Family history of cancer and risk of Pancreatic Cancer: A Pooled Analysis from the Pancreatic Cancer Cohort Consortium (PanScan)

doi:10.1002/ijc.25148

Int J Cancer. Author manuscript; available in PMC 2011 September 1.

Published in final edited form as:

Int J Cancer. 2010 September 1; 127(6): 1421–1428.

doi: 10.1002/ijc.25148

PMCID: PMC2926939

NIHMSID: NIHMS216877

Family history of cancer and risk of Pancreatic Cancer: A Pooled Analysis from the Pancreatic Cancer Cohort Consortium (PanScan)

Eric J. Jacobs,^*^†¹ Stephen J. Chanock,^*^2,³ Charles S. Fuchs,^*^†^4,⁵ Andrea LaCroix,^*^†⁶ Robert R. McWilliams,^*⁷ Emily Steplowski,^*⁸ Rachael Z. Stolzenberg-Solomon,^*^†³ Alan A. Arslan,^†^9,^10,¹¹ H. Bas Bueno-de-Mesquita,^†¹² Myron Gross,^†¹³ Kathy Helzlsouer,^†¹⁴ Gloria Petersen,^†¹⁵ Wei Zheng,^†¹⁶ Ilir Agalliu,¹⁷ Naomi E. Allen,¹⁸ Laufey Amundadottir,^2,³ Marie-Christine Boutron-Ruault,¹⁹ Julie E. Buring,^20,²¹ Federico Canzian,²² Sandra Clipp,²³ Miren Dorronsoro,²⁴ J. Michael Gaziano,²⁵ Edward L. Giovannucci,^5,^26,²⁷ Susan E. Hankinson,^5,²⁷ Patricia Hartge,³ Robert N. Hoover,³ David J. Hunter,^5,²⁷ Kevin B. Jacobs,^3,^28,²⁹ Mazda Jenab,³⁰ Peter Kraft,²⁷ Charles Kooperberg,⁶ Shannon M. Lynch,³¹ Malin Sund,³² Julie B. Mendelsohn,³ Tracy Mouw,³³ Christina C. Newton,¹ Kim Overvad,³⁴ Domenico Palli,³⁵ Petra H.M. Peeters,^33,³⁶ Aleksandar Rajkovic,³⁷ Xiao-Ou Shu,¹⁶ Gilles Thomas,³ Geoffrey S. Tobias,³ Dimitrios Trichopoulos,^27,³⁸ Jarmo Virtamo,³⁹ Jean Wactawski-Wende,⁴⁰ Brian M. Wolpin,^4,⁵ Kai Yu,³ and Anne Zeleniuch-Jacquotte^10,¹¹

¹Department of Epidemiology, American Cancer Society, Atlanta, Georgia

²Laboratory of Translational Genomics, Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD

³Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD

⁴Department of Medical Oncology, Dana-Farber Cancer Institute. Boston, MA

⁵Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA

⁶Fred Hutchinson Cancer Research Center, Division of Public Health Sciences

⁷Department of Oncology, College of Medicine, Mayo Clinic, Rochester, MN

⁸Information Management Services, Silver Spring, MD

⁹Department of Obstetrics and Gynecology, New York University School of Medicine

¹⁰Department of Environmental Medicine, New York University School of Medicine

¹¹NYU Cancer Institute, New York, NY

¹²National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands, and Department of Gastroenterology and Hepatology, University Medical Center Utrecht, Utrecht, The Netherlands

¹³Department of Laboratory Medicine/Pathology, School of Medicine, University of Minnesota, Minneapolis, MN

¹⁴Prevention and Research Center, Mercy Medical Center, Baltimore, MD

¹⁵Department of Health Sciences Research, Mayo Clinic, Rochester, MN

¹⁶Department of Medicine and Vanderbilt-Ingram Cancer Center, Vanderbilt University, Nashville, TN

¹⁷Dept. of Epidemiology and Population Health, Albert Einstein College of Medicine, Bronx, NY

¹⁸Cancer Epidemiology Unit, University of Oxford, Oxford, UK

¹⁹Inserm, (Institut National de la Santé et de la Recherche Médicale) and Institut Gustave Roussy, Villejuif, France

²⁰Divisions of Preventive Medicine and Aging, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA

²¹Department of Ambulatory Care and Prevention, Harvard Medical School, Boston, MA

²²Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

²³George W. Comstock Center for Public Health Research and Prevention, Hagerstown, MD

²⁴Public Health Division of Gipuzkoa and Ciberespana, Spain

²⁵Physicians' Health Study, Divisions of Aging, Cardiovascular Disease, and Preventive Medicine, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA and Massachusetts Veterans Epidemiology Research and Information Center, Veterans Affairs Boston Healthcare System, Boston, MA

²⁶Department of Nutrition, Harvard School of Public Health, Boston, MA

²⁷Department of Epidemiology, Harvard School of Public Health, Boston, MA

²⁸BioInformed LLC, Gaithersburg, MD

²⁹Core Genotyping Facility, Advanced Technology Program, SAIC-Frederick Inc., NCI-Frederick, Frederick, MD

³⁰International Agency for Research on Cancer, Lyon, France

³¹Division of Cancer Control and Population Science, National Cancer Institute, National Institutes of Health, Bethesda, MD

³²Department of Surgical and Perioperative Sciences, Surgery, Umeå University, Umeå, Sweden

³³Department of Epidemiology and Public Health, Imperial College, London, UK

³⁴Department of Cardiology and Department of Clinical Epidemiology, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark

³⁵Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute - ISPO Florence, Italy

³⁶Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands

³⁷Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, TX

³⁸Department of Hygiene and Epidemiology, University of Athens Medical School, Athens, Greece

³⁹Department of Chronic Disease Prevention, National Institute for Health and Welfare,Helsinki, Finland

⁴⁰Department of Social and Preventive Medicine, University at Buffalo, SUNY, Buffalo, NY

^*Writing committee member

^†Steering Committee Member

The publisher's final edited version of this article is available at Int J Cancer

See other articles in PMC that cite the published article.

Abstract

A family history of pancreatic cancer has consistently been associated with increased risk of pancreatic cancer. However, uncertainty remains about the strength of this association. Results from previous studies suggest a family history of select cancers (i.e. ovarian, breast, and colorectal) could also be associated, although not as strongly, with increased risk of pancreatic cancer. We examined the association between a family history of five types of cancer (pancreas, prostate, ovarian, breast, and colorectal) and risk of pancreatic cancer using data from a collaborative nested case-control study conducted by the Pancreatic Cancer Cohort Consortium. Cases and controls were from cohort studies from the United States, Europe, and China, and a case-control study from the Mayo Clinic. Analyses of family history of pancreatic cancer included 1,183 cases and 1,205 controls. A family history of pancreatic cancer in a parent, sibling, or child was associated with increased risk of pancreatic cancer (multivariate-adjusted OR = 1.76, 95% CI 1.19–2.61). A family history of prostate cancer was also associated with increased risk (OR = 1.45, 95% CI 1.12–1.89). There were no statistically significant associations with a family history of ovarian cancer (OR = 0.82, 95% CI 0.52–1.31), breast cancer (OR = 1.21, 95% CI 0.97–1.51), or colorectal cancer (OR = 1.17, 95% CI 0.93–1.47). Our results confirm a moderate sized association between a family history of pancreatic cancer and risk of pancreatic cancer and also provide evidence for an association with a family history of prostate cancer worth further study.

INTRODUCTION

A family history of pancreatic cancer is considered an established risk factor for pancreatic cancer¹. However, uncertainty remains about the size of the relative risk associated with a family history of pancreatic cancer in first degree relatives. Relative risks less than 2.0 were reported by the two largest studies to date, a U.S. cohort study (RR=1.7, 95% CI 1.4–1.9)² and a Swedish registry-based analysis (RR=1.7, 95% CI 1.2–2.4).³ However, nearly all other studies have reported relative risks of 2.0 or more. An Icelandic registry-based analysis reported a relative risk of 2.3 (90% CI 1.8–3.0)⁴ and relative risk estimates from case-control studies have ranged from 1.9 to 5.0.⁵^–¹²

A family history of ovarian, colorectal, or breast cancer has also been associated with increased risk of pancreatic cancer in some studies. Two⁴^,⁶ of seven studies²^–⁴^,⁶^,⁸^–¹⁰ that examined family history of ovarian cancer reported a statistically significant increased risk of pancreatic cancer, as did four²^,⁴^,⁶^,¹⁰ of eight studies²^–⁴^,⁶^,⁸^–¹⁰^,¹³ that examined family history of colorectal cancer. Only one¹⁰ of eight studies²^–⁴^,⁶^,⁸^–¹⁰^,¹³ that examined family history of breast cancer found significantly increased risk, although there was some suggestion of modestly increased risk in two others.⁶^,¹³ A family history of prostate cancer has not been associated with increased risk of pancreatic cancers in previous studies.²^–⁴^,⁶^,⁸^–¹⁰ It should be noted that with the exception of the U.S. cohort study,² which included 7,306 pancreatic cancer deaths, and the Swedish registry-based analysis,³ which included 1,254 cases of pancreatic cancer, all of these studies of family history of cancer included fewer than 1,000 cases. Given the low prevalence of a family history of each specific type of cancer, some of these studies likely had limited statistical power to detect modest associations. Identifying associations between family history of specific cancers and risk of pancreatic cancer could provide clues to both environmental and genetic factors contributing to pancreatic cancer etiology.

We pooled data from several prospective cohort studies and one rapid ascertainment case-control study participating in the Pancreatic Cancer Cohort Consortium (PanScan) in order to examine the association between a family history of pancreatic and certain other cancers and risk of pancreatic cancer.

MATERIALS AND METHODS

Study populations

The PanScan study was assembled in order to conduct a genome-wide association study, and includes pancreatic cancer cases and matched controls from twelve prospective cohort studies and one ultra-rapid ascertainment case-control study. This analysis includes cases and controls in PanScan that were from studies that had family history data available, namely the Mayo Clinic Molecular Epidemiology of Pancreatic Cancer Study case-control study (Mayo),¹⁴ and the following ten cohorts; the Alpha-Tocopherol, Beta-Carotene Prevention Study (ATBC),¹⁵ the Give Us a Clue to Cancer and Heart Disease Study (Clue II),¹⁶ the Cancer Prevention Study Nutrition Cohort (CPS II),¹⁷ the European Prospective Investigation into Cancer and Nutrition (EPIC),¹⁸ the Health Professionals Follow-up Study (HPFS),¹⁹ the Nurses' Health Study (NHS),²⁰ the New York University Women's Health Study (NYU-WHS),²¹ the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO),²² the Shanghai Men's and Women's Health Study (SMWHS),²³^,²⁴ and the Women's Health Initiative (WHI).²⁵

Six studies (CPS-II, Clue II, Mayo, NHS, PLCO and SMWHS) had sufficient family history data to be included in analyses of all of the types of family history of cancer that we examined. Other studies could be included only in the specific analyses for which appropriate family history data was available. ATBC was included in analyses of family history of pancreatic, colorectal and prostate cancer, HPFS was included in analyses of family history of breast and prostate cancer, NYU-WHS was included in analyses of family history of breast and colorectal cancer, WHI was included in analyses of family history of ovarian, prostate, breast and colorectal cancer, and specific EPIC study centers that had collected information about family history of breast cancer were included in analyses of family history of breast cancer. We did not examine family history of cancers other than pancreatic, prostate, ovarian, colorectal and breast cancer because few studies had information on family history of other cancers.

Cases from the Mayo case-control study included in this pooled analysis were included in a previous analysis of family history and pancreatic cancer.⁹ In addition, most cases from the CPS-II Nutrition Cohort included in this analysis were included in a previous analysis of family history and pancreatic cancer mortality.² However, the previous analysis used family history information reported by CPS-II participants in 1982 whereas this analysis used more up to date family history information reported between 1997 and 2001.

The Special Studies Institutional Review Board (SSIRB) of the National Cancer Institute approved the pooled PanScan study. Each study was approved by its local institutional review board (IRB).

Case ascertainment and control selection

Cases included all incident primary pancreatic adenocarcinoma (ICD-O-3 code C250–C259 or C25.0–C25.3, C25.7–C25.9). We excluded known non-exocrine pancreatic tumors (C25.4, histology type, 8150, 8151, 8153, 8155, 8240, 8246). Cases were confirmed through cancer registries (ATBC, CPS II, EPIC, HPFS, NHS, NYU-WHS, SMWHS), death certificates (CPS II, EPIC, HPFS, NHS), or review of medical records by medical personnel (ATBC, CPS II, EPIC, HPFS, Mayo, NHS, NYU-WHS, PLCO, WHI, SMWHS).

Controls were matched in a 1:1 ratio to cases within each study on calendar year of birth (± 5 years), sex, and race and were alive and free of pancreatic cancer on the incidence date of the matched case. Some studies also matched on other factors including age at baseline or age at blood draw (±5 years), date/time of day of blood draw, fasting status at blood draw, and smoking status.

Ascertainment of family history

Participants in all studies self-reported family history of different cancers on self-administered written questionnaires. Some studies ascertained family history only in parents and siblings (ATBC, EPIC, NHS, HPFS, NYU-WHS), others also included children (CPS II, CLUE II, Mayo, PLCO, SMWHS, WHI). The Mayo study also ascertained family history beyond first degree relatives (e.g. grandparents, aunts and uncles). For this analysis, we considered only family history in first degree relatives (parents, siblings, and children).

Statistical analysis

Our primary analyses were based on a dichotomous variable (yes/no) for family history of each type of cancer. We also examined family history by number of affected relatives (one/two or more) and by age at diagnosis of the youngest affected relative. Age at relative's diagnosis was not ascertained in the ATBC study, and was sometimes left missing by participants in other studies. Therefore, analyses by age at relative's diagnosis include considerably fewer subjects than other analyses. For family history of breast and colorectal cancer we used category cut-points of 45 and 55 years, respectively, in order to be able to include data from the WHI study in our analysis. The WHI questionnaire did not ask for exact age of relative's cancer diagnosis, but instead asked participants to report whether or not their relative was diagnosed before or after age 45 (for breast cancer), or before or after age 55 (for colorectal cancer).

We calculated odds ratios (OR) and 95% confidence intervals (95% CI) for pancreatic cancer risk using unconditional logistic regression. We used unconditional logistic regression rather than a matched pair analysis using conditional logistic regression in order to avoid losing information from matched sets in which one subject had missing family history data. Results, however, were similar in analyses using conditional logistic regression. Unless otherwise specified, models were adjusted for age (< 50, 50 –54, 55–59, 60–64, 65–69, 70–74–75–79, ≥ 80), study, sex, cigarette smoking status (never, former quit ≥10 years, former quit <10 years, current < 20 cigarettes/day, current 20 cigarettes/day, current > 20 cigarettes/day, unknown), race (Caucasian, African, Asian, other, unknown), BMI (< 25, 25 – < 30, ≥ 30, unknown), and self-reported diabetes (yes, no, missing). Due to small numbers, models that included only one study (study specific models) included the adjustment variables above except that they were not adjusted for race, and used less detailed categories for age (10 year categories) and cigarette smoking(never/former/current).

We modeled multiplicative interaction terms between a dichotomous variable for family history of each cancer and age at diagnosis (< 60, ≥ 60), BMI (< 25, ≥ 25), sex, and cigarette smoking status (current, not current), all common and established risk factors for pancreatic cancer. We also evaluated heterogeneity across studies by fitting models with and without multiplicative interaction terms between family history and study and calculating a p-value for heterogeneity using the likelihood ratio statistic.²⁶

RESULTS

Demographic characteristics and numbers of cases and controls are shown by study in Table 1. The numbers of cases and controls from each study shown in Table 1 are the numbers included in analyses of family history of pancreatic cancer, or (for studies that did not collect information on family history of pancreatic cancer) the numbers that were included in analyses of family history of breast cancer. Similar numbers of cases and controls from each study were generally included in analyses of other types of family history of cancer. Most cases and controls were from the United States and were Caucasian. The overall prevalence of a family history of pancreatic cancer among controls in whom this information was available was 3.6 %, although prevalence varied considerably by study, with the highest prevalence (6.9 %) observed in the CPS-II Nutrition Cohort, the oldest study population in PanScan.

Table 1

Characteristics of study populations included in Pancreatic Cancer Cohort Consortium (PanScan) family history analyses

A family history of pancreatic cancer in a first degree relative was associated with moderately increased risk of pancreatic cancer (OR = 1.76, 95% CI 1.19–2.61) (Table 2). The population attributable fraction associated with this odds ratio is 3.0 %.²⁷ It should be noted that the population attributable fraction for family history does not fully represent the population effect of genetic factors because higher risk genetic variants may often be present in individuals with no family history. Results adjusted only for age, sex, race, and study were similar to multivariate-adjusted results (OR=1.82, 95% CI 1.24–2.68). Having two relatives with a family history of pancreatic cancer was rare, so the risk associated with having two affected relatives could not be assessed with precision. Results are shown by study in Figure 1. There was no evidence of heterogeneity across studies (p=0.27).

Table 2

Risk of pancreatic cancer by family history of various cancers and by number and age at diagnosis of affected relatives¹

Figure 1

Odds Ratios for Pancreatic Cancer Associated with Family History of Pancreatic Cancer by Study¹

A family history of prostate cancer was also associated with increased risk of pancreatic cancer (OR = 1.45, 95% CI 1.12–1.89) (Table 2). There was evidence of heterogeneity across studies in the odds ratio for family history of prostate cancer (p = 0.03 for heterogeneity). In analyses by study (not shown in table), a family history of prostate cancer was associated with a statistically significant increase in risk in PLCO (OR = 2.99, 95% CI 1.28–6.98) and WHI (OR = 2.88, 95% CI 1.48–5.59), a suggestion of increased risk in HPFS (OR = 2.46, 95% CI 0.88–6.92), and no suggestion of increased risk in the remaining six studies (odds ratios ranging from 0.4 to 1.2). The increase in risk of pancreatic cancer associated with a family history of prostate cancer differed by age at pancreatic cancer diagnosis (p = 0.01 for interaction), with a statistically significant increase in risk of pancreatic cancer diagnosed at age 60 or older (RR = 1.65, 95% CI 1.25–2.18), but no suggestion of increased risk of pancreatic cancer diagnosed before age 60 (OR = 0.64, 95% CI 0.29–1.44). However, numbers for this analysis were limited since relatively few pancreatic cancer cases were diagnosed before age 60. There was no evidence of particularly increased risk with having two affected relatives or with having a relative diagnosed with prostate cancer before age 60 (Table 2).

A family history of ovarian cancer was not associated with risk of pancreatic cancer (OR = 0.82, 95% CI 0.52–1.31). There was a suggestion of increased risk of pancreatic cancer associated with a family history of colorectal cancer (OR = 1.17, 95% CI 0.93–1.47), and breast cancer (OR = 1.21, 95% CI 0.97–1.51), but neither of these associations was statistically significant.

We also examined results specifically among the cohort studies (excluding the Mayo case-control study). Among the cohort studies, odds ratios associated with of risk of pancreatic cancer were 1.78 (95% CI 1.07–2.97) for a family history of pancreatic cancer, 1.51 (95% CI 1.10–2.08) for a family history of prostate cancer, 0.70 (95% CI 0.42–1.18) for a family history of ovarian cancer, 1.10 (95% CI 0.84–1.44) for a family history of colorectal cancer, and 1.19 (95% CI 0.92–1.55) for a family history of breast cancer. We also calculated p-values for heterogeneity of associations across all studies, as noted in the methods section. With the exception of family history of prostate cancer (p=0.03 as noted above) we found no statistical evidence for heterogeneity across studies. P values for heterogeneity across studies for family history of pancreatic, ovarian, colorectal and breast cancer were 0.27, 0.35, 0.57 and 0.43, respectively.

Because we hypothesized family history associations might be substantially stronger for earlier onset pancreatic cancer (diagnosed before age 60), we examined results specifically among those diagnosed under age 60, although numbers for this analysis were limited. Odds ratios for pancreatic cancer before the age of 60 were 0.86 (95% CI 0.25–3.00) for a family history of pancreatic cancer, 0.64 (95% CI 0.29–1.44) for a family history of prostate cancer, 1.21(95% CI 0.57–2.54) for a family history of breast cancer, 6.20 (95% CI 0.69–55.54) for a family history of ovarian cancer, and 1.41(95% CI 0.64–3.09) for a family history of colorectal cancer.

We found no statistically significant interactions between sex, smoking status, or body mass index and family history of any of the cancers we examined. Odds ratios for family history of pancreatic cancer were 2.11 (95% CI 1.15–3.85) among never smokers, 1.25 (95% CI 0.67–2.35) among former smokers, and 2.84 (95% CI 1.06–7.59) among current smokers.

DISCUSSION

This large pooled analysis confirms that a family history of pancreatic cancer is associated with moderately increased risk of pancreatic cancer. The size of the increase in risk we observed (OR = 1.8) is consistent with results from the two largest studies to date, a large U.S. cohort study² and a Swedish registry based analysis,³ which both reported relatives risk of 1.7. Our results are less consistent with the relative risk estimates of 2.5 or greater observed in some previous case-control studies.⁵^–⁸^,¹⁰^,¹¹

The association between a family history of pancreatic cancer and risk of pancreatic cancer is likely due to genetic and/or other risk factors that are shared by family members. Relatively little is known about the specific genetic factors that might contribute to associations between family history of cancer and risk of pancreatic cancer. Inherited truncating mutations in the BRCA2 gene increase risk of pancreatic cancer,²⁸ but these mutations have been detected in only about 6 percent of pancreatic cancer patients with a family history of pancreatic cancer.²⁹ Inherited mutations in the DNA mismatch-repair genes responsible for hereditary non-polyposis colorectal cancer (HNPCC) are also associated with increased risk of pancreatic cancer,²⁸^,³⁰ but these mutations are rare in the general population. Peutz-Jeghers syndrome (caused by an inherited mutation in LKB1), hereditary pancreatitis (caused by an inherited mutation in PRSS1), and inherited mutations in CDKN2A (responsible for familial atypical multiple mole melanoma syndrome), increase risk of pancreatic cancer, but are very rare.²⁸ In addition to these rare mutations, common polymorphisms in the ABO gene have recently been shown to be associated with modestly increased risk of pancreatic cancer.³¹ Nongenetic risk factors that are shared between family members could also have contributed to the association we observed. However, in our analysis, adjustment for nongenetic pancreatic risk factors, including obesity, diabetes and smoking, did not attenuate the association between a family history of pancreatic cancer and risk of pancreatic cancer, indicating these risk factors are unlikely to explain the association with family history.

A family history of prostate cancer was associated with a notable increase in risk of pancreatic cancer in this analysis (OR = 1.45, 95% CI 1.12–1.89). The magnitude of this association was unexpected, given that a family history of prostate cancer has not been associated with significant increases in risk of pancreatic cancer in seven previous studies.²^–⁴^,⁶^,⁸^–¹⁰ It should be noted, however, that previous studies do not preclude a small increase in risk associated with a family history of prostate cancer. In the three largest previous studies,²^–⁴ a family history of prostate cancer was associated with nonstatistically significant relative risks ranging between 1.06 and 1.13. In this analysis, the association between a family history of prostate cancer and risk of pancreatic cancer was most apparent in the WHI and PLCO studies. We know of no clear explanation why results from these two studies with respect to a family history of prostate cancer would differ from those of other studies included in this pooled analysis. Like the WHI and PLCO cohorts, the NHS and CPS-II cohorts are composed mostly of well educated older adults from the United States, but there was no suggestion of any association between a family history of prostate cancer and risk of pancreatic cancer in either NHS or CPS-II. Chance may be at least partly responsible for the association with a family history of prostate cancer observed in our study. Further examination of this association between a family history of prostate cancer and risk of pancreatic cancer in other large studies may be informative.

A family history of ovarian cancer was not associated with increased risk of pancreatic cancer in our analysis (OR = 0.82, 95% CI 0.52–1.31). Our null results are similar to those from a large Swedish registry-based analysis,³ a large U.S. cohort study,² and two U.S. case-control studies.⁸^,⁹ However, a strong increase in risk of pancreatic cancer was observed in a U.S. population-based case-control study (OR = 5.3, 95% CI 1.4–20.2),⁶ and in a large registry-based analysis from Iceland (RR =1.7, 90% CI 1.3–2.1).⁴ Family history of ovarian cancer may be associated with increased risk of pancreatic cancer in the Icelandic population because Icelanders have a unusually high prevalence of BRCA2 truncating mutations,³² which increase risk of both ovarian³³ and pancreatic cancer.³⁴^,³⁵

A family history of colorectal cancer was not associated with a statistically significant increase in risk of pancreatic cancer in our study. However, our risk estimate (OR = 1.2, 95% CI 0.9–1.5) does not preclude a weak association. Previous studies of a family history of colorectal have yielded mixed results. Significantly increased risk of pancreatic cancer was observed in a two U.S. case-control studies,⁶^,¹⁰ and in a registry-based analysis from Iceland.⁴ A statistically significant increase in risk was also observed in a large U.S. cohort study, but the increase in risk appeared quite small (RR=1.12, 95% CI 1.01–1.23).² No association was observed in a large registry-based analysis from Sweden.³ To date, the totality of epidemiologic evidence does not support an important association between a family history of colorectal cancer and risk of pancreatic cancer.

A family history of breast cancer was also not associated with a statistically significant increase in risk in our study, although our risk estimate (OR = 1.2, 95% CI 1.0–1.5) does not preclude a weak association. A family history of breast cancer was associated with increased risk in one U.S. case-control study.¹⁰ However, the absence of increased risk in our study is consistent with results from most previous studies of family history of breast cancer.²^–⁴^,⁶^,⁸^,⁹^,¹³

A strength of this analysis is its relatively large size, which is important given the low prevalence of family histories of individual types of cancer. Although not as large as a recent analysis of family history and pancreatic cancer mortality in a U.S. cohort,² this pooled analysis is among the largest studies of family history to date, and includes a somewhat geographically diverse population, with participants from Europe, China, and the United States. Limitations of this analysis include the fact that family history of cancer was self-reported and some misclassification of family history can be expected as a result.³⁶ In cohort studies, which contributed most of the cases and controls in this analysis, misclassification of family history would be expected to be nondifferential with respect to later development of pancreatic cancer, and would therefore contribute to underestimating the strength of associations with family history. In addition, participants in the cohort studies included in this analysis may have reported family history several years before their diagnosis date. A small proportion of participants who had no family history of a particular type of cancer at the time family history was assessed would have acquired a family history of this type of cancer during study follow-up. However, this proportion is likely to have been too small to have meaningfully influenced our relative risk estimates. An additional limitation is that numbers were limited for analyses by number and age of affected relatives.

In summary, results of this large pooled analysis indicate that a family history of pancreatic cancer is associated with moderately increased risk of pancreatic cancer. The association between family history of prostate cancer and increased risk of pancreatic cancer observed in our study should be interpreted cautiously and requires careful examination in future analyses of family history.

References

1. Hart AR, Kennedy H, Harvey I. Pancreatic cancer: a review of the evidence on causation. Clin Gastroenterol Hepatol. 2008;6:275–82. [PubMed]

2. Jacobs EJ, Rodriguez C, Newton CC, Bain EB, Patel AV, Feigelson HS, Thun MJ, Calle EE. Family history of various cancers and pancreatic cancer mortality in a large cohort. Cancer Causes Control. 2009;20:1261–69. [PubMed]

3. Hemminki K, Li X. Familial and second primary pancreatic cancers: a nationwide epidemiologic study from Sweden. Int J Cancer. 2003;103:525–30. [PubMed]

4. Amundadottir LT, Thorvaldsson S, Gudbjartsson DF, Sulem P, Kristjansson K, Arnason S, Gulcher JR, Bjornsson J, Kong A, Thorsteinsdottir U, Stefansson K. Cancer as a complex phenotype: pattern of cancer distribution within and beyond the nuclear family. PLoS Med. 2004;1:e65. [PMC free article] [PubMed]

5. Fernandez E, La Vecchia C, D'Avanzo B, Negri E, Franceschi S. Family history and the risk of liver, gallbladder, and pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 1994;3:209–12. [PubMed]

6. Silverman DT, Schiffman M, Everhart J, Goldstein A, Lillemoe KD, Swanson GM, Schwartz AG, Brown LM, Greenberg RS, Schoenberg JB, Pottern LM, Hoover RN, et al. Diabetes mellitus, other medical conditions and familial history of cancer as risk factors for pancreatic cancer. Br J Cancer. 1999;80:1830–7. [PMC free article] [PubMed]

7. Schenk M, Schwartz AG, O'Neal E, Kinnard M, Greenson JK, Fryzek JP, Ying GS, Garabrant DH. Familial risk of pancreatic cancer. J Natl Cancer Inst. 2001;93:640–4. [PubMed]

8. Ghadirian P, Liu G, Gallinger S, Schmocker B, Paradis AJ, Lal G, Brunet JS, Foulkes WD, Narod SA. Risk of pancreatic cancer among individuals with a family history of cancer of the pancreas. Int J Cancer. 2002;97:807–10. [PubMed]

9. McWilliams RR, Rabe KG, Olswold C, De Andrade M, Petersen GM. Risk of malignancy in first-degree relatives of patients with pancreatic carcinoma. Cancer. 2005;104:388–94. [PubMed]

10. Hassan MM, Bondy ML, Wolff RA, Abbruzzese JL, Vauthey JN, Pisters PW, Evans DB, Khan R, Chou TH, Lenzi R, Jiao L, Li D. Risk factors for pancreatic cancer: case-control study. Am J Gastroenterol. 2007;102:2696–707. [PMC free article] [PubMed]

11. Del Chiaro M, Zerbi A, Falconi M, Bertacca L, Polese M, Sartori N, Boggi U, Casari G, Longoni BM, Salvia R, Caligo MA, Di Carlo V, et al. Cancer risk among the relatives of patients with pancreatic ductal adenocarcinoma. Pancreatology. 2007;7:459–69. [PubMed]

12. Inoue M, Tajima K, Takezaki T, Hamajima N, Hirose K, Ito H, Tominaga S. Epidemiology of pancreatic cancer in Japan: a nested case-control study from the Hospital-based Epidemiologic Research Program at Aichi Cancer Center (HERPACC) Int J Epidemiol. 2003;32:257–62. [PubMed]

13. Poole CA, Byers T, Calle EE, Bondy J, Fain P, Rodriguez C. Influence of a family history of cancer within and across multiple sites on patterns of cancer mortality risk for women. Am J Epidemiol. 1999;149:454–62. [PubMed]

14. McWilliams RR, Bamlet WR, Cunningham JM, Goode EL, de Andrade M, Boardman LA, Petersen GM. Polymorphisms in DNA repair genes, smoking, and pancreatic adenocarcinoma risk. Cancer Res. 2008;68:4928–35. [PMC free article] [PubMed]

15. The alpha-tocopherol, beta-carotene lung cancer prevention study: design, methods, participant characteristics, and compliance. The ATBC Cancer Prevention Study Group. Ann Epidemiol. 1994;4:1–10. [PubMed]

16. Gallicchio L, Chang H, Christo DK, Thuita L, Huang HY, Strickland P, Ruczinski I, Hoffman SC, Helzlsouer KJ. Single nucleotide polymorphisms in inflammation-related genes and mortality in a community-based cohort in Washington County, Maryland. Am J Epidemiol. 2008;167:807–13. [PubMed]

17. Calle EE, Rodriguez C, Jacobs EJ, Almon ML, Chao A, McCullough ML, Feigelson HS, Thun MJ. The American Cancer Society Cancer Prevention Study II Nutrition Cohort - rationale, study design and baseline characteristics. Cancer. 2002;94:2490–501. [PubMed]

18. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M, Charrondiere UR, Hemon B, Casagrande C, Vignat J, Overvad K, Tjonneland A, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr. 2002;5:1113–24. [PubMed]

19. Wei EK, Giovannucci E, Fuchs CS, Willett WC, Mantzoros CS. Low plasma adiponectin levels and risk of colorectal cancer in men: a prospective study. J Natl Cancer Inst. 2005;97:1688–94. [PubMed]

20. Colditz GA, Hankinson SE. The Nurses' Health Study: lifestyle and health among women. Nat Rev Cancer. 2005;5:388–96. [PubMed]

21. Zeleniuch-Jacquotte A, Gu Y, Shore RE, Koenig KL, Arslan AA, Kato I, Rinaldi S, Kaaks R, Toniolo P. Postmenopausal levels of sex hormones and risk of breast carcinoma in situ: results of a prospective study. Int J Cancer. 2005;114:323–7. [PubMed]

22. Hayes RB, Sigurdson A, Moore L, Peters U, Huang WY, Pinsky P, Reding D, Gelmann EP, Rothman N, Pfeiffer RM, Hoover RN, Berg CD. Methods for etiologic and early marker investigations in the PLCO trial. Mutat Res. 2005;592:147–54. [PubMed]

23. Zheng W, Chow WH, Yang G, Jin F, Rothman N, Blair A, Li HL, Wen W, Ji BT, Li Q, Shu XO, Gao YT. The Shanghai Women's Health Study: rationale, study design, and baseline characteristics. Am J Epidemiol. 2005;162:1123–31. [PubMed]

24. Xu WH, Zhang XL, Gao YT, Xiang YB, Gao LF, Zheng W, Shu XO. Joint effect of cigarette smoking and alcohol consumption on mortality. Prev Med. 2007;45:313–9. [PMC free article] [PubMed]

25. Anderson GL, Manson J, Wallace R, Lund B, Hall D, Davis S, Shumaker S, Wang CY, Stein E, Prentice RL. Implementation of the Women's Health Initiative study design. Ann Epidemiol. 2003;13:S5–17. [PubMed]

26. Hosmer DW, Lemeshow S. Applied Logistic Regression. John Wiley and Sons; New York: 1989.

27. Rockhill B, Weinberg CR, Newman B. Population attributable fraction estimation for established breast cancer risk factors: considering the issues of high prevalence and unmodifiability. Am J Epidemiol. 1998;147:826–33. [PubMed]

28. Habbe N, Langer P, Sina-Frey M, Bartsch DK. Familial pancreatic cancer syndromes. Endocrinol Metab Clin North Am. 2006;35:417–30. xi. [PubMed]

29. Couch FJ, Johnson MR, Rabe KG, Brune K, de Andrade M, Goggins M, Rothenmund H, Gallinger S, Klein A, Petersen GM, Hruban RH. The prevalence of BRCA2 mutations in familial pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 2007;16:342–6. [PubMed]

30. Geary J, Sasieni P, Houlston R, Izatt L, Eeles R, Payne SJ, Fisher S, Hodgson SV. Gene-related cancer spectrum in families with hereditary non-polyposis colorectal cancer (HNPCC) Fam Cancer. 2008;7:163–72. [PubMed]

31. Amundadottir L, Kraft P, Stolzenberg-Solomon RZ, Fuchs CS, Petersen GM, Arslan AA, Bueno-de-Mesquita HB, Gross M, Helzlsouer K, Jacobs EJ, LaCroix A, Zheng W, et al. Genome-wide association study identifies variants in the ABO locus associated with susceptibility to pancreatic cancer. Nat Genet. 2009;41:986–90. [PMC free article] [PubMed]

32. Thorlacius S, Sigurdsson S, Bjarnadottir H, Olafsdottir G, Jonasson JG, Tryggvadottir L, Tulinius H, Eyfjord JE. Study of a single BRCA2 mutation with high carrier frequency in a small population. Am J Hum Genet. 1997;60:1079–84. [PubMed]

33. Johannesdottir G, Gudmundsson J, Bergthorsson JT, Arason A, Agnarsson BA, Eiriksdottir G, Johannsson OT, Borg A, Ingvarsson S, Easton DF, Egilsson V, Barkardottir RB. High prevalence of the 999del5 mutation in icelandic breast and ovarian cancer patients. Cancer Res. 1996;56:3663–5. [PubMed]

34. Lal G, Liu G, Schmocker B, Kaurah P, Ozcelik H, Narod SA, Redston M, Gallinger S. Inherited predisposition to pancreatic adenocarcinoma: role of family history and germ-line p16, BRCA1, and BRCA2 mutations. Cancer Res. 2000;60:409–16. [PubMed]

35. Thorlacius S, Olafsdottir G, Tryggvadottir L, Neuhausen S, Jonasson JG, Tavtigian SV, Tulinius H, Ogmundsdottir HM, Eyfjord JE. A single BRCA2 mutation in male and female breast cancer families from Iceland with varied cancer phenotypes. Nat Genet. 1996;13:117–9. [PubMed]

36. Murff HJ, Spigel DR, Syngal S. Does this patient have a family history of cancer? An evidence-based analysis of the accuracy of family cancer history. Jama. 2004;292:1480–9. [PubMed]