Cost-effectiveness analysis of elective endovascular repair compared with open surgical repair of abdominal aortic aneurysms for patients at a high surgical risk: A 1-year patient-level analysis conducted in Ontario, Canada
Article Outline
Background
Abdominal aortic aneurysm (AAA) is a prevalent health condition affecting up to 14% of men and 6% of women. The objective of this study was to estimate the cost-effectiveness and cost-utility of elective endovascular aneurysm repair (EVAR) compared with open surgical repair (OSR) in patients at a high risk of surgical complications.
Methods
Patient-level cost and outcome data from a 1-year prospective observational study conducted at London Health Sciences Centre, London, Ontario, Canada, was used to determine the incremental cost per life-year gained and the incremental cost per quality-adjusted life year (QALY) gained of EVAR compared with OSR in patients with an AAA >5.5 cm and a high risk of surgical complications. The analysis was taken from a societal perspective and the time horizon was 1 year. To measure sampling uncertainty on costs and effects, nonparametric bootstrap techniques were applied. Uncertainty results were expressed using cost-effectiveness acceptability curves. Extrapolations of the 1-year results to a 5-year time horizon were conducted in sensitivity analyses.
Results
Between August 11, 2003, and April 3, 2005, 192 patients at a high risk of surgical complications were enrolled: 140 received EVAR and 52 OSR. Point estimates during a 1-year period showed that EVAR dominated OSR for high-risk patients in terms of incremental cost per life-year gained and incremental cost per QALYs. However, bootstrap estimates for the two cost-effectiveness measures indicated there was a great deal of uncertainty regarding the costs and the QALYs and less uncertainty regarding life-years gained. If society was willing to pay $50,000 per life-year gained or per QALY gained, the probability of EVAR being cost-effective was found to be 0.76 and 0.55, respectively. Five-year extrapolations indicated that EVAR was cost-effective compared with OSR.
Conclusions
According to this 1-year observational study, EVAR may be a cost-effective strategy compared with OSR for high-risk patients. Longer-term data are needed to decrease the uncertainty associated with the results.
Open surgical repair (OSR) is currently the primary method of repair of abdominal aortic aneurysm (AAA) in Canada; however, in some jurisdictions, endovascular aneurysm repair (EVAR) is becoming the predominate method of managing AAA.1, 2 The Canadian Society for Vascular Surgery (CSVS) “recommends that EVAR should be the procedure of choice for patients with suitable vascular anatomy who are at intermediate or high risk (6%-10%) for perioperative morbidity or death with open repair. For patients at low risk (2%-4%), open repair remains the current standard…”3 The guidelines also emphasize that all treatment options should be discussed with the patient before a treatment decision is made.
Because EVAR is more expensive but potentially more effective than OSR, several economic evaluations have been conducted in recent decades. In 2007 for example, Jonk et al4 reviewed 20 economic articles and concluded that compared with OSR, EVAR was not cost-effective. However, none of these studies evaluated the cost-effectiveness of EVAR compared with OSR for patients at a high risk of postoperative complications.
The recent economic evaluations conducted by Epstein et al5 and Prinssen et al,6 both published in 2007, used mainly data from the EVAR trial 1 (EVAR1) and Dutch Randomized Endovascular Aneurysm Management (DREAM) trials, respectively, and did not provide any new information on the comparative value of EVAR and OSR in high-risk patients. However, results of a 4-year nonrandomized prospective trial conducted in Ireland among high-risk patients recently suggested similar outcomes and lower costs when EVAR in 66 patients was compared with OSR in 52.7 Unfortunately, the uncertainty associated with the data (eg, sampling variability) was not dealt with in that study, which may limit our confidence in the results. This article presents a Canadian cost-effectiveness analysis of EVAR against OSR for patients at a high surgical risk, using data from a 1-year observational study conducted in Ontario at London Health Science Centre (LHSC).
The LHSC Endovascular Program, launched in December 1997, has been previously described.8 The LHSC endovascular team currently consists of four experienced endovascular surgeons, a dedicated interventional radiologist, specially trained operating room nurses, anesthesiologists, and radiology technicians. The patients are initially assessed by the vascular/endovascular surgeon, and the preoperative planning for each patient is done by the surgeon and the radiologist.
Postoperatively, the patients are recovered in the regular postanesthetic recovery unit before being transferred to the regular vascular surgery ward before discharge home. In terms of postprocedural surveillance, EVAR patients have a computed tomography (CT) scan and are seen by the surgeon at 1, 3, 6, and 12 months postoperatively. No routine follow-up is conducted with OSR patients beyond a single postoperative visit at 4 to 6 weeks, unless necessary. Indications for remedial therapy are a new type 1 endoleak, type 3 endoleak, or a persistent type 2 endoleak with an enlarging aneurysm sac.
Evaluation of three scoring systems for selecting patients for OSR or EVAR were recently reported on the basis of 310 EVAR and 561 OSR patients who underwent elective repair of an infrarenal aortic aneurysm, by the LHSC team, from September 1999 to December 2004.9 The results of this retrospective database analyses confirmed that objective scoring systems may help identify patients at a high risk for OSR but a low risk for EVAR and patients who may not benefit from EVAR (low risk). The following presents a cost-effectiveness analysis of EVAR compared with OSR in high-risk patients using data from a 1-year prospective observational study conducted at LHSC.
Methods
Study participants
All patients requiring elective repair of an intact AAA >5.5 cm were invited to participate in this 1-year nonrandomized, prospective observational study conducted at LHSC between August 11, 2003, and April 3, 2005. The study was conducted on an intention-to-treat principle, and received ethics approval by the University of Western Ontario Ethics Review Board, London, Ontario, Canada.
Treatment algorithm
The method of intervention to repair AAA was determined according to LHSC clinical criteria for EVAR and OSR and by discussion with the patient. Individuals not willing to accept surgical options were offered best medical treatment. Patients considering surgical options were evaluated, and their surgical risk and suitability was assessed. Patients were considered high risk according to the extent of their comorbidities, which were objectively assessed using the American Society of Anesthesiologists (ASA) and Society for Vascular Surgery/International Society for Cardiovascular Surgery (SVS/ISCVS) score.10, 11, 12
According to this algorithm, the choice of EVAR or OSR was presented as surgical options for patients anatomically suitable for EVAR. The repair of AAAs was completed using OSR for low-risk patients or high-risk patients not anatomically suitable for EVAR. For comparability purposes, only OSR high-risk patients were included in this economic evaluation because all EVAR study patients were high-risk patients.
Resource utilization and costs
The costs related to the initial hospitalization were directly derived from the LHSC patient-specific data-costing system. Procedural costs (including the cost of the endograft) and nonprocedurally related costs were itemized. Subsequent postoperative resource utilization data, such as hospital admissions, physician visits, procedures, and medications, were obtained for each patient at 30 days and every 3 months for 1 year through the administration of economic questionnaires administered by a study coordinator. To capture productivity losses after AAA repair, patients were asked to indicate the mean number of paid days they took off work as well as the average number of hours of care provided by others. Unit costs of health care resource utilization for the initial hospitalization and diagnostic tests were derived from the LHSC and from the Ontario Schedule of Benefits for Physician Fees. Costs were expressed in 2006 Canadian dollars. The Canadian average hourly wage was used to determine the cost of productivity losses.13
Effectiveness measures
Vital status and cause of death were collected during the 1-year study period. The mean number of life-years during the 1-year study period was estimated using Kaplan-Meier survival curves, and the observed life-years for EVAR and OSR for high-risk patients were calculated as the area under each survival curve.
Patients' utilities were calculated at baseline and at regularly scheduled intervals using results from the European quality of life (EuroQol) instrument (EQ-5D), which was included in the study questionnaires to measure health-related quality of life. The EQ-5D includes a visual analogue scale and five questions related to mobility, self-care, usual activity, pain/discomfort, and anxiety/depression, for which answers can be used to generate a utility value ranging from 0 (death) to 1 (perfect health). Health utility summary scores for the EQ-5D were estimated using the quality-adjusted survival for EVAR and OSR in high-risk patients by combining Kaplan-Meier survival curves with utility estimates over time.
Because of the observational nature of this trial, regression techniques were used to adjust utilities to take into account potential differences in the mean baseline utility between EVAR and OSR. At each assessment point (discharge, 1, 3, 6, 9, and 12 months), ordinary least squares regressions were used to estimate the mean change in utility from baseline for the two treatment groups. In each regression model (eg, 1, 3 months), patient-specific utility values were used as the dependant variable. Patient baseline utility value and a treatment indicator were used as independent variables, allowing at each assessment point the calculation of an estimate of the mean change from baseline. Utility values were then calculated at each time period assuming a common baseline utility value of 0.77 for each group, the mean baseline utility value.
Cost-effectiveness and statistical analyses
Cost-effectiveness analyses of EVAR compared with OSR were performed in terms of incremental cost per life-year gained and incremental cost per quality-adjusted life-year (QALY) gained. Incremental cost-effectiveness ratios were not calculated if one treatment strategy dominated the other (ie, lower costs, better outcomes). The analysis was taken from a societal perspective and the time horizon was 1 year.
To measure uncertainty on costs and effects due to sampling variability associated with the trial, nonparametric bootstrap techniques were applied allowing the calculations of 95% confidence intervals (CIs). Bootstrapping consists of drawing a sample with replacement and equal of size of the original data set (eg, 140 for EVAR and 52 for OSR). For this new sample, the mean costs and effects associated with each group are calculated to derive the incremental cost-effectiveness ratios. This process is repeated several times (eg, hundreds of times) to determine the incremental cost-effectiveness ratio and the 95% CIs associated with the differences in costs and effects between EVAR and OSR. The bootstrap differences in cost and effect pairs were also plotted using a cost-effectiveness plane to get a better understanding of the sampling distribution of the cost and effect pairs of EVAR compared with OSR in high-risk patients. Uncertainty results were also expressed using cost-effectiveness acceptability curves to show for several threshold values (eg, willingness to pay $100,000 to save 1 year of life) the probability that EVAR is cost-effective compared with OSR, when uncertainty is taken into account. To test differences between the two groups in terms of baseline characteristics, costs and outcomes, statistical significance was conducted using χ2 tests for categoric variables and t tests for continuous variables.
Sensitivity analyses
The long-term impact of the lower operative mortality of EVAR on costs and outcomes may be underestimated due to the 1-year time horizon of the study; therefore, modeling techniques were used to extend the time horizon 5 years in sensitivity analysis. Because patients in the study may be at increased risk of death compared with the general population, the 1-year within-trial mortality rate for subjects alive after 30 days was first compared with Canadian life tables of 75-year-old Canadian men. According to this analysis, the risk of death in study patients was 30% higher compared with the general population of 75-year-old men. This relative risk (ie, 1.3) was applied to the Statistics Canada mortality rates during the course of the extrapolation to account for the increased comorbidity and risk of death in the study population.
Routine follow-up costs for EVAR patients were assigned in the time horizon sensitivity analyses. From the LHSC experience, it was assumed that EVAR patients would have two CT scans and specialist consults annually for the first 2 years after initial treatment and one CT scan and specialist consult yearly thereafter. No routine follow-up cost for OSR patients was assumed for the extrapolations, according to LHSC clinical practice.
Reintervention rates of 5%, 10%, and 20% were applied in sensitivity analyses assuming that 75% of reinterventions would be embolizations and 25% would be complete EVAR reinterventions, according to LHSC clinical experience. The average cost of embolization was assumed to be $2000 on the basis of LHSC costs. The observed study EVAR costs were assumed for the EVAR reinterventions. The same utility rate was applied to both treatment groups in the sensitivity analyses. The utility rate was set equal to the lowest average observed 12-month utility rate between EVAR and OSR patients. A discount rate of 5% was applied to all costs and outcomes beyond 1 year.
The 5-year extrapolation analysis was run under three mortality scenarios: (1) cumulative mortality converges after 2 years; (2) cumulative mortality converges after 3 years; and (3) cumulative mortality converges after 5 years. Specifically in these calculations, survival for EVAR patients alive at the end of the trial was extrapolated over 5 years using adjusted age-specific survival rates from Canadian life tables. To run the three scenarios, the cumulative survival rate for OSR patients was forced to converge to the cumulative survival rate estimated for EVAR at 2 years (scenario 1), 3 years (scenario 2), and 5 years (scenario 3).
Results
Patients and baseline characteristics
Between August 11, 2003, and April 3, 2005, 351 patients with an AAA >5.5 cm required elective AAA repair, and 342 patients participated in this study. Of the 192 patients classified as high risk, 140 were treated with EVAR and 52 with OSR, of whom four (8%) were anatomically suitable for EVAR and chose OSR. Seven patients were treated with best medical treatment.
As reported in Table I, the two groups were similar in terms of age, smoking status, AAA diameter, SVS/ISCVS score, ASA grade, history of myocardial infarction, congestive heart failure (CHF), angina, renal function, and chronic obstructive pulmonary disease (COPD). The Leiden Risk Score9 indicated they had a predicted mortality rate of approximately 7%, and the two groups were comparable (Table I). The proportion of EVAR procedures completed in men was greater (85.7% vs 73.1%, P = .04). The EVAR and OSR patients were similar in other cardiac (eg, valvular heart disease, 15.7% vs 9.6%; P = .28), vascular characteristics (eg, stroke, 12.9% vs 5.8%; P = .16), and diagnosed diabetes (eg, 19.3% vs 19.2%, P = .99).
Table I. Baseline patient characteristics
| Variable | EVAR | OSR | Pa |
|---|---|---|---|
| Patients, No. | 140 | 52 | |
| Age, mean (SD) years | 75.6 (7.8) | 74.0 (7.9) | .24 |
| Male gender, % | 85.7 | 73.1 | .04 |
| Work full or part time, % | 5.0 | 6.0 | .73 |
| Smoking status, % | .07 | ||
 Current | 22.8 | 34.6 | |
| Ever | 63.6 | 61.5 | |
| Never | 13.6 | 3.9 | |
| AAA size, mean (SD) cm | 6.2 (0.9) | 6.5 (1.0) | .10 |
| SVS/ISCVS grade, % | .97 | ||
| I | 34.3 | 34.6 | |
| II | 65.7 | 65.4 | |
| ASA grade, % | .69 | ||
| I | 0 | 0 | |
| II | 1.4 | 0 | |
| III | 32.1 | 33.3 | |
| IV | 66.5 | 66.7 | |
| MI <6 months previous, % | 2.1 | 3.9 | .61 |
| MI >6 months previous, % | 43.9 | 40.4 | .66 |
| Congestive heart failure | 9.3 | 9.6 | .28 |
| Angina | 35.7 | 42.3 | .40 |
| Abnormal renal function | 1.4 | 0 | >.99 |
| COPD | 35.7 | 41.2 | .49 |
| Leiden Score (raw), mean (SD) | 9.8 (6.2) | 7.8 (8.1) | .10 |
| Leiden Score (% predicted mortality), mean (SD) | 6.9 (4.3) | 7.2 (10.0) | .76 |
aEVAR vs OSR. |
Initial hospitalization and postoperative outcomes
From admission to discharge, EVAR patients spent significantly less time in the hospital than OSR patients (7.7 days vs 16.1 days, respectively), which was significant. Specifically, average lengths of stay were 7.0 days for EVAR and 11.2 days for OSR in the absence of postoperative complications and 16.7 days for EVAR and 24.0 days for OSR in the presence of complications. Overall, the median length of stay was 6 days for EVAR and 11 days OSR, and the postoperative length of stay was 5.7 days for EVAR and 14.4 for OSR. Intensive care unit admission was required for 4% EVAR patients compared with 31% of OSR patients. The 30-day mortality rates were 9.6% for OSR and 0.7% for EVAR, and this difference was statistically significant. The differences in postoperative complications were primarily attributable to higher rates of complications in the OSR group (Table II).
Table II. Initial hospitalization and postoperative outcomes
| Variable | EVAR | OSR | P |
|---|---|---|---|
| Patient, No. | 140 | 52 | |
| Initial hospitalization | |||
| LOS, mean (SD) days | 7.7(5.8) | 16.1(16.0) | <.01 |
| LOS, median (range) days | 6.0(4-36) | 11.0(6-92) | |
| ICU admission, % | 3.6 | 30.8 | <.01 |
| ICU LOS, mean (SD), days | 0.2(1.7) | 3.2(8.3) | <.01 |
| Post-op complications (30 days), % | |||
| Death | 0.7 | 9.6 | <.01 |
| Myocardial infarction | 4.3 | 9.6 | .17 |
| CHF/pulmonary edema | 3.4 | 17.3 | <.01 |
| Arrhythmia | 3.6 | 9.6 | .14 |
| Stroke | 0.7 | 0 | >.99 |
| Renal failure | 3.6 | 11.5 | .07 |
| Pneumonia | 0 | 7.7 | <.01 |
| Sepsis | 0 | 5.8 | .02 |
Resource utilization and costs
Table III presents the total average 1-year cost of EVAR and OSR for high-risk patients by main categories, including initial hospitalization costs, follow-up medical costs, and productivity costs. The total average initial costs of hospitalization were $28,139 for EVAR and $31,181 for OSR, a difference not statistically significant. Total average procedural costs, including the cost of the endograft, were statistically significantly higher for the EVAR patients ($18,326) then for the OSR patients ($6162), mainly because EVAR patients spent statistically less time in hospital than OSR patients and had fewer complications. The average nonprocedurally related costs were, however, statistically significantly greater for the OSR patients ($25,029) then for the EVAR patients ($9813).
Table III. Total average 1-year costs by treatment group
| Costs | EVAR (n = 140) | OSR high-risk (n = 52) | EVAR vs OSR high-risk |
|---|---|---|---|
| Initial hospitalization costs | |||
| Procedural | $18,326 | $6162 | $12,164a |
| Nonprocedural | $9813 | $25,029 | −$15,216a |
| Subtotals initial hospitalization | $28,139 | $31,181 | −$3042 |
| Follow-up medical costs | |||
| Hospital admissions | $2318 | $1250 | $1069 |
| Tests and procedures | $1372 | $90 | $1282a |
| Emergency department | $115 | $38 | $77* |
| GP visits | $349 | $300 | $49 |
| Specialist visits | $272 | $145 | $127a |
| Other health care professionals | $746 | $348 | $407 |
| Subtotal follow-up cost | $5172 | $2171 | $3010a |
| Total health care costs | $33,311 | $33,352 | −$32 |
| Follow-up productivity costs | |||
| Mean paid days taken off of work | $523 | $433 | $90 |
| Mean hours of care provided by others | $311 | $385 | −$74 |
| Subtotal | $835 | $818 | $17 |
| Total | $34,146 | $34,170 | −$24 |
aIndicates significance at the 5% level. |
In the absence of postoperative complications during the initial hospitalization (EVAR, 131 of 140; OSR, 30 of 52), the mean hospitalization cost of EVAR was approximately $10,000 more than OSR in these high-risk patients ($26,985 vs $17,411), a difference close to the average cost of the graft. Talent (W. L. Gore and Assoc, Flagstaff, Ariz) and Zenith (Cook, Bloomington, Ind) grafts were used in this study, with unit costs of $9600 to $10,200. In the presence of postoperative complications, the costs associated with the initial hospitalization were similar between OSR ($49,942) and EVAR ($44,942).
The average 1-year medical cost of follow-up was statistically higher in the EVAR group ($5181) than in the OSR group ($2171), because of more frequent rehospitalizations, tests, procedures, and physician visits in the EVAR group. No reinterventions occurred within the 1-year study period. One OSR patient required treatment for an abdominal wall hernia. Indirect costs associated with days off work and hours of care provided by others were estimated at $837 for EVAR and $818 for OSR, which were 2% of the total costs. Overall, the total 1-year costs were calculated at $34,146 for EVAR and $34,170 for OSR high-risk patients, an almost negligible difference of $24.
Life-years gained
In our high-risk population, an increased all-cause mortality rate of 17.3% was observed in OSR patients compared with 7.1% in EVAR individuals at 1 year (P = .04). The estimated life-years gained, as determined from the Kaplan Meier survival curves, were 0.959 for EVAR and 0.848 for OSR high-risk patients (Fig 1).
Fig 1.
Kaplan-Meier survival is shown for endovascular aneurysm repair (EVAR, blue line) and open surgical repair (OSR, green line) in high-risk patients at up to 365 days of follow-up.
Utilities and QALYs
When adjusted by baseline values, the patients' EQ-5D utility was lower at discharge compared with baseline utilities and then increased over time for the two treatment groups, as shown in Fig 2, A. However, although EVAR patients returned to a level similar to their baseline values, the utilities after 1 year were higher than the baseline values for OSR patients, as shown in this figure. The resulting quality-adjusted survival curves, when adjusted utilities over time were adjusted for survival over time, are presented in Fig 2, B. The resulting QALYs were calculated to be 0.713 for EVAR and 0.688 OSR high-risk patients.
Fig 2.
A, Mean utility scores for the EuroQol quality of life instrument (EQ-5D) and (B) estimated quality-adjusted life-years (QALYS) are shown for high-risk patients undergoing endovascular aneurysm repair (EVAR, blue lines) and open surgical repair (OSR, green lines) during 365-days of follow-up.
Cost-effectiveness results
The trial point estimates indicated that EVAR has slightly a lower 1-year cost of $24 (Table III) and provides more benefits. However, the 95% CIs associated with the differences in costs ranged from –$11,582 to $9165 when calculated using bootstrap techniques. In terms of outcomes, EVAR had 0.111 more life-years compared with OSR for high-risk patients (95% CI, 0.022-0.213). More QALYs (0.025) were also associated with EVAR, but the differences were not statistically significant (95% CI –0.075 to 0.128). On the basis of point estimates only, EVAR dominated OSR in terms of incremental cost per life-year gained and incremental cost per QALYs (Table III). However, as indicated by the CIs, the differences in the costs and QALYs were not significant at the 5% level.
The 1-year costs and effects pairs generated by the bootstrap resampling of the trial data are presented in Fig 3. Here, costs are plotted on the y axis (“EVAR more/less expensive”) against QALYs (Fig 3, A) or life-years gained (Fig 3, B) represented on the x axis (“EVAR more/less effective”). The origin represents OSR, the treatment of reference. As shown by these cost-effectiveness planes, there was considerable uncertainty regarding the costs and the QALYs (Fig 3, A) and less uncertainty regarding life-years gained (Fig 3, B) because there was an almost 100% chance that EVAR would save additional lives compared with OSR. For each quadrant of the cost-effectiveness planes, the probability that EVAR was either dominant (less costly and more effective), dominated (more costly and less effective), or involved a trade-off vs OSR (more costly and more expensive or less costly and less effective) is given along with the bootstrap representation of uncertainty. Each dot in these graphs represents a bootstrap sample.
Fig 3.
Incremental cost and effect pairs for endovascular aneurysm repair (EVAR) vs open surgical repair (OSR) are shown for (A) incremental costs and quality-adjusted life-years (QALYs) and (B) incremental costs and life-year (LY) gained. Each dot represents a bootstrap sample.
Fig 4 presents the cost-effectiveness acceptability curves for cost per life-year gained and QALY gained, respectively. Not suggesting any particular threshold, but it may be worthwhile to consider two commonly quoted thresholds of $50,000 and $100,000 per QALY gained.14 The way to interpret these curves is to consider a threshold that decision makers might be willing to pay for a unit of effect, that is, willingness to pay per QALY gained, along with the horizontal axis and read along the vertical axis the probability that the treatment is cost-effective after accounting for uncertainty. If society was willing to pay $50,000 per life-year gained, the probability of EVAR being cost-effective was 0.76 compared with OSR in high-risk patients (Fig 4, A). This probability increased to 0.9 when a threshold of $100,000/life-year gained was used. In terms of incremental cost per QALY, the probability of EVAR being cost-effective was lower, with probabilities of 0.55 and 0.58 when using thresholds of $50,000 and $100,000, respectively (Fig 4, B).
Fig 4.
Cost-effectiveness acceptability curves are shown for (A) incremental costs per life-year (LY) gained and (B) incremental cost per quality-adjusted life-year (QALY) gained.
Sensitivity analyses
The results of the sensitivity analyses examining several mortality scenarios when the data were extrapolated an additional 5 years indicated that EVAR was cost-effective. The most favorable cost-effectiveness occurs when it is assumed that survival converges after 5 years and the EVAR reintervention rate was 5%. In this scenario the incremental cost-effectiveness ratios of EVAR against OSR were $10,167/QALY gained and $5584/life-year gained. The assumption least favorable to EVAR was convergence of cumulative survival between EVAR and OSR after 2 years along with a reintervention rate of 20%. With this assumption, the incremental cost per QALY gained and life-year gained were $38,720 and $14,968, respectively.
Discussion
In this 1-year observational study, a significant reduction in mortality was associated with EVAR compared with OSR in patients at a high risk of surgical complications. The 30-day mortality rates observed in this study were comparable with the rates reported in a recent review of randomized controlled trials for EVAR (0.7% vs 1.6%)15 but higher for OSR (9.6% vs 4.7%),15 which may be caused by the high-risk nature of the patients included in our study. Our definition of high-risk patients was similar to the definition used by Sicard et al16 in their analysis of high-risk patients from the investigational device exemption trials, including age ≥60 years, AAA ≥5.5 cm, and at least one of the following comorbidities: symptomatic CHF, valvular heart disease, cardiac arrhythmia, COPD, chronic renal failure, or serum creatinine value >2.6 mg/dL. However, our 30-day and 1-year mortality rates were lower for EVAR and higher for OSR. As noted by Nagpal et al,17 however, the absence of a standardized reporting between studies of high-risk patients makes comparisons difficult. Almost two-third of patients in our study were at SVS/ISCVS grade II and ASA grade 4. In both groups, the baseline Leiden predicted mortality rate was 7%, which is also aligned with the Canadian guidelines regarding the use of EVAR and the definition of patients at an intermediate or high risk of risk of postoperative morbidity or death (6% to 10%).
Although it is difficult to compare the findings of our 1-year prospective observational study with randomized clinical trials of EVAR, the pattern of the utility scores over time observed in our study (Fig 2, A) was similar to results of the DREAM trial.6p886 When compared with the DREAM trial, our results indicated that high-risk patients undergoing EVAR experienced the same number of QALYs (0.71 vs 0.72 in DREAM), whereas the OSR population derived less QALYs (0.69 vs 0.72, respectively). The QALY differences were not significant in either study. Another Canadian study indicated no differences in health-related quality of life between EVAR and OSR.18
Consistent with other evaluations, EVAR patients had a shorter hospital and ICU length of stay than OSR patients. The relatively long length of stay observed in our study is a reflection of the Ontario health system and the high-risk patient population included in the study. In Ontario and in Canada in general, postoperative care occurs within the primary care setting, and transfer to rehabilitation facilities is infrequent. An earlier study of 552 patients undergoing nonelective or elective AAA repair in four Canadian hospitals between 1997 and 2000 reported 90% were discharged to self-care vs 66% for the United States (US). Almost one-third of US patients were either discharged to home care (16.0% vs 0.5% in Canada) or to institutional care (16.8% vs 8.0% in Canada).19 In their study, the median length of stay for AAA repair was 9.0 days in Canada compared with 7.0 days in the United States. In comparison, we observed median lengths of stay of 6 days for EVAR and 11 days for OSR. Based on a relatively small number of patients, previous evaluations of EVAR and OSR in Canada have reported average lengths of stay of 5 days for EVAR and 11 days for OSR,18, 20 which is also consistent with our findings. Although our two groups had comparable baseline characteristics and risk assessments (ie, Leiden Score), more postoperative complications were observed in the OSR group, leading to longer hospitalizations, more ICU admissions, and higher hospitalization costs. Despite the additional cost of the endograft, the 1-year health-related costs for EVAR patients were found to be nearly identical to the costs for OSR patients.
On the basis of the point estimate, the EVAR treatment group was found to be dominant in terms of cost-effectiveness (ie, cheaper and more effective). However, uncertainty regarding cost-effectiveness was found when bootstrap techniques were used with the 1-year data. These differences between EVAR and OSR were not statistically significant in terms of costs and QALYs, but a significant mortality benefit effect was observed for EVAR.
The 1-year economic analysis indicated that if society or decision makers were willing to pay $100,000/life-year gained, the probability of EVAR being cost-effective was 0.9. Five-year extrapolations indicated that EVAR might be cost-effective. However, the extrapolations do not account for quality of life differences or cost-effects associated with long-term comorbidities such as stroke, renal failure, and myocardial infarction and therefore should be interpreted accordingly. It is also important to note that this sensitivity analysis was meant to extrapolate the 1-year mortality rates observed in our trial conducted amongst high-risk patients. As such, sensitivity analyses with different 1-year mortality rates were not conducted. We did not consider scenarios with different costs for endovascular devices for the initial procedure or the reinterventions because we assumed that the prices would remain constant.
Several limitations were associated with our study. This prospective, nonrandomized study evaluated consecutive patients and compared outcomes in patients with a high surgical risk of death and postoperative complications. As with all nonrandomized studies, there may be concern about the comparability of treatment groups (ie, selection bias). This concern was partially addressed through the risk stratification of patients, and results indicated that the baseline demographics, comorbidities, and Leiden scores were similar between the two treatment groups. We also adjusted for baseline utility values in our QALY calculations to account for the nonrandomized aspect of the study.
The potential effect of anatomic suitability for EVAR on outcomes was not examined within this study because all patients treated with EVAR were anatomically suitable for the procedure. Almost none of the OSR patients were anatomically suitable for EVAR, which may limit the comparability of the results. Although the baseline characteristics of our two groups were similar, our OSR patients might have had a more complicated anatomy, which could also explain the high mortality rate observed in the OSR group. However, for these high-risk OSR patients willing to undergo surgery and not anatomically suitable for EVAR, OSR is the only available option. The relatively small number of patients in the OSR group compared with the EVAR group remains a concern, but sampling variability was addressed with nonparametric bootstrap techniques, and cost-effectiveness acceptability curves were used to present uncertainty.
The results may not be generalizable to lower-risk patients or randomized clinical trials evaluating mixed-risk patient populations, as are commonly found in published economic studies comparing EVAR with OSR. In addition, the results of the study are based on a single center in Ontario, and this should be considered if results are to be applied to other jurisdictions. LHSC was the primary referral center for EVAR in Ontario during the study period; thus, the proportion of low- and high-risk patients observed in our study may not reflect the case mix of other hospitals in Canada or elsewhere.
Finally, patient recall was used to collect follow-up information, which may have introduced some bias. Although we evaluated the long-term costs and effects of EVAR and OSR in high-risk patients, the 5-year extrapolation results were based on point estimates and, as such, these results should be interpreted with caution.
Conclusions
Despite these limitations, our results provide a second insight in the cost-effectiveness of EVAR and OSR in high-risk patients. Patients in this study were treated according to current medical management of AAA patients at LHSC, in which those at a low risk of surgery are treated with OSR. As such, this cost-effectiveness analysis is not comparable with previous economic evaluations that included patients at low risk. In contrast to these studies, but in agreement with results from Ireland,7 our results indicated that EVAR may be a cost-effective strategy for patients at a high risk of surgical complications and anatomically suitable for EVAR. Follow-up is still being conducted with the study participants, and the future analyses of midterm outcomes will provide more information on the midterm costs and effects of EVAR and OSR in these high-risk patients.
Author contributions
We would like to thank all the patients who participated in this study, without whom this research could not have been accomplished. We would also like to acknowledge the Ontario Health Technology Advisory Committee, Ontario Ministry of Health and Long-Term Care (MOHLTC), for assistance and helpful comments during the conduct of this study. We are also grateful to the following individuals: Dr Les Levin, Dr Birthe Jorgensen, and Shirley Lee at the Medical Advisory Secretariat, MOHLTC, for their support of the study; Dr Thomas L. Forbes, Dr. Kenneth A. Harris and Dr. D. Kirk Lawlor from the Division of Vascular Surgery, Department of Surgery at London Health Sciences Centre, for their assistance and clinical input; and Randy Welch and Jennifer McCallum from London Health Sciences Centre, for their assistance with the management and retrieval of the case-costing and clinical data. Special thanks to Dr Feng Xie, Christine Henderson, and Jan Watson from the Programs for Assessment of Technology in Health (PATH) Research Institute, St. Joseph's Healthcare Hamilton.
References
- A statewide experience with endovascular abdominal aortic aneurysm repair: rapid diffusion with excellent early results. J Vasc Surg. 2004;39:10–19
- . Endovascular repair compared with open surgical repair of abdominal aortic aneurysm: Canadian practice and a systematic review (Technology Report 33). Ottawa, ON, Canada: Canadian Coordinating Office for Health Technology Assessment; 2002;
- . Canadian Society for Vascular Surgery consensus statement on endovascular aneurysm repair. CMAJ. 2005;172:867–868
- . Cost-effectiveness of abdominal aortic aneurysm repair: a systematic review. Int J Technol Assess Health Care. 2007;23:205–215
- Modelling the long-term cost-effectiveness of endovascular or open repair for abdominal aortic aneurysm. Br J Surg. 2008;95:183–190
- Cost-effectiveness of conventional and endovascular repair of abdominal aortic aneurysms: results of a randomized trial. J Vasc Surg. 2007;46:883–890
- . A prospective clinical, economic, and quality-of-life analysis comparing endovascular aneurysm repair (EVAR), open repair, and best medical treatment in high-risk patients with abdominal aortic aneurysms suitable for EVAR: the Irish patient trial. J Endovasc Ther. 2007;14:763–776
- . Establishing an endovascular abdominal aortic program–decisions, decisions, decisions: The London Health Sciences Centre experience. J Vasc Nurs. 2001;19:10–13
- . Objective scoring systems of medical risk: a clinical tool for selecting patients for open or endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2007;45:1102–1108
- Identifying and grading factors that modify the outcome of endovascular aortic aneurysm repair. J Vasc Surg. 2002;35:1061–1066
- . Guidelines for the treatment of abdominal aortic aneurysms (Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery). J Vasc Surg. 2003;37:1106–1117
- . Recommended indications for operative treatment of abdominal aortic aneurysms (Report of a subcommittee of the Joint Council of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery). J Vasc Surg. 1992;15:1046–1056
- Statistics Canada. http://www.statcan.ca/2006;
- . How attractive does a new technology have to be to warrant adoption and utilization? (Tentative guidelines for using clinical and economic evaluations). CMAJ. 1992;146:473–481
- . Systematic review of recent evidence for the safety and efficacy of elective endovascular repair in the management of infrarenal abdominal aortic aneurysm. Br J Surg. 2005;92:937–946
- . Endovascular abdominal aortic aneurysm repair: long-term outcome measures in patients at high-risk for open surgery. J Vasc Surg. 2006;44:229–236
- Midterm results of endovascular infrarenal abdominal aortic aneurysm repair in high-risk patients. Vasc Endovascular Surg. 2007;41:301–309
- Pain and quality of life assessment after endovascular versus open repair of abdominal aortic aneurysms in patients at low risk. J Vasc Interv Radiol. 2005;16:1093–1100
- In-hospital cost of abdominal aortic aneurysm repair in Canada and the United States. Arch Intern Med. 2003;163:2500–2504
- . A cost-effectiveness analysis of standard versus endovascular abdominal aortic aneurysm repair. Can J Surg. 2002;45:420–424
Competition of interest: none.
This project was funded by the Ontario Ministry of Health & Long-term Care (Contract No. 06129) to address the 2002 recommendations of the Ontario Health Technology Advisory Committee regarding EVAR. The final study results were presented to the Ontario Health Technology Advisory Committee on December 15, 2006. Daria O'Reilly and Jean-Eric Tarride each hold a 2007 Career Scientist Award, Ontario Ministry of Health and Long-Term Care.
PII: S0741-5214(08)00867-7
doi:10.1016/j.jvs.2008.05.064
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
