The intersection of community socioeconomic factors with gender on outcomes after thoracic aortic surgery

Objective: We evaluated the relationship among community socioeconomic factors (poverty, income, and education), gender, and outcomes in patients who underwent ascending aortic, root, and arch surgery. Methods: For 2634 consecutive patients, we associated patients ’ ZIP codes with community socioeconomic factors. The composite adverse outcome comprised death, persistent neurological injury, and renal failure necessitating dialysis at discharge. Multivariable analysis and Kaplan – Meier survival curves were used. Men and women from the full cohort and from the elective patients were propensity matched. Results: Median follow-up was 3.6 years (interquartile range, 1.2-9.3). Men lived in areas characterized by less poverty ( P ¼ .03), higher household income ( P ¼ .01), and more education ( P ¼ .02) than women; likewise, in the elective cohort, all community socioeconomic factors favored men ( P (cid:2) .009). Female gender predicted composite adverse outcome ( P ¼ .006). In the propensity-matched women and men (820 pairs), the composite adverse outcome rates were 14.2 % and 11 % , respectively ( P ¼ .06). In 583 propensity-matched pairs of elective patients, men had less composite adverse outcome ( P ¼ .02), operative mortality ( P ¼ .04), and renal ( P ¼ .02) and respiratory failure ( P ¼ .0006). The 5-and 10-year survivals for these men and women were 74.2 % versus 71.4 % and 50.2 % versus 48.2 % ,

We analyzed data from a large group of patients who underwent ascending aortic, root, and arch surgery.We attempted to determine the implications of 3 community socioeconomic (CSE) factors-education, income, and poverty-and their association with gender for short-and long-term outcomes (Figure 1 and Video 1).

PATIENTS AND METHODS
From 2000 to 2020, 2634 consecutive patients underwent urgent/emergency or elective proximal aortic surgery-defined as ascending aortic, aortic root, or partial or total aortic arch repair-at our center.Baylor College of Medicine's Institutional Review Board approved our clinical research protocol (#18095) in February 2006.In December 2021, the approval was renewed until December 2026.The board also approved waiver of consent for patients who could not provide consent or whose family members were not available.The data were abstracted from a large aortic database that included prospectively maintained data from 2006 to 2020.For patients operated on before 2006, data were collected retrospectively from medical records with waiver of consent from these patients.Informed consent had been obtained when possible since 2006.
The end point, composite adverse outcome (CAO), comprised death (operative mortality, 30-day mortality, and in-hospital mortality), persistent neurological injury (stroke, paraplegia, and paraparesis), and renal failure necessitating dialysis at discharge.Secondary outcomes included shortterm postoperative outcomes, hospital length of stay (LOS), intensive care unit (ICU) LOS, and long-term survival.Patients were considered symptomatic if they presented with symptoms related to the aortic pathology (eg, chest pain or discomfort not related to coronary artery disease, back pain).Asymptomatic patients were those with known aortic pathology that was under surveillance or whose aortic pathology was an incidental finding from a recent routine physical or imaging performed for unrelated reasons.Data were collected according to previously defined descriptions of preoperative variables (Table 1), operative data and times (Table E1), and outcome variables (Table 2). 17Follow-up was obtained as previously described 9 and relied heavily on data from direct telephone communications, medical records, outpatient physician visits, and the Social Security Death Index (when it was available).
The latest UnitedStatesZipCodes.org version was used as the resource for US ZIP code data at the time of the patient's operation.Because our SE variables were based on data from the ZIP codes where each patient lived and not on individual income or education level, we refer to these variables as "CSE factors."These factors, derived from the publicly available ZIP code data, were percentage at or below the poverty line, median household income, and education level measured as percentage with at least a high school diploma, percentage with at least a bachelor's degree, and percentage with a graduate degree.

Intraoperative Details
8][19] Over time, we have changed some of our strategies; early on, femoral cannulation was used often, but later we started using right axillary cannulation, and for the past 10 years or more, we have preferred innominate cannulation.When hypothermic circulatory arrest is required, we use moderate hypothermia (20.1-28C).Early in the study period, we targeted the lower end of the moderate hypothermia range, but over the past decade or more, the target temperature has been approximately 24 C. Antegrade cerebral perfusion at flows of 10 to 15 mL/kg/min has been our main brain-protection strategy for the past 15 years.Bilateral antegrade cerebral perfusion, when feasible, is our preference.Regarding myocardial protection, we prefer to use both antegrade cardioplegia-administered through the aortic root or directly via the coronary ostia-and retrograde cardioplegia (with a 4:1 blood-to-crystalloid solution at 4 C) when feasible.Usually, cardioplegia is given every 20 to 30 minutes.For the past 2 to 3 years, we have used del Nido cardioplegia (same dose, interval, and temperature).Table E1 shows the intraoperative details.

Statistical Analysis
SAS Version 9.4 (SAS Institute Inc) was used for all statistical analyses.All tests were 2-sided.P values were not adjusted for multiplicity.A 1:1 greedy match algorithm without replacement, based only on the preoperative characteristics shown in Table 1, was used for propensity matching.The CSE factors were not adjusted for in any analysis.Continuous data were expressed as median and interquartile range (IQR), and categorical data were expressed as number (percentage).Potential differences in the distribution of preoperative and intraoperative data and in the short-term outcomes of the unmatched patients were tested with the chi-square test for categorical data and the Wilcoxon 2-sample test for continuous data because of their non-normal distribution.For the propensity-matched patients, potential differences between the genders in the preoperative data were tested by the standardized mean difference.The McNemar test for the categorical outcomes and the Wilcoxon signed-ranks test for the continuous outcomes were used to evaluate differences between the genders.
Multivariable analysis was performed to determine the predictors for our primary end point (CAO) and the other short-term outcomes.The modeling was done with logistic regression (PROC LOGISTIC with STEP-WISE selection).The multivariable analysis for hospital and ICU LOS implemented generalized linear models (PROC HPGENSELECT with a Poisson distribution function and STEPWISE selection) because of the non-normal distribution of the residuals.The variables considered for these models were gender and all the preoperative characteristics except body mass index (which had too many missing values), year of the procedure, median household income, percentage of population living below the poverty level, percentage with at least a high school diploma, percentage with at least a bachelor's degree, aortic root bioprosthetic replacement, and all the intraoperative characteristics shown in Table E1.
Because preliminary descriptive analysis revealed that the association of CSE factors with type of admission differed between men and women, interaction terms between gender and CSEs were included in the model to account for this effect modification.In addition, the models were stratified by gender and run on the male and female patients separately.Odds ratios (ORs) and 95% confidence intervals (CIs) are used to report the regression results for the PROC LOGISTIC models, and parameter estimates with 95% confidence limits are used to report results of the PROC HPGENSE-LECT models.Multicollinearity was verified to not be an issue upon inspection of the variance inflation factors.To assess long-term survival differences between women and men, Kaplan-Meier survival curves were used (PROC LIFETEST).A Cox proportional hazards model (PROC PHREG) using the survival data, demographic characteristics, preoperative and intraoperative variables, and CSEs was done to explain the effect of these variables on the hazard rate.The proportional hazards assumption was checked with the Kolmogorov-type supremum test.For the propensity-matched patients, a robust sandwich variance estimate was used to test for significant differences between the survival curves.A subset analysis of acuity (urgent/emergency vs elective) for women and men prompted a propensity match on preoperative clinical characteristics to examine the association among CSE factors, gender, and outcomes for only the elective patients.Sensitivity analyses included comparing the CSE factors by acuity and by acuity sorted by gender, as well as considering the CSE factors for those with or without the adverse outcomes.

Unadjusted Data
In the overall cohort of 2634 patients, men lived in areas with higher median household income (P ¼ .01)and a lower percentage of people living below the poverty level (P ¼ .03).Men also lived in areas characterized by more education, because people in these areas more often had at least a high school diploma (P ¼ .02)or at least a bachelor's degree (P ¼ .01)than women (Table 3 and Figure 2).
All CSE factors were significantly associated with acuity; lower percentage of population living below the poverty level, higher median household income, and greater education (higher percentage of high school diploma, bachelor's degree, and graduate degree) were significantly more common in patients with elective versus urgent and emergency cases (P <.0001 for all) (Table 4).

Conclusion
The associations among CSE factors (poverty, income, education), gender, and outcomes in aortic surgery patients were: • Female gender was associated with adverse outcome • Men had better outcomes in elective cases, and all CSE factors favored men (non-significantly) in propensity-matched groups We conclude that patient-specific CSE information should be included in national databases.

ADULT
In the unmatched elective cohort, significantly fewer men than women had the CAO (102 [8.0%] vs 75 [12.4%];P ¼ .0022).In contrast, in the urgent/emergency cohort, the CAO rate did not differ significantly between men and women (107 [20.6%] for men, 46 [19.7%] for women, P ¼ .78).Table 5 shows the CSE factors and their interactions with gender and acuity in the overall unmatched group.Tables E2-E9 show the association between CSE and outcomes (CAO, operative mortality, stroke, renal failure, pericardial effusion requiring drainage, respiratory failure, myocardial infarction, and long-term survival).

Propensity-Matched Data
Standardized mean differences showed that propensity matching balanced the female and male patients regarding their preoperative characteristics, both in the patients drawn from the total cohort (Table 1) and in those drawn from the elective cohort only (Table E10).In the 820 propensitymatched pairs of women and men drawn from the total cohort, the CAO was less frequent in men (n ¼ 90, 11%) than in women (n ¼ 116, 14.2%) (P ¼ .06)(Table 2).Among all patients in these propensity-matched groups, CSE factors were more favorable in men, but not significantly so (Table 3).CSE factors in patients who experienced a CAO are shown in Table E11.

Multivariable Analysis
Composite adverse outcome.Female gender was an independent risk factor for the CAO (OR, 1.47, 95% CI, 1.12-1.93,P ¼ .006)(Table E12).Preoperative pulmonary disease and greater age at admission were independent risk factors for men (OR, 1.72, 95% CI, 1.17-2.52,P ¼ .006;OR, 1.04, 95% CI, 1.02-1.05,P .0001)and women (OR, 1.66, 95% CI, 1.04-2.65,P ¼ .03;OR, 1.04, 95% CI, 1.03-1.06,P .0001).Acute symptoms independently predicted the CAO in women (OR, 2.42, 95% CI, 1.49-3.93,P ¼ .0004),but not in men.CSE factors were not associated with the CAO in either men or women in the overall cohort.Operative mortality.Risk factors for mortality and all multivariable results are shown in Table E12.CSE factors were not associated with operative mortality in either men or women.Other outcomes and community socioeconomic factors.CSE factors were not associated with rate of stroke or postoperative renal dysfunction.Among the CSE factors, respiratory failure was associated with the interaction of gender and lower education level (P ¼ .007)and, for the men only, with percentage of the population below the poverty level (OR, 1.03, 95% CI, 1.01-1.04,P ¼ .0004).The interaction between gender and lower education significantly predicted ICU LOS and hospital LOS (P <.0001).The interactions between gender and both percentage below the poverty level and median household income were also associated with ICU LOS (P <.0001).

Presentation and Community Socioeconomic Factors
All CSE factors were negatively associated with symptoms (P<.0001).In asymptomatic patients, the median percentage of population living below the poverty level was 11.4% (IQR, 6. 4   E13).

Follow-up and Survival
For all patients, median follow-up was 3.6 years (IQR, 1.2-9.3)and maximum follow-up was 21.1 years.For the elective patients, median follow-up was 3.4 years (IQR, 1.1-8.9)and maximum follow-up was 21.1 years.The unadjusted 5-and 10-year survivals for the overall cohort were 70.2% and 50.8% for the male patients versus 68.1% and 48.1% for the female patients, respectively (P ¼ .02).The Cox regression analysis (verified to satisfy the proportional hazards assumption with the supremum test P ¼ .42)validated these results, showing that long-term survival differed significantly between men and women (hazard ratio, 0.856, 95% confidence limit, 0.734-0.997).For the 820 propensity-matched pairs drawn from the full cohort (ie, both elective and urgent/emergency cases), 5-and 10-year survivals were 66.7% and 43.2% for the men versus 69.2% and 49.1% for the women, respectively (P ¼ .73).For the 583 pairs of propensity-matched elective patients, 5-and 10-year survivals were 74.2% and 50.2% for the male patients versus 71.4% and 48.2% for the female patients, respectively (P ¼ .05). Figure 3 shows Kaplan-Meier survival curves for the propensity-matched pairs drawn from the full cohort and from the elective cases only.

DISCUSSION
Because SE factors are vital determinants of health, there is increasing focus on minimizing healthcare disparities caused by SE differences, with the goal of improving outcomes. 20,21ecent reports emphasize the judicious and thoughtful use of social factors in the Society of Thoracic Surgeons risk models. 22,23Another report describes the importance of social factors such as low income and low education level, and their association with mortality. 24For patients undergoing aortic surgery specifically, the implications of SE factors for outcomes and how these factors interact with gender are not well understood.
In this study, we found that in unadjusted data from patients who underwent ascending aortic, root, and arch surgery, male patients more often lived in areas with a lower percentage living below the poverty level and a higher median household income than female patients.In addition,   men lived in areas characterized by higher levels of education, having populations with higher percentages of high school diploma, bachelor's degree, and graduate degree recipients.This is not surprising, given that 2015 US Census Bureau data show that women had a higher percentage living below the poverty level than men. 25he rates of the CAO, renal insufficiency, respiratory failure, and hemodialysis dependence, along with hospital and ICU LOS, were all significantly lower in male versus female patients.When men and women who had elective operations were propensity matched, better outcomes were seen in the men.The multivariable analysis did not associate any of the CSE factors with the CAO or operative mortality, although a trend toward more favorable CSE factors for men was seen in the propensity-matched groups.Female gender was an independent risk factor for the CAO.In our previous study, 9 which spanned a decade more than the current study, we similarly found less favorable results for women regarding respiratory failure, hospital LOS, and ICU LOS, but female gender was not an independent predictor of the CAO as it was in the present study.The reason for this discrepancy is unclear, but in both studies, there was a trend toward worse outcomes for women in the propensity-matched data.
Living in an area with a lower education level was associated with respiratory failure, which can lead to prolonged hospital and ICU LOS.This association may be due to greater tobacco use among those who did not graduate from high school or earn a college degree, 26,27 or to greater environmental pollution in distressed neighborhoods.
Health literacy varies among patients and affects health outcomes within the broader sphere of healthcare delivery. 28Tasks such as reading an informed consent document and understanding all the risks and benefits of a proposed surgical procedure can be challenging for patients with a lower literacy level, which is associated with poorer comprehension.Restructuring informed consent documents as educational material has been suggested. 29Most institutional review boards require that informed consent documents be written at the seventh-grade reading level.
An especially important finding of our study is the interaction among CSE factors, acuity, and gender.A lower percentage of the population living below the poverty level, higher income level, and higher education level were significantly associated with elective (vs urgent/emergency) status in the overall group.A report from our center showed that patients with lower SE status presented with significantly higher degrees of acuity for thoracoabdominal aortic aneurysm repair. 16In a recent report by Lin and colleagues 30 examining potential disparities in access to   been shown. 8Mehaffey and colleagues 2 searched the Society of Thoracic Surgeons Adult Cardiac Surgery Database for patients who underwent isolated coronary artery bypass surgery and found that patients from distressed communities were at increased risk for adverse events and death.However, the authors reported no specific associations with gender.
Our study showed that among elective patients, men had significantly fewer adverse outcomes than women.In addition, a trend toward more favorable CSE factors in men was noticed.Access to care, regular follow-up with the patient's doctor's office, and following physician advice and medication plans can be of crucial importance for elective surgery patients.Higher percentage living below the poverty level, lower income, and poor education can interfere with these activities, thus creating barriers to scheduling and having elective procedures.Adding to these other social circumstances that disproportionately affect female patients, such as lack of affordable childcare, lack of access to capital, and single-mother families, can explain why elective procedures have better outcomes for men versus women.
Understanding how outcomes are affected by gender and social determinants of health can be valuable as we transition into value-based payment models and care.Adjusting payment and reimbursement, and not physicians' performance scores, for social factors means that society rewards providers for caring for socioeconomically vulnerable and disadvantaged patients.This is important, but it must be done without lowering quality standards.In addition, focus on postdischarge care or on more detailed discharge planning based on the patient's SE condition could potentially prevent unnecessary readmissions or emergency department visits.Creating neighborhood clinics and off-campus initiatives could help individuals living in disadvantaged communities. 32Making such adjustments shows that we understand that favorable outcomes can be more difficult to achieve or, more important, may require different resources for patients with SE disadvantages.It is a step toward investment in addressing the social determinants of health, which could potentially reduce spending on traditional care delivery.

Study Limitations
Our study has several limitations.We do not have specific patient-level data regarding living below the poverty level, income, and education level.In addition, other CSE factors, such as unemployment, housing vacancy rate, and change in employment, were not available, and marital status and insurance information were not included.
Furthermore, the relationship between SE markers and cardiovascular outcomes changes substantially over an individual's life course.The data used in the current analysis-CSE factors based on ZIP codes-represent a snapshot in time and are not contemporary point-of-care data, which would incorporate all the SE markers for individual patients.The CSE factors associated with the patient's ZIP code were obtained during the latest year of the study and thus do not account for potential changes of these factors during the study period.Also, these factors were not adjusted for inflation.No additional analysis was performed to confirm that the SE status of neighborhoods did not change over time.However, according to a post on the unitedstateszipcodes.org website, "Over a 5 year period.anindividual ZIP code is likely to vary by less than 10%.That leaves a small amount of room for error."In addition, the census is taken every 10 years.Another limitation of our study is that the denominator is missing regarding how many patients with aortic disease and multiple comorbidities were not sent to surgeons and how many such patients were not offered surgery because of their high risk.In addition, whether and how many patients were not referred to our aortic center because of SE or insurance factors could not be known.Finally, potential gender differences in referral could not be analyzed.
This study examined the interaction among CSE factors, gender, and outcomes in aortic surgery in a high-income country; results may not be generalizable to low-or middle-income countries.Furthermore, the state of Texas did not participate in Medicaid expansion; the results might have been different in another state.In addition, race and its interaction with gender and CSE factors were not examined.
Nevertheless, this is one of the first studies to examine the impact of SE factors along with gender on outcomes in patients who undergo surgery in the proximal aorta.Furthermore, there is a clear need for a sturdy SE index for every patient in the various databases, especially the national databases.Our study highlights the convoluted interaction among gender, poverty, income, and education and its effects on outcomes in patients who undergo complex thoracic proximal aortic surgery.

CONCLUSIONS
Among patients who underwent surgery in the ascending aorta, aortic root, or aortic arch, CSE factors such as percentage of the population living below the poverty level, household income, and level of education were more favorable in men than in women.These CSE factors were more favorable in elective and asymptomatic patients.Female gender was an independent risk factor for the CAO.Living in an area with a lower education level independently predicted respiratory failure and prolonged LOS.These data can be used as a benchmark and as a basis for more complete and granular studies that will help us use SE factors to identify patients who might benefit from targeted mea- Community Socioeconomic (CSE) Factors with Gender on Outcomes after Thoracic Aortic Surgery CSE Factors Favored Men / P = .01)( CSE Factors Favored Men / Non-significant) ( CSE Factors Favored Men / Non-significant) (

FIGURE 1 .
FIGURE 1.A visual summary of the study and its findings.USD, United States dollars; AE, adverse outcome; ICU, intensive care unit; LOS, length of stay.
and Men with Graduate Degree

FIGURE 2 .
FIGURE 2. Box-and-whisker plots showing that on average, men lived in areas characterized by higher median household income (top) and a higher median percentage of people with graduate degrees (bottom).The upper and lower borders of each box represent the lower and upper quartiles (25th and 75th percentile), the middle horizontal line represents the median, the lower and upper whiskers represent the minimum and maximum values of nonoutliers, and the extra dots represent outliers.USD, United States dollars.

FIGURE 3 .
FIGURE 3. Kaplan-Meier survival curves for propensity-matched pairs of men and women drawn from the full cohort (A; 820 pairs) and from the elective patients only (B; 583 pairs).
sures and specialized protocols intended to minimize the risk of adverse outcome.The Journal of Thoracic and Cardiovascular Surgery c December 2023 Adult: Aorta Preventza et al ADULT Webcast You can watch a Webcast of this AATS meeting presentation by going to: https://www.aats.org/resources/2158.Conflict of Interest Statement O.P. is a consultant for Terumo Aortic and WL Gore & Associates.S.C. has served on advisory boards for Edwards Lifesciences, La Jolla Pharmaceutical, Baxter Pharmaceuticals, and Eagle Pharmaceuticals.S.A.L. is supported in part by the Jimmy and Roberta Howell Professorship in Cardiovascular Surgery at Baylor College of Medicine, serves as a consultant for Terumo Aortic and Cerus, and serves as a principal investigator for clinical studies sponsored by Terumo Aortic and CytoSorbents.J.S.C. serves as principal investigator, consults for, and receives royalties and a departmental educational grant from Terumo Aortic; consults and participates in clinical trials for Medtronic, Inc, and WL Gore & Associates; and participates in clinical trials for Abbott Laboratories, CytoSorbents, Edwards Lifesciences, and Artivion.All other authors reported no conflicts of interest.The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest.The editors and reviewers of this article have no conflicts of interest.The Journal of Thoracic and Cardiovascular Surgery c December 2023 Adult: Aorta Preventza et al ADULT

TABLE 3 .
Community socioeconomic factors in the overall cohort and in the matched patients stratified by gender in all cases and in elective cases Data are presented as median (25%-75% IQR).SMD, Standardized mean difference; USD, United States dollars.*Wilcoxon2-sampletest.1576TheJournal of Thoracic and Cardiovascular Surgery c December 2023

TABLE 5 .
Community socioeconomic factors stratified by gender and acuity of surgery

TABLE 6 .
Outcomes for the propensity-matched elective patients 31ta are presented as n (%) or median (25%-75% IQR).MI, Myocardial infarction; ICU, intensive care unit; LOS, length of stay.*McNemar or Wilcoxon signed-ranks test.The Journal of Thoracic and Cardiovascular Surgery c December 2023 ADULT elective surgical procedures during the recovery phase of the Coronavirus Disease 2019 pandemic, the number of patients with procedure requests was lowest in those with disadvantaged SE status.Patient access to elective surgery was related to insurance, SE status, distance from care, language, and marital status.Rozental and colleagues31showed that after abdominal aortic aneurysm repair, uninsured and Medicaid patients had higher mortality rates than private insurance beneficiaries.Similar associations between insurance and outcomes in cardiac surgery have

TABLE E5 .
Association between community socioeconomic factors and postoperative renal failure

TABLE E6 .
Association between community socioeconomic factors and pericardial effusion Data are presented as median (25%-75% IQR).USD, United States dollars.*Wilcoxon 2-sample test.The Journal of Thoracic and Cardiovascular Surgery c Volume 166, Number 6 1582.e3

TABLE E10 .
Preoperative variables for propensity-matched elective patients only Data are presented as n (%) or median (25%-75% IQR).SMD, Standardized mean difference.The Journal of Thoracic and Cardiovascular Surgery c Volume 166,Number 61582.e5

TABLE E11 .
Community socioeconomic factors in propensity-matched patients who had the composite adverse outcome Data are presented as median (25%-75% IQR).SMD, Standardized mean difference; USD, United States dollars.1582.e6The Journal of Thoracic and Cardiovascular Surgery c December 2023 The Journal of Thoracic and Cardiovascular Surgery c Volume 166,Number 61582.e7