Author + information
- Received July 23, 2002
- Revision received November 13, 2002
- Accepted November 27, 2002
- Published online May 7, 2003.
- Duminda N Wijeysundera, MD*,
- W.Scott Beattie, MD, PhD†,* (, )
- Vivek Rao, MD, PhD‡ and
- Jacek Karski, MD†
- ↵*Reprint requests and correspondence:
Dr. W. Scott Beattie, Associate Professor, Department of Anaesthesia, University of Toronto, EN 3-453, Toronto General Hospital, University Health Network, 200 Elizabeth Street, Toronto, Ontario, M5G 2C4, Canada.
These results were presented at the annual meeting of the Canadian Anesthesiologists’ Society, Victoria, British Columbia, Canada, June 24, 2002.
Objectives We sought to determine the efficacy of calcium antagonists (CAs) in reducing death, myocardial infarction (MI), ischemia, and supraventricular tachyarrhythmia (SVT) after cardiac surgery.
Background Calcium antagonists may reduce complications after cardiac surgery—namely, death, MI, and renal failure. However, they are underused, possibly due to the results from previous observational studies.
Methods Both MEDLINE (1966 to December 2001) and EMBASE (1980 to December 2001) were searched, with supplementation by reference list searches. No language restrictions were applied. Included studies were randomized, controlled trials (RCTs) evaluating preoperative, intraoperative, or postoperative (first 48 h) CA use (intravenous or oral) during aortocoronary bypass or valve surgery. Studies were excluded if they exclusively recruited transplant recipients, individuals <18 years old, or patients with pre-existing SVT. Two reviewers independently evaluated study quality by using the Jadad score; a minimal score of 1/5 was required. Forty-one studies, encompassing 3,327 patients, were included. No studies assessed treatment exclusively with short-acting oral nifedipine. Treatment effects were calculated using the random-effects model. Heterogeneity was assessed using the Q test.
Results Calcium antagonists significantly reduced MI (odds ratio [OR] 0.58, 95% confidence interval [CI] 0.37 to 0.91; p = 0.02) and ischemia (OR 0.53, 95% CI 0.39 to 0.72; p < 0.001). Non-dihydropyridines significantly reduced SVT (OR 0.62, 95% CI 0.41 to 0.93; p = 0.02). Calcium antagonists were associated with trends toward decreased mortality during aortocoronary bypass (OR 0.66, 95% CI 0.26 to 1.70, p = 0.4).
Conclusions Use of CAs during cardiac surgery significantly reduced rates of MI, ischemia, and SVT. Further study using large RCTs is justified.
Coronary artery bypass graft surgery (CABG), the most common cardiac surgical procedure (1), reduces morbidity (2)and mortality (3)from coronary artery disease. However, cardiac surgery remains associated with significant complications, including death, myocardial infarction (MI), neurocognitive dysfunction, and acute renal failure. Strategies that reduce these events should therefore improve overall outcomes.
Calcium antagonists (CAs) may reduce these complications. They improve balance between myocardial oxygen supply and demand through negative chronotropic, negative inotropic, afterload-reducing, and coronary vasodilatory properties. An imbalance between myocardial oxygen supply and demand causes ischemia, thereby potentially leading to MI (4). Calcium antagonist–mediated vasodilation may reduce post-CABG graft spasm, another cause of postoperative ischemia (5). Calcium antagonists prevent supraventricular tachyarrhythmias (SVTs) (6), therefore potentially reducing postoperative atrial fibrillation, which is associated with neurocognitive dysfunction (7)and prolonged hospitalization (8). Also, CAs may limit renal damage by decreasing renal vascular resistance and increasing the glomerular filtration rate (9). These benefits vary by CA class: benzothiazepines (e.g., diltiazem), phenylalkylamines (e.g., verapamil), and dihydropyridines (e.g., nifedipine, amlodipine).
Despite these benefits, CAs are underused (10). The reasons for this are unclear, although there are at least three possibilities: CAs have shown no perioperative benefit in previous observational studies (10–12). Second, CAs’ negative inotropic properties may have led to concerns regarding exacerbating left ventricular dysfunction. Third, short-acting oral nifedipine, a dihydropyridine, has been associated with increased mortality in non-surgical studies (13).
Given this discrepancy between theoretical benefits and clinical practice, a systematic review of randomized, controlled trials (RCTs) evaluating CAs during cardiac surgery is justified.
This review was conducted according to the Quality of Reports of Meta-Analyses of Randomized Controlled Trials (QUOROM) recommendations (14).
Inclusion and exclusion criteria
Eligible studies were published RCTs that evaluated CAs (administered immediately preoperatively, intraoperatively, or postoperatively within 48 h) during CABG or valve surgery and reported any of the following outcomes: death, MI, ischemia, SVT, or creatinine clearance. Perioperative outcomes were defined as occurring between the initiation of surgery and postoperative day 30. Ischemia was defined as ST-segment deviation on the electrocardiogram or new wall motion abnormalities on the transesophageal echocardiogram. The SVTs included atrial fibrillation, atrial flutter, and supraventricular tachycardia. We did not strictly define MI, given the lack of uniform criteria in the literature. Studies were ineligible if they exclusively recruited organ transplant recipients, individuals <18 years old, or individuals with pre-existing SVT. Studies were also ineligible if CAs were exclusively administered as a cardioplegic additive.
Search strategy, methodologic assessment, and data abstraction
We identified RCTs by searching MEDLINE (1966 to December 2001) [calcium channel blockers AND(postoperative complications ORperioperative care ORintraoperative complications)] and EMBASE (1980 to December 2001) [calcium channel blocking agent AND(postoperative complication ORpostoperative period ORperoperative period ORintraoperative period ORperoperative care ORperoperative complication)], without language restriction. Titles and abstracts were screened to exclude ineligible studies. Two reviewers (Drs. Wijeysundera and Beattie) independently read the remaining studies and rated their quality using the Jadad score (15). This validated five-point scale assesses RCT quality based on the adequacy of randomization, blinding, and follow-up. The minimal score required was 1. Bibliographies of included studies were surveyed.
The following were abstracted independently by two reviewers (Drs. Wijeysundera and Beattie) onto standardized forms: patients, surgery, treatments, death, MI, ischemia, SVT, creatinine clearance, previous medications, low cardiac output syndrome, inotropic support, pacing, and blood loss. We did not strictly define low cardiac output syndrome, given the heterogeneity of definitions in the literature. Inotropic support was defined as the need for inotropes or intra-aortic balloon pump support. Where possible, we abstracted data only for comparisons of CAs against placebo or nitroglycerin. All disagreements were resolved by consensus.
Treatment effects for dichotomous outcomes were expressed as odds ratios (ORs), with 95% confidence intervals (CIs). Treatment effects on continuous outcomes were expressed as weighted mean differences. We employed the random-effects model and Q test to calculate pooled treatment effects and heterogeneity, respectively. All calculations were performed using Review Manager version 4.1 (Cochrane Collaboration, Oxford, U.K.). Statistical significance for treatment effects and heterogeneity were defined by p values <0.05 and <0.1, respectively.
The effects of CAs on death, MI, ischemia, SVT, and postoperative creatinine clearance were calculated. We determined CAs’ effects on several adverse events: low cardiac output syndrome, inotropic support, pacing, and blood loss. We also compared previous beta-blocker and CA use between the CA and non-CA arms, given that differences in medication use may have affected estimates of treatment effects. We employed the fixed-effects model for comparisons of adverse events and previous medication use.
Subgroup analyses were performed for each CA class (diltiazem, verapamil, dihydropyridines). We also compared CAs specifically against nitroglycerin, which has been considered a superior prophylaxis against post-CABG ischemia (16). Calcium antagonists may differentially impact patients with coronary artery disease; therefore, we performed a subgroup analysis for patients who underwent CABG alone.
We performed several sensitivity analyses to determine the robustness of our findings. We repeated the meta-analyses after successively withdrawing trials with the most favorable CA treatment effects. To assess the effect of study quality on estimates of treatment effect, we repeated the meta-analyses in subgroups of trials with Jadad scores >0, >1, and >2. Funnel plots were performed to assess for a publication bias.
Forty-one studies, encompassing 3,327 patients, were included (17–57)(Table 1). The search results are presented in Figure 1. The median Jadad score was 1 (range 1 to 4). A single study assessed concurrent CA treatment given systemically and as a cardioplegic additive (44); it was the only study that assessed short-acting oral nifedipine. A list of excluded studies is available from the authors.
Mortality and cardiovascular outcomes
Eleven studies reported deaths, with an incidence among 1,001 patients of 3.1%. The CAs did not affect mortality (OR 1.01; p = 1) (Fig. 2), without significant heterogeneity (chi-squared statistic = 7.99; p = 0.63). A nimodipine study (47)found that it significantly increased mortality, largely due to excessive bleeding. When nimodipine studies (43,47)were excluded, the OR for mortality was 0.66 (95% CI 0.26 to 1.70; p = 0.4).
Twenty-two trials reported MIs, with an incidence among 1,853 patients of 4.7%. The CAs significantly reduced MI (OR 0.58, 95% CI 0.37 to 0.91; p = 0.02) (Fig. 3), without significant heterogeneity (chi-square = 7.56; p = 1). Subgroup analyses suggested that diltiazem (OR 0.55, 95% CI 0.26 to 1.20; p = 0.13) and dihydropyridines (OR 0.46, 95% CI 0.22 to 0.96; p = 0.04) reduced MI to a greater degree than did verapamil (OR 0.83, 95% CI 0.35 to 1.98; p = 0.7). The definition of perioperative MI varied between studies. A post hoc analysis of studies strictly defining MI, a priori, as significant new Q waves and/or elevated creatine kinase, MB fraction (>5% or 50 U/l) (18,21,23,25,28,31–33,41,46,52–54)showed an improvedreduction in perioperative MI (OR 0.49, 95% CI 0.28 to 0.86; p = 0.01).
Twenty studies reported ischemia, with an incidence among 1,522 patients of 17.1%. The CAs significantly reduced ischemia (OR 0.53, 95% CI 0.39 to 0.72; p < 0.001) (Fig. 4), without significant heterogeneity (chi-square = 16.62; p = 0.62). Subgroup analyses showed that diltiazem (OR 0.60, 95% CI 0.35 to 1.05; p = 0.07) and dihydropyridines (OR 0.47, 95% CI 0.32 to 0.71; p < 0.001) reduced perioperative ischemia.
Fifteen studies reported SVT, with an incidence of 17.3% among 1,756 patients. Overall, the CAs’ effect on perioperative SVT was non-significant (OR 0.73, 95% CI 0.48 to 1.12; p = 0.15) (Fig. 5). The Q test, however, indicated significant heterogeneity among these studies. Subgroup analyses suggested that dihydropyridines and non-dihydropyridines had opposite effects on SVT (Fig. 4). Non-dihydropyridines significantly reduced SVT (OR 0.62, 95% CI 0.41 to 0.93; p = 0.02), whereas dihydropyridines non-significantly increased SVT (OR 2.69, 95% CI 0.57 to 12.64). When considered individually, diltiazem (OR 0.54, 95% CI 0.28 to 1.04) and verapamil (OR 0.69, 95% CI 0.42 to 1.15) similarly reduced SVT.
The effects of CAs on mortality, MI, ischemia, or SVT were not qualitatively affected by temporal trends, indications for CA treatment (ischemia, hypertension, SVT), or exclusion of the study that concurrently administered CAs orally with a cardioplegic solution (44).
Five studies (17,36,38,50,56), encompassing 161 patients, reported creatinine clearance. The CAs non-significantly increased postoperative creatinine clearance (7.65 ml/min increase; 95% CI −4.21 to 19.51 ml/min; p = 0.2). However, there was significant heterogeneity (chi-square = 19.63; p = 0.0006).
We performed post hoc analyses to explain this heterogeneity. The CAs significantly improved postoperative creatinine clearance if preoperative creatinine clearance was <95 ml/min (13.12 ml/min increase; 95% CI 9.16 to 17.07; p < 0.001), without significant heterogeneity (chi-square = 1.30; p = 0.52). The CAs also non-significantly worsened postoperative renal function if preoperative creatinine clearance was >95 ml/min (5.03 ml/min decrease; 95% CI −12.38 to 2.33; p = 0.18), without significant heterogeneity (chi-square = 0.20; p = 0.66). Subgroup analyses based on CA class did not remove the statistical heterogeneity within each subgroup.
Compared with nitroglycerin, CAs reduced MI (OR 0.51, 95% CI 0.25 to 1.06; p = 0.07), ischemia (OR 0.65, 95% CI 0.39 to 1.09; p = 0.10), and SVT (OR 0.52, 95% CI 0.26 to 1.14; p = 0.10). There was no significant heterogeneity for the effects on MI (chi-square = 2.23; p = 0.95) and ischemia (chi-square = 10.48; p = 0.23); however, there was significant heterogeneity for effects on SVT (chi-square = 14.62; p = 0.012). Relative to nitroglycerin, CAs had no effect on mortality (OR 1.18, 95% CI 0.37 to 3.79; p = 0.8).
Among patients who underwent CABG, CAs reduced mortality (OR 0.66, 95% CI 0.26 to 1.70; p = 0.4), MI (OR 0.58, 95% CI 0.37 to 0.91; p = 0.02), ischemia (OR 0.53, 95% CI 0.39 to 0.73; p < 0.001), and SVT (OR 0.76, 95% CI 0.48 to 1.21; p = 0.3). There was no significant heterogeneity for effects on mortality (chi-square = 3.17; p = 0.92), MI (chi-square = 7.56; p = 1), and ischemia (chi-square = 16.17; p = 0.58); however, there was significant heterogeneity for effects on SVT (chi-square = 31.07; p = 0.0033).
Previous medication use
Nineteen trials reported previous beta-blocker use (18,19,23,24,26–28,31,33,36–38,41,46,49,51,54,55,57). Patients in the CA arm of these trials were significantly less likely to have been on beta-blockers preoperatively (OR 0.76, 95% CI 0.60 to 0.96; p = 0.02), without significant heterogeneity (chi-square = 8.36; p = 0.94). Eighteen trials reported previous CA use (18,19,23,24,26–28,31,33,36–38,41,46,49,54,55,57). Patients assigned to the CA arm were non-significantly less likely to have been on CAs preoperatively (OR 0.81, 95% CI 0.64 to 1.03; p = 0.08), without significant heterogeneity (chi-square = 9.41; p = 0.90).
Three studies reported low cardiac output syndrome (19,32,34). There was no difference between the CA and non-CA arms (OR 1.01, 95% CI 0.25 to 4.11; p = 1), without significant heterogeneity (chi-square = 2.28; p = 0.32). Eleven trials reported inotropic support (18,19,21,27,31,32,34,36,41,49,57). There was no difference between the CA and non-CA arms (OR 0.98, 95% CI 0.62 to 1.53; p = 0.9), without significant heterogeneity (chi-square = 13.73; p = 0.19). Three trials reported postoperative pacing (32,34,57). The CAs were associated with a significant increase in pacing (OR 6.57, 95% CI 3.54 to 12.18; p < 0.001), without significant heterogeneity (chi-square = 0.50; p = 0.78).
Three studies reported postoperative blood loss (29,39,40). The difference between the CA and non-CA arms was non-significant (difference −0.47 ml; p = 1), without significant heterogeneity (chi-square = 1.70; p = 0.43). A nimodipine trial (47)found that significant bleeding, as defined by requirement of >10 U blood during the operative period or chest drainage of >2,400 ml within 24 h, was increased in the CA arm (OR 3.64; p = 0.04).
The effect on MI became non-significant when the four trials with the smallest ORs (i.e., most favorable CA treatment effects) were excluded. Excluding the nine most favorable trials resulted in a non-significant effect on ischemia. Excluding the two most favorable trials removed non-dihydropyridines’ significant effect on SVT. The effects of CAs on MI, ischemia, and SVT were not affected by study quality (Table 2). Funnel plots for mortality and MI revealed no obvious publication bias.
To our knowledge, this is the first systematic review in the literature to examine CA use during cardiac surgery. Overall, CAs decreased perioperative MI and ischemia; in subgroup analyses, diltiazem and verapamil decreased perioperative SVT. This meta-analysis followed current guidelines (14). The literature search was extensive, encompassing all languages. We employed the conservative random-effects model to calculate treatment effects. These estimates remained significant even when 20% to 45% of the most favorable studies were removed.
The CAs significantly reduced MI, an important clinical benefit for cardiac surgical patients. Perioperative MI is associated with increased in-hospital mortality (58). Its effects on long-term survival are more controversial; however, it is associated with decreased three-year survival among individuals with previous MI or ventricular dysfunction (58). We did not apply a strict definition of MI, given the heterogeneity in the literature. This is unlikely to have affected our results. Patients were directly compared only within the same study. Furthermore, in a post hoc analysis, the treatment effect was improved when analyses were restricted to trials employing common definitions of MI.
Diltiazem and verapamil significantly reduced SVT. Overall, the CAs are unlikely to significantly affect SVT, given the varying chronotropic properties of CA classes. Perioperative SVT reduction is clinically important. Postoperative atrial fibrillation is associated with worsened postoperative outcomes (7,8).
The CAs significantly increased postoperative creatinine clearance among individuals with decreased preoperative renal function. The renal-protective properties of CAs are likely to vary with preoperative renal function. However, these findings reflect a post hoc analysis of <200 patients. Nonetheless, these results justify further study among patients with pre-existing renal insufficiency.
The CAs did not affect overall mortality. However, the analysis was greatly affected by the negative study of Legault et al. (47), who found that perioperative nimodipine significantly increased mortality, largely due to increased postoperative bleeding. These patients all underwent hypothermic cardiopulmonary bypass, however. Hypothermia impairs platelet activity (60), reduces coagulation factor function (61), and increases perioperative blood loss (62). Furthermore, only 46% received aminocaproic acid, an antifibrinolytic agent that reduces blood loss during cardiac surgery by 30% to 40% (63). The mortality benefit of CAs may apply specifically to patients undergoing CABG. In this subgroup, CAs non-significantly reduced mortality.
The effects of CAs on bleeding warrant further discussion. The present analysis lacks sufficient power to examine CAs’ overall and class-specific effects on perioperative bleeding, because only three included studies reported the outcome. In vitro, dihydropyridines (64), verapamil (65), and diltiazem (66)all reduce platelet aggregation. However, these in vitro effects are unlikely to have clinically significance. An abstract, referred to in a review of CAs’ adverse effects (67), found no increased risk of bleeding among 5,157 cardiac surgical patients. However, this 1996 abstract has not been subsequently published in full. The same review (67)concluded that most of the clinical data linking CAs and bleeding point against an increased risk.
The CAs did not increase the incidence of low cardiac output syndrome or inotropic support. Although postoperative pacing was increased, there were no associated adverse hemodynamic effects. These benign chronotropic and inotropic effects are in contrast to CA cardioplegic additives, which were associated with increased inotropic support (68)and prolonged electromechanical arrest (69).
Our results should be interpreted cautiously. The quality of included studies affects the magnitude of pooled treatment effects (70). The majority of included trials were unblinded. However, we did conduct sensitivity analyses to examine the effect of poorer study quality on our results. The treatment effects on MI and ischemia were essentially unchanged when lower quality studies were excluded.
Blinding was not employed during study evaluation and data abstraction. This did not significantly affect our results (71,72). As with all meta-analyses, our review may have been affected by a publication bias. Our analyses did include trials where CAs had neutral or negative effects, however. Language restrictions were not applied, therefore removing that component of publication bias. Unpublished data were excluded; however, the importance of this in meta-analyses is still debatable (73).
There certainly was clinical heterogeneity among the studies with regard to patient characteristics, drug dose, and duration of therapy. However, we employed statistical tests that indicated that most pooled treatment effects were unaffected by heterogeneity. Furthermore, p values for these tests were consistently >0.60 for overall analyses pertaining to mortality, MI, and ischemia.
Therefore, our results justify evaluating perioperative CAs in a large, simple, double-blinded RCT, the best method of estimating CAs’ true efficacy. We suggest that this trial should evaluate CAs among patients undergoing CABG, the subgroup most likely to benefit from CAs.
Our meta-analysis indicated that CA use during cardiac surgery significantly reduced perioperative MI and ischemia. Furthermore, non-dihydropyridines significantly reduced perioperative SVT. Further study is needed to determine the true effects of CAs on the aforementioned outcomes, as well as their effect on perioperative mortality.
We thank Dr. Terence M. Yao, Division of Cardiac Surgery, University of Toronto, Ontario, Canada, for his excellent review of this manuscript.
☆ Dr. Wijeysundera is the recipient of the Allan K. Laws Clinician Scientist Fellowship from the University of Toronto, Ontario, Canada.
- calcium antagonist
- coronary artery bypass graft surgery
- myocardial infarction
- randomized, controlled trial
- supraventricular tachyarrhythmia
- Received July 23, 2002.
- Revision received November 13, 2002.
- Accepted November 27, 2002.
- American College of Cardiology Foundation
- American Heart Association
- Talajic M.,
- Papadatos D.,
- Villemaire C.,
- Nayebpour M.,
- Nattel S.
- Furberg C.D.,
- Psaty B.M.,
- Meyer J.V.
- Babin-Ebell J.,
- Keith P.R.,
- Elert O.
- Hannes W.,
- Fasol R.,
- Zajonc H.,
- et al.
- Hannes W.,
- Seitelberger R.,
- Christoph M.,
- et al.
- Hirnle T.,
- Stachurski A.,
- Negrusz-Kawecka M.,
- Halawa B.,
- Bross T.
- Keilich M.,
- Kulinna C.,
- Seitelberger R.,
- Fasol R.
- Malhotra R.,
- Mishra M.,
- Kler T.S.,
- Kohli U.M.,
- Mehta Y.,
- Trehan N.
- Schoneberger A.,
- Raschka C.,
- Reifart N.,
- Hoffmann S.,
- Satter P.
- Tschirkov A.,
- Mishev B.,
- Natschev G.,
- Petkov R.,
- Alexandro V.,
- Jurukova Z.
- Fang W.,
- Mao L.
- Forsman M.,
- Olsnes B.T.,
- Semb G.,
- Steen P.A.
- Legault C.,
- Furberg C.D.,
- Wagenknecht L.E.,
- et al.
- Seitelberger R.,
- Zwolfer W.,
- Huber S.,
- et al.
- Donmez A.,
- Ergun F.,
- Kayhan Z.,
- Tasdelen A.,
- Dogan S.
- van Wezel H.B.,
- Bovill J.G.,
- Schuller J.,
- Gielen J.,
- Hoeneveld M.H.
- Force T.,
- Hibberd P.,
- Weeks G.,
- et al.
- Kiyomoto A.,
- Sasaki Y.,
- Odawara A.,
- Morita T.