tcamcsaaeididtec

CmrJ

0d

Influence of Age on Revascularization Related Costs ofHospitalization Among Patients of Stable Coronary Artery Disease

Shikhar Agarwal, MD, MPHa, Sudeep Banerjee, MS, MPHc, E. Murat Tuzcu, MDb, andSamir R. Kapadia, MDb,*

The legitimacy of expensive revascularization procedures in patients with stable coronaryartery disease (CAD) remains to be seen. In this study, resource utilization was comparedacross age-group categories of patients with stable CAD who underwent revascularization.Fiscal year 2006 discharge data maintained by the Maryland Health Services Cost ReviewCommission were examined. Current Procedural Terminology codes 36.1x and 36.0x wereused to identify patients who underwent coronary artery bypass grafting (CABG) andpercutaneous coronary intervention (PCI), respectively. Patients with acute myocardialinfarctions (International Classification of Diseases, Ninth Revision, Clinical Modificationcode 410.xx) were excluded to limit the study group to patients with stable CAD. Totals of1,066, 2,909, 4,272, 4,514, and 1570 hospitalizations per 100,000 hospitalizations wereobserved in the age groups 40 to 50, 50 to 60, 60 to 70, 70 to 80, and >80 years, respectively.The costs of hospitalization were driven by length of stay and choice of revascularizationprocedure (CABG vs PCI). There was trend toward increasing total costs and longerhospital stays with increasing age (p for trend <0.001) in patients who underwent CABG.Although the adjusted costs and adjusted length of stay were significantly higher in thepatients who underwent PCI who were aged >80 years than in those aged 40 to 50 years,the differences among other age-group categories were not statistically significant. Inconclusion, older patients with stable CAD who undergo interventions represent a sub-stantial expenditure. Hospitalization costs are driven by longer hospital stays and thechoice of CABG compared to PCI. With data supporting aggressive medical managementof such patients, it is advisable to rethink management strategies in these patients. © 2010

Published by Elsevier Inc. (Am J Cardiol 2010;105:1549–1554)

tcp

M

pv(gafiapHrnc

wd3nPCfi

The recent Clinical Outcomes Utilizing Revasculariza-ion and Aggressive Drug Evaluations (COURAGE) trialomparing the clinical efficacy and cost-effectiveness ofdding percutaneous coronary intervention (PCI) to optimaledical therapy in the management of patients with stable

oronary artery disease (CAD) failed to show any directurvival benefit or cost-effective potential of adding PCI toggressive medical management.1,2 In contrast, Claude etl3 demonstrated that an invasive strategy for stable CAD inlderly patients was indeed cost effective compared to med-cal therapy in the medium and long terms, suggesting thatirect costs of revascularization should not withhold annvasive strategy for the treatment of stable angina in el-erly patients. The legitimacy of expensive revasculariza-ion procedures in patients with stable CAD remains to bestablished. With an increasing aging population in ourommunity, health care expenditures continue to increase. In

Departments of aInternal Medicine and bCardiovascular Medicine,leveland Clinic, Cleveland, Ohio; and cUniversity of Alabama at Bir-ingham School of Public Health, Birmingham, Alabama. Manuscript

eceived November 23, 2009; revised manuscript received and acceptedanuary 5, 2010.

*Corresponding author: Tel: 216-444-6735; fax: 216-445-6176.

tE-mail address: kapadis@ccf.org (S.R. Kapadia).

002-9149/10/$ – see front matter © 2010 Published by Elsevier Inc.oi:10.1016/j.amjcard.2010.01.012

his study, we attempted to determine the economic burden ofhronic stable CAD in elderly patients compared to youngeratients who undergo revascularization procedures.

ethods

We conducted a cross-sectional study of Maryland hos-ital discharges maintained by the Maryland Health Ser-ices Cost Review Commission during fiscal year 2006July 2005 to June 2006). In addition to baseline demo-raphic data, extensive data were available on the primarydmission diagnosis, coded using the International Classi-cation of Diseases, Ninth Revision, Clinical Modification,s well as up to 14 secondary diagnoses and the codes forrocedures performed during each hospitalization. Theealth Services Cost Review Commission uses diagnosis-

elated groups to set case mix–adjusted limits on the reve-ue per discharge for inpatient services, similar to Medi-are’s inpatient prospective payment nationally.

The study sample consisted of adults aged �18 yearsith stable CAD who underwent revascularization proce-ures. Current Procedural Terminology codes 36.1x and6.0x were used to identify patients who underwent coro-ary artery bypass grafting (CABG) and PCI, respectively.atients with acute myocardial infarctions (Internationallassification of Diseases, Ninth Revision, Clinical Modi-cation code 410.xx) were excluded to limit the study group

o patients with chronic CAD. Total costs and length of stay

www.AJConline.org

(tmvust

(ccw

cfcebpseosC

TB

C

NAMWBAOD

P

PP

tion.

TR

C

NP

LIT

ion as i

1550 The American Journal of Cardiology (www.AJConline.org)

LOS) were used as primary outcome variables. We usedhe Deyo-adapted Charlson index to predict patient co-orbidities. For predicting Deyo-adapted Charlson index

alues, there are 17 diagnotic indicators identifiable bynique International Classification of Diseases, Ninth Revi-ion, Clinical Modification codes, assigned weights from 1hrough 6, giving the index a possible range of 0 to 33.4,5

The analysis was carried out using Stata version 10.0StataCorp LP, College Station, Texas). The bivariate asso-iations between categorical variables and the age-groupategories were tested using the Kruskal-Wallis test. One-

able 1aseline characteristics

haracteristic

40–50 50–6

umber of patients 995 2,85ge (years) 45.7 � 2.6 55.2 �en 734 (73.7%) 2,062 (7hites 714 (71.8%) 2,109 (7

lacks 191 (19.2%) 521 (1sians 15 (1.5%) 55 (1thers 75 (7.5%) 171 (6eyo-adapted Charlson index0 552 (55.5%) 1,346 (41 302 (30.4%) 1,018 (32 85 (8.5%) 312 (1�3 56 (5.6%) 180 (6Mean � SD 0.69 � 1.1 0.81 �rimary payerMedicare 80 (8.0%) 247 (8Medicaid 87 (8.7%) 198 (6Private 435 (43.7%) 1,411 (4Self 50 (5.0%) 92 (3HMO 331 (33.3%) 867 (3Other 12 (1.2%) 41 (1revious myocardial infarction 224 (22.5%) 576 (2atient dispositionHome 972 (97.7%) 2,779 (9Continued care (another service/hospital) 1 (0.1%) 8 (0Died 7 (0.7%) 5 (0Left against medical advice 15 (1.5%) 62 (2

Data are expressed as number (percentage) or as mean � SD.ANOVA � analysis of variance; HMO � health maintenance organiza

able 2esource utilization among the age-group categories

haracteristic Age Gr

40–50 50–60 6

umber of patients 995 2,856 3rocedure performedCABG 149 (15.0%) 605 (21.2%) 1,006PCI 841 (85.0%) 2,246 (78.8%) 2,881OS (days) 2.3 � 2.7 2.7 � 3.6 3.3

CU/CCU stay (days) 0.3 � 1.1 0.4 � 1.7 0.6otal costs $15,622 � $10,694 $17,298 � $13,637 $19,145

Data are expressed as number (percentage) or as mean � SD.* For continuous variables only.CCU � coronary care unit; ICU � intensive care unit. Other abbreviat

ay analysis of variance was used to compare means of t

ontinuous variable across the age-group categories. Bon-erroni’s correction was applied while performing pairwiseomparisons. We hypothesized that the distribution of totalxpenditure and LOS across the age-group categories maye confounded by gender, race, co-morbidities, patient dis-osition, and primary payment source. We used boot-trapped estimates (1,000 replicates) from multivariate lin-ar regression modeling to determine adjusted comparisonsf cost and LOS across the age-group categories. Regres-ion analyses were performed separately for the PCI andABG patients with logarithm-transformed costs and LOS,

Age Group (years) p Value(ANOVA)

60–70 70–80 �80

3,907 4,418 153264.4 � 2.8 73.1 � 3.1 82.8 � 2.9 �0.001

2,636 (67.5%) 2,908 (65.8%) 774 (50.5%) �0.0012,964 (75.9%) 3,406 (77.1%) 1,343 (87.7%) �0.001

637 (16.3%) 721 (16.3%) 131 (8.6%)100 (2.6%) 49 (1.1%) 12 (0.8%)206 (5.3%) 242 (5.5%) 46 (3.0%)

�0.0011,586 (40.6%) 1,943 (44.0%) 688 (44.9%)1,402 (35.9%) 1,688 (38.2%) 516 (33.7%)

552 (14.1%) 484 (11.0%) 197 (12.9%)367 (9.4%) 303 (6.9%) 131 (8.6%)0.97 � 1.1 0.85 � 1.0 0.90 � 1.1 �0.001

�0.0011,793 (45.9%) 4,035 (91.3%) 1,502 (98.0%)

147 (3.8%) 29 (0.7%) 5 (0.3%)1,181 (30.2%) 146 (3.3%) 12 (0.8%)

52 (1.3%) 9 (0.2%) 0696 (17.8%) 193 (4.4%) 12 (0.8%)36 (0.9%) 6 (0.1%) 1 (0.1%)

692 (17.7%) 930 (21.1%) 312 (20.4%) 0.001�0.001

3,673 (94.1%) 4,107 (93.2%) 1,315 (86.0%)19 (0.5%) 16 (0.4%) 18 (1.2%)4 (0.1%) 4 (0.1%) 0

208 (5.3%) 282 (6.4%) 196 (12.8%)

ars) p Valuefor Trend*

p Value(ANOVA)

70–80 �80

4,418 1532�0.001

) 754 (17.1%) 268 (17.6%)) 3,658 (82.9%) 1,259 (82.5%)

3.0 � 4.8 3.6 � 5.3 �0.001 �0.0010.5 � 2.7 0.7 � 2.9 �0.001 �0.001

,595 $17,550 � $17,493 $19,482 � $19,116 �0.001 �0.001

n Table 1.

0

62.8

2.2%)3.8%)8.2%).9%).0%)

7.1%)5.6%)0.9%).3%)1.0

.7%)

.9%)9.4%).2%)0.3%).4%)0.7%)

7.4%).3%).2%).2%)

oup (ye

0–70

,907

(25.9%(74.1%� 5.2� 3.1� $21

o satisfy the constant variance assumption for multiple

lw

miaptii

R

firBTe7

s4gca

teh7pIdprtwte

Fay

Fay

TR

C

P

C

1551Coronary Artery Disease/Costs of Stable Coronary Artery Disease

inear regression (homoskedascity). The significance levelas set at 0.05 for all statistical tests.Secondary analysis was performed to develop a predictor

odel of total hospitalization costs in these patients, includ-ng confounders and mediators of the proposed statisticalssociation. Because LOS and type of revascularizationrocedure were hypothesized to significantly affect hospi-alization costs, this predictor model was developed afterncorporation of these variables into the regression model-ng, in addition to the aforementioned confounders.

esults

A total of 638,447 unique hospitalizations were identi-ed in the data set. We found 13,830 revascularization-elated hospitalizations that matched the inclusion criteria.aseline characteristics of the study groups are listed inable 1. Significant differences in racial distribution werevident among the study groups. Elderly patients aged 60 to

igure 1. Odds ratios (ORs) for total costs and LOS comparing subsequentge-group categories with the reference age-group category (40 to 50ears) for patients who underwent CABG.

able 3esource utilization and costs compared among the age-group categories

haracteristic Age G

40–50 50–60 6

CINumber of patients 841 2,246LOS (days) 1.5 � 1.6 1.5 � 1.9 1.6ICU/CCU stay

(days)0.1 � 0.5 0.1 � 1.0 0.1

Total costs $13,230 � $7,778 $13,584 � $9,314 $13,459ABGNumber of patients 149 605 1LOS (days) 6.3 � 3.3 7.2 � 4.5 8.2ICU/CCU stay

(days)1.3 � 1.5 1.7 � 2.7 2.1

Total costs $27,580 � $12,465 $30,904 � $17,765 $33,758

Data are expressed as number or as mean � SD.Abbreviations as in Tables 1 and 2.

0 years and those aged �80 years were observed to have t

ignificantly higher co-morbidities compared to those aged0 to 50 years. In addition, it was noted that a significantlyreater proportion of older patients aged �80 years requiredontinued care other services (1.2%) compared to any otherge group.

Tables 2 and 3 list unadjusted resource utilization acrosshe age-group categories. The highest proportion of PCI wasncountered in patients aged 40 to 50 years (85%), and theighest proportion of CABG was seen in patients aged 60 to0 years (25.9%). There was a trend toward a longer hos-ital stay in the CABG and PCI groups with increasing age.n the CABG arm, the mean LOS ranged from 6.3 � 3.3ays in patients aged 40 to 50 years to 10.4 � 7.9 days inatients aged �80 years. In the PCI arm, the mean LOSanged from 1.5 � 1.6 days in patients aged 40 to 50 yearso 2.1 � 2.7 days in those aged �80 years. Mean total costsere significantly higher in the CABG patients compared to

he PCI patients in each age-group category (p �0.001 forach comparison). There was a statistically significant trend

igure 2. Odds ratios (ORs) for total costs and LOS comparing subsequentge-group categories with the reference age-group category (40 to 50ears) for patients who underwent PCI.

d by type of revascularization procedure

ears) p Valuefor Trend

p Value(ANOVA)

70–80 �80

3,658 12591.7 � 2.1 2.1 � 2.7 �0.001 �0.0010.1 � 0.6 0.1 � 0.9 0.3 0.09

749 $13,375 � $7,857 $14,412 � $9,169 0.05 0.002

754 2689.4 � 7.9 10.4 � 7.9 �0.001 �0.0012.6 � 5.9 2.9 � 5.7 �0.001 �0.001

,896 $37,426 � $31,455 $42,115 � $29,729 �0.001 �0.001

stratifie

roup (y

0–70

2,881� 1.6� 0.5

� $7,

,006� 6.6� 4.9

� $25

oward increasing total costs in the CABG arm with increas-

ip

j((asactyhc

tpchsiat1btp

tmawo�opcptiimALtm

va

cldwww4ueswtcFliLhp

D

ucirpcChAs

stfcmpLH

TU

C

NP

1552 The American Journal of Cardiology (www.AJConline.org)

ng age (p �0.001). However, this trend was not apparent inatients who underwent PCI (p � 0.05).

A statistically significant trend toward increasing ad-usted costs and LOS was apparent in the CABG armp �0.001; Figure 1). Patients aged �80 years incurred a 20%95% confidence interval [CI] 9% to 32%) higher total costnd a 22% (95% CI 9% to 37%) longer duration of hospitaltay compared to patients aged 40 to 50 years. In the PCIrm, no significant differences in adjusted hospitalizationosts or LOS were observed in the age groups 50 to 60, 60o 70, and 70 to 80 years compared to patients aged 40 to 50ears. Only patients aged �80 years incurred a significantlyigher total cost and a longer duration of hospital stayompared to those aged 40 to 50 years (Figure 2).

Table 4 demonstrates total hospitalization costs acrosshe age-group categories stratified according to primaryayer groups. With the exception of the youngest age-groupategory, Medicaid beneficiaries had significantly higherospitalization costs compared to Medicare beneficiaries. Atatistically significant trend toward increasing costs withncreasing age was observed in the Medicaid beneficiariesnd self-pay group of patients. Adjusted analysis revealedhat Medicaid beneficiaries incurred a 7% (95% CI 5% to2%) higher hospitalization cost compared to the Medicareeneficiaries. No significant differences were observed inhe adjusted hospitalization costs between other primaryayer categories.

We developed a prediction model for total hospitaliza-ion costs on the basis of the multivariate linear regressionethod. Because the total costs were affected by the LOS

nd the type of revascularization procedure, these variablesere included in this secondary analysis. The model dem-nstrated a good fit, with an Akaike information criterion of11,049 and adjusted R2 of 58.6%, implying that 58.6%

f the variability in total costs was explained by theredictor model. LOS and type of revascularization pro-edure (CABG vs PCI) emerged as the most importantredictors of the total costs. The elimination of LOS fromhe predictor model reduced the adjusted R2 to 45.8% andncreased the Akaike information criterion to �7,369. Sim-larly, elimination of the procedure type from the predictorodel reduced the adjusted R2 to 52.4% and increased thekaike information criterion to �8,931. Elimination ofOS and procedure type from the predictor model reduced

he adjusted R2 to 20.2% and increased the Akaike infor-

able 4nadjusted total costs and ratio of adjusted total costs in comparison with

haracteristic Age Gr

40–50 50–60 6

umber of patients 995 2,856 3rimary payerMedicare $18,269 � $12,681 $17,828 � $17,439 $19,502Medicaid $17,422 � $13,768 $22,438 � $16,800 $21,555Private $15,594 � $10,418 $16,812 � $12,929 $18,796Self $16,469 � $10,196 $17,354 � $12,504 $17,129HMO $14,253 � $9,214 $16,713 � $12,623 $18,148Other $20,194 � $15,711 $18,273 � $12,919 $25,647

Abbreviations as in Table 1.

ation criterion to �1,875, indicating that LOS and inter- w

ention type together explained 66% of the total predictivebility of the model.

Because LOS emerged as the strongest predictor of totalosts, we performed a multivariate regression analysis withogarithm-transformed LOS as the dependent variable toetermine the role of other variables in governing LOS. Itas observed that choice of revascularization procedureas the strongest determinant of LOS. Patients who under-ent CABG were observed to have 4.7 (95% CI 4.61 to.84) times longer hospital stays compared to those whonderwent PCI. Also, patient disposition had a strong influ-nce on LOS. In comparison to patients discharged home,ignificantly longer hospital stays were observed in patientsho died during their hospital stays (39% longer, 95% CI 2%

o 90%) or were discharged to different health care facilities forontinued medical care (70% longer, 95% CI 60% to 80%).urthermore, black patients had a 15% (95% CI 12% to 19%)

onger hospital stay compared to white patients. In addition,ncreasing co-morbidities were associated with increasingOS; patients with Deyo-adapted Charlson index values �2ad 47% (95% CI 40% to 55%) longer hospitalizations com-ared to those without concomitant co-morbidities.

iscussion

We observed that the old patients with stable CAD whonderwent revascularization incurred significantly higherosts than younger patients, even after adjustment for clin-cal and demographic characteristics. LOS and choice ofevascularization procedure emerged as the most importantredictors of the hospitalization related costs. A signifi-antly longer LOS was observed in patients who underwentABG and patients who died or were discharged to otherealth care facilities for continued medical care. Africanmerican race and increasing co-morbidities were other

ignificant determinants of a longer hospital stay.Patients who underwent CABG were observed to have

ignificantly higher adjusted expenditures in comparison tohose who underwent PCI. A notable age-varying cost dif-erential was evident between the CABG and PCI arms. Theost differential between the CABG and PCI arms wasarkedly higher in the older compared to the younger

opulations. There was a trend toward increasing costs andOS with increasing age in patients who underwent CABG.owever, such a trend was absent in patients who under-

est age group stratified according to primary payer groups

ars) p Valuefor Trend

p Value(ANOVA)

70–80 �80

4,418 1532

947 $17,460 � $17,657 $19,279 � $18,602 0.7 �0.001993 $25,607 � $24,015 $71,306 � $74,480 �0.001 0.001829 $19,664 � $15,984 $24,179 � $10,645 0.06 �0.001482 $28,419 � $16,997 — 0.007 0.04281 $15,790 � $13,109 $15,010 � $7,728 0.9 0.005600 $27,830 � $10,270 $63,431 � $0 0.003 0.01

young

oup (ye

0–70

,907

� $19,� $17,� $24,� $10,� $21,� $16,

ent PCI, and health care resource utilization remained

cp

b2grCthtpimypcwpTeicdt

sprtFct2wcva

apiCdigesYtDcsaPfse

or

ipptvmvstdivcfrwneo

nfwpttstTdte

1553Coronary Artery Disease/Costs of Stable Coronary Artery Disease

onstant across age-group categories, with the exception ofatients aged �80 years.

Elderly patients currently pose a significant economicurden to our health care system. We observed totals of 995,,856, 3,907, 4,418, and 1,532 hospitalizations in the ageroups 40 to 50, 50 to 60, 60 to 70, 70 to 80, and �80 years,espectively. On the basis of Health Services Cost Reviewommission annual population estimates, these observa-

ions translate into 1,066, 2,909, 4,272, 4,514, and 1,570ospitalizations per 100,000 hospitalizations annually forhe respective age categories. In the United States, theroportion of the population aged �65 years is projected toncrease from 12.4% in 2000 to 19.6% in 2030.6 Approxi-

ately 85% of those who die from CAD are aged �65ears.7 Although the elderly represent only 13% of theopulation, they consume approximately 36% of all healthare expenditures. The average health care expense in 2002as $11,089 per year for elderly individuals but only $3,352er year for working-age individuals (aged 19 to 64 years).8

he total direct and indirect costs attributed to CAD werestimated at $156.4 billion in 2007 and 2008.9 These costs arenclusive of direct health care costs ($87.6 billion), namely, theosts of physicians, hospitals, medications, diagnostic proce-ures, surgeries, and indirect costs such as lost productivity dueo morbidity and mortality ($68.8 billion).9

The recent COURAGE trial provided an important in-ight into stable CAD management. The investigators re-orted that PCI added to routine medical therapy does noteduce the risk for death or myocardial infarction on long-erm follow-up compared to routine medical therapy alone.2

ollow-up economic analysis revealed that PCI plus medi-al therapy was significantly more expensive than medicalherapy alone.1 The cumulative cost difference between the

arms was sustained, implying that the costs associatedith initial PCI are never recovered. Economic outcomes in

onjunction with clinical outcomes led the COURAGE in-estigators to conclude that it is safe to defer PCI in favor ofggressive medical therapy.

The PCI cost-effectiveness data in comparison to CABGre divided and unclear. Cost analysis in the Emory Angio-lasty Versus Surgery Trial (EAST) demonstrated a largenitial cost advantage in the PCI group compared to theABG group.10 However, after 8 years of follow-up, theifference in total cumulative expenditure was not discern-ble between the 2 strategies. The data from the registryroup in the same study, however, indicated that the totalxpenditure at the end of 8 years in the PCI arm wasignificantly lower than the expenditure in the CABG arm.10

ock et al11 did not find any significant differences in theotal lifetime estimates between the PCI and CABG arms.ata from a more recent study suggest that initial hospital

osts and costs at 2 years in patients who undergo PCI areignificantly less in comparison to the CABG group, despitehigher rate of repeat revascularization procedures in the

CI arm.12 Preliminary short-term and medium-term datarom studies assessing the cost-effectiveness of drug-elutingtents have shown a sizable economic benefit of drug-luting stents over other revascularization strategies.13,14

With a questionable benefit of revascularization strategyver routine medical therapy, the legitimacy of spending

esources on revascularizations in patients with stable CAD

s put to question as well. It is universally known that not allatients with CAD are in definite need of revascularizationrocedures. Symptomatic patients with chronic stable symp-oms should be managed with optimal medical therapy. Re-ascularization needs to be considered only after failure ofedical therapy. In older patients in absolute need of re-

ascularization, PCI may serve as a more cost-effectivetrategy compared to CABG. It has been previously shownhat after primary PCI, clinical data and the catheterizationata may help identify low-risk patients with myocardialnfarctions who can safely forgo intensive care and nonin-asive testing, leading to quicker discharge and substantialost savings.15 The adoption of this management strategyor stable CAD, especially in the elderly, may serve as aeasonable cost-cutting alternative. In symptomatic patientsith CAD for whom PCI and CABG are reasonable alter-atives, PCI may be preferable because of higher cost-ffectiveness. In CABG patients, further effort is needed toptimize the conditions to decrease LOS.

The limitations of our study arise from its cross-sectionalature and its vulnerability to biases related to unmeasuredactors. The follow-up costs and the indirect cost estimatesere not available, which are vital to bring about changes inractice or policy. Although it was not possible to determinehe exact reason for the choice of revascularization therapy,he striking differences in the costs associated with the 2trategies must always be considered in conjunction withhe clinical scenario before deciding on the choice of therapy.he need for repeat intervention, especially in those who un-ergo PCI, was impossible to assess in this database. However,he annual rate of repeat revascularization in this current drug-luting stent era is estimated to be low (2% to 4%).16

1. Weintraub WS, Boden WE, Zhang Z, Kolm P, Zefeng Z, Spertus JA,Hartigan P, Veledkar E, Jurkovit C, Bowen J, Maron DJ, O’Rourke R,Dada M, Teo KK, Georee R, Barnett PG. Cost-effectiveness of per-cutaneous coronary intervention in optimally treated stable coronarypatients. Circ Cardiovasc Qual Outcomes 2008;1:12–20.

2. Boden WE, O’Rourke RA, Teo KK, Hartigan PM, Maron DJ, Kostuk WJ,Knudtson M, Dada M, Casperson P, Harris CL, Chaitman BR, Shaw L,Gosselin G, Nawaz S, Title LM, Gau G, Blaustein AS, Booth DC, BatesER, Spertus JA, Berman DS, Mancini GB, Weintraub WS; COURAGETrial Research Group. Optimal medical therapy with or without PCI forstable coronary disease. N Engl J Med 2007;356:1503–1516.

3. Claude J, Schindler C, Kuster GM, Schwenkglenks M, Szucs T, BuserP, Osswald S, Kaiser C, Grädel C, Estlinbaum W, Rickenbacher P,Pfisterer M; Trial of Invasive Versus Medical Therapy in the Elderly(TIME) Investigators. Cost-effectiveness of invasive versus medicalmanagement of elderly patients with chronic symptomatic coronaryartery disease. Findings of the randomized trial of invasive versusmedical therapy in elderly patients with chronic angina (TIME). EurHeart J 2004;25:2195–2203.

4. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method ofclassifying prognostic comorbidity in longitudinal studies: develop-ment and validation. J Chronic Dis 1987;40:373–383.

5. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidityindex for use with ICD-9-CM administrative databases. Clin Epide-miol 1992;45:613–619.

6. United States Census Bureau. International database. Table 094. Mid-year population, by age and sex. Available at: http://www.census.gov/population/www/projections/natdet-D1A.html. Accessed on July 20,2009.

7. Maycock CA, Muhlestein JB, Horne BD, Carlquist JF, Bair TL,Pearson RR. Statin therapy in very elderly patients with CAD. J Am

Coll Cardiol 2002;40:1780.

1

1

1

1

1

1

1

1554 The American Journal of Cardiology (www.AJConline.org)

8. Stanton MW, Rutherford MK. The High Concentration of U.S. HealthCare Expenditures. AHRQ Pub. No. 06-0060. Rockville, Maryland:Agency for Healthcare Research and Quality; 2005.

9. American Heart Association. Heart Disease and Stroke Statistics—2008 Update. Dallas, Texas: American Heart Association; 2008.

0. Weintraub WS, Becker ER, Mauldin PD, Culler S, Kosinski AS, KingSB III. Costs of revascularization over eight years in the randomizedand eligible patients in the Emory Angioplasty Versus Surgery Trial(EAST). Am J Cardiol 2000;86:747–752.

1. Yock CA, Boothroyd DB, Owens DK, Garber AM, Hlatky MA. Costeffectiveness of bypass surgery versus stenting in patients with mul-tivessel coronary artery disease. Am J Med 2003;115:382–389.

2. Reynolds MR, Neil N, Ho KK. Clinical and economic outcomes ofmultivessel coronary stenting compared with bypass surgery: a single-

center US experience. Am Heart J 2003;145:334–342.

3. Gorenoi V, Dintsios CM, Schonermark MP Hagen A. Health economicmodeling on the use of drug eluting stents versus coronary artery bypassgraft surgery in coronary heart disease. Herz 2009;34:231–239.

4. Groeneveld PW, Matta MA, Greenhut AP, Yang F. The cost ofdrug-eluting coronary stents among Medicare beneficiaries. AmHeart J 2008;155:1097–1105.

5. Grines CL, Marsalese DL, Brodie B, Griffin J, Donohue B, CostantiniCR, Balestrini C, Stone G, Wharton T, Esente P, Spain M, Moses J,Nobuyoshi M, Ayres M, Jones D, Mason D, Sachs D, Grines LL,O’Neill W; PAMI-II Investigators. Safety and cost-effectiveness ofearly discharge after primary angioplasty in low risk patients withacute myocardial infarction. J Am Coll Cardiol 1998;31:967–972.

6. Leon MB, Allocco DJ, Dawkins KD, Baim DS. Late clinical eventsafter drug-eluting stents: the interplay between stent-related and nat-

ural history-driven events. JACC Cardiovasc Interv 2009;2:504–512.