br To date aside from cigarette smoking
To date, aside from cigarette smoking [2,3], firmly established risk factors for pancreatic cancer are few. Obesity [4e7], a sedentary lifestyle [8e10], and diabetes [3,10,11] have been investigated as potential risk factors. However, of these hypothesized risk factors, only obesity shows consistent evidence of an association with the disease, and the association is weak . Central adiposity in adulthood [12e14], as well as adiposity earlier in adulthood [12e14], may also be risk factors [6,13]; however, fewer studies have addressed these variables. It is challenging to determine whether diabetes is an independent risk factor for pancreatic cancer owing to the inter-related effects of obesity, sedentary
behavior, and hyperglycemia and hyperlipidemia and the fact that diabetes can be an early symptom of pancreatic cancer . In an umbrella review of meta-analyses of studies of type II diabetes and cancer , pancreatic cancer was not among the 5 cancers for which the association with diabetes was judged to be robust.
We used data from the Women's Health Initiative (WHI) cohort to examine the associations of diabetes, measured adiposity (overall and abdominal), and self-reported weight earlier in life with risk of incident pancreatic cancer.
Materials and methods
The WHI is a large, multicenter, multifaceted study designed to advance understanding of the determinants of major chronic dis-eases in postmenopausal women. It is composed of a clinical trial component (CT, n ¼ 68,132) and an observational study component (OS, n ¼ 93,676) . The clinical trial component included three randomized controlled interventions: hormone therapy, low-fat diet modification, and calcium-vitamin D supplementation. Women be-tween the ages of 50 and 79 years and representing major racial groups were recruited from the general Jasplakinolide at 40 clinical centers throughout the United States between 1993 and 1998. De-tails of the design and reliability of the baseline measures have been published elsewhere [16,17]. Written informed consent was obtained from participants at all WHI centers in accordance with recognized ethical guidelines, and the study was approved by the institutional review board of each center, as well as by that of the Coordinating Center at the Fred Hutchinson Cancer Research Center.
Exposures and covariates ascertainment
At study entry, self-administered questionnaires were used to collect information on demographics, medical (including history of diabetes), reproductive, and family history, and on dietary and lifestyle factors, including smoking history, alcohol consumption and recreational physical activity. In relation to diabetes, women in the OS and CT groups were asked if diastole had ever been diagnosed with diabetes by a doctor while they were not pregnant [18,19]. Women were categorized as having diabetes, if they answered yes to this question.
All participants had their weight, height, and waist and hip circumferences measured by trained staff at the baseline. Weight was measured to the nearest 0.1 kg, whereas height and waist and hip circumferences were measured to the nearest 0.1 cm.
Body mass index (BMI) was computed as weight in kilograms divided by the square of height in meters. We used the World Health Organization classification of BMI (18.5e<25.0, 25.0e<30, 30e<35, 35 kg/m2). Participants in the OS were also asked about their weight at ages 18, 35, and 50 years, as well as about the number of times that they had experienced substantial weight loss. BMI at earlier ages was computed using self-reported weight at those ages and height measured at enrollment. Given the small number of women whose BMI fell in the overweight and obese categories (particularly those with measurements at ages 18 and 35 years), the self-reported BMI variables were categorized into quartiles based on the distribution among the noncases. For waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR), we also created quartiles based on the distribution among the noncases.
Clinical outcomes (including new cancer diagnoses) were updated semiannually in the CT and annually in the OS using in-person, mailed, or telephone questionnaires. Self-reports of pancreatic cancer were verified by centralized review of medical
records and pathology reports by trained physician adjudicators . As of September 20, 2016, a total of 1078 incident cases of pancreatic cancer (occurring as the first cancer) had been ascer-tained among the 161,808 participants in the OS and CT after a median of 17.9 years of follow-up (interquartile range: 9.0e19.4).