Are meat and heme iron intake associated with pancreatic cancer? Results from the NIH-AARP diet and health cohort

International Journal of CancerVolume 138, Issue 9 p. 2172-2189 Cancer EpidemiologyFree Access Are meat and heme iron intake associated with pancreatic cancer? Results from the NIH-AARP diet and health cohort Pulkit Taunk, Corresponding Author Pulkit Taunk Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL Division of Cancer Epidemiology and Genetics, Department of Health and Human Services, National Cancer Institute, National Institutes of Health, Rockville, MDCorrespondence to: Rachael Stolzenberg-Solomon, PhD, 9609 Medical Center Drive, Room 6E420, Rockville, MD 20850, USA, Tel.: +240-276-7224, Fax: +240-276-7837, E-mail: rs221z@nih.gov or Pulkit Taunk, 4050 Presidential Drive, Palm Harbor, FL 34685, USA, Tel.: +727-385-6725, E-mail: pt44@med.miami.eduSearch for more papers by this authorEric Hecht, Eric Hecht Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FLSearch for more papers by this authorRachael Stolzenberg-Solomon, Corresponding Author Rachael Stolzenberg-Solomon Division of Cancer Epidemiology and Genetics, Department of Health and Human Services, National Cancer Institute, National Institutes of Health, Rockville, MDCorrespondence to: Rachael Stolzenberg-Solomon, PhD, 9609 Medical Center Drive, Room 6E420, Rockville, MD 20850, USA, Tel.: +240-276-7224, Fax: +240-276-7837, E-mail: rs221z@nih.gov or Pulkit Taunk, 4050 Presidential Drive, Palm Harbor, FL 34685, USA, Tel.: +727-385-6725, E-mail: pt44@med.miami.eduSearch for more papers by this author Pulkit Taunk, Corresponding Author Pulkit Taunk Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL Division of Cancer Epidemiology and Genetics, Department of Health and Human Services, National Cancer Institute, National Institutes of Health, Rockville, MDCorrespondence to: Rachael Stolzenberg-Solomon, PhD, 9609 Medical Center Drive, Room 6E420, Rockville, MD 20850, USA, Tel.: +240-276-7224, Fax: +240-276-7837, E-mail: rs221z@nih.gov or Pulkit Taunk, 4050 Presidential Drive, Palm Harbor, FL 34685, USA, Tel.: +727-385-6725, E-mail: pt44@med.miami.eduSearch for more papers by this authorEric Hecht, Eric Hecht Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FLSearch for more papers by this authorRachael Stolzenberg-Solomon, Corresponding Author Rachael Stolzenberg-Solomon Division of Cancer Epidemiology and Genetics, Department of Health and Human Services, National Cancer Institute, National Institutes of Health, Rockville, MDCorrespondence to: Rachael Stolzenberg-Solomon, PhD, 9609 Medical Center Drive, Room 6E420, Rockville, MD 20850, USA, Tel.: +240-276-7224, Fax: +240-276-7837, E-mail: rs221z@nih.gov or Pulkit Taunk, 4050 Presidential Drive, Palm Harbor, FL 34685, USA, Tel.: +727-385-6725, E-mail: pt44@med.miami.eduSearch for more papers by this author First published: 15 December 2015 https://doi.org/10.1002/ijc.29964Citations: 37 Conflicts of interest: : Nothing to report AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Several studies on pancreatic cancer have reported significant positive associations for intake of red meat but null associations for heme iron. We assessed total, red, white and processed meat intake, meat cooking methods and doneness and heme iron and mutagen intake in relation to pancreatic cancer in the NIH-AARP Diet and Health Study cohort. A total of 322,846 participants (187,265 men and 135,581 women) successfully completed and returned the food frequency questionnaire between 1995 and 1996. After a mean follow-up of 9.2 years (up to 10.17 years), 1,417 individuals (895 men and 522 women) developed exocrine pancreatic cancer. Cox proportional hazard models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs), and trends were calculated using the median value of each quantile. Models incorporated age as the time metric and were adjusted for smoking history, body mass index, self-reported diabetes and energy-adjusted saturated fat. Pancreatic cancer risk significantly increased with intake of total meat (Q5 vs. Q1: HR = 1.20, 95% CI 1.02–1.42, p-trend = 0.03), red meat (HR = 1.22, 95% CI 1.01–1.48, p-trend = 0.02), high-temperature cooked meat (HR = 1.21, 95% CI 1.00–1.45, p-trend = 0.02), grilled/barbequed meat (HR = 1.24, 95% CI 1.03–1.50, p-trend = 0.007), well/very well done meat (HR = 1.32, 95% CI 1.10–1.58, p-trend = 0.005) and heme iron from red meat (Q4 vs. Q1: HR = 1.21, 95% CI 1.01–1.45, p-trend = 0.04). When stratified by sex, these associations remained significant in men but not women except for white meat intake in women (HR = 1.33, 95% CI 1.02–1.74, p-trend = 0.04). Additional studies should confirm our findings that consuming heme iron from red meat increases pancreatic cancer risk. Abstract What's new? How would you like your steak? A new study examines whether meat cooking techniques boost pancreatic cancer risk. The comprehensive investigation also looked at total meat intake; meat intake by type; as well as heme iron and meat mutagen intake, using data from a questionnaire given to more than 567,000 persons. Of these, 1417 developed pancreatic cancer. Analysis showed that higher risk of pancreatic cancer accompanies total meat consumption, high temperature cooking, grilled meat, well done meat, red meat, and heme iron intake. This is the largest study yet to evaluate the link between meat cooking methods, heme iron, and pancreatic cancer. Abbreviations AARP American Association of Retired Persons BaP benzo(a)pyrene CI confidence interval DiMeIQx 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline FFQ food frequency questionnaire HCAs heterocyclic amines HR hazard ratio MeIQx 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline NIH National Institutes of Health PAHs polycyclic aromatic hydrocarbons PhIP 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine RFQ risk factor questionnaire SAS Statistical Analysis Systems WCFR/AICR World Cancer Research Fund and American Institute for Cancer Research Pancreatic cancer is the fourth leading cause of cancer-related mortality in men and women of all ages in the United States.1 It is rapidly fatal and has a 5-year survival rate of <7%.1 Cigarette smoking, history of diabetes mellitus and overweight and obesity are consistent potentially modifiable risk factors for pancreatic cancer.2 The association between diet and pancreatic cancer risk is unclear mostly because of inconsistent study findings. In 2012, a panel from the World Cancer Research Fund and American Institute for Cancer Research (WCRF/AICR) concluded that the evidence that red and processed meats contribute to pancreatic cancer was suggestive. Dose-response meta-analyses of eight prospective cohort studies revealed that higher intake of red meat was positively, but insignificantly, associated with pancreatic cancer risk with some heterogeneity between studies.3 In addition, two meta-analyses on five case–control studies and 11 prospective cohort studies similarly found significant positive associations for red meat which were stronger in case–control studies and among men, respectively.4, 5 Processed meat was also positively associated with pancreatic cancer; however, the association was only significant in men.3, 5 Furthermore, these associations could potentially be explained by other factors in meat particularly confounded by compounds that are generated with meat cooking methods and doneness levels.6-10 Our prior examination of the NIH-AARP Diet and Health Study cohort revealed significant positive associations between total, red and high-temperature cooked meats and pancreatic cancer in men but not women.8 The study also noted that associated meat mutagens specifically overall mutagenic activity in men and the heterocyclic amine (HCA) 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (DiMeIQx) in both men and women may contribute to pancreatic carcinogenesis. Therefore, we examined the association between meat, meat cooking methods and compounds in meat in the large NIH-AARP Diet and Health Study cohort. In contrast to our previous examination of these hypotheses, this study has longer follow-up time (6 additional years) and more than three times as many pancreatic cancer cases (n > 1,400 cases) with meat cooking methods and related mutagens. The greater number of cases increases the power of our study and also enables us to look at interactions by other exposures. In addition to the meat mutagens, our study also examines heme iron intake as a risk factor for pancreatic cancer. Subjects and Methods Study population The NIH-AARP Diet and Health Study is a large prospective study of AARP members established in 1995–1996.8, 11 In total, 567,169 AARP members aged 50–71 years living in six US states (California, Florida, Louisiana, New Jersey, North Carolina and Pennsylvania) and in two metropolitan areas (Atlanta, GA and Detroit, MI) successfully completed and returned the self-administered questionnaires8, 11 that assessed demographic characteristics, dietary intake over the previous year and health-related factors.8 Six months after this baseline questionnaire was sent, participants who responded to the baseline questionnaire received a second risk factor questionnaire (RFQ) eliciting information on meat cooking methods.8, 11, 12 In total, 332,913 participants completed and returned the second questionnaire.8 Informed consent was obtained from all study participants and the study was approved by the National Cancer Institute Special Studies Institutional Review Board.8 For this study, we only included participants who completed the meat module portion of the second questionnaire. We excluded subjects who had questionnaires filled out by proxy respondents (n = 6,959), who had prevalent cancers as determined by the cancer registry data (n = 2,361) and whose energy consumption lay outside the normal sex-specific distribution for energy intake by two interquartile ranges above the 75th or below the 25th percentile on the logarithmic scale (n = 2,701; Ref. 8). We also excluded those with ≤0 years of follow-up (n = 38). Our final analytic cohort consisted of 322,846 individuals (187,265 men, 135,581 women; Ref. 8). Cohort follow-up and case ascertainment Pancreatic cancer cases were ascertained by linking cohort members to state cancer registries where the study participants reside, as well as Arizona, Texas and Nevada and to the US National Death Index from 1995 to 2006.8, 12, 13 Vital status of cohort participants was also ascertained by linkage to the Social Security Administration Death Master File. We included incident adenocarcinoma of the exocrine pancreas as the primary outcome in our analysis [International Classification of Disease for Oncology, Third Edition (code C250-C259); Refs. 8 and 12]. Our case definition excluded pancreatic endocrine tumors, sarcomas and lymphomas (histology types, 8150, 8151, 8153, 8155, 8240) as the etiologies may differ.8, 12 Dietary assessment and meat variable The baseline questionnaire gathered information on demographic characteristics, medical history and health-related behaviors and also contained a grid-based food frequency questionnaire (FFQ) that assessed the frequency and portion size of 124 food items consumed over the past year.8, 11 These line items were constructed using over 5,000 individual food codes found in the United States Department of Agriculture (USDA)'s Continuing Survey of Food Intake by Individuals database.14 Intake of total, red, white, processed and high-temperature cooked meats, heme iron as well as energy and other nutrients was determined from the baseline FFQ. While the specific items in each category have been described in detail elsewhere,8 a few categories have been altered. We included all forms of poultry products in the white meat category. Processed meat from red meat and poultry contributed to the processed meat category. The validity of both questionnaires was discussed in an earlier study.8 The RFQ included a meat module that elicited meat-cooking methods (barbequed/grilled, pan-fried, oven-broiled and microwaved) and doneness levels (rare/medium and well/very well done; Refs. 8, 15 and 16). Meat samples cooked by different methods and to varying levels of doneness were analyzed for levels of HCAs, polycyclic aromatic hydrocarbons (PAHs) and overall mutagenicity. Overall mutagenic activity was quantified using the Ames assay.8, 9, 15, 17 These meat mutagens were entered into the CHARRED database.8, 15, 18-21 The RFQ was linked to the CHARRED database to estimate the daily intake of meat mutagens including HCAs DiMeIQx, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) and benzo(a)pyrene (BaP), a marker for PAHs.8, 15, 18-21 Statistical analysis We used generalized linear models to calculate the means within each total meat intake quintile for the continuous variables and frequency proportions for the categorical variables in Table 1. Table 1. Selected baseline characteristics of NIH-AARP Diet and Health Study cohort participants by total meat intakea Characteristics Quintile of daily total meat intake (g/1,000 kcal) in men Men (n = 187,265) ≤43.6 >43.6 and ≤58.5 >58.5 and ≤72.9 >72.9 and ≤92.3 >92.3 Meat intake (g/1,000 kcal) Total meat 30.9 51.3 65.6 81.8 115.9 Red meat 17.0 28.8 36.8 45.1 59.0 White meat 13.9 22.5 28.8 36.7 56.9 Age (years) 63.2 63.0 62.7 62.2 61.6 Education, college graduate or postgraduate (%) 46.5 47.0 47.6 48.0 47.6 Race (%) African American 3.7 3.1 3.0 3.0 3.5 Non-Hispanic white 90.8 93.0 93.3 93.5 92.5 Smoking historya (%) Never 30.4 29.5 29.7 29.6 29.0 Former smoker 57.0 58.0 57.9 57.4 57.5 Current smoker 8.9 9.0 9.1 9.6 10.0 BMI (kg/mb) (%) 26.2 26.7 27.1 27.5 28.1 <25 39.5 33.4 30.2 27.0 22.12 25–<30 44.9 48.3 49.2 49.3 49.0 ≥30 13.8 16.7 19.2 22.3 27.4 Heavy physical activity, ≥5 times/week (%) 26.9 23.2 21.2 20.4 19.4 Self-reported diabetes (%) 6.6 7.8 9.2 11.2 14.2 Dietary intake Energy (kcal) 2,032 1,985 1,987 2,004 2,024 Total fat (g/1,000 kcal/day) 28.5 32.5 34.0 35.4 37.3 Saturated fat (g/1,000 kcal/day) 8.9 10.2 10.6 11.0 11.6 Heme iron (mg/1,000 kcal/day) 86.1 153.5 202.5 255.8 358.9 Advanced glycation end products 333.2 1274.6 1290.5 1444.8 1985.1 Alcohol intake (# drinks/day) 2.0 1.3 1.1 1.0 0.8 Alcohol intake ≥ 3 drinks/day (%) 15.9 12.3 10.8 9.2 6.8 Characteristics Quintile of daily total teat intake (g/1,000 kcal) in women Women (n = 135,581) ≤38.0 >38.0 and ≤53.4 >53.4 and ≤68.4 >68.4 and ≤88.3 >88.3 Meat intake (g/1,000 kcal) Total meat 25.7 45.9 60.7 77.5 113.7 Red meat 12.6 22.6 29.4 35.7 45.0 White meat 13.2 23.3 31.4 41.8 68.6 Age (years) 62.5 62.4 62.2 61.9 61.4 Education, college graduate or postgraduate (%) 33.2 30.6 31.1 32.1 31.8 Race (%) African American 2.7 2.2 2.0 2.0 2.2 Non-Hispanic white 91.5 94.1 94.3 94.6 94.0 Smoking historyb (%) Never 46.3 44.8 44.5 43.4 40.9 Former smoker 38.5 38.4 39.1 39.8 41.8 Current smoker or having quit <1 year ago 11.8 13.6 13.6 13.8 14.0 BMI (kg/mb) (%) 25.5 26.3 26.8 27.2 27.8 <25 53.8 46.8 43.8 40.3 36.4 25–<30 27.6 31.1 31.7 33.0 33.1 ≥30 15.2 19.3 21.9 24.3 27.8 Heavy physical activity, ≥5 times/week (%) 21.1 17.2 15.8 15.3 15.3 Self-reported diabetes (%) 4.8 5.7 6.5 7.8 10.0 Dietary intake Energy (kcal) 1,553 1,557 1,563 1,573 1,567 Total fat (g/1,000 kcal/day) 28.9 32.1 33.7 34.8 36.1 Saturated fat (g/1,000 kcal/day) 9.1 10.0 10.4 10.7 10.9 Heme iron (mg/1,000 kcal/day) 59.5 114.6 154.5 196.5 270.7 Advanced glycation end products 372.7 1364.4 1382.8 1684.4 1698.5 Alcohol intake (drinks/day) 0.5 0.5 0.4 0.4 0.4 Alcohol intake ≥3 drinks/day (%) 3.4 3.4 3.0 2.4 1.7 a Generalized linear models were used to estimate mean values for the continuous variables and frequencies for dichotomous proportions within each total meat intake quintile. b A total of 6,599 (3.52%) men and 4,214 (3.11%) women having missing smoking history data; 2,878 (1.54%) men and 3,762 (2.77%) women have missing BMI data. Person-years were determined by the date of receipt of the RFQ and date of pancreatic cancer diagnosis, death, emigration out of the registry area or December 2006, whichever occurred first.8, 13, 16 Cox proportional hazard models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs). Trends were estimated using the median value of each quantile. As the dietary variables were correlated with energy, we adjusted for energy using the density method. The distribution of the meat-related variables differed by sex (e.g., men consumed higher levels of red meat, women consumed higher levels of white meat); therefore, the quintiles for the meat-related variables were based on the distribution of each variable in the cohort by sex. For foods that were consumed by <20% of the cohort (pan-fried, oven-broiled and microwaved meats, DiMeIQx), sex-specific categories were created using zero intake as the referent and remaining subjects categorized as quartiles or tertiles. Subjects were merged in sex-specific quantile categories for sex-combined risk estimates. Relevant meat groups were controlled simultaneously in our models (red and white; processed and nonprocessed; high- and low-temperature cooked; barbequed/grilled, pan-fried, oven-broiled and microwaved; rare/medium and well/very well done). Our crude models included age at study entry, energy (kcal) and sex (in sex-combined models). Additional variables were included in our models if they were confounders that altered the risk estimate ≥10% or were putative risk factors for pancreatic cancer. Variables tested as potential confounders included smoking history, body mass index (BMI), self-reported diabetes, saturated fat intake and alcohol use. Smoking was a confounder in most of the models and saturated fat in our model for total meat. Our final model included age, smoking (never, quit ≥10 years ago, quit 5–9 years ago, quit 1–4 years ago, quit <1 year ago or current and smokes ≤20 cigarettes/day, quit <1 year ago or current and smokes >20 cigarettes/day and missing), BMI (kg/m2, <18.5, ≥18.5 and <25, ≥25 and <30, ≥30 and <35, ≥35 and missing), self-reported diabetes (yes, no) and energy-adjusted saturated fat (continuous) as well as sex (in sex-combined models). The meat mutagens, total heme iron, heme iron from red meat and advanced glycation end products were all correlated with each other as they are all present in meat and mutual adjustment of all attenuated associations; therefore, we show associations without mutual adjustment. We evaluated interactions by sex, diabetes, BMI and smoking status (never, former and current) on the association between meat exposure and pancreatic cancer by including cross-product terms for the meat trend variable with the respective effect modifier in the multivariable model. Significant interactions were further stratified. We additionally performed sensitivity analyses of the baseline FFQ meat variables (total, red, white and high-temperature cooked meats, and heme iron) on the full baseline cohort to assess internal consistency with associations observed for the RFQ cohort. Statistical Analysis Systems (SAS) software was used to perform all statistical analyses, and p values for all tests were two-tailed.8, 12 Results During follow-up up to 10.17 years (median 10.07 years, 2,974,128 person-years of observation), 1,417 individuals (895 men and 522 women) in the risk factor cohort developed exocrine pancreatic cancer.12 The characteristics of our study population according to total meat intake quintile are shown in Table 1. Men and women with higher total meat intake were more likely to be younger, current smokers, of higher BMI, less physically active and diabetic; they also consumed more red and white meat, total and saturated fat and heme iron but less alcohol.12 A greater proportion of men consumed more red than white meat while women consumed more white than red meat. In men and women combined, total, red and high-temperature cooked meat intake were significantly associated with pancreatic cancer with significant trends across quintiles (Table 2, Q5 vs. Q1: HR = 1.20, 95% CI 1.02–1.42, p-trend = 0.03; HR = 1.22, 95% CI 1.01–1.48, p-trend = 0.02; HR = 1.21, 95% CI 1.00–1.45, p-trend = 0.02, respectively). These associations tended to be stronger in men than in women; however, the interaction by sex was only significant for red meat (p-interaction = 0.03; Q5 vs. Q1 men HR = 1.36, 95% CI 1.07–1.73, p-trend = 0.004; women HR = 1.01, 95% CI 0.74–1.38, p-trend = 0.91). White meat intake showed a significant positive association in women (Q5 vs. Q1 HR = 1.33, 95% CI 1.02–1.74, p-trend = 0.04) but not in men (HR = 1.05, 95% CI 0.85–1.30, p-trend = 0.63, p-interaction = 0.26). Men and women consuming moderate amounts of processed meat (quintiles 2–4 compared to quintile 1) had significantly elevated risk for pancreatic cancer, although extreme intake was not associated. We also conducted substitution model analyses to examine the associations of subgroups of meat while holding total meat constant and observed associations similar to those obtained by the partition method. For example, holding total meat constant the association between red meat and pancreatic cancer still remained significant. Therefore, our associations were robust with both the partition and substitution methods. Table 2. Hazard ratios (HRs) and 95% confidence intervals (CIs) for baseline meat intake in the NIH-AARP Diet and Health Cohorta Quintile of daily meat intake (g/1,000 kcal) Variable 1 2 3 4 5 p trendb Total meat Men and women combinedc Cases/person-years 265/593,355 286/595,543 271/595,177 282/595,934 313/594,119 Age-adjusted HR (95% CI) 1.00 (reference) 1.09 (0.92–1.29) 1.06 (0.89–1.25) 1.14 (0.96–1.34) 1.32 (1.12–1.56) 0.0007 Multivariable HR (95% CI)d 1.00 (reference) 1.05 (0.89–1.24) 1.01 (0.85–1.19) 1.06 (0.89–1.25) 1.20 (1.02–1.42) 0.03 Men Cases/person-years 162/340,658 188/342,044 173/342,094 165/342,570 207/341,577 Age-adjusted HR (95% CI) 1.00 (reference) 1.18 (0.95–1.45) 1.11 (0.89–1.37) 1.09 (0.88–1.36) 1.44 (1.17–1.77) 0.002 Multivariable HR (95% CI)d 1.00 (reference) 1.12 (0.91–1.39) 1.04 (0.83–1.29) 1.00 (0.80–1.24) 1.27 (1.02–1.57) 0.08 Women Cases/person-years 103/252,697 98/253,499 98/253,082 117/253,369 106/252,543 Age-adjusted HR (95% CI) 1.00 (reference) 0.96 (0.73–1.26) 0.98 (0.74–1.29) 1.20 (0.92–1.56) 1.14 (0.87–1.49) 0.13 Multivariable HR (95% CI)d 1.00 (reference) 0.93 (0.71–1.23) 0.95 (0.72–1.26) 1.15 (0.88–1.51) 1.08 (0.82–1.42) 0.27 Red meat Men and women combinedc Cases/person-years 243/598,789 268/596,590 284/595,847 306/593,567 316/589,339 Age-adjusted HR (95% CI) 1.00 (reference) 1.11 (0.94–1.33) 1.21 (1.02–1.44) 1.35 (1.14–1.60) 1.47 (1.24–1.74) <0.0001 Multivariable HR (95% CI)d 1.00 (reference) 1.06 (0.88–1.26) 1.11 (0.92–1.32) 1.18 (0.99–1.42) 1.22 (1.01–1.48) 0.02 Men Cases/person-years 148/344,177 159/342,803 180/342,009 191/341,440 217/338,513 Age-adjusted HR (95% CI) 1.00 (reference) 1.09 (0.87–1.37) 1.27 (1.02–1.58) 1.38 (1.11–1.71) 1.66 (1.34–2.05) <0.0001 Multivariable HR (95% CI)d 1.00 (reference) 1.02 (0.81–1.28) 1.14 (0.91–1.44) 1.20 (0.95–1.51) 1.36 (1.07–1.73) 0.004 Women Cases/person-years 95/254,612 109/253,788 104/253,838 115/252,127 99/250,826 Age-adjusted HR (95% CI) 1.00 (reference) 1.16 (0.88–1.52) 1.12 (0.85–1.48) 1.29 (0.98–1.69) 1.16 (0.87–1.54) 0.3224 Multivariable HR (95% CI)d 1.00 (reference) 1.11 (0.84–1.47) 1.05 (0.78–1.40) 1.17 (0.87–1.56) 1.01 (0.74–1.38) 0.91 White meat Men and women combinedc Cases/person-years 284/587,575 294/593,947 272/595,811 272/597,199 295/599,601 Age-adjusted HR (95% CI) 1.00 (reference) 1.01 (0.86–1.19) 0.94 (0.79–1.11) 0.96 (0.81–1.13) 1.09 (0.93–1.29) 0.25 Multivariable HR (95% CI)d 1.00 (reference) 1.04 (0.88–1.23) 0.98 (0.83–1.16) 1.02 (0.86–1.20) 1.15 (0.98–1.36) 0.08 Men Cases/person-years 185/336,636 182/340,946 183/342,711 171/343,262 174/345,387 Age-adjusted HR (95% CI) 1.00 (reference) 0.96 (0.78–1.17) 0.96 (0.78–1.18) 0.92 (0.75–1.14) 0.99 (0.81–1.22) 0.96 Multivariable HR (95% CI)d 1.00 (reference) 0.99 (0.81–1.22) 1.01 (0.82–1.24) 0.99 (0.80–1.22) 1.05 (0.85–1.30) 0.63 Women Cases/person-years 99/250,939 112/253,001 89/253,100 101/253,936 121/254,214 Age-adjusted HR (95% CI) 1.00 (reference) 1.11 (0.85–1.46) 0.89 (0.67–1.19) 1.03 (0.78–1.36) 1.28 (0.98–1.67) 0.07 Multivariable HR (95% CI)d 1.00 (reference) 1.13 (0.86–1.49) 0.92 (0.69–1.23) 1.07 (0.81–1.42) 1.33 (1.02–1.74) 0.04 Processed meat Men and women combinedc Cases/person-years 228/597,449 308/597,036 301/595,084 313/593,677 267/590,886 Age-adjusted HR (95% CI) 1.00 (reference) 1.36 (1.14–1.62) 1.33 (1.12–1.59) 1.38 (1.15–1.64) 1.17 (0.98–1.40) 0.73 Multivariable HR (95% CI)d 1.00 (reference) 1.29 (1.08–1.53) 1.23 (1.02–1.46) 1.24 (1.04–1.49) 1.02 (0.85–1.24) 0.23 Men Cases/person-years 141/343,575 200/342,861 181/341,653 200/341,213 173/339,640 Age-adjusted HR (95% CI) 1.00 (reference) 1.43 (1.15–1.77) 1.30 (1.04–1.62) 1.42 (1.14–1.77) 1.23 (0.98–1.55) 0.49 Multivariable HR (95% CI)d 1.00 (reference) 1.33 (1.07–1.66) 1.17 (0.93–1.47) 1.26 (1.00–1.58) 1.05 (0.83–1.33) 0.44 Women Cases/person-years 87/253,874 108/254,176 120/253,431 113/252,464 94/251,246 Age-adjusted HR (95% CI) 1.00 (reference) 1.25 (0.94–1.66) 1.39 (1.05–1.84) 1.30 (0.98–1.74) 1.06 (0.79–1.43) 0.60 Multivariable HR (95% CI)d 1.00 (reference) 1.22 (0.91–1.62) 1.33 (1.00–1.76) 1.23 (0.91–1.65) 0.98 (0.72–1.33) 0.25 Meat cooked at high temperatures Men and women combinedc Cases/person-years 245/598,389 264/596,728 293/595,287 301/593,045 314/590,684 Age-adjusted HR (95% CI) 1.00 (reference) 1.08 (0.91–1.29) 1.23 (1.03–1.46) 1.29 (1.09–1.54) 1.43 (1.20–1.70) <0.0001 Multivariable HR (95% CI)d 1.00 (reference) 1.02 (0.86–1.22) 1.12 (0.94–1.34) 1.15 (0.96–1.37) 1.21 (1.00–1.46) 0.02 Men Cases/person-years 146/344,080 161/343,002 174/341,836 210/340,482 204/339,542 Age-adjusted HR (95% CI) 1.00 (reference) 1.12 (0.89–1.40) 1.24 (0.99–1.55) 1.54 (1.24–1.91) 1.58 (1.27–1.97) <0.0001 Multivariable HR (95% CI)d 1.00 (reference) 1.04 (0.83–1.31) 1.11 (0.88–1.40) 1.33 (1.06–1.68) 1.30 (1.03–1.66) 0.009 Women Cases/person-years 99/254,309 103/253,725 119/253,451 91/252,563 110/251,142 Age-adjusted HR (95% CI) 1.00 (reference) 1.04 (0.78–1.37) 1.20 (0.92–1.57) 0.94 (0.70–1.26) 1.20 (0.91–1.59) 0.30 Multivariable HR (95% CI)d 1.00 (reference) 0.99 (0.75–1.32) 1.13 (0.86–1.49) 0.87 (0.64–1.17) 1.07 (0.79–1.45) 0.85 a Cohort that successfully completed the meat module on the RFQ: n = 322,846 subjects (1,417 cases), 187,265 men (895 cases) and 135,581 women (522 cases). b p trend calculated using median values for each quintile. c Sex-combined models additionally adjusted for sex. d Cox proportional hazard models used to calculate hazard ratios with age as the time metric. All nutrients are adjusted for energy by the density method with energy also in the model. Models are additionally adjusted for smoking (never, quit ≥10 years, quit 5–9 years ago, quit 1–4 years ago, quit <1 year ago or current and smoked ≤20 cigarettes/day, quit <1 year ago or current and smoked >20 cigarettes/day and missing), BMI (kg/m2, <18.5, ≥18.5 and <25, ≥25 and <30, ≥30 and <35, ≥35 and missing), self-reported diabetes (yes, no) and energy-adjusted saturated fat (continuous). p Values for interaction by sex: total meat = 0.46, red meat = 0.03, white meat = 0.26, processed meat = 0.34 and high-temperature cooked meat = 0.12. All values in bold represent significant HRs or p-values. Certain meat preparation methods were associated with pancreatic cancer (Table 3). In sex-combined models, those consuming the most barbequed meat compared to the least had a significant 24% (95% CI 1.03–1.50) increased pancreatic cancer risk with significant trends across the quintiles (p-trend = 0.007). In sex-stratified analysis, men in the highest quintile of barbequed meat intake and quartile of broiled meat intake had a 33% (95% CI 1.05–1.68) and 34% (95% CI 1.09–1.64) greater risk, respectively, than those with the lowest intake. This risk increased across the quintiles and quartile (p-trends = 0.004 and 0.01), respectively. There was a borderline nonsignificant interaction for broiled meat by sex (p-interaction = 0.07). Consumption of pan-fried and microwaved meats was not associated with pancreatic cancer. With regard to doneness levels, well/very well done meat was associated with a significantly elevated risk for pancreatic cancer in sex-combined and stratified models (Q5 vs. Q1 men and women HR = 1.32, 95% CI 1.10–1.58, p-trend = 0.005; men HR = 1.33, 95% CI 1.05–1.67, p-tren