ABSTRACT
Background
Chronic inflammation is associated with ovarian carcinogenesis; yet, the impact of inflammatory-related exposures on outcomes has been understudied.
Objective
Given the poor survival of women diagnosed with ovarian cancer, especially African-Americans, we examined whether diet-associated inflammation, a modifiable source of chronic systemic inflammation measured by the dietary inflammatory index (DII), was associated with all-cause mortality among African-American women with ovarian carcinoma.
Methods
Data were available from 490 ovarian carcinoma patients enrolled in a population-based case-control study of African-American women with ovarian cancer, the African-American Cancer Epidemiology Study. Energy-adjusted DII (E-DII) scores were calculated based on prediagnostic dietary intake of foods alone or foods and supplements, which was self-reported using the 2005 Block Food Frequency Questionnaire. Cox proportional hazards regression was used to estimate risk of mortality overall and for the most common histotype, high-grade serous carcinoma. Additionally, we assessed interaction by age at diagnosis and smoking status.
Results
Women included in this study had a median age of 57 y, and the majority of women were obese (58%), had late-stage disease (Stage III or IV, 66%), and had high-grade serous carcinoma (64%). Greater E-DII scores including supplements (indicating greater inflammatory potential) were associated with an increased risk of mortality among women with high-grade serous carcinoma (HR1-unit change: 1.08; 95% CI: 1.01, 1.17). Similar associations were observed for the E-DII excluding supplements, although not statistically significant (HR1-unit change: 1.07; 95% CI: 0.97, 1.17). There was an interaction by smoking status, where the positive association with mortality was present only among ever smokers (HRQuartile 4/Quartile 1: 2.36; 95% CI: 1.21, 4.60) but not among never smokers.
Conclusions
Greater inflammatory potential of prediagnostic diet may adversely impact prognosis among African-American women with high-grade serous carcinoma, and specifically among ever smokers.
Keywords: ovarian cancer, diet, inflammation, dietary inflammatory index, African-American women, gynecologic malignancies, race/ethnicity, cancer epidemiology
Introduction
Ovarian cancer is the fifth deadliest malignancy among women in the US, with only 47% of patients surviving 5 y after diagnosis (1). Markedly different survival patterns for women with this disease exist by race/ethnicity, where African-American women with ovarian cancer have a worse 5-y relative survival in comparison to white women, 37% compared with 47%, respectively (1). Moreover, since the 1970s, white women with ovarian carcinoma have shown improved survival, yet African-American women have experienced little to no improvement in outcomes during this time period (2). The causes for these poor survival patterns by race/ethnicity are not well understood due to the paucity of studies with well-annotated epidemiologic and clinical data on African-American women with ovarian cancer.
Chronic inflammation is implicated in ovarian carcinogenesis (3–5), and studies suggest a link between inflammation and ovarian cancer outcomes (6–8). Dietary intake is a modifiable source of chronic inflammation (9) that may influence cancer outcomes. The literature-derived dietary inflammatory index (DII) was developed (10, 11) to measure the inflammatory potential of one's diet and has been validated with circulating inflammatory markers (12–14). Four case-control studies investigating the DII in ovarian cancer noted a positive association with risk for a more proinflammatory diet prior to diagnosis overall (15–17), or specifically, among postmenopausal women (18). By contrast, Tabung et al. (19) used an empirical-based index of inflammatory dietary potential and found no association with ovarian cancer risk among the Nurses’ Health Study cohort. The only evidence as to whether dietary inflammatory potential is associated with outcomes in ovarian cancer patients is from an Australian case-control study, where no association was observed with overall or ovarian cancer-specific survival (17). To date, no study has investigated the influence of dietary inflammatory potential on mortality among African-American ovarian cancer patients in the US. Therefore, we sought to evaluate whether the prediagnostic DII score is associated with risk of all-cause mortality among African-American women with ovarian carcinoma. As ovarian carcinoma is a heterogeneous disease, we also evaluated this association restricted to the most common histotype, high-grade serous carcinoma.
Methods
Study population
We used data from a multi-center, population-based case-control study of African-American women with ovarian carcinoma, the African-American Cancer Epidemiology Study (AACES), which has been described in detail elsewhere (20). Briefly, cases were identified by a rapid case ascertainment approach through cancer registries, hospitals, and gynecologic oncology departments. The eligibility criteria for cases included a primary diagnosis of ovarian carcinoma during December 2010 to December 2015, aged 20–79 y at diagnosis, a resident of 1 of 11 geographic locations (Alabama, Georgia, Illinois, Louisiana, metropolitan Detroit, Michigan, New Jersey, North Carolina, Ohio, South Carolina, Tennessee, Texas), and the ability to speak English. Controls were identified using random digit dialing and were frequency matched to cases by 5-y age categories and geographic location. All AACES participants completed a telephone survey at baseline to collect information on demographics and well-established and suspected ovarian cancer risk factors (e.g. reproductive history, exogenous hormone use). A shortened version of this survey was offered to participants who would have otherwise refused. A centralized pathologic review was conducted for the majority of participants to confirm diagnosis and characterize histotype. The study protocol was approved by the Institutional Review Board at each participating site. A total of 593 eligible ovarian cancer cases were enrolled in the AACES.
Dietary Inflammatory Index
AACES participants completed a mailed version of the Block 2005 FFQ, which queried their usual dietary intake of 110 foods and beverages in the year before diagnosis. NutritionQuest, formerly known as Block Dietary Data Systems, derived individual nutrient and total energy intake from the FFQ data. Using these data, the energy-adjusted DII score (E-DII; i.e. the DII per 1,000 calories of food consumed) was calculated using the methods previously described by Peres et al. (15) for this study population. Two E-DII scores were examined, one for nutrients from food sources alone and one for nutrients from food sources plus dietary supplements, noted as the E-DII excluding supplements and the E-DII including supplements, respectively. Higher DII scores indicate a more proinflammatory diet. Data were available for 27 of the 45 possible food parameters used to calculate the DII (carbohydrates; protein; fat; alcohol; fiber; cholesterol; saturated, monounsaturated, and polyunsaturated fatty acids; ω-3 and ω-6 polyunsaturated fatty acids; trans-fat; niacin; thiamin; riboflavin; vitamins A, B-6, B-12, C, D, E; iron; magnesium; zinc; selenium; folic acid; β-carotene; and isoflavones), and data on 19 dietary supplements were available (vitamins A, B-6, B-12, C, D, E; iron; calcium; zinc; β-carotene; folic acid; copper; selenium; thiamin; riboflavin; magnesium; niacin; and ω-3 and ω-6 fatty acids).
Analyses conducted to examine the relation between E-DII scores and risk of mortality focus on the 490 cases who completed the FFQ, representing 83% of the 593 AACES cases. E-DII scores were analyzed as both a continuous variable and categorized by quartiles. For the E-DII including supplements, the range of values included in each quartile are as follows: Quartile 1 (−5.22, −3.50), Quartile 2 (−3.49, −2.23), Quartile 3 (−2.22, 0.02), and Quartile 4 (0.03, 3.11); and for the E-DII excluding supplements: Quartile 1 (−4.15, −2.16), Quartile 2 (−2.15, −0.48), Quartile 3 (−0.47, 1.12), and Quartile 4 (1.13, 3.11).
Follow-up and outcome ascertainment
Vital status was ascertained annually through a systematic collection of follow-up data from cancer registries and annual contact with the patients, and supplemented by data from LexisNexis Accurint, which offers access to current public records. Follow-up time was calculated as the number of days from the date of diagnosis to the date of death or the date of last contact. Although we can be fairly certain that patient deaths were accurately determined in this study population, the date of last contact for survivors may not be as current, and accuracy may vary by study site. Due to this potential violation of independent censoring, we estimated the date of last contact as the date of the latest vital status update for each site (≤9 May, 2018). Among the 490 ovarian carcinoma cases, there were 223 deaths (46%) and the median time of follow-up was 3.5 y.
Statistical analysis
First, we used chi-square and t-tests to examine whether the distribution of the E-DII scores (both including and excluding supplements) differed by participant characteristics, where appropriate. Unadjusted Kaplan–Meier survival curves were estimated by quartiles of the E-DII (both including and excluding supplements), and log-rank tests were used to compare survival patterns by E-DII quartiles. Cox proportional hazards regression models were used to estimate HRs and 95% CIs for the association between the E-DII scores (both including and excluding supplements) and risk of overall mortality. Additionally, the median value of each quartile was modeled as a continuous variable to test for linear trends. Due to >10% missing data for important covariates of interest (e.g. stage at diagnosis, household income), we used multiple imputation methods as described in White and Royston (21) to impute missing data. The imputation was repeated 10 times before pooling to generate a single set of estimates. In addition to the analyses using multiple imputation, we also provide results restricted to cases with complete data for all covariates for comparison. Two models were evaluated with different adjustment sets. In Model 1, we adjusted for study design variables, age at diagnosis (y) and study site (Alabama, Georgia and Tennessee combined due to sample size and geographic similarities, Illinois and metropolitan Detroit, Michigan combined due to sample size and geographic similarities, Louisiana, New Jersey, North Carolina, Ohio, South Carolina, Texas), and the most important prognostic factor for ovarian carcinoma, stage (Stage I–II, Stage III–IV). For Model 2, we additionally adjusted for a priori confounders: smoking status (current, former, and never smokers; current and former smokers were combined for modeling due to small numbers of current smokers), BMI in the year prior to diagnosis (BMI; <25 kg/m2 or normal weight, 25–29.9 or overweight, and ≥30 or obese), comorbidities as defined by the modified Charlson Comorbidity Index (0, 1, ≥2 comorbid conditions), and household income (≤$25,000, >$25,000). For the models using the E-DII excluding supplements, we also adjusted for supplement use in the year prior to diagnosis (yes, no). These analyses were repeated after restricting to cases with the most common histotype, high-grade serous ovarian carcinoma. Due to our finding of a stronger association between the DII and ovarian cancer risk among women aged ≥60 y in Peres et al. (15), interaction by age was evaluated using Wald tests based on joint tests of the interaction parameters combined by multiple imputation (e.g. age times E-DII scores). The same strategy was also used to examine whether there was interaction by smoking status. The proportional hazard assumption was tested by examining interaction terms for survival time and each covariate (e.g. survival time in years multiplied by stage) individually and collectively. We did not observe any violation of the proportional hazard assumption in this study. SAS, Version 9.4 (SAS Institute, Inc.) was used to complete all analyses except for the graphics, which were produced in R, Version 3.5.0.
Sensitivity analysis including clinical variables
The medical records from the diagnostic and treatment facilities of each patient were requested and abstracted for data on surgery, chemotherapy (adjuvant and neoadjuvant), residual disease after primary debulking surgery, and CA-125 at initial presentation and after adjuvant chemotherapy. Medical record abstractions were completed for 444 of the 490 women included in this study; however, the data were incomplete (e.g. missing data ranged from 4% to 23% for key clinical variables). As these clinical variables are important prognostic factors for ovarian cancer, we completed a sensitivity analysis, repeating the primary analyses adjusting for the clinical variables that were associated with outcome in this population, neoadjuvant chemotherapy (yes, no) and debulking status (optimal, suboptimal). Debulking status was defined as optimal for women who had a residual tumor diameter of <1 cm after debulking surgery. If data were not available on debulking status, we utilized CA-125 after adjuvant chemotherapy (when available) as a proxy and designated women with a CA-125 ≥35 (abnormal CA-125) as suboptimally debulked (22–24).
Results
The distribution of AACES patient characteristics by the E-DII quartiles is provided in Table 1. For the E-DII including supplements, women who were younger at diagnosis, had a higher total energy intake, currently smoked, and had a lower household income were more likely to have more proinflammatory E-DII scores, P < 0.05 (Table 1). The distribution of stage at diagnosis, BMI, comorbidities, and histotype did not differ across quartiles of E-DII scores including supplements. For the E-DII excluding supplements, we observed similar patterns as shown for the E-DII including supplements although not as pronounced; women who used supplements within the year prior to diagnosis were more likely to have more anti-inflammatory E-DII scores.
TABLE 1.
Distribution of African-American Cancer Epidemiology Study (AACES) patient characteristics by the E-DII including and excluding supplement intake1, n = 490 cases
| E-DII including supplement intake | E-DII excluding supplement intake | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q1 | Q2 | Q3 | Q4 | |||
| n = 123 | n = 122 | n = 122 | n = 123 | P | n = 122 | n = 123 | n = 122 | n = 123 | P | |
| Age, y | 59.3 ± 9.5 | 57.8 ± 10.8 | 59.5 ± 10.1 | 55.6 ± 11.1 | 0.01 | 59.0 ± 9.8 | 59.1 ± 10.5 | 58.3 ± 10.8 | 55.8 ± 10.6 | 0.06 |
| Total energy intake, kcal/d | 1,304 ± 617 | 1,747 ± 1,021 | 1,933 ± 1911 | 2,140 ± 1,539 | <0.0001 | 1,358 ± 693 | 1,893 ± 1,920 | 1,729 ± 990 | 2,139 ± 1,527 | 0.0001 |
| Stage | ||||||||||
| I and II | 39 (34) | 43 (37) | 39 (35) | 33 (28) | 0.56 | 42 (37) | 36 (32) | 40 (34) | 36 (31) | 0.71 |
| III and IV | 75 (66) | 73 (63) | 74 (65) | 83 (72) | 70 (63) | 76 (68) | 77 (66) | 82 (69) | ||
| Unknown | 9 | 6 | 9 | 7 | 10 | 11 | 5 | 5 | ||
| BMI (kg/m2)2 | ||||||||||
| <25 | 20 (16) | 23 (19) | 15 (12) | 16 (13) | 0.05 | 21 (17) | 17 (14) | 18 (15) | 18 (15) | 0.74 |
| 25–29.9 | 41 (34) | 24 (20) | 39 (32) | 25 (20) | 34 (28) | 38 (31) | 29 (24) | 28 (23) | ||
| ≥30 | 61 (50) | 75 (62) | 68 (56) | 81 (66) | 66 (55) | 68 (55) | 74 (61) | 77 (63) | ||
| Missing | 1 | 0 | 0 | 1 | 1 | 0 | 1 | 0 | ||
| Smoking status | ||||||||||
| Never smoker | 74 (60) | 72 (59) | 62 (51) | 65 (53) | 0.006 | 71 (58) | 69 (56) | 67 (55) | 66 (54) | 0.32 |
| Former smoker | 42 (34) | 42 (34) | 43 (35) | 33 (27) | 42 (34) | 44 (35) | 39 (32) | 36 (29) | ||
| Current smoker | 7 (6) | 8 (7) | 17 (14) | 25 (20) | 9 (7) | 11 (9) | 16 (13) | 21 (17) | ||
| Charlson Comorbidity Index | ||||||||||
| 0 conditions | 52 (42) | 41 (34) | 45 (37) | 40 (32) | 0.37 | 49 (40) | 42 (34) | 46 (38) | 41 (33) | 0.13 |
| 1 condition | 20 (16) | 31 (25) | 32 (26) | 34 (28) | 19 (16) | 36 (29) | 25 (20) | 37 (30) | ||
| ≥2 conditions | 51 (42) | 50 (41) | 45 (37) | 49 (40) | 54 (44) | 45 (37) | 51 (42) | 45 (37) | ||
| Income | ||||||||||
| <$25,000 | 32 (28) | 44 (39) | 58 (49) | 68 (59) | <0.0001 | 41 (36) | 47 (40) | 55 (47) | 59 (52) | 0.07 |
| ≥$25,000 | 83 (72) | 70 (61) | 60 (51) | 48 (41) | 74 (64) | 70 (60) | 62 (53) | 55 (48) | ||
| Missing | 8 | 8 | 4 | 7 | 7 | 6 | 5 | 9 | ||
| Supplement intake3 | ||||||||||
| No | 13 (11) | 38 (31) | 43 (35) | 48 (39) | <0.0001 | |||||
| Yes | 109 (89) | 85 (69) | 79 (65) | 75 (61) | ||||||
| Histology | ||||||||||
| High-grade serous | 78 (64) | 79 (65) | 80 (66) | 76 (62) | 0.98 | 74 (61) | 81 (66) | 75 (61) | 83 (68) | 0.83 |
| Low-grade serous | 4 (3) | 4 (3) | 2 (2) | 6 (5) | 4 (3) | 3 (3) | 2 (2) | 7 (6) | ||
| Endom-etrioid | 13 (11) | 10 (8) | 9 (7) | 13 (11) | 15 (12) | 8 (7) | 13 (11) | 9 (7) | ||
| Clear cell | 5 (4) | 6 (5) | 6 (5) | 3 (2) | 6 (5) | 5 (4) | 4 (3) | 5 (4) | ||
| Mucinous | 5 (4) | 7 (6) | 5 (4) | 9 (7) | 5 (4) | 6 (5) | 8 (7) | 7 (6) | ||
| Other epithelial | 17 (14) | 15 (12) | 20 (16) | 15 (10) | 17 (14) | 19 (16) | 20 (16) | 11 (9) | ||
| Missing | 1 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | ||
Values are n (%) or mean ± SD. E-DII, energy-adjusted dietary inflammatory index.
1 y prior to diagnosis date.
In the year prior to diagnosis.
Figure 1 provides the Kaplan–Meier survival curves by quartiles of the E-DII, for both the E-DII excluding (Panel A) and including supplements (Panel B). No statistically significant differences in survival were observed by quartiles of the DII excluding supplements. For the E-DII including supplements, a worse survival was observed for women who had more proinflammatory E-DII scores compared with more anti-inflammatory E-DII scores (P = 0.03).
FIGURE 1.
Unadjusted Kaplan–Meier survival curves for ovarian carcinoma by quartiles of the E-DII excluding supplements (Panel A) and the E-DII including supplements (Panel B). Quartile 1 is the most anti-inflammatory and Quartile 4 is the most proinflammatory. E-DII, energy-adjusted dietary inflammatory index.
Table 2 provides the adjusted HRs and 95% CIs for the association between the E-DII and risk of mortality using multiple imputation. For the E-DII including supplements, a more proinflammatory diet prior to diagnosis was associated with an increased risk of mortality in the minimally adjusted model (Model 1). After adjustment for additional epidemiologic characteristics in Model 2, the association was attenuated, where for every one-unit increase in the E-DII, a 6% increased risk of mortality (HR: 1.06; 95% CI: 0.99, 1.13) was observed. We did not observe an association between the E-DII excluding supplements and risk of mortality, yet the trend was similar to the findings for the E-DII including supplements. The complete case analysis approach (n = 434) resulted in fairly similar, yet less precise, findings (Supplemental Table 1).
TABLE 2.
Adjusted HRs and 95% CIs for all-cause deaths in women with ovarian carcinoma and in women with high-grade serous ovarian carcinoma by E-DII1
| All ovarian carcinoma (n = 490) | High-grade serous ovarian carcinoma (n = 313) | |||||
|---|---|---|---|---|---|---|
| Model 12 | Model 23 | Model 12 | Model 23 | |||
| E-DII | Cases (deaths) | HR (95% CI) | HR (95% CI) | Cases (deaths) | HR (95% CI) | HR (95% CI) |
| Including supplements | ||||||
| Quartile 1 (−5.22, −3.50) | 123 (49) | 1.00 (Referent) | 1.00 (Referent) | 78 (32) | 1.00 (Referent) | 1.00 (Referent) |
| Quartile 2 (−3.49, −2.23) | 122 (48) | 1.06 (0.71, 1.58) | 0.98 (0.65, 1.47) | 79 (36) | 1.17 (0.72, 1.90) | 1.08 (0.66, 1.76) |
| Quartile 3 (−2.22, 0.02) | 122 (62) | 1.46 (1.00, 2.13) | 1.34 (0.90, 1.97) | 80 (46) | 1.79 (1.13, 2.84) | 1.62 (1.01, 2.61) |
| Quartile 4 (0.03, 3.11) | 123 (64) | 1.59 (1.09, 2.32) | 1.35 (0.90, 2.03) | 76 (47) | 1.87 (1.19, 2.95) | 1.58 (0.98, 2.56) |
| P-trend | 0.005 | 0.06 | 0.002 | 0.03 | ||
| Per 1 unit of E-DII | 490 (223) | 1.08 (1.02, 1.15) | 1.06 (0.99, 1.13) | 313 (161) | 1.11 (1.04, 1.19) | 1.08 (1.01, 1.17) |
| Excluding supplements4 | ||||||
| Quartile 1 (−4.15, −2.16) | 122 (46) | 1.00 (Referent) | 1.00 (Referent) | 74 (31) | 1.00 (Referent) | 1.00 (Referent) |
| Quartile 2 (−2.15, −0.48) | 123 (58) | 1.26 (0.84, 1.88) | 1.26 (0.84, 1.89) | 81 (38) | 1.22 (0.75, 1.99) | 1.23 (0.75, 2.01) |
| Quartile 3 (−0.47, 1.12) | 122 (59) | 1.24 (0.83, 1.85) | 1.16 (0.78, 1.74) | 75 (43) | 1.34 (0.83, 2.16) | 1.24 (0.76, 2.02) |
| Quartile 4 (1.13, 3.11) | 123 (60) | 1.38 (0.92, 2.06) | 1.28 (0.85, 1.92) | 83 (49) | 1.63 (1.01, 2.61) | 1.53 (0.95, 2.46) |
| P-trend | 0.17 | 0.37 | 0.04 | 0.09 | ||
| Per 1 unit of E-DII | 490 (223) | 1.05 (0.97, 1.13) | 1.03 (0.95, 1.11) | 313 (161) | 1.08 (0.99, 1.19) | 1.07 (0.97, 1.17) |
E-DII, energy-adjusted dietary inflammatory index.
Adjusted for age at diagnosis, site, and stage at diagnosis.
Adjusted for age at diagnosis, site, stage at diagnosis, smoking status, BMI, comorbidities, and household income.
Additionally adjusted for any supplement use within the year prior to diagnosis.
When the analyses were restricted to women with high-grade serous carcinoma, we observed differences in survival patterns by quartiles of the E-DII including supplements (P = 0.01), but not the E-DII excluding supplements (Figure 2). In comparison to the findings for overall ovarian carcinoma, we observed a more pronounced magnitude of effect for most associations (Table 2). For the E-DII including supplements, the more proinflammatory quartiles of the E-DII scores were positively associated with risk of mortality in comparison to the most anti-inflammatory quartile (HRQuartile 3 compared with 1: 1.62; 95% CI: 1.01, 2.61 and HRQuartile 4 compared with 1: 1.58; 95% CI: 0.98, 2.56, respectively), and a one-unit change in the E-DII score was associated with an 8% increase in the risk of mortality (HR: 1.08; 95% CI: 1.01, 1.17). Again, similar trends were observed for the E-DII excluding supplements, but less pronounced than the results for the E-DII including supplements.
FIGURE 2.
Unadjusted Kaplan–Meier survival curves for high-grade serous ovarian carcinoma by quartiles of the E-DII excluding supplements (Panel A) and the E-DII including supplements (Panel B). Quartile 1 is the most anti-inflammatory and Quartile 4 is the most proinflammatory. E-DII, energy-adjusted dietary inflammatory index.
We observed an interaction by smoking status for the association between the E-DII and risk of mortality; P-interaction = 0.03 for the continuous measure of E-DII including supplements (Table 3). Among ever smokers, a >2-fold increased risk of mortality was observed for women consuming the most proinflammatory diet compared with those with the most anti-inflammatory diet (HRQuartile 4 compared with 1: 2.36; 95% CI: 1.21, 4.60), whereas among never smokers, no association between the E-DII and mortality was observed (HRQuartile 4 compared with 1: 0.89; 95% CI: 0.52, 1.53). This finding was consistent for the E-DII excluding supplements, although not as pronounced as the E-DII including supplements. An interaction by age at diagnosis was not found in this study (data not shown).
TABLE 3.
Adjusted HRs and 95% CIs for all-cause deaths in women with ovarian carcinoma by E-DII and stratified by smoking status1
| Never smokers (n = 273) | Ever smokers (n = 217) | |||||
|---|---|---|---|---|---|---|
| Model 12 | Model 23 | Model 12 | Model 23 | |||
| E-DII | Cases (deaths) | HR (95% CI) | HR (95% CI) | Cases (deaths) | HR (95% CI) | HR (95% CI) |
| Including supplements | ||||||
| Quartile 1 (−5.22, −3.50) | 74 (34) | 1.00 (Referent) | 1.00 (Referent) | 49 (15) | 1.00 (Referent) | 1.00 (Referent) |
| Quartile 2 (−3.49, −2.23) | 72 (32) | 0.95 (0.58, 1.56) | 0.90 (0.54, 1.49) | 50 (16) | 1.09 (0.52, 2.26) | 0.92 (0.44, 1.96) |
| Quartile 3 (−2.22, 0.02) | 62 (31) | 1.04 (0.63, 1.72) | 0.95 (0.57, 1.59) | 60 (31) | 2.33 (1.23, 4.42) | 2.30 (1.18, 4.48) |
| Quartile 4 (0.03, 3.11) | 65 (29) | 1.05 (0.63, 1.75) | 0.89 (0.52, 1.53) | 58 (35) | 2.75 (1.46, 5.15) | 2.36 (1.21, 4.60) |
| P-trend | 0.75 | 0.76 | 0.0002 | 0.002 | ||
| Per 1 unit of E-DII | 273 (126) | 1.02 (0.94, 1.10) | 0.99 (0.91, 1.08) | 217 (97) | 1.19 (1.08, 1.31) | 1.16 (1.05, 1.29) |
| Excluding supplements4 | ||||||
| Quartile 1 (−4.15, −2.16) | 71 (30) | 1.00 (Referent) | 1.00 (Referent) | 51 (16) | 1.00 (Referent) | 1.00 (Referent) |
| Quartile 2 (−2.15, −0.48) | 69 (35) | 1.13 (0.68, 1.89) | 1.09 (0.64, 1.83) | 54 (23) | 1.58 (0.82, 3.05) | 1.75 (0.89, 3.45) |
| Quartile 3 (−0.47, 1.12) | 67 (29) | 0.83 (0.48, 1.42) | 0.77 (0.44, 1.32) | 55 (30) | 2.06 (1.08, 3.93) | 2.13 (1.11, 4.11) |
| Quartile 4 (1.13, 3.11) | 66 (32) | 1.10 (0.64, 1.89) | 1.02 (0.58, 1.77) | 57 (28) | 1.98 (1.03, 3.78) | 2.00 (1.02, 3.93) |
| P-trend | 0.92 | 0.67 | 0.03 | 0.04 | ||
| Per 1 unit of E-DII | 273 (126) | 0.99 (0.89, 1.10) | 0.97 (0.87, 1.08) | 217 (97) | 1.13 (1.01, 1.27) | 1.13 (1.00, 1.28) |
E-DII, energy-adjusted dietary inflammatory index.
Adjusted for age at diagnosis, site and stage at diagnosis.
Adjusted for age at diagnosis, site, stage at diagnosis, BMI, comorbidities, and household income.
Additionally adjusted for any supplement use within the year prior to diagnosis.
In the sensitivity analysis evaluating whether clinical variables (neoadjuvant chemotherapy and debulking status) influenced the association between the E-DII score and risk of mortality (Supplemental Table 2), we found that controlling for these clinical variables resulted in an attenuation of the majority of the HRs, and the overall trends were similar to the findings not adjusting for these clinical variables.
Discussion
In a population of African-American women with ovarian carcinoma, we found evidence that a more proinflammatory diet in the year prior to diagnosis was positively associated with risk of all-cause mortality among women with high-grade serous ovarian carcinoma, but not among overall ovarian carcinoma, although suggestive. We also found that the positive risk of mortality was present only among ever smokers. This analysis is the first to evaluate how dietary inflammatory potential impacts outcomes in African-American women with ovarian cancer, and by using the DII, we were able to capture the inflammatory potential of the entire diet, thereby overcoming challenges typically observed when individual dietary components are the focus of an analysis, such as collinearity and loss of power. Our results are also strengthened by the fact that any misclassification of exposure would be nondifferential and would not explain the elevated HRs.
At present, only one study has evaluated the influence of dietary inflammatory potential on ovarian cancer outcomes (17). In contrast to our findings, Nagle et al. (17) did not find any evidence to suggest that dietary inflammatory potential (as measured by both the DII and the empirical-based measure of dietary inflammatory potential) was associated with all-cause or ovarian cancer-specific survival. Given that the prior study was among Australian women, most of whom are white, their findings may not be generalizable to nonwhite women living in the US due to likely differences in food availability and dietary patterns. Considering other cancer types, two studies (25, 26) have evaluated the impact of the prediagnostic DII on the risk of mortality among cancer patients. Similar to our findings in ovarian cancer, a prediagnostic proinflammatory diet conferred a greater risk of mortality in men with prostate cancer but only for those with a higher Gleason score (25). By contrast, no association was found between the DII and survival in breast cancer patients (26). Although data examining the impact of an inflammatory diet on risk of mortality among women with ovarian carcinoma is limited to the present study and Nagle et al. (17), four studies (27–30) have investigated prediagnostic dietary intake and/or dietary quality in ovarian cancer survivors, noting inverse associations with mortality for higher vegetable consumption (27–30) and higher overall diet quality as measured by the Healthy Eating Index (29). Although these prior studies do not directly measure dietary inflammatory potential, the foods and nutrients studied have anti- or proinflammatory properties (9); however, the mechanism underlying how these foods and nutrients may impact ovarian cancer outcomes is unclear.
Previous studies have shown that patients tend to make changes in their diet after a cancer diagnosis (31–34), which has the potential to impact outcomes. In the AACES, data were not collected on postdiagnostic diet, and we were, therefore, unable to assess the impact of postdiagnostic diet or dietary changes on survival in this population. Moreover, at present, no data exist on changes in dietary patterns across the course of diagnosis to treatment of ovarian cancer patients. A randomized, controlled trial, Lifestyle Intervention for Ovarian Cancer Enhanced Survival (LIVES), is currently enrolling women with Stage II–IV ovarian cancer to test whether a diet and physical activity intervention after treatment will impact progression-free survival (35). Hopefully, the findings from this study will shed light on whether postdiagnostic dietary changes can have a positive impact on ovarian cancer outcomes.
Our findings highlight an interaction by smoking status in the association between the E-DII and risk of mortality, where a positive association with mortality was present only among ever smokers but not among never smokers. Due to the small number of current smokers in this study population, we were unable to adequately assess this association among current and former smokers separately. Although a definitive explanation for these results is unclear, we provide a few possibilities for speculation. Residual confounding of the E-DII and risk of mortality association by smoking status may still be present. When we additionally adjust for pack-years among ever smokers, we observed slightly stronger HRs, but the results were virtually the same (data not shown), suggesting that residual confounding by smoking status does not account for these findings. Residual confounding by another variable correlated with smoking status could also be present; however, it is unlikely that this would completely explain the observed association between the E-DII and risk of mortality among ever smokers. Another possible explanation for our findings is a synergistic effect of the inflammation-related consequences of both diet and smoking status. Tobacco smoke is known to alter concentrations of systemic inflammatory markers (36). Although this can occur through nonoxidative mechanisms (37, 38), the most important mechanism is likely through the production of reactive oxygen species (ROS) (39), which cause cellular damage and have been linked to worse outcomes and therapeutic resistance in cancer patients (40). The harmful effects of ROS can be counterbalanced by antioxidants (41) that are usually a part of an anti-inflammatory diet. Thus, the positive association between a more proinflammatory diet and mortality among ever smokers may be due to oxidative stress through an accumulation of ROS and a depletion of antioxidants.
The majority of associations between the E-DII and risk of all-cause mortality were found only for the E-DII including supplements; however, trends were similar for the E-DII excluding supplements. As there is a wider range and more variability of E-DII scores for the E-DII including supplements than the E-DII excluding supplements, it is possible that we did not have adequate power to detect an association for the E-DII excluding supplements. Nevertheless, the association between dietary supplements and outcomes in cancer patients is controversial and findings across studies are largely inconsistent (42–44). To our knowledge, no studies have directly evaluated supplement use and mortality among ovarian cancer patients, but one study observed that selenium supplementation during treatment reduced adverse events from chemotherapy in ovarian cancer patients (45). The high prevalence of supplement use among cancer patients (46) coupled with the uncertainty of whether supplement use is beneficial or harmful for cancer patients underscores the need for additional research in this area to inform clinical care guidelines on the use of supplements throughout the disease course.
The attenuation of the HRs after adjustment for a priori confounders was driven largely by household income. This is not entirely surprising given the known differences in the receipt of adequate and quality healthcare for cancer patients by income, which can affect survival outcomes. In ovarian cancer, Long et al. (47) noted that patients with a low socioeconomic status were less likely to receive adequate treatment. Additionally, racial/ethnic and socioeconomic influences on dietary intake through food availability and access through the built environment are complex (48). Several studies have shown that healthier food items, such as fruits and vegetables, which generally have more anti-inflammatory properties, are more expensive (49, 50). Moreover, low-income or predominantly black neighborhoods often entirely lack or have less access to food outlets with high-quality foods (e.g. supermarkets, grocery stores, farmer's markets) and have a greater number of fast-food restaurants than more affluent or predominantly white neighborhoods (51–53).
A number of AACES subjects (n = 55) were missing data on two important confounders, stage at diagnosis and household income. To avoid a loss of precision and power in our analyses, we utilized multiple imputation under the assumption that data were missing at random or completely at random. We believe this assumption is reasonable for household income as income was not asked in the short version of the AACES questionnaire, and thus, the missingness of that variable can be explained by the characteristics related to why subjects elected to complete the short versus the long version of the questionnaire (e.g. severity of the disease, age at diagnosis) which are adjusted for in these analyses. Moreover, in these analyses, multiple imputation was preferred over the complete case analysis approach because the latter likely introduced a selection bias. We noted statistically significant differences between the women included versus excluded from the complete case analysis by age (P = 0.01), where women excluded from the complete case analysis were older than those included. In this study population, older women tended to consume a more anti-inflammatory diet and had a greater risk of mortality; therefore, the validity of the complete case analysis findings is questionable.
There are a few notable limitations in the present study. Even though cases were identified using a rapid case ascertainment approach, the median time from diagnosis to enrollment was ∼6 mo, resulting in an inability to enroll rapidly fatal cases in this study. Therefore, the results of the present study are generalizable to ovarian cancer patients that survive ≥6 mo after diagnosis. Follow-up time was defined as the time from the date of diagnosis to the date of death or last contact; however, because no outcomes could have occurred from the date of diagnosis to the date of entry into the study, this approach may have introduced a bias (54). Using the time from the interview date (date of entry into the study) to the date of death or last contact to calculate follow-up time (left censoring), we repeated the analyses and the results were virtually unchanged (data not shown). As we estimated the date of last contact based on the date of the latest vital status update from each study site, we repeated the analyses using the recorded date of last contact and again, observed minimal to no changes in the results, with the conclusions remaining the same (data not shown). We were unable to evaluate 18 of the 45 food parameters utilized in the DII calculations because they were not measured in the FFQ. Previous research has shown that a similar drop off in parameter availability had no material effect on risk estimates (12). In the analyses evaluating interaction by smoking status, we used the cut-points for the E-DII quartiles from the entire study population for comparative purposes between ever and never smokers; however, utilizing stratum-specific cut-points for the E-DII quartiles resulted in similar findings (data not shown). Although there are differences in outcomes by ovarian carcinoma histotype (55), we were limited by small numbers to evaluate the association between the DII and risk of mortality in the less common histotypes.
In summary, the results of the present study suggest that greater inflammatory potential of a prediagnostic diet may adversely impact prognosis among African-American women with ovarian carcinoma, especially among women with high-grade serous ovarian carcinoma and only among ever smokers. Given the paucity of research on the relation between diet and ovarian cancer survival, it is critical that these results are confirmed in other studies, preferably with inclusion of inflammatory biomarkers as well as diverse racial/ethnic populations. Moreover, observational studies which longitudinally follow ovarian cancer survivors from diagnosis through treatment are especially needed to determine how the inflammatory potential of postdiagnostic diet or dietary changes may play a role in prognosis.
Supplementary Material
Acknowledgments
We would like to acknowledge the AACES interviewers, Christine Bard, LaTonda Briggs, Whitney Franz (North Carolina), and Robin Gold (Detroit). We also acknowledge the individuals responsible for facilitating case ascertainment across the 10 sites including: Christie McCullum-Hill (Alabama); the Metropolitan Detroit Cancer Surveillance System staff (Detroit); Rana Bayakly, Vicki Bennett, Judy Andrews, and Debbie Chambers (Georgia); the Louisiana Tumor Registry; Lisa Paddock and Manisha Narang (New Jersey); Diana Slone, Yingli Wolinsky, Steven Waggoner, Anne Heugel, Nancy Fusco, Kelly Ferguson, Peter Rose, Deb Strater, Taryn Ferber, Donna White, Lynn Borzi, Eric Jenison, Nairmeen Haller, Debbie Thomas, Vivian von Gruenigen, Michele McCarroll, Joyce Neading, John Geisler, Stephanie Smiddy, David Cohn, Michele Vaughan, Luis Vaccarello, Elayna Freese, James Pavelka, Pam Plummer, William Nahhas, Ellen Cato, John Moroney, Mark Wysong, Tonia Combs, Marci Bowling, Brandon Fletcher (Ohio); Susan Bolick, Donna Acosta, Catherine Flanagan (South Carolina); Martin Whiteside (Tennessee) and Georgina Armstrong and the Texas Registry, Cancer Epidemiology and Surveillance Branch, Department of State Health Services.
The authors’ contributions were as follows—LCP, JMS, EVB, and BQ: contributed to the design of the study. JMS, AJA, EVB, MLB, MLC, JSB-S, EF, PGM, ESP, AGS, and PDT: contributed to data acquisition. JRH and NS: computed the E-DII including and excluding supplements, and LCP, DC, and TFC: performed all statistical analyses. LCP: drafted the manuscript, and all authors: contributed to the interpretation of results and reviewed and approved the final manuscript.
Notes
Supported by the National Cancer Institute (R01CA142081, K99CA218681,and R00CA218681 to LCP). Additional support was provided by the Metropolitan Detroit Cancer Surveillance System with funding from the National Cancer Institute, NIH, and the Department of Health and Human Services (Contract HHSN261201000028C), and the Epidemiology Research Core, supported in part by the National Cancer Institute (P30CA22453) to the Karmanos Cancer Institute, Wayne State University School of Medicine. The New Jersey State Cancer Registry, Cancer Epidemiology Services, New Jersey Department of Health, is funded by the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute under contract HHSN261201300021I, the National Program of Cancer Registries (NPCR), Centers for Disease Control and Prevention under grant 5U58DP003931-02 as well as the State of New Jersey and the Rutgers Cancer Institute of New Jersey.
Author disclosures: JRH owns controlling interest in Connecting Health Innovations LLC (CHI), a company planning to license the right to his invention of the dietary inflammatory index (DII®) from the University of South Carolina in order to develop computer and smart phone applications for patient counseling and dietary intervention in clinical settings. NS is an employee of CHI. The subject matter of this paper will not have any direct bearing on that work, nor has that activity exerted any influence on this project. All other authors, no conflicts of interest.
Supplemental Tables 1 and 2 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.
Abbreviations used: AACES, African-American Cancer Epidemiology Study; DII, dietary inflammatory index; E-DII, energy-adjusted dietary inflammatory index; LIVES, Lifestyle Intervention for Ovarian Cancer Enhanced Survival; ROS, reactive oxygen species.
References
- 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68:7–30. [DOI] [PubMed] [Google Scholar]
- 2. Jemal A, Ward EM, Johnson CJ, Cronin KA, Ma J, Ryerson AB, Mariotto A, Lake AJ, Wilson R, Sherman RL et al.. Annual Report to the Nation on the Status of Cancer, 1975–2014, Featuring Survival. J Natl Cancer Inst. 2017;109:djx030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ness RB, Cottreau C.. Possible role of ovarian epithelial inflammation in ovarian cancer. J Natl Cancer Inst. 1999;91:1459–67. [DOI] [PubMed] [Google Scholar]
- 4. Trabert B, Ness RB, Lo-Ciganic W-H, Murphy MA, Goode EL, Poole EM, Brinton LA, Webb PM, Nagle CM, Jordan SJ et al.. Aspirin, nonaspirin nonsteroidal anti-inflammatory drug, and acetaminophen use and risk of invasive epithelial ovarian cancer: a pooled analysis in the Ovarian Cancer Association Consortium. J Natl Cancer Inst. 2014;106:djt431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Poole EM, Lee IM, Ridker PM, Buring JE, Hankinson SE, Tworoger SS. A prospective study of circulating C-reactive protein, interleukin-6, and tumor necrosis factor alpha receptor 2 levels and risk of ovarian cancer. Am J Epidemiol. 2013;178:1256–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wang Y-Q, Jin C, Zheng H-M, Zhou K, Shi B-B, Zhang Q, Zheng F-Y, Lin F. A novel prognostic inflammation score predicts outcomes in patients with ovarian cancer. Clin Chim Acta. 2016;456:163–9. [DOI] [PubMed] [Google Scholar]
- 7. Williams KA, Labidi-Galy SI, Terry KL, Vitonis AF, Welch WR, Goodman A, Cramer DW. Prognostic significance and predictors of the neutrophil-to-lymphocyte ratio in ovarian cancer. Gynecol Oncol. 2014;132:542–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Hefler LA, Concin N, Hofstetter G, Marth C, Mustea A, Sehouli J, Zeillinger R, Leipold H, Lass H, Grimm C et al.. Serum C-reactive protein as independent prognostic variable in patients with ovarian cancer. Clin Cancer Res. 2008;14:710–4. [DOI] [PubMed] [Google Scholar]
- 9. Galland L. Diet and inflammation. Nutr Clin Pract. 2010;25:634–40. [DOI] [PubMed] [Google Scholar]
- 10. Cavicchia PP, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS, He JR. A new dietary inflammatory index predicts interval changes in serum high-sensitivity C-reactive protein. J Nutr. 2009;139:2365–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Shivappa N, Steck SE, Hurley TG, Hussey JR, Hébert JR. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. 2014;17:1689–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Shivappa N, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS, Tabung FK, Hebert JR. A population-based dietary inflammatory index predicts levels of C-reactive protein in the Seasonal Variation of Blood Cholesterol Study (SEASONS). Public Health Nutr. 2014;17:1825–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Tabung FK, Steck SE, Zhang J, Ma Y, Liese AD, Agalliu I, Hingle M, Hou L, Hurley TG, Jiao L et al.. Construct validation of the dietary inflammatory index among postmenopausal women. Ann Epidemiol. 2015;25:398–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Wirth MD, Shivappa N, Davis L, Hurley TG, Ortaglia A, Drayton R, Blair SN, Hébert JR. Construct validation of the dietary inflammatory index among African Americans. J Nutr Health Aging. 2017;21:487–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Peres LC, Bandera EV, Qin B, Guertin KA, Shivappa N, Hebert JR, Abbott SE, Alberg AJ, Barnholtz-Sloan JS, Bondy ML et al.. Dietary inflammatory index and risk of epithelial ovarian cancer in African-American women. Int J Cancer. 2017;140:535–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Shivappa N, Hebert JR, Rosato V, Rossi M, Montella M, Serraino D, La Vecchia C. Dietary inflammatory index and ovarian cancer risk in a large Italian case-control study. Cancer Causes Control. 2016;27(7):897–906. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Nagle CM, Ibiebele T, Shivappa N, Hebert JR, DeFazio A, Webb PM; Australian Ovarian Cancer Study . The association between the inflammatory potential of diet and risk of developing, and survival following, a diagnosis of ovarian cancer. Eur J Nutr. 2018. doi: 10.1007/s00394-018-1779-x. [DOI] [PubMed] [Google Scholar]
- 18. Shivappa N, Hebert JR, Paddock LE, Rodriguez-Rodriguez L, Olson SH, Bandera EV. Dietary inflammatory index and ovarian cancer risk in a New Jersey case-control study. Nutrition. 2018;46:78–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tabung FK, Huang T, Giovannucci EL, Smith-Warner SA, Tworoger SS, Poole EM. The inflammatory potential of diet and ovarian cancer risk: results from two prospective cohort studies. Br J Cancer. 2017;117:907–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Schildkraut JM, Alberg AJ, Bandera EV, Barnholtz-Sloan J, Bondy M, Cote ML, Funkhouser E, Peters E, Schwartz AG, Terry P et al.. A multi-center population-based case-control study of ovarian cancer in African-American women: the African American Cancer Epidemiology Study (AACES). BMC Cancer. 2014;14:688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. White IR, Royston P.. Imputing missing covariate values for the Cox model. Stat Med. 2009;28:1982–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Zwakman N, van de Laar R, Van Gorp T, Zusterzeel PL, Snijders MP, Ferreira I, Massuger LF, Kruitwagen RF. Perioperative changes in serum CA125 levels: a prognostic factor for disease-specific survival in patients with ovarian cancer. J Gynecol Oncol. 2017;28:e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Zorn KK, Tian C, McGuire WP, Hoskins WJ, Markman M, Muggia FM, Rose PG, Ozols RF, Spriggs D, Armstrong DK. The prognostic value of pretreatment CA 125 in patients with advanced ovarian carcinoma: a Gynecologic Oncology Group study. Cancer. 2009;115:1028–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Pelissier A, Bonneau C, Chereau E, de La Motte Rouge T, Fourchotte V, Darai E, Rouzier R. CA125 kinetic parameters predict optimal cytoreduction in patients with advanced epithelial ovarian cancer treated with neoadjuvant chemotherapy. Gynecol Oncol. 2014;135:542–6. [DOI] [PubMed] [Google Scholar]
- 25. Zucchetto A, Gini A, Shivappa N, Hébert JR, Stocco C, Dal Maso L, Birri S, Serraino D, Polesel J. Dietary inflammatory index and prostate cancer survival. Int J Cancer. 2016;139:2398–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Zucchetto A, Serraino D, Shivappa N, Hébert JR, Stocco C, Puppo A, Falcini F, Panato C, Dal Maso L, Polesel J. Dietary inflammatory index before diagnosis and survival in an Italian cohort of women with breast cancer. Br J Nutr. 2017;117:1456–62. [DOI] [PubMed] [Google Scholar]
- 27. Nagle CM, Purdie DM, Webb PM, Green A, Harvey PW, Bain CJ. Dietary influences on survival after ovarian cancer. Int J Cancer. 2003;106:264–9. [DOI] [PubMed] [Google Scholar]
- 28. Dolecek TA, McCarthy BJ, Joslin CE, Peterson CE, Kim S, Freels SA, Davis FG. Prediagnosis food patterns are associated with length of survival from epithelial ovarian cancer. J Am Diet Assoc. 2010;110:369–82. [DOI] [PubMed] [Google Scholar]
- 29. Thomson CA, Crane TE, Wertheim BC, Neuhouser ML, Li W, Snetselaar LG, Basen-Engquist KM, Zhou Y, Irwin ML. Diet quality and survival after ovarian cancer: results from the Women's Health Initiative. J Natl Cancer Inst. 2014;106: pii:dju314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Playdon MC, Nagle CM, Ibiebele TI, Ferrucci LM, Protani MM, Carter J, Hyde SE, Neesham D, Nicklin JL, Mayne ST et al.. Pre-diagnosis diet and survival after a diagnosis of ovarian cancer. Br J Cancer. 2017;116:1627–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Alfano CM, Day JM, Katz ML, Herndon JE, Bittoni MA, Oliveri JM, Donohue K, Paskett ED. Exercise and dietary change after diagnosis and cancer-related symptoms in long-term survivors of breast cancer: CALGB 79804. Psychooncology. 2009;18:128–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Vance V, Campbell S, McCargar L, Mourtzakis M, Hanning R. Dietary changes and food intake in the first year after breast cancer treatment. Appl Physiol Nutr Metab. 2014;39:707–14. [DOI] [PubMed] [Google Scholar]
- 33. Avery KNL, Donovan JL, Horwood J, Neal DE, Hamdy FC, Parker C, Wade J, Lane A. The importance of dietary change for men diagnosed with and at risk of prostate cancer: a multi-centre interview study with men, their partners and health professionals. BMC Family Practice. 2014;15:81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Blanchard CM, Denniston MM, Baker F, Ainsworth SR, Courneya KS, Hann DM, Gesme DH, Reding D, Flynn T, Kennedy JS. Do adults change their lifestyle behaviors after a cancer diagnosis?. Am J Health Behav. 2003;27:246–56. [DOI] [PubMed] [Google Scholar]
- 35. Thomson CA, Crane TE, Miller A, Garcia DO, Basen-Engquist K, Alberts DS. A randomized trial of diet and physical activity in women treated for stage II–IV ovarian cancer: rationale and design of the Lifestyle Intervention for Ovarian Cancer Enhanced Survival (LIVES): an NRG Oncology/Gynecologic Oncology Group (GOG-225) study. Contemp Clin Trials. 2016;49:181–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Shiels MS, Katki HA, Freedman ND, Purdue MP, Wentzensen N, Trabert B, Kitahara CM, Furr M, Li Y, Kemp TJ et al.. Cigarette smoking and variations in systemic immune and inflammation markers. J Natl Cancer Inst. 2014;106(11):dju294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Tetley TD, Smith SF, Burton GH, Winning AJ, Cooke NT, Guz A. Effects of cigarette smoking and drugs on respiratory tract proteases and antiproteases. Eur J Respir Dis Suppl. 1987;153:93–102. [PubMed] [Google Scholar]
- 38. Laurent P, Bieth JG.. Cigarette smoke decreases the rate constant for the association of elastase with α1-proteinase inhibitor by a non-oxidative mechanism. Biochem Biophys Res Commun. 1985;126:275–81. [DOI] [PubMed] [Google Scholar]
- 39. Valavanidis A, Vlachogianni T, Fiotakis K. Tobacco smoke: involvement of reactive oxygen species and stable free radicals in mechanisms of oxidative damage, carcinogenesis and synergistic effects with other respirable particles. Int J Environ Res Public Health. 2009;6:445–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Kumari S, Badana AK, G MM, G S, Malla R. Reactive oxygen species: a key constituent in cancer survival. Biomark Insights. 2018;13:1177271918755391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Birben E, Sahiner UM, Sackesen C, Erzurum S, Kalayci O. Oxidative stress and antioxidant defense. World Allergy Organ J. 2012;5:9–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Rock CL, Doyle C, Demark-Wahnefried W, Meyerhardt J, Courneya KS, Schwartz AL, Bandera EV, Hamilton KK, Grant B, McCullough M et al.. Nutrition and physical activity guidelines for cancer survivors. CA Cancer J Clin. 2012;62:243–74. [DOI] [PubMed] [Google Scholar]
- 43. Harvie M. Nutritional supplements and cancer: potential benefits and proven harms. Am Soc Clin Oncol Educ Book. 2014:e478–86. [DOI] [PubMed] [Google Scholar]
- 44. Hardy ML. Dietary supplement use in cancer care: help or harm. Hematol Oncol Clin North Am. 2008;22:581–617., vii. [DOI] [PubMed] [Google Scholar]
- 45. Sieja K, Talerczyk M.. Selenium as an element in the treatment of ovarian cancer in women receiving chemotherapy. Gynecol Oncol. 2004;93:320–7. [DOI] [PubMed] [Google Scholar]
- 46. Velicer CM, Ulrich CM. Vitamin and mineral supplement use among US adults after cancer diagnosis: a systematic review. J Clin Oncol. 2008;26:665–73. [DOI] [PubMed] [Google Scholar]
- 47. Long B, Chang J, Ziogas A, Tewari KS, Anton-Culver H, Bristow RE. Impact of race, socioeconomic status, and the health care system on the treatment of advanced-stage ovarian cancer in California. Am J Obstet Gynecol. 2015;212:468.e1–.e9. [DOI] [PubMed] [Google Scholar]
- 48. Walker RE, Keane CR, Burke JG. Disparities and access to healthy food in the United States: a review of food deserts literature. Health & Place. 2010;16:876–84. [DOI] [PubMed] [Google Scholar]
- 49. Monsivais P, Drewnowski A.. The rising cost of low-energy-density foods. J Am Diet Assoc. 2007;107:2071–6. [DOI] [PubMed] [Google Scholar]
- 50. Darmon N, Drewnowski A.. Contribution of food prices and diet cost to socioeconomic disparities in diet quality and health: a systematic review and analysis. Nutr Rev. 2015;73:643–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Powell LM, Slater S, Mirtcheva D, Bao Y, Chaloupka FJ. Food store availability and neighborhood characteristics in the United States. Prev Med. 2007;44:189–95. [DOI] [PubMed] [Google Scholar]
- 52. Block JP, Scribner RA, DeSalvo KB. Fast food, race/ethnicity, and income. Am J Prev Med. 2004;27:211–7. [DOI] [PubMed] [Google Scholar]
- 53. Zenk SN, Powell LM, Rimkus L, Isgor Z, Barker DC, Ohri-Vachaspati P, Chaloupka F. Relative and absolute availability of healthier food and beverage alternatives across communities in the United States. Am J Public Health. 2014;104:2170–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Lash TL, Cole SR.. Immortal person-time in studies of cancer outcomes. J Clin Oncol. 2009;27:e55–e6. [DOI] [PubMed] [Google Scholar]
- 55. Peres LC, Cushing-Haugen KL, Köbel M, Harris HR, Berchuck A, Rossing MA, Schildkraut JM, Doherty JA. Invasive epithelial ovarian cancer survival by histotype and disease stage. J Natl Cancer Inst. 2019;111(1):60–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.