Abstract
Chronic inflammation is implicated in pancreatic cancer, and can be modulated by diet and other lifestyle factors. We examined the association between Dietary Inflammatory Index (DII) scores and pancreatic cancer risk in the NIH-AARP Diet and Health Study, and examined effect modification by inflammation-related lifestyle factors, including body mass index, cigarette smoking, diabetes, alcohol drinking, and use of non-steroidal anti-inflammatory drugs.
Energy-adjusted DII scores (E-DII) were computed on the basis of food frequency questionnaire responses for foods and dietary supplements. Cox proportional hazards models were fitted and effect modification was examined by adding a cross-product of each effect modifier with E-DII quintile in the multivariable-adjusted model.
There were 2,824 primary incident pancreatic cancers diagnosed during a median of 13.4 years follow-up, and there was no association between E-DII scores and pancreatic cancer risk among either men [HRQ5vsQ1, 1.00; 95% confidence interval (CI), 0.86–1.16] or women (HRQ5vsQ1, 1.00; 95% CI, 0.82–1.21) in the multivariable-adjusted model, and no association was detected by any cancer stage. The E-DII and pancreatic cancer association was not modified by any of the inflammation-related lifestyle factors examined.
Results from this large prospective study did not support an association between inflammatory potential of diet and pancreatic cancer, or effect modification by other inflammation-related lifestyle factors.
Inflammatory potential of diet may not be related to pancreatic cancer risk. Future cohort studies with more frequent dietary measures could be useful in determining the appropriate timing of dietary intake in relation to pancreatic cancer etiology.
Introduction
Chronic inflammation is implicated in pancreatic carcinogenesis and can be modulated by diet and other lifestyle factors (1, 2). The Dietary Inflammatory Index (DII) assesses dietary inflammatory potential (3) and has been examined in relation to pancreatic cancer in both case–control and prospective studies (4–7), with inconsistent results. In the NIH—American Association of Retired Persons (NIH–AARP) Diet and Health Study, the largest diet and health prospective cohort study in the United States, we aimed to examine and verify the findings of the relationship between the E-DII and pancreatic cancer risk and effect modifications by important inflammation-related lifestyle factors including body mass index (BMI), cigarette smoking, alcohol consumption, non-steroidal anti-inflammatory drugs use, and diabetes history. We also aimed to explore the E-DII and pancreatic cancer association by cancer stage.
Materials and Methods
The NIH-AARP study (8) was a prospective cohort of 617,119 American adults ages 50 to 71 years old who responded to the baseline questionnaire (BQ) between 1995 and 1996. In the current analysis, participants were excluded if they had unreliable responses on the BQ (n = 50,721), had BQs filled out by proxy (n = 15,760), had a personal history of cancer (n = 8,828), reported an extreme energy intake or BMI (2 interquartile ranges below the sex-specific 25th percentile or above the 75th percentile of log-transformed energy intake or BMI; n = 4,261 and 4,214, respectively), or moved out of the study area or died before baseline (n = 79), bringing the analytic sample to 533,256.
The E-DII score (3) was calculated by summing across the product of energy-adjusted and global dietary database-standardized intake of each of the 33 available DII components, derived from self-reports of previous-year intake using the 124-item NCI-developed food frequency questionnaire (FFQ), with the DII component-specific inflammatory effect scores (3, 9). Incident pancreatic cancer cases were mainly identified through linkage with state cancer registries (9). The HRs (95% confidence intervals, 95% CI) of cancer risk associated with each one-unit increase of a standard deviation of the continuous E-DII score or with the E-DII quintiles (the first E-DII quintile as the referent group representing the most anti-inflammatory diets) were estimated using Cox proportional hazards models. We evaluated associations for men, women, and both combined. Effect modification was assessed by adding a cross-product of each categorical effect modifier with E-DII quintile in the multivariable-adjusted model. In addition, each pancreatic cancer stage was analyzed as a separate outcome. Multivariable models controlled for potential confounders listed in Table 1 and included as a footnote of Table 2. Sensitivity analyses included: (i) excluding subjects with follow-up <5 years; (ii) removing subjects who self-reported diabetes history at baseline; (iii) imputing cases' diagnosis dates by subtracting survival time (estimated on the basis of SEER data) from the death date if diagnosis date was same as death date.
Results
A total of 2,824 primary pancreatic cancers occurred during a median follow-up of 13.4 years. Population characteristics across E-DII quintiles are shown in Table 1. The E-DII was not associated with any pancreatic cancer outcome among both men and women (Table 2). The E-DII and pancreatic cancer association was not modified by any of the inflammation-related lifestyle factors examined (Table 2). All sensitivity analyses produced null results.
Discussion
The null association between the E-DII and pancreatic cancer risk found in this large prospective cohort study did not support previous case–control studies, which observed an approximate 2.5-fold increased risk among subjects consuming the most pro-inflammatory diets (4–6). Likewise, in contrast to our null findings, two case–control studies found evidence of effect modification by smoking status, BMI, or diabetes (4, 5). One study identified significant positive associations of DII with pancreatic cancer in all cancer stages (4). A pooled analyses of five nested case–control studies within the Pancreatic Cancer Cohort Consortium (PanScan) reported a modest positive association with E-DII (ORQ5 vs. Q1, 1.23; 95% CI, 0.92–1.66; Ptrend = 0.008; ref. 6). Our findings agreed with previously conducted analyses using the prospective PLCO study, which reported an overall null association among 101,449 participants with baseline demographics similar to this study (7). Possible selection bias, frequent use of proxy responses, and other sources of information bias in pancreatic cancer case–control studies that may contribute to differential misclassification could help to explain differences in results from cohort studies.
The PLCO cohort suggested reverse causality for E-DII and pancreatic cancer association, where an inverse association between pro-inflammatory diets and pancreatic cancer risk was observed in follow up of <4 years (7). The current study did not support this phenomenon, even in the lag time analysis. The differences in study follow-up time (PLCO median follow-up is 8.5 years) and use of a FFQ with different components for computing the DII score may partially explain differing results. The effect of preclinical disease on dietary intake should be further explored in future studies. In addition to the possible true null association, it is possible that a one-time dietary assessment many years prior to diagnosis may not be an etiologically relevant timeframe for assessing exposure. Therefore, a large prospective study with more frequent dietary measures could be useful in determining the appropriate timing of dietary intake in relation to pancreatic cancer etiology.
Disclosure of Potential Conflicts of Interest
M.D. Wirth is a senior research scientist at Connecting Health Innovations, LLC. J.R. Hebert is a president and has ownership interest (including stock, patents, etc.) in Connecting Health Innovations LLC. No potential conflicts of interest were disclosed by the other authors.
Disclaimer
The views expressed herein are solely those of the authors and do not necessarily reflect those of the FCDC or FDOH.
Authors' Contributions
Conception and design: J. Zheng, J.R. Hebert, S.E. Steck
Development of methodology: J. Zheng, M.D. Wirth, J. Zhang, J.R. Hebert, S.E. Steck
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.Z. Stolzenberg-Solomon, S.E. Steck
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Zheng, M.D. Wirth, A.T. Merchant, J. Zhang, R.Z. Stolzenberg-Solomon, S.E. Steck
Writing, review, and/or revision of the manuscript: J. Zheng, M.D. Wirth, A.T. Merchant, J. Zhang, N. Shivappa, R.Z. Stolzenberg-Solomon, J.R. Hebert, S.E. Steck
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Zheng, R.Z. Stolzenberg-Solomon, S.E. Steck
Study supervision: S.E. Steck
Acknowledgments
This research was supported in part by the Intramural Research Program of the NIH, NCI. Cancer incidence data from the Atlanta metropolitan area were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia. Cancer incidence data from California were collected by the California Cancer Registry, California Department of Public Health's Cancer Surveillance and Research Branch, Sacramento, California. Cancer incidence data from the Detroit metropolitan area were collected by the Michigan Cancer Surveillance Program, Community Health Administration, Lansing, Michigan. The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System (Miami, Florida) under contract with the Florida Department of Health, Tallahassee, Florida. Cancer incidence data from Louisiana were collected by the Louisiana Tumor Registry, Louisiana State University Health Sciences Center School of Public Health, New Orleans, Louisiana. Cancer incidence data from New Jersey were collected by the New Jersey State Cancer Registry, The Rutgers Cancer Institute of New Jersey, New Brunswick, New Jersey. Cancer incidence data from North Carolina were collected by the North Carolina Central Cancer Registry, Raleigh, North Carolina. Cancer incidence data from Pennsylvania were supplied by the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, Pennsylvania. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations or conclusions. Cancer incidence data from Arizona were collected by the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services, Phoenix, Arizona. Cancer incidence data from Texas were collected by the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services, Austin, Texas. Cancer incidence data from Nevada were collected by the Nevada Central Cancer Registry, Division of Public and Behavioral Health, State of Nevada Department of Health and Human Services, Carson City, Nevada. We are indebted to the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation. We also thank Sigurd Hermansen and Kerry Grace Morrissey from Westat for study outcomes ascertainment and management and Leslie Carroll at Information Management Services for data support and analysis. J. Zheng was supported by University of South Carolina Support to Promote Advancement of Research and Creativity (SPARC) graduate research grant (PI: J. Zheng) and Cancer Prevention and Research Institute of Texas grant PR 170259 (PI: S. Chang). N. Shivappa, M.D. Wirth, and J.R. Hébert were supported by grant number R44 DK103377 from the National Institute of Diabetes and Digestive and Kidney Diseases (PI: N. Shivappa).
RSS Feeds