Review of Recent US Cost Estimates of Revascularization

Objective: To review recent US cost estimates ofrevascularization and discuss their implications forthird-party payers.

Study Design and Methods: A literature reviewwas performed using MEDLINE. The review was limitedto English-language articles published betweenJanuary 2000 and September 2003. The most recentlypublished articles that included US-derived clinicaloutcomes and costs of revascularization wereselected for review. Cost estimates were abstractedand updated to 2003.

Results: Coronary revascularization procedures,including percutaneous coronary interventions(PCIs) and coronary artery bypass graft (CABG), arecommonly performed in the United States. Theseprocedures are costly. Costs for PCI in single-vesseldisease are lower than costs for PCI in multivesseldisease. Although initial estimated costs are lowerfor multivessel PCI (with or without stenting) thanCABG, longer-term costs and lifetime costs are similar.Drug-eluting stents have the potential to altertreatment and economics dramatically, although it istoo early to draw definitive conclusions about theircosts. It is imperative that cost comparisons beplaced in the appropriate context.

Conclusion: Revascularization procedures arecostly, and short-term cost differences in proceduresmay not exist when considered long term.Importantly, recent cost data may be conservativegiven the rapid innovation in revascularization proceduresand technology and the lag in publication ofcost data reflecting these advances.

(Am J Manag Care. 2004;10:S370-S376)

Cardiovascular disease (CVD) exactsan enormous clinical and economictoll on the population of the UnitedStates. CVD affects 64.4 million Americansand ranks as the leading cause of death inthe United States, accounting for 38.5% ofmore than 2.4 million total US deaths in2001.¹ In 2004, the predicted direct andindirect cost of CVD is $368.4 billion.

Coronary revascularization is commonlyperformed in the United States on a varietyof patient populations who have CVD. In2001, an estimated 1.051 million angioplastyprocedures were performed in the UnitedStates, including 571 000 percutaneoustransluminal coronary angioplasties (PTCAs)performed on 559 000 patients, and 475 000stenting procedures.¹ From 1987 to 2001,the number of PTCA procedures increased266%. In addition, 516 000 coronary bypassprocedures were performed on 305 000patients in the United States.

charges

cost

These procedures are costly. Data fromthe 2001 National Healthcare Cost andUtilization Project show mean of$28 558 for PTCA and $60 853 for coronaryartery bypass graft (CABG).¹ Estimates of include more than $30 000 per procedurefor CABG and approximately $12 000per percutaneous coronary intervention(PCI) procedure,² giving total US direct costsfor coronary revascularization of more than$28 billion annually.

While the clinical advantages and disadvantagesof both PCIs and CABG are welldocumented in both randomized and nonrandomizedtrials, older published studyoutcomes and cost and cost-effectivenessestimates may not be based on current treatmentapproaches because of the rapidimprovements in PCI and CABG technologyand procedures.^2-4 PCIs, including PTCA,atherectomy, laser angioplasty, and implantationof intracoronary stents and othercatheter devices (including the use ofplatelet inhibitors) are undergoing continuousrefinement and modification. CABG procedureshave also advanced. Treatmentchoice today is complex, and cost and cost-effectivenessare important considerations.Incorporating the most current economicand clinical information is critical to makinginformed treatment decisions. To gain a betterunderstanding of the cost of revascularizationprocedures, we examined some ofthe more recent literature on relevant costs.

Current Costs of Revascularizationin the United States

It is critical that the current costs andcost effectiveness of revascularization beanalyzed in the context of specific patient subpopulationsand the alternative treatmentsavailable. Treatments that may demonstratelower costs and/or higher value in one contextmay not provide either lower costs orhigher value relative to available options inanother context.^2-4 The literature wereviewed generally reports cost data in thecontext of either single-vessel or multivesseldisease and often compares PCI (with orwithout stenting) with CABG. Because ofthis, we report findings in 2 major categories:PCI with stenting in single-vessel diseaseand PCI versus CABG in multivesseldisease. Finally, we report the limited availablecost data on drug-eluting stents.

PCI With Stenting in Single-Vessel Disease

Two recently published articles evaluatedstenting in single-vessel disease. Neil andcolleagues⁵ compared the economic andhealth status impacts of primary coronarystenting versus optimal PTCA with provisionalstenting using the resource utilizationdata from 479 patients randomly assignedbetween 1996 and 1998 in the OptimalPTCA Utilization versus Stent (OPUS-I)trial. The study showed that primary stentingin patients undergoing single-vesselangioplasty reduced the 6-month compositeincidence of death, myocardial infarction(MI), cardiac surgery, or target vessel revascularizationcompared with the strategy ofinitial PTCA with provisional stenting.Itemized hospital cost estimates, exclusiveof professional fees, for procedure- and nonprocedure-related costs were derived fromprimary hospital charge data gathered in 2previous multicenter, prospective clinicaltrials evaluating coronary stents. Costs werederived by adjusting charges using Medicarecost to charge ratios specific to each center.Procedure-related costs included the costs ofequipment and use of the catheterizationlaboratory. Nonprocedure-related costs werecalculated using a multivariate log-linearregression. Costs of initial hospitalizationwere higher in the primary stent group thanprovisional stenting (Table 1). Mean 6-monthcosts were similar between the groups, witha slightly reduced cost associated with primarystenting. The authors concluded that,for patients undergoing single-vessel angioplasty,routine stenting improves clinicaloutcomes at a comparable cost versus angioplastywith provisional stenting.

The most recent study that describes thecosts of PCI with stenting in single-vesseldisease is a retrospective economic analysisof the Enhanced Suppression of the PlateletIIb/IIIa Receptor with Integrilin Therapy(ESPRIT) trial.⁶ ESPRIT was a randomized,double-blind, crossover-permitted trial ofthe study drug compared with placebo. Theeconomic analysis included 2064 patients.In-hospital costs were estimated for eachpatient on the basis of hospital resourceconsumption, occurrence of adverse events,and treatment received. Procedural costswere estimated based on resource consumption,using a published regression model.Length of stay and adverse event costs wereestimated from a second regression model,which was based on hospital costs for 3241patients who underwent PCI at 89 US hospitals,as part of a number of PCI multicenterclinical trials. Cost data from these trialswere obtained using a variety of standardtechniques, including bottom-up cost accounting,conversion of charges to costsbased on department-level cost to chargeratios, and the Medicare fee schedule forprofessional services. The study found mediantotal initial hospital costs of $10 452(Table 1) for PCI with planned stent implantation(placebo group). Studies are underway to determine longer-term costs.

PCI Compared With CABG inMultivessel Disease

Two recent articles have analyzed long-termcosts of PTCA versus CABG for multivesseldisease. Weintraub and colleagues⁷analyzed costs in the Emory Angioplasty versusSurgery Trial (EAST), a randomized trialthat compared, by intent-to-treat analysis, theclinical outcomes and costs of PTCA andCABG for multivessel coronary artery diseasein patients without prior coronary arteryrevascularization. The single-center studyincluded 392 randomly assigned patients and450 registry patients, entered between 1987and 1991 and followed up annually for 8 years.For randomized patients with comparable diseasewho were suitable for revascularizationby CABG or PTCA, there was an initial clinicaladvantage to CABG in angina status butlong-term outcomes were similar. For registrypatients, who had their care directed by theirclinician, sicker patients (eg, patients whowere older and patients with 3-vessel CAD)were treated with CABG and less sick patientswere treated with PTCA. Patients in the randomizedand registry groups were similaroverall in baseline characteristics and longtermoutcome. Costs were assessed for initialhospitalization, cumulative costs of the initialhospitalization, and additional revascularizationprocedures. CABG and PTCA mean costsof initial hospitalization at 3 and 8 years areshown in Table 2. Despite a large initial costadvantage to PTCA in the randomly assignedgroup, this advantage is reduced by 3 yearsand not discernable by 8 years because additionalrevascularization procedures arerequired for patients treated with PTCA. Inthe registry group, costs for PTCA werelower than for CABG, but these procedureswere used to treat different patients, asnoted above. The authors comment thatlower costs in the registry group at 3 and 8years suggest that physician decision makingcan contribute to lower costs. Thesecost data illustrate the observation that therelative costs of PTCA and CABG dependon the population concerned (eg, randomlyplaced vs registry populations).

Yock and colleagues⁸ modeled the costeffectiveness of surgical and angioplasty-basedrevascularization for multivessel diseasein patients without prior coronaryartery revascularization. Outcomes and costdata from the Study of Economics andQuality of Life (SEQOL), a substudy of therandomized, multicenter Bypass AngioplastyRevascularization Investigation (BARI), wereupdated to reflect technology changes sincethe time of enrollment for BARI. Lifetimecosts and quality-adjusted life-years (QALYs)were projected using a Markov decisionmodel. Total lifetime costs for CABG withprovisional stenting in follow-up angioplastyprocedures, CABG without stenting in follow-up angioplasty procedures, CABG withprimary stenting in follow-up angioplastyprocedures, initial angioplasty with provisionalstenting, and initial angioplasty withprimary stenting are shown in Table 2. Theauthors found that despite short-termadvantages with stents, the improvementsare not sufficient to make primary stentingless costly and more effective than CABG forrelieving angina in this patient population.

Data from a recent study showed a significantdifference in stenting versus CABGcosts both in initial hospitalization and at2 years. In the single center, retrospective,matched cohort study, Reynolds and colleagues⁹ compared the clinical and economicoutcomes of multivessel stenting (n = 100)and CABG (n = 200, n_Total = 300) over amedian of 2.8 years of follow up (enrollmentbetween 1994 and 1998). Initial hospitalizationcosts and total costs at 2 years of multivesselstenting versus CABG (Table 2)demonstrate a significant cost advantage tomultivessel stenting over CABG at 2 years offollow up (27% relative reduction), despite ahigher rate of revascularization in the stentgroup. The 2 treatment strategies resulted incomparable risks of death and MI for thestudy period. The authors note that, whileclinical findings were consistent with previouslyreported randomized trials, cost differenceswere more robust. The authorsattribute this, at least partially, to theirstudy being more reflective of current USpractice patterns at the time. The authorsnote that recent developments in both stentingand CABG procedures were incorporatedinto the study to a limited extent becauseof the study's relatively long-term follow-up.

Drug-Eluting Stents

Drug-eluting stents have garnered significantattention because of their potential fordecreasing restenosis and their increasedcosts over bare metal stents.^10-12 Drug-elutingstent cost data from a substudy of theSirolimus-Eluting Stent in Coronary Lesions(SIRIUS) trial were presented at the 52ndAnnual Scientific Session of the AmericanCollege of Cardiology (2003).^11,12 In SIRIUS,1058 patients (42% with multivessel disease)with a newly diagnosed native coronaryartery lesion were randomly assigned to PCItreatment of the target lesion with either theCypher sirolimus-eluting stent or a baremetal stent (1 or 2 stents per target lesion).¹³The substudy analyzed the costs and costeffectiveness of these procedures. The economicend points of the substudy includedin-hospital and 1-year costs. The cost of thesirolimus-eluting stent was considered to be$3000 compared with $1000 for each baremetal stent. Median costs for the initial procedurefor the sirolimus-eluting stent were$2857 higher than for the bare metal stent(Table 3). However, because of the differencein the need for repeat revascularizationprocedures, 12-month follow-up costsfor bare metal stents were $2571 higherthan sirolimus-eluting stents, resulting inonly a slightly higher total 1-year cost forsirolimus-eluting stents (a difference of$309). The investigators concluded that thesirolimus-eluting stent is highly cost effectivefor the trial's target population and that longerstents and improved technique shouldenhance the stent's cost effectiveness.

A recent economic model of stent use forWilliam Beaumont Hospital's fiscal year 2003budget showed that the hospital will lose $3.8million as it adopts drug-eluting stents.¹⁰ Keymodel assumptions included the following:commercial availability in April 2003, stentavailability in 2.5- to 4.0-mm diameters, 1.43stents per case (current usage), diagnosis-relatedgroup revenue increase of $1800 fordrug-eluting stent codes, 10% reduction insurgical volume, 50% reduction in coronaryrestenosis interventions, $3500 per drug-elutingstent, and 50% usage. The anticipated lossstems from fewer repeat revascularizationprocedures plus fewer CABGs resulting in lessrevenue, while at the same time supply costsincrease because of the costs of drug-elutingstents. It is acknowledged that cost concernshave slowed acceptance of drug-eluting stentsin Europe and that a similar situation mayoccur in the United States.¹⁰

Of particular interest to payers may be amodel developed to project the impact ofdrug-eluting stents on individual hospitalbudgeting.¹⁴ Based on PTCA and CABG costand reimbursement data for 214 000patients from 198 hospitals in 7 states andon expectations that drug-eluting stents willdecrease repeat PTCA procedures, be widelyadopted, and reduce the use of CABG, it wasestimated that managed care organizationscould provide "budget neutral" rate relief tohospitals of approximately $3600 per drug-elutingstent case.

Discussion

The most recently published articles onthe costs of revascularization have beenreviewed on the basis of outcomes and costsderived in the United States. Costs for PCI insingle-vessel disease are lower than costs forPCI in multivessel disease. Although initialestimated costs are lower for multivessel PCI(with or without stenting) than CABG,longer-term costs and lifetime costs are similar.Drug-eluting stents have the potential toalter treatment and economics dramatically,although it is too early to draw definitiveconclusions about their costs.

Revascularization procedures, includingCABG and PCI with or without stenting, arecostly, and new technologies have the potentialto affect managed care organizations substantially.However, cost data are not onlydifficult to interpret but must also be placedin a suitable context, particularly if comparisonsare to be made. It is important that thecontext of any cost analysis for revascularizationprocedures include the patient population,the length of follow up, the treatmentpractices in the country or region under consideration,the type of study on which theoutcomes are based, and the current state-of-the-art treatment, among others.^2-4,6-9

The most obvious examples herein liewithin the Weintraub et al study.⁷ Patientsin the randomized and registry groups weresimilar overall in baseline characteristicsand long-term outcome, although within theregistry group sicker patients were morelikely to undergo CABG. PTCA registrygroup mean costs are lower than those forthe PTCA randomized group for initial hospitalization,total 3-year costs and total 8-year costs, while the opposite is true for theCABG groups. This is likely due to patientselection bias, as sicker registry patientswere treated with CABG, making CABG registrycosts higher than CABG randomizedcosts, and less sick registry patients weretreated with PTCA, making PTCA registrycosts lower than PTCA randomized costs.Additionally, while registry group meancosts for PTCA are 48% to 58% of those ofCABG at the time intervals, the long-termcosts in the randomized groups are comparableat 8 years likely due to repeat revascularizationsnecessary for patients undergoinginitial PTCA.

The cost-related findings of Weintrauband colleagues may be at odds with the findings of Reynolds and colleagues.⁹ AlthoughReynolds and colleagues found no significantdifferences in mortality outcomes over amedian 2.8-year follow up and that PCIpatients required more frequent repeatrevascularization, the difference in costs ofmultivessel stenting versus CABG was statisticallysignificant at initial hospitalizationand at 2 years in favor of multivessel stenting.It is possible that Reynolds and colleagues'stenting and CABG costs wouldconverge over a longer term, as was the casewith the randomized results from Weintrauband colleagues, or that advances such asminimally invasive CABG, not reflected ineither study, would cause costs to convergeover a shorter term. However, there arenotable differences in the Reynolds et alstudy methods that may support a sustaineddifference in costs. For example, all PCIpatients were treated with stenting, whilestenting is represented to only a limitedextent in the Weintraub et al study. This isan important consideration given that today,70% to 90% of PCIs involve placing 1 or morestents, resulting in improved proceduralsafety and reduced rates of restenosis comparedwith balloon angioplasty.¹ Furthermore,while baseline patient characteristicswere similar between groups in the Reynoldset al study, similar to the randomized groupsof the Weintraub et al study, many patientsunderwent PCI immediately after diagnosticcoronary angiography and this resulted in 1less diagnostic catheterization procedureand a reduction of preprocedure length ofstay for many patients, which is perhapsmore reflective of real-world practices. Thecost differences in stenting versus CABG inthe Reynolds et al study may be muted fordrug-eluting stents given their increasedcosts, or may be enhanced because ofdecreased repeat revascularization procedures,when viewed in the context of thestate-of-the-art treatment today. Lastly, the2 studies used different approaches to derivingcosts, which may affect comparisonefforts.

The examples of issues noted abovebecome even more important when comparingcost-related findings between geographies.A number of cost articles based onEuropean outcomes and cost data^15-17 arefairly widely cited in the literature; however,these may not be relevant to US audiencesbecause of differing practice patterns andresource costs.⁹ Variation in practice patternsmay limit the applicability of cost andcost-effectiveness analyses.

Of particular concern in evaluating revascularizationcosts is how to balance long-termoutcomes and cost data withprocedures undergoing rapid technologyimprovement and advancement. The burdenon managed care from technology and procedureadvances may increase if theseadvances come at a premium price orincrease the volume of use. However, premiumprices and increased volume may be offsetif these advances are able to saveresources and costs.

Even recently published studies includelong-term data from periods during whichstents were just beginning to be adopted.However, as previously noted, it is estimatedthat 70% to 90% of current angioplastiesinvolve placing 1 or more stents, and it ispossible that many or the majority of thesestents will be drug-eluting stents in the verynear future. Recent clinical and economicmodels, such as those of O'Neill and Leon,¹⁰and Chumer and colleagues,¹⁴ provide insightinto the potential impact of drug-elutingstents on hospitals. Though not capturedin our original literature review because ofits recent publication, data from a diseasestate-transition computer model simulatedclinical and economic consequences 5 yearsafter the introduction of drug-eluting stentsto hospitals. The results showed that introducingthese stents would change a $2.29million annual hospital profit (2003 dollars)to a loss of $5.41 million in the first year and$6.38 million in annual losses in later years,based on an annual volume of 3112 patientswith 85% of stent procedures shifted to drug-elutingstents in year 1, and assuming proposedchanges in Medicare reimbursementpolicy.¹⁸ This would divert over $28 millionfrom the hospital over a 5-year period.

Revascularization procedures are costly,and short-term cost differences in proceduresmay disappear when considered overthe long term. Recent cost data do not likelyreflect current treatment practices given therapid innovation in revascularization procedures and technology and the lag in publicationof cost data reflecting these advances.Modeling provides insight into the potentialeffect of these advances, although models'assumptions have yet to bear out. It is commonfor physicians and payers to try tomake informed clinical decisions with aneye toward cost containment. The costs ofthe procedures must be understood andplaced in the appropriate context.

A limitation to this review is that thecosts discussed may not reflect real worldcosts. Most of the costs referenced are basedon resource utilization from protocol-drivenclinical trials. Clinical trials often requireresource utilization that may not be reflectiveof real-world practice. This is importantfor payers relying on such cost data to makeinformed decisions.

Acknowledgment

The present review was supported by CVTherapeutics, Inc.

Heart Disease andStroke Statistics—2004 Update.

1. American Heart Association. Dallas, Tex: AmericanHeart Association; 2003.

Circulation.

2. Mark DB, Hlatky MA. Medical economics and theassessment of value in cardiovascular medicine: part II. 2002;106:626-630.

Circulation.

3. Smith SC Jr, Dove JT, Jacobs AK, et al. ACC/AHAguidelines for percutaneous coronary intervention (revisionof the 1993 PTCA guidelines): executive summaryand recommendations: a report of the American Collegeof Cardiology/American Heart Association Task Force onPractice Guidelines (Committee to revise the 1993 guidelinesfor percutaneous transluminal coronary angioplasty)endorsed by the Society for Cardiac Angiography andInterventions. 2001;103:3019-3041.

Circulation.

4. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHAguidelines for coronary artery bypass graft surgery: executivesummary and recommendations: a report of theAmerican College of Cardiology/American HeartAssociation Task Force on Practice Guidelines (Committeeto revise the 1991 guidelines for coronary arterybypass graft surgery). 1999;100:1464-1480.

J Interv Cardiol.

5. Neil N, Ramsey SD, Cohen DJ, Every NR, Spertus JA,Weaver WD. Resource utilization, cost, and health statusimpacts of coronary stent versus "optimal" percutaneouscoronary angioplasty: result from the OPUS-I trial. 2002;15:249-255.

Am JCardiol.

6. Cohen DJ, O'Shea JC, Pacchiana CM, et al. In-hospitalcosts of coronary stent implantation with and withouteptifibatide (the ESPRIT trial). Enhanced Suppression ofthe Platelet IIb/IIIa Receptor with Integrilin. 2002;89:61-64.

Am J Cardiol.

7. Weintraub WS, Becker ER, Mauldin PD, Culler S,Kosinski AS, King SB 3rd. Costs of revascularizationover eight years in the randomized and eligible patientsin the Emory Angioplasty versus Surgery Trial (EAST). 2000;86:747-752.

Am J Med.

8. Yock CA, Boothroyd DB, Owens DK, Garber AM,Hlatky MA. Cost-effectiveness of bypass surgery versusstenting in patients with multivessel coronary artery disease. 2003;115:382-389.

Am Heart J.

9. Reynolds MR, Neil N, Ho KK, et al. Clinical and economicoutcomes of multivessel coronary stenting comparedwith bypass surgery: a single-center USexperience. 2003;145:334-342.

Circulation.

10. O'Neill WW, Leon MB. Drug-eluting stents: costsversus clinical benefit. 2003;107:3008-3011.

11. Gruberg L. Cost-effectiveness of sirolimus drugelutingstents for the treatment of complex coronarystenoses: results from the randomized SIRIUS trial.Selected Session Coverage of the American College ofCardiology 52nd Annual Scientific Session. Presentedby David J. Cohen. March 30, 2003-April 2, 2003,Chicago, Ill. Medscape. Available at: http://www.medscape.com/viewarticle/451827. AccessedOctober 1, 2003.

12. Prescott LM. The case for drug-eluting stents growsstronger. Drug and market development 2003. Availableat: www.bioportfolio.com/reports/DMD_stent.htm.Accessed October 1, 2003.

N Engl J Med.

13. Moses JW, Leon MB, Popma JJ, et al. Sirolimus-elutingstents versus standard stents in patients with stenosisin a native coronary artery. 2003;349:1315-1323.

J Cardiovasc Manag.

14. Chumer K, DeCerce J, Vaul J. Managing technologicalexplosion in the medical management of coronaryartery disease, part II. DES economic modeling: budgetingstrategies for drug-eluting stent technology. 2002;13:16-20.

N Engl J Med.

15. Serruys PW, Unger F, Sousa JE, et al. Comparisonof coronary-artery bypass surgery and stenting for thetreatment of multivessel disease. 2001;344:1117-1124.

Circulation.

16. Zhang Z, Mahoney EM, Stables RH, et al. Disease-specifichealth status after stent-assisted percutaneouscoronary intervention and coronary artery bypass surgery:one-year results from the stent or surgery trial. 2003;108:1694-1700.

Drugs.

17. Ibbotson T, McGavin JK, Goa KL. Abciximab: anupdated review of its therapeutic use in patients withischaemic heart disease undergoing percutaneous coronaryrevascularisation. 2003;63:1121-1163.

Am Heart J.

18. Kong DF, Eisenstein EL, Sketch MH Jr, et al.Economic impact of drug-eluting stents on hospital systems:a disease-state model. 2004;147:449-456.

Download PDF: Review of Recent US Cost Estimates of Revascularization