Journal of the American College of Cardiology 2006 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 47, No. 3, 2005 ISSN 0735-1097/05/$32.00 doi:10.1016/j.jacc.2005.08.070

The Role of Risk Stratication in the Decision to Provide Upstream Versus Selective Glycoprotein IIb/IIIa Inhibitors for Acute Coronary Syndromes
A Cost-Effectiveness Analysis
Ruchira Glaser, MD,* Henry A. Glick, PHD, Howard C. Herrmann, MD, FACC,* Stephen E. Kimmel, MD, MSCE, FACC* Philadelphia, Pennsylvania
We endeavored to determine under what conditions a strategy of upstream use of small molecule platelet glycoprotein (GP) IIb/IIIa inhibitors for all acute coronary syndromes (ACS) patients is cost effective compared to that of selective use of abciximab in only those patients requiring percutaneous coronary intervention (PCI). BACKGROUND Small molecule GP IIb/IIIa inhibitors have shown benet in ACS, but abciximab, the more expensive GP IIb/IIIa inhibitor, may be more effective during PCI. However, abciximab does not have proven efcacy in medical management. No prior study has attempted to balance these competing benets. METHODS A decision analysis was performed to examine two strategies: 1) treat all ACS patients upstream with a small molecule GP IIb/IIIa inhibitor and continue through medical management and PCI, if performed; or 2) wait, and selectively use abciximab only in patients who ultimately undergo PCI. Applicable randomized controlled trial data were used for the principal analysis. RESULTS The strategy of upstream use of a small molecule GP IIb/IIIa inhibitor was superior to selective use, and economically acceptable, with a cost-effectiveness ratio of $18,000 per year of life gained. The superiority of the upstream use strategy persisted over the majority of sensitivity analyses. When stratied by risk according to Thrombolysis in Myocardial Infarction risk score, a strategy of upstream use was only cost effective in those patients with moderate or high risk. CONCLUSIONS Upstream use of small molecule GP IIb/IIIa inhibition in ACS patients with moderate or high risk for cardiovascular events is a cost-effective approach that should be considered in this subset of patients. (J Am Coll Cardiol 2005;47:529 37) 2006 by the American College of Cardiology Foundation OBJECTIVES

Randomized trials of patients with acute coronary syndromes (ACS) have demonstrated a signicant reduction in myocardial infarction (MI) and death with the use of the small molecule platelet glycoprotein (GP) IIb/IIIa receptor inhibitors, tiroban and eptibatide (1 4). In contrast, the See page 538 alternative GP IIb/IIIa inhibitor, abciximab, was not benecial when administered to a predominantly medically managed ACS population (5). However, during percutaneous coronary intervention (PCI) abciximab was superior to the small molecule GP IIb/IIIa inhibitor, tiroban (6). These ndings suggest that small molecule GP IIb/IIIa inhibitors may provide some benet during medical management of ACS, but may not be as effective during PCI as abciximab, a drug that is more costly and provides no benet during medical management.
From the *Division of Cardiovascular Medicine, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and the Center for Clinical Epidemiology and Biostatistics and Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania. Manuscript received December 3, 2004; revised manuscript received July 22, 2005, accepted August 23, 2005.

No randomized controlled trial has addressed the key question that clinicians must answer when a patient with ACS presents for care; namely, whether a strategy of upstream use of a small molecule platelet GP IIb/IIIa inhibitor, with continuation of the drug in those patients undergoing PCI, or a strategy of selective use of abciximab in only those patients undergoing PCI, is the optimal therapeutic option. In addition, validated measures of risk, such as the Thrombolysis In Myocardial Infarction (TIMI) risk score (7), have not been applied to upstream versus selective GP IIb/IIIa inhibition (8 10). Decision analysis allows clinicians to compare the possible outcomes of different therapeutic strategies, particularly in the face of uncertainty, when no appropriate randomized clinical trial is available (11). The purpose of this study was to use decision analysis and primary data from randomized controlled trials to determine whether one strategy is superior to the other, to calculate the cost-effectiveness of that strategy, and to assess the value of the TIMI risk score in choosing the most cost-effective strategy.

METHODS
Decision tree. We performed a decision analysis (DATA version 4.0, TreeAge Software Inc., Williamstown, Massa-

530

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

Abbreviations and Acronyms ACS acute coronary syndromes CABG coronary artery bypass graft GP glycoprotein ICER incremental cost-effectiveness ratio MI myocardial infarction OR odds ratio PCI percutaneous coronary intervention QALY quality-adjusted life years RR relative risk TACTICS Treat angina with Aggrastat and determine Cost of Therapy with Invasive or Conservative Strategy

chusetts) of patients presenting with unstable angina and nonST-segment elevation MI ACS (Fig. 1). The rst option, upstream small molecule, represents the strategy in which a small molecule GP IIb/IIIa inhibitor is given upstream to all patients upon presentation with ACS and continued during PCI in those patients receiving PCI. The second option, selective abciximab, represents the strategy in which no GP IIb/IIIa inhibitor is given upon presentation, and abciximab is selectively given only to patients who ultimately undergo PCI. In each strategy, patients can get one of three therapies: PCI, coronary artery bypass graft (CABG), or neither PCI nor CABG (Fig. 1). For all three of these possible therapies, they can survive without MI (either with or without a major bleed), they can have a nonfatal MI (either with or without a major bleed), or they can die. Data assumptions. Point estimates and ranges for the probabilities used in the model are summarized in Table 1. The probabilities of PCI and CABG were obtained from the invasive arm of a contemporary trial of invasive versus conservative management of ACS (12). It was assumed that the rates of revascularization therapy would not be inuenced by whether or not a GP IIb/IIIa inhibitor was used up front (13). Relative risks (RR) and odds ratios (OR) from randomized controlled trials of GP IIb/IIIa inhibitors in ACS patients were used to derive all further probabilities as follows. Where possible, RRs or ORs were obtained from direct randomized comparison data representative of each particular branch in the tree. For PCI with abciximabrelated risks, the absolute probability was calculated as a function of the probability of the event in medical management patients and the RR with PCI using abciximab. It was assumed that probability of MI in placebo patients with medical management would approximate the probability of MI in placebo patients with PCI. This assumption enabled the use of one set of baseline risks for the entire tree, as opposed to separate risks across potentially different patient populations. For small molecule PCI-related risks, the absolute probability was calculated as a function of the absolute probability of the event with PCI using abciximab (calculated as above), and the RR for the event with small molecule versus abciximab during PCI. The ORs for death

and MI with small molecule GP IIb/IIIa inhibitors in patients who do not undergo PCI were derived from the pooled ORs of death and MI versus placebo for the ve major randomized trials comparing small molecule GP IIb/IIIa inhibitors versus placebo in a predominantly medically managed population. A total of 24% of these patients underwent PCI (13). However, because the exclusion of a subgroup of patients based on their subsequent care could bias our estimates (14), we did not exclude those receiving PCI from the pooled estimate. Major bleeding rates were based on pooled estimates of bleeding rates in randomized controlled ACS trials (13). Bleeding after CABG was estimated using the published subanalyses of these trials (12,15,16). Outcomes. The primary outcome measure was the incremental cost-effectiveness ratio (ICER), measured as the difference in cost of the two strategies divided by the difference in life expectancy (both discounted at a rate of 3% per year; life expectancy values and life years that are described throughout the Results section reect such discounting). The secondary end point was difference in major bleeding rate.
COSTS. Estimates were obtained for 30-day, 6-month, and lifetime costs, with lifetime cost used to calculate the ICER. The cost assigned to each 10-mg vial of abciximab was $450, with an assumption of a mean of three vials used per patient (i.e., $1,350) (1720). The cost ($1,060) assigned to upstream small molecule GP IIb/IIIa inhibitor was based on the mean cost of drug in the Treat angina with Aggrastat and determine Cost of Therapy with Invasive or Conservative Strategy (TACTICS) trial (median duration of therapy 48 h in the invasive arm) (21). The 30-day and 6-month costs of procedures and hospital stay were derived by adjustment of costs reported for the invasive arm of the TACTICS trial and for the Evaluation in PTCA to Improve Long-term Outcome with abciximab GP IIb/IIIa blockade (EPILOG) trial (18,21). These costs include the costs of MI, bleeding, and of increased length of stay. All costs were further adjusted to 2002 dollars. Costs after six months (for the lifetime cost estimate) were derived from projections for a typical 60-year-old MI patient from the Cholesterol And Recurrent Events (CARE) trial population (17,22). EFFECTIVENESS.

Effectiveness was measured in terms of additional life expectancy. Life expectancies were modeled for four main outcomes in the decision tree: survivors who did not require PCI or CABG; survivors who required PCI or CABG and did not have a subsequent nonfatal MI; survivors who had a subsequent nonfatal MI; and those who died within 30 days (life expectancy 0). Life expectancy for patients with ACS who did not require PCI or CABG was derived from vital statistics data (23). Life expectancy for those with ACS who had a nonfatal MI was derived from modeling from Kaplan-Meier estimates in patients with creatine kinase elevation after PCI; life expectancy for those who underwent PCI without nonfatal MI was simi-

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

531

Figure 1. Decision tree comparing upstream early use of tiroban or eptibatide in all patients versus selective use of abciximab in patients undergoing percutaneous coronary intervention (PCI). Squares decision; circles chance events; triangles outcomes. ACS acute coronary syndrome; CABG coronary artery bypass grafting; MI myocardial infarction.

larly derived from patients without creatine kinase elevation (24). The mean age in the cohorts used to derive life expectancy values was 61 years, and 76% of patients were men. Life expectancy for survivors who did not require PCI or CABG was 17.592 years; for survivors who required PCI or CABG and did not have a nonfatal MI was 15.535 years; for all survivors who had a subsequent nonfatal MI was 13.524 years. Life expectancy estimates matched closely the estimates used in other cost-effectiveness analyses of GP IIb/IIIa inhibitors, including those derived from the Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PUR-

SUIT) trial and Duke database (25,26). Effects of the treatments on the patients quality of life were examined separately in a sensitivity analysis (27). Assessment of the role of risk stratication. Baseline patient risk was assessed by published risk for death and MI according to TIMI risk score at six weeks (9) and categorized as low, intermediate, and high risk (7). These analyses provided the cost-effectiveness of upstream use of GP IIb/IIIa inhibition in low-, intermediate-, and high-risk score patients. Sensitivity analyses. Sensitivity analyses were performed for important clinical questions. Stability of the models predictions was assessed by varying all parameters in the tree across a

532

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

Table 1. Probability of Outcomes Used in the Decision Model

Base Case Probability or Relative Risk 0.4 References Base Case TACTICS, invasive arm (12) Additional References for Range or 95% CI Source GUSTO IV, FRISC II, RITA 3, PRISM PLUS, PURSUIT (1,3,5,40,41) FRISC II (40)

Factor PCI

Range 0.20.8

CABG

0.15

0.110.35

PRISM, PRISM PLUS, PURSUIT, TACTICS, RITA 3, GUSTO IV (13,5,12,41) Boersma et al. (placebo arm) (13) Boersma et al. (placebo arm) (13) Boersma et al. (placebo arm) (13) PURSUIT (placebo group) (1) Boersma et al. (placebo arm) (13) Boersma et al. (placebo arm) (13) CAPTURE (4) CAPTURE (4) CAPTURE (4)

Selective use abciximab during PCI only No PCI, no drug, bleed No PCI, no drug, MI No PCI, no drug, death CABG, bleed CABG, MI CABG, death PCI with abciximab, bleed PCI with abciximab, MI PCI with abciximab, death Upstream use small molecule GP IIb/IIIa inhibitor in all ACS patients No PCI, small molecule, bleed No PCI, small molecule, MI

0.01 0.081 0.037 RR 0.93 0.21 0.045 RR 2.05 RR 0.48 RR 0.75

0.010.03 0.0390.135 0.0290.051

GUSTO IV ACS (placebo arm) (5) Boersma et al. (placebo arm) (13) Boersma et al. (placebo arm) (13)

0.060.28 0.0450.058 1.034.09 0.300.78 0.272.09

PURSUIT, PRISM (placebo group) (1,2) PURSUIT, PRISM (placebo group) (1,2) CAPTURE (4) CAPTURE (4) CAPTURE (4)

RR 1.4 RR 0.88

RR 1.32.07 RR 0.470.95

Boersma et molecule Boersma et molecule

al. (small analysis) (13) al. (small analysis) (13)

No PCI, small molecule, death

RR 0.88

RR 0.410.95

Boersma et al. (small molecule analysis) (13)

CABG, bleed CABG, MI

CABG, death

PCI, small molecule, bleed PCI, small molecule, MI PCI, small molecule, death

0.58 RR 1.0 (vs. CABG no GPIIb/IIIa inhibitor MI) RR 1.0 (vs. CABG no GP IIb/IIIa inhibitor death) RR 1.0 (vs. PCI bleed abciximab) RR 1.5 (vs. PCI MI abciximab) RR 1.0 (vs. PCI death abciximab)

0.20.8 RR 0.653.01

PURSUIT (1) Empiric

Boersma et al. (small molecule analysis) (13) Boersma et al. (meta-analysis of pre-PCI events), PRISM no PCI cohort, PRISM PLUS rst 48 h (medical therapy), PURSUIT (time until PCI or surgery) (13,42) Boersma et al. (meta-analysis of pre-PCI events), PRISM no PCI cohort, PRISM PLUS rst 48 h (medical therapy), PURSUIT (time until PCI or surgery) (13,42) TACTICS, EPIC (12,42) PURSUIT, PRISM (1,2)

RR 0.492.5

Empiric

PURSUIT, PRISM (1,2)

RR 0.82.7 RR 02.5 RR 02.5

TARGET (6) TARGET ACS cohort (44) TARGET ACS cohort (44)

TARGET (6) TARGET ACS cohort (44) TARGET ACS cohort (44)

ACS acute coronary syndromes; CABG coronary artery bypass grafting; CAPTURE C7E3 Anti Platelet Therapy in Unstable REfractory angina trial; CI condence interval; EPIC Evaluation of c7E3 for Prevention of Ischemic Complications trial; FRISC II Fragmin and Fast Revascularization during Instability in Coronary Artery disease trial; GP glycoprotein; GUSTO IV Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries IV trial; MI myocardial infarction; PCI percutaneous coronary intervention; PRISM PLUS Platelet Receptor Inhibition in ischemic Syndrome Management in Patients Limited by Unstable Signs and symptoms trial; PURSUIT Platelet glycoprotein IIb/IIIa in Unstable angina: Receptor Supression Using Integrellin Therapy trial; RITA 3 Randomized Intervention Trial of unstable Angina; RR relative risk; TACTICS Treat Angina with Aggrastat and determine Cost of Therapy with an Invasive or Conservative Strategy; TARGET Do Tiroban and Reo-Pro Give Similar Efcacy Outcome Trial.

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937 Table 2. Costs, Effectiveness, and Incremental Cost-Effectiveness Ratio for Upstream Use Strategy Compared With Selective Use Strategy
Selective Use Abciximab During PCI Only 30-day cost per patient 6-month cost per patient Lifetime cost per patient* Incremental lifetime cost per patient Incremental lifetime cost per 100,000 patients* Years of life per 100,000 patients Incremental years of life gained per 100,000 patients Incremental cost per year of life gained $17,631 $23,463 $81,124 1,564,000 Upstream Use Small Molecule GPIIb/IIIa Inhibitor in All ACS Patients $18,210 $24,092 $81,844 $720 $72,000,000 1,568,000 4,000

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

533

$18,000

*Used for calculation of incremental cost-effectiveness ratio. ACS acute coronary syndromes; GP glycoprotein; PCI percutaneous coronary intervention.

range dened either by clinical plausibility or by the variables 95% condence intervals observed in the randomized trials.

RESULTS
Probabilities. The probability of death or MI at 30 days was 11.9% in the upstream strategy and 12.1% in the selective abciximab strategy. The probability of major bleeding was 10.2% with upstream use and 9.7% in the selective abciximab strategy. The rates of non-CABG-related major bleeding were 3.6% and 3.0%, with upstream and selective abciximab use, respectively. Costs, life expectancy, quality-adjusted life years (QALY), and cost-effectiveness. Despite the higher drug cost of abciximab ($1,350 per patient for abciximab vs. $1,060 for

small molecule), an upstream strategy of small molecule GP IIb/IIIa inhibition for all ACS patients was more costly than selective abciximab use, with an incremental lifetime cost of $720 per patient (Table 2) because, in large part, of the higher proportion of patients receiving a GP IIb/IIIa inhibitor in the upstream strategy. The average life expectancy was 15.68 years with the upstream strategy, compared with 15.64 years for selective abciximab. This yielded a gain of 4,000 years of life per 100,000 patients treated with the strategy of upstream small molecule GP IIb/IIIa inhibition. Thus, the use of upstream small molecule GP IIb/IIIa inhibitors resulted in an ICER of $18,000 per year of life gained (Table 2). Quality adjustment to years of life gained yielded an average life expectancy of 14.80 QALYs with upstream use versus 14.76 with selective use of abciximab. This yielded a cost-effectiveness ratio of $18,000 per QALY gained. The role of risk stratication. Increasing patient baseline risk for adverse events amplied the benets of the upstream use strategy. Patients with low TIMI risk scores (0 to 2) had a gain of only 1,000 years of life per 100,000 patients treated with upstream use, with an ICER of $58,300 per year of life gained (Fig. 2). In contrast, patients with moderate or higher TIMI risk scores (3 to 4 and 5 to 7) had more years of life saved with an upstream GP IIb/IIIa inhibitor strategy, resulting in economically favorable cost-effectiveness ratios ($24,730 and $14,444 per year of life gained, in moderate and high TIMI risk score patients, respectively, Fig. 2). Further sensitivity analyses of high-risk patients revealed an extremely favorable cost-effectiveness ratio ($1,276 per year of life gained) for the upstream GP IIb/IIIa strategy if the probability of PCI was increased to 80%. However, the upstream use strategy became less cost effective if the probability of CABG was also extremely high. For instance, if the probability of CABG was 30% and the probability of

Figure 2. Cost effectiveness of upstream glycoprotein IIb/IIIa use according to Thrombolysis In Myocardial Infarction (TIMI) risk score of patient.

534

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

Table 3. Incremental Cost-Effectiveness Ratios in Sensitivity Analyses

Principal Analysis Estimate 0.40 Estimates Used for Sensitivity Analyses (threshold 0.63) 0.2 0.8 0.11 0.35 No PCI, MI 0.081 0.039 0.135 No PCI, death 0.037 0.029 0.051 CABG, MI 0.21 0.06 0.28 CABG, death 0.045 0.045 0.058 PCI with abciximab, MI RR 0.48 0.30 0.78 PCI with abciximab, death RR 0.75 0.27 2.09 No PCI, MI with upstream small molecule in all ACS patients RR 0.88 0.47 1.0 No PCI, death with upstream small molecule in all ACS patients RR 0.88 0.41 1.0 PCI, MI with upstream small molecule in all ACS patients RR 1.5 1.0 2.5 PCI, death with upstream small molecule in all ACS patients RR 1.0 1.0 2.5 Life expectancy discount rate 3% 1% 5% 1,961,800 1,289,000 1,957,500 1,286,000 4,300 3,000 16,700 24,000 1,569,000 1,564,000 1,565,000 1,564,000 (threshold RR 1.2 LE 1,565,000) 1,568,000 1,544,000 1,564,000 1,564,000 5,000 1,000 14,400 72,000 1,582,000 1,565,000 1,564,000 1,564,000 18,000 1,000 4,000 72,000 1,574,000 1,566,000 1,564,000 1,564,000 10,000 2,000 7,200 36,000 1,579,000 1,539,000 1,575,000 1,536,000 4,000 3,000 18,000 24,000 1,570,000 1,565,000 1,565,000 1,563,000 5,000 2,000 14,400 36,000 1,568,000 1,565,000 1,564,000 1,561,000 4,000 4,000 18,000 18,000 1,572,000 1,566,000 1,569,000 1,562,000 3,000 4,000 24,000 18,000 1,577,000 1,551,000 1,575,000 1,546,000 2,000 5,000 36,000 14,400 1,577,000 1,556,000 1,574,000 1,552,000 3,000 4,000 24,000 18,000 Years of Life With Upstream Small Molecule Use* 1,568,000 1,603,000 1,498,000 1,577,000 1,522,000 Years of Life With Selective Abciximab Use* 1,564,000 1,596,000 1,501,000 1,573,000 1,520,000 Years of Life Gained With Upstream Use* 4,000 7,000 3,000 4,000 2,000 Cost Effectiveness of Upstream Use (Dollars per Year Life Gained) 18,000 5,800 18,000 36,000

Subject of Sensitivity Analysis Principal simulation PCI rate

CABG rate

0.15

4,000 20,000

18,000

*Represents years of life per 100,000 patients; Represents relative risk versus PCI with abciximab. ACS acute coronary syndromes; CABG coronary artery bypass grafting; LE life expectancy; MI myocardial infarction; PCI percutaneous coronary intervention; RR relative risk.

PCI was 50%, upstream GP IIb/IIIa use had a borderline favorable cost-effectiveness ratio of $55,571. Additional sensitivity analyses. The superior effectiveness of upstream use of GP IIb/IIIa inhibition persisted in most cases when applying a wide range of plausible probabilities in reduction of death and MI for both strategies during sensitivity analyses (Table 3). The assumption that small molecule GP IIb/IIIa inhibitors and abciximab have equal efcacy during PCI resulted in further effectiveness of an upstream use strategy (5,000 years of life gained per 100,000 patients treated).

The use of the same, small molecule GP IIb/IIIa inhibitor selectively or upstream resulted in 5,000 years of life saved per 100,000 patients treated with upstream versus selective use; there was also a higher incremental lifetime cost of $848 for upstream use. Thus, the ICER was $16,960 per year of life gained. If the probability of ACS patients undergoing PCI was 65% or greater, selective use of abciximab was more effective than upstream use of small molecule GP IIb/IIIa inhibitors (Fig. 3). If the probability of PCI was greater than 60%, upstream use of small molecule GP IIb/IIIa inhibitors was

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

535

Figure 3. Benets of upstream versus selective glycoprotein (GP) IIb/IIIa use by rates of percutaneous coronary intervention (PCI). CE cost effectiveness.

not cost effective (ICER of $100,000 or greater). When the effect of GP IIb/IIIa inhibitors in medically managed patients was varied from an OR of 0.47 to 1.0 (to examine ORs from other published subanalyses), an upstream strategy was still superior (Table 2). This also held true when ORs were simultaneously varied for death and MI in two-way sensitivity analyses. The OR death or MI with small molecule versus placebo would need to be 0.97 for the two strategies to break even. Sensitivity analyses in which the costs of abciximab and small molecule GP IIb/IIIa inhibitors were equal yielded a higher incremental cost of upstream use of $848, and a cost-effectiveness ratio of $21,200. In sensitivity analyses of bleeding, rates were 5.0% in upstream use versus 4.0% in selective use of abciximab at the lower limit of the 95% condence interval of CABG-related bleeding. Bleeding rates were 14.0% with a strategy of upstream use versus 13.0% with selective use of abciximab when CABG-related bleeding was at the upper limit of the 95% condence interval.

DISCUSSION
The use of upstream GP IIb/IIIa inhibition versus selective use in only those patients undergoing PCI is a key question that clinicians face when caring for the ACS patient. It has been argued that a randomized clinical trial examining this

question is needed, but no such trial has been completed yet (28). In the absence of a specic randomized trial, a decision analysis may provide formal insight into what conditions may make one strategy potentially more benecial than another. Through an analysis that utilized randomized trial data, we found that upstream use of small molecule GP IIb/IIIa inhibitors in ACS patients is superior to selective use of abciximab. The ICER of $18,000 per year of life gained per patient with upstream use is economically attractive compared with other widely accepted cardiovascular treatments (21,29 31). Importantly, though, when using TIMI risk score as a measure of a patients baseline risk for adverse outcome, the benets of upstream GP IIb/IIIa inhibition were marginal, and thus no longer cost effective, in those patients with the lowest risk. In contrast, patients with moderate or high risk showed benet and favorable cost effectiveness with upstream GP IIb/IIIa inhibition. These ndings support the American College of Cardiology and American Heart Association recommendation to manage ACS patients with upstream GP IIb/IIIa inhibition based on level of risk (32). There has been considerable controversy about the extent of reduction in death or MI with GP IIb/IIIa inhibitors during medical management. Both the post-randomization categorization of patients into medically managed versus PCI subgroups, and the use of meta-analysis have been

536

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

criticized (14,3335). By developing a decision tree where the effects of GP IIb/IIIa inhibition in medical management may be more precisely examined, we could estimate that as long as upstream therapy reduced risk of death or MI by more than 3%, this strategy was superior. Our analysis enabled us to examine additional important questions that have not been formally tested as of yet in clinical trials. For one, it has been suggested that at proper doses of tiroban, or with the use of eptibatide, the small molecule class of GP IIb/IIIa inhibitors may be equally effective as abciximab during PCI (36 38). Our analysis suggests superiority of an upstream use strategy if one assumes all GP IIb/IIIa inhibitors have equal efcacy during PCI, but that small molecules have greater efcacy than abciximab during medical management. Second, it is not known whether the use of upstream eptibatide for all ACS patients is superior to using eptibatide selectively in only PCI patients. We found that a strategy of upstream use of eptibatide was superior to selective eptibatide, and still had a favorable ICER. The results of this analysis may also help to guide under which clinical conditions one particular therapy is more appropriate. For instance, patients who are highly (65% or greater) likely to undergo PCI may benet from a selective abciximab strategy. However, there are few cases in which a clinician is that certain the patient will undergo percutaneous revascularization. Contemporary trials of ACS have shown that a relatively small proportion (64%) of patients with ACS will subsequently undergo PCI, even with aggressive revascularization strategies (12,39). In addition, if one assumes an equivalent effect of small molecules and abciximab during PCI, then the upstream use strategy remains superior, even if PCI rates are very high. The major strength of the decision model in this study is that each comparison was based directly on randomized controlled trial data (1 6,12,13,40 44). However, there are also limitations to our analysis. First, despite using randomized trial data, we may not be able to fully account for variable patient presentation. Randomized trials may not represent the full spectrum of ACS patients; those patients at high risk for adverse outcomes may have been excluded. However, increasing patient risk of death or MI in sensitivity analyses further improved the cost-effectiveness of a strategy of upstream use. Second, our study was not designed to examine the potential benets of novel antiplatelet therapies including clopidigrel, which may reduce the baseline risk of the ACS patient and/or the risk of MI and death during PCI (45). Sensitivity analyses may provide a rough estimate of the effects of clopidigrel. For instance, if clopidigrel reduces further the overall risk for death and MI in the ACS patient by 30%, upstream GP IIb/IIIa inhibition still has a favorable ICER of $26,500 per year of life gained. However, the effect of clopidigrel in reducing potential efcacy of GP IIb/IIIa inhibitors is not known and therefore could not be studied in our analysis. Third, we did not directly take into account the timing of PCI. Markedly

earlier PCI in both strategies may diminish the benet of upstream use (39), but no randomized trial has simultaneously compared GP IIb/IIIa inhibition versus placebo and early PCI versus late PCI. Fourth, RRs from the Chimeric 7E3 Antiplatelet Therapy in Unstable Angina REfractory to Standard Treatment (CAPTURE) trial were used to estimate the effect of abciximab during PCI, but, in that trial, abciximab was administered for a period before PCI, unlike in the present model. Finally, our cost analysis did not account for the incremental costs of drug-eluting stent use, which would increase overall costs of both strategies, but should not change the incremental cost of upstream use. Indirect and long-term costs may not be fully incorporated. Further, cost and utilization data are limited to that from randomized trials, and this is a potentially signicant limitation to all cost-effectiveness analyses performed using randomized trials. A systematic approach to the appropriate selection of therapy becomes increasingly important when multiple therapeutic options, each with individually proven benets and incremental costs, are available. This analysis quanties under what conditions a strategy of upstream small molecule use is superior to selective use of abciximab, and conrms the importance of risk stratication in guiding therapy. It further suggests that the upstream use of small molecule GP IIb/IIIa inhibitors is a superior, cost-effective strategy that should be considered for those moderate and high-risk patients presenting with ACS.
Reprint requests and correspondence: Dr. Ruchira Glaser, University of Pennsylvania School of Medicine, Medicine-Cardiology, 9 Founders Cardiology-HUP, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. E-mail: ruchira.glaser@uphs.upenn.edu.

REFERENCES
1. The PURSUIT Investigators. Inhibition of platelet glycoprotein IIb/IIIa with eptibatide in patients with acute coronary syndromes. N Engl J Med 1998;339:436 43. 2. PRISM Study Investigators. A comparison of aspirin plus tiroban with aspirin plus heparin for unstable angina. N Engl J Med 1998; 338:1498 505. 3. PRISM-PLUS Study Investigators. Inhibition of the platelet glycoprotein IIb/IIIa receptor with tiroban in unstable angina and nonQwave myocardial infarction. N Engl J Med 1998;338:1488 97. 4. CAPTURE Investigators. Randomised placebo-controlled trial of abciximab before and during coronary intervention in refractory unstable angina: the CAPTURE study. Lancet 1997;349:1429 35. 5. Simoons ML and GUSTO IV Investigators. Effect of glycoprotein IIb/IIIa receptor blocker abciximab on outcome in patients with acute coronary syndromes without early coronary revascularisation: the GUSTO IV-ACS randomised trial. Lancet 2001;357:191524. 6. Topol EJ, Moliterno DJ, Herrmann HC, et al. Comparison of two platelet glycoprotein IIb/IIIa inhibitors, tiroban and abciximab, for the prevention of ischemic events with percutaneous coronary revascularization. N Engl J Med 2001;344:1888 94. 7. Antman EM, Cohen M, Bernink PJ, et al. The TIMI risk score for unstable angina/non ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;284:835 42. 8. Sabatine MS, Morrow DA, Giugliano RP, et al. Implications of upstream glycoprotein IIb/IIIa inhibition and coronary artery stenting in the invasive management of unstable angina/non ST elevation myocardial infarction: a comparison of the Thrombolysis in Myo-

JACC Vol. 47, No. 3, 2005 February 7, 2005:52937

cardial Infarction (TIMI) IIIB trial and the Treat angina with Aggrastat and determine Cost of Therapy with Invasive or Conservative Strategy (TACTICS)-TIMI 18 trial. Circulation 2004; 109:874 80. Scirica BM, Cannon CP, Antman EM, et al. Validation of the Thrombolysis In Myocardial Infarction (TIMI) risk score for unstable angina pectoris and nonST elevation myocardial infarction in the TIMI III registry. Am J Cardiol 2002;90:3035. Morrow DA, Antman EM, Snapinn SM, et al. An integrated clinical approach to predicting the benet of tiroban in nonST elevation acute coronary syndromes. Application of the TIMI risk score for UA/NSTEMI in PRISM-PLUS. Eur Heart J 2002;23:2239. Kasner SE, Kimmel SE. Accuracy of initial stroke subtype diagnosis: a decision analysis. Cerebrovasc Dis 2000;10:18 24. Cannon CP, Weintraub WS, Demopoulos LA, et al. Comparison of early invasive and conservative strategies in patients with unstable coronary syndromes treated with the glycoprotein IIb/IIIa inhibitor tiroban. N Engl J Med 2001;344:1879 87. Boersma E, Harrington RA, Moliterno DJ, et al. Platelet glycoprotein IIb/IIIa inhibitors in acute coronary syndromes: a meta-analysis of all major randomised clinical trials. Lancet 2002;359:189 98. Pieper KS, Tsiatis AA, Davidian M, et al. Differential treatment benet of platelet glycoprotein IIb/IIIa inhibition with percutaneous coronary intervention versus medical therapy for acute coronary syndromes: exploration of methods. Circulation 2004;109:641 6. Lincoff AM, LeNarz LA, Despotis GJ, et al. Abciximab and bleeding during coronary surgery: results from the EPILOG and EPISTENT trials. Improve Long-term Outcome with abciximab GP IIb/IIIa blockade. Ann Thor Surg 2000;70:516 26. Dyke CM, Bhatia D, Lorenz TJ, et al. Immediate coronary artery bypass surgery after platelet inhibition with eptibatide: results from PURSUIT. Ann Thorac Surg 2000;70:866 71. Bakhai A, Stone GW, Grines CL, et al. Cost-effectiveness of coronary stenting and abciximab for patients with acute myocardial infarction: results from the CADILLAC trial. Circulation 2003;108:2857 63. Lincoff AM, Mark DB, Tcheng JE, et al. Economic assessment of platelet glycoprotein IIb/IIIa receptor blockade with abciximab and low-dose heparin during percutaneous coronary revascularization: results from the Evaluation in PTCA to Improve Long-term Outcome with abciximab GP IIb/IIIa blockade (EPILOG) randomized trial. Circulation 2000;102:29239. Mark DB, Talley JD, Topol EJ, et al. Economic assessment of platelet glycoprotein IIb/IIIa inhibition for prevention of ischemic complications of high-risk coronary angioplasty. Circulation 1996;94:629 35. McCollam PL, Lage MJ, Bala M. A comparison of total hospital costs for percutaneous coronary intervention patients receiving abciximab versus tiroban. Catheter Cardiovasc Interv 2001;54:1527. Mahoney EM, Jurkovitz CT, Chu H, et al. Cost and cost-effectiveness of an early invasive vs conservative strategy for the treatment of unstable angina and nonST-segment elevation myocardial infarction. JAMA 2002;288:1851 8. Tsevat J, Kuntz KM, Orav EJ, et al. Cost-effectiveness of pravastatin therapy for survivors of myocardial infarction with average cholesterol levels. Am Heart J 2001;141:72734. National Center for Health Statistics. Available at: http://www. cdc.gov/nchs/nvss.htm. Accessed May 24, 2004. Abdelmeguid AE, Topol EJ, Whitlow PL, et al. Signicance of mild transient release of creatine kinase-MB fraction after percutaneous coronary interventions. Circulation 1996;94:1528 36. Mark DB, Harrington RA, Lincoff AM, et al. Cost-effectiveness of platelet glycoprotein IIb/IIIa inhibition with eptibatide in patients with nonST-elevation acute coronary syndromes. Circulation 2000; 101:366 71. Halon DA, Flugelman MY, Merdler A, et al. Long-term (10-year) outcome in patients with unstable angina pectoris treated by coronary balloon angioplasty. J Am Coll Cardiol 1998;32:16039.

Glaser et al. Cost-Effectiveness of Upstream GP IIb/IIIa

537

9.

10.

11. 12.

13. 14.

15.

16. 17. 18.

19. 20. 21.

22. 23. 24. 25.

26.

27. Catalog of Preference Scores. Available at: http://www.hcra. harvard.edu/pdf/preferencescores.pdf. Accessed April 4, 2004. 28. Braunwald E. Application of current guidelines to the management of unstable angina and nonST-elevation myocardial infarction. Circulation 2003;108:III28 37. 29. Gaspoz JM, Coxson PG, Goldman PA, et al. Cost effectiveness of aspirin, clopidogrel, or both for secondary prevention of coronary heart disease. N Engl J Med 2002;346:1800 6. 30. Delea T, Vera-Llonch M, Richner RE, et al. Cost effectiveness of carvedilol for heart failure. Am J Cardiol 1999;83:890 6. 31. Prosser L, Stinnett AA, Goldman PA, et al. Cost effectiveness of cholesterol lowering therapies according to selected patient characteristics. Ann Intern Med 2000;132:769 79. 32. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guidelines for the management of patients with unstable angina and nonSTsegment elevation myocardial infarction. A report of the American College of Cardiology/American Heart Association task force on practice guidelines. J Am Coll Cardiol 2000;36:970 1062. 33. Rof M, Chew DP, Mukherjee D, et al. Platelet glycoprotein IIb/IIIa inhibition in acute coronary syndromes. Gradient of benet related to the revascularization strategy. Eur Heart J 2002;23:1441 8. 34. Antman EM. I can see clearly now: a new view on the use of IV GP IIb/IIIa inhibitors in acute coronary syndromes. Eur Heart J 2002;23: 1408 11. 35. Kleiman NS, Lincoff AM, Flaker GC, et al. Early percutaneous coronary intervention, platelet inhibition with eptibatide, and clinical outcomes in patients with acute coronary syndromes. Circulation 2000;101:7517. 36. Harding SA, Boon NA, Flapan AD. Antiplatelet treatment in unstable angina: aspirin, clopidogrel, glycoprotein IIb/IIIa antagonist, or all three? Heart 2002;88:11 4. 37. Herrmann HC, Swierkosz TA, Kapoor S, et al. Comparison of degree of platelet inhibition by abciximab versus tiroban in patients with unstable angina pectoris and nonQ-wave myocardial infarction undergoing percutaneous coronary intervention. Am J Cardiol 2002;89: 12937. 38. Schneider DJ, Herrmann HC, Lakkis N, et al. Enhanced early inhibition of platelet aggregation with an increased bolus of tiroban. Am J Cardiol 2002;90:14213. 39. Neumann FJ, Kastrati A, Pogatsa-Murray G, et al. Evaluation of prolonged antithrombotic pretreatment (cooling-off strategy) before intervention in patients with unstable coronary syndromes: a randomized controlled trial. JAMA 2003;290:15939. 40. FRagmin and Fast Revascularisation during InStability in Coronary artery disease (FRISC) Investigators. Long-term low-molecular-mass heparin in unstable coronary-artery disease: FRISC II prospective randomised multicentre study. Lancet 1999;354:7017. 41. Fox KA, Poole-Wilson PA, Henderson RA, et al. Interventional versus conservative treatment for patients with unstable angina or nonST-elevation myocardial infarction: the British Heart Foundation Randomized Intervention Trial of unstable Angina (RITA 3) randomised trial. Lancet 2002;360:74351. 42. Boersma E, Akkerhuis KM, Theroux P, et al. Platelet glycoprotein IIb/IIIa receptor inhibition in nonST-elevation acute coronary syndromes: early benet during medical treatment only, with additional protection during percutaneous coronary intervention. Circulation 1999;100:2045 8. 43. The EPIC Investigators. Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high-risk coronary angioplasty. N Engl J Med 1994;330:956 61. 44. Stone GW, Moliterno DJ, Bertrand M, et al. Impact of clinical syndrome acuity on the differential response to 2 glycoprotein IIb/IIIa inhibitors in patients undergoing coronary stenting: the TARGET trial. Circulation 2002;105:234754. 45. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494 502.