Racial disparities in thoracic aortic surgery: Myth or reality?

Objective: We examined the relationship between Black or White race and adverse outcomes in patients who underwent surgery of the ascending aorta, aortic root, or aortic arch at our center. Methods: We analyzed 2335 consecutive patients who identi ﬁ ed as Black (n ¼ 217, 9.3 % ) or White (n ¼ 2118, 90.7 % ) and underwent proximal aortic surgery. Patient zip codes were used to determine community socioeconomic (CSE) characteristics. The composite adverse outcome comprised mortality, persistent neurologic injury, and renal failure necessitating dialysis at discharge. We performed multivariable analysis, Kaplan – Meier analysis, and propensity score matching adjusted for CSE factors. Results: Median follow-up time was 3.7 years. Compared with White patients, Black patients lived in areas characterized by a higher percentage living below poverty level, lower income, and lower education level ( P < .0001). Black patients had higher rates of emergency presentation ( P < .0001) and lower 5-and 10-year survival rates ( P ¼ .0002). Short-term outcomes were similar between groups, except for respiratory failure and length of stay ( P < .0001), which were higher in the Black population. After propensity score matching adjusted for CSE factors, Black and White patients (n ¼ 204 each) had similar short-term outcomes and 5-and 10-year survival rates ( P ¼ .30). Multivariable analysis strati ﬁ ed by race showed that CSE factors independently predicted adverse outcomes in Black but not White patients. Conclusions: This is among few studies

socioeconomic factors 3,[5][6][7][8] ; however, whether race is specifically associated with adverse outcomes after cardiac surgery remains uncertain.][17] In this study, we investigated the relationship between Black or White race and early outcomes in a cohort of consecutive patients who underwent thoracic surgery of the ascending aorta, aortic root, or aortic arch at our academic center over a span of 20 years.In addition, we evaluated the association between race and long-term survival in this patient cohort (Figure 1, Video 1).

PATIENTS AND METHODS Study Design
The Institutional Review Board at Baylor College of Medicine approved this study and waived the requirement of consent from patients or families of patients who were not able to provide consent (#18095, approved and renewed until December 2026).We reviewed data for 2335 consecutive patients who underwent surgery of the aortic root, ascending aorta, or aortic arch during a recent 20-year period (2000-2020) at our center; all of these patients identified their race as either Black or White.Patients from other racial groups were excluded.Data were extracted from our large database of aortic procedures that includes prospectively collected data from 2006 to present and retrospective data collected before 2006 from medical records.The index procedure for these patients was proximal aortic surgery, defined as ascending aortic surgery procedures, aortic root surgery, or aortic arch surgery.The overall composite adverse outcome was defined as operative mortality or 30-day death, persistent neurologic event (eg, paraplegia, stroke, and paraparesis), and renal failure requiring hemodialysis at hospital discharge.The secondary outcomes were short-term postoperative outcomes (eg, operative mortality, stroke, renal failure, bleeding requiring reoperation, pericardial effusion requiring drainage, respiratory failure, myocardial infarction, hospital length of stay [LOS], and intensive care unit [ICU] LOS) and long-term survival at 5 and 10 years.Outcomes definitions and definitions of preoperative and intraoperative variables were previously described. 18Follow-up data were obtained from patient visits, medical records, communication with families and doctors' offices, and the Social Security Death Index, when available.The latest version of UnitedStatesZipCodes.org was used as the resource for identifying the following socioeconomic factors: percent of the population living below the poverty level, education level (ie, percent with a graduate degree, percent with at least a bachelor's degree, and percent with at least a high school diploma), and median household income.Because these socioeconomic factors were based on zip codes and provided information only about where the patient lived and not about the individual's income or education level, we referred to the related factors as community socioeconomic factors.

Intraoperative Details
Over the span of 20 years, our surgical strategies have evolved; cannulation strategies have shifted from the femoral artery to the axillary artery to the innominate artery.Innominate artery cannulation has been our preferred cannulation strategy for more than a decade. 19Step-by-step approaches on redo scenarios or on acute type A (DeBakey I or II) aortic dissection have been previously described. 20,21Brain protection strategies used during the last 15 years of the study were based on antegrade cerebral perfusion (mainly bilaterally when feasible) and on temperature management with moderate hypothermia (20.1-28C). 22 For myocardial protection, we use antegrade and retrograde cardioplegia every 20 to 30 minutes (with a 4:1 blood-tocrystalloid solution at 4 C).During the last 2 to 3 years, we have used del Nido cardioplegia (same temperature, interval, and dose).

Statistical Analysis
Statistical analyses were performed by using SAS version 9.4 (SAS Institute Inc).All tests were 2-sided.P values were not adjusted for multiplicity.To balance the preoperative risk factors between the groups, we performed propensity score matching on the basis of the preoperative characteristics listed in Table 1, including the community socioeconomic factors, as well as the year of the surgery.A 1:1 greedy matching algorithm was used without replacement.Body mass index was excluded from propensity matching because too many patient values were missing.All continuous data were expressed as the median with a 25% to 75% interquartile range (IQR) due to their predominantly nonnormal distribution.Categorical variables were expressed as actual numbers and percentages of the total.Differences in the distribution of preoperative and intraoperative characteristics, as well as short-term outcomes for the unmatched cohort, were tested with the c 2 test, or the Fisher exact test when necessary, for the categorical variables and the Wilcoxon 2-sample test for the continuous variables.For the propensity-matched cohort, differences in the distribution of preoperative and intraoperative characteristics were measured with the standardized mean difference, whereas the paired t test and Wilcoxon signed-rank test were used on the outcomes.
To examine the predictors for our primary end point of the composite adverse outcome and other short-term outcomes, multivariable analysis was conducted with logistic regression (PROC LOGISTIC in SAS with STEPWISE selection).Because of the nonnormal distribution of the residuals, the multivariable analysis for ICU and overall LOS was conducted with generalized linear models (PROC HPGENSELECT in SAS with a Poisson distribution function and STEPWISE selection).The variables considered for the multivariable analysis models were as follows: intervention year, race, gender, patient age at admission, acute subdissection, chronic dissection, heritable thoracic aortic disease, urgent/emergency surgery, hyperlipidemia, hypertension, prior stroke, diabetes, chronic renal insufficiency, tobacco use, pulmonary disease, redo sternotomy, peripheral vascular disease, symptomatic, acute symptoms only, percent of the population living below poverty level, median household income, percent with at least a bachelor's degree, percent with at least a high school diploma,

Abbreviations and Acronyms
ICU ¼ intensive care unit IRAD ¼ International Registry of Acute Aortic Dissection LOS ¼ length of stay Scanning this QR code will take you to the table of contents to access supplementary information.The Journal of Thoracic and Cardiovascular Surgery c January 2024 Adult: Aorta Preventza et al ADULT aortic root mechanical, aortic root bioprosthetic, aortic root valve-sparing procedure, aortic valve replacement bioprosthetic, aortic valve replacement mechanical, concomitant mitral and tricuspid valve repair, concomitant coronary artery bypass grafting, elephant trunk graft replacement, full aortic graft replacement, and cardiopulmonary bypass time.Upon inspecting the variance inflation factors, we verified that multicollinearity was not an issue.For sensitivity analysis, the models were stratified by race to allow different models for the Black and White patient groups.Kaplan-Meier survival curves were used to assess long-term survival rate differences between Black and White patient groups.We performed regression analysis of the survival data based on the Cox proportional hazards model to explain the effect of explanatory variables on the hazard rate (PROC PHREG).We also performed a subgroup analysis of elective-only surgeries for Black and White patients to examine the effect of socioeconomic factors along with race on outcomes without the emergency cases included.

Unadjusted Data
Preoperative and intraoperative variables for the entire cohort.Patient characteristics are shown in  1).The areas in which Black and White patients lived also differed with respect to resident education level.The median percent of the population living in areas with residents who had at least a high school diploma, a bachelor's degree, or a graduate degree was higher for White than Black patients (Table 1).

Conclusion
The associations between race and outcomes in aortic surgery patients were: The complex role of race in health-related outcomes requires examining the patient's environment, its interaction with genetic factors, and the medical care received.
• Emergency status and acute symptomatology more frequent in Black pts, emphasizing the importance of primary care and preventive health care   Multivariable analysis for the entire cohort.For the overall cohort, race was not a predictor of the composite adverse outcome or of operative mortality, but acuity (P <.0001 for both outcomes), preoperative chronic renal insufficiency (P <.0001 and P ¼ .002,respectively), and pulmonary disease (P ¼ .001and P ¼ .0003,respectively) were.Increasing age (P < .0001)and female gender (P ¼ .009)were independently associated with the composite adverse outcome.Follow-up and survival analysis for the entire cohort.
For the entire cohort of Black and White patients (n ¼ 2335), the median follow-up time was 3.7 years (IQR, 1.2-9.7 days).In Figure E1, a scatterplot is shown depicting the completeness of follow-up.In our long-term survival analysis, survival was 53.5% for Black patients versus 70.4% for White patients at 5 years, and 33.7% for Black patients versus 49.6% for White patients at 10 years (P ¼ .0002)(Figure 2).

Data after Propensity Score Matching
We performed propensity score matching on the basis of preoperative risk factors that included community socioeconomic factors (Table 1, preoperative variables matched).Propensity score matching generated 204 pairs of Black and White patients.After this match, the overall rate of the composite adverse outcome was 14.7% (n ¼ 30) for Black patients and 12.3% (n ¼ 25) for White patients (P ¼ .44).All adjusted outcomes are shown in Table 3 (patient outcomes matched).Table 5 also shows all adjusted outcomes for the elective cases only.For the elective cases, the rate of respiratory failure was different between groups even after adjusting for preoperative characteristics and community socioeconomic factors (38.7% for Black patients vs 23.7% for White patients; P ¼ .02).After propensity score matching, survival was 56.2% for Black patients versus 66.6% for White patients at 5 years, and 35.4% for Black patients versus 43.4% for White patients at 10 years (P ¼ .30)(Figure 3).

DISCUSSION
Despite a decrease in the rate of death from heart disease in the United States in recent decades, notable racial disparities have persisted in cardiovascular medicine. 23Different factors have contributed to these disparities.Compared with White patients, Black patients have shown an increased  24 All of these factors, in addition to deeprooted opinions and perceptions, are by nature interconnected, making it difficult to define the role of race in cardiovascular health. 25n this study, we analyzed the association between race (Black or White) and outcomes after surgery of the ascending aorta, aortic root, or aortic arch at our large academic tertiary center.We used the UnitedStatesZipCodes.org to determine community socioeconomic factors that could potentially affect health outcomes such as the percent of the population living below poverty level, household income, and education level.In our overall cohort of patients who underwent surgery of the ascending aorta, aortic root, or aortic arch, we found that cases were emergencies in half of the Black group compared with only one-quarter in the White group.Furthermore, acute symptomatology and the overall presence of symptoms was significantly higher in Black patients than in White patients.In addition, comorbid conditions were observed in more Black patients than in White patients.The 5-and 10-year survival rates were significantly better in White patients than in Black patients.We also observed a nonsignificant trend of more frequent adverse outcomes in Black patients than in White patients.Significant differences were seen in outcomes such as the rate of respiratory failure and the ICU and total LOS overall.
In reports from the International Registry of Acute Aortic Dissection (IRAD), investigators showed that race other than White was associated with operative delay 16 and that Black patients had a mortality rate similar to that of White patients, despite a higher likelihood of having symptoms. 15he similar mortality rates and outcomes observed between Black and White patients in the IRAD report can be explained by the fact that these patients generally received care in aortic centers of excellence (ie, centers participating in IRAD). 15This is in accordance with our study findings showing similar short-term outcomes between Black and White patients but a higher likelihood of acute symptomatology and emergency situations in Black patients than in White patients.
In our unadjusted analysis, we found that Black patients lived in areas characterized by disadvantaged community socioeconomic factors.These areas had a higher percent of the population living below the poverty level, a lower household income, and fewer residents educated at all levels (ie, high school diploma, bachelor's degree, and graduate degree) than did the areas inhabited by White patients.These community socioeconomic factors may in part explain the differences in urgent/emergency presentation, acute symptomatology, respiratory failure, tobacco use, and prolonged hospital stay.In addition, these factors may partially explain the differences in the 5-and 10-year survival rates between Black and White patients.These factors also emphasize the importance of primary and preventive health care.Outreach programs, such as those providing imaging (eg, mobile computed tomography scans and cardiac ultrasound) to patients from distressed communities, are of utmost importance.Distressed socioeconomic factors can lead to reduced preventive health resources and diminished access to care.Furthermore, they can cause work and household instability that increases stress and psychological instability, leading to unhealthy behavior.All of these can increase hospital-related morbidity and mortality.In addition, the absence of insurance, family support, or family stability; living below the poverty level; and reduced household income can directly or indirectly affect prolonged LOS.
When we adjusted for preoperative characteristics and for community socioeconomic factors (ie, percent of the population living below the poverty level, median household income, and percent of patients with different levels of education), similar short-term outcomes and long-term survival rates were seen between the Black and White populations.This implies that community socioeconomic factors may supersede race in the contribution to healthrelated outcomes and survival.Results from the multivariable analysis confirmed these findings.However, this may be a result of the study being underpowered.Our study also showed that unfavorable community socioeconomic factors were associated with adverse outcomes in Black patients but not White patients.In Black patients only, living in an area characterized by a higher percent of the population living below the poverty level, lower household income, or less education independently predicted the composite adverse outcome of operative mortality, renal failure, and respiratory failure.Jeter and colleagues 26 found that living in poverty can lead to a higher mortality rate for the Black population but not for the White population.A possible explanation for this is that White Americans are more likely to draw benefits from living close to areas with better resources than from living in isolated inner-city neighborhoods, which is where poor Black Americans predominantly live.Additionally, economic segregation occurs more in the Black community than in the White community, which may explain why socioeconomic factors influence Black patients more than White patients. 26ecause of the significant differences noted in acuity and presentation between White and Black patients in our study cohort, we performed a subgroup analysis of the patients who underwent elective surgery of the ascending aorta, aortic root, or aortic arch.In the overall elective cohort, we observed a trend that was not consistent with higher complication rates in Black patients than in White patients.When we adjusted for preoperative characteristics and community socioeconomic factors, this trend persisted, with the exception of the respiratory failure rate, which continued to be significantly different between Black and White patients.Despite the fact that we do not have a firm explanation for this finding, we believe that components other than the community socioeconomic factors analyzed in this study may have played a role, such as health behavior and biologic differences between races.

ADULT Limitations
One limitation of our study is its retrospective nature and the potential for inherited bias.In addition, although we adjusted for three major community socioeconomic factors, other factors that may affect health, such as change in employment, insurance status, housing vacancy, and migration status, were missing.We also did not account for some potential factors that uniquely affect Black patients, such as specific heritable cardiovascular biomarkers and responses to some cardiovascular medications that are known to affect the Black population more than the White population.In addition, we used the zip codes of patients to extrapolate data for the 3 major community socioeconomic factors that we accounted for in our analysis.These zip codes represented the social status of an individual's community at a given moment in time and may not reflect the events in the course of a patient's lifetime.Specifically, the community socioeconomic factors used in the analysis were correlated with the patient's zip code from the 2020 Census for the income data and the 2013 Census for the education data.In most cases, these were not the year of surgery.As a result, potential changes in community socioeconomic factors during the study period were not accounted for.However, according to US population estimates from the 2010 US Census that covers our study period, the population within an individual zip code was unlikely to vary more than 10% over a 5-year period.Furthermore, even though intervention year was a variable, the effect of time and era on the community socioeconomic factors was not examined in our study.We also did not examine whether patients had to travel to obtain care and whether this was related to race.Further, our surgical technique has evolved over time and changes in referral patterns may have occurred.Additionally, although the overall cohort included a large number of patients, the number of patients who identified as Black was relatively small, which may have contributed to our study being underpowered.Our study also lacked the power to detect differences in the outcomes of greatest interest (adverse outcomes and mortality).In addition, we excluded patients from other racial and ethnic groups.Finally, we recognize that we compared outcomes between Black and White patients in a high-income country, which does not necessarily reflect the scenario in a middle-or low-income country.
Undoubtedly, questions about racial disparities are better answered using national and regional databases that can provide data for larger sample sizes and avoid the referral bias that exists in the present study.Despite these limitations, our study is among the first to our knowledge in which the association between race and health-related outcomes has been examined in patients with aortic disease undergoing surgery in the ascending aorta, aortic root, or aortic arch, either electively or as an emergency.Furthermore, the issue of health care equity was examined in patients requiring specialized surgery for high-risk disease.By adjusting for important community socioeconomic factors, we attempted to mitigate the known influence of socioeconomic disparities in health-related outcomes.However, we acknowledge that the environmental factors where individuals live, work, and learn are modifiable and can be manipulated to minimize the social determinants of health.Furthermore, it is apparent that, to determine causation and quantify the effect of race on outcomes between Black and White patients with cardiovascular disease, other factors such as specific human genome and biomarkers specific to Black patients are important to consider.

CONCLUSIONS
Emergency status and acute symptomatology at the time of the index operation for the ascending aorta, aortic root, or aortic arch were more prominent in Black patients than in White patients, and acuity was associated with both operative mortality and composite adverse outcome in the overall cohort.In addition, community socioeconomic factors, such as the percent of the population living below the poverty level, median household income, and different levels of education, were more favorable in White patients.The trend toward more frequent adverse short-term outcomes in Black patients and differences in 5-and 10-year survival seen in our unadjusted analysis was not seen in the propensity-score matched cohort adjusted for community socioeconomic factors, possibly because our study was underpowered.Community socioeconomic factors were associated with the overall rates of the composite adverse outcome, mortality, and renal or respiratory failure in Black patients but not in White patients.As we move toward precision medicine and seek to produce better outcomes in patients with aortic disease, more granular data collected at the level of the individual are needed to better serve patients in the future.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they have a conflict of interest.The editors and reviewers of this manuscript have no conflicts of interest.The Journal of Thoracic and Cardiovascular Surgery c January 2024 (database used to study interaction of race and community socioeconomic (CSE) factors) 2118 White and 217 Black pts who underwent primary or redo aortic operations (ascending, root, or arch) Thoracic Aortic Surgery: Myth or reality?• Unfavorable CSE factors predicted adverse outcome in Black but not in White pts • Short-term outcomes and 5-and 10-year survival similar in the propensity matched analysis after adjusting for CSE factors.CSE factors may supersede race in health-related outcomes.

FIGURE 1 .
FIGURE 1. Overview of the study and its key findings.

FIGURE
FIGURE E1.A scatterplot depicting the completeness of follow-up.

TABLE 1 .
Preoperative variables Values are presented as n (%) or median (25%-75% interquartile range).SMD, Standardized mean difference.*c 2 or Wilcoxon 2-sample test.yWeight in kilograms (or pounds) divided by the square of height in meters (or feet).The Journal of Thoracic and Cardiovascular Surgery c January 2024 ADULT Multivariable analysis for the entire cohort stratified according to race.When we stratified patients according to race, the percent of the population residing in areas living below poverty level independently predicted renal failure (P ¼ .01)andrespiratoryfailure (P ¼ .02)forBlackpatients but not White patients.Similarly, living in an area characterized by lower household income independently predicted the composite adverse outcome (P ¼ .01) in Black patients but not White patients.Furthermore, living in an area with a lower percentage of residents with a high school diploma predicted operative mortality (P ¼ .04) in Black patients but not White patients.Results of the multivariable analysis stratified by race are shown in Table4.Subgroup analysis of elective-only cases in the entire cohort.Elective operations on the ascending aorta, aortic root, or aortic arch were performed in 1695 patients; of

TABLE 4 .
Results of multivariable analysis stratified by race Values are presented as n (%) or median (25%-75% interquartile range).MI, Myocardial infarction; ICU, intensive care unit; LOS, length of stay.*c 2 or Wilcoxon 2-sample test.yPaired t test or Wilcoxon signed-rank test.zFisher exact test.The Journal of Thoracic and Cardiovascular Surgery c January 2024