Problems in the classification of CMML--dysplastic versus proliferative type.

The FAB group proposed to distinguish two subgroups of chronic myelomonocytic leukemia (CMML). Depending on the total leukocyte count, a myelodysplastic type (MDS-CMML) (< or = 13,000 microl(-1)) was separated from a myeloproliferative type (MPD-CMML) (> 13,000 microl(-1)). Based on retrospective analyses of 158 patients with CMML, we compared the presenting clinical and hematological features of both disorders and examined whether the refined classification is important in terms of prognosis. There were 81 patients with MDS-CMML and 77 patients with MPD-CMML. Median age of patients at diagnosis (70 versus 72 years) was not different. The sex ratio showed a preponderance of males in the MPD group (m:f; 2.1:1). Splenomegaly was more common in MPD-CMML (54 versus 30%; P = 0.002). With regard to laboratory findings, patients with MPD-CMML presented with significantly higher LDH values (medians 295 versus 231 U ml(-1); P = 0.008) and higher serum deoxythymidine kinase levels (medians 150 versus 41 U microl(-1); P = 0.0025). Except for white blood cell count (WBC), peripheral blood counts were not different. Median percentage of bone marrow blasts was 9% and cumulative survival rates were similar in both disorders. Two years after diagnosis, actuarial survival for patients with MPD-CMML was 33%, as compared to 50% for patients with MDS-CMML (P = 0.31). The probability of transformation to AML was higher in MDS-CMML (32 versus 17% after 5 years), but this difference also did not reach statistical significance. The survival of patients with MDS-CMML was similar to that of other MDS patients (RAEB) who had corresponding medullary blast counts. Using the Düsseldorf-score, we could define two risk groups within MDS-CMML with a median survial of 12 versus 40 months (P = 0.001). None of the known scoring systems could define risk groups within the MPD-CMML group. In summary, these data suggest that MDS-CMML and MPD-CMML are clinically distinguishing conditions, but the separation provides little prognostic information. Further studies are needed to clarify whether response to therapy is different in MDS-CMML and MPD-CMML.