- Email: [email protected]
Is Thrombolysis of Lower Extremity Acute Arterial Occlusion Cost-Effective?
Journal of Surgical Research 83, 106 –112 (1999) Article ID jsre.1999.5575, available online at http://www.idealibrary.com on
Is Thrombolysis of Lower Extremity Acute Arterial Occlusion Cost-Effective? Sheela T. Patel, M.D., Paul B. Haser, M.D., Harry L. Bush, Jr., M.D., and K. Craig Kent, M.D. 1 Division of Vascular Surgery, New York Presbyterian Hospital, Cornell University Medical College, New York, New York 10021 Presented at the Annual Meeting of the Association for Academic Surgery, Seattle, Washington, November 18 –22, 1998
Background. The TOPAS (thrombolysis or peripheral artery surgery) trial randomized 544 patients with acute lower extremity ischemia to either surgery or thrombolysis. Although statistically equivalent 1-year morbidities and mortalities were demonstrated, the comparative cost-effectiveness of these two interventions has not been explored. Materials and methods. We constructed a Markov decision-analytic model to determine the cost-effectiveness of thrombolysis relative to surgery for a hypothetical cohort of patients with acute lower extremity arterial occlusion. Our measure of outcome was the costeffectiveness ratio (CER), defined as the incremental lifetime cost per quality-adjusted life year gained. Estimates of 1-year outcomes were based on the TOPAS trial: mortality (lysis, 20%; surgery, 17%), amputation (lysis, 15%; surgery, 13%), the number of additional interventions required following the initial procedure (lysis, 544; surgery, 439). Procedural costs were estimated from the cost accounting system at the New York Presbyterian Hospital as well as from the literature. Results. Operative intervention for acute lower extremity arterial occlusion extended life and was less costly compared to thrombolysis. The projected life expectancy for patients who underwent initial surgery was 5.04 years versus 4.75 years for initial thrombolysis. The lifetime costs were $57,429 for surgery versus $76,326 for thrombolysis. In performing sensitivity analyses, a threshold CER of $60,000 was considered what society would pay for accepted medical interventions. Thrombolysis became cost-effective if the 1-year mortality rate for lysis was lowered from 20 to 10.7%, if the amputation rate for lysis diminished from 1 To whom correspondence should be addressed at New York Presbyterian Hospital, Cornell University Medical College, 525 East 68th Street, Room F1909, New York, NY 10021. Fax: (212) 746-5812. E-mail: [email protected].
0022-4804/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
15 to 3.9%, or if the 1-year cost of lysis could be reduced to a level below $13,000. Conclusions. Initial surgery provides the most efficient and economical utilization of resources for acute lower extremity arterial occlusion. The high cost of thrombolysis is related to the expense of the lytic agents, the need for subsequent interventions in patients treated with initial lysis, and the long-term costs of amputation in patients who fail lytic therapy. © 1999 Academic Press Key Words: vascular; thrombolysis; cost-effectiveness. INTRODUCTION
Acute lower extremity arterial occlusion is associated with significant morbidity and mortality. Although surgery is the traditional treatment for patients who have lower extremity ischemia secondary to arterial occlusion, catheter-directed thrombolysis has become an increasingly recognized modality for treating this condition. Thrombolysis has the potential advantage of minimizing endothelial trauma, lysing thrombus in distal branch and collateral vessels, allowing the cause of occlusion to be identified, simplifying surgery, and allowing lesions to be treated with endovascular therapy. Moreover, because of its less invasive nature, thrombolysis might theoretically reduce hospital length of stay and costs relative to surgical interventions. There have been two large randomized, prospective comparisons of thrombolysis and surgery for the treatment of lower extremity arterial occlusion [1, 2]. In the STILE (surgery versus thrombolysis for ischemia of the lower extremity) trial, patients with nonembolic, native artery occlusions occurring within 6 months were randomized to thrombolysis or surgery [1]. This trial was terminated early when it was found that initial
106
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
107
FIG. 1. Simplified Markov decision-analytic model. Hypothetical cohort of patients with acute lower extremity arterial occlusion may undergo either an initial strategy of thrombolysis or surgery. The outcomes of either strategy are limb salvage, amputation, stroke, or death.
surgery was more effective and durable than lysis. At 1 year, the incidence of recurrent ischemia and major amputation was greater with lysis, although there was no difference in the 1-year mortality rate. The largest study to date comparing treatment of acute arterial occlusion with surgery or catheterdirected lysis is the TOPAS (thrombolysis or peripheral arterial surgery) trial [2]. Five hundred and forty-four patients with thrombotic or embolic events of less 14 days’ duration were randomized. Patients with both native arterial and bypass graft occlusions were included. The authors found a slightly greater rate of amputation and mortality at 1 year in the thrombolysis group. However, from a statistical standpoint, the outcomes of these two interventions were equivalent. The authors concluded that lytic therapy, because it is “less invasive,” should be the first-line approach for patients who present with acute arterial occlusion. Not considered in the TOPAS trial, however, were the costs associated with these two interventions. Substantial costs can be incurred in patients treated with thrombolysis. Urokinase is expensive. Moreover, patients undergoing thrombolysis require serial angiographic imaging and monitoring in an intensive care unit. Although equivalent morbidity and mortality were reported in the TOPAS trial, the cost-effectiveness of thrombolysis versus surgery remains unclear. Using data from the TOPAS trial, we developed a comprehen-
sive decision-analytic Markov model that incorporated not only the cost of the initial hospitalization but also the cost of complications and subsequent interventions for both treatment modalities. The objective of this analysis was to determine whether thrombolysis is a cost-effective alternative to operative intervention in the treatment of acute lower extremity ischemia. METHODS
The Decision-Analytic Model We developed a decision-analytic model which reflected the possible clinical outcomes and costs associated with a hypothetical 65year-old cohort presenting with acute (,14 days) lower extremity ischemia. Patients were selected to receive either surgery or thrombolysis (Fig. 1). We assumed that patients included within this cohort were medically suitable to undergo either procedure. Each strategy is associated with numerous possible outcomes; these various outcomes form the “branches” of a decision tree. A probability of occurrence and cost is assigned to each branch. Using a computerized Markov decision-analytic tree (SMLTREE software, version 2.9, James P. Hollenberg, Roslyn, NY), all of the possible clinical events and outcomes occurring in both hypothetical cohorts of patients were tracked and compared [3]. The Markov model follows these patients until they have died. Endpoints for this model include: (1) long-term survival in quality-adjusted life years, and (2) the lifetime treatment costs incurred by each strategy. Our measure of outcome was the incremental cost-effectiveness ratio, defined as the additional cost per quality-adjusted life year (QALY) gained by an intervention. The lower the cost-effectiveness ratio, the more efficiently a medical intervention utilizes economic resources. Although a traditional
108
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
Markov model incorporates outcomes and costs from the time of intervention until a patient’s death, data from the TOPAS trial was available for only the first year following intervention. Thus, in assigning probabilities for morbidity and deriving costs associated with interventions, it was assumed that no additional morbidity or cost was incurred after 1 year. In our initial base-case analysis, the single best estimates of the probabilities, costs, and quality-adjustment factors were used. In sensitivity analyses, we systematically varied these assumptions through a wide plausible range and the effect of these variations on the base-case analysis was measured. Because most widely accepted medical interventions have incremental cost-effectiveness ratios of less than $60,000, we considered this value to be the threshold for our sensitivity analyses [4]. All costs were converted to 1997 U.S. dollars using the medical care component of the Consumer Price Index for All Urban Consumers. In accordance with standard principles of economic analysis, costs and life expectancies were discounted at 3% per year to reflect the greater value of current dollars and life years compared to that of the future [5].
Assumptions Probabilities. The mortality in the TOPAS trial at 1 year was 20% for the thrombolysis group and 17% for the operative group. Mortality rates beyond 1 year were determined by adjusting U.S. age-specific annual mortality rates for patients with peripheral vascular disease [6]. An excess mortality rate of 8% per year was assigned to these patients [7, 8]. The amputation rate in the TOPAS trial was 15% for the lysis group and 13.1% for the operative group. Amputees were assigned an excess annual mortality of 13.6% per year [9, 10]. The rate of intracranial hemorrhage was 1.1% for lysis and 0% for surgery based on data from TOPAS. It was assumed that after 1 year, no further amputations or strokes occurred and that no additional procedures were required in these patients. Costs. We determined the cost, not the charge or reimbursement, for the various interventions and complications. Costs were derived from our hospital cost accounting system (Transitions Systems, Inc., Boston, MA) as well as from the literature. Procedural costs. The procedural cost of thrombolysis was determined to be $21,918 based on the average cost incurred by 25 patients admitted to New York Presbyterian Hospital over a 6-month period who received thrombolytic treatment and no adjunctive open surgical procedure. In the TOPAS trial, 10 of 272 patients in the thrombolytic arm had failed attempts to place a catheter or guidewire. These patients were not assigned the full cost of thrombolysis but rather the cost of an abdominal arteriogram ($1358). Open surgical procedures performed in the TOPAS trial were divided into major, moderate, and minor types. A major procedure was defined as the insertion of a new bypass graft, replacement of an existing graft, or excision or repair of an aneurysm. We assigned an average cost of $20,271 for patients undergoing a major reconstruction. This value was derived from a study by Jansen and colleagues [11] where the costs associated with 467 major bypass procedures, which included aortobifemoral, iliofemoral, femorofemoral, femoropopliteal, femoroinfrapopliteal, and popliteoinfrapopliteal bypasses, were evaluated. This cost is comparable to that reported by others for major lower extremity reconstructions [12, 13]. In the TOPAS trial, the number of major surgical procedures performed was 116 in the thrombolysis group and 193 in the surgery group. A moderate open surgical procedure was defined as graft revision, endarterectomy, profundaplasty, exploratory vascular procedure, or a transmetatarsal amputation. Based upon data from Jansen et al. [11], we assumed that the cost of endarterectomy ($12,596) would be representative of a moderate procedure. In the TOPAS trial, the number of moderate procedures performed was 98 in the thrombolysis group and 145 in the surgery group. A minor surgical procedure (137 lysis, 252 sur-
gery) was defined as thromboembolectomy or embolectomy, amputation of digits, and fasciotomy. It was assumed that the cost of thromboembolectomy would represent that of a minor procedure. Because there are no data available in the study by Jansen et al. regarding the cost of thromboembolectomy, we used the cost of surgical thrombectomy for the treatment of dialysis graft thrombosis in a study by Vesely et al. ($5938) [14]. A percutaneous cathether-based procedure (135 lysis, 70 surgery) was assigned a cost of $11,020 based upon 96 percutaneous procedures (from Jansen et al. [11]), which included iliac percutaneous transluminal angioplasty with and without stent placement and femoropopliteal angioplasty. A major amputation was assigned a cost of $32,191 based on data from Hunink et al. [15]. There is substantial support for this cost estimate in the literature [16, 17]. There were 58 amputations performed in the lysis group and 51 in the surgery group. Because of the lack of data in the literature, we did not differentiate between aboveknee and below-knee amputations. Long-term costs. Stroke and amputation have both immediate and long-term associated costs. We estimated a cost of $51,150 for the first year after a stroke and an annual cost of $26,880 for subsequent years [18]. Assuming that 29% of amputees require long-term nursing care, the annual cost of amputation was estimated at $39,735 [15]. Quality Adjustment. A quality-adjustment factor is assigned for each year of survival that a patient lives with a major morbidity based on the patient’s perception of quality of life. This qualityadjustment factor may range from 0 (death) to 1 (perfect health). We used a quality-adjustment factor of 0.40 for patients surviving with a stroke [18]. Thus, for each year a patient survives, he is credited with only 0.40 of a quality-adjusted year of life. Amputation was assigned a quality-adjustment factor of 0.80 [16].
RESULTS
Average Procedural Cost After summing the costs of all procedures performed in patients treated with an initial strategy of thrombolysis or surgery, we found that patients randomized to thrombolysis incurred an average procedural cost of $49,106, whereas the patient randomized to surgery incurred an average cost of $35,471. Base-Case Analysis For a hypothetical 65-year-old cohort of patients with acute lower extremity arterial occlusion, an initial strategy of thrombolysis provided an average life expectancy of 4.75 years at a lifetime cost of $76,326. Initial surgery provided an average life expectancy of 5.04 years at a lifetime cost of $57,429. No costeffectiveness ratio could be determined since a strategy of initial surgery was both less costly and extended life (a dominant strategy). Sensitivity Analysis Baseline assumptions were tested in sensitivity analyses by substituting a wide range of values for each variable. This process allows determination of the variables that have the greatest influence on the outcome of our analysis.
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
109
Other variables. Alterations in age, disutilities, excess mortalities, discount factor, and cost of morbidity had insignificant effects on the base-case conclusion. DISCUSSION
FIG. 2. Cost-effectiveness of thrombolysis as a function of mortality. The dotted line indicates the $60,000 threshold for accepted medical interventions. QALY, quality-adjusted life year.
Mortality. We first evaluated the mortality associated with thrombolysis and its impact on the costeffectiveness (Fig. 2). We used 20% at the base-case mortality for thrombolysis. If the mortality rate for thrombolysis was reduced to less than 10.7%, then lysis became cost-effective (i.e., cost-effectiveness ratio less than $60,000). Amputation rate. We next evaluated the effect on cost-effectiveness of diminishing the amputation rate in lysis patients. We used 15% at the base-case amputation rate for thrombolysis. If the amputation rate for lysis patients diminished to less than 3.9%, then lysis became cost-effective (i.e., ratio less than $60,000). Cost of lysis. The average cost of lysis at 1 year was determined to be $49,106 per patient. A reduction in the cost of lysis to less than $13,000 at 1 year would be necessary to make lysis cost-effective. Two- and three-way sensitivity analyses. A two-way sensitivity analysis was performed varying the 1-year rates of amputation and mortality simultaneously (Fig. 3). If the amputation rate for lysis was reduced from 15 to 13% and the mortality rate for lysis was reduced from 20 to 12%, then lysis became cost-effective. If the rate of amputation for lysis was halved from 15 to 7.5%, then the mortality rate for lysis only had to be reduced from 20 to 17% to make lysis a cost-effective alternative to surgery. If the cost of lysis at 1 year could be halved from $49,106 to $24,553, then the amputation rate for lysis only had to be reduced from 15 to 11% or the mortality rate reduced from 20 to 17% to make lysis costeffective. In a three-way sensitivity analysis, if at 1 year, the mortality rate associated with lysis lowered to 18%, the amputation rate lowered to 13.7%, and the cost of lysis diminished by 50% to $24,553, then lysis became cost-effective.
Because of the current emphasis in medicine on costcontainment, both outcomes and costs must be considered when comparing two alternative forms of treatment. Although, there have been several studies comparing the outcome of patients treated for acute lower extremity arterial occlusion with thrombolysis versus surgery, the cost of these two interventions relative to their morbidity has not been previously investigated. The most revealing study to date is the TOPAS trial in which 544 patients with potentially reversible limb-threatening ischemia of less than 14 days’ duration were randomized to either thrombolysis or surgery [2]. Since the rate of amputation and the mortality at 1 year of patients treated by both methods were equivalent, the TOPAS investigators concluded that thrombolysis was the preferred strategy because of its less invasive nature. In this analysis, we readdress the TOPAS trial giving consideration not only to morbidity and mortality but also to costs. Decision-analytic modeling is used to create a comparison with regard to cost between two alternative treatment modalities. This relationship is reported as the incremental cost-effectiveness ratio, which is a standard of measure by which the cost-effectiveness of various interventions can be compared. A costeffectiveness ratio can be calculated only under the circumstance where an intervention reduces morbidity but is more expensive than an alternative form of therapy. In other words, the cost-effectiveness ratio depicts the additional expense imposed on society by a procedure that can reduce morbidity or extend life. In our
FIG. 3. Three-way sensitivity analysis of the cost-effectiveness of thrombolysis varying the mortality and amputation rate associated with thrombolysis, and the 1-year cost of lysis using $60,000 as the threshold for accepted medical interventions.
110
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
base-case analysis, we found that the average lifetime cost for patients treated with thrombolysis was $76,326, whereas the cost for patients treated with initial surgery was $57,429. Thus, the lifetime cost of lytic therapy was actually greater than that of surgery. Moreover, we also found that the average life expectancy of patients treated with thrombolysis was less than that of patients treated with surgery (4.75 years for thrombolysis, 5.04 years for surgery). Therefore, in our base-case analysis, surgery was both less expensive and provided a greater life expectancy. No costeffectiveness ratio could be calculated since surgery was clearly the preferred strategy in terms of both outcome and cost. There are several aspects of the TOPAS trial that lend insight as to why surgical treatment is the more economically favorable strategy. Although the lysis group was initially treated with catheter-based urokinase, many of these patients, within a year, required one or more open surgical procedures. Thus, lysis was not always less invasive since in some patients, multiple procedures rather than a singular intervention were required. The total number of secondary procedures performed in the lysis group was 544 whereas in the surgical group only 439 secondary procedures were performed. Although a procedure that is “less invasive” is usually associated with diminished cost, this was not the case with catheter-based thrombolysis. The cost of the initial hospitalization for a patient treated with thrombolysis was $21,918 whereas the cost of a major surgical intervention for acute arterial occlusion such as a femoral-tibial bypass was nearly equivalent at $20,271. A similarity in the initial hospital costs of these two strategies has been reported by other investigators. Ouriel et al. [19] in the Rochester study found that treatment costs for the initial hospitalization did not differ significantly between patients initially randomized to receive urokinase versus those randomized to operation ($22,171 for lysis versus $19,775 for surgery). The observation that thrombolysis generates costs that are equivalent to surgery has also been made in patients treated for thrombosed dialysis access grafts [14]. The reasons for the high cost of lysis are multiple. Urokinase is expensive; the cost of urokinase is approximately $1.50/1000 units. Since the average patient in TOPAS received 3.5 million units, the cost of the drug alone in these patients was $5250. Treatment with thrombolysis also requires repeated utilization of the angiography suite. In addition to the diagnostic angiogram that accompanies the initiation of lysis, patients must return multiple times to the angiography suite both for repositioning of catheters and to evaluate the progress of their treatment. At our own institution, the average patient undergoing lysis was evaluated in the angiography suite on three separate occasions. The
total costs are high since the fee for maintenance of the angiography suite is approximately $500 per hour [20]. During treatment, patients are monitored in an intensive care unit setting. This is also expensive; the cost of an intensive care unit bed at our institution is approximately $1500 per day versus a cost of $491 for a ward bed. An additional reason for the high cost of lysis relates to its long-term morbidity. The rate of amputation in TOPAS was greater for thrombolysis than for surgery (15% versus 13%). Although not statistically significant, this difference has profound economic implications because of the high cost of rehabilitative and nursing care in survivors of amputation. In the Rochester study, Ouriel and colleagues [19] found an amputation rate that was equivalent for patients treated either with thrombolysis or with surgery. Thus, it is not clear that thrombolysis uniformly results in greater rates of amputation. However, it is important to realize that the costs incurred in a patient who is treated with amputation are significant and small differences in the amputation rate can have a profound effect on the cost-effectiveness of a procedure. One benefit of decision-analytic models is that the assumptions (variables) can be varied and the effect of these variations on the base-case conclusion can be tested. In sensitivity analyses, we evaluate the effect of three variables (the mortality of patients treated with thrombolysis, the amputation rate for patients treated with thrombolysis, and the lifetime costs of thrombolysis) on our conclusion that surgery is a cost-effective alternative to thrombolysis. We found that thrombolysis became cost-effective if its mortality rate diminished from 20 to 10.7%, if the amputation rate was reduced from 15 to 3.9%, or if the cost of thrombolysis at one year diminished dramatically from $49,106 to $13,000. A great reduction in any of these individual variables was necessary to make thrombolysis cost-effective. However, thrombolysis became cost-effective with only minor changes in these variables if all three were simultaneously adjusted. For example, thrombolysis became costeffective if the associated mortality was diminished from 20 to 16%, the amputation rate was diminished from 15 to 13%, and the cost was diminished from $49,106 to $35,000. These sensitivity analyses also demonstrate that if the costs associated with thrombolysis remain high, substantial reductions in morbidity and mortality must be achieved to allow this strategy to become a cost-effective alternative to surgery. We did not adjust our analysis to account for patient preference for either lysis or surgery. The convalescence period after thrombolysis may be shorter than that of a surgical procedure. Therefore, patients might be willing to accept additional risk or society might accept additional cost as a trade-off for this diminished period of convalescence. Techniques are available that
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
allow incorporation of patient preference into a decisionanalytic model; however, these techniques require prospective data that were not accrued as part of the TOPAS trial. To determine the cost-effectiveness of thrombolysis for acute lower extremity arterial occlusion, we relied solely on data from the TOPAS trial. Because these data were limited it was necessary that we make several important assumptions. First, since the follow-up from TOPAS was only 1 year, the long-term fate of a large randomized group of patients treated with either surgery or thrombolysis is unknown. We assumed for this analysis that after the first year, the clinical course was stable and equivalent for patients treated with lysis or surgery. We also assumed that there were no additional costs incurred by patients in either group after 1 year. These assumptions were made so as to not bias the analysis in favor of either thrombolysis or surgery. However, only after the long-term consequences of both forms of treatment are available will we truly understand their comparative economic value. A major criticism of the TOPAS trial was the heterogeneity of the patients who were studied [21]. Patients were included who presented with both embolic and thrombotic events as well as native arterial and graft occlusions. Although we found it more costeffective to intervene for acute arterial occlusion with surgery versus lysis, because we relied upon TOPAS for our data, our analysis is also affected by the lack of uniformity in the patient population. The most costeffective treatment of acute lower extremity arterial occlusion may vary with the patient’s clinical presentation. For example, a patient with an undiseased arterial system who embolizes to the common femoral artery as a consequence of atrial fibrillation might be most efficiently treated with a surgical embolectomy under local anesthesia. Alternatively, a patient with a history of claudication and an iliac artery stenosis who presents with an iliac artery thrombosis might be most efficiently treated with lysis and iliac artery angioplasty. More data regarding the outcomes and costs of patients within these individual categories will be required before specific cost comparisons can be made. There is to date no data that demonstrate a superior outcome for patients treated with thrombolysis versus surgery for acute lower extremity arterial occlusion. Furthermore, the costs associated with these two modalities of treatment favor the initial use of surgery. The high cost of lysis, the need for secondary interventions, and the long-term costs of amputation make lysis, at least in the patient cohort treated in TOPAS, an economically unfavorable alternative to surgery. Since patients who present with acute lower extremity arterial ischemia are heterogeneous, we recommend that treatment of these
111
patients still be individualized, with careful attention paid to costs as well as outcomes. REFERENCES 1.
Weaver, F. A., Comerota, A. J., Youngblood, M., Froehlich, J., Hosking, J. D., and Papanicolaou, G. Surgical revascularization versus thrombolysis for nonembolic lower extremity native artery occlusions: Results of a prospective randomized trial. J. Vasc. Surg. 24: 513, 1996.
2.
Ouriel, K., Veith, F. J., and Sasahara, A. A. A comparison of recombinant urokinase with vascular surgery as initial treatment for acute arterial occlusion of the legs. N. Engl. J. Med. 338: 1105, 1998.
3.
Sonnenberg, F. A., and Beck, J. R. Markov models in medical decision making: A practical guide. Med. Decis. Making 13: 322, 1993.
4.
Goldman, L., Gordon, D. J., Rifkind, B. M., Hulley, S. B., Detsky, A. S., Goodman, D. W., Kinosian, B., and Weinstein, M. C. Cost and health implications of cholesterol lowering. Circulation 85: 1960, 1992.
5.
Gold, M. R., Siegel, J. S., Russell, L. B., Weinstein, M. C., and the Panel on Cost-Effectiveness in Health and Medicine. CostEffectiveness in Health and Medicine. New York: Oxford Univ. Press, 1996.
6.
National Center for Health Statistics. Vital Statistics of the United States, 1988, Vol. II, Sect. 6, Life Tables. Washington, DC: U.S. Public Health Service, 1990.
7.
Abou-Zamzam, A. M., Jr., Lee, R. W., Moneta, G. L., Taylor, L. M., Jr., and Porter, J. M. Functional outcome after infrainguinal bypass for limb salvage. J. Vasc. Surg. 25: 287, 1997.
8.
Kalman, P. G., and Johnston, K. W. Predictors of long-term patient survival after in situ vein leg bypass. J. Vasc. Surg. 25: 899, 1997.
9.
Dawson, I., Keller, P. J. A., Brand, R., Pesch-Batenburg, J., and van Bockel, J. H. Late outcomes of limb loss after failed infrainguinal bypass. J. Vasc. Surg. 21: 613, 1995.
10.
Luther, M. Surgical treatment of chronic critical leg ischemia: A five-year follow-up of survival, mobility, and treatment level. Eur. J. Surg. 164: 35, 1998.
11.
Jansen, R. M. G., de Vries, S. O., Cullen, K. A., Donaldson, M. C., and Hunink, M. G. M. Cost-identification analysis of revascularization procedures on patients with peripheral arterial occlusive disease. J. Vasc. Surg. 28: 617, 1998.
12.
Roddy, S. P., O’Donnell, T. F., Iafrati, M. D., Isaacson, L. A., Bailey, V. E., and Mackey, W. C. Reduction of hospital resources utilization in vascular surgery: A four-year experience. J. Vasc. Surg. 27: 1066, 1998.
13.
Gupta, S. K., and Veith, F. J. Inadequacy of diagnosis related group (DRG) reimbursements for limb salvage lower extremity arterial reconstructions. J. Vasc. Surg. 11: 348, 1990.
14.
Vesely, T. M., Idso, M. C., Audrain, J., Windus, D. W., and Lowell, J. A. Thrombolysis versus surgical thrombectomy for the treatment of dialysis graft thrombosis: Pilot study comparing costs. JVIR 7: 507, 1996.
15.
Hunink, M. G. M., Wong, J. B., Donaldson, M. C., Meyerovitz, M. F., de Vries, J., and Harrington, D. P. Revascularization for femoropopliteal disease: A decision and cost-effectiveness analysis. J. Am. Med. Assoc. 274: 165, 1995.
16.
Eckman, M. H., Greenfield, S., Mackey, W. C., Wong, J. B., Kaplan, S., Sullivan, L., Dukes, K., and Pauker, S. G. Foot
112
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
infections in diabetic patients: Decision and cost-effectiveness analyses. J. Am. Med. Assoc. 273: 712, 1995. 17. Raviola, C. A., Nichter, L. S., Baker, J. D., Busuttil, R. W., Machleder, H. I., and Moore, W. S. Cost of treating advanced leg ischemia: Bypass graft vs primary amputation. Arch. Surg. 123: 495, 1988. 18. Kuntz, K. M., and Kent, K. C. Is carotid endarterectomy costeffective? An analysis of symptomatic and asymptomatic patients. Circulation 94: II-194, 1996.
19.
Ouriel, K., Kolassa, M., DeWeese, J. A., and Green, R. M. Economic implications of thrombolysis or operation as the initial treatment modality in acute peripheral arterial occlusion. Surgery 118: 810, 1995. 20. Bell, G. K., Bettman, M. A., Wiener, T., Lucas, C., Krone, S., and McDaniel, M. D. Peripheral vascular disease intervention cost analysis. Circulation 94: I-300, 1996. 21. Porter, J. M. Thrombolysis for acute arterial occlusion of the legs. N. Engl. J. Med. 338: 1148, 1998.
Is Thrombolysis of Lower Extremity Acute Arterial Occlusion Cost-Effective? Sheela T. Patel, M.D., Paul B. Haser, M.D., Harry L. Bush, Jr., M.D., and K. Craig Kent, M.D. 1 Division of Vascular Surgery, New York Presbyterian Hospital, Cornell University Medical College, New York, New York 10021 Presented at the Annual Meeting of the Association for Academic Surgery, Seattle, Washington, November 18 –22, 1998
Background. The TOPAS (thrombolysis or peripheral artery surgery) trial randomized 544 patients with acute lower extremity ischemia to either surgery or thrombolysis. Although statistically equivalent 1-year morbidities and mortalities were demonstrated, the comparative cost-effectiveness of these two interventions has not been explored. Materials and methods. We constructed a Markov decision-analytic model to determine the cost-effectiveness of thrombolysis relative to surgery for a hypothetical cohort of patients with acute lower extremity arterial occlusion. Our measure of outcome was the costeffectiveness ratio (CER), defined as the incremental lifetime cost per quality-adjusted life year gained. Estimates of 1-year outcomes were based on the TOPAS trial: mortality (lysis, 20%; surgery, 17%), amputation (lysis, 15%; surgery, 13%), the number of additional interventions required following the initial procedure (lysis, 544; surgery, 439). Procedural costs were estimated from the cost accounting system at the New York Presbyterian Hospital as well as from the literature. Results. Operative intervention for acute lower extremity arterial occlusion extended life and was less costly compared to thrombolysis. The projected life expectancy for patients who underwent initial surgery was 5.04 years versus 4.75 years for initial thrombolysis. The lifetime costs were $57,429 for surgery versus $76,326 for thrombolysis. In performing sensitivity analyses, a threshold CER of $60,000 was considered what society would pay for accepted medical interventions. Thrombolysis became cost-effective if the 1-year mortality rate for lysis was lowered from 20 to 10.7%, if the amputation rate for lysis diminished from 1 To whom correspondence should be addressed at New York Presbyterian Hospital, Cornell University Medical College, 525 East 68th Street, Room F1909, New York, NY 10021. Fax: (212) 746-5812. E-mail: [email protected].
0022-4804/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
15 to 3.9%, or if the 1-year cost of lysis could be reduced to a level below $13,000. Conclusions. Initial surgery provides the most efficient and economical utilization of resources for acute lower extremity arterial occlusion. The high cost of thrombolysis is related to the expense of the lytic agents, the need for subsequent interventions in patients treated with initial lysis, and the long-term costs of amputation in patients who fail lytic therapy. © 1999 Academic Press Key Words: vascular; thrombolysis; cost-effectiveness. INTRODUCTION
Acute lower extremity arterial occlusion is associated with significant morbidity and mortality. Although surgery is the traditional treatment for patients who have lower extremity ischemia secondary to arterial occlusion, catheter-directed thrombolysis has become an increasingly recognized modality for treating this condition. Thrombolysis has the potential advantage of minimizing endothelial trauma, lysing thrombus in distal branch and collateral vessels, allowing the cause of occlusion to be identified, simplifying surgery, and allowing lesions to be treated with endovascular therapy. Moreover, because of its less invasive nature, thrombolysis might theoretically reduce hospital length of stay and costs relative to surgical interventions. There have been two large randomized, prospective comparisons of thrombolysis and surgery for the treatment of lower extremity arterial occlusion [1, 2]. In the STILE (surgery versus thrombolysis for ischemia of the lower extremity) trial, patients with nonembolic, native artery occlusions occurring within 6 months were randomized to thrombolysis or surgery [1]. This trial was terminated early when it was found that initial
106
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
107
FIG. 1. Simplified Markov decision-analytic model. Hypothetical cohort of patients with acute lower extremity arterial occlusion may undergo either an initial strategy of thrombolysis or surgery. The outcomes of either strategy are limb salvage, amputation, stroke, or death.
surgery was more effective and durable than lysis. At 1 year, the incidence of recurrent ischemia and major amputation was greater with lysis, although there was no difference in the 1-year mortality rate. The largest study to date comparing treatment of acute arterial occlusion with surgery or catheterdirected lysis is the TOPAS (thrombolysis or peripheral arterial surgery) trial [2]. Five hundred and forty-four patients with thrombotic or embolic events of less 14 days’ duration were randomized. Patients with both native arterial and bypass graft occlusions were included. The authors found a slightly greater rate of amputation and mortality at 1 year in the thrombolysis group. However, from a statistical standpoint, the outcomes of these two interventions were equivalent. The authors concluded that lytic therapy, because it is “less invasive,” should be the first-line approach for patients who present with acute arterial occlusion. Not considered in the TOPAS trial, however, were the costs associated with these two interventions. Substantial costs can be incurred in patients treated with thrombolysis. Urokinase is expensive. Moreover, patients undergoing thrombolysis require serial angiographic imaging and monitoring in an intensive care unit. Although equivalent morbidity and mortality were reported in the TOPAS trial, the cost-effectiveness of thrombolysis versus surgery remains unclear. Using data from the TOPAS trial, we developed a comprehen-
sive decision-analytic Markov model that incorporated not only the cost of the initial hospitalization but also the cost of complications and subsequent interventions for both treatment modalities. The objective of this analysis was to determine whether thrombolysis is a cost-effective alternative to operative intervention in the treatment of acute lower extremity ischemia. METHODS
The Decision-Analytic Model We developed a decision-analytic model which reflected the possible clinical outcomes and costs associated with a hypothetical 65year-old cohort presenting with acute (,14 days) lower extremity ischemia. Patients were selected to receive either surgery or thrombolysis (Fig. 1). We assumed that patients included within this cohort were medically suitable to undergo either procedure. Each strategy is associated with numerous possible outcomes; these various outcomes form the “branches” of a decision tree. A probability of occurrence and cost is assigned to each branch. Using a computerized Markov decision-analytic tree (SMLTREE software, version 2.9, James P. Hollenberg, Roslyn, NY), all of the possible clinical events and outcomes occurring in both hypothetical cohorts of patients were tracked and compared [3]. The Markov model follows these patients until they have died. Endpoints for this model include: (1) long-term survival in quality-adjusted life years, and (2) the lifetime treatment costs incurred by each strategy. Our measure of outcome was the incremental cost-effectiveness ratio, defined as the additional cost per quality-adjusted life year (QALY) gained by an intervention. The lower the cost-effectiveness ratio, the more efficiently a medical intervention utilizes economic resources. Although a traditional
108
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
Markov model incorporates outcomes and costs from the time of intervention until a patient’s death, data from the TOPAS trial was available for only the first year following intervention. Thus, in assigning probabilities for morbidity and deriving costs associated with interventions, it was assumed that no additional morbidity or cost was incurred after 1 year. In our initial base-case analysis, the single best estimates of the probabilities, costs, and quality-adjustment factors were used. In sensitivity analyses, we systematically varied these assumptions through a wide plausible range and the effect of these variations on the base-case analysis was measured. Because most widely accepted medical interventions have incremental cost-effectiveness ratios of less than $60,000, we considered this value to be the threshold for our sensitivity analyses [4]. All costs were converted to 1997 U.S. dollars using the medical care component of the Consumer Price Index for All Urban Consumers. In accordance with standard principles of economic analysis, costs and life expectancies were discounted at 3% per year to reflect the greater value of current dollars and life years compared to that of the future [5].
Assumptions Probabilities. The mortality in the TOPAS trial at 1 year was 20% for the thrombolysis group and 17% for the operative group. Mortality rates beyond 1 year were determined by adjusting U.S. age-specific annual mortality rates for patients with peripheral vascular disease [6]. An excess mortality rate of 8% per year was assigned to these patients [7, 8]. The amputation rate in the TOPAS trial was 15% for the lysis group and 13.1% for the operative group. Amputees were assigned an excess annual mortality of 13.6% per year [9, 10]. The rate of intracranial hemorrhage was 1.1% for lysis and 0% for surgery based on data from TOPAS. It was assumed that after 1 year, no further amputations or strokes occurred and that no additional procedures were required in these patients. Costs. We determined the cost, not the charge or reimbursement, for the various interventions and complications. Costs were derived from our hospital cost accounting system (Transitions Systems, Inc., Boston, MA) as well as from the literature. Procedural costs. The procedural cost of thrombolysis was determined to be $21,918 based on the average cost incurred by 25 patients admitted to New York Presbyterian Hospital over a 6-month period who received thrombolytic treatment and no adjunctive open surgical procedure. In the TOPAS trial, 10 of 272 patients in the thrombolytic arm had failed attempts to place a catheter or guidewire. These patients were not assigned the full cost of thrombolysis but rather the cost of an abdominal arteriogram ($1358). Open surgical procedures performed in the TOPAS trial were divided into major, moderate, and minor types. A major procedure was defined as the insertion of a new bypass graft, replacement of an existing graft, or excision or repair of an aneurysm. We assigned an average cost of $20,271 for patients undergoing a major reconstruction. This value was derived from a study by Jansen and colleagues [11] where the costs associated with 467 major bypass procedures, which included aortobifemoral, iliofemoral, femorofemoral, femoropopliteal, femoroinfrapopliteal, and popliteoinfrapopliteal bypasses, were evaluated. This cost is comparable to that reported by others for major lower extremity reconstructions [12, 13]. In the TOPAS trial, the number of major surgical procedures performed was 116 in the thrombolysis group and 193 in the surgery group. A moderate open surgical procedure was defined as graft revision, endarterectomy, profundaplasty, exploratory vascular procedure, or a transmetatarsal amputation. Based upon data from Jansen et al. [11], we assumed that the cost of endarterectomy ($12,596) would be representative of a moderate procedure. In the TOPAS trial, the number of moderate procedures performed was 98 in the thrombolysis group and 145 in the surgery group. A minor surgical procedure (137 lysis, 252 sur-
gery) was defined as thromboembolectomy or embolectomy, amputation of digits, and fasciotomy. It was assumed that the cost of thromboembolectomy would represent that of a minor procedure. Because there are no data available in the study by Jansen et al. regarding the cost of thromboembolectomy, we used the cost of surgical thrombectomy for the treatment of dialysis graft thrombosis in a study by Vesely et al. ($5938) [14]. A percutaneous cathether-based procedure (135 lysis, 70 surgery) was assigned a cost of $11,020 based upon 96 percutaneous procedures (from Jansen et al. [11]), which included iliac percutaneous transluminal angioplasty with and without stent placement and femoropopliteal angioplasty. A major amputation was assigned a cost of $32,191 based on data from Hunink et al. [15]. There is substantial support for this cost estimate in the literature [16, 17]. There were 58 amputations performed in the lysis group and 51 in the surgery group. Because of the lack of data in the literature, we did not differentiate between aboveknee and below-knee amputations. Long-term costs. Stroke and amputation have both immediate and long-term associated costs. We estimated a cost of $51,150 for the first year after a stroke and an annual cost of $26,880 for subsequent years [18]. Assuming that 29% of amputees require long-term nursing care, the annual cost of amputation was estimated at $39,735 [15]. Quality Adjustment. A quality-adjustment factor is assigned for each year of survival that a patient lives with a major morbidity based on the patient’s perception of quality of life. This qualityadjustment factor may range from 0 (death) to 1 (perfect health). We used a quality-adjustment factor of 0.40 for patients surviving with a stroke [18]. Thus, for each year a patient survives, he is credited with only 0.40 of a quality-adjusted year of life. Amputation was assigned a quality-adjustment factor of 0.80 [16].
RESULTS
Average Procedural Cost After summing the costs of all procedures performed in patients treated with an initial strategy of thrombolysis or surgery, we found that patients randomized to thrombolysis incurred an average procedural cost of $49,106, whereas the patient randomized to surgery incurred an average cost of $35,471. Base-Case Analysis For a hypothetical 65-year-old cohort of patients with acute lower extremity arterial occlusion, an initial strategy of thrombolysis provided an average life expectancy of 4.75 years at a lifetime cost of $76,326. Initial surgery provided an average life expectancy of 5.04 years at a lifetime cost of $57,429. No costeffectiveness ratio could be determined since a strategy of initial surgery was both less costly and extended life (a dominant strategy). Sensitivity Analysis Baseline assumptions were tested in sensitivity analyses by substituting a wide range of values for each variable. This process allows determination of the variables that have the greatest influence on the outcome of our analysis.
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
109
Other variables. Alterations in age, disutilities, excess mortalities, discount factor, and cost of morbidity had insignificant effects on the base-case conclusion. DISCUSSION
FIG. 2. Cost-effectiveness of thrombolysis as a function of mortality. The dotted line indicates the $60,000 threshold for accepted medical interventions. QALY, quality-adjusted life year.
Mortality. We first evaluated the mortality associated with thrombolysis and its impact on the costeffectiveness (Fig. 2). We used 20% at the base-case mortality for thrombolysis. If the mortality rate for thrombolysis was reduced to less than 10.7%, then lysis became cost-effective (i.e., cost-effectiveness ratio less than $60,000). Amputation rate. We next evaluated the effect on cost-effectiveness of diminishing the amputation rate in lysis patients. We used 15% at the base-case amputation rate for thrombolysis. If the amputation rate for lysis patients diminished to less than 3.9%, then lysis became cost-effective (i.e., ratio less than $60,000). Cost of lysis. The average cost of lysis at 1 year was determined to be $49,106 per patient. A reduction in the cost of lysis to less than $13,000 at 1 year would be necessary to make lysis cost-effective. Two- and three-way sensitivity analyses. A two-way sensitivity analysis was performed varying the 1-year rates of amputation and mortality simultaneously (Fig. 3). If the amputation rate for lysis was reduced from 15 to 13% and the mortality rate for lysis was reduced from 20 to 12%, then lysis became cost-effective. If the rate of amputation for lysis was halved from 15 to 7.5%, then the mortality rate for lysis only had to be reduced from 20 to 17% to make lysis a cost-effective alternative to surgery. If the cost of lysis at 1 year could be halved from $49,106 to $24,553, then the amputation rate for lysis only had to be reduced from 15 to 11% or the mortality rate reduced from 20 to 17% to make lysis costeffective. In a three-way sensitivity analysis, if at 1 year, the mortality rate associated with lysis lowered to 18%, the amputation rate lowered to 13.7%, and the cost of lysis diminished by 50% to $24,553, then lysis became cost-effective.
Because of the current emphasis in medicine on costcontainment, both outcomes and costs must be considered when comparing two alternative forms of treatment. Although, there have been several studies comparing the outcome of patients treated for acute lower extremity arterial occlusion with thrombolysis versus surgery, the cost of these two interventions relative to their morbidity has not been previously investigated. The most revealing study to date is the TOPAS trial in which 544 patients with potentially reversible limb-threatening ischemia of less than 14 days’ duration were randomized to either thrombolysis or surgery [2]. Since the rate of amputation and the mortality at 1 year of patients treated by both methods were equivalent, the TOPAS investigators concluded that thrombolysis was the preferred strategy because of its less invasive nature. In this analysis, we readdress the TOPAS trial giving consideration not only to morbidity and mortality but also to costs. Decision-analytic modeling is used to create a comparison with regard to cost between two alternative treatment modalities. This relationship is reported as the incremental cost-effectiveness ratio, which is a standard of measure by which the cost-effectiveness of various interventions can be compared. A costeffectiveness ratio can be calculated only under the circumstance where an intervention reduces morbidity but is more expensive than an alternative form of therapy. In other words, the cost-effectiveness ratio depicts the additional expense imposed on society by a procedure that can reduce morbidity or extend life. In our
FIG. 3. Three-way sensitivity analysis of the cost-effectiveness of thrombolysis varying the mortality and amputation rate associated with thrombolysis, and the 1-year cost of lysis using $60,000 as the threshold for accepted medical interventions.
110
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
base-case analysis, we found that the average lifetime cost for patients treated with thrombolysis was $76,326, whereas the cost for patients treated with initial surgery was $57,429. Thus, the lifetime cost of lytic therapy was actually greater than that of surgery. Moreover, we also found that the average life expectancy of patients treated with thrombolysis was less than that of patients treated with surgery (4.75 years for thrombolysis, 5.04 years for surgery). Therefore, in our base-case analysis, surgery was both less expensive and provided a greater life expectancy. No costeffectiveness ratio could be calculated since surgery was clearly the preferred strategy in terms of both outcome and cost. There are several aspects of the TOPAS trial that lend insight as to why surgical treatment is the more economically favorable strategy. Although the lysis group was initially treated with catheter-based urokinase, many of these patients, within a year, required one or more open surgical procedures. Thus, lysis was not always less invasive since in some patients, multiple procedures rather than a singular intervention were required. The total number of secondary procedures performed in the lysis group was 544 whereas in the surgical group only 439 secondary procedures were performed. Although a procedure that is “less invasive” is usually associated with diminished cost, this was not the case with catheter-based thrombolysis. The cost of the initial hospitalization for a patient treated with thrombolysis was $21,918 whereas the cost of a major surgical intervention for acute arterial occlusion such as a femoral-tibial bypass was nearly equivalent at $20,271. A similarity in the initial hospital costs of these two strategies has been reported by other investigators. Ouriel et al. [19] in the Rochester study found that treatment costs for the initial hospitalization did not differ significantly between patients initially randomized to receive urokinase versus those randomized to operation ($22,171 for lysis versus $19,775 for surgery). The observation that thrombolysis generates costs that are equivalent to surgery has also been made in patients treated for thrombosed dialysis access grafts [14]. The reasons for the high cost of lysis are multiple. Urokinase is expensive; the cost of urokinase is approximately $1.50/1000 units. Since the average patient in TOPAS received 3.5 million units, the cost of the drug alone in these patients was $5250. Treatment with thrombolysis also requires repeated utilization of the angiography suite. In addition to the diagnostic angiogram that accompanies the initiation of lysis, patients must return multiple times to the angiography suite both for repositioning of catheters and to evaluate the progress of their treatment. At our own institution, the average patient undergoing lysis was evaluated in the angiography suite on three separate occasions. The
total costs are high since the fee for maintenance of the angiography suite is approximately $500 per hour [20]. During treatment, patients are monitored in an intensive care unit setting. This is also expensive; the cost of an intensive care unit bed at our institution is approximately $1500 per day versus a cost of $491 for a ward bed. An additional reason for the high cost of lysis relates to its long-term morbidity. The rate of amputation in TOPAS was greater for thrombolysis than for surgery (15% versus 13%). Although not statistically significant, this difference has profound economic implications because of the high cost of rehabilitative and nursing care in survivors of amputation. In the Rochester study, Ouriel and colleagues [19] found an amputation rate that was equivalent for patients treated either with thrombolysis or with surgery. Thus, it is not clear that thrombolysis uniformly results in greater rates of amputation. However, it is important to realize that the costs incurred in a patient who is treated with amputation are significant and small differences in the amputation rate can have a profound effect on the cost-effectiveness of a procedure. One benefit of decision-analytic models is that the assumptions (variables) can be varied and the effect of these variations on the base-case conclusion can be tested. In sensitivity analyses, we evaluate the effect of three variables (the mortality of patients treated with thrombolysis, the amputation rate for patients treated with thrombolysis, and the lifetime costs of thrombolysis) on our conclusion that surgery is a cost-effective alternative to thrombolysis. We found that thrombolysis became cost-effective if its mortality rate diminished from 20 to 10.7%, if the amputation rate was reduced from 15 to 3.9%, or if the cost of thrombolysis at one year diminished dramatically from $49,106 to $13,000. A great reduction in any of these individual variables was necessary to make thrombolysis cost-effective. However, thrombolysis became cost-effective with only minor changes in these variables if all three were simultaneously adjusted. For example, thrombolysis became costeffective if the associated mortality was diminished from 20 to 16%, the amputation rate was diminished from 15 to 13%, and the cost was diminished from $49,106 to $35,000. These sensitivity analyses also demonstrate that if the costs associated with thrombolysis remain high, substantial reductions in morbidity and mortality must be achieved to allow this strategy to become a cost-effective alternative to surgery. We did not adjust our analysis to account for patient preference for either lysis or surgery. The convalescence period after thrombolysis may be shorter than that of a surgical procedure. Therefore, patients might be willing to accept additional risk or society might accept additional cost as a trade-off for this diminished period of convalescence. Techniques are available that
PATEL ET AL.: THROMBOLYSIS AND ARTERIAL OCCLUSION
allow incorporation of patient preference into a decisionanalytic model; however, these techniques require prospective data that were not accrued as part of the TOPAS trial. To determine the cost-effectiveness of thrombolysis for acute lower extremity arterial occlusion, we relied solely on data from the TOPAS trial. Because these data were limited it was necessary that we make several important assumptions. First, since the follow-up from TOPAS was only 1 year, the long-term fate of a large randomized group of patients treated with either surgery or thrombolysis is unknown. We assumed for this analysis that after the first year, the clinical course was stable and equivalent for patients treated with lysis or surgery. We also assumed that there were no additional costs incurred by patients in either group after 1 year. These assumptions were made so as to not bias the analysis in favor of either thrombolysis or surgery. However, only after the long-term consequences of both forms of treatment are available will we truly understand their comparative economic value. A major criticism of the TOPAS trial was the heterogeneity of the patients who were studied [21]. Patients were included who presented with both embolic and thrombotic events as well as native arterial and graft occlusions. Although we found it more costeffective to intervene for acute arterial occlusion with surgery versus lysis, because we relied upon TOPAS for our data, our analysis is also affected by the lack of uniformity in the patient population. The most costeffective treatment of acute lower extremity arterial occlusion may vary with the patient’s clinical presentation. For example, a patient with an undiseased arterial system who embolizes to the common femoral artery as a consequence of atrial fibrillation might be most efficiently treated with a surgical embolectomy under local anesthesia. Alternatively, a patient with a history of claudication and an iliac artery stenosis who presents with an iliac artery thrombosis might be most efficiently treated with lysis and iliac artery angioplasty. More data regarding the outcomes and costs of patients within these individual categories will be required before specific cost comparisons can be made. There is to date no data that demonstrate a superior outcome for patients treated with thrombolysis versus surgery for acute lower extremity arterial occlusion. Furthermore, the costs associated with these two modalities of treatment favor the initial use of surgery. The high cost of lysis, the need for secondary interventions, and the long-term costs of amputation make lysis, at least in the patient cohort treated in TOPAS, an economically unfavorable alternative to surgery. Since patients who present with acute lower extremity arterial ischemia are heterogeneous, we recommend that treatment of these
111
patients still be individualized, with careful attention paid to costs as well as outcomes. REFERENCES 1.
Weaver, F. A., Comerota, A. J., Youngblood, M., Froehlich, J., Hosking, J. D., and Papanicolaou, G. Surgical revascularization versus thrombolysis for nonembolic lower extremity native artery occlusions: Results of a prospective randomized trial. J. Vasc. Surg. 24: 513, 1996.
2.
Ouriel, K., Veith, F. J., and Sasahara, A. A. A comparison of recombinant urokinase with vascular surgery as initial treatment for acute arterial occlusion of the legs. N. Engl. J. Med. 338: 1105, 1998.
3.
Sonnenberg, F. A., and Beck, J. R. Markov models in medical decision making: A practical guide. Med. Decis. Making 13: 322, 1993.
4.
Goldman, L., Gordon, D. J., Rifkind, B. M., Hulley, S. B., Detsky, A. S., Goodman, D. W., Kinosian, B., and Weinstein, M. C. Cost and health implications of cholesterol lowering. Circulation 85: 1960, 1992.
5.
Gold, M. R., Siegel, J. S., Russell, L. B., Weinstein, M. C., and the Panel on Cost-Effectiveness in Health and Medicine. CostEffectiveness in Health and Medicine. New York: Oxford Univ. Press, 1996.
6.
National Center for Health Statistics. Vital Statistics of the United States, 1988, Vol. II, Sect. 6, Life Tables. Washington, DC: U.S. Public Health Service, 1990.
7.
Abou-Zamzam, A. M., Jr., Lee, R. W., Moneta, G. L., Taylor, L. M., Jr., and Porter, J. M. Functional outcome after infrainguinal bypass for limb salvage. J. Vasc. Surg. 25: 287, 1997.
8.
Kalman, P. G., and Johnston, K. W. Predictors of long-term patient survival after in situ vein leg bypass. J. Vasc. Surg. 25: 899, 1997.
9.
Dawson, I., Keller, P. J. A., Brand, R., Pesch-Batenburg, J., and van Bockel, J. H. Late outcomes of limb loss after failed infrainguinal bypass. J. Vasc. Surg. 21: 613, 1995.
10.
Luther, M. Surgical treatment of chronic critical leg ischemia: A five-year follow-up of survival, mobility, and treatment level. Eur. J. Surg. 164: 35, 1998.
11.
Jansen, R. M. G., de Vries, S. O., Cullen, K. A., Donaldson, M. C., and Hunink, M. G. M. Cost-identification analysis of revascularization procedures on patients with peripheral arterial occlusive disease. J. Vasc. Surg. 28: 617, 1998.
12.
Roddy, S. P., O’Donnell, T. F., Iafrati, M. D., Isaacson, L. A., Bailey, V. E., and Mackey, W. C. Reduction of hospital resources utilization in vascular surgery: A four-year experience. J. Vasc. Surg. 27: 1066, 1998.
13.
Gupta, S. K., and Veith, F. J. Inadequacy of diagnosis related group (DRG) reimbursements for limb salvage lower extremity arterial reconstructions. J. Vasc. Surg. 11: 348, 1990.
14.
Vesely, T. M., Idso, M. C., Audrain, J., Windus, D. W., and Lowell, J. A. Thrombolysis versus surgical thrombectomy for the treatment of dialysis graft thrombosis: Pilot study comparing costs. JVIR 7: 507, 1996.
15.
Hunink, M. G. M., Wong, J. B., Donaldson, M. C., Meyerovitz, M. F., de Vries, J., and Harrington, D. P. Revascularization for femoropopliteal disease: A decision and cost-effectiveness analysis. J. Am. Med. Assoc. 274: 165, 1995.
16.
Eckman, M. H., Greenfield, S., Mackey, W. C., Wong, J. B., Kaplan, S., Sullivan, L., Dukes, K., and Pauker, S. G. Foot
112
JOURNAL OF SURGICAL RESEARCH: VOL. 83, NO. 2, MAY 15, 1999
infections in diabetic patients: Decision and cost-effectiveness analyses. J. Am. Med. Assoc. 273: 712, 1995. 17. Raviola, C. A., Nichter, L. S., Baker, J. D., Busuttil, R. W., Machleder, H. I., and Moore, W. S. Cost of treating advanced leg ischemia: Bypass graft vs primary amputation. Arch. Surg. 123: 495, 1988. 18. Kuntz, K. M., and Kent, K. C. Is carotid endarterectomy costeffective? An analysis of symptomatic and asymptomatic patients. Circulation 94: II-194, 1996.
19.
Ouriel, K., Kolassa, M., DeWeese, J. A., and Green, R. M. Economic implications of thrombolysis or operation as the initial treatment modality in acute peripheral arterial occlusion. Surgery 118: 810, 1995. 20. Bell, G. K., Bettman, M. A., Wiener, T., Lucas, C., Krone, S., and McDaniel, M. D. Peripheral vascular disease intervention cost analysis. Circulation 94: I-300, 1996. 21. Porter, J. M. Thrombolysis for acute arterial occlusion of the legs. N. Engl. J. Med. 338: 1148, 1998.