Methods

We conducted an extensive search of published studies using multiple online databases (PubMed, Scopus, and Ovid) for studies evaluating the association between female gender and short-term, midterm, and long-term outcomes after isolated CABG. The search was limited through May 31, 2012 and was performed using the Medical Subject Headings search string “gender AND female AND women AND coronary artery bypass grafting AND outcomes.” We included peer-reviewed and published studies for our analysis. We excluded meeting abstracts, studies in languages other than English when translations were not available, studies that did not provide comparative outcomes, and studies reporting outcomes of CABG with any other concomitant surgery (e.g., valve repair or replacement, aortic root repair or replacement). Figure 1 shows the flowchart of the search method of studies included in the meta-analysis. Once inclusion criteria were met, each of the studies was thoroughly reviewed, and baseline and outcomes data were extracted independently by 2 reviewers (MA and SJB). The quality of included studies was assessed independently by 2 reviewers (MA and SJB) using the Newcastle-Ottawa Quality Assessment Scale (NOQAS), and disparities were resolved by concensus. Twenty studies that included ≥1 of the following outcomes were included in the meta-analysis: postoperative mortality, 30-day mortality, and mortality at 1, 5, or 10 years. Short-term mortality was defined as a composite of postoperative and 30-day mortality. Outcome definitions among the individual studies were found to be comparable.

Flowchart describing the details of the search method.

Percentages, means, and SDs were calculated to describe the distributions of categorical and continuous variables, respectively. Continuous variables were compared using 2-tailed independent-samples Student’s t tests. Categorical variables were compared using chi-square tests with Yates’ correction. Measures of heterogeneity, including Cochran’s Q statistic the and I ² index, were computed. An I ² value ≥25% was considered to indicate significant heterogeneity. Publication bias was assessed using funnel-plot analysis for short-term mortality. We used a random-effects model to summarize the effect size for clinical outcomes using odds ratios (ORs) and their 95% confidence intervals (CIs). The baseline data were analyzed using SPSS version 19.0 (SPSS, Inc., Chicago, Illinois). A p value ≤0.05 was considered significant. The meta-analysis was performed using Review Manager version 5.1 (Cochrane Collaboration, Copenhagen, Denmark). Various subgroup analyses were performed. The first subgroup analysis included 7 prospective studies only. The second subgroup analysis included 4 propensity score–matched studies only. We also performed additional analyses on the basis of study quality assessment. All included studies were stratified on the basis of NOQAS score (≥7 = high quality, 6 = intermediate quality, ≤5 = lower quality). These stratified analyses are provided in the online supplement. Supplemental Table 1 lists the details of NOQAS scores for included studies. A time-sensitive analysis to ascertain the impact of contemporary practice on clinical outcomes was performed. For this analysis, all studies were divided into 2 groups: studies published before 2001 and contemporary studies (published in 2001 and later).

Results

A cumulative 966,492 patients were included in the meta-analysis (688,709 men [71%] and 277,783 women [29%]). A list of the included studies is provided in Table 1 . As listed in Table 2 , women were more likely to be older at presentation and to have multiple co-morbidities, including diabetes mellitus, hypertension, hyperlipidemia, peripheral vascular disease, unstable angina, congestive heart failure, and New York Heart Association class III or IV symptoms. In contrast, men had a higher prevalence of smoking and history of cerebrovascular accident and myocardial infarction. Angiographic data showed that men had a higher burden of coronary artery disease than women, with triple-vessel disease being more prevalent in men (68% vs 58%, p <0.01). However, women were more likely to undergo urgent surgery (51% vs 44%, p <0.01) ( Table 2 ).

Figure 2 shows a funnel plot of short-term all-cause mortality. This funnel plot shows the standard error of log(OR) on the y axis and the OR for short-term mortality on the x axis. A review of this plot shows a strong treatment effect, as most of the studies with narrow standard errors for log(OR) lie around the mean treatment effect, which itself lies to the right of the no-effect line (i.e., 1.0). However, as all except 1 study lie to the right of the no-effect line, it is possible that negative studies may not have been published, and therefore publication bias cannot be excluded.

Funnel-plot analysis for outcome of short-term mortality (including postoperative and 30-day all-cause mortality). This funnel plot shows the standard error (SE) of log(OR) on the y axis and the OR for short-term mortality on the x axis. A review of this plot shows a strong treatment effect, as most studies with narrow SEs for log(OR) lie around the mean treatment effect, which itself lies to the right of the no-effect line (i.e., 1.0). However, as all except 1 study lie to the right of the no-effect line, it is possible that negative studies may not have been published, and therefore publication bias cannot be excluded.

The impact of female gender on all-cause mortality after isolated CABG is listed in Table 3 . Female gender was associated with a 77% increase in the relative risk for short-term mortality (4.28% vs 2.53%, OR 1.77, 95% CI 1.67 to 1.88) after CABG. This was reflected in the postoperative period (4.36% vs 2.47%, OR 1.85, 95% CI 1.66 to 2.05) and at 30 days (4.17% vs 2.53%, OR 1.66, 95% CI 1.59 to 1.74). Mortality rates in women remained higher compared with men at 1-year and 5-year follow-up (2.55% vs 1.95%, OR 1.31, 95% CI 1.18 to 1.45, and 9.26% vs 8.16%, OR 1.14, 95% CI 1.08 to 1.20, respectively). Figure 3 displays forest plots showing higher mortality in women after isolated CABG at short-term (postoperative or 30-day), midterm (1-year), and long-term (5-year) follow-up. Long-term follow-up at 10 years was reported in only 1 study; therefore, meta-analysis could not be performed. However, in this study, women had higher mortality compared with men as well (14.78% vs 12.76%, OR 1.19, 95% CI 1.13 to 1.25). Figure 4 shows ORs for all-cause mortality at various follow-up intervals from the immediate postoperative period up to 10-year follow-up. At longer term follow-up, the OR for all-cause mortality in women decreased but remained significantly higher than in men.

Association of female gender with (A) short-term mortality (composite of postoperative or 30-day mortality), (B) 1-year mortality, and (C) 5-year mortality after isolated CABG. (D) Association of female gender with postoperative or 30-day mortality after isolated CABG in patients with propensity score matching. M-H = Mantel-Haenszel.

Line graph with ORs and 95% CIs showing the association between female gender and all-cause mortality after isolated CABG up to 10 years of follow-up.

In the first subgroup analysis of all prospective data (7 studies; Table 4 ) reported for 41,500 patients, women exhibited high short-term mortality (postoperative and/or 30-day) compared with men (3.12% vs 1.71%, OR 1.83, 95% CI 1.59 to 2.12). A second subgroup analysis including 4 studies ( Table 5 ) that performed propensity score matching (3 studies reporting postoperative mortality and 1 reporting 30-day mortality) was performed. This included 11,522 patients well matched for baseline demographics and equally divided between men and women ( Table 6 ). The analysis revealed an increased risk for short-term mortality after CABG in women compared with men (6.14% vs 4.48%, OR 1.36, 95% CI 1.04 to 1.78). Details of this propensity score–matched analysis are listed in Table 7 .

Stay updated, free articles. Join our Telegram channel

Join