False Claims of Benefit in Cholesterol Trials

     In the pre-statin era of the 1970-1980s many trials were conducted in an attempt to prove the validity of the “lipid hypothesis”, the theory that dietary fat and cholesterol were causative in heart disease, and that reductions in dietary and blood cholesterol would translate into reductions in risk for heart disease and cardiovascular mortality.


The Framingham Study was the first major project designed for the purpose of establishing the underlying cause for the high rate of cardiac related deaths in the USA. It was commenced in 1948 and conducted over a 30 year period, to explore the relationship between cholesterol and coronary heart disease (CHD), and the results were published in a 30 year follow up paper in 1987.

The results from Framingham did not reveal an association between cholesterol and all cause mortality across all age groups, showing only an association  for males aged between 30 and 49 years of age. Beyond these years there was no evidence for  such an association. (1)

Association does not imply causation.

A significant finding from the Framingham data was that high levels of cholesterol were, in fact, predictive of increased survival and greater longevity for those over 49 years of age, and lower levels of cholesterol were associated with an increase in death from all causes, including those from CHD. (2) (3) (4)

The Framingham Study showed that for every 1 mg./ dl. drop in cholesterol during the first 14 years, there was a 14%  INCREASE in cardio-vascular disease, and a 11% INCREASE in overall mortality during the subsequent 18 years.

These facts were ignored by the American Heart Association (AHA) and the National Heart, Lung and Blood Institute (NHLBI), the trial sponsors, who conveniently altered the data to give credence to their ideology, and announced that “for every 1 % reduction in cholesterol there was a corresponding 2% reduction in CHD. (5)

Clearly, this exaggerated claim of benefit was unfounded, and did not reflect what the Framingham authors recorded !

In fact, the Framingham researchers went further in their support of cholesterol as providing a benefit, rather than risk.

Their Director, Dr. William Castelli, admitted that “in Framingham, Mass, the more saturated fat one ate, the more cholesterol one ate, the more calories one ate, the lower the person’s serum cholesterol…we found that the people who ate the most cholesterol, ate the most saturated fat, ate the most calories, weighed the least, were the most physically active” (6)

    MRFIT STUDY: (The Multiple Risk Factor Intervention Trial)

In their pursuit of the causative factors underling heart disease the  NHLBI  considered the role of 3 perceived risk factors, namely: high cholesterol, smoking and high blood pressure, and recruited over 360,000 males to test the effects of these risk factors in heart disease.

From this number, 12,000 were selected, aged 35-57, who were considered to be especially at risk for heart attack, because they had high cholesterol levels and smoked. Of note, is the fact that many of these screenees had followed the “healthy” diet as recommended by the AMA, but yet had “high cholesterol” level.     The screenees were divided into 2 groups, the “treatment” group being guided into an adopted lifestyle programme, incorporating dietary choices, anti-smoking programmes, and advice about exercise, while the “control” group received no such advice, and were permitted to follow their normal pattern of lifestyle and diet.

Blood samples and lifestyle counselling was conducted every 4 months for the “treatment “ group, while the “control” group had their blood tested annually, over a 7 year period.

After 7 years the results were tabled and compared, to evaluate the effects of lifestyle intervention and cholesterol lowering on mortality.

The researchers were excited to note the reductions in risk factors in the treatment” group, with a 7% reduction in cholesterol levels, and Blood Pressure reductions also.  The “control” group, who had not made major dietary changes, recorded a 5% reduction in cholesterol, but many more had continued to smoke. (3456 vs 2842)  (7)

In the “treatment” group 115 had died of CHD while 124 had died in the “control” group, a difference of 9 deaths, and deaths from ALL CAUSES were 265 in the “treatment” group” and 260 in the “control” group, a difference of only 5 deaths, but MORE deaths in the treatment group !

After a further 4 years, a total of 202 deaths from heart disease was recorded in the “treatment” group,  and 226 deaths were recorded in the “control” group.

It is unlikely that the small difference of 2% in cholesterol reduction accounted for the very small difference in death rates between the 2 groups, and more likely that the higher rate of smoking in the “control” group was a contributing factor, aided by the fact that more participants (991) who had been counselled in the “treatment” group had given up smoking, than in the “control” group (374). (7)

Details of the reductions in dietary fats and cholesterol were recorded by the American Dietetic Association (8)  as being quite significant  (25%-50%), yet despite these reductions the differences in blood cholesterol were insignificant.

From the data presented it is clear that survival benefits in the “treatment” group  were not significantly better, yet the researchers, after a review and revision of the study design, confidently proclaimed  a benefit for cholesterol reduction, stating that a “strong, continuous and graded “ relationship existed between blood cholesterol and CHD. (9) JAMA Nov, 28, 1986; 256; 200: 2823-2828

These conclusions were derived from a comparative analysis of  sub-groups in the study, and reported again in 1990. (10) JAMA 263, 1795-1801, 1990.

It is also noteworthy that the researchers failed to report any data for overall mortality. Had they done so they may have been less enthusiastic about their claims of benefit.

Neither did they suggest that the greater reduction in smoking in the “treatment” group might have contributed to the very small, but insignificant, difference in deaths from cardiac causes.

Some years later the study  data was reviewed by Dr. Hiroyasu Iso and his team, and reported  in the New England Journal of  Medicine, as being U-Shaped in character, demonstrating that the highest all-cause death rate was present amongst those with the LOWEST cholesterol levels, under 3.6 Mmol\L (140 mg/dl) , and the lowest mortality rates occurred across a range of 4.1Mmol/l  to  5.6 Mmol/l ( 160 – 219 mg/dl). (11)

The data derived from this large scale interventive study failed to support the hypothesis that dietary cholesterol reduction could reduce the rate of cardiovascular deaths.

It did, however, suggest that smoking cessation might reduce the risk for CHD.

In a methodological study by Prof. Lars Werko in 2009, he concluded that the often repeated statement that “ the MRFIT screenees constitute the largest and most exact data base regarding the relation of risk factors to mortality in the healthy male US population” has no foundation. (12)


In 1984 the 6th study in a series of costly attempts was initiated, to once again find evidence for the “cholesterol hypothesis”, the theory that cholesterol was a causative factor in heart disease

These earlier studies, between 1980 – 1985, were funded by the NHLBI in its endeavour to prove its hypothesis, that had  hitherto remained unproven.

In this study, at a cost of €140 million, the benefit of a cholesterol lowering drug,  cholestyramine (Questran) , was tested, as opposed to the early dietary intervention trials, and 4000 participants were recruited for enrolment, after testing blood cholesterol levels in half a million middle aged males.

As before, a treatment group, using cholestyramine, was tested against a control group, using a placebo powder, and the number of heart attacks in each group assessed after 7 years of treatment.

A non-fatal heart attack was recorded in 190 men (10%) in the treatment group, compared with 212 in the control group (11.1%), a difference of 1.1%

Fatal heart attacks occurred in 30 patients (1.6%) in the treatment group, and in 38 patients  (2%) in the control group, a difference of 0.4 %.

Despite an absolute difference of only 0.4% the researchers quoted a “relative risk” reduction of 24%, which is clearly an exaggerated figure used to artificially inflate the perception of benefit.

To further exaggerate the figures the researchers abandoned their previous decision to use a stringent statistical test for determining accuracy, known as a “two-tailed” test, or a probability level of 0.01 (error probability 1:100).

Their 1979 protocol, as reported  (13) was conveniently substituted for a less stringent  “one-tailed” test, to obtain statistical significance !

When the final figures for total mortality (deaths from all causes) were calculated, the figures for both groups  were not largely different, with 68 deaths (3.6%) in the treatment group, and 71 deaths (3.7%) in the control group.

This insignificant  difference between the 2 groups  in  did not deter the researchers from publishing  their unfounded claim, of “strong evidence” for a causal role for cholesterol in CHD. and a beneficial role for dietary cholesterol lowering.

Further unsubstantiated claims of predicted benefit were made by the researchers for women and other groups, even though no women were included in this study.

Only middle-aged males who had higher than normal ( the upper 0.8% of those recruited) levels of cholesterol, which in all likelihood represented those with inborn errors of metabolism, were selected. These high levels were an unlikely representation of average middle aged males.

The extrapolation of these results, even though unconvincing, to the general public and to all population groups, would not be appropriate.

Once again, at great cost, a large scale trial was conducted to substantiate claims of mortality benefit for cholesterol reduction, through a drug intervention strategy.

Clearly, as with the previous trials, there was no evidence in this study to support the thesis that cholesterol lowering conferred a significant  benefit, either for all cause mortality, or for deaths from coronary artery disease.

Dr. Neville Wilson.

The Leinster Clinic – Medical Suite.

July, 2011.



  1. JAMA 1987; 257: 2176-2180 Anderson, K.M, et al
  2. Athero Thromb & Vascular Biol., 1997; 17: 1224-1232.
  3. Lancet, 1997; 350: 1778 – 1779. (females)
  4. Lancet, 1997; 350: 119 – 1123. (elderly)
  5. Circ 81; 1721-1723, 1994.
  6. Arch of Int Med, July 1992; 152:1371-1372 Castelli, W.P.
  7. Jama 248, 1465 – 1477, 1982
  8. American Dietetic Ass 86, 744 – 758, 1986.
  9. JAMA Nov 28, 1986; 256 : 200: 2833-2828)
  10. JAMA 263, 1795 – 1801, 1990
  11. NEJM, Apr, 1989; 320 (14) 904 – 910.
  12. J of Int Med 237, 507 –518, May 1995.
  13. J of Chronic Diseases, 1979, 52: 619 – 631.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>