RESUMO
A multivariate analysis of clinical, biochemical and hematologic data was performed in 138 patients with myelodysplastic syndromes (MDS) in order to evaluate their prognostic significance. The most important individual variables, isolated in a previous univariate analysis, were placed in a multiple regression modeling procedure to identify major prognostic factors. Multivariate analysis tends to identify prognostic variables containing significant predictive information. Characteristics were examined on both continuous and binary bases. The FAB classification was the first parameter entered in regression equations of both models, followed by platelet count, systemic symptoms, bone marrow blast and infection. Our analysis confirms FAB classification as the best prognostic factor in MDS. It supports the previously predictive value of platelet count, hemoglobin level and bone marrow blast and recognizes the importance of systemic symptoms and infection as prognostic characteristics in MDS.