to harm (NNH) with their respective 95% CIs for each outcome based on the results of our multivariable models to further quantify the association between BB withdrawal versus BB continuation and outcomes.

While our study was designed to retain as many outcomes of interest and remove as much potential confounding as possible in our analysis cohort, our analysis may still have been underpowered. To evaluate our potential for type II error, we conducted a power calculation and sample size analysis to determine our ability to detect impactful NNT/NNH for the primary outcome of all-cause mortality. We felt that if the detectable NNT/NNH was <500, the decision to continue or withdraw patients from BBs would be significant at a system level of any large community-based anesthesiology group practice. In our propensity-matched BB continuation group, we observed a 48-hour all-cause mortality rate of 0.10%, and given that rate, a 4:1 matched study of 19,145 cases would reach 80% power to detect a NNT/NNH of 500 for the primary outcome, which corresponds to a 0.20% increase in risk of all-cause mortality. Smaller absolute risk differences were beyond the scope of our study, but a future study of this rare outcome in a 1:1 matched cohort would require 47,050 patients (23,525 patients in each BB pattern group) to reach 80% power to detect a NNT/NNH of 1000 (corresponds to a 0.10% increase in risk of all-cause mortality).

Propensity matching was performed using the “MatchIt” package (https://cran.r-project.org/web/packages/MatchIt/index.html),19 and standardized differences were calculated using the “tableone” package in R (version 3.1.1; http://www.R-project.org/).20 All modeling analyses were performed using SAS Version 9.4 (SAS Institute, Inc, Cary, NC).

RESULTS

Before propensity matching, there were several notable differences between cases withdrawn from BB therapy and cases continued on BB therapy, including age; ASA Physical Status Classification System scores; frequency of emergency surgery; history of smoking; obesity; hypertension; diabetes mellitus; frequency of low-, intermediate-, and high-risk surgery; and type of anesthesia. We included the following demographic and clinical patient characteristics in the propensity score model that were available in the current data set: age, gender, ASA physical status classification, obesity, hypertension, coronary artery disease, diabetes mellitus, type of anesthesia, and surgical procedure associated risk (Table 1). Histories of stroke and transient ischemic attack were only available on a subset of cases, and therefore, they could not be considered in our propensity score model for propensity score matching. History of renal insufficiency was well balanced between the groups of cases withdrawn from BB therapy versus cases continued on BB therapy in the full cohort, and hence, there was no indication that this clinical patient characteristic would act as a confounding variable in our subsequent analyses. Current smoking status was not included due to its detrimental effect on the propensity score model stability.

Table 1.
Table 1.:
Demographic and Clinical Characteristics of the Full and Matched Study Population

After propensity matching, 3829 cases withdrawn from BB therapy and 15,316 cases continued on BB therapy were included in subsequent analyses. Although our 1:4 matching did not use a caliper, the propensity scores covered a narrow range (0.29 for withdrawal cases and 0.28 for continuation) and were well balanced between the matched groups with median (Q1–Q3) of 0.06 (0.05–0.07) in both groups. Demographic and clinical characteristics of the study cohort before and after propensity score matching are presented in Table 1. Based on the standardized difference of our propensity score-matched cohorts, demographic and clinical characteristics of cases continued on BB therapy were well matched to demographic and clinical characteristics of cases withdrawn from BB therapy.

Association Between BB Withdrawal and Outcomes According to the Revised Cardiac Risk Index

Figure 1 shows the incidence of all-cause mortality and cardiovascular-related CQIs according to the risk score category of the modified Revised Cardiac Risk Index, in cases withdrawn from BB therapy versus cases continued on BB therapy. Generally, the incidence of all-cause mortality and cardiovascular complications increased significantly with increasing clinical risk score category. In particular, cases withdrawn from BB therapy had an incrementally higher incidence of all-cause mortality versus cases continued on BB therapy (Figure 1 and Table 2). However, cases withdrawn from BB therapy had a lower incidence of cardiovascular-related CQIs such as need for vasopressor infusion for blood pressure support, unplanned admission to the intensive care unit, or longer stay in a postanesthesia care unit (>2 hours; Figure 1 and Table 2).

Table 2.
Table 2.:
Event Rates by β-Blocker Use and Multivariable Analysis in the Matched Cohort
Figure 1.
Figure 1.:
Incidence of mortality and cardiovascular-related CQIs within 48 hours after surgery by the modified Revised Cardiac Risk Index categories and patterns of β-blocker use. P values shown in the figure are from the reported outcome models for all-cause mortality and cardiovascular-related CQIs and are also reported in Table 2. BB indicates β-blockers; CQI, critical quality indicator.

After adjusting for the modified Revised Cardiac Risk Index, the multivariable conditional logistic regression analysis showed that the odds of all-cause mortality up to 48 hours after surgery were 3.5-fold higher among cases withdrawn from BB therapy versus cases continued on BB therapy (adjusted OR, 3.52; 95% CI, 1.71–7.24; P = .0006; Table 2). The multivariable model for all-cause mortality had good discriminatory ability and excellent goodness of fit (c-index, 0.725; overall Hosmer–Lemeshow goodness-of-fit test, χ2 = 0.13; P = .9997). In separate multivariable analyses of the cardiovascular-related CQIs, the estimated odds of requiring vasopressor infusion for blood pressure support or an extended stay in a postanesthesia care unit (>2 hours) in cases withdrawn from BB therapy versus cases continued on BB therapy were 0.85 times (95% CI, 0.77–0.93) and 0.69 times (95% CI, 0.54–0.89) lower, respectively (Table 2). The odds of electrocardiogram (EKG) changes requiring treatment in cases withdrawn from BB therapy versus cases continued on BB therapy was not statistically significant (Table 2).

Figure 2.
Figure 2.:
Log ORs of all-cause mortality and cardiovascular-related CQIs within 48 hours after surgery by the modified Revised Cardiac Risk Index categories in cases withdrawn from BB therapy compared to cases continued on BB therapy. P values shown in the figure are for an interaction between patterns of BB use and the number of risk scores of the modified Revised Cardiac Risk Index for all-cause mortality and cardiovascular-related CQIs. BB indicates β-blockers; CQI, critical quality indicator; EKG, electrocardiogram; OR, odds ratio.

We also studied whether there was an interaction between patterns of BB use and the Revised Cardiac Risk Index on all-cause mortality and cardiovascular-related CQIs. As shown in Figure 2, we found no evidence of an interaction between patterns of BB use and the number of risk scores of the modified Revised Cardiac Risk Index for mortality (P = .97) or any of the cardiovascular-related CQIs outcomes we studied (Figure 2).

Number Needed to Treat and Number Needed to Harm

Table 3.
Table 3.:
Risk Differences and Number Needed to Treat and Number Needed to Harm in the Matched Cohort

In a subsequent analysis, absolute risk differences, NNT, and NNH were assessed in our matched cohort of cases withdrawn from BB therapy and cases continued on BB therapy (Table 3). We found that continuation of BB therapy after surgery was needed in 383 (95% CI, 228–715) cases to prevent one all-cause mortality up to 48 hours after surgery. Further, 40 (95% CI, 33–52) cases needed to continue on BB therapy after surgery for 1 case to occur requiring vasopressor infusion for blood pressure support; and 119 (95% CI, 100; 162) cases needed to continue on BB therapy after surgery for 1 case to occur requiring an extended stay in a postanesthesia care unit (>2 hours; Table 3). In our multivariable analysis, EKG changes requiring treatment did not differ significantly between cases withdrawn from BB therapy and cases continued on BB therapy; however, continuation of BB therapy after surgery was needed in 349 (95% CI, 286–589) cases to prevent 1 case that required treatment for EKG changes (Table 3).

DISCUSSION

In this study of patients from a large, multiinstitutional, community-based anesthesiology group practice who underwent emergent or elective noncardiac surgery, perioperative withdrawal of BB therapy was associated with an increased risk for mortality, but a decreased risk for cardiovascular-related CQIs such as need for vasopressor use and extended postanesthesia care unit stay within 48 hours after surgery. Further our findings demonstrated that the association between BB withdrawal and outcomes did not vary by the cardiac risk status of cases, thus indicating that cases irrespective of their cardiac risk profile based on the modified Revised Cardiac Risk Index were equally affected by BB withdrawal.

Earlier studies reported that perioperative BB therapy is associated with a lower risk for cardiovascular complications and mortality after noncardiac surgery21,22 and that this beneficial association is more pronounced in patients at intermediate and high risk for perioperative cardiovascular complications and mortality. However, when it came to the interpretation of the findings of these studies, it was subsequently highlighted that in the study by Mangano et al,21 some patients had their existing BB treatment withdrawn before randomization to atenolol or placebo, the effect of which on perioperative cardiovascular outcomes were not studied, whereas there were some concerns raised over the reported findings of the study by Poldermans et al.22 Despite limited evidence from these earlier randomized trials, perioperative treatment with BB therapy become widely advocated, and in among many different health care agencies, the Agency for Healthcare Research and Quality in its 2001 report identified perioperative BB therapy as a key opportunity for safety improvement in patients undergoing noncardiac surgery,23 and the National Quality Forum included the use of BBs in high-risk noncardiac surgery patients among its top safe practices for better health care in its 2003 report.24

Two subsequent randomized clinical trials, however, reported no benefit from perioperative BB therapy and thus called into question the generalizability of the earlier findings.25,26 Since the earlier clinical trials were less likely to represent routine clinical practice for BB therapy in noncardiac surgery, Lindenauer et al27 conducted a large observational study and found that perioperative administration of BBs that occurred on the first or second hospital day is associated with clear and clinically significant reductions in mortality in intermediate- and high-risk patients undergoing major noncardiac surgery. These investigators also observed that only a minority of high-risk patients received BBs, thus highlighting the need for following the recommendation of the Agency for Healthcare Research and Quality on the perioperative use of BBs.

In an effort to resolve the conflicting results from prior BB trials and observational studies, Devereaux et al conducted a large-scale clinical trial. Their findings indicated that high-dose BB therapy with extended-release metoprolol started 2–4 hours before surgery and continued for 30 days is associated with a reduced risk for myocardial infarction and cardiac revascularization compared with placebo, but that BB therapy with extended-release metoprolol also increases risk for death, stroke, and, similar to our findings, clinically significant hypotension and bradycardia.6 Indeed, bradycardia and hypotension were the most common reasons for temporary suspension of metoprolol use in the Devereaux trial.

These striking findings prompted additional investigators to explore the safety and effectiveness of BB therapy during the perioperative period. In a large-scale observational study of patients undergoing noncardiac surgery, Wallace et al5 found that BB therapy administered according to the Perioperative Cardiac Risk Reduction protocol is associated with reduced 30-day mortality and, but similar to our study, found that withdrawal of BBs is associated with increased mortality. In a more recent study, London et al4 found that exposure to BB therapy on postoperative day 0 or 1 is associated with lower rates of perioperative mortality in patients at intermediate and high risk according to the Revised Cardiac Risk Index scores, suggesting that decisions about administering and continuing BB therapy could be based on a cumulative number of Revised Cardiac Risk Index predictors. Again, similar to our findings, these investigators also observed that patients withdrawn from BB therapy are at a higher risk for mortality. Nevertheless, these previous randomized clinical trials and observational studies had different regimens for initiation and/or withdrawal of perioperative BB therapy and thus making it particularly challenging for practicing physicians to assess the safety and efficacy of perioperative BB therapy across the spectrum of patients at risk for perioperative cardiovascular complications and mortality undergoing noncardiac surgery.

Such concerns surrounding the initiation and/or withdrawal of perioperative BB therapy in patients undergoing noncardiac surgery prompted the ACC/AHA Task Force on Perioperative Cardiovascular Evaluation and Management Practice Guidelines to reinforce its recommendation to continue BB therapy in noncardiac surgery patients who have been receiving BBs chronically.1 This recommendation is consistent with the SCIP-Card-2 measures that mandate that all surgery patients taking BBs before admission, continue BB therapy from 24 hours before surgical incision through discharge from postanesthesia/recovery area.9 Further, a statewide education campaign promoting adherence to the SCIP-2-Card measures and deployment of a Surgical Care and Outcomes Assessment Program surgical checklist, which includes a review of the perioperative plan on BB therapy by the surgical team, have substantially increased the rate that BBs are continued perioperatively.7

Importantly, the ACC/AHA Guidelines and SCIP-Card-2 measures do not provide specific recommendations for modifying or temporarily discontinuing BB therapy when circumstances dictate. Nevertheless, BB withdrawal is associated with significantly higher rates of mortality, cardiac complications, and combined adverse events at 90 days and 1 year after intermediate risk surgery.7 Our study corroborated some of these findings, but we also uncovered evidence that strengthens not only the potential implications on mortality risk of withdrawing BB therapy but also the potential relationship between continuing BB therapy and cardiovascular-related CQIs such as perioperative hemodynamic perturbations and postoperative use of health care resources. Thus, our findings in agreement with the recommendations of the most recent update of the Agency for Healthcare Research and Quality on use of BBs to prevent perioperative cardiac events28 that highlights the need for future studies and clinical metrics to guide perioperative BB therapy toward minimizing the risk for BB therapy-related complications.

Given its retrospective design, our study has certain limitations that should be taken into consideration when interpreting the results.

  1. Available data were limited to the first 48 hours after the procedure was completed and thus only partially reflect patient outcomes. Nevertheless, this large, multiinstitutional, community-based anesthesiology group practice applies a system-wide quality assurance initiative that requires completion of a detailed quality assurance sheet on all consecutive anesthesia cases and independent verification by quality assurance nurses at each practice site of the validity and accuracy of the data entered.
  2. The data used from this quality assurance database were from consecutive anesthetic procedures and not consecutive patients. Thus, some patients may have had multiple anesthetic procedures, and our inference may be based on biased variance estimates. However, given the size and nature of this data set, this is unlikely to have significantly impacted our findings.
  3. We adjusted for the Revised Cardiac Risk Index to avoid overfitting in our covariate-adjusted analyses. Since some of the clinical risk factors in the Revised Cardiac Risk Index were not available in the study database, we made some modifications, which are described in the Methods section. Thus, the impact of some of the key clinical risk factors and procedures included in the Revised Cardiac Risk Index may not have been considered or were underestimated. Nevertheless, the magnitude and direction of the estimate of the Revised Cardiac Risk Index score in our analyses were similar to previous reports.29
  4. Data on within-practice and between-practice variability by clinicians regarding perioperative BB withdrawal or continuation practices were not collected and assessed. We also observed some inherent differences in baseline and clinical characteristics between cases withdrawn from BB therapy and cases continued on BB therapy that could have resulted in a biased estimate of the association between BB withdrawal and outcomes. However, by applying propensity matching of cases withdrawn from BB therapy and cases continued on BB therapy, we were able to account for such differences.
  5. There is a possibility that the observed shorter stay in a postanesthesia care unit (<2 hours) in cases withdrawn from BB therapy was in part due to higher all-cause mortality. Nevertheless, in order for this to be the case, all-cause mortality attributed to BB withdrawal would have had to occur within 2 hours of surgery and outside the operating room. While we cannot determine the rate of this kind of event and may not be able to completely rule it out, we expect the rate of such events to be low. In addition, extended stay in a postanesthesia care unit (>2 hours) occurred at a much higher frequency than 48-hour all-cause mortality, so the difference is likely robust to the rate of mortality differences.
  6. The database analyzed in the present study did not include cardiovascular medications other than BBs nor type of BBs administered. Hence, we were not able to determine the association between other cardiovascular medications and outcomes nor their interactions with BB therapy; and we could not evaluate the relative impact of different types of BBs. Further, the database did not collect specific information on why BBs (eg, hypotension or bradycardia) were discontinued in the postoperative period. Also, no specific data were collected on the number of cases who, with no BB therapy before admission, were administered BB therapy during the perioperative period. Thus, we could not study the potential association between initiation of BB therapy and perioperative outcomes. Future studies are urgently needed to determine the association between type of BB medication and perioperative outcomes in patients receiving other cardiovascular medications and the association between initiating of BB therapy during the perioperative period and outcomes in patients with no BB therapy before admission.
  7. The diagnosis of myocardial infarction practice wide was established by a qualified physician during the perioperative period up to 24 hours after surgery (see Supplemental Digital Content, Table 1, http://links.lww.com/AA/C88). However, it has been observed that perioperative myocardial infarction is silent in most instances.30 Since no routine protocol was instituted for serial blood sampling of cardiac enzymes and EKGs, it is likely that many more cases that developed myocardial infarction could have been undetected, especially as cardiac enzyme measurements were limited to the first 24 hours. Similarly, future studies are needed to determine the relationship of BB withdrawal with the occurrence of perioperative myocardial infarction.

In conclusion, perioperative withdrawal of BB therapy on the day of surgery and after surgery was associated with an increased risk for mortality, but a reduced risk for cardiovascular-related CQIs such as need for vasopressor infusion for blood pressure support and extended stay in a postanesthesia care unit (>2 hours). Thus, perioperative continuation of BB therapy is an advisable quality improvement target to prevent mortality, but could be associated with higher risk for perioperative hypotension and increased use of immediate postoperative health care resources. Our findings indicate that in addition to following the SCIP-Card-2 measures for perioperative BB therapy to reduce risk for mortality, differential care focused on hemodynamic management and immediate postoperative care should be devoted to high cardiac risk patients who are on BB therapy and could be at risk for BB therapy-related hemodynamic perturbations with subsequent increased use of immediate postoperative health care resources.

DISCLOSURES

Name: Miklos D. Kertai, MD, PhD.

Contribution: This author helped design the study, analyze the data, interpret the data, and prepare the manuscript.

Name: Mary Cooter, MSc.

Contribution: This author helped design the study, analyze the data, and critically review the manuscript.

Name: Richard J. Pollard, MD.

Contribution: This author helped with acquisition of the data, design the study, and critically review the manuscript.

Name: William Buhrman, MD.

Contribution: This author helped with acquisition of the data.

Name: Solomon Aronson, MD, MBA.

Contribution: This author helped design the study and review the manuscript.

Name: Joseph P. Mathew, MD, MHSc.

Contribution: This author helped design the study and review the manuscript.

Name: Mark Stafford-Smith, MD, FRCPC.

Contribution: This author helped design the study, analyze the data, interpret the data, and prepare the manuscript.

This manuscript was handled by: W. Scott Beattie, PhD, MD, FRCPC.

REFERENCES

1. Fleisher LA, Fleischmann KE, Auerbach AD, et al. 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;130:2215–2245.
2. Andersson C, Mérie C, Jørgensen M, et al. Association of β-blocker therapy with risks of adverse cardiovascular events and deaths in patients with ischemic heart disease undergoing noncardiac surgery: a Danish nationwide cohort study. JAMA Intern Med. 2014;174:336–344.
3. Lau WC, Froehlich JB, Jewell ES, et al. Impact of adding aspirin to beta-blocker and statin in high-risk patients undergoing major vascular surgery. Ann Vasc Surg. 2013;27:537–545.
4. London MJ, Hur K, Schwartz GG, Henderson WG. Association of perioperative β-blockade with mortality and cardiovascular morbidity following major noncardiac surgery. JAMA. 2013;309:1704–1713.
5. Wallace AW, Au S, Cason BA. Association of the pattern of use of perioperative β-blockade and postoperative mortality. Anesthesiology. 2010;113:794–805.
6. Devereaux PJ, Yang H, Yusuf S, et al. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet. 2008;371:1839–1847.
7. Kwon S, Thompson R, Florence M, et al.; Surgical Care and Outcomes Assessment Program (SCOAP) Collaborative. β-Blocker continuation after noncardiac surgery: a report from the surgical care and outcomes assessment program. Arch Surg. 2012;147:467–473.
8. Shammash JB, Trost JC, Gold JM, Berlin JA, Golden MA, Kimmel SE. Perioperative beta-blocker withdrawal and mortality in vascular surgical patients. Am Heart J. 2001;141:148–153.
9. Surgical Care Improvement Project. SCIP-Card-2: surgery patients on beta-blocker therapy prior to admission who received a beta-blocker during the perioperative period. Available at: https://manual.jointcommission.org/releases/archive/TJC2010B/MIF0059.html. Accessed September 7, 2017.
10. Hoeks SE, Scholte Op Reimer WJ, van Urk H, et al. Increase of 1-year mortality after perioperative beta-blocker withdrawal in endovascular and vascular surgery patients. Eur J Vasc Endovasc Surg. 2007;33:13–19.
11. Austin PC. Unisng the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Comm Stat Simulat Comput. 2009;38:1228–1234.
12. Flury BK, Riedwyl H. Standard distance in univariate and multivariate analysis. Am Statistician. 1986;40:249–251.
13. Yang D, Dalton JE. Paper 335-2012: A unified approach to measuring the effect size between two groups using SAS. SAS Global Forum 2012. 2012 Cary SAS:335i. http://support.sas.com/resources/papers/proceedings12/335-2012.pdf. Accessed September 7, 2017.
14. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10:150–161.
15. Lee TH, Marcantonio ER, Mangione CM, et al. Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery. Circulation. 1999;100:1043–1049.
16. Kristensen SD, Knuuti J, Saraste A, et al.; Authors/Task Force Members. 2014 ESC/ESA Guidelines on non-cardiac surgery: cardiovascular assessment and management: The Joint Task Force on non-cardiac surgery: cardiovascular assessment and management of the European Society of Cardiology (ESC) and the European Society of Anaesthesiology (ESA). Eur Heart J. 2014;35:2383–2431.
17. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29–36.
18. Lemeshow S, Hosmer DW Jr. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol. 1982;115:92–106.
19. Ho DE, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Software. 2011;42:1–28.
20. Yoshida K, Bohn J. Create “table 1” to describe baseline characteristics [computer program]. R package version 0.7.3. 2015. Available at: http://CRAN.R-project.org/package=tableone. Accessed September 7, 2017.
21. Mangano DT, Layug EL, Wallace A, Tateo I. Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. Multicenter Study of Perioperative Ischemia Research Group. N Engl J Med. 1996;335:1713–1720.
22. Poldermans D, Boersma E, Bax JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery. Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study Group. N Engl J Med. 1999;341:1789–1794.
23. Auerbach AD, Shojania KG, Duncan BW, McDonald KM, et al. Beta-blockers and reduction of perioperative cardiac events. In: Making health care safer: a critical analysis of patient safety practices. Evidence Report/Technology Assessment No. 43. AHRQ Publication No. 01-E058. Rockville, MD: Agency for Healthcare Research and Quality; 2001:271–279.
24. Safe Practices for Better Healthcare: A Consensus Report. 2003.Washington, DC: National Quality Forum (Publication no. NQFCR-05-03.)
25. Juul AB, Wetterslev J, Gluud C; DIPOM Trial Group. Effect of perioperative beta blockade in patients with diabetes undergoing major non-cardiac surgery: randomised placebo controlled, blinded multicentre trial. BMJ. 2006;332:1482.
26. Yang H, Raymer K, Butler R, Parlow J, Roberts R. The effects of perioperative beta-blockade: results of the Metoprolol after Vascular Surgery (MaVS) study, a randomized controlled trial. Am Heart J. 2006;152:983–990.
27. Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med. 2005;353:349–361.
28. Ranji SR, Shekelle PGShekelle PG, Wachter RM, Pronovost PJ, et al. Use of beta-blockers to prevent perioperative cardiac events: brief update review. In: Making Health Care Safer II: An Updated Critical Analysis of the Evidence for Patient Safety Practices. Comparative Effectiveness Review No. 211. Rockville, MD: Agency for Healthcare Research and Quality; 2013:169–171.
29. Boersma E, Kertai MD, Schouten O, et al. Perioperative cardiovascular mortality in noncardiac surgery: validation of the Lee cardiac risk index. Am J Med. 2005;118:1134–1141.
30. Devereaux PJ, Chan MT, Alonso-Coello P, et al. Association between postoperative troponin levels and 30-day mortality among patients undergoing noncardiac surgery. JAMA. 2012;307:2295–304.

Supplemental Digital Content

Copyright © 2017 International Anesthesia Research Society