In a prospective, randomized trial the effectiveness of dibromomannitol (DBM), a brominated sugar alcohol derivative, was compared to busulfan in previously untreated patients. One hundred thirty-one patients were evaluated for response and survival. The effective dose of DBM was 4 mg/kg. The persistence of sensitivity to either DBM or busulfan was shown in a quantified fashion. Despite initial reports of the alleged unique effectiveness of DBM in treating chronic myeloid leukemia (CML), this study did not disclose any advantage of DBM over busulfan. Multivariate analysis investigating the importance of prognostic factors in CML indicated that age, sex, splenomegaly, and initial platelet count were important for predicting survival. Determination of such factors in CML are valuable in deciding those patients who could benefit from alternative forms of therapy. It also insures that study results are related to treatment rather than selection of patients with favorable or unfavorable prognostic factors.