Azacitidine Versus Decitabine for the Treatment of Myelodysplastic Syndrome: A Propensity Score-Matched Retrospective Cohort Study Using Trinetx

Introduction

Azacitidine (AZA) and decitabine (DAC) are hypomethylating agents (HMAs), widely employed as monotherapy for myelodysplastic syndrome (MDS) either as primary treatment or as a bridge to transplant. There is a lack of studies in the literature comparing the efficacy and safety of AZA versus DAC for the treatment of MDS. The clinical choice between these two agents remains debated. This propensity-matched retrospective cohort study aimed to compare the efficacy (i.e overall survival) and safety (i.e., risk of infection or progression to Acute Myeloid Leukemia [AML]) between DAC and AZA for the treatment of MDS. The efficacy of the two agents was also compared in allogeneic hematopoietic stem cell transplantation (alloHSCT) patients.

Methods

This multicenter retrospective cohort study included patients with a diagnosis of MDS in the TriNetX Network, a database of deidentified electronic medical record data from 89 large healthcare organizations with over 100 million patient records. Patients included had a diagnosis of MDS and received either AZA or DAC for treatment. The primary outcomes analyzed were overall survival, survival in alloHSCT recipients, progression to AML, and risk of infection (bacterial, fungal, COVID-19). Kaplan-Meier analysis, hazard ratios (HR), risk ratios (RR), and 95% confidence intervals (CI) were used to assess the primary outcomes.

Results

Among 9,907 eligible patients with MDS who received either AZA or DAC for treatment, 6,503 patients received AZA and 3,404 patients received DAC. Following propensity score matching for demographics, laboratory values, and body mass index (BMI), there were 3,386 patients in each cohort.

In the AZA cohort, the mean current age was 74.5 years old, 74.5% were White, 7.6% were Black, 3.0% were Hispanic or Latino, 3.8% were Asian, and 59.9% were male. Comparably, in the DAC cohort, the mean current age was 74.7 years old, 73.8% were White, 8.1% were Black, 3.1% were Hispanic or Latino, 3.8% were Asian, and 60% were male.

AZA showed a statistically significant reduced overall risk of death and superior overall survival probability compared to DAC at the primary endpoint (442 days vs. 390 days median survival, 30.05% vs. 28.39% survival probability at 3 years; HR, 0.910 [95% CI, 0.853-0.971]). There was no statistically significant difference in 3-year survival in patients who underwent alloHSCT between the AZA and DAC cohorts (71.87% vs. 71.55% survival probability; HR, 1.010 [95% CI, 0.907-1.125]).

There was no statistically significant difference in progression to AML between the AZA and DAC groups (RR, 0.971 [95% CI, 0.876 -1.106]).

The overall risk of infection was significantly lower amongst patients receiving treatment with AZA compared to DAC (RR, 0.896 [95% CI, 0.837-0.960]). Specifically, the risk of bacterial infections was lower amongst patients receiving AZA versus DAC (RR, 0.901 [95% CI, 0.820-0.989]). AZA demonstrated statistically significant increased survival in patients with infection compared to DAC (454 days vs. 288 days median survival, 34.75% vs. 27.77% survival probability at 3-year; HR, 0.770 [95% CI, 0.700-0.846]). The risk of fungal (RR, 0.910 [95% CI, 0.810 -1.023]), intestinal infection (RR, 1.013 [95% CI, 0.863 -1.189]) or COVID-19 infection (RR, 0.989 [95% CI, 0.879 -1.112]) between the two groups was comparable.

Discussion and Conclusions

HMAs are common treatment options for high-risk MDS. Both AZA and DAC have shown superior outcomes when compared to supportive care alone; however, the two agents have not been compared directly in head-to-head randomized control trials. Our multicenter propensity score-matched retrospective cohort study found that AZA showed a statistically significant greater overall survival probability and reduced risk of death compared to DAC. Additionally, DAC showed a higher risk of bacterial infection and a significantly lower survival probability amongst patients with infection when compared to AZA.

This study demonstrates that the choice of HMA for the treatment of MDS may have significant impact on patient outcomes, while in the past they have been utilized nearly interchangeably. If these agents are used for bridging to alloHSCT, then both agents can be used interchangeably based on our data. Patients who received alloHSCT had the overall highest survival probability, demonstrating that alloHSCT remains the best potentially curative method of treatment.

No relevant conflicts of interest to declare.