| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Occupational and Environmental Medicine, National Heart and Lung Institute, Imperial College, London, UK.
Correspondence to: Jessica Harris, MSc, Department of Occupational and Environmental Medicine, Imperial College (National Heart and Lung Institute), 1b Manresa Rd, London SW3 6LR, UK; e-mail: jessica.harris{at}ic.ac.uk
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Design and setting: Data on occupation and address (postal code) were obtained from certificates of men and women dying as a result of CFA between 1981 and 1990 and were compared with national mortality statistics. Place of birth data were extracted from certificates for deaths between 1993 and 1995, the only available years, and were compared with national birth statistics.
Measurements and results: Standardized mortality ratios (SMRs) were raised (p < 0.05) in the following four occupational groups: members of the armed forces (SMR, 217.8); miners and quarrymen (SMR, 142.0); service, sports, and recreation workers (SMR, 118.6); and electrical and electronic workers (SMR, 146.6). Of these four groups, the latter group might be worth testing in a future study. There was statistical evidence of geographic clustering in postal code sectors for the recorded place of death, but the high-rate areas were different in men and women. Deaths were increased for those subjects born in urban areas, although these did not follow a clear geographic pattern.
Conclusions: Overall, these analyses provide little evidence of any important contribution from environmental factors to the etiology of CFA and suggest that more consideration be given to alternative concepts of causation.
Key Words: cryptogenic fibrosing alveolitis • geographic distribution • occupation
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Despite the lack of any marked difference between male and female mortality rates, which is the usual evidence of occupational etiology, several investigators have searched for an occupational clue to the cause of CFA. For example, a 1996 case-control study3 of CFA reported an increased risk in men from occupational exposures to metal and wood dusts, findings that were in line with results from earlier studies.4 5 A more recent report6 found evidence of increased exposure to antidepressants among a series of CFA cases compared to age-matched, sex-matched, and community-matched control subjects. Few other hypotheses have been tested formally.
In order to investigate further the possible role of occupational and other environmental exposures, we have carried out a detailed analysis of death certificates in England and Wales between the years 1981 and 1990. We also have examined the recorded places of birth for subjects whose deaths occurred between 1993 and 1995 for any evidence of clustering, which might suggest early life exposures, such as childhood infections or a common geologic factor, having a role in this disease.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We obtained from ONS details from the 5,941 death certificates in which "idiopathic fibrosing alveolitis" (International Classification of Disease, ninth revision [ICD-9] code, 516.3) had been coded as the underlying cause of death for the years 1981 to 1990. Information on sex, date of birth, usual occupation, and postal code of usual residence from the entry completed at death registration was extracted. From the same source, we also obtained all coded information on 1951 death certificates in England and Wales in 1985 and 1986, which mentioned other alveolar and parietoalveolar pneumopathy (ICD-9 code, 516). These 2 years were the only ones in which coded information on the mentioned causes of death was available. CFA (ICD-9 code, 516.3) accounted for 91% of the deaths for each sex in which ICD-9 code 516 was coded and 95% of deaths occurred in subjects > 30 years of age.
Data from a further set of certificates were extracted; we obtained details from all 2,954 certificates in which CFA had been coded as the underlying cause of death for the years 1993 to 1995. These were the only 3 years for which information on the deceased person’s place of birth was available.
Occupational Analyses (Certificates From 1981 to 1990)
Occupations were recoded by the authors to the classification
used at the time of the 1971 census. The analysis of occupation was
restricted to the 2,938 men who were between the ages of 15 and
64 years in 1971 and to those subjects who had an occupation coded on
their death certificate (n = 2,132 [73%]). The expected numbers of
deaths for each occupational code were calculated using age-specific
and sex-specific rates from 1981 to 1990 applied to occupational
denominators obtained from the 1971 census, which were adjusted to take
into account the proportion for which no occupational information was
available. In this way, standardized mortality ratios (SMRs
[ie, the ratio of observed to expected deaths]) with 95%
confidence intervals (CIs) were produced for each of 223 possible
occupational codes and the corresponding 27 occupational groups
identified by ONS.
Registrars of death are instructed that the occupation of a woman should not be recorded unless she had been in paid employment most of her life.7 Consequently, occupational details were only available for 339 women, 20% of those women between 15 and 64 years of age in 1971. These details were analyzed, but clearly no reliable comparison with expected rates could be made.
Geographic Analyses
Place of Usual Residence (Certificates From 1981 to 1990):
We examined the clustering of male and female deaths from CFA by an
analysis of geographic sectors derived from the first
four characters of the recorded postal code; each sector contains about
2,000 addresses. The numbers of deaths in each sector were compared
with those that would be expected from a Poisson distribution. To
investigate the extent of any misclassification with asbestosis, we
examined also deaths in the following nine areas in England and Wales
with a history of shipbuilding and repair: Merseyside, Tyne and Wear,
Barrow-in-Furness, Plymouth, Portsmouth, Southampton, Gillingham,
Rochester, and Swansea.
Birthplace (Certificates From 1993 to 1995): We compared the observed distribution of the places of birth of the deceased to the number expected, based on information extracted from the 1921 census. As the mean age at death was 72 years, this census was the nearest to the average year of birth. The place of birth, which was recoded by the authors to the nearest of 110 county and metropolitan boroughs and 1,451 urban and rural districts that had been detailed in the 1921 census, was recorded in a text format on the death entry. The county and metropolitan boroughs include the larger towns and cities of the United Kingdom and accounted for 46% of the population and 48% of births.
Information on place of birth was available for all but three of the
2,954 deaths from CFA between 1993 and 1995. The analysis was
restricted to those who died at 
45 years of age and were born in
England or Wales (n = 2,615). It was possible to recode 2,424
certificates (93%) directly. One hundred sixty-seven did not contain
complete information, and it was not possible to code the remaining 24
places.
The expected numbers of births for each area were weighted by the population number of births and were adjusted to account for the deaths with birthplace incomplete, or missing, or outside England and Wales. SMRs were calculated for each county and metropolitan borough and standard region. SMRs also were calculated separately for all county and metropolitan boroughs and compared to other towns within each standard region.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
55 years. The previously reported secular increases in
age and sex-specific death rates for each code were confirmed and
observed to continue beyond 1988. The mortality rate from CFA for men
was slightly higher than that for women across all ages, and CFA
accounted for a slightly higher proportion of all deaths among men (Fig 1
). There was no seasonal variation in the proportion of deaths for each
month of birth, and this lack of variation remained after stratifying
for age and sex (data not displayed).
|
|
Geographic Analyses
Place of Death:
Postal codes were available for 5,933
certificates (99.9%). Assuming a Poisson distribution, 0.46 male
deaths and 0.30 female deaths would be expected in each postal
code sector; the geographic analysis of deaths resulting from ICD-9
code 516.3 revealed a greater frequency in certain sectors than would
have been expected by random assortment (p < 0.001). This was true
separately for men and women, but there were few areas overlapping
between the sexes (Table 2
). A separate analysis of rural postal codes also suggested a tendency
to cluster, although with no discernable geographic pattern. The number
of deaths from CFA in areas known for shipbuilding, and therefore due
to possible asbestos exposure, were higher in four of nine areas. In
Tyne and Wear, deaths were more frequent than expected (deaths, 103;
SMR, 130.1; 95% CI, 107 to 158). Ratios also were increased, although
not significantly, in Merseyside (deaths, 107; SMR, 107.6), Plymouth
(deaths, 23; SMR, 135.2), and Barrow-in-Furness (deaths, 7; SMR,
129.8). Ratios were not increased in the other five areas
(ie, Portsmouth, Southampton, Gillingham, Rochester, and
Swansea).
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The inaccuracies associated with information from death certificates are well known, particularly for a disease such as CFA for which the diagnostic criteria are less than specific. Even so, we see no reason to think that a large case series based on death certificates is likely to be seriously biased epidemiologically. It is possible that asbestosis was wrongly diagnosed as CFA in a small number of cases, but our findings provided little evidence of this. Other pneumoconioses such as those due to coal dust or silica are more readily diagnosed radiologically than asbestosis and are unlikely to be an important source of error.
Listed occupations and places of residence may be those nearest to the time of death rather than the usual ones. We have in part accounted for long-latency diseases by calculating occupation-specific SMRs using occupational denominator values from between 11 and 19 years prior to death. Diagnostic or coding inaccuracies in death certificates also have been reported in persons with CFA.1 8 Despite these problems, we believe that the study of death certificates is a useful means of generating and testing etiologic hypotheses in a disease in which access to the services of specialists is an important factor in diagnosis.
Of the 27 occupational categories used by ONS in 1971, the SMR among men was raised in 4 categories to a statistically important extent. The excess in miners and quarrymen was entirely explained by the data for coal miners, who are subject to frequent chest radiographs and therefore to a higher probability of diagnosis, whether it is correct or incorrect. The excess among persons in the armed forces might well be explained in the same way. Of greater interest are those members of the remaining two groups who are at increased risk (specifically, electricians, electrical engineers, firemen, and cleaners), all of whom were exposed to a variety of fumes and potentially toxic airborne chemicals. These findings deserve, therefore, to be tested in other populations.
No increases were found among laundry workers, wood workers, and hairdressers and beauticians, occupations suggested by Iwai et al5 as having an increased risk of CFA. Although not formally tested, there were more deaths among women employed as metalworkers than the occupational data collected in the 1971 census would have suggested.
The analyses by place of death and by place of birth, although not wholly negative, provided no consistent or promising leads. Any undue clustering by date of death would be as likely to reflect ascertainment bias as causation, but it was not apparent. Place and season of birth might relate to infections in early life or to nutritional or geologically related factors, but these analyses also yielded nothing suggestive.
Clearly, it would be wrong and premature to conclude from these almost wholly negative findings that environmental factors play no part in the causation of this complex and puzzling disease, but they do suggest that we should consider quite different hypotheses. Could it be that we have here an example of a common pathologic process induced not by a few specific environmental agents but perhaps by many, in subjects who for primarily genetic reasons react in this seriously abnormal way? There is evidence of a dominant genetic pattern in a small proportion of patients (< 2%),9 but clearly this cannot explain the great majority of cases.
A mechanism of this kind would not exclude the possibility that long-term multifactorial environmental changes might lead to an increasing incidence of disease or that such changes might be found more frequently in some occupations than in others, particularly in those entailing exposure to fibrogenic dusts. The elucidation of such a hypothesis would require the investigation of genetic susceptibility by molecular techniques alongside classic environmental epidemiology.
|
Footnotes |
|---|
Received for publication March 15, 2000. Accepted for publication August 2, 2000.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. MIYAKE, S. SASAKI, T. YOKOYAMA, K. CHIDA, A. AZUMA, T. SUDA, S. KUDOH, N. SAKAMOTO, K. OKAMOTO, G. KOBASHI, et al. Occupational and Environmental Factors and Idiopathic Pulmonary Fibrosis in Japan Ann. Hyg., April 1, 2005; 49(3): 259 - 265. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or low (
) numbers of deaths from
CFA from 1993 to 1995 by place of birth, as indicated by the 95%
CIs. SMRs for deaths from CFA by places of birth categorized
into standard regions, England and Wales, from 1993 to 1995. Values are
given as SMRs for county and metropolitan boroughs (in bold) and other
towns (in italics). * = 95% CI excludes 100.