Day 033 - 10 Oct 94 - Page 31
1 association. It did not find a statistical association
2 between fat and breast cancer. This study has been widely
3 criticised on the grounds that the range of fat intake was
4 quite narrow.
5
6 To put this in context, when Japan's breast cancer rates
7 were quite low -- they are still low but they have risen
8 quite dramatically as the fat intake has gone up in Japan
9 -- when they were at their low point (I am speaking of
10 1940s, around that time) the fat intake in Japan was seven
11 to ten per cent of calories; in the United States it is 37
12 per cent.
13
14 Well, in the nurses' study there was no group that looked
15 at fat intakes of seven per cent or 17 per cent or 20 per
16 cent. If memory serves -- I do not have the study in
17 front of me -- but my recollection is that the median fat
18 intake of the lowest quintile in that study was 27 per
19 cent of calories. They did not look lower than that. The
20 high end was somewhat higher than average, but this is a
21 rather narrow band. One would not have expected to see
22 differences in breast cancer rates within that band. In
23 other words, it is rather like if you have -----
24
25 Q. Are you saying that the -----
26
27 MR. JUSTICE BELL: I understand what he is saying.
28 A. If we have a group of smokers, some smoke three packs
29 a day and some smoke two packs a day ----
30
31 Q. The difference in intake was not pronounced enough, you
32 contend, to draw any satisfactory conclusion from it?
33 A. Yes, I would contend that. That has been a common
34 criticism of Willett's study.
35
36 MR. MORRIS: It is not just the range of fat, is it? It is
37 also the fact that the range was actually at a high level
38 in any case?
39 A. Both are true. The range was narrow and the range was
40 quite high; rather like, as I said, with cigarette
41 smoking, if you have a group that smokes three packs a day
42 of cigarettes and another that smokes two packs a day of
43 cigarettes, one might not be very surprised to find very
44 little difference in the risk of lung cancer between the
45 two groups. In fact, if you look at the cancer data, it
46 is exactly what you find.
47
48 One would argue -- I would certainly argue -- that the
49 range in Willett's study was quite narrow; that all the
50 subjects were Americans following a typical American diet,
51 some slightly higher fat intakes, some slightly lower; but
52 the differences were not enough to demonstrate any effect
53 on breast cancer risk.
54
55 MR. JUSTICE BELL: So the lower line was just within the
56 recommendations?
57 A. If memory serves, their median intake was about 27 per
58 cent, and many cancer research groups have suggested that
59 the diet should be below 30 per cent of calories from fat,
60 although I should say that that limit is -- at least
