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
Between 1985 and 1988, 21 men and one woman who were referred to the principal investigator agreed to take part in the study. This number was determined by the physician as the most manageable, given his other professional obligations. Five participants left the study within the first 2 years and six maintained the diet but did not complete the data-collection portion of the study. The results of the remaining 11 participants are reported here. These participants had completed a mean follow-up period of 5.5 years as of 1992. All participants had severe, progressive triple-vessel coronary heart disease documented by angiography. In the 8 years before the present study began, these 11 participants, who were receiving state-of-the-art cardiac care at The Cleveland Clinic Foundation, collectively experienced 37 cardiovascular events, including 15 cases of increased angina, 6 cases of angiographically determined disease progression, 6 cases of coronary artery bypass surgery (2 others had had bypass surgery more than 8 years before the study), 4 myocardial infarctions, 3 strokes, 2 angioplasty procedures, and 1 abnormal (worsening) stress test.
All participants were nondiabetic, nonhypertensive, and nonsmokers. They ranged in age from 43 to 67 years (mean age, 56 years).
Baseline angiograms werc obtained within 60 days of the start of the study for 10 of the 11 participants analyzed here. The remaining participant’s baseline angiogram was taken 1 year into the study. Every participant had a follow-up angiogram.
Among the 11 participants, the two angiographers identified 38 lesions associated with more than 20% stenosis. Of these lesions, three were excluded from analysis, two of which had been treated before and one that was treated after entry into the study with percutaneous transluminal coronary angioplasty (PTCA). Another four lesions in native vessels immediately proximal to bypass grafts were excluded from analysis a priori because accelerated disease is known to occur in this circumstance,29-32 and their inclusion would confound the analysis by introducing a new variable. As expected, these lesions progressed during our study. Of the remaining 31 lesions, 18 (5 8%) produced stenosis of greater than 50%. Of these 18 lesions, six could not be accurately assessed at follow-up because of discrepancies in the angle of projection or visual overlap, which prevented the evaluation of changes. The final analysis, therefore, involved 25 lesions in 11 patients.
MEAN PERCENT STENOSIS. The mean percent stenosis (±standard deviation) on all 25 lesions decreased from 53.4% (±14.8%) to 46.2% (±16.8%) from baseline to follow-up. The mean percent reduction in stenosis estimated by the mixed-effects model, which adjusted for correlations between lesions in the same participant, was 7% (95% CI, 3.3 to 10.7, P<.05). The average standard deviation of the three stenosis measurements was 4.4% at baseline and 4.9% at follow-up, which is consistent with the values reported by Hambrecht et al.12
According to this method of analysis, of the 25 lesions, 14 remained unchanged and 11 regressed. Thus, stenosis decreased in 8 of the 11 participants, indicating regression of the disease, and was unchanged in 3.
MINIMAL LUMEN DIAMETER. Mean minimal lumen di ameter (±standard deviation) of the 25 lesions was 1.3 mm (±0.6) at baseline and 1.4 mm (±0.6 mm) at follow- up. The mean increase in diameter as estimated by the mixed-effects model was 0.08 mm (95% CI, –0.06 to 0.22, P=NS). The average standard deviation of the three diameter measurements was 0.14 mm at baseline and 0.15 mm at follow-up, which again is consistent with the values reported by Hambrecht et al.12
Of the 25 lesions, 14 remained unchanged, 6 re gressed, and 5 progressed. Of the lesions that progressed, 4 were arterial (see below) and 1 was venous. Among the 11 participants, stenosis generally decreased in 5, remained stable in 1, progressed in 4 (although the change was not statistically significant), and both regressed and progressed in 1.
The four arterial lesions that progressed according to the MLD method occurred in four participants. In one participant, the lesion progressed 0.03 mm beyond the cutpoint, although another lesion showed marked regression. In another participant, the lesion also progressed only 0.03 mm beyond the cutpoint but was accompanied by the complete recanalization of the left circumflex artery. This recanalization was not defined as regression because this phenomenon can occur spontaneously.’2 In a third participant, one lesion progressed 0.04 mm beyond the cutpoint, whereas another remained unchanged. Finally, in a fourth participant, the lesion progressed 0.07mm beyond the cutpoint, making it the most advanced progression. Two other lesions in the same participant remained unchanged. The regression of coronary artery disease is illustrated in the Figure.
The mean number of lipid analyses performed on each participant was 126 (range 74 to 156). The mean baseline (fasting) total cholesterol level for the 11 participants was 246 mg/dL (6.36 mmol/L). The mean total cholesterol level between the onset of the medication effect and the follow-up angiogram was 132.4 mg/dL (3.42 mmol/L). The range was 109.9 mg/dL (2.84 mmol/L) to 149.9 mg/dL (3.88 mmol/L). Thus, each participant had a mean total cholesterol level under the target level of 150 mg/dL (3.88 mmol/L). Individual and average lipid values are shown in Tables 1 and 2. For all patients, the mean follow-up high-density lipoprotein (HDL) value was 36.3 mg/dL (0.94 mmol/L), and the mean low-density lipo protein (LDL) was 71.6 mg/dL (1.85 mmol/L). Other Outcomes Mean baseline and follow-up weight and blood pressure values are shown in Table 2. Angina, initially reported by nine participants, was eliminated in two and reduced in seven. In six patients the angina decreased by one grade, and in three patients it decreased by two grades, as determined by the Canadian Cardiovascular Society Scale.27
Two participants required coronary interventions during the study period. A 67-year-old man with 70% stenosis of the left anterior descending artery at baseline suffered from angina. Although the lesion and the angina were stable, the patient desired to be free of angina, so he underwent PTCA 20 months into the study. The artery restenosed 12 months later, and he underwent a second angioplasty. An angiogram 22 months later showed no restenosis. This lesion was not included in the analysis.A 64-year-old man began the study 2 weeks after an anterior myocardial infarction that occurred during an unsuccessful angioplasty. Severe left ventricular dysfunction was documented on an angiogram 7 months later. A thallium test confirmed anterior myocardial ischemia, and he underwent coronary artery bypass surgery. Four and one-half years into the study, four lesions remote from the bypass were examined angiographically. Two of the lesions were unchanged and two had regressed. The left ventricular ejection fraction was estimated to be less than 20%. The patient died of cardiac arrhythmia 10 months later, and no new occlusions or infarctions were found at autopsy. Analysis of Dropouts and Those Not Completing the 5-Year Follow-up Of the initial 22 participants, 11 left the study within the first 2 years: 3 moved from the area; 4 were employed in jobs that prevented them from making regular clinic visIts, 2 of whom returned to their prestudy diet; 3 were unable to maintain a very low-fat diet; and 1 patient who had stable angina and adhered to the diet underwent bypass surgery at another institution to eliminate his angina, which this highly competitive patient contended inhibited his ability to play tennis aggressively.Ten of these 11 patients were contacted by the primary investigator in 1995, 9 years after the study began. Five participants still adhered to the diet: two kept their serum cholesterol below 150 mg/dL (3.88 mmol/L), one to below 175 mg/dL (4.53 mmol/L), one had not had a cholesterol measurement for 7 years, and one had undergone elective coronary bypass surgery. These patients reported no new cardiac events. One patient who continued the study diet has not been heard from since 1992. All five dropouts who resumed their prestudy diet reported new cardiac events. These included four in stances of increased angina, two episodes of ventricular tachycardia, one coronary arterial bypass operation, one angioplasty, one case of congestive heart failure, and one death from complications of arrhythmia. Current Participant Status Follow-up angiogram and lipid measurements were taken in 1992. In the 3 years since, other than the one death from arrhythmia, no new coronary events have been reported among these 10 participants. The remaining six participants also continue to follow the diet and drug- monitoring program and remain free of clinical progression of disease as they complete their 9th year of study. No accidents, suicides, cancer deaths, or aggressive behavior 33 have been observed in any participants.