Clinical Characteristics and Prognosis in South-American Patients with Myelodysplastic Syndrome: A Multicenter Study

There are previous reported data describing differences between European and Asian patients with Myelodysplastic Syndromes (MDS), and little is known about South-American (SA) patients. Our aim was to describe clinical characteristic of SA MDS population, to compare our series with diverse ethnicity, to evaluate scoring systems and prognostic factors.

We retrospectively analysed a series of 1080 patients with de novo MDS (1981-2014) from Argentine (Ar-635), Brazil (Br-345), and Chile (Ch-100). Patients were classified following FAB and WHO criteria, excluding patients with bone marrow (BM) blasts >30% and CMML with white blood cells count >12x10⁹/L. Patients undergoing hematopoietic stem cell transplantation (5%) or hypomethylating therapy (12%) were censored at treatment initiation.

Chilean patients were younger (mean age: 59 vs 66-Ar, p<.001 and vs 65 years old-Br, p=.003), with a female preponderance (ratio M/F: 0.8-Ch, 1.3-Ar, 1.3-Br, p=.061). Since Ch did not include patients with CMML, distributions among FAB and WHO categories were different (p<.001). Br series showed a higher predominance of RARS (17% vs 9%-Ar vs 1%-Ch), and a lower percentage of RCMD (38% vs 43%-Ar vs 49%-Ch), while the frequency of the other subtypes were comparable. BM blast categories according to IPSS and IPSS-R were similar (p=.626 and =.508). Hemoglobin (Hb) level was significantly higher in Ar series (9.6g/dL vs 8.6g/dL-Ch, p<.001, and vs 8.7g/dL-Br, p=.002), with no differences in ANC and platelets (p=.842 and .763). There were no differences in the distribution of cytogenetic groups of risk (CGR) according to IPSS (p=.157). With respect to IPSS-R CGR distribution, Ar series showed a higher frequency of Very Good and Intermediate findings (4% and 19%, vs 2% and 15%–Br, 0% and 9%-Ch, respectively, p=.037).

The distribution among IPSS categories was similar with a higher predominance of the high risk group in Ch (17%, 9%-Ar, 9%-Br, p=.056). IPSS-R distribution confirmed a higher frequency of very high risk in Ch series (20%, 13%-Br, 10%-Ar) and a lower predominance of very low risk (5%, 15%-Br and 22%-Ar) (p<.001).

We also evaluated different prognostic factors in each MDS population and in the whole population (table). Overall survival (OS) was lower in Ch patients, consistent with the higher prevalence of higher risk group categories (36 months vs 56m-Ar; 63m-Br, p=.026). Hb, BM blast, CGR, IPSS and IPSS-R were useful to predict survival in the three series and in the overall SA MDS population. Age was not useful for Br patients and showed a borderline influence for the whole SA series. The ANC was neither a predictive variable for Br nor for Ch patients, and platelets count was not useful in Ch series.

Epidemiological and clinical characteristics, distribution among prognostic subgroups and OS were similar for Ar and Br compared to Ch MDS series. This will need further analysis in a larger group of patients. Nevertheless, the IPSS-R system and its variables showed a good reproducibility to predict clinical outcome for the whole SA population.