Advertisement

Midterm survival after thoracic endovascular aortic repair in more than 10,000 Medicare patients

Open ArchivePublished:October 17, 2014DOI:https://doi.org/10.1016/j.jtcvs.2014.10.036

      Objective

      Aneurysms and dissections of the descending thoracic aorta represent a complex substrate with a variety of therapeutic options. The introduction of thoracic endovascular aortic repair (TEVAR) has revolutionized the treatment of thoracic aortic disease. However, longitudinal analyses of post-TEVAR outcomes appropriately stratified by aortic disease remain limited.

      Methods

      A total of 11,996 patients undergoing TEVAR from 2005-2010 were identified from the Medicare/Centers for Medicare and Medicaid Services database. Patients were stratified by underlying aortic disease and the presence of Current Procedural Terminology (CPT) codes. Survival was assessed using Kaplan-Meier analysis. Cox proportional hazards analysis determined predictors of survival from TEVAR.

      Results

      After TEVAR, patients had a median survival of 57.6 months (95% confidence interval, 54.9-61.3 months). Although patients without CPT codes had significantly fewer recorded comorbidities, TEVAR survival was comparable between patients with and without CPT codes (56.3 vs 59.5 months, P = .54). The early and late incidence of death varied significantly by aortic disease. Patients with aortic rupture, acute aortic dissection, and aortic trauma had the highest early incidence of death, whereas late survival was highest in patients with acute aortic dissection, aortic trauma, and isolated thoracic aortic aneurysm. Although hospital TEVAR volume was not associated with survival, an independent hospital effect (determined by using a mixed-effect Cox model) associated certain hospitals with a hazard for death 50% of what it was at other hospitals.

      Conclusions

      TEVAR has been applied to a multitude of aortic diseases in the Medicare population; early and late post-TEVAR survival varies by aortic disease. The late incidence of death remains high in TEVAR recipients, although certain aortic diagnoses such as acute aortic dissection, aortic trauma, and isolated thoracic aortic aneurysm were associated with improved late survival. An independent hospital effect, but not hospital volume, is correlated with post-TEVAR survival. Future analyses of TEVAR outcomes using the Medicare database should adjust for underlying aortic diagnoses and the presence of CPT codes.

      CTSNet classification

      Abbreviations and Acronyms:

      CPT (Current Procedural Terminology), FDA (US Food and Drug Administration), ICD-9 (International Classification of Diseases, ninth revision), TEVAR (thoracic endovascular aortic repair)
      See related commentary on pages 823-4.
      Since its introduction in 1992, thoracic endovascular aortic repair (TEVAR) has seen rapid adoption as a treatment modality for disease involving the descending thoracic aorta.
      • Dake M.D.
      • Miller D.C.
      • Semba C.P.
      • Mitchell R.S.
      • Walker P.J.
      • Liddell R.P.
      Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms.
      • Scali S.T.
      • Goodney P.P.
      • Walsh D.B.
      • Travis L.L.
      • Nolan B.W.
      • Goodman D.C.
      • et al.
      National trends and regional variation of open and endovascular repair of thoracic and thoracoabdominal aneurysms in contemporary practice.
      • Walker K.L.
      • Shuster J.J.
      • Martin T.D.
      • Hess Jr., P.J.
      • Klodell C.T.
      • Feezor R.J.
      • et al.
      Practice patterns for thoracic aneurysms in the stent graft era: health care system implications.
      Initial enthusiasm for TEVAR, spawned by its low procedural morbidity, has been tempered by sobering midterm outcomes, particularly in older patients with multiple comorbidities.
      • Demers P.
      • Miller D.C.
      • Mitchell R.S.
      • Kee S.T.
      • Sze D.
      • Razavi M.K.
      • et al.
      Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts.
      • Shah A.A.
      • Barfield M.E.
      • Andersen N.D.
      • Williams J.B.
      • Shah J.A.
      • Hanna J.M.
      • et al.
      Results of thoracic endovascular aortic repair 6 years after United States Food and Drug Administration approval.
      Initially approved by the US Food and Drug Administration (FDA) only for the treatment of descending thoracic aortic aneurysms, TEVAR devices were often used off-label to treat catastrophes of the descending thoracic aorta—such as acute, complicated, type B aortic dissections and acute, traumatic aortic tears—with reasonable success.
      • Dake M.D.
      • Kato N.
      • Mitchell R.S.
      • Semba C.P.
      • Razavi M.K.
      • Shimono T.
      • et al.
      Endovascular stent-graft placement for the treatment of acute aortic dissection.
      • White R.A.
      • Miller D.C.
      • Criado F.J.
      • Dake M.D.
      • Diethrich E.B.
      • Greenberg R.K.
      • et al.
      Report on the results of thoracic endovascular aortic repair for acute, complicated, type B aortic dissection at 30 days and 1 year from a multidisciplinary subcommittee of the Society for Vascular Surgery Outcomes Committee.
      • Cambria R.P.
      • Crawford R.S.
      • Cho J.S.
      • Bavaria J.
      • Farber M.
      • Lee W.A.
      • et al.
      A multicenter clinical trial of endovascular stent graft repair of acute catastrophes of the descending thoracic aorta.
      • Tehrani H.Y.
      • Peterson B.G.
      • Katariya K.
      • Morasch M.D.
      • Stevens R.
      • DiLuozzo G.
      • et al.
      Endovascular repair of thoracic aortic tears.
      • Moainie S.L.
      • Neschis D.G.
      • Gammie J.S.
      • Brown J.M.
      • Poston R.S.
      • Scalea T.M.
      • et al.
      Endovascular stenting for traumatic aortic injury: an emerging new standard of care.
      Several recently published guidelines favor TEVAR over surgery for the treatment (when feasible) of acute catastrophe of the descending thoracic aorta.
      • Svensson L.G.
      • Kouchoukos N.T.
      • Miller D.C.
      • Bavaria J.E.
      • Coselli J.S.
      • Curi M.A.
      • et al.
      Expert consensus document on the treatment of descending thoracic aortic disease using endovascular stent-grafts.
      • Hiratzka L.F.
      • Bakris G.L.
      • Beckman J.A.
      • Bersin R.M.
      • Carr V.F.
      • Casey Jr., D.E.
      • et al.
      2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/SIR/STS/SVM guidelines for the diagnosis and management of patients with Thoracic Aortic Disease: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, American Association for Thoracic Surgery, American College of Radiology, American Stroke Association, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, Society of Thoracic Surgeons, and Society for Vascular Medicine.
      An attempt to further broaden the indications for TEVAR to include uncomplicated chronic type B aortic dissection was undertaken with the Investigation of Stent Grafts in Patients with Type B Aortic Dissection trial (INSTEAD; comparing TEVAR to optimal medical management), the only randomized controlled trial of TEVAR ever completed; however, no improvement in 2- or 5-year all-cause mortality was shown, probably because of the high incidence of nonaneurysmal death (and thus the low overall survival benefit of TEVAR) in patients with uncomplicated type B aortic dissection.
      • Nienaber C.A.
      • Rousseau H.
      • Eggebrecht H.
      • Kische S.
      • Fattori R.
      • Rehders T.C.
      • et al.
      Randomized comparison of strategies for type B aortic dissection: the INvestigation of STEnt Grafts in Aortic Dissection (INSTEAD) trial.
      • Nienaber C.A.
      • Kische S.
      • Rousseau H.
      • Eggebrecht H.
      • Rehders T.C.
      • Kundt G.
      • et al.
      Endovascular repair of type B aortic dissection: long-term results of the randomized investigation of stent grafts in aortic dissection trial.
      Several commercial devices have been approved for TEVAR, which have all been associated with good freedom from aneurysm-related death, although midterm all-cause mortality remains suboptimal.
      • Makaroun M.S.
      • Dillavou E.D.
      • Kee S.T.
      • Sicard G.
      • Chaikof E.
      • Bavaria J.
      • et al.
      Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis.
      • Go M.R.
      • Cho J.S.
      • Makaroun M.S.
      Mid-term results of a multicenter study of thoracic endovascular aneurysm repair versus open repair.
      • Matsumura J.S.
      • Cambria R.P.
      • Dake M.D.
      • Moore R.D.
      • Svensson L.G.
      • Snyder S.
      International controlled clinical trial of thoracic endovascular aneurysm repair with the Zenith TX2 endovascular graft: 1-year results.
      • Fairman R.M.
      • Criado F.
      • Farber M.
      • Kwolek C.
      • Mehta M.
      • White R.
      • et al.
      Pivotal results of the Medtronic Vascular Talent Thoracic Stent Graft System: the VALOR trial.
      • Foley P.J.
      • Criado F.J.
      • Farber M.A.
      • Kwolek C.J.
      • Mehta M.
      • White R.A.
      • et al.
      Results with the Talent thoracic stent graft in the VALOR trial.
      • Fairman R.M.
      • Tuchek J.M.
      • Lee W.A.
      • Kasirajan K.
      • White R.
      • Mehta M.
      • et al.
      Pivotal results for the Medtronic Valiant Thoracic Stent Graft System in the VALOR II trial.
      Since FDA approval made TEVAR available for general clinical use in the United States, analyses of large US registries such as the Nationwide Inpatient Sample and US Medicare database have shown a dramatic rise in the use of TEVAR, although the rate of open repair of descending thoracic aortic aneurysms did not decrease appreciably.
      • Scali S.T.
      • Goodney P.P.
      • Walsh D.B.
      • Travis L.L.
      • Nolan B.W.
      • Goodman D.C.
      • et al.
      National trends and regional variation of open and endovascular repair of thoracic and thoracoabdominal aneurysms in contemporary practice.
      • Walker K.L.
      • Shuster J.J.
      • Martin T.D.
      • Hess Jr., P.J.
      • Klodell C.T.
      • Feezor R.J.
      • et al.
      Practice patterns for thoracic aneurysms in the stent graft era: health care system implications.
      In particular, the US Medicare database provides unprecedented access to patient demographics, comorbidities, and operative characteristics while also providing long-term survival data; several studies have used this database to assess midterm outcomes after TEVAR and open repair of both descending thoracic aortic aneurysm and type B dissection.
      • Goodney P.P.
      • Travis L.
      • Lucas F.L.
      • Fillinger M.F.
      • Goodman D.C.
      • Cronenwett J.L.
      • et al.
      Survival after open versus endovascular thoracic aortic aneurysm repair in an observational study of the Medicare population.
      • Jones D.W.
      • Goodney P.P.
      • Nolan B.W.
      • Brooke B.S.
      • Fillinger M.F.
      • Powell R.J.
      • et al.
      National trends in utilization, mortality, and survival after repair of type B aortic dissection in the Medicare population.
      However, although specific subsets of TEVAR recipients have been studied, a broad overview is lacking of how TEVAR has been applied in US Medicare patients since the first TEVAR devices received FDA approval, particularly with respect to which types of aortic disease are being treated with TEVAR. Furthermore, prior analyses of the US Medicare database have not fully taken advantage of the variables available in the Medicare database to evaluate surgical complexity (eg, how arterial access was obtained, whether the left subclavian artery required coverage, whether additional procedures were required). The aim of our study was to evaluate the application of TEVAR in the US Medicare population since its approval by the FDA, with a particular focus on operative characteristics and making sure to stratify patients by aortic disease.

      Methods

       Data Collection and Study Population

      We retrospectively reviewed data from the Centers for Medicare and Medicaid Services administrative database from 2005 to 2010. Patient demographics and survival data were obtained from the Beneficiary Summary file; International Classification of Diseases, ninth revision (ICD-9) diagnosis codes pertaining to the descending thoracic aorta and ICD-9 procedural codes pertaining to TEVAR were obtained from the MedPar file; patient comorbidities were obtained from the Chronic Conditions file; and surgeon billed Current Procedural Terminology (CPT) codes pertaining to prior and current surgical procedures were obtained from the Carrier file. Using data from the Chronic Conditions file to define patient comorbidities has been validated, particularly for conditions requiring regular physician follow-up, which is the case for TEVAR-treated descending thoracic aortic disease.
      • Gorina Y.
      • Kramarow E.A.
      Identifying chronic conditions in Medicare claims data: evaluating the Chronic Condition Data Warehouse algorithm.
      Patients with an ICD-9 procedural code of 39.73 or a CPT code of 33880, 33881, 33883, 33884, 33886, 75956, 75957, 75958, or 75959 were designated as having undergone TEVAR and were included in our analysis (N = 11,996).
      • Seabrook G.R.
      Current Procedural Terminology (CPT) coding for descending thoracic aorta endovascular repair.
      • Seabrook G.R.
      Current Procedural Terminology (CPT) coding for endovascular intervention in the descending thoracic aorta.

       Determination of Aortic Disease

      An algorithm was used to determine each patient's aortic disease from the ICD-9 diagnosis codes recorded at the index admission and at prior admissions. We classified patients as having 1 of 8 underlying aortic diseases: descending thoracic aortic rupture at the current admission (441.1, 441.3, 441.5, 441.6), thoracic aortic trauma at the current admission (901.0, 902.0), chronic aortic dissection (ie, aortic dissection diagnosed before the current admission) (441.0, 441.00, 441.01, 441.02, 441.03), acute aortic dissection (ie, aortic dissection not diagnosed before the current admission), thoracoabdominal aortic aneurysm (diagnosed either previously or at the current admission) (441.7), thoracic aortic aneurysm with concomitant abdominal aortic aneurysm (diagnosed either previously or at the current admission) (441.2, 441.4), isolated thoracic aortic aneurysm (441.2, in current or prior diagnoses), and descending thoracic aortic disease without a recorded diagnosis (441.9 or no other ICD-9 code previously described).

       Statistical Analysis

      We identified 999 US hospitals that performed between 1 and 477 TEVARs in Medicare patients during the study period. Because rigorous studies of both center volume and center-specific variation have shown them to be independently associated with postprocedural outcomes,
      • Thabut G.
      • Christie J.D.
      • Kremers W.K.
      • Fournier M.
      • Halpern S.D.
      Survival differences following lung transplantation among US transplant centers.
      we included both variables in our analysis. A hospital was considered high volume if it performed ≥100 TEVARs in Medicare patients (15 hospitals met this criteria), and moderate volume if it performed 20 to 99 TEVARs in Medicare patients (122 hospitals met this criteria) over the study period (hospital Medicare TEVAR volume was used as a surrogate for overall hospital volume). Hospital-specific variation with respect to post-TEVAR survival was modeled as a random effect in a multivariate mixed-effect Cox model
      • Therneau T.M.
      • Grambsch P.M.
      • Pankratz V.S.
      Penalized survival models and frailty.
      that included all variables listed in Table 1 as covariates; this interhospital variation was statistically significant (likelihood-ratio test of θ = 0: 10.78; P = .001), implying that hospital-specific variation is associated with post-TEVAR survival. In our multivariable analysis, in an effort to show the importance of hospital-specific variation, an independent hazard for postprocedural death was calculated for all 999 hospitals, and a dichotomous variable representing hospitals among the top third of post-TEVAR survival was generated to account for hospital variation in subsequent analyses.
      Table 1Characteristics of thoracic endovascular aortic repair (TEVAR) recipients, both as a whole and stratified by whether Current Procedural Terminology (CPT) codes were available in the Medicare database
      CharacteristicEntire cohort (N = 11,966)With CPT codes (n = 7895)Without CPT codes (n = 4071)P value
      Demographic
       Age at operation, y74.3 ± 8.874.7 ± 8.973.7 ± 8.7<.001
       Gender, male6611 (55.3)4270 (54.1)2341 (57.5)<.001
       Race, white9919 (82.9)6600 (83.6)3319 (81.5).004
       Race, African American1415 (11.8)900 (11.4)515 (12.7).05
       Race, Asian224 (1.9)145 (1.8)79 (1.9).69
       Race, Hispanic172 (1.4)106 (1.3)66 (1.6).23
       Insurance in addition to Medicare2008 (16.8)206 (2.6)1802 (44.3)<.001
      Aortic disease
       Chronic aortic dissection1637 (13.7)1165 (14.8)472 (11.6)<.001
       Acute aortic dissection1217 (10.2)700 (8.9)517 (12.7)<.001
       Thoracic aortic aneurysm3751 (31.4)2467 (31.3)1284 (31.5).74
       Thoracoabdominal aortic aneurysm2329 (19.5)1587 (20.1)742 (18.2).01
       Thoracic and abdominal aortic aneurysm925 (7.7)665 (8.4)260 (6.4)<.001
       Rupture1055 (8.8)710 (9.0)345 (8.5).34
       Trauma317 (2.6)216 (2.7)101 (2.5).41
       Missing diagnosis pertaining to the descending thoracic aorta735 (6.1)385 (4.9)350 (8.6)<.001
      Comorbidities
       History of myocardial infarction733 (6.1)556 (7.0)177 (4.4)<.001
       Ischemic heart disease8417 (70.3)6158 (78.0)2259 (55.5)<.001
       Congestive heart failure4829 (40.4)3612 (45.8)1217 (29.9)<.001
       Atrial fibrillation2315 (19.4)1758 (22.3)557 (13.7)<.001
       Hyperlipidemia8581 (71.7)6261 (79.3)2320 (57.0)<.001
       Hypertension10,178 (85.1)7361 (93.2)2817 (69.2)<.001
       Chronic obstructive pulmonary disease5390 (45.0)4077 (51.6)1313 (32.3)<.001
       Chronic kidney disease3893 (32.5)2986 (37.8)907 (22.3)<.001
       End-stage renal disease510 (4.3)367 (4.7)143 (3.5).004
       Diabetes3403 (28.4)2548 (32.3)855 (21.0)<.001
       Anemia6566 (54.9)4921 (62.3)1645 (40.4)<.001
       Stroke/transient ischemic attack2265 (18.9)1722 (21.8)543 (13.3)<.001
       Alzheimer's/dementia967 (8.1)755 (9.6)212 (5.2)<.001
       History of cancer1801 (15.1)1349 (17.1)452 (11.1)<.001
      Operative center characteristics
       Year of operation2008 (2007-2009)2008 (2007-2009)2008 (2007-2010)
       High-volume TEVAR hospital
      More than 100 TEVARs in Medicare patients from 2005-2010.
      2798 (23.4)1980 (25.1)818 (20.1)<.001
       Medium- or high-volume TEVAR hospital
      More than 20 TEVARs in Medicare patients from 2005-2010.
      7511 (62.8)5185 (65.7)2326 (57.1)<.001
       High-performing TEVAR hospital5328 (44.5)3620 (45.8)1708 (42.0)<.001
      Values are presented as n (%) or median (interquartile range). Boldface type indicates P values <.05. CPT, Current Procedural Terminology; TEVAR, thoracic endovascular aortic repair.
      More than 100 TEVARs in Medicare patients from 2005-2010.
      More than 20 TEVARs in Medicare patients from 2005-2010.
      The primary end point was all-cause mortality. Patients' vital status and date of death were validated with National Death Index data from 2005-2008 and with an internal Medicare determination of death (which itself informs the Social Security Death Master File) in patients for whom National Death Index data was unavailable; agreement between the National Death Index and Medicare death composite for vital status and date of death was >99%.
      • Morales D.L.
      • McClellan A.J.
      • Jacobs J.P.
      Empowering a database with national long-term data about mortality: the use of national death registries.
      Post-TEVAR survival distributions were estimated with the nonparametric Kaplan-Meier method
      • Kaplan E.
      • Meier P.
      Nonparametric Estimation from Incomplete Observations.
      ; the log-rank test
      • Bland J.M.
      • Altman D.G.
      The logrank test.
      was used to compare differences between survival distributions. Post-TEVAR survival for the entire cohort was compared to an age-/sex-/race-matched general US population cohort (data from the National Center for Health Statistics) using the 1-sample log-rank test as described by Finkelstein and colleagues.
      • Finkelstein D.M.
      • Muzikansky A.
      • Schoenfeld D.A.
      Comparing survival of a sample to that of a standard population.
      Univariate and multivariable Cox proportional hazards regression analyses
      • Cox D.R.
      Regression Models and Life-Tables.
      assessed the effect of demographics, descending thoracic aortic disease, clinical comorbidities, prior operations, and surgical complexity on post-TEVAR survival. The proportional hazards assumption was tested with the Grambsch/Therneau method of plotting scaled Schoenfeld residuals
      • Grambsch P.M.
      • Therneau T.M.
      Proportional hazards tests and diagnostics based on weighted residuals.
      ; no significant deviations from the proportional hazards assumption were noted. Interactions between variables were explored, without any significant findings. Purposeful selection of covariates was used to create a multivariable model
      • Hosmer D.W.
      • Lemeshow S.
      • May S.
      Applied survival analysis: regression modeling of time-to-event data.
      ; variables hypothesized or previously shown to have clinical significance in TEVAR recipients were included along with novel variables that were plausibly significant (P ≤ .20) on univariate analysis. Variables insignificant (P > .05) by the Wald test in our multivariable models but that were plausible predictors of survival were included in our final models; covariate selection was also guided by optimizing the Akaike information criterion.
      • Akaike H.
      A new look at the statistical model identification.
      Two separate models were created: 1 for the entire cohort (N = 11,966), and 1 for all patients for whom CPT codes were available (n = 7895). Covariates used in the final model for the entire cohort included age, gender, race, insurance status, previous myocardial infarction, congestive heart failure, atrial fibrillation, hypertension, chronic obstructive pulmonary disease, chronic kidney disease, end-stage renal disease necessitating hemodialysis, diabetes, stroke or transient ischemic attack, Alzheimer's disease or dementia, history of cancer, aortic disease, aortic rupture, hospital volume, and hospital effect. Covariates used in the final model for patients with CPT codes included the aforementioned demographic, comorbidity, and hospital variables, along with method of stent graft deployment, whether the stent graft covered the left subclavian artery, concomitant cardiac surgery, concomitant thoracic aortic surgery, concomitant abdominal aortic surgery, concomitant visceral abdominal artery reimplantation, concomitant head- or neck-vessel surgery, and lumbar drain placement.
      Means are presented with standard deviations, and hazard ratios are presented with 95% confidence intervals. Because this study was exploratory, no adjustments were made for multiple comparisons.
      • Rothman K.J.
      No adjustments are needed for multiple comparisons.
      Analyses were conducted with Stata software (version 13, StataCorp LP, College Station, Tex).

      Results

       Patient Characteristics

      Of 11,966 patients who underwent TEVAR (Table 1), 1055 had an acute aortic rupture, 317 had traumatic aortic injury, 1637 had chronic aortic dissection, 1217 had acute aortic dissection, 2329 had thoracoabdominal aneurysm, 925 had thoracic aortic aneurysm with concomitant abdominal aortic aneurysm, and 3751 had thoracic aortic aneurysm alone. In the remaining 735 patients, we could not ascertain a diagnosis of the descending thoracic aorta.
      The overall cohort was stratified according to whether patients' CPT codes were available (n = 7895) or unavailable (n = 4071). These 2 groups differed substantially with regard to most of the patient characteristics we assessed; for example, patients without CPT codes had fewer comorbidities and were more likely to have other health insurance in addition to Medicare.

       Post-TEVAR Survival

      After TEVAR, patients had a median survival of 57.6 months (95% confidence interval, 54.9-61.3 months) (Figure 1, A). Post-TEVAR survival was comparable between patients with and without CPT codes (56.3 vs 59.5 months; P = .54 by log-rank) (Figure 1, B). In contrast, different diagnoses were associated with clear differences in survival (Figure 2). The late hazard of post-TEVAR death also differed by descending thoracic aortic disease type (Figure 3). The highest perioperative death rates were noted in patients with aortic rupture, acute aortic dissection, and aortic trauma. However, acute aortic dissection and aortic trauma were also associated with the lowest late postoperative death rates.
      Figure thumbnail gr1
      Figure 1Survival after thoracic endovascular aortic repair (TEVAR) in Medicare patients, modeled according to the Kaplan-Meier method. A, Post-TEVAR survival for the entire cohort compared with an age-/sex-/race-matched general US population cohort. B, Post-TEVAR survival for patients with and without Current Procedural Terminology (CPT) codes.
      Figure thumbnail gr2ad
      Figure 2A comparison of Kaplan-Meier survival among the different diseases of the descending thoracic aorta treated with thoracic endovascular aortic repair (TEVAR). For each graph, the survival of patients with the diagnosis of interest is compared with that of all other patients combined. A, Post-TEVAR survival for chronic aortic dissection. B, Post-TEVAR survival for acute aortic dissection. C, Post-TEVAR survival for thoracic aortic aneurysm. D, Post-TEVAR survival for thoracoabdominal aortic aneurysm. E, Post-TEVAR survival for thoracic aortic aneurysm with concomitant abdominal aortic aneurysm. F, Post-TEVAR survival for aortic rupture. G, Post-TEVAR survival for aortic trauma. H, Post-TEVAR survival for descending thoracic aortic disease but no recorded diagnosis.
      Figure thumbnail gr2eh
      Figure 2A comparison of Kaplan-Meier survival among the different diseases of the descending thoracic aorta treated with thoracic endovascular aortic repair (TEVAR). For each graph, the survival of patients with the diagnosis of interest is compared with that of all other patients combined. A, Post-TEVAR survival for chronic aortic dissection. B, Post-TEVAR survival for acute aortic dissection. C, Post-TEVAR survival for thoracic aortic aneurysm. D, Post-TEVAR survival for thoracoabdominal aortic aneurysm. E, Post-TEVAR survival for thoracic aortic aneurysm with concomitant abdominal aortic aneurysm. F, Post-TEVAR survival for aortic rupture. G, Post-TEVAR survival for aortic trauma. H, Post-TEVAR survival for descending thoracic aortic disease but no recorded diagnosis.
      Figure thumbnail gr3ad
      Figure 3A comparison of conditional Kaplan-Meier survival (survival among patients who survived the perioperative period; that is, alive at 180 days) among the different diseases of the descending thoracic aorta treated with thoracic endovascular aortic repair (TEVAR). For each graph, the survival of patients with the diagnosis of interest is compared with that of all other patients combined. A, Post-TEVAR 180-day conditional survival for patients with chronic aortic dissection. B, Post-TEVAR 180-day conditional survival for patients with acute aortic dissection. C, Post-TEVAR 180-day conditional survival for patients with thoracic aortic aneurysm. D, Post-TEVAR 180-day conditional survival for patients with thoracoabdominal aortic aneurysm. E, Post-TEVAR 180-day conditional survival for patients with thoracic aortic aneurysm with concomitant abdominal aortic aneurysm. F, Post-TEVAR 180-day conditional survival for patients with aortic rupture. G, Post-TEVAR 180-day conditional survival for patients with aortic trauma. H, Post-TEVAR 180-day conditional survival for patients with descending thoracic aortic disease but no recorded diagnosis.
      Figure thumbnail gr3eh
      Figure 3A comparison of conditional Kaplan-Meier survival (survival among patients who survived the perioperative period; that is, alive at 180 days) among the different diseases of the descending thoracic aorta treated with thoracic endovascular aortic repair (TEVAR). For each graph, the survival of patients with the diagnosis of interest is compared with that of all other patients combined. A, Post-TEVAR 180-day conditional survival for patients with chronic aortic dissection. B, Post-TEVAR 180-day conditional survival for patients with acute aortic dissection. C, Post-TEVAR 180-day conditional survival for patients with thoracic aortic aneurysm. D, Post-TEVAR 180-day conditional survival for patients with thoracoabdominal aortic aneurysm. E, Post-TEVAR 180-day conditional survival for patients with thoracic aortic aneurysm with concomitant abdominal aortic aneurysm. F, Post-TEVAR 180-day conditional survival for patients with aortic rupture. G, Post-TEVAR 180-day conditional survival for patients with aortic trauma. H, Post-TEVAR 180-day conditional survival for patients with descending thoracic aortic disease but no recorded diagnosis.

       Predictors of Post-TEVAR Survival for Entire Cohort

      Table 2 shows univariate and multivariable hazard ratios for death after TEVAR for all variables not involving CPT codes. Twenty variables were included in our multivariable model of predictors of survival after TEVAR. Factors associated with worse survival included older age, additional insurance (eg, a health maintenance organization), history of myocardial infarction, congestive heart failure, atrial fibrillation, chronic obstructive pulmonary disease, chronic kidney disease, end-stage renal disease necessitating hemodialysis, anemia, history of transient ischemic attack or stroke, Alzheimer's disease and/or dementia, history of cancer, and aortic rupture. Predictors of improved survival included a diagnosis of hypertension, isolated thoracic aortic aneurysm as the underlying aortic disease, and undergoing TEVAR at a high-performing TEVAR hospital.
      Table 2Predictors of survival following thoracic endovascular aortic repair (TEVAR) for the entire cohort (N = 11,966)
      PredictorMultivariable hazard ratio (95% CI)P value
      P value obtained from multivariable Cox proportional hazards regression model of predictors of post-TEVAR survival, as described in the Methods.
      Univariate hazard ratio (95% CI)P value
      P value obtained from univariate Cox proportional hazards regression model of post-TEVAR survival.
      Demographic characteristic
       Age at operation, y1.029 (1.025-1.034)<.0011.034 (1.030-1.038)<.001
       Gender, male1.06 (0.99-1.14).061.03 (0.96-1.09).44
       Race, white0.95 (0.87-1.03).20
       Race (African American)1.09 (0.98-1.20).101.13 (1.03-1.24).01
       Race (Asian)0.93 (0.74-1.19).58
       Race (Hispanic)0.90 (0.68-1.20).47
       Insurance in addition to Medicare1.21 (1.09-1.34)<.0010.97 (0.89-1.07).59
      Comorbidities
       History of myocardial infarction1.14 (1.01-1.28).031.42 (1.27-1.60)<.001
       Ischemic heart disease1.30 (1.21-1.40)<.001
       Congestive heart failure1.24 (1.15-1.33)<.0011.64 (1.54-1.75)<.001
       Atrial fibrillation1.14 (1.05-1.23).0011.45 (1.35-1.56)<.001
       Hyperlipidemia1.06 (0.99-1.14).12
       Hypertension0.83 (0.74-0.93).0021.32 (1.19-1.45)<.001
       Chronic obstructive pulmonary disease1.24 (1.16-1.33)<.0011.39 (1.31-1.48)<.001
       Chronic kidney disease1.36 (1.27-1.47)<.0011.74 (1.63-1.85)<.001
       End-stage renal disease1.93 (1.70-2.20)<.0012.27 (2.02-2.56)<.001
       Diabetes0.97 (0.90-1.04).421.17 (1.10-1.25)<.001
       Anemia1.21 (1.12-1.30)<.0011.61 (1.50-1.72)<.001
       Stroke/transient ischemic attack1.13 (1.05-1.22).0021.44 (1.33-1.54)<.001
       Alzheimer's/dementia1.21 (1.09-1.34)<.0011.72 (1.56-1.90)<.001
       History of cancer1.11 (1.02-1.21).011.25 (1.15-1.36)<.001
      Diagnosis
       Chronic aortic dissection0.85 (0.77-0.94).001
       Acute aortic dissection1.06 (0.95-1.18).28
       Thoracic aortic aneurysm0.78 (0.72-0.84)<.0010.66 (0.61-0.71)<.001
       Thoracoabdominal aortic aneurysm1.04 (0.96-1.12).32
       Concomitant thoracic aortic aneurysm and abdominal aortic aneurysm0.76 (0.67-0.86)<.001
       Rupture2.52 (2.31-2.76)<.0012.87 (2.64-2.13)<.001
       Aortic trauma0.96 (0.78-1.17).66
       Missing thoracic aortic diagnosis1.26 (1.12-1.42)<.001
      Operative center characteristics
       Year of operation1.01 (0.98-1.03).61
       High-volume TEVAR hospital
      More than 100 TEVARs in Medicare patients from 2005-2010.
      1.02 (0.95-1.10).581.10 (1.02-1.18).01
       Medium- to high-volume TEVAR hospital
      More than 20 TEVARs in Medicare patients from 2005-2010.
      1.05 (0.99-1.13).12
       High-performing TEVAR hospital0.51 (0.48-0.55)<.0010.55 (0.52-0.59)<.001
      Boldface indicates P values <.05. CI, Confidence interval; TEVAR, thoracic endovascular aortic repair.
      P value obtained from multivariable Cox proportional hazards regression model of predictors of post-TEVAR survival, as described in the Methods.
      P value obtained from univariate Cox proportional hazards regression model of post-TEVAR survival.
      More than 100 TEVARs in Medicare patients from 2005-2010.
      § More than 20 TEVARs in Medicare patients from 2005-2010.

       Characteristics of Patients With CPT Codes, Stratified by Aortic Disease

      Of the 7895 patients with CPT codes (Table 3), 710 had aortic rupture, 216 sustained a traumatic aortic injury, 1165 had chronic aortic dissection, 700 had acute aortic dissection, 1587 had thoracoabdominal aortic aneurysm, 665 had thoracic aortic aneurysm with concomitant abdominal aortic aneurysm, and 2467 had thoracic aortic aneurysm. Significant differences existed among patients with different descending thoracic aortic diseases, particularly with respect to patient comorbidities, previous operations, and the need for concurrent procedures.
      Table 3Demographic, comorbid, and operative characteristics of subcohort of thoracic endovascular aortic repair (TEVAR) recipients with Current Procedural Terminology codes (n = 7895)
      CharacteristicChronic aortic dissection (n = 1165)Acute aortic dissection (n = 700)Thoracic aortic aneurysm (n = 2467)Thoraco-abdominal aneurysm (n = 1587)Thoracic aortic aneurysm plus abdominal aortic aneurysm (n = 665)Rupture of thoracic aorta (n = 710)Trauma to thoracic aorta (n = 216)Missing diagnosis (n = 385)
      Demographic characteristic
       Age at operation, y73.0 ± 10.272.6 ± 10.375.1 ± 8.375.2 ± 7.476.0 ± 6.877.0 ± 8.370.3 ± 13.374.5 ± 8.8
       Gender, male625 (53.7)380 (54.3)1337 (54.2)824 (51.9)402 (60.5)351 (49.4)124 (57.4)227 (59.0)
       Race, white914 (78.5)530 (75.7)2079 (84.3)1381 (87.0)589 (88.6)589 (83.0)187 (86.6)331 (86.0)
       Race, African American176 (15.1)138 (19.7)247 (10.0)142 (9.0)53 (8.0)84 (11.8)19 (8.8)41 (10.7)
       Race, Asian24 (2.1)11 (1.6)51 (2.1)19 (1.2)10 (1.5)16 (2.3)5 (2.3)9 (2.3)
       Race, Hispanic22 (1.9)7 (1.0)37 (1.5)17 (1.1)8 (1.2)10 (1.4)2 (0.9)3 (0.8)
       Insurance in addition to Medicare25 (2.2)18 (2.6)66 (2.7)44 (2.8)20 (3.0)14 (2.0)5 (2.3)14 (3.6)
      Comorbidities
       History of myocardial infarction74 (6.4)52 (7.4)161 (6.5)119 (7.5)45 (6.8)52 (7.3)12 (6.0)40 (10.4)
       Ischemic heart disease929 (79.7)464 (66.3)1946 (78.9)1290 (81.3)558 (83.9)536 (75.5)114 (52.8)231 (83.4)
       Congestive heart failure576 (49.4)288 (41.1)1056 (42.8)744 (46.9)309 (46.5)353 (49.7)58 (26.9)228 (59.2)
       Atrial fibrillation310 (26.6)140 (20.0)500 (20.3)340 (21.4)151 (22.7)168 (23.7)34 (15.7)115 (29.9)
       Hyperlipidemia901 (77.3)483 (69.0)2003 (81.2)1328 (83.7)582 (87.5)518 (73.0)140 (64.8)306 (79.5)
       Hypertension1125 (96.6)626 (89.4)2298 (93.2)1499 (94.5)628 (94.4)664 (93.5)158 (73.2)363 (94.3)
       Chronic obstructive pulmonary disease528 (45.3)241 (34.4)1313 (53.2)945 (59.6)419 (63.0)346 (48.7)60 (27.8)225 (58.4)
       Chronic kidney disease478 (41.0)205 (29.3)832 (33.7)685 (43.2)261 (39.3)299 (42.1)47 (26.8)179 (46.5)
       End-stage renal disease63 (5.4)30 (4.3)82 (3.3)78 (4.9)20 (3.0)57 (8.0)3 (1.4)34 (8.8)
       Diabetes368 (31.6)192 (27.4)810 (32.8)500 (31.5)238 (35.8)220 (31.0)74 (34.3)146 (37.9)
       Anemia764 (65.6)379 (54.1)1479 (60.0)995 (62.7)411 (61.8)498 (70.1)119 (55.1)276 (71.7)
       Stroke/transient ischemic attack270 (23.2)110 (15.7)531 (21.5)339 (21.4)158 (23.8)181 (25.5)36 (16.7)97 (25.2)
       Alzheimer's/dementia119 (10.2)65 (9.3)229 (9.3)126 (7.9)57 (8.6)96 (13.5)14 (6.5)49 (12.7)
       History of cancer161 (13.8)96 (13.7)461 (18.7)292 (18.4)140 (21.1)117 (16.5)22 (10.2)60 (15.6)
      Previous operations
       Cardiac surgery
      Coronary artery bypass grafting/valve/mechanical circulatory support/transplant.
      184 (15.8)44 (6.3)231 (9.4)198 (12.5)89 (13.4)64 (9.0)14 (6.5)52 (13.5)
       Open thoracic aortic surgery275 (23.6)15 (2.1)140 (5.7)207 (13.0)60 (9.0)48 (6.8)6 (2.8)42 (10.9)
       TEVAR25 (2.2)8 (1.1)21 (0.9)14 (0.9)7 (1.1)9 (1.3)2 (0.9)7 (1.8)
       Open abdominal aortic surgery146 (12.5)1 (0.1)15 (0.6)245 (15.4)206 (31.0)45 (6.3)0 (0.0)101 (26.2)
       Endovascular aortic repair33 (2.8)1 (0.1)14 (0.6)36 (2.3)47 (7.1)9 (1.3)1 (0.5)17 (4.4)
       Visceral abdominal aorta branch surgery58 (5.0)3 (0.4)14 (0.6)198 (12.5)51 (7.7)12 (1.7)0 (0.0)22 (5.7)
       Head/neck artery operation
      Subclavian artery/common carotid artery/vertebral artery.
      178 (15.3)39 (5.6)262 (10.6)158 (10.0)62 (9.3)18 (2.5)10 (4.6)36 (9.4)
       Upper-extremity artery operation
      Axillary/brachial/radial/ulnar artery.
      8 (0.7)5 (0.7)1 (0.1)7 (0.4)1 (0.2)1 (0.1)1 (0.5)3 (0.8)
       Lower-extremity artery operation
      Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      26 (2.2)9 (1.3)30 (1.2)76 (4.8)33 (5.0)18 (2.5)1 (0.5)24 (6.2)
      Operative center characteristics
       Year of operation2008 (2007-2009)2008 (2007-2009)2009 (2007-2010)2008 (2007-2009)2008 (2007-2009)2008 (2007-2009)2009 (2007-2010)2008 (2007-2009)
       High-volume TEVAR hospital
      More than 100 TEVARs in Medicare patients from 2005-2010.
      375 (32.2)178 (25.4)508 (20.6)473 (29.8)140 (21.1)192 (27.0)36 (16.7)78 (20.3)
       Medium- to high-volume TEVAR hospital
      More than 20 TEVARs in Medicare patients from 2005-2010.
      851 (73.1)470 (67.1)1522 (61.7)1100 (69.3)390 (58.7)499 (70.3)131 (60.7)222 (57.7)
      Operation characteristics
       TEVAR with subclavian artery coverage458 (39.3)271 (38.7)837 (33.9)393 (24.8)192 (28.9)230 (32.4)81 (37.5)115 (29.9)
       Prior or concomitant head/neck bypass232 (19.9)82 (11.7)405 (16.4)176 (11.1)81 (12.2)61 (8.6)18 (8.3)51 (13.3)
       TEVAR without subclavian artery coverage693 (59.5)434 (62.0)1655 (67.1)1145 (72.2)468 (70.4)488 (68.7)140 (64.8)216 (56.1)
       Prior or concomitant head/neck bypass63 (5.4)23 (3.3)107 (4.3)59 (3.7)24 (3.6)20 (2.8)4 (1.9)17 (4.4)
       Any additional endoprosthesis
      Use of TEVAR extension endoprosthesis or endovascular aortic repair.
      382 (32.8)192 (27.4)712 (28.9)714 (45.0)219 (32.9)238 (33.5)40 (18.5)153 (39.7)
       Use of TEVAR extension endoprosthesis322 (27.6)154 (22.0)649 (26.3)538 (33.9)169 (25.4)203 (28.6)32 (14.8)124 (32.2)
       Use of endovascular aortic repair endoprosthesis93 (8.0)49 (7.0)94 (3.8)287 (18.1)61 (9.2)54 (7.6)9 (4.2)69 (17.9)
       Open femoral artery access for deployment807 (69.3)489 (69.9)1674 (67.9)1002 (63.1)450 (67.7)487 (68.6)171 (79.2)246 (63.9)
       Open iliac artery access for deployment144 (12.4)70 (10.0)407 (16.5)333 (21.0)92 (13.8)110 (15.5)13 (6.0)64 (16.6)
       Nonfemoral, noniliac artery access
      Percutaneous access or antegrade from carotid/proximal aorta.
      251 (21.6)163 (23.3)488 (19.8)323 (20.4)147 (22.1)147 (20.7)36 (16.7)96 (24.9)
       Additional use of open brachial artery access33 (2.8)20 (2.9)36 (1.5)63 (4.0)8 (1.2)16 (2.3)3 (1.4)13 (3.4)
       Cardiac surgery
      Coronary artery bypass grafting/valve/mechanical circulatory support/transplant.
      28 (2.4)27 (3.9)60 (2.4)19 (1.2)6 (0.9)8 (1.1)1 (0.5)17 (4.4)
       Use of cardiopulmonary bypass31 (2.7)29 (4.1)55 (2.2)25 (1.6)10 (1.5)13 (1.8)3 (1.4)10 (2.6)
       Thoracic aortic surgery82 (7.0)68 (9.7)141 (5.7)87 (5.5)30 (4.5)59 (8.3)7 (3.2)27 (7.0)
       Abdominal aortic surgery45 (3.9)39 (5.6)120 (4.9)178 (11.2)37 (5.6)55 (7.8)0 (0.0)29 (7.5)
       Visceral abdominal aorta branch surgery59 (5.1)29 (4.1)120 (4.9)258 (16.3)34 (5.3)46 (6.5)3 (1.4)37 (9.6)
       Head/neck artery operation
      Subclavian artery/common carotid artery/vertebral artery.
      204 (17.5)94 (13.4)369 (15.0)137 (8.6)75 (11.3)70 (9.9)13 (6.0)54 (14.0)
       Upper-extremity artery operation
      Axillary/brachial/radial/ulnar artery.
      14 (1.2)13 (1.9)19 (0.8)15 (1.0)5 (0.8)10 (1.4)0 (0.0)5 (1.3)
       Lower-extremity artery operation
      Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      168 (14.4)102 (14.6)379 (15.4)247 (15.6)110 (16.5)118 (16.6)20 (9.3)65 (16.7)
      Adjunct procedures
       Lumbar drain placement225 (19.3)101 (14.4)409 (16.6)390 (24.6)135 (20.3)83 (11.7)5 (2.3)59 (15.3)
      Data are presented as n (%), mean ± standard deviation, or median (interquartile range). TEVAR, Thoracic endovascular aortic repair.
      Coronary artery bypass grafting/valve/mechanical circulatory support/transplant.
      Subclavian artery/common carotid artery/vertebral artery.
      Axillary/brachial/radial/ulnar artery.
      § Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      More than 100 TEVARs in Medicare patients from 2005-2010.
      More than 20 TEVARs in Medicare patients from 2005-2010.
      # Use of TEVAR extension endoprosthesis or endovascular aortic repair.
      ∗∗ Percutaneous access or antegrade from carotid/proximal aorta.

       Predictors of Post-TEVAR Survival for the Subcohort With CPT Codes

      Table 4 shows univariate and multivariable hazard ratios for death after TEVAR for the subcohort of patients with CPT codes. Thirty variables were included in our multivariable model of predictors of survival after TEVAR (Table 3). Factors associated with worse survival included older age; male gender; history of myocardial infarction; congestive heart failure; atrial fibrillation; chronic obstructive pulmonary disease; chronic kidney disease; end-stage renal disease necessitating hemodialysis; anemia; history of transient ischemic attack or stroke; Alzheimer's disease or dementia; history of cancer; previous TEVAR or endovascular aneurysm repair; and concurrent cardiac, thoracic aortic, abdominal aortic, visceral abdominal aorta branch-artery, and head or neck artery surgery. Predictors of improved survival included having an isolated thoracic aortic aneurysm, having undergone a previous open thoracic aortic procedure, undergoing TEVAR at a high-performing TEVAR hospital, and receiving an aortic graft that did not cover the left subclavian artery.
      Table 4Predictors of survival following thoracic endovascular aortic repair (TEVAR) for subcohort with available Current Procedural Terminology codes (n = 7895)
      PredictorMultivariable hazard ratio (95% CI)P value
      P value obtained from multivariable Cox proportional hazards regression model of predictors of post-TEVAR survival, as described in the Methods.
      Univariate hazard ratio (95% CI)P value
      P value obtained from univariate Cox proportional hazards regression model of post-TEVAR survival.
      Demographic characteristic
       Age at operation, y1.031 (1.025-1.036)<.0011.032 (1.027-1.037)<.001
       Gender, male1.09 (1.01-1.18).041.05 (0.97-1.13).21
       Race, white0.98 (0.89-1.09).77
       Race, African American1.03 (0.91-1.17).591.08 (0.96-1.21).23
       Race, Asian1.02 (0.76-1.36).90
       Race, Hispanic0.80 (0.55-1.17).26
       Insurance in addition to Medicare0.86 (0.66-1.12).270.82 (0.63-1.07).14
      Comorbidities
       History of myocardial infarction1.18 (1.03-1.36).021.39 (1.21-1.59)<.001
       Ischemic heart disease1.32 (1.19-1.46)<.001
       Congestive heart failure1.19 (1.09-1.30)<.0011.65 (1.52-1.78)<.001
       Atrial fibrillation1.14 (1.04-1.25).0051.44 (1.32-1.57)<.001
       Hyperlipidemia1.03 (0.93-1.13).61
       Hypertension0.94 (0.78-1.13).521.51 (1.26-1.80)<.001
       Chronic obstructive pulmonary disease1.26 (1.16-1.36)<.0011.39 (1.29-1.51)<.001
       Chronic kidney disease1.38 (1.26-1.50)<.0011.77 (1.64-1.92)<.001
       End-stage renal disease2.08 (1.73-2.42)<.0012.47 (2.15-2.83)<.001
       Diabetes0.97 (0.89-1.06).541.15 (1.06-1.25).001
       Anemia1.22 (1.11-1.34)<.0011.63 (1.50-1.77)<.001
       Stroke/transient ischemic attack1.08 (0.98-1.20).111.40 (1.28-1.53)<.001
       Alzheimer's/dementia1.32 (1.17-1.48)<.0011.80 (1.61-2.02)<.001
       History of cancer1.09 (0.98-1.20).111.18 (1.07-1.30).001
      Diagnosis
       Chronic aortic dissection0.86 (0.76-0.96).008
       Acute aortic dissection0.96 (0.84-1.11).52
       Thoracic aortic aneurysm0.85 (0.77-0.93)<.0010.71 (0.65-0.78)<.001
       Thoracoabdominal aortic aneurysm1.01 (0.92-1.11).73
       Concomitant thoracic aortic aneurysm and abdominal aortic aneurysm0.73 (0.63-0.85)<.001
       Rupture2.57 (2.30-2.87)<.0012.80 (2.52-3.11)<.001
       Aortic trauma0.93 (0.72-1.19).55
       Missing thoracic aortic diagnosis1.41 (1.20-1.65)<.001
      Previous surgery
       Cardiac surgery
      CABG/Valve/MCS/transplant.
      1.07 (0.95-1.21).25
       Open thoracic aortic0.86 (0.75-0.98).0020.92 (0.80-1.05).20
       Open abdominal aortic0.99 (0.87-1.13).86
       TEVAR or endovascular aortic repair1.26 (1.03-1.54).031.44 (1.18-1.76)<.001
       Visceral abdominal aorta branch1.20 (1.01-1.43).04
       Head/neck artery
      Subclavian artery/common carotid artery/vertebral artery.
      0.88 (0.77-1.02).08
       Upper-extremity artery
      Axillary/brachial/radial/ulnar artery.
      1.96 (1.14-3.39).02
       Lower-extremity artery
      Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      1.38 (1.12-1.71).003
      Operative center characteristics
       Year of operation1.01 (0.98-1.04).57
       High-volume TEVAR hospital
      More than 100 TEVARs in Medicare patients from 2005-2010.
      1.06 (0.97-1.15).221.13 (1.04-1.23).005
       Medium- to high-volume TEVAR hospital
      More than 20 TEVARs in Medicare patients from 2005-2010.
      1.05 (0.97-1.14).22
       High-performing TEVAR hospital0.53 (0.49-0.57)<.0010.56 (0.51-0.60)<.001
      Operation characteristics
       TEVAR with subclavian artery coverage1.15 (1.06-1.24).001
       TEVAR with subclavian artery coverage and prior or concomitant head/neck bypass1.10 (0.98-1.23).09
       TEVAR without subclavian artery coverage0.87 (0.79-0.95).0020.88 (0.81-0.95).001
       Any additional endoprosthesis
      Use of TEVAR extension endoprosthesis or endovascular aortic repair.
      1.14 (1.05-1.24).001
       Use of TEVAR extension endoprosthesis1.10 (1.01-1.20).02
       Use of endovascular aortic repair endoprosthesis1.21 (1.07-1.38).003
       Open femoral artery access for deployment0.93 (0.86-1.01).07
       Open iliac artery access for deployment1.06 (0.95-1.18).281.12 (1.01-1.24).03
       Nonfemoral, noniliac artery access
      Percutaneous versus antegrade from carotid/proximal aorta.
      1.09 (0.99-1.19).09
       Open brachial artery access for deployment1.27 (1.00-1.61).05
       Cardiac surgery
      CABG/Valve/MCS/transplant.
      1.57 (1.22-2.03).0011.62 (1.27-2.05)<.001
       Use of cardiopulmonary bypass1.34 (1.05-1.73).02
       Thoracic aortic surgery1.46 (1.25-1.70)<.0011.62 (1.41-1.86)<.001
       Abdominal aortic surgery1.29 (1.12-1.49).0011.61 (1.41-1.85)<.001
       Visceral abdominal aorta branch surgery1.46 (1.27-1.68)<.0011.63 (1.43-1.85)<.001
       Head/neck artery operation
      Subclavian artery/common carotid artery/vertebral artery.
      1.22 (1.08-1.38).0011.25 (1.12-1.40)<.001
       Upper-extremity artery operation
      Axillary/brachial/radial/ulnar artery.
      1.24 (0.89-1.80).19
       Lower-extremity artery operation
      Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      1.29 (1.16-1.42)<.001
      Adjunct procedures
       Lumbar drain placement0.99 (0.89-1.10).880.93 (0.84-1.03).19
      Boldface indicates P value <.05. CI, Confidence interval; TEVAR, thoracic endovascular aortic repair.
      P value obtained from multivariable Cox proportional hazards regression model of predictors of post-TEVAR survival, as described in the Methods.
      P value obtained from univariate Cox proportional hazards regression model of post-TEVAR survival.
      CABG/Valve/MCS/transplant.
      § Subclavian artery/common carotid artery/vertebral artery.
      Axillary/brachial/radial/ulnar artery.
      Profunda/superficial femoral/popliteal/tibial trunk/popliteal artery.
      # More than 100 TEVARs in Medicare patients from 2005-2010.
      ∗∗ More than 20 TEVARs in Medicare patients from 2005-2010.
      †† Use of TEVAR extension endoprosthesis or endovascular aortic repair.
      ‡‡ Percutaneous versus antegrade from carotid/proximal aorta.

      Discussion

      Our study of predictors of midterm survival in Medicare TEVAR patients produced four particularly noteworthy findings. First, hospital volume was not associated with survival. However, there was a clear and significant hospital effect: The hazard of death at high-performing hospitals was about 50% of what it was at other hospitals. Whether patient selection, operative competence, postoperative clinical excellence, high-quality long-term management, or some combination of these factors underlies the hospital effect is unknown and warrants further study.
      Second, the early and late hazard of post-TEVAR death varied by disease type. For example, although patients who underwent TEVAR for acute aortic dissection or aortic trauma had a high incidence of periprocedural death, patients with these diseases also had the lowest incidence of late death (≥180 days); meanwhile, isolated thoracic aortic aneurysm was associated with the lowest early incidence of death, but a comparatively higher incidence of late death. Aortic rupture was associated with the worst incidence of both early and late hazard of death.
      Third, greater operative complexity was associated with worse postprocedural survival. The simplest aortic disease (isolated thoracic aortic aneurysm) was associated with the best early postprocedural survival, particularly if treatment did not require covering the left subclavian artery.
      Fourth and finally, records of Medicare patients without CPT codes had fewer documented comorbidities, even though these patients had similar post-TEVAR survival to patients with CPT codes. Thus, studies that do not differentiate between the patients with and without CPT codes are subject to selection bias.

       Interhospital Variability in Post-TEVAR Survival

      Somewhat surprisingly, we found no correlation between hospital TEVAR volume (at least in Medicare beneficiaries) and post-TEVAR survival. However, there was clear and significant interhospital variability with respect to survival, even when rigorous multivariable analysis was performed to control for patient comorbidities and operative characteristics. The importance of interhospital variability has been previously shown for other surgical procedures.
      • Thabut G.
      • Christie J.D.
      • Kremers W.K.
      • Fournier M.
      • Halpern S.D.
      Survival differences following lung transplantation among US transplant centers.
      Unfortunately, the Medicare database includes no data that could inform us whether patient selection, operative competence, the quality of postoperative clinical care or long-term management, or some combination of these factors underlies this important hospital effect.

       Aortic Diseases Treated by TEVAR

      TEVAR was applied to the spectrum of aortic disease, from isolated thoracic aortic aneurysm to aortic rupture, chronic aortic dissection, and even thoracoabdominal aneurysm. As expected, patients with acute catastrophes of the aorta such as acute aortic dissection or aortic trauma had fewer comorbidities than patients with aneurysms or chronic aortic dissection. The prevalence of prior cardiac, aortic, or peripheral arterial surgery also varied by aortic disease; patients with thoracoabdominal aneurysm, thoracic aortic aneurysm with concomitant abdominal aortic aneurysm, or chronic aortic dissection were more likely to have undergone a previous endovascular aneurysm repair, open abdominal aortic operation, or lower-extremity bypass than patients with other aortic diseases. Meanwhile, patients with chronic aortic dissection were more likely than others to have undergone a previous open thoracic aortic repair or TEVAR. Patients with more complex aortic disease also required more complex procedures. For example, patients with chronic aortic dissection were most unlikely to undergo TEVAR requiring subclavian artery coverage and concomitant head-neck artery bypass, whereas patients with thoracoabdominal aneurysm were most likely to require EVAR, open abdominal aortic repair, or abdominal aorta visceral artery branch surgery. Patients with acute aortic dissection, rupture of the aorta, or aortic trauma were less likely than other patients to require a lumbar drain, underscoring the more emergent nature of these diseases.

       Early and Late Hazard of Death Among TEVAR Recipients

      The median survival of 57.6 months (1- and 5-year survival of 76% and 48%) in this cohort is consistent with those found in previous analyses of midterm post-TEVAR outcomes.
      • Demers P.
      • Miller D.C.
      • Mitchell R.S.
      • Kee S.T.
      • Sze D.
      • Razavi M.K.
      • et al.
      Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts.
      • Shah A.A.
      • Barfield M.E.
      • Andersen N.D.
      • Williams J.B.
      • Shah J.A.
      • Hanna J.M.
      • et al.
      Results of thoracic endovascular aortic repair 6 years after United States Food and Drug Administration approval.
      Perioperative mortality varied by diagnosis. More complex or acute diagnoses were associated with worse perioperative survival, whereas the least complex condition—isolated thoracic aortic aneurysm that did not require stent-graft coverage of the left subclavian artery—was associated with the highest rate of perioperative survival. As has been previously been shown,
      • Goodney P.P.
      • Travis L.
      • Lucas F.L.
      • Fillinger M.F.
      • Goodman D.C.
      • Cronenwett J.L.
      • et al.
      Survival after open versus endovascular thoracic aortic aneurysm repair in an observational study of the Medicare population.
      aortic rupture was associated with the worst early, midterm, and late survival. Interestingly, although patients with acute aortic dissection and aortic trauma had poor periprocedural outcomes, they also had the lowest late postoperative death rates. We suggest that these patients represent a fundamentally less sick population than patients with chronic aortic dissection, thoracic aortic aneurysm, or thoracoabdominal aortic aneurysm; this intuition is supported by the younger ages and smaller number of comorbidities noted in acute aortic dissection and aortic trauma patients.
      The high late mortality rate of Medicare TEVAR recipients was perhaps the most disconcerting finding. In patients who survived to 180 days (assessed by using conditional survival) (Figure 3), well beyond the perioperative risk period associated with TEVAR, 6% to 12% of patients died per year, depending on aortic diagnosis. In nonacute aortic dissection and nonaortic trauma patients, this rate was even worse (9%-12%). The utility of TEVAR in patients without acute disease (acute aortic dissection, trauma, or rupture), who have an annualized death rate of ≥10% per year after the early periprocedural hazard phase of TEVAR, is worth debating.

       Comparing Medicare Patients With and Without CPT Codes

      A currently underutilized strength of the Medicare database is the CPT codes provided in the Carrier file. When CPT codes are available, linking them to a patient's hospitalization provides a wealth of data regarding the actual operation performed. CPT codes were available for 66% of our overall TEVAR cohort.
      Comparing patients with and without CPT codes yielded several notable differences. First, patients without CPT codes had a lower prevalence of all documented comorbidities than patients with CPT codes. We do not attribute this finding to a genuine difference in comorbidity but to the greater tendency for patients without CPT codes to have incomplete data in the Medicare database. Our finding that post-TEVAR survival was similar between patients with and without CPT codes seems to support this attribution; if patients without CPT codes actually had fewer comorbidities, we would expect their post-TEVAR survival to exceed that of patients with CPT codes. These findings underscore the risk associated with including patients with and without CPT codes together in an analysis of Medicare patients; that is, comorbidities are likely to be underrepresented in patients without CPT codes.
      Also significant was the difference in insurance status between patients with and without CPT codes. Nearly 50% of patients without CPT codes had insurance in addition to Medicare (compared with only 2.6% of patients with CPT codes), implying that for many patients with additional insurance, CPT codes are not billed directly to Medicare. Analyses of Medicare data must be account for these differences.

       Limitations

      Our study was limited by its retrospective design; its susceptibility to selection bias; the variability of ICD-9 and CPT code documentation across hospitals and regions; and the fact that the data from this administrative database are not audited or adjudicated for comorbidities, diagnoses, or procedures performed. We attempted to address these limitations by stratifying patients by whether they had CPT codes, and by using multivariable regression analyses to identify variables that predicted post-TEVAR survival. Nevertheless, unmeasured variables and the observational design of this study leave ample room for residual bias to confound our results. Also, ICD-9 codes do not fully capture the spectrum of thoracic aortic disease, and our aortic disease selection algorithm is imperfect in its ability to stratify patients appropriately. As well, it is possible that interactions exist between aortic disease and other patient variables that we were unable to explore in this analysis. Finally, given the size of the Medicare database, it is impossible to validate data entered into it, and some amount of inaccuracy should be expected, although random misclassification of data would not be expected to alter our findings.

      Conclusions

      TEVAR has been applied to a multitude of aortic diseases in the Medicare population, and early and late post-TEVAR survival varies significantly by disease. Less complex aortic disease is associated with less procedural complexity and improved post-TEVAR survival. Even among those patients who survive 6 months after their initial TEVAR, 6% to 12% die per year thereafter, depending on aortic disease. In this era of cost-effective medicine, the high late hazard for death of TEVAR recipients must be considered when one is deciding whether to offer this therapy to patients with nonacute conditions. Age, male gender, patient comorbidities, previous operations on the descending thoracic aorta, aortic disease, operative complexity, and hospital effect all predict post-TEVAR survival. Hospital effect is probably a surrogate for some combination of patient selection, operative competence, and the quality of postoperative clinical care and long-term management. Analyses of outcomes in TEVAR recipients using the Medicare database should adjust for underlying diagnoses and the presence of CPT codes, as well as use CPT codes to appropriately classify patients by procedural details to avoid selection bias. Finally, the post-TEVAR outcomes noted in this study reflect the daily practice of US surgeons, and the survival rates can serve as a real world benchmark for comparison with outcomes associated with newer technologies and therapies.
      The authors thank Bruce A. Reitz and Michael D. Dake for funding the acquisition of Medicare data. Justin Schaffer and Bharathi Lingala had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

      References

        • Dake M.D.
        • Miller D.C.
        • Semba C.P.
        • Mitchell R.S.
        • Walker P.J.
        • Liddell R.P.
        Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms.
        N Engl J Med. 1994; 331: 1729-1734
        • Scali S.T.
        • Goodney P.P.
        • Walsh D.B.
        • Travis L.L.
        • Nolan B.W.
        • Goodman D.C.
        • et al.
        National trends and regional variation of open and endovascular repair of thoracic and thoracoabdominal aneurysms in contemporary practice.
        J Vasc Surg. 2011; 53: 1499-1505
        • Walker K.L.
        • Shuster J.J.
        • Martin T.D.
        • Hess Jr., P.J.
        • Klodell C.T.
        • Feezor R.J.
        • et al.
        Practice patterns for thoracic aneurysms in the stent graft era: health care system implications.
        Ann Thorac Surg. 2010; 90: 1833-1839
        • Demers P.
        • Miller D.C.
        • Mitchell R.S.
        • Kee S.T.
        • Sze D.
        • Razavi M.K.
        • et al.
        Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts.
        J Thorac Cardiovasc Surg. 2004; 127: 664-673
        • Shah A.A.
        • Barfield M.E.
        • Andersen N.D.
        • Williams J.B.
        • Shah J.A.
        • Hanna J.M.
        • et al.
        Results of thoracic endovascular aortic repair 6 years after United States Food and Drug Administration approval.
        Ann Thorac Surg. 2012; 94: 1394-1399
        • Dake M.D.
        • Kato N.
        • Mitchell R.S.
        • Semba C.P.
        • Razavi M.K.
        • Shimono T.
        • et al.
        Endovascular stent-graft placement for the treatment of acute aortic dissection.
        N Engl J Med. 1999; 340: 1546-1552
        • White R.A.
        • Miller D.C.
        • Criado F.J.
        • Dake M.D.
        • Diethrich E.B.
        • Greenberg R.K.
        • et al.
        Report on the results of thoracic endovascular aortic repair for acute, complicated, type B aortic dissection at 30 days and 1 year from a multidisciplinary subcommittee of the Society for Vascular Surgery Outcomes Committee.
        J Vasc Surg. 2011; 53: 1082-1090
        • Cambria R.P.
        • Crawford R.S.
        • Cho J.S.
        • Bavaria J.
        • Farber M.
        • Lee W.A.
        • et al.
        A multicenter clinical trial of endovascular stent graft repair of acute catastrophes of the descending thoracic aorta.
        J Vasc Surg. 2009; 50 (e1-4): 1255-1264
        • Tehrani H.Y.
        • Peterson B.G.
        • Katariya K.
        • Morasch M.D.
        • Stevens R.
        • DiLuozzo G.
        • et al.
        Endovascular repair of thoracic aortic tears.
        Ann Thorac Surg. 2006; 82 (discussion 7-8): 873-877
        • Moainie S.L.
        • Neschis D.G.
        • Gammie J.S.
        • Brown J.M.
        • Poston R.S.
        • Scalea T.M.
        • et al.
        Endovascular stenting for traumatic aortic injury: an emerging new standard of care.
        Ann Thorac Surg. 2008; 85 (discussion 9-30): 1625-1629
        • Svensson L.G.
        • Kouchoukos N.T.
        • Miller D.C.
        • Bavaria J.E.
        • Coselli J.S.
        • Curi M.A.
        • et al.
        Expert consensus document on the treatment of descending thoracic aortic disease using endovascular stent-grafts.
        Ann Thorac Surg. 2008; 85: S1-S41
        • Hiratzka L.F.
        • Bakris G.L.
        • Beckman J.A.
        • Bersin R.M.
        • Carr V.F.
        • Casey Jr., D.E.
        • et al.
        2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/SIR/STS/SVM guidelines for the diagnosis and management of patients with Thoracic Aortic Disease: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, American Association for Thoracic Surgery, American College of Radiology, American Stroke Association, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, Society of Thoracic Surgeons, and Society for Vascular Medicine.
        Circulation. 2010; 121: e266-e369
        • Nienaber C.A.
        • Rousseau H.
        • Eggebrecht H.
        • Kische S.
        • Fattori R.
        • Rehders T.C.
        • et al.
        Randomized comparison of strategies for type B aortic dissection: the INvestigation of STEnt Grafts in Aortic Dissection (INSTEAD) trial.
        Circulation. 2009; 120: 2519-2528
        • Nienaber C.A.
        • Kische S.
        • Rousseau H.
        • Eggebrecht H.
        • Rehders T.C.
        • Kundt G.
        • et al.
        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: 407-416
        • Makaroun M.S.
        • Dillavou E.D.
        • Kee S.T.
        • Sicard G.
        • Chaikof E.
        • Bavaria J.
        • et al.
        Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis.
        J Vasc Surg. 2005; 41: 1-9
        • Go M.R.
        • Cho J.S.
        • Makaroun M.S.
        Mid-term results of a multicenter study of thoracic endovascular aneurysm repair versus open repair.
        Perspect Vasc Surg Endovasc Ther. 2007; 19: 124-130
        • Matsumura J.S.
        • Cambria R.P.
        • Dake M.D.
        • Moore R.D.
        • Svensson L.G.
        • Snyder S.
        International controlled clinical trial of thoracic endovascular aneurysm repair with the Zenith TX2 endovascular graft: 1-year results.
        J Vasc Surg. 2008; 47 (discussion 57): 247-257
        • Fairman R.M.
        • Criado F.
        • Farber M.
        • Kwolek C.
        • Mehta M.
        • White R.
        • et al.
        Pivotal results of the Medtronic Vascular Talent Thoracic Stent Graft System: the VALOR trial.
        J Vasc Surg. 2008; 48: 546-554
        • Foley P.J.
        • Criado F.J.
        • Farber M.A.
        • Kwolek C.J.
        • Mehta M.
        • White R.A.
        • et al.
        Results with the Talent thoracic stent graft in the VALOR trial.
        J Vasc Surg. 2012; 56: 1214-1221.e1
        • Fairman R.M.
        • Tuchek J.M.
        • Lee W.A.
        • Kasirajan K.
        • White R.
        • Mehta M.
        • et al.
        Pivotal results for the Medtronic Valiant Thoracic Stent Graft System in the VALOR II trial.
        J Vasc Surg. 2012; 56: 1222-1231.e1
        • Goodney P.P.
        • Travis L.
        • Lucas F.L.
        • Fillinger M.F.
        • Goodman D.C.
        • Cronenwett J.L.
        • et al.
        Survival after open versus endovascular thoracic aortic aneurysm repair in an observational study of the Medicare population.
        Circulation. 2011; 124: 2661-2669
        • Jones D.W.
        • Goodney P.P.
        • Nolan B.W.
        • Brooke B.S.
        • Fillinger M.F.
        • Powell R.J.
        • et al.
        National trends in utilization, mortality, and survival after repair of type B aortic dissection in the Medicare population.
        J Vasc Surg. 2014; 60: 11-19
        • Gorina Y.
        • Kramarow E.A.
        Identifying chronic conditions in Medicare claims data: evaluating the Chronic Condition Data Warehouse algorithm.
        Health Serv Res. 2011; 46: 1610-1627
        • Seabrook G.R.
        Current Procedural Terminology (CPT) coding for descending thoracic aorta endovascular repair.
        J Vasc Surg. 2006; 43: 106A-110A
        • Seabrook G.R.
        Current Procedural Terminology (CPT) coding for endovascular intervention in the descending thoracic aorta.
        J Vasc Surg. 2010; 52: 103S-106S
        • Thabut G.
        • Christie J.D.
        • Kremers W.K.
        • Fournier M.
        • Halpern S.D.
        Survival differences following lung transplantation among US transplant centers.
        JAMA. 2010; 304: 53-60
        • Therneau T.M.
        • Grambsch P.M.
        • Pankratz V.S.
        Penalized survival models and frailty.
        J Comput Graph Stat. 2003; 12: 156-175
        • Morales D.L.
        • McClellan A.J.
        • Jacobs J.P.
        Empowering a database with national long-term data about mortality: the use of national death registries.
        Cardiol Young. 2008; 18: 188-195
        • Kaplan E.
        • Meier P.
        Nonparametric Estimation from Incomplete Observations.
        J Am Stat Assoc. 1958; 53: 457-481
        • Bland J.M.
        • Altman D.G.
        The logrank test.
        BMJ. 2004; 328: 1073
        • Finkelstein D.M.
        • Muzikansky A.
        • Schoenfeld D.A.
        Comparing survival of a sample to that of a standard population.
        J Natl Cancer Inst. 2003; 95: 1434-1439
        • Cox D.R.
        Regression Models and Life-Tables.
        J R Stat Soc Series B Stat Methodol. 1972; 34: 187-220
        • Grambsch P.M.
        • Therneau T.M.
        Proportional hazards tests and diagnostics based on weighted residuals.
        Biometrika. 1994; 81: 515-526
        • Hosmer D.W.
        • Lemeshow S.
        • May S.
        Applied survival analysis: regression modeling of time-to-event data.
        2nd ed. Wiley-Interscience, Hoboken2008
        • Akaike H.
        A new look at the statistical model identification.
        IEEE Trans Automat Contr. 1974; 19: 716-723
        • Rothman K.J.
        No adjustments are needed for multiple comparisons.
        Epidemiology. 1990; 1: 43-46