Population Pharmacokinetics of Oral Azacitidine, and Exposure-Response Analysis in Acute Myeloid Leukemia

Oral azacitidine (oral-AZA) maintenance is approved for adults with acute myeloid leukemia (AML) in remission post-intensive chemotherapy, not proceeding to hematopoietic stem cell transplantation. This study aimed to develop a population pharmacokinetic (PopPK) model to characterize oral-AZA concentration-time profiles in patients with AML, myelodysplastic syndrome, or chronic myelomonocytic leukemia. PopPK-estimated exposure parameters were used to evaluate exposure-response relationships in the phase III QUAZAR AML-001 study. The PopPK dataset comprised 286 patients with 1,933 evaluable oral-AZA concentration records. The final PopPK model was a one-compartment model with first-order absorption incorporating an absorption lag time and first-order elimination. Regression analyses identified two oral-AZA exposure parameters (area under the plasma concentration-time curve at steady state (AUC_ss ); maximum plasma concentration (C_max )) as statistically significant predictors for relapse-free survival (hazard ratio (HR) = 0.521, P < 0.001; HR = 0.630, P = 0.013; respectively), and AUC_ss as a significant predictor for overall survival (HR = 0.673, P = 0.042). The probability of grade ≥ 3 neutropenia was significantly increased with increases in AUC_ss (odds ratio (OR) = 5.71, 95% confidence interval (CI) = 2.73-12.62, P < 0.001), cumulative AUC through cycles 1 to 6 (OR = 2.71, 95% CI = 1.76-4.44, P < 0.001), and C_max at steady-state (OR = 2.38, 95% CI = 1.23-4.76, P = 0.012). A decreasing trend was identified between AUC_ss and relapse-related schedule extensions, vs. an increasing trend between AUC_ss and event-related dose reductions. As the majority (56.8%) of patients required no dose modifications, and the proportions requiring schedule extension (19.4%) or dose reduction (22.9%) were almost equal, oral-AZA 300 mg once daily for 14 days is the optimal dosing schedule balancing survival benefit and safety risk.