Retrospective, real‐life study of venetoclax plus azacitidine or low‐dose cytarabine in French patients with acute myeloid leukemia ineligible for intensive chemotherapy

Abstract Background Recently, the combination of venetoclax plus a hypomethylating agent (HMA; azacitidine ordecitabine) or low‐dose cytarabine (LDAC) showed promise in Phase III trials in previously untreated AML. In France at the time of this study, venetoclax was not yet approved for AML and there were therefore no formal usage recommendations. Here we report the first study in a French cohort that assessed venetoclax in combination with existing treatments for AML under real‐life conditions. Method This retrospective, real‐life study collected data on venetoclax use and management in a French cohort with acute myeloid leukemia (AML) ineligible for intensive chemotherapy. Result Of 118 patients, 81 were in second line/beyond (71.6% also hypomethylating agent [HMA]; 23.5% lowdose cytarabine [LDAC]) and 37 in first line. For venetoclax initiation, 57.3% underwent ramp up and 74.6% were hospitalized. Median venetoclax duration was 2.5 months (range 0.03‐16.2). With all treatment lines and regimens, most common grade 3/4 adverse events were hematologic (overall 96.4% of patients) and infections (57.1%). Dosage adjustments for drug interactions and safety varied between centers. In second‐line/beyond, median progression‐free survival was 4.0 months (95% confidence interval [CI] 2.7‐12.8) with venetoclax‐HMA and 3.4 months (1.3‐8.9) with venetoclax‐LDAC; overall response rate was 51.9% and 41.2%, respectively. Thus, we showed that venetoclax‐based treatment yields promising findings in patients with AML, but to address treatment complexity, practice harmonization is needed.


| INTRODUCTION
Acute myeloid leukemia (AML) mainly affects older adults, 1 who often present with comorbidities and are ineligible for intensive chemotherapy. Recently, the combination of venetoclax plus a hypomethylating agent (HMA; azacitidine or decitabine) or low-dose cytarabine (LDAC) showed promise in Phase III trials in previously untreated AML. 2,3 DiNardo et al 2 reported median overall survival (OS) of 14.7 months for venetoclax-azacitidine versus 9.6 months for placebo-azacitidine. Wei et al 3 reported median OS of 8.4 months for venetoclax-LDAC versus 4.1 months for placebo-LDAC. In France at the time of this study, venetoclax was not yet approved for AML and there were therefore no formal usage recommendations.
Here we report the first study in a French cohort that assessed venetoclax in combination with existing treatments for AML under real-life conditions.

| MATERIALS AND METHODS
This retrospective, observational study included patients with AML (World Health Organization 2016 classification) 4 who were ineligible for intensive chemotherapy and received venetoclax, from 14 French centers, between February 2017 and May 2020. The study was conducted in compliance with reference methodology MR004 (V3.1 23/01/2020) and all patients gave informed consent.
The primary objective was to collect pseudonymized data on venetoclax use and management. We also investigated safety and tolerability ( overall response rate was 51.9% and 41.2%, respectively. Thus, we showed that venetoclax-based treatment yields promising findings in patients with AML, but to address treatment complexity, practice harmonization is needed.

K E Y W O R D S
acute myeloid leukemia, azacitidine, low-dose cytarabine, venetoclax Statistical analyses used R software (v3.6.1; see Appendix S1). For descriptive data expressed as %, missing data were subtracted from the denominator (n/N is given where this applies).
Antifungal prophylaxis was used in 68.1% (79/116) of patients, most commonly azoles. Patients receiving antifungal prophylaxis had a higher rate of dosage adjustment at initiation due to drug interaction (p = 0.031) and a lower final venetoclax dose (p < 0.001) than those without antifungal prophylaxis. For initiation, the dosage was adjusted in 50.4% (59/117) of patients, due to drug interaction (n = 49, 43 receiving antifungal prophylaxis) and safety (n = 10). After the first cycle, dosage was adjusted in 41.5% of patients, due to safety (n = 35), drug interaction (n = 12, 8 receiving antifungal prophylaxis), and being in the COVID-19 setting (n = 1). At the time of analysis, 22.0% of patients were still receiving venetoclax. Of 90 patients who discontinued venetoclax, primary reasons were progressive disease (disease progression or relapse; 51.1% of discontinuations), toxicity (21.1%), death (18.9%), and undergoing stem cell transplantation (8.9%).
Overall, 94.9% of patients experienced at least one treatment-emergent grade 3/4 adverse event (Table S2) (Table S3) were neutropenia, thrombocytopenia, and anemia. Dosage adjustments for predominantly hematologic toxicity included adjusting the venetoclax dose and modifying the schedule. Red blood cell and platelet transfusions occurred in 87.3% and 76.3% of patients, respectively. Antibiotics were used in 77.1% of patients. Hematopoietic growth factors, such as granulocyte colony-stimulating factor, were used in 22.6% (26/115) of patients.

| DISCUSSION
Our observations around the real-life use, management, and safety of venetoclax in patients with AML in French centers were in line with previous findings. 2,3,[7][8][9] Since venetoclax is metabolized by CYP3A4, dosage adjustments for drug interactions were expected, 10-12 including when using azole CYP3A4 inhibitors for antifungal prophylaxis. 10 Rates of cytopenia were similar to those in other real-life studies. 8,9 Of note, we found that contrary to standard practice a ramp-up phase was not always used for venetoclax initiation, and that incompliance with a ramp-up phase increased the likelihood of TLS. We also observed variations between our centers in dosage adjustments made both for drug interactions and safety. Algorithms to address dosage adjustments are now present in recommendations for venetoclax-based treatment in patients with AML, [13][14][15] which may help to harmonize practices.
Our heterogeneous real-life population makes comparisons with the Phase III trials difficult, with no restrictions on treatment lines, prior hematologic disorders, prior HMA and hydroxyurea exposure, and cytogenetic risk groups. Moreover, our sample size is limited, especially the first-line cohort of only 37 patients. Nonetheless, our findings for venetoclax-based treatment in patients ineligible for intensive chemotherapy were promising, including in relapsed/refractory AML and when combined with an HMA. Our second-line/beyond ORR of 51.9% with venetoclax-HMA and 41.2% with venetoclax-LDAC, and CR/CRi rate of 33.3% and 23.5%, respectively, were comparable to those in other studies. 8,9,[16][17][18] However, our first-line CR/CRi rate of 43.8% with venetoclax-HMA was lower than in a Phase III trial (66%) 2 and our first-line median response duration of 105 days was also relatively short (17.8 months in Phase III trial).
Our findings showed evidence of a multidisciplinary approach, although this was not consistent. Clinical pharmacists play key roles in venetoclax management, including patient education, building links between hospital and ambulatory medicine, and collaboration around drug interactions. First developed in France for lymphoid disease, ambulatory medical assistance (AMA) is a multidisciplinary program that uses phone calls from nurses to monitor patients who are undergoing active home treatment. 19 During active-phase treatment in patients with hematologic malignancies, the AMA program was demonstrated to ensure safe care through avoidance of adverse events (hematologic or nonhematologic). 20 Patients with AML treated with new molecules are thus good candidates for an AMA program.
In conclusion, this real-life study in France showed that venetoclax-based treatment yields promising findings in patients with AML ineligible for intensive chemotherapy, but its use and management are complex, particularly in terms of dosage adjustments. Practice harmonization, along with greater clinical pharmacist and nurse involvement, could potentially be future improvements.