| Summary: | 博士 === 國防醫學院 === 生命科學研究所 === 91 === Abstract
In the early June of 1998, the media uncovered a serious contamination of underground water due to improper disposal of chlorinated hydrocarbon solvents from an electronics factory. A series of epidemiological studies are intended to detect the possible association between cancer incidence of workers and chemical exposures in their environment.
The results of these studies are as following:
Exposure assessment and establishment of the occupational cohorts
The environmental exposure data were retrieved from the records of company registration, periodical industrial inspections, and chemical import information from several governmental departments. The exposed and comparison worker cohorts were constructed by tracing them back through the insurance records from the Bureau of Labor Insurance (BLI). The solvents most frequently detected and with high concentrations were tetrachloroethylene (PCE), trichloroethylene (TCE) and vinyl chloride (VC). The average concentration of PCE and TCE were about 10 times higher than the groundwater quality standard. However PCE was not used as solvent in this index factory until 1981. TCE had never been used as solvent in this factory. The numbers of cohorts were 86,868, 11,178, 14,418, and 17,960, respectively, in the exposed group, textile industry comparison group A, B and electronic industry comparison group C. Females were predominant in all of the factories.
The proportionate cancer morbidity ratio study among the index electronics factory employees
To obtain information on the incidence of the various types of cancer among the exposed and comparison groups, the cohort populations were merged with the National Mortality Database, the National Cancer Registry Database, and the National Insurance Hospitalization, as well as the Labor Insurance Hospitalization Database. The proportionate cancer morbidity ratio (PCMR) was used to estimate the cancer risk of the exposed workers in comparison with either textile workers or electronics workers.
After adjustment for age, only the PCMR for breast cancer in the exposed female employees was significantly elevated when compared with the two comparison groups. No consistent finding could be drawn for the male workers. The increased risk of breast cancer was mainly found in the category of 1989-1997 when stratified by calendar year. However, there was no dose-response relationship between female breast cancer risk and duration of employment. Although some PCMRs for the cancers were also increased in the exposed group, no consistent results could be found when similar analysis were conducted using different exclusion criteria.
The standardized mortality ratio study for cancer among the index electronics factory employees
Basic demographic and employment data of the cohort were retrieved from the Bureau of Labor Insurance (BLI) computer database from 1978 through December 31, 1997. To ensure complete recruitment of the exposed cohort, BLI also provided hard copies of insurance records of the exposed group who had their first employment from 1973 to 1978 (no computerized records were available before 1978). In order to determine the SMRs for all types of cancer, the Taiwanese general population was selected as comparison group.
Vital status and causes of death of study subjects were determined from January 1, 1985 to December 31, 1997 by linking cohort data with the National Mortality Database. Cause-specific standardized mortality ratios (SMRs) based on comparisons with the general population in Taiwan were calculated for the entire cohort, by gender, age and calendar time period.
The observed deaths for all cancer in the males was significantly fewer than expected (SMR: 0.65, 95% C.I. 0.50-0.83), while the observed number of deaths for all cancer was equal to the expected in the females (SMR: 1.00, 95% C.I. 0.88-1.24). For the male workers, no SMR was found to be significantly increased. The SMRs of small intestine cancer, cancer of other respiratory organs, and cancer of other lymphatic and hemopoietic tissue were non-significantly increased (SMR>1.0). For the female workers, there was also no SMR significantly increased. SMRs for cancers of stomach, colon, rectum, pancreas, other digestive organs, other respiratory organs, bone and articular cartilage, breast, kidney and other urinary organs, and other lymphatic and hemopoietic tissue, as well as Hodgkin’s disease were non-significantly increased. The SMRs for several cancers found to be non-significantly increased were based on small numbers of deaths. In addition, the analysis of ovarian cancer suggested an upward trend by length of employment, the risk for the entire female cohort was not elevated.
The standardized incidence ratio study for cancer among the index electronics factory employees
Data on the exposed cohort covering the period from 1973 through December 31, 1997, were retrieved from the computer database of the BLI. The vital status and causes of disease among the study subjects for the period from January 1, 1979 to December 31, 1997, were determined by linking cohort data with data from the National Cancer Registry database. SIRs were calculated, with cancer incidence among the general population of Taiwan serving as a basis for comparison. No SIRs for male workers were significantly elevated. As for the females, only the SIR for breast cancer was significantly above expectation (SIR:1.19, 95% C.I. 1.03-1.36). Further SIR analysis of breast cancer, stratified by age at which diagnosis occurred, detected no significantly elevated age-specific SIR, although most of the age-specific SIRs were greater than 1.0. Within the 40-45 age stratum, the SIR for breast cancer was found to be marginally higher than expected (SIR:1.34, 95% C.I. 1.00-1.77).
Further analysis was conducted with minimum employment durations of 6 months, 1 year, and 5 years, and minimum latency periods of 5 years and 10 years, respectively. After adjustment for age and calendar year, no SIRs were calculated that were significantly higher than expected. The SIR for breast cancer showed no significant increase. A trend analysis was separately conducted with (3-interval) duration of employment and calendar year as variables. No significant trend relationship was discovered between the incidence of male cancers and the two time variables, and no female cancer showed an exposure dose-response relationship with duration of employment. Female cancers of all types, and the trachea, bronchus, lung, and breast revealed an increasing trend with calendar year, with SIRs highest for the period 1991-1997.
Conclusion
No significant association due to lacking of clear dose-response relationship and inconsistence between two comparison groups were also support the non-causal relationship found in our previous PCMR study. It was also found that no significant elevation of SMR for any cancer in this exposed cohort in our previous mortality study. Increased risks of breast cancer and all forms of cancer were found for the total cohort. However, it should be noted that the majority of these workers had an employment duration of less than 1 year, and that a major proportion of the female workers with a diagnosis of breast cancer had been employed for less than 3 months. When workers with duration of employment less then 3 months were excluded, no elevated risk of cancer was detected. In other words, increased risk of breast cancer and all cancers was concentrated among short-term employees. When the data were stratified by age, the risk of breast cancer was not found to be elevated for any age group. No dose-response relationship was discovered between exposure and duration of employment, and the only significant elevation of SIR was found among workers with an employment duration of less than 1 year. When the data were stratified by calendar year, a significant increasing trend relationship was found to exist between incidence of breast cancer and 3 calendar-year intervals; however, this trend was similar to that found among the population of Taiwan as a whole.
This is the first instance of an epidemiological study conducted in Taiwan in order to solve the controversies of cancer risk among electronics workers. This is also the first endeavor to establish the worker cohort at a specific factory using records of the labor insurance system. Both computerized database and paper manuscripts were used to trace the detailed information to construct the cohort. The results of this epidemiological study will have a large impact on the workers’ compensation of occupational cancers among the exposed workers.
Key words: chlorinated organic solvents, proportionate cancer morbidity ratio, standardized mortality ratio, standardized incidence ratio, breast cancer.
|
|---|
