A phase II study was undertaken to evaluate the efficacy and toxicity of a new schedule of cladribine administration (10 mg/m2 orally daily for 3 d every 3 weeks) in 107 patients with B-cell chronic lymphocytic leukaemia (CLL). To minimize toxicity, treatment withdrawal criteria were defined. The results of the 63 previously untreated patients were retrospectively compared with 63 from an earlier study using a 5-d monthly schedule. The compiled data were analysed for prognostic factors for survival. No significant difference regarding response were seen in the two cohorts of the 126 previously untreated patients. The complete response (CR), nodular partial response (nPR) and partial response (PR) rates were 15%, 21% and 41%. Quality of response had no impact on survival. The 3- and 5-year overall survival for previously untreated patients was 73% and 58%, respectively, with a median follow-up of 54 months. Pretreatment haemoglobin <11.0 g/dl and elevated beta-2-microglobulin had a negative influence on survival. Major infections occurred in 21% of patients in the 3-d study compared with 35% in the 5-d study. The overall response (OR) and CR rates in the 40 previously treated patients were 34% and 5% respectively. Median overall survival was 24 months and median progression-free survival for responding patients was 14 months. Cladribine used as a single agent is an effective treatment with an acceptable safety profile for pretreated and untreated B-CLL. The achievement of complete remission was not a prerequisite for long-term survival.