Post-cardiac surgery atrial fibrillation and sex differences in clinical outcomes: a systematic review and meta-analysis

Abstract

There is conflicting literature on sex differences and clinical outcomes in patients who develop atrial fibrillation (AF) post-cardiac surgery. Our aim was to compare clinical outcomes between females and males with post-cardiac surgery AF. A systematic search was conducted for studies published until 27 September 2024 in MEDLINE, Embase, and CENTRAL. Included studies compared mortality and stroke in females vs. males who developed AF after cardiac surgery. Outcomes of interest were mortality and stroke. Pooled prevalence was used to compare comorbidities. Raw event rates were used to calculate odds ratios (ORs), which were pooled with a fixed-effects model. 10 422 studies were identified and 5 studies met inclusion criteria. 14 970 patients who developed AF after cardiac surgery were included, of whom 3748 were females. The length of follow-up was up to 10 years. The weighted average age was 70.4 ± 2.9 years in females and 66.7 ± 2.4 years in males (P = 0.32). At baseline, females had a higher prevalence of hypertension, diabetes, dyslipidemia, stroke, and peripheral vascular disease. The odds of in-hospital mortality among females were higher compared to males (5.5 vs. 3.0%; OR 2.04, 95% CI 1.42–2.91, P < 0.001, I² = 57%). There were no significant differences in post-discharge mortality (45.6 vs. 42.9%; OR 1.05, 95% CI 0.97–1.15, P = 0.23, I² = 0%) or in-hospital stroke (2.5 vs. 1.9%; OR 1.30, 95% CI 0.79–2.13, P = 0.30, I² = 57%) in females vs. males. In conclusion, females with post-cardiac surgery AF had a higher prevalence of comorbidities at baseline. The odds of in-hospital mortality were twice as high among females. There were no significant differences in post-discharge mortality or in-hospital stroke. Future studies are warranted to understand the mechanisms of increased in-hospital mortality in females and to develop effective monitoring strategies and interventions.

Introduction

Cardiovascular disease is a leading cause of death among females, but remains suboptimally investigated, diagnosed, and treated.^1,2 Atrial fibrillation (AF) is more common in males until the age of 75, at which point females are disproportionately affected.³ AF in females is associated with a higher risk of mortality, stroke, symptom burden, and AF recurrence.^1,4 Post-cardiac surgery AF is common, where the highest risk procedure is combined coronary artery bypass (CABG) and valvular surgery, with the risk of AF at 60–80%.⁵ Post-cardiac surgery AF is associated with increased mortality, stroke, length of stay, and readmission.^5,6 Literature on sex-related differences in post-cardiac surgery AF is inconsistent. In a prospective cohort study that followed 181 patients with AF after CABG for up to 6 years,⁷ there were no significant differences in post-discharge mortality between females and males (P = 0.92).⁷ In contrast, a more recent retrospective study of 8499 patients with AF after CABG and/or valvular surgery found a higher rate of mortality among females who developed post-operative AF (50.4 vs. 48.9%).⁸ Therefore, it remains unclear if females who develop post-cardiac surgery AF have worse outcomes than males.

Accordingly, we conducted a systematic review and meta-analysis to compare clinical outcomes between females and males who developed AF following cardiac surgery. Our outcomes of interest were mortality and stroke.

Methods

Registration

This study was prospectively registered on PROSPERO (CRD42024609115).

Literature search and data sources

A systematic search was conducted in MEDLINE, Embase, and Cochrane Central Register of Controlled Trials (CENTRAL) for comparative studies published until 27 September 2024. The MEDLINE search strategy is available in Supplementary material online, Appendix S1. The search strategy was adapted for other databases. A grey literature search was conducted, including conference proceedings, relevant databases, and manual web searches. The reference lists of pertinent articles were hand-searched.

Study selection and quality assessment

Eligible studies must (1) include adults who developed AF after cardiac surgery and (2) compare mortality (in-hospital or post-discharge) or stroke (in-hospital or post-discharge) in females vs. males. Cardiac surgery was defined as CABG and/or surgical valve replacement/repair. Non-comparative studies were excluded, such as case reports and case series. Abstracts, systematic reviews, meta-analyses, and clinical guidelines were also excluded. Articles were independently screened in duplicate by five investigators (F.M., T.Q., M.A., S.B.L., and R.S.). Disagreements were resolved through discussion, and when necessary, in consultation with a sixth author (M.S.).

The quality of primary studies was assessed in duplicate (F.M. and T.Q.). The Newcastle-Ottawa Scale was used to evaluate cohort studies, where a study may receive up to nine points.⁹

Data extraction

Two investigators (F.M. and T.Q.) independently extracted the following data in duplicate from included studies: study design, sample size, demographics (age, sex, comorbidities, surgery type, and procedural details), outcome data (length of follow-up, rate of in-hospital and post-discharge all-cause mortality, and rate of in-hospital stroke), and definitions of outcomes. Disagreements were resolved through discussion, and when necessary, in consultation with a third author (M.S.).

Statistical analysis

The RevMan (version 5) software package by the Cochrane Collaboration was used for statistical analysis. The pooled prevalence and 95% confidence interval (CI) of each comorbidity was estimated separately for females and males using a random-effects model based on inverse variance. A random-effects model was performed to estimate the differences in the pooled prevalence of each comorbidity between females and males. The outcome was expressed as the risk difference with 95% CI. Outcomes were pooled with a fixed-effects model as described by DerSimonian and Laird.¹⁰ Summary estimates and 95% CI were reported for dichotomous variables as odds ratios (ORs). Heterogeneity was assessed with Cochran’s Q X² and I². An I² above 50% represented substantial heterogeneity.¹¹ A P-value below 0.05 indicated statistical significance. Results of the meta-analysis and summary estimates were presented in forest plots.

Results

Literature search and characteristics of included studies

The literature search yielded 10 422 studies (4064 from MEDLINE, 5352 from Embase, and 1006 from CENTRAL). Duplicates were removed (n = 2008), and 8397 studies were excluded during title and abstract review. The remaining 17 studies were assessed in full text. Five studies met the inclusion criteria and were included in this meta-analysis. Figure 1 summarizes the literature search and study selection according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards.

We included two prospective cohort studies and three retrospective cohort studies. Length of follow-up was up to 10 years (Table 1). Three studies included patients post-CABG, one study followed patients with surgical aortic valve replacement, and the remaining study was on CABG and/or surgical mitral/aortic valve replacement/repair. Based on the Newcastle-Ottawa scale, all included studies scored a minimum of six points, out of a maximum of nine points (Table 2).⁹

Baseline characteristics of included patients

We included 14 970 patients, of whom 3748 were females. The weighted average age was 70.4 ± 2.9 years among females and 66.7 ± 2.4 years among males (P = 0.32). At baseline, females had a higher prevalence of hypertension, diabetes, dyslipidemia, stroke, and peripheral vascular disease. The prevalences of chronic kidney disease and pre-operative beta-blockade were similar between groups. Average aortic cross-clamp time and mean bypass time were similar between females and males. The pooled prevalence of each comorbidity for females and males and the risk difference between females and males are summarized in Table 3. Additional information is available in Supplementary material online, Appendix S2.

Outcomes

In-hospital mortality occurred in 5.5% of females with post-cardiac surgery AF, compared with 3.0% of males. The odds of in-hospital mortality were twice as high in females vs. males (OR 2.04, 95% CI 1.42–2.91, P < 0.001, I² = 57%; Figure 2). To explore possible sources of heterogeneity, influencer analysis was performed and suggested heterogeneity was driven by Akça et al.¹²

Post-discharge mortality occurred in 45.6% of females and 42.9% of males. There were no statistically significant differences in post-discharge mortality between females and males (OR 1.05, 95% CI 0.97–1.15, P = 0.23, I² = 0%; Figure 3).

In-hospital stroke was not found to be different between females (2.5%) and males (1.9%) (OR 1.30, 95% CI 0.79–2.13, P = 0.30, I² = 57%; Figure 4).

Random-effects model

Analyses were repeated using a random-effects model. The odds of in-hospital mortality were increased in females compared with males but did not reach statistical significance (OR 1.73, 95% CI 0.79–3.80, P = 0.17, I² = 57%; Supplementary material online, Appendix S3). Post-discharge mortality (OR 1.05, 95% CI 0.97–1.15, P = 0.23, I² = 0%; Supplementary material online, Appendix S4) and in-hospital stroke (OR 0.81, 95% CI 0.23–2.80, P = 0.74, I² = 57%; Supplementary material online, Appendix S5) remain similar between groups.

Discussion

Our meta-analysis demonstrated higher odds of in-hospital mortality among females who developed AF post-cardiac surgery, compared with males. This finding may be related to females having a higher comorbidity burden at baseline. There were no significant differences in post-discharge mortality or in-hospital stroke.

In our study, females with post-cardiac surgery AF had twice the odds of in-hospital mortality compared to males (OR 2.04, 95% CI 1.42–2.91, P < 0.001). In contrast, the included retrospective cohort study by Akça et al. found higher rates of in-hospital mortality in males, rather than females (8.0 vs. 6.6%).¹² However, this study is limited by imprecision with a small sample size of only 368 patients. Additionally, there were four males who required re-exploration after the initial CABG for bleeding/tamponade, compared with no females, which may have influenced the higher risk of in-hospital mortality among males.¹² The higher odds of in-hospital mortality among females in our meta-analysis may be due to increased comorbidity burden, influencing perioperative risk. At baseline, females had a higher prevalence of hypertension, diabetes, dyslipidemia, stroke, and peripheral vascular disease. A previous prospective study of 27 239 patients demonstrated the association between mortality after CABG and certain comorbidities, including diabetes, vascular disease, and end-stage renal disease.¹⁵ It has been previously demonstrated that females are older and more comorbid at the time of CABG, relative to male counterparts.¹⁶ Optimization of cardiovascular risk factors prior to cardiac surgery is imperative and may help modify mortality risk. Additionally, timely management of post-cardiac surgery AF is warranted. Future studies may explore the risk factors for in-hospital mortality, allowing for individualized management and closer inpatient monitoring of female patients.

There were no differences in post-discharge mortality between females and males (45.6 vs. 42.9%; OR 1.05, 95% CI 0.97–1.15, P = 0.23). Length of follow-up ranged from 4 to 10 years.^7,8,13 A large prospective cohort study of 89 213 patients with non-valvular AF, outside of the context of cardiac surgery, found higher mortality rates among females compared with males (33.5 vs. 32.0%, P < 0.001).¹⁷ The adjusted risk of mortality was reported to vary with age.¹⁷ It remains unclear whether mortality associated with post-cardiac surgery AF differs with age due to a lack of reporting among primary studies.

The odds of in-hospital stroke were not different between females and males (OR 1.30, 95% CI 0.79–2.13, P = 0.30). Sex differences in the risk of in-hospital stroke of AF are not well-studied. Historically, females with AF have been found to be at a higher risk of long-term stroke compared with males.¹⁸ However, this was confounded by lower rates of anticoagulation among females.¹⁹ More contemporary data have shown that the excess stroke risk for female vs. male patients has been declining, with limited sex differences in recent years.^20,21 Accordingly, the CHA₂DS₂-VA score, which excludes sex, is the recommended risk stratification tool in the European Society of Cardiology 2024 guidelines, where anticoagulation is recommended for a score of ≥2.²² Similarly, the Canadian Cardiovascular Society has recommended the CHADS-65 algorithm since 2014, which also excludes a sex variable.¹⁴ Our study found similar odds of in-hospital stroke between males and females. This is in line with the contemporary data that showed less apparent sex differences in stroke risk and informed the exclusion of sex from stroke risk prediction tools. However, the lower event rates in our analysis support the need for additional studies to further investigate sex-related differences in stroke risk post-cardiac surgery.

Our meta-analysis has a number of strengths. This is the first meta-analysis to compare clinical outcomes among females and males who develop AF following cardiac surgery. We identified 5 studies with a total of 14 970 patients, making this the most comprehensive review of the topic to date. The included studies were of good quality, with all scoring a minimum of six out of nine possible points on the Newcastle-Ottawa scale.⁹ Finally, our meta-analysis is clinically relevant and may trigger further studies to better understand sex differences in post-cardiac surgery AF.

Limitations

The limitations of our meta-analysis are related to the primary studies. All included studies are observational, three of which are retrospective in nature. Such studies are inherently prone to residual confounding and selection bias, especially when examining in-hospital mortality. Low comparability between studies regarding study design, AF detection methods, and follow-up duration affects the validity of our meta-analysis. Our finding of increased in-hospital mortality among females may be confounded by the increased prevalence of comorbidities at baseline. Publication bias is a potential limitation of our study. One study did not report baseline characteristics.⁸ There was limited reporting of perioperative risk scores and characteristics that can affect AF risk and outcomes, such as proportion done on-pump, inotrope use, and intra-aortic balloon pump use.²³ Studies did not report the specific cause of death for patients with in-hospital mortality.

Conclusions

Our meta-analysis found statistically significant higher odds of in-hospital mortality in females who develop AF post-cardiac surgery. At baseline, females had higher rates of comorbidities, compared with males. Post-discharge mortality and in-hospital stroke were similar between groups.

Lead author biography

Faith Michael received her MD from McMaster University and subsequently completed an Internal Medicine residency with NOSM University. She is currently at Queen’s University for a General Internal Medicine fellowship. During her training, she developed a research interest in atrial fibrillation and greatly benefited from the mentorship of Dr Mohammed Shurrab, who is an electrophysiologist, the Cardiovascular Research Chair at Health Sciences North Research Institute, and an adjunct scientist at ICES. Her research interests include women’s cardiovascular health, with the goal of improving health outcomes among women.

Supplementary material

Supplementary material is available at European Heart Journal Open online.

Funding

None declared.

References

Author notes