Abstract
Background: Chronic inflammation is one of the mechanisms involved in carcinogenesis. Diet is a major source of pro- and anti-inflammatory compounds. The Dietary Inflammatory Index (DII) was designed to estimate its overall inflammatory potential.
Objective: Our objective was to investigate the associations between the DII and overall, breast, and prostate cancer risks.
Methods: This prospective study included 6542 participants [3771 women and 2771 men with a mean ± SD age of 49.2 ± 6.4 y and a BMI (in kg/m2) of 24.0 ± 3.6 at baseline] from the Supplémentation en VItamines et Minéraux AntioXydants (SU.VI.MAX) cohort who completed at least six 24-h dietary records during the first 2 y of follow-up. The DII was based on 36 food variables. Higher scores corresponded to more proinflammatory diets. A total of 559 incident cancers were diagnosed (median follow-up, 12.6 y), including 158 female breast and 123 prostate cancers (the 2 main cancer sites in this cohort). Associations were characterized by multivariable Cox proportional hazards models. Stratified analyses were performed according to the median of usual daily alcohol intake.
Results: Sex-specific quartiles of the DII were positively associated with prostate cancer risk [quartile (Q) 4 compared with Q1, HR: 2.08; 95% CI: 1.06, 4.09] but not with overall or breast cancer risks. There was an interaction between the DII and alcohol intake (grams per day) on overall cancer risk (P-interaction = 0.02): the DII was positively associated with overall cancer risk in low-to-moderate alcohol drinkers (Q4 compared with Q1 HR: 1.75; 95% CI: 1.15, 2.68; P-trend = 0.02), whereas no association was detected in higher consumers of alcohol (P-trend = 0.8). This interaction was also observed for breast cancer (P-interaction = 0.001).
Conclusion: Consistent with mechanistic data, findings from this study indicated that proinflammatory diets are associated with increased prostate cancer risk and, in low-to-moderate alcohol drinkers, with increased overall and breast cancer risk. The SU.VI.MAX trial was registered at clinicaltrials.gov as NCT00272428.
Keywords: cancer, diet, Dietary Inflammatory Index, inflammation, prospective study
Introduction
The involvement of chronic inflammation in carcinogenesis is increasingly acknowledged. In many experimental studies, inflammatory mechanisms involving inflammatory cells and signaling pathways also have been described as promoting carcinogenesis (1). Moreover, as reported by 2 meta-analyses (2, 3), epidemiologic studies have shown associations between inflammatory biomarkers such as C-reactive protein (CRP)10 and IL-6 and an increased risk of overall and lung cancer. A previous prospective study nested in the Supplémentation en VItamines et Minéraux AntioXydants (SU.VI.MAX) cohort found that baseline high-sensitivity CRP was associated with an increased overall cancer and prostate cancer risk (4).
A growing body of literature has investigated the influence of diet on chronic inflammation (5–8). Indeed, diet is a major source of bioactive compounds for which either pro- or anti-inflammatory properties have been identified (9). The Dietary Inflammatory Index (DII) is a literature-derived scoring system that was developed to estimate the pro- and anti-inflammatory potential of the overall diet and to investigate its possible relations with chronic diseases (9).
Globally, cancer is one of the leading causes of death, with about one-third of cancers attributed to unhealthy lifestyle and dietary behaviors, i.e., modifiable risk factors (10). To our knowledge, few studies have investigated the pro- or anti-inflammatory potential of the diet, measured with the DII, in relation to cancer risk (11–23). Moreover, only 4 of these studies were based on a prospective design and focused on colorectal cancer (16, 18, 19) or lung cancer (14). The associations between DII and female breast cancer (13) or prostate cancer (12, 17) have only been explored in nonprospective case-control studies.
Therefore, our objective was to investigate the relations between the DII and overall, female breast, and prostate cancer risks in a large French prospective cohort.
Methods
Study population.
The SU.VI.MAX study is a population-based double-blind, placebo-controlled, randomized trial (NCT00272428) initially designed to assess the effect of a daily antioxidant supplementation on coronary heart disease and cancer incidence. All participants took a single daily capsule containing a combination of 120 mg of ascorbic acid, 30 mg of vitamin E, 6 mg of β-carotene, 100 μg of selenium, and 20 mg of zinc, or a placebo (24). Participants were advised against taking any self-prescribed supplements during the trial. The study was conducted according to the Declaration of Helsinki guidelines and was approved by the Ethics Committee for Studies with Human Subjects at the Paris-Cochin Hospital (CCPPRB No. 706) and the Commission Nationale de l’Informatique et des Libertés (French National Commission for Computed Data and Individual Freedom) (CNIL No. 334641). During the first-year study, 13,017 subjects were enrolled (1994–1995) and health events monitoring was pursued until September 2007. Written informed consent was obtained from all participants.
Baseline data collection.
At enrollment, all participants completed questionnaires on sociodemographics, smoking status, physical activity, and medication use. They also underwent a clinical examination and anthropometric measurements by study nurses and physicians. Age at menopause (natural or artificial) was self-reported by women during follow-up.
Dietary data were collected with the use of the Minitel Telematic Network (an Internet prototype), which was widely used in France as an adjunct to the telephone at the beginning of the study. During the trial phase, participants were invited to provide a 24-h dietary record every 2 mo (randomly distributed over weekend and week days). In order to comply with the prospective design, dietary records from the first 2 y of follow-up were used in the present study. The participants selected the foods consumed at breakfast, lunch, dinner, and/or any other occasion throughout the day from a predefined list of ∼1000 items. Portion sizes were estimated with the use of a validated picture booklet distributed at enrollment (25). Energy, alcohol (grams per day), and nutrient intake was estimated with the use of a published French food composition table (26).
DII computation.
The computation of the DII was previously described in detail (9). Steps involved in the DII calculation can be found in Supplemental Figure 1. Briefly, the DII is a score initially designed with 45 dietary variables determined from a literature review of 1943 articles published up to 2010. A literature-derived inflammatory effect score was assigned to every micronutrient, macronutrient, or food variable associated with an increase (+1), a decrease (−1), or no effect (0) on 6 major inflammatory biomarkers, including IL-1β, IL-4, IL-6, IL-10, TNF-α, and CRP.
Mean intake of every food variable was transformed with standardized values from a world database into a z score, then converted to a percentile and centered. Finally, the centered percentile score for each food variable was multiplied by its associated coefficient, then these were summed across the 45 dietary variables, thus providing an individual DII score. In the present study, the DII was based on 36 dietary variables available in the database. Total intake of energy, protein, carbohydrate, total fat, SFAs, cholesterol, vitamin B-12, and iron were computed into the score as proinflammatory factors, whereas dietary intake of MUFAs, PUFAs, n–3 FAs, n–6 FAs, alcohol, fiber, magnesium, niacin, thiamin, riboflavin, vitamin B-6, vitamin A, vitamin C, vitamin D, vitamin E, folic acid, β-carotene, anthocyanidins, flavan-3-ol, flavonols, flavanones, flavones, isoflavones, garlic, ginger, pepper, onions, and tea were entered as anti-inflammatory factors. Theoretical DII scores ranged from −8.9 (maximally anti-inflammatory diet) to +8.0 (maximally proinflammatory diet). Higher DII scores indicated a more proinflammatory diet.
Case ascertainment.
Major health events were self-reported by subjects during follow-up (between inclusion in 1994 and September 2007). Investigations allowed collection of medical case data from participants themselves, physicians, and/or hospitals. An independent expert committee of physicians reviewed the medical information. Cases were validated through a pathologic report and classified with the use of the International Classification of Diseases, 10th Revision, Clinical Modification (27). All first-incident primary cancers were considered as cases in this study, except for basal cell carcinoma (not considered to be cancer).
Statistical analysis.
Of the 13,017 participants of the SU.VI.MAX study, 161 were excluded because of a cancer diagnosis reported before the start of the follow-up. Of the remaining participants, 6542 provided ≥6 valid 24-h dietary records within the first 2 y of follow-up, and thus were available for analysis. For all covariates, <5% of values were missing and were replaced by the respective mode value.
Baseline characteristics of participants were compared between cases and noncases with the use of χ2 tests or Student’s t tests, as appropriate.
HRs and 95% CIs obtained from Cox proportional hazards models with age as the primary time variable were used to characterize the association between sex-specific quartiles of the DII and overall, female breast, and prostate cancer risk. Participants contributed person-time until the date of diagnosis of the studied cancer, the date of the last completed questionnaire, the date of death, or 30 September 2007, whichever occurred first. Participants who reported a cancer other than the one studied during the follow-up were included and censored at the date of that diagnosis (except for basal cell carcinoma, which was not considered to be cancer). We confirmed that the assumptions of proportionality were satisfied through examination of the log–log (survival) compared with log–time plots. The SAS macro %RCS_Reg (28) was used to test for a possible threshold effect and to select the most appropriate cutoff.
Multivariable models were adjusted for age (time-scale in the Cox model), sex, intervention group of the initial SU.VI.MAX trial (antioxidant or placebo), number of dietary records (continuous), BMI (in kg/m2) (<25, ≥25 to <30, or ≥30), height (continuous), physical activity (irregular, <1 h/d, or ≥1 h walking or equivalent/d), smoking status (never, former, or current), educational level (primary, secondary, or university), family history of cancer in first-degree relatives (yes or no), daily mean energy intake without alcohol (kilocalories per day; continuous), and daily mean alcohol intake (grams per day; continuous). For breast cancer analyses, additional adjustments were performed for menopausal status at baseline (yes or no), use of hormonal treatment for menopause at baseline (yes or no), number of live births (continuous), and family history of breast cancer in first-degree relatives (yes or no). For prostate cancer analyses, additional adjustments were applied for prostate-specific antigen concentration at baseline (continuous) and family history of prostate cancer in first-degree relatives (yes or no).
Because alcohol intake is an important risk factor for several cancer sites (29) and is considered to be an anti-inflammatory factor in formulating the DII (9), we tested the interaction between alcohol intake (sex-specific population median) and DII (sex-specific quartiles). Analyses were conducted overall and then stratified by alcohol intake. Two-factor interactions were also tested between the DII and the intervention group of the initial SU.VI.MAX trial, BMI, and sex.
Models also were computed separately for premenopausal and postmenopausal breast cancer (women contributed to the premenopausal model until their age of menopause, or from their age of menopause, respectively). Sensitivity analyses were carried out while excluding cancer cases diagnosed during the first year of follow-up. All tests were 2-sided, and P < 0.05 was considered to be statistically significant. SAS software version 9.3 was used for analyses.
Results
Median follow-up was 12.6 y. DII scores ranked from −5.1 to 6.0 for women (mean ± SD: 1.0 ± 1.8) and −5.0 to 5.3 for men (mean ± SD: 0.3 ± 1.8). A total of 5.2% of the subjects were lost to follow-up. During follow-up, a total of 559 incident cancers (512 invasive and 47 in situ) were diagnosed, including 158 female breast, 123 prostate, and 278 other cancers (including 51 skin, 32 colorectal, 19 thyroid, 15 lung, and 161 other cancers). Regarding female breast cancer cases, 59% were estrogen receptor–positive and 43% were progesterone receptor–positive. Among prostate cancer cases, 5% had a Gleason score between 2 and 4, 87% between 5 and 7, and 8% between 8 and 10. Study population characteristics are described in Table 1. Compared with those without cancer, those with cancer (overall) were older and taller; were more frequently male; had a slightly higher number of children, higher alcohol intake, and a higher baseline plasma prostate-specific antigen concentration (in men); and had more family members with a history of cancer.
TABLE 1.
Baseline characteristics of the study population, SU.VI.MAX cohort, France, 1994–20071
| Noncases(n = 5983) | Overall cancer cases (n = 559) | Breast cancer cases(n = 158) | Prostate cancer cases(n = 123) | P2 | |
| DII | 0.7 ± 1.9 | 0.6 ± 1.8 | 0.9 ± 1.8 | 0.3 ± 1.5 | 0.4 |
| Age, y | 49.0 ± 6.3 | 51.7 ± 6.3 | 49.7 ± 6.4 | 54.9 ± 4.7 | <0.0001 |
| Sex | 0.02 | ||||
| M | 2508 (41.9) | 263 (47.0) | 123 (100) | ||
| F | 3475 (58.1) | 296 (53.0) | 158 (100) | ||
| Intervention group | 0.9 | ||||
| Antioxidants | 2977 (49.8) | 276 (49.4) | 73 (46.2) | 61 (49.6) | |
| Placebo | 3006 (50.2) | 283 (50.6) | 85 (53.8) | 62 (50.4) | |
| Educational level | 0.7 | ||||
| Primary | 1226 (20.5) | 109 (19.5) | 33 (20.9) | 24 (19.5) | |
| Secondary | 2264 (37.8) | 221 (39.5) | 53 (33.5) | 47 (38.2) | |
| University | 2493 (41.7) | 229 (41.0) | 72 (45.6) | 52 (42.3) | |
| Smoking status | 0.09 | ||||
| Never smoker | 2891 (48.3) | 250 (44.7) | 87 (55.1) | 47 (38.2) | |
| Former smoker | 2318 (38.7) | 220 (39.4) | 42 (26.6) | 64 (52.0) | |
| Current smoker | 774 (12.9) | 89 (15.9) | 29 (18.4) | 12 (9.8) | |
| Physical activity | 0.7 | ||||
| Irregular | 1457 (24.4) | 142 (25.4) | 50 (31.6) | 29 (23.6) | |
| <1 h walking or equivalent/d | 1810 (30.3) | 160 (28.6) | 53 (33.5) | 31 (25.2) | |
| ≥1 h walking or equivalent/d | 2716 (45.4) | 257 (46.0) | 55 (34.8) | 63 (51.2) | |
| Children, n | 2.1 ± 1.1 | 2.2 ± 1.2 | 2.1 ± 1.2 | 2.4 ± 1.2 | 0.008 |
| Height, cm | 167 ± 8.4 | 167 ± 8.3 | 162 ± 6.2 | 173 ± 6.3 | 0.03 |
| BMI, kg/m2 | 0.2 | ||||
| <25 | 3950 (66.0) | 360 (64.4) | 123 (77.8) | 65 (52.8) | |
| ≥25 to <30 | 1679 (28.1) | 156 (27.9) | 20 (12.7) | 45 (36.6) | |
| ≥30 | 354 (5.9) | 43 (7.7) | 15 (9.5) | 13 (10.6) | |
| 24-h dietary records, n | 10.6 ± 2.4 | 10.6 ± 2.4 | 10.4 ± 2.4 | 11.0 ± 2.3 | 0.8 |
| Energy intake (without alcohol), kcal/d | 1979 ± 538 | 2020 ± 552 | 1778 ± 414 | 2276 ± 544 | 0.08 |
| Alcohol intake, g/d | 18.6 ± 20.6 | 21.3 ± 21.5 | 12.9 ± 12.1 | 30.2 ± 22.1 | 0.003 |
| Family history of cancer3 | |||||
| All cancers | 2091 (34.9) | 220 (39.4) | 0.04 | ||
| Breast cancer4 | 292 (8.4) | 36 (12.2) | 24 (15.2) | 0.08 | |
| Prostate cancer | 117 (4.7) | 23 (8.7) | 16 (13.0) | 0.03 | |
| Plasma PSA, μg/L | 1.2 ± 1.3 | 2.3 ± 2.8 | 3.4 ± 3.7 | <0.0001 |
Values are means ± SDs for quantitative variables and n (%) for qualitative variables. DII, Dietary Inflammatory Index; PSA, prostate-specific antigen; SU.VI.MAX, Supplémentation en VItamines et Minéraux AntioXydants.
For the comparison between noncases and overall cancer cases, by Student’s t test or χ2 test, as appropriate.
In first-degree relatives.
In women only.
The associations between sex-specific quartiles of the DII and risk of overall, female breast, prostate, and nonbreast/nonprostate cancers are represented in Table 2. Higher sex-specific quartiles of the DII were positively associated with prostate cancer risk [quartile (Q) 4 compared with Q1 HR: 2.08; 95% CI: 1.06, 4.09]. The observed results suggested a possible threshold effect, confirmed in a dose–response analysis with the use of restricted cubic splines (see Supplemental Figure 2). A cutoff of −1.0 was selected. An increased risk of prostate cancer was observed for a DII value greater than or equal to −1.0 compared with a DII less than −1.0 (HR: 2.31; 95% CI: 1.35, 3.95; P = 0.002). No significant relation was found regarding overall, female breast, and nonbreast/nonprostate cancer risks.
TABLE 2.
Associations between sex-specific quartiles of the DII and risk of overall, breast, and prostate cancers from multivariable Cox proportional hazards models, SU.VI.MAX cohort, France, 1994–20071
| Cases, n/noncases, n | HR (95%CI) | P-trend | |
| All cancers (n = 559)2 | 0.2 | ||
| Q1 | 139/1495 | 1.00 | |
| Q2 | 145/1491 | 1.12 (0.88, 1.42) | |
| Q3 | 139/1497 | 1.11 (0.87, 1.43) | |
| Q4 | 136/1500 | 1.23 (0.94, 1.62) | |
| Breast cancer (n = 158)3 | 0.9 | ||
| Q1 | 45/897 | 1.00 | |
| Q2 | 34/909 | 0.78 (0.50, 1.23) | |
| Q3 | 46/897 | 1.10 (0.71, 1.70) | |
| Q4 | 33/910 | 0.85 (0.52, 1.41) | |
| Prostate cancer (n = 123)4 | 0.2 | ||
| Q1 | 23/669 | 1.00 | |
| Q2 | 44/649 | 2.76 (1.58, 4.82) | |
| Q3 | 28/665 | 1.90 (1.02, 3.54) | |
| Q4 | 28/665 | 2.08 (1.06, 4.09) | |
| Other cancers (n = 278)2 | 0.2 | ||
| Q1 | 71/1563 | 1.00 | |
| Q2 | 67/1569 | 1.02 (0.72, 1.43) | |
| Q3 | 65/1571 | 1.04 (0.72, 1.48) | |
| Q4 | 75/1561 | 1.34 (0.92, 1.95) |
Sex-specific cutoffs for quartiles of the DII were −0.98, 0.23, and 1.5 for men and −0.33, 1.0, and 2.3 for women. DII, Dietary Inflammatory Index; Q, quartile; SU.VI.MAX, Supplémentation en VItamines et Minéraux AntioXydants.
Adjusted for age (time-scale in the Cox model), sex, intervention group of the initial SU.VI.MAX trial, number of 24-h dietary records, BMI, height, physical activity, smoking status, educational level, family history of cancer in first-degree relatives, energy intake without alcohol, and alcohol intake.
Adjusted for age (time-scale in the Cox model), sex, intervention group of the initial SU.VI.MAX trial, number of 24-h dietary records, BMI, height, physical activity, smoking status, educational level, energy intake without alcohol, and alcohol intake, in addition to menopausal status at baseline, use of hormonal treatment for menopause at baseline, number of live births, and family history of breast cancer (instead of overall cancers) in first-degree relatives.
Adjusted for age (time-scale in the Cox model), sex, intervention group of the initial SU.VI.MAX trial, number of 24-h dietary records, BMI, height, physical activity, smoking status, educational level, energy intake without alcohol, and alcohol intake, in addition to baseline plasma prostate-specific antigen concentration and family history of prostate cancer (instead of overall cancers) in first-degree relatives.
The interaction between sex-specific quartiles of the DII and alcohol intake (sex-specific median, i.e., 6.6 g/d for women and 24.1 g/d for men) on overall cancer risk was statistically significant (P-interaction = 0.02) (Table 3). In stratified analyses, the DII was positively associated with overall cancer risk in low-to-moderate alcohol consumers (Q4 compared with Q1 HR: 1.75; 95% CI: 1.15, 2.68; P-trend = 0.02), but not in higher consumers of alcohol (P-trend = 0.8). This interaction was also observed for breast cancer (P-interaction = 0.001): DII was positively associated with breast cancer risk in low-to-moderate alcohol consumers (Q4 compared with Q1 HR: 3.82; 95% CI: 1.51, 9.64; P-trend = 0.002). An inverse association was detected in higher consumers of alcohol (Q4 compared with Q1 HR: 0.31; 95% CI: 0.15, 0.65; P-trend = 0.005). The P-interaction between the DII and alcohol intake was 0.8 for prostate cancer.
TABLE 3.
Associations between sex-specific quartiles of DII and risk of overall and breast cancers from multivariable Cox proportional hazards models stratified by daily alcohol intake, SU.VI.MAX cohort, France, 1994–20071
| Cases, n/noncases, n | HR (95%CI) | P-trend | P-interaction2 | |
| All cancers (n = 559) | 0.02 | |||
| Alcohol intake < median | 243/3027 | 0.02 | ||
| Q1 | 45/697 | 1.00 | ||
| Q2 | 66/679 | 1.60 (1.09, 2.36) | ||
| Q3 | 62/754 | 1.52 (1.02, 2.28) | ||
| Q4 | 70/897 | 1.75 (1.15, 2.68) | ||
| Alcohol intake ≥ median | 316/2956 | 0.8 | ||
| Q1 | 94/798 | 1.00 | ||
| Q2 | 79/812 | 0.86 (0.63, 1.17) | ||
| Q3 | 77/743 | 0.88 (0.64, 1.22) | ||
| Q4 | 66/603 | 0.96 (0.66, 1.38) | ||
| Breast cancer (n = 158) | 0.001 | |||
| Alcohol intake < median | 59/1826 | 0.002 | ||
| Q1 | 7/432 | 1.00 | ||
| Q2 | 12/411 | 1.92 (0.75, 4.94) | ||
| Q3 | 18/439 | 3.17 (1.28, 7.85) | ||
| Q4 | 22/544 | 3.82 (1.51, 9.64) | ||
| Alcohol intake ≥ median | 99/1787 | 0.005 | ||
| Q1 | 38/465 | 1.00 | ||
| Q2 | 22/498 | 0.55 (0.32, 0.95) | ||
| Q3 | 28/458 | 0.69 (0.41, 1.17) | ||
| Q4 | 11/366 | 0.31 (0.15, 0.65) |
Sex-specific cutoffs for quartiles of the DII were −0.98, 0.23, and 1.5 for men and −0.33, 1.0, and 2.3 for women. Adjusted for age (time-scale in the Cox model), sex, intervention group of the initial SU.VI.MAX trial, number of 24-h dietary records, BMI, height, physical activity, smoking status, educational level, family history of cancer in first-degree relatives, energy intake without alcohol, and alcohol intake; for breast cancer, additionally adjusted for: menopausal status and use of hormonal treatment for menopause at baseline, number of live births, and family history of breast cancer. The sex-specific population median of daily alcohol intake was 24.1 g/d for men and 6.6 g/d for women. DII, Dietary Inflammatory Index; Q, quartile; SU.VI.MAX, Supplémentation en VItamines et Minéraux AntioXydants.
Between the DII and alcohol intake.
The interaction test between sex-specific quartiles of the DII and the antioxidant supplementation group from the initial SU.VI.MAX trial was not statistically significant (P-interaction = 0.9 for overall cancer, P-interaction = 0.4 for breast cancer, and P-interaction = 0.4 for prostate cancer). DII scores ranked from −5.0 to 5.8 in the antioxidant supplementation arm (mean ± SD: 0.7 ± 1.8) and from −5.1 to 6.0 in the placebo arm (mean ± SD: 0.7 ± 1.9). No interaction was observed between the DII and sex (P-interaction = 0.3 for overall cancer) or BMI (P-interaction = 0.4 for overall cancer, P-interaction = 0.1 for breast cancer, and P-interaction = 0.6 for prostate cancer). Exclusion of in situ cancer cases (n = 47) or exclusion of cancer cases diagnosed during the first year of follow-up (n = 11) provided similar findings (data not shown). Breast cancer results were similar in pre- and postmenopausal women (data not shown).
Discussion
In this prospective study, a proinflammatory diet corresponding to higher DII scores was associated with an increased prostate cancer risk. In addition, when analyses were stratified by alcohol intake, which is a major risk factor for several cancer sites, the DII also was associated with increased overall and breast cancer risk in low-to-moderate drinkers of alcohol. These results support a role of inflammation in mechanisms underlying established nutrition–cancer relations (29).
To our knowledge, few studies have explored the relations between the DII and cancer risk. Most of them were nonprospective case-control studies: 1 showed positive associations of the DII with pancreatic cancer (15); 2 with prostate cancer, in line with our results (12, 17); 2 with colorectal cancer (20, 21); and 2 with esophageal cancer (11, 22). A recent case-control study in postmenopausal women observed no association between the DII and breast cancer (13). Only 4 prospective studies investigated the relations between the DII and cancer risk: 1 found a direct association with lung cancer risk (14), and 3 found a direct association with colorectal cancer risk (16, 18, 20). All of these studies used an FFQ (11–19, 21, 22) or a dietary history questionnaire (20), but none used repeated 24-h records or recalls. To our knowledge, no prospective study has previously investigated the relation between the DII and overall, female breast, and prostate cancer risks.
Our analysis resulted in a positive association between the proinflammatory potential of the individual diets assessed with higher DII scores and increased prostate and, in low-to-moderate alcohol drinkers, overall and breast cancer risks. These associations were statistically significant from the Q2 of the DII compared with Q1, suggesting a threshold effect. These results are consistent with epidemiologic studies that demonstrated 1) correlations between the DII and inflammatory biomarkers, and 2) associations between these inflammatory biomarkers and cancer risk. Indeed, the DII has been significantly and positively associated with several inflammatory biomarkers, such as IL-6 and CRP (5, 7, 30, 31). Moreover, in a meta-analysis published in 2013, Guo et al. (3) concluded that elevated concentrations of CRP were associated with an increased risk of all cancers, lung cancer, and possibly breast, prostate, and colorectal cancer. Since then, several prospective studies corroborated these findings. We previously showed that baseline high-sensitivity CRP concentrations were positively associated with subsequent overall and prostate cancer risk in the SU.VI.MAX cohort (4). In a nested case-control study from the European Prospective Investigation into Cancer and Nutrition cohort, colorectal cancer cases were characterized by a “reactive oxygen metabolites and CRP” pattern (32), and several inflammatory biomarkers were associated with elevated endometrial cancer risk in postmenopausal women (33). Zuo et al. (34) found that plasma neopterin, kynurenine-to-tryptophan ratio, and CRP were interrelated and positively associated with overall cancer risk. Shiels et al. (35) observed a positive association between inflammatory markers and lung cancer risk. However, this point remains debated because recent studies also reported no association between CRP and/or IL-6 and prostate (36) and epithelial ovarian cancer risk (37).
Experimental evidence strongly supports the involvement of inflammation in carcinogenesis (1, 38–40). Reactive oxygen and nitrogen intermediates, prostaglandins, and inflammatory cytokines were found to be associated with chronic inflammation (39). These molecules are involved in neoplastic tumor processes through immune responses (1). Genetic polymorphisms in inflammation-related genes [IL1B, IL10, tumor necrosis factor (TNF), IL6, IL8, and cyclooxygenase 2 (COX2)] and inflammatory molecules expressed may be implicated in prostate cancer initiation and progression (38).
Based on the results of numerous experimental and observational studies (9), alcohol intake contributes to the DII computation as an anti-inflammatory marker. Nevertheless, alcohol is acknowledged to be a major risk factor in many cancers and to be involved in several procarcinogenic mechanisms (DNA damage, generation of reactive oxygen species, increase in circulating estrogens, etc.) (29, 41). For these reasons, we tested the interaction between the DII and alcohol intake, which was statistically significant for overall and breast cancer risk. Thus, we stratified our analysis by the sex-specific median of daily alcohol intake. A higher DII was associated with increased overall and breast cancer risk in low-to-moderate drinkers only. Several hypotheses can be formulated to explain these results. First, the pro- or anti-inflammatory potential of the diet may influence cancer risk in low-to-moderate drinkers while being exceeded by the deleterious effect of alcohol in heavier drinkers. Second, because of the DII design, higher alcohol intake makes DII scores lower and may scramble the results because alcohol simultaneously increases cancer risk. This also may explain the inverse association between the DII and breast cancer risk observed in higher alcohol drinkers, which is unlikely to be causal. Indeed, although analyses were adjusted for alcohol intake, it may be difficult to disentangle the effect of alcohol. In addition, in the SU.VI.MAX cohort, we previously showed an interaction between several classes of polyphenols and alcohol intake on breast cancer risk (42). Although an inverse association was observed between polyphenols and breast cancer risk in women with an alcohol intake below the median, a direct association was observed with higher consumption of alcohol. Because several classes of polyphenols are included in the DII calculation as anti-inflammatory factors, this adds further complexity to the association between the DII, alcohol, and breast cancer risk.
Strengths of our study included its prospective design with long follow-up and the implementation of a literature-derived index to assess the overall inflammatory potential of the diet. In addition, usual dietary intake was assessed with a high degree of accuracy by using repeated 24-h dietary records (with a mean of 10 records/subject) that accounted for intraindividual day-to-day and seasonal variability. Nevertheless, some limitations should be acknowledged. First, the limited number of cases may have impaired our ability to detect some of the hypothesized associations. Moreover, specific cancer sites other than the 2 primarily represented in the cohort (i.e., breast and prostate) could not be investigated separately. Second, subjects were volunteers involved in a long-term nutrition study and thus were more highly educated, less likely to smoke, and more health-conscious than the general French population. Despite this healthier profile (low baseline risk), the number of incident cancers was larger than in the general population. The study protocol included a yearly check-up that may have encouraged early diagnosis (in particular for breast cancer in women, for whom screening mammograms were performed during follow-up). Therefore, caution is needed when extrapolating these results to the entire French population. Third, all self-reported cases were validated by an expert committee. However, it is possible that some individuals with cancer did not report their cancer events. Next, although most of the 45 DII variables were taken into account (n = 36), some items (n = 9) were missing in our database and thus could not be included, such as caffeine, eugenol, saffron, selenium, turmeric, zinc, thyme or oregano, and rosemary (anti-inflammatory factors), and trans FAs (proinflammatory factor). However, it is unlikely that these missing variables drastically influenced the results, because it has been shown that the ability of the DII to predict inflammation remained the same when the number of food variables was decreased from 45 to 28 (5). Finally, residual confounders could not be ruled out. However, a large range of confounding factors was taken into account, thus limiting potential bias.
To our knowledge, this study was the first to investigate the prospective association between the overall pro- or anti-inflammatory potential of the diet, assessed with the DII, and overall, female breast, and prostate cancer risk. Consistent with mechanistic data, our results suggest that a proinflammatory diet may be associated with increased prostate cancer risk. When models were stratified by alcohol intake, a major cancer risk factor, a proinflammatory diet also was associated with increased overall and breast cancer risk in low-to-moderate drinkers. These results provide interesting insights for the understanding of the relations between diet and cancer risk, and inflammation-based underlying mechanisms. They suggest that promoting an anti-inflammatory dietary pattern may contribute to cancer prevention.
Acknowledgments
We thank Nathalie Arnault, statistician, for her technical contribution to the Supplémentation en VItamines et Minéraux AntioXydants study. LG and MT designed the research; SH, PG, EK-G, and MT conducted the research; NS, JRH, and MDW designed and computed the Dietary Inflammatory Index score; LG performed the statistical analysis and wrote the paper; MD, FM, LN, NS, JRH, MDW, PL-M, SH, PG, CJ, EK-G, and MT contributed to the data interpretation and revised each draft for important intellectual content; and MT supervised the study and had primary responsibility for final content. All authors read and approved the final manuscript.
Footnotes
Abbreviations used: COX2, cyclooxygenase 2; CRP, C-reactive protein; DII, Dietary Inflammatory Index; Q, quartile; SU.VI.MAX, Supplémentation en VItamines et Minéraux AntioXydants; TNF, tumor necrosis factor.
References
- 1.Coussens LM, Werb Z. Inflammation and cancer. Nature 2002;420:860–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Heikkilä K, Harris R, Lowe G, Rumley A, Yarnell J, Gallacher J, Ben-Shlomo Y, Ebrahim S, Lawlor DA. Associations of circulating C-reactive protein and interleukin-6 with cancer risk: findings from two prospective cohorts and a meta-analysis. Cancer Causes Control 2009;20:15–26. [DOI] [PubMed] [Google Scholar]
- 3.Guo YZ, Pan L, Du CJ, Ren DQ, Xie XM. Association between C-reactive protein and risk of cancer: a meta-analysis of prospective cohort studies. Asian Pac J Cancer Prev 2013;14:243–8. [DOI] [PubMed] [Google Scholar]
- 4.Touvier M, Fezeu L, Ahluwalia N, Julia C, Charnaux N, Sutton A, Mejean C, Latino-Martel P, Hercberg S, Galan P, et al. Association between prediagnostic biomarkers of inflammation and endothelial function and cancer risk: a nested case-control study. Am J Epidemiol 2013;177:3–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shivappa N, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS, Tabung F, 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]
- 6.Nanri A, Yoshida D, Yamaji T, Mizoue T, Takayanagi R, Kono S. Dietary patterns and C-reactive protein in Japanese men and women. Am J Clin Nutr 2008;87:1488–96. [DOI] [PubMed] [Google Scholar]
- 7.Shivappa N, Hebert JR, Rietzschel ER, De Buyzere ML, Langlois M, Debruyne E, Marcos A, Huybrechts I. Associations between dietary inflammatory index and inflammatory markers in the Asklepios Study. Br J Nutr 2015;113:665–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ahluwalia N, Andreeva VA, Kesse-Guyot E, Hercberg S. Dietary patterns, inflammation and the metabolic syndrome. Diabetes Metab 2013;39:99–110. [DOI] [PubMed] [Google Scholar]
- 9.Shivappa N, Steck SE, Hurley TG, Hussey JR, Hebert 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]
- 10.WHO Media centre. Cancer fact sheet No. 297. 2015. [Google Scholar]
- 11.Shivappa N, Zucchetto A, Serraino D, Rossi M, La VC, Hebert JR. Dietary inflammatory index and risk of esophageal squamous cell cancer in a case-control study from Italy. Cancer Causes Control 2015;26:1439–47. [DOI] [PubMed] [Google Scholar]
- 12.Shivappa N, Jackson MD, Bennett F, Hebert JR. Increased Dietary Inflammatory Index (DII) is associated with increased risk of prostate cancer in Jamaican men. Nutr Cancer 2015;67:941–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ge I, Rudolph A, Shivappa N, Flesch-Janys D, Hebert JR, Chang-Claude J. Dietary inflammation potential and postmenopausal breast cancer risk in a German case-control study. Breast 2015;24:491–6. [DOI] [PubMed] [Google Scholar]
- 14.Maisonneuve P, Shivappa N, Hebert JR, Bellomi M, Rampinelli C, Bertolotti R, Spaggiari L, Palli D, Veronesi G, Gnagnarella P. Dietary inflammatory index and risk of lung cancer and other respiratory conditions among heavy smokers in the COSMOS screening study. Eur J Nutr. In press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Shivappa N, Bosetti C, Zucchetto A, Serraino D, La Vecchia C, Hebert JR. Dietary inflammatory index and risk of pancreatic cancer in an Italian case-control study. Br J Nutr 2014;113:292–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shivappa N, Prizment AE, Blair CK, Jacobs DR Jr, Steck SE, Hebert JR. Dietary inflammatory index and risk of colorectal cancer in the Iowa Women’s Health Study. Cancer Epidemiol Biomarkers Prev 2014;23:2383–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Shivappa N, Bosetti C, Zucchetto A, Montella M, Serraino D, La VC, Hebert JR. Association between dietary inflammatory index and prostate cancer among Italian men. Br J Nutr 2014;113:278–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tabung FK, Steck SE, Ma Y, Liese AD, Zhang J, Caan B, Hou L, Johnson KC, Mossavar-Rahmani Y, Shivappa N, et al. The association between dietary inflammatory index and risk of colorectal cancer among postmenopausal women: results from the Women’s Health Initiative. Cancer Causes Control 2015;26:399–408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wirth MD, Shivappa N, Steck SE, Hurley TG, Hebert JR. The dietary inflammatory index is associated with colorectal cancer in the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Br J Nutr 2015;113:1819–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zamora-Ros R, Shivappa N, Steck SE, Canzian F, Landi S, Alonso MH, Hebert JR, Moreno V. Dietary inflammatory index and inflammatory gene interactions in relation to colorectal cancer risk in the Bellvitge colorectal cancer case-control study. Genes Nutr 2015;10:447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Shivappa N, Zucchetto A, Montella M, Serraino D, Steck SE, La VC, Hebert JR. Inflammatory potential of diet and risk of colorectal cancer: a case-control study from Italy. Br J Nutr 2015;114:152–8. [DOI] [PubMed] [Google Scholar]
- 22.Lu Y, Shivappa N, Lin Y, Lagergren J, Hebert JR. Diet-related inflammation and oesophageal cancer by histological type: a nationwide case-control study in Sweden. Eur J Nutr. In press. [DOI] [PubMed] [Google Scholar]
- 23.Shivappa N, Sandin S, Lof M, Hebert JR, Adami HO, Weiderpass E. Prospective study of dietary inflammatory index and risk of breast cancer in Swedish women. Br J Cancer 2015;113:1099–103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briancon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med 2004;164:2335–42. [DOI] [PubMed] [Google Scholar]
- 25.Le Moullec N, Deheeger M, Preziosi P, Montero P, Valeix P, Rolland-Cachera M, Potier de Courcy G, Christides J, Galan P, Hercberg S. Validation du manuel photo utilisé pour l’enquête alimentaire de l’étude SU.VI.MAX. [Validation of the food portion size booklet used in the SU.VI.MAX study.] Cah Nutr Diet 1996;31:158–64 (in French). [Google Scholar]
- 26.Hercberg S. Table de composition SU.VI.MAX des aliments. [SU.VI.MAX food composition table.] Paris: Les éditions INSERM/Economica; 2005 (in French). [Google Scholar]
- 27.WHO. ICD-10, International classification of diseases and related health problems. 10th revision. Geneva (Switzerland): World Health Organization; 2010. [Google Scholar]
- 28.Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med 2010;29:1037–57. [DOI] [PubMed] [Google Scholar]
- 29.WCRF/AICR. Food, nutrition, physical activity and the prevention of cancer: a global perspective. Washington (DC): AICR; 2007. [Google Scholar]
- 30.Wirth MD, Burch J, Shivappa N, Violanti JM, Burchfiel CM, Fekedulegn D, Andrew ME, Hartley TA, Miller DB, Mnatsakanova A, et al. Association of a dietary inflammatory index with inflammatory indices and metabolic syndrome among police officers. J Occup Environ Med 2014;56:986–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cavicchia PP, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS, Hebert 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]
- 32.Aleksandrova K, Jenab M, Bueno-de-Mesquita HB, Fedirko V, Kaaks R, Lukanova A, van Duijnhoven FJ, Jansen E, Rinaldi S, Romieu I, et al. Biomarker patterns of inflammatory and metabolic pathways are associated with risk of colorectal cancer: results from the European Prospective Investigation into Cancer and Nutrition (EPIC). Eur J Epidemiol 2014;29:261–75. [DOI] [PubMed] [Google Scholar]
- 33.Dossus L, Lukanova A, Rinaldi S, Allen N, Cust AE, Becker S, Tjonneland A, Hansen L, Overvad K, Chabbert-Buffet N, et al. Hormonal, metabolic, and inflammatory profiles and endometrial cancer risk within the EPIC cohort–a factor analysis. Am J Epidemiol 2013;177:787–99. [DOI] [PubMed] [Google Scholar]
- 34.Zuo H, Tell GS, Vollset SE, Ueland PM, Nygard O, Midttun O, Meyer K, Ulvik A, Eussen SJ. Interferon-gamma-induced inflammatory markers and the risk of cancer: the Hordaland Health Study. Cancer 2014;120:3370–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Shiels MS, Pfeiffer RM, Hildesheim A, Engels EA, Kemp TJ, Park JH, Katki HA, Koshiol J, Shelton G, Caporaso NE, et al. Circulating inflammation markers and prospective risk for lung cancer. J Natl Cancer Inst 2013;105:1871–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Toriola AT, Laukkanen JA, Kurl S, Nyyssonen K, Ronkainen K, Kauhanen J. Prediagnostic circulating markers of inflammation and risk of prostate cancer. Int J Cancer 2013;133:2961–7. [DOI] [PubMed] [Google Scholar]
- 37.Ose J, Schock H, Tjonneland A, Hansen L, Overvad K, Dossus L, Clavel-Chapelon F, Baglietto L, Boeing H, Trichopolou A, et al. Inflammatory markers and risk of epithelial ovarian cancer by tumor subtypes: the EPIC cohort. Cancer Epidemiol Biomarkers Prev 2015;24:951–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Sfanos KS, Hempel HA, De Marzo AM. The role of inflammation in prostate cancer. Adv Exp Med Biol 2014;816:153–81. [DOI] [PubMed] [Google Scholar]
- 39.Shacter E, Weitzman SA. Chronic inflammation and cancer. Oncology (Williston Park). 2002;16:217–26, 229. [PubMed] [Google Scholar]
- 40.Nakai Y, Nonomura N. Inflammation and prostate carcinogenesis. Int J Urol 2013;20:150–60. [DOI] [PubMed] [Google Scholar]
- 41.Chhim AS, Fassier P, Latino-Martel P, Druesne-Pecollo N, Zelek L, Duverger L, Hercberg S, Galan P, Deschasaux M, Touvier M. Prospective association between alcohol intake and hormone-dependent cancer risk: modulation by dietary fiber intake. Am J Clin Nutr 2015;102:182–9. [DOI] [PubMed] [Google Scholar]
- 42.Touvier M, Druesne-Pecollo N, Kesse-Guyot E, Andreeva VA, Fezeu L, Galan P, Hercberg S, Latino-Martel P. Dual association between polyphenol intake and breast cancer risk according to alcohol consumption level: a prospective cohort study. Breast Cancer Res Treat 2013;137:225–36. [DOI] [PubMed] [Google Scholar]