Does sex impact outcomes after mitral valve surgery? A systematic review and meta-analysis

Background and objective: The published literature investigating the impact of sex on outcomes after mitral valve (MV) surgery has demonstrated inferior outcomes for females over males. However, the true relationship between sex and outcomes after MV surgery continues to be poorly understood. Materials: PubMed, Medline, and Embase were systematically searched for articles published from 1 January 2005 to 1 August 2021. This systematic review included retrospective and prospective studies investigating the relationship between sex and outcomes after MV surgery. In all, 2068 articles were initially screened and 12 studies were included in this review. Results: Few studies were adequately powered or structured to investigate this topic. Few studies propensity matched patients or isolated for surgical approach. In individual studies, females experienced increased rates of short-term and long-term mortality and increased 1-year mortality in the pooled data. Males experienced increased rates of required pacemaker insertion. The remaining rates of morbidity and mortality did not differ significantly between males and females. Conclusions: This review identified increased rates of 1-year mortality in the pooled data for females, while males had increased rates of pacemaker insertion. Despite this, the absence of propensity matching and isolating for surgical approach has introduced confounding variables that impair the ability of the included studies to interpret the results found in the current literature. Studies isolating for surgical approach, propensity matching patients, and examining outcomes with long-term follow-up are required to elucidate the true nature of this relationship.


Introduction
Discrepancies in outcomes between males and females following various medical and surgical interventions are prevalent in the literature. Across the majority of cardiovascular disease states and interventions, males tend to experience superior outcomes. Despite copious theories pertaining to the root of this inequity, the casual factors have yet to be identified.
One particular cardiovascular disease state in which this discrepancy in outcomes has been well established is valvular heart disease (VHD). VHD is a highly prevalent condition affecting 2.5% of the world population and with over 250,000 patients requiring and receiving heart Does sex impact outcomes after mitral valve surgery? A systematic review and meta-analysis valve surgery each year. 1,2 Elevated postoperative morbidity and mortality rates have consistently been reported in females undergoing surgical valve interventions as compared to males. [3][4][5][6][7] The discrepancy between males and females in VHD is not limited to postoperative outcomes. Sex-specific differences in the presentation and etiology of VHD have also been noted. [8][9][10] Despite females having a higher likelihood than males of developing severe mitral regurgitation (MR) and subsequently experiencing higher mortality rates with MR, they are less likely to receive surgery. [10][11][12][13] Various other explanations/factors have been postulated to explain the inequity, including females being underrepresented in various cardiovascular clinical trials; presenting with more advanced disease, with a higher preoperative risk profile; and often receiving more invasive surgical interventions than their male equivalents. 9 -14 It is imperative that all patients be provided with the best possible care. Consequently, there is an increasing acknowledgment of the need to investigate the inequity in outcomes between males and females. However, this discrepancy still requires elucidation.

Objectives
The objective of this systematic review and meta-analysis is to review previous studies comparing males and females undergoing surgical mitral valve (MV) interventions to provide a contemporary perspective on the state of MV surgery and the postoperative differences experienced between the sexes.

Data sources
PubMed, Medline, and Embase were systematically searched using the terms "sex, gender, heart valve, heart valve repair, heart valve replacement, mitral valve, mitral valve repair, mitral valve replacement, surgical mitral valve repair, and surgical mitral valve replacement" individually or in combination for articles from 1 January 2005 to 1 August 2021. This study period was chosen for data collection in an attempt to create a contemporary review mitigating the influence of older valve technologies and surgical techniques where possible. Data were extracted from included articles by two authors based on prespecified outcomes. This review was not registered prior to the commencement of data collection. A search of the PROSPERO database did not return results for ongoing reviews investigating sex differences in MV surgery.

Study selection
The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines and a published explanation of the PRISMA guidelines were used to perform this systematic review and meta-analysis. 15,16 In all, 2068 English manuscripts were initially screened, including the reference lists of reviewed manuscripts. All retrospective and prospective studies that compared adult males and females undergoing MV surgery with or without concomitant procedures were included in this review. Articles were excluded if data were not provided comparing the outcomes of males and females, if MV surgery was not investigated, if the article was a published abstract without associated full-length text accessible from any source, if insufficient preoperative or postoperative data were provided, or if <50% of the patients included in a single study underwent MV surgery. In total, 2056 articles were excluded and 12 articles were included in the final analysis (Fig. 1). The initial screening of the literature was performed by two authors and all full texts were reviewed when felt to meet the inclusion criteria. Statistical analysis of the pooled data was performed using RevMan 5.4 software (Cochrane Training, London, UK).

Risk of bias
All included manuscripts had a risk of bias assessment created to address confounding variables, selection bias, performance bias, detection bias, and reporting bias. Assessments were based on the ROBINS-I tool (Cochrane, London, UK) and previous descriptions in publications. [17][18][19] The risk of bias assessments were illustrated using RevMan 5.4 software (Cochrane Training) (Supplemental Fig. 1). Of the 12 included studies, 3 were assessed to be at high risk of bias, 4 at unclear risk of bias, and 5 at low risk of bias. Areas such as group selection, inclusion, exclusion, and statistical analysis were often felt to be performed with a low risk of bias. Areas of concern included patient matching, sample sizes, and follow-up.

Statistical analysis
A comparative analysis was performed between males and females undergoing MV surgery. Rates of in-hospital and 1-year mortality, MI, stroke, reoperation, acute kidney injury (AKI), and required pacemaker insertion were compared between the groups. In comparisons with considerable heterogeneity between studies, a random effect model was used. In comparisons where heterogeneity was not significant or substantial, a fixed-effect model was used. Odds ratio (OR) and 95% confidence intervals (CI) were reported using the Inverse Variance method in comparison with substantial heterogeneity and the Mantel-Haenszel method where heterogeneity was not substantial. Forest plots were used to represent the individual and pooled data. Heterogeneity between studies was measured using the chi-square statistic, and the I 2 statistic was used to identify degrees of difference thought to be due to effects size. A p value <0.05 was considered to be statistically significant for all analyses. The RevMan 5.4 software (Cochrane Training) was used to perform statistical analysis and in the creation of the forest plots. Table 1 summarizes the characteristics of the individual included studies, Table 2 summarizes mortality rates, and rates of postoperative morbidity are indicated in Table 3. Postoperative outcomes included mortality rates in-hospital, at 30 days, at 1 year, and at other follow-up times. Measures of postoperative morbidity include myocardial infarction (MI), stroke, reoperation rates, AKI, requirement of postoperative pacemaker implantation, readmission for heart failure (HF), and major adverse cardiac and cerebrovascular events (MACCE). Rates of morbidity and mortality were collected and averaged from each included study. In studies where either specific values or percentages were not provided, they were calculated based on the data provided and added to the overall average. Adjusted statistics were utilized when provided, and in studies where patients were propensity-matched, the propensity-matched data were used. Rates of morbidity and mortality were pooled, and meta-analysis was performed comparing males and females. Forest plots of the included outcomes are provided in Figs 2 and 3.

Overview of MV studies
In all, 12 studies were included in this review that explored the outcomes of surgical MV repair or replacement ± other procedures. A total of 129,479 patients were included in all studies with data collection dates ranging from 1999 to 2018. Included study types were retrospective and prospective studies. In all, four studies included propensity matching 20,21,23,24 while eight did not. Among studies that did not propensity match patients, Giustino et al. 5 found that males were more likely to have a history of smoking and ventricular arrhythmia and less likely to have diabetes, hypertension, and    (3) Females (%) P 51 (4) 0.17 Male MVR (%) 33 (7) Female MVR (%) P 40 (7) 25 found that males were younger, with higher rates of endocarditis and previous coronary artery bypass grafting compared to females. Namazi et al. 27 found higher BSA and rates of ischemic HF for males and higher rates of non-ischemic HF for females. Vassileva et al. found that among patients undergoing mitral valve repair (MVr), males were more likely to be younger, to have renal failure, and to have end-stage liver disease, while females were more likely to have hypertension, diabetes, previous stroke or transient ischemic attack, HF, history of respiratory failure, renal failure, atrial fibrillation, and anemia. In patients undergoing mitral valve replacement (MVR), males had higher rates of chronic obstructive pulmonary disease and renal failure, while females were more likely to have hypertension, diabetes, HF, history of respiratory failure, atrial fibrillation, and anemia. 28 Overall, in the unmatched patient populations, the
The 2-year mortality was found to be significantly increased in females in one study with 17.4% and 26.0% (p = 0.03) mortality for males and females, respectively. 5 The 5-year mortality ranged from 7.3% to 42% for males and 3.1% to 33% for females with no significant differences identified between the groups. 6,23,27 One study reported significantly different 10-year mortality with 28% mortality for males and 42% mortality for females at 10 years (p < 0.0001). 31

Myocardial Infarction
Two studies reported rates of MI postsurgical MV intervention. 20,21 One study reported significant differences with 7.9% of males and 4.7% of females (p = 0.04) exhibiting postoperative MI. 20

Postoperative pacemaker implantation
Pacemaker insertion was reported in four studies following MV surgery with no significant differences between the groups. 20,21,24,26 Pooled estimates of required postoperative pacemaker insertion favored females over males with an OR of 1.33 and 95% CI of 1.11-1.60, z-score = 3.15, p = 0.002 (Fig. 3D). A fixed-effects model was used due to insignificant heterogeneity identified between the included studies (χ 2 = 3.24, I 2 = 7%, p = 0.36) (Fig. 3D).

Readmission for Heart Failure
Readmission for HF was reported in three individual studies with no significant differences between males and females. 5,20,21 Major Adverse Cardiovascular and Cerebrovascular Events MACCE was reported in two studies for patients undergoing MV surgical interventions. 5,26 Females experienced higher rates of MACCE, with rates of 38.1% and 49% (p = 0.02) for males and females, respectively. 5

Discussion
The results of this review have revealed several noteworthy findings regarding the postoperative differences between males and females undergoing surgical MV interventions. First, consistent with previously established literature, a subset of individual studies found that females tended to have increased mortality in the short-and long-term following MV surgery. 4,23,28 Second, rates of postoperative morbidity varied, with females having significantly higher rates of stroke and MACCE in individual studies and males having increased rates of MI in another. 5,20,22 Individual studies did not report significant differences in reoperation, AKI, postoperative pacemaker implantation, or readmission for HF between males and females. Finally, pooled estimates of mortality found 1-year mortality rates in favor of males and pooled estimates of postoperative pacemaker insertion in favor of females (Figs 2 and 3, Supplemental Fig. 2). Pooled estimates of in-hospital mortality, postoperative stroke, reoperation, and AKI did not differ significantly between groups.
There is no shortage of literature on the discrepancies in outcomes between males and females undergoing surgical heart valve interventions, with most studies reporting inferior morbidity and mortality for females. 4,5,11,14,29,30 Despite this, the pooled data of this meta-analysis demonstrated contrary results with comparable outcomes for males and females. In addition, the results of the systematic review discovered that few studies were adequately powered or appropriately structured in order to make inferences about the relationship between sex and outcomes after MV surgery. Many studies investigating this relationship were excluded from this study due to numerous confounding variables such as concomitant procedures or MV surgery being performed in a minority of included patients. Of the included studies, only four propensity-matched males and females controlled for preoperative characteristics. In addition, only three of the propensitymatched studies controlled for surgical approach isolating for either MV repair or MV replacement. 20,21,23 Moreover, aside from the aforementioned articles, the studies that have isolated surgical approach, examining data exclusively on patients receiving either MV replacement or MV repair, did not propensity match males and females. 22,26 As previous literature has illustrated, females are more likely to receive MV replacement than males, and MV replacement is associated with increased rates of morbidity and mortality over MV repair. Therefore, it is essential that the surgical approach be controlled in addition to propensity matching in order to determine the true impact of sex on outcomes after MV surgery.
Attempting to address specific confounding variables, our research group has recently published the two aforementioned manuscripts examining sex differences in MV surgery by propensity-matching males and females and isolating for surgery type. 20,21 Trends in postoperative morbidity and mortality between males and females undergoing MVR and MVr were examined independently. Contrary to previously established literature, males and females demonstrated equivalent shortand long-term mortality rates in both studies. Furthermore, increased rates of postoperative morbidity, demonstrated by rates of MI and sepsis, were experienced by males following MVR. The main limitation of these two studies is relatively small sample sizes. 20,21 The etiology of this discrepancy in outcomes is still unclear with the following variables proposed-females are underrepresented in cardiovascular trials, present with VHD at an older age with more advanced disease, are with a worse preoperative risk profile, are referred for surgical intervention less often and/ or later in the disease process, and are more likely to receive replacement than their male counterparts. 8 -11,13,14,20,21,29,30,31 Therefore, determining whether the major risk factor for inferior outcomes following MV surgery is sex itself or the treatment that patients receive remains challenging. Large retrospective and prospective studies isolating for surgical approach, matching patients based on preoperative characteristics and comorbidities, and examining outcomes with long-term follow-up are required to elucidate this relationship.
The results of this review are of particular importance given the increasing prevalence of minimally invasive and transcatheter MV interventions. Trials will be needed to not only compare the safety and efficacy of these interventions but also compare trends and variations in postoperative outcomes between males and females. It is imperative that the aforementioned biases, including female underrepresentation, surgical technique, and preoperative comorbidities, be mitigated in these future trials and studies. In doing so, the factors contributing to the inequity in outcomes between males and females undergoing surgical valve interventions can be identified. Knowledge of the variables underlying the inequity may then permit individualized treatments for males and females, ultimately providing patients with the best possible care.

Limitations
First, the vast majority of studies included were retrospective and, therefore, had inherent limitations in controlling for confounding variables. Pooled estimates depicted by forest plots were based on data accumulated primarily from these retrospective studies. Finally, not all the retrospective studies included in the review propensity-matched males and females, further compounding the difficulty in disentangling the factors underlying the observed inequity in outcomes.

Conclusion
The results of this review contrast the established literature regarding sex differences in MV surgery which have shown inferior outcomes for females when compared to males. In this meta-analysis, females were found to have increased rates of short-term mortality in the pooled data while males had increased rates of pacemaker insertion. The remaining rates of morbidity and mortality did not differ significantly between males and females. In addition, this systematic review has found that relatively few studies were adequately powered or structured in order to investigate this topic. Few studies propensity matched males and females or isolated for MV repair or replacement. These confounding variables impair the ability to interpret the results found in the current literature and the true impact of sex on postoperative outcomes after MV surgery is still uncertain. Studies that propensity matched patients and isolated for surgical approach have found relatively equivalent outcomes between males and females after MV surgery indicating that sex alone may not explain the differences in outcomes experienced by males and females. Large retrospective and prospective studies isolating for surgical approach, matching patients based on preoperative characteristics and comorbidities, and examining outcomes with long-term follow-up are required to elucidate the true nature of this relationship.

Author contributions
Ryaan EL-Andari: Study design, literature search, data collection, data analysis, data interpretation, generation of figures, writing of the manuscript, approval of final manuscript.
Sabin J. Bozso: Study design, data analysis, data interpretation, generation of figures, writing of the manuscript, proof reading and revisions, approval of final manuscript.
Nicholas M. Fialka: Study design, literature search, data collection, data analysis, data interpretation, writing of the manuscript, approval of final manuscript.
Jimmy J.H. Kang: Study design, data collection, data analysis, data interpretation, writing of the manuscript, approval of final manuscript.
Jeevan Nagendran: Study design, data interpretation, writing of the manuscript, proof reading and revisions, approval of final manuscript.

Data availability statement
The data underlying this article are available in the article and in its online supplementary material.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the University Hospital Foundation, University of Alberta.

Supplemental material
Supplemental material for this article is available online.