684
Editorial
Nonfatal Myocardial Infarction Is, by Itself,
an Inappropriate End Point in Clinical Trials
in Cardiology
Joseph L. Fleiss, PhD, J. Thomas Bigger Jr., MD, Michael McDermott, PhD,
J. Philip Miller, AB, Thomas Moon, PhD, Arthur J. Moss MD, David Oakes, PhD,
Linda M. Rolnitzky, MS, and Terry M. Therneau, PhD
Downloaded from http://circ.ahajournals.org/ by guest on July 31, 2017
In a number of double-blinded, clinical trials in
cardiology,1-4 the occurrence of a nonfatal myocardial infarction has served as a secondary end
point and, in at least one,5 as the primary end point.
We believe that this practice may be inappropriate.
An assumption underlying the separate analysis of
nonfatal infarction is that there may be pathophysiological differences between fatal and nonfatal infarctions so that the effect of treatment on one of these
outcomes may be different from the effect of treatment on the other. However, the difference between a
fatal and a nonfatal myocardial infarction is often the
result of chance factors (e.g., having an ambulance
called immediately after the ischemic attack, being
brought to a hospital where first-rate coronary intensive care is offered), not of biological ones.
Under a competing assumption that the difference
between a fatal and a nonfatal infarction is mainly a
matter of severity, inferences about the effect of
treatment on nonfatal infarctions may be ambiguous.
For example, a reduction in the incidence of nonfatal
infarction could represent a harmful effect if the
intervention has no effect on the overall incidence of
infarctions but increases their severity so that more
infarctions are fatal. On the other hand, a reduction
in the incidence of nonfatal infarction could represent a beneficial effect if the intervention reduces the
overall incidence of infarctions without necessarily
affecting those that are fatal. These conflicting possibilities suggest that when there is research interest
in nonfatal infarctions, they should be analyzed in
tandem with fatal infarctions and not by themselves.
If a group of patients experience a reduction in the
The opinions expressed in this editorial are not necessarily those
of the editors or of the American Heart Association.
From Columbia University (J.L.F., J.T.B., L.M.R.), New York,
New York, University of Rochester (M.M., A.J.M., D.O.), Rochester, New York, Washington University (J.P.M.), St. Louis,
Missouri, University of Arizona (T.M.), Tuscon, Arizona, and
Mayo Clinic (T.M.T.), Rochester, Minnesota.
Address for correspondence: Joseph L. Fleiss, PhD, Division of
Biostatistics, Columbia University School of Public Health, 600
West 168 Street, New York, NY 10032.
incidence of nonfatal infarctions, it must be ascertained that the reduction was not at the expense of an
increase in the incidence of fatal events.
A problem exists in the statistical analysis of
nonfatal infarctions. Consider a clinical trial in which
patients who have survived a myocardial infarction
are randomized to receive one of two treatments.
Each patient is followed until study termination, loss
to follow-up, or the occurrence of the study's end
point, whichever comes first. If the end point under
consideration is a new nonfatal myocardial infarction, patients who die before experiencing it are
censored at the time of death; they are included in
the analysis as alive and at risk of experiencing the
end point until that time and are then withdrawn
from the analysis.
Available methods of survival analysis assume that
a patient who experiences a censoring event is at the
same risk of subsequently experiencing the end point
(nonfatal reinfarction) as a prognostically similar
patient who survives as long but is not censored.6
This may be an appropriate assumption for deaths
due to causes such as cancer or cirrhosis of the liver,
but it is likely to be erroneous for sudden deaths. In
one study,7 two thirds of postinfarction patients who
died within 24 hours of the onset of symptoms were
found on autopsy to have suffered a new myocardial
infarction. The assumption concerning the independence between censoring and experiencing the
study's end point is completely erroneous, finally,
when deaths due to a documented new myocardial
infarction are censored.
A practical reason for considering nonfatal infarctions alone as an end point is that information is
frequently lacking to classify a death as being due to
a definite new myocardial infarction. In contrast, a
patient whose infarct is nonfatal has the opportunity
to report the clinical symptoms and to undergo the
laboratory and electrocardiographic tests required
for diagnosis. Some deaths, however, occur late
enough after the patient's arrival in the hospital to
permit a diagnosis of definite myocardial infarction.
Fleiss et al Nonfatal Myocardial Infarction, an Inappropriate Endpoint
Downloaded from http://circ.ahajournals.org/ by guest on July 31, 2017
Rather than censor these cases, it is more sensible to
count them as occurrences of a slightly broadened
end point, definite myocardial infarction, whether
fatal or not. This suggestion is consistent with standard practice in clinical trials in neurology: nonfatal
stroke seems never to be analyzed as a separate end
point but is instead combined with fatal stroke.
Other deaths due to coronary heart disease, especially those that occur soon after the onset of ischemic symptoms but without detailed medical information, may or may not be due to a myocardial
infarction. Caution suggests that these cases be analyzed twice, once by censoring them and once by
counting them along with the definite infarctions as
occurrences of a yet broader end point, definite or
probable myocardial infarction, whether fatal or not.
We suggest that in trials aimed at testing whether
an experimental agent reduces the risk of myocardial
infarction, the primary end point be a definite infarction, fatal or not, and that a secondary end point be
a definite or probable infarction, fatal or not. An
alternative primary end point that is in wide use is the
composite event defined as a nonfatal myocardial
infarction or death due to coronary heart disease,
whichever occurred first.12'4'8-10 This end point is
referred to as "coronary incidence," "coronary heart
disease event," and "first recurrent cardiac event."
Nonfatal myocardial infarction should not be used by
itself as an end point. After all, it is the infarction that
one is seeking to prevent, not just the nonfatal one.
(Circulation 1990;81:684-685)
685
References
1. Aspirin Myocardial Infarction Study Research Group: A
randomized, controlled trial of aspirin in persons recovered
from myocardial infarction. JAm MedAssoc 1980;243:661-669
2. Beta-Blocker Heart Attack Trial Research Group: A randomized trial of propranolol in patients with acute myocardial
infarction: II. Morbidity results. JAA4 1983;250:2814-2819
3. Coronary Drug Project Research Group: The Coronary Drug
Project: Design, methods, and baseline results. Circulation
1973;47(suppl I):I-1-I-50
4. Persantine-Aspirin Reinfarction Study Research Group: Persantine and aspirin in coronary heart disease. Circulation
1980;62:449-461
5. Gibson RS, Boden WE, Theroux P, Strauss HD, Pratt CM,
Gheorghiade M, Capone RJ, Crawford MH, Schlant RC,
Kleiger RE, Young PM, Schechtman K, Perryman B, Roberts
R, Diltiazem Reinfarction Study Group: Diltiazem and reinfarction in patients with non-Q-wave myocardial infarction:
Results of a double-blind, randomized, multicenter trial. N
EnglJ Med 1986;315:423-429
6. Cox DR, Oakes D: Analysis of Survival Data. London, Chapman and Hall, 1984
7. Vedin A, Wilhelmsson C, Elmfeldt D, Save-Sdderbergh J,
Tibblin G, Wilhelmsen L: Deaths and nonfatal reinfarctions
during two years' follow-up after myocardial infarction: A
follow-up study of 440 men and women discharged alive from
hospital. Acta Med Scand 1975;198:353-364
8. Coronary Drug Project Research Group: Clofibrate and
niacin in coronary heart disease. J Am Med Assoc 1975;
231:360-381
9. Lipid Research Clinics Program: The Lipid Research Clinics
coronary primary prevention trial results: I. Reduction in
incidence of coronary heart disease. JAMA 1984;251:351-364
10. Multicenter Diltiazem Postinfarction Trial Research Group:
The effect of diltiazem on mortality and reinfarction after
myocardial infarction. N Engl J Med 1988;319:385-392
Nonfatal myocardial infarction is, by itself, an inappropriate end point in clinical trials in
cardiology.
J L Fleiss, J T Bigger, Jr, M McDermott, J P Miller, T Moon, A J Moss, D Oakes, L M
Rolnitzky and T M Therneau
Downloaded from http://circ.ahajournals.org/ by guest on July 31, 2017
Circulation. 1990;81:684-685
doi: 10.1161/01.CIR.81.2.684
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1990 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/81/2/684.citation
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/