Correspondence

More on the Survival and Ventricular Enlargement Trial

N Engl J Med 1993; 329:1204-1206October 14, 1993

Article

To the Editor:

In describing the results of a placebo-controlled study of captopril in patients with left ventricular dysfunction after myocardial infarction, Pfeffer et al. (Sept. 3, 1992, issue)1 reported that significantly fewer patients given captopril had at least one recurrent infarction (170 patients in the placebo group and 133 in the captopril group). This result is not as clear-cut as the report suggests. The favorable finding arose from a post hoc reanalysis of the data.

The protocol of the Survival and Ventricular Enlargement (SAVE) trial precisely defined the criteria for recurrent myocardial infarction: 1) new Q or QS waves in contiguous electrocardiographic leads, with or without clinical symptoms; 2) typical symptoms of acute myocardial infarction, with an increase in the creatine kinase level to 1.5 times the upper limit of normal and elevation of the MB isoform value; 3) a typical history of acute myocardial infarction, with an increase in the creatine kinase level to twice the upper limit of normal and either new Q waves or T-wave changes consistent with non-Q-wave infarction; and 4) a classification of death due to acute myocardial infarction by an end-point review committee. For the first three criteria to be met, the events had to be confirmed by an electrocardiography core laboratory that reviewed cases in blinded fashion. According to the protocol definition, 136 patients given captopril and 146 given placebo had recurrent myocardial infarction (P = 0.39). After reviewing the data, however, the SAVE investigators modified the protocol definition in two ways and described the results as “clinically reported” recurrent myocardial infarction. First, they excluded 28 patients given captopril and 17 given placebo who fulfilled the criteria for new Q or QS waves but had no symptoms -- i.e., patients with clinically silent myocardial infarction. Second, they added 25 patients given captopril and 41 given placebo whom the investigators considered to have had recurrent myocardial infarction but who did not meet any one of the criteria for recurrent myocardial infarction according to the core laboratory.

Any of the definitions of recurrent myocardial infarction might be considered reasonable, but the post hoc development of a new criterion raises the possibility that the choice was not unbiased. The protocol definition specifically allowed silent myocardial infarctions to be detected. No explanation for subsequently excluding them was offered. The decisions to include investigator-identified infarctions that did not fulfill the protocol-specified criteria is not necessarily unreasonable, but it is clearly at odds with the protocol's attempt to define recurrent infarction rigorously. Although the SAVE trial identified several important benefits of post-infarction administration of captopril, we believe that it did not demonstrate that captopril reduces the risk of recurrent myocardial infarction.

Charles J. Ganley, M.D.
H.M. James Hung, Ph.D.
Robert Temple, M.D.
Food and Drug Administration, Rockville, MD 20857

1 References

1. 1

   Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction -- results of the Survival and Ventricular Enlargement Trial. N Engl J Med 1992;327:669-677
   Full Text | Web of Science | Medline

Author/Editor Response

The authors reply:

To the Editor: Recurrent myocardial infarction was one of the important prospectively defined secondary end points in the SAVE protocol prepared in 1986. Three operational definitions of myocardial infarction were developed before unblinding: 1) clinical myocardial infarction: the diagnosis of a myocardial infarction as reported by center investigators on a specific form, or a death attributed to a myocardial infarction identified by the mortality committee; 2) clinical myocardial infarction, confirmed: cases of nonfatal clinical myocardial infarctions were evaluated (with blinding to treatment assignment) for the prespecified enzyme levels described in the study; and 3) clinical myocardial infarction, confirmed, or isolated new Q waves: in the absence of notification of a myocardial infarction by a study center, a patient with a new Q wave on the annual electrocardiogram was added to the total of patients with confirmed clinical myocardial infarctions.

As anticipated, regardless of their treatment assignment, patients who had a clinical myocardial infarction, whether or not confirmed (categories 1 and 2, above), were at a much greater risk for subsequent cardiovascular death, development of heart failure, and survival with a decrease in left ventricular ejection fraction by 9 units or more (Table 1Table 1Relative Risk of End Points in the SAVE Study, According to Classification of Recurrent Myocardial Infarction.). In sharp contrast to the patients with these clinical events reported from the centers, patients identified because of isolated Q-wave changes were not at increased risk for any of these cardiovascular events as compared with other patients in the SAVE population who did not have any indicators of a myocardial infarction (Table 1). Although we thought that the addition of information from review of annual electrocardiograms would result in the broadest definition of myocardial infarction, from the above analysis conducted after unblinding but without regard to therapy, it was obvious that in our post-myocardial infarction population with abnormal base-line electrocardiograms the annual readings -- in isolation from reports from the centers -- did not add clinically meaningful information.

In our article we therefore chose to describe the effects of captopril in relation to the end points of clinical myocardial infarction and confirmed clinical myocardial infarction, both of which are strong predictors of adverse outcomes (death, heart failure, and a major reduction in the ejection fraction) and both of which were reduced in incidence by captopril (patients with clinical myocardial infarction, 170 given placebo vs. 133 given captopril -- reduction in risk, 25 percent; P = 0.015; patients with confirmed myocardial infarction, 129 given placebo vs. 108 given captopril -- reduction in risk, 19 percent; P = 0.102). We stand behind this decision and emphasize that these clinically important definitions were not developed after reviewing the data, but rather were formed very early in the conduct of the trial, and indeed were the only data on myocardial infarction reported to the Data and Safety Monitoring Board during the trial. We believe that our conclusion that there was definitive improvement in survival with captopril therapy in the SAVE study is strongly supported by the reductions in the number of relevant clinical coronary ischemic events.

Marc A. Pfeffer, M.D., Ph.D.
Brigham and Women's Hospital, Boston, MA 02115

Lemuel A. Moye, M.D., Ph.D.
University of Texas Health Science Center at Houston, Houston, TX 77004

John Rutherford, M.D.
University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75235

Eugene Braunwald, M.D.
Brigham and Women's Hospital, Boston, MA 02115

Citing Articles (2)

Citing Articles

1. 1

   James B. Young. (1995) Reduction of ischemic events with angiotensin-converting enzyme inhibitors: Lessons and controversy emerging from recent clinical trials. Cardiovascular Drugs and Therapy 9:1, 89-102
   CrossRef

2. 2

   Philip A. Poole Wilson. (1994) When to start an ACE inhibitor and in whom. Cardiovascular Drugs and Therapy 8:1, 111-114
   CrossRef