Increased risk of preterm delivery among people living near the three oil refineries in Taiwan Chun-Yuh Yang a, * , Chih-Ching Chang a , Hung-Yi Chuang a , Chi-Kung Ho b , Trong-Neng Wu b , Po-Ya Chang b a Institute of Public Health, Kaohsiung Medical University, 100 Shih-Chuan 1st RD, Kaohsiung 80708, Taiwan, ROC b Institute of Occupational Safety and Health, Kaohsiung Medical University, Kaohsiung, Taiwan, ROC Received 18 April 2003; accepted 21 July 2003 Abstract The petrochemical and petroleum industries are two of the main sources of industrial air pollution in Taiwan. Data used in this study concern outdoor air pollution and the health of individuals living in communities in close proximity to oil refinery plants. The prevalence of delivery of preterm birth infants was significantly higher in mothers living near the oil refinery plants than in control mothers in Taiwan. After controlling for several possible confounders (including maternal age, season, marital status, maternal education, and infant sex), the adjusted odds ratio was 1.14 (95% CI = 1.01 – 1.28) for delivery of preterm infants for mothers living near oil refinery plants. These data provide further support for the hypothesis that air pollution can affect the outcome of pregnancies. D 2003 Elsevier Ltd. All rights reserved. Keywords: Preterm; Air pollution; Oil refinery; Epidemiology Oil refining and other petrochemical processes may lead to low level emissions of compounds with possible carci- nogenic potential including benzene (Austin et al., 1988) and 1,3-butadiene (Landrigan, 1990). Excess mortality from various cancers has been found among workers in the petroleum industry (Wong and Raabe, 1989; Rushton, 1993). If exposure to pollutants for petroleum industry workers can cause excess cancer mortality, it is reasonable to believe that air pollution from the petroleum industry might also cause excess cancer mortality, particularly for lung cancer, among residents of the neighborhoods exposed to those pollutants. However, only a few studies have examined the relationship between petroleum air emissions and the cancer risk for surrounding communities (Wilkinson et al., 1999; Yang et al., 2000a). The petroleum and petrochemical industries are consid- ered to be the main source of industrial air pollution in Taiwan (EPA/ROC, 1993). In Taiwan, exposure to petro- chemical air pollution has been found to be associated with increased rates of acute irritative symptoms in adults (Yang et al., 1997a) and upper respiratory symptoms and asthma in school children (Yang et al., 1998a). In addition, studies have also shown that residential petrochemical air pollution may increase the risk of liver cancer in males (Yang et al., 1997b) and lung cancer in females (Yang et al., 1999). Further petrochemical air pollution was associated with altered sex ratios at birth (Yang et al., 2000b) and an excess risk of delivery of low-birth weight babies at term (Lin et al., 2001a). During the past decade, there has been growing concern about reproductive health hazards caused by air pollution (Sram, 1999). At present, the relationship between pregnancy outcome and residential proximity to a source of petroleum emissions is not known. A pilot study conducted by Lin et al. (2001b) seemed to show a significant higher prevalence of delivery of preterm births in people living in the vicinity of the oldest petroleum refinery plant. This study expands the earlier study populations by adding people living in the vicinity of the other two oil refineries, thus adding substantial power to the original study, to determine if the earlier observed patterns of preterm delivery have continued. This paper is one in a series of studies evaluating the potential hazards posed by exposure to air pollution near industrial installa- tions in Taiwan. 0160-4120/$ - see front matter D 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0160-4120(03)00180-6 * Corresponding author. Fax: +886-7-3110811. E-mail address: chunyuh@kmu.edu.tw (C.-Y. Yang). www.elsevier.com/locate/envint Environment International 30 (2004) 337 – 342