University of Minnesota

Cooperative Study of Lipoproteins and Atherosclerosis

Study Category: The Cohort Studies (1947-1972)
Year Begun: 1952
Location: United States of America
Principal Investigator(s): Cooperative Lipoprotein Study Group


The seminal publication from this early cooperative study of CVD epidemiology was in Circulation in 1956. It begins: “In 1950 the attention of the National Advisory Heart Council was called to studies by Gofman and his colleagues at the Donner Laboratory of the University of California,” a strange way to describe the initiation of an NIH-supported study. An equally unusual footnote reads: “Dr C.J. Van Slyke, who was then director of the National Heart Institute, was active in the initiation of the Cooperative Study and the organization of this committee.”

Much of this dramatic story can be found in the essays and oral history contents of this history website and in the companion book, but all this was unusual, even for that time, when initiation of major NIH studies had fewer circumlocutions than now. In essence, “big wheels” turned other “big wheels,” and the ambitious study soon got under way in four U.S. laboratories. Ambiguous beginnings begat, however, an ambiguous ending.


John Gofman had identified separate lipoprotein (LP) fractions using the same sorts of instruments, preparatory and analytical ultracentrifuges, he had used in isolating radioactive isotopes during World War II. This opened a major new field of research on the structure and function of the protein fractions that carry lipids in soluble form in the blood.

Four laboratories were delegated to recruit 15,000 subjects from whom 4,914 men, aged 40-59, found “clinically normal,” were enlisted for prospective study of the predictive value of the newly defined LP fractions. Extensive organization was required for doing the centrifugal analyses at each center, for their quality control, and for central analysis of the data.

These were the several larger goals of the study:

  • Determine the range of Sf 12-20 LP in “normal men and women.”
  • Study associations of LP with incidence of coronary events
  • Measure effects on LP of a low-fat, low cholesterol diet
  • Compare the relative discrimination of cases by LP fractions versus total serum cholesterol level
  • Study the distribution of LP in hypertensives and diabetics.


This study was one of the first to explore these questions. The 1956 report dealt principally with characteristics of the 82 validated coronary cases compared to the remainder of the group. The Report provided an unprecedented majority and minority statement of the investigators. The group agreed that there was predictive value in the lipid measures. It diverged in interpretation.


Discussion (A) by the Gofman-Donner Group gave results as an Atherogenic Index representing the significant relation of the LP densities to cardiac events. It concluded that the Sf 12-20 fraction discriminated new coronary events better than total cholesterol level: “This Cooperative Study established clearly for the first time that elevation of blood lipids preceded clinical coronary disease and predicts it, rather than being a metabolic result of coronary disease.”

Discussion (B) by all the other clinics and the chief statistician concluded that: “The present study does not confirm the hypothesis that lipid levels can successfully be used to predict those individuals who will develop coronary heart disease.” These discussants go on to say that this does not preclude the atherogenic properties of lipid fractions or their use in characterizing population differences and mass causes, but that the study of individual prediction might better be carried out in more diverse populations or in youth.

“The use of Sf 12-20 and Sf 20-100 LP measures, or the related Atherogenic Index (A.I.), had no advantage over the simpler measurement of cholesterol in the characterization of men prone to develop coronary heart disease.” Finally, “The LP measurements are so complex that it cannot be reasonably expected that they could be done reliably in hospital laboratories.”

Thus, the larger group did not accept the statistical manipulations of the Gofman Atherogenic Index. More salient, however, was the negative reaction of the larger group to Gofman’s introducing new LP measures midway in the study and cumulative “attitudes” that soured relations among the investigators. The sequelae of this unhappy conclusion were prolonged and painful. John Gofman’s findings and concepts, concerned as much with function of the LP as with prediction, were fully justified by subsequent epidemiological analyses from more powerful cohort studies, particularly the Framingham Study, from which the total cholesterol/HDL ratio is now standard procedure for preventive cardiology. Moreover, Gofman’s LP fractionation paved the way for elaboration of the hepatic receptors to LP and other strides in understanding lipid metabolism and atherogenesis.

Following this official rejection of his conclusions, Gofman turned his energies back to earlier pursuits and a lifetime of research and activism over the noxious effects of irradiation on cells and on individual and population risk of cancer and cardiovascular diseases. (HB)


Gofman, John W., Hanig, M., Jones, H.B., “Evaluation of Serum Lipoprotein and Cholesterol Measurements as Predictors of Clinical Complications of Atherosclerosis: Report of a Cooperative Study of Lipoproteins and Atherosclerosis” (1956) Circulation 14:689-741.

Henry Blackburn interview with John Gofman 2004. CVD History Archive, School of Public Health, University of Minnesota.