Pre-entry characteristics of participants in the Lipid Research Clinics' Coronary Primary Prevention Trial.
JOURNAL OF CHRONIC DISEASES 1983;
36:467-79. [PMID:
6863470 DOI:
10.1016/0021-9681(83)90138-8]
[Citation(s) in RCA: 27] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 01/22/2023]
Abstract
The Coronary Primary Prevention Trial (CPPT), a double-blind, randomized clinical trial being conducted by 12 Lipid Research Clinics in North America, was initiated in 1973 to test whether long-term reduction of plasma total cholesterol in individuals with Type II hyperlipoproteinemia would reduce the incidence of coronary heart disease. The trial is scheduled to conclude in 1983. Here we document that a major requirement of the CPPT--the comparability of the control and experimental groups prior to cholesterol-lowering treatment--has been achieved. The 3810 men participating in the CPPT were allocated to either treatment group at each clinic on the basis of low density lipoprotein cholesterol, S-T segment response to graded exercise testing, and the logistic risk function of age, smoking and diastolic blood pressure. Randomization was performed separately within each of the 96 cells thereby defined. An extensive battery of tests and questionnaires also yielded other laboratory measurements, data on health habits and family history, and sociodemographic information. Mean plasma total cholesterol levels were 291.8 mg/dl for the placebo group, and 291.5 mg/dl for the cholestyramine group. The other lipid and lipoprotein cholesterol levels were very similar in the two treatment groups. Among the 83 variables examined, only five differed significantly (P less than 0.05)--height, weight, albumin, 2-hr-post-challenge glucose, and serum aspartate amino-transferase (AST, SGOT)--and these differences were deemed to be so small as to have little or no clinical importance. Inasmuch as chance alone could produce four statistically different variables, we concluded that it seemed highly likely that the observed differences were attributable to random fluctuation. Thus, we have determined that the randomization process has resulted in two very nearly identical groups.
Collapse