Abstract
Background: Two recent meta-analyses demonstrated that angiotensin converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) reduce the risk of developing new-onset atrial fibrillation (AF) by nearly 50%. However, the ability of ACEIs or ARBs to prevent postoperative atrial fibrillation (POAF) after cardiac surgery has not been adequately evaluated. Objective: To evaluate the impact of preoperative ACEI or ARB use on the incidence of POAF after cardiac surgery. Methods: Patients undergoing coronary artery bypass grafting and/or valvular surgery from the (atrial fibrillation suppression trials II and III (AFIST II and III) randomized, controlled trials were evaluated in this cohort evaluation. Data in respect to patient demographics, surgical characteristics, medication utilization and the incidence of POAF (defined as AF lasting at least 5 min in duration documented by telemetry) were all uniformly and prospectively collected as part of AFIST II and III. Multivariate logistic regression was utilized to calculate adjusted odds ratios with 95% confidence intervals. Results: A total of 338 patients were evaluated of which 175 (51.8%) received an ACEI or ARB preoperatively and 163 (48.2%) did not. The study population was 65.7 ± 9.1 years of age, 77.8% were male, 11.2% underwent valve surgery, 3.6% had prior AF, 10.1% had heart failure and 84.0 and 37.9% received postoperative beta-blockade and prophylactic amiodarone, respectively. In total, 110 (32.5%) patients developed POAF. Upon multivariate logistic regression, the preoperative use of an ACEI or ARB was not found to be associated with a statistically significant reduction in POAF (adjusted odds ratio; 0.71, 95% CIs 0.42–1.20). Conclusions: Although preoperative ACEI or ARB use reduced the odds of developing POAF by 29%, this association with not found to be statistically significant. A study with approximately 600 subjects would be needed to discern if ACEIs or ARBs truly impact POAF.
1 Introduction
Approximately 800,000 patients will undergo cardiac surgery this year [1]. One of the most common complications following cardiac surgery is the development of postoperative atrial fibrillation (POAF) [2]. Unfortunately, the incidence of POAF remains high, occurring in up to 40% of patients, even when proven and recommended preventative drug therapy such as beta-blockers and amiodarone are employed [2]. Numerous studies have demonstrated that when POAF develops, patients are at increased risk of developing hemodynamic instability and cerebrovascular complications [2]. It is likely that these complications – at least partially – contribute to the increased hospital length of stay and treatment costs observed in patients with POAF [2–4].
In two recent meta-analyses not involving cardiac surgery patients, it was determined that chronic therapy with an angiotensin converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB) could reduce patients’ odds of developing new-onset atrial fibrillation (AF) by nearly one-half [5–6]. Various potential explanations exist for an ACEIs’ or ARBs’ ability to prevent AF, including their ability to decrease left atrial stretching (secondary to afterload reduction and stroke volume increases), preserve serum potassium concentrations or reduce pathogenic atrial remodeling [7–8].
Over the past several years, our research group has conducted two randomized and controlled clinical trials evaluating different modalities to prevent POAF [9–10]. Together, the most recent atrial fibrillation suppression trials II and III (AFIST II and III) have evaluated over 300 cardiac surgery patients in a prospective and scientifically rigorous manner and consequently could serve as an excellent source of data to conduct a cohort study. Since there is a paucity of data evaluating an ACEI or ARB to prevent POAF, we sought to determine whether the preoperative utilization of these drugs could prevent the development of atrial fibrillation following cardiac surgery.
2 Methods
2.1 Design and population
This was a prospective cohort evaluation of the 338 patients undergoing coronary artery bypass grafting (CABG) and/or valvular surgery from the AFIST II and III randomized, controlled trials [9–10]. AFIST II [9] was a 2 × 2 factorial designed study evaluating the effect of two active treatments, amiodarone and atrial septal pacing, on the incidence of POAF in CABG and/or valvular surgery patients (n = 160). Patients were randomized to amiodarone or placebo and were then randomized to active atrial septal pacing or no active atrial septal pacing. AFIST III [10] was a randomized controlled trial evaluating the impact of aortic fat pad (AFP) maintenance on the incidence of POAF in patients (n = 178) undergoing first-time CABG surgery. The use of beta-blockers in both AFST II and III was directed by the hospital critical pathway. The treatment of POAF once it occurred was up to the patient’s physician as was the decision to use amiodarone in AFIST III.
Data in respect to patient demographics, surgical characteristics, medication utilization and the incidence of POAF were all prospectively and uniformly collected as part of the AFIST II and III trials. In both trials, the primary endpoint for all patients was the development of POAF within 30 days of cardiac surgery and was defined as any AF lasting more than 5 min in duration and documented by telemetry either during the index hospital admission or at any follow-up visit [9–10].
2.2 Statistical analysis
Continuous variables are presented as means with standard deviations and were compared between groups using a Student’s t-test. Dichotomous variables are presented as percentages and were compared between groups using a Chi-square test or Fisher’s exact test, when appropriate.
As this was an observation study, the investigators had no control over which patients received an ACEI or ARB and which did not. As a result, significant differences in important observed demographic, surgical and medication utilization characteristics were likely to occur which could lead to a biased estimate of treatment effect. Therefore, we conducted multivariate logistic regression to control for potential confounders in our evaluation. We first conducted univariate analysis to examine the association between the occurrence of the endpoint of interest (dependant variable; development of POAF) and demographic, surgical and medication utilization characteristics (independent variables). All variables with a p-value of ≤0.2 in the univariate analysis were entered into a multivariate logistic regression model. In the multivariate model, variables were selected by backward elimination, and a p-value ≪0.05 was considered significant. Adjusted odds ratios and 95% confidence intervals were calculated for all independent predictors. Statistical analysis was performed with SPSS version 11.0 (SPSS Inc., Chicago, IL).
3 Results
A total of 338 patients were evaluated in this cohort evaluation, of which 175 (51.8%) received an ACEI or ARB preoperatively and 163 (48.2%) did not. The study population averaged 65.7 ± 9.1 years of age, was 77.8% male, 11.2% underwent valve surgery, 3.6% had prior AF, 10.1% had heart failure and 84.0 and 37.9% received postoperative beta-blockade and prophylactic amiodarone, respectively. Comparisons of demographic, surgical and medication utilization characteristics between the preoperative ACEI or ARB cohort and the no preoperative ACEI or ARB cohort as well as in those developing and not developing POAF are provided in Tables 1 and 2 .
Patient demographic, surgical and medication utilization characteristics within study cohorts
Patient demographic, surgical and medication utilization characteristics within study cohorts
Patient demographic, surgical and medication utilization characteristics by POAF status
Patient demographic, surgical and medication utilization characteristics by POAF status
In total, 110 (32.5%) patients developed POAF. POAF occurred in 51 of 175 patients (29.1%) receiving a preoperative ACEI or ARB and in 59 of 163 patients (36.2%) whom did not (crude odds ratio; 0.73, 95% CIs 0.46–1.14). Upon multivariate logistic regression, the preoperative use of an ACEI or ARB was not found to be associated with a statistically significant reduction in POAF (adjusted odds ratio; 0.71, 95% CIs 0.42–1.20) (Table 3 ).
Identified independent predictors of postoperative atrial fibrillation
Identified independent predictors of postoperative atrial fibrillation
4 Discussion
Although preoperative ACEI or ARB use was found to reduce the odds of developing POAF by 29% in our evaluation, this association with not found to be statistically significant (p = 0.20). Upon conducting a post-hoc sample size calculation (α = 0.05; β = 0.20), it was determined that a study enrolling approximately 600 subjects would be needed to discern if ACEIs or ARBs truly impact POAF.
It is important to note that this 29% reduction in POAF with ACEIs and ARBs was seen in a population that already had a high background utilization of both beta-blockers (84%) and prophylactic amiodarone (38%), both of which are highly efficacious drugs in preventing POAF following cardiac surgery. Whether, an ACEI or ARB when administered alone would exert a larger or statistically significant effect is not known.
The safety of ACEI or ARB utilization in the setting of cardiac surgery has also not been adequately evaluated. Both classes of drugs may have the potential to increase patients’ risk of developing a pre-renal azotemia via vasodilatation of the efferent renal arterioles and resultant reductions in glomerular filtration. Renal endpoint data was not collected as part of the AFIST II or III trials and therefore we were unable to assess this important safety endpoint. We recommend any future trials evaluating ACEIs or ARBs to prevent POAF also evaluate the incidence of renal insufficiency, acute renal failure and the need for hemodialysis.
This study does have some limitations. First, while we were able to assign patients to either the ACEI or ARB cohort or the no ACEI or ARB cohort, sufficient data was not available to assess the effect of duration of therapy on POAF. Secondly, as this was a cohort study, investigators did not dictate the use of ACEIs or ARBs. This lack of randomization in observational studies could introduce both confounding and biases that can reduce a study’s internal validity. A number of characteristics of our study may have helped to limit the effects of confounding and biases on our results. Our use of multivariate logistic regression to calculate adjusted odds ratios likely minimized much of the effect of confounding during statistical analysis. Data on most of the commonly accepted independent predictors of POAF identified in the American Heart Association’s and American College of Cardiology’s cardiac AF guidelines [11] were available for use in multivariate adjustment. Additionally, the use of data from the previously conducted, prospective, randomized and controlled AFIST II and III trials likely prevented or reduced the unwanted effects of many common types of biases including selection, misclassification, surveillance and researcher bias [9–10].
5 Conclusion
Although preoperative ACEI or ARB use reduced the odds of developing POAF by 29%, this association with not found to be statistically significant. We approximate that a randomized and controlled trial with approximately 600 subjects would be needed to discern if ACEIs or ARBs truly impact POAF.
