National trends and regional variation of open and endovascular repair of thoracic and thoracoabdominal aneurysms in contemporary practice

Abstract

Objectives

Successful surgical management of thoracic aortic aneurysms (TAA) and thoracoabdominal aortic aneurysms (TAAA) has historically relied upon open surgical repair (OSR). More recently, the advent and application of thoracic endovascular stent graft aneurysm repair (TEVAR) permutations have become increasingly performed in contemporary practice. To better determine the effect of TEVAR techniques on OSR, we examined national and regional trends in treatment use.

Methods

All Medicare patients from 1998 through 2007 undergoing isolated TAA and TAAA repair were analyzed using a clinically validated algorithm using diagnostic International Classification of Disease 9th revision (ICD-9; 441.1, 441.2, 441.6, 441.7, 441.9) codes and procedural (ICD-9 OSR: 38.35, 38.45 and TEVAR: 39.73, 39.79) codes. Differential rates of OSR and TEVAR were compared across census tract regions during the study interval.

Results

Total complex aortic repairs increased by 60%, from 10.8 to 17.8/100,000, between 1998 and 2007 (P <.001). A dramatic increase occurred in TEVAR (not performed in 1998, 5.8/100,000 in 2007) during the study period, but OSR rates remained stable during the same interval (10.7 to 12.0/100,000 in 2007, P < NS). There was substantial regional variation for both OSR and TEVAR. This regional variation was greater in OSR (range, 8.8–16.7/100,000) than in TEVAR (range, 4.5–6.9/100,000).

Conclusions

Degenerative TAA and TAAA aneurysms are being repaired in the United States at an increasing rate. This reflects the rapid acceptance of TEVAR, which apparently supplements rather than supplants OSR. There appears to be greater regional variation in OSR compared with TEVAR. These data may have significant implications for those interested in the effect of new technologies on health care and cost containment.

Open surgical repair (OSR) has been the main stay of therapy for the treatment of degenerative aneurysmal disease in the thoracic aorta (TAA) and thoracoabdominal aorta (TAAA) for nearly 4 decades.^1–3 An increase in the incidence of thoracic aortic pathology has been documented during this interval, likely related to the widespread application of improved diagnostic modalities.^4,5 This has resulted in an increased demand for thoracic aortic intervention. However, despite the advent of high-volume centers of excellence, operative mortality and associated morbidities remain in the 10% range.^1,4–6 The adoption of thoracic endovascular aortic aneurysm repair (TEVAR) has been reported to decrease morbidity, paraplegia risk, and short-term mortality rates. Such favorable outcomes have resulted in the rapid and widespread application of TEVAR in contemporary practice.^7–10

Associated costs referable to OSR of TAAA have been previously estimated to be as high as $119,000.¹¹ Earlier series have documented net hospital losses, particularly in TAAA patients undergoing OSR, primarily due to increased lengths of stay.¹² The initial cost-analysis data for the use of TEVAR demonstrate that either equivalent or only slightly reduced (range, $10,000–$20,000) associated costs to the patient and the insurers can be anticipated.^13,14 This partly reflects the high associated costs for currently available commercial devices.

The rapid adoption of endovascular therapies in other anatomic locations for a variety of vascular disease processes, including lower extremity revascularization, has had a significant effect on decreasing the rate of OSR.¹⁵ The introduction of endoluminal therapies in the mesenteric circulation for acute mesenteric ischemia was heralded with an initial decrease in OSR over time.¹⁶ In a similar fashion, the effect of EVAR for infrarenal AAA disease has been accompanied by a significant decrease in the number of patients undergoing OSR.¹⁷ The purpose of this study was to determine the effect of TEVAR on rates of OSR to document current national trends in treatment use.

METHODS

Database

The Medicare Provider Analysis and Review (MEDPAR) file was interrogated. Part A claims data analysis was performed on all Medicare beneficiaries under-going open and endovascular treatment of TAA and TAAA from 1998 through 2007. The data in this file include demographic information, admission-related diagnostic International Classification of Diseases, 9th Revision (ICD-9) codes, operative Current Procedural Terminology (CPT) codes (American Medical Association, Chicago, Ill), discharge data, and hospital charges on Medicare admissions to certified hospitals and skilled nursing facilities. Procedures were identified by using a combination of CPT codes, diagnostic, and procedural ICD-9 codes using a clinically validated algorithm. The clinical algorithm consisted of a record review of a random sample of 20 records from two institutions of preidentified patients with a known history of TAA or TAAA repair.

Coding strategy

The coding strategy included capture of all ruptured and nonruptured TAA and TAAA open (ICD-9: 38.35, 38.45; CPT: 33875, 33877, 35091, 35092) and endovascular repairs (ICD-9: 39.79; CPT: 33880, 33881, 33883, 33884, 33886). In addition, open and endovascular repair of thoracic and thoracoabdominal dissection were analyzed if a code for aneurysm was also found because many of these are likely for chronic dissection with aneurysmal degeneration (ICD-9: 441, 441.1, 441.2, 441.7, 441.03).

Excluded were patients with concomitant codes for valve replacement, coronary artery bypass, grafting cardio-plegia, aorta-to-great vessel bypass, and hypothermic arrest. An inclusionary code (ICD-9 441.9) for all aortic aneurysms of unspecified site, without mention of rupture, aneurysm, dilatation of the aorta, or hyaline necrosis of the aorta, was also arrogated (Fig 1). This strategy was used in an attempt to eliminate patients with ascending aortic aneurysms and type A dissection while maximizing capture of the TAA and TAAA repairs. This algorithm was piloted at Dartmouth-Hitchcock Medical Center and the University of Florida and was accurate in identifying hospital admissions for thoracic endovascular and open procedures.

Fig. 1.

International Classification of Diseases, 9th edition and Current Procedural Terminology coding strategy for thoracic aortic aneurysms (TAA) and thoracoabdominal aortic aneurysms (TAAA) TEVAR, Thoracic endovascular aortic aneurysm repair.

Geographic analysis

After the inclusion criteria were established, information derived from the Medicare Denominator file was analyzed to ascertain the geographic variation in incidence of each procedure over time between 1998 and 2007. Data were analyzed between nine census tract regions of the United States, as defined by the U.S. Census Bureau (Fig 2). Census tract regions are census areas or districts that often coincide with the limits of cities, towns, or other administrative areas, and several tracts commonly exist within a county. Census tracts represent the smallest territorial units for which population data are available in many countries, including the United States (http://www.census.gov).

Fig. 2.

Map shows the nine census tract regions of the United States, as designated by the U.S. Census Bureau. Adapted from http://www.census.gov/factsheet.

Statistics

We used t tests to compare rates between regions. Nonparametric tests of trend were used to test significance across years; values of P <.05 were considered significant. All analyses were performed using SAS (SAS Institute, Cary, NC), and STATA 10 software (StataCorp, College Station, Tex).

RESULTS

Changes in use of OSR and TEVAR

During the study interval from 1998 to 2007, the total repair rate of TAA and TAAAs increased substantially. In 1998, 10.8 procedures were performed per 100,000 Medicare beneficiaries (Fig 3). By 2007, however, this rate increased by 60% to 17.8/100,000 (P <.001). This increase in the repair rate was due almost entirely to a rapid increase in the use of TEVAR. OSR procedures for thoracic and TAAA modestly increased by 10.8%, from 10.7/100,000 beneficiaries in 1998 to 12/100,000 beneficiaries in 2007 (P = NS).

Fig. 3.

Rates of total thoracic and thoracoabdominal aortic aneurysm repair (diamonds), open repair (squares) and endovascular repair (triangles) in Medicare patients, 1998–2007. CI, Confidence interval.

In contrast, a dramatic increase was demonstrated in the use of endovascular repair, which was not performed in 1998. As shown in Fig 3, a more rapid application of endovascular repair of TAA and TAAA appears to have occurred after 2004. This likely coincides with the publication of the GORE TAG trial results and availability of the TEVAR CPT code in 2005. During a 3-year interval from 2004 to 2007, the proliferation of TEVAR in contemporary practice resulted in a greater than threefold increase in use across the United States. By 2007, the national rate of TEVAR use for isolated thoracic and TAAA repairs had increased to 5.8 procedures per 100,000.

In 2000, only 1.6% of all thoracic procedures in the study cohort were performed using TEVAR, and 98% were performed with OSR. Among OSRs, nearly 40% were performed for descending TAA (DTAA), 21% for TAAA, and 29% for aortic dissection–related etiologies. Ruptured aneurysms constituted 6% of the descending TAA cohort and 4% of the TAAA group. In contrast, by 2007, repairs were endovascular in 31% of the study group and open in 69%. Among patients undergoing TEVAR repair, 51% had descending TAA, 22% had TAAA, 17% were performed for aortic dissection–related pathology, such as aneurysmal degeneration, 7% were performed for ruptured DTAA, and 2% were performed for ruptured TAAA. Among those patients undergoing open repair, 54% were performed for DTAA, 12% for TAAA, 28% for aortic dissection-related pathology, 3% for ruptured DTAA, and 2% for ruptured TAAA. Because of coding imprecision, these groups were collectively aggregated to better reflect treatment use over time (Table).

Table.

Aggregated pathology and method of repair for thoracic aortic and thoracoabdominal aortic aneurysms among Medicare beneficiaries from 2000 to 2007

Variable	Total cohort (%)	OSR (%)	TEVAR (%)
Nonruptured
Descending thoracic aneurysms	49	83	17
Thoracoabdominal aneurysms	17	81	19
Dissection	26	90	10
Other (thoracic aortic pathology)	0.3	55	45
Ruptured
Descending thoracic aneurysms	5	77	23
Thoracoabdominal aneurysms	3	89	11

OSR, Open surgical repair; TEVAR, thoracic endovascular aneurysm repair.

Regional variation

To better understand the most recent effect and use of OSR and TEVAR on regional variation, a geographic analysis of the adaptation of the two therapies across nine census tract regions was compared within the year 2007 (Fig 4). Relatively comparable rates of TEVAR use across census tract regions were seen, with a range of 4.5 to 6.9/100,000 Medicare beneficiaries. The extent of variation in repair rates across regions was greater for OSR (range 8.8–16.7/100,000) than TEVAR (range 4.5–6.9/100,000) in both absolute and relative terms.

Fig. 4.

Regional variation in use of open and endovascular repair of thoracic and thoracoabdominal aneurysms (TEVAR) among Medicare beneficiaries in 2007.

DISCUSSION

This study provides a view of the current trends in treatment of TAA and TAAA in Medicare patients. These findings confirm TEVAR has increased significantly on a national level.^8,18–20 Unlike other vascular disease processes in the extracranial carotid, mesenteric, or infrarenal aortic territories, endovascular treatment of degenerative aneurysms in the thoracic and thoracoabdominal aorta appears to have increased the total number of repairs.^18–20 This rapid increase is largely due to TEVAR, whereas rates of OSR have remained relatively stable. In addition, TEVAR rates vary much less across regions than OSR. This suggests that the relative technical simplicity and reduction in resource utilization of TEVAR compared with OSR have promoted rapid and more uniform adoption across the country.

Previously, the advent of new endoluminal therapies has been associated with its rapid substitution for its open surgical counterpart, as seen in the cerebrovascular circulation with carotid endarterectomy (CEA) and carotid artery stenting (CAS). The dissemination of CAS into 95% of hospital referral regions has occurred with apparent substitution of CEA, explaining an overall decline in open surgical reconstruction in the extracranial carotid circulation. This change in carotid revascularization patterns occurred before the comparative effectiveness of the two therapies was adequately analyzed in the surgical literature.²¹

This precedent is again seen in other index vascular procedures, such as treatment of mesenteric arterial occlusive disease. Schermerhorn et al¹⁶ described an overall decrease in the number of open surgical procedures for the treatment of acute mesenteric ischemia alone, while a rapid increase in angioplasty, with or without stenting, was observed. In a similar fashion, endovascular management of aneurysm disease localized to the infrarenal aorta has been associated with substantial change in treatment paradigms. EVAR has decreased rates of OSR while leading to increased variability of procedure rates across the United States. This may be consistent with the hypothesis that broadening the indications—perhaps to patients at higher risk for OSR—has led to greater variability in clinical decision making.¹⁷

Perhaps the most dramatic effect of catheter-based therapies on the management of vascular disease is demonstrated in infrainguinal arterial reconstruction. Between 1996 and 2006, the national rate of lower extremity peripheral interventions increased nearly threefold, whereas open surgical bypass decreased by 42%.¹⁵

The reason why endoluminal therapy for TAA and TAAA appears to supplement rather than supplant open repair may be due to a number of factors. Traditional open repair of these types of aneurysms requires formidable skill and enormous clinical resources driving a concentration on centers of excellence as a cardinal concept in the field.³ Current OSR of DTAA and TAAA can anticipate good short-term and long-term technical and functional out-comes.^3,22 Although recent reports have begun to espouse similar notions in the endovascular management of acute aortic syndromes, this perception is not fully proven.^4,5 This is likely reflected in the relatively consistent use of TEVAR across census tract regions compared with the significant variability with OSR (Fig 4). Within 3 years of U.S. Food and Drug Administration (FDA) approval, TEVAR rates across a variety of geographic regions had effectively tripled and have very little variance in utilization. This appears to support the notion of widespread acceptance of this new technology in modern vascular surgery practice.

Another potential explanation for the additive nature of endovascular treatment in the thoracic and thoracoabdominal aorta to overall utilization rates is that a variety of authors have demonstrated good short-term and midterm results with TEVAR and an attendant low morbidity and mortality rate. These early experiences may have incited enthusiasm amongst a number of providers across different disciplines, leading to the rapid development of TEVAR programs in hospital systems that may not routinely perform open TAA or TAAA repair. In addition, it is quite possible that treatment thresholds or operative risk profiles are different between patients treated with TEVAR vs OSR because these trends were initially seen with the introduction of EVAR in the management of aneurysms isolated to the infrarenal aorta.^23,24

An interesting question is specifically which types of aneurysms are being treated with TEVAR? Unlike a number of aortic pathologies, such as dissection, penetrating aortic ulcer, or traumatic disruption, TAAAs have the added complexity and routine prerequisite of branch vessel preservation as part of the definitive repair strategy. Currently available devices in the United States do not readily accommodate this requirement unless branched, fenestrated, or chimney techniques are used.²⁵ Given the added complexity of treating visceral-segment aneurysm disease, one could hypothesize that national rates for OSR for such lesions would remain stable. In addition, this may explain the greater regional variation with OSR because volume-outcome relationships and tertiary medical centers with designated interest for TAAA will affect referral patterns.

This finding of stable rates for OSR for aneurysmal disease in the DTAA and the TAAA brings into sharp relief the current management of isolated, degenerative DTAA. Several clinical trials have documented clinical equipoise between open repair and endovascular repair of DTAA.^8,9,26–28 These data, combined with the demonstration of lower short-term morbidity, have resulted in a paradigm shift in the management of these lesions²⁹ that has occurred despite a lack of level 1 evidence to support the widespread use of this technology. Moreover, a consensus statement by the Society of Thoracic Surgeons Endovascular Surgery Task Force recently described the lack of prospective, randomized data comparing the two types of repairs in the same aortic pathology.³⁰

Another point sustaining debate about the appropriate application of these two treatments is that many patients who undergo TEVAR for isolated TAA require lifelong surveillance and often require reintervention, leading to higher downstream costs compared with OSR.³¹ Issues of graft migration, collapse, and endoleak are not infrequent complications of endovascular therapies in this region of the aorta.^32,33 Despite newly found enthusiasm for the adoption of endovascular techniques for the management of disease traditionally repaired by OSR, lessons can be learned from the potentially negative implications of regional variation in the application of other index vascular procedures, such as stroke risk and CEA.³⁴

This study has several limitations, including the potential coding errors that are encountered with an administrative database. The procedural ICD-9 and CPT codes specific for TEVAR (39.79; 33880, 33881, 33883, 33884, 33886) were not routinely used throughout the study interval by Centers for Medicare and Medicaid Services. Additionally, endovascular therapies are often used off-label, and some patients may have undergone intervention for nonaneurysmal disease. It is unclear if these misclassifications did occur or whether it would be systematically different year to year. This may possibly have introduced unmeasured confounding into the analysis. Cost data were not analyzed in this study, so definitive conclusions cannot be drawn about how open and endovascular repair compare.

Another limitation of this analysis is that there are no current FDA-approved therapies for TAAAs in the United States, so drawing definitive conclusions about the effect TEVAR has on these aneurysms is not possible at this time. Although one can speculate that TEVAR is strictly affecting the management of isolated DTAAs nationally, the full effect of coding errors, debranching, chimney techniques, and homemade fenestrated endografting is not known.

To gain a complete perspective on the current national management of all TAA and TAAA during the initial phases of TEVAR utilization, these pathologies were queried together. It is likely that the boundaries between isolated DTAAs and TAAA will become increasingly blurred as newer endovascular technologies are developed. These data will provide a reference point against which future studies can be compared once fenestrated branched technology begins to permeate into the U.S. treatment paradigm of degenerative aneurysmal disease of the thoracic and thoracoabdominal aorta.

Lastly, this study demonstrates differences in use of OSR and TEVAR for TAA and TAAA, but whether this variation represents overuse or underuse of the two therapies across regions remains uncertain. To answer this question, a study of indications and a comparison of outcomes data are needed to fully understand the clinical effect of TEVAR vs OSR in the treatment of TAA and TAAA disease. Unfortunately, the relatively widespread and uniform acceptance of TEVAR nationally would likely make a randomized, prospective trial virtually impossible to perform. Some practitioners may even believe such a trial would be unethical in light of the contemporary data supporting the superiority of TEVAR to OSR with regard to short-term morbidity and, possibly, in-hospital mortality.^24,31

CONCLUSIONS

Repair rates in the United States for degenerative TAA and TAAA increased from 1998 to 2007. Unlike in alternate anatomic locations, endovascular management of thoracic aortic pathology has not supplanted but has supplemented historical treatment paradigms. These data may have significant implications for the application of new technologies and health care cost containment. Because national practice patterns can change rapidly in response to innovation, rapidly performed, accurate analyses comparing clinical efficacy and cost-effectiveness are needed to keep these changes rational.