Day 017 - 25 Jul 94 - Page 30


     
     1        with no allowance for wastage.  Furthermore, since
              averaging is done over the entire population, no account
     2        is taken of differing age and sex patterns.  A more
              important problem is the fact that comparisons of food
     3        balance data with individual-based nutritional studies
              have shown that the former are not equally reliable (or
     4        unreliable) in all countries, or within a given country,
              for each food item.
     5
              The imperfections are not confined to the nutritional
     6        details, since the mortality data include countries with
              up to 14% of deaths from senility and other ill-defined
     7        causes at ages 35-64.
 
     8        Fat is sometimes singled out in these analyses because of
              the magnitude of the correlation coefficient with breast
     9        cancer mortality (0.89).  In fact, an economic variable,
              gross national product also gives a high correlation with
    10        breast cancer (0.72).  Indeed, for breast cancer
              incidence, which in an aetiological context is the more
    11        relevant" -- is it more relevant than mortality?
              A.  It is because obviously some people are "cured", in
    12        inverted commas, of their cancers.  So that if you are
              measuring purely mortality data, you are not getting a
    13        true measure of how frequently the cancer develops in that
              particular society.  You are getting a lower estimate.
    14
         Q.   "For breast cancer incidence ... the value is higher
    15        (0.83)   with GNP than it is with fact (0.79)".  That is
              again Armstrong and Doll.  "It is therefore possible that
    16        some contribution to the high correlation comes from
              economic effects on the quality of the incidence data; for
    17        these operate not only within a vital statistics office or
              cancer registry, but touch many aspects of national life
    18        including the availability and organisation of medical
              services.
    19
              The problem is that fat consumption and national wealth
    20        are so highly correlated that it is difficult to
              distinguish the effects of fat per se from those of other
    21        indirect effects of wealth that influence breast cancer
              risk.  These include early age at menarche, delayed first
    22        birth and low family size.  One of the most distinctive
              aspects of rich developed countries is their
    23        overconsumption of food in general.  This is highly
              relevant sine obesity (and in certain studies, body size)
    24        is another determinant of breast cancer risk consistent
              with an effect of the oestrogen produced in adipose tissue
    25        from" -- you say it, I do not know.
  
    26        A. "Androstenedione". 
  
    27   Q.   Thank you.  Is that what you were telling us earlier?
              A.  That is what I was telling you earlier, that oestrogen
    28        like substances can be produced in adipose tissue and,
              therefore, may be changing the hormone environment of an
    29        individual, but he mentions these other factors which we
              have discussed already, such as menarche, family size and
    30        the age at which one has one's first child.  These are
              certainly socially, at least in part, socially related

Prev Next Index