Azacitidine for Patients with Vexas Syndrome: Data from the French Vexas Registry

Background

VEXAS syndrome (Vacuoles, E1 Enzyme, X-linked, Autoinflammatory, Somatic syndrome) is an an often life threatening autoinflammatory disease resulting from UBA1 gene somatic mutations, associated in 40-50% patients (pts) with MDS. While we and others found Ruxolitinib to have frequent efficacy in VEXAS (Heiblig, Blood 2022;140:927-931), we also suggested a significant role of azacitidine (AZA) in a retrospective study (n=11, Comont, Br J Haematol 2022 ;196:969-974) and a prospective study (n=12, Mekinian, Leukemia 2022 ;36:2739-2742). We report here the efficacy, safety and prognostic factors of AZA treatment in a larger cohort of VEXAS pts, most of whom had concomitant MDS.

Patients

In this retrospective study, we selected typical VEXAS pts from the French nationwide registry (n=326 pts with UBA1 mutation: Met41 or alternative splicing site) who received at least one full cycle of AZA. Major response to AZA on VEXAS was defined as (i) complete control of clinical inflammatory symptoms (ii) C-reactive protein (CRP) <10 mg/L (iii) daily steroid dosage <10 mg/day, and no other immunosuppressive treatment administered for VEXAS. Minor response referred to significant improvement, not meeting all major response criteria

Results

57 male pts (median age 71, range 54-86, at AZA onset) with severe (clinical and/or biological) phenotype were included. 50 (88%) of them were considered to have associated MDS based on morphology in 46 (WHO 2016 criteria): MDS-MLD (n=31), MDS-EB1 (n=8), MDS-SLD (n=3), MDS-RS (n=2), MDS-EB2 (n=1) and CMML-2 (n=1), of whom 94 had normal karyotype; and based on presence of clonal cytogenetic abnormality in 4 (del 9q, +15, t(12;15), del 7q). R-IPSS was very low/low/intermediate/high in 5/24/9/6 cases respectively. NGS mutational profile before AZA, available in 44 pts, found TET2 mutation in 10 (22.7%), DNMT3A mutation in 13 (29.6%) and at least another myeloid mutation in 12 (27.2%) pts.

At AZA onset, all pts except 3 were receiving steroids, with a median dose of 20 mg/day, and median duration of steroid exposure was 21 months (range 1-173). Apart from steroids, pts had received a median of 2 immunosuppressive/modulatory drugs (range 0-8).

The median number of AZA cycles administered was 11 (range 1-111), and 27 (47%) pts had discontinued AZA at the time of analysis, due to serious infection (n=6), persisting response (n=4), primary or secondary failure (n=5), local or systemic allergic manifestations (n=4), allogeneic stem cell transplantation (n=4), unrelated causes (n=4). Two pts progressed to high-risk MDS, but none to AML. Median follow-up from AZA onset was 29 months, and 16/ 57 patients had died, 8 of whom while still on AZA. Cause of death in those 8 pts was COVID-19 (n=1), other infection (n=3), VEXAS (n=1), unrelated (n=3).

During AZA treatment, serious adverse events were reported in 30 (53%) pts, including mostly infections (n=25).

51 pts (46/50 with, and 5/7 without MDS), who received at least 3 cycles of AZA, were considered evaluable for response. A VEXAS response was achieved in 34/46 (74%) pts with MDS, major in 23 (50%). In responders, prednisone could be discontinued in 15 pts, while the remaining 19 patients received a median dose of 7 mg/day. Best response was obtained after 3-6/ 7-9/10-12/>12 cycles of AZA in 20/5/3/4 pts, respectively (2 unknown). After a median follow up of 35 months (range 5-163) from AZA onset, 21/34 (62%) were still responders.

All 5 pts without MDS achieved a major response (during 4 to 37+ months, with no relapse).

In univariate analysis on 46 MDS pts receiving at least 3 cycles of AZA (including VEXAS phenotype and MDS characteristics), no predictive factor for the response to AZA were identified.

Of the 51 pts who received at least 3 cycles of AZA, 32/41 (78%) corrected anemia (including 22/31 (71%) who achieved RBC-transfusion independence), 19/28 (68%) thrombocytopenia, and 3/6 (50%) neutropenia.

UBA1 mutation VAF monitoring during treatment, currently available in 6 responders, showed mutation negativation in 4 patients, and reduction to 89 and 62 % in 2 and more responders are being analyzed.

Conclusion

We confirmed the efficacy of AZA in this large cohort of VEXAS pts, even in those with no demonstrated underlying MDS, where the treatment appears to directly impact clonal evolution in at least some cases. The infectious risk, in these pts who have in particular received steroids and immunosuppressive drugs, requires close monitoring.