The Journal of Nutrition. First published ahead of print January 27, 2016 as doi: 10.3945/jn.115.221481. The Journal of Nutrition Nutritional Epidemiology
Weight Status and Alcohol Intake Modify the Association between Vitamin D and Breast Cancer Risk1–3 M´elanie Deschasaux,4* Jean-Claude Souberbielle,5 Paule Latino-Martel,4 Angela Sutton,6 Nathalie Charnaux,6 Nathalie Druesne-Pecollo,4 Pilar Galan,4 Serge Hercberg,4,7 Sigrid Le Clerc,8 Emmanuelle Kesse-Guyot,4 Khaled Ezzedine,4,9 and Mathilde Touvier4 Downloaded from jn.nutrition.org at UNIVERSITY OF PITTSBURGH FALK LIBRARY on February 7, 2016
4 Sorbonne Paris Cit´e Epidemiology and Statistics Research Center (CRESS), French National Institute of Health and Medical Research (Inserm U1153), French National Institute for Agricultural Research (INRA U1125), French National Conservatory of Arts and Crafts (CNAM), Paris 13 University, Nutritional Epidemiology Research Team (EREN), Bobigny, France; 5Physiology Department, Necker Hospital, Inserm U845, Paris, France; 6Biochemistry Department, Jean Verdier Hospital, Inserm U698, Paris 13 University, Bondy, France; 7Public Health Department, Avicenne Hospital, Bobigny, France; 8Conservatoire National des Arts et M´etiers (CNAM), Genomics, Bioinformatics and Applications Team (EA4627), Paris, France; and 9Dermatology Department, Saint Andr´e Hospital, Bordeaux, France
Abstract Background: Mechanistic hypotheses suggest that vitamin D may contribute to the prevention of breast cancer. However, epidemiologic evidence is inconsistent, suggesting a potential effect modification by individual factors. Objective: Our objective was to perform exploratory analyses on the prospective associations between the plasma 25-hydroxyvitamin D [25(OH)D] concentration, polymorphisms of genes encoding for the vitamin D receptor (VDR) and vitamin D-binding protein (also known as gc-globulin or group-specific component, GC), and breast cancer risk, along with 2 potential modifiers: body mass index (BMI; in kg/m2) and alcohol intake. ´ ´ Methods: A nested case-control study was set up in the SUpplementation en VItamines et Mineraux Anti-oXydants (SU. VI.MAX) cohort (1994–2007), involving 233 women with breast cancer and 466 matched controls (mean 6 SD age: 49 6 6 y). The plasma total 25(OH)D concentration and gene polymorphisms were assessed on samples obtained at baseline. Conditional logistic regression models were computed. Results: A higher plasma 25(OH)D concentration was associated with a decreased risk of breast cancer for women with a BMI < the median of 22.4 [OR quartile (Q)4 compared with Q1: 0.46; 95% CI: 0.23, 0.89; P-trend = 0.01, P-interaction = 0.002], whereas it was associated with an increased risk for women with a BMI $ the median (OR Q4 compared with Q1: 2.45; 95% CI: 1.13, 5.28; P-trend = 0.02, P-interaction = 0.002). A plasma 25(OH)D concentration $ 10 ng/mL was associated with a decreased risk of breast cancer for women with alcohol intakes $ the median of 7.1 g/d (OR $10 compared with 0.5) or calcium intake (P-interaction > 0.2), associations were modified by BMI (P-interaction = 0.002 for all; 0.02 for all after multipletesting adjustment) and alcohol intake ($10 compared with
Weight Status and Alcohol Intake Modify the Association between Vitamin D and Breast Cancer Risk1–3 M´elanie Deschasaux,4* Jean-Claude Souberbielle,5 Paule Latino-Martel,4 Angela Sutton,6 Nathalie Charnaux,6 Nathalie Druesne-Pecollo,4 Pilar Galan,4 Serge Hercberg,4,7 Sigrid Le Clerc,8 Emmanuelle Kesse-Guyot,4 Khaled Ezzedine,4,9 and Mathilde Touvier4 Downloaded from jn.nutrition.org at UNIVERSITY OF PITTSBURGH FALK LIBRARY on February 7, 2016
4 Sorbonne Paris Cit´e Epidemiology and Statistics Research Center (CRESS), French National Institute of Health and Medical Research (Inserm U1153), French National Institute for Agricultural Research (INRA U1125), French National Conservatory of Arts and Crafts (CNAM), Paris 13 University, Nutritional Epidemiology Research Team (EREN), Bobigny, France; 5Physiology Department, Necker Hospital, Inserm U845, Paris, France; 6Biochemistry Department, Jean Verdier Hospital, Inserm U698, Paris 13 University, Bondy, France; 7Public Health Department, Avicenne Hospital, Bobigny, France; 8Conservatoire National des Arts et M´etiers (CNAM), Genomics, Bioinformatics and Applications Team (EA4627), Paris, France; and 9Dermatology Department, Saint Andr´e Hospital, Bordeaux, France
Abstract Background: Mechanistic hypotheses suggest that vitamin D may contribute to the prevention of breast cancer. However, epidemiologic evidence is inconsistent, suggesting a potential effect modification by individual factors. Objective: Our objective was to perform exploratory analyses on the prospective associations between the plasma 25-hydroxyvitamin D [25(OH)D] concentration, polymorphisms of genes encoding for the vitamin D receptor (VDR) and vitamin D-binding protein (also known as gc-globulin or group-specific component, GC), and breast cancer risk, along with 2 potential modifiers: body mass index (BMI; in kg/m2) and alcohol intake. ´ ´ Methods: A nested case-control study was set up in the SUpplementation en VItamines et Mineraux Anti-oXydants (SU. VI.MAX) cohort (1994–2007), involving 233 women with breast cancer and 466 matched controls (mean 6 SD age: 49 6 6 y). The plasma total 25(OH)D concentration and gene polymorphisms were assessed on samples obtained at baseline. Conditional logistic regression models were computed. Results: A higher plasma 25(OH)D concentration was associated with a decreased risk of breast cancer for women with a BMI < the median of 22.4 [OR quartile (Q)4 compared with Q1: 0.46; 95% CI: 0.23, 0.89; P-trend = 0.01, P-interaction = 0.002], whereas it was associated with an increased risk for women with a BMI $ the median (OR Q4 compared with Q1: 2.45; 95% CI: 1.13, 5.28; P-trend = 0.02, P-interaction = 0.002). A plasma 25(OH)D concentration $ 10 ng/mL was associated with a decreased risk of breast cancer for women with alcohol intakes $ the median of 7.1 g/d (OR $10 compared with 0.5) or calcium intake (P-interaction > 0.2), associations were modified by BMI (P-interaction = 0.002 for all; 0.02 for all after multipletesting adjustment) and alcohol intake ($10 compared with
