PS1395 COMPARATIVE EFFECTIVENESS OF FRONT-LINE TREATMENTS FOR PATIENTS WITH NEWLY DIAGNOSED MULTIPLE MYELOMA WHO ARE TRANSPLANT INELIGIBLE

Background: Carfilzomib (CFZ) is a proteasome inhibitor associated with cardiovascular (CV) adverse events (AEs) such as hypertension (HTN), cardiac dysfunction and thrombotic events such as venous thromboemolism and thrombotic microangiopathy (TMA). Endothelial dysfunction is a major mediating mechanism in these complications and may be affected by proteasome inhibition caused by CFZ, but have not been explored in humans. Aims: To assess whether indices of vascular dysfunction can provide prognostic information regarding the development of CV AEs in patients treated with carfilzomib Methods: We prospectively evaluated markers of vascular function as potential predictors of CFZ-associated CV complications and assessed associations with CFZ-mediated inhibition of proteasome activity in 46 relapsed/refractory myeloma (MM) patients treated with Kd [CFZ 20/56 mg/m2 and dexamethasone](NCT03543579). At baseline and at pre-specified time-points during Kd therapy, cardiac echo and hemodynamic and vascular function parameters were non-invasively assessed [aortic blood pressure and arterial wave reflections using pulse wave analysis, aortic stiffness using pulse wave velocity and endothelial function using flow-mediated dilatation of the brachial artery (FMD)] while proteasome activity (PrA) was measured in peripheral blood mononuclear cells (PBMC). The incidence of CV events [hypertension, acute coronary syndrome (ACS) and heart failure(HF)] was defined as the primary end-point and of both CV events and extra-coronary thrombotic events (pulmonary embolism (PE) and TMA) was defined as the secondary end-point. Results: Median follow up is 10 months. The prevalence of risk factors for CV toxicity was high (median 3 factors). CV AEs were recorded in 23 (50%) patients [HTN(Gr3/4): 29%, HF: 10%, ACS (Gr3): 4%, PA (Gr3): 2%, TMA: 10%]. Median time to first CV event was 1.87 months. Kd was discontinued in 3(6.25%) and dose was reduced in 2 patients (4.2%) due to cardiotoxicity; no CFZ dose reduction or discontinuation was needed for HTN. Low FMD [adjusted HR = 3.68] and high aortic systolic BP [adjusted HR = 4.57] (but not peripheral systolic BP), were the strongest independent predictors of the primary end-point, even after adjustment for baseline risk factors. Low FMD (adjusted HR = 4.01) was also an independent predictor the secondary end-point. Finally, increased baseline aortic, but not peripheral, SBP was the strongest independent predictor of grade ≥3 HTN events [adjusted HR = 11.23] after adjustment for risk factors for CV toxicity. FMD decreased acutely within 2 hours from 1st CFZ dose (from 5.2% to 3.6%, p = 0.008), and partially recovered before and after 2nd CFZ infusion; the decrease was more pronounced among patients who had lower recovery rate of PBMC PrA 24hpost 1st CFZ (5.12%to 2.97%, p = 0.002) but, was less pronounced in those who had higher PrA recovery rate (5.25% to 4.5%, p = 0.197), indicating that PrA could be implicated in CFZ induced endothelial dysfunction. Summary/Conclusion: Impaired FMD and increased aortic SBP at baseline were strong predictors of CV and extracoronary thrombotic AEs during CFZ therapy and could serve as clinical markers to identify patients at high risk for CFZ-associated CV toxicity.