Introduction. IMIDs occur in about 15% of MDS and CMML (Mekinian et al, Rheumatology (2016;55:291-300), but the precise relationship between those disorders remains uncertain. The impact of IMIDs on OS in those patients is also still unclear. Published data may include methodological biases, especially the consideration of IMIDs as a time-dependent event, as it can be diagnosed before, concomitantly, or after MDS/CMML diagnosis. We studied the impact of IMIDs on OS in MDS/CMML, considering this time-dependent relationship.

Methods. We constructed a multicenter European cohort of non-VEXAS patients, to which 5 centers participated (University Hospitals of Bordeaux, Vall d'Hebron (Barcelona), Sassari (Italy), Paris Saint-Louis, and Poitiers (France)). A first analysis compared MDS patients who, at MDS/CMML diagnosis, were already diagnosed with IMID (irrespective of diagnostic interval between the 2 disorders) versus those without IMID; a second analysis was performed in the patient subgroup without IMIDs at MDS/CMML diagnosis to estimate the impact of developing IMIDs as a time-dependent event. Cox multivariate regressions included R-IPSS variables, age, and gender. An updated Cox regression was used for the second analysis. Models were stratified on the center. Multiple imputation was performed to manage missing values. HR and confidence intervals (CI) were calculated by bootstrapping the imputed datasets.

Results. Data from 1132 patients (810 MDS and 322 CMML) diagnosed between 2000 and 2022 were collected. Median age was 74 (range 21-95), with 61% males; R-IPSS was very high in 6.6%, high in 12%, intermediate in 29%, low in 35%, and very low risk in 17%. Molecular biology was unavailable in close to 50% of the patients, diagnosed before NGS techniques were available, so that the WHO 2016 classification was used. An IMID was identified in 196 patients (17.3%), including autoimmune cytopenias (20%), connective diseases (6.6%), inflammatory arthritis (32%), neutrophilic dermatosis (7.7%), vasculitis (21%), and other less frequent IMIDs (13%). In some male patients diagnosed before 2020 and in whom UBA1mutation analysis was not performed subsequently, VEXAS could however not be excluded. No difference was found in aseline characteristics between those with and without IMIDs, except for neutrophils, which were higher in the IMIDs group (median 3.0 versus 2.3 G/L, p=0.007).

When comparing patients with or without IMID at MDS/CMML diagnosis (first analysis, n=1003) presence of an IMID had no impact on OS in the multivariate analysis (HR 1.02; 95% CI 0.77-1.35), while prognostic factors for OS were, as expected, lower Hb level (HR 1,11), higher log(neutrophils) (HR 1.14), platelets < 50 G/L (HR 1.66,), marrow blasts 5-9% (HR 1.68), and 10-19% (HR 1.66) high-risk (HR 2.41) and very-high risk karyotype (HR 3.88).

In the second analysis (n=966), however, the occurrence of IMID after MDS/CMML diagnosis had a negative impact on OS (HR 1.84; 95% CI 1.05-3.20) in the multivariate time-dependent analysis. Other prognostic factors remained generally similar including older age (HR 1.03), lower Hb level (HR 1.12), higher log(neutrophils) (HR 1.26), platelets < 50 G/L (HR 1.69), marrow blasts 10-19% (HR 1.68), high-risk (HR 2.11; 95% CI 1.35-3.42) and very-high risk karyotype (HR 3.83; 95% CI 2.39-6.26). No specific type of IMID was associated with worse survival.Conclusion. Development of an IMID during the course of MDS/CMML negatively impacts OS. We are currently reviewing in detail possible worsening of MDS/CMML characteristics (regarding WHO classification, IPSS-R, karyotype, and, in more recent patients, somatic mutations) just before or after diagnosis of IMID, potentially explaining poorer outcome. This would possibly also contribute to better knowledge of the physiopathological link between IMIDs and MDS/CMML, and in particular whether occurrence of an inflammatory disease can lead to MDS/CMML progression or if clonal evolution of MDS/CMML can trigger or worsen an IMID, as suggested by our group (Zhao et al. Leukemia. 2023; 37:1186-1190). These two explanations may not be exclusive.

This content is only available as a PDF.
2025
Sign in via your Institution