To the Editor: Gaspoz et al. (June 6 issue)1 present an interestingperspective on the problem of escalating health care costs.Their comparison between the cost effectiveness of aspirin andthat of clopidogrel is commendable, given the increasing focusby the public on the costs of newer drugs. In their analysis,the authors' assumptions about the costs of the drugs do nottake into consideration future costs that would be expectedto be lower for both brand-name and generic versions of clopidogrel.
Estimates of the cost of developing a new drug vary, with somefigures as high as $800 million.2 The need to recoup these expensesis one of many reasons for the price of new drugs. Without themarketing of new drugs, it is doubtful whether lower-pricedgeneric versions would become available once the patents hadexpired; if they did not, the public would be deprived of therapeuticallysuperior medications. Clopidogrel has been shown to be moreeffective than aspirin alone in reducing the incidence of cardiovascularevents.3 The authors' conclusion that the use of clopidogrelis financially unattractive appears to sound a death knell fortherapeutic innovation and to mark the beginning of a managed-careera for the pharmaceutical industry.
Bolanle K. Akinlade, M.D. Community Medical Centers Stockton, CA 95202
References
Gaspoz J-M, 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;346:1800-1806. [Free Full Text]
Tufts Center for the Study of Drug Development pegs cost of a new prescription medicine at $802 million. News release of the Tufts Center for the Study of Drug Development, Boston, November 30, 2001.
CAPRIE Steering Committee. A randomised, blinded trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329-1339. [CrossRef][ISI][Medline]
To the Editor: We believe that the estimates of incrementaleffectiveness used by Gaspoz et al. are incorrect for each ofthe strategies evaluated. First, the 31 percent relative riskreduction associated with aspirin therapy is based on a meta-analysis1of trials performed before statins, angiotensin-converting–enzymeinhibitors, glycoprotein IIb/IIIa inhibitors, and coronary stentswere used routinely after acute coronary syndromes. A more recentreport from the same group suggests that there was a 27.7 percentrelative risk reduction for nonfatal myocardial infarction andonly a 21 percent relative risk reduction for major cardiovascularevents.2
Second, it appears that the incremental effects of clopidogreland of clopidogrel plus aspirin were underestimated. Assuminga 31 percent relative risk reduction from aspirin, an incremental20 percent relative risk reduction associated with the additionof clopidogrel should yield a net 44.8 percent relative riskreduction, as opposed to the 37.2 percent used by Gaspoz etal. Similarly, the relative risk reduction with clopidogrelalone should be 37 percent, rather than the 33.7 percent reported.The correct formula is RRRcombined = 1 – [(1 – RRRaspirin)x (1 – RRRaspirin+clopidogrel)], where RRR is the relativerisk reduction; the formula used by Gaspoz et al. was RRRcombined= RRRaspirin + (RRRaspirin x RRRaspirin+clopidogrel).
The assumptions of efficacy and safety used in this model aretaken from a short-term study involving patients treated earlyafter the onset of an acute coronary syndrome, for a maximumof 12 months.3 The extrapolation of such results to long-termtherapy is inherently flawed. The rate of adverse ischemic eventsis highest during the period immediately after an acute coronarysyndrome and decreases thereafter. Episodes of bleeding, onthe other hand, would probably follow a more "linear" time course.Because the risks and benefits of 25 years of treatment withthe combination of clopidogrel and aspirin have not been established,the cost effectiveness of this strategy is irrelevant.
James A. de Lemos, M.D. Darren K. McGuire, M.D. University of Texas Southwestern Medical Center Dallas, TX 75390-9047 james.delemos{at}utsouthwestern.edu
Editor's note: Drs. de Lemos and McGuire report having receivedspeaker's honorariums from Sanofi, which manufactures clopidogrel.Dr. de Lemos reports having received speaker's honorariums fromBristol-Myers Squibb, which markets clopidogrel.
References
Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy. I. Prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994;308:81-106. [Erratum, BMJ 1994;308:1540.] [Free Full Text]
Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002;324:71-86. [Erratum, BMJ 2002;324:141.] [Free Full Text]
The Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. 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. [Erratum, N Engl J Med 2001;345:1506, 1716.] [Free Full Text]
To the Editor: I use the Clopidogrel versus Aspirin in Patientsat Risk of Ischaemic Events (CAPRIE) study1 to teach residentsthe importance of avoiding the use of relative risk reductionsin describing treatment effects and instead using absolute riskreductions and numbers needed to treat. Treatment with clopidogrelresulted in an 8.7 percent relative risk reduction for the compositeprimary end point of ischemic stroke, myocardial infarction,or death from vascular causes, as compared with aspirin. Atfirst glance, this reduction seems impressive, but closer scrutinyreveals an absolute risk reduction of only 0.9 percent and anumber needed to treat of 115 (95 percent confidence interval,58 to 8647). This huge 95 percent confidence interval, whichranges from a number needed to treat that would be worthwhileto one that would offer no advantage at all, calls into questionwhether clopidogrel is truly superior to aspirin. On the basisof the number needed to treat of 115, 114 patients would haveto be treated for 730 days (2 years) at a cost of $3.22 pertablet in order to prevent 1 patient from having an adverseevent. The cost would be $267,968 per adverse event prevented.A Cochrane review2 that included the findings of the CAPRIEtrial presented data for single outcomes rather than for a clusterof clinical outcomes. According to these analyses, the differencesbetween clopidogrel and aspirin in terms of the outcomes oftotal strokes and total deaths are nonsignificant. For totalmyocardial infarctions, the absolute risk reduction is 0.7 percent,giving a number needed to treat of 143, and the cost per myocardialinfarction prevented over a two-year period is a staggering$333,785.
Coronary heart disease is a major killer of adults in the UnitedStates. An estimated 40 million Americans are uninsured, andthose who are insured are paying more but getting less. In thiscontext, I would replace the somewhat euphemistic term "unattractive"used by Gaspoz et al. with the term "prohibitive" in describingthe cost effectiveness of routine use of clopidogrel for secondaryprevention of coronary heart disease.
Eamon C. Armstrong, M.D. Lehigh Valley Hospital Allentown, PA 18105
References
CAPRIE Steering Committee. A randomised, blinded trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329-1339. [CrossRef][ISI][Medline]
Hankey GJ, Sudlow CLM, Dunbabin DW. Thienopyridine derivatives (ticlopidine, clopidogrel) versus aspirin for preventing stroke and other serious vascular events in high vascular risk patients. Cochrane Database Syst Rev 2000;2:CD001246-CD001246. [Medline]
To the Editor: In his commentary on the study by Gaspoz et al.,Wood1 states that he finds it "deeply troubling" that the authorsinterpret $130,000 per quality-adjusted year of life for clopidogrelas an "unattractive" value for expenditure. He then cites thecost of drug development as reasonable justification for theprice of the drug. Unfortunately, his link between drug pricingand drug-development costs — a link often made by pharmaceuticalmanufacturers — is spurious. The $3.22 price of a 75-mgclopidogrel tablet has very little to do with the cost of developingthe drug. Rather, it is what the managers at the companies thatmarket clopidogrel (Bristol-Myers Squibb and Sanofi Pharmaceuticals)believe the market will bear for this product. What troublesme is that Wood appears to reject the fact that health valuefor expenditure should play any part in the argument, particularlywhen the quality-adjusted years of life are his. Public andprivate health insurance plans face budget constraints and severepressure to curb future increases in health care expenditures.Those who make the difficult decisions for these organizationsknow that the incremental cost of adding clopidogrel to theirformularies will force them to cut costs elsewhere, perhapsby restricting access to much more cost-effective therapies.Even if the lives saved by those treatments are not his, whichones would he choose?
Scott D. Ramsey, M.D., Ph.D. Fred Hutchinson Cancer Research Center Seattle, WA 98109 sramsey{at}fhcrc.org
Editor's note: Dr. Ramsey reports having received a researchgrant from Bristol-Myers Squibb.
References
Wood AJJ. When increased therapeutic benefit comes at increased cost. N Engl J Med 2002;346:1819-1821. [Free Full Text]
To the Editor: Wood seems to be arguing that we should be willingto pay any increase in cost, no matter how large, in returnfor any increase in benefit, no matter how minuscule. Such anattitude would certainly cost more lives than it would save.Forty million Americans lack health insurance and thereforelack access to even basic health care. As medical costs continueto rise, more and more employers are forced to drop health carecoverage. Nobody knows how many people die because lack of insurancecauses them to delay seeking medical care until it is too late,but the number is certainly larger than the number of livessaved by adding clopidogrel to aspirin therapy.
Robert J. Yaes, M.D. 15 Quantum Pl. Gaithersburg, MD 20877
The authors reply: We disagree with Dr. Akinlade. Pharmaceuticalcompanies should be able to achieve profits at reasonable cost-effectivenessratios so that beneficial drugs are affordable. Future drugprices are unpredictable, and the patents on clopidogrel rununtil 2019.
Dr. Armstrong calculates the costs per event avoided to reachconclusions that are stronger than ours. We prefer the term"unattractive" but understand his preference for "prohibitive."
Drs. de Lemos and McGuire raise three issues. First, our estimated31 percent reduction in the risk of cardiovascular events withaspirin is consistent with the 30 percent reduction in the oddsof nonfatal myocardial infarction among patients with previousinfarction1 and the 32 percent reduction in the odds of vascularevents among patients receiving moderate doses of aspirin2 inthe most recent overview.
Second, we appreciate their carefulness in detecting our mathematicalerror and apologize for it. A corrected version of our articleis now available at http://www.nejm.org. With the correctionof this error and a similar error in estimating the benefitsin reducing the risk of stroke, the use of clopidogrel aloneinstead of aspirin is associated with a cost-effectiveness ratioof $110,000 per quality-adjusted year of life saved and remainsunattractive except for patients with the highest risk. Forthe combination of clopidogrel and aspirin, the cost-effectivenessratio changed from $130,000 to $61,000 per quality-adjustedyear of life saved, on the basis of our original assumptionsthat had been purposely tilted to favor clopidogrel in orderto ensure the robustness of our conclusion that clopidogrelwas unattractive from the perspective of cost effectivenessdespite the most favorable set of estimates. These assumptions,which were unlikely to be accurate, were that the 20 percentbenefit of combination therapy for the prevention of nonfatalmyocardial infarction that was found during the first year inpatients with acute coronary syndromes would be maintained for25 years and that the same benefit would apply to all cardiovascularevents. However, the observed relative benefits of combinationtherapy declined by about 50 percent during months 9 through12 of the trial, and the reductions in the rates of stroke,fatal myocardial infarction, and death from cardiovascular causeswere much lower than 20 percent, even during the first year.3If the relative benefit of combination therapy in terms of allevents after the first year were similar to what was seen duringmonths 9 through 12, the cost-effectiveness ratio would be about$120,000 per quality-adjusted year of life saved — nearlyidentical to the $130,000 we estimated.
Third, when the actual event-specific results reported for 12months of combination therapy3 were applied to our originalquestion, which was about the cost effectiveness of therapybeginning 30 days after the onset of symptomatic coronary disease,the 25-year cost-effectiveness ratio for 1 year of combinationtreatment was $180,000 per quality-adjusted year of life gained.As a result, we stand by our original conclusion that the long-termuse of clopidogrel, despite its apparent effectiveness, is financiallyunattractive for patients who can tolerate aspirin, unless itsprice is reduced substantially.
Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002;324:71-86. [Erratum, BMJ 2002;324:141.]
The Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. 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. [Erratum, N Engl J Med 2001;345:1506, 1716.] [Free Full Text]
Dr. Wood replies: Gaspoz et al. demonstrated that the additionof clopidogrel to aspirin therapy would produce a benefit of1,437,000 quality-adjusted years of life over a 25-year periodas compared with aspirin therapy alone — hardly a "minuscule"benefit, as Dr. Yaes suggests. The cost of these years of lifewill fall with the price of clopidogrel, as the authors acknowledged.Dr. Ramsey takes issue with the ways in which drugs are priced.Although the pricing of drugs and the fact that 40 million Americanslack health insurance are serious issues, the fundamental issueis whether we should ration care, and if so, how that rationingshould be carried out. I would argue that our expenditures forhealth care, which, as a fraction of the gross domestic product,are higher than those of any other developed country, couldcertainly support therapies that have been proved to resultin substantial reductions in mortality in well-controlled clinicaltrials. Surely, the first therapies to be eliminated shouldbe those that we know to be ineffective or those that have notbeen demonstrated to be effective. Too many such therapies arestill in widespread use. If society deems it appropriate torestrict care further, it will be critical for physicians tobe explicit about such restrictions in talking to their patients,who might choose to deploy their economic resources differently.Patients expect us to be frank and open about the options werecommend to them. To deny patients potentially lifesaving therapywithout offering them the option of receiving it (even if itmust be at their own expense) seems unethical. Reasonable peoplecan differ in their judgments of economic value; if you doubtit, look at the variety of cars in any large parking lot.
Alastair J.J. Wood, M.D. Vanderbilt University Nashville, TN 37232-6602 alastair.wood{at}vanderbilt.edu
Correction
to
Gaspoz et al., 
Previous
