Females have an increased risk of short‐term mortality after cardiac surgery compared to males: Insights from a national database

Abstract Objectives Female sex is considered a risk factor for mortality and morbidity following cardiac surgery. This study is the first to review the UK adult cardiac surgery national database to compare outcomes following surgical coronary revascularisation and valvular procedures between females and males. Methods Using data from National Adult Cardiac Surgery Audit, we identified all elective and urgent, isolated coronary artery by‐pass grafting (CABG), aortic valve replacement (AVR) and mitral valve replacement/repair (MVR) procedures from 2010 to 2018. We compared baseline data, operative data and outcomes of mortality, stroke, renal failure, deep sternal wound infection, return to theater for bleeding, and length of hospital stay. Multivariable mixed‐effect logistical/linear regression models were used to assess relationships between sex and outcomes, adjusting for baseline characteristics. Results Females, compared to males, had greater odds of experiencing 30‐day mortality (CABG odd ratio [OR] 1.76, confidence interval [CI] 1.47−2.09, p < .001; AVR OR 1.59, CI 1.27−1.99, p < .001; MVR OR 1.37, CI 1.09−1.71, p = .006). After CABG, females also had higher rates of postoperative dialysis (OR 1.31, CI 1.12−1.52, p < .001), deep sternal wound infections (OR 1.43, CI 1.11−1.83, p = .005) and longer length of hospital stay (β 1.2, CI 1.0−1.4, p < .001) compared to males. Female sex was protective against returning to theater for postoperative bleeding following CABG (OR 0.76, CI 0.65−0.87, p < .001) and AVR (OR 0.72, CI 0.61−0.84, p < .001). Conclusion Females in the United Kingdom have an increased risk of short‐term mortality after cardiac surgery compared to males. This highlights the need to focus on the understanding of the causes behind these disparities and implementation of strategies to improve outcomes in females.


| MATERIALS AND METHODS
The NICOR registry prospectively collects demographic, pre and postoperative clinical information, including mortality, for all major adult cardiac surgery procedures performed in the United Kingdom. The flow of the data from surgeon-input to analysis has been described elsewhere. 13 Briefly, data entered locally by surgeons are validated at the unit-level by database managers before upload NICOR. At this stage, further validation is performed according to logical rules and missing data reports are generated for primary variables (e.g., EuroSCORE risk factors, patient identifiers and outcome data). The data are cleaned by the academic healthcare informatics department. The complete data cleaning process has been previously described. 13 Duplicate records are removed, transcriptional discrepancies re-coded and clinical and temporal conflicts resolved. Missing data are resolved during the validation stages of the data transfer from individual centers. Missing and conflicting data for inhospital mortality status are backfilled and validated via record linkage to the Office for National Statistics census database. The overall percentage of missing data for baseline information is very low (1.7%). Missing categorical or dichotomous variable data were imputed with the mode while missing continuous variables data imputed with the median.
For this study, we used the NACSA data set to include all adult patients who underwent cardiac surgery between 2010 and 2018.
We included elective and urgent isolated coronary revascularization and valvular procedures (AVR, mitral valve replacement/repair [MVR]). We excluded minor procedures, aortic arch surgery, heart transplantation and emergency, and salvage surgery.

| Statistical analysis
A Shapiro−Wilks test was used to assess normality of distribution of continuous data. Data of normal distribution was averaged as a mean with standard deviation and analyzed using student t-test. Nonnormally distributed data was averaged as a median with interquartile range and analyzed using a rank sum test. Categorical data is presented as frequencies and compared using a χ 2 test. For binary outcomes, a logistical regression model was used. For continuous outcomes, a linear regression model was used. Multivariable analysis was used to assess relationships between sex and our outcomes, adjusting for baseline characteristics including age, body mass index (BMI), smoking status, diabetes, chronic obstructive pulmonary disease (COPD), renal failure, cerebrovascular disease, peripheral vascular disease, atrial fibrillation, and hypertension. p < .05 was considered significant in all the analysis. Statistical analysis was performed using R version 4.0.0 using the packages sjplot, lme4, lmertest, gtsummary, and ggplot2.

| Patient characteristics
CABG and AVR were more commonly performed in males than females whereas more females underwent MVR than males ( Figure 1 shows the proportion of each intervention per sex).

| CABG
During the study period, 121,319 males (82.3%) and 26,157 females (17.7%) underwent an elective or urgent isolated CABG (see Table 1). Females tended to be older with more comorbidities such as diabetes, COPD and hypertension. Females were more likely than males to have an urgent rather than elective CABG and females scored more highly in New York Heart Association and Canadian Cardiovascular Society scoring systems. Females had shorter cross clamp times and cardiopulmonary bypass (CPB) times (see Table 2). Male were more likely to receive internal mammary artery grafts including left internal mammary artery (LIMA), right internal mammary artery (RIMA) or bilateral internal mammary arteries (BIMA) than females. However, a higher proportion of females received total arterial grafts than males.

| AVR
Twenty-six thousand seven hundred forty-two males (58.3%) and 19,168 females (41.7%) underwent an elective or urgent isolated AVR (see Table 3). Females tended to be older but were more likely to have an elective rather than urgent AVR compared to males (see supplementary

| MVR
Seven thousand nine hundred ninety-one males (47.7%) and 8778 females (52.3%) underwent an elective or urgent isolated MVR (see Table 4). Females were more likely to have an elective rather than urgent MVR compared to males. Females tended to be older with higher rates of COPD and diabetes.
Females had shorter cross clamp times and CPB times than males. Males were more likely than females to have a MV repair rather than a MV replacement (see Supporting Information: Table 2).

| Outcomes
The rate of 30-day mortality by procedure and within each sex stratum is presented in Figure 2.

| CABG
Females, compared to males, had greater odds of experiencing  Table 6).

| DISCUSSION
This study is the first to review current practice of UK national data to compare sex-related differences in outcomes following surgical coronary revascularisation and valvular cardiac procedures. symptoms to those thought of as "typical" 14 The present study supports the well reported claim that females undergoing coronary revascularisation surgery are often older and with more comorbidities than males. 3 Furthermore, we found that women were also more likely to need urgent, as opposed to elective, revascularisation than men, which may be responsible for some of the poorer outcomes reported. Our findings somewhat agree with a nationwide study from Sweden that females had higher risk of mortality following CABG due to poorer preoperative risk profiles, 7 however our study additionally found female sex was an independent risk factor after multivariable analysis. This difference may be due to the Sweden inclusion criteria of adults under 50 years of age whereas our study included all adult patients.
It is also suggested that sex-related differences in operative strategy decisions and techniques may explain sex-related differences in cardiac surgery outcomes. 14  Nevertheless, the multivariable regression analysis used in our study adjusted for differences in baseline and operative differences, including revascularisation strategy and still a sexrelated difference remained in short-term mortality. These findings suggest that female sex is an independent risk factor for short-term mortality following CABG which supports the consensus of the current literature. 6,19 The idea of female sex being an independent risk factor for worse outcomes following CABG is speculated to be related to the more challenging anatomy of female patients, such as smaller coronary artery targets for grafting, narrower conduits and more diffuse patterns of coronary disease. 3,20 This is consistent with the findings of national level database review from Italy who ascribed the sexrelated difference in mortality to the differing operative strategies adopted for the more technically challenging coronary revascularisation of female patients. 9 T A B L E 4 (Continued) Our study also evaluated other post-CABG outcomes. We did not find an increased risk of stroke following CABG as other national studies have reported. 8,21 This may be related to the fact that in our cohort of females a significant 15% underwent off pump revascularisation which has been reported to be particularly beneficial in women because of its effect to reduce the risk of stroke. 22 Sternal wound complications were more common in females than males following CABG in our study. A risk prediction tool developed in the United Kingdom identifies female sex as one of six independent predictors of surgical site infection following cardiac surgery. 23  with significantly higher short-term mortality following both isolated AVR and MVR procedures.
Our finding of increased mortality following AVR in females is reflected from other nationally representative databases such as the USA 24,25 and a previous UK database analysis. 26 Our findings of poorer preoperative health status in females was also shown in data from United States and similarly, even after adjusting for these differences, females still had higher mortality rates following AVR. 25 However, other national studies did not report sex-related differences in AVR mortality. 12,27 In our study, men were more likely to receive a mechanical aortic valve than women which may reflect the differences in age and comorbidities between the sexes at time of surgery and their influence on the management planning. It is known that women with severe aortic stenosis are diagnosed at a later stage of the disease process 28 but even when adjusting for preoperative difference women are less likely to be referred for surgical AVR than men. 25,29 There is no clear explanation for why women have worse outcomes compared to men following AVR but several mechanisms have been implicated. For similar degrees of aortic stenosis, females tend to have higher transvalvular pressure gradients, thicker ventricle walls and smaller end-systolic and end-diastolic chamber sizes than males. 30 Second, females on average receive smaller valves than males, the outcomes of patient-prosthesis mismatch seem more severe in smaller size valves 31 and therefore may effect women disproportionately. Furthermore, females are also more likely to require additional aortic annular enlargement than males leading to increased operative risk associated with the annular enlargement procedure. 5 As with the other procedures, females in the United Kingdom