Sign up for free updates from Dr. Esselstyn
Home / Methods


Abstract | Methods | Results | Discussion | Conclusions | References

A Strategy to Arrest and Reverse Coronary Artery Disease: A 5-Year Longitudinal Study of a Single Physician’s Practice
Caldwell B. Esselstyn, Jr, MD; Stephen G. Ellis, MD; Sharon V. Medendorp, MPH; and Timothy D. Crowe


Several patients who were not considered to have immediately life-threatening coronary artery disease or who had declined or were not considered candidates for intervention by their cardiologists were referred to the principal investigator, a general surgeon, by colleagues in the De partment of Cardiology, The Cleveland Clinic Foundation. Some patients had heard of the study and asked to be involved. The only inclusion criteria were angiographically documented evidence of severe coronary heart disease, and the willingness to attempt the dietary changes described below and to be examined periodically so that progress could be monitored. All participants were informed about the purpose and nature of the study and consented to the follow-up activities. The study was conducted with the approval of The Cleveland Clinic Foundation’s Institutional Review Board.

Therapeutic Interventions

Participants were asked to adhere to a diet that derived less than 10% of its calories from fat. They were to avoid oils, meat, fish, fowl, and dairy products, except for skim milk and nonfat yogurt. Grains, legumes, lentils, vegeta bles, and fruit comprised the major portion of the diet, which contained adequate amounts of vitamins, minerals, protein, and iron.’6~’7 A list of fat-free recipes taken from cookbooks and other resources that focus on weight loss, nutrition, and lifestyle changes to improve heart health18-26 was provided to each participant. Daily food diaries were kept and reviewed with the physician during biweekly visits. Participants were also encouraged to take a daily multivitamin and were asked, but not required, to moderate their consumption of alcohol and caffeine. There were no prescribed exercise requirements.

Each participant also received an individualized prescription for a cholesterol-lowering drug. The most frequent regimen included cholestyramine, 4 g twice daily, and lovastarin, 40 mg to 60 mg daily. Time-release niacin was prescribed for a short while but was discontinued when many patients reported nausea, vomiting, and swollen ankles.

Participants also received several hours of instruction in relaxation and meditation techniques through the Cleveland Clinic’s Department of Psychology. However, all participants discontinued these techniques after only a few weeks of sporadic use, so this part of the intervention was abandoned.

Measurements of Outcome Variables

The degree of coronary artery disease was determined from angiograms using both the percent stenosis and minimal lumen diameter (MLD) methods.5’12 These methods were applied independently by two senior an giography laboratory technicians with angiographic evaluation experience in national multicenrer cardiac trials. The technicians were masked as to whether the angio grams were from the initial or follow-up catheterization. Data were obtained for the initial angiogram, which helped identify patients who would be appropriate to participate in the study, and for a minimum of one follow-up angiogram. Most follow-up angiograms were taken in 1992.

One technician evaluated all lesions from both the baseline and the follow-up angiograms on two separate occasions, the second time at least 6 months after the first to prevent recall bias. The second technician independently evaluated both the baseline and the follow-up angiograms. Thus, each lesion was measured three times. The mean of the three measurements for each lesion on each angiogram was used in the statistical analysis, and the standard deviation of each mean was used to check for consistency with published guidelines for lesion progression or regression. The percent stenosis and minimal lumen diameter were measured for each lesion identified on the angiogram as described above. Disease status was defined by the criteria of Hambrecht et al ‘2 for each method. For the percent stenosis method, an increase in stenosis of 10% or more was considered to indicate disease progression; a decrease in stenosis of 10% or more was considered to indicate disease regression; and a change of less than 10% in either direction was considered to indicate no change in disease status. For the MLD method, a decrease of more than 0.18 mm was considered to indicate disease progression; an increase of more than 0.18 mm was considered to indicate disease regression; and a change of less than 0.18 mm in either direction was considered to indicate no change in disease status.

Baseline angiograms were reviewed before the follow-up angiograms were made so that the projection angles, cardiac segments, and lesions could be matched with those present at baseline. Previously dilated lesions (n=3) were not studied. Fasting serum lipid levels were analyzed biweekly during the first 5 years of the study and monthly thereafter. Weight and blood pressure were also recorded biweekly. Angina was graded on the Canadian Cardiovas cular Society Scale.27 During the second year of the study (1987), participants’ food diaries were analyzed with the 3-Day Nutrient Intake computer program28 to determine the fat content of the diet.

Promoting Adherence to the Diet

The strategy used to foster adherence to the diet had four components. First, the physician conducted a 1-hour in rerview with the participant and his or her spouse and reviewed the history, epidemiology, angiographic data, research findings, and therapeutic options of coronary heart disease. The purpose of the interview was to explain the rationale for the study, to impress upon participants the importance of the diet and of keeping a food diary, and to establish informed consent with each participant.

Second, the physician saw each participant every 2 weeks for 5 years and every month thereafter for an additional 5 years. At each visit, the participant’s food diary was discussed, blood pressure and weight were recorded, and blood samples were drawn so that lipid profiles could be determined.

Third, during the first year of the study, the physician called each participant on the evening after his or her clinic visit (every 2 weeks) to report the results of the blood test and discuss any adjustments to the diet or medication.

Fourth, three or four times each year, participants met at the home of the physician or one of the participants to discuss the low-fat diet program, trade recipes, and socialize. The physician was personally committed to adhering to a low-fat diet, as well as to supporting patients in their efforts to adopt the diet.

The costs of the study were limited to blood tests and a follow-up angiogram, which were paid for by each participant (or the insurance carrier) as part of regular medical care. During the first 3 years of the study, the costs of the lipid profiles were borne by the Cleveland Clinic.

Statistical Analysis

Disease progression was analyzed with each method (per cent stenosis and MLD) by averaging the three measurements taken on each lesion of each angiogram and comparing the averages of the initial and follow-up angiograms. Changes in lesion characteristics were examined using mixed-effects models, which take into account that most participants had multiple lesions and that some data were therefore correlated. Results are reported by average degree of arterial patency, by change in individual lesions, and by overall change in disease status for each patient.

Three reference points were chosen for analyzing cholesterol data: baseline (the value measured just be fore or just after beginning the study), onset of medi cation effect (3 weeks after participants began taking the cholesterol-lowering drugs, which were prescribed soon after entering the study), and follow-up (the date of the most recent angiogram). Cholesterol data for each visit between onset of medication effect and the follow-up angiogram were averaged to create a single cholesterol value for each participant for the entire study period.

Analyses were performed with the SAS software. Alpha was set at .05, and all tests were two-tailed.

Abstract | Methods | Results | Discussion | Conclusions | References

Follow Dr. Esselstyn on :