Abstract A26: Dietary intake and C-peptide levels in adolescent girls: Results from the Dietary Intervention Study in Children (DISC)

Abstract

Abstract Background: Adolescence is the time period of greatest breast development. Diet during this time has been hypothesized to be associated with later breast cancer risk through its effect on insulin resistance. C-peptide is a good indicator of insulin production and higher levels in adulthood have been associated with insulin resistance. Information on dietary predictors of C-peptide levels in adolescence, however, is limited. Objectives: To examine associations between dietary intake of selected nutrients [fat, animal and vegetable protein, carbohydrate, lactose, fructose, glucose, galactose, sucrose, starch, fiber, calcium, zinc, iron, and sodium] and serum C-peptide in adolescent girls. Methods: The current study includes 155 postmenarcheal adolescent girls 14.6-to-18.9 yrs (mean: 16.6 ± 0.9 yrs) who attended the last visit of the DISC. Intervention and control group data are combined. Dietary intake was assessed via 3 averaged, nonconsecutive 24–hour dietary recalls with nutrient intakes estimated using Nutrition Data System v20. Serum C-peptide levels were determined by electrochemiluminescence immunoassay. Nutrients were selected based on those associated in the literature with insulin or C-peptide and availability in our data set. Associations between nutrients and C-peptide were estimated using unadjusted and adjusted Pearson's correlation analyses. Adjusted geometric means of C-peptide were also calculated within quartiles of each nutrient with linear trends calculated across quartiles. Adjusted analyses included age, time since menarche and energy intake. Associations were also examined using forward-selection multivariable linear regression (p<0.05) including age, time since menarche, body mass index (BMI)-for-age-percentile, total fat, energy intake, starch, and either fiber or vegetable protein intake in analyses. To mitigate skewness, nutrients and C-peptide were log transformed; thus, β coefficients represent percentage changes in C-peptide for 1% increases in nutrient intakes. Results: In correlation analyses the following dietary variables were significant predictors of C-peptide (p<0.05): total fat, vegetable protein, starch and fiber. Adjusted Pearson's correlation coefficients were 0.17, -0.19, -0.17 and -0.20, respectively. In adjusted analyses, trends in geometric means of C-peptide across quartiles of nutrients were significant for vegetable protein (p=0.048) and fiber (p=0.027). For the lowest to the highest quartiles of dietary fiber intake, geometric means of C-peptide were 643 pmol/L (95% CI: 547, 755), 593 pmol/L (95% CI: 501, 702), 592 pmol/L (95% CI: 504, 695), and 490 pmol/L (95% CI: 417, 577). In multivariable linear regression analyses, vegetable protein and fiber intake were examined in separate models due to collinearity (r=0.749, p<0.0001). In forward selection multivariable analyses, only BMI-for-age percentile (β=0.004, p=0.004) and fiber intake (β=-0.305, p=0.009) remained as significant predictors of C-peptide. In a similar model, vegetable protein was not a significant predictor. Conclusion: This study suggests dietary fiber intake in adolescent girls is inversely associated with serum C-peptide, a marker of insulin resistance, after adjustment for BMI-for-age-percentiles. To our knowledge, no previous studies have examined this association in adolescence. Citation Format: Zhenzhen Zhang, Jean M. Kerver, Joseph C. Gardiner, Joanne F. Dorgan, Ellen M. Velie. Dietary intake and C-peptide levels in adolescent girls: Results from the Dietary Intervention Study in Children (DISC). [abstract]. In: Proceedings of the Eleventh Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2012 Oct 16-19; Anaheim, CA. Philadelphia (PA): AACR; Cancer Prev Res 2012;5(11 Suppl):Abstract nr A26.