The fallacies of the lipid hypothesis

Abstract

Most researchers to-day consider that a high intake of saturated fat and elevated LDL cholesterol are the most important causes of atherosclerosis and coronary heart disease. The lipid hypothesis has dominated cardiovascular research and prevention for almost half a century although the number of contradictory studies may exceed those that are supportive. The harmful influence of a campaign that ignores much of the science extends to medical research, health care, food production and human life. There is an urgent need to draw attention to the most striking contradictions, many of which may be unknown to most doctors and researchers.

Full text

INVITED ARTICLE
The fallacies of the lipid hypothesis
UFFE RAVNSKOV
Abstract
Most researchers to day consider that a high intake of saturated fat and elevated LDL cholesterol are the most important
causes of atherosclerosis and coronary heart disease. The lipid hypothesis has dominated cardiovascular research and
prevention for almost half a century although the number of contradictory studies may exceed those that are supportive.
The harmful influence of a campaign that ignores much of the science extends to medical research, health care, food
production and human life. There is an urgent need to draw attention to the most striking contradictions, many of which
may be unknown to most doctors and researchers.
Key words: Cholesterol, saturated fat, cardiovascular disease
The lipid hypothesis consists of two main postulates.
The first says that a high intake of saturated fat raises
blood cholesterol an d the second that high choles-
terol leads to atherosclerosis and cardiovascular
disease. The philosopher Karl Popper is usually
cited as codifying the principle of falsifiability that
scientific theories must meet. The lipid hypothesis
conforms to this requirement: if saturated fat pro-
motes cardiovascular disease by raising choles terol, a
high intake should be followed by high total and/or
low-density-lipoprotein (LDL) cholesterol and
should be associated with an increased risk of
achieving cardiovascular disease; a decreased intak e
should reduce that risk. If high total or LDL
cholesterol causes atherosclerosis, people with high
cholesterol should be more atherosclerotic and run a
greater risk of achieving cardiovascular disease than
people with low cholesterol, and a lowering of total
and/or LDL cholesterol should lead to regress, or at
least to a slower progress of atherosclerosis and to a
lower risk of cardiovascular disease. If these out-
comes are not found, that is, if saturated fat does not
reliably increase cholesterol, if cholesterol does not
reliably predict atherosclerosis and cardiovascular
disease, the theory must be accepted as false. In the
following I shall demonstrate that the postulates have
been falsified effectively in many studies.
Saturated fat
According to all official guidelines a diet rich in
carbohydrates and poor in fat, in particular saturated
fat, is said to be the best non-pharmaceutical way to
prevent cardiovascular disease. The main argument
against saturated fat is that a high intake raises blood
cholesterol. A strong contradiction of this statement
is that the lowest cholesterol concentrations ever
seen have been measured in African tribes whose
diet consists almost entirely of animal food (1). Even
if this is considered an exception ! in such a complex
system, the standards of falsifiability are not absolute
! the cholesterol-raising effect cannot be considered
highly predictable, because in at least ten recent
controlled low-carbohydrate trials, where intakes of
saturated fat were 3!7 times higher than the recom-
mended upper limit, total or LDL cholesterol
remained unchanged. As the concentration of small
dense LDL particles is a stronger risk factor for
cardiovascular disease than LDL cholesterol itself it
is also a contradiction to the theory that LDL size is
inversely associated with intake of saturated fat (2).
Even if a high intake of saturated fat raised
cholesterol to any extent in more balanced diets,
this is surrogate outcome. The crucial question is
whether a high intake leads to cardi ovascular disease,
but few studies suppor t that notion. At least
Correspondence: Uffe Ravnskov, Magle Stora Kyrkogata 9, S 22350 Lund, Sweden. Tel and Fax: "46 46145022. E-mail: ravnskov@tele2.se
Scandinavian Cardiovascular Journal. 2008; 42: 236!239
(Received 8 February 2008; accepted 12 Febr uary 2008)
ISSN 1401-7431 print/ISSN 1651-2006 online # 2008 Informa UK Ltd.
DOI: 10.1080/14017430801983082

30 cohort and case-control studies including more
than 300 000 individuals have found that coronary
patients have not eaten more saturated fat before
their first heart attack than others (1). More at odds
is that in at least seven cohort studies stroke patients
had eaten significantly less. There is no support from
autopsy studies either: low-consumers are just as
atherosclerotic as high-c onsumers (1), and a meta-
analysis of the controlled, randomised dietary clin-
ical trials found no effect of reducing saturated fat,
either on coronary morbidity, coronary mortality, or
total mortality (1).
High cholesterol
Contrary to the widely held belief among doctors
and researchers, there is little evidence that high
cholesterol leads to cardiovascular diseas e. Autopsy
studies of individuals who have died from non-
medical causes have confirmed Lande´ and Sperry?s
observation from 1936 of an absence of an associa-
tion between total or LDL cholesterol and degree of
atherosclerosis, measured before death or immedi-
ately after (Figure 1) (2). A few autopsy studie s of
patients with cardiovascular disease have found a
weak association, probably because such studies
always include proportionally more patients with
familial hypercholesterolaemia. As the latter on
average always are more atherosclerotic than others,
and as their cholesterol is higher, their inclusion
automatically creates an association between choles-
terol and atherosclerosis, although, as explained
below, their higher degree of atherosclerosis may
not be due to their high cholesterol (Figure 2) (3).
Some studies of young and middle-aged men have
found high cholesterol to be a risk factor for
coronary disease, whereas others have not (4). On
the other hand, high cholesterol has not been found
to be a risk factor in patients with diabetes (4),
patients with established heart disease (3), patients
with terminal renal failure, (5) or in women (3), and,
most surprising, rarely in old people (6), although at
least in Sweden more than 90% of all coronary
deaths occur in people above age 65. Also contra-
dictory is that neither the concentration of LDL nor
total cholesterol predict the degree of prog ress of
angiographic changes (7), and neither are associated
with peripheral atherosclerosis or intermittent clau-
dication (8). In cohorts of people with familial
hypercholesterolaemia, LDL or total cholesterol do
not predict future coronary heart disease or per iph-
eral atheroscl erosis; those with moderately elevated
cholesterol run the same risk as those whose
cholesterol is 2!3 times higher than the mean value
in normal people (9!12). Indeed, in one study those
with the highest cholesterol had the lowest risk of
heart dis ease (12). Even Brown and Goldstein were
aware of the lack of an association between choles-
terol and cardiovascular disease among these people.
Thus, in a 1983 paper they wrote: ‘‘Among FH
patients (both heterozygous and homozygous), there
is considerable variation in the rate of progression of
atherosclerosis, despite uniformly elevated LDL
levels.’’(13). The explanation may be that other
hereditary abnormalities are seen in some of these
people, for ins tance a predisposition to abnormalities
of the coagulation system, which is a strong risk
factor for coronary heart disease in familial hyperch-
olesterolaemia (14).
Cholesterol lowering does not prevent
cardiovascular disease
Meta-analyses of the cholesterol lowering trials
before the introduction of the statins found no effect
0
500
1000
1500
2000
2500
3000
3500
4000
0 50 100 150 200 250 300 350 400
Cholesterol; mg/dl
Lipid content in aorta.
Figure 1. Association between degree of atherosclerosis and total
cholesterol concentration in the blood in 40 men and women
between 50 69 years, who had died violently without preceding
disease. The figure is constructed using data from Lande´ and
Sperry (see ref. 7).
0
10
20
30
40
50
60
70
80
90
100
150 200 250 300 350 400
Cholesterol; mg/dl
% surface involved
Figure 2. Association between degree of atherosclerosis and total
cholesterol at autopsy. It is obvious that the weak association
disappears after exclusion of individuals with cholesterol above
350 mg/dl. (9 mmol/l). Redrawn from Solberg et al. (see ref. 7).
The lipid hypothesis 237

of cholesterol lowering on heart mortality; indeed,
total mortality increased (15). Statin treatment is
able to lower heart mortality, but the effect is trivial,
it is only presen t in high-risk, young and middle-age
male patients, and the small effect achieved is
probably due to the statins? pleiotropic effects, not
to cholesterol lowering, because there is no associa-
tion between initial cholesterol, or degree of choles-
terol lowering, and the clinical or angiographic
outcome (3,6).
Unfortunately, the statins have many side effects,
such as muscular problems, liver damage, renal
failure, depression, amnesia and nerve damage, as
well as impotence, abortion and severe birth def ects
(16!19). Ac cording to the reports from the statin
trials, all of which have been sponsored by the drug
companies, side effects are mild and rare, but
underreporting is prevalent. Muscular symptoms
for instance are said to occur in less than one
percent, but researchers independent on the drug
companies have found the frequency to be 64% (20)
and 75% (21). This side effect may not only be
painful, it also hampers exercising, the most im-
portant, the cheapest and the least harmful measure
in the prevention of heart disease.
High cholesterol may be beneficial
By 1992, a meta-analysis of 19 cohort studies
including more than 600 000 men and women
from many countries had found that cholesterol
was inversely associated with mortality from respira-
tory and digestive diseases (22), most of which were
of an infectious origin. The observation was in line
with a large number of epidemiological, laboratory
and experimental studies showing that high choles-
terol protects against infections (6). The main effect
seems to be exerted by the LDL molecule. For
instance, mice with FH challenged with bacterial
endotoxin by injection with Gram-negative bacteria,
had an 8-fold increased LD50, and a significantly
delayed and overall lower mortality than control
mice. Also, pharmacologically induced hypocholes-
terolaemia led to a markedly increased endotoxin
mortality that was returned to normal after injection
of exoge nous lipoproteins; and the haemolytic effects
of Staphylococcus aureus a-toxin was prevented by
adding purified LDL to the test tube (6). In
agreement with this idea, individua ls with famili ar
hypercholesterolaemia had a longer life expectancy
than others before year 1900, where the main cause
of mortality was infectious diseases (23). A recent,
large follow-up study found that their average life-
span was normal because the increased coronary
mortality was balanced by a decreased mortality
from other diseases (24). Older people with high
cholesterol live longer than older people with low
cholesterol, and this author has tabulated the many
clinical studies that have confirmed the benefits of
high choles terol in various types of infectious dis-
eases (6).
Conclusions
The cholesterol hypothesis has failed to stand up to
the standard of falsifiability. In a med ical theory, it
takes more than one counter-example, but this
review has pointed to the numerous studies that
contradict basic predictions. Although I have fol-
lowed the literature in this field meticulou sly for
almost 18 years I may have overlooked supportive
studies. However, a few supportive studies cannot
outweigh the large number of contradictory ones.
But the cholesterol campaign continues, inappropri-
ate studies are funded and controversial studies are
either ignored or cited as if they were supportive
(15,26). There is, however, increasing awareness
that the diet-heart hypothesis is sustained by a
number of soci al, political and financial factors,
most of which have little to do with science or any
established success in public health.
References
1. Ravnskov U. The questionable role of saturated and poly
unsaturated fatty acids in cardiovascular disease. J Clin
Epidemiol. 1998;51:443 60.
2. Krauss RM, Blanche PJ, Rawlings RS, Fernstrom HS,
Williams PT. Separate effects of reduced carbohydrate intake
and weight loss on atherogenic dyslipidemia. Am J Clin Nutr.
2006;83:1025 31.
3. Ravnskov U. The Cholesterol Myths. Washington: New
Trends Publishing; 2000.
4. Roselli della Rovere G, Lapolla A, Sartore G, Rossetti C,
Zambon S, Minicuci N, et al. Plasma lipoproteins, apopro
teins and cardiovascular disease in type 2 diabetic patients. A
nine year follow up study. Nutr Metab Cardiovasc Dis. 2003;
13:46 51.
5. Bellomo G, Lippi G, Saronio P, Reboldi G, Verdura C, Timio
F, et al. Inflammation, infection and cardiovascular events in
chronic hemodialysis patients: a prospective study. J Nephrol.
2003;16:245 51.
6. Ravnskov U. High cholesterol may protect against infections
and atherosclerosis. Q J Med. 2003;96:927 34.
7. Ravnskov U. Is atherosclerosis caused by high cholesterol. Q J
Med. 2002;95:397 403.
8. Mo
¨
lgaard J, von Schenk H, Kuusi T, Holme I, Lassvik C,
Taskinen MR, et al. Plasma lipoprotein abnormalities and
apolipoptotein E phenotypes in intermittent claudication.
A multivariate analysis of randomly selected subjects. Nutr
Metab Cardiovasc Dis. 1996;6:114 23.
9. Miettinen TA, Gylling H. Mortality and cholesterol metabo
lism in familial hypercholesterolemia. Ar teriosclerosis. 1988;
8:163 7.
10. Neil HAW, Seagroatt V, Betteridge DJ, Cooper MP, Dur
rington PN, Miller JP, et al. Established and emerging
238 U. Ravnskov

coronary risk factors in patients with heterozygous familial.
hypercholesterolaemia. Heart. 2004;90:1431 7.
11. Rodriguez G, Bertolini S, Nobili F, Arrigo A, Masturzo P,
Elicio N, et al. Regional cerebral blood flow in f amilial
hypercholesterolaemia. Stroke. 1994;25:831 6.
12. Hopkins PN, Stephenson S, Wu LL, Riley WA, Xin Y, Hunt
SC. Evaluation of coronary risk factors in patients with
heterozygous familial hypercholesterolemia. Am J Cardiol.
2001;87:547 53.
13. Brown MS, Goldstein JL. Lipoprotein metabolism in the
macrophage: Implications for cholesterol deposition in ather
osclerosis. Ann Rev Biochem. 1983;52:223 61.
14. Jansen AC, van Aalst Cohen ES, Tanck MW, Cheng S,
Fontecha MR, Li J, et al. Genetic determinants of cardiovas
cular disease risk in familial hypercholesterolemia. Arterios
cler Thromb Vasc Biol. 2005;25:1475 81.
15. Ravnskov U. Cholesterol lowering trials in coronary heart
disease: frequency of citation and outcome. BMJ. 1992;305:
15 9.
16. Ravnskov U, Rosch PJ, Sutter MC, Houston MC. Should
we lower cholesterol as much as possible? BMJ. 2006;332:
1330 2.
17. Solomon H, Samarasinghe YP, Feher MD, Man J, Rivas Toro
H, Lumb PJ, et al. Erectile dysfunction and statin treatment
in high cardiovascular risk patients. Int J Clin Pract. 2006;60:
141 5.
18. Kenis I, Tartakover Matalon S, Cherepnin N, Drucker L,
Fishman A, Pomeranz M, et al. Simvastatin has deleterious
effects on human first trimester placental explants. Hum
Reprod. 2005;20:2866 72.
19. Edison RJ, Muenke M. Central nervous system and limb
anomalies in case reports of first trimester statin exposure.
N Engl J Med. 2004;350:1579 82.
20. Marzoa Rivas R, Crespo Leiro MG, Paniagua Marin MJ,
Llinares Garcia D, Muniz Garcia J, Aldama Lopez G, et al.
Safety of statins when response is carefully monitored: a
study of 336 heart recipients. Transplant Proc. 2005;37:
4071 3.
21. Langsjoen PH, Langsjoen JO, Langsjoen AM, Lucas LA.
Treatment of statin adverse effects with supplemental Coen
zyme Q10 and statin dr ug discontinuation. Biofactors. 2005;
25:147 52.
22. Sinzinger H, O’Grady J. Professional athletes suffering from
familial hypercholesterolaemia rarely tolerate statin treatment
because of muscular problems. Br J Clin Pharmacol. 2004;57:
525 8.
23. Jacobs D, Blackburn H, Higgins M, Reed D, Iso H, McMillan
G, Neaton J, et al. Report of the Conference on Low Blood
Cholesterol: Mortality Associations. Circulation. 1992;86:
1046 60.
24. Sijbrands EJG, Westendorp RGJ, Defesche JD, de Meier
PHEM, Smelt AHM, Kastelein JJP. Mortality over two
centuries in large pedigree with familial hypercholesterolae
mia: family tree mortality study. Br Med J. 2001;322:
1019 23.
25. Neil HAW, Hawkins MM, Durrington PN, Betteridge DJ,
Capps NE, Humphries SE; Simon Broome Familial Hyperli
pidaemia Register Group and Scientific Steering Committee.
Non coronary heart disease mortality and risk of fatal cancer
in patients with treated heterozygous familial hypercholester
olaemia: a prospective registry study. Atherosclerosis.
2005;179:293 7.
26. Ravnskov U. Quotation bias in reviews of the diet heart idea.
J Clin Epidemiol. 1995;48:713 9.
The lipid hypothesis 239