|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Background Both aspirin and clopidogrel reduce the rate of cardiovascular events in patients with coronary heart disease. We estimated the cost effectiveness of the increased use of aspirin, clopidogrel, or both for secondary prevention in patients with coronary heart disease.
Methods We used the Coronary Heart Disease Policy Model, a computer simulation of the U.S. population, to estimate the incremental cost effectiveness (in dollars per quality-adjusted years of life gained) of four strategies in patients over 35 years of age with coronary disease from 2003 to 2027: aspirin for all eligible patients (i.e., those who were not allergic to or intolerant of aspirin), aspirin for all eligible patients plus clopidogrel for patients who were ineligible for aspirin, clopidogrel for all patients, and the combination of aspirin for all eligible patients plus clopidogrel for all patients.
Results The extension of aspirin therapy from the current levels of use to all eligible patients for 25 years would have an estimated cost-effectiveness ratio of about $11,000 per quality-adjusted year of life gained. The addition of clopidogrel for the 5 percent of patients who are ineligible for aspirin would cost about $31,000 per quality-adjusted year of life gained. Clopidogrel alone in all patients or in routine combination with aspirin had an incremental cost of more than $130,000 per quality-adjusted year of life gained and remained financially unattractive across a wide range of assumptions. However, clopidogrel alone or in combination with aspirin would cost less than $50,000 per quality-adjusted year of life gained if its price were reduced by 70 to 82 percent, to $1.00 and $0.60 per day, respectively.
Clopidogrel, a thienopyridine derivative, was shown to reduce the relative risk of ischemic stroke, myocardial infarction, or death from vascular causes in patients with prior cardiovascular disease by 8.7 percent as compared with aspirin,16 and the addition of clopidogrel to aspirin for patients with acute coronary syndromes reduced the risk of death from cardiovascular causes, reinfarction, and stroke by 20 percent as compared with aspirin alone.17 The purpose of the present study was to perform an incremental cost-effectiveness analysis of the long-term use of aspirin, clopidogrel, or both for secondary prevention in patients with known coronary disease.
Methods
The Coronary Heart Disease Policy Model18,19,20,21,22,23,24,25,26,27 is a state-transition computer simulation that predicts the incidence of coronary disease and mortality from noncoronary causes among subjects without coronary disease, stratified according to age, sex, smoking status, diastolic blood pressure, serum cholesterol level, and high-density lipoprotein level. Each year, persons without coronary disease may die of noncoronary causes, they may reach 85 years of age as survivors without coronary disease and leave the model, they may remain alive and younger than 85 years of age without coronary disease, or coronary disease may develop. When coronary disease develops in a person, the model classifies the presentation as cardiac arrest, acute myocardial infarction, or angina, and it includes deaths and health care costs during the first 30 days. Then, the model tracks patients who survive the first month with coronary disease and categorizes them according to whether they are in their first or subsequent year after the initial event and whether their history includes one or more cardiac arrests, myocardial infarctions, or coronary-revascularization procedures. Each year, patients with coronary disease have a defined risk of cardiac arrest, acute myocardial infarction, or coronary revascularization (or any combination of these events). Each event has a specific case fatality rate tailored to the condition in which the person started that year. Each patient is assigned an annual cost on the basis of his or her history and on additional costs related to any new events.
Sources of Data and Calibration of the Model
Data for the initial model were obtained from a review of the literature, the National Vital Statistics reports, the National Hospital Discharge Survey, the National Health Interview Survey, the second and third Health and Nutrition Examination Surveys, the Framingham Heart Study, and a variety of clinical trials and observational studies.18,19 The model has been updated with many revised or newly estimated variables.20,21,22,23 The model is based on the Framingham Heart Study, which has been shown to predict the benefits found in cholesterol-lowering trials.24 Using the cholesterol changes in the Scandinavian Simvastatin Survival Study,25 our model nearly perfectly reproduces the observed reduction in the rates of coronary events in that trial and provides cost-effectiveness ratios in the same general range as those estimated for that trial26 and for the Cholesterol and Recurrent Events study.27
Health-related quality-of-life weights for coronary disease are based on whether patients have angina, heart failure, or both.24 Noncoronary health-related quality-of-life weights are based on observational data.28
Interventions
Our principal simulations modeled U.S. patients, 35 to 84 years of age, in whom coronary disease developed during or before 2003 to 2027 and who survived their first month with it. Their currently expected (no-intervention) costs and quality-adjusted years of life over this 25-year period were calculated and compared with what would be expected with four strategies based on pooled data from randomized trials for secondary prevention of coronary events in patients with prior coronary disease1,16,17,29: aspirin for all eligible patients, aspirin for all eligible patients plus clopidogrel for patients ineligible for aspirin, clopidogrel alone for all patients, and the combination of aspirin for all eligible patients plus clopidogrel for all patients.
For aspirin, the 31 percent reduction in the odds of nonfatal myocardial infarction reported in the pooled trials1 was applied to myocardial infarction, cardiac arrest, and death from chronic coronary disease (Table 1). The 19 percent reduction in fatal stroke1 was used to derive a 2.8 percent reduction for the rate of death from noncoronary causes in the model. To model the effects of clopidogrel, additional relative reductions were assumed for the rates of coronary events (8.7 percent) and deaths from noncoronary causes (5.0 percent) on the basis of randomized data that directly compared aspirin with clopidogrel.16 Combination treatment with aspirin and clopidogrel was assumed to yield a 20 percent relative reduction in the rates of coronary events as compared with aspirin alone.17
|
Drug costs were estimated to be $0.04 for one 325-mg tablet of enteric-coated aspirin per day and $3.22 for one 75-mg tablet of clopidogrel.30 The cost of the combination of aspirin and clopidogrel was assumed to be the sum of the two costs.
We assumed the incidence of gastrointestinal adverse effects and rash to be as reported for aspirin and clopidogrel.16 In 1989, the cost of one major episode of gastrointestinal bleeding and the cost of one minor episode of gastrointestinal bleeding were estimated as $6,866 and $733, respectively.33 The yearly incidence of the other, less serious complications was multiplied by the cost of one office visit at $44.20, as in a prior analysis.34
In the 14 secondary-prevention trials involving high-risk patients, there was a 24 percent decrease in fatal or disabling strokes (P<0.01) and a 17 percent decrease in nondisabling strokes (P<0.09) for patients receiving aspirin. The incidence of stroke in the population with coronary disease was assumed to be the incidence reported in pooled secondary statin trials,29 with the relative distribution according to age group derived from studies conducted in Rochester, Minnesota, from 1980 to 1984.31 The in-hospital mortality from stroke (18 percent), the percentage of hospital survivors who went directly to a nursing home (15 percent), and the percentage of patients transferred to a nursing home after a rehabilitation center (8 percent) were derived from Dobkin,35 whereas the percentage of survivors discharged to a rehabilitation center from acute-care hospitals (6.8 percent) was derived from Oster et al.34 The cost of acute care for stroke (hospital costs plus physicians' fees) was reported to be $7,026 in 1991.34 From work of the same authors, we derived the costs for one stay in a rehabilitation service ($40,793), the cost for one year in a nursing home ($26,620), the yearly costs for outpatient services and home care ($1,212), and the yearly costs for recurrent strokes ($624).
Total costs were calculated as the sum of costs of coronary disease, costs of noncoronary disease (an annual estimate based on data from the National Medical Expenditure Survey), and the costs of the specific intervention being studied, and were summed from 2003 to 2027 with the use of a discount rate of 3 percent per year. All costs were converted to year-2000 U.S. dollars with the use of the medical care component of the Consumer Price Index.
Sensitivity Analyses
Lower and upper bounds of the percent reductions in the odds of coronary events with aspirin were based on the Antiplatelet Trialists' reported standard deviation.1 For clopidogrel as compared with aspirin and the combination of the two, we used the 95 percent confidence intervals of the relative reductions.16,17
Because the median follow-up time in the secondary-prevention trials for high-risk patients was three years,1 we modeled interventions with benefits limited to three years, whereas drug-related complications and costs continued for 25 years or just 3 years. We examined cost effectiveness in subgroups of differing risk according to age and clinical characteristics, and we assessed the cost effectiveness of the interventions, assuming that they might have as great an effect on reducing coronary revascularization procedures as on reducing other coronary events.
We varied the health care costs of noncoronary disease by up to 100 percent and assessed the effect of excluding them from our analysis. We simulated a higher annual discount rate of 5 percent. The cost effectiveness of clopidogrel as compared with aspirin was assessed for a wide range of drug costs.
Results
As compared with the estimated current utilization of aspirin, extension of aspirin therapy to all eligible patients would result in an additional $189 million in drug costs and $8 billion in overall costs from 2003 to 2027 in patients 35 to 84 years of age (Table 2). The benefits, however, would be substantial, with the avoidance of about 155,000 myocardial infarctions and a gain of an additional 682,000 quality-adjusted years of life over the same period. As compared with no aspirin, the use of aspirin in all eligible patients would save an estimated 6.9 million quality-adjusted years between 2003 and 2027.
|
According to these projections, the estimated cost effectiveness of extending aspirin therapy to all eligible patients is favorable by any measure: with our base-line estimates, the ratio would be about $11,000 per quality-adjusted year of life saved. The addition of clopidogrel for the estimated 5.7 percent of patients who are ineligible for aspirin is also associated with a reasonable cost-effectiveness ratio of about $31,000 per quality-adjusted year of life saved. By comparison, either the strategy of routine use of clopidogrel alone in all patients or the strategy of combined aspirin plus clopidogrel in patients who are eligible for aspirin and clopidogrel alone in patients who are ineligible for aspirin would be associated with cost-effectiveness ratios of well over $100,000 as compared with aspirin alone or with the routine use of aspirin complemented by the use of clopidogrel in patients who are ineligible for aspirin.
With aspirin therapy, the costs of coronary heart disease would decline substantially in the first several years (Figure 1). However, the costs of noncoronary disease and later costs related to coronary disease would increase, because more patients would be alive with coronary disease and susceptible to recurrent coronary events. In analyses that considered only patients with prevalent coronary disease in 2002 and did not include patients with incident cases each year, the cost-effectiveness ratios over the 25-year simulation were very similar.
|
If the rate of aspirin use in eligible patients were only 42 percent instead of 85 percent, all cost-effectiveness ratios would remain the same, but the absolute benefits of current aspirin use would be about 50 percent of those reported in Table 2. Aspirin has a more favorable cost-effectiveness ratio ($3,000 per quality-adjusted year of life gained) if the health care costs of noncoronary disease are not considered (Table 3). The use of aspirin and the use of clopidogrel in patients who are intolerant of aspirin would save money as well as lives if these strategies reduce the rate of revascularization as much as they reduce the rate of myocardial infarction. Results were similar according to sex and age, even if treatment continued beyond the age of 85. Conversely, if the benefits of therapy persisted for only 3 years even though therapy was continued for 25 years, all options would become much less attractive.
|
The substitution of clopidogrel for aspirin or the addition of clopidogrel to aspirin in patients who are eligible for aspirin would become attractive, however, if the cost of clopidogrel declined substantially. For example, in our base-line analysis, the cost-effectiveness ratio of clopidogrel as compared with aspirin would fall to $50,000 per quality-adjusted year of life gained if the cost of clopidogrel were reduced by 82 percent, from $3.22 daily to $0.60 daily. For the combination of aspirin plus clopidogrel, the daily price of clopidogrel would have to fall by 70 percent, to about $1 daily, for a cost-effectiveness ratio of $50,000 per quality-adjusted year of life gained.
Discussion
The prescription of aspirin until death or for 25 years has an attractive cost effectiveness in men and women with coronary disease across all age ranges and despite varying assumptions about the efficacy of treatment. For patients with contraindications to aspirin treatment, clopidogrel had a reasonably attractive cost-effectiveness ratio as compared with no antiplatelet treatment. By comparison, the incremental cost-effectiveness ratio of clopidogrel as compared with aspirin for patients who are eligible for aspirin was unattractive across a wide range of assumptions, because of the higher daily costs of the drug itself. Clopidogrel reached favorable cost-effectiveness ratios only when its costs were reduced to about $0.60 per day. Clopidogrel used in combination with aspirin for all patients who were eligible for aspirin also had unattractive cost-effectiveness ratios, even if the health benefits described for patients with acute coronary syndromes17 were maintained in the long term. To date, available data have not clearly demonstrated an increased risk of thrombotic thrombocytopenic purpura with clopidogrel treatment.36,37 If such an association exists, clopidogrel would become even less attractive.
Though favorable, the annual overall cost effectiveness of aspirin therapy was not as favorable as might have been expected given the very low cost of aspirin itself. The main explanation is that the health care costs of noncoronary disease would be estimated to increase substantially, because patients whose cardiac events were prevented by aspirin would survive to have other medical costs. In the first several years of therapy, these other medical costs would be offset by the savings generated from the prevention of coronary events. Subsequently, however, costs related to coronary disease would also increase, because the prevalence of persons alive with coronary disease, and hence susceptible to coronary events, would be greatly increased because of deaths prevented by aspirin therapy.
Our findings are much less favorable for clopidogrel than those of Sarasin et al.,38 who reported a cost-effectiveness ratio of about $27,000 per quality-adjusted year of life gained for secondary prevention in patients with prior strokes or transient ischemic attacks. Those authors modeled clopidogrel use in highly selected patients who were 65 years of age and were not candidates for carotid surgery. They assumed an additional 14 percent reduction in vascular events with clopidogrel as compared with aspirin, a benefit that was 1.6 times as high as current data suggest. They did not consider downstream coronary costs, however, other than for myocardial infarction, or the costs of noncoronary disease, other than direct adverse effects of antiplatelet treatment. If we eliminated the costs considered in our study but not theirs, estimates of the cost effectiveness of clopidogrel in the two analyses would be similar.
Our findings represent a conservative assessment of the benefits of aspirin for secondary prevention of coronary disease. First, we modeled the effects of aspirin during long-term use when given to patients 30 days after they had survived an initial coronary event. Large, randomized trials39 have also shown short-term benefits of aspirin for patients in the acute phase of myocardial infarction, in particular when combined with thrombolysis. The administration of aspirin in the acute phase of myocardial infarction has been estimated to cost $2,800 per year of life saved.40 Data also suggest that the long-term benefits of aspirin, when administered with thrombolysis, may be substantially greater than previously reported.41 Second, we assumed that the daily dose of aspirin was 325 mg per day, because that regimen was the one most commonly used in the United States. There is good evidence that 100 mg per day could be as effective and safer.5 Third, we used the cost of the enteric-coated aspirin tablets, which may trigger fewer gastrointestinal complications, rather than other, less costly formulations.
Aspirin for secondary prevention of coronary disease is attractive from a cost-effectiveness perspective under a wide range of assumptions. Clopidogrel, as currently priced, has an attractive cost-effectiveness ratio for patients with contraindications to aspirin but not for patients who can tolerate aspirin, whether used alone or in combination with aspirin. The gap between proven effectiveness and unattractive projected cost effectiveness could be eliminated by reductions in the price of clopidogrel.
Supported in part by grants from the Agency for Health Care Policy and Research (RO1 HS06258) and the National Heart, Lung, and Blood Institute (RO1 HL46315).
Source Information
From the Clinique de Médecine II and the Division of Cardiology, Hôpitaux Universitaires, Geneva (J.-M.G.); the Department of Medicine, University of California, San Francisco, School of Medicine, San Francisco (P.G.C., L.G.); the Department of Health Policy and Management, Harvard School of Public Health, Boston (P.A.G., L.W.W., K.M.K., M.G.M.H.); and the Department of Epidemiology and Biostatistics and the Department of Radiology, Erasmus University Medical School, Rotterdam, the Netherlands (M.G.M.H.).
Address reprint requests to Dr. Gaspoz at the Clinique de Médecine II, Department of Medicine, Hôpitaux Universitaires, 24 rue Micheli-du-Crest, 1211 Geneva 14, Switzerland, or at jean-michel.gaspoz{at}hcuge.ch.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Aspirin, Clopidogrel, or Both for Secondary Prevention of Coronary Disease
Akinlade B. K., de Lemos J. A., McGuire D. K., Armstrong E. C., Ramsey S. D., Yaes R. J., Gaspoz J.-M., Coxson P., Goldman L., Wood A. J.J.
Extract |
Full Text |
PDF
N Engl J Med 2003;
348:560-563, Feb 6, 2003.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |