Abstract P192: Metabolomic Profiles Associated with Total and CVD Mortality in Women
Background: Using novel metabolomic markers to identify individuals at increased risk for total and cardiovascular mortality may help inform therapeutic or prevention strategies.
Methods: A total of 2306 women in the Women’s Health Initiative (WHI) Observational Study (WHI-OS) and Hormone Trials (WHI-HT) had 370 plasma metabolites measured at baseline by liquid chromatography tandem mass spectroscopy. The analysis cohort included 2298 women (943 WHI-OS; 1355 WHI-HT) with complete covariate data. Multivariate cox proportional hazards models were constructed to examine the association of metabolites with mortality, adjusting for age, body mass index, systolic blood pressure, hypertension treatment, diabetes, smoking, total and HDL cholesterol. The WHI-OS data were used for discovery analyses; metabolites with a multivariate-adjusted false discovery rate (FDR) adjusted p value <0.05 were considered significant. These candidate metabolites were included for backwards selection (p for exclusion > 0.05) to identify independent metabolites for inclusion in a composite score. The score of these significant mutually-adjusted metabolites was created based on the quartile score for each metabolite, with higher values indicating higher likelihood of all cause mortality and was tested in the WHI-HT cohort.
Results: During a median follow-up of 16.7 years, 1,102 women died (417 WHI-OS; 685 WHI-HT), with 601 cardiovascular deaths. At baseline, the median age was 68 years (interquartile range 62-72) and median time to death was 9.9 years. In the WHI-OS discovery cohort, 56 metabolites were significantly associated with multivariate-adjusted total mortality (FDR p< 0.05). When all 56 metabolites were mutually adjusted in the same model and subjected to backwards selection, 15 metabolites remained significantly associated with total mortality after multivariate adjustment and were used to create a composite score. In the separate WHI-HT validation cohort, women in the highest score quartile had a hazards ratio (HR) of 2.36 (95% CI: 1.88-3.21) for death. The mortality score was also associated with CVD death (HR 3.27 [2.54-4.22] highest vs. lowest quartile) for the combined WHI-OS and WHI-HT sample. Six metabolites were significantly associated with mortality in both samples with multivariate mutual adjustment; docosatrienoic acid an omega-3 fatty acid, histamine a pro-inflammatory compound , and the alpha-amino acids histidine and tryptophan were protective while 3-ureidopropionic acid an intermediate in the metabolism of uracil, and 18:1 cholesterol ester were associated with increased risk.
Conclusions: Using a robust discovery and validation design, metabolite profiles were associated with significantly increased total and CVD mortality, even after adjustment for traditional risk factors.
Author Disclosures: K.M. Rexrode: None. R. Balasubramanian: None. N. Paynter: None. F. Giulianini: None. J. Chen: None. J.E. Manson: None. M. Vitolins: None. C.M. Albert: None. C. Clish: None.
- © 2017 by American Heart Association, Inc.