Abstract A046: Metabolomic profiles associated with breastfeeding and subsequent ovarian cancer risk

Abstract Introduction: While inverse associations have been reported in epidemiologic studies of breast feeding and ovarian cancer risk, little is known about the biologic pathways impacted by breastfeeding that leads to risk reduction. Therefore, we aimed to identify individual metabolites and pathways associated with breastfeeding among parous women in the Nurses’ Health Study (NHS) (n=7, 111) and NHSII (n=2, 793), and ultimately develop a breastfeeding metabolite score to quantify the association of this score with ovarian cancer risk using a nested case control study within NHS/NHSII (n=504 cases and controls). Methods: To identify individual metabolites associated with breastfeeding, we conducted a cross-sectional analysis using data from all nested studies of metabolites and various disease outcomes (e. g. , ovarian cancer, type 2 diabetes, stroke) within the NHS/NHSII. Liquid chromatography tandem mass spectrometry was used to measure 717 known metabolites. We excluded metabolites that did not pass the processing delay pilot study or were below the limit of detection for 25% of participants. Other values below the limit of detection were imputed to be ½ the minimum value for that metabolite, resulting in 606 metabolites. All metabolite values were transformed to probit scores to achieve normality. Associations of having ever breastfed (yes, no) and breastfeeding duration (never, ≤6 months, 6 to 12 months, ≥12 months) with metabolite levels were examined using multivariable linear regression adjusting for age, date of blood draw, time of blood draw, fasting status, and menopausal status. Metabolites were considered associated with breastfeeding if p≤0. 10 after Bonferroni correction. Results: Of the 9, 904 parous participants included in analyses, 73% (n=7, 225) reported having ever breastfed. When comparing having ever breastfed to never, the average difference of indole-3-propionate was 19% of one standard deviation (SD) higher (β=0. 19; 95% CI: 0. 11, 0. 26), 4-pyridoxate was 15% of one SD higher (β=0. 15; 95% CI: 0. 08, 0. 21), proline betaine was 12% of one SD higher (β=0. 12; 95% CI: 0. 07, 0. 16), and both N-acetylornithine (β=0. 09; 95% CI: 0. 05-0. 14), and pantothenic acid (β=0. 09; 95% CI: 0. 05-0. 14) were increased by 9% of one standard deviation. The average difference of cotinine was 15% of one SD lower (β=-0. 15; 95% CI: -0. 19, -0. 10), threonine was 12% of one SD lower (β=-0. 12; 95% CI: -0. 16, -0. 07), and hydroxyproline was 9% of one SD lower (β=-0. 09; 95% CI: -0. 13, -0. 04) among those who ever breastfed compared to never. After considering duration, only having breastfed for 12 months, or longer was associated with these metabolites. Conclusion: This study suggests prior breastfeeding is associated with altered levels of several metabolites in parous women. Additional analyses in progress include metabolite set enrichment analysis and development of a breastfeeding metabolite score to better understand the underlying mechanisms that may mediate the relationship between breastfeeding and ovarian cancer risk. Citation Format: Jennifer M. Mongiovi, Julian Avila Pachecoc, Clary B. Clish, A. Heather Eliassen, Mary K. Townsend, Shelley S. Tworoger, Oana A. Zeleznik, Naoko Sasamotob. Metabolomic profiles associated with breastfeeding and subsequent ovarian cancer risk abstract. In: Proceedings of the AACR Special Conference on Ovarian Cancer; 2023 Oct 5-7; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84 (5 Suppl₂): Abstract nr A046.