Day 036 - 13 Oct 94 - Page 17
1 distorted if they do not exclude these special factors?
2 A. It is important to be as specific as possible.
3
4 Q. Tab 24, please, Dr. Barnard: We have moved forward five
5 years to 1992. I hope that I am going to find -- I am
6 doing this chronologically because I think that may be
7 important -- a paper by Lawrence Kushi?
8 A. Pronounced Kushi.
9
10 Q. Is he a well known figure in this field?
11 A. Yes, he is.
12
13 Q. He is from the University of Minnesota; is that right?
14 A. Yes, he is.
15
16 Q. Though the whole team does not come from Minnesota; one of
17 them it looks as though he came from the Iowa College of
18 Medicine, University of Iowa; roughly speaking, the same
19 part of America; never mind that. Can I read the
20 introduction: "Background: Although the results of
21 animal studies and cross-cultural comparisons generally
22 support a role for dietary fat in the aetiology of breast
23 cancer, results of analytic epidemiology studies are
24 equivocal. Purpose: The association between dietary fat
25 and subsequent breast cancer was examined in a cohort of
26 34,388 postmenopausal women from Iowa. Methods: Dietary
27 habits were assessed by a food-frequency questionnaire
28 mailed in January 1986. Through December 31, 1989, 459
29 incident cases of breast cancer occurred in this cohort.
30 Proportional hazards regression was used to examine the
31 dietary fat-breast cancer association while adjusting for
32 potential confounders. The effects on this association of
33 four analytic approaches to adjustment for energy intake
34 were also considered."
35
36 Does that mean, perhaps, we can see as we go through some
37 parts of this document, what they did was to apply four
38 different kinds of analytic methodology to the results
39 which they obtained?
40 A. Yes, that is correct.
41
42 Q. "Results: After adjustment for known determinants of
43 breast cancer, a modest positive association of total fat
44 intake with risk of breast cancer was seen.
45 Polyunsaturated fat intake was also positively associated
46 with breast cancer (relative risks from lowest to highest
47 intake, 1.0, 1.25, 1.31, and 1.49 ...). Different
48 approaches to adjustment for energy intake, however,
49 provided different impressions of the dietary fat-breast
50 cancer association. One method, involving categorization
51 of crude fat intake and inclusion of total energy intake
52 in regression analysis, gave relative risk estimates from
53 low to high fat intake of 1.0, 1.17, 1.25, and 1.38 ...
54 Another method, based on categorization of fat intake
55 residuals in which the variation in fat due to total
56 energy intake was removed, gave corresponding estimates of
57 1.0, 1.24, 1.30, and 1.16 ... The former suggests
58 increasing breast cancer risk with increasing fat intake;
59 the latter suggests no association. Conclusions: These
60 results are consistent with other cohort studies that have
