Published ahead of print on June 12, 2008, doi:10.1164/rccm.200801-063OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200801-063OC
Mapping the Risk of Mesothelioma Due to Neighborhood Asbestos Exposure1 Department of Community Health and Epidemiology, Nara Medical University School of Medicine, Kashihara; and 2 Department of Environmental Health, Osaka Prefectural Institute of Public Health, Osaka, Japan Correspondence and requests for reprints should be addressed to Norio Kurumatani, M.D., Department of Community Health and Epidemiology, Nara Medical University School of Medicine, 840 Shijyocho, Kashihara, Nara 634-8521, Japan. E-mail: knorio{at}naramed-u.ac.jp
Rationale: Little is known about neighborhood exposure to asbestos and mesothelioma risk among residents around an industrial source of asbestos. Objectives: To investigate the magnitude of the risk among residents by asbestos exposure levels and to determine the range of the area affected by asbestos. Methods: We calculated standardized mortality ratios of mesothelioma from 1995 to 2006 among the estimated population at risk that lived around a former large asbestos cement pipe plant in Amagasaki City, Japan, between 1957 and 1975, the time when the plant had used crocidolite and chrysotile. The distance between the plant and homes and relative asbestos concentrations obtained by diffusion equations involving meteorological conditions were used to determine asbestos exposure levels among residents. Measurements and Main Results: We identified 73 mesothelioma deaths of 35 men and 38 women who had no occupational exposure to asbestos. Among persons who had lived within a 300-m radius of the plant, the standardized mortality ratio of mesothelioma was 13.9 (95% confidence interval, 5.6–28.7) for men and 41.1 (95% confidence interval, 15.2–90.1) for women. When the study area was divided into five regions by relative asbestos concentration, standardized mortality ratios of mesothelioma declined, for both sexes, in a linear dose-dependent manner with concentration. The regions with a significantly elevated standardized mortality ratio reached 2,200 m from the plant in the same direction in which the wind predominantly blew. Conclusions: Neighborhood exposure to asbestos can pose a serious risk to residents across a wide area.
Key Words: relative asbestos concentration • crocidolite • environmental exposure • pleural mesothelioma
On the basis of a case series analysis of 33 patients with pleural mesothelioma, Wagner and coworkers (1) first reported, in 1960, a strong relationship between mesothelioma and exposure to crocidolite, including occupational, paraoccupational, and neighborhood exposures. Thereafter, several epidemiologic studies have reported on mesothelioma among residents with neighborhood exposure to asbestos from industrial activities (2–12). Asbestos has been used widely since the early 20th century, and many active and closed factories worldwide that handled asbestos and asbestos-containing products could have spread asbestos in surrounding communities. To interpret a causal inference about mesothelioma encountered among residents and to protect the community from asbestos-related health hazards, public health professionals and policymakers would like to know the size and shape of at-risk neighborhoods. This information is available from few studies (2, 9–11), which investigated relations between residential distance and mesothelioma deaths. However, they did not consider meteorological factors, such as wind direction, that can influence diffusion of asbestos fibers from a point of source. In Japan, a newspaper article published on June 29, 2005 (13), reported that five residents who had lived near a now-closed asbestos cement pipe plant (called "the plant" hereafter) in Amagasaki City, Hyogo Prefecture, developed pleural mesothelioma. The plant used crocidolite and chrysotile between 1957 and 1975. These deaths attracted much social concern, and the mass media covered specific and general asbestos-related problems daily afterwards. Kubota Corporation, a major machinery maker that ran the plant, established in April 2006 a compensation system for residents (14) who lived or worked within a 1-km radius of the plant when it used asbestos, developed asbestos-related diseases including mesothelioma, and were not exposed to asbestos occupationally. The company pays 25 to 46 million yen ($220,000–420,000) per person to eligible residents and their bereaved families. A total of 164 residents with mesothelioma have applied for the compensation program as of April 2007, and most applicants have been compensated. On the basis of these 164 cases, we describe here the epidemiologic features of mesothelioma due to neighborhood exposure to asbestos, we determine the risk of mesothelioma related to residential distance from the plant by methods described in previous studies (2, 9–11), and we estimate the range of the area affected by asbestos after taking account of meteorological conditions. We have also investigated the relationship between estimated exposure and risk of mesothelioma in the community. Some of the results of this study have been previously reported in the form of an abstract (15).
Subjects The initial study population consisted of 162 of 164 residents with mesothelioma who had contacted the agent of Kubota Corporation by the end of April 2007 and for whom consent to our interview study was obtained (138 relatives of deceased subjects and 24 subjects under treatment). This study was approved by the Ethics Committee of Nara Medical University (Kashihara, Japan).
Data Collection and Calculation of Relative Asbestos Concentration Because actual concentrations of airborne asbestos fibers in the surroundings of the plant, during the period when crocidolite was used, were unknown, we estimated "relative asbestos concentrations" (unit, m–3) in each of 2,500 grid units (each grid unit, 100 m x 100 m). Briefly, the method assumed that an emission point of asbestos was at the center of the plant premises and used diffusion equations that account for meteorological conditions (17–20). We considered that airborne asbestos emitted from the plant was the sole industrial source of exposure among study subjects (see DISCUSSION). Additional details are provided in the online supplement.
Data Analysis We calculated 95% confidence intervals of SMRs, using the exact probabilities of the Poisson distribution (22). Multiple comparisons of average values among several groups were performed by one-way analysis of variance followed by Dunnett's procedure (23). The relation between relative asbestos concentration and the SMR of mesothelioma was evaluated by linear regression analysis (23). All statistical tests were conducted with SPSS version J11.5 for the PC (SPSS, Inc., Chicago, IL).
As shown in Figure 1, 121 of the 162 study subjects were judged as not having been exposed to asbestos directly or indirectly at work. Twelve of the 121 subjects worked at a location closer to the plant than their residence was, and the other 109 had no chance of occupational exposure to asbestos and had lived for at least 1 year near the plant when it used crocidolite. We then eliminated six subjects whose diagnosis was pleural cancer on the death certificate but turned out to be mesothelioma, to maintain comparability with the expected deaths calculated from national vital statistics. Further exclusion of seven deaths before the ICD-10 went into effect in Japan resulted in 79 mesothelioma deaths plus 17 patients under medical treatment for mesothelioma for further analysis. All of the 17 patients were diagnosed histologically. Of the 79 deaths, the diagnosis was confirmed by histology in 62 (78.5%), by cytology in 12 (15.2%), and by clinical findings including chest computed tomography scan in 5 (6.3%). Mean age at death was 56.9 years (SD 10.8) for men and 66.3 years (SD 11.8) for women. All 79 deceased subjects had pleural mesothelioma except for 1 patient with one peritoneal mesothelioma.
Table 1 summarizes the SMRs of mesothelioma and demographic characteristics of the 90 subjects who lived within a 1,500-m radius of the center of the plant. Residential distance was divided by intervals of 300 m. The 73 mesothelioma deaths showed a significantly increased SMR of 4.3 (95% confidence interval, 3.4–5.4). The highest SMR, 41.4 (95% confidence interval, 15.2–90.1), occurred among women who had lived within a 300-m radius of the center of the plant, and the SMR of women remained significantly elevated up to the group within 1,200 to less than 1,500 m. The SMRs of men were one-half or less than those of women. As for the women, the highest SMR among the men was found for those living within a 300-m radius of the plant; but only one other SMR for men, that is, the SMR of men living 300 to less than 600 m from the plant, was significantly elevated.
When data from the 73 deceased and 17 surviving subjects were combined, their residential period averaged 145 (range, 24 to 227) months, or 12.1 years, during the use of crocidolite in the plant. No statistically significant difference in time existed between men and women, but for both sexes the residential exposure in the 300-m radius group was 20 to 90 months shorter than the exposure of the other distance groups. The latent period, defined as the time from the subject's first year living near the plant to the appearance of the initial symptom related to mesothelioma, varied from 265 to 595 months with an average of 520 months, that is, about 43.3 years. We found no significant differences in the latent period (1) between men and women or (2) among the groups by distance from the plant. Figure 2 depicts relative asbestos concentrations classified into nine levels in each of the 100 m x 100 m grid units with places of residence of the 90 deceased and surviving subjects. We assumed the center of the plant premises was a single emission point of asbestos. The relative asbestos concentrations decrease rapidly as the distance from the plant increases. Also, grid units with higher concentrations occur more frequently to the south–southwest of the plant, coincident with the study area's predominant wind direction of blowing from the north–northeast (see Figure E1 in the online supplement). Mesothelioma case homes were scattered more north–south than east–west and more south than north relative to the plant. We identified three mesothelioma deaths (not depicted) among subjects who worked in a factory in a grid unit south of the plant whose relative asbestos concentration was the highest in the study.
Figure 3 illustrates regression lines of the SMR of mesothelioma among the deceased against relative asbestos concentrations. For this analysis, we determined five areas with the following relative asbestos concentration levels: 2–4.9 (3.2 on average), 5–9.9 (6.9), 10–19.9 (13.7), 20–49.9 (30.4), and 50–199.9 (89.8). This grouping was not done according to quintiles of asbestos exposure or population exposed, but the bands were defined so as to have at least five deaths and a high enough population at risk to allow a statistical analysis. (An area with a relative concentration of 200 m–3 or greater [Figure 2] was excluded from the analysis because it contained only the plant and the other big factories, but no residences. Also excluded was an area with a concentration of less than 2 m–3 because the area was too wide to limit its boundary and, therefore, we were unable to enumerate the population at risk; four deaths [two men and two women] lived in that area.) The sex-specific SMR of each area was calculated and the relation between relative asbestos concentrations and these SMRs was investigated by linear regression analysis. As shown in Figure 3, the highest SMR, 47.7 (95% confidence interval, 20.8–105.7) occurred for women in the area that had the highest relative asbestos concentration. Significantly increased SMRs were found in the area with concentrations of more than 5 m–3. The area reached about 2,200 m south–southwest and 900 m north–northeast from the plant's center (Figure 2). Although the SMR regression line of men is significantly less steep than that of women (P < 0.05), both lines show a linear dose–response relation between relative asbestos concentration and risk of mesothelioma. These results did not change after exclusion of 11 subjects who had died while living with household members who could have been exposed to asbestos at work (see Figure E2).
From 1957 to 1975, our study plant used an annual average of 4,670 tons of crocidolite and an annual average of 4,600 tons of chrysotile to produce cement pipes (24) (see Figure E3). The amount of asbestos used was about 5–10% of all asbestos imported to Japan in those years. According to local government information (25), none of the other 135 closed and current companies identified in Amagasaki City where asbestos products were made or used reported consuming more than 10 tons of crocidolite per year, far less than the amount used by the study plant. Packed raw asbestos transported to the plant by vehicle was separated by fluffing and then was sent upstairs through pneumatic ducts, collected, mixed with other materials to mold cement pipes, dried, and cut (24). The plant operated around the clock. We believe that fluffed asbestos was emitted from the outlet of the ducts, dispersed into the surrounding air, and inhaled by neighborhood residents. Previous studies suggested that (1) transport of raw asbestos with a loosely attached cover (11, 26), (2) improper use of asbestos residuals (1, 11, 12) such as for thermal home insulation, gardening, and creating a hard pavement, and (3) playing on piles of asbestos (12) were other asbestos exposure routes in the environment around a plant. In our study, these routes were not found except for a single narrow 30-m-length road with asbestos residuals located 500 m southwest of the plant. We did not find naturally occurring asbestos, which can cause mesothelioma (27), in the study area, and houses were not constructed with asbestos-contaminated soil mixtures (28). In the study plant, 46 employees had mesothelioma in the past 20 years, and 40 had asbestosis or asbestos-related lung cancer (24). Mesothelioma, especially pleural mesothelioma, is highly specific to asbestos (29, 30). All these facts inspired us to investigate the occurrence of the disease in the study area under the assumption that the plant was the only emitting source of crocidolite, which has much greater carcinogenicity than chrysotile (31). Our study interviewees were individuals or their families who voluntarily contacted the agent of Kubota Corporation to seek compensation for death or disease related to environmental exposure to asbestos. We could not enroll all deaths due to mesothelioma in the study area during the study period because no mesothelioma registry existed. Accordingly, a selection bias must exist. For example, deceased persons without relatives and people who were unaware of the Kubota event did not contact the compensation agent. Bias from such unreported cases would result in underestimation of the SMRs. We determined carefully whether or not the subjects had been exposed to asbestos at their workplace. However, misclassification was inevitable. The list of occupations (16) to which we referred includes all jobs with possible exposure to asbestos, which tends to make more false positives than false negatives, resulting in underestimation of the SMRs. On the other hand, underestimation of the population at risk may lead to overestimations of the SMRs, but it is unlikely that large errors occurred in the population estimates and it is even less likely that such errors varied systematically with exposure. The present study showed higher SMRs among women than men. We excluded cases with occupational exposure when counting observed deaths, but could not do so for expected deaths because no relevant data were obtained from the national vital statistics office. This bias also leads to underestimation of the SMRs. Such bias could be greater for men than for women because more men than women had been employed and therefore were exposed to asbestos occupationally. Consequently, our results show more directly for women than for men the effects of environmental asbestos exposure from the plant. Few studies have been reported, to our knowledge, on the risk of mesothelioma in relation to the residential distance from an industrial source of asbestos. A significantly increased risk of mesothelioma was observed within half a mile of an asbestos factory in London (2), within 1,000 and 2,000 m of asbestos mines, asbestos factories, or shipyards in South Africa (9) and some European countries (10), respectively, and beyond 2,500 m of an asbestos cement factory in Italy (11). Our results on residential distance are consistent with these previous findings. None of the previous studies considered meteorological conditions. Asbestos concentrations in the air surrounding the emission point depend on wind direction and velocity (32), which determine the direction and distance that asbestos fibers travel. Actually, relative asbestos concentration levels differed greatly in different directions from our study plant even when the distance from the plant was the same. For example, among women, the area with relative concentration levels showing significantly elevated SMRs extended as far as 2,200 m south–southwest of the plant, toward which the wind predominantly blew, but only 900 m in the opposite direction, north–northeast. Thus, a parameter including meteorological conditions is a better proxy as exposure dose than residential distance only, and it should be useful to investigate more accurately the effects of asbestos exposure among residents in communities. When relative asbestos concentration was used as dose of asbestos exposure, a clear linear relation was observed for both sexes. Such relations have been established by epidemiologic findings in occupational settings (33), where asbestos exposure is higher than it is in neighborhoods. The present results suggest that the linear model can apply to settings, with much lower exposure levels, such as those resulting from environmental contamination. Nearly 100 subjects with mesothelioma were identified as victims of neighborhood exposure in the present study. This number is one of the largest reported in studies of mesothelioma among residents around industrial sources of asbestos (1–12). Our subjects lived near the plant for 12 years on average, and developed the disease after a latency of 43 years on average. This duration of residential exposure is comparable to that of subjects who developed mesothelioma after living within half a mile of an asbestos factory in London (2) (14.4 yr on average) and is more than double that of subjects with mesothelioma who lived in the neighborhood of a crocidolite mine in Australia, 89% of whom, however, also lived with an asbestos worker (34). Two features of our study cases are noteworthy. First, all subjects but one had pleural mesothelioma. In the plant itself, peritoneal mesothelioma occurred in 28 workers and pleural mesothelioma occurred in 18 workers (24). This contrast is consistent with knowledge that peritoneal mesothelioma occurs more often in persons exposed to higher levels of asbestos (35, 36). Second, workers exposed to asbestos occupationally may be younger at the time of death . Our subjects who died were 10 years younger (men) and 5 years younger (women) compared with all Japanese who died of mesothelioma (37). Because workers are not exposed to asbestos until they are actually working with it, whereas residents of areas polluted with asbestos are exposed from the time they start to live there, sometimes from birth (38), it is understandable that asbestos-related deaths occur at younger ages after residential exposure than after occupational exposure. In conclusion, we believe the mesothelioma outbreak among residents was causally associated with asbestos, in particular crocidolite, that the plant had used, and that the affected area spread as far as 2,200 m from the center of the plant in a dose-dependent way. Public health policymakers, health professionals, and companies should recognize the serious risk to health of neighborhood exposure to asbestos across a wide area.
The authors deeply thank all the interviewees for responding to the long interview, which triggered grief for many. The authors thank Mr. Mitsuyoshi Hanaoka and Ms. Kazumi Yoshizaki for arranging interviews with the patients and their family members, and editing study data and other miscellaneous clerical work; and Mr. Hiroshi Iida and Akihiko Kataoka for accepting our study proposal and acting as intermediaries between the interviewees and the authors. Five reviewers and an associate editor of the Journal greatly contributed to the revision of the present manuscript through many invaluable comments and suggestions.
Supported by Nara Medical University and the Ministry of Education, Culture, Sports, Science, and Technology of Japan. This article has an online supplement, which is accessible from this issue's table of contents at www.astjournals.org Originally Published in Press as DOI: 10.1164/rccm.200801-063OC on June 12, 2008 Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form January 10, 2008; accepted in final form June 11, 2008
Related articles in AJRCCM:
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
