Prospective Pharmacokinetic Evaluation of Venetoclax in AML Supports Re-Evaluation of Recommended Dose Adjustments With Azole Antifungals

Abstract

Incorporation of the BCL-2 inhibitor, venetoclax (VEN), into the treatment paradigm of acute myelogenous leukemia (AML) has led to a dramatic improvement in outcomes for older and unfit patients, demonstrating an overall survival benefit when added to azacitidine in patients with newly diagnosed (ND) AML ≥ 75 years, or otherwise ineligible for intensive chemotherapy [1]. Posaconazole prophylaxis has improved overall survival in patients with ND AML undergoing remission induction chemotherapy anticipated to experience neutropenia for > 7 days [2]. As a result, prophylaxis with any mold-active triazole antifungal is recommended [3].

The triazole antifungals inhibit CYP3A4, the enzyme responsible for VEN metabolism, to varying degrees. As a result, specific VEN dose reductions are recommended when co-administered with CYP3A4 inhibitors (CYP3A4i). A pharmacokinetic (PK) analysis of 11 patients receiving VEN with posaconazole 300 mg daily demonstrated an increase in mean C_max by 53% and AUC by 76% with VEN 50 mg daily (VEN50), and a 93% increase in C_max and 155% increase in AUC with VEN 100 mg (VEN100) [4]. As a result, the FDA recommends VEN 70 mg daily in combination with posaconazole. Notably, in the VIALE-A trial, all patients receiving any strong CYP3A4i were dose-reduced to VEN50 [1].

We and others have reported a delay in time to platelet count recovery among patients with ND AML receiving increased VEN exposure as a result of increased VEN doses or when VEN is given in combination with azole antifungals, particularly posaconazole [5, 6]. As a result, we hypothesized that VEN serum levels may be supratherapeutic when given in combination with posaconazole and that higher VEN levels could be associated with prolonged myelosuppression.

As part of an ongoing phase 2 study of VEN combined with cladribine and LDAC in older patients with ND AML (NCT03586609), we prospectively characterized VEN pharmacokinetics including C_max, AUC, C_trough and clearance when given with or without a strong CYP3A4i during induction. VEN100 was administered with voriconazole, VEN50 and VEN100 with posaconazole, and VEN 400 mg (VEN400) with caspofungin. Steady state VEN PK analysis was conducted on day 8. Blood samples were collected prior to the dose, and 2, 4, 8, and 24 h post dose. Trough (24-h post dose) levels were collected on days 12 and 16 of cycle 1 (Figure S1). We also evaluated the association between VEN trough levels and AUC as well as clinical outcomes. Complete methodology is in the Data S1.

Thirty-nine patients, median age 68 years (range, 61–77), were included for PK analysis (Table S1), of whom 33 (85%) achieved CR (n = 29) or CRi (n = 4) after induction. Among responders, 28 patients (85%) achieved MRD-negativity. Eighteen patients received VEN100 (n = 11) or VEN50 (n = 7) with posaconazole, 12 patients received VEN100 with voriconazole, and 9 received VEN400 with caspofungin (Table S1). Median time to ANC > 1 × 10³/μL and PLT > 100 × 10⁶/μL during induction was 27 and 25 days, respectively.

Venetoclax pharmacokinetics were altered most significantly when administered at the 100 mg dosage with posaconazole (Figure 1A–D; Table S2). Average C_max and C_trough values were highest among patients receiving VEN100 and posaconazole (1.55 μg/mL and 1.09 μg/mL) or voriconazole (1.36 μg/mL and 0.99 μg/mL) compared with VEN400 with caspofungin (0.95 and 0.31 μg/mL) (Figure 1B). VEN clearance was severely delayed when combined with posaconazole, to 3.89 mL/h with VEN100 and 4.71 mL/h with VEN50 (Figure 1D). As a result, the average VEN AUC with posaconazole and VEN100 was 30.17 μg*h/mL, a 103% increase compared to caspofungin and VEN400 (Figure 1C). With voriconazole, average AUC was increased by 77% to 26.4 μg*h/mL (Figure 1C). Reducing VEN to 50 mg with posaconazole resulted in a lower average C_max of 1.16 μg/mL, C_trough of 0.73 μg/mL (Figure 1A–D), and AUC of 25.53 μg*h/mL, a 72% increase in AUC compared to VEN400 and caspofungin (Figure 1C). Due to posaconazole's significant impact on clearance, the calculated accumulation index was 12.5 with VEN100 and 9.2 with VEN50 versus 3.7 with voriconazole and 1.7 with caspofungin (Table S2). A significant correlation between VEN AUC and VEN trough levels was observed (p < 0.001, r² = 0.89) (Figure 1E). VEN trough levels collected on days 8–16 remained consistent for each patient regardless of antifungal group, indicating little intra-patient variability; significant interpatient PK variability was observed (Figure S2).

FIGURE 1

Open in figure viewerPowerPoint

Venetoclax pharmacokinetic results. (A) Mean (Std. dev) venetoclax plasma concentration-time profiles in different antifungal groups (caspofungin, voriconazole or posaconazole) and venetoclax (400 mg, 100 mg or 50 mg) doses; comparison of venetoclax pharmacokinetic parameters (B) C_trough, (C) area under the curve (AUC), (D) clearance among various antifungal and venetoclax doses (caspofungin + venetoclax 400 mg, voriconazole + venetoclax 100 mg, posaconazole + venetoclax 100 mg, and posaconazole + venetoclax 50 mg) groups. Wide interpatient pharmacokinetic variability observed within each group; (E) linear correlation observed between venetoclax AUC_tau and C_trough levels. Comparison of venetoclax pharmacokinetic parameters (F) C_trough, (G) area under the curve (AUC), and (H) Clearance among patients who died at 12 weeks versus those who did not.

A high CR/CRi rate prevented significant differences in PK parameters to be determined according to clinical outcomes including achievement of CR/CRi and ANC or PLT recovery after cycle 1. Median C_trough was 0.73 μg/mL (range, 0.09–2.71) overall. Two patients had C_trough < 0.1 μg/mL, neither of whom achieved a response after induction. Assuming a consistent VEN concentration was maintained after achieving steady state, we utilized trough levels obtained during cycle 1 to represent anticipated VEN exposure during subsequent cycles. Eighteen patients who achieved CR after induction and proceeded to cycle 2 with the same VEN dose and azole combination were analyzed for blood count recovery after cycle 2 (Table S3). Median C_max and C_trough were significantly higher among those who did not achieve (n = 12) ANC > 1 × 10³/μL by day 28 of cycle 2 compared to those (n = 6) who did (C_max: 1.56 μg/mL vs. 0.97 μg/mL, p = 0.042; C_trough: 1.12 μg/mL vs. 0.6 μg/mL, p = 0.05). Seven of these 18 patients had C_trough > 0.85 μg/mL, all of whom were receiving VEN100 with posaconazole (n = 4) or voriconazole (n = 3). Among these 7 patients, 1 achieved PLT > 100 × 10⁶/μL by day 35 and 3 achieved ANC > 1 × 10³/μL by day 35. No patient with C_trough > 0.85 μg/mL achieved ANC > 1 × 10³/μL by day 28 of cycle 2.

Three patients (8%) died within 12 weeks due to gram-negative sepsis (n = 1), intracranial hemorrhage (n = 1), and COVID-19 (n = 1). These 3 patients had a significantly higher median VEN AUC of 42.13 μg*h/mL (range, 24.24–64.43 μg*h/mL) compared with 21.99 μg*h/mL (range, 4.34–48.83 μg*h/mL) (p = 0.044) and C_trough of 1.53 μg/mL (range, 0.86–2.71 μg/mL) versus 0.73 μg/mL (range, 0.09–2.04 μg/mL) (p = 0.039); and significantly lower clearance of 2.06 mL/h (range, 1.55–2.34 mL/h) versus 4.33 mL/h (range, 1.5–34.95 mL/h); p = 0.019 compared to those who did not experience early mortality (Figure 1F–H; Table S4). All 3 patients were receiving concomitant posaconazole with VEN100 (n = 2) and VEN50 (n = 1).

Overall, our pharmacokinetic analysis revealed an increase in VEN exposure when given concomitantly with either voriconazole or posaconazole despite VEN dose reductions. Concomitant voriconazole with VEN at the recommended dose of 100 mg resulted in a higher maximum VEN concentration (C_max) and total VEN exposure (AUC) compared to that observed without a concomitant CYP3A4i. Posaconazole demonstrated an even greater impact on both C_max and AUC when administered with the same dose of VEN100 likely due to its more potent CYP3A4 inhibition, supporting the need for a more significant dose reduction with this combination. Reducing VEN to 50 mg with posaconazole still resulted in increased VEN exposure overall, but C_max, C_trough, and AUC values were most similar to those observed with VEN without a concomitant CYP3A4i. The highest VEN concentrations were reached with voriconazole or posaconazole combined with VEN100 and occurred in patients without timely ANC recovery after cycle 2. The combination of VEN50 with posaconazole resulted in comparatively lower VEN concentrations and did not appear to contribute to prolonged myelosuppression during course 2.

Posaconazole significantly delayed VEN clearance, resulting in more than a 5- and 2.5-fold higher accumulation index compared with caspofungin and voriconazole, respectively. Although no relationship was observed between any PK parameter and blood count recovery in induction, prolonged myelosuppression has been observed clinically during consolidation (cycle 2+) despite reduced chemotherapy and VEN dosing. No patient with a C_trough > 0.85 μg/mL achieved ANC recovery by day 28 of cycle 2, and only 1 achieved PLT recovery by this time. Based on our modeling, we hypothesize the impact on myelosuppression is more significant during consolidation as a result of reduced VEN clearance by strong CYP3A4i, leading to accumulation and prolonged VEN exposure.

Venetoclax trough levels correlate with AUC, therefore can serve as a surrogate for VEN exposure and be considered for therapeutic drug monitoring (TDM) in the future. Our data demonstrated consistent trough values with little intra-patient variability; therefore, a single true trough level collected after achieving steady state (≥ 7 days of consecutive treatment) would be adequate. We observed an increase in VEN trough following an increase in posaconazole dose. Consequently, any changes to interacting medications warrants obtaining an additional VEN trough level. An ideal TDM parameter must also correlate with a safety or efficacy outcome. Elevated C_trough levels were observed in 3 patients who died within 12 weeks and among those with suboptimal blood count recovery after cycle 2, therefore further investigation into a goal C_trough range to optimize efficacy and safety is necessary.

In conclusion, VEN C_max, C_trough, and AUC values were greater when administered with voriconazole or posaconazole compared to caspofungin, even with recommended VEN dose reductions. Posaconazole significantly delays VEN clearance, leading to accumulation. Venetoclax 50 mg in combination with posaconazole resulted in PK parameters most similar to that observed with VEN without a concomitant CYP3A4i and is an appropriate dose reduction. Venetoclax trough levels correlated well with AUC and may be a suitable variable for TDM in the future. Further PK analyses evaluating VEN trough levels and clinical outcomes are warranted.