Myelodysplastic Syndromes and Acute Myelogenous Leukemia Resulting from Therapy for Autoimmune Disease, a Case-Control Cohort Study

Background

Cytotoxic agents are a well-established risk factor for therapy related myeloid neoplasms (t-MN). The incidence of t-MN in patients (pts) with autoimmune diseases (AID) treated with cytotoxic and immuno-modulatory agents are not well described in the literature. Most case series assess an MDS/AML population and co-occurrence with an AID. To our knowledge, there exists no large case-control studies evaluating a primary AID population that developed therapy-related MDS or AML, and the relationship to detailed therapeutic intervention for AID. Our study queries the association of cytotoxic and immune-modulating agents to treat AID as risk factor for developing MDS or AML.

Methods

Between 2004-2014, 23,671 pts with an ICD-9 coded AID were seen at Mayo Clinic Arizona; 148 pts from this population had a concomitant coded diagnosis of MDS/AML. In an IRB approved, retrospective chart review we confirmed 61 cases meeting strict inclusion criteria. We performed a case-control match from the same primary AID population for age, gender, and autoimmune specific diagnosis at a 2:1 ratio. AML/MDS cases were compared to controls by use of chi-square test for frequency data. Odds ratios with 95% confidence interval were constructed using conditional logistic regression. Case incidence was compared to SEER data for the general population

Results

Of 61 cases 33 were MDS, 20 de-novo AML, and 8 MDS transformed to AML. Psoriasis 16 (26.2%) and rheumatoid arthritis 14 (23.0%) were the most common autoimmune diagnoses, followed by systemic lupus 11 (18%), Crohn's disease 7 (11.5%), and ulcerative colitis 5 (8.2%). 39 out of 61 cases (63.9%) received either a cytotoxic or immune modulating agent for AID, compared to 64/122 (52.5%) in the control group; p=0.140. 7/61 cases (11.5%) received both a cytotoxic and immune modulating agent as compared to 16/122 (13.1%) in controls; p=0.753.

Overall, the case population received more agents than controls. No cytotoxic or immune-modulating treatment was documented in 16% of cases and 36% controls. Exposures to cytotoxic and immune-modulating agents were similar in both case and control cohorts. Exposure time was not statistically significant. Azathioprine was the only statistically significant agent exposure observed more often in the case cohort than control. Per SEER data, over a 10-year period in a general age-matched population that is comparable to our AID population cohort, we would expect 54 MDS and 35 AML cases.

Discussion

Treatment of AID does not seem to increase the incidence of MDS or AML. Furthermore, the incidence of myeloid neoplasm occurring in association with treatment of AID is not higher than expected in the general SEER population. However, the case cohort was exposed to more agents overall and more likely to have exposure to azathioprine than controls. Limitations of our study are associated with its retrospective nature, possible associated underrepresentation of t-MN, and autoimmune population profile. Nevertheless, this is the largest case study with detailed drug exposures with timelines for pts with AID, contributing to a better understanding regarding possible risk of developing MDS/AML in pts treated for AID. Confirmation of the same study at Mayo Clinic Rochester and Florida are ongoing.