Date of Award


Document Type

Open Access Dissertation


Epidemiology and Biostatistics

First Advisor

Susan E. Steck


Introduction: The dietary inflammatory index (DII) assesses an individual’s overall diet quality with regards to its inflammatory potential on a continuum from maximally anti-inflammatory (lower or healthier DII scores) to maximally pro-inflammatory (higher or unhealthy DII scores). The DII measured at one point in time has been associated with cancer risk in previous studies; however, data are lacking regarding the change in DII over time and how these changes impact cancer risk. We assessed changes in the DII, and evaluated associations between cumulative history, and changes over time in dietary inflammatory potential, and risk of colorectal cancer (CRC) and breast cancer (BRCA). Methods: Study participants were women aged 50-79 years recruited from 1993-1998 into the Women’s Health Initiative (WHI) Observational Study (OS) and Dietary Modification Trial (DMT), and followed through September 30, 2010. The DII was calculated from repeated food frequency questionnaires (FFQ) data in the OS (n=76,671) at baseline and Year 3, and in the DMT (n=48,482) at up to 11 time points. Univariate generalized estimating equations were used to compare mean DII over time, adjusting for multiple comparisons. We calculated ten cumulative averages of DII, incrementally from baseline to Year 10, categorized each average into quintiles, and estimated hazard ratios (HR) and 95% confidence intervals (95%CI) for CRC, colon, rectal cancer and invasive BRCA incidence by DII quintiles in multivariable-adjusted Cox regression models. We also derived patterns of changes in DII between baseline and Year 3; and calculated HR for CRC, colon, rectal, and breast cancer incidence including molecular and histologic BRCA subtypes, using multivariable-adjusted Cox regression models. Results: In the OS, mean DII decreased from -1.14 (±2.58) at baseline to -1.50 (±2.60) at Year 3. In the DMT, DII decreased from -0.40 (±2.54) to its lowest point of -1.70 (±2.63) at Year 3 in the intervention arm and from -0.38 (±2.55) to its lowest point of -1.04 (±2.60) at Year 3 in the control arm. These changes were influenced by BMI, education, and race/ethnicity. During an average 11.7 years, 1,240 cases of CRC and 4,242 cases of BRCA were identified. HR for the association between high DII scores and CRC were consistently significantly elevated in the first seven years of follow up, for colon cancer with multivariable-adjusted HR ranging from 1.30 to 1.58 in quintile 3 vs. 1, while no significant associations were observed for rectal cancer. Compared to participants in the anti-inflammatory stable category, risk was increased in participants with a pro-inflammatory stable diet, for CRC (HR, 1.18; 95%CI, 0.99, 1.41), and for rectal cancer (HR, 1.53; 95%CI, 1.01, 2.32). HR revealed no significant association between changes in DII and risk of invasive BRCA or its subtypes. Conclusion: In this large prospective study of postmenopausal women, dietary inflammatory potential was relatively stable in OS participants, but decreased significantly over time in women enrolled in the DMT. DII changes were modified by BMI, education, and race/ethnicity. Long-term pro-inflammatory diets increased the risk of colon cancer, while shorter-term stable pro-inflammatory diets increased the risk of rectal cancer but not breast cancer or its subtypes. Lowering the inflammatory potential of diet could be a means for colon cancer, and potentially rectal cancer prevention in postmenopausal women.

Included in

Public Health Commons