Real-World Treatment Patterns, Resource Utilization, and Cost of Care in Medicare Beneficiaries Newly Diagnosed with Acute Myeloid Leukemia and Treated with Venetoclax-Based Regimens

Background: Previous research has demonstrated substantial healthcare resource utilization (HCRU) and poor survival outcomes in older patients diagnosed with acute myeloid leukemia (AML). Venetoclax-based regimens are a standard of care for patients not eligible for intensive chemotherapy (NIC). This study describes characteristics, treatment, and outcomes among US Medicare beneficiaries newly diagnosed with AML and receiving a first line (1L) venetoclax-based regimen. Methods: This retrospective cohort study used Medicare Fee-For-Service (FFS) Parts A/B/D claims and enrollment data to identify patients diagnosed with AML not yet in remission between September 1, 2016, and September 30, 2022, who initiated a 1L venetoclax-based regimen. Outcomes included duration of therapy and follow-up, time to next treatment (TTNT), per-beneficiary per-month (PBPM) costs (adjusted to 2022 USD), and HCRU, including hospitalizations, emergency department (ED) visits, outpatient services, blood transfusions, and pharmacy. Outcomes were measured from 1L initiation to first occurrence of AML relapse, second-line initiation, stem cell transplant, death, disenrollment, or end of study. Time to transfusion independence was defined as time from 1L start until beginning of first 2-month period without transfusion. TTNT was modeled using cumulative incidence functions (CIF) with death as a competing event and was reported overall and across age groups and Deyo-Charlson Comorbidity Index scores of 0-2, 3-4, and ≥5. Results: Of 55,033 beneficiaries diagnosed with AML not yet in remission, 3,184 received 1L treatment, met minimum continuous enrollment, and had no confounding diagnoses. Among these, 2,765 (86.8%) received a 1L venetoclax-based regimen (study population). Study patients were followed for a median of 6.9 months (interquartile range [IQR]=8.0). Median age at diagnosis was 76 years (IQR=8), and majority were male (n=1,528, 55.3%) and white (n=2,439, 88.2%). The most common 1L venetoclax regimen included administration of unspecified antineoplastics (n=702, 25.4%) followed by combination with azacitidine (n=695, 25.1%); while 445 received venetoclax monotherapy (16.1%). Median venetoclax dosage per day was 140.4 mg (IQR=182.8). Median duration of treatment was 5.2 months (IQR=7.2) and 30.9% of patients had a documented diagnosis of AML relapse as the reason for 1LOT termination. During 1L treatment, patients had a per-month median of 5.3 (IQR=4.3) office visits, 1 transfusion day (IQR=2.2), 0.3 (IQR=0.6) ED visits, and 0.2 (IQR=0.4) hospitalizations [with median hospital length of stay of 10 days (IQR=23.0)]. The median time to transfusion independence was 5.6 weeks (IQR=6.4). Mean total healthcare costs were $18,092 (SD=$14,289) PBPM, with $12,042 (SD=$12,147) PBPM attributable to medical services and $6,050 (SD=$7,841) PBPM attributable to Part D prescription drugs. AML-specific care constituted 78.1% of total healthcare costs. Across patients receiving 1L venetoclax-based regimens, there was a 46.4% (95% confidence interval=44.4-48.3%) probability of progressing to next treatment within one year of initiating treatment; a further 22.0% of patients died within the year without receiving a next treatment. Patients under 65 years of age had a 66.8% probability of progressing to next treatment within one year, versus 63.5% for ages 65-69, 55.9% for ages 70-74, 42.2% for ages 75-79, and 33.7% for ages 80+ (p<0.0001). The probability of progressing to the next treatment within one year was 50.3% for patients with CCI 0-2, 47.1% with CCI 3-4, and 43.7% with CCI≥5 (p=0.0028). Discussion: Within this Medicare study population, relapse was the most common reason for discontinuation of 1L venetoclax, aligning with high rates of relapse seen in older populations. Older patients and those with complex comorbidities were less likely to receive subsequent treatment. As clinicians shift towards evaluating both age and overall fitness when making AML therapy decisions, these treatment patterns are likely to evolve. Our findings provide initial insights into outcomes for an increasingly common AML therapy and can inform future hypothesis-driven research on clinical and economic outcomes.