| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Evidence-based Practice |
1 From the Julius Center for Health Sciences and Primary Care (E.B., P.J.N., T.B.v.d.W., Y.v.d.G.) and Depts of Radiology (W.P.T.M.M.), Neurology (L.J.K.), and Vascular Surgery (B.C.E.), Univ Medical Center Utrecht, Heidelberglaan 100, Rm D01.335, 3584 CX Utrecht, the Netherlands; Program for Assessment of Radiologic Technology, Dept of Epidemiology and Biostatistics, and Dept of Radiology, Erasmus MC, Rotterdam, the Netherlands (M.G.M.H.); and Dept of Health Policy and Management, Harvard School of Public Health, Boston, Mass (M.G.M.H.). Received Jun 10, 2003; revision requested Aug 25; revision received Jan 6, 2004; accepted Feb 2. Supported by grant OG/030 from the Dutch Health Insurance Council, mandated by the Ministry of Health, Welfare and Sports. Address correspondence to E.B. (e-mail: e.buskens@umcutrecht.nl).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
MATERIALS AND METHODS: From 1997 through 2000, 350 patients were included in a multicenter blinded consecutive cohort study. The sensitivities and specificities of duplex ultrasonography (US), magnetic resonance (MR) angiography, and these two examinations combined were estimated by using digital subtraction angiography (DSA) as the reference standard. The actual costs (from a societal perspective) of performing imaging and endarterectomy were estimated. The survival, quality of life, and costs associated with stroke were based on data reported in the literature. Markov modeling was used to predict long-term outcomes. Subsequently, a decision model was used to calculate costs, quality-adjusted life-years (QALYs), and incremental costs per QALY gained for 62 examination-treatment strategies. Extensive sensitivity analyses were performed.
RESULTS: Duplex US had 88% sensitivity and 76% specificity with use of conventional cutoff criteria. MR angiography had comparable values: 92% sensitivity and 76% specificity. Combined concordant duplex US and MR angiography had superior diagnostic performance: 96% sensitivity and 80% specificity. Duplex US alone was the most efficient strategy. Adding MR angiography led to a marginal increase in QALYs gained but at prohibitive costs (cost-effectiveness ratio > 
1 500 000 per QALY gained). Performing DSA owing to discordant duplex US and MR angiographic findings and to confirm duplex US and MR angiographic findings led to extra costs and QALY loss owing to complications. Sensitivity analyses revealed that duplex US as a stand-alone examination remained the preferred strategy while estimates and assumptions were varied across plausible ranges.
CONCLUSION: Duplex US performed without additional imaging is cost-effective in the selection of symptomatic patients suitable for endarterectomy. Adding MR angiography increases effectiveness slightly at disproportionately high costs, whereas DSA is inferior because of associated complications.
© RSNA, 2004
Index terms: Carotid arteries, MR, 172.12142, 904.12942 • Carotid arteries, stenosis or obstruction, 172.721, 904.721 • Carotid arteries, US, 172.12981, 172.12984, 904.12981, 904.12984 • Cost-effectiveness • Economics, medical
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
In the NASCET and the European Carotid Surgery Trial, the degree of stenosis was confirmed by performing digital subtraction angiography (DSA), which has since become the standard of reference. DSA, however, is associated with a 4% risk of TIA or minor stroke, a 1% risk of major stroke, and even a small (<1%) risk of death (4). Also, nonapparent infarctions have been identified in patients who did not show a clear neurologic deficit at DSA (5). Exposing all patients to a potentially harmful diagnostic examination to lower the future risk of stroke or death in only a subgroup of patients has questionable benefit. Overall, the potential benefit to symptomatic patients is diminished by the invasive nature of DSA. This diminished benefit has increasingly been recognized during the past decade. At present, noninvasive diagnostic examinations rather than DSA are recommended. In fact, noninvasive examinations are frequently used to select patients for surgery.
Studies of diagnostic duplex ultrasonographic (US) and magnetic resonance (MR) angiographic examinations have been performed, and promising results have been reported (6–12). However, critical comments regarding the substandard results of the noninvasive examination alternatives also have been reported (13). Furthermore, despite the seemingly obvious advantages of performing noninvasive examinations, the actual outcomes after these procedures in terms of strokes prevented, quality-adjusted life-years (QALYs) gained, and cost-effectiveness remain to be determined.
Adding further to the discussion, combination strategies such as duplex US combined with MR angiography are conceivable. The combination of an initial duplex US examination with results indicating severe (70%–99%) stenosis and a subsequent MR angiographic examination with results confirming this finding might be considered sufficient for deciding to perform surgery. In cases of discordant results of duplex US and MR angiography, however, clinicians may feel compelled to perform DSA. Also, in accordance with current standard practice, routine confirmation of severe stenosis by means of DSA is an option. Finally, numerous other examination strategies that take into account variable cutoff points for duplex US and MR angiography are conceivable—that is, they involve the use of alternative flow velocities or degrees of stenosis to define a positive examination result.
The purpose of the current study was to assess the cost-effectiveness of various noninvasive diagnostic imaging strategies in patients who had had a TIA or a minor stroke and were suspected of having significant carotid artery stenosis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
In accordance with the results obtained in the NASCET and the European Carotid Surgery Trial, patients with 70%–99% stenosis (according to NASCET criteria) were treated with CEA (3,15). On the basis of data pertaining to the carotid arteries ipsilateral to the symptoms, the sensitivities and specificities of variable duplex US and MR angiographic criteria positive for severe stenosis were calculated. We also calculated the performance values for combinations of these examinations.
Current Cost Analysis Study
Parallel to the clinical study (14), the current investigation was focused on our estimation of the costs of the diagnostic examinations—including hospitalization, if applicable—and of CEA from a societal perspective, originally in terms of 1998 euros (in 1998, 1.00 equaled 2.20371 Dutch florins, which equaled approximately $1.11). Actual costs, which included the expenditures for personnel, equipment, materials, maintenance, housing, cleaning, administration, and overhead, were estimated (16). Cost data were recorded at one university hospital (University Medical Center Utrecht) and one general hospital (Enschede Medical Center, Enschede, the Netherlands). On the basis of the distribution of patients across various types of hospitals in the Netherlands, weighted average costs were calculated. Additional estimates of costs associated with hospitalization and stroke in the Dutch setting were derived from data published in the literature (17).
The actual short- and long-term costs associated with minor and major strokes, including those for diagnostic work-up, medication, hospitalization, rehabilitation, and nursing home admission, have been previously reported (18–21). The costs associated with a TIA and with death were estimated by using input from experts in the field and taking into account the costs of consultations with a general practitioner and/or a specialist, diagnostic testing, therapeutic procedures, and hospitalization.
Cost-effectiveness Analysis
In this study, we considered patients who had had a TIA or a minor stroke and at initially performed duplex US with use of a low peak systolic velocity threshold were suspected of having carotid artery stenosis. The cost-effectiveness of various examination-treatment strategies was compared by using a decision model created with a computer software program (Data 3.5; TreeAge Software, Williamstown, Mass). Short-term outcomes were based on the clinical study results (14). Long-term outcomes were estimated by combining published clinical trial data with age- and sex-specific Dutch survival statistics (1–3,22,23).
The major strategies evaluated were defined according to the diagnostic examinations or combinations of examinations performed. Sometimes MR angiography was not feasible: For example, in our clinical study (14), 11.5% of the eligible patients had contraindications to MR angiography or claustrophobia and thus underwent another imaging examination. For most strategies involving the use of MR angiography, we assumed that these patients would undergo DSA instead. We also assessed a strategy in which the duplex US results alone determined the indication for surgery in patients who were unable to undergo MR angiography and did not undergo DSA.
When the combination strategies yielded discordant results—that is, the duplex US and MR angiographic findings were not the same—MR angiography was considered the decisive examination or DSA was performed, depending on the strategy definition. All single-examination and combination strategies were also considered in combination with DSA to either rule in or rule out severe (ie, 70%–99%) stenosis.
Substrategies were defined according to variable threshold criteria for positive duplex US and MR angiographic results and two criteria for endarterectomy. For the individual noninvasive examinations, we considered two thresholds for positive results—a strict positivity criterion and a lenient positivity criterion—and two criteria for the indication for surgery—70%–99% stenosis and 50%–99% stenosis. Thus, four possible strategies per noninvasive examination strategy were considered. For combinations of duplex US and MR angiography, we considered either a lenient positivity criterion for duplex US followed by a lenient or strict positivity criterion for MR angiography, or strict criteria for both duplex US and MR angiography (three combinations), and again the two described criteria as indications for surgery. Thus, six strategies per combination of noninvasive examinations were considered. Finally, for DSA, we considered the two criteria for the indication for surgery. Thus, a total of 62 diagnostic strategies were analyzed.
The degree of carotid artery stenosis was categorized as 0%–49%, 50%–69%, or 70%–99% stenosis or as occlusion. Also, the initial conditions (TIA or minor stroke) and complications that occurred as a result of DSA or surgery were modeled. Next, a Markov process model with a 1-year cycle length was developed to extrapolate and evaluate the long-term outcomes of the diagnostic work-ups and subsequent treatments offered. The initial Markov health state reflected the outcome after the diagnostic work-up and treatment and was determined on the basis of survival length, stenosis degree, treatment given (nonsurgical vs endarterectomy), and neurologic disease (TIA, minor stroke, or major stroke as complication or finding at presentation) severity. Disease progression and death were modeled by allowing the health condition of patients to advance to more severe states during follow-up in the Markov model.
For each year (ie, one cycle) spent in a given health state, the associated quality-adjusted equivalent of 1 year and the associated costs during 1 year were accumulated. In accordance with current Dutch guidelines, the time preference was accounted for by using a 4% discount rate for costs and effects (24).
In symptomatic patients, besides optimal medical care, including aspirin therapy, surgery is generally considered to be indicated for treatment of 70%–99% stenosis. According to relatively recent study results (3), however, patients with 50%–69% stenosis—particularly men who have had a TIA or a minor stroke—may also expect limited benefits from surgery. For lower grade (<50%) stenoses, optimal nonsurgical care alone is recommended.
In the Markov model, optimal nonsurgical care was always assumed to have been given, whereas we considered two criteria as indications for endarterectomy: 70%–99% stenosis and 50%–99% stenosis. In addition, patients who had had a minor stroke were assumed to incur costs and to have ongoing limited disability as a result of the initial event. Therefore, the initial event—that is, TIA or minor stroke—was also distinguished in the model. Moreover, prognosis depends on the extent of the underlying vascular disorder. In symptomatic patients with low-grade stenosis, the risk of future events is lowest, whereas in patients with high-grade stenosis, the risk is highest, particularly during the period early after the initial event.
The associated risk estimates used in the model were derived from data published in the literature (3), and if these data were not reported in detail, they were refined with the help of experienced clinicians and scientists. In the Appendix (A1 and A2), we present details about the risks of stroke over time and about the risk differences used to update survival probabilities.
Analyses
With respect to future cerebrovascular events, CEA will predominantly affect the prognosis related to the symptomatic carotid artery. Accordingly, only ipsilateral cerebrovascular events were modeled. Fatal contralateral cerebrovascular events and other cardiovascular mortalities, however, were accounted for by adjusting age- and sex-specific mortality rates for the Dutch population by using a disease-specific rate ratio (23). Utility weights for the health states of well, minor stroke, major stroke, and death—1.0, 0.8, 0.2, and 0.0, respectively—were derived from data published in the literature (25,26). We assumed that the disutility for TIA was equivalent to 2 days with major stroke.
The Markov model was used to estimate long-term outcomes in terms of life-years, QALYs, and costs associated with all possible health states, with the assumption that these outcomes would pertain to a symptomatic 55-year-old male patient. The results obtained by using the prognostic Markov model were used as input for the diagnosis and treatment decision model to perform a comprehensive long-term comparison of the diagnostic strategies (Fig 1). The expected lifetime costs and QALYs related to the various health states that occurred after the diagnostic strategies were compared by using the decision model. Similarly, incremental cost-effectiveness ratios for the successive strategies were calculated.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
|
Additionally performing DSA in cases of discordant examination results—that is, when the results of one of the noninvasive examinations indicated the presence of a greater degree of stenosis than the degree threshold indication for surgery and the results of the other examination indicated the presence of a degree of stenosis below this threshold—also led to improved sensitivity: to 98.4%. However, in the scenario in which DSA was performed after duplex US and MR angiography yielded discordant results, in addition to the 11.5% of patients who were unable to undergo MR angiography, an additional 16.0% of patients had to undergo DSA because of discordant results. With other strategies, varying proportions of patients—up to 75% or 100% of the patients considered for surgery in the reference strategy—had to undergo DSA (Table 1).
Corresponding calculations of the diagnostic performance of individual examinations performed with the assumption of a 50% stenosis threshold for the indication for surgery resulted in small (ie, up to about 5%) decreases in sensitivity but substantial (ie, approximately 25%) increases in specificity. The effects on the diagnostic characteristics of the combination strategies were smaller: a 1% decrease in sensitivity and a 10% increase in specificity.
A final finding was that with all of the noninvasive examinations (ie, duplex US, MR angiography, and combination examinations), there tended to be a slight overestimation of the stenosis degree compared with the stenosis degree determined by using DSA. This finding implies that when DSA is not used to confirm noninvasive examination results, some patients—in the worst case, up to about 26%—might be considered "overtreated" (Table 1). Additional details on the probabilistic data used in the diagnostic decision model are presented in the Appendix (A3 and A4).
Cost Analysis
The costs for duplex US, MR angiography, DSA, and CEA estimated in 1998 at the university hospital and the general hospital did not differ substantially (Table 2). The actual costs associated with stroke and subsequent rehabilitation were retrieved from the literature (17–21). The costs incurred after TIA or death were estimated with the help of experts in the field (Table 2).
|
|
|
|
|
Sensitivity Analysis
We took into account the variability of criteria used to indicate positive examination results and of criteria indicating the need for surgery in a sensitivity analysis of coherent combinations of sensitivity, specificity, and indication to operate. Within the major categories of the diagnostic strategies, varying the above criteria resulted in a move along the negatively sloped lines observed in Figures 2–4. Figure 4 illustrates this finding for the strategy of performing duplex US alone. Using both a lower diagnostic threshold and a lower threshold indication to perform surgery would lead to lower costs and improved health outcomes. The rank of strategies in terms of incremental cost-effectiveness ratios, however, would remain the same.
Additionally performed sensitivity analyses revealed that the results were robust. The ranges for the sensitivity analysis are presented in Table 4. The utilities for minor and major stroke were varied by ±0.1, and this variability did not alter the order of preference. Similarly, the age of symptomatic patients was varied between 45 and 65 years, and the discount rate was varied between 0% and 10%, and these variances had no effect on the order of preference. The cost estimates of the diagnostic examinations were varied by a factor of 0.5–2.0, and these variances did not alter the order of preference among the strategies. As expected, increasing the cost of MR angiography resulted in even more unfavorable incremental cost-effectiveness ratios. Likewise, varying the costs of surgery did not cause a substantial altering of the results. However, with high surgery costs (>$6550 per procedure), strategies involving the use of higher threshold values for positive examination results—that is, those involving the use of strict positivity criteria implying lower sensitivity and higher specificity—became slightly more favorable.
|
The proportion of patients who were unfit or unwilling to undergo MR angiography, which varied between 3% and 30%, had no effect on the rank of the strategies. At high estimates of this range of proportions, the incremental cost-effectiveness ratio for the combination strategies became increasingly unfavorable. Similarly, the proportion of strokes classified as major had little effect. Only at implausibly high proportions of major stroke (>50%) did the combination strategy reach an acceptable incremental cost-effectiveness ratio (<$27 750 per QALY gained).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
Another important finding was that the use of various positivity criteria had limited effect on the overall results of the study. At higher stenosis thresholds, specificity increased, with the result being a lower proportion of patients who underwent surgery. Since patients with moderate-grade stenosis may benefit from surgery, high thresholds for both the diagnostic examinations and the policy to perform surgery actually yield less favorable cost-effectiveness ratios. Accordingly, warnings regarding the misclassifications and overtreatments that may result from using a "nonperfect" examination (ie, examination that, according to strict criteria to perform surgery for 70%–99% stenosis, would result in overtreatment) in symptomatic patients may not be clinically relevant (28).
Also, sensitivity analyses of the likelihood of complications resulting from DSA or surgery, and the cost estimates in general, revealed that the rank of the strategies was barely altered. The conventional strategy to always perform DSA as the final examination after duplex US was clearly inferior. Various other combined examination strategies, including performing DSA in cases of discrepancy between duplex US and MR angiographic findings or to confirm or rule out severe stenosis, also were inferior.
A comparable study by Kent et al (8) was published before the results regarding intermediate-grade (50%–69%) stenoses and the long-term results from the NASCET became available. The authors used a less complex model, assessed fewer strategies, and based the examination diagnostic characteristics on findings in a small sample of patients. Kent et al found that a combination strategy of duplex US and MR angiography, including DSA performed in cases of discordant findings, was favorable. However, they assumed a constant high risk of stroke in patients with severe stenosis—and consequently projected overly optimistic long-term benefits of endarterectomy—and underestimated the benefits for patients with intermediate-grade stenoses. Both the assumed high stroke risk and the underestimated benefits call for high specificity and justify the small risk associated with the diagnostic work-up. Thus, their acceptable incremental cost-effectiveness ratio for the combination strategy, including DSA, can be explained.
An interesting finding that is not often observed was the apparent negative correlation between costs and QALYs across the strategies. The ranks of the diagnostic strategies were determined on the basis of the prevention of incident (ie, major) strokes. Strokes are associated with both high costs and a reduced quality of life, which explain the negative correlation. In the analysis of costs and life-years, strokes still direct the costs incurred, but their effect on longevity is less apparent, leading to a scattered, noncorrelated distribution of costs and effects.
A point of surgical and technical interest that we cannot substantiate or reject on the basis of the available data pertains to anatomy—for example, that of the extracranial or intracranial part of the carotid arteries. Surgeons who perform surgery on the basis of duplex US results do so without detailed anatomic information. Some may argue that this information has no relevance in terms of the surgical approach or outcome, whereas others may challenge this opinion. The fact is that in all of the CEAs performed as a part of this study, the surgeons had detailed anatomic information because DSA was routinely performed. Single-examination duplex US will remain the optimal examination only if neither the complication rate nor the effectiveness is altered in the absence of anatomic information. We have not come across publications addressing this issue specifically; however, at present, numerous surgeries are performed solely on the basis of duplex US information.
With regard to the reliability of the above findings and the associated inferences made, we believe that although considerable extrapolation occurred while using the modeling techniques described in the Methods section, the results do not indicate a specific "sensitive" parameter that when used alters the ranking of strategies. The described diagnostic study was performed by using a large representative and consecutive sample of symptomatic patients. Moreover, the diagnostic values of the examinations were calculated on the basis of findings in the ipsilateral (symptomatic) carotid artery only, and, thus, the overestimation of specificity that would have resulted from also including the asymptomatic carotid arteries was avoided.
Some limitations of our study should be recognized. The included patients were selected on the basis of duplex US findings. Accordingly, the prevalence of severe stenosis and the diagnostic values of the examinations might be biased because not all of the symptomatic patients underwent MR angiography and DSA. Selective verification could have resulted in an overestimation of the sensitivity, particularly that of duplex US but also that of MR angiography, and an underestimation of the specificity. However, because these possible results apply to both duplex US (in this study, with duplex US considered a stringent "final" examination) and MR angiography, it is very likely that the examination characteristics would have changed proportionally. Accordingly, the ranking of the diagnostic strategies in terms of clinical outcomes would not have been affected. Because cost estimates were not affected, the ranking of cost-effectiveness ratios remained stable. Similarly, the combination of duplex US and MR angiography would still be associated with high additional costs that might well be considered unacceptable.
Also recall that the use of less stringent positivity criteria (which are associated with higher sensitivity and lower specificity) results in a move toward lower costs and better outcome along the cost-effectiveness line. A comparable but opposite effect would have been observed had we been able to obtain unbiased estimates of sensitivity (somewhat lower) and specificity (higher). The strategies that included DSA were inferior as a result of complications. Adjusting for verification bias would not have altered the results obtained in this respect.
Finally, we would like to stress that although strictly speaking the presented examination diagnostic values may not be correct, they do apply to the population of interest—that is, patients who are preselected and considered for surgery after undergoing an initial duplex US examination. We argue that in spite of possibly imperfect point estimates of examination diagnostic values, interpretations would remain the same. Also, a relatively recent report by a U.S. group who used data from a comparable setting supports our findings and adds to the credibility and robustness of our study results (28).
A limitation of our analyses is that our event-free survival estimates—for both the patients treated nonsurgically and those who underwent surgery—were based on trial results. The treatment results in a trial setting generally are better than those in routine practice, and this is especially true for nonsurgical therapies. However, since such "trial bias" would have affected all diagnostic strategies in a similar way, it is unlikely that it affected our cost-effectiveness results in a major way.
An additional limitation might be the fact that unrelated medical and nonmedical costs, such as the costs associated with and the effects of nonfatal contralateral strokes, were not accounted for. We reason, however, that regardless of whether endarterectomy was performed, the long-term prognosis associated with contralateral strokes or other cardiovascular outcomes would not have been altered. This reasoning obviously does not apply to periprocedural complications, but these were taken into account. We recognize that high-grade contralateral stenosis and associated high-competing morbidity and mortality might reduce the yield from CEA. We reason that this effect was accounted for, however, because differential cardiovascular risks based on the ipsilateral stenosis grade were incorporated. Moreover, in an incremental comparison of strategies, the long-term outcomes that are equivalent across strategies will "drop out" of the equation.
U.S. cost estimates may be different from European and particularly Dutch cost estimates. However, a more or less global increase in costs, in accordance with differences in U.S. and Dutch cost estimates, would not change the overall conclusions. In fact, if anything, the incremental cost-effectiveness ratio for the combination strategy would probably be even higher, and this higher ratio would strengthen our argument that adding MR angiography to the work-up is not cost-effective. Furthermore, our estimates reflect actual costs rather than reimbursements or charges, such as those available from Medicare, which may differ considerably. Moreover, our sensitivity analysis of cost estimates revealed that the costs of diagnostic examinations can vary across a wide range of settings, including U.S. clinical settings, without affecting the overall ranking of strategies.
Finally, we chose not to perform an extensive sensitivity analysis in the prognostic model. The results obtained in the prognostic Markov model were used as input for the diagnostic model. We believe this was an acceptable simplification because the underlying estimates used in the prognostic model are based on the results of large-scale multicenter trials. Taking into account the very certain trial results would have added substantially to the complexity of the modeling. However, one would not expect a change in overall results. We also did not perform any multivariable probabilistic sensitivity analyses and therefore cannot present formal measures of uncertainty. However, given the unambiguous results of the current univariate sensitivity analyses, this simplification seems justified.
An issue that has not been validated and that was alluded to earlier in this article is the usefulness of a strict stenosis grade threshold for surgery. Results from the NASCET (3) indicate that this threshold has limited benefit, particularly for male symptomatic patients with 50%–69% carotid stenosis who have had a TIA or minor stroke. This conclusion is based on results that show favorable cost-effectiveness ratios with use of a 50% stenosis threshold as compared with those associated with the use of a 70% stenosis threshold. Sensitivity analyses specifically addressing the complication rates of endarterectomy, however, revealed that when periprocedural stroke rates surpass 10%, the high stenosis threshold will yield more favorable outcomes. This finding is in accordance with previously expressed concerns that the advantages of surgery may be outweighed by complications (3,29). CEA, as well as the more recently available percutaneous interventions, should be performed only in centers that are able to maintain a stable and low complication rate (30). In our opinion, this requires continued monitoring of complication rates and specific (ie, local) deliberations regarding optimal diagnostic and treatment strategies.
Finally, the emergence of contrast material–enhanced MR angiography should be addressed. Around the time that our study began, time-of-flight MR angiography was considered a state-of-the-art examination. By enabling one to avoid so-called flow voids, which at times may result in poor-quality images and difficulties in determining the stenosis grade, contrast enhancement could contribute to even higher accuracy of MR angiography, especially in terms of specificity. Yet, as our analysis results indicated, such a gain in accuracy would result in only a marginal gain in QALYs. Furthermore, contrast enhancement would increase the cost of MR angiography substantially, and, therefore, a very unfavorable incremental cost-effectiveness ratio would be expected.
We conclude that the use of duplex US as a single-examination strategy results in an optimal trade-off between costs and effectiveness and that only if society is willing to pay extraordinarily high sums of money per QALY gained should one consider performing MR angiography additionally. We believe that owing to the risk of complications, DSA should no longer be routinely performed in the process of selecting patients for CEA.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
|
|
|
|
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, E.B., W.P.T.M.M., L.J.K., B.C.E., Y.v.d.G., M.G.M.H.; study concepts and design, E.B., W.P.T.M.M., L.J.K., B.C.E., Y.v.d.G., M.G.M.H.; literature research, E.B.; clinical studies, P.J.N.; data acquisition, P.J.N., E.B., T.B.v.d.W.; data analysis/interpretation, E.B., P.J.N.; statistical analysis, E.B.; manuscript preparation and editing, E.B.; manuscript definition of intellectual content, E.B., P.J.N., Y.v.d.G., M.G.M.H.; manuscript revision/review, E.B., W.P.T.M.M., L.J.K., B.C.E., Y.v.d.G., M.G.M.H.; manuscript final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P. Greving, E. Buskens, H. Koffijberg, and A. Algra Cost-Effectiveness of Aspirin Treatment in the Primary Prevention of Cardiovascular Disease Events in Subgroups Based on Age, Gender, and Varying Cardiovascular Risk Circulation, June 3, 2008; 117(22): 2875 - 2883. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Sassi and M. McKee Do clinicians always maximize patient outcomes? A conjoint analysis of preferences for carotid artery testing J Health Serv Res Policy, April 1, 2008; 13(2): 61 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Saba, G. Caddeo, R. Sanfilippo, R. Montisci, and G. Mallarini CT and Ultrasound in the Study of Ulcerated Carotid Plaque Compared with Surgical Results: Potentialities and Advantages of Multidetector Row CT Angiography AJNR Am. J. Neuroradiol., June 1, 2007; 28(6): 1061 - 1066. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. R. van Beek and D. E. Malone Evidence-based Practice in Radiology Education: Why and How Should We Teach It? Radiology, June 1, 2007; 243(3): 633 - 640. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. F. Layton, J. Huston 3rd, H. J. Cloft, T. J. Kaufmann, K. N. Krecke, and D. F. Kallmes Specificity of MR Angiography as a Confirmatory Test for Carotid Artery Stenosis: Is It Valid? Am. J. Roentgenol., April 1, 2007; 188(4): 1114 - 1116. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. R. Bates, C. J. D. Babb, D. E. Casey, C. U. Cates, G. R. Duckwiler, T. E. Feldman, W. A. Gray, K. Ouriel, E. D. Peterson, K. Rosenfield, et al. ACCF/SCAI/SVMB/SIR/ASITN 2007 Clinical Expert Consensus Document on Carotid Stenting: A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents (ACCF/SCAI/SVMB/SIR/ASITN Clinical Expert Consensus Document Committee on Carotid Stenting) Vascular Medicine, February 1, 2007; 12(1): 35 - 83. [PDF]
|
|
|
|
 |
|
 American Society of Interventional & Therapeutic N, Society for Cardiovascular Angiography and Interve, Society for Vascular Medicine and Biology, Society of Interventional Radiology, E. R. Bates, J. D. Babb, D. E. Casey Jr, C. U. Cates, G. R. Duckwiler, T. E. Feldman, et al. ACCF/SCAI/SVMB/SIR/ASITN 2007 Clinical Expert Consensus Document on Carotid Stenting: A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents (ACCF/SCAI/SVMB/SIR/ASITN Clinical Expert Consensus Document Committee on Carotid Stenting) J. Am. Coll. Cardiol., January 2, 2007; 49(1): 126 - 170. [Full Text] [PDF]
|
|
|
|
|
|
M. Lell, C. Fellner, U. Baum, T. Hothorn, R. Steiner, W. Lang, W. Bautz, and F.A. Fellner Evaluation of Carotid Artery Stenosis with Multisection CT and MR Imaging: Influence of Imaging Modality and Postprocessing AJNR Am. J. Neuroradiol., January 1, 2007; 28(1): 104 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Nikolaou, H. Kramer, C. Grosse, D. Clevert, O. Dietrich, M. Hartmann, P. Chamberlin, S. Assmann, M. F. Reiser, and S. O. Schoenberg High-Spatial-Resolution Multistation MR Angiography with Parallel Imaging and Blood Pool Contrast Agent: Initial Experience Radiology, December 1, 2006; 241(3): 861 - 872. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. Heijenbrok-Kal, E. Buskens, P. J. Nederkoorn, Y. van der Graaf, and M. G. M. Hunink Optimal Peak Systolic Velocity Threshold at Duplex US for Determining the Need for Carotid Endarterectomy: A Decision Analytic Approach Radiology, December 21, 2005; (2005) 2381041078. [Abstract] [Full Text]
|
|
|
|
 |
|
 M. H. Heijenbrok-Kal, P. J. Nederkoorn, E. Buskens, Y. van der Graaf, and M.G. Myriam Hunink< |
