A Study Comparing Dosing Regimens and Efficacy of Subcutaneous to Intravenous Azacitidine (AZA) for the Treatment of Myelodysplastic Syndromes (MDS).

AZA, the most commonly used disease-modifying therapy for MDS in the US, demonstrates a survival advantage in higher-risk (International Prognostic Scoring System [IPSS] Int-2, High, or excess blasts) MDS subtypes. It was approved by the US FDA in 2004 as a 7-day per 28-day, subcutaneous (SC) regimen, based on response rates, with the 7-day intravenous (IV) route approved in 2007 based on efficacy data from the CALGB 8421 study and pharmacokinetic data. This registry allows for the evaluation of community based practices for dosing and effectiveness of IV compared with SC administration in MDS.

AVIDA, a prospective, longitudinal, multicenter US patient (pt), registry collects data from community-based hematology clinics on the history and management of pts with MDS treated with AZA. Interim analyses were conducted on data collected from October 2006 – May 2009 with investigator chosen regimens according to the most commonly used per pt. Responses used 2006 International Working Group criteria, quantified as hematologic improvement (HI) or better and were assessed centrally. ANCOVA models were used for dose per cycle analyses and generalized estimating equations were used for analyses of cycle delay. Cox models incorporating survival were used for multivariate analyses.

Of 380 registry pts (median age, 75 years [range, 29-91], 104 (31%), female), 331 had MDS or oligoblastic leukemia, of whom 190 (57%) most commonly received AZA via IV and 141 (43%) by SC (both for a median of 4 cycles). Median time since MDS diagnosis was 2 months (range, 0-207), and 10% of pts had secondary MDS. Baseline bone marrow blast % was: <5% (46%); 5-10% (26%); 11-20% (14%); 21-30% (3%); and unknown (11%); IPSS cytogenetic groups were: good (61%); intermediate (10%); poor (17%); unknown (11%), with 6% and 9% of pts having a chromosome 7 and 7q abnormalities, respectively. IPSS risk groups were: Low (11%); Int-1 (36%); Int-2 (15%); High (6%); unknown (33%). ECOG performance status was: 0 (25%); 1 (53%); 2 (16%); 3 (6%). Only 17% of pts received the FDA-approved continuous 7-day dosing schedule; 51% received <7 days; 30% 7 days with breaks; and 2% >7 days. There were no significant differences between SC and IV AZA recipients for any of the above baseline parameters. At the interim analysis time point of 600 days, there were 21 deaths (15%) in the SC group and 32 deaths (17%) in the IV group. In univariate analyses, the following predicted for worse survival: black race (versus white, P = .02); higher-risk MDS (P = .02); cytogenetic abnormalities (P = .004); IPSS poor risk cytogenetic group (P = .04); low body mass index (BMI, P = .03), high blast % (P = .01)and baseline low hemoglobin (hgb, P = .007), or low platelets (plt, P = .05), all analyzed continuously. SC vs. IV dosing had no differing effect on HI rate (24% overall). In multivariate analyses, significant variables included low hgb (P = .006), low plt (P = .05), high blast % (P = .04), cytogenetic abnormalities (P = .05), and low BMI (P = .002). Delays in cycle start times (>28 days from previous cycle start) were related in multivariate analyses to higher blast % (P = .004), male gender (P = .03), dosing schedule (P = .02), a trend with IV dosing (P = .1), cytogenetic risk score (P = .1) and del (5q) (P = .09). Lower AZA doses per cycle were related in multivariate analyses to IV dosing (P = .001, on average 12 mg lower); older age (P = .01, on average .6mg less per year); BMI (P < .001), female gender (P < .001), and, as expected, dosing schedule (P = .04).

In this interim analysis of the ongoing AVIDA Registry, IV AZA appears equi-efficacious to SC, although dose, schedule, and treatment differed between groups. Though the FDA-approved continuous dosing schedule of AZA was infrequently used, this has not negatively impacted efficacy to date. In multivariate analyses, traditional IPSS predictors of survival are still relevant in pts treated with AZA. These analyses also reveal other potentially significant factors, such as low BMI, the effect of gender, and age on treatment intensity and clinical outcomes in patients receiving AZA.

Baker:Celgene: Employment, Equity Ownership. Sullivan:Celgene: Employment, Equity Ownership.