Predictors of vascular disease in myelodysplastic syndromes

Abstract The escalating link between somatic mutations commonly seen in myelodysplastic syndromes (MDS) and atherosclerotic vascular disease has increased the interest in management and associations of these conditions. We present a retrospective study examining clinical and molecular variables associated with vascular disease in patients with MDS. This study included a comprehensive evaluation of 236 patients with MDS. Our study has multiple findings. Mutations in ASXL1 correlated with increased risk of vascular disease for the entire cohort (P = .013). Though this has been replicated in other studies, there are no guidelines at this time for preventing vascular events in these patients. Our study also showed that lower ferritin levels may be linked to increased vascular events (P = .043), therefore the optimal use of supportive red blood cell transfusions in patients with MDS and the overall impact of inflammatory markers such as erythrocyte sedimentation rate and c‐reactive protein should be re‐addressed. Furthermore, our study showed that patients with Trisomy 8 in the low‐risk MDS cohort (based on IPSS‐R scores) were protected from vascular events (P = .036). Our findings of lower ferritin being linked with increased risk of vascular events as well as patients with Trisomy 8 being protected from vascular events may impact patient care. There do not appear to be any prior studies with these findings. In addition, given the connection between MDS and atherosclerotic vascular disease, we believe guideline‐based management of cardiac risk factors among MDS patients may improve overall outcomes. Further studies with larger patient cohorts are needed to further investigate these findings.

Mutations in ASXL1, TET2, and DNMT3A have been linked to 1.9 times greater risk of coronary heart disease compared to non-carriers of CHIP mutations [4].
Myelodysplastic Syndromes (MDS) and CHIP share common genomic events and are strongly associated with aging [5]. In fact, a subset of MDS cases is believed to arise from a preexisting CHIP state. A recent SEER study demonstrated an increased incidence of vascular events for MDS patients living 5 years beyond their MDS diagnosis [6]. These individuals had a likelihood of death attributed to vascular disease that approximated that of MDS itself [6]. Patients with low risk MDS were shown to have an elevated risk of death from cardiovascular disease [6]. As the link between vascular disease and clonal hematopoiesis strengthens, management of vascular disease in patients with MDS has become essential. It is key to establish which subset of MDS patients are at risk of death from vascular disease.
Hence, we investigated both clinical and genetic factors predictive of vascular disease in patients with MDS.

METHODS
This retrospective analysis included 236 adult patients with a con-  [7]. IPSS-R has five risk groups; we used the five groups to define two main groups for this study. We defined low risk MDS as an IPSS-R score <4.5 while defining high risk MDS as an IPSS-R score of 5 and higher.
Conventional karyotype information was present in 99% of the cohort.
Next generation sequencing-based multi-gene sequencing results were evaluated in 37% (n = 87), with a majority using FoundationOne Heme assay (Foundation Medicine Inc, Cambridge, MA, USA), which detects mutations at minor allele frequency threshold of at least 5% [8]. An additional 66 patients (28%) had limited specific mutational data accounted for in our study.

RESULTS
Median age at diagnosis was 69 (range: 20-91) years; 61% were men in this cohort. Patient characteristics are shown in  Reducing variability by using the square root of the ferritin value has been proposed by prior authors [11] and resulted in a statistically significant difference (P = .043). Treatment with drugs typically associated with prothrombotic states, such as erythropoietin-stimulating agents and lenalidomide, as well as other therapies such as hypomethylating agents were not associated with increased vascular events albeit limited sample size.
In the univariate analysis for the entire cohort, ferritin levels (OR: 2.54, P = .016), male gender (OR: 3.572, P < .01), and mutations in ASXL1 gene (OR: 3.58, P = .013) were the variables predictive for vascular disease that are not traditional cardiac risk factors (Table 2A). We assessed low and high risk MDS populations for vascular disease risk factors, as separate cohorts. Among low risk MDS patients, male sex and hyperlipidemia were associated with vascular disease (Table 2B).
Trisomy 8 was associated with a lower incidence of vascular disease in the low risk cohort (P = .036) as was elevated ferritin (P = .013).
In the high-risk disease group, traditional risk factors such as hyperlipidemia, hypertension and age were the most strongly predictive for vascular events (  Two retrospective studies (N = 77 [13] and N = 566 [14]) were conducted at outside institutions to study the link between MDS and TA B L E 2 Associations of MDS variables with vascular events. (A) shows data for the entire cohort of MDS patients, (B) shows data for patients with MDS with low risk of progression, and (C) shows data for patients with MDS with high risk of progression  [13] as well as DNMT3A mutations with a history of myocardial infarction (14% vs 6%, P = .03) [14]. Our study did not replicate these findings, as only ASXL1 mutations were significantly linked to cardiovascular events.
Our results confirm that management of traditional cardiac risk factors among MDS patients such as hyperlipidemia and hypertension may improve their overall outcomes. However, heightened attention may need to be given to MDS patients with older age and somatic mutations in ASXL1. The optimal use of supportive red blood cell transfusions in patients with MDS may also be re-addressed given our data that lower ferritin levels are linked to increased events. The impact of inflammatory markers such as ferritin as well as molecular events needs to be explored and validated in larger patient cohort.