XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Vascular OnlineFirst, published on April 28, 2015 as doi:10.1177/1708538115584507
Original Article
Renal artery stenting may not be appropriate for patients with atherosclerotic renal artery stenosis
Vascular 0(0) 1–10 ! The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1708538115584507 vas.sagepub.com
Anahita Dua1, Sachin Desai2, Gilbert R Upchurch Jr3 and Sapan S Desai4
Abstract Introduction: This study compared aortorenal bypass to renal artery stenting to determine the most efficacious and financially sound method for treating patients with atherosclerotic renal artery stenosis (RAS). Methods: A decision analysis using direct and indirect costs, and value of statistical life (VSL) was completed. Direct costs were obtained using the Nationwide Inpatient Sample (NIS), indirect costs from the National Institute of Diabetes and Digestive and Kidney Diseases, and VSL from the Department of Transportation. A variance-based sensitivity analysis was completed to assess the accuracy of the decision analysis. Results: Aortorenal bypass has a 95% five-year patency, a 98% 30-day survival, a 26% rate of overall complications, and a 70% five-year dialysis-free survival. Renal artery stenting has a 56% five-year patency, a 99% 30-day survival, a 40% rate of complications, and a 65% five-year dialysis-free survival. Renal artery stenting has an overall cost of $305,370 and aortorenal bypass has an overall cost of $103,453 per patient. After accounting for VSL, renal artery stenting has a negative value of $182,270 and aortorenal bypass has a value of $415,881. Conclusions: Lower five-year patency and higher rate of complications from renal artery stenting that ultimately lead to significantly lower five-year dialysis-free survival.
Keywords Renal artery stenosis, renal artery stenting
Introduction Renal artery stenosis (RAS) affects up to 32% of patients with hypertension, and 22% of patients with end stage renal disease.1,2 Approximately 10% of lesions with over 60% stenosis will progress to occlusion at three years.3,4 Proper management of RAS may contribute to improving two major public health problems: hypertension and kidney disease. Medical management, endovascular procedures (including angioplasty and stenting), and open surgery are used to treat patients with RAS. Angioplasty with balloon-expandable stent is the preferred therapeutic option for RAS. This approach has a 56% five-year patency with a 17% restenosis rate. Morbidity varies between 0% and 43%, and mortality between 0% and 5%. Patients have a 63% three-year dialysis-free survival, however up to 25% of patients experience a decline in renal function with negligible long-term improvement in hypertension.5–9
Open surgical options include aortorenal bypass, iliorenal bypass, hepatorenal bypass, and splenorenal bypass. Other options include endarterectomy with patch angioplasty, and reimplantation of the renal artery. Of these modalities, aortorenal bypass has the highest five-year patency at 95%. Morbidity varies between 7% and 24%, and mortality between 1% and 3% in recent series. Patients have a 70% fiveyear dialysis-free survival, and 85% have full resolution of their hypertension at that time.10–12 1 Department of Surgery, Medical College of Wisconsin, Milwaukee, WI, USA 2 Private Industry, Chicago, IL, USA 3 Department of Surgery, University of Virginia, Charlottesville, VA, USA 4 Department of Vascular Surgery, Southern Illinois University, Springfield, IL, USA
Corresponding author: Anahita Dua, 8701 W Watertown Plank Rd, Milwaukee, WI 53226, USA. Email: [email protected]
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
2
Vascular 0(0)
The optimal management of atherosclerotic RAS is controversial. The cardiovascular outcomes in renal atherosclerotic lesions (CORAL) trial demonstrates equivalency in outcome between renal artery stenting and medical management, and that both of these are poor options for treating RAS.5 While aortorenal bypass appears to be a viable alternative to both renal artery stenting and optimal medical management in patients with atherosclerotic RAS, the relative efficacy of aortorenal bypass has not been evaluated. Such a clinical trial would require many years to complete and would have issues in obtaining sufficient power; the CORAL trial enrolled 947 patients and had to be extended by an additional year due to similar concerns. Due to the limitations in completing a head-to-head clinical trial, we sought to compare the outcomes and relative value of aortorenal bypass with renal artery stenting using a decision analysis.
Methods Data sources A decision analysis was completed using direct costs, indirect costs, and overall value. Direct costs include the actual cost of care incurred by the payer or insurance at the time of care. Examples of direct costs include the cost of renal artery stenting or aortorenal bypass, and the cost of the inpatient hospitalization along with any inpatient complications. Direct costs were obtained from the Nationwide Inpatient Sample (NIS). The NIS is a part of the Health Care Utilization Project (HCUP) that is maintained by the Agency for Healthcare Research and Quality (AHRQ). It is the largest all-payer inpatient database and includes a stratified 20% random sample of all nonfederal inpatient hospital admissions throughout the United States. Clinical records were obtained with the use of International Classification of Diseases, Ninth Revision (ICD-9) diagnosis and procedure codes to ensure that the sample included only patients who had a primary diagnosis of RAS and underwent either aortorenal bypass or renal artery stenting. Renal angiography was used as an adjunct procedure code in order to more accurately select patients who underwent stenting (Table 1). No patient consent was required as this was a database study and hence institutional review board (IRB) exempt. Patients who underwent primary renal artery stenting (endovascular) were compared to those who underwent aorto-renal bypass (open). In order to compare best practices, we sought to compare the open approach with the highest overall patency (aorto-renal bypass) with the mainstream and most efficacious endovascular option (angioplasty þ stent). As we intended to compare the most common endovascular approach
Table 1. ICD-9 diagnosis and procedure codes used to select patients from the NIS. ICD-9 code Diagnosis codes 440.1 403.00 Procedure codes 39.24 39.90, 00.45, 00.46, 00.47, 00.48 88.45 Exclusion criteria 403.01 404.01, 404.02, 404.03 442.1 443.23 593.81 747.62
Description Renal artery atherosclerosis Renal hypertension without failure Aortorenal bypass Renal artery stent Renal angiography Renal failure Renal failure with congestive heart failure Renal artery aneurysm Renal artery dissection Fibromuscular dysplasia Fibromuscular dysplasia
ICD-9: International Classification Revision; NIS: National Inpatient Sample.
of
Diseases,
Ninth
to the most common open technique for this pathology, all other variations of RAS management were excluded from this analysis. Patient who underwent open procedures after endovascular failure were excluded. Both the open and endovascular groups included only patients with clinically significant atherosclerotic RAS that led to medically refractory hypertension. Patients who had progressed to the point of renal failure were excluded. Indirect costs are the actual cost of care incurred by the payer or insurance as a result of clinical management. An example of an indirect cost includes the cost of dialysis due to renal failure. Indirect costs were obtained from the National Institute of Diabetes and Digestive and Kidney Diseases. The annual cost of hemodialysis in 2014 USD is $91,455.13 Value is the amount an individual is willing to pay for a reduction in mortality as a result of clinical management. A surrogate for this value is the value of statistical life (VSL). The VSL is based on three meta-analyses completed by the Environmental Protection Agency using wage data and workplace risk data from the Bureau of Labor Statistics.14 Similar to a QALY analysis, the VSL is used by the US government when determining public health policy change, implementing new safety measures, and deciding whether to fund a new Medicare / Medicaid measure.15 The VSL in 2014 USD is $50,462. Population estimates were made using discharge sampling weights from the NIS and the US Census
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
3
Bureau. All monetary calculations are reported in 2014 USD adjusted using the Consumer Price Index. A US Treasury 3% discount rate was used for future value calculations. An average lifespan of 78.6 years was obtained from the US Census Bureau and used for VSL calculations.
Patient selection and cost calculation Patients were identified using ICD-9 diagnosis and procedure codes for atherosclerotic RAS (ICD-9 440.1 or 403.00) from the 1998 through 2011 (the most recent year available) NIS. The exclusion criteria for this study resulted in the inclusion of patients who had clinically-significant atherosclerotic renal artery disease with hypertension refractory to medical management. Patients who had concomitant renal failure, renal artery aneurysm, renal artery dissection, fibromuscular dysplasia, had risk factors for dialysis (i.e. chronic kidney disease (CKD)), diseases that led to renal disease (congestive heart failure (CHF)) or who underwent an intervention as part of another operation were excluded. Patient with CKD were excluded due to the higher likelihood of treatment failure in this group. Finally, any patients who crossed over from endovascular to open surgery were excluded from the study with the exception of an initial cost calculation used for the decision analysis. Patients were separated into those who underwent aortorenal bypass or renal artery stent (Table 1). Patient covariates included demographics, number of chronic conditions, diagnosis-related group (DRG) mortality and severity of illness scores, length of stay (LOS), incidence of complications, cost, and inpatient mortality (Table 2). The DRG mortality score and severity of illness score are validated methods of assigning values from 0 to 4 and estimates the risk of inpatient mortality or loss of function.16 Outpatient
renal artery stenting was included in this study as part of the sensitivity analysis by assuming an inpatient cost of $0 and utilizing the best-case estimates for morbidity and mortality due to the lower presumed risk of these patients. Complications were identified through the use of specific ICD-9 codes for surgical site infection (SSI), bleeding / hematoma formation, pulmonary embolism (PE), myocardial infarction (MI), cerebrovascular accident (CVA), limb occlusion, renal artery complication (such as dissection or rupture), acute kidney injury (AKI), permanent dialysis, nephrectomy, pseudoaneurysm, conversion to open (for renal artery stent patients), and secondary procedures including redo stent placement. Costs were determined for patients who underwent each procedure and for each subgroup of patients who experienced a complication. Costs were calculated using cost-to-charge ratios provided by the NIS.
Statistics Statistical analysis was completed using analysis of variance (ANOVA) for continuous variables (i.e. age and costs). The Mann-Whitney U-test was used for comparing total costs and overall value. Data analysis and management were completed with the IBM SPSS software package (SPSS version 22.0; SPSS Inc., Chicago, IL). Statistical significance was set at a level of p < 0.05. Values are presented as mean standard deviation or as median with interquartile range (IQR) in parentheses. LOS is presented a median standard deviation. The 95% confidence intervals were determined for all cost and value calculations. Individual cost calculations were made for patients who experienced complications and weighted samples used to determine an overall cost for this node (Table 3).
Table 2. Demographics and hospital covariates for patients who underwent aortorenal bypass compared to renal artery stenting. Variable
Aortorenal bypass
Renal artery stent
Significance
Sample size Age Female gender Caucasian African American Number of chronic conditions DRG risk of mortality score DRG severity of illness score Length of stay
482 62.3 15.5 54% 86.6% 4.1% 6.2 3.1 2.2 1.1 3.0 0.6 7.0 7.1
50,419 69.3 9.9 55% 83.6% 7.2% 6.8 2.6 1.9 0.9 2.3 0.7 2.0 4.8
p < 0.0001 NS NS NS NS p < 0.05 p < 0.001 p < 0.001
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
4
Vascular 0(0)
Table 3. Cost and overall incidence of complications for patients who underwent either aortorenal bypass or renal artery stentinga. Aortorenal bypass
Renal artery stent
Complication
Incidence
Direct cost
Incidence
Direct cost
Surgical site infection Bleeding/hematoma Pulmonary embolism Myocardial infarction Cerebrovascular accident Limb occlusion Renal artery complication Acute kidney injury Permanent dialysis (inpatient) Nephrectomy Pseudoaneurysm Conversion to open Secondary procedures/redo surgery Total
0.70% 3.03% 0.54% 2.15% 1.08% 0.54% 1.39% 0.54% 1.00% 1.40% N/A N/A 10.26% 23.48%
$44,813 $32,001 $42,227 $20,432 $35,356 $50,377 $46,405 $50,960 $63,022 $24,186 $5,628 $38,552 $22,236 $44,073 $45,760 $44,073 $45,760 $54,113 $72,084 N/A N/A $19,407 $15,407 $21,116 $15,401
0.00% 2.90% 0.11% 6.50% 2.60% 1.90% 0.93% 15.70% 1.90% 0.04% 1.10% 0.01% 13.69% 47.38%
$57,237 $54,470 $25,593 $21,451 $49,027 $45,273 $31,590 $22,709 $44,848 $41,449 $23,324 $16,842 $29,247 $24,840 $24,671 $18,254 $24,671 $18,254 $38,611 $19,936 $43,533 $26,175 $29,938 $25,142 $26,460 $22,623 $29,341 $25,131
a Costs are provided as median standard deviation in 2014 USD as obtained from the NIS. The range of percentages used for incidence of complications is provided in parenthesis with the data sources in superscript. Indirect costs for permanent dialysis following aortorenal bypass is $1,332,060, and for renal artery stent is $849,824.
Table 4. Data sources and the relative level of evidence utilized for the decision analysis. Variable
Data source
Level of evidence
Direct costs
National Inpatient Sample using cost/charge ratios and charges adjusted for 2014 USD National Institute of Diabetes and Digestive and Kidney Diseases, adjusted for 2014 USD Department of Transportation and Environmental Protection Agency, adjusted for 2014 USD CORAL, ASTRAL, STAR trials Meta-analysis of the literature National Inpatient Sample Reference values published in Rutherford’s textbook Consumer price index, Bureau of Labor Statistics
IIA
Indirect costs Value of statistical life Risk/benefit ratios Complication rates Outcomes data Inflation
Decision analysis A decision analysis was completed using Excel OM/ QM (Excel OM/QM version 4.2.012; Pearson Inc., New York, NY). The objective was to identify the tree that maximized savings while minimizing cost. For the decision analysis, the number given in the top left is the overall costs. The overall value is listed in the top right. Direct and indirect costs are in the bottom left, while costs and VSL are in the bottom right. Negative numbers are in red. The rate of the event is given as a percentage on the left side (Tables 4 and 5, Figure 1a and b). The initial decision was the procedure (aortorenal bypass vs. renal artery stent) for patients with a diagnosis of atherosclerotic RAS. The next two
IIA IA IB IA-IIB IIA-IIC IA-IV IIC
branches evaluated inpatient outcomes. The first bifurcation was for inpatient mortality. Patients who survived were then split into patients who experienced no complications and those who experienced one or more complications. The final branches were for five-year outcomes and included bypass/stent patency and whether the patient initiated hemodialysis or died. A composite end point of hemodialysis and death was used as the cost associated with each event was similar, and the VSL lost from either event was similar.17
Sensitivity analysis Variability in costs and value for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
5
Table 5. Types of costs. Type of cost
Description
Example
Direct cost
Actual cost of inpatient care incurred by payer/insurance
Indirect cost Value
Post-discharge cost of care incurred by payer/insurance Amount an individual is willing to pay to reduce mortality risk
Cost of procedure, inpatient hospitalization, and any complications Cost of dialysis due to renal failure Value of statistical life
Figure 1. a and b: Decision analysis tree for patients with atherosclerotic RAS who were treated with either aortorenal bypass or renal artery stent. Thirty-day outcomes and five-year outcomes were determined for each group. Thirty-day outcomes included inpatient mortality and whether the patient experienced one of 13 different complications. Five-year outcomes included overall patency of the bypass or stent, and whether the patient required hemodialysis or died. For each node of the decision analysis tree, costs, value, and rate of the event were calculated. The row above the node description includes overall cost or costs plus value as calculated at that point within the tree. The row below the node description includes direct and indirect cost or costs plus value for that particular event. The number to the left of the node description is the rate of the actual event.
of complications, five-year patency, five-year mortality, five-year dialysis-free survival, and life expectancy was accounted for by using known values published within the literature and normalized around the mean. Normal distributions for aortorenal bypass and renal artery stent placement taking into cost (Figure 2a) and costþ value (Figure 2b) were generated through a variancebased sensitivity analysis using reference values within
the literature and standard deviations obtained from the NIS.
Results There were 482 patients who underwent aortorenal bypass for atherosclerotic RAS between 1998 and 2011. Over the same time period, 50,419 patients
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
6
Vascular 0(0)
Figure 2. a. Normal distribution plot generated from a variance-based sensitivity analysis of direct costs and indirect costs for renal artery stent (left) and aortorenal bypass (right). This plot takes into account the variation in costs for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate of complications, five-year patency, five-year mortality, the rate of hemodialysis, and life expectancy. The horizontal axis is presented as 2014 USD with negative values to the left of the vertical axis (N ¼ sample size as based on the NIS). There is minimal overlap between the 95% confidence intervals (p < 0.001). b. Normal distribution plot generated from a variance-based sensitivity analysis of direct costs, indirect costs, and value for renal artery stent (left) and aortorenal bypass (right). This plot takes into account the variation in costs and value for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate of complications, five-year patency, five-year mortality, the rate of hemodialysis, and life expectancy. The horizontal axis is presented as 2014 USD with negative values to the left of the vertical axis (N ¼ sample size as based on the NIS). There is no significant overlap between the 95% confidence intervals (p < 0.0001).
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
7
underwent renal artery stenting. Patients who underwent aortorenal bypass were younger than those who underwent stent placement (62.3 15.5 vs. 69.3 9.9, p < 0.0001), but were otherwise similar with regard to gender, race, and number of chronic conditions. Aortorenal bypass patients had a higher DRG mortality score (2.2 1.1 vs. 1.9 0.9, p < 0.05) and severity of illness score (3.0 0.6 vs. 2.3 0.7, p < 0.001) compared to renal artery stent patients. As expected, the median LOS was greater for aortorenal bypass compared to renal artery stenting (7.0 7.1 vs. 2.0 4.8, p < 0.001) (Table 2). Patients who underwent aortorenal bypass had 2.5% inpatient mortality, compared to 0.6% for renal artery stenting (p < 0.01) (Figure 1). Of the aortorenal bypass patients who survived, 23.5% experienced one or more complications, compared to 47.4% for renal artery stent patients. The most significant complications with regard to incidence and cost were determined, including SSI, bleeding/hematoma formation, PE, MI, CVA, limb occlusion, renal artery complication, AKI, permanent dialysis, nephrectomy, pseudoaneurysm, conversion to open procedure, and need for secondary procedures or redo surgery. The total direct cost for these complications was $21,116 $15,401 for aortorenal bypass patients and $29,341 $25,131 for renal artery stent patients (p < 0.001) (Table 3). A subgroup analysis completed for patients with asymptomatic lesions (ICD-9 440.1), renovascular hypertension (403.00) revealed no significantly different findings with regard to relative incidence, number or type of complications, and overall costs or value, thereby permitting these two groups to be pooled. Values for five-year outcomes were determined from a review of the literature. Five-year patency following aortorenal bypass is 95.0%, compared to 55.6% for renal artery stenting.7,10,18–20 The five-year patency was assumed to be similar between patients who had no complications and those who experienced one or more complications. A primary patency of 75% was identified in the literature and used for aortorenal bypass patients who developed complications to determine whether this significantly influenced cost or value for aortorenal bypass compared to renal artery stenting, but none was found (Figure 2).18,19 Similarly, a maximum five-year secondary patency of 87% was used for renal artery stent patients, but no significant difference in cost or value was found.19 A composite endpoint of dialysis and death from all causes was used and as a basis of comparison to patients who were alive and dialysis-free at five years. Individual analysis of the costs of dialysis and all-cause mortality did not introduce significant variation in the results (Figure 2). Of those patients who undergo aortorenal bypass and have a preoperative estimated
glomerular filtration rate (EGFR) of 25 mL/min/m2, 70% are alive and dialysis-free at five years. Of those who have an EGFR of 39 mL/min/m2 or greater, between 80% and 95% are alive and dialysis-free at five years.12,21 For the purposes of this decision analysis, the minimum value was used as the preoperative EGFR of all of the patients in this study could not be ascertained. Of those patients who had a failed aortorenal bypass, only 18% are dialysis-free at five years.10 Patients who undergo optimal management through a combination of renal artery stent placement and best medical management have a 64.9% five-year dialysis-free survival.5,22,23 Specific data on five-year dialysis-free survival are not available for patients who fail renal artery stenting, and values for aortorenal bypass were substituted. While greater variation in the cost/cost þ value calculations are introduced for renal artery stenting if an assumption of 64.9% dialysis-free survival is made even for failed renal artery stenting, there is no overlap within the confidence intervals of the normal distribution for both cost and cost þ value (Figure 2). In order to test the validity of the decision tree, variability within the costs and costs þ value for procedure, inpatient stay, and complications was introduced using standard deviations from the NIS. Variability in inpatient mortality, incidence of complications, five-year patency, five-year mortality, and five-year dialysis-free survival was introduced using reference values from the literature.5,7,10,12,18–27 Variation in life expectancy was introduced using reference values in the literature and the US Census Bureau.10,19,21 Normal distribution plots were generated from a variance-based sensitivity analysis of these factors for cost (Figure 2a) and cost þ value (Figure 2b). The results of the decision analysis indicate that the cost for aortorenal bypass is $103,453 $596,782 (N ¼ 482, 95% CI $156,864 to $50,042). The cost for renal artery stent is $305,370 $300,519 (N ¼ 50,419, 95% CI $307,993 to $302,747). When including the VSL, aortorenal bypass has an overall value of $415,881 $760,021 (N ¼ 482, 95% CI $347,860– $483,902). Renal artery stenting has an overall negative value of $182,270 $710,304 (N ¼ 50,419, 95% CI $188,470 to $176,070). Both cost and cost þ value for aortorenal bypass compared to renal artery stenting are significantly different (p < 0.0001).
Discussion Several trials have since been conducted on comparing angioplasty, angioplasty and stent placement, and best medical management.5,28–33 Each of the earlier trials had significant limitations in patient selection, degree of RAS, utilization of a stent and/or distal embolic
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
8
Vascular 0(0)
protection device, optimal medical management, and sufficient power to determine a difference in the groups. These issues were corrected with the CORAL trial, which demonstrated equivalency between optimal medical management and stent placement plus optimal medical management.5 While multiple trials have been completed on percutaneous renal artery intervention, there is only one study comparing percutaneous renal intervention with surgical revascularization.34 Weibull et al. compared percutaneous transluminal renal angioplasty with surgical reconstruction of atherosclerotic RAS in a prospective randomized study in 1993.34 At that time, primary patency following angioplasty was 75% at two years compared to 96% for bypass in a group of 58 patients, with 17% of patients who underwent angioplasty requiring surgical intervention. In addition to the higher re-intervention rate following angioplasty, other limitations of this study was that long-term dialysis-free survival was not evaluated, exclusion criteria that limited the applicability of the results, crossover from the angioplasty cohort to the surgical management group, and that stents were not used in conjunction with angioplasty. Apart from this limited study more than two decades ago, there has not been a suitably controlled trial evaluating surgical revascularization with percutaneous renal intervention or optimal medical management for atherosclerotic RAS. It is possible to compare the costs and relative value of aortorenal bypass with renal artery stenting. Using level IIA or greater data from the literature, reference values for efficacy, patency, and survival for each of the procedures can be determined. Cost data can be obtained from the NIS, along with any missing values regarding the incidence of complications and inpatient survival. While individual studies have evaluated longterm dialysis-free survival, no study has thus far placed these long-term results in the context of complications and overall efficacy using a cost analysis. Aortorenal bypass has an overall cost of $103,453 compared to $305,370 for renal artery stenting. While the initial cost of the procedure is nearly $10,000 less for renal artery stenting ($29,938 for aortorenal bypass compared to $19,407 for renal artery stenting), escalating costs are incurred from the significantly lower fiveyear patency rate for renal artery stenting (Figure 1). The annual cost of dialysis is $91,455 and the proportion of patients who require dialysis at five years is greater for renal artery stent patients than it is for aortorenal bypass patients. This leads to a nearly threefold increase in costs associated with renal artery stenting, even though these patients tend to be older (69.3 vs. 62.3 years) and less ill (Table 2). Furthermore, while most individual costs associated with complications from renal artery stenting are less than that for aortorenal bypass, the overall cost is nearly $8000 greater
($29,341 vs. $21,116) due to the increased incidence of individual complications (Table 3). The value of these interventions was determined using a VSL analysis. The VSL was assigned to patients who survived their procedure and had a five-year dialysis-free survival. Limited VSL is expected for patients on dialysis due to the substantially poorer quality of life, but even assigning a 50% rate of VSL to these patients as part of a sensitivity analysis yields no significant overlap in the normal distribution for costþ value (Figure 2b). When incorporating cost þ value to assist in decision-making, aortorenal bypass has an overall value of $415,881, compared to a negative value of $182,270 for renal artery stenting. Overall, there is a marginal improvement of $519,334 for aortorenal bypass patients and $123,100 for renal artery stent patients when incorporating VSL analysis. If a future value calculation is completed with the $519,334 improvement, aortorenal bypass is expected to lead to an additional 10 years of dialysis-free survival compared to just 2.4 years for patients who undergo renal artery stenting, similar to what has been previously reported.5,10,12,19,21,23
Limitations The use of an inpatient database to obtain cost information, incidence of complications, and inpatient mortality relies upon the accuracy of the data entry. Multiple assumptions are made with regard to patency of each of the approaches and five-year outcomes. While reference values are used from level 2 A or better sources within the literature, defaulting only to the NIS when these values are not available, there is some bias that is introduced into the analysis. To compensate for this bias, a variance-based sensitivity analysis was completed using maximum variability in all of the input data. Normal distributions were created based off this analysis, and there is very little overlap between these normal distributions. This indicates a high level of confidence with the data (p < 0.0001). The influence of individual or hospital volume was not included within this analysis and may be a source of bias. Another limitation of the database utilized is that there is no data on the type of renal lesions in these patients and how the lesion characteristics impacted management choices; furthermore we do not have any data on long term follow-up in these patients given the limitations of the NIS.
Conclusion Patients who fail best medical management for atherosclerotic RAS and who are candidates for open
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
9
revascularization should undergo aortorenal bypass in lieu of renal artery stenting. Acknowledgments Dr Sapan S Desai, the principal investigator, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. There are no conflicts of interest or other pertinent disclosures. The dialysis cost data reported here have been supplied by the United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy or interpretation of the US government.
Conflict of interest No conflicts of interest.
Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References 1. Appel RG, Bleyer AJ, Reavis S, et al. Renovascular disease in older patients beginning renal replacement therapy. Kidney Int 1995; 48: 171–176. 2. Davis RP, Pearce JD, Craven TE, et al. Atherosclerotic renovascular disease among hypertensive adults. J Vasc Surg 2009; 50: 564–570. (571 e1-3; discussion 571). 3. Dean RH, Kieffer RW, Smith BM, et al. Renovascular hypertension: anatomic and renal function changes during drug therapy. Arch Surg 1981; 116: 1408–1415. 4. Schreiber MJ, Pohl MA and Novick AC. The natural history of atherosclerotic and fibrous renal artery disease. Urol Clin North Am 1984; 11: 383–392. 5. Cooper CJ, Murphy TP, Cutlip DE, et al. Stenting and medical therapy for atherosclerotic renal-artery stenosis. N Engl J Med 2014; 370: 13–22. 6. Corriere MA, Pearce JD, Edwards MS, et al. Endovascular management of atherosclerotic renovascular disease: early results following primary intervention. J Vasc Surg 2008; 48: 580–587; discussion 587–588. 7. Kashyap VS, Sepulveda RN, Bena JF, et al. The management of renal artery atherosclerosis for renal salvage: does stenting help? J Vasc Surg 2007; 45: 101–108. (discussion 108–109). 8. Zeller T, Frank U, Mu¨ller C, et al. Predictors of improved renal function after percutaneous stent-supported angioplasty of severe atherosclerotic ostial renal artery stenosis. Circulation 2003; 108: 2244–2249. 9. Zeller T, Frank U, Mu¨ller C, et al. Stent-supported angioplasty of severe atherosclerotic renal artery stenosis preserves renal function and improves blood pressure control: long-term results from a prospective registry of 456 lesions. J Endovasc Ther 2004; 11: 95–106. 10. Hansen KJ, Deitch JS, Oskin TC, et al. Renal artery repair: consequence of operative failures. Ann Surg 1998; 227: 678–689. (discussion 689–690).
11. Hansen KJ, Cherr GS, Craven TE, et al. Management of ischemic nephropathy: dialysis-free survival after surgical repair. J Vasc Surg 2000; 32: 472–481. (discussion 481– 482). 12. Cherr GS, Hansen KJ, Craven TE, et al. Surgical management of atherosclerotic renovascular disease. J Vasc Surg 2002; 35: 236–245. 13. System, U.S.R.D., USRDS 2013 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. United States Renal Data System, 2014 Annual Data Report: Epidemiology of Kidney Disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2014. 14. Kip Viscusi W and Aldy JE. The value of a statistical life: a critical review of market estimates throughout the world. J Risk Uncertainty 2003; 27: 5–76. 15. Polly Trottenberg, RSR. Guidance on Treatment of the Economic Value of a Statistical Life (VSL) in U.S. Department of Transportation Analyses. Washington DC: Department of Motor Vehicles, 2013. 16. Baram D, Daroowalla F, Garcia R, et al. Use of the All Patient Refined-Diagnosis Related Group (APR-DRG) Risk of Mortality Score as a severity adjustor in the medical ICU. Clin Med Circ Respir Pulm Med 2008; 2: 19–25. 17. Lee CP, Chertow GM and Zenios SA. An empiric estimate of the value of life: updating the renal dialysis costeffectiveness standard. Value Health 2009; 12: 80–87. 18. Feng R, Wei XL, Zhao ZQ, et al. Aortorenal bypass with autologous saphenous vein in Takayasu arteritis-induced renal artery stenosis. Eur J Vasc Endovasc Surg 2011; 42: 47–53. 19. Ham SW, Kumar SR, Wang BR, et al. Late outcomes of endovascular and open revascularization for nonatherosclerotic renal artery disease. Arch Surg 2010; 145: 832–839. 20. Kennedy DJ, Colyer WR, Brewster PS, et al. Renal insufficiency as a predictor of adverse events and mortality after renal artery stent placement. Am J Kidney Dis 2003; 42: 926–935. 21. Hansen KJ, Cherr GS and Dean RH. Dialysis-free survival after surgical repair of ischemic nephropathy. Cardiovasc Surg 2002; 10: 400–4. 22. Arthurs Z, Starnes B, Cuadrado D, et al. Renal artery stenting slows the rate of renal function decline. J Vasc Surg 2007; 45: 726–731. (discussion 731–732). 23. Bates MC, Campbell JE, Stone PA, et al. Factors affecting long-term survival following renal artery stenting. Catheter Cardiovasc Interv 2007; 69: 1037–1043. 24. Hagino RT, Valentine RJ and Clagett GP. Supraceliac aortorenal bypass. J Vasc Surg 1997; 26: 482–489. (discussion 490–491). 25. Hoshino Y, Nakamura T, Nakano A, et al. Successful treatment of renovascular hypertension due to fibromuscular dysplasia by intravascular ultrasound-guided atherectomy. Nephron 2002; 91: 521–525. 26. Kimura H, Sato O, Deguchi JO, et al. Surgical treatment and long-term outcome of renovascular hypertension in children and adolescents. Eur J Vasc Endovasc Surg 2010; 39: 731–737.
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
10
Vascular 0(0)
27. Schweiger H, Raithel D, Seyferth W, et al. Surgical treatment of renal artery occlusive disease: long term results. J Cardiovasc Surg (Torino) 1984; 25: 111–114. 28. van Jaarsveld BC, Derkx FH, Krijnen P, et al. Hypertension resistant to two-drug treatment’ is a useful criterion to select patients for angiography: the ‘Dutch Renal Artery Stenosis Intervention Cooperative’ (DRASTIC) study. Contrib Nephrol 1996; 119: 54–58. 29. Plouin PF, Chatellier G, Darne´ B, et al. Blood pressure outcome of angioplasty in atherosclerotic renal artery stenosis: a randomized trial. Essai Multicentrique Medicaments vs Angioplastie (EMMA) Study Group. Hypertension 1998; 31: 823–829. 30. Webster J, Marshall F, Abdalla M, et al. Randomised comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous renal artery stenosis. Scottish and Newcastle Renal Artery Stenosis Collaborative Group. J Hum Hypertens 1998; 12: 329–335.
31. Ives NJ, Wheatley K, Stowe RL, et al. Continuing uncertainty about the value of percutaneous revascularization in atherosclerotic renovascular disease: a meta-analysis of randomized trials. Nephrol Dial Transplant 2003; 18: 298–304. 32. Tobe SW, Atri M, Perkins N, et al. Renal athersosclerotic revascularization evaluation (RAVE study): study protocol of a randomized trial [NCT00127738]. BMC Nephrol 2007; 8: 4. 33. ASTRAL Investigators, Wheatley K, Ives N, et al. Revascularization versus medical therapy for renalartery stenosis. N Engl J Med 2009; 361: 1953–1962. 34. Weibull H, Bergqvist D, Bergentz SE, et al. Percutaneous transluminal renal angioplasty versus surgical reconstruction of atherosclerotic renal artery stenosis: a prospective randomized study. J Vasc Surg 1993; 18: 841–850. (discussion 850–852).
Downloaded from vas.sagepub.com by guest on November 15, 2015
(VAS)
[1–10] [PREPRINTER stage]
Vascular OnlineFirst, published on April 28, 2015 as doi:10.1177/1708538115584507
Original Article
Renal artery stenting may not be appropriate for patients with atherosclerotic renal artery stenosis
Vascular 0(0) 1–10 ! The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1708538115584507 vas.sagepub.com
Anahita Dua1, Sachin Desai2, Gilbert R Upchurch Jr3 and Sapan S Desai4
Abstract Introduction: This study compared aortorenal bypass to renal artery stenting to determine the most efficacious and financially sound method for treating patients with atherosclerotic renal artery stenosis (RAS). Methods: A decision analysis using direct and indirect costs, and value of statistical life (VSL) was completed. Direct costs were obtained using the Nationwide Inpatient Sample (NIS), indirect costs from the National Institute of Diabetes and Digestive and Kidney Diseases, and VSL from the Department of Transportation. A variance-based sensitivity analysis was completed to assess the accuracy of the decision analysis. Results: Aortorenal bypass has a 95% five-year patency, a 98% 30-day survival, a 26% rate of overall complications, and a 70% five-year dialysis-free survival. Renal artery stenting has a 56% five-year patency, a 99% 30-day survival, a 40% rate of complications, and a 65% five-year dialysis-free survival. Renal artery stenting has an overall cost of $305,370 and aortorenal bypass has an overall cost of $103,453 per patient. After accounting for VSL, renal artery stenting has a negative value of $182,270 and aortorenal bypass has a value of $415,881. Conclusions: Lower five-year patency and higher rate of complications from renal artery stenting that ultimately lead to significantly lower five-year dialysis-free survival.
Keywords Renal artery stenosis, renal artery stenting
Introduction Renal artery stenosis (RAS) affects up to 32% of patients with hypertension, and 22% of patients with end stage renal disease.1,2 Approximately 10% of lesions with over 60% stenosis will progress to occlusion at three years.3,4 Proper management of RAS may contribute to improving two major public health problems: hypertension and kidney disease. Medical management, endovascular procedures (including angioplasty and stenting), and open surgery are used to treat patients with RAS. Angioplasty with balloon-expandable stent is the preferred therapeutic option for RAS. This approach has a 56% five-year patency with a 17% restenosis rate. Morbidity varies between 0% and 43%, and mortality between 0% and 5%. Patients have a 63% three-year dialysis-free survival, however up to 25% of patients experience a decline in renal function with negligible long-term improvement in hypertension.5–9
Open surgical options include aortorenal bypass, iliorenal bypass, hepatorenal bypass, and splenorenal bypass. Other options include endarterectomy with patch angioplasty, and reimplantation of the renal artery. Of these modalities, aortorenal bypass has the highest five-year patency at 95%. Morbidity varies between 7% and 24%, and mortality between 1% and 3% in recent series. Patients have a 70% fiveyear dialysis-free survival, and 85% have full resolution of their hypertension at that time.10–12 1 Department of Surgery, Medical College of Wisconsin, Milwaukee, WI, USA 2 Private Industry, Chicago, IL, USA 3 Department of Surgery, University of Virginia, Charlottesville, VA, USA 4 Department of Vascular Surgery, Southern Illinois University, Springfield, IL, USA
Corresponding author: Anahita Dua, 8701 W Watertown Plank Rd, Milwaukee, WI 53226, USA. Email: [email protected]
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
2
Vascular 0(0)
The optimal management of atherosclerotic RAS is controversial. The cardiovascular outcomes in renal atherosclerotic lesions (CORAL) trial demonstrates equivalency in outcome between renal artery stenting and medical management, and that both of these are poor options for treating RAS.5 While aortorenal bypass appears to be a viable alternative to both renal artery stenting and optimal medical management in patients with atherosclerotic RAS, the relative efficacy of aortorenal bypass has not been evaluated. Such a clinical trial would require many years to complete and would have issues in obtaining sufficient power; the CORAL trial enrolled 947 patients and had to be extended by an additional year due to similar concerns. Due to the limitations in completing a head-to-head clinical trial, we sought to compare the outcomes and relative value of aortorenal bypass with renal artery stenting using a decision analysis.
Methods Data sources A decision analysis was completed using direct costs, indirect costs, and overall value. Direct costs include the actual cost of care incurred by the payer or insurance at the time of care. Examples of direct costs include the cost of renal artery stenting or aortorenal bypass, and the cost of the inpatient hospitalization along with any inpatient complications. Direct costs were obtained from the Nationwide Inpatient Sample (NIS). The NIS is a part of the Health Care Utilization Project (HCUP) that is maintained by the Agency for Healthcare Research and Quality (AHRQ). It is the largest all-payer inpatient database and includes a stratified 20% random sample of all nonfederal inpatient hospital admissions throughout the United States. Clinical records were obtained with the use of International Classification of Diseases, Ninth Revision (ICD-9) diagnosis and procedure codes to ensure that the sample included only patients who had a primary diagnosis of RAS and underwent either aortorenal bypass or renal artery stenting. Renal angiography was used as an adjunct procedure code in order to more accurately select patients who underwent stenting (Table 1). No patient consent was required as this was a database study and hence institutional review board (IRB) exempt. Patients who underwent primary renal artery stenting (endovascular) were compared to those who underwent aorto-renal bypass (open). In order to compare best practices, we sought to compare the open approach with the highest overall patency (aorto-renal bypass) with the mainstream and most efficacious endovascular option (angioplasty þ stent). As we intended to compare the most common endovascular approach
Table 1. ICD-9 diagnosis and procedure codes used to select patients from the NIS. ICD-9 code Diagnosis codes 440.1 403.00 Procedure codes 39.24 39.90, 00.45, 00.46, 00.47, 00.48 88.45 Exclusion criteria 403.01 404.01, 404.02, 404.03 442.1 443.23 593.81 747.62
Description Renal artery atherosclerosis Renal hypertension without failure Aortorenal bypass Renal artery stent Renal angiography Renal failure Renal failure with congestive heart failure Renal artery aneurysm Renal artery dissection Fibromuscular dysplasia Fibromuscular dysplasia
ICD-9: International Classification Revision; NIS: National Inpatient Sample.
of
Diseases,
Ninth
to the most common open technique for this pathology, all other variations of RAS management were excluded from this analysis. Patient who underwent open procedures after endovascular failure were excluded. Both the open and endovascular groups included only patients with clinically significant atherosclerotic RAS that led to medically refractory hypertension. Patients who had progressed to the point of renal failure were excluded. Indirect costs are the actual cost of care incurred by the payer or insurance as a result of clinical management. An example of an indirect cost includes the cost of dialysis due to renal failure. Indirect costs were obtained from the National Institute of Diabetes and Digestive and Kidney Diseases. The annual cost of hemodialysis in 2014 USD is $91,455.13 Value is the amount an individual is willing to pay for a reduction in mortality as a result of clinical management. A surrogate for this value is the value of statistical life (VSL). The VSL is based on three meta-analyses completed by the Environmental Protection Agency using wage data and workplace risk data from the Bureau of Labor Statistics.14 Similar to a QALY analysis, the VSL is used by the US government when determining public health policy change, implementing new safety measures, and deciding whether to fund a new Medicare / Medicaid measure.15 The VSL in 2014 USD is $50,462. Population estimates were made using discharge sampling weights from the NIS and the US Census
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
3
Bureau. All monetary calculations are reported in 2014 USD adjusted using the Consumer Price Index. A US Treasury 3% discount rate was used for future value calculations. An average lifespan of 78.6 years was obtained from the US Census Bureau and used for VSL calculations.
Patient selection and cost calculation Patients were identified using ICD-9 diagnosis and procedure codes for atherosclerotic RAS (ICD-9 440.1 or 403.00) from the 1998 through 2011 (the most recent year available) NIS. The exclusion criteria for this study resulted in the inclusion of patients who had clinically-significant atherosclerotic renal artery disease with hypertension refractory to medical management. Patients who had concomitant renal failure, renal artery aneurysm, renal artery dissection, fibromuscular dysplasia, had risk factors for dialysis (i.e. chronic kidney disease (CKD)), diseases that led to renal disease (congestive heart failure (CHF)) or who underwent an intervention as part of another operation were excluded. Patient with CKD were excluded due to the higher likelihood of treatment failure in this group. Finally, any patients who crossed over from endovascular to open surgery were excluded from the study with the exception of an initial cost calculation used for the decision analysis. Patients were separated into those who underwent aortorenal bypass or renal artery stent (Table 1). Patient covariates included demographics, number of chronic conditions, diagnosis-related group (DRG) mortality and severity of illness scores, length of stay (LOS), incidence of complications, cost, and inpatient mortality (Table 2). The DRG mortality score and severity of illness score are validated methods of assigning values from 0 to 4 and estimates the risk of inpatient mortality or loss of function.16 Outpatient
renal artery stenting was included in this study as part of the sensitivity analysis by assuming an inpatient cost of $0 and utilizing the best-case estimates for morbidity and mortality due to the lower presumed risk of these patients. Complications were identified through the use of specific ICD-9 codes for surgical site infection (SSI), bleeding / hematoma formation, pulmonary embolism (PE), myocardial infarction (MI), cerebrovascular accident (CVA), limb occlusion, renal artery complication (such as dissection or rupture), acute kidney injury (AKI), permanent dialysis, nephrectomy, pseudoaneurysm, conversion to open (for renal artery stent patients), and secondary procedures including redo stent placement. Costs were determined for patients who underwent each procedure and for each subgroup of patients who experienced a complication. Costs were calculated using cost-to-charge ratios provided by the NIS.
Statistics Statistical analysis was completed using analysis of variance (ANOVA) for continuous variables (i.e. age and costs). The Mann-Whitney U-test was used for comparing total costs and overall value. Data analysis and management were completed with the IBM SPSS software package (SPSS version 22.0; SPSS Inc., Chicago, IL). Statistical significance was set at a level of p < 0.05. Values are presented as mean standard deviation or as median with interquartile range (IQR) in parentheses. LOS is presented a median standard deviation. The 95% confidence intervals were determined for all cost and value calculations. Individual cost calculations were made for patients who experienced complications and weighted samples used to determine an overall cost for this node (Table 3).
Table 2. Demographics and hospital covariates for patients who underwent aortorenal bypass compared to renal artery stenting. Variable
Aortorenal bypass
Renal artery stent
Significance
Sample size Age Female gender Caucasian African American Number of chronic conditions DRG risk of mortality score DRG severity of illness score Length of stay
482 62.3 15.5 54% 86.6% 4.1% 6.2 3.1 2.2 1.1 3.0 0.6 7.0 7.1
50,419 69.3 9.9 55% 83.6% 7.2% 6.8 2.6 1.9 0.9 2.3 0.7 2.0 4.8
p < 0.0001 NS NS NS NS p < 0.05 p < 0.001 p < 0.001
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
4
Vascular 0(0)
Table 3. Cost and overall incidence of complications for patients who underwent either aortorenal bypass or renal artery stentinga. Aortorenal bypass
Renal artery stent
Complication
Incidence
Direct cost
Incidence
Direct cost
Surgical site infection Bleeding/hematoma Pulmonary embolism Myocardial infarction Cerebrovascular accident Limb occlusion Renal artery complication Acute kidney injury Permanent dialysis (inpatient) Nephrectomy Pseudoaneurysm Conversion to open Secondary procedures/redo surgery Total
0.70% 3.03% 0.54% 2.15% 1.08% 0.54% 1.39% 0.54% 1.00% 1.40% N/A N/A 10.26% 23.48%
$44,813 $32,001 $42,227 $20,432 $35,356 $50,377 $46,405 $50,960 $63,022 $24,186 $5,628 $38,552 $22,236 $44,073 $45,760 $44,073 $45,760 $54,113 $72,084 N/A N/A $19,407 $15,407 $21,116 $15,401
0.00% 2.90% 0.11% 6.50% 2.60% 1.90% 0.93% 15.70% 1.90% 0.04% 1.10% 0.01% 13.69% 47.38%
$57,237 $54,470 $25,593 $21,451 $49,027 $45,273 $31,590 $22,709 $44,848 $41,449 $23,324 $16,842 $29,247 $24,840 $24,671 $18,254 $24,671 $18,254 $38,611 $19,936 $43,533 $26,175 $29,938 $25,142 $26,460 $22,623 $29,341 $25,131
a Costs are provided as median standard deviation in 2014 USD as obtained from the NIS. The range of percentages used for incidence of complications is provided in parenthesis with the data sources in superscript. Indirect costs for permanent dialysis following aortorenal bypass is $1,332,060, and for renal artery stent is $849,824.
Table 4. Data sources and the relative level of evidence utilized for the decision analysis. Variable
Data source
Level of evidence
Direct costs
National Inpatient Sample using cost/charge ratios and charges adjusted for 2014 USD National Institute of Diabetes and Digestive and Kidney Diseases, adjusted for 2014 USD Department of Transportation and Environmental Protection Agency, adjusted for 2014 USD CORAL, ASTRAL, STAR trials Meta-analysis of the literature National Inpatient Sample Reference values published in Rutherford’s textbook Consumer price index, Bureau of Labor Statistics
IIA
Indirect costs Value of statistical life Risk/benefit ratios Complication rates Outcomes data Inflation
Decision analysis A decision analysis was completed using Excel OM/ QM (Excel OM/QM version 4.2.012; Pearson Inc., New York, NY). The objective was to identify the tree that maximized savings while minimizing cost. For the decision analysis, the number given in the top left is the overall costs. The overall value is listed in the top right. Direct and indirect costs are in the bottom left, while costs and VSL are in the bottom right. Negative numbers are in red. The rate of the event is given as a percentage on the left side (Tables 4 and 5, Figure 1a and b). The initial decision was the procedure (aortorenal bypass vs. renal artery stent) for patients with a diagnosis of atherosclerotic RAS. The next two
IIA IA IB IA-IIB IIA-IIC IA-IV IIC
branches evaluated inpatient outcomes. The first bifurcation was for inpatient mortality. Patients who survived were then split into patients who experienced no complications and those who experienced one or more complications. The final branches were for five-year outcomes and included bypass/stent patency and whether the patient initiated hemodialysis or died. A composite end point of hemodialysis and death was used as the cost associated with each event was similar, and the VSL lost from either event was similar.17
Sensitivity analysis Variability in costs and value for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
5
Table 5. Types of costs. Type of cost
Description
Example
Direct cost
Actual cost of inpatient care incurred by payer/insurance
Indirect cost Value
Post-discharge cost of care incurred by payer/insurance Amount an individual is willing to pay to reduce mortality risk
Cost of procedure, inpatient hospitalization, and any complications Cost of dialysis due to renal failure Value of statistical life
Figure 1. a and b: Decision analysis tree for patients with atherosclerotic RAS who were treated with either aortorenal bypass or renal artery stent. Thirty-day outcomes and five-year outcomes were determined for each group. Thirty-day outcomes included inpatient mortality and whether the patient experienced one of 13 different complications. Five-year outcomes included overall patency of the bypass or stent, and whether the patient required hemodialysis or died. For each node of the decision analysis tree, costs, value, and rate of the event were calculated. The row above the node description includes overall cost or costs plus value as calculated at that point within the tree. The row below the node description includes direct and indirect cost or costs plus value for that particular event. The number to the left of the node description is the rate of the actual event.
of complications, five-year patency, five-year mortality, five-year dialysis-free survival, and life expectancy was accounted for by using known values published within the literature and normalized around the mean. Normal distributions for aortorenal bypass and renal artery stent placement taking into cost (Figure 2a) and costþ value (Figure 2b) were generated through a variancebased sensitivity analysis using reference values within
the literature and standard deviations obtained from the NIS.
Results There were 482 patients who underwent aortorenal bypass for atherosclerotic RAS between 1998 and 2011. Over the same time period, 50,419 patients
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
6
Vascular 0(0)
Figure 2. a. Normal distribution plot generated from a variance-based sensitivity analysis of direct costs and indirect costs for renal artery stent (left) and aortorenal bypass (right). This plot takes into account the variation in costs for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate of complications, five-year patency, five-year mortality, the rate of hemodialysis, and life expectancy. The horizontal axis is presented as 2014 USD with negative values to the left of the vertical axis (N ¼ sample size as based on the NIS). There is minimal overlap between the 95% confidence intervals (p < 0.001). b. Normal distribution plot generated from a variance-based sensitivity analysis of direct costs, indirect costs, and value for renal artery stent (left) and aortorenal bypass (right). This plot takes into account the variation in costs and value for the procedure, cost of inpatient stay, rate of inpatient mortality, type and rate of complications, five-year patency, five-year mortality, the rate of hemodialysis, and life expectancy. The horizontal axis is presented as 2014 USD with negative values to the left of the vertical axis (N ¼ sample size as based on the NIS). There is no significant overlap between the 95% confidence intervals (p < 0.0001).
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
7
underwent renal artery stenting. Patients who underwent aortorenal bypass were younger than those who underwent stent placement (62.3 15.5 vs. 69.3 9.9, p < 0.0001), but were otherwise similar with regard to gender, race, and number of chronic conditions. Aortorenal bypass patients had a higher DRG mortality score (2.2 1.1 vs. 1.9 0.9, p < 0.05) and severity of illness score (3.0 0.6 vs. 2.3 0.7, p < 0.001) compared to renal artery stent patients. As expected, the median LOS was greater for aortorenal bypass compared to renal artery stenting (7.0 7.1 vs. 2.0 4.8, p < 0.001) (Table 2). Patients who underwent aortorenal bypass had 2.5% inpatient mortality, compared to 0.6% for renal artery stenting (p < 0.01) (Figure 1). Of the aortorenal bypass patients who survived, 23.5% experienced one or more complications, compared to 47.4% for renal artery stent patients. The most significant complications with regard to incidence and cost were determined, including SSI, bleeding/hematoma formation, PE, MI, CVA, limb occlusion, renal artery complication, AKI, permanent dialysis, nephrectomy, pseudoaneurysm, conversion to open procedure, and need for secondary procedures or redo surgery. The total direct cost for these complications was $21,116 $15,401 for aortorenal bypass patients and $29,341 $25,131 for renal artery stent patients (p < 0.001) (Table 3). A subgroup analysis completed for patients with asymptomatic lesions (ICD-9 440.1), renovascular hypertension (403.00) revealed no significantly different findings with regard to relative incidence, number or type of complications, and overall costs or value, thereby permitting these two groups to be pooled. Values for five-year outcomes were determined from a review of the literature. Five-year patency following aortorenal bypass is 95.0%, compared to 55.6% for renal artery stenting.7,10,18–20 The five-year patency was assumed to be similar between patients who had no complications and those who experienced one or more complications. A primary patency of 75% was identified in the literature and used for aortorenal bypass patients who developed complications to determine whether this significantly influenced cost or value for aortorenal bypass compared to renal artery stenting, but none was found (Figure 2).18,19 Similarly, a maximum five-year secondary patency of 87% was used for renal artery stent patients, but no significant difference in cost or value was found.19 A composite endpoint of dialysis and death from all causes was used and as a basis of comparison to patients who were alive and dialysis-free at five years. Individual analysis of the costs of dialysis and all-cause mortality did not introduce significant variation in the results (Figure 2). Of those patients who undergo aortorenal bypass and have a preoperative estimated
glomerular filtration rate (EGFR) of 25 mL/min/m2, 70% are alive and dialysis-free at five years. Of those who have an EGFR of 39 mL/min/m2 or greater, between 80% and 95% are alive and dialysis-free at five years.12,21 For the purposes of this decision analysis, the minimum value was used as the preoperative EGFR of all of the patients in this study could not be ascertained. Of those patients who had a failed aortorenal bypass, only 18% are dialysis-free at five years.10 Patients who undergo optimal management through a combination of renal artery stent placement and best medical management have a 64.9% five-year dialysis-free survival.5,22,23 Specific data on five-year dialysis-free survival are not available for patients who fail renal artery stenting, and values for aortorenal bypass were substituted. While greater variation in the cost/cost þ value calculations are introduced for renal artery stenting if an assumption of 64.9% dialysis-free survival is made even for failed renal artery stenting, there is no overlap within the confidence intervals of the normal distribution for both cost and cost þ value (Figure 2). In order to test the validity of the decision tree, variability within the costs and costs þ value for procedure, inpatient stay, and complications was introduced using standard deviations from the NIS. Variability in inpatient mortality, incidence of complications, five-year patency, five-year mortality, and five-year dialysis-free survival was introduced using reference values from the literature.5,7,10,12,18–27 Variation in life expectancy was introduced using reference values in the literature and the US Census Bureau.10,19,21 Normal distribution plots were generated from a variance-based sensitivity analysis of these factors for cost (Figure 2a) and cost þ value (Figure 2b). The results of the decision analysis indicate that the cost for aortorenal bypass is $103,453 $596,782 (N ¼ 482, 95% CI $156,864 to $50,042). The cost for renal artery stent is $305,370 $300,519 (N ¼ 50,419, 95% CI $307,993 to $302,747). When including the VSL, aortorenal bypass has an overall value of $415,881 $760,021 (N ¼ 482, 95% CI $347,860– $483,902). Renal artery stenting has an overall negative value of $182,270 $710,304 (N ¼ 50,419, 95% CI $188,470 to $176,070). Both cost and cost þ value for aortorenal bypass compared to renal artery stenting are significantly different (p < 0.0001).
Discussion Several trials have since been conducted on comparing angioplasty, angioplasty and stent placement, and best medical management.5,28–33 Each of the earlier trials had significant limitations in patient selection, degree of RAS, utilization of a stent and/or distal embolic
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
8
Vascular 0(0)
protection device, optimal medical management, and sufficient power to determine a difference in the groups. These issues were corrected with the CORAL trial, which demonstrated equivalency between optimal medical management and stent placement plus optimal medical management.5 While multiple trials have been completed on percutaneous renal artery intervention, there is only one study comparing percutaneous renal intervention with surgical revascularization.34 Weibull et al. compared percutaneous transluminal renal angioplasty with surgical reconstruction of atherosclerotic RAS in a prospective randomized study in 1993.34 At that time, primary patency following angioplasty was 75% at two years compared to 96% for bypass in a group of 58 patients, with 17% of patients who underwent angioplasty requiring surgical intervention. In addition to the higher re-intervention rate following angioplasty, other limitations of this study was that long-term dialysis-free survival was not evaluated, exclusion criteria that limited the applicability of the results, crossover from the angioplasty cohort to the surgical management group, and that stents were not used in conjunction with angioplasty. Apart from this limited study more than two decades ago, there has not been a suitably controlled trial evaluating surgical revascularization with percutaneous renal intervention or optimal medical management for atherosclerotic RAS. It is possible to compare the costs and relative value of aortorenal bypass with renal artery stenting. Using level IIA or greater data from the literature, reference values for efficacy, patency, and survival for each of the procedures can be determined. Cost data can be obtained from the NIS, along with any missing values regarding the incidence of complications and inpatient survival. While individual studies have evaluated longterm dialysis-free survival, no study has thus far placed these long-term results in the context of complications and overall efficacy using a cost analysis. Aortorenal bypass has an overall cost of $103,453 compared to $305,370 for renal artery stenting. While the initial cost of the procedure is nearly $10,000 less for renal artery stenting ($29,938 for aortorenal bypass compared to $19,407 for renal artery stenting), escalating costs are incurred from the significantly lower fiveyear patency rate for renal artery stenting (Figure 1). The annual cost of dialysis is $91,455 and the proportion of patients who require dialysis at five years is greater for renal artery stent patients than it is for aortorenal bypass patients. This leads to a nearly threefold increase in costs associated with renal artery stenting, even though these patients tend to be older (69.3 vs. 62.3 years) and less ill (Table 2). Furthermore, while most individual costs associated with complications from renal artery stenting are less than that for aortorenal bypass, the overall cost is nearly $8000 greater
($29,341 vs. $21,116) due to the increased incidence of individual complications (Table 3). The value of these interventions was determined using a VSL analysis. The VSL was assigned to patients who survived their procedure and had a five-year dialysis-free survival. Limited VSL is expected for patients on dialysis due to the substantially poorer quality of life, but even assigning a 50% rate of VSL to these patients as part of a sensitivity analysis yields no significant overlap in the normal distribution for costþ value (Figure 2b). When incorporating cost þ value to assist in decision-making, aortorenal bypass has an overall value of $415,881, compared to a negative value of $182,270 for renal artery stenting. Overall, there is a marginal improvement of $519,334 for aortorenal bypass patients and $123,100 for renal artery stent patients when incorporating VSL analysis. If a future value calculation is completed with the $519,334 improvement, aortorenal bypass is expected to lead to an additional 10 years of dialysis-free survival compared to just 2.4 years for patients who undergo renal artery stenting, similar to what has been previously reported.5,10,12,19,21,23
Limitations The use of an inpatient database to obtain cost information, incidence of complications, and inpatient mortality relies upon the accuracy of the data entry. Multiple assumptions are made with regard to patency of each of the approaches and five-year outcomes. While reference values are used from level 2 A or better sources within the literature, defaulting only to the NIS when these values are not available, there is some bias that is introduced into the analysis. To compensate for this bias, a variance-based sensitivity analysis was completed using maximum variability in all of the input data. Normal distributions were created based off this analysis, and there is very little overlap between these normal distributions. This indicates a high level of confidence with the data (p < 0.0001). The influence of individual or hospital volume was not included within this analysis and may be a source of bias. Another limitation of the database utilized is that there is no data on the type of renal lesions in these patients and how the lesion characteristics impacted management choices; furthermore we do not have any data on long term follow-up in these patients given the limitations of the NIS.
Conclusion Patients who fail best medical management for atherosclerotic RAS and who are candidates for open
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
Dua et al.
9
revascularization should undergo aortorenal bypass in lieu of renal artery stenting. Acknowledgments Dr Sapan S Desai, the principal investigator, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. There are no conflicts of interest or other pertinent disclosures. The dialysis cost data reported here have been supplied by the United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy or interpretation of the US government.
Conflict of interest No conflicts of interest.
Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References 1. Appel RG, Bleyer AJ, Reavis S, et al. Renovascular disease in older patients beginning renal replacement therapy. Kidney Int 1995; 48: 171–176. 2. Davis RP, Pearce JD, Craven TE, et al. Atherosclerotic renovascular disease among hypertensive adults. J Vasc Surg 2009; 50: 564–570. (571 e1-3; discussion 571). 3. Dean RH, Kieffer RW, Smith BM, et al. Renovascular hypertension: anatomic and renal function changes during drug therapy. Arch Surg 1981; 116: 1408–1415. 4. Schreiber MJ, Pohl MA and Novick AC. The natural history of atherosclerotic and fibrous renal artery disease. Urol Clin North Am 1984; 11: 383–392. 5. Cooper CJ, Murphy TP, Cutlip DE, et al. Stenting and medical therapy for atherosclerotic renal-artery stenosis. N Engl J Med 2014; 370: 13–22. 6. Corriere MA, Pearce JD, Edwards MS, et al. Endovascular management of atherosclerotic renovascular disease: early results following primary intervention. J Vasc Surg 2008; 48: 580–587; discussion 587–588. 7. Kashyap VS, Sepulveda RN, Bena JF, et al. The management of renal artery atherosclerosis for renal salvage: does stenting help? J Vasc Surg 2007; 45: 101–108. (discussion 108–109). 8. Zeller T, Frank U, Mu¨ller C, et al. Predictors of improved renal function after percutaneous stent-supported angioplasty of severe atherosclerotic ostial renal artery stenosis. Circulation 2003; 108: 2244–2249. 9. Zeller T, Frank U, Mu¨ller C, et al. Stent-supported angioplasty of severe atherosclerotic renal artery stenosis preserves renal function and improves blood pressure control: long-term results from a prospective registry of 456 lesions. J Endovasc Ther 2004; 11: 95–106. 10. Hansen KJ, Deitch JS, Oskin TC, et al. Renal artery repair: consequence of operative failures. Ann Surg 1998; 227: 678–689. (discussion 689–690).
11. Hansen KJ, Cherr GS, Craven TE, et al. Management of ischemic nephropathy: dialysis-free survival after surgical repair. J Vasc Surg 2000; 32: 472–481. (discussion 481– 482). 12. Cherr GS, Hansen KJ, Craven TE, et al. Surgical management of atherosclerotic renovascular disease. J Vasc Surg 2002; 35: 236–245. 13. System, U.S.R.D., USRDS 2013 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. United States Renal Data System, 2014 Annual Data Report: Epidemiology of Kidney Disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2014. 14. Kip Viscusi W and Aldy JE. The value of a statistical life: a critical review of market estimates throughout the world. J Risk Uncertainty 2003; 27: 5–76. 15. Polly Trottenberg, RSR. Guidance on Treatment of the Economic Value of a Statistical Life (VSL) in U.S. Department of Transportation Analyses. Washington DC: Department of Motor Vehicles, 2013. 16. Baram D, Daroowalla F, Garcia R, et al. Use of the All Patient Refined-Diagnosis Related Group (APR-DRG) Risk of Mortality Score as a severity adjustor in the medical ICU. Clin Med Circ Respir Pulm Med 2008; 2: 19–25. 17. Lee CP, Chertow GM and Zenios SA. An empiric estimate of the value of life: updating the renal dialysis costeffectiveness standard. Value Health 2009; 12: 80–87. 18. Feng R, Wei XL, Zhao ZQ, et al. Aortorenal bypass with autologous saphenous vein in Takayasu arteritis-induced renal artery stenosis. Eur J Vasc Endovasc Surg 2011; 42: 47–53. 19. Ham SW, Kumar SR, Wang BR, et al. Late outcomes of endovascular and open revascularization for nonatherosclerotic renal artery disease. Arch Surg 2010; 145: 832–839. 20. Kennedy DJ, Colyer WR, Brewster PS, et al. Renal insufficiency as a predictor of adverse events and mortality after renal artery stent placement. Am J Kidney Dis 2003; 42: 926–935. 21. Hansen KJ, Cherr GS and Dean RH. Dialysis-free survival after surgical repair of ischemic nephropathy. Cardiovasc Surg 2002; 10: 400–4. 22. Arthurs Z, Starnes B, Cuadrado D, et al. Renal artery stenting slows the rate of renal function decline. J Vasc Surg 2007; 45: 726–731. (discussion 731–732). 23. Bates MC, Campbell JE, Stone PA, et al. Factors affecting long-term survival following renal artery stenting. Catheter Cardiovasc Interv 2007; 69: 1037–1043. 24. Hagino RT, Valentine RJ and Clagett GP. Supraceliac aortorenal bypass. J Vasc Surg 1997; 26: 482–489. (discussion 490–491). 25. Hoshino Y, Nakamura T, Nakano A, et al. Successful treatment of renovascular hypertension due to fibromuscular dysplasia by intravascular ultrasound-guided atherectomy. Nephron 2002; 91: 521–525. 26. Kimura H, Sato O, Deguchi JO, et al. Surgical treatment and long-term outcome of renovascular hypertension in children and adolescents. Eur J Vasc Endovasc Surg 2010; 39: 731–737.
Downloaded from vas.sagepub.com by guest on November 15, 2015
XML Template (2015) [23.4.2015–4:53pm] //blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/VASJ/Vol00000/150020/APPFile/SG-VASJ150020.3d
(VAS)
[1–10] [PREPRINTER stage]
10
Vascular 0(0)
27. Schweiger H, Raithel D, Seyferth W, et al. Surgical treatment of renal artery occlusive disease: long term results. J Cardiovasc Surg (Torino) 1984; 25: 111–114. 28. van Jaarsveld BC, Derkx FH, Krijnen P, et al. Hypertension resistant to two-drug treatment’ is a useful criterion to select patients for angiography: the ‘Dutch Renal Artery Stenosis Intervention Cooperative’ (DRASTIC) study. Contrib Nephrol 1996; 119: 54–58. 29. Plouin PF, Chatellier G, Darne´ B, et al. Blood pressure outcome of angioplasty in atherosclerotic renal artery stenosis: a randomized trial. Essai Multicentrique Medicaments vs Angioplastie (EMMA) Study Group. Hypertension 1998; 31: 823–829. 30. Webster J, Marshall F, Abdalla M, et al. Randomised comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous renal artery stenosis. Scottish and Newcastle Renal Artery Stenosis Collaborative Group. J Hum Hypertens 1998; 12: 329–335.
31. Ives NJ, Wheatley K, Stowe RL, et al. Continuing uncertainty about the value of percutaneous revascularization in atherosclerotic renovascular disease: a meta-analysis of randomized trials. Nephrol Dial Transplant 2003; 18: 298–304. 32. Tobe SW, Atri M, Perkins N, et al. Renal athersosclerotic revascularization evaluation (RAVE study): study protocol of a randomized trial [NCT00127738]. BMC Nephrol 2007; 8: 4. 33. ASTRAL Investigators, Wheatley K, Ives N, et al. Revascularization versus medical therapy for renalartery stenosis. N Engl J Med 2009; 361: 1953–1962. 34. Weibull H, Bergqvist D, Bergentz SE, et al. Percutaneous transluminal renal angioplasty versus surgical reconstruction of atherosclerotic renal artery stenosis: a prospective randomized study. J Vasc Surg 1993; 18: 841–850. (discussion 850–852).
Downloaded from vas.sagepub.com by guest on November 15, 2015
