Re-analysis of prognosis in refractory anemia alone, with ring sideroblasts, and with excess of blasts

Based on 473 primary cases of refractory anemia (RA), RA with ring sideroblasts (RARS), or with an excess of blasts (RAEB), new staging systems for overall survival, leukemic transformation, and nonleukemic death were developed. The reproducibility of these new staging systems, our former staging systems for leukemic transformation and nonleukemic death, and that for overall survival provided by Sanz et al. (Blood 1989;74:395-408), were estimated using a total of 132 patients with RA, RARS, or RAEB (test sample). The correlation of risk ratios of the test sample from our system and that of Sanz et al. was poor (tau = 0.24; Kendall rank correlation). Our risk ratio was a significant (p = 0.0012) continuous covariate of the Cox model life-table analysis in the test sample but that of Sanz et al. was not. The new and previous staging systems for leukemic transformation were highly similar (tau = 0.7) and risk ratios of both systems were significant in the test sample (p = 0.00046 and 0.0017, respectively). Based upon the new system, patients at high risk for leukemic transformation RA or RARS were identified as males, with bone marrow myeloblasts higher than 2.5%, abnormal granules in the granulocytes, and circulating erythroblasts. Their cumulative leukemic transformation rate exceeded 50% within 2 years from presentation. The new and previous systems for nonleukemic death were moderately correlated (tau = 0.47). The risk ratios of both systems were significant in the test sample (p = 0.00012 and 0.000024, respectively) but failed to stratify test sample patients into three distinct risk groups. In conclusion, our new staging systems for overall survival and leukemic transformation were reproducible in RA, RARS, or RAEB patients and will be useful especially in identifying high-risk patients with RA or RARS, who are considered to be at lower risk by the FAB classification.