Clinical Alert: Bypass Over Angioplasty for Patients with Diabetes

The National Heart, Lung, and Blood Institute (NHLBI), part of the National Institutes of Health, today announced that, as a first revascularization procedure, coronary artery bypass graft (CABG) surgery has been shown to have a markedly lower 5-year death rate than angioplasty for persons with diabetes mellitus (Type I or II) who are on oral hypoglycemic agents or insulin.

The finding came from the NHLBI-funded Bypass Angioplasty Revascularization Investigation (BARI). The multicenter, international, randomized trial studied patients who needed a first revascularization because of severe ischemia with obstructions in two or more major coronary arteries. Over 5 years, patients with diabetes mellitus who were on drug therapy had a significantly lower (p=0.002) mortality rate with CABG, compared with percutaneous transluminal coronary angioplasty (PTCA). The 5-year CABG mortality rate was 19 percent, compared with 35 percent for PTCA. By contrast, in patients without diabetes and in those with diabetes but not on drug treatment, the 5-year mortality rates for CABG and PTCA were both about 9 percent.

The higher death rate for PTCA was not due to complications of the procedure itself. Those with diabetes are known to have an excessive cardiovascular risk and a higher mortality rate was expected for them, regardless of revascularization procedure. However, the excess mortality with PTCA had not been anticipated.

BARI's results indicate that CABG should be the preferred treatment for patients with diabetes on drug or insulin therapy who have multivessel coronary artery disease and need a first coronary revascularization. These results have a significant impact on the clinical care of these patients.

Coronary revascularization plays an important role in the treatment of clinically severe coronary artery disease. The two most commonly used methods of revascularization are CABG and PTCA. PTCA is a catheter-based nonsurgical approach that directly targets coronary obstructions by dilation of the vessel at the point of obstruction. The process is accompanied by local vascular injury and subsequent healing. The extent of injury and the healing process may be different in diabetic and nondiabetic patients. Not all lesions can be dilated, due largely to technical reasons. CABG is a major operation, requiring opening of the chest. It provides a new channel, with a lumen frequently larger than the native, diseased lumen. There is no instrumentation of the local lesion and, therefore, no related vascular injury. While both treatments alleviate the effects of coronary artery disease, they do not correct or alter the natural course of the disease.


Before PTCA, CABG was the traditional revascularization strategy. But PTCA, first performed in the United States in 1977, has grown rapidly in use. In 1993, about 362,000 PTCAs were done in the United States, compared with about 309,000 CABG surgeries.

However, PTCA use has expanded not just in number but also in the type of patient treated. Initially, PTCA was done on patients with one obstructed vessel. But, with increased physician expertise and an improved technology, PTCA use has rapidly expanded to include patients with more complex, multivessel coronary obstructions, once treated exclusively with CABG. This has led to uncertainties about the long-term effectiveness and safety of PTCA compared with CABG and prompted the NHLBI to fund a rigorous evaluation of the two methods.

That rigorous investigation -- the BARI study -- tests whether the use of PTCA as an initial treatment compromised the clinical outcome for patients with multivessel coronary artery disease who needed revascularization and could be treated by either PTCA or CABG. BARI did not test outcomes for repeat procedures. It also studied only PTCAs performed with the standard balloon technique.

The trial is the largest randomized study of its type, with enough patients to be able to address key endpoints, both overall and in predetermined patient subgroups. The subgroups were based on patients' anginal status, number of diseased vessels, and left ventricular function. Also studied are various demographic factors such as gender, age, race, and the presence of diabetes.

The primary endpoint is mortality after 5 years of follow-up. Other important endpoints include the occurrence of a myocardial infarction, need of repeat revascularization procedures, angina, functional status, quality of life, and utilization of health-care resources. (Both quality of life and utilization of health-care resources are studied in detail in a separately funded ancillary study.)

Patients were eligible for the trial if they had coronary artery disease with a 50 percent or more luminal obstruction (as measured by calipers) in at least two of the coronary vessels supplying two or three major coronary territories. They also had to have clinically severe ischemia but not a prior revascularization. Patients were ineligible for the trial if they had, for example, insufficient angina or ischemia, required emergency revascularization, left main stenosis of 50 percent or greater, a noncardiac illness expected to result in limited survival, primary coronary spasm, or a poor quality angiogram. Baseline angiograms were reviewed and classified by the Central Radiographic Laboratory (Dr. Edwin L. Alderman, director) at the Stanford University Medical Center in Palo Alto, CA.


Between August 1988 and August 1991, 18 clinical centers randomized 1,829 patients, ages 17 to 80 and including 353 on drug treatment for diabetes. Half of the patients were randomly assigned to PTCA and the other half to CABG. At baseline, the mean age of the randomized patients was 61 years. Thirty-nine percent of the randomized patients were age 65 or older, 27 percent were women, 25 percent were classified as having diabetes (of these, 76 percent were being treated with oral hypoglycemic agents or insulin), 60 percent had two-vessel disease and 40 percent had three-vessel disease, and 98 percent had angina (of these, 64 percent had unstable angina and 17 percent had class 3-4 angina). At the time of this alert, 66 percent of patients had completed follow-up. Patients will be followed for a minimum of 7 years. The trial is expected to finish follow-up on all patients by November 1998.

Follow-up includes annual functional status assessments and an electrocardiogram (ECG), and a biennial exercise stress test. As required by the protocol, risk factor modification was initiated for all patients after their enrollment. This includes help with smoking cessation, exercise, and diet. Patients also were treated as needed for hypertension, elevated blood cholesterol, and diabetes. Risk factors and medical problems were managed by each patient's primary care physician.

The trial has been closely monitored by both the study chairman (Dr. Robert Frye, Mayo Clinic Foundation), the Clinical Coordinating Center (Dr. Katherine Detre, University of Pittsburgh), and the independent Data and Safety Monitoring Board (chaired by Dr. J. David Bristow, Oregon Health Sciences University). The Data and Safety Monitoring Board is composed of PTCA experts, cardiovascular surgeons, clinical cardiologists, and experts in biostatistics and ethics. The Board regularly reviews the monitoring reports. ECG analyses are being done by the Central ECG Laboratory (Dr. Bernard R. Chaitman, director) at the St. Louis University Medical Center. An independent Mortality and Morbidity Classification Committee (chaired by Dr. Ronald Prineas, University of Miami) categorizes fatal events in the trial.

On September 13, 1995, the Data and Safety Monitoring Board held an urgent session to review the 5-year mortality data. The Board concluded that the unfavorable mortality results for the patients on drug treatment for diabetes and first treated with PTCA were unlikely to be due to chance. The Board recommended to the National Institutes of Health that physicians and other health care professionals and the public be promptly informed of the results.


In summary:

BARI's findings should not be applied to all persons with diabetes. They apply only to those on oral hypoglycemic agents or insulin for diabetes and who have multivessel coronary artery disease and are undergoing an initial revascularization procedure. The data offer the following guidelines for such patients: They will probably fare better with CABG than PTCA as an initial treatment. For patients who have already had a PTCA and are asymptomatic -- experiencing no ischemia, angina, or other symptom -- they should take no special action but continue their regular care. Alternatively, if they have already had a PTCA and had their ischemia return (e.g., reappearance of angina), they should consult their physicians to assess their current health status and review optimal strategies for further care. Close physician monitoring is particularly important for patients with diabetes who have coronary artery disease, since they may not experience symptoms during periods of ischemia.

Finally, all patients who have evidence of coronary artery disease, with or without a prior PTCA or CABG, should receive an aggressive approach to medical management of known risk factors for coronary artery disease, including smoking cessation for smokers, appropriate control of elevated blood pressure or serum cholesterol, and optimal control of diabetes.

The full text of this alert has been mailed to all libraries that are members of the National Network of Libraries of Medicine.

WebMD Public Information from the U.S. National Institutes of Health


"The National Heart, Lung, and Blood Institute of The National Institutes of Health. Press Release: Clinical Alert: Bypass Over Angioplasty for Patients with Diabetes. September 21, 1995. Last updated March 15, 1996. (Online)"