Epidemiological studies of the etiology of pancreatic cancer

Sammanfattning: Pancreatic cancer is one of the most lethal human cancers with less than 5% of patients surviving 5 years. Despite an urgent need for primary prevention, little is known about the etiology of this cancer. This thesis was aimed at increasing our understanding of the etiology of pancreatic cancer with regard to lifestyle and infectious agents. We first analyzed the secular trends of pancreatic cancer in Sweden and then focused on studying risk factors. Firstly, we retrieved all incident cases (46,257) of pancreatic cancer from the Swedish Cancer Register and all mortality cases (53,686) from the Causes of Death Register between 1960 and 2003, to investigate the underlying temporal trends of pancreatic cancer by both incidence and mortality rates. A similar pattern of trends was observed irrespective of whether incidence or mortality was used. The age-standardized rates of pancreatic cancer increased during the first decade and then peaked for both sexes (the male peak occurred in the early 1970s and the female peak in the 1980s) followed by a steady decline in both groups. The close agreement between the incidence and mortality and the gender disparity suggest a true decline in pancreatic cancer incidence in recent years in Sweden, and gender-specific trends in exposure to environmental risk factors. Secondly, we used the large Swedish Construction Workers Cohort to investigate the association between Swedish moist snuff (snus) use and the risk of pancreatic cancer. Overall, 279,897 men who were construction workers in 1978-92 were included and followed up until the end of 2004. In order to better control the strong confounding effect of smoking, we mainly focused on 83 pancreatic cancer cases in the 125,576 never-smokers and assessed the pure effect of snus use using Cox proportional hazards regression. The main finding in this study is that snus users who had never smoked had a doubled risk (95% confidence interval [CI]: 1.2-3.3) of pancreatic cancer compared with men who never used any tobacco, with some evidence of a dose-risk association. Thus, despite its comparably low levels of tobacco-specific nitrosamines, the Swedish snus may not be an entirely safe product as an alternative to cigarette smoking. Thirdly, we took advantage of the large prospective Women s Health Initiative in the United States with measured anthropometric factors and detailed potential confounders, to examine the role of obesity, especially central obesity (measured as waist-to-hip ratio [WHR] ) on the risk of pancreatic cancer. Again, Cox proportional hazards regression models were used to estimate relative risk. In total, 251 incident cases of pancreatic cancer were identified over an average 7.7 years of follow-up among 138,503 postmenopausal women. Among all tested anthropometric variables, only WHR was significantly associated with the risk of pancreatic cancer. The risk of pancreatic cancer increased 27% (95% CI: 7%-50%) per 0.1 increase of WHR. In addition, we noted that women who were current smokers with the highest WHR had a 3.7-fold (95% CI: 2.1-6.4) elevated risk of pancreatic cancer compared with never smoking women with the lowest WHR (p for interaction = 0.03). Findings from this study suggest that the WHR may be a better predictor for the disease risk in postmenopausal women. Finally, we used two types of peptic ulcer (gastric and duodenal) to test two proposed models (increased intragastric N-nitrosamine formation or hyperchlohydria) for the observed link between Helicobacter pylori (H. pylori) infection and pancreatic cancer, since gastric ulcer is primarily associated with corpus colonization of H. pylori, atrophic gastritis and formation of N-nitrosamines, while duodenal ulcer is usually related to antral colonization and hyperacidity release. We retrieved 88,338 patients hospitalized for gastric ulcer and 70,516 patients for duodenal ulcer from the Swedish Inpatient Register between 1965 and 2003. Following operation for peptic ulcer, 14,887 patients who underwent gastric resection and 8,205 with vagotomy were analyzed separately. Standardized incidence ratios (SIRs) were used to estimate the risk for pancreatic caner. During 3-38 years of follow-up, we observed a small excess risk (SIR=1.2, 95% CI 1.1-1.4) for pancreatic cancer among unoperated gastric ulcer patients comparing to the Swedish general population, and the risk increased with increasing duration of follow-up (p for trend = 0.03). Patients who underwent gastric resection had 50% (95% CI: 10%-110%) excess risk of pancreatic cancer comparing to the Swedish general population. The risk increased with time since surgery, reaching 2.1-fold (95% CI 1.4-3.1) 20 years after gastric resection. Unoperated duodenal ulcer was not associated with pancreatic cancer risk, nor was vagotomy. Findings from this study lend indirect support to the nitrosamine hypothesis, but not to the hyperacidity hypothesis, in the etiology of pancreatic cancer.