Correspondence
Thrombolytic Therapy for Acute Myocardial Infarction: GUSTO Criticized
N Engl J Med 1994; 330:504-506February 17, 1994
- Article
To the Editor:
After an intensive marketing campaign by Genentech, the report by the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) investigators (Sept. 2 issue)1 has finally appeared. In view of the small survival advantage conferred by tissue plasminogen activator (t-PA) (0.9 percent -- an effect evident only when it is administered within four hours after the onset of symptoms), the increased risk of stroke, and the striking difference in the expense of therapy (a difference of $220,000 to save a single life: $2,300 - $320 × 100 divided by 0.9), the fact that t-PA is the most frequently used thrombolytic agent in the United States is indeed puzzling. Sharp differences in patient grouping and the lack of a consistent number of patients in the groups render the statistical differences suspect. The conclusions about the effects of treatment according to age are also of concern: the 0.9 percent increase in the incidence of hemorrhagic stroke observed with t-PA in patients older than 75 years had no effect on the resulting differences in the outcomes of death or nonfatal disabling stroke, whereas in patients ≤ 75 years a 0.1 percent increase in this complication resulted in a 0.1 percent reduction in the combined outcome. The occurrence of revascularization in 24 percent of the patients may be an important confounder; however, the distribution of revascularization according to groups has been pointedly omitted: in fact, 0.5 percent more patients had coronary angioplasty and 0.8 percent more had coronary-bypass surgery in the t-PA group than in the two streptokinase groups combined. Also, even though more than 75 percent of the study group received thrombolytic therapy within four hours of the onset of symptoms, the median time for a patient with an acute myocardial infarction to arrive at a treatment facility exceeds the length of time in which treatment can confer a possible advantage. Indeed, if t-PA was the only drug used to treat myocardial infarction, it might result in no overall benefit and possibly in a worsening of outcome.
This investigation was designed to test treatment strategies for t-PA when previous studies had shown no advantage. If the object had been to determine the proper application of these agents, “front-loading” strategies for streptokinase and late administration of these drugs -- situations in which streptokinase has been shown to be beneficial -- might also have been tried. In response to the prepublication publicity, many institutions have begun to use t-PA. Such an action appears to have been entirely premature. Because of numerous scientific issues, the purported marginal benefits of t-PA over streptokinase remain uncertain. What is incontrovertible is the marketing success of the purveyors of t-PA.
1 ReferencesHoward S. Friedman, M.D.
Long Island College Hospital, Brooklyn, NY 112011
The GUSTO Investigators
Full Text | Web of Science | Medline
To the Editor:
The GUSTO trial compared conventional thrombolytic regimens with treatment with accelerated t-PA and intravenous heparin. As reported, a statistically significant survival benefit was observed with accelerated t-PA and heparin. What are the clinical implications of this finding?
This trial serves as a reminder of the difficulty clinicians face when considering the clinical impact of large multicenter trials with several treatment groups. The large sample provides statistical power to detect small, yet significant, differences whose clinical importance may in fact be difficult to assess. Feinstein has suggested that a valuable approach to presenting such data is to estimate the total number of patients who will require treatment in order to produce the observed therapeutic difference in one person1. The GUSTO results suggest that 91 to 111 patients (with streptokinase groups used as controls) would require treatment with accelerated t-PA to produce one additional survivor 30 days after myocardial infarction. This manner of presenting findings is more informative than one that presents findings in terms of the percent difference between treatment and control groups, because small differences may be magnified and perceived as profoundly important clinically.
In the current health care environment, an evaluation of costs associated with expensive therapies and forms of technology is warranted2. Such an analysis was not performed in the GUSTO trial. Using the GUSTO results and our hospital's pharmacy costs of $297 for a typical course of treatment with streptokinase and $2,184 for a course of t-PA, we calculated that the added cost of producing one additional survivor would be $188,000. This estimate is similar to the $200,000 estimate by Brody et al.3. Furthermore, the GUSTO study presumes that the additional survivor at 30 days has a life expectancy similar to that of those who survived in the control groups. If, however, the benefit diminishes after 30 days, the cost to society per life saved increases further. Given the natural history of ischemic heart disease and acute myocardial infarction, it is by no means certain that survival at 30 days confers any long-term benefit.
Physicians are faced daily with the reality of balancing appropriate and justifiable patient care needs with the social imperative to consider the economic impact of clinical practice. To make informed clinical decisions about new or modified treatment regimens, information on costs and long-term benefits is crucial. We, therefore, implore all journal editors to insist that data be presented in more useful formats and that investigators in large trials such as GUSTO include an evaluation of the resources required to implement positive findings.
3 ReferencesLakshmipathi Chelluri, M.D.
Carl A. Sirio, M.D.
Derek C. Angus, M.B., Ch.B., M.P.H.
University of Pittsburgh Medical Center, Pittsburgh, PA 152131
Feinstein AR . Invidious comparisons and unmet clinical challenges. Am J Med 1992;92:117-120
CrossRef | Web of Science | Medline2
Welch HG . Valuing clinical strategies early in their development. Ann Intern Med 1992;116:263-264
Web of Science | Medline3
Brody B ,Wray N ,Bame S ,Ashton C ,Petersen N ,Harward M . The impact of economic considerations on clinical decisionmaking: the case of thrombolytic therapy. Med Care 1991;29:899-910
CrossRef | Web of Science | Medline
To the Editor:
The recommendation by Dr. Fuster regarding the use of heparin following streptokinase therapy (Sept. 2 issue)1 deserves clarification. Although the role of heparin as an adjunct to thrombolysis may be controversial, its use in reducing the occurrence of left ventricular thrombosis after anterior myocardial infarction is established2,3. Although streptokinase has been shown to reduce ventricular thrombosis,4 the incidence of left ventricular thrombosis and the morbidity and long-term mortality associated with it in the thrombolytic era are not negligible5.
Since ventricular thrombosis is best prevented by the initiation of high-dose heparin soon after the diagnosis of anterior myocardial infarction, an unqualified suggestion that heparin therapy after streptokinase therapy is of little benefit may lead to erroneous generalizations that ignore its long-term benefits for the prevention of stroke and arterial emboli.
5 ReferencesKofo Ogunyankin, M.D.
Nassau County Medical Center, East Meadow, NY 115541
Fuster V . Coronary thrombolysis -- a perspective for the practicing physician. N Engl J Med 1993;329:723-725
Full Text | Web of Science | Medline2
Turpie AG . Prevention of left ventricular mural thrombus. Am J Cardiol 1989;64:41B-43B
CrossRef | Web of Science | Medline3
The SCATI (Studio sulla Calciparina nell'Angina e nella Trombosi Ventricolare nell'Infarto) Group
Web of Science | Medline4
Destro G ,Barbieri E ,Bicego D ,Zanolla L ,Franceschini L ,Zardini P . Acute anterior myocardial infarction: streptokinase prevents ventricular thrombosis independently of its effect on infarct size. Clin Cardiol 1990;13:789-793
CrossRef | Web of Science | Medline5
Cregler LL . Antithrombotic therapy in left ventricular thrombosis and systemic embolism. Am Heart J 1992;123:4:1110-1114
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: We were surprised by many of the comments of Dr. Friedman, who has not separated the research of a large-scale, multinational effort from the interests of industry. We took special care to distance ourselves from the five sponsors of the GUSTO trial, and particularly the company with much to gain (or lose), by unanimously adopting strict conflict-of-interest guidelines for the conduct of the trial1. The importance of a 14 percent relative reduction in mortality should not be minimized, in view of the fact that nearly 200,000 patients receive myocardial reperfusion therapy each year in the United States. The 1 percent absolute reduction represents a major stepwise increment in survival beyond that of the established standard (streptokinase), for which there was a 2.6 percent absolute reduction in mortality2. It is critical to remember that we are no longer in a placebo-controlled era and cannot expect as large a benefit of a new strategy as was encountered a decade ago.
The difference in the number of patients per treatment strategy is explained in the article: the treatment arm consisting of streptokinase and subcutaneous heparin was instituted as soon as possible after the results of the Third International Study of Infarct Survival became known. There was no effect of revascularization on the differences in mortality. The treatment benefit of accelerated t-PA remained significant (P = 0.001) after an adjustment for the effects of coronary-artery bypass surgery. There are no data supporting the rapid achievement of complete patency of infarcted vessels with accelerated streptokinase therapy.
With respect to subgroup analyses, we must point out that these are underpowered comparisons that are unidimensional and fraught with the possibility of false negative results. More important than performing a subgroup analysis is considering the direction of benefit and whether there is consistency. As it turned out in the GUSTO analyses, there was hardly a subgroup that did not benefit from accelerated t-PA; it was only a matter of degree. For example, patients over the age of 75 years had an absolute benefit (12 lives saved per 1000 patients treated) that was similar to that of the overall study population. The reason for the net benefit in the presence of a higher stroke rate was that in the elderly, stroke was much more commonly a fatal event and thus already “counted” in the mortality rate. With the overwhelming consistency of benefit that we observed, the only reason to use subgroup analyses would be to promote health care rationing, with the use of probability estimates to concentrate the use of the more expensive strategy in patients who will gain the most.
We agree with the calculations of Dr. Friedman and Dr. Chelluri and colleagues of the initial cost of saving a life with accelerated t-PA -- approximately $200,000. However, these calculations do not take into account the fact that the average survivor's life expectancy is 9 to 11 years, producing the more appropriate calculation of approximately $20,000 per quality-adjusted year of life3. This is relatively inexpensive when one compares it with the cost of dialysis or hypolipidemic or antihypertensive therapies, which range from $35,000 to $180,000 per quality-adjusted year of life4. Although we have no intention of defending the high price of this specific thrombolytic strategy, we have demonstrated the superior recanalization profile of accelerated t-PA as compared with streptokinase5. Most important, the achievement of rapid, complete, and sustained recanalization of the infarcted vessel has been validated by the GUSTO trial as the therapeutic goal of myocardial reperfusion. We hope that this vital confirmation of the underlying pathophysiologic process and treatment advances will not be overshadowed by legitimate concern over health care costs.
5 ReferencesEric J. Topol, M.D.
Cleveland Clinic Foundation, Cleveland, OH 44195Robert M. Califf, M.D.
Duke University Medical Center, Durham, NC 27710Frans Van de Werf, M.D.
Catholic University, 3000 Leuven, Belgiumfor the GUSTO Steering Committee
1
Topol EJ ,Armstrong P ,Van de Werf F , et al. Confronting the issues of patient safety and investigator conflict of interest in an international clinical trial of myocardial reperfusion. J Am Coll Cardiol 1992;19:1123-1128
CrossRef | Web of Science | Medline2
Topol EJ. Thrombolytic intervention. In: Topol EJ, ed. Textbook of interventional cardiology. 2nd ed. Philadelphia: W.B. Saunders, 1994:68-111.
3
Goel V ,Naylor CD . Potential cost effectiveness of intravenous tissue plasminogen activator versus streptokinase for acute myocardial infarction. Can J Cardiol 1992;8:31-38
Web of Science | Medline4
Mark DB. Medical economics and health policy issues for interventional cardiology. In: Topol EJ, ed. Textbook of interventional cardiology. 2nd ed. Philadelphia: W.B. Saunders, 1994:1323-53.
5
The GUSTO Angiographic Investigators
Full Text | Web of Science | Medline
Author/Editor Response
I agree with Dr. Ogunyankin that although the role of heparin as an adjunct to thrombolysis, particularly when used with streptokinase or anisoylated plasminogen-streptokinase activator complex (APSAC), may be controversial,1 its use in reducing the occurrence of left ventricular thrombosis after anterior myocardial infarction is established. I think the following points are important.
The role of heparin as adjunctive therapy to streptokinase and APSAC is open to debate2. However, new information from the GUSTO angiographic substudy shows in all the treatment groups, including those with streptokinase, the lowest reocclusion rates (only 5.3 to 6.9 percent from 90 minutes to 5 to 7 days after infarction) for a large trial of aspirin (which was given to all patients) plus continuous heparin3. It is also possible that new antithrombins, such as hirudin and hirulog, will replace heparin in the future in patients treated for myocardial infarction, regardless of the thrombolytic agent to be used. My editorial focused on the role of heparin in coronary reperfusion and not on adjunctive therapy for other cardiac processes at the time of infarction, such as thromboembolism related to left ventricular thrombosis. In cases of large anterior myocardial infarction, a high dose of subcutaneous or intravenous heparin should indeed be used during hospitalization for the prevention of left ventricular thrombosis and systemic thromboembolism; nevertheless, at the time of discharge, the questions of whether thrombus is activated if warfarin therapy is not overlapped properly with heparin therapy4 and of which patients need long-term anticoagulation for the prevention of left ventricular thrombus5 are now open to debate.
5 ReferencesValentin Fuster, M.D., Ph.D.
Massachusetts General Hospital, Boston, MA 021141
Fuster V . Coronary thrombolysis -- a perspective for the practicing physician. N Engl J Med 1993;329:723-725
Full Text | Web of Science | Medline2
Ridker PM ,Hebert PR ,Fuster V ,Hennekens CH . Are both aspirin and heparin justified as adjuncts to thrombolytic therapy for acute myocardial infarction? Lancet 1993;341:1574-1577
CrossRef | Web of Science | Medline3
The GUSTO Angiographic Investigators
Full Text | Web of Science | Medline4
Kontny F ,Dale J ,Abildgaard U ,Hegrenaes L ,Lem P ,Morstol T . Adverse effect of warfarin in acute myocardial infarction: increased left ventricular thrombus formation in patients not treated with high-dose heparin. Eur Heart J 1993;14:1040-1043
CrossRef | Web of Science | Medline5
Tanne D ,Goldbourt U ,Zion M ,Reicher-Reiss H ,Kaplinsky E ,Behar S . Frequency and prognosis of stroke/TIA among 4808 survivors of acute myocardial infarction. Stroke 1993;24:1490-1495
CrossRef | Web of Science | Medline
- Citing Articles (1)
Citing Articles
1
P.P. Hujoel, G.C. Armitage, R.I. García. (2000) A Perspective on Clinical Significance. Journal of Periodontology 71:9, 1515-1518
CrossRef
