A Phase II Study of the Efficacy and Tolerance of Azacytidine (AZA) in Steroid Dependent/Refractory Systemic Autoimmune and Inflammatory Disorders (SAID) Associated with MDS or CMML (GFM- AZA-SAID -trial)

Background: Systemic autoimmune and inflammatory diseases (SAIDs) occur in 10-15% of MDS or CMML patients. They are often severe, difficult to treat and steroid resistant/dependent disorders. In a retrospective report of 22 patients with steroid-dependent MDS/CMML associated SAID, we observed promising results with azacytidine (AZA) on the SAID component (Fraison et al, Leuk Res, 2016). In this prospective French nationwide trial, we aimed at confirming this effect prospectively.

Patients and Methods: This open-label, single-arm multicenter, phase II study included MDS/CMML patients with IPSS/CPSS int 2 or high, or lower IPSS/CPSS but with significant cytopenias, especially ESA resistant anemia (or, for CMML, proliferative features) and SAID with steroid dependence and/or resistance. The primary endpoint was response rate (RR, including CR and PR) of SAID after 6 cycles of AZA. Secondary endpoints included RRof SAID after 3 cycles of AZA, response of MDS, time to MDS progression (including to AML), overall survival (OS) and safety. AZA was used at 75 mg/m²/d x 7 days, every 4 weeks for a minimum of 6 cycles (unless overt disease progression occurred before). At the onset of AZA, prednisone (PRED) was administered at 1 mg/kg during 1 month and then decreased progressively over 6 months.

Results: 30 patients were included between July 2017 and June 2020 , 29 of whom received AZA and were considered evaluable: median age 72 [range 68-78], 69% males; WHO: MDS ML (n=11; 41%), CMML (n=11, 41%), MDS-RS (n=2, 7%), EB-1 (n=1) and EB-2 (n=1), unclassified (n=3); IPSS-R : good (n=15, 54%), very good (n=3; 11%, each), intermediate (n=7; 25%) and very poor profile (n=3; 11%) (all CMML had WBC <13G/l and could be classified by IPSS-R). SAIDs features included non-infectious fever (n=5; 17%), skin lesions (n=14; 48%), joint involvement (n=18; 68%), chondritis (n=6; 21%) and peripheral nervous system impairment (n=3; 10%). 11 (38%) /28 tested patients had UBA1 mutation (VEXAS syndrome), while inin the remaining cases SAID was vasculitis (n=5), rheumatic disease (n=6), neutrophilic dermatosis (n=2) and others (n=5).

19 patients (65.5%, IC95% [45.7.82.1]) obtained a SAID response after 6 cycles (including 8 CR and 11 PR), while 17 (58.6% (IC95% [38.9; 76.5]) patients achieved an IWG 2006 hematological response (7 CR, 9 SD with HI and 1 marrow CR). All the 13 patients who received 12 cycles of AZA achieved SAID CR or PR at 12 months, and 10 of them had a hematological response (6 CR, 1PR, 2 HI, 1 marrow CR). The SAID response rate at 12 month was at 72.4% (IC95% : [51.3;85.6]). SAIDs complete/partial response occurred in 3/9 (33%) UBAI-positive versus 8/19 (42%) non-mutated cases. No significant prognostic factor of SAID response after 6 cycles, including sex, age, WHO classification, IPSS-R, UBA1 and TET-2 mutation status was found. With a median follow-up of 18.5 months [11.8; 24.3], 9 (31%) patients had died, with a 1-year OS of 82.3% (IC95% [69.4; 97.7]. At least one SAE occurred in 23 patients, with a median of 2 events/patient [1.5; 5.5]. Death was due to infection (n=5, two of them in patients with good IPSS-R, but severe steroid dependent SAID), MDS progression (n=2), or unrelated causes (n=2).

Conclusion We confirmed prospectively an effect of AZA on the autoimmune/inflammatory component in two thirds of MDS/CMML patients with steroid dependent / refractory concomitant systemic autoimmune/inflammatory disease. No significant difference in AZA response was noted between UBA1 mutated and unmutated MDS patients. The adverse events and mortality rates, often from infectious origin, appear to be those expected in such an MDS population, potentially worsened by the associated SAID and background of immunosuppressive treatment (especially steroids)