Personal exposure assessment in the epidemiology of air pollutants

Document Sample
Personal exposure assessment in the epidemiology of air pollutants Powered By Docstoc
					458                                                                                                            Occup Environ Med 2003;60:458–460

PostScript                           ..............................................................................................

                                                  at that time, predicted more than twice as              Medicine,1 Kromhout and van Tongeren ad-
                LETTERS                           many mesotheliomas by 2020 than an elimi-               monish us for paying insufficient attention to
                                                  nation rate of 15% per year. Preliminary                the earlier literature on occupational pollutant
                                                  results to 1999 were given by Musk and                  exposures.2 While no doubt an element of
                                                  colleagues,7 and it is now known that the               their criticism is justified, we feel that the
 If you have a burning desire to respond to       number of mesothelioma deaths in men in                 exposure situation for the general public is
 a paper published in Occupational and            the period 1987 to 2000 was similar to the              sufficiently different that it should not be
 Environmental Medicine, why not make             lowest predictions made based on the number             assumed that findings in the occupational
 use of our “rapid response” option?              up to 1986. This result is evidence that models         environment can necessarily be extrapolated
    Log      on       to       our     website    of mesothelioma incidence that take account             to environmental exposures of the general
 (, find the paper              of a gradual elimination of crocidolite from            public. A large component of environmental
 that interests you, and send your response       the lungs after exposure are more realistic.            exposure arises from diffuse sources and may
 via email by clicking on the “eLetters”          There is strong evidence from other sources             therefore be very spatially homogeneous at
 option in the box at the top right hand          that such elimination does occur and that for           locations such as people’s homes which are
 corner.                                          crocidolite the rate of elimination is in the           often relatively remote from outdoor pollution
    Providing it isn’t libellous or obscene, it   range of 10–15% a year.8                                sources.
                                                     There is a high continuing toll from the use            There has been some controversy in the lit-
 will be posted within seven days. You can
                                                  of asbestos in Europe and from the mining               erature regarding the extent to which meas-
 retrieve it by clicking on “read eLetters”       and milling of crocidolite at Wittenoom in              urements at fixed central urban background
 on our homepage.                                 Australia, but fortunately recent evidence              monitoring locations can reflect the exposures
    The editors will decide as before             strongly suggests that the number of mes-               of large urban populations who spend much
 whether to also publish it in a future           otheliomas will not be as high as earlier               of their time indoors at locations relatively
 paper issue.                                     predictions.                                            remote from the monitoring station.3 4 It has
                                                                                                          been typical to find that for an individual,
                                                                                             G Berry      daily personal exposures correlate with con-
                                                       School of Public Health, University of Sydney,     centrations at the monitoring station, while if
                                                                 New South Wales 2006, Australia;
Predictions of mortality from                                  
                                                                                                          data are pooled from many individuals, the
                                                                                                          exposures appear to be uncorrelated with
mesothelioma                                                                             A W Musk         ambient air data.5 6 This finding suggests that
The update of predictions of mortality from                Sir Charles Gairdner Hospital, Nedlands,       the diffuse background as represented by the
pleural mesothelioma in the Netherlands1                                         Western Australia        central urban monitor does account for a sub-
provides welcome news that the peak number                                                                stantial proportion of variance in the expo-
and the total during 2000–28 are now                                                  N H de Klerk
                                                                        Department of Public Health,
                                                                                                          sure of an individual, and this conclusion is
predicted to be only a little more than half of                                                           supportive of causality in the time series epi-
                                                                      University of Western Australia
the figures predicted only four years earlier.2                                                            demiological studies, which would appear
This marked change in prediction has oc-                                  A Johnson, D H Yates            implausible if the monitoring data were unre-
curred because the known decrease in asbes-                Dust Diseases Board (NSW) Research and         lated to human exposures. The finding of our
tos use after 1984 and a ban in 1993 were                                    Education Unit, Sydney       paper that microenvironment measurements
taken into account in the modelling, and                                                                  do, in general, well represent individual
there were five extra years of data (1994–98).
                                                    1 Segura O, Burdorf A, Looman C. Update of
                                                                                                          personal exposures in that microenvironment
   Since most mesotheliomas are caused by                                                                 (except for the personal cloud of PM10) is far
                                                      predictions of mortality from pleural
asbestos the pattern of use during different          mesothelioma in the Netherlands. Occup              from self-evident from much of the earlier lit-
periods of time has a marked influence on the          Environ Med 2003;60:50–5.                           erature and is a useful addition to knowledge.
risk in cohorts whose working lives covered         2 Peto J, Decarli A, La Vecchia C, et al. The         The fact that cigarette smokers were outliers
different periods. The marked effect of the           European mesothelioma epidemic. Br J                in the regression analysis shows not unex-
discontinuation of crocidolite importation by         Cancer 1999;79:666–72.                              pectedly that they generate strong local
1970 into the United Kingdom on the amount          3 McDonald JC, Armstrong BG, Edwards CW,              concentration gradients and would therefore
                                                      et al. Case-referent survey of young adults
of crocidolite found in the lungs of men born                                                             need to be treated differently in any modelling
                                                      with mesothelioma: I. Lung fibre analyses. Ann
in 1943 or later, who developed a mesothe-            Occup Hyg 2001;45:513–18.                           of personal exposures. In the absence of such
lioma between 1990 and 1996 when aged               4 Berry G. Asbestos lung fibre analysis in the        local sources of pollution, our study supports
36–52 years,3 was clear when compared with            United Kingdom, 1976–96. Ann Occup Hyg              the concept that were sufficient microenvi-
the lung contents of mesotheliomas in 1976            2002;46:523–6.                                      ronment measurement data available, it
and 1977.4 The average amount of crocidolite        5 Berry G. Prediction of mesothelioma, lung           would be perfectly feasible to model personal
in the lungs of the men with mesothelioma in          cancer, and asbestosis in former Wittenoom          exposures with some degree of reliability.
1990–96 was about a tenth of that in the lungs        asbestos workers. Br J Ind Med
                                                                                                             Kromhout and van Tongeren advocate the
of the 1976–77 cases. This difference was a         6 de Klerk NH, Armstrong BK, Musk AW, et al.          use of personal exposure measurements in
consequence of the shorter period of exposure         Predictions of future cases of asbestos-related     environmental epidemiological studies. In
and the longer time since exposure during             disease among former miners and millers of          doing so, they fail to acknowledge the magni-
which elimination of fibres occurred in the            crocidolite in Western Australia. Med J Aust        tude of such studies. For example, in the large
absence of new exposure.                              1989;151:616–20.                                    North American cohort studies, 8111 subjects
   The importance of the effect of elimination      7 Musk AW, de Klerk NH, Olsen NJ, et al.              were recruited in the Harvard Six Cities Study
is illustrated by results from former workers         Mortality in miners and millers of crocidolite in   and over one million in the American Cancer
at the Wittenoom crocidolite mine and mill in         Western Australia: follow-up to 1999. Ann
                                                                                                          Society Study. Were it possible to reconstruct
                                                      Occup Hyg 2002;46(suppl 1):90–2.
Western Australia. Exposure of this group           8 Berry G. Models for mesothelioma incidence          the exposure environments of those individu-
ceased in 1966 when the mine and mill were            following exposure to fibers in terms of timing     als, even in a rather general way from time
closed. Based on an analysis of the number of         and duration of exposure and the                    activity diaries, a considerable refinement
deaths with mesothelioma in men to the end            biopersistence of the fibers. Inhal Toxicol         would have been achieved. Even in panel
of 1986, predictions were made of the number          1999;11:111–30.                                     studies, which typically recruit a far smaller
occurring up to 2020.5 6 The model of the mes-                                                            number of individuals, the subjects are
othelioma death rate used was that the rate                                                               frequently drawn from susceptible groups and
increased with a power of time since expo-        Personal exposure assessment in                         therefore not willing to be encumbered with
sure, moderated by a factor representing          the epidemiology of air                                 troublesome and heavy sampling equipment.
elimination of crocidolite fibres over time                                                                It must be remembered that concentrations in
since exposure. Rates of elimination from zero    pollutants                                              environmental samples are typically orders of
to 15% per year were considered. Assuming no      In commenting on our paper published                    magnitude lower than in occupational sam-
elimination, which was the usual model used       recently in Occupational and Environnmental             ples, therefore requiring higher flow rates
PostScript                                                                                                                                            459

(hence bigger pumps) and longer sampling              107 (101 to 113), and a respiratory disease        intervals overlap between strata, so while
intervals. In some instances it may be possible       SMR of 120 (101 to 141). Increased mortality       there is not a significant increase, there is no
to use passive samplers such as diffusion             of similar magnitude from these causes was         inverse latency response relation. In addition,
tubes and badges, but these are typically             observed among the smaller number of               much the largest portion of this long latency
rather imprecise and are available only for a         women, and the SMRs for both genders com-          group must have come from two of the 11
very limited range of pollutants.                     bined were significantly increased. Raised          plants (factories 3 and 4 in table 1) where
   In summary, therefore, while the measure-          SMRs for all-cause and respiratory disease         exposures may have been different to those in
ment of personal exposure in environmental            mortality are hardly ever seen in occupational     the other nine facilities.
epidemiology is highly desirable, it is in reality    cohorts except for foundry and asbestos               In summary, this study has found a highly
very unlikely to be practicable in most studies.      workers. Typically, the SMR for all-cause mor-     unusual and statistically significant increase
Thus, our study of the feasibility of recon-          tality is about 80, and the SMR for most can-      in all cause mortality and respiratory disease
structing exposures from microenvironment             cer causes about 90 in the absence of exposure     mortality in the cohort as a whole, consist-
data is well justified and has thrown useful           to a carcinogen at the site.5 We have observed     ently in both men and women. For women
light on the problem, for example, in illustrat-      SMRs for all-cause mortality as low as 60 in       alone, and men and women combined, there
ing the lower exposures of members of certain         UAW vehicle assembly and stamping
                                                                                                         was significantly increased mortality for lung
of our susceptible groups.                            cohorts.6 We are surprised that these authors
                                                                                                         cancer, and for both genders combined,
                                                      mentioned a deficit for all-causes in the
              R M Harrison, R P Kinnersley,                                                              pancreatic cancer. This is certainly “some evi-
                                                      abstract of their previous study of this cohort,
                             R G Lawrence             but make no mention of the excess in the           dence” for work related mortality, respiratory
          Division of Environmental Health and Risk   present paper.7 For lung cancer, the authors       disease mortality, and cancer mortality in the
           Management, University of Birmingham,      noted a significant SMR of 181 (126 to 251)         exposure circumstance of polyurethane foam
             Edgbaston, Birmingham B15 2TT, UK;       for lung cancer among women. They discount         production. The nature of the cohort gave
                 this partly because the SMR of 107 (90 to 227)     little prospect for observing an exposure
                                                      among men was only slightly increased com-         response relation, if it were there.
                                       J G Ayres
                                                      pared to the general population, without also         We also note that, unlike all the other
                            University of Aberdeen
                                                      noting that the combined SMR was 117 (101          papers in this edition of the journal, these
                                         D Mark       to 136) and statistically significantly in-         authors have neglected to acknowledge their
                        Health & Safety Laboratory    creased. The authors also fail to mention that     funding source.
                                                      the SMRs for pancreatic cancer were in-
                                                      creased to a similar degree in both genders,                                             F E Mirer
                                                      and the combined SMR was 147 (102 to 212)                     Director, Health and Safety Department,
  1 Harrison RM, Thornton CA, Lawrence RG, et
    al. Personal exposure monitoring of particulate
                                                      and significantly increased. We believe that           International Union, UAW, 8000 East Jefferson
    matter, nitrogen dioxide, and carbon              consistency in direction of effect is more          Avenue, Detroit, MI 48214, USA;
    monoxide, including susceptible groups.           important than statistical significance, espe-
    Occup Environ Med 2002;59:671–9.                  cially in view of the healthy worker effect bias
  2 Kromhout H, van Tongeren M. How                   against seeing an effect if it were there.         References
    important is personal exposure assessment in         These findings apply to an exposure cir-
    the epidemiology of air pollutants? Occup                                                              1 Sorahan T, Nichols L. Mortality and cancer
                                                      cumstance with several suspect agents. The             morbidity of production workers in the UK
    Environ Med 2003;60:143–4.                        isocyanates are most prominently associated
  3 Linaker CH, Chauthan AJ, Inskip HM, et al.                                                               flexible polyurethane foam industry: updated
    Personal exposures of children to nitrogen
                                                      with non-malignant respiratory disease.                findings, 1958–98. Occup Environ Med
    dioxide relative to concentrations in outdoor     Therefore, the increase in mortality from this         2002;59:751–8.
    air. Occup Environ Med 2000;57:472–6.             cause is of distinct interest. In addition,          2 McMichael AJ. Standardized mortality ratios
  4 Mark D, Upton SL, Lyons CP, et al. Personal       pancreatic cancer was noted in gavage studies          and the “healthy worker effect”: scratching
    exposure measurements of the general public       of toluene diisocyanate.8 In our experience,           beneath the surface. J Occup Med
    to atmospheric particles. Ann Occup Hyg           the most substantial exposure with carcino-            1976;18:165–8.
    1997;41(suppl 1):700–6.                           genic risk in foam moulding is methylene             3 Schnorr TM, Steenland K, Egeland GM, et al.
  5 Janssen NAH, Hoek G, Brunekreef B, et al.                                                                Mortality of workers exposed to toluene
                                                      chloride,9 although brominated and chlorin-
    Personal sampling of particles in adults:                                                                diisocyanate in the polyurethane foam
    relation among personal, indoor and outdoor
                                                      ated alcohol flame retardants, and formalde-            industry. Occup Environ Med
    air concentrations. Am J Epidemiol                hyde are usually present in foam moulding              1996;53:703–7.
    1998;147:537–47.                                  operations. Catalyst amines may also be              4 Hagmar L, Welinder H, Mikoczy Z. Cancer
  6 Wallace L. Correlations of personal exposure      absorbed through the skin in physiologically           incidence and mortality in the Swedish
    to particles with outdoor air measurements: a     significant amounts.10 These multiple expo-             polyurethane foam manufacturing industry. Br
    review of recent studies. Aerosol Sci Technol     sures, often in different parts of the process,        J Ind Med 1993;50:537–43.
    2000;32:15–25.                                    including off-gassing from stored foam, un-          5 Park RM, Maizlish NA, Punnett L, et al. A
                                                      dermine the ability to see an effect of                comparison of PMRs and SMRs as estimators
                                                      isocyanates alone.                                     of occupational mortality. Epidemiology
                                                         We now turn to the exposure response por-           1991;2:49–59.
Mortality results for                                                                                      6 Park R, Krebs J, Mirer F. Mortality at an
                                                      tion of the study. Health related termination
                                                                                                             automotive stamping and assembly complex.
polyurethane manufacture                              of exposure has previously been noted as an            Am J Ind Med 1994;26:449–63.
                                                      obstacle to finding an exposure response              7 Sorahan T, Pope D. Mortality and cancer
understated                                           effect based on duration.11 Those with highest         morbidity of production workers in the United
Sorahan and Nichols,1 writing in this journal,        exposure to isocyanates would be expected to           Kingdom flexible polyurethane foam industry.
incorrectly understate the strength of evi-           be sensitised and migrate into lower exposure          Br J Ind Med 1993;50:528–36.
dence for work related increased mortality            jobs; in any event, there were only 19 of 1652       8 National Toxicology Program. TR-251:
among their cohort of production workers in           deceased workers with more than five years in           toxicology and carcinogenesis studies of
the UK flexible polyurethane foam industry.            higher exposed jobs, no lung cancer victims,           commercial grade 2,4 (80%)- and 2,6 (20%)-
Their study actually found “some” evidence            and only two respiratory disease victims. In           toluene diisocyanate (CAS No. 26471-62-5)
                                                                                                             in F344/N rats and B6C3F1 mice (gavage
for a work related increase in all-cause              our view, the absence of an exposure response          studies). 1986.
mortality, respiratory disease mortality, and         relation in a cohort with such a small higher        9 NIEHS. TR-306: toxicology and
lung cancer mortality in this exposure cir-           exposed group detracts little from our concern         carcinogenesis studies of dichloromethane
cumstance, especially taking into account the         for occupational cause of an observed excess.          (methylene chloride) (CAS No. 75-09-2) in
healthy worker effect.2 We are concerned to              More damaging to the evidence of occupa-            F344/N rats and B6C3F1 mice (inhalation
correct this error, because the United Auto-          tional causation is the absence of a monotonic         studies). 1986.
mobile, Aerospace and Agricultural Imple-             increasing trend with latency from first expo-       10 Baker EL, Christiani DC, Wegman DH, et al.
ment Workers of America (UAW) represents              sure: The general trend of increased risk in           Follow-up studies of workers with bladder
substantial numbers of workers exposed to             exposure strata greater than 10 years latency,         neuropathy caused by exposure to
                                                                                                             dimethylaminopropionitrile. Scand J Work
this process, and the UK data provide the first        clearly significant for all-cause mortality, is
                                                                                                             Environ Health 1981;7(suppl 4):54–9.
evidence of a mortality hazard in this indus-         not seen in those with greater than 30 years        11 Steenland K, Deddens J, Salvan A, et al.
try, in contrast to two previous, perhaps             latency. However, we note that the all-cause           Negative bias in exposure-response trends in
weaker studies.3 4                                    and respiratory disease SMRs are at unity or           occupational studies: modeling the healthy
   The authors observed an all-cause stand-           above for this long latency strata, itself a           workers survivor effect. Am J Epidemiol
ardised mortality ratio (SMR) among men of            highly unusual observation. The confidence              1996;143:202–10.


Shared By: