Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery: a systematic review and meta-analysis

Transcription

1 British Journal of Anaesthesia, 8 (): 2 (207) doi: 093/bja/aew380 Review Article REVIEW ARTICLE Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery: a systematic review and meta-analysis S. Hajibandeh, *, S. Hajibandeh, S. A. Antoniou 2, F. Torella 3, and G. A. Antoniou 4 Department of General Surgery, Royal Blackburn Hospital, Blackburn, UK, 2 Department of General Surgery, University Hospital of Heraklion, University of Crete, Heraklion, Greece, 3 Department of Mathematical Sciences, School of Physical Sciences, University of Liverpool, Liverpool, UK and 4 Department of Vascular and Endovascular Surgery, The Royal Oldham Hospital, Pennine Acute Hospitals NHS Trust, Manchester, UK *Corresponding author. shahab_hajibandeh@yahoo.com Abstract Background. To investigate the role of perioperative beta-blocker use in vascular and endovascular surgery. Methods. We performed a systematic review in accordance with Preferred Reporting Items for Systematic Reviews and Meta- Analyses statement standards. The review protocol was registered with International Prospective Register of Systematic Reviews (registration number:crd ). We searched electronic databases to identify all randomized controlled trials and observational studies investigating outcomes of patients undergoing vascular and endovascular surgery with or without perioperative beta blockade. We used the Cochrane tool and the Newcastle-Ottawa scale to assess the risk of bias of trials and observational studies, respectively. Random-effects models were applied to calculate pooled outcome data. Results. We identified three randomized trials, five retrospective cohort studies, and three prospective cohort studies, enrolling a total of 32,602 patients. Our analyses indicated that perioperative use of beta-blockers did not reduce the risk of all-cause mortality [odds ratio (OR).0, 95% confidence interval (CI) , P ¼ 0.77], cardiac mortality (OR 2.62, 95% CI , P ¼ 0.09), myocardial infarction (OR 0.89, 95% CI , P ¼ 0.58), unstable angina (OR.34, 95% CI , P ¼ 0.63), stroke (OR 2.45, 95% CI , P ¼ 0.08), arrhythmias (OR 0.76, 95% CI , P ¼ 0.40), congestive heart failure (OR.2, 95% CI , P ¼ 0.56), renal failure (OR.48, 95% CI , P ¼ 3), composite cardiovascular events (OR 0.88, 95% CI , P ¼ 0.58), rehospitalisation (OR 0.86, 95% CI , P ¼ 0.60), and reoperation (OR.7, 95% CI , P ¼ 0.77) in vascular surgery. Conclusions. Beta-blockers do not improve perioperative outcomes in vascular and endovascular surgery. Key words: beta blocker; perioperative; vascular Perioperative cardiac mortality and morbidity are the most frequent adverse events in vascular surgery. There is a strong relationship between perioperative cardiac and noncardiac complications and subsequent mortality; nearly half of the patients experiencing cardiac morbidity will develop other types of noncardiac complications and mortality. 2 Beta-adrenoceptor blocking agents are traditionally used to treat hypertension and are the primary treatment choice after Editorial decision September 3, 206; Accepted: October 2, 206 VC The Author 206. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. For Permissions, please journals.permissions@oup.com

2 2 Hajibandeh et al. Editor s Key Points Vascular surgical patients are at particular risk of adverse cardiac events The effects of perioperative beta-blocker therapy may be different in vascular surgery This pooled analysis could not identify any beneficial or harmful effects of beta-blockers in vascular surgery myocardial infarction or for chronic angina. 34 In addition, betablockers have been shown to reduce morbidity and mortality in patients with mild, moderate and severe chronic heart failure. 5 7 It has been proposed that beta- blockers reduce the risk of perioperative cardiac complications by slowing heart rate, decreasing blood pressure, and moderating haemodynamic stress responses. 8 The effectiveness and safety of perioperative beta-blockers for patients undergoing noncardiac surgery remains controversial. Some authors have reported that perioperative beta-blockers started within one day or less of noncardiac surgery prevent nonfatal myocardial infarction but increase the risk of stroke, death, hypotension, and bradycardia. 9 Others have found no clear evidence for the effect of perioperative beta-blockers on all-cause mortality, cerebrovascular 0 events, myocardial infarction, or arrhythmias. We thus conducted a comprehensive systematic review and meta-analysis to investigate the role of perioperative betablockers in noncardiac vascular and endovascular surgery. Methods This systematic review was performed according to an agreed predefined protocol which was registered with the International Prospective Register of Systematic Reviews (registration number: CRD ). The review was conducted and presented according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement standards. 2 Eligibility criteria We included all observational studies and randomized controlled trials (RCTs) investigating outcomes of patients undergoing vascular and endovascular surgery with or without perioperative beta blockade. Perioperative beta-blockers of any dose, titration, duration and mode of administration were considered as intervention of interest, and placebo or no treatment was considered as comparator. Adults more than 8 yr of age undergoing noncardiac vascular and endovascular surgery were considered as participants of interest. We defined noncardiac vascular and endovascular surgery as any surgical or endovascular intervention for treatment of vascular disease in carotid or vertebral arteries, other supra-aortic arteries, aortoiliac arteries, renal and visceral arteries, upper or lower extremity arteries, and for vascular access for haemodialysis. Outcome measures All-cause mortality was considered the primary outcome measure. The secondary outcome measures included cardiac mortality, myocardial infarction (MI), unstable angina, stroke, arrhythmias, congestive heart failure (CHF), renal failure, composite cardiovascular events (MI, unstable angina, stroke, dysrhythmia or cardiac death), rehospitalization, and reoperation. Literature search strategy Two authors (S.H., S.H.) independently searched the following electronic databases: MEDLINE, EMBASE, CINAHL, and the Cochrane Central Register of Controlled Trials (CENTRAL). The last search was run on 30 April 206. Thesaurus headings, search operators and limits in each of the above databases were adapted accordingly. The literature search strategy is outlined in Appendix I. In addition, World Health Organization International Clinical Trials Registry ( search/), ClinicalTrials.gov ( and ISRCTN Register ( were searched for details of ongoing and unpublished studies. The bibliographic lists of relevant articles and reviews were searched for further potentially eligible studies. Moreover, leading journals in vascular and endovascular surgery were hand-searched. No language restrictions were applied in our search strategies. Study selection The title and abstract of articles identified from the literature searches were assessed independently by two authors (S.H., S.H.). The full-texts of relevant reports were retrieved and those articles that met the eligibility criteria of our review were selected. Any discrepancies in study selection were resolved by discussion between the authors. An independent third author (S.A.A.) was consulted in the event of disagreement. Data collection We created an electronic data extraction spreadsheet which was pilot-tested in randomly selected articles and was adjusted accordingly. Our data extraction spreadsheet included: studyrelated data (first author, yr of publication, country of origin of the corresponding author, journal in which the study was published, study design, study size, clinical condition of the study participants); baseline patient characteristic and clinical information of the study populations (age, gender, diabetes mellitus, coronary heart disease, hypertension, cerebrovascular disease, and smoking status); and primary and secondary outcome data. Data collection was performed independently by two authors (S.H., S.H.), and disagreements were resolved by discussion. If no agreement could be reached, a third author (SAA) was consulted. Methodological quality and risk of bias assessment Two authors (S.H. and S.H.) independently assessed the methodological quality and risk of bias of the included articles, using the Cochrane tool and the Newcastle-Ottawa scale (NOS) 3 for assessing the risk of bias of randomized trials and observational studies, respectively. The Cochrane s tool assesses domains including selection bias, performance bias, detection bias, attrition bias, reporting bias, and other sources of bias and, for each individual domain, classifies studies into low, unclear, and high risk of bias. The NOS uses a star system with a maximum of nine stars to evaluate a study in three domains (8 items): the selection of the study groups, the comparability of the groups, and the ascertainment of outcome of interest. For each item of the scale, we judged each study as low risk (one star awarded) or high risk (no star awarded). We determined studies that received a score of nine stars to be of low risk of bias, studies that scored seven or eight stars to be of moderate risk, and those that scored six or less to be of high risk of bias. Disagreements were resolved by discussion between the reviewers. If no

3 Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery 3 agreement could be reached, a third author (G.A.A.) acted as an adjudicator. A risk of bias graph was constructed to present the results. Data synthesis and statistical analyses The primary and secondary outcome measures in our study were dichotomous variables; therefore, we calculated the odds ratio (OR) as the summary measure. Individual patient was used as the unit of analysis. Information about dropouts, withdrawals, and other missing data were recorded and, if not reported, we contacted the study authors where possible. We based our analysis on intention-totreat data from the individual clinical studies. We used the Review Manager 5.3 software for data synthesis. Because of the anticipated clinical between-study heterogeneity, we used the random effects model for analysis, and the results were reported in a forest plot with 95% confidence intervals (CIs). Heterogeneity among the studies was assessed using the Cochran Q test (v 2 ). We quantified inconsistency by calculating I 2 and interpreted it using the following guide: 0% to 50% may represent low heterogeneity; 50% to 75% may represent moderate heterogeneity; and 75% to 00% may represent high heterogeneity. We planned to construct funnel plots and evaluate their symmetry to visually assess publication bias for outcomes reported by at least 0 studies. Also, in order to quantify the bias captured by the funnel plot and to formally assess reporting bias, we planned to calculate the Egger s regression intercept for outcomes reported by at least 0 studies using the Comprehensive Meta-Analysis (CMA) software (Biostat, Englewood, NJ). We planned to calculate the intercept from a linear regression of normalized effect estimate (estimate divided by its standard error) against precision (reciprocal of the standard error of the estimate). Sensitivity and sub-group analyses In order to explore potential sources of heterogeneity and assess the robustness of our results, we performed additional analyses for outcomes that were reported by at least four studies. For each outcome, we repeated the primary analysis using the random effects and fixed effect model. In addition, we calculated the risk ratio (RR) and risk difference (RD) for each dichotomous variable. We assessed the effect of each study on the overall effect size and heterogeneity by repeating the analysis after removing one study at a time. Also, we planned to perform separate analyses for RCTs with low risk of selection bias in terms of randomization and allocation concealment, and for observational studies with low or moderate risk of bias to assess the change in direction of the effect size. We also performed separate analyses for randomized trials and observational studies. Trial sequential analysis We performed trial sequential analysis for outcomes that were reported by at least four studies. In order to control the risk of type error, we adjusted the thresholds for the Z-values using O Brien-Fleming a-spending function; allowing the type I error risk to be restored to the desired maximum risk. Crossing the O Brien-Fleming a-spending boundaries by a Z-curve indicates statistical significance. 4 Moreover, we penalized the Z values according to the strength of the available evidence and the number of repeated significance tests, as defined by the law of the iterated logarithm. 4 The risk of type 2 error was controlled using the b-spending function and futility boundaries. Crossing the futility boundaries by a Z-curve indicates that the two interventions do not differ more than the anticipated intervention effect. 4 We used random effects models for the analyses. We handled the zero event trials by constant continuity correction, which involved adding a continuity correction factor () to the number of events and non events in each intervention group. A two-sided CI with 95% confidence level was used to indicate statistical significance in all analyses. We estimated the information size for the analyses based on achievement of 80% power and 20% relative risk reduction between beta-blocker and control groups. Results Searches of electronic databases identified 357 articles of which were deemed as relevant. After retrieving the full text of the relevant studies, four studies were excluded; three because they did not report the outcomes defined in our review, and another one 29 because its principal investigator had been dismissed for alleged scientific misconduct that included failing to obtain written patient consent and negligent data collection. 29 Therefore, eleven studies 5 25 were eligible for this review (Fig. ). These included three RCTs, five retrospective cohort studies, and three prospective cohort studies, enrolling a total of 32,602 patients. All patients underwent elective primary vascular or endovascular surgical procedures including abdominal aortic aneurysm repair; carotid endarterectomy; axillofemoral revascularization; and infrainguinal and suprainguinal revascularization procedures. The baseline characteristics of the included studies and study populations are demonstrated in Table and Table 2, respectively. Methodological quality and risk of bias The summary and results of methodological quality assessment of the eight observational studies and three RCTs are demonstrated graphically in Figure 2. The rationale for judgment for each Newcastle-Ottawa scale item is reported in supplementary Table. Outcome synthesis All-cause mortality. All-cause mortality was reported in seven studies, including 4,352 patients (Fig. 3). No significant difference in all-cause mortality between the beta-blocker and control groups was found (OR.0, 95% CI , P ¼ 0.77). A moderate level of heterogeneity among the studies existed (I 2 ¼70%, P ¼ 0.003). Cardiac mortality. Cardiac mortality was reported in four studies, including 2,406 patients (Fig. 3). No significant difference in the risk of cardiac mortality between the betablocker and control groups was found (OR 2.62, 95% CI , P ¼ 0.09). A low level of heterogeneity among the studies existed (I 2 ¼3%, P ¼ 0.32). Myocardial infarction (MI). MI was reported in seven studies, including 6,380 patients (Fig. 3). No significant difference in the risk of perioperative MI between the betablocker and control groups was found (OR 0.89, 95% CI , P ¼ 0.58). A low level of heterogeneity among the studies existed (I 2 ¼48%, P ¼ 0.07). Unstable angina. Unstable angina was reported in three studies, including 69 patients (Fig. 3). No significant difference in the risk of perioperative unstable angina between the

4 4 Hajibandeh et al. 357 of records identified through database searching 338 of records after duplicates removed 338 of records screened 5 of full-text articles assessed for eligibility of studies included in qualitative and quantitative synthesis 323 of records excluded 4 of full-text articles excluded: 3 Did not report the outcomes considered in our review Principal investigator had been dismissed for alleged scientific misconduct 8 9 Renal failure. Renal failure was reported in three studies, 2 including 3,256 patients (Fig. 3). No significant difference in the risk of perioperative renal failure between the beta-blocker and control groups was found (OR.48, 95% CI , P ¼ 3). A low level of heterogeneity among the studies existed (I 2 ¼42%, P ¼ 8). Composite cardiovascular events. Composite cardiovascular events were reported in two studies, 2 23 including 593 patients. No significant difference in perioperative composite cardiovascular event rate between the beta-blocker and control groups was found (OR 0.88, 95% CI , P ¼ 0.58). There was no evidence of heterogeneity among the studies (I 2 ¼ 0%, P ¼ 0.73). Rehospitalisation. Rehospitalisation was reported in one study, 2 including 496 patients (Fig. 3). No significant difference in rehospitalisation rate between the beta-blocker and control groups was found (OR 0.86, 95% CI , P ¼ 0.60). Reoperation. Reoperation was reported in one study, 2 including 496 patients (Fig. 3). No significant difference in reoperation rate between the beta-blocker and control groups was found (OR.7, 95% CI , P ¼ 0.77). Sensitivity analyses We conducted additional analyses for the outcomes that were reported by at least four studies (all-cause mortality, cardiac mortality, MI, and arrhythmias) (Supplementary Table 2). The direction of the effect size for all of the outcomes remained unchanged when RRs or RDs were calculated. Removing one study at a time did not change the direction of the effect size for any of the outcomes except for cardiac mortality. In fact, after excluding the study of Yang and colleagues 2 the difference in cardiac mortality became significant in favour of the control group [OR: 3.60, (95% CI ), P ¼ 0.006]. Moreover, the use of random-effects or fixed-effect models did not affect the direction of the effect size in any of the outcomes except cardiac mortality, where the direction of the effect size was changed in favour of the control group. The separate analyses for studies with low or moderate risk of bias did not affect the direction of the effect size. Figure Study flow diagram. beta-blocker and control groups was found (OR.34, 95% CI , P ¼ 0.63). There was no evidence of heterogeneity among the studies (I 2 ¼0%, P ¼ 0.78). Stroke. Stroke was reported in three studies, including 2,394 patients (Fig. 3). No significant difference in the risk of perioperative stroke between the beta-blocker and control groups was found (OR 2.45, 95% CI , P ¼ 0.08). A low level of heterogeneity among the studies existed (I 2 ¼ 6%, P ¼ 0.34). 5 9 Arrhythmias. Arrhythmias were reported in five studies, including 4,553 patients (Fig. 3). No significant difference in the risk of perioperative arrhythmias between the betablocker and control groups was found (OR 0.76, 95% CI , P ¼ 0.40). A high level of heterogeneity among the studies existed (I 2 ¼75%, P ¼ 0.003). Congestive heart failure (CHF). CHF was reported in three studies, including 4,373 patients (Fig. 3). No significant difference in the risk of perioperative CHF between the beta-blocker and control groups was found (OR.2, 95% CI , P ¼ 0.56). A low level of heterogeneity among the studies existed (I 2 ¼ 0%, P ¼ 0.33). Subgroup analyses Randomized controlled trials. Three RCTs, enrolling a total of 69 patients, were included. There was no significant difference in the risk of all-cause mortality (OR 0.6, 95% CI , P ¼ 0.76), cardiac mortality (OR 0.34, 95% CI -8.32, P ¼ 0.5), MI (OR 0.79, 95% CI , P ¼ 0.43), unstable angina (OR.34, 95% CI , P ¼ 0.63), stroke (OR.4, 95% CI , P ¼ 0.58), arrhythmia (OR 0.97, 95% CI , P ¼ 0.92), CHF (OR.7, 95% CI , P ¼ 0.47), renal failure (OR.44, 95% CI , P ¼ 0.54), composite cardiovascular event rate (OR 0.88, 95% CI , P ¼ 0.58), reoperation rate (OR.7, 95% CI , P ¼ 0.77), and rehospitalisation rate (OR 0.86, 95% CI , P ¼ 0.60) between the beta-blocker and control groups Observational studies. Eight observational studies, enrolling a total of 3,983 patients, were included. There was no significant difference in the risk of all-cause mortality (OR.4, 95% CI , P ¼ 0.68), MI (OR 0.93, 95% CI , P ¼ 0.80), arrhythmia (OR 0.63, 95% CI , P ¼ 0.33), CHF (OR.09, 95% CI , P ¼ 0.73), renal failure (OR.4, 95% CI , P ¼ 0.36) between the beta blocker and non-beta blocker groups. The risks of cardiac mortality (OR 3.53, 95% CI , P ¼ 0.006) and stroke (OR 5.89, 95% CI ,

5 Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery 5 Table Baseline characteristics of included studies. AAA: abdominal aortic aneurysm; RCT: Randomized Control Trial Author Yr Country Journal Type of study Procedure performed Scali 205 USA J Vasc Surg Retrospective cohort Elective open infrainguinal lower extremity bypass, aortofemoral bypass, open AAA repair Høgh 203 Denmark Eur J Vasc Endovasc Surg Retrospective cohort Primary vascular surgical or endovascular reconstruction as a result of atherosclerotic disease London 203 USA JAMA Retrospective cohort Primary vascular surgical procedures Le Manach 202 UK Anesthesiology Prospective cohort Elective infrarenal aortic reconstructive surgery USA J Cardiothorac Vasc Anesth Retrospective cohort Primary supra- and infrainguinal vascular surgery Welten 2007 Netherlands Kidney Int Retrospective cohort Elective open infrarenal AAA repair, lower limb arterial revascularization procedures Yang 2006 Canada Am Heart J RCT Abdominal aortic surgery, infrainguinal/axillofemoral revascularizations Feringa 2006 Netherlands Eur J Vasc Endovasc Surg Prospective cohort Elective AAA repair, lower extremity revascularisation, carotid artery surgery POBBLE Trial 2005 UK J Vasc Surg RCT Major elective infrarenal vascular surgery under general anaesthesia Raby 999 USA Anesth Analg RCT Aortic aneurysm repair, infrainguinal arterial bypass, carotid endarterectomy Pasternac 987 USA Circulation Prospective cohort AAA resection P ¼ 0.03) were higher in the beta-blocker group. Unstable angina, composite cardiovascular event rate, reoperation rate, and rehospitalization rate were not reported by observational studies. Considering the available data, we were not able to perform subgroup analyses based on type of vascular procedure performed, or type of beta-blocker used. Trial sequential analysis We performed trial sequential analysis for four outcomes: allcause mortality, cardiac mortality, MI, and arrhythmias. All-cause mortality. There was no significant difference in the risk of all-cause mortality between beta-blocker and control groups, as the Z-curve did not cross the alpha-spending boundaries and the absolute number for penalized Z value remained smaller than.96 in both sides. The risk of type 2 error was not significant in all-cause mortality cumulative analysis as the Z- curve crossed the futility boundaries. The information size of 3565 was reached for all-cause mortality (Supplementary Fig. ). Cardiac mortality. The risk of cardiac mortality was higher in beta-blocker group. The Z-curve did not cross the alphaspending boundaries and the absolute number for penalized Z value remained smaller than.96 in both sides. The Z-curve crossed the conventional boundary in favour of non-beta blocker after the information size was reached; therefore, the results of original analysis are not subject to type error. The information size of 986 was reached for cardiac mortality (Supplementary Fig. ). Myocardial infarction (MI). There was no significant difference in the risk of MI between beta blocker and control groups, as the Z-curve did not cross the alpha-spending boundaries and the absolute number for penalized Z value remained smaller than.96 in both sides. The risk of type 2 error was not significant in MI cumulative analysis, as the Z-curve crossed the futility boundaries. The information size of 2,339 was reached for allcause mortality (Supplementary Fig. ). Arrhythmias. There was no significant difference in the risk of arrhythmias between beta blocker and non-beta blocker groups, as the Z-curve did not cross the alpha-spending boundaries and the absolute number for penalized Z value remained smaller than.96 in both sides. The risk of type 2 error was significant in arrhythmias cumulative analysis, as the Z-curve did not cross the futility boundaries and the information size of 0,879 was not reached (Supplementary Fig. ). Discussion There is no comprehensive review and analysis on the role of beta-blocker use in vascular and endovascular surgery specifically. We performed a systematic review of the literature and meta-analysis of reported outcomes to evaluate the effect of beta blockers on perioperative outcomes in patients undergoing vascular and endovascular surgery. We included three RCTs, five retrospective cohort studies, and three prospective cohort studies, enrolling a total of 32,602 patients. The results of our analyses indicated that perioperative beta-blocker use did not reduce the risk of all-cause mortality, cardiac mortality, MI, unstable angina, stroke, arrhythmias, CHF, composite cardiovascular events, renal failure, rehospitalisation, or reoperation in vascular surgery. We used randomeffects models for analysis of the outcomes, because of the observed clinical heterogeneity among the study populations and the variability in interventional treatments and the pathology treated. Overall, a low to moderate level of between-study statistical heterogeneity was identified. The directions of the effect sizes remained consistent throughout our sensitivity analyses for all of the outcomes except for cardiac mortality, making our results statistically robust for the rest of the

6 6 Hajibandeh et al. Table 2 Baseline characteristics of included population. NR: Not reported Author No of patients Follow up Mean age Male gender Smoking Diabetes mellitus Hypertension Coronary artery disease Cerebrovsacular disease Scali months NR NR Beta-blocker: 89% No beta-blocker: 90.3% Høgh months NR Beta-blocker:52.9% No beta-blocker: 54.3% Beta-blocker:8.3% No beta-blocker: 85.3% Beta-blocker:30.8% No beta-blocker: 32.% Beta-blocker:4.6% No beta-blocker: 2.0% Betablocker:77.5% No beta-blocker: 78.8% NR NR NR Beta-blocker:5.% No beta-blocker: 6% London days NR NR NR NR NR NR NR NR Le Manach days Beta-blocker:67 No beta-blocker: 68 Beta-blocker:88% No beta-blocker: 88% NR Beta-blocker:6% No beta-blocker: 6% Beta-blocker:80% No beta-blocker: 79% Matyal days All patients: 69 All patients: 62% All patients: 9.50% All patients: 2.70% All patients: 7.0% Welten months Beta-blocker:70.7 No beta-blocker: 72.4 Yang days Beta-blocker:66.4 No beta-blocker: 65.9 Feringa 5 84 months Beta-blocker:64 No beta-blocker: 65 Brady days Beta-blocker:73 No beta-blocker: 74 Raby 26 NR Beta-blocker:69 No beta-blocker: 67 Beta-blocker:70% No beta-blocker: 72% Beta-blocker:79% No beta-blocker: 74% Beta-blocker:70% No beta-blocker: 77% Beta-blocker:75% No beta-blocker: 80% Beta-blocker:53% No beta-blocker: 36% Beta-blocker:29% No beta-blocker: 24% Beta-blocker:87.8% No beta-blocker: 90% Beta-blocker:30% No beta-blocker: 34% Beta-blocker:9% No beta-blocker: 9% Beta-blocker:2% No beta-blocker: 5% Beta-blocker:22% No beta-blocker: 4.8% Beta-blocker:2% No beta-blocker: 2% Beta-blocker:9% No beta-blocker: 8% NR Beta-blocker:20% No beta-blocker:36% Beta-blocker:7% No beta-blocker: 43% Beta-blocker:33% No beta-blocker: 29% Beta-blocker::2.3% No beta-blocker: 9.8% All patients: 0.90% All patients: 2.80% Beta-blocker:46% No beta-blocker: 27% NR Beta-blocker:5% No beta-blocker: 2% Beta-blocker:63% No beta-blocker: 3% Beta-blocker:6% No beta-blocker: 49% NR NR NR NR Beta-blocker:40% No beta-blocker: 36% Pasternac days NR NR NR NR NR NR NR Beta-blocker:6% No beta-blocker: 7% Beta-blocker:0.6% No beta-blocker: 2.0% Beta-blocker:9% No beta-blocker: 5% NR

7 Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery 7 A Random sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Other bias Brady ? Random sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Other bias Raby 999 Yang 2006?? ?? % 25% 50% 75% 00% Low risk of bias Unclear risk of bias High risk of bias B Feringa 2006 Hogh 203 Le Manach 202 London 203 Pasternac 987 Scali 205 Welten 2007 Representativeness of the exposed cohort Selection of the non-exposed cohort Ascertainment of exposure Demonstration that outcome of interest was not present at start of study Comparability of cohorts on the basis of the design or analysis Comparability of cohorts on the basis of the design or analysis 2 Assessment of outcome Was follow-up long enough for outcomes to occur Adequacy of follow up of cohorts Representativeness of the exposed cohort Selection of the non-exposed cohort Ascertainment of exposure Demonstration that outcome of interest was not present at start of study Comparability of cohorts on the basis of the design or analysis Comparability of cohorts on the basis of the design or analysis 2 Assessment of outcome Was follow-up long enough for outcomes to occur Adequacy of follow up of cohorts 0% 25% 50% 75% 00% Low risk of bias Unclear risk of bias High risk of bias Figure 2 Risk of bias summary and graph showing authors judgements about each risk of bias item for: A) Randomized trials B) Observational studies.

8 8 Hajibandeh et al. A Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Pasternac 987 Brady 2005 Yang 2006 Le Manach 202 London 203 Scali % 5.7% 3.9% 7.9% 2.6% 24.9% 2.9% Total (95% CI) % 23 7 Heterogeneity: Tau 2 =0.37; χ 2 =9.88, df=6 (P=0.003); I 2 =70% Test for overall effect: Z=0.29 (P=0.77) B.6 [0, 26.73] 2.58 [0.26, 25.72] [, 2.07] 0.39 [6, 0.9] 2.82 [.53, 5.20] 0.89 [0.6,.28].43 [0.79, 2.58].0 [0.59, 2.04] All-cause mortality Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Pasternac 987 Raby 999 Yang 2006 Le Manach %.4% 74.%.6 [0, 26.73] Not estimable 0.34 [, 8.32] 3.95 [.5, 0.34] Total (95% CI) % 5 8 Heterogeneity: Tau 2 =0.20; χ 2 =2.30, df=2 (P=0.32); I 2 =3% Test for overall effect: Z=.69 (P=0.09) C Pasternac 987 Raby 999 Brady %.5% 6.3% 2.62 [0.86, 8.05] Cardiac mortality Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr.5 [0.02,.25] 0.23 [, 6.09] 0.47 [, 2.08] Yang % 0.9 [0.48,.74] 2006 Le Manach 202 Scali % 23.5% 23.7% 0.66 [0.39,.2].03 [0.63,.70].67 [.02, 2.74] Total (95% CI) % 24 3 Heterogeneity: Tau 2 =3; χ 2 =.49, df=6 (P=0.07); I 2 =48% Test for overall effect: Z=0.55 (P=0.58) D 0.89 [0.59,.35] Myocardial Infarction Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Raby % 2.38 [0.09, 64.05] 999 Brady %.04 [0.26, 4.4] 2005 Yang % 3.06 [2, 75.5] 2006 Total (95% CI) %.34 [0.4, 4.38] 7 4 Heterogeneity: Tau 2 =0.00; χ 2 =0.50, df=2 (P=0.78); I 2 =0% Test for overall effect: Z=0.49 (P=0.63) E Unstable angina Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Brady % 2.54 [0, 63.99] 2005 Yang %.28 [0.34, 4.8] 2006 Le Manach % 5.89 [.22, 28.43] 202 Total (95% CI) % 3 6 Heterogeneity: Tau 2 =0.06; χ 2 =2.3, df=2 (P=0.34); I 2 =6% Test for overall effect: Z=.73 (P=0.08) 2.45 [0.89, 6.75] Stroke Figure 3a Forest plots of the comparisons: A) All-cause mortality; B) Cardiac mortality; C) Myocardial infarction; D) Unstable stable; E) Stroke.

9 Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery 9 F Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Pasternac % [0.03, 0.35] 987 Brady %.38 [0.49, 3.94] 2005 Yang % 0.7 [0.26,.88] % 0.99 [0.59,.69] 2008 Scali %.24 [0.9,.70] 205 Total (95% CI) % 0.76 [0.4,.43] Heterogeneity: Tau 2 =0.35; χ 2 =6.24, df=4 (P=0.003); I 2 =75% Test for overall effect: Z=0.84 (P=0.40) G Arrhythmia Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Yang %.7 [0.40, 7.23] % 0.86 [0.53,.40] 2008 Scali %.4 [0.84, 2.37] 205 Total (95% CI) %.2 [0.77,.63] Heterogeneity: Tau 2 =; χ 2 =2.23, df=2 (P=0.33); I 2 =0% Test for overall effect: Z=0.59 (P=0.56) H Congestive heart failure Beta Blocker No Beta Blocker Odds Ratio Study or Subgroup Events Total Events Total Weight Le Manach %.9 [.40, 2.60] % 0.88 [0.4,.90] Yang %.44 [0.45, 4.58] Total (95% CI) %.48 [0.90, 2.45] 8 04 Heterogeneity: Tau 2 =0.09; χ 2 =3.42, df=2 (P=8); I 2 =42% Test for overall effect: Z=.53 (P=3) I Renal failure Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Brady % 0.99 [0.43, 2.3] 2005 Yang % 0.83 [0.47,.46] 2006 Total (95% CI) % 0.88 [0.55,.40] Heterogeneity: Tau 2 =0.00; χ 2 =2, df= (P=0.73); I 2 =0% Test for overall effect: Z=0.55 (P=0.58) J Composite cardiovascular event Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yang % 0.86 [0.48,.52] Total (95% CI) Heterogeneity: Not applicable Test for overall effect: Z=0.52 (P=0.60) K % 0.86 [0.48,.52] Rehospitalisation Beta blocker No beta blocker Study or subgroup Events Total Events Total Weight Yr Yang %.7 [0.42, 3.27] 2006 Total (95% CI) Heterogeneity: Not applicable Test for overall effect: Z=0.29 (P=0.77) %.7 [0.42, 3.27] Reoperation Figure 3b F) Arrythmia; G) Congestive heart failure; H) Renal failure; I) Composite cardiovascular event; J) Rehospitalisation; and K) Reoperation. The solid squares denote the odds ratios (ORs), the horizontal lines represent the 95% confidence intervals (CIs), and the diamond denotes the pooled effect size. M-H, Mantel Haenszel test.

10 20 Hajibandeh et al. outcomes included in sensitivity analyses. Separate analyses of randomized trials showed no benefits of perioperative betablocker use for any of the outcomes. We found no reduction in all-cause mortality associated with perioperative beta-blocker use. This is consistent with results reported by other authors. Mostafaie and colleagues pooled data from two RCTs, which are also included in our review, and found no benefit of perioperative beta-blocker use in vascular surgery in terms of all-cause mortality. The incidence of perioperative cardiovascular complications and mortality may be decreasing as a result of other improvements in perioperative care, including case selection, optimization, use of statins, 30 improved postoperative care. Furthermore, improved management of perioperative cardiac complications may also have improved mortality for those who suffer complications. On this basis, it is becoming increasingly more difficult to demonstrate the efficacy of a single intervention, such as betablockade, on the outcome measures. There is a logical reason why beta-blockers may reduce the incidence of type II MIs but not type I MIs. Any reduction of type II MIs as a result of beta-blocker capacity to lower heart rate and reduce oxygen demand may, however, be offset by their tendency to cause hypotension (thus malperfusion of the myocardium or other organs). The efficacy of beta-blockers may thus be negatively balanced by their safety profile. We found no difference in the risk of MI between beta blocker and control groups. Consistent with our finding, Mostafaie and colleagues reported no reduction in MI in vascular surgery patients treated with beta blockers. This is contradictory to results reported in non-cardiac surgery. This suggests that the prognostic effect of perioperative beta-blocker use may vary according to type of noncardiac surgery (vascular vs non-vascular). Our results showed no difference in risk of stroke between the beta-blocker and control groups in vascular surgery. Consistent with our results, Mostafaie and colleagues did not find any effect on risk of stroke with beta-blockers. The effect of perioperative beta-blockers on postoperative renal failure in vascular surgery has not been assessed by previous systematic reviews in vascular or non-vascular noncardiac surgeries. Our results suggest that perioperative beta blockers do not affect the risk of perioperative renal failure. The definition of renal failure was consistent among the included studies (defined as an increase in serum creatinine of.8 mg dl or greater). Considering that only three studies reported renal failure as an outcome and that the potential mechanisms of renal impairment vary among different vascular and endovascular interventions, future studies are required to further evaluate a potential association. It is possible that beta-blockade may still be efficacious in particular subgroups (for example, those with severe IHD), or that specific drugs (cardioselective) and/or regimens (long leading period before surgery, intensive perioperative monitoring to avoid hypotension) may be more effective and safer. This cannot be ascertained by our meta-analysis. Clinicians should not initiate beta-blockers in patients who do not have an independent indication to this treatment, for the sole purpose of managing perioperative risk. Patients on betablockers may require extended perioperative monitoring in ICU/ HDU to prevent hypotension-related complications. The reported outcomes of our review and analysis should be viewed and interpreted in the context of inherent limitations. Most of the included studies in our review were nonrandomized observational studies, which are inevitably subject to selection bias. Most of the outcomes were reported by less than five studies, not providing a robust basis for definite conclusions. None of the outcomes were reported by more than 0 studies and we were not able to formally assess the publication bias; therefore, reporting bias cannot be excluded in our study. The risk of bias in the included observational studies, which was assessed using NOS, might have been underestimated as traditional quality assessment tools for observational studies have been found to underestimate the risk of bias in nonrandomized studies, compared with more sophisticated tools. 3 Some of the included studies in our review included few participants; occurrence of few outcome events in these studies might have led to imprecise effects estimates. This is reflected in the wide confidence intervals for some of the outcomes, such as cardiac mortality, stroke, unstable angina, and reoperation. Moreover, lack of outcome data stratified by type or duration of beta-blocker or type of vascular procedure limits the applicability of our findings for these subgroups. Conclusions Our analysis demonstrated that perioperative use of betablockers does not reduce the risk of cardiovascular and/or cerebrovascular complications. Perioperative use of beta-blockers should not be routinely used in vascular and endovascular surgery. Further studies should be performed to investigate effect of perioperative beta-blocker in selected groups such as patients with known ischaemic heart disease with preserved left ventricular function. Authors contributions Shahab H. and Shahin H. equally contributed to this paper and joined first authorship is proposed. Study design/planning: S.H., S.H., G.A.A., F.T. Study conduct: S.H., S.H. Data analysis: S.H., S.H., S.A.A., F.T., G.A.A. Writing paper: S.H., S.H. Revising paper: all authors Supplementary material Supplementary material is available at British Journal of Anaesthesia online. Declaration of interest None declared. References. Eagle KA, Coley CM, Newell JB, et al. Combining clinical and thallium data optimizes preoperative assessment of cardiac risk before major vascular surgery. Ann Intern Med 989; 0: Fleischmann KE, Goldman L, Young B, Lee TH. Association between cardiac and noncardiac complications in patients undergoing noncardiac surgery: outcomes and effects on length of stay. Am J Med 2003; 5: Abrams J. Clinical practice. Chronic stable angina. N Engl J Med 2005; 352: 2524e33 4. Law MR, Morris JK, Wald NJ. Use of blood pressure lowering drugs in the prevention of cardiovascular disease: metaanalysis of 47 randomized trials in the context of

11 Effect of beta-blockers on perioperative outcomes in vascular and endovascular surgery 2 expectations from prospective epidemiological studies. Br Med J 2009; 338: b Foody JM, Farrell MH, Krumholz HMB. blocker therapy in heart failure: scientific review. JAMA 2002; 287: Farrell MH, Foody JM, Krumholz HMB. blockers in heart failure: clinical applications. JAMA 2002; 287: Packer M, Coats AJ, Fowler MB, et al. Carvedilol Prospective Randomized Cumulative Survival Study Group. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med 200; 344: Fleisher LA, Beckman JA, Brown KA, Calkins H, Chaikof EL, Fleischmann KE, et al. ACC/AHA 2007 guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery. J Am Coll Cardiol 2007; 50: e Wijeysundera DN, Duncan D, Nkonde-Price C, et al. Perioperative beta blockade in noncardiac surgery: a systematic review for the 204 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. Circulation 204; 30: Blessberger H, Kammler J, Domanovits H, Schlager O, Wildner B, Azar D, et al. Perioperative beta-blockers for preventing surgery-related mortality and morbidity. Cochrane Database Syst Rev 204; 9: CD Mostafaie K, Bedenis R, Harrington D. Beta-adrenergic blockers for perioperative cardiac risk reduction in people undergoing vascular surgery. Cochrane Database Syst Rev 205; : CD Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. Br Med J 2009; 339: b Wells GA, Shea B, O connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomized studies in meta-analyses. Available at: clinical_epidemiology/oxford.asp. (accessed May 5, 204) 4. Thorlund K, Engstrøm J, Wetterslev J, Brok J, Imberger G, Gluud C, User manual for trial sequential analysis (TSA). Copenhagen Trial Unit, Centre for Clinical Intervention Research, Copenhagen, Denmark. 20. p. -5. Available from (accessed July 0, 206) 5. Scali S, Patel V, Neal D, et al. Preoperative b-blockers do not improve cardiac outcomes after major elective vascular surgery and may be harmful. J Vasc Surg 205; 62: e2 doi: 06/j.jvs Høgh A, Lindholt JS, Nielsen H, Jensen LP, Johnsen SP. Betablocker use and clinical outcomes after primary vascular surgery: a nationwide propensity score-matched study. Eur J Vasc Endovasc Surg 203; 46: doi: 06/j.ejvs London MJ, Hur K, Schwartz GG, Henderson WG. Association of perioperative b-blockade with mortality and cardiovascular morbidity following major noncardiac surgery. JAMA 203; 309: doi: 00/jama Erratum in: JAMA. 205;34(2): Le Manach Y, Collins GS, Ibanez C, et al. Impact of perioperative bleeding on the protective effect of b-blockers during infrarenal aortic reconstruction. Anesthesiology 202; 7: 203 doi: 097/ALN.0b03e3825adaea. 9. Matyal R, Mahmood F, Panzica P, et al. Sex-related differences in outcome after high-risk vascular surgery after the administration of beta-adrenergic-blocking drugs. J Cardiothorac Vasc Anesth 2008; 22: doi: 053/ j.jvca Welten GM, Chonchol M, Hoeks SE, et al. Beta-blockers improve outcomes in kidney disease patients having noncardiac vascular surgery. Kidney Int 2007; 72: Epub 2007 Sep 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; 52: Feringa HH, Bax JJ, Schouten O, et al. Beta-blockers improve in-hospital and long-term survival in patients with severe left ventricular dysfunction undergoing major Vascular Surgery. Eur J Vasc Endovasc Surg 2006; 3: Brady AR, Gibbs JS, Greenhalgh RM, Powell JT, Sydes MR. POBBLE trial investigators. Perioperative beta-blockade (POBBLE) for patients undergoing infrarenal vascular surgery: results of a randomized double-blind controlled trial. J Vasc Surg 2005; 4: Raby KE, Brull SJ, Timimi F, et al. The effect of heart rate control on myocardial ischemia among high-risk patients after vascular surgery. Anesth Analg 999; 88: Pasternack PF, Imparato AM, Baumann FG, et al. The hemodynamics of beta-blockade in patients undergoing abdominal aortic aneurysm repair. Circulation 987; 76: III Fleisher LA, Corbett W, Berry C, Poldermans D. Cost-effectiveness of differing perioperative beta-blockade strategies in vascular surgery patients. J Cardiothorac Vasc Anesth 2004; 8: Katznelson R, Djaiani G, Mitsakakis N, et al. Delirium following vascular surgery: increased incidence with preoperative beta-blocker administration. Can J Anaesth 2009; 56: doi: 007/s Yeager RA, Moneta GL, Edwards JM, Taylor LM Jr, McConnell DB, Porter JM. Reducing perioperative myocardial infarction following vascular surgery. The Potential Role of Beta-Blockade. Arch Surg 995; 30: discussion Poldermans D, Boersma E, Bax JJ, et al. Bisoprolol reduces cardiac death and myocardial infarction in high-risk patients as long as 2 years after successful major vascular surgery. Eur Heart J 200; 22: Antoniou GA, Hajibandeh S, Hajibandeh S, Vallabhaneni SR, Brennan JA, Torella F. Meta-analysis of the effects of statins on perioperative outcomes in vascular and endovascular surgery. J Vasc Surg 205; 6: e 3. Bilandzic A, Fitzpatrick T, Rosella L, Henry D. Risk of bias in systematic reviews of non-randomized studies of adverse cardiovascular effects of thiazolidin ediones and cyclooxygenase-2 inhibitors: application of a new cochrane risk of bias tool. PLoS Med 206; 3: e00987 Handling editor: P. S. Myles