Learning Center
Plans & pricing Sign in
Sign Out

Routes of asbestos exposure and the development of mesothelioma


									Occupational and Environmental Medicine 1997;54:403-409                                                                       403

                              Routes of asbestos exposure and the development
                              of mesothelioma in an English region

                              Denise Howel, Lorna Arblaster, Layinka Swinburne, Martin Schweiger, Edward Renvoize,
                              Paul Hatton

                              Abstract                                           30-40 years.' The death rate is expected to
                              Objectives-To investigate the contribu-            rise in Britain over the next 15-25 years.4
                              tion of exposure to asbestos through               Although the risks of developing the disease
                              different routes in the development of             from occupational exposure to asbestos are
                              mesothelioma.                                      well established, those due to non-occupational
                              Methods-Case-control study. 185 con-               exposure are of increasing concern. This study
                              firmed cases of mesothelioma and 160               investigated the contribution of the different
                              controls were identified, when death had           routes of exposure to asbestos (table 1) in the
                              occurred between 1979 and 1991 in four             development of mesothelioma; concentrating
                              health districts in Yorkshire. The surviv-         on occupational, paraoccupational, and resi-
                              ing relatives were interviewed to ascertain        dential exposure. Over recent years, there has
                              lifetime exposure to asbestos. Adjusted            been public concern about the relatively high
                              odds ratios (ORs) of exposure to asbestos          incidence of mesothelioma locally5: occupa-
                              (through occupational, paraoccupational,           tional risks in three factories which used asbes-
                              and residential routes) were calculated for        tos in Leeds, Calderdale, and York have been
                              cases and were compared with controls.             highlighted by the media. This study arose
                              Results-Likely or possible occupational            from further publicity about cases labelled as
                              exposure to asbestos was more common in            environmental by Her Majesty's Coroner or
                              cases than in controls (OR 5.6, 95% confi-         the local media and which were said to be
                              dence interval (95% CI) 3.1 to 10.1). After        linked to the Leeds factory. This factory closed
                              excluding those with likely or possible            in 1958, but considerable concern remains
                              occupational exposure, likely or possible          about continuing residential exposure to asbes-
                              paraoccupational exposure was more                 tos in the vicinity of the factory.6 There has
                              common in cases than controls (OR 5.8,             been recent legal action by some former local
                              95% CI 1.8 to 19.2). Only six cases of mes-        residents who have developed mesothelioma
                              othelioma were identified as being solely          without occupational exposure to asbestos.
                              exposed to asbestos through their resi-            However, many local factories used asbestos,
                              dence, compared with nine controls. The            and this study considers exposure to asbestos
Department of                 OR for residential exposure to asbestos            from all potential sources in the study area.
Epidemiology and              varied between 1.5 and 6.6, depending on
Public Health,                which potential industrial sources were
University of                 included, but the 95% CIs were so wide             Subjects and methods
Newcastle upon Tyne           that slightly reduced or greatly increased         STUDY SUBJECTS
D Howel                                                                          Potential cases of mesothelioma were sought
                              odds comparing cases with controls could
                              not be excluded.                                   from the Health and Safety Executive's Na-
Formerly: Department                                                             tional Mesothelioma Register, the Yorkshire
of Public Health,             Conclusion-Study results support previ-
University of Leeds           ous evidence that occupational and                 Regional Cancer Registry, and local postmor-
L Arblaster                   paraoccupational exposure to asbestos is           tem records, where a previous diagnosis of
                              associated with developing mesothelioma.           mesothelioma had been made by a local
St James's University                                                            pathologist. Subjects who had died from
Hospital, Leeds               Despite a rigorous search, purely residen-
                              tial exposure seemed to account for only           mesothelioma between January 1979 and
L Swinburne                                                                      December 1991 were considered for inclusion,
                              3% of identified cases. No firm conclusion
Leeds Health                  can be drawn about the risks from                  if the address at death was in the Yorkshire dis-
Authority, Blenheim           residential exposure alone, as many of the         tricts of Leeds, Calderdale, York, Wakefield, or
House, Leeds                  study subjects could also have been occu-          Pontefract: these are largely urban areas in the
M Schweiger                                                                      north of England. Two pathologists then inde-
P Hatton                      pationally or paraoccupationally exposed
                              to asbestos.                                       pendently examined available histological sec-
United Leeds Teaching                                                            tions from each potential case. Results were
Hospitals, Leeds              (Occup Environ Med 1997;54:403-409)                classified into definite or possible malignant
E Renvoize                                                                       mesothelioma, or other tumour: only cases
                              Keywords: malignant mesothelioma; asbestos; non-   agreed to be definite mesothelioma were stud-
Correspondence   to:
                              occupational exposure                              ied further.
Denise Howel, Department
of Epidemiology and Public                                                          Necropsy records were used to identify con-
Health, The Medical School,                                                      trols for the cases. Sets of cases and controls
Framlington Place,            Malignant mesothelioma is an uncommon              were matched for sex, age at death (to within
Newcastle upon Tyne NE2       cancer, often associated with exposure to          10 years), and year of death (to within two
                              asbestos, with a latent interval between first     years): the sets ranged in size from one case
Accepted 18 December 1996     exposure and tumour development averaging          matched with one control, to six cases matched
404                                                                   Howel, Arblaster, Swinburne, Schweiger, Renvoize, Haton

      Table 1 Routes of asbestos exposure
      Route                                 Description of exposure
      Occupational                          Takes place at work
      Paraoccupational (domestic)           Other people working with asbestos who shed it from their clothes or person
      Incidental exposure                   Through hobbies or visits
      Residential exposure                  Living near a particular source of asbestos (mine or factory)
      General environmental exposure        In the environment without a particular source

      with five controls. The choice of matched sets                  determined the occupational codes of the sub-
      rather than matched pairs was considered to                     jects in the study. Also, the occupations of
      make best use of scarce subjects, when some                     members of the subject's household were simi-
      will be dropped from later analyses. Potential                  larly coded, and the subject was assigned an
      controls were excluded if they had died from                    overall code which reflected the highest level of
      mesothelioma, or diseases which could have                      paraoccupational exposure. Any incidental
      been confused with mesothelioma (bronchial                      exposure to asbestos-for example, home
      or ovarian cancer), or in circumstances in                      improvements-was coded likewise.
      which gathering information would be difficult                     Factories active from 1900 onwards in north
      (suicide; or if they had been homeless at the                   and west Yorkshire considered to use asbestos
      time of death; or if they had spent most of their               and which might have emitted asbestos into the
      adult life in an institution). This did not give                environment-for example, through loading,
      rise to many exclusions. Specimens of non-                      ventilation, or waste disposal-were identified
      tumorous lung tissue were obtained when                         through trade directories and local knowledge.
      available for cases and controls, and were sent                 Altogether 278 were identified: many were only
      for mineral fibre analysis. These results will be               active for short periods, and most were within
      discussed elsewhere.                                            the Leeds or Bradford area. Only 83 of these
         Experience suggested that about 300 cases of                 factories ever had study cases or controls living
      mesothelioma would be identified in the study                   within 0.5 km of them. The factories were clas-
      period and area; however, we could only obtain                  sified as either group 1 (those where goods
      both histological material to confirm the diag-                 containing asbestos were known to be
      nosis and a later exposure history on half of                   manufactured-for example, engine packing,
      these. Assuming one third of these cases had                    brake and clutch linings, joints and gaskets,
      not been exposed to asbestos through occupa-                    railway carriages, or insulation materials) or
      tional or paraoccupational routes, there was an                 group 2 (those where this was in some doubt-
      80% power to detect an odds ratio (OR) =3                       that is, the site may have been a warehouse, or
      linking residential exposure and mesothelioma                   asbestos may not have been the material used).
      at a significance level of 5%.                                  Appendix 2 shows the activities at these facto-
      EXPOSURE HISTORIES                                                 A residential history, including schools
      Attempts were made to contact the next of kin                   attended, was obtained at interview. The
      of subjects, and a semistructured questionnaire                 addresses of subjects within Yorkshire and
      was used to interview them to obtain a detailed                 potential industrial sources were converted,
      life history of the deceased. It was explained                  where possible, into the grid references ap-
      that we were seeking information about the                      proximating to current postcodes. Distances
      deceased's life history, and they were encour-                  were calculated between every residential
      aged to ask other family members or friends for                 address and every potential source, and then
      further details. Interviewees were unaware of                   the residential exposure of a subject was coded
      the main concerns of the study: only the final                  as: likely if residence was within 0.5 km of one
      few questions mentioned exposure to asbestos                    or more sources at some time; and unlikely if it
      explicitly. When it was not possible to interview               was not known whether residence was within
      relatives, information was sought from coro-                    0.5 km of any of these sources during life.
      ner's records. Permission for the study was                        Residential exposure was coded firstly for all
      obtained from local ethics committees.                          potential sources, and secondly for those where
         Coding of exposure histories was performed                   manufacture of asbestos goods was more
      blind to case-control status, and was based on                  certain (group 1 only). The figure shows the
      responses obtained before asking explicitly                     grid references both of addresses where
      about asbestos exposure. It was assumed that                    subjects lived and the factories which were
      asbestos exposure in the last 15 years of life (by              potential sources of asbestos to the environ-
      whatever route) would not have contributed to                   ment.
      the development of mesothelioma, and this
      period was excluded when coding information.                    STATISTICAL ANALYSIS
      Occupational and paraoccupational exposures                     The main technique used for the analysis of the
      of less than one month were excluded, as were                   comparative risks of exposure was conditional
      residential exposures of less than one year.                    logistic regression.8 Occupational and paraoc-
         Occupations were coded with an existing                      cupational exposure were coded to three levels
      classification' as likely, possible, or unlikely to             -likely, possible, and unlikely. The analyses
      have led to exposure to asbestos. Each subject                  were carried out with two sets of groupings.
      was then assigned an overall code, which was                    Firstly, by comparing likely and possible with
      the highest level of occupational exposure ever                 the unlikely group, and secondly, by comparing
      experienced. Appendix 1 lists the jobs which                    likely with the possible and unlikely groups.
Routes of asbestos exposure and the development of mesothelioma                                                                             405

       4800 r

       4700 H              0 Subjects' addresses                                                        0
                           + Industrial sources of asbestos
       4600 F-
                                                                                          90                                York
       4500 F-
       4400 e
       4300                                                                                             0


       4100                                                  Huddersfield

                    3900           4000           4100            4200             4300    4400         4500        4600       4700        4800
                                                                         Grid reference east
Position of subject addresses and potential industrial sources of asbestos
                                 The conditional logistic model incorporated                   These were divided into 55 sets of cases and
                                 terms to further adjust for the age, year of                  controls. Table 2 shows the distribution of age,
                                 death, and district; as the matching criteria                 sex, and place of death for cases and controls.
                                 were quite wide. The effects of exposure to
                                 asbestos by a particular route was evaluated by               COMPARING EXPOSURE TO ASBESTOS IN CASES
                                 successively excluding subjects who had been                  AND CONTROLS FOR EACH EXPOSURE ROUTE
                                 exposed to asbestos by routes which were likely               Overall, 103 (56%) of the cases were classified
                                 to be at higher levels-for example, excluding                 as likely to have been occupationally exposed to
                                 occupational exposure while considering                       asbestos, compared with 22 (14%) of the con-
                                 paraoccupational exposure. The same matched                   trols. Inclusion of those who had possibly been
                                 sets were used for all analyses, but exclusions               exposed, gives corresponding figures of 150
                                 were made as already described: if all cases in a             (81%) and 80 (50%). The source of infor-
                                 set were excluded for this reason, the controls               mation about occupational history had little
                                 in the set were not used.                                     association with the exposure category. Among
                                                                                               those whose work history came from interview
                                 Results                                                       55% had likely occupational exposure, 26%
                                 There were 316 potential cases of mesothe-                    possible, and 19% unlikely, with corresponding
                                 lioma identified: 71 of these could not be con-               proportions in the cases in which the infor-
                                 firmed, it was agreed that 15 were not                        mation came from coroner's records of 58%,
                                 mesothelioma, and four were possible meso-                    25%, and 17%. In most matched sets the cases
                                 thelioma, leaving 226 confirmed cases. The                    had the same or a greater chance of being
                                 relatives of 133 of these cases were interviewed              occupationally exposed than the controls: there
                                 and information on a further 52 cases was                     were 41 sets where there were a higher
                                 obtained from the coroners' records. Interview                proportion of cases than controls who had been
                                 information was obtained for all 159 controls.                likely or possibly occupationally exposed, 11
                                                                                               sets where the proportions were equal, and
Table 2 Characteristics of cases and controls                                                  three sets where there was a higher proportion
                                                                                               of controls. Table 3 shows the condtional logis-
                                             Cases (n=185)            Controls (n=159)
                                                                                               tic regression results, and show that the odds
Age at death (median(range))                 66 (38-89)               67 (38-85)               on having occupational exposure were consid-
Men (n(%))                                   137 (74)                 118 (74)                 erably higher in cases than controls.
Died in (n(%)):
  Leeds                                      135 (73)                 122 (77)                    It was considered inappropriate to compare
  Calderdale                                 22 (12)                  20 (13)                  paraoccupational exposure of subjects who
  Wakefield                                  19 (10)                  13 (8)                   have also been occupationally exposed to
  York                                       9 (5)                    4 (3)
                                                                                               asbestos. Two analyses have therefore been
406                                                                       Howel, Arblaster, Swinburne, Schweiger, Renvoize, Haton

      Table 3 Adjusted ORs of asbestos exposure and mesothelioma
                                                                                                                   OR (95% CI)
      Occupational exposure:
        Likely v possible and unlikely                                                                             9.1   (4.8 to 17.1)
        Likely and possible v unlikely                                                                             5.6   (3.1 to 10.1)
      Paraoccupational exposure:
        Excluding subjects with likely occupational exposure:
           Likely v possible and unlikely                                                                          5.6 (1.9 to 16.5)
           Likely and possible v unlikely                                                                          1.8 (0.87 to 3.6)
        Excluding those with likely or possible occupational exposure:
           Likely v possible and unlikely                                                                          61.7 (3.4 to 1104)
           Likely and possible v unlikely                                                                          5.8 (1.7 to 19.2)
      Residential exposure:
        Excluding subjects with likely occupational exposure or likely paraoccupational exposure:
           Likely v unlikely (using group 1 sources)                                                               1.7 (0.78 to 3.8)
           Likely v unlikely (using all sources)                                                                   2.0 (0.9 to 4.2)
        Excluding subjects with likely or possible occupational exposure and likely paraoccupational exposure:
           Likely v unlikely (using group 1 sources)                                                               1.9 (0.51 to 7.1)
           Likely v unlikely (using all sources)                                                                   1.5 (0.46 to 5.1)
        Excluding subjects with likely or possible occupational exposure and likely or possible paraoccupational
           Likely v unlikely (using group 1 sources)                                                               6.6 (0.86 to 50)
           Likely v unlikely (using all sources)                                                                   2.3 (0.54 to 9.7)

      done: excluding subjects with likely occupa-                          It was considered inappropriate to compare
      tional exposure, and excluding those with likely                   the residential exposure of cases and controls if
      or possible occupational exposure. Altogether                      they had also been exposed to asbestos by other
      45 matched sets remained, comprising 81 cases                      routes likely to have been at higher exposure
      and 124 controls, for the first analysis. Eighteen                 levels. Therefore three analyses were carried
      of these cases had likely and five had possible                    out: excluding subjects with likely occupational
      paraoccupational exposure: five of the available                   exposure or likely paraoccupational exposure;
      controls had likely and 16 had possible paraoc-                    excluding subjects with likely or possible occu-
      cupational exposure. Table 3 shows the ad-                         pational exposure and likely paraoccupational
      justed OR estimates. The odds of paraoccupa-                       exposure; and excluding subjects with likely or
      tional exposure were higher in cases than                          possible occupational exposure and likely or
      controls, whether the possibles were included                      possible paraoccupational exposure. Table 3
      with the likely or unlikely exposure groups.                       shows the adjusted OR estimates.
      However, when the possible group was com-                             Thirty five matched sets comprising 46 cases
      bined with the likely group, the 95% confi-                        and 96 controls were usable for the first analy-
      dence interval (95% CI) for the OR showed                          sis. Of the 46 cases, 25 had lived within 0.5 km
      that data were also consistent with equal odds                     of any potential source, and 20 in group 1 had
      of paraoccupational exposure.                                      lived within 0.5 kun of a potential source. The
         Only 27 matched sets, comprising 34 cases                       corresponding figures for the 1 17 controls were
      and 58 controls, were usable for the second                        32 and 24. There were only 17 matched sets
      analysis. Of these cases 13 had likely and four                    comprising 18 cases and 40 controls who were
      had possible paraoccupational exposure; three                      usable for the second analysis. Of the 18 cases,
      of the controls had likely and eight had possible                  eight had lived within 0.5 km of any potential
      paraoccupational exposure. By excluding a                          source, and six in group 1 had lived within 0.5
      further group of subjects who might have been                      km of a source. The corresponding figures for
      occupationally exposed, we should obtain a less                    the 69 controls were 13 and nine. There were
      biased estimate of the strength of the associ-                     only 14 matched sets comprising 14 cases and
      ation; but the estimate was much less precise,                     29 controls who were usable for the third
      because of the reduced numbers of subjects                         analysis. Of the 14 cases, six had lived within
      available. The point estimates of ORs sug-                         0.5 km of any potential source, and five in
      gested a much stronger association, but were                       group 1 had lived within 0.5 km of a potential
      consistent with ORs both little greater than                       source. The corresponding figures for the 56
      unity and very high indeed. So whereas the                         controls were nine and five.
      odds on paraoccupational exposure were                                The estimated OR on residential exposure to
      higher for cases of mesothelioma, we are                           asbestos comparing cases with controls rose as
      unable to say with any precision by how much                       each of the other sources of exposure was
      they were raised.                                                  excluded. However, the few subjects remaining
         Forty four cases and 43 controls reported                       without occupational and paraoccupational
      some activities which could be coded as possi-                     exposure do not allow us to estimate the
      ble incidental exposure. These included car or                     strength of the relation between residential
      motorbike repair, removing fireplaces or walls                     exposure and mesothelioma with sufficient
      at home, installing central heating or lagging                     precision to rule out either no association or a
      pipes, dismantling radiators, demolishing a                        strong association.
      coke oven, and unspecified home improvement
      tasks. There were three controls for whom this                     RELATIVE CONTRIBUTIONS OF THE DIFFERENT
      seemed to be the only asbestos exposure, and                       TYPES OF ASBESTOS EXPOSURE
      two cases and five more controls for whom only                    Altogether 45% of the exposure histories
      incidental and residential exposure were iden-                    collected in this study reported asbestos expo-
      tified. The vague exposure descriptions pre-                      sure by more than one route. It is impossible to
      cluded formal analysis.                                           be certain which was the dominant exposure
Routes of asbestos exposure and the development of mesothelioma                                                                                          407

Table 4 Main route of exposure for cases and controls (n(%))
Category            Description                                                                                      Cases        Controls    Total
Occupational        Likely or possibly occupationally exposed                                                         150 (82)     80 (50)    230 (67)
Paraoccupational    Likely or possibly paraoccupationally exposed and not occupationally exposed (likely or possible) 17 (9)       14 (9)      31 (9)
Incidental          Likely or possible incidentally exposed but not occupationally or paraoccupationally exposed
                    (likely or possible)                                                                                2 (1)       8 (5)      10 (3)
Residential         Likely residential exposure to any source but not occupationally, paraoccupationally, or
                    incidentally exposed (likely or possible)                                                           6 (3)      14 (9)      20 (6)
None or unknown     Subjects not thought to be exposed to asbestos by any of the above routes (this will include
                    subjects whose exposure history has missing information)                                           10 (5)      43 (27)     53 (15)
Total                                                                                                                 185 (100)   159 (100)   344 (100)

                                  route for each subject; nevertheless, a hierarchy  and 9.5 % of controls had lived in the vicinity of
                                  of exposure categories has been suggested. It is   an asbestos friction materials plant.7 These
                                  assumed that occupational exposure to asbes-       results are not comparable with each other, or
                                  tos has the highest level: lesser exposures are    with this study, because of the varying
                                  deemed to be paraoccupational, incidental,         definitions. Outdoor airborne asbestos concen-
                                  and residential. Subjects have been grouped        trations were found to be raised near a factory
                                  into five categories, and table 4 shows the        that made asbestos slate board, and near an
                                  number of cases and controls in each of these.     asbestos-cement plant,' '4 but in general, little
                                  Occupational exposure was the most common          is known about the likely increase in airborne
                                  route in both cases and controls. Only 5% of       asbestos concentrations near factories that
                                  cases seemed to have no exposure to asbestos,      used asbestos, or about the area over which this
                                  whereas the figure was 27% for the controls.       increase might be expected.
                                                                                        Given the established risks of occupational
                                  Discussion                                         and paraoccupational exposure, a major focus
                                  The hazards of occupational exposure to of our study was the hypothesis that living in
                                  asbestos have been documented for decades. the vicinity of a factory that used asbestos is a
                                  Wagner et aP first established the link between risk factor for mesothelioma. No study has
                                  working at, or in close association with, an looked at residential exposure so comprehen-
                                  asbestos mine and development of mesothe- sively. Many potential industrial sources of
                                  lioma, and subsequent papers have described asbestos were identified in the study area, as
                                  the risks associated with other occupations. Of well as the sources which provoked public and
                                  study cases 82% had likely or possible occupa- media concern. However, only a few sources
                                  tional exposure, compared with 50% of the were near the residences of study subjects, and
                                  controls. Comparisons with other case-control very few after the exclusion of subjects with
                                  studies are complicated by differing occupa- occupational and paraoccupational exposure.
                                  tional coding schemes, but there were similar There were more subjects with occupational or
                                  proportions of occupational exposure in the paraoccupational exposure than expected from
                                  cases and controls in a North American study other studies: this may be because there was
                                  (70% and 51% respectively),10 and fewer so genuinely more exposure by these routes in
                                  exposed in a London study (41% and 1 1 %). " Yorkshire in the study period, or because we
                                  Our study confirmed that the odds of having a obtained more complete exposure histories, or
                                  likely or possible occupational exposure to differences in definitions. We had assumed that
                                  asbestos are significantly higher in cases than one third of the cases would have been
                                  controls.                                          unexposed through occupational or paraoccu-
                                     It is generally accepted that there is a link pational routes, but the actual figure was 9.7%,
                                  between paraoccupational exposure to asbestos leaving relatively few cases available for that
                                  and mesothelioma. Case-control studies in part of the study that concentrated on residen-
                                  London," New York,'2 and North Americal0 tial exposure. No firm conclusion can therefore
                                  found a history of paraoccupational exposure be drawn from this study about the risks of
                                  to asbestos in more cases than controls, residential exposure, because of the small
                                  although the numbers were small. This study number of suitable subjects. The estimated OR
                                  supports previous evidence about paraoccupa- for residential exposure to asbestos varied
                                  tional exposure, but the size and precision of between 1.5 and 6.6, depending on which sub-
                                  the OR depends on whether we include those jects and sources were included, but the 95%
                                  possibly exposed with those likely or unlikely to CIs were so wide that slightly reduced or
                                  be paraoccupationally exposed.                     greatly increased odds comparing cases with
                                     The risk of residential exposure is less clear. controls could not be excluded.
                                  Some case-control studies have investigated           The 0.5 km radius from a potential indus-
                                  the role of residential exposure after excluding trial source of asbestos is arbitrary; it is a proxy
                                  other asbestos exposures. The London study for exposure from a particular source. It is not
                                  found 11 cases compared with five controls liv- assumed that airborne asbestos from the
                                  ing within 0.5 miles of a particular asbestos potential source is necessarily found at any
                                  factory." The New York study found one case residence within 0.5 km. Living close to a
                                  living 1.2 km from an asbestos factory, and no source may mean that the subject often walked
                                  controls nearby.'2 The North American study past it, or played in the factory yard, etc. Also,
                                  found that one case and four controls had lived the identified sources carried out a wide range
                                  within 20 miles of chrysotile mines.'0 A of industrial activities with asbestos; some were
                                  Connecticut study found that 9.2% of cases known to process large amounts of raw fibre,
408                                                         Howel, Arblaster, Swinburne, Schweiger, Renvoize, Hatton

      whereas others may only have stored asbestos             This case-control study has brought together
      goods. The use of trade directories and local         detailed histories of asbestos exposures, and
      knowledge to identify premises used by the            has provided valuable information on the risks
      asbestos industry may be insufficiently precise.      of occupational and paraoccupational expo-
      If residents in the vicinity of some of the poten-    sure, and made a useful contribution to the
      tial sources were not exposed to asbestos, or         study of residential exposure to asbestos. A
      only at distances <0.5 km, their inclusion will       high proportion of both cases and controls
      underestimate the risk from larger industrial         could have been exposed to asbestos at work.
      sources of asbestos. However, more detailed           Although heavy exposure through the manu-
      information about the manufacturing proc-             facture of asbestos products has fallen over the
      esses at the various factories is rarely available,   years, there is still scope for exposure in work-
      preventing more precise estimates of their            ers engaged in repair, renovation, and demoli-
      potential for releasing asbestos into the envi-       tion of buildings containing asbestos, the so
      ronment. Nor was it possible to determine             called "third wave of asbestos disease".'7
      which type of asbestos was used in many facto-        Prevention requires rigorous enforcement of
      ries, although the Leeds factory, which had           existing legislation, workforce education, and
      provoked concern, was known to use mainly             surveillance to ensure that existing standards
      crocidolite, and the Calderdale factory used          are adequate to identify other potential causes
      crocidolite and brown asbestos (amosite) for          of mesothelioma. Only 3% of the cases were
      long periods.6 15                                     associated solely with residential exposure, and
         This and other case-control studies suffer         it is not possible to conclude from this study
      from the fact that retrospective asbestos expo-
      sure histories are likely to be incomplete: rela-     alone Whether such exposure is a risk factor for
      tives providing information may not be aware          development ofmesothelioma. A larger study is
      of, or be able to recollect, important occupa-        needed to do this. It is important that all possi-
      tions or residences, particularly if these were of    ble sources of asbestos exposure should be
      short duration. The quality of information var-       considered and thoroughly investigated in
      ied: some informants were able to provide very        cases of mesothelioma. For instance, 25 of the
      detailed life histories, others were vague about      cases of mesothelioma in Leeds between 1971
      particular periods or aspects of the subject's        and 1987 were labelled as due to environmen-
      life.                                                 tal exposure to asbestos by Her Majesty's
         There are also difficulties in trying to ascer-    Coroner or the local media, but further investi-
      tain the key routes of asbestos exposure when a       gation established that 20 of these could have
      life history suggests exposure from more than         been occupationally or paraoccupationally
      one route. For example, whereas a weaver of           exposed.'8 At present, such investigations are
      asbestos products would almost certainly have         performed on an ad hoc basis, but it would be
      been exposed to very high levels of asbestos          useful to determine whose responsibility this
      which would dominate those from other                 should be. Combining the results of such
      routes, an electrician (coded as possible             inquiries would provide a means of obtaining a
      occupational exposure) might have limited             sufficiently large database from which the role
      asbestos exposure at work, but higher exposure        of residential exposure in the development of
      as the result of living near an asbestos factory,     mesothelioma in the community could be
      or from a family member working there. We             reconsidered.
      have assumed that occupational exposure
      dominates that from other sources in this study.
      The few studies which have measured airborne          We are particularly grateful to the contacts of the cases of mes-
                                                            othelioma and those who acted as controls, for their help and
      asbestos concentrations in non-occupational           interest; and to Dr Philip E da Costa, consultant pathologist,
      settings indicate that paraoccupational expo-         Seacroft and Killingbeck Hospital, for carrying out histological
                                                            confirmation of the cases. We also acknowledge support given
      sure can lead to higher concentrations than           by staff at local pathology laboratories and medical records
      residential exposure'6, but we cannot be sure         departments, staff at local offices of Her Majesty's Coroner and
      that this is the case for all the subjects in this    staff at the University of Leeds Medical Library, Leeds City
                                                            Library, and Bradford City Library. The study was funded by
      study.                                                the Colt Foundation. There are no conflicts of interest.

      Appendix 1: List of jobs that determined the classification of subjects into each occupational class of
      exposure to asbestos.
      Asbestos factory worker (fitter, sprayer, carding, packer, mattress maker, maintenance, spinner,
      etc), boilermaker, bricklayer, builder's labourer, built gas retorts, built railway coaches, display
      work using asbestos, dockwork involving asbestos, fitter on ships, fitter using asbestos, foreman
      using asbestos, gas fitter, heating engineer or plumber, joiner using asbestos, laboratory technician
      using asbestos, lagger, maintenance involving asbestos, metal worker using asbestos, munitions
      work involving asbestos, naval engineer using asbestos, power station worker, roofer or slater,
      shipyard worker.
      Aircraft mechanic, air raid warden, architect, boilerman or stoker, car mechanic, chemist, clerk at
      asbestos goods factory, cleaner in asbestos goods factory, driver of asbestos loads, electrician,
      engineer, fireman, fitter, foreman, foundry worker, Hoffman presser, industrial brush repair,
Routes of asbestos exposure and the development of mesothelioma                                                                                               409

                               joiner, laundry worker (ironing), maintenance work, metal worker, munition worker, painter,
                               railway worker, scaffolder, stonemason, tank gunner, telephone engineer, welder.
                               Baker, barman, catering, cinema manager, cleaner, clerk, clothing industry worker (machinist,
                               tailor's cutter, finisher, etc), colliery worker, draughtsman, driver, dyeworker, factory work, farm-
                               work, film editor, groundsman, hairdresser, laundry worker, letter sorter, manager, navy, nurse,
                               paver, postman, priest, sales assistant, shoemaker, teacher, tiler, upholsterer, warehouseman,
                               Appendix 2: Activities at potential industrial sources of asbestos near Yorkshire residences of study
                                                                                                                                Number of sources within 0. 5 km of
                                                                                                                                3 1 study subject
                               Group 1-likely manufacture using asbestos:
                                 Asbestos goods manufacturers                                                                   17
                                 Engine packing, steam packing, or metallic packing manufacturers                               13
                                 Brake and clutch manufacturers                                                                 7
                                 Jointing and gasket manufacturers                                                              1
                                 Railway carriage and locomotive manufacturers                                                  7
                                 Insulation, boiler covering, or packing manufacturers                                          1
                                 Fibrous plaster or fibrous cement manufacturers                                                9
                               Group 2-may be storage only, or asbestos not used:
                                 Asbestos building materials, asbestos cement, asbestos roofing                                 10
                                 Insulation specialists, engine packing, fireproofing, steam products manufacturer   18
                               Total                                                                                            83

                                1 Talcott JA, Antman KH. Asbestos related malignancy. Current     10 McDonald AD, McDonald JC. Malignant mesothelioma in
                                   problems in cancer. Chicago year book. Chicago: Medical            North America. Cancer 1980;46:1650-6.
                                   Publishers, 1988.                                              11 Newhouse M, Thompson H. Mesothelioma of the pleura
                                2 De Klerk NH, Armstrong BK. The epidemiology of asbes-               and peritoneum following exposure to asbestos in the Lon-
                                   tos and mesothelioma. In: Henderson DW, Shilkin KB,                don area. BrJInd Med 1965;22:261-9.
                                   Langlois SL, Whitaker D, eds. Malignant mesothelioma.          12 Vianna NJ, Polan AK. Non-occupational exposure to asbes-
                                   New York: Hemisphere, 1992:223-50.                                 tos and malignant mesothelioma in females. Lancet 1978;i:
                                3 Polnar PV. Further evidence of non-asbestos related                 1061-3.
                                   mesothelioma. A review of the literature. Scand Jf Work        13 Lebel. Review offibre concentrations in Quebec asbestos mining
                                   Environ Health 1988;14:141-4.                                      towns. Quebec: Quebec Asbestos Mining Association,
                                4 Peto J, Hodgson JT, Matthews FE, Jones JR. Continuing               1984.
                                   increase in mesothelioma mortality in Britain. Lancer 1995;    14 Kohyama N. Airborne asbestos levels in non-occupational
                                   345:535-9.                                                         environments in Japan. In: Bignon J, ed. Non-occupational
                                5 Gardner MJ, Acheson ED, Winter PD. Mortality from mes-              exposure to mineralfibres. Lyon: IARC, 1989:262-76.
                                   othelioma of the pleura during 1968-78 in England and          15 Edward AT, Whitaker D, Browne K, Pooley FD, Gibbs AR.
                                   Wales. BrJ Cancer 1982;46:81-8.                                    Mesothelioma in the north of England. Occup Environ Med
                                6 Arblaster L, Hatton P, Renvoize ER, Schweiger MS. Leeds
                                   mesothelioma deaths 1971-87. Leeds: Leeds Western Health           1996;53;547-52.
                                   Authority, Leeds Eastern Health Authority, and Leeds City      16 Nicholson WJ, Rohl AN, Weisman I, Selikoff IJ. Environ-
                                   Council, January 1990.                                             mental asbestos concentrations. In: Wagner JC, ed. Biologi-
                                7 Teta MJ, Lewinsohn HC, Meigs J, Wister V, Romeo A,                  cal effects of mineralfibres. Lyon: IARC, 1980:823-7.
                                   Mowad LZ, et al. Mesothelioma in Connecticut, 1955-77.         17 Landrigan P. The third wave of asbestos disease: exposure to
                                   J Occup Med 1983;25:749-56.                                        asbestos in place [preface]. Ann New York Acad Sci
                                8 Collett D. Modelling binary data. London: Chapman and               1991;643:15-6.
                                   Hall, 1991.                                                    18 Arblaster L, Hatton P. Howel D, Renvoize E, Schweiger M,
                                9 Wagner J, Sleggs C, Marchand P. Diffuse pleural mesothe-            Swinburne LM. Occupational and environmental links to
                                   lioma and asbestos exposure in the North Western Cape              mesothelioma deaths occuring in Leeds during 1971-87. J
                                   Province. BrJ7Ind Med 1960;17:260-71.                              Public Health Med 1995;17:297-304.

To top