Reimbursement of long-term postplacement costs after endovascular abdominal aortic aneurysm repair
Article Outline
Objective
Postplacement cost of surveillance and secondary procedures over 5 years increases the global cost of endovascular aortic aneurysm repair (EVAR) by nearly 50%. This study identified and assessed the reimbursement received for long-term postplacement costs after EVAR.
Methods
Between December 1995 and June 2007, 360 patients underwent EVAR at a single institution. The reimbursement collected from charges of postplacement surveillance and secondary procedures related to the aneurysmal disease was evaluated and compared against the actual costs. All amounts were converted to year 2007 dollars. To minimize costs associated with the early learning curve, the initial 50 EVAR patients between December 1995 and 1998 were excluded. Patients with <1 year follow-up were also excluded. Data are expressed as mean ± standard error.
Results
The mean follow up after EVAR for 152 patients was 38.8 ± 1.8 months. Medicare, capitated insurance, and commercial insurance provided coverage for 85 (56.0%), 49 (32.2%), and 18 (11.8%) patients, respectively. The cumulative 5-year postplacement reimbursement received per patient was $9792 meeting 81.4% of the cumulative cost of $12,027 for a net loss of $2235 per patient. Although 123 (80.9%) patients without secondary procedures generated a 5-year cumulative gain of $1830 per patient, 29 (19.1%) patients with secondary procedures averaged a 5-year cumulative loss of $9378 per patient. The average reimbursement rate over the 5-year period was 35.8% ± 0.6%, with the lowest reimbursement rate seen in patients with Medicare at 31.6% ± 0.7%.
Conclusion
Current reimbursement is not sufficient to meet the costs associated with long-term surveillance and needed secondary procedures after EVAR. Inadequate reimbursement of costs associated with secondary procedures was the primary driver for the net institutional loss. Reimbursement for outpatient radiological procedures generated a modest surplus.
Decreased perioperative morbidity and mortality of endovascular abdominal aortic aneurysm repair (EVAR) offers an attractive alternative to open surgical repair.1, 2 However, the financial commitment involved with the high cost of the endograft device and compulsatory long-term surveillance may limit the widespread application of this procedure as a cost-effective alternative.3, 4 Initial enthusiasm that the cost savings from shortened operative time, intensive care unit stay, and length of overall hospital stay would compensate for the high endograft device cost has been tempered by the failure of reimbursement to meet the initial cost of EVAR.5, 6, 7, 8 The noticeable increase in price of endograft devices since the approval of their use in 1999 by the US Food and Drug Administration further exacerbates this disparity.8
The uncertainty of long-term durability of EVAR necessitates vigilant surveillance. Long-term postplacement surveillance, diagnostic studies, and possible secondary procedures has been shown to increase the global cost of EVAR by nearly 50%.9 However, the adequacy of reimbursement received for these costs remains unknown. This study analyzed the reimbursement received for long-term postplacement costs after EVAR from Medicare, capitated insurance, and commercial insurance at a single institution.
Methods
Between December 1995 and June 2007, 360 consecutive patients underwent EVAR of infrarenal aortic aneurysms at the Ochsner Clinic Foundation (OCF). The first 50 patients undergoing EVAR from December 1995 to December 1998 were excluded to minimize possible increased costs associated with the learning curve. Patients with less than 12 months of follow-up were also excluded.
The standard surveillance protocol after EVAR at OCF during the study period consisted of an outpatient visit at 2 weeks, then 1, 6, and 12 months, and yearly thereafter. All outpatient visits after the initial 2-week postoperative visit included four-view plain abdominal radiographs (anteroposterior, lateral, left and right oblique) and computed tomography (CT) scan of the abdomen and pelvis with and without intravenous contrast with 2.5 to 3.0 mm axial images. Duplex ultrasonography was substituted for CT in patients with renal insufficiency, defined as serum creatinine greater than 1.4 mg/dL. Noncontrasted CT scan was not utilized. If a secondary procedure was performed, the follow-up timetable was reset to 1, 6, and 12 months, and yearly thereafter with surveillance studies at each visit.
Cost analysis and reimbursement
The method of cost determination have been described previously.9 The previously published cost data has been updated for this report and the steps in our cost accounting system are briefly described here. Direct expenses are first consolidated monthly at a departmental (cost center) level and then are classified by the cost components of labor, supply, equipment, and facility. Next, components are divided into fixed expenses independent of patient volume and variable expenses dependent on patient volume. Individual patients are assigned cost codes for every chargeable service or supply item. Departmental managers and Hospital Decision Support then develop allocation statistics (relative value units) for each component of every cost code so that a final cost can be assigned. Individual patient costs are the sum of costs from each department utilized. The average cost calculations from January to June 2007 were used to adjust prior years' costs for the sake of data uniformity.
Institutional overhead expenses, costs of professional services including fees for anesthesia, radiology, and surgeon's fees, and outpatient visits were also included in cost determinations. Charges, including professional fees, were calculated by using department specific cost-to-charge ratios based on a fixed fee schedule. Costs for outpatient appointments were determined by using a time-based formula. Patient appointments were categorized as 10, 20, and 30 minutes according to evaluation and management codes. Outpatient department expenses were then allocated by the time coded for the follow-up appointment.
Overhead costs for the institution were added to the direct costs by a factor of 30% for each patient. This factor has been uniformly decided on by Hospital Decision Support and varies 1% to 2% monthly. This method allows for equitable distribution among patients: those patients with higher overall cost absorb more overhead cost. Overhead expense includes costs from departments not directly involved in patient care (eg, housekeeping, computer technology, finance), equipment and facility depreciation, and interest expense on borrowings.
All costs were converted to year 2007 dollars calculated with a 3.5% yearly inflation rate. The costs, charges, and reimbursement associated with the preoperative evaluation and initial endograft placement and hospitalization were excluded from the analysis. Only those costs incurred after discharge from the original EVAR and the reimbursement received for those costs were included in this study. Two accounting software databases were utilized (Eclipsys TSC, Atlanta, Ga; and Data Warehouse, Oracle, Redwood Shores, Calif).
Payors were classified into one of three categories: Medicare, capitated insurance, and commercial insurance. Medicare reimbursement was dependent upon inpatient or outpatient generated costs. For inpatient visits, reimbursement was calculated by multiplying the OCF-specific hospital base rate by the relative weight of the diagnostic related groups (DRG) assigned to the patient. Hospital base rates are influenced by a number of factors, including geographic location, local labor markets, rural or urban hospital status, and teaching or nonteaching hospital designation.10 Additionally, Medicare compensates for hospitals that serve a disproportional mix of indigent patients. Unlike hospital base rates which can be influenced by numerous factors, the DRG relative weights are set by Medicare and do not change.
For outpatient visits and procedures, reimbursement was calculated using a predetermined group of services known as ambulatory payment classification (APC) groups. Each APC organized patients into separate categories based on similar clinical characteristics and utilization of similar types and amounts of resources at similar costs, and was associated with a status indicator, payment rate, relative weight, national unadjusted coinsurance amount, and minimum unadjusted coinsurance amount.11 The relative weight was calculated on the basis of median cost of the services included in the APC group and the payment rate applied to all services bundled under the APC.
In contrast to Medicare's inpatient prospective payment system, which assigns a single DRG to each patient visit, the outpatient system may assign multiple APCs to a single visit. If a patient undergoes multiple unrelated significant procedures on the same day, separate APCs may be assigned, but the APC with the highest payment will be paid in full, with the remaining APCs reimbursed at a discounted rate. Multiple related procedures are assigned only one APC. None of the patients requiring secondary procedures within the study cohort had additional unrelated secondary procedures on the same outpatient visit that qualified to be billed with a separate, discounted APC.
Capitated insurance was defined as any insurance contract where the hospital received a prenegotiated and fixed stipend per month to encompass the global cost of that patient's care independent of the number of services utilized. To calculate the actual total reimbursement received per patient per year, the year-specific reimbursement rate was first identified by dividing the cumulative prenegotiated stipend received for that year by the total submitted charges. This reimbursement rate was then applied to each individual charge during that fiscal year to calculate the patient-specific reimbursement amount received per procedure.
Commercial insurance included all private health insurance plan providers, including for-profit and not-for-profit organizations. Reimbursement was predetermined by an annually negotiated contract rate established by OCF with each individual payor.
No patients in this study cohort had Armed Forces related coverage (TRICARE). The one uninsured private-pay patient in this cohort arranged a private contract with OCF and was included in the commercial insurance group.
Definitions
A secondary procedure was defined as any intervention (including diagnostic angiograms) performed due to the original aneurysm or for a complication of EVAR during the follow-up period.
Endoleak was identified as persistent blood flow outside the endograft. The types of endoleak have been described previously.12, 13 Indications for secondary procedures were the presence of a type I or type III endoleak, or a type II endoleak with significant aneurysm sac expansion (>5 mm).
Statistical analysis
The data are expressed as mean ± standard error. P < .05 was considered statistically significant. Continuous variables were compared between groups using the unpaired Student t test and analysis of variance (ANOVA) methods. Statistical analysis was performed using SAS software, version 9.1 (SAS Institute Inc, Cary, NC).
Results
Three-hundred and ten patients were treated with EVAR in the study period. Reasons for exclusion from the study included perioperative death (n = 2), incomplete reimbursement data (n = 3), follow-up at another institution (n = 16), patients who refused further follow-up (n = 12), and less than 1 year of follow-up (n = 125). These exclusions left a patient cohort of 152 patients.
The mean follow-up was 38.8 ± 1.8 months. Sixty-seven of 152 patients were compliant with all of their recommended outpatient visits and follow-up imaging. One or more outpatient radiologic studies were missed by 65 patients, and 57 patients missed one or more outpatient visits. Overall, 88.1% of outpatient visits and 84.1% of radiological imaging studies were completed. Greater than 1-year lapse between follow-up visits occurred in 20 patients (mean 17.4 ± 2.2 months).
The mean age of patients was 73.0 ± 0.6 years. Spinal/regional anesthesia was used in 126 of 152 patients (82.9%). Modular bifurcated endografts were deployed in 146 patients. Endograft types were Zenith (Cook Medical Inc, Bloomington, Ind) in 92, AneuRx (Medtronic, Minnesota, Minn) in 49, and Excluder (W.L. Gore & Associates, Flagstaff, Ariz.) in five patients. The other six patients were treated with endocuffs (AneuRx, n = 5; Zenith, n = 1) alone for saccular aneurysms (n = 4) and pseudoaneurysms (n = 2).
The cumulative 5-year reimbursement for all payors per patient was $9792, meeting 81.4% of the $12,027 in incurred cost (Table I). This reimbursement resulted in a net loss of $2235 for each treated patient. Medicare, capitated insurance, and commercial insurance provided coverage for 85 (56.0%), 49 (32.2%), and 18 (11.8%) patients, respectively. Of the three payors, commercial insurance had the highest reimbursement at $17,153 over 5 years (Table II), and was the only payor to provide reimbursement that exceeded the incurred costs of long-term surveillance (Fig 1).
Table I. Five-year cumulative postplacement costs and reimbursement per patient
| Follow-up year | No. | Postplacement cost | Reimbursement received |
|---|---|---|---|
| 1 | 152 | $2936 ±$515 | $2317±$216 |
| 2 | 109 | $1693±$454 | $1117±$208 |
| 3 | 66 | $2282±$730 | $2047±$628 |
| 4 | 54 | $1769±$631 | $1531±$371 |
| 5 | 32 | $3346±$1748 | $2780±$1184 |
| Cumulative | $12,027 | $9792 |
Table II. Five-year cumulative postplacement reimbursement by individual payor
| Medicare | Capitated insurance | Commercial insurance | ||||
|---|---|---|---|---|---|---|
| Follow-up year | No. | Reimbursement received | No. | Reimbursement received | No. | Reimbursement received |
| 1 | 85 | $1615±$266 | 49 | $2626±$318 | 18 | $5160±$750 |
| 2 | 60 | $765±$144 | 35 | $1022±$145 | 14 | $2861±$1412 |
| 3 | 34 | $1640±$830 | 23 | $1224±$319 | 9 | $5689±$3175 |
| 4 | 31 | $1770±$629 | 15 | $1384±$258 | 8 | $849±$179 |
| 5 | 16 | $3683±$2276 | 9 | $2275±$1193 | 7 | $2594±$1811 |
| Cumulative | $9473 | $8531 | $17,153 | |||
Fig 1.
Five-year cumulative cost, charges, and reimbursement by payor for all patients. The values in the figure represent the respective sum of the absolute costs, charges, and reimbursement per patient for each payor group.
The cumulative 5-year charges submitted for all patients, patients with secondary procedures, and patients without secondary procedures are presented in Fig 1, Fig 2, Fig 3, respectively. The overall reimbursement rate from all payors over 5 years for submitted charges was 35.9% ± 0.6% (Table III). The reimbursement rate for Medicare patients was the lowest in every patient cohort.
Fig 2.
Five-year cumulative cost, charges, and reimbursement by payor for patients needing a secondary procedure. The values in the figure represent the sum of the absolute costs, charges, and reimbursement per patient for each payor group.
Fig 3.
Five-year cumulative cost, charges, and reimbursement by payor for patients without secondary procedures. The values in the figure represent the sum of the absolute costs, charges, and reimbursement per patient for each payor group.
Table III. Five-year cumulative mean reimbursement rates relative to charge
| All patientsa | Secondary procedureb | No secondary procedurea | |
|---|---|---|---|
| n = 152 | n = 29 | n = 123 | |
| All patients | 35.9%±0.6% | 41.7%±2.0% | 35.1%±0.7%c |
| Medicare | 31.6%±0.7% | 35.3%±2.3% | 31.2%±0.7% |
| Capitated insurance | 40.7%±1.1% | 42.0%±2.0% | 40.4%±1.2% |
| Commercial insurance | 41.8%±2.1% | 49.5%±2.6% | 38.3%±2.6%c |
aThe reimbursement rate was similar between capitated insurance and commercial insurance (P > .05) but were both significantly greater than that of Medicare (P < .05). |
bThe reimbursement rate of each payor was significantly different from each other (P < .05). |
cStatistically significant (P < .05) vs the respective secondary procedure group. |
Fifty-three secondary procedures were performed in 29 (19.1%) patients at a mean time after EVAR of 27.9 ± 2.9 months (range 1 month to 58 months, Table IV). The overall 5-year cumulative cost per patient with secondary procedures was $33,314 with a reimbursement of $23,936, which compensated 71.8% of the incurred cost (Fig 2). Similar to the findings of the overall patient cohort, only those patients with commercial insurance received sufficient reimbursement to exceed the costs incurred from secondary procedures.
Table IV. Secondary procedures
| Procedure | Total | Mean costa |
|---|---|---|
| Endograft explant/EABb | 2 | $59,700±$6690 |
| Iliac limb placement/AECc | 1 | $30,511 |
| Open conversion (delayed) | 4 | $22,130±$3240 |
| Aorto-uni-iliac device | 1 | $21,717 |
| Iliac endocuff | 6 | $13,498±$2354 |
| Aortic endocuff | 5 | $9027±$2112 |
| Femoral-femoral bypass | 1 | $8505 |
| Laparoscopic IMA ligation | 1 | $8302 |
| Thrombolysis | 2 | $7697±$743 |
| PTA±stent | 2 | $5333±$346 |
| Translumbar glue embolization | 3 | $5020±$796 |
| Coil microembolization | 4 | $4802±$474 |
| Diagnostic angiogram | 20 | $2771±$391 |
| Abscess drainage | 1 | $1560 |
aData are mean ± standard error. |
bOne procedure was staged, with the costs of both procedures combined. |
cStaged placement of iliac limb after the initial endograft device malfunctioned. |
For patients without secondary procedures (n = 123, 80.9%) the 5-year cumulative cost was $4021 with an overall reimbursement of $5851 resulting in a net gain of $1830 per patient (Fig 3). All payors returned reimbursements that exceeded the incurred costs.
Discussion
This study demonstrates that current reimbursement for long-term follow-up surveillance and needed secondary procedures after EVAR is inadequate to cover the costs. Importantly, the average net loss of $2235 per patient was driven by the very poor reimbursement for needed secondary procedures in 19.1% of patients. Combined with the average loss of $2100 to $3800 that may be seen with the initial cost of EVAR,5, 6 the necessity of vigilant lifetime surveillance continues to increase the overall cost disparity between EVAR and open surgical repair.9 This shortfall in reimbursement for long-term surveillance after EVAR further exacerbates the economic pressure on health care facilities.
The initial hospital cost of EVAR exceeds that of open surgical repair mostly due to the endograft device cost, which can account for 47% to 78% of the total hospital cost dependent upon the number of endograft devices, catheters, balloons, and other devices utilized for each specific patient.5, 6, 7, 8, 14 In some cases, the reimbursement received for EVAR was lower than that for open surgical repair.15
With postplacement surveillance, one factor for the financial deficit may stem from the lower ratio of submitted charges to incurred costs for those patients with secondary procedures compared to those patients without secondary procedures (Fig 2, Fig 3). Patients with secondary procedures submitted an average charge 1.9-fold greater than the incurred cost while those patients without secondary procedures submitted an average charge 4.3-fold greater than the incurred cost (P < .05). While patients with secondary procedures generated an 8.3-fold increase in costs over those without secondary procedures ($33,314 vs $4021, respectively, P < .05), they had a disproportionately lower 3.7-fold increase in submitted charges ($62,768 vs $17,137, respectively, P < .05). Despite the modestly higher reimbursement rate for patients with secondary procedures vs that of patients without secondary procedures (41.7% ± 2.0% vs 35.1% ± 0.7%, respectively, P < .05), it was insufficient to compensate for the disproportionately smaller increase in charges, consequently resulting in an insufficient reimbursement and an overall deficit.
The factors responsible for the disparity between cost and charges seen between the cohort of patients with and without secondary procedures are difficult to isolate, but are likely based on the billing and accounting practices beyond the scope of this study. Billing regulations and restrictions, especially in association with outpatient Medicare reimbursement for multiple same-day procedures and nonallowable hospital expenses, may have accounted for the lower reimbursement seen in those patients with secondary procedures. However, it is difficult to definitively explain the lower charge-to-cost ratio seen in patients with secondary procedures compared to patients without secondary procedures, or the consistently higher charges per cost seen for patients with commercial insurance vs those with Medicare. In addition to this disparity, the lower reimbursement rate seen with Medicare in the majority of the study population may have also contributed to the overall reimbursement deficit.
Of the three factors identified as driving the overall deficit (unequal cost-to-charge ratio between patient cohorts, lower reimbursement rate of Medicare patients, and secondary procedure rate), decreasing the secondary procedure rate is the one factor that can be most directly influenced by the surgeon, and will likely have the largest impact in improving overall reimbursement.
The secondary procedure rate of 19.1% at 38.8 months at OCF falls within the range of 12% to 28% at 36 months published in recent literature.16, 17, 18 In the Zenith US multicenter trial, the incidence of secondary procedures was 19.5% at 48 months.17 With the current distribution of payors, reimbursement rates, and device costs identified in this study, a secondary procedure rate of 16.3% would be required for the reimbursement received to equal the costs of long-term surveillance after EVAR at OCF.
Further improvements in technical skill, patient selection, endograft device durability, and endograft manufacturing to decrease the incidence of endograft failure, migration, or endoleak will all be instrumental in decreasing the rate of secondary procedures and eliminating a deficit between reimbursement and cost. Improved accuracy in documentation resulting in more accurate billing may provide a greater degree of reimbursement.
A shift in the surveillance protocol to monitoring with the use of abdominal ultrasound imaging instead of CT scan19 may decrease the cost of long-term surveillance. Use of the EndoSure AAA Wireless Pressure Measurement System (CardioMEMS, Atlanta, Ga) within the aneurysmal sac during EVAR as a surveillance tool to identify an endoleak20 is another potential alternative to CT scans. However, it is unknown whether this surveillance regimen would be cost effective, as the initial expense of the pressure sensor is $3500, and the device reader cost is $25,000. While these methods may result in potential cost savings of long-term surveillance, this may not translate to an increase in reimbursement. In most health care settings, reimbursement for CT scanning exceeds the cost, resulting in an incremental profit. A shift away from routine use of CT scanning may actually increase the overall net loss for surveillance after EVAR.
Limitations
Several limitations exist in this study. First, the analysis of costs and reimbursement received at a single institution is not universally applicable. Multiple factors will influence the hospital-specific reimbursement rate, including and not limited to geographic location, the local labor market, and the population demographic served by the hospital. In addition, billing and accounting methods will likely differ between institutions, and the base contract rates negotiated by the hospital with each individual payor varies annually and could even vary between hospitals within the same geographic region.
The relatively small sample size of patients with secondary procedures may have limited the ability for meaningful statistical analysis due to the low power, and may explain why much of the differences in the cost, charge, and reimbursement in this patient cohort failed to reach statistical significance.
Secondly, the distribution of patient payors at OCF is not reflective of the national average. Patients with Medicare comprised 56.0% of the patients undergoing EVAR at OCF. Nationally, 76.0% to 82.2% of patients undergoing EVAR had Medicare coverage within the same time period.21 The lower reimbursement rate by Medicare vs commercial insurance seen in this study is particularly concerning for hospitals whose payor population more closely resembles the national average.
Additionally, less than half of the overall cohort of EVAR patients at OCF obtained complete follow-up of all scheduled outpatient visits and surveillance imaging per protocol. The calculated cost and reimbursement presented in this study does not represent all of the potential costs and reimbursement of a patient after EVAR. Instead the data reflects the values to be expected from a “real-world” patient population.
Lastly, there may also be device-specific costs unique to the brand of endograft used in EVAR. Other institutions utilizing different endograft devices may experience different rates of secondary procedures and subsequent variables in cost and reimbursement.
Conclusion
Reimbursement for the cost of long-term surveillance after EVAR is inadequate. Although the majority of patients was free from secondary procedures and generated a profit, a net deficit created by the minority of patients requiring secondary procedures resulted in the net loss of income post EVAR.
Author contributions
The authors thank P. Joseph Paul, BA, MBA, for cost data retrieval and cost analysis, and Xiaozhang Jiang, MS, with statistical analysis.
References
- A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351:1607–1618
- . Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day mortality results: randomized control trial. Lancet. 2004;364:843–848
- . The cost-effectiveness of endovascular repair vs open surgical repair of abdominal aortic aneurysms: a decision analysis model. J Vasc Surg. 1999;29:958–972
- . Abdominal aortic aneurysms: cost-effectiveness of elective endovascular and open surgical repair. Radiology. 2002;225:337–344
- Current hospital costs and Medicare reimbursement of endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2003;37:272–279
- . Update on open repair of abdominal aortic aneurysms: the challenges for endovascular repair. J Am Coll Surg. 2004;199:946–959
- . Hospital costs for endovascular and open repair of abdominal aortic aneurysm. J Am Coll Surg. 2003;197:64–70
- . Follow-up costs increase the cost disparity between endovascular and open abdominal aortic aneurysm repair. J Vasc Surg. 2005;42:912–918
- . Long-term postplacement cost after endovascular aneurysm repair. J Vasc Surg. 2007;46:9–15
- . http://www.cms.hhs.gov/AcuteInpatientPPS/Accessed September 21, 2007.
- . http://www.cms.hhs.gov/HospitalOutpatientPPS/Accessed September 21, 2007.
- . Type I and type II endoleak: a more useful classification for reporting results of endoluminal AAA repair. [Letter] J Endovasc Surg. 1998;5:189–191
- . Type III and type IV endoleak: toward a complete definition of blood flow in the sac after AAA repair. J Endovasc Surg. 1998;5:305–309
- Open vs endovascular repair of abdominal aortic aneurysms: what does each really cost?. Ann Vasc Surg. 2004;18:612–618
- . Improving aneurysm-related outcomes: nationwide benefits of endovascular repair. J Vasc Surg. 2006;43:446–452
- Secondary interventions after endovascular abdominal aortic aneurysm repair. Am J Surg. 2005;190:787–794
- . Zenith abdominal aortic aneurysm endovascular graft. J Vasc Surg. 2008;48:1–9
- . Outcomes of secondary interventions after abdominal aortic aneurysm endovascular repair. J Vasc Surg. 2004;39:298–305
- . Redefining postoperative surveillance after EVAR: recommendations based on 5-year follow-up in the US Zenith multi-center trial. J Vasc Surg. 2008;48:278–285
- Initial results of wireless pressure sensing for endovascular aneurysm repair: the APEX Trial – Acute Pressure Measurement to Confirm Aneurysmal Sac EXclusion. J Vasc Surg. 2007;45:236–242
- . United States Department of Health and Human Services, Rockville, MD. http://hcupnet.ahrq.govAccessed Nov 14, 2007.
Competition of interest: none.
PII: S0741-5214(08)01223-8
doi:10.1016/j.jvs.2008.07.064
© 2008 Published by Elsevier Inc.
