### 23060 ### Validation Of Circulating MicroRnas As Biomarkers In Heart Failure In Two Large Independent Cohorts Antoni Bayes-Genis1, David Lanfear2, Maurice de Ronde3, Josep Lupon1, Joost Leenders4, Elisabet Zamora1, L. Keoki Williams2, De Antonio Marta5, Zhen Liu4, Koos Zwinderman3, Sara-Joan Pinto-Sietsma3, Yigal Pinto3; 1Universitat Autonoma de Barcelona, Barcelona, Spain; 2Henry Ford Hospital, Detroit, MI; 3Academic Med Cntr, Univ of Amsterdam, Amsterdam, Netherlands; 4Mirnext BV, Amsterdam, Netherlands; 5Germans Trias i Pujol Hosp, Badalona, Spain Background: Small studies suggested circulating microRNAs (circmiRs) as biomarkers for Heart Failure (HF). However, standardized approaches and quality assessment are not established, and results have been inconsistent, with little replication between studies. We aimed to implement quality standards to enable comparison between cohorts and assess which circmiRs may add prognostic information in HF. Methods: We measured 15 circmiRs in two independent cohorts totaling >2000 subjects. Cohort I (Barcelona) comprised of n=843 chronic HFrEF patients. Cohort II from Detroit comprised n= 1384 chronic HF patients (892 HFrEF, 492 HFpEF). Each sample was measured in duplicate, and normalized to an abundant and stable circmiR (miR-486-5p). Algorithms were installed to define each circmiR measurement as “valid”, “unmeasurable” or “invalid”. This allowed inclusion of valid low-level circmiR measurements while reducing noise from false amplification signals. Results: In general, between 20–40% of measurements were “invalid”, while miR-499a\_5p and -208a were “unmeasurable” in the majority of patients in both cohorts. Higher levels of circmiRs-133b, -1254, -622, -208a and -499a\_5p were significantly associated with risk of death in both cohorts, with hazard ratios ranging from 1.103 to 1.365 per log increase (p-values 0.001 to 0.05). However, adding these circmiRs to established predictors (age, renal function and NTproBNP) did not further augment the c-stat beyond 0.71 (cohort I) or 0.78 (cohort II). …
Read full abstract