Interferon-a Therapy in Chronic Myelogenous Leukemia: Questions Related to the German Randomized Trial

We read with interest the report of Hehlmann et all in which 133 patients were randomized to interferon-a (IFN-a) therapy (5 X lo6 U/m2/d). The incidence of complete hematologic response. (CHR) was 31% and of complete cytogenetic response was 7%. Their median survival was 5.5 years, which was superior to that of 186 patients treated with busulfan (median survival, 3.8 years; P = .008), but similar to that of 194 patients treated with hydroxyurea (median survival, 4.7 years). The study has raised several important questions. The first question regards the incidence of CHR with IFN-a therapy, which was considerably lower than that in other studies, including ours, despite similar response criteria (in our studies, normalization of peripheral counts also requires a decrease in platelet counts to <450 X IO3/pL).’ The incidences of major and complete cytogenetic responses were inferior to those obtained in our studies of 274 patients treated with IFN-a regimens (Table l).’ A complete cytogenetic response was achieved in 26% of patients (versus 7% in the current study), and a major cytogenetic response was achieved in 38% (versus <lo% in the current study). This is important because survival improvement is noted mostly among patients achieving cytogenetic response at 12 to 24 months3“ or a major cytogenetic response at any time.’ This is in line with the concept that achieving minimal tumor burden (complete or major cytogenetic response) in chronic myeloid leukemia ( C m ) prolongs survival. Thus, studies in which the incidence of major cytogenetic response is low (eg, < 10%. as in this study) may not be associated with survival prolongation over conventional (hydroxyurea) therapy in the total population. The second question regards why there are such differences in response rates with the same IFN-a therapy? Several factors may account for the differences. These factors include (1) the actual dose schedule of IFN-a delivery and the expertise of the treating physicians and (2) different population characteristics. A physician “learning curve” may exist with newer therapies, and physicians may be reluctant to use the investigational treatment (eg, IFN-a) at the full intended dose or even at all, particularly when it is more