Predicting the toxicity-efficacy ratio of venetoclax in real-world patients

Abstract

This study investigates the pharmacokinetic variability and exposure-response relationships of Venetoclax (VEN) in adult patients with acute myeloid leukemia (AML) who are ineligible for intensive chemotherapy. The study was conducted in a real-world clinical setting and included 48 patients who were treated with VEN in combination with azacytidine. We found significant inter-individual variability of 68% and intra-individual variability of 39% in plasma VEN concentrations. In addition, higher VEN concentrations were associated with better hematological responses. Median overall survival for the entire cohort was 17.8 months, with 1- and 2-year survival rates of 51% and 36.4%, respectively. A comparison of the 14-day VEN and 28-day VEN protocols showed that patients benefited from longer treatment durations, which resulted in more courses being administered. Plasma concentrations in the 14-day VEN protocol were higher than in the 28-day VEN protocol (2330 + 1675 ng/mL vs. 1503 + 966 ng/mL, respectively) without an increase in toxicities. The optimal protocol would be 400 mg/14d if we consider survival as a function of the 1818ng/mL cutoff. These results highlight the importance of considering therapeutic drug monitoring (TDM) as a strategy to optimize treatment outcomes by balancing efficacy and safety.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00277-025-06531-7.

Introduction

Acute myeloid leukemia (AML) is the most common type of acute leukemia [1]. Survival rates are particularly poor in older patients aged 65–70 years. In addition, elderly people with frequent comorbidities account for 50% of AML patients [2, 3]. For years, the canonical treatment for AML in the elderly has been azacytidine (AZA). Recently, the combination of Venetoclax (VEN) with azacytidine (i.e., VENAZA) has been the new backbone for the treatment of AML in elderly or frail patients who are ineligible for chemotherapy. The phase III VIALE-A trial demonstrated a clinical benefit of the combination of VEN and AZA over single agent AZA in terms of overall survival (66.7% vs. 28.3%) [4]. VEN is administered after an induction phase from 100 to 400 mg over 3 days in cycle 1, followed by 400 mg once daily in combination with azacytidine as recommended in the VIALE-A study [4] for the first 7 days of each 28-day cycle until progression or unacceptable toxicity [4]. This combination results in significant haematological toxicity and high infection rates ranging from 32 to 42%, leading to empirical adjustments in the dosing and scheduling of VEN [4]. While the standard treatment is based on a continuous course of 400 mg daily, many adjustments are made to the protocol in real-world settings, particularly with regard to the duration of exposure to VEN, which ranges from a minimum of 7 days to a maximum of 28 days [5]. For example, it is now recommended to reduce VEN exposure to 21 days every 28 days in responders with grade IV cytopenia lasting > 7 days [6, 7]. This reduction applies to more than half of patients. Dose adjustments are frequent in the VENAZA protocol. Indeed, in the Vacchini et al. study, venetoclax dose adjustments were required in 60% of patients [5]. Due to haematological toxicity, venetoclax dose adjustments were required in 61% of patients who achieved CR or Cri [8]. Shorter duration of venetoclax administration to 14 days has same efficacy and better safety profile in treatment of acute myeloid leukemia [9]. All of this is done empirically by adjusting how much to dose.

As with all oral targeted therapies, food intake, gene polymorphisms affecting liver enzymes or ABC-family transporters, and changes in CYP P450 metabolism such as drug-drug interactions are important sources of both inter and intra-individual pharmacokinetic variability. Compared with other oral targeted therapies [10], there are few data on therapeutic drug monitoring (TDM) with VEN, with most studies focusing on the influence of ethnicity on PK [11, 12]. In particular, exposure-response relationships have not been fully investigated. The primary objective of this real-world study was to evaluate the pharmacokinetic variability of VEN in adult AML patients. Secondary objectives were to investigate a possible relationship between plasma exposure and toxicities and efficacy.

Materials and methods

Study design and patients

This was a retrospective, non-interventional, observational study. The study was approved by our Review Board (Cellule d’Evaluation, at the Direction de la Recherche Clinique & de l’Innovation (DRCI, APHM) and were conducted in accordance with the European recommendations regarding the Clinical Trials Regulation (EU) 536/2014 and the General Data Protection Regulation (GDPR, EU) 2016/679. This study has been granted French MR-004 legal status for non-interventional studies. Due to its non-interventional nature, the review board waived the requirement for patient informed consent. This study was registered as MR-004 RH7VH5 [13]. All adult patients with AML [14] ineligible for intensive chemotherapy and receiving VEN at the University Hospital of Marseille, Marseille, France between August 2022 and June 2024 were included. Most patients received VEN 400 mg orally once daily on days 1–14 or 1–28 and azacytidine 75 mg/m² subcutaneously once daily on days 1–7 once every 28-day cycle. We collected pseudo-anonymized data on VEN dosing, pharmacist analysis and TDM values of co-medications, biological and clinical data from our institute database.

Outcomes and assessments

Disease response i.e. OS (Overall Survival), response, cytogenetic risk categorization, was recorded using the ELN 2022 criteria [15]. CR was defined as absolute neutrophils count > 1G/L, platelets > 100G/L, red cell transfusion independence, and BM with < 5% blasts. CRi was defined as all criteria for CR, except for neutropenia ≤ 1G/L or thrombocytopenia ≤ 100G/L. Bone marrow assessments were performed at screening and at the end of cycle 1. OS was defined as the time from random assignment to the date of death from any cause.

Safety analyses

Adverse events (AEs) were graded according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE) at the end of cycle 1 [16]. AEs occurring from the first cycle onwards were recorded.

Venetoclax Pharmacokinetic analysis

Blood samples were collected on Day 8 of the first course, or Day 6 for subsequent courses, to ensure that the drug had reached steady state. Blood samples were shipped at 4 °C to the analytical laboratory for centrifugation and freezing within 5 h to avoid degradation. The trough blood concentration of VEN (Ctrough) was determined by liquid chromatography-mass spectrometry (LC/MS), validated according to EMA and ISO 15,189 guidelines, with both accuracy and precision below 15% [17]. If the sampling time missed the trough concentration, concentrations were then extrapolated to T24h using the formula of Van Eerden et al. [18].

Statistical analysis

Demographics were analyzed using descriptive statistics. Pairwise comparison was used to show Ctrough as a function of dose administered and the Cmin distribution in the first dose. Inter- and intra-individual variabilities were calculated using the coefficient of variation of each patient’s first residual concentration. The best cut-off was determined using a Maximally Selected Rank Statistics method [19, 20]. OS was estimated using the Kaplan-Meier method. Statistical analyses were performed using R^® software (version 4.3.1). A value of p ≤ 0.05 was considered statistically significant.

Results

Patient’s baseline characteristics

A total of 48 patients received VEN treatment, including 28 men and 20 women (Table 1). The median age at VEN initiation was 77 years, ages range from 40 to 86 years. The majority of patients had de novo AML (n = 25). Risk stratification according to the current ELN criteria 2022 classified patients into favorable (2 patients), intermediate (9 patients) and unfavorable risk groups (37 patients) (Table 1). Regarding molecular alterations, TP53 mutations affected 27%, FLT3 mutations 21%, IDH1/2 18% and NPM1 17% of the patients.

Table 1.

Baseline demographic and clinical characteristics of the patients

		Patients
Characteristics		N		Percentage
Patient’s numbers		48		100%
Sex
Female		20		42%
Male		28		58%
Age (years)*			77 (40–86)
Acute Leukemia Myeloid
Primary		24		50%
Secondary		24		50%
	MRC	12		50%
	Chemotherapy	12		50%
European Leukemia Network
Favorable		2		4%
Intermediate		9		19%
Adverse		37		77%
Karyotype
Normal		24		50%
Monosomic		9		19%
Complex		15		31%
Mutations
FLT3		10		21%
IDH1		2		4%
IDH2		7		14%
NRAS		3		6%
KRAS		1		2%
NPM1		8		17%
TP53		13		27%
Treatment lines
First line		33		69%
Second line or more		15		31%
Baseline blood cells
Hemoglobin*		85.5 (62–124)
Platelets*		53 (6-356)
Neutrophil*		0.6 (0-10.2)
White blood cell*		2.8 (0.6–42.1)
Bone marrow blasts (%)
≤ 30%		17		38%
]30%; 50%]		10		22%
> 50%		18		40%
Median*			42 (8–98)
Baseline cytopenia grade ≥ 3
Anemia		12		25%
Thrombopenia		23		48%
Neutropenia		28		58%

*Median (Minimum-Maximum)

Treatment

Patients were treated with the combination of VEN and azacytidine, except for some patients for whom azacytidine was replaced by cytarabine (n = 3). The median duration of VEN administration during this period was 7 cycles [1–42]. At the beginning of treatment, 11 patients started at 400 mg for 28 days (Table 2). Another 26 patients started at 400 mg as well, but due to frail condition, the duration of treatment was reduced to 7 days in 4 patients, 14 days in 16 patients and 21 days in 6 patients. Posaconazole was used as antifungal prophylaxis in 7 patients; these patients benefited from dose adjustment as recommended by the RCP and received a dose of 100 mg of VEN [21]. Of the remaining 4 patients, 2 started at 200 mg and 2 at 100 mg without antifungal prophylaxis, based on clinician choice. The median interval between first and second treatment was 33.5 days.

Table 2.

Population protocols description

	Patients
VEN Protocol	N	Percentage
Initial dosage (mg)
100	2	4%
100*	7	15%
200	2	4%
400	37	77%
Initial duration (j)
7	5	10%
14	21	44%
21	10	21%
28	12	25%
Standard Protocol
400 mg for 28 days	11	23%
Start protocol modification
Dosage	4	8%
Duration	5	10%
Dosage and duration	1	2%
Cycle VEN duration
[1–3]	23	48%
[4–9]	15	31%
> 9	10	21%

* Combined with posaconazole

Pharmacokinetics of VEN in AML Patients.

A total of 105 blood samples were analyzed for VEN concentration (mean 2.2 samples per patient). The Ctrough in patients treated with the standard dose of 400 mg VEN was 1951 ± 1394 ng/ml (CV = 71%). The difference in VEN concentration between VEN 400 mg and VEN 100 mg plus posaconazole was not statistically significant (p = 0.24; Pairwise test) (Fig. 1).

Fig. 1.

Comparison of Ctrough as a Function of the Administered Dose

The dose-related Ctrough was then examined as a function of the administered dose (Appendix Fig. 1). There were no significant differences between the mean concentrations of 6.57, 5.58 and 4.88 µg/L for the 100, 200 and 400 mg dose related Ctrough, respectively. There was a statistically significant difference between VEN 100 mg plus posaconazole and VEN 400 mg (p = 0.06).

The inter-individual pharmacokinetic variability of Ctrough at first treatment was 68%, and in 12 patients with at least 3 assays, the observed intra-individual variability was 39% (Appendix Tables 1 and 2). The inter-individual variability was lower when VEN was co-administered with posaconazole (24%) (Fig. 2).

Fig. 2.

Violin plot showing Cmin distribution in the first dose

Adverse events

Myelosuppression was the most common AEs. All patients received at least one transfusion, with an average of 6 red cell transfusions and 3 platelet transfusions. In the first cycle, for protocol 28 days (n = 11), grade 3 and 4 hematologic side effects, including neutropenia, thrombocytopenia and anemia were 69%, 23% and 27% respectively (Appendix Table 3). In term of toxicities, we observed no difference between protocols 14 and 28 days. The rate of febrile aplasia did not differ between the 2 groups (p = 0.7; Fisher’s exact test). Finally, there were no differences in delayed cures between the protocols (p = 0.5; Fisher’s exact test).

On the other hand, when based on concentration values, the low concentration group (concentration below < 1818 ng/mL) had fewer delayed cures (0.9 ± 1.4) than the high concentration group with concentration above > 1818ng/mL (4.2 ± 5.2) (p = 0.092; Welch two-sample t-test).

Exposure-Response relationships

In the group of patients treated with VEN 400 mg, we identified 2 different protocols in terms of treatment duration. Twenty-six patients were evaluable for all cycles. Seventeen patients were treated with the 14-day VEN protocol and 9 with the 28-day VEN protocol at a dose of 400 mg once daily. The 14-day group received an average of 8 cycles, while the 28-day group received an average of 4.5 cycles. A total of 50 blood samples were analysed. When comparing the concentrations and observing a variability according to the protocol used, the mean Ctrough was 1503 +/- 966 ng/ml in the 28-day VEN protocol (CV = 64%) and 2330 ng/ml +/- 1675 (CV = 72%) in the 14-day VEN protocol. The difference in concentration between the two groups was statistically significant (p = 0.032; Welch two-sample t-test). Patients on the 14-day VEN protocol had fewer bone marrow blasts (9 ± 10%) than patients on the 28-day protocol (19 ± 9%) (p = 0.03; Welch two-sample t-test). Maximally Selected Rank Statistics method show a survival benefit for patients with a threshold concentration of 1818 ng/mL in cycle 1 (p = 0.36) (Appendix Fig. 2).

We focused on patients with more than 3 samples (n = 12) for a total of 60 samples (Fig. 3). We found that a high mean concentration (µ = 2462 ng/mL) was associated with fewer deaths than a low mean concentration (µ = 1280 ng/mL) (p = 1.16e-04; Welch two-sample t-test). We found extended OS in patients with higher concentrations.

Fig. 3.

Disease related to concentration

CR, CRi and PD were 52%, 10% and 38% respectively. With a median follow-up of 7.3 months (1.0-40.7) as the majority of patients died due to relapse, progression, or infection leading to treatment discontinuation. In the overall population, 1- and 2-year OS were 51% and 36.4%, respectively. The median OS for the entire cohort was 17.8 months (n = 45; 3 lost to follow-up) (Fig. 4).

Fig. 4.

Overall survival

Survival curves for the 14- and 28-day VEN protocols show that, the median survival for the 14-day protocol was not reached, whereas it is reached with the 28-day protocol (5.4 months, p = 0.0072) (Fig. 5).

Fig. 5.

Overall survival versus protocol

Survival curves by protocol and 1818 ng/mL cutoff show that, the median survival for the 14-day protocol was not reached, whereas it is reached with the 28-day protocol (p = 0.033) (Fig. 6).

Fig. 6.

Overall survival by protocol and 1818ng/mL cutoff. OS red corresponds below threshold protocol 14 day. OS green corresponds below threshold protocol 28 day. OS blue corresponds above threshold protocol 14 day. OS purple corresponds above threshold protocol 28 day

Figure 7 shows the risk as a function of protocol. The 14-day protocol results in an HR of approximately 0.20 compared to 28 days. The result is significant for the 14-day protocol (p = 0.009). The risk is multiplied by 0.20, which corresponds to an 80% reduction in risk when using the 14-day protocol.

Fig. 7.

Hazard ratio as a function of protocol

Discussion

For elderly patients with AML who are ineligible for intensive chemotherapy, treatment with VEN plus azacytidine is now recommended as the new standard of care. Observations of the use, management and safety of VEN in real-world patients have shown that the dose and duration of use differ from those in clinical trials. Compared to the recommendations of DiNardo et al., only 11 patients were started with the standard protocol, i.e. at a dose of 400 mg for 28 days. Taking this into account, as well as the frailty of the patients, the clinicians at our institute determine the standard protocol a priori.

This explains why both duration and dose have been modified to anticipate toxicities and preserve quality of life as much as possible. However, there were few data on the effectiveness of these changes in dose and duration in the population. We observed significant variability in exposure levels, both between patients (about 70%) and within patients (about 40%). This calls for the implementation of Therapeutic Drug Monitoring to optimize VEN, particularly because of exposure-response relationships on efficacy and toxicity endpoints. A few studies in the literature have focused on the pharmacokinetics of VEN [11, 12, 22]. Given the pharmacokinetic characteristics of VEN and the significant limitations of studying exposure by AUC, we chose to focus here on residual concentrations which are much more likely to be implemented in real-world setting because they do not require PK parameters identification to derive AUC from individual clearance. Our data show that our patients are more exposed, with a mean Ctrough of 1951 ± 1394 ng/ml compared to a mean Ctrough of 1018 ± 729.4 ng/mL in the Asian patients of Yang et al. [12]. However, another publication comparing the exposure of Asians and non-Asians shows no difference in exposure [22]. As already shown by Yang et al., our study confirmed that VEN pharmacokinetics shows a high inter-individual variability of > 68%, with Ctrough ranging from 150 to 7743 ng/mL. In addition, repeated dosing in different patients showed intra-individual variability of VEN of 39%. Studies have shown that diet has a significant effect on the bioavailability of VEN [23]. Here patients were taking VEN with meals according to the prescribing information, limiting the effect of food intake. In addition, pharmaceutical analysis and search for drug-drug interactions helped to limit causes of variability. Our study shows a slight difference in concentration between patients treated and not treated with posaconazole, which is consistent with the dose recommendation in the study by Agarwal et al. There was less variability (24%) when VEN was combined with posaconazole at a dose of 100 mg than at 400 mg. Which is expected given the recommended 75% dose reduction with the enzyme inhibitor. One hypothesis is that there is either a high degree of variability in the absorption phase or a smoothing of variability at the hepatic level with posaconazole, which inhibits CYP3A4 thus reducing its own variability.

One of the main issues with VENAZA is that patients are at risk for hematologic toxicities, which leads to additional infections requiring further adjustment in dose or duration of treatment, with some patients having to discontinue treatment for 7 to 14 days [4]. These dose or duration adjustments may apply to patients with myelodysplastic syndromes, for whom an increase in cytopenias has been observed. In our cohort, MDS patients accounted for 19% of the VEN14 group and 37% of the VEN28 group. Although the VEN14 group was found to be more exposed than the VEN28 group, our data showed no difference in toxicity between the two groups. Brackmann et al. found no association between exposure and toxicity [24]. Additional data support this, showing that changing the duration and dose of VEN does not alter patient outcomes [5]. In this real-world study, we first demonstrated a Ctrough threshold associated with hematological response.

In the DiNardo et al. cohort, 42% of patients had grade 3–4 neutropenia, whereas in our cohort, 69% had grade > 3 neutropenia (with 37% of febrile aplasia) after the first treatment all protocols combined. Median survival was 17.8 months, in line with published results. We hypothesize that the median survival of 5.4 months in the 28-day group is explained by the high proportion of secondary AML compared to the overall population. The complete response rate (CR + Cri) in our study was 62%, which was comparable to data in the literature [4].

In practice, we have observed a reduction in the duration of VEN treatment due to the occurrence of severe toxicities. According to various studies, the duration of VEN treatment is not predictive of response [5, 7]. However, our results show an association between plasma exposure and response. Looking at plasma exposure as a function of treatment duration, we found a significant difference in plasma concentrations after prolonged continuous exposure for 28-day VEN compared to exposure for 14-day VEN. Furthermore, when looking at survival as a function of 1818ng/mL concentration, the best protocol would be 400 mg/14d if the concentration is above the cut-off. However, due to the small number of patients and the fact that the trajectories are not the same, we must remain cautious.

This observation, correlated with an 89% response rate in the subset of patients treated with VEN 400 mg/14 days at the end of cycle 1, paves the way for this alternative regimen. In addition, a significant difference in overall survival was seen in the 14-day VEN protocol compared to the 28-day VEN protocol (p = 0.0072).

One hypothesis we propose to explain this decrease in plasma concentration with chronic use of the VEN 28 protocol is an increase in the expression of intestinal efflux pumps, leading to a reduction in absorption [25].

Given the lower response rates with the 28-day administration protocol and its reported association with similar toxicities to the 14-day VEN protocol, our results support that intermittent/discontinuous administration of VEN could improve the toxicity-efficacy ratio. The main limitation of our study is its single-center nature, and the fact that consequently only a small number of patients were monitored. It will thus be necessary to confirm our findings with a multi-centric, larger study. Because exposure-response relationships have been found both on efficacy and safety endpoints, this could pave the way for PK-guided dosing of VEN, so as to ensure an optimal toxicity-efficacy ratio at bedside.

Conclusion

Our study showed marked inter- and intra-individual variability with VEN. We cannot conclude that the 400/14-day protocol is superior to the 400/28-day protocol due to the small number of patients. However, in view of the results, a treatment duration of 14 days does not appear to be unfavorable in terms of patient survival. The article examines the pharmacokinetics of VENAZA in real-world patients, finding a link between exposure and patient response rate that suggests the clinical relevance of TDM in AML patients treated with VEN. At present, the duration of the VENAZA protocol varies significantly between centers and patients. Our primary objective was to ascertain whether a protocol such as 400/14-day would yield superior outcomes for patients in terms of response.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Loïc Osanno performed, designed the research and wrote the paper Lucy Brocque performed the research Carla Delpech, Jacques Chiaroni, Caroline Izard, Mathilde Dacos, Chems Djezzar contributed to the data Laurent Bourguignon analysed the data Laure Farnault, Julien Colle, Pauline Roche, Régis Costello, Caroline Solas, Joseph Ciccolini, Thomas Cluzeau read the research paper and made corrections Geoffroy Venton and Raphaëlle Fanciullino designed the research study, wrote the paper and made corrections.

Funding

No funding declared.

Data availability

No datasets were generated or analysed during the current study.

Competing interests

The authors declare no competing interests.