PT-007 - EVALUATION OF CAUSAL INFERENCE OF BILE ACIDS ON CARFILZOMIB RELATED CARDIOVASCULAR ADVERSE EVENTS IN PATIENTS WITH MULTIPLE MYELOMA BY MENDELIAN RANDOMIZATION

S. Shabnaz¹, L. Shaffer², E. Farber-Eger², M. Tantawy³, R. Alshammari¹, T. Garrett¹, S. Rubinstein⁴, R. Cornell⁵, D. Lenihan⁶, Q. Wells², Q. Lu⁷, Y. Gong⁷; ¹University of Florida, Gainesville, FL, USA, ²Vanderbilt University Medical Center, TN, USA, ³University of Florida, Gainesville, FL, USA, ⁴NC, NC, USA, ⁵UNC, TN, USA, ⁶Saint Francis Medical Center, MO, USA, ⁷University of Florida, Gainesville, FL, USA.

Background: Carfilzomib (CFZ) is an efficacious proteasome inhibitor for the treatment of patients with multiple myeloma (MM) but has been associated with cardiovascular adverse events (CVAE). In this study, we aim to assess the causal inference of previously identified bile acids, tauro- and glycol-ursodeoxycholic acid (TUDCA & GUDCA), & CFZ-CVAE using a Mendelian randomization (MR) approach.

Methods: 49 MM patients enrolled in the observational ‘Prospective study of cardiac events during proteasome inhibitor therapy (PROTECT)’ study were analyzed. Genotyping was performed using the Illumina multi-ethnic genotyping array and imputation was performed using the TOPMED reference panel. Metabolome-wide association studies (mGWAS) were conducted for TUDCA & GUDCA using PLINK. To infer causality, two-sample MR by Inverse variance weighted method (IVW) was performed in MR Base using top SNPs (p < 5*e-6) from the mGWAS as instrumental variables. Sensitivity analyses were conducted using MREgger, weighted median, and weighted mode methods. Horizontal pleiotropy was assessed by MR-Egger regressions and heterogeneity by Cochran’s Q. For the outcome analysis of CVAE, one of the largest UK Biobank datasets available in MR-Base (cardiovascular disease case = 177923) was utilized.

Results: In mGWAS, 97 SNPs were associated with TUDCA & 11 SNPs were associated with GUDCA with p< 5e-6. The IVW showed significant causal link for both TUDCA (p = 0.03) and GUDCA (p = 0.0038). Sensitivity analyses confirmed these findings with no evidence of pleiotropy (p=0.087) or heterogeneity (p=0.68) for GUDCA. However, TUDCA displayed significant pleiotropy (p = 0.0052) & heterogeneity (p = 3.81e-13) (Fig1), suggesting potential alternative pathways. Among the top genes related to TUDCA, MYOCD & ZFHX3 might be directly involved with cardiac function, indicating a violation of the MR assumption. The MYOCD may play a key role in heart development, while ZFHX3 is identified as the gene responsible for atrial fibrillation and related heart anomalies.

Conclusion: MR analysis revealed a significant causal relationship between GUDCA and CVAE. These findings suggest that GUDCA may be a potential biomarker for guiding therapeutic decisions in MM patients undergoing carfilzomib treatment. However, further validation in larger cohorts is necessary to confirm its clinical utility.