The Lifestyle Heart Trial: five years, one combined intervention
The five-year follow-up tested an intensive lifestyle package, not a diet in isolation.

Illustrative food image, not a photograph of the study intervention.
What was actually randomized
The five-year Lifestyle Heart Trial report involved 48 people with moderate to severe coronary heart disease. They were randomized to an intensive lifestyle program or usual care using a randomized invitational design. The intervention combined a whole-food vegetarian diet providing 10% of energy from fat with aerobic exercise, stress-management training, smoking cessation and group psychosocial support. That combination is central to interpreting the findings. The experiment did not compare an isolated food with a placebo, and it cannot tell us how much of the observed change came from diet rather than the other components of the program.
Ornish D, Scherwitz LW, Billings JH, et al.. JAMA 280(23):2001–2007, 1998. PMID 9863851 · DOI
Reading the artery measurements
The main anatomical measure was coronary artery percent diameter stenosis, assessed with quantitative coronary arteriography. In the intervention group, average stenosis decreased by 1.75 absolute percentage points after one year and by 3.1 points after five years. The authors described these as relative improvements of 4.5% and 7.9%, respectively. In the usual-care group, stenosis increased by 2.3 points after one year and 11.8 points after five years. Absolute and relative changes are different descriptions of the same measurements; neither should be presented as the percentage of all arterial plaque that disappeared. This was an imaging endpoint in a small trial.
Ornish D, Scherwitz LW, Billings JH, et al.. JAMA 280(23):2001–2007, 1998. PMID 9863851 · DOI
Events were not unique patients
During five years of follow-up, the investigators recorded 25 cardiac events among 28 intervention-group participants and 45 events among 20 control-group participants. The reported risk ratio for any event in the control group was 2.47, with a 95% confidence interval of 1.48 to 4.20. Those event totals must not be rewritten as numbers of people having heart attacks: the abstract reports cardiac events, and the totals exceed the numbers of participants. The results favor the comprehensive intervention, but the distinction between events and individual patients matters when describing how large the benefit appeared in this particular study.
Ornish D, Scherwitz LW, Billings JH, et al.. JAMA 280(23):2001–2007, 1998. PMID 9863851 · DOI
The follow-up boundary
Only 35 of the original 48 participants completed the five-year follow-up coronary arteriography: 20 of 28 in the intervention group and 15 of 20 in the control group. The report therefore combines an unusually long lifestyle intervention with a small imaging sample and incomplete follow-up. Its stated objective concerned intensive lifestyle changes without lipid-lowering drugs; that describes the study question, not an instruction to change a prescription. The supported conclusion is that this combined program produced greater regression of measured coronary narrowing than usual care among the studied patients. It does not establish that any single vegan meal reproduces those results.
Ornish D, Scherwitz LW, Billings JH, et al.. JAMA 280(23):2001–2007, 1998. PMID 9863851 · DOI