Day 036 - 13 Oct 94 - Page 32
1 cancer, a positive association between amounts of fat and
2 incidence of breast cancer?
3 A. Yes. What Dr. Willett argues is that even at very
4 high intakes of fat, such as was found in the nurses
5 study, one should be still able to see a little bit more
6 breast cancer among those women on astronomical amounts of
7 fat versus those on only moderately astronomical amounts
8 of fat.
9
10 Q. How explain the contrasting findings of Dr. Willett's work
11 in relation to colon cancer on the one hand and breast
12 cancer on the other?
13 A. Well, there are a number of different points that we
14 have gone over already. One has to say that when a clear
15 and significant relationship is found, if that is only
16 found in, say, one in 20 studies, or thereabouts it may be
17 perhaps by chance alone, but aside from that, that it does
18 not mean something, if one has investigated in a
19 population and not found an association, that certainly
20 does not disprove it. What seems quite likely (and as
21 Dr. Campbell's data suggest) is that, indeed, what is
22 probably most important is dietary intake earlier in life,
23 although as the studies of Gregorio, Verreault and others
24 have indicated, there never seems to be a time where fat
25 has no effect. But what is earlier in life may be
26 important. It may well be, as Dr. Willett suggests, that
27 lower fat intakes are where the threshold effect lies.
28 That has been the suggestion of Dr. Campbell.
29
30 I would take issue with what Dr. Willett says here, such
31 as below 20 per cent of energy being what is relevant.
32 That may be true. Regretfully, Dr. Willett's data do not
33 get even anywhere near that level. My recollection, I
34 have not re-read this in a while, is that his lowest
35 quintile is 27 per cent average fat intake. Again, it is
36 a short-term study. Eight years is longer than four, but
37 it is rather short. In carcinogenesis eight years is in
38 fact the time, an average time, between the beginning of
39 cancer and before it is even detected. So any woman who
40 was diagnosed as, say, having cancer eight years out,
41 already had cancer at the beginning of the study.
42
43 So, we have a very narrow range. We have a short-term
44 study and, as we described yesterday, his interpretation
45 is something to which he is entitled. I think he is quite
46 capable of doing a well-designed study. But that
47 interpretation is not one which is supported by other
48 authorities, or certainly most other authorities, although
49 some do agree with Dr. Willett, it is not a majority view.
50
51 Q. I am going to be quite blunt with you, Dr. Barnard. You
52 have not succeeded in showing us (and I do not know of any
53 authority certainly in the last eight or 10 years that has
54 felt able to assert, whether it be a public authority or a
55 professional scientist, with any confidence) that these
56 associations we have been discussing are, indeed, causal?
57 A. The associations have not been proved.
58
59 Q. That is not what I put.
60
