Meta-analysis of antiarrhythmic therapy in the prevention of postoperative atrial fibrillation and the effect on hospital length of stay, costs, cerebrovascular accidents, and mortality in patients undergoing cardiac surgery

Autor: Wassim Choucair, Michael Greenberg, Pamela Karasik, Steven N. Singh, Jeffrey Southard, Peter Kokkinos, John C. Pezzullo, Jennifer Zimmer
Rok vydání: 2003
Předmět:
Zdroj: The American Journal of Cardiology. 91:1137-1140
ISSN: 0002-9149
DOI: 10.1016/s0002-9149(03)00168-1
Popis: T have been 13 randomized controlled trials of prophylactic antiarrhythmic therapy in patients undergoing cardiac surgery that have assessed its effects on hospital length of stay.1–13 These have shown consistent and marked decreases in the incidence of atrial fibrillation (AF), but the effects on hospital stay have been less concordant. Correlation of a decreased incidence of this arrhythmia with a reduction in hospital length of stay, costs, morbidity, or mortality would help determine how much continuing effort should be placed on its prevention. To determine whether this decreased incidence translates into clinically important outcomes, we conducted a meta-analysis of various antiarrhythmic therapies and their effects on the length of hospitalization, costs, stroke, and mortality. • • • We conducted a review of reports (in English) in the MEDLINE database, using the keywords “antiarrhythmics” and “postoperative atrial fibrillation” between January 1977 and March 2001. Published reviews, computerized literature search, and analysis of references identified potentially eligible studies. Only published data were included in this analysis. Unpublished studies and results reported in abstracts were excluded. Studies were included if they met the following criteria: (1) randomized comparison of an intervention to placebo; (2) evaluation of pacing modality or drug administration excluding calcium channel blockers, digoxin, and magnesium; (3) reported hospital length of stay; (4) AF identified as the arrhythmia. Thirteen randomized trials met the criteria and are included in this analysis. Forty trials were excluded because they did not include data on hospital length of stay, 14 because they did not include a placebo group, 7 because they evaluated atrial flutter or supraventricular arrhythmias as a group, and 4 because they were retrospective studies. Data were also extracted on 3 additional outcomes when available (costs, stroke, and death). Six studies had data on costs, 5 had data on the incidence of stroke, and 9 had data on mortality. All trial analyses were double-blinded and performed on an intention-to-treat basis. In the study by Dorge et al,8 the 2 different dosing regimens of amiodarone were considered as a single treatment group when evaluated for outcomes in the meta-analysis. In cases where different pacing modalities resulted in no significantly different effect on the incidence of atrial fibrillation, the results for outcomes were averaged before statistical analysis.3 In trials of different pacing modalities that showed superiority of 1 treatment group in decreasing the incidence of atrial fibrillation, only this group was included in the analysis of outcomes.10,11 The trials by Guanieri et al5 and Fan et al9 did not have SDs for the cost analysis; therefore, we approximated the SD to be 70% to 75% of the total cost, based on an evaluation of other studies in this meta-analysis that included data on SDs. Our primary outcomes for analysis were the effects on duration of hospitalization, defined as the number of days from surgery to hospital discharge, as well as hospital costs. The clinically important outcomes of the incidence of stroke and mortality were also assessed. For quantitative outcomes, the difference in the outcomes between the patient and the intervention groups was computed along with 95% confidence intervals. These were checked for consistency with an analysis of variance. By taking a weighted average of the differences, the results from the separate studies were combined, the weight being inversely proportional to the square of the width of the confidence interval. The values were combined using a method described by Fleiss.14 For dichotomous outcome variables, odds ratios and confidence intervals were computed, and these were combined on the basis of the weighted logarithms of the odds ratio, as described in the aforementioned source.14 In the 13 studies, a total of 1,783 patients were enrolled and included in the meta-analysis. Of these patients, 1,038 were assigned to the antiarrhythmic group and 745 were assigned to the control group. One thousand five hundred and sixty-nine patients underwent isolated coronary artery bypass grafting, 87 underwent only valvular surgery, 117 underwent both coronary artery bypass grafting and valvular surgery, and 10 underwent other types of cardiac surgery. The incidence of AF varied from 8% to 37% in the treatment groups and 29% to 53% in the control groups, From the Department of Cardiology, Veterans Affairs Medical Center; and Georgetown University Medical Center, Washington, DC. Dr. Singh’s address is: Department of Cardiology, Veterans Affairs Medical Center, 50 Irving Street, NW, Room 1E301, Washington, DC 20422. E-mail: singh@va.med.gov. Manuscript received September 24, 2002; revised manuscript received and accepted January 22, 2003.
Databáze: OpenAIRE
Externí odkaz: