Day 036 - 13 Oct 94 - Page 18


     
     1        shown a weak association or no association between dietary
     2        fat and breast cancer.  They are also consistent with
     3        studies suggesting that fat intake is a determinant of
     4        breast cancer, particularly after accounting for
     5        inaccuracies in dietary assessment.  The effects of
     6        different energy-adjustment methods may account in part
     7        for the varying interpretations of four previous cohort
     8        studies of dietary fat and breast cancer.  Implications:
     9        Further work is needed to clarify not only the nature of
    10        the dietary fat-breast cancer association, but also the
    11        impact of different analytic methods used in the
    12        investigation of diet-disease associations."
    13
    14        Would you agree that this study appears to have, as it
    15        were, tested its information in at least two different
    16        ways?
    17        A.  Yes, that is right.
    18
    19   Q.   We can see the results of that in a table, I think it is
    20        in this paper, (alas no page numbers) -- it must be 1098,
    21        I think; it is after 1097, which you will find in the
    22        bottom right-hand corner.  Sorry, it has 512 stamped at
    23        the bottom.  I do not have that.  There are two tables
    24        there.  Do you see that?
    25        A.  Yes, I do.
    26
    27   Q.   You see, do you not, that according to which principle you
    28        use, the first one is "Multivariate-adjusted relative risk
    29        of breast cancer according to quartile of macronutrient
    30        intake"; the second one (table 4) is the same thing:
    31        "... according to quartile of macronutrient intake
    32        residuals".  Could you explain the difference between
    33        those two, Dr. Barnard?
    34        A.  Not without reading the study.
    35
    36   Q.   Maybe it does not matter.  What is interesting though, if
    37        you look at the figures in the table, is that in table 3
    38        total fat there is an upward line, is there not -- there
    39        are four quartiles -- 1 is low and 4 is high; the relative
    40        risk goes up.
    41        A.  That is correct.
    42
    43   Q.   That is right, is it not?  If you look at saturated fat it
    44        has a bump at 3 and then it descends again, does it not?
    45        A.  You could say that, yes.
    46
    47   Q.   When you look at the other method of calculation (table
    48        4), total fat residual, you go up from 1 to 2, slightly up
    49        from 2 to 3 and then from 3 to 4 you go down again, do you
    50        not? 
    51        A.  Yes. 
    52 
    53   Q.   The same sort of effect for saturated fat residual, except
    54        it is even more like a switch back; it goes from 1 to
    55        1.08, then to 0.96, then back up to 1.10.  Can I draw your
    56        attention to some of what the authors said about these
    57        results?
    58        A.  Well, it should perhaps be pointed out that the
    59        right-hand column indicates that the differences are not
    60        significant in the saturated fat line that you just

Prev Next Index