Education And Debate

Risk factor thresholds: their existence under scrutiny

BMJ 2002; 324 doi: https://doi.org/10.1136/bmj.324.7353.1570 (Published 29 June 2002) Cite this as: BMJ 2002;324:1570

Association does not prove causation and is not the same as a dose-response relation.

Law and Wald posit that physiological variables believed to be
associated with a disorder should be reduced irrespective of their initial
value if a dose-response relation can be demonstrated between the variable
and the disease (1). This is a puzzling proposal that primarily appears to
be an ad hoc attempt to discredit numerous observations that contradict
current concepts about the cause of cardiovascular and other diseases. It
is not only in conflict with the literature but also violates the
principles of medical science for the following reasons.

By selecting biochemical and biophysical variables that allegedly induce
illness they have confused causation with association. For instance,
while hypertension is an established risk factor for coronary heart
disease at all ages and in both sexes, most high blood pressure-lowering
trials have shown no reduction in coronary or total mortality in women and
younger individuals (2) and in many studies, no effect has been
demonstrated on male mortality either (3). This clearly indicates that
hypertension is not a causa vera of coronary heart disease and "risk
marker" would be a more appropriate description.

Selecting the MRFIT study to support evidence for a causal role of high
cholesterol for coronary heart disease is similarly a poor choice since
more careful analysis of the data does not support this conclusion (4). In
addition, Law and Wald have chosen to ignore the many studies that found
an absence, and in some instances, an inverse correlation between blood
cholesterol levels and the incidence of coronary events (5).

Perhaps the most egregious error is that Law and Wald use results from
cohort studies as evidence of dose-response relations. As they themselves
admit, dose-response implies that a given change in the variable should
reduce the likelihood of a disease by a constant proportion of the
existing risk. But in most of the studies they have chosen as evidence of
dose-response, no risk factor was changed. What these studies recorded
were instant observations of risk factors and disease and therefore they
cannot prove or disprove dose-response.

A recent review has shown that true dose-response, or preferably, exposure
-response, was absent in almost all cholesterol-lowering angiographic
trials as assessed by the degree of atherosclerosis growth (5). Changes in
blood cholesterol bore no significant relationship to the degree of
atherosclerotic progression, clearly indicating that the development of
atherosclerotic plaques has nothing to do with blood cholesterol. Also, in
the only clinical statin trial where exposure-response was analysed, the
extent of LDL-reduction was not a significant predictor of the coronary
event rate, whether expressed as an absolute amount (p=0.97) or a
percentage (p=0.76)(7)

Law and Wald think that an important cause of heart disease that is highly
correlated with serum cholesterol is not known. May we remind them that
several potential etiologic influences satisfy this description. Smoking,
lack of exercise, obesity, hypertension, emotional stress (6) and
hyperhomocysteinemia (8) are factors that are associated with high
cholesterol. All of these may contribute to cardiovascular disease by
mechanisms other than raising serum cholesterol or lipid levels.

1. Law MR, Wald NJ. Risk factor thresholds: their existence under
scrutiny. BMJ 2002;324:1570-6.

2. Måttligt förhöjt blodtryck. SBU-rapport nr. 121. Stockholm 1994.

3. Hebert PR, Moser, M, Mayer J, Hennekens CH. Recent evidence on drug
therapy of mild to moderate hypertension and decreased risk of coronary
heart disease. Arch Int Med 1993;153:578-81.

4. Werkö L. Analysis of the MRFIT screenees: a methodological study. J
Intern Med 1995;237:507-18.

5. Ravnskov U. The Cholesterol Myths. New Trends Publishing, Washington DC
2002, p. 58-64.

6. Ravnskov U. Is atherosclerosis caused by high cholesterol? QJM 2002;
95:397-403.

7. Sacks FM, Moye LA, Davis BR, Cole TG, Rouleau JL, Nash DT, Pfeffer MA,
Braunwald E. Relationship between plasma LDL concentrations during
treatment with pravastatin and recurrent coronary events in the
Cholesterol and Recurrent Events trial. Circulation. 1998;97:1446-52

8. Refsum H, Ueland PM, Nygard O, Vollset SE.. Homocysteine and
cardiovascular disease. Annu Rev Med. 1998;49:31-62.

Competing interests: No competing interests

16 July 2002
Uffe Ravnskov
Joel M. Kauffman, Peter H. Langsjoen, Kilmer S. McCully and Paul J. Rosch
Magle Stora Kyrkogata 9, S-22350 Lund, Sweden