Spontaneous abortion risk in women textile workers in Shanghai, China

Previous studies have suggested associations between spontaneous abortions (SAB) and occupational exposures. We conducted a retrospective study of SAB among 1,752 women who were participants in a randomized trial of breast self-examination (BSE) in the Shanghai textile industry and subsequently participated in nested casecontrol studies of breast diseases. Subjects were interviewed in person from 1994--2001 to obtain a detailed reproductive history. Their occupational work history was abstracted from factory personnel records, then combined with factory-specific information and an industry-specific job-exposure matrix to ascertain occupational exposures. Industry-specific job-exposure matrices were developed for exposure to chemical agents and physical activity. Exposure reconstruction for cotton dust, endotoxin, and electromagnetic fields, informed by contemporary measurements, was used to develop a quantitative job-exposure matrix. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by logistic regression for the outcome of the first pregnancy, with adjustment for age, education level, smoking status of the woman and the spouse, and woman's use of alcohol. Analyses considering the entire reproductive history were conducted using generalized estimating equations. For the 109 women who experienced a SAB in the first pregnancy, an elevation in risk of SAB was observed associated with exposure to synthetic fibers (1.89, 95% CI: 1.20--3.00) and mixed synthetic and natural fibers (3.31, 95% CI: 1.30--8.42). Compared with women employed in sedentary jobs, a reduced risk of SAB in the first pregnancy was found for women working in jobs with either light physical activity (0.18, 95% CI: 0.07--0.50) or medium physical activity (0.24, 95% CI: 0.08--0.66). An elevation in risk with frequent crouching in the job was also observed. No increased risks were observed by textile process of employment or for women working with solvents, electric and magnetic fields, or in jobs with rotating shift work. Analyses including all pregnancies of all women yielded similar conclusions. Although misclassification of exposure is a potential limitation, this study relied upon detailed characterizations of the textile industry and thousands of historical and contemporary measurements to reconstruct exposures and inform job-exposure matrices.