Concurrent use of nucleoside reverse transcriptase inhibitors and purine analogs is associated with worse survival for patients with myeloid neoplasms: A single-center retrospective analysis

Abstract

Abstract Background: Nucleoside reverse transcriptase inhibitors (NRTIs) are integral to therapy for human immunodeficiency virus (HIV) and hepatitis B virus (HBV) infections. Data from pre-clinical work and case series suggest that these medications mimic the structure of natural purines and thus diminish the activity of purine analogs (i.e., cladribine or fludarabine), which has implications for treatment for patients with concurrent HIV/HBV infection and acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS). Therefore, we sought to investigate outcomes for patients at our institution who received purine analogs for treatment of AML/MDS with concurrent NRTI treatment for either HIV or HBV. Methods: Using an IRB-approved institutional database, we retrospectively reviewed all patients treated for high-grade neoplasms (>10% blasts) at Fred Hutchinson Cancer Center from 2008-2025 and identified 59 patients as having a concurrent diagnosis of HIV (n=10) or HBV (n=49) infection. Using direct review of electronic medical records, these 59 patients were divided into those who received a purine analog and an NRTI at the same time (“drug-drug interaction” group), and those without this drug interaction. Patients in the latter group either 1) received a chemotherapy regimen lacking a purine analog or 2) were not treated with an NRTI. Statistical analysis included descriptive analysis, Kaplan-Meier modeling, and multivariable analyses for event-free survival (EFS) and overall survival (OS), with the multivariable analysis constructed using treatment-related mortality (TRM) score. Results: In this cohort, 21 patients (30%) were treated with both a purine analog (17 with cladribine, 4 fludarabine) and an NRTI (9 with entecavir, 6 emtricitabine/tenofovir, 4 lamivudine, 2 tenofovir monotherapy), with 13 of these patients receiving the purine analog for induction therapy. Amongst the 38 patients who did not have the drug-drug interaction, for induction therapy, 22 (58%) received cladribine, cytarabine, G-CSF, and mitoxantrone (CLAG-M), 10 (26%) received a hypomethylating agent, 5 (13%) received cytarabine and an anthracycline (7+3), and 2 (5%) received fludarabine, cytarabine, G-CSF, and idarubicin (FLAG-IDA). Patients with and without the drug-drug interaction had similar diagnoses [n=14 (67%) vs n=33 (87%) for AML, p = 0.17], sex [n=5 (24%) vs n=8 (21%) for female, p=0.46], performance status [n=15 (71%) vs n=28 (74%) for ECOG 0-1, p=0.47], and disease risk (n=11 (52%) vs n=22 (58%) for ELN2022 adverse risk, p=0.5]. TRM score also did not differ between groups (overall median TRM score was 5.1, range 0.6-43.2), corresponding approximately to a 5% risk of death in the first 28 days after initiation of induction chemotherapy. In the subset of patients with HIV, no differences in viral loads or CD4 counts were identified between the two groups. Patients with and without drug-drug interactions achieved similar rates of remission and had a similar event-free survival (EFS) and overall survival (OS). However, multivariable time-dependent regression found patients with drug-drug interactions had a significantly worse EFS (HR 2.14, 95% CI 1.12-4.08, p = 0.021) and trended towards worse OS, though this comparison is limited due to a small number of OS events (HR 2.02, 95% CI 0.96-4.26, p = 0.064). Conclusions: Among patients with high-grade myeloid disease and HIV+ or HBV+ infection, concomitant therapy with purine analogs and NRTIs was associated with significantly worse survival compared to similar patients without this drug-drug interaction. These findings suggest that patients receiving NRTIs for HIV and HBV should be considered for alternative antiviral therapy if purine analog-based chemotherapy is planned. Next steps include developing larger prospective studies in this patient population, possibly through multi-center collaboration, to better assess the strength of this effect and determine whether similar interactions occur in the treatment of other malignancies.