Impact of venetoclax trough levels on safety and efficacy in the treatment of acute myeloid leukemia

Abstract

Venetoclax-azacitidine chemotherapy is a key treatment for older or medically unfit patients with acute myeloid leukemia (AML). As venetoclax is metabolized via cytochrome P450 3 A, dose adjustments are required when combined with antifungal agents. However, data on venetoclax concentrations and their impact on safety and efficacy remain limited. This study analyzed the association between venetoclax trough levels and treatment outcomes in 152 AML patients (median age: 70 years). The median trough level was 1,518 ng/mL (range: 60–11,328 ng/mL), with significant interindividual variability, even after adjusting for antifungal use. High venetoclax trough levels were significantly associated with elevated creatinine and total bilirubin levels. Patients with trough levels below 1857.3 ng/mL had lower rates of hematologic toxicity during the first treatment cycle (92.9% vs. 100%, P = 0.041). In the second cycle, hematologic toxicity was lower in patients with concentrations below 1,299.3 ng/mL (68% vs. 90%, P = 0.009). In the treatment-naïve group, trough levels above 1857.3 ng/mL were associated with a higher composite complete remission rate (77.3% vs. 47.8%, P = 0.042). These findings highlight the need for venetoclax dose optimization based on drug levels to improve both safety and efficacy in AML treatment.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38587-4.

Introduction

Venetoclax is an orally administered pharmaceutical agent that exerts its antitumor effects through the selective inhibition of the B-cell lymphoma-2 protein, an apoptosis suppressor protein overexpressed in tumor mitochondria^1–5. Initially, venetoclax was employed for the treatment of chronic lymphocytic leukemia^2,6, and its indications have been expanded to include acute myeloid leukemia (AML)^7–9. Currently, venetoclax combined with a hypomethylating agent (HMA) is a key treatment for older or medically unfit patients with AML^5,7–11. Furthermore, the indications for venetoclax administration continue to expand^12–15.

However, the clinical use of venetoclax is challenging because of its susceptibility to drug interactions^16,17. This challenge arises because venetoclax is primarily metabolized by CYP3A, a key enzyme involved in the metabolism of several medications^16–18. In particular, concomitant administration of antifungal agents may result in significant drug-drug interactions, potentially affecting venetoclax blood levels and, consequently, treatment safety and efficacy^17,19,20. In the Asian population, including Japanese individuals, venetoclax exposure is reported to be approximately 67% higher than in non-Asians^18,20,21. Additionally, CYP3A poor metabolizers are relatively prevalent in the Asian population^20,21. Consequently, the Japanese regulatory authorities recommend a more stringent dose reduction—specifically to 50 mg once daily when combined with strong CYP3A inhibitors—to ensure safety while maintaining efficacy. Given these pharmacogenetic differences, characterizing the relationship between drug levels and outcomes in this specific population is of clinical importance. However, evidence regarding the association among venetoclax blood levels, safety, and efficacy remains limited. Available evidence is primarily derived from steady-state area under the curve (AUC) measurements, such as the 24-h AUC (₀–₂₄ AUC)^1,18,20,22, which are difficult to measure in clinical practice. Adaptive clinical monitoring is crucial, particularly in settings requiring prompt drug adjustments when venetoclax levels are suspected to influence treatment response and adverse effects. A pharmacokinetic study conducted by our group demonstrated a strong correlation between 0 and 24 AUC and venetoclax trough levels, suggesting that trough level measurements may serve as an alternative to AUC-based assessments (under submission). Accordingly, the present multicenter study examined the significance of venetoclax trough levels in actual clinical practice.

Material and methods

Data collection

The study included patients who received venetoclax treatment at Kyoto University Hospital and its collaborating institutions between June 2020 and July 2023 and who provided informed consent. Venetoclax blood concentrations were quantified using liquid chromatography–tandem mass spectrometry as previously reported²³. The inter-assay precision and accuracy required for venetoclax quantification were 15% and 100 ± 15%, respectively (Supplemental methods). Basic information, including age, sex, and diagnosis, was collected at the time of venetoclax trough measurement. Additional information, including adverse effects and response to the initial three cycles of venetoclax-azacitidine therapy, was obtained using questionnaires completed by the respective centers for cases in which the diagnosis was AML and the patients were adults (aged ≥ 18 years). In instances where multiple venetoclax trough-level measurements were conducted, the values obtained from the initial trough-level measurements were utilized. Trough values were analyzed exclusively for cases in which the same venetoclax dose was administered continuously for a minimum of 5 day prior to trough measurement. Laboratory data, including serum creatinine, total bilirubin, AST, and ALT, were collected simultaneously with the venetoclax trough blood sampling to assess the concurrent organ function during treatment.

Inclusion and exclusion criteria

The inclusion criteria were as follows: adult patients (≥ 18 years) with newly diagnosed or relapsed/refractory AML who received venetoclax and azacitidine chemotherapy and had trough levels measured within the first three cycles of treatment. Additionally, patients received their first therapy between June 2020 and July 2023. Patients diagnosed with conditions other than AML and those treated with venetoclax monotherapy were excluded. Patients who did not receive the venetoclax dose as per the Japanese prescribing information—which mandates a reduction to 50 mg for strong CYP3Ai (voriconazole, itraconazole, and posaconazole) and 200 mg for moderate CYP3Ai (fluconazole)—were excluded from the analysis to ensure data consistency. Hematologic toxicity and venetoclax trough levels were assessed for each treatment cycle. To ensure the accuracy of the toxicity-exposure correlation, assessments were only conducted if the venetoclax dose remained consistent between the time of trough measurement and the cycle being analyzed. For the assessment of treatment response, we included patients who were not in complete remission (CR) at the initiation of venetoclax treatment and who received a constant dose throughout the observation period. A total of 207 questionnaires were distributed based on the information available at the time of enrollment. The research protocol was approved by the Kyoto University Graduate School and Faculty of Medicine, Ethics Committee. The registry and registration number of the study/trial was R2915. All patients provided informed consent. All methods were performed in accordance with the relevant guidelines and regulations. This study was performed in accordance with the Declaration of Helsinki.

Registration and trial number: This study was registered with the UMIN Clinical Trials Registry (UMIN-CTR), which is a primary register in the WHO International Clinical Trial Registry Platform. The registration number is UMIN000060501, and the date of first registration was 28/01/2026.

Venetoclax doses were adjusted according to Japanese regulatory approval: 50 mg once daily with strong CYP3Ai (voriconazole, itraconazole, and posaconazole), 200 mg with moderate inhibitors (fluconazole), and 400 mg without inhibitors (micafungin or no prophylaxis). These adjustments account for the increased venetoclax exposure typically observed in Asian individuals. Antifungal agents were continued for the entire duration of venetoclax administration for infection prophylaxis. Isavuconazole was not included in this cohort due to its late approval in Japan (December 2022).

Definition

Hematologic toxicity was defined if the hemoglobin level was less than 8.0 g/dL or anemia necessitating a transfusion was present, white blood cell count was less than 1000/mm³, neutrophil count was less than 500/mm³, or platelet count was less than 50,000/mm³. Grade ≥ 3 adverse events other than hematologic toxicity were reported in accordance with the Common Terminology Criteria for Adverse Events version 5.0. Hematologic toxicity was monitored throughout the entire duration of each treatment cycle, spanning from the initiation of venetoclax–azacitidine therapy until the commencement of the subsequent cycle. In the typical clinical practice of the 22 participating centers, complete blood counts were performed at least once or twice per week, and the toxicity assessment was based on the nadir values or the requirement for transfusions recorded during each cycle. A composite response was defined as CR or CR with incomplete blood count recovery¹⁰.

Endpoint

The primary objective of this study was to evaluate the correlation between venetoclax trough levels and overall hematologic toxicity. Secondary endpoints included the correlation between venetoclax trough levels and response rate. Adverse effects other than hematologic toxicity were also reported.

Statistics

This study evaluated patients’ pre-administration complications, liver function levels, creatinine levels, and antifungal medications co-administered with venetoclax. This study investigated the optimal cutoff values for trough levels and presence or absence of hematologic toxicity using receiver operating characteristic (ROC) curves. The association of trough levels with response and hematologic toxicity was analyzed using a logistic regression model. Furthermore, the patients were grouped into three according to the trisection of the trough values, and the groups were compared using Pearson’s chi-square test for categorical variables and Kruskal–Wallis rank test for continuous variables, as appropriate. Overall survival was estimated using the Kaplan–Meier method, and survival curves were compared using the log-rank test. Statistical significance was set at P < 0.05.

Result

Patient enrollment

A total of 207 questionnaires were collected from the participating centers. As shown in the CONSORT flow diagram (Fig. 1), patients were excluded based on predefined criteria: non-AML diagnoses (n = 8), regimens using combinations other than azacitidine (n = 11), trough measurements taken after the fourth cycle (n = 18), and dosing that did not adhere to the prescribing information regarding dose adjustments for concomitant CYP3A inhibitors (n = 18). Consequently, 152 Japanese patients were included in the final analysis. For toxicity analysis, 140 patients during cycle 1, 87 patients during cycle 2, and 46 patients during cycle 3 were included. For response analysis, 145 (95.4%) were either treatment-naïve or in non-CR at baseline. Among these, 110 patients (75.9%) were evaluable for their best treatment response within the first three cycles.

Fig. 1.

CONSORT Flow diagram.

Basic characteristics

Table 1 summarizes the patients’ basic characteristics. The median patient age was 70 years (range: 21–89 years). The median weight was 55.2 kg (range: 35.7–96.6 kg). The median body mass index was 21.1 kg/m² (range: 14.4–35.6 kg/m²). Of the patients included in the study, 89 were men (58.6%). The Eastern Cooperative Oncology Group performance status was 0–1 in 133 of the 152 patients (88.1%). Sixty-two patients (40.8%) were newly diagnosed and treatment-naïve. Seven patients (4.6%) received venetoclax treatment as a consolidation therapy. The remaining 83 patients (54.6%) had relapsed or were refractory, including 13 patients (8.6%) who underwent hematopoietic stem cell transplantation before the commencement of venetoclax-azacitidine therapy. According to the 2017 World Health Organization criteria, 64 patients (42.1%) were diagnosed with AML with myelodysplasia-related changes, 25 patients (16.4%) with AML with recurrent genetic abnormalities, 47 patients (30.9%) with AML/not otherwise specified, and 13 patients (8.6%) with therapy-related myeloid neoplasm.

Table 1.

Baseline characteristics.

	Median (range)
Age (years)	70 (21–89)
Weight (kg)*	55.2 (35.7–96.6)
Height (cm)*	162 (137.6–184.5)
Body mass index (kg/m 2 )*	21.1 (14.4–35.6)
Creatinine (mg/dL)†	0.79 (0.3–3.6)
T-bilirubin (mg/dL)*	0.72 (0.06–2.2)
AST (U/L)†	19 (7–63)
ALT (U/L)‡	17.5 (2–96)
	Number (%)
Male patient	89 (58.6)
ECOG performance status 0–1†	133 (88.1)
Disease Status
Treatment-naïve	62 (40.8)
Consolidation (in CR after prior therapy):	7 (4.6)
Relapsed or Refractory	83 (54.6)
Lines of treatment before VEN-based therapy†
Never treated	62 (41.1)
1	22 (14.6)
2	30 (19.9)
3–5	37 (24.5)
History of hematopoietic stem cell transplantation	13 (8.6)
Diagnosis
AML with MRC	64 (42.1)
AML with recurrent genetic abnormalities	25 (16.4)
AML, not otherwise specific	47 (30.9)
Therapy related myeloid neoplasm	13 (8.6)
Other types of AML	3 (2.0)
Gene mutations and cytogenetic aberrations
FLT3-ITD	18 (11.8)
TP53	15 (9.9)
Complex karyotype	25 (16.5)
Best response
Composite CR (CR/CRi)	78 (51.3)
MLFS	14 (9.2)
Non-CR	48 (31.6)
Missing	12 (7.9)

Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase; ECOG, Eastern Cooperative Oncology Group; CR, complete remission; AML, acute myelogenous leukemia; MRC, myelodysplasia-related changes; CRi, CR with incomplete blood cell recovery; MLFS, morphological leukemia-free state. Data missing in *three cases; †one missing; ‡two missing.

In the first cycle, the strong CYP3Ai were used for 40% of patients (n = 61), moderate CYP3Ai, fluconazole, were used for 52% (n = 78), and micafungin or no antifungal agents were used for 8% (n = 12). In the second cycle, the strong CYP3Ai were used for 40% of patients (n = 42), moderate CYP3Ai, fluconazole, were used for 54% (n = 57), and micafungin or no antifungal agents were used for 6% (n = 7). In the third cycle, the strong CYP3Ai were used for 32% of patients (n = 20), moderate CYP3Ai, fluconazole, were used for 51% (n = 32), and micafungin or no antifungal agents were used for 16% (n = 10). The duration of venetoclax administration in the first, second, and third cycles is shown in Supplemental Table 1.

Trough of venetoclax, dose of venetoclax, and concomitant antifungal agents

The median trough level was 1,518 ng/mL (range: 60–11,328 ng/mL) (Fig. 2). Of the total number of patients, 60 (39.5%), 79 (52.0%), and 13 (8.6%) received a dosage of 50 mg, 200 mg, and 400 mg venetoclax as per the prescribing information due to concomitant use of azole agents with CYP3Ai, respectively (Supplemental Table 2). The antifungal drugs used in the trough measurements were the strong CYP3Ai group; voriconazole (VRCZ, n = 31, 20.3%), itraconazole (ITCZ, n = 16, 10.5%), and posaconazole (PSCZ, n = 13, 8.6%), moderate CYP3Ai group; fluconazole (FLCZ, n = 79, 52.0%), and other group; micafungin (MCFG, n = 4, 2.6%) and no antifungal agents (n = 9, 5.9%).

Fig. 2.

Distribution of venetoclax trough levels.

Median trough levels (IQR) were 1,400 (958-2,019) ng/mL, 1,740 (1,098 − 2,897) ng/mL, and 1,243 (583-1,652) ng/mL for the strong CYP3Ai group, the moderate CYP3Ai group, and the other groups, respectively, and a notable decline in trough levels was evident in the strong CYP3Ai and no CYP3Ai groups compared with those in the moderate CYP3Ai groups (Table 2, P = 0.028). However, the coefficient of variation was smaller in the strong CYP3Ai group (0.509) than in the moderate CYP3Ai (0.721) and other groups (1.420). The strong CYP3Ai group was divided into three subgroups according to the antifungal agent. Median trough levels (IQR) were 1,137 (848-1,937) ng/mL, 1,834 (1,288-2,687) ng/mL, and 1,459 (906-2,171) ng/mL for voriconazole, posaconazole, and itraconazole, respectively. The trough tended to be low in the voriconazole and itraconazole group as compared to posaconazole group (P = 0.083).

Table 2.

Venetoclax trough levels according to Azole agents with CYP3A inhibition.

	Strong CYP3A inhibition	Moderate CYP3A inhibition	Others/No*	Pvalue
	N = 60	N = 79	N = 13
Trough by fungal drug (ng/mL), median (IQR)	1,400 (958-2,019)	1,740 (1,098 − 2,897)	1,243 (583-1,652)	0.028

IQR, interquartile range.

The patients were divided into three groups based on the trisection of trough levels. No statistically significant correlations were observed between aspartate aminotransferase or alanine aminotransferase levels and trough levels. However, significant increases in creatinine (P < 0.001) and total bilirubin levels (P = 0.035) were observed in the high-trough group (Table 3).

Table 3.

Patient characteristics according to the venetoclax trough levels.

	Low (n = 51)	Intermediate (n = 51)	High (n = 50)	P value
Trough (ng/mL, median, range)	848 (60–1,177)	1,520 (1,201–2155)	2,952 (2,175–11,328)	-
Venetoclax 50 mg	23 (45%)	25 (49%)	12 (24%)	0.033
200 mg	22 (43%)	22 (43%)	35 (70%)
400 mg	6 (12%)	4 (8%)	3 (6%)
Creatinine (mg/dL)	0.71 (0.3–3.62)	0.76 (0.43–1.54)	0.9 (0.47–2.06)*	< 0.001
T-bilirubin (mg/dL)	0.69 (0.25–1.80)*	0.80 (0.33–2.10)	0.87 (0.06–2.20)†	0.035
AST (IU/L)	17 (8–56)	20 (7–54)	20 (8–63)*	0.164
ALT (IU/L)	16 (2–96)	16 (4–95)*	21 (7–96)*	0.265

AST, aspartate aminotransferase; ALT, alanine aminotransferase. Creatinine/T-bilirubin/AST/ALT data from trough measurements are presented. *Missing data: one missing, †three missing.

Table 4.

Venetoclax dose adjustments according to concomitant antifungal agents.

CYP3A inhibition category	Antifungal agent	Venetoclax daily dose	Number of patients (n = 152)
Strong	Voriconazole, Itraconazole, Posaconazole	50 mg	60 (39.5%)
Moderate	Fluconazole	200 mg	79 (52.0%)
None / others	Micafungin or None	400 mg	13 (8.6%)

AML, acute myeloid leukemia; CR, complete remission.

Toxicity of venetoclax- Azacitidine therapy and venetoclax trough levels

Hematologic toxicity occurred at a high frequency during the initial cycle of chemotherapy (95.7%, n = 134/140), followed by the subsequent cycle (80.5%, n = 70/87) and third cycle (60.9%, n = 28/46). Trough levels are shown in the first, second, and third cycles according to the presence of hematologic toxicities (Fig. 3). The median (IQR) venetoclax trough levels for the absence and presence of hematologic toxicities were 1,584 (674-1,740) and 1,507 (1,032 − 2,440) ng/mL in the first cycle, 1,072 (884-1,499) and 1,573 (1,053 − 2,223) ng/mL in the second cycle, and 1,508 (1,098 − 2,292) and 1,466 (923-1,959) ng/mL in the third cycle, respectively. The association between hematologic toxicity and trough values was evaluated using ROC curves, and the cutoff trough levels were calculated as 1857.3 ng/mL for the first cycle (area under the curve (AUC), 0.71), 1299.3 ng/mL for the second cycle (AUC, 0.67), and 2975.5 ng/mL for the third cycle (AUC, 0.56). In the initial cycle of treatment, the incidence of hematologic toxicity was lower in the venetoclax group with concentrations of < 1857.3 ng/mL (n = 84) than in the group with concentrations of ≥ 1857.3 ng/mL (n = 56) (92.9% vs. 100%, P = 0.041). All 32 patients with concentrations of ≥ 2636.5 ng/mL exhibited hematologic toxicity. In the second cycle, the frequency of hematologic toxicity was lower in the group with venetoclax levels < 1299.3 ng/mL (n = 37) than in the group with levels ≥ 1299.3 ng/mL (n = 50) (68% vs. 90%, P = 0.009). All ten patients with venetoclax levels > 3,000 ng/mL exhibited hematologic toxicity. In the third cycle, the frequency of hematologic toxicity was lower in the group with venetoclax levels < 2975.5 ng/mL (n = 40) than in the group with levels ≥ 2975.5 ng/mL (n = 6) (58% vs. 83%, P = 0.227), although the difference was not statistically significant. Even after excluding patients who received venetoclax treatment in CR status as a consolidation therapy, the results were not changed (first cycle, 93% vs. 100%, P = 0.038; second cycle, 67% vs. 91%, P = 0.004; third cycle, 56% vs. 83%, P = 0.210). Then, we compared toxicity separately for the treatment-naïve, relapse and refractory groups according to each cutoff point. In the treatment-naïve group, the frequencies of hematologic toxicity were 93.6% versus 100% (P = 0.196) in the first cycle; 47% versus 88% (P = 0.006) in the second cycle; and 52% versus 75% (P = 0.396) in the third cycle, respectively, for the low and high cutoffs. In the relapse and refractory group, the incidence of toxicity was 92% versus 100% (P = 0.196) in the second cycle and 81% versus 96% (P = 0.125) in the third cycle, respectively. The frequency of hematologic toxicity was generally higher in the relapse and refractory group.

Fig. 3.

Distribution of trough levels according to hematologic toxicity. Box plots were superimposed to the distribution of trough levels.

These associations between the first and second cycles and trough levels were also shown in the logistic regression model (Fig. 4). A reduction in the incidence of reported hematologic toxicities was observed for each successive cycle (Supplemental Table 3). Further details regarding adverse events of grade ≥ 3 can be found in Supplemental Table 3.

Fig. 4.

Logistic regression model predicting the probability of toxicity in the first and second, cycles and response to venetoclax-azacitidine therapy.

Response of venetoclax-azacitidine therapy and venetoclax trough levels

Regarding treatment response, among the 145 non-CR or untreated patients at the commencement of venetoclax-azacitidine treatment, 78 (51.3%) achieved composite CR (Table 1). For the treatment-naïve and relapse/refractory group, 40 (64.5%) and 38 (45.8%) achieved composite CR, respectively. The relation between trough level and response rate was evaluated in 110 patients who received the same target dose of venetoclax throughout the observation period. The median (IQR) trough level for patients who achieved composite CR was 1,485 (1,044 − 2,292) ng/mL, whereas the median (IQR) trough level for patients who did not achieve composite CR was 1,397 (851-2,176) ng/mL (Fig. 5A). Among the treatment-naïve group, the median (IQR) trough level for patients who achieved composite CR was 1,764 (1,073 − 2,235) ng/mL, whereas the median (IQR) trough level for patients who did not achieve composite CR was 1,313 (1,072 − 1,937) ng/mL (Fig. 5B). Among the relapse and refractory group, the median (IQR) trough level for patients who achieved composite CR was 1,330 (943-2,893) ng/mL, whereas the median (IQR) trough level for patients who did not achieve composite CR was 1,553 (782-2,177) ng/mL (Fig. 5C). The association is also shown based on the logistic regression model (Fig. 4). The analysis of the ROC curve did not yield an optimal cutoff point, and no statistically significant correlation was observed between venetoclax trough levels and response (P = 0.36) among patients with AML who were not in CR. However, if we focused on the treatment-naïve group, the significant cutoff trough levels were calculated as 1430.4 ng/mL (AUC, 0.66). The incidence of composite CR was lower in the venetoclax group with concentrations of < 1430.4 ng/mL (n = 23) than in the group with concentrations of ≥ 1430.4 ng/mL (n = 22) (composite CR rate; 47.8% vs. 77.3%, P = 0.042).

Fig. 5.

Distribution of trough levels according to treatment response for (A) total patients, (B) treatment-naïve patients, and (C) relapsed/refractory patients. Box plots were superimposed to the distribution of trough levels. CR, complete remission; CRi, CR with incomplete blood cell recovery; MLFS, morphological leukemia-free state.

Overall survival and venetoclax trough levels

As part of our secondary analysis, we examined the impact of venetoclax trough levels on overall survival for 146 patients with available data. We used the same cutoff level (1,857.3 ng/mL) as in the first-cycle analysis of hematologic toxicity. However, we did not observe any statistically significant differences in overall survival based on venetoclax trough levels (Low vs. high; 96% vs. 89% at 60 days and 80% vs. 63% at 180 days, respectively; log-rank test, P = 0.101).

Discussion

This study aimed to examine the association between venetoclax trough levels and the safety and efficacy of venetoclax-azacitidine treatments for AML. In the present study, there was significant interindividual variability in trough levels, even after accounting for the potential impact of CYP3A inhibition due to concurrent antifungal medications. High bilirubin and creatinine levels were significantly associated with high trough levels. In this study, higher venetoclax trough levels were significantly associated with an increased risk of severe hematologic toxicity, particularly during the first and second cycles. Regarding treatment response, there was no significant association between venetoclax trough levels and the composite CR rate in the overall cohort. However, higher venetoclax trough levels were associated with a higher composite CR rate specifically in the treatment-naïve group. These findings suggest the potential importance of monitoring venetoclax trough levels to optimize both clinical response and management of hematologic toxicities.

Venetoclax therapy was indicated not only for newly diagnosed AML but also for high-risk patients with induction failure, post-chemotherapy relapse, posttransplant relapse, and prior to planned transplantation. The results of this retrospective study reflect the real-world use of venetoclax-azacitidine regimens for AML across participating multicenter centers. The present study demonstrated a notable variation in blood venetoclax levels, even when the recommended doses of venetoclax were based on the level of CYP3A inhibition by antifungal drugs. This suggests that, regardless of the use of potent CYP3A inhibitor antifungal drugs, it may be challenging to anticipate venetoclax trough levels. Another finding was that the mean trough levels at the recommended dose of venetoclax were reasonable, even though the dose was reduced to one-eighth in the strong CYP3A group. Although the mean trough level was lower in the strong CYP3A group, the coefficient of variation was notably lower. Regarding hepatic and renal function, bilirubin and creatinine levels were significantly high in the high-trough level group. Therefore, caution should be exercised in cases of impaired hepatic and renal function. Significant differences in trough levels and hematologic toxicities were observed in the initial two cycles of treatment. Further, higher trough levels were associated with the better CR rate in the patient who received venetoclax-based regimen as the treatment-naïve therapy. Based on the findings of this study, although preliminary, maintaining venetoclax trough levels below the range of approximately 1,400 to 1,800 ng/mL may reduce the risk of severe hematologic toxicity while maintaining a favorable treatment response. Therapeutic drug monitoring during the early phase of treatment, such as the first or second cycle, may be beneficial for dose optimization. However, given the retrospective nature of our analysis and the variability in dosing durations, these thresholds should be validated in future prospective trials to establish standardized clinical guidelines. It is important to acknowledge the limitations of this study to fully understand the implications of these findings. The specific criteria for adjusting the duration of venetoclax administration and the exact timing of antifungal agent discontinuation in subsequent cycles were left to the discretion of individual treating physicians. Although most of the patients continued to use the same antifungal agents in the subsequent cycles, the adjustments may have influenced the observed trends in hematologic toxicity across successive cycles. The full details of venetoclax administration duration, azacitidine dose duration, and reduction were not adequately provided in the present study, which could potentially weaken the actual effect of venetoclax trough levels on safety and efficacy. Furthermore, it should be noted that in patients who are not in remission, hematologic toxicity may be influenced by the underlying leukemia itself. Although our subgroup analysis confirmed a consistent trend between high trough levels and toxicity in non-CR patients, the baseline disease activity remains a confounding factor that is difficult to isolate in real-world clinical data.

Interestingly, our data revealed that while venetoclax trough levels tended to increase in subsequent cycles, the incidence of severe hematologic toxicity paradoxically decreased. This phenomenon may be explained by the recovery of normal hematopoiesis in responding patients. In cycle 1, the bone marrow is typically occupied by leukemic blasts and has a very limited hematologic reserve, making patients highly susceptible to even moderate drug exposure. As patients achieve remission (CR/CRi) and the tumor burden decreases, the recovery of normal marrow function may enhance tolerance to higher venetoclax concentrations in later cycles. Therefore, while stringent therapeutic drug monitoring is particularly critical during cycle 1 to prevent early severe complications, the ‘optimal’ trough range may shift as the treatment transitions from the induction to the consolidation phase, reflecting the changing resilience of the bone marrow.

As mentioned in the Introduction, Asian populations exhibit higher venetoclax exposure, which led to the adoption of the 50 mg dose for strong CYP3A inhibitors in this study^18,20,21. Our results demonstrated that although this dose resulted in lower mean trough levels, it provided the most consistent exposure with the smallest coefficient of variation. This finding suggests that the Japanese regulatory dose-reduction strategy effectively stabilizes drug concentration in patients receiving potent CYP3A inhibitors.

In conclusion, there were significant interindividual variations in venetoclax trough levels, and decreased liver and kidney function might have contributed to the high venetoclax blood levels. Furthermore, elevated blood venetoclax levels may increase the risk of hematologic toxicity. Consequently, actual measurements may facilitate safe and effective management of venetoclax treatment, thereby avoiding unexpected fluctuations in drug levels. As venetoclax chemotherapy is increasingly incorporated into various cytotoxic regimens, objective measurement of blood venetoclax levels may be essential to ensure safe and effective treatment. It is imperative that prospective studies be conducted in homogeneous populations to gain a deeper understanding of this subject.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization and design of the study: H.Ha. and J.K. Statistical analysis: H.Ha. and J.K. Data acquisition: T.Y., T.I., M.S., Y.M., T.Ta., M.W., Y.U., Y.Ts., K.I., S.F., T.To., K.Yo., H.Hi., H.K., M.T., S.W., H.Y., T.Ki., K.Mu., Y.Ka., E.K., S.K., R.Y., K.Mi., D.N., A.I., T.Ka., Y.O., T.S., C.M., T.Kit., K.Ya., A.Y., T.Kaw., S.N., R.T., T.Te., and A.T.K. Interpretation of data: H.Ha., T.Y., J.K., T.I., M.S., Y.M., T.Ta., M.W., Y.U., Y.Ts., K.I., S.F., T.To., K.Yo., H.Hi., H.K., M.T., S.W., H.Y., T.Ki., K.Mu., Y.Ka., E.K., S.K., R.Y., K.Mi., D.N., A.I., T.Ka., Y.O., T.S., C.M., T.Kit., K.Ya., A.Y., T.Kaw., S.N., R.T., T.Te., and A.T.K. Drafting of the manuscript: H.Ha. and J.K. Revision of the manuscript: H.Ha., T.Y., J.K., T.I., M.S., Y.M., T.Ta., M.W., Y.U., Y.Ts., K.I., S.F., T.To., K.Yo., H.Hi., H.K., M.T., S.W., H.Y., T.Ki., K.Mu., Y.Ka., E.K., S.K., R.Y., K.Mi., D.N., A.I., T.Ka., Y.O., T.S., C.M., T.Kit., K.Ya., A.Y., T.Kaw., S.N., R.T., T.Te., and A.T.K. Final approval of the manuscript: All authors.

Data availability

The data that support the findings of this study may be available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Disclosure

Financial disclosure: This work was supported in part by JSPS KAKENHI Grant Number 24K11515 (JK) and the Takeda Science Foundation (JK).