home   author    contact   web site dedication



Specific guide to this web site for:

 1.  Medical School
      in Statistics

 2.  Medical Students

 3.  Science media writers

 4.  High School & College
     Statistic Teachers


1. Harvard led MI study

2. JACC study 

   (J. of Amer. Coll.

3. NEJM cath study

4. Amer. J. of Cardio.
    review of literature


Oat bran study

Pregnancy & Alcohol

Are Geminis really
9. Columbia 'Miracle' Study  

Additional Topics:


Limitations of Meta-Analyses

Large Randomized Clinical Trials

Tale of Two Large

Advocate meta-analyses

Network meta-analyses





This is the initial letter written to trial investigators bringing to their attention the multiple problems with the trial protocol and suggestions for improvement. This letter was sent to the directors of the study and some other leaders in the field of cardiology.
March 7, 1990

Eugene Braunwald, M.D.
Chairman, Department of Medicine Brigham and Women's Hospital
75 Francis Street
Boston, Massachusetts 02115

Dear Dr. Braunwald,

I am writing because of my concern about the ongoing TIMI IIIB trial. As you know the TIMI IIIB trial is designed to evaluate the comprehensive management of patients with unstable angina and non-Q-wave myocardial infarction. An invasive revascularization strategy is compared to a conservative management strategy. Both treatment strategies have a protocol arm with and without TPA.

The TIMI IIIB trial appears to have potential limitations in protocol design that may negatively impact the value of the trial. results.  One problem is that many of the patients undergoing angioplasty in the invasive group will not have received adequate pretreatment with aspirin. Another potential limitation in protocol design is that the endpoints of the trial are analyzed earlier (at six weeks and one year) than may be ideal for a comparison of a revascularization strategy with a medical management strategy. Though this time frame may be adequate for evaluating the effects of TPA, it may not be sufficient for evaluating a surgical intervention. Finally, the TIMI IIIB trial data will be used to help evaluate the appropriate role of coronary angiography in patients presenting with unstable angina and non-Q-wave myocardial infarction. The protocol should include plans to report the percentage of patients undergoing cardiac catheterization in the conservative strategy at six weeks, one year, and all subsequent planned follow up endpoints,

I. Inadequate Pretreatment with Aspirin Prior to Angioplasty in TIMI IIIB Trial

There is a major potential problem with the TIMI IIIB protocol which may have a substantial negative impact on the outcome of the patients treated with the invasive strategy. The patients in the invasive strategy by protocol design will in large numbers not be pretreated with aspirin for adequate periods of time prior to undergoing angioplasty. The protocol states that aspirin will not be administered during the first 24 hours after randomization and will be started on day two at a 80mg. dose.  The protocol also calls for the patients in the invasive strategy to undergo angiography 24 to 48 hours after randomization. Percutaneous transluminal angioplasty, if indicated, will be performed at the same time as the cardiac catheterization.

This results in many patients in the invasive treatment strategy not be pretreated with aspirin for an acceptable period of time prior to undergoing angioplasty. As an example, a patient who is randomized to the invasive strategy and undergoes cardiac catheterization and percutaneous transluminal angioplasty 24 hours after randomization will not have received aspirin. At the far end of the spectrum, patients with unstable angina in the invasive strategy needing percutaneous transluminal angioplasty at the most will have received 80 mg. of aspirin 24 hours prior to the procedure. All the patients in this group fall somewhere between these two extremes.

This differs from the current standard of optimal care which is the treatment with aspirin and heparin prior to performing percutaneous transluminal angioplasty after a patient presents with unstable angina. In addition, the low dose of 80 mg. of aspirin may conceivably not be adequate given the short duration between the time it is given to the patient and the time the angioplasty is performed. Though an 80 mg. aspirin dose has positive clinical benefits in patients with unstable angina, this smaller dose of aspirin may possibly take a longer time period to produce clinically reliable antiplatelet effects needed to reduce the acute reocclusion rate occurring with angioplasty. Modification of the TIMI IIIB trial needs to be evaluated and addressed before it progresses further from its current stages of patient enrollment.

II. Premature Evaluation of "points at Six Weeks when Comparing Invasive and Noninvasive Strategies for Unstable Angina and Non-Q-wave MI in TIMI IIIB

The TIMI IIIB trial is currently designed to have as the primary endpoint an assessment at six weeks of the differences between the treatment groups for myocardial infarction, death, and exercise treadmill test response. Follow up at one year is planned as well. This may not be an adequate time frame to compare the clinical outcome of a revascularization strategy to a conservative strategy particularly in this type of patient.

Unstable angina is associated with a high frequency of subsequent cardiac events for the first six to twelve months after the initial presentation. Dr. Robert Califf presented the Duke experience (ACCEL December 1989) which indicated that the risk of dying from unstable angina did not return to the baseline of a patient with stable angina for six to twelve months. The VA Cooperative Study of Unstable Angina documented a high frequency of crossover from medical management to surgical therapy during the first sixteen months following the initial randomization for unstable angina.

Non-Q-wave myocardial infarctions have a significant subsequent event rate occurring the first two years after the initial event. Though the initial mortality is lower than a Q-wave myocardial infarction, the overall mortality and morbidity for a non-Q-wave myocardial infarction becomes similar to a Q-wave infarction two years after the initial infarction because of a higher rate of subsequent cardiac events. Hence, both unstable angina and non-Q-wave myocardial infarction continue to have a relatively high frequency of events after presentation that persists beyond six weeks and extends up to one year if not two years. The TIMI IIIB trial should have as its primary endpoint a comparison of cardiac events at one year with subsequent follow up at two years and three years. The study should continue follow up through the time period when a high rate of subsequent cardiac events and crossovers from medical to surgical treatment are expected.

Comparing endpoints at six weeks clearly has potential problems particularly when comparing a medical treatment strategy to a revascularization strategy. Both coronary artery bypass surgery and angioplasty have many of their untoward clinical events occurring early in tithe course of treatment as perioperative myocardial infarctions and fatalities. The differences in outcome between medical treatment and surgical treatment can take several years to develop.

The CASS study as an example showed a significant reduction in the mortality of patients undergoing surgery who had decreased left ventricular function and triple vessel disease. This difference in outcome was not present at one year, but developed in later subsequent follow up. If longer follow up of the treatment groups had not been continued, this important difference in treatment outcomes would not have been substantiated. Though trials such as the VA Cooperative Study of Unstable Angina in patients with abnormal left ventricular function have shown a difference between treatment groups developing by one year , this does not provide a guarantee that significant differences will be manifest within the initial first year. Analysis at six weeks or even at one year of a trial comparing medical versus surgical therapy may miss important differences in outcome.

The publication of the TIMI II trial data at one year does not provide reassuring information that a one year time period of follow up is sufficient. One week after randomization, the invasive group had a statistically significant higher incidence of death and myocardial infarction than the noninvasive group (8.9% cardiac event rate invasive group vs. 6.3% noninvasive group, p !5.05). However, after the initial first week of randomization and treatment there has been a contrasting trend of lower cardiac events in the invasive strategy group. Because of this trend, there was a cumulative lower incidence of cardiac events in the invasively treated group after one year of follow up (14.5% cardiac event rate invasive group vs. 15.1% noninvasive group, NS).

The subsequent lower cardiac event trend for the invasively treated group persisted not only in the one week to six week follow up time period, but continued as well in the six weeks to one year follow up time frame. Does this trend of lower cardiac events in the invasive group continue after one year?  If such a trend did continue, it is conceivable that a statistically significant difference may develop between the groups. There is no way of knowing if this trend persists unless the TIMI II trial publishes results at two years and three years of follow up. Judging from prior trials comparing surgical versus medical therapy, differences may be seen after the first year of randomization. Particularly with a sustained apparent difference in cardiac event rates between treatment groups, it would seem reasonable to continue follow up analysis past one year. Hence, the TIMI II trial does not provide convincing evidence that six weeks or even one year of follow up is definitely sufficient for assessing a revascularization versus a medical treatment strategy even in the setting of a Q-wave myocardial infarction (which has many of its subsequent cardiac events occurring relatively early after initial presentation).


III. Reporting the Percentage of Patients Undergoing Cardiac Catheterization in the Conservative Strategy Group

 Attempts to make recommendations for the appropriate use of cardiac catheterization in patients following thrombolytic therapy are being formulated on the basis of the TIMI II trial. Similarly, the results of the TIMI IIIB study are almost certain to be used in future discussions concerning the role of cardiac angiography in the setting of unstable angina and non-Q-wave myocardial infarction. A more complete reporting than was given in the TIMI II study of the percentage of patients undergoing cardiac catheterization in the conservative treatment strategy group will benefit the TIMI IIIB trial.

Though the TIMI II results have played prominently in discussions concerning the appropriate role of coronary angiography after thrombolytic therapy, important information regarding this issue from the TIMI II trial has not yet been released. What percentage of the patients assigned to the initial conservative management approach underwent cardiac catheterization? The only published data at this point indicates that 32.7% underwent cardiac catheterization after two weeks of randomization. What percentage of these patients underwent cardiac catheterization after six weeks of therapy and after completion of one year of follow up? If the number requiring cardiac catheterization in the conservatively treated group by one year is high in the 50 to 75% range, this has different implications for the role of coronary arteriography in this study than if only a relatively limited number of patients required cardiac catheterization by the end of twelve months.

There is another important reason for making known the data on the percentage of patients undergoing cardiac catheterization at six weeks, twelve months, and all subsequent follow up endpoints. When a study is analyzed to assess the impact of cardiac catheterization on morbidity and mortality, what is actually being evaluated is a particular invasive revascularization treatment strategy compared to a particular noninvasive strategy. The particular conservative patient management strategy used in the TIMI II trial is not fully described until the percentage of patients who underwent coronary angiography is known. The percentage of patients undergoing cardiac catheterization helps define the threshold for proceeding with angiography used to obtain the excellent morbidity and mortality results that were achieved in the TIMI II study. The TIMI IIIB trial would be similarly benefited by reporting the percentage of patients undergoing cardiac angiography at all planned follow up time intervals.


The TIMI trials have yielded important information regarding thrombolytic therapy. The TIMI IIIB trial addresses the comprehensive management of unstable angina and non-Q-wave myocardial infarction. The results of this trial will be used to help guide patient management for this group of patients for years to come.

Funding allotment for medical research and clinical trials appears to be diminishing at a time when controlled clinical trials will increasingly direct the way medicine in practiced in this country. It becomes even more imperative that the trials which are undertaken address the questions they attempt to answer in an optimum fashion, even if this requires modification of a trial already in progress.

Obviously, there will be more expense to the TIMI IIIB trial if the primary cardiac endpoints are changed to one year and the patient outcomes are followed for at least three years. However, the incremental cost to follow and report the endpoints for a longer duration would be relatively small compared to the total initial outlay of our national research funds being used for this study. I know the time demands on basic and clinical researchers are heavy, so please forgive the length of this letter. Thank you for your patience and consideration in this matter.

Sincerely yours,

Eric F. Roehm, M.D.,

David O. Williams, M.D.
Robert Roberts, M.D.
James H. Chesebro, M.D.
Eugene R. Passamani, M.D.
Bernard R. Chaitman, M.D.
Harold T. Dodge, M.D..
Genell L. Knatterud, Ph.D.
James Forrester, M.D.
Shahbudin H. Rahimtoola, M.D.
Robert A. O'Rourke, M.D.
Sylvan Lee Weinberg, M.D.
Eric J. Topol, M.D.
William O'Neill, M.D.


1. TIMI IIIB protocol. Dec 7,1989.

2. Califf, Robert. Unstable angina long-term results. AHA 62nd Scientific Session: Plenary Session VIII, Nov 15, 1989.

3. Mulcahy, et al. Natural history and prognosis of unstable angina. Am Heart J 1985; 109:753-8. 

4. Wallentin, et al. Aspirin 75 mg. after an episode of unstable coronary artery disease; effects on angina and need for revascularization. Circulation 1989; 80 (Suppl II):II-419.5. Scott, et al. Veterans Administration Cooperative Study for treatment of patients with unstable angina; results in patients with abnormal left ventricular function. Circulation 1988; 78 (Suppl I): 1-113-1-121.

6. Williams, et al. The TIMI Trial; outcome at one year of patients randomized to either invasive or conservative management. Circulation 1989; 80 (Suppl II);II-519.

7. Passamani, et al. A randomized trial of coronary artery bypass surgery: survival of patients with a low ejection fraction. N Eng J Med 1985; 312:1665-71.

8. TIMI Study Group. Comparison of invasive and conservative strategies after treatment with intravenous TPA in acute myocardial infarction: results of the TIMI Phase II Trial. N Eng J Med 1989; 320:618-27.