Re: World Health Organization results: Passive Smoking in Childhood Prevents Lung Cancer

On Sat, 12 May 2007 21:11:52 +0000 (UTC), DZ
<15137@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:

Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx> wrote:
DZ wrote:
[snip]
ru > >
when most odds ratios for potential 'intervening factors' -- ones
that might account for the coincidence -- are so much smaller.
DZ >
If you have any specific criticism about the genetic study (or my
interpretation of it) discussed in this thread, this would be a good
time to do so.

The study gave estimates that about 10% of population carry one or two
copies of an allele that increases lung cancer risk 12 times (for
heterozygote carriers), and 62.5 times (for homozygotes) in heavy
smokers. In non-smokers, the risk is increased 14.5 times (for
heterozygotes), and 2245.6 times (for homozygotes).

Okay. Specific criticism. For what you show above, the problem
seems to be in the numbers. If this is what the researchers are
reporting, then you are right, and the gene does account for a
large fraction of cases. And, in that case, I'm surprised that I
haven't heard of it.


I've tried to dummy up some numbers.
If 10% of the population of smokers have 12-fold risk,
then allmost 40% of the smokers who get lung cancer
ought to have the allele.

Total population 300.
100 smokers, 16 cancers for an overall 6%-of-300 rate.
90 without allele, 10 get cancer 80+10
10 with allele (1 homozygous), 6 get cancer 4+6
That's an odds ratio of 12 to 1, which is close, but ignores
the separate odds for homozygous.

These numbers are a little too big, since they overstate
the rate. But that is a large number of gene-related cancers.


200 Nonsmokers: 2 are homozygous; both probably
get lung cancer. Are the alleles assorted by 'race' so there
would be more?
Maybe there are 2 cases among the 180 non-allele non-smokers
and another 2 cases among the heterozygous -- 6
cases among nonsmokers, at least 4 with the allele.


That is a strong enough genetic effect that family
studies ought to find it. I have no reason to think
that "randomizing twins" to conditions - where you started-
should be more revealing that the observational data.

Is this what they are reporting, numbers like those above?
Has this gene been identified?

--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.

Relevant Pages

  • Re: OT:Thanksgiving
    ... And smoking tobacco doesn't cause cancer either, ... which reports smokers got lung cancer ... The confounder is poverty, thus the high fat diet of poor people. ...
    (comp.lang.cobol)
  • Re: OT:Thanksgiving
    ... That experiments do not work on animals suggests a confounder unique to ... with lung cancer. ... One pack a day ... smokers get the most lung cancer; three pack a day smokers are back to the ...
    (comp.lang.cobol)
  • Re: OT:Thanksgiving
    ... And smoking tobacco doesn't cause cancer either, ... which reports smokers got lung cancer ... The confounder is poverty, thus the high fat diet of poor people. ...
    (comp.lang.cobol)
  • *Article: Smokers sue tobacco firm for cancer scans
    ... A GROUP of long-term American smokers are suing tobacco giant Philip Morris - demanding not money but hi-tech screenings to catch lung cancer early. ... "Philip Morris is obligated morally, not just legally, to make a health investment to help prevent death and suffering caused by the Marlboro smoke," the smokers' attorney Jerome Block said. ...
    (alt.support.stop-smoking)
  • Re: Dom is a cock
    ... simply sticking up for the rights of smokers! ... Also, people can obviously die of lung cancer without having smoked, so any ... Aside from smoking, there are other ways to contract ... provable link but I've perused the BMJ a few times ...
    (uk.media.tv.misc)
Quantcast