Regarding the statistical analysis, first, we calculated basic descriptive statistics of the anthropometric, sociodemographic, and lifestyle characteristics for all women and by categories of mammographic density. Normally distributed continuous variables were described using the mean±standard deviation. Differences across categories of mammographic density were tested with analysis of variance tests. Nonnormally distributed continuous variables were described using the median (interquartile interval) and differences by mammographic density were tested with nonparametric Kruskal-Wallis tests. Categorical variables were described using the number of cases and corresponding percentages, and differences by mammographic density were tested with χ2 tests.
Associations between adherence to either dietary pattern and mammographic density were evaluated using ordinal logistic regression models with random center-specific intercepts including center as a random effect. As fixed-effects terms, age, body mass index (BMI, calculated as weight (kg)/[height (m)]2), parity, menopausal and smoking status, family history of breast cancer, use of hormonal therapy, and calorie and alcohol intake were considered as potential confounders. Three mixed models were adjusted to explore the confounding effect of different sets of variables. Model 1 was adjusted for only age and BMI (in addition to the random-effects term); model 2 also included parity, menopausal and smoking status, family history of breast cancer, and use of hormonal therapy. Finally, calorie and alcohol intake were added to model 3. Both categorical (grouping the scores of adherence into quartiles) and continuous (one-standard deviation increase) associations with the scores were examined with all three models. For model 3, nonlinear associations between the adherence to each pattern and mammographic density were assessed by fitting fractional polynomials.
With regard to the sample size, 22.8% of women had a mammographic density of greater than 50%. Therefore, our data allowed us to detect differences of 8% or more in the percentage of women classified in this category between extreme quartiles of adherence to each dietary pattern with a power of 80%.
Finally, when significant associations were found, separate analyses were performed by categories of all potential confounders mentioned and represented in forest plots. Heterogeneity of effects was tested in model 3 by including an interaction term between the score of adherence and the corresponding variable.
Analyses were performed using STATA/MP 14.0, and statistical significance was set at two-sided P<.05.
The protocol study “Determinants of Mammographic Density in Spain” was formally approved by the bioethics and animal welfare committee at the Carlos III Institute of Health and all participants signed informed consent, including permission to publish the results from the research.
Thirty-six participants were excluded from analyses: 10 women who developed breast cancer within 6 months of study entry and mammography, 16 who did not have mammographic density assessment, two who did not have BMI information, and eight who reported a daily kilocalorie intake under 750 or above 4,500. Therefore, analyses included data from 3,548 women for whom we had complete information regarding all the variables of interest. As expected, premenopausal and perimenopausal women showed a higher percentage of dense tissue (higher mammographic density). An elevated mammographic density was also associated with a family history of breast cancer, tobacco use, high calorie and alcohol intake, younger age, and lower BMI and parity (Table 2).
Multivariable analyses supported these findings confirming that, although breast density did not differ by level of adherence to the Mediterranean dietary pattern (adjusted odds ratio [OR]Q4vsQ1 [95% CI] 0.99 [0.81–1.21] and adjusted OR1-standard deviation increase [95% CI] 1.02 [0.95–1.09]), those with a high adherence to the Western dietary pattern had higher mammographic density (adjusted ORQ4vsQ1 [95% CI] 1.25 [1.03–1.52] and adjusted OR1-standard deviation increase [95% CI] 1.09 [1.02–1.18]) (Table 3). No statistically significant departure from linearity was observed in this association when the analysis with fractional polynomials was performed (data not shown).
Stratified analysis by subgroups revealed that the effect of the Western dietary pattern on mammographic density was confined to women with BMIs higher than 25 (adjusted ORQ4vsQ1 [95% CI] 1.41 [1.13–1.76], heterogeneity P value=.068). Our results also suggested some differences according to parity, calorie intake, and tobacco consumption, but none of the interaction terms reached statistical significance (Fig. 1).
Our results suggest that, whereas the Mediterranean diet was not related to mammographic density, a higher adherence to the Western dietary pattern was associated with higher mammographic density. Subgroup analyses suggest that this effect may be confined to overweight and obese women and to be stronger among parous, nonsmokers, and women with elevated calorie intake. However, our tests for heterogeneity approached significance at best, probably for lack of power. Thus, larger studies are needed to confirm these potential differential effects of diet on mammographic density.
Taking into account that high mammographic density is considered one of the key risk factors for breast cancer,3 we expected to identify associations between dietary patterns and mammographic density similar in direction to those found for dietary patterns and breast cancer by Castelló et al.6 However, although we found a positive association for the Western dietary pattern, mammographic density was not influenced by adherence to the Mediterranean dietary pattern.
Our findings support previous studies exploring the association between mammographic density and specific nutrients or foods included in the Western dietary pattern that reported positive associations with total energy16; high-density foods17; total, saturated, and cholesterol fats18,19; proteins18; and meat.20 Not surprisingly, a Western-type diet contrasts with the recommendations issued by the World Cancer Research Fund and the American Association for Cancer Research to reduce cancer burden. Adherence to these recommendations has been positively associated with a reduction of breast cancer risk21 and mammographic density17 in our context.
On the other hand, a weak inverse association with the Mediterranean dietary pattern8 or with some of its main components such as olive oil,16 vegetables, and fiber22 has been previously reported. Others have found an absence or even a positive association of some of these items with mammographic density.19,23,24 These inconclusive findings suggest that a reduction in mammographic density may not be one of the key mechanisms through which the Mediterranean diet lowers breast cancer risk. A possible explanation for the contradictory effect of a Mediterranean diet on mammographic density and breast cancer is that this diet could be influencing the fat deposit of the breast without altering the percentage of dense tissue. Obesity, a condition inversely associated with mammographic density, increases breast cancer risk through several mechanisms, including the inflammatory effect of adipokines,25 whereas the Mediterranean diet seems to counteract an inflammatory state.26
It is worth mentioning that the effect of the Western dietary pattern on mammographic density was observed only among overweight and obese women. Adipocytes are potent endocrine cells that produce hormones and growth factors; obesity strongly influences this endocrine millieu.27 Our results may reflect a synergic effect of this dietary pattern and the local adipose tissue on the fibroglandular component of the breast.
For its kind, this is a fairly large and carefully conducted study on risk factors and mammographic density; however, it presents some limitations. First, the sample size was insufficient to detect significant interactions even when some differences by subgroups are observed. Second, the representativeness of the selected sample might be slightly biased because healthy screening participants might be more concerned about their health than nonparticipants. However, participation rates in Spanish breast cancer screening programs are high28 and women in our study are very similar to the women in the Spanish National Health Survey in terms of age range, socioeconomic level, prevalence of smoking, and physical activity.29 Third, the visual assessment of breast density by a single radiologist may imply a degree of subjectivity. However, the radiologist had very high intraobserver concordance,30 and we have confirmed that the visual scale used here is a predictor of subsequent breast cancer development risk.3 Additionally, the collection of data with different mammographic devices and interviewers in different centers might introduce some heterogeneity. These unmeasured sources of variability were taken into account by including random center-specific intercepts in our regression models. Finally, it should be noted that the cross-sectional design of the current study precludes the establishment of causal relationships between adherence to dietary patterns and mammographic density. However, it is hard to think that this association is acting in the other direction because information on diet was collected before the mammographic exploration.
1. WHO. Global health risks: mortality and burden of disease attributable to selected major risks. Geneva (Switzerland): World Health Organization; 2009.
2. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, Rosso S, Coebergh JW, Comber H, et al. Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer 2013;49:1374–403.
3. Pollán M, Ascunce N, Ederra M, Murillo A, Erdozain N, Alés-Martinez J, et al. Mammographic density and risk of breast cancer according to tumor characteristics and mode of detection: a Spanish population-based case-control study. Breast Cancer Res 2013;15:R9.
4. Li J, Humphreys K, Eriksson L, Edgren G, Czene K, Hall P. Mammographic density reduction is a prognostic marker of response to adjuvant tamoxifen therapy in postmenopausal patients with breast cancer. J Clin Oncol 2013;31:2249–56.
5. Tice JA, Cummings SR, Smith-Bindman R, Ichikawa L, Barlow WE, Kerlikowske K. Using clinical factors and mammographic breast density to estimate breast cancer risk: development and validation of a new predictive model. Ann Intern Med 2008;148:337–47.
6. Castelló A, Pollán M, Buijsse B, Ruiz A, Casas AM, Baena-Cañada JM, et al. Spanish Mediterranean diet and other dietary patterns and breast cancer risk: case-control EpiGEICAM study. Br J Cancer 2014;111:1454–62.
7. Takata Y, Maskarinec G, Park SY, Murphy SP, Wilkens LR, Kolonel LN. Mammographic density and dietary patterns: the multiethnic cohort. Eur J Cancer Prev 2007;16:409–14.
8. Voevodina O, Billich C, Arand B, Nagel G. Association of Mediterranean diet, dietary supplements and alcohol consumption with breast density among women in South Germany: a cross-sectional study. BMC Public Health 2013;13:203.
9. Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 1985;122:51–65.
10. Vioque J, Navarrete-Muñoz EM, Gimenez-Monzó D, García-de-la-Hera M, Granado F, Young IS, et al. Reproducibility and validity of a food frequency questionnaire among pregnant women in a Mediterranean area. Nutr J 2013;12:26.
11. Vioque J, Weinbrenner T, Asensio L, Castelló A, Young IS, Fletcher A. Plasma concentrations of carotenoids and vitamin C are better correlated with dietary intake in normal weight than overweight and obese elderly subjects. Br J Nutr 2007;97:977–86.
12. Boyd NF, Byng JW, Jong RA, Fishell EK, Little LE, Miller AB, et al. Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study. J Natl Cancer Inst 1995;87:670–5.
13. Ramón Y Calal T, Chirivella I, Miranda J, Teule A, Izquierdo, Balmaña J, et al. Mammographic density and breast cancer in women from high risk families. Breast Cancer Res 2015;17:93.
14. Burt C. Factor analysis and canonical correlations. Br J Math Stat Psychol 1948;1:95–106.
15. Castello A, Buijsse B, Martin M, Ruiz A, Casas AM, Baena-Canada JM, et al. Evaluating the applicability of data-driven dietary patterns to independent samples with focus on measurement tools for pattern similarity. J Acad Nutr Diet 2016 Jun 28 [Epub ahead of print].
16. García-Arenzana N, Navarrete-Muñoz EM, Lope V, Moreo P, Vidal C, Laso-Pablos S, et al. Calorie intake, olive oil consumption and mammographic density among Spanish women. Int J Cancer 2014;134:1916–25.
17. Castelló A, Prieto L, Ederra M, Salas-Trejo D, Vidal C, Sánchez-Contador C, et al. Association between the adherence to the international guidelines for cancer prevention and mammographic density. PLoS One 2015;10:e0132684.
18. Nagata C, Matsubara T, Fujita H, Nagao Y, Shibuya C, Kashiki Y, et al. Associations of mammographic density with dietary factors in Japanese women. Cancer Epidemiol Biomarkers Prev 2005;14:2877–80.
19. Qureshi SA, Couto E, Hilsen M, Hofvind S, Wu AH, Ursin G. Mammographic density and intake of selected nutrients and vitamins in Norwegian women. Nutr Cancer 2011;63:1011–20.
20. Sala E, Warren R, Duffy S, Welch A, Luben R, Day N. High risk mammographic parenchymal patterns and diet: a case-control study. Br J Cancer 2000;83:121–6.
21. Castelló A, Martin M, Ruiz A, Casas AM, Baena-Cañada JM, Lope V, et al. Lower breast cancer risk among women following the World Cancer Research Fund and American Institute for Cancer Research lifestyle recommendations: EpiGEICAM case-control study. PLoS One 2015;10:e0126096.
22. Nagel G, Mack U, von Fournier D, Linseisen J. Dietary phytoestrogen intake and mammographic density—results of a pilot study. Eur J Med Res 2005;10:389–94.
23. Mishra GD, dos Santos Silva I, McNaughton SA, Stephen A, Kuh D. Energy intake and dietary patterns in childhood and throughout adulthood and mammographic density: results from a British prospective cohort. Cancer Causes Control 2011;22:227–35.
24. Thomson CA, Arendell LA, Bruhn RL, Maskarinec G, Lopez AM, Wright NC, et al. Pilot study of dietary influences on mammographic density in pre- and postmenopausal Hispanic and non-Hispanic white women. Menopause 2007;14:243–50.
25. Renehan AG, Zwahlen M, Egger M. Adiposity and cancer risk: new mechanistic insights from epidemiology. Nat Rev Cancer 2015;15:484–98.
26. Medina-Remón A, Casas R, Tressserra-Rimbau A, Ros E, Martinez-González MA, Fito M, et al. Polyphenol intake from a Mediterranean diet decreases inflammatory biomarkers related to atherosclerosis: a sub-study of the PREDIMED trial. Br J Clin Pharmacol 2016 Apr 21 [Epub ahead of print].
27. McCready J, Arendt LM, Rudnick JA, Kuperwasser C. The contribution of dynamic stromal remodeling during mammary development to breast carcinogenesis. Breast Cancer Res 2010;12:205.
28. Castells X, Sala M, Ascunce N, Salas D, Zubizarreta R, Casamitjana M. Descripción del cribado del cáncer en España: Proyecto DESCRIC. Madrid (Spain): Plan de Calidad para el Sistema Nacional de Salud, Ministerio de Sanidad y Consumo, Agència d'Avaluació de Tecnologia i Recerca Mèdiques de Cataluña; 2007. Informes de Evaluación de Tecnologías Sanitarias, AATRM núm: 2006/01; 2007.
29. García-Arenzana N, Navarrete-Muñoz EM, Peris M, Salas D, Ascunce N, Gonzalez I, et al. Diet quality and related factors among Spanish female participants in breast cancer screening programs. Menopause 2012;19:1121–9.
30. Garrido-Estepa M, Ruiz-Perales F, Miranda J, Ascunce N, González-Román I, Sánchez-Contador C, et al. Evaluation of mammographic density patterns: reproducibility and concordance among scales. BMC Cancer 2010;10:485.