REVIEW 673 CONFIDENTIA
    Subiect ref, 8d
    "~!s passive smoking a cause of asthma in childhood?"
    R Z*-rlich, 4 Kat:zan and D E Lilienfeld
    Zournal of Smoking Related Diseases (19911), !~, 91-99
    40 This review paper concludes chat "there is growingevidence,
    although not wholly consistent,  that maternal smoking is a risk factor
    for asthma and whee---ms illness in childhood ... sufficient evidence for
    public health purposes to target maternal smoking as a modifiable risk
    factor for -6iheezing i
    f  illness in childhood".
    Altho~;zh the reference list is extensive, with 100 papers cited, the
    - -*ew is actually not very impressive. It is really too short for such a
    ,ev.
    complex subject.
    One is given no information as to whether any of the cited studies
    have any s--*Pificant weaknesses (or strengths), and there areno
    relevant s-=mary tables to allow any impression to be gained of the
    overall data, let alone results of any. meta-analyses to allow some
    quantitative estimate of the overall association from the studies
    considered.
    The statements in the conclusion section also do not seem to arise
    naturally fram the preceding text. For example, the conclusion section
    starts by noting that "there is reasonable consistency among the cc
    C1%
    co
    r~
    BATCo document for Legal Services : Health Canada 19 October 1999

    




    -2-
    studies", a fact seems to belie -the existence of incons is tent
    findings relat-ing t:; -.--:-.-.;ally all the zy-es of studies considered
    (general population szudiss, studies of astl-Mlatics, clinical laboratory
    studies). It alss s::s:es  'I
    the conclusion secticn that "an
    exposure -response re'ac'onship has been demonstrated", but the preceding
    discussion did not. as far as I could see, mention dzse-resporse
    relationships a= a!*-.
    In the discuss-on s--ction =Inere is men=-on of a number of potential
    sources of confounding, buz from what s presented it is 4_:npossible to
    judge how much fa4_'_-_-re :o take confounders into account might have
    affected conclusions from the va:4-ous studies.
    Generally. _: re=a-ns unclear from the review ho-.: much the
    associazion observed bat~een parental smoking and aschma in some, but not
    all, studies (mors o---:=-n in relation to maternal smoking, the author
    claims) is due to the var-ous possible explanations, namely ETS exposure,
    maternal smoking in ;regnancy, cross-infection, confounding by various
    factors, and c1hance.
    In view of the cla-ms of the recent EFA report that ----7S exposure
    worsens the conditicn. of _M,000 to.1,000,000 asthmatic children, and is
    a risk factor for 8,000 to 26,000 new cases of asthma a year in children
    who have not previously displayed symptoms, the whole area demands more
    serious attention. I 'aalieve there would be value in conducting a
    detailed review of the e%idence. Such a review, which wculd need to
    Uri
    involve both a s=azist-c-ari/epidemiologist an-- a clinician fa.-.iliar with CD
    CD
    co
    C\
    oc~
    r\J
    110
    BATCo document for Legal Services : Health Canada 19 October 1999

    




    asthma, would seek to do mors than si=?Iy list studies reporting and not
    reporting associations - it should see'.; to g---.,e a more quanritacive view
    of the overall data, taking accounz of the magnitude of the various
    sources of bias, and gi,.a a cric-cal assessment of the individual
    studies. Because of the extensive nature of' the evidence, this would be a
    major exercise.
    P N Lee
    11.11.93
    0
    Co
    CI\
    BATCo document for Legal Services : Health Canada 19 October 1999

    




    -5-
    in asthma score and lung function over the period in children of
    smokers, it anDears to be mainly -due to reasons other Chan reduce
    smoking bv the Darents in the child's Presence. It is notable in
    Table 11L -that t7ne:e are quite substantial improvements in lung
    function (though -.~t in asthma score) over the time period in
    children whose parents do not smoke.
    Generally, the pa;er is unconvincing. and parts of the data
    inconsistent wizin t~he authors' hypothesis are nor. really brought to the
    reader's attenzion.
    P N Lee
    11.11.93
    ui
    C)
    CD
    CN
    Co
    CI',
    BATCo document for Legal Services : Health Canada 19 October 1999
    M