November 2002 I wish to follow up on the issue of herbal remedies and low-fat diets as referred to by Dr Montemuro (BCMJ 2001;43:452-457).
It may be of interest to note that the Nurses Health Study referred to in Dr Montemuro’s reply does indeed address fat consumption and cardiovascular mortality. However, total fat intake was not, in fact, related to cardiovascular mortality.
As the authors of the relevant paper (Hu, F., Stampfer M., Manson J., et al. Dietary fat intake and the risk of coronary heart disease in women. N Engl J Med 1997;337:1491-1499) state: “In age adjusted analyses, a higher total fat intake was significantly associated with increased risk [of coronary heart disease]… However, the association virtually disappeared in the multivariate analysis, primarily because of confounding by smoking… When total fat was entered into the multivariate model as a continuous variable, the relative risk was 1.02 (95% confidence interval, 0.97 to 1.07; P=0.55)…”
A couple of points are illustrated by this data: Observational studies may provide useful information about possible sources of disease but are often plagued by confounders, and these studies cannot establish causation.
Even if the 83% reduction in coronary artery events referred to in Dr Montemuro’s letter truly were due to low fat intake (as opposed to smoking and other confounders as the authors suggest) the benefits of this are likely to be inflated simply because the Nurses Health Study dealt with a population in which the baseline risk of coronary artery disease was quite low (women of mean age 47 years at outset). In this sense, the absolute risk reduction (ARR) would be a more appropriate measure than relative risk reduction. In this case the 939 cases of non-fatal MI or death from coronary artery disease occurred in a group of 80082 women over 14 years. The risk here for an individual is 0.00084 per year. An 83% risk reduction would yield an annual individual risk of 0.00014 per year with an ARR of 0.00070 per annum.
As mentioned earlier, even this tiny reduction of individual risk is not supported by the data in the Nurses Health Study which, again, indicates that smoking is the more likely factor related to cardiovascular mortality rather than total fat intake per se.
—Lloyd Oppel, MD
Vancouver
However, it should be noted that there is a “big picture” here and we should not forget that individuals with significantly elevated lipids, irrespective of other confounders, are indeed at increased risk of vasculopathy. Dietary modifications +/- pharmacotherapy for these individuals can be lifesaving and needs to continue to be part of the interventional armamentarium. —ED.