Predictors of hypomethylating agent discontinuation among patients with higher-risk myelodysplastic syndromes.

Abstract

7043 Background: Real-world studies have shown that persistence with intravenous (IV) and subcutaneous (SC) hypomethylating agents (HMAs) among patients (pts) with higher-risk myelodysplastic syndromes (MDS) is poor, with over one-third of treated pts receiving <4 cycles or having a ≥90-day gap in therapy, despite recommendations for at least 4-6 cycles to elicit response in absence of progression/unacceptable toxicity. Survival outcomes have also been shown to be worse, and direct medical costs higher, among HMA non-persistent vs persistent pts. We explored factors associated with early discontinuation of HMA therapy in this population. Methods: This was a retrospective cohort study among pts from the 2010-2016 SEER-Medicare linked database with a diagnosis of refractory anemia with excess blasts (RAEB; a surrogate for higher-risk MDS) from 2011-2015. Included pts had to have received HMA therapy and have ≥12 months’ continuous follow-up after diagnosis. Discontinuation was defined as stopping HMA therapy before 4 cycles. Multivariable logistic regression was used to assess predictors of HMA discontinuation. Results: In total, 664 pts with RAEB and treated with HMAs were included. Overall, 193 (29%) discontinued before 4 cycles; of these, 91 (47%) discontinued after 1 cycle, 57 (30%) 2 cycles, and 45 (23%) 3 cycles. Compared with pts continuing for ≥4 cycles, pts discontinuing before 4 cycles were generally older and more likely to be single/separated/divorced/widowed, have more comorbidities, and have poor performance status (PS) (Table). These trends were most pronounced among pts discontinuing HMA therapy after only 1 cycle vs ≥4 cycles (Table). In multivariable analysis, age 71-75 vs ≥80 y (odds ratio [OR] 0.556, p=0.017) and poor PS (OR 1.585, p=0.019) remained significant predictors of HMA discontinuation. Among treatment-related factors, the most statistically significant association with HMA discontinuation was observed for GCSF use (OR 0.453, p<0.001). Number of pills/day was not a predictor of HMA discontinuation (OR 1.009, p=NS). Conclusions: In this real-world study, almost one-third of RAEB pts treated with IV/SC HMAs discontinued before 4 cycles, with almost half of these pts discontinuing after only 1 cycle. Predictors of HMA discontinuation included older age and poor PS. Novel approaches are needed to improve persistence with HMA therapy, particularly among these higher-risk groups.[Table: see text]