To examine the possible role of body size and reproductive factors in pancreatic cancer, data were analyzed from a population-based case-control study conducted in Shanghai, China.Cases (n = 45 I) were permanent residents of Shanghai, 30-74 years of age, newly diagnosed with pancreatic cancer between October I, 1990, and June 30, 1993. Deceased cases (I 9Oh) were excluded from the study. Controls (n = 1,552) were randomly selected from permanent Shanghai residents and frequencymatched to cases by gender and age. Information on body size and reproductive and other possible risk factors was collected through personal interviews. After adjustment for age, income, smoking and other confounders, a positive dose-response relation between body mass index and risk of pancreatic cancer was observed in both sexes. Among women, the risk of pancreatic cancer was significantly associated with number of pregnancies and live births. Compared with 0-2 pregnancies or live births, the odds ratio (OR) for 8 or more pregnancies was I .90. while that for 5 or more births was I .88. A modest elevation in risk, independent of parity, was associated with early age at first birth. Risk increased over 40% among women with a first birth at or before age 19 years relative to those at age 26 years or older. Ever use of oral contraceptives was associated with excess risk, though based on small numbers of users. Our findings suggest that, in Shanghai, obesity, gravidity, parity and perhaps use of oral contraceptives are associated with moderate increases in risk of pancreatic cancer, indicating that hormonal determinants deserve further investigation.o 1996 Wiley-Liss, Inc. * Epidemiologic studies have suggested a role for reproductive factors and steroidal hormones in the development of pancreatic cancer (Lin and Kessler, 1981;Bueno de Mesquita et al., 1992), but the evidence is limited. In addition, a positive association has been reported with body weight or obesity (Friedman and van den Eeden, 1993; M~l l e r et al., 1994). However, other studies have implicated height, but not weight, among women (Bueno de Mesquita et al., 1992), while still others have found no effect of weight (Kalapothaki et al., 1993).Since previous studies of pancreatic cancer have been carried out mainly in Western countries, we conducted a population-based case-control study of pancreatic cancer in Shanghai, China, where the incidence rates have been steadily rising over time (Jin et al., 1993). In earlier reports of this study, we have shown a positive relation with smoking and protective effects of fresh fruits and vegetables (Ji et al., 1995a, b). In this report, we examined the association with anthropometric and reproductive factors.
MATERIAL AND METHODSData were derived from a population-based case-control study of gastro-intestinal cancers (pancreas, esophagus, colon and rectum) in urban Shanghai. The design has been described in detail elsewhere (Ji et al., 19956). Briefly, pancreatic cancer cases newly diagnosed between October 1, 1990, and June 30, 1993...