Long-term survival of older patients with MDS treated with HMA therapy without subsequent stem cell transplantation

TO THE EDITOR:

The approval of the hypomethylating agents (HMAs) azacitidine and decitabine for therapy of myelodysplastic syndromes (MDS) has been celebrated by the MDS community as a major milestone in improving the clinical outcomes of patients.^1,2  Although both drugs lead to hematologic responses and improvement in quality of life, only azacitidine prolonged overall survival (OS) among patients with International Prognostic Scoring System (IPSS) intermediate-2 and high risk categories (ie, higher-risk MDS) compared with conventional care regimens (CCR) in a randomized trial (AZA-001).¹  With a median follow-up of 21.1 months, azacitidine prolonged OS (24.5 vs 15 months, P = .0001) and doubled the 2-year OS probability (50.8% vs 26.2%) compared with CCR.¹  Allogeneic stem cell transplantation (alloSCT) is recommended for eligible higher-risk MDS patients as the only potentially curative intervention.³  For the selected minority of patients undergoing alloSCT, 30% to 40% long-term OS has been reported with relapse being the leading cause of death.⁴ 

Although the 2- to 3-year outcomes with HMAs for higher-risk MDS patients who do not undergo alloSCT are well described in clinical trials and real-life analyses,^1-4  long-term survival prospects and whether there is a “tail” at the end of the survival curve is not clear. Most published trials of HMAs among MDS patients have limited duration of follow-up, rendering the 5-year OS rates unreliable. Furthermore, many of the retrospective analyses of HMA use among MDS patients did not observe and/or censor for alloSCT and therefore the impact of HMA use on long-term survival rates cannot be ascertained. As patients with higher-risk MDSs are counseled on the benefits and alternatives of HMA use and contemplate the decision to undergo alloSCT, an accurate description of the long-term outcomes expected with HMA use (without undergoing alloSCT) would be very important information. Additionally, whether the choice of HMA used for higher-risk patients affects the probability of long-term survival is currently unknown. To address these questions, we analyzed the Surveillance, Epidemiology, and End Results Medicare-linked database to assess the long-term OS rates for MDS patients treated with HMAs who did not undergo alloSCT.^5,6 

MDS patients who initiated HMAs between 1 January 2004 and 31 December 2009 were identified using the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) codes⁷  and were followed until death or end of follow-up on 31 December 2013. We limited eligibility to patients initiating HMAs before 2010 to allow at least 4 years of potential follow-up for every patient in the study to optimize reliability of long-term survival assessment. This analysis focused on patients with the ICD-O-3 code for refractory anemia with excess blasts (RAEB), a histologic designation that often overlaps with IPSS higher-risk MDS. Patients had to be 66 years of age or older at diagnosis, and had to have continuous Medicare Parts A and B coverage from 12 months before diagnosis until the end of the study period (defined as censoring, death, or 31 December 2013, whichever came first). Patients were excluded if they did not receive at least 10 days of HMA therapy, to minimize the effect on survival estimates of patients who deteriorated before having the opportunity to experience potential benefits of HMA therapy. Patients who received both azacitidine and decitabine at any point during observation and patients who underwent alloSCT were excluded.

Healthcare Common Procedural Coding System codes in Medicare claims identified treatments including HMAs and red blood cell (RBC) and platelet transfusions during an 8-week period preceding the first HMA treatment. Cycles of HMA therapy were defined by at least 3 individual days of HMA use and a gap of at least 2 weeks between cycles. We measured OS from the first day of HMA until death or end of study, whichever came first. Median household income and proportion of adults with high school education or less were derived from census tract-level measures (or zip code–level if not available). Comorbid conditions during the 12 months prior to initiation of HMA were used to measure the Elixhauser Comorbidity Index.⁸ 

We used χ² tests for categorical variables and Student t tests for continuous variables to compare covariates associated with use of either HMA. Kaplan-Meier methods were used to analyze unadjusted survival, whereas Cox proportional hazards models were used to compare survival between azacitidine and decitabine. We tested proportionality assumptions and observed no deviations. SAS version 9.4 (SAS Institute, Inc., Cary, NC) was used for all analyses. We used 2-sided statistical tests with an α of 0.05. This study was determined as not directly involving human subjects by Yale Human Investigation Committee.

Of 1187 MDS patients initiating an HMA, 336 (28.3%) had RAEB at diagnosis. The median OS for the entire 1187 patient cohort was 14 months (95% confidence interval [CI], 13-15 months). The 5-year OS probability was 8% (95% CI, 7%-10%). Of those with RAEB, 266 patients (79.2%) received azacitidine whereas 70 (20.8%) received decitabine. A higher proportion of decitabine-treated RAEB patients received ≥4 cycles (78.6% vs 64.3%, P = .03) compared with azacitidine-treated patients, but there was no difference in proportion of those receiving ≥6 cycles (51.4% vs 46.2%, P = .46). There were no significant differences based on HMA used in proportion who received RBC or platelet transfusions or in the median number of HMA cycles received (median of 5 cycles for azacitadine and 6 cycles for decitabine, respectively).

Almost all patients (n = 325, 96.7%) with RAEB had died by the end of study. Median OS was 11 months (95% CI, 10-12 months) and the 5-year OS probability for RAEB patients was a dismal 4% (95% CI, 2%-6%; Figure 1A). There was no significant survival difference for azacitidine-treated (n = 266) compared with decitabine-treated (n = 70, P = .80) patients, and the 5-year OS probability for the 2 groups was 4% (95% CI, 2%-6%) vs 4% (95% CI, 1%-11%; Figure 1B). As expected, the 5-year OS probability for patients receiving ≥6 HMA cycles (n = 159) was significantly higher than for those receiving <6 HMA cycles (n = 177) (6% [95% CI, 3-11] vs 1% [95% CI, 0-3], P < .001; Figure 1C). In multivariate survival analysis, the choice of HMA was still not associated with increased probability of long-term survival among RAEB patients (decitabine vs azacitidine, hazard ratio = 0.96, 95% CI, 0.70-1.31, P = .78; Table 1). On the other hand, receipt of RBCs or platelet transfusions remained significantly associated with increased risk of death.

In summary, in this retrospective cohort study of HMA-treated older patients with RAEB who had at least 4 years of follow-up from time of initiation of HMA therapy and who did not undergo alloSCT, we observed dismal 5-year OS probability rates of 4% (95% CI, 2%-6%), which did not differ by type of HMA used. These results offer a realistic estimate of the long-term survival probability for patients older than 65 years of age who wish to be treated indefinitely with HMAs instead of alloSCT. Further research should attempt to identify biologic and/or clinical markers that could predict prolonged survival with HMA therapy. Furthermore, our findings add to the accumulating body of evidence regarding the suboptimal performance of HMAs for treatment of MDS in the real-life setting^9-13  and provide a rationale for strong consideration of alternative management strategies including alloSCT or clinical trials for all patients with higher-risk MDS.¹⁴