Calcium Channel Blockers for Reducing Cardiac Morbidity After Noncardiac Surgery: A Meta-Analysis Duminda N. Wijeysundera,

MD*,

and W. Scott Beattie,

MD, PhD, FRCPC†

From the *Department of Anesthesia, University of Toronto, and the †Department of Anesthesia, University Health Network, University of Toronto, Toronto, ON

Cardiac complications are the leading cause of death after noncardiac surgery. Despite theoretical benefits, calcium channel blockers (CCB) are not widely used in the perioperative setting. This systematic review assessed the efficacy of CCBs during noncardiac surgery. MEDLINE, EMBASE, Science Citation Index, PubMed, and reference lists were searched without language restriction for randomized controlled trials (RCT) evaluating CCBs during noncardiac surgery. Two reviewers independently abstracted data on death, myocardial infarction (MI), ischemia, supraventricular tachyarrhythmia (SVT), and congestive heart failure (CHF). Treatment effects were calculated as relative risks (RR) with 95% confidence intervals (CI). Eleven studies (1007 patients) were included. CCBs significantly reduced ischemia (RR, 0.49; 95% CI, 0.30 – 0.80; P ⫽

C

ardiac complications are the leading cause of perioperative death after noncardiac surgery (1). Perioperative myocardial infarction (MI) is predominately silent, non-Q-wave, and preceded by long periods of myocardial ischemia (2,3). Perioperative ischemia is itself mediated in part through imbalance between myocardial oxygen supply and demand. Calcium channel blockers (CCB) theoretically improve this balance through their negative inotropic, negative chronotropic, afterload reducing, and coronary vasodilator properties. Despite these benefits, CCBs are not widely used in the perioperative setting. The basis for this under-use may be observational studies showing that CCBs have no beneficial effect on perioperative cardiac complications (4 –7). However, these analyses were generally small (fewer than 500 patients), and incorporated minimal (6,7) to no (4,5) risk-adjustment techniques. Risk adjustment in these Presented at the Annual Scientific Sessions of the American Heart Association, Chicago, IL, November 18, 2002. Accepted for publication May 29, 2003. Address correspondence and reprint requests to Dr. W. Scott Beattie, MD, PhD, Associate Professor, Department of Anesthesia, University of Toronto, EN 3– 453, Toronto General Hospital, 200 Elizabeth Street, Toronto, ON, M5G 2C4, Canada. Address email to [email protected].

634

Anesth Analg 2003;97:634–41

0.004) and SVT (RR, 0.52; 95% CI, 0.37– 0.72; P ⬍ 0.0001). CCBs were associated with trends towards reduced death and MI. In post hoc analyses, CCBs significantly reduced death/MI (RR, 0.35; 95% CI, 0.15– 0.86; P ⫽ 0.02) and major morbid events (MME), defined as death, MI, or CHF (RR, 0.39; 95% CI, 0.17– 0.89; P ⫽ 0.02). In subgroup analyses, diltiazem significantly reduced ischemia, SVT, death/MI, and MMEs. This meta-analysis shows CCBs significantly reduced ischemia, SVT, and combined end-points in the setting of noncardiac surgery. The majority of these benefits are attributable to diltiazem, suggesting the need for further evaluation of this drug in a large RCT. (Anesth Analg 2003;97:634 –41)

observational studies is important. Patients on CCBs are more likely to have specific comorbidities that may increase postoperative complications, including diabetes mellitus, peripheral vascular disease, and chronic obstructive pulmonary disease (8). The failure to fully adjust for such comorbidities may prevent identification of the potential benefits of CCBs. Indeed, in contrast to observational studies, meta-analyses of randomized controlled trials (RCT) in cardiac surgery have shown CCBs to significantly reduce MI, myocardial ischemia, and supraventricular tachyarrhythmias (SVT) (non-dihydropyridines only) (9). An adequate assessment of the effects of CCBs during noncardiac surgery will therefore necessitate an appropriately powered RCT. As a prelude to such a study, we conducted a systematic review and meta-analysis of RCTs that evaluated CCBs during noncardiac surgery for the prevention of cardiac complications.

Methods We conducted our systematic review according to the Quality of Reporting of Meta-analyses (10) recommendations for improving the quality of meta-analyses. ©2003 by the International Anesthesia Research Society 0003-2999/03

ANESTH ANALG 2003;97:634 –41

Eligible studies were published RCTs that evaluated CCBs (administered immediately preoperatively, intraoperatively, or postoperatively within 48 h) during noncardiac surgery, and reported any of the following outcomes: death, MI, ischemia, or SVT. Acceptable definitions for ischemia were ST segment deviation on an electrocardiogram or new wall-motion abnormalities on a transesophageal echocardiogram. SVTs included atrial fibrillation, atrial flutter, and supraventricular tachycardia. We did not use a standard definition of perioperative MI given the lack of standardized criteria in the literature. Studies were excluded if they exclusively recruited prior organ transplant recipients, individuals younger than 18 yr, patients who had already developed SVT, or patients undergoing surgery for subarachnoid hemorrhage. When required, authors of included studies were contacted to provide additional data. We identified published RCTs by searching MEDLINE (1966 to December 2001) for [Calcium channel blockers AND (Postoperative complications OR Perioperative care OR Intraoperative complications)] and EMBASE (1980 to December 2001) for [Calcium channel blocking agent AND (Postoperative Complication OR Postoperative period OR Perioperative period OR Intraoperative period OR Perioperative care OR Perioperative complication)] without language restriction. Titles and abstracts were screened to exclude obviously ineligible studies. Two reviewers independently read the remaining papers in full to determine final eligibility. Reasons for exclusion were documented for all excluded studies. Bibliographies were surveyed to identify any further eligible papers. Included papers were entered into the Science Citation Index and PubMed (Related Articles search) to identify other relevant studies. The reviewers evaluated the quality of included studies with regard to the adequacy of randomization, allocation concealment, blinding, and handling of dropouts. The following items were independently abstracted by two reviewers onto standardized data collection forms: patients, surgery, treatments, death, MI, ischemia, SVT, congestive heart failure, prior medications (␤-blockers), hypotension, and bradycardia. Bradycardia and hypotension were defined by the need for pharmacological intervention. When possible, we abstracted data only for comparisons of CCBs against placebo or nitrates. All disagreements were resolved by consensus. Statistical analyses were performed using RevMan 4.2 (Cochrane Collaboration, Oxford, UK). Treatment effects were expressed as pooled relative risks (RR) with 95% confidence intervals (CI). Initially, we assessed for heterogeneity using the Q-statistic, with the cutoff for statistically significant heterogeneity set at P ⬍ 0.1. Statistically significant heterogeneity was defined as greater variation between the results of trials

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

635

than would be expected by chance (assuming a single underlying treatment effect for all included trials). In the absence of statistically significant heterogeneity, pooled RRs were calculated under the fixed effects model. If there was statistically significant heterogeneity, the random effects model was used instead; in addition, we performed post hoc analyses to explain the observed heterogeneity. Statistical significance for treatment effects was defined by P ⬍ 0.05. In the primary analyses, pooled RRs were calculated for the effects of CCBs on death, MI, ischemia, SVT, and adverse events (e.g., congestive heart failure, hypotension, bradycardia). Several secondary analyses were planned a priori. We calculated treatment effects of each CCB class (diltiazem, verapamil, dihydropyridines) on mortality, MI, ischemia, and SVT. Given that prior medication use may influence perioperative outcomes, we also compared prior ␤-blocker use in the CCB and control arms using meta-analytic methods. Medication use differences were expressed as pooled RRs under the fixed effects model. Given the relatively small total number of outcome events in this meta-analysis, we performed post hoc analyses that used combined outcomes: death and/or MI, and major morbid events (MME). A MME was defined as a combined outcome of death, MI, or congestive heart failure. The results of individual RCTs may have been biased if study subjects who dropped out after randomization were excluded from analyses. We therefore performed a series of sensitivity analyses to quantify the influence by such violations of the intention-totreat principle. In cases where dropouts were excluded from analyses, we repeated the meta-analyses after imputing values for the dropouts, using both best-case and worst-case scenarios. Several additional sensitivity analyses were also planned a priori. The first sensitivity analysis assessed the relationship between study quality and estimated treatment effects. The meta-analyses were repeated in subgroups of double-blinded studies. The second sensitivity analysis examined the influence of statistical model on estimated treatment effects. Analyses that used the fixed effects model were repeated using the random effects model.

Results Eleven studies (11–21), encompassing 1007 patients, were included (Table 1). Six were double blinded. The search results are presented in Figure 1. Eight studies evaluated diltiazem, two evaluated verapamil, and two evaluated dihydropyridines. A single study incorporated three arms: control, diltiazem, and nifedipine (21). Of the 11 studies, only 2 were conducted in North America. Approximately 55% were published in the English language.

636

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

ANESTH ANALG 2003;97:634 –41

Table 1. Characteristics of Included Studies Study

Drug & Dose

Shimada (17) Verapamil Lindgren (18)

van Mieghem (19)

Dihydropyridines du Toit (20)

Multiple Treatment Arms Retamal (21)

Double- Concealed Blinded Allocation Language

n

Drop-Outs

Placebo

Major noncardiac surgery

66

0

Ischemia

⫹

⫺

French

France

Control

Major urologic procedures

58

0

Ischemia

⫺

⫺

Greek

Greece

Placebo

Major noncardiac surgery

60

14

Ischemia

⫹

⫺

Spanish

Spain

Placebo

Vascular surgery

30

0

Ischemia

⫹

⫺

English

France

Digoxin 1 mg IV total dose over 24 hours, then 0.125– 0.25 mg orally once daily for 30 days Placebo

Pneumonectomy

70

0

SVT

⫺

⫺

English

USA

1.7 ␮g 䡠 kg⫺1 䡠 min⫺1 IV postoperatively for 18–24 h, then 120 mg orally twice daily for 14 days. 1–3 ␮g 䡠 kg⫺1 䡠 min⫺1 Control IV postoperatively.

Pneumonectomy 336 and lobectomy

6

SVT

⫹

⫹

English

USA

Esophagectomy

52

0

SVT

⫺

⫺

Japanese

Japan

0.17 Placebo ␮g 䡠 kg⫺1 䡠 min⫺1 IV postoperatively in recovery room, then 80 mg orally every 8 h for 7 days. 0.125 mg/min IV Placebo postoperatively for 3 days.

Thoracotomy

25

SVT

⫹

⫺

English

Finland

5 ␮g 䡠 kg⫺1 䡠 min⫺1 IV intraoperatively until 3–12 h after surgery. Hatziantoniou (12) 3.3 ␮g 䡠 kg⫺1 䡠 min⫺1 IV, then 240 mg orally once daily while in-hospital. Garcia-Guasch (13) 3 ␮g 䡠 kg⫺1 䡠 min⫺1 IV intraoperatively until 3 h after surgery. Godet (14) 3 ␮g 䡠 kg⫺1 䡠 min⫺1 IV intraoperatively until 2 h after surgery. Amar (15) 10 mg IV postoperatively every 4 hours for 24–36 h, then 180– 240 mg orally once daily for 1 mo

Amar (16)

Primary Outcome

Surgery Type

Diltiazem Caramella (11)

Control

Site

Pneumonectomy 199 and lobectomy

0

SVT

⫺

⫺

English

Belgium

Nifedipine 10 mg every 8 h for 3 days, beginning 2 days before surgery.

Placebo

Total hip replacement

50

0

Ischemia

⫹

⫺

English

South Africa

Treatment arm 1: diltiazem 3 ␮g 䡠 kg⫺1 䡠 min⫺1 IV postoperatively until extubation. Treatment arm 2: intranasal nifedipine 10 mg postoperatively

Control

Carotid 63 endarterectomy

0

Perioperative hypertension

⫺

⫺

French

France

SVT ⫽ supraventricular tachyarrhythmia; IV ⫽ intravenous. n represents the total number randomized. “Drop-Outs” represents the number withdrawn after randomization.

The number of studies reporting mortality was 5 (Fig. 2), with an overall incidence among 692 patients of 2% (n ⫽ 17). The observed reduction in mortality did not achieve statistical significance (RR, 0.40; 95%

CI, 0.14 –1.16; P ⫽ 0.09). There was no statistically significant heterogeneity in this analysis (P ⫽ 0.54). Six papers reported MI (Fig. 2), with an incidence among 486 patients of 1% (n ⫽ 5). CCBs caused a

ANESTH ANALG 2003;97:634 –41

Figure 1. Flow diagram of meta-analysis.

reduction in MI that did not achieve statistical significance (RR, 0.25; 95% CI, 0.05–1.18; P ⫽ 0.08) without statistically significant heterogeneity (P ⫽ 0.99). Six papers reported ischemia (Fig. 3), with an incidence of 21% (n ⫽ 54) among 263 patients. Overall, CCBs significantly reduced perioperative ischemia (RR, 0.49; 95% CI, 0.30 – 0.80; P ⫽ 0.004) with borderline statistical heterogeneity (P ⫽ 0.10). Subgroup analyses explained this heterogeneity. Diltiazem significantly reduced ischemia (RR, 0.34; 95% CI, 0.18 – 0.63; P ⫽ 0.0005) in a subgroup analysis with reduced heterogeneity (P ⫽ 0.39). In contrast, the single dihydropyridine study, which used short-acting sublingual nifedipine, showed an increase in ischemia that did not achieve statistical significance (RR, 1.85; 95% CI, 0.64 –5.35; P ⫽ 0.26). The single verapamil study showed a trend towards reduced ischemia (RR, 0.15; 95% CI, 0.01–2.70; P ⫽ 0.20). Five studies reported perioperative SVT (Fig. 4), with an incidence of 19% (n ⫽ 127) among 680 patients. CCBs significantly reduced SVT (RR, 0.52; 95% CI, 0.37– 0.72; P ⬍ 0.0001) without statistically significant heterogeneity (P ⫽ 0.65). Both diltiazem (RR, 0.54; 95% CI, 0.37– 0.78; P ⬍ 0.001) and verapamil (RR, 0.44; 95% CI, 0.20 – 0.94; P ⫽ 0.03) reduced SVT. There was no statistically significant heterogeneity in the subgroup analyses for diltiazem (P ⫽ 0.49) or verapamil (P ⫽ 0.31). Four studies reported hypotension (11,13,14,19), with an incidence of 8% (n ⫽ 28) among 341 patients. CCBs were associated with trends towards increased hypotension (RR, 1.74; 95% CI, 0.28 –10.81; P ⫽ 0.55); however, this estimate had both wide confidence intervals and statistically significant heterogeneity (P ⫽ 0.05). In subgroup analyses, diltiazem (11,13,14) was not associated with increased hypotension (RR, 0.99;

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

637

95% CI, 0.39 –2.49; P ⫽ 1) without statistically significant heterogeneity (P ⫽ 0.8). In contrast, the single verapamil study (19) found a statistically significant increase in perioperative hypotension (RR, 28.71; P ⫽ 0.02). Four studies reported bradycardia (13,14,16,19), with an incidence of 6% (n ⫽ 35) among 605 patients. CCBs caused an increase in bradycardia that did not achieve statistical significance (RR, 3.32; 95% CI, 0.70 – 15.66; P ⫽ 0.13). This estimate had statistically significant heterogeneity (P ⫽ 0.07). In subgroup analyses, diltiazem was associated with trends towards increased bradycardia (RR, 1.76; 95% CI, 0.61–5.07; P ⫽ 0.3) without statistically significant heterogeneity (P ⫽ 0.27). The single verapamil study (19) found a statistically significant increase in perioperative bradycardia (RR, 18.81; P ⫽ 0.04). Six studies reported congestive heart failure, with an incidence of 0.4% (n ⫽ 3) among 720 patients. CCBs were not associated with statistically significant changes in congestive heart failure (RR, 0.60; 95% CI, 0.08 – 4.81; P ⫽ 0.63). There was no statistically significant heterogeneity for this analysis (P ⫽ 0.63). CCBs significantly reduced MME (RR, 0.39; 95% CI, 0.17– 0.89; P ⫽ 0.02) without significant heterogeneity (P ⫽ 0.91). In subgroup analyses, diltiazem significantly reduced major morbid events (RR, 0.31; 95% CI, 0.11– 0.88; P ⫽ 0.03) without statistically significant heterogeneity (P ⫽ 0.91). When death and MI were combined into a single outcome, CCBs significantly reduced death/MI (RR, 0.35; 95% CI, 0.15– 0.86; P ⫽ 0.02) without statistically significant heterogeneity (P ⫽ 0.90). In subgroup analyses, diltiazem significantly reduced death/MI (RR, 0.26; 95% CI, 0.08 – 0.83; P ⫽ 0.02). There was no statistically significant heterogeneity for this analysis (P ⫽ 1). Four studies reported preoperative ␤-blocker use (11,14,16,21) with an overall prevalence of 13% (n ⫽ 62) among 493 patients. Beta-blocker use was a specific exclusion criterion for three studies (15,18,19). Patients assigned to CCB arms were as likely to receive ␤-blockers as patients assigned to control arms (RR, 1.03; 95% CI, 0.65–1.63; P ⫽ 0.90). There was no statistically significant heterogeneity for this analysis (P ⫽ 0.93). A single included study, which reported effects on ischemia alone, required that all antianginal and antihypertensive medications be held for 24 h before surgery (13). Exclusion of this study did not qualitatively alter the overall effects of CCBs on ischemia (RR, 0.56; 95% CI, 0.34 – 0.93; P ⫽ 0.03). Two studies excluded selected randomized patients from the final analyses (13,16). In a 2000 publication,

638

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

ANESTH ANALG 2003;97:634 –41

Figure 2. Effect of calcium channel blockers (CCBs) on mortality (above) and myocardial infarction (MI) (below), with a combined analysis of these results. Squares represent point estimates. The area of a square correlates with its contribution towards the weighted summary estimate. Horizontal lines denote 95% confidence intervals (CI), some of which extend beyond the limits of the scale. The diamonds represent overall summary estimates. RR, relative risk.

Figure 3. Effect of calcium channel blockers (CCBs) on myocardial ischemia, with a combined analysis of these results. Squares represent point estimates. The area of a square correlates with its contribution towards the weighted summary estimate. Horizontal lines denote 95% confidence intervals (CI), some of which extend beyond the limits of the scale. The diamond represents the overall summary estimate. RR, relative risk.

Amar et al. (16) excluded six patients after randomization. Two excluded patients had been incorrectly enrolled despite the presence of exclusion criteria. The remaining four patients met stipulated exclusion criteria after randomization but before diltiazem or placebo were administered in a double-blinded manner. Given the nature of these exclusions, the reviewers concluded that Amar et al. had adhered to a reasonable intentionto-treat analysis.

After randomization, Garcı´a-Guasch et al. (13) excluded 14 of 60 enrolled patients; none met the exclusion criteria that had been defined a priori. As Garcı´aGuasch et al. only reported treatment effects on ischemia, we recalculated pooled effects on ischemia after imputing values for the excluded patients (using best-case and worst-case scenarios). In the best-case scenario, the reduction of perioperative ischemia improved in magnitude (RR, 0.25; 95% CI, 0.07– 0.88; P ⫽

ANESTH ANALG 2003;97:634 –41

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

639

Figure 4. Effect of calcium channel blockers (CCBs) on supraventricular tachyarrhythmia (SVT), with a combined analysis of these results. Squares represent point estimates. The area of a square correlates with its contribution towards the weighted summary estimate. Horizontal lines denote 95% confidence intervals (CI), some of which extend beyond the limits of the scale. The diamond represents the overall summary estimate. RR, relative risk.

0.03). In the worst-case scenario, the reduction of ischemia did not change in magnitude, but instead lost statistical significance (RR, 0.50; 95% CI, 0.19 –1.33; P ⫽ 0.17). The overall treatment effects of CCBs were generally sensitive to changes in study quality and statistical model (Table 2). Treatment effects on myocardial ischemia and SVT were associated with the most robust overall estimates.

Discussion This systematic review demonstrates that CCB use during noncardiac surgery results in a significantly reduced myocardial ischemia and SVT. When the major clinical end-points of congestive heart failure, MI, and death were combined, there was also a significant reduction with CCB use. Reduction in perioperative ischemia is clinically important. Postoperative myocardial ischemia may be silent and clinically unrecognized, and may take place without aberrations in hemodynamics (22,23). It is associated with a fourfold increase in postoperative MI. The duration of the ischemia is itself an independent predictor of MI (2,3). The reduction of perioperative SVT by CCBs is also clinically important. SVT after noncardiac surgery is associated with increased MI, stroke, and postoperative length-of-stay (24). Although CCBs were associated with trends towards reduced MI and death, this review lacked sufficient statistical power to show significant effects on these clinical end-points. Most existing RCT evidence points to a beneficial effect, specifically from diltiazem. Eight of the 11 included studies assessed diltiazem, predominantly at IV doses between 1 and 5 ␮g · kg⫺1 · min⫺1. Diltiazem was found to significantly reduce ischemia, SVT, death/MI, and MME without any statistically significant increase in perioperative hypotension or bradycardia. The lack of stronger evidence for other classes

of CCBs is likely a result of the paucity of relevant RCTs. However, the available data do suggest limitations for short-acting oral nifedipine and verapamil. The former was associated with nonsignificant increases in perioperative ischemia (20), echoing findings in the nonsurgical literature (25). Although verapamil reduced perioperative SVT, it was associated with significantly increased incidences of hypotension and bradycardia. The major limitations to this review are the number and quality of included studies. Given the relatively small number of outcome events identified, the effects of CCBs on perioperative outcomes were generally not robust. Furthermore, the current study lacks sufficient statistical power to adequately examine the effects of CCBs on specific outcomes (e.g., death, MI, congestive heart failure, bradycardia, hypotension). With a mortality rate in the control arm of 2%, an individual RCT will need at least 4636 subjects to detect a 50% RR reduction (80% power). The included studies were also generally of small to moderate quality. Just over half the included studies were double blinded, and only one provided explicit description of allocation concealment. The quality of RCTs is known to influence the estimated treatment effects (26). Aside from demonstrating CCBs’ potential for improving clinical outcomes, this review raises important questions regarding prior research in this area. The published worldwide experience of RCTs of perioperative CCB use is relatively small, at only 1007 patients. Furthermore, there is a suggestion that geographic bias exists against research in this field. Nine of the eleven studies included in our analyses were conducted outside North America. Five studies, of which three reported favorable effects on perioperative MI (11,12,21), were not published in the English language. Any search strategy restricted to English language studies (27) would therefore have missed these positive studies. This bias is confirmed by the

640

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

ANESTH ANALG 2003;97:634 –41

Table 2. Influence of Study Quality and Statistical Model on Estimated Treatment Effects

Study Quality All Double-Blinded All

Statistical Model Fixed Effects Fixed Effects Random Effects

Death RR (95% CI) 0.40 (0.14–1.16) 0.32 (0.03–3.12) 0.43 (0.14–1.29)

MI RR (95% CI)

Ischemia RR (95% CI)

SVT RR (95% CI)

0.25 (0.05–1.18) 0.49 (0.30–0.80) 0.51 (0.37–0.71) 0.21 (0.01–4.26) 0.56 (0.34–0.93) 0.58 (0.37–0.90) 0.25 (0.05–1.19) 0.47 (0.21–1.05) 0.52 (0.38–0.73)

Death and/or MI RR (95% CI)

MME RR (95% CI)

0.35 (0.15–0.86) 0.28 (0.05–1.66) 0.37 (0.15–0.93)

0.39 (0.17–0.89) 0.39 (0.09–1.64) 0.41 (0.17–0.98)

CI ⫽ confidence interval; RR ⫽ relative risk; SVT ⫽ supraventricular tachyarrhythmias; MME ⫽ major morbid events (death, MI, or congestive heart failure); MI ⫽ myocardial infarction; CHF ⫽ congestive heart failure.

experience at our North American institution, which has seen a steady reduction in preoperative CCB use over the past three years (8). This bias likely stems from the results of previous publications. Short-acting oral nifedipine has been associated with increased mortality among patients with coronary artery disease in the nonsurgical setting (25). As mentioned previously, several observational studies have shown no perioperative benefit for patients receiving CCBs (4 –7). These observational studies were, however, small and they incorporated minimal or no risk adjustment techniques. Adequate adjustment for preoperative medical status is a critical issue for observational studies. We recently demonstrated (8) that among cardiac surgical patients, CCB use was associated with increased age, more severe coronary artery disease, and more comorbidities (e.g., diabetes mellitus, peripheral vascular disease, renal dysfunction, chronic obstructive pulmonary disease). Furthermore, CCB use was associated with reduced use of ␤-blockers, which improve perioperative outcomes (28,29). The results of this systematic review suggest that CCBs may be beneficial and that the previous perceptions are not justified. Previous research has shown that ␤-blockers also reduce perioperative ischemia, in-hospital death, and long-term morbidity (28,29). The positive effects of CCBs seen in this meta-analysis were not attributable to concurrent ␤-blocker use. Beta-blocker use in the included studies was minimal (13%) and equally distributed between the CCB and control arms. The limited concurrent use of ␤-blockers and CCBs does mean that this review cannot provide insight into the safety or efficacy of combining the two drugs. Importantly, many high-risk patients either have contraindications (absolute or relative) to ␤-blockers or are intolerant of them. Furthermore, subpopulations of high-risk patients continue to have relatively frequent cardiac events despite receiving ␤-blocker therapy (30). The search for therapies to reduce perioperative cardiac complications should therefore not be limited to any single drug. This meta-analysis shows that CCBs, specifically diltiazem, reduce perioperative ischemia and further suggests that they may reduce major cardiac events (MME, death/MI). Contrary to a seeming North American bias, this review

indicates that the addition of diltiazem to the perioperative treatment regimen of high-risk patients deserves further evaluation in a well-designed, adequately powered, double-blinded RCT (31).

References 1. Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990;72:153– 84. 2. Landesberg G, Luria MH, Cotev S, et al. Importance of longduration postoperative ST-segment depression in cardiac morbidity after vascular surgery. Lancet 1993;341:715–9. 3. Fleisher LA, Rosenbaum SH, Nelson AH, Barash PG. The predictive value of preoperative silent ischemia for postoperative ischemic cardiac events in vascular and nonvascular surgery patients. Am Heart J 1991;122:980 – 6. 4. Chung F, Houston PL, Cheng DC, et al. Calcium channel blockade does not offer adequate protection from perioperative myocardial ischemia. Anesthesiology 1988;69:343–7. 5. Slogoff S, Keats AS. Does chronic treatment with calcium entry blocking drugs reduce perioperative myocardial ischemia? Anesthesiology 1988;68:676 – 80. 6. Sear JW, Foex P, Howell SJ. Effect of chronic intercurrent medication with beta-adrenoceptor blockade or calcium channel entry blockade on postoperative silent myocardial ischaemia. Br J Anaesth 2000;84:311–5. 7. Sear JW, Howell SJ, Sear YM, et al. Intercurrent drug therapy and perioperative cardiovascular mortality in elective and urgent/emergency surgical patients. Br J Anaesth 2001;86: 506 –12. 8. Wijeysundera DN, Beattie WS, Rao V, et al. Calcium antagonists are associated with reduced mortality after cardiac surgery: a propensity analysis. J Thorac Cardiovasc Surg. In press. 9. Wijeysundera DN, Beattie WS, Rao V, Karski J. Calcium antagonists reduce cardiovascular complications after cardiac surgery: a meta-analysis. J Am Coll Cardiol 2003;41:1496 –505. 10. Moher D, Cook DJ, Eastwood S, et al. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Lancet 1999;354:1896 –900. 11. Caramella JP, Goursot G, Carcone B, et al. Prevention of perand postoperative myocardial ischemia in non-cardiac surgery by intravenous diltiazem. Ann Fr Anesth Reanim 1988;7:245–50. 12. Hatziantoniou GP, Tassiopoulos PA, Fakiolas CN, et al. Perioperative risk of coronary patients undergoing major urological operations. Hellenic J Cardiol 1997;38:28 –34. 13. Garcı´a-Guasch R, Llubia C, Preciado MJ, et al. Diltiazem compared with placebo in the prevention of myocardial ischemia during non-cardiac surgery. Rev Esp Anestesiol Reanim 1998; 45:409 –15. 14. Godet G, Coriat P, Baron JF, et al. Prevention of intraoperative myocardial ischemia during noncardiac surgery with intravenous diltiazem: a randomized trial versus placebo. Anesthesiology 1987;66:241–5. 15. Amar D, Roistacher N, Burt ME, et al. Effects of diltiazem versus digoxin on dysrhythmias and cardiac function after pneumonectomy. Ann Thorac Surg 1997;63:1374 – 82.

ANESTH ANALG 2003;97:634 –41

16. Amar D, Roistacher N, Rusch VW, et al. Effects of diltiazem prophylaxis on the incidence and clinical outcome of atrial arrhythmias after thoracic surgery. J Thorac Cardiovasc Surg 2000;120:790 – 8. 17. Shimada M, Namai H, Morisaki H, et al. Preventive use of diltiazem to suppress supraventricular tachyarrhythmia in the patients after esophagectomy. Masui 1997;46:658 – 63. 18. Lindgren L, Lepantalo M, von Knorring J, et al. Effect of verapamil on right ventricular pressure and atrial tachyarrhythmia after thoracotomy. Br J Anaesth 1991;66:205–11. 19. Van Mieghem W, Tits G, Demuynck K, et al. Verapamil as prophylactic treatment for atrial fibrillation after lung operations. Ann Thorac Surg 1996;61:1083–5. 20. du Toit HJ, Weich HF, Weymar HW, Przybojewski JZ. Effects of nifedipine on the peri-operative ECG, as determined by continuous Holter monitoring. S Afr Med J 1986;69:427–31. 21. Retamal O, Coriat P, Pamela F, et al. Prevention of hypertensive attacks after carotid surgery: the value of nifedipine and diltiazem. Ann Fr Anesth Reanim 1986;5:278 – 86. 22. Beattie WS, Warriner CB, Etches R, et al. The addition of continuous intravenous infusion of ketorolac to a patient-controlled analgetic morphine regime reduced postoperative myocardial ischemia in patients undergoing elective total hip or knee arthroplasty. Anesth Analg 1997;84:715–22. 23. Badner NH, Beattie WS, Freeman D, Spence JD. Nitrous oxideinduced increased homocysteine concentrations are associated with increased postoperative myocardial ischemia in patients undergoing carotid endarterectomy. Anesth Analg 2000;91: 1073–9.

CARDIOVASCULAR ANESTHESIA WIJEYSUNDERA AND BEATTIE CALCIUM CHANNEL BLOCKERS IN NONCARDIAC SURGERY

641

24. Polanczyk CA, Goldman L, Marcantonio ER, et al. Supraventricular arrhythmia in patients having noncardiac surgery: clinical correlates and effect on length of stay. Ann Intern Med 1998;129:279 – 85. 25. Furberg CD, Psaty BM, Meyer JV. Nifedipine: dose-related increase in mortality in patients with coronary heart disease. Circulation 1995;92:1326 –31. 26. Moher D, Pham B, Jones A, et al. Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses? Lancet 1998;352:609 –13. 27. Eagle KA, Berger PB, Calkins H, et al. ACC/AHA guideline update on perioperative cardiovascular evaluation for noncardiac surgery. Circulation 2002;105:1257– 67. 28. Mangano DT, Layug EL, Wallace A, Tateo I. Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 1996;335:1713–20. 29. Poldermans D, Boersma E, Bax JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in highrisk patients undergoing vascular surgery. N Engl J Med 1999; 341:1789 –94. 30. Boersma E, Poldermans D, Bax JJ, et al. Predictors of cardiac events after major vascular surgery: role of clinical characteristics, dobutamine echocardiography, and beta-blocker therapy. JAMA 2001;285:1865–73. 31. Yusuf S, Collins R, Peto R. Why do we need some large, simple randomized trials? Stat Med 1984;3:409 –22.