Page 39 - Spring2008

ANALYSIS OF VEGETARIAN STUDIES BY RUSSELL SMITH

Russell Smith, PhD, was a statistician and critic of the lipid heart theory of heart disease. He is the author if
the massive Diet, Blood Cholesterol and Coronary Heart Disease: A Critical Review of the Literature (1991, Vector
Enterprises), as well as The Cholesterol Conspiracy (Warren H. Green, Inc., 1991). As part of his efforts to reveal the
flimsiness of the theoretical basis for the lipid hypothesis, he also looked at studies on vegetarianism in the scientific
literature.
In a review of some 3,000 articles, Smith found only two that compared mortality data for vegetarians and non-
vegetarians. One was a 1978 study of Seventh Day Adventists (SDAs) to which the above unreferenced claim probably
refers. Two very poor analyses of the data were published in 1984, one by H. A. Kahn and one by D. A. Snowden.
3
The publication by Kahn rather arbitrarily threw out most of the data and considered only subjects who indicated
very infrequent or very frequent consumption of the various foods. The author then computed “odds ratios” which
showed that mortality increased as meat or poultry consumption increased (but not for cheese, eggs, milk or fat at-
tached to meat). When Smith analyzed total mortality rates from the study as a function of the frequencies of consuming
cheese, meat, milk, eggs and fat attached to meat, he found that the total death rate decreased as the frequencies of
consuming cheese, eggs, meat and milk increased. He called the Kahn publication “yet another example of negative
results which are massaged and misinterpreted to support the politically correct assertions that vegetarians live longer
lives.”
The Snowden analysis looked at mortality data for coronary heart disease (CHD), rather than total mortality data,
for the 21-year SDA study. Since he did not eliminate the intermediate frequencies of consumption data on meat, but
did so with eggs, cheese and milk, this analysis represents further evidence that both Kahn and Snowden based their
results on arbitrary, after-the-fact analysis and not on pre-planned analyses contingent on the design of their ques-
tionnaire. Snowden computed relative risk ratios and concluded that CHD mortality increased as meat consumption
increased. However, the rates of increase were trivial at 0.04 percent and 0.01 percent respectively for males and
females. Snowden, like Kahn, also found no relationship between frequency of consumption of eggs, cheese and milk
and CHD mortality “risk.”
Citing the SDA study, other writers have claimed that nonvegetarians have higher all-cause mortality rates than
vegetarians and that, “There seems little doubt that SDA men at least experience less total heart disease than do
4
others. . .” The overpowering motivation to show that a diet low in animal products protects against CHD (and other
5
diseases) is no better exemplified than in the SDA study and its subsequent analysis. While Kahn and Snowden both
used the term “substantial” to describe the effects of meat consumption on mortalities, it is obvious that “trivial” is
the appropriate descriptor. It is also interesting to note that throughout their analyses, they brushed aside their totally
negative findings on foods which have much greater quantities of fat, saturated fat and cholesterol.
The second study was published by Burr and Sweetnam in 1982. It was shown that annual CHD death rate
6
among vegetarians was only 0.01 percent lower than that of nonvegetarians, yet the authors indicated that the differ-
ence was “substantial.”
The table below presents the annual death rates for vegetarians and nonvegetarians which Smith derived from
the raw data in the seven-year Burr and Sweetnam study. As can be seen, the “marked” difference between vegetar-
ian and nonvegetarian men in Ischemic Heart Disease (IHD) was only .11 percent. The difference in all-cause death
rate was in the opposite direction, a fact that Burr and Sweetnam failed to mention. Moreover, the IHD and all-cause
death rates among females were actually slightly greater for heart disease and substantially greater for all causes in
vegetarians than in nonvegetarians.

ANNUAL DEATH RATES OF VEGETARIANS AND NONVEGETARIANS

IHD All-Cause
Male vegetarians .22% .93%
Male nonvegetarians .33% .88%
Female vegetarians .14% .86%
Female nonvegetarians .10% .54%

These results are absolutely not supportive of the proposition that vegetarianism protects against either heart
disease or all-cause mortalities. They also indicate that vegetarianism is more dangerous for women than for men.

SPRING 2008 Wise Traditions 39