Treatment outcomes in patients with VEXAS syndrome: a retrospective cohort study.

BACKGROUND: Vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic (VEXAS) syndrome is a recently described autoinflammatory disorder with little therapeutic evidence. We compared treatment outcomes of targeted therapies versus prednisolone alone in the largest UK cohort of patients with VEXAS syndrome to date. METHODS: In this retrospective cohort study, we analysed the outcomes of targeted therapies in patients with VEXAS syndrome in six tertiary referral centres across the UK between July 22, 2014, and Oct 19, 2024. The inclusion criteria were genetically confirmed VEXAS syndrome and receipt of at least one targeted therapy or prednisolone alone. Patients without clinical information at all timepoints after baseline were excluded. Data collection forms were used to record clinical and biochemical data at the following timepoints: time of diagnosis, initiation of treatment, and follow-up at 3 months, 6 months, and 12 months from the initiation of treatment (±28 days). Laboratory parameters, including C-reactive protein (CRP) and haemoglobin, and glucocorticoid doses were collected at each timepoint and compared between timepoints. Primary outcomes were complete response (ie, clinical remission, CRP ≤10 mg/L, and prednisolone ≤10 mg per day) and partial response (ie, clinical remission with ≥50% reductions in both CRP and glucocorticoid dose from baseline) to treatment. Treatment discontinuation and adverse events were documented for each treatment. Due to the high prevalence of cytopenias in VEXAS syndrome, these were only recorded as adverse events when necessitating treatment change. People with lived experience were not involved in the study. FINDINGS: We analysed 71 targeted therapies in 59 patients with genetically confirmed VEXAS syndrome. Of the 59 patients, 58 (98%) were male and one (2%) was female, with a mean age of 71 years (SD 8), and 27 (46%) had myelodysplastic syndrome. The treatments included tocilizumab (n=19), anakinra (n=13), azacitidine (n=13), baricitinib (n=11), and prednisolone only (n=10). At 6 months, in those who continued therapy, ten (91%) of 11 patients receiving azacitidine showed a response (three [27%] complete responses), as well as did seven (64%) of 11 receiving tocilizumab (four [36%] complete responses), three (100%) of three receiving anakinra (one [33%] complete response), and two (40%) of five receiving baricitinib (no complete responses). Although all patients who tolerated anakinra had a response, the discontinuation rate was high (eight [62%] of 13), mostly due to severe injection-site reactions (n=5). Patients were more likely to respond to azacitidine than to other therapies at 6 months (risk ratio 2·47, 95% CI 1·18-5·20; p=0·018). Absence of fever or thromboembolism at diagnosis was associated with better outcomes. By 6 months, median CRP concentrations had decreased in patients receiving tocilizumab (from 30 mg/L [IQR 13-45] to 4 mg/L [3-37]) or anakinra (from 18 mg/L [11-52] to 2 mg/L [1-28]), whereas azacitidine showed the greatest increase in haemoglobin (from mean concentration 104 g/L [SD 17·5] to 120 g/L [14·4]). 28 (39%) of 71 treatments were discontinued, most commonly due to serious adverse events (12 [17%]) and death (nine [13%]). Infections were most frequent with azacitidine (eight [62%] of 13) and tocilizumab (nine [47%] of 19). INTERPRETATION: In this UK cohort of patients with VEXAS syndrome, azacitidine and tocilizumab showed superior effectiveness compared with anakinra, baricitinib, and prednisolone only. Treatment selection should consider individual risk factors and tolerability. Prospective studies are needed to confirm optimal treatment strategies and develop standardised protocols. FUNDING: None.