This cohort study examines the association of initial thoracic endovascular aortic repair (TEVAR) with mortality and morbidity among patients undergoing acute uncomplicated type B aortic dissection.
Key Points
Question
Is initial thoracic endovascular aortic repair (TEVAR) for acute uncomplicated type B aortic dissection (uTBAD) associated with reduced mortality and morbidity compared with medical therapy alone?
Findings
In this cohort study of 7105 patients, initial TEVAR compared with medical therapy alone for uTBAD was not associated with improved survival or reduced aorta-related hospitalizations, aortic interventions, or cardiovascular hospitalizations over 1, 2, and 5 years in a propensity score–weighted analysis.
Meaning
The findings suggest that initial TEVAR for uTBAD is not associated with reduced aortic rupture, aneurysmal degeneration, or mortality.
Abstract
Importance
Thoracic endovascular aortic repair (TEVAR) has increasingly been used for uncomplicated type B aortic dissection (uTBAD) despite limited supporting data.
Objective
To assess whether initial TEVAR following uTBAD is associated with reduced mortality or morbidity compared with medical therapy alone.
Design, Setting, and Participants
This cohort study included Centers for Medicare & Medicaid Services inpatient claims data for adults aged 65 years or older with index admissions for acute uTBAD from January 1, 2011, to December 31, 2018, with follow-up available through December 31, 2019.
Exposures
Initial TEVAR was defined as TEVAR within 30 days of admission for acute uTBAD.
Main Outcomes and Measures
Outcomes included all-cause mortality, cardiovascular hospitalizations, aorta-related and repeated aorta-related hospitalizations, and aortic interventions associated with initial TEVAR vs medical therapy. Propensity score inverse probability weighting was used.
Results
Of 7105 patients with eligible index admissions for acute uTBAD, 1140 (16.0%) underwent initial TEVAR (623 [54.6%] female; median age, 74 years [IQR, 68-80 years]) and 5965 (84.0%) did not undergo TEVAR (3344 [56.1%] female; median age, 76 years [IQR, 69-83 years]). Receipt of TEVAR was associated with region (vs South; Midwest: adjusted odds ratio [aOR], 0.66 [95% CI, 0.53-0.81]; P < .001; Northeast: aOR, 0.63 [95% CI, 0.50-0.79]; P < .001), Medicaid dual eligibility (aOR, 0.76; 95% CI, 0.63-0.91; P = .003), hypertension (aOR, 1.26; 95% CI, 1.03-1.54; P = .03), peripheral vascular disease (aOR, 1.24; 95% CI, 1.02-1.49; P = .03), and year of admission (2012, 2013, 2014, and 2015 were associated with greater odds of TEVAR compared with 2011). After inverse probability weighting, mortality was similar for the 2 strategies up to 5 years (hazard ratio [HR], 0.95; 95% CI, 0.85-1.06), as were aorta-related hospitalizations (HR, 1.12; 95% CI, 0.99-1.27), aortic interventions (HR, 1.01; 95% CI, 0.84-1.20), and cardiovascular hospitalizations (HR, 1.05; 95% CI, 0.93-1.20). In a sensitivity analysis that included deaths within the first 30 days, initial TEVAR was associated with lower mortality over a period of 1 year (adjusted HR [aHR], 0.86; 95% CI, 0.75-0.99; P = .03), 2 years (aHR, 0.85; 95% CI, 0.75-0.96; P = .008), and 5 years (aHR, 0.87; 95% CI, 0.80-0.96; P = .004).
Conclusions and Relevance
In this study, 16.0% of patients underwent initial TEVAR within 30 days of uTBAD, and receipt of initial TEVAR was associated with hypertension, peripheral vascular disease, region, Medicaid dual eligibility, and year of admission. Initial TEVAR was not associated with improved mortality or reduced hospitalizations or aortic interventions over a period of 5 years, but in a sensitivity analysis that included deaths within the first 30 days, initial TEVAR was associated with lower mortality. These findings, along with cost-effectiveness and quality of life, should be assessed in a prospective trial in the US population.
Introduction
Type B aortic dissection (TBAD) affects approximately 3 people per 100 000,1 of whom more than 60% present without signs of rupture or malperfusion (termed uncomplicated TBAD [uTBAD]).2,3,4,5 Thoracic endovascular aortic repair (TEVAR) is standard of care for patients with complicated TBAD, but its role in treating uTBAD is uncertain because optimal medical therapy is recommended but prophylactic TEVAR to prevent late aortic events and mortality can be considered.3 This management question, highlighted in available TBAD management guidelines,3,6 previously motivated 2 randomized clinical trials: Investigation of Stent Grafts in Patients With Type B Aortic Dissection (INSTEAD)7,8 and A European Study on Medical Management Vs TAG Device + Medical Management for Acute Uncomplicated Type B Dissection (ADSORB).9 However, these trials were small, underpowered for clinical outcomes, and conducted in a homogenous European population. A number of observational studies have attempted to fill this evidence gap,4,5,10,11,12,13,14,15,16,17 but they have either analyzed data from procedures performed before the approval by the US Food & Drug Administration (FDA) of the first dissection-specific thoracic stent graft in 2013 (Gore TAG Conformable Thoracic Endoprosthesis) or included heterogeneous or limited patient cohorts. Therefore, even modern practice patterns concerning the use of TEVAR for uTBAD and associated rates of aortic and cardiovascular events remain unclear, and a recent survey study demonstrated near-complete equipoise about the role of TEVAR for uTBAD.18 We aimed to describe patterns of TEVAR use within 30 days of hospital admission for acute uTBAD, characterize subsequent clinical events, and compare outcomes between patients treated with a strategy of initial TEVAR vs initial medical therapy.
Methods
Data Source and Cohort Identification
This cohort study used 100% Centers for Medicare & Medicaid Services (CMS) Limited Data Set inpatient claims data and corresponding Master Beneficiary Summary File data between January 1, 2011, and December 31, 2019. The study cohort consisted of adult patients aged 65 years or older with a nonelective index admission for acute uTBAD between January 1, 2011, and December 31, 2018. This study was determined to be exempt by the Duke University institutional review board, and no patient consent was required because the analysis used deidentified patient data. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The presence of TBAD was identified using the occurrence of International Classification of Diseases, Ninth Revision (ICD-9) and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) diagnosis codes in the primary position (ICD-9, Clinical Modification codes 441.01 and 441.03 or International Statistical Classification of Diseases, Tenth Revision, Clinical Modification I71.01 and I71.03) before applying a series of exclusion criteria (eFigure in the Supplement). To ensure a cohort of patients with acute uTBAD and because billing codes used in CMS analyses do not allow for discrimination between conditions present at admission and those that develop during hospitalization, patients with certain conditions (aortic rupture, arterial malperfusion, stroke, and paraplegia or paraparesis) during their index admission were excluded. To allow for ascertainment of the exposure (TEVAR within 30 days), patients who died within 30 days of the index admission or had fewer than 30 days of fee-for-service Medicare coverage were excluded from the primary analyses. Patients with nonacute presentations of TBAD, hospitalizations for aneurysmal degeneration of previous dissections, open thoracic aortic dissection repairs, and type A aortic dissection were excluded according to previous literature3,19 (exclusion criteria codes are given in eTable 1 in the Supplement). Patients needed to have had fee-for-service Medicare coverage for at least 1 year before their index date to allow for comorbidity ascertainment.
Outcomes
Our primary outcome was all-cause mortality, which was derived from the death date in the Master Beneficiary Summary File. Key secondary outcomes derived from Medicare inpatient claims included aorta-related hospitalizations and repeated aorta-related hospitalizations, aortic interventions, and cardiovascular hospitalizations. Aorta-related hospitalizations were defined to include hospitalizations with primary diagnosis codes for aortic rupture, malperfusion syndromes, hypertensive urgency or emergency, stroke, paraplegia or paraparesis, extension or degeneration of dissection, or retrograde aortic dissection (outcome definition codes are given in eTable 2 in the Supplement). Aortic interventions included procedures associated with aforementioned dissection extension or degeneration or retrograde dissection. Procedure codes in any position (not only primary) were considered in ascertaining outcomes. Cardiovascular hospitalizations included hospitalizations with diagnosis codes in the primary position for conditions that might reasonably be associated with dissection, untreated hypertension, or both (eTable 2 in the Supplement). When available, the cardiovascular hospitalization codes were drawn from the American Heart Association’s Heart Disease and Stroke Statistics–2021 Update.20 Other secondary outcomes included components of aorta-related and cardiovascular hospitalizations, repeated aorta-related and cardiovascular hospitalizations, and all-cause hospitalizations.
Covariates and Exposure
Covariates of interest included age, sex, race (Black, White, or other [Asian, Hispanic, and North American Native individuals and those with other race] or unknown), geographical region, and Medicaid dual eligibility from the Master Beneficiary Summary File; comorbidities derived from inpatient claims based on validated algorithms21,22; and the number of TEVARs performed at the index admission institution in the year before the index admission. Race was obtained from the Master Beneficiary Summary File and was included as a covariate because it has previously been shown to be associated with receipt of TEVAR and TEVAR outcomes after TBAD.23 Frailty was also included as a covariate using the Claims-Based Frailty Indicator derived from inpatient claims.24,25 For the primary analyses, frailty was considered as a binary variable, with a cutoff of 0.25 or more considered to indicate frailty; in a sensitivity analysis, frailty was considered as a continuous variable.26 The exposure of interest was a procedure code in any position on an inpatient claim indicating TEVAR within 30 days of the index hospitalization for acute uTBAD (ICD-9, Procedure Coding System codes 39.73 and 39.79; ICD-10-Procedure Coding System codes 02UW3JZ, 02QW3ZZ, 02VW3DZ, 02VW0DZ, and 02VW3ZZ).
Statistical Analysis
Baseline characteristics for the cohort were described using median (IQR) for continuous variables and frequencies with percentages for categorical variables, stratified by receipt of TEVAR. Group differences were tested using Wilcoxon rank sum or χ2 tests as well as standardized mean differences. For the primary analyses, outcomes were landmarked at 30 days to allow for exposure determination and outcome calculations without survival time bias. Observed mortality rates were calculated using Kaplan-Meier estimates and compared between groups using log-rank tests. Observed rates of the secondary outcomes were calculated using the cumulative incidence function censoring for death, end of available follow-up, and loss of fee-for-service Medicare coverage and were compared using the Gray test.27,28 Repeated hospitalizations were the cumulative incidence of a second hospitalization after the index.
A multivariable logistic propensity model was used to obtain each patient’s probability of receiving TEVAR conditional on observed covariates and to assess associations between baseline factors and receipt of TEVAR. We incorporated generalized estimating equations to account for clustering of patients at the index facility. We tested for multicollinearity using the variance inflation factor of covariates and removed variables when this value exceeded 3. In the propensity model, receipt of TEVAR was the dependent variable and age, sex, race, region, Medicaid dual eligibility, comorbidities, year of diagnosis, site volume, and frailty were independent variables. Patient data were weighted using inverse probability weighting derived from propensity scores. Visual inspection of the distribution of propensity scores and assessment of standardized mean differences were used to ensure valid and adequate weighting. In addition, 2 falsification end points (hip fractures and urinary tract infections) were chosen to reveal any residual or unmeasured confounding between the groups. Estimated weights were incorporated into Cox proportional hazards regression models for the primary, secondary, and falsification end points in which the only independent variable was TEVAR status. Repeated aorta-related hospitalization was modeled using the Andersen-Gill method with robust SEs with direct covariate adjustment.29,30
Two sensitivity analyses were performed. In the first analysis, frailty was included as a continuous rather than binary variable in the logistic regression model with generalized estimating equations used to derive propensity scores, and Cox proportional hazards regression models were rerun. In the second analysis, TEVAR status was used as a time-dependent variable in a series of covariate-adjusted Cox proportional hazards regression models for all-cause mortality and the 2 falsification end points using variables from the propensity model as covariates. This analysis allowed inclusion of patients who died within the first 30 days and was performed to account for possible immortal time bias. Similar time-varying sensitivity analyses for hospitalizations and aortic interventions were not possible because patients could experience hospitalization or aortic intervention before or at the same time as the exposure. All aforementioned Cox proportional hazards regression models accounted for clustering of patients at the index facility with robust SEs. All statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc). Two-sided P < .05 was considered significant.
Results
Baseline Cohort Characteristics
There were 7883 patients with index admissions for acute uTBAD between 2011 and 2018, of whom 778 died or lost fee-for-service Medicare coverage within 30 days and were excluded from the primary analyses. Of the remaining 7105 patients, 1140 (16.0%) underwent TEVAR within 30 days of the index date (623 [54.6%] female; 517 [45.4%] male) and 5965 (84.0%) did not undergo initial TEVAR (3344 [56.1%] female; 2621 [43.9] male). Patients who underwent initial TEVAR were slightly younger (median age, 74 years [IQR, 68-80 years] vs 76 years [IQR, 69-83 years]; P < .001) and less likely to have congestive heart failure (112 [9.8%] vs 758 [12.7%]; P = .007), valvular heart disease (78 [6.8%] vs 561 [9.4%]; P = .006), or frailty (72 [6.3%] vs 530 [8.9%]; P < .001) compared with patients who did not undergo initial TEVAR (Table 1). Patients who underwent initial TEVAR were from different regions, with 523 (45.9%) from the South undergoing TEVAR (vs 2239 [37.5%] without initial TEVAR), but only 222 (19.5% vs 1445 [24.2%]) and 180 (15.8% vs 1242 [20.8%]) from the Midwest and Northeast, respectively, underwent TEVAR (overall P < .001). The proportion of patients with uTBAD undergoing initial TEVAR ranged from 114 (11.8%) to 163 (19.4%) per year (Figure). In addition, patients who underwent initial TEVAR were admitted to institutions with higher yearly TEVAR volumes than were patients who did not undergo initial TEVAR (median TEVARs per year, 19 [IQR, 6-43] vs 12 [IQR, 4-30]; P < .001). TEVARs were performed at 2308 institutions over the study period, with the number of TEVARs per institution ranging from fewer than 11 to 1067, with a median of 18 (IQR, <11 to 52); 907 institutions (39.3%) performed fewer than 11 TEVARs. Patients with index admissions in 2011 or 2012, before FDA approval of the first TBAD-specific device, were similar to those with index admissions in 2013 or after (eTable 3 in the Supplement).
Table 1. Unweighted Baseline Characteristics by Initial Treatment Strategy.
| Characteristic | Patientsa | SMD, % | P value | |
|---|---|---|---|---|
| With TEVAR (n = 1140) | Without TEVAR (n = 5965) | |||
| Age, median (IQR), y | 74.0 (68.0-80.0) | 76.0 (69.0-83.0) | 10.8 | <.001 |
| Sex | ||||
| Female | 623 (54.6) | 3344 (56.1) | 2.8 | .38 |
| Male | 517 (45.4) | 2621 (43.9) | ||
| Race | ||||
| Black | 169 (14.8) | 963 (16.1) | 5.0 | .30 |
| White | 888 (77.9) | 4625 (77.5) | ||
| Other or unknownb | 83 (7.3) | 377 (6.3) | ||
| Medicaid dual eligibility | 190 (16.7) | 1138 (19.1) | 6.3 | .06 |
| End-stage renal disease | 50 (4.4) | 317 (5.3) | 4.3 | .19 |
| US census region | ||||
| South | 523 (45.9) | 2239 (37.5) | 20.6 | <.001 |
| Midwest | 222 (19.5) | 1445 (24.2) | ||
| Northeast | 180 (15.8) | 1242 (20.8) | ||
| West | 215 (18.9) | 1039 (17.4) | ||
| Comorbidities | ||||
| Hypertension | 411 (36.1) | 2124 (35.6) | 0.9 | .77 |
| Diabetes | 82 (7.2) | 531 (8.9) | 6.3 | .06 |
| Ischemic heart disease | 181 (15.9) | 1093 (18.3) | 6.5 | .049 |
| Congestive heart failure | 112 (9.8) | 758 (12.7) | 9.1 | .007 |
| Cerebrovascular disease | 52 (4.6) | 324 (5.4) | 4.0 | .23 |
| COPD | 172 (15.1) | 959 (16.1) | 2.7 | .40 |
| Cardiac arrhythmias | 189 (16.6) | 1091 (18.3) | 4.5 | .17 |
| Peripheral vascular disease | 238 (20.9) | 1116 (18.7) | 5.4 | .09 |
| Kidney disease | 106 (9.3) | 649 (10.9) | 5.3 | .11 |
| Valvular heart disease | 78 (6.8) | 561 (9.4) | 9.4 | .006 |
| Frailtyc | 72 (6.3) | 530 (8.9) | 13.3 | <.001 |
| Year of index admission | ||||
| 2011 | 123 (10.8) | 754 (12.6) | 22.8 | <.001 |
| 2012 | 155 (13.6) | 740 (12.4) | ||
| 2013 | 160 (14.0) | 680 (11.4) | ||
| 2014 | 157 (13.8) | 640 (10.7) | ||
| 2015 | 163 (14.3) | 679 (11.4) | ||
| 2016 | 147 (12.9) | 749 (12.6) | ||
| 2017 | 121 (10.6) | 869 (14.6) | ||
| 2018 | 114 (10.0) | 854 (14.3) | ||
| Site TEVAR volume in year before index year, median (IQR) | 19.0 (6.0-43.0) | 12.0 (4.0-30.0) | 22.8 | <.001 |
Abbreviations: COPD, chronic obstructive pulmonary disease; NA, not applicable; SMD, standardized mean difference; TEVAR, thoracic endovascular aortic repair.
Data are reported as number (percentage) of patients unless otherwise indicated.
Other includes Asian, Hispanic, and North American Native individuals and those with other race.
Frailty is a dichotomous variable derived by applying 0.25 or greater as the cutoff using the Segal et al24 Claims-Based Frailty Index with predicted probability from 0 to 1.
Figure. Hospital Admissions for Acute Uncomplicated Type B Aortic Dissection (uTBAD), Percentage of Patients Undergoing Thoracic Endovascular Aortic Repair (TEVAR), and Associated Outcomes.
Observed cumulative incidence of primary outcomes was unweighted. CV indicates cardiovascular.
Factors Associated With Undergoing Initial TEVAR
Multivariable analysis showed that receipt of initial TEVAR was inversely associated with age (each additional year older: adjusted odds ratio [aOR], 0.99; 95% CI, 0.98-0.99; P < .001) and region of admission (vs South; Midwest: aOR, 0.66 [95% CI, 0.53-0.81]; P < .001; Northeast: aOR, 0.63 [95% CI, 0.50-0.79]; P < .001) (Table 2). Medicaid dual eligibility was associated with lower odds of initial TEVAR (aOR, 0.76; 95% CI, 0.63-0.91; P = .003). Certain comorbidities, including hypertension (aOR, 1.26; 95% CI, 1.03-1.54; P = .03) and peripheral vascular disease (aOR, 1.24; 95% CI, 1.02-1.49; P = .03), were also associated with initial TEVAR. Compared with year of admission in 2011, admission in 2012, 2013, 2014, or 2015 was associated with greater odds of TEVAR. TEVAR volume at the admission site in the year before the index date was also associated with a small but significant increase in initial TEVAR (per 1 procedure; aOR, 1.01; 95% CI, 1.01-1.01; P < .001). Frailty was not associated with initial TEVAR when considered as a binary variable.
Table 2. Factors Associated With Receiving Initial Thoracic Endovascular Aortic Repair.
| Baseline characteristic | Adjusted OR (95% CI) | P value |
|---|---|---|
| Age, y | 0.99 (0.98-0.99) | <.001 |
| Sex | ||
| Female | 1.02 (0.90-1.16) | .75 |
| Male | 1.00 [Reference] | NA |
| Race | ||
| Black | 0.86 (0.71-1.04) | .12 |
| White | 1.00 [Reference] | NA |
| Other or unknowna | 1.24 (0.94-1.63) | .13 |
| US census region | ||
| South | 1.00 [Reference] | NA |
| Midwest | 0.66 (0.53-0.81) | <.001 |
| Northeast | 0.63 (0.50-0.79) | <.001 |
| West | 0.91 (0.73-1.14) | .41 |
| Medicaid dual eligibility | 0.76 (0.63-0.91) | .003 |
| Comorbidities | ||
| End-stage renal disease | 0.87 (0.60-1.25) | .44 |
| Hypertension | 1.26 (1.03-1.54) | .03 |
| Diabetes | 0.85 (0.65-1.10) | .21 |
| Ischemic heart disease | 0.84 (0.69-1.03) | .09 |
| Congestive heart failure | 0.87 (0.68-1.12) | .29 |
| Cerebrovascular disease | 0.89 (0.65-1.22) | .47 |
| COPD | 0.98 (0.80-1.21) | .87 |
| Cardiac arrhythmias | 0.94 (0.76-1.16) | .55 |
| Peripheral vascular disease | 1.24 (1.02-1.49) | .03 |
| Kidney disease | 0.94 (0.71-1.23) | .63 |
| Valvular heart disease | 0.73 (0.56-0.95) | .02 |
| Frailtyb | 0.83 (0.63-1.11) | .21 |
| Year of index admission | ||
| 2011 | 1.00 [Reference] | NA |
| 2012 | 1.30 (1.02-1.66) | .04 |
| 2013 | 1.40 (1.07-1.84) | .01 |
| 2014 | 1.40 (1.07-1.83) | .02 |
| 2015 | 1.33 (1.03-1.72) | .03 |
| 2016 | 1.16 (0.88-1.52) | .29 |
| 2017 | 0.96 (0.72-1.27) | .76 |
| 2018 | 0.95 (0.71-1.27) | .73 |
| Site TEVAR volume per 1 procedure in year before index year | 1.01 (1.01-1.01) | <.001 |
Abbreviations: COPD, chronic obstructive pulmonary disease; NA, not applicable; OR, odds ratio; TEVAR, thoracic endovascular aortic repair.
Other includes Asian, Hispanic, and North American Native individuals and those with other race.
Frailty is a dichotomous variable derived by applying 0.25 or greater as the cutoff using the Segal et al24 Claims-Based Frailty Index with predicted probability from 0 to 1.
In the sensitivity analysis considering frailty as a continuous variable, many of the same factors were associated with receipt of initial TEVAR; age was not associated with initial TEVAR, but the aOR values for the remaining associated factors changed only minimally (eTable 4 in the Supplement). When frailty was considered as a continuous variable, it was significantly associated with reduced odds of initial TEVAR (aOR, 0.09; 95% CI, 0.03-0.31; P < .001).
Outcomes at 1, 2, and 5 Years
Median follow-up for the primary landmarked analysis was 2.55 years (IQR, 1.22-4.45 years). There were significant unweighted differences in observed 5-year rates of mortality (404 [44.5%] with initial TEVAR vs 2199 [49.0%] without initial TEVAR; P = .02) and aorta-related hospitalizations (333 [34.7%] with initial TEVAR vs 1504 [30.7%] without initial TEVAR; P = .04) (Table 3 and Figure). There were significant differences in some unweighted secondary outcomes as well (eTable 5 in the Supplement). The observed stroke rate was higher in the initial TEVAR group than in the group that did not undergo initial TEVAR at each time point (1 year: 24 [2.2%] vs 73 [1.2%]; P = .02; 2 years: 41 [3.9%] vs 125 [2.3%]; P = .002; and 5 years: 64 [7.0%] vs 196 [4.4%]; P < .001). The proportion of all-cause hospitalizations was higher in the initial TEVAR group at each time point (1 year: 524 [46.8%] vs 2474 [42.1%]; P = .002; 2 years: 640 [58.6%] vs 3122 [54.8%]; P = .006; and 5 years: 773 [76.6%] vs 3755 [73.9%]; P = .006). The rates of hospitalization for abdominal aortic aneurysm were lower in the initial TEVAR group at each time point (1 year: 63 [5.6%] vs 435 [7.4%]; P = .03; 2 years: 76 [6.9%] vs 530 [9.3%]; P = .01; and 5 years: 98 [9.7%] vs 622 [11.9%]; P = .03). There was no difference in need for aortic intervention by 5 years (167 [16.8%] with initial TEVAR vs 776 [15.1%] without initial TEVAR; P = .29)
Table 3. Observed Cumulative Incidence of Primary Outcomes.
| Outcome | Patients, No. (%) | P value | |
|---|---|---|---|
| With initial TEVAR (n = 1140) | Without initial TEVAR (n = 5965) | ||
| All-cause mortality | |||
| 1 y | 175 (15.7) | 967 (16.5) | .56 |
| 2 y | 245 (22.7) | 1413 (25.3) | .12 |
| 5 y | 404 (44.5) | 2199 (49.0) | .02 |
| Aorta-related hospitalizations | |||
| 1 y | 191 (17.1) | 907 (15.4) | .17 |
| 2 y | 237 (21.8) | 1173 (20.7) | .38 |
| 5 | 333 (34.7) | 1504 (30.7) | .04 |
| Repeated aorta-related hospitalizations | |||
| 1 y | 32 (2.9) | 146 (2.5) | .46 |
| 2 y | 57 (5.4) | 252 (4.6) | .27 |
| 5 y | 99 (11.1) | 425 (9.8) | .25 |
| Aortic interventions | |||
| 1 y | 106 (9.5) | 536 (9.1) | .72 |
| 2 y | 126 (11.5) | 646 (11.3) | .84 |
| 5 y | 167 (16.8) | 776 (15.1) | .29 |
| Cardiovascular hospitalizations | |||
| 1 y | 156 (14.0) | 837 (14.2) | .78 |
| 2 y | 212 (19.6) | 1101 (19.5) | .99 |
| 5 y | 296 (31.3) | 1418 (29.5) | .47 |
Abbreviation: TEVAR, thoracic endovascular aortic repair.
In unweighted analyses, receipt of initial TEVAR was associated with a lower hazard of mortality over a period of 5 years (hazard ratio [HR], 0.88; 95% CI, 0.79-0.97; P = .01) (Table 4). In inverse probability-weighted Cox proportional hazards regression model analysis (or adjusted in the case of repeated aorta-related hospitalizations), receipt of initial TEVAR was not associated with any key primary or secondary outcome (Table 4). After inverse probability weighting, mortality was similar for the 2 strategies up to 5 years (HR, 0.95; 95% CI, 0.85-1.06), as were aorta-related hospitalizations (HR, 1.12; 95% CI, 0.99-1.27), aortic interventions (HR, 1.01; 95% CI, 0.84-1.20), and cardiovascular hospitalizations (HR, 1.05; 95% CI, 0.93-1.20). This lack of association between initial TEVAR and outcomes remained in the sensitivity analysis considering frailty as a continuous variable (eTable 6 in the Supplement).
Table 4. Unweighted and Weighted Associations Between Receipt of Initial Thoracic Endovascular Aortic Repair and Outcomes.
| Outcome | Unweighted HR (95% CI) | P value | Weighted HR (95% CI)a | P value |
|---|---|---|---|---|
| Primary outcomes | ||||
| Mortality | ||||
| 1 y | 0.95 (0.82-1.11) | .55 | 1.05 (0.89-1.24) | .58 |
| 2 y | 0.90 (0.78-1.03) | .12 | 0.99 (0.86-1.14) | .87 |
| 5 y | 0.88 (0.79-0.97) | .01 | 0.95 (0.85-1.06) | .36 |
| Aorta-related hospitalizations | ||||
| 1 y | 1.11 (0.95-1.31) | .19 | 1.15 (0.97-1.37) | .11 |
| 2 y | 1.06 (0.92-1.22) | .45 | 1.08 (0.93-1.26) | .32 |
| 5 y | 1.11 (0.98-1.24) | .09 | 1.12 (0.99-1.27) | .08 |
| Repeated aorta-related hospitalizations | ||||
| 1 y | 1.09 (0.94-1.26) | .26 | 1.10 (0.95-1.28)b | .21 |
| 2 y | 1.02 (0.90-1.16) | .74 | 1.05 (0.93-1.20)b | .43 |
| 5 y | 1.05 (0.94-1.16) | .38 | 1.10 (0.99-1.22)b | .08 |
| Aortic interventions | ||||
| 1 y | 1.04 (0.84-1.28) | .72 | 0.96 (0.77-1.20) | .71 |
| 2 y | 1.02 (0.84-1.23) | .87 | 0.96 (0.78-1.18) | .71 |
| 5 y | 1.08 (0.91-1.27) | .38 | 1.01 (0.84-1.20) | .94 |
| Cardiovascular hospitalizations | ||||
| 1 y | 0.98 (0.83-1.16) | .79 | 0.98 (0.82-1.18) | .87 |
| 2 y | 1.00 (0.86-1.15) | .96 | 1.01 (0.86-1.18) | .91 |
| 5 y | 1.02 (0.91-1.15) | .68 | 1.05 (0.93-1.20) | .43 |
| Falsification outcomes | ||||
| Hip fracture | ||||
| 1 y | 1.95 (1.11-3.43) | .02 | 1.86 (1.04-3.32) | .04 |
| 2 y | 1.50 (0.89-2.53) | .13 | 1.43 (0.84-2.43) | .19 |
| 5 y | 1.42 (0.89-2.25) | .14 | 1.47 (0.92-2.36) | .11 |
| Urinary tract infection | ||||
| 1 y | 1.09 (0.69-1.73) | .70 | 1.05 (0.64-1.71) | .86 |
| 2 y | 1.24 (0.88-1.73) | .21 | 1.23 (0.87-1.75) | .25 |
| 5 y | 1.22 (0.94-1.58) | .14 | 1.23 (0.93-1.62) | .15 |
Abbreviation: HR, hazard ratio.
Inverse probability-weighted Cox proportional hazards regression model analysis was used.
The Andersen-Gill method and direct covariate model adjustment were used instead of inverse probability weighting.
Time-Varying Sensitivity Analysis
There were 752 deaths (9.54%) in the overall sample within 30 days of index admission. When receipt of TEVAR was considered as a time-varying exposure allowing for inclusion of deaths within 30 days or less in a multivariable analysis, initial TEVAR was associated with lower mortality over a period of 1 year (adjusted HR [aHR], 0.86; 95% CI, 0.75-0.99; P = .03), 2 years (aHR, 0.85; 95% CI, 0.75-0.96; P = .008), and 5 years (aHR, 0.87; 95% CI, 0.80-0.96; P = .004). Patients who received initial TEVAR had a higher adjusted hazard of the hip fracture falsification end point over a period of 5 years in the time-varying analysis (5-year aHR, 1.63; 95% CI, 1.00-2.65; P = .049). There was a higher adjusted hazard of urinary tract infection over a period of 5 years in the initial TEVAR group (aHR, 1.38; 95% CI, 1.06-1.79; P = .02).
Discussion
This analysis provided several novel insights into the demographics of patients with uTBAD as well as uTBAD treatment strategies and their subsequent outcomes in a large US sample. First, only 16.0% of patients underwent initial TEVAR within 30 days of uTBAD, and receipt of initial TEVAR appeared to be associated with regional and national practice patterns and patient fitness (as measured by comorbidities and frailty) for intervention. Second, we found that morbidity and mortality among patients with uTBAD were significantly associated with aortic pathology (1837 [31.4%] overall had an aorta-related hospitalization, and 943 [15.4%] an aortic intervention by 5 years) and cardiovascular health (2603 [48.2%] died by 5 years, and 1714 [29.8%] had a cardiovascular hospitalization). In addition, 524 (10.0%) of all patients with uTBAD had recurrent aorta-related hospitalizations, and 596 (11.3%) had recurrent cardiovascular hospitalizations over 5 years. Third, initial TEVAR was not associated with improved mortality at 1, 2, and 5 years, although results of the sensitivity analysis including early deaths suggest the possibility of an early mortality benefit that cannot be totally excluded in an observational analysis.
Previously available data supporting the prevalence, general population outcomes, and treatment strategies for uTBAD are sparse. The 2 available randomized clinical trials (INSTEAD7,8 and ADSORB9) were small, with somewhat limited follow-up and end points, and featured highly selected patient cohorts. For instance, the INSTEAD trial randomized only 140 patients with a mix of acute and nonacute uTBAD.7 This sample size was generated from power calculations based on unpublished outcome data that yielded 80% power to detect an 18% absolute difference in mortality rates at 1 year.31 The ADSORB trial randomized only 61 patients and took as its primary end point freedom from a composite of predominantly radiographic findings (incomplete or no false lumen thrombosis, aortic dilation of ≥5 mm, and aortic rupture).9 The INSTEAD and ADSORB trials also enrolled predominantly male, European cohorts, which raises questions about the broader applicability of their findings. In addition, a number of previous studies reported TBAD analyses from single and multiple institutions,4,5,12,14,15 registries (ie, the International Registry of Acute Aortic Dissection and the Vascular Quality Initiative),11,13 or regional and national data sets,16,17,32 but these were limited by size, years of data analyzed, available follow-up, or inclusion of complicated TBAD and uTBAD together. Although some of these studies attempted to assess the association of TEVAR with event rates, none were able to clearly delineate outcomes in a large, nationally representative cohort. This lack of data has led to confusion about the natural history of uTBAD with and without initial TEVAR, which has in turn hampered evidence-based discussions about the indications for initial TEVAR. This confusion is reflected in weak recommendations to consider initial TEVAR for uTBAD among available guidelines statements.3,6
Mortality rates following uTBAD have previously been reported to approach 30% at 3 to 5 years,5,8,12,13,15,33,34 but we found 5-year mortality of 44.5%, which may, in part, reflect the older age of the cohort as well as complete outcome ascertainment. We also found hospitalization rates that were higher than previously reported primarily in single-institution or institution-linked registries, highlighting the limitations of institution-centered analyses.5,13 Moreover, the rate of aortic aneurysmal degeneration in the absence of TEVAR has been reported to be as high as 30% to 50%, but we found that only 15.1% in the group without initial TEVAR required aortic intervention.4,9,12,35,36 Although this finding may underestimate the rate of radiographic aneurysmal degeneration, we were able to capture the manifestations of clinically significant aneurysmal degeneration whether appropriately treated (aortic interventions) or inappropriately treated (hospitalizations and mortality).
The present analysis allowed for evaluation of the arguments for and against initial TEVAR. Some researchers have reported that initial (prophylactic) TEVAR could stabilize the aorta and prevent aortic rupture or aneurysmal degeneration.14,15,36,37 Others have reported that TEVAR exposes patients to the risk of retrograde dissections, strokes or paraplegia, and the need for endoleak-related reinterventions possibly unnecessarily.13,14,15,16,38 Our analysis suggests that there is some evidence in support of both sets of findings. Patients without initial TEVAR were more frequently admitted for abdominal aortic aneurysm at 1, 2, and 5 years, while patients with initial TEVAR more commonly experienced strokes. We did not observe a difference in aortic interventions, but this may be because the 2 groups required different interventions that offset each other: reinterventions for TEVAR complications vs interventions for aneurysmal degeneration. Fattori et al13 and Xiang et al15 similarly did not find differences in the rate of aortic interventions between initial TEVAR and medical therapy groups for up to 5 years, suggesting that avoiding later intervention may not be the strongest indication for initial TEVAR. Of note, some of the most severe sequelae of either initial TEVAR (paraplegia or retrograde type A dissection) or conservative management (aortic rupture) were rare and did not differ between groups in our analysis. In contrast, the mortality benefit previously reported to be associated with TEVAR14,15 was not shown in our analysis, although when early deaths were included in a time-varying sensitivity analysis, the possibility of an early mortality benefit could not be excluded. This finding may reflect observations made in previous studies about early, precipitous conversion from uTBAD to complicated TBAD.12
Altogether, the high event rates and lack of significant weighted differences between groups in this analysis suggest the need to weigh not only clinical events but also patient preferences in making uTBAD treatment decisions. It appeared that patient-specific factors, most notably frailty, played a role in decision-making but that nonpatient health status-related factors, including institutional practice patterns, Medicaid dual eligibility, region, and year of admission, were also important. Other authors have similarly documented patient characteristics associated with receipt of initial TEVAR,10,33 but this analysis is, to our knowledge, the first to show that institutional volume was associated with receipt of TEVAR. There was a lower likelihood of initial TEVAR receipt among dual Medicaid-eligible patients, and previous research suggests that aortic dissection is more common among people with Medicaid.39 It is unlikely (although possible) that anatomic or physiologic eligibility varied by institution, insurance status, region, or year, suggesting that variability in the likelihood of initial TEVAR receipt may be due to individual biases introduced by clinician training, personal experiences, or external incentives that are filling the gap left by a lack of high-quality evidence. Better data regarding appropriate use of TEVAR for uTBAD are needed because a recent survey of vascular and aortic surgeons in Australia, New Zealand, and Europe indicated that half would recommend initial TEVAR for uTBAD while the other half would not.18 In addition, in both our primary analyses and our time-varying sensitivity analyses, the falsification end points of hip fracture and urinary tract infection were more likely to occur in the TEVAR group, which suggests that remaining, unmeasured confounding may have contributed to higher mortality in the initial TEVAR group and that mortality in the TEVAR group was estimated as being higher than in reality due to unmeasured confounding. This unmeasured confounding may have caused estimated differences in event rates to be smaller between the 2 groups than in reality. Given the apparent adequacy of our propensity score weighting (with minimal standardized mean differences) based on elements available in claims data, it is likely that comprehensive, carefully collected prospective data will be needed to fully characterize any treatment effect of initial TEVAR.
Limitations
There are some limitations to our analysis. Retrospective analyses are fundamentally limited by difficulties in understanding decision-making, especially because CMS claims do not offer any anatomical data. Similarly, retrospective claims-based analyses cannot provide information on the adequacy of medical therapy, which may affect outcomes for all patients with uTBAD regardless of TEVAR status; however, we found that there were no differences in admissions for inadequate blood pressure or pain control between the groups. The possibility cannot be excluded that some patients received TEVAR only after experiencing anti-impulse therapy failure, which might introduce unaccounted bias resulting in higher estimates of post-TEVAR mortality compared with reality. Furthermore, in attempting to analyze a cohort with uTBAD, it is likely that we excluded some patients who had TEVAR-related complications rather than TBAD-related complications; therefore, our analysis did not attempt to characterize TEVAR-related complications. In addition, the generalizability of our analytical cohort is limited by using patients with Medicare fee-for-service coverage, which may not be generalizable to the entire population of patients with uTBAD. Therefore, it may be that factors associated with receipt of initial TEVAR differ among younger patients and that the risks and benefits of initial TEVAR differ as well.
Conclusions
Our findings demonstrated that 16.0% of patients with uTBAD underwent initial TEVAR, which was not associated with reduced mortality at 1, 2, and 5 years but was associated with increased odds of stroke and all-cause hospitalizations. The associations and the reasons why some patients received initial TEVAR while others did not remain incompletely characterized. Our findings and limitations of the study suggest that a high-quality, adequately powered randomized clinical trial is needed to conclusively define optimal treatment of uTBAD, especially taking into consideration anatomic features, periprocedural complications, adequacy of medical therapy, patient preferences, and cost-effectiveness of initial TEVAR compared with initial medical therapy strategies.
eFigure. Cohort Construction Through Application of Exclusion Criteria
eTable 1. Exclusion Criteria Codes
eTable 2. Outcome Definitions
eTable 3. Unweighted Baseline Characteristics by Year of Treatment Relative to 2013
eTable 4. Sensitivity Analysis of Association Between Baseline Characteristics and Upfront TEVAR With Frailty as a Continuous Variable
eTable 5. Observed Cumulative Incidences of All Secondary Outcomes
eTable 6. Sensitivity Analysis of Association Between Upfront TEVAR and Outcomes With Frailty as a Continuous Variable in Weighted Analysis
References
- 1.McClure RS, Brogly SB, Lajkosz K, Payne D, Hall SF, Johnson AP. Epidemiology and management of thoracic aortic dissections and thoracic aortic aneurysms in Ontario, Canada: a population-based study. J Thorac Cardiovasc Surg. 2018;155(6):2254-2264.e4. doi: 10.1016/j.jtcvs.2017.11.105 [DOI] [PubMed] [Google Scholar]
- 2.Garbade J, Jenniches M, Borger MA, et al. Outcome of patients suffering from acute type B aortic dissection: a retrospective single-centre analysis of 135 consecutive patients. Eur J Cardiothorac Surg. 2010;38(3):285-292. doi: 10.1016/j.ejcts.2010.02.038 [DOI] [PubMed] [Google Scholar]
- 3.MacGillivray TE, Gleason TG, Patel HJ, et al. The Society of Thoracic Surgeons/American Association for Thoracic Surgery clinical practice guidelines on the management of type B aortic dissection. J Thorac Cardiovasc Surg. 2022;163(4):1231-1249. doi: 10.1016/j.jtcvs.2021.11.091 [DOI] [PubMed] [Google Scholar]
- 4.Sailer AM, Nelemans PJ, Hastie TJ, et al. Prognostic significance of early aortic remodeling in acute uncomplicated type B aortic dissection and intramural hematoma. J Thorac Cardiovasc Surg. 2017;154(4):1192-1200. doi: 10.1016/j.jtcvs.2017.04.064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Afifi RO, Sandhu HK, Leake SS, et al. Outcomes of patients with acute type B (DeBakey III) aortic dissection: a 13-year, single-center experience. Circulation. 2015;132(8):748-754. doi: 10.1161/CIRCULATIONAHA.115.015302 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Riambau V, Böckler D, Brunkwall J, et al. ; ESVS Guidelines Committee . Editor’s choice—management of descending thoracic aorta diseases: clinical practice guidelines of the European Society for Vascular Surgery (ESVS). Eur J Vasc Endovasc Surg. 2017;53(1):4-52. doi: 10.1016/j.ejvs.2016.06.005 [DOI] [PubMed] [Google Scholar]
- 7.Nienaber CA, Rousseau H, Eggebrecht H, et al. ; INSTEAD trial . Randomized comparison of strategies for type B aortic dissection: the Investigation of Stent Grafts in Aortic Dissection (INSTEAD) trial. Circulation. 2009;120(25):2519-2528. doi: 10.1161/CIRCULATIONAHA.109.886408 [DOI] [PubMed] [Google Scholar]
- 8.Nienaber CA, Kische S, Rousseau H, et al. ; INSTEAD-XL trial . Endovascular repair of type B aortic dissection: long-term results of the randomized investigation of stent grafts in aortic dissection trial. Circ Cardiovasc Interv. 2013;6(4):407-416. doi: 10.1161/CIRCINTERVENTIONS.113.000463 [DOI] [PubMed] [Google Scholar]
- 9.Brunkwall J, Kasprzak P, Verhoeven E, et al. ; ADSORB trialists . Endovascular repair of acute uncomplicated aortic type B dissection promotes aortic remodelling: 1 year results of the ADSORB trial. Eur J Vasc Endovasc Surg. 2014;48(3):285-291. doi: 10.1016/j.ejvs.2014.05.012 [DOI] [PubMed] [Google Scholar]
- 10.Weissler EH, Osazuwa-Peters OL, Greiner MA, et al. National trends in repair for type B aortic dissection. Clin Cardiol. 2021;44(8):1058-1068. doi: 10.1002/clc.23672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tsai TT, Fattori R, Trimarchi S, et al. ; International Registry of Acute Aortic Dissection . Long-term survival in patients presenting with type B acute aortic dissection: insights from the International Registry of Acute Aortic Dissection. Circulation. 2006;114(21):2226-2231. doi: 10.1161/CIRCULATIONAHA.106.622340 [DOI] [PubMed] [Google Scholar]
- 12.Durham CA, Cambria RP, Wang LJ, et al. The natural history of medically managed acute type B aortic dissection. J Vasc Surg. 2015;61(5):1192-1198. doi: 10.1016/j.jvs.2014.12.038 [DOI] [PubMed] [Google Scholar]
- 13.Fattori R, Montgomery D, Lovato L, et al. Survival after endovascular therapy in patients with type B aortic dissection: a report from the International Registry of Acute Aortic Dissection (IRAD). JACC Cardiovasc Interv. 2013;6(8):876-882. doi: 10.1016/j.jcin.2013.05.003 [DOI] [PubMed] [Google Scholar]
- 14.Qin YL, Wang F, Li TX, et al. Endovascular repair compared with medical management of patients with uncomplicated type B acute aortic dissection. J Am Coll Cardiol. 2016;67(24):2835-2842. doi: 10.1016/j.jacc.2016.03.578 [DOI] [PubMed] [Google Scholar]
- 15.Xiang D, Kan X, Liang H, et al. Comparison of mid-term outcomes of endovascular repair and medical management in patients with acute uncomplicated type B aortic dissection. J Thorac Cardiovasc Surg. 2021;162(1):26-36.e1. doi: 10.1016/j.jtcvs.2019.11.127 [DOI] [PubMed] [Google Scholar]
- 16.Shah TR, Rockman CB, Adelman MA, Maldonado TS, Veith FJ, Mussa FF. Nationwide comparative impact of thoracic endovascular aortic repair of acute uncomplicated type B aortic dissections. Vasc Endovascular Surg. 2014;48(3):230-233. doi: 10.1177/1538574413518122 [DOI] [PubMed] [Google Scholar]
- 17.Mody PS, Wang Y, Geirsson A, et al. Trends in aortic dissection hospitalizations, interventions, and outcomes among Medicare beneficiaries in the United States, 2000-2011. Circ Cardiovasc Qual Outcomes. 2014;7(6):920-928. doi: 10.1161/CIRCOUTCOMES.114.001140 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Munshi B, Doyle BJ, Ritter JC, et al. Surgical decision making in uncomplicated type B aortic dissection: a survey of Australian/New Zealand and European surgeons. Eur J Vasc Endovasc Surg. 2020;60(2):194-200. doi: 10.1016/j.ejvs.2020.04.005 [DOI] [PubMed] [Google Scholar]
- 19.Carroll BJ, Schermerhorn M, Kennedy KF, et al. Readmissions after acute type B aortic dissection. J Vasc Surg. 2020;72(1):73-83.e2. doi: 10.1016/j.jvs.2019.08.280 [DOI] [PubMed] [Google Scholar]
- 20.Virani SS, Alonso A, Aparicio HJ, et al. ; American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee . Heart disease and stroke statistics-2021 update: a report from the American Heart Association. Circulation. 2021;143(8):e254-e743. doi: 10.1161/CIR.0000000000000950 [DOI] [PubMed] [Google Scholar]
- 21.Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care. 2005;43(5):480-485. doi: 10.1097/01.mlr.0000160417.39497.a9 [DOI] [PubMed] [Google Scholar]
- 22.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-1139. doi: 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
- 23.Sorour AA, Kirksey L, Laczynski DJ, et al. Racial disparities in presentation and short-term outcomes for patients with acute type B aortic dissection. J Vasc Surg. 2022;75(6):1855-1863.e2. doi: 10.1016/j.jvs.2022.01.010 [DOI] [PubMed] [Google Scholar]
- 24.Segal JB, Chang HY, Du Y, Walston JD, Carlson MC, Varadhan R. Development of a claims-based frailty indicator anchored to a well-established frailty phenotype. Med Care. 2017;55(7):716-722. doi: 10.1097/MLR.0000000000000729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kundi H, Coskun N, Yesiltepe M. Association of entirely claims-based frailty indices with long-term outcomes in patients with acute myocardial infarction, heart failure, or pneumonia: a nationwide cohort study in Turkey. Lancet Reg Health Eur. 2021;10:100183. doi: 10.1016/j.lanepe.2021.100183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Butala NM, Raja A, Xu J, et al. Association of frailty with treatment selection and long-term outcomes among patients with chronic limb-threatening ischemia. J Am Heart Assoc. 2021;10(24):e023138. doi: 10.1161/JAHA.121.023138 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16(3):1141-1154. doi: 10.1214/aos/1176350951 [DOI] [Google Scholar]
- 28.Austin PC, Fine JP. Practical recommendations for reporting Fine-Gray model analyses for competing risk data. Stat Med. 2017;36(27):4391-4400. doi: 10.1002/sim.7501 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Andersen PK, Gill RD. Cox’s regression model for counting processes: a large sample study. Ann Stat. 1982;10(4):1100-1120. doi: 10.1214/aos/1176345976 [DOI] [Google Scholar]
- 30.Amorim LD, Cai J. Modelling recurrent events: a tutorial for analysis in epidemiology. Int J Epidemiol. 2015;44(1):324-333. doi: 10.1093/ije/dyu222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nienaber CA, Zannetti S, Barbieri B, Kische S, Schareck W, Rehders TC; INSTEAD Study Collaborators . Investigation of Stent Grafts in Patients With Type B Aortic Dissection: design of the INSTEAD trial—a prospective, multicenter, European randomized trial. Am Heart J. 2005;149(4):592-599. doi: 10.1016/j.ahj.2004.05.060 [DOI] [PubMed] [Google Scholar]
- 32.Sachs T, Pomposelli F, Hagberg R, et al. Open and endovascular repair of type B aortic dissection in the nationwide inpatient sample. J Vasc Surg. 2010;52(4):860-866. doi: 10.1016/j.jvs.2010.05.008 [DOI] [PubMed] [Google Scholar]
- 33.Iannuzzi JC, Stapleton SM, Bababekov YJ, et al. Favorable impact of thoracic endovascular aortic repair on survival of patients with acute uncomplicated type B aortic dissection. J Vasc Surg. 2018;68(6):1649-1655. doi: 10.1016/j.jvs.2018.04.034 [DOI] [PubMed] [Google Scholar]
- 34.Hysa L, Khor S, Starnes BW, et al. Cause-specific mortality of type B aortic dissection and assessment of competing risks of mortality. J Vasc Surg. 2021;73(1):48-60.e1. doi: 10.1016/j.jvs.2020.04.499 [DOI] [PubMed] [Google Scholar]
- 35.Fattori R, Cao P, De Rango P, et al. Interdisciplinary expert consensus document on management of type B aortic dissection. J Am Coll Cardiol. 2013;61(16):1661-1678. doi: 10.1016/j.jacc.2012.11.072 [DOI] [PubMed] [Google Scholar]
- 36.Durham CA, Aranson NJ, Ergul EA, et al. Aneurysmal degeneration of the thoracoabdominal aorta after medical management of type B aortic dissections. J Vasc Surg. 2015;62(4):900-906. doi: 10.1016/j.jvs.2015.04.423 [DOI] [PubMed] [Google Scholar]
- 37.Famularo M, Meyermann K, Lombardi JV. Aneurysmal degeneration of type B aortic dissections after thoracic endovascular aortic repair: a systematic review. J Vasc Surg. 2017;66(3):924-930. doi: 10.1016/j.jvs.2017.06.067 [DOI] [PubMed] [Google Scholar]
- 38.Neuhauser B, Greiner A, Jaschke W, Chemelli A, Fraedrich G. Serious complications following endovascular thoracic aortic stent-graft repair for type B dissection. Eur J Cardiothorac Surg. 2008;33(1):58-63. doi: 10.1016/j.ejcts.2007.10.010 [DOI] [PubMed] [Google Scholar]
- 39.Andersen ND, Hanna JM, Ganapathi AM, et al. Insurance status predicts acuity of thoracic aortic operations. J Thorac Cardiovasc Surg. 2014;148(5):2082-2086. doi: 10.1016/j.jtcvs.2014.03.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eFigure. Cohort Construction Through Application of Exclusion Criteria
eTable 1. Exclusion Criteria Codes
eTable 2. Outcome Definitions
eTable 3. Unweighted Baseline Characteristics by Year of Treatment Relative to 2013
eTable 4. Sensitivity Analysis of Association Between Baseline Characteristics and Upfront TEVAR With Frailty as a Continuous Variable
eTable 5. Observed Cumulative Incidences of All Secondary Outcomes
eTable 6. Sensitivity Analysis of Association Between Upfront TEVAR and Outcomes With Frailty as a Continuous Variable in Weighted Analysis