Between January 1991 and December 1997, 103 patients, 97 with typical hairy cell leukemia (HCL) and 6 with HCL-variant (HCL-V) were treated with 2-chlorodeoxyadenosine (2-CdA) given as 2-h infusion for 5 consecutive d at a daily dose 0.12 mg/kg. To our knowledge this is the largest cohort of HCL patients treated with this type of regimen. Median follow-up amounted to 36 months. Fifty-six of 97 patients with typical HCL were newly diagnosed and 41 were relapsed after previous treatment. Splenectomy as a first-line therapy was performed in 23 patients and 18 remaining patients received prednisone, chlorambucil or interferon-alpha (IFN-alpha) alone or in combinations. Seventy-five (77.3%) patients entered CR and 18 (18.6%) achieved PR, giving an overall response rate of 95.9%. The mean time of first CR duration amounting to 32 months (range 3-72) did not correlate to the number of 2-CdA cycles. 2-CdA was equally effective in treatment of newly diagnosed patients and patients who relapsed after previous therapeutic procedures. Relapse of the disease occurred in 20 of 75 patients who achieved CR after 2-CdA and was usually manifested by very discrete changes in peripheral blood counts (neutropenia and/or relative lymphocytosis). The mean progression-free survival (PFS) time in this group was 37.4 (range 10-66) months. Ten of 20 relapsed patients were retreated with 2-CdA given an identical course to the first one. Seven patients entered second CR lasting 19+ (range 8-47) months and 3 experienced PR. This confirms the previous observations that 2-CdA gives no resistance to leukemic clone. Ten remaining patients have not required retreatment so far and remain in a good clinical and hematological state. The results of HCL-V treatment with 2-CdA were poor. Only 2 patients achieved PR and 4 patients did not respond to this drug. Seven patients (5 with typical HCL and 2 with HCL-V) died, 3 of causes unrelated to the disease. Second neoplasms were noted in 5 patients. 2-CdA-related side effects resulted mainly from myelosuppression and infectious complications. In conclusion we confirm the effectiveness of 2-CdA in inducing CR in patients with typical HCL, but this drug is unable to completely eradicate the leukemic clone which results in the relapse of the disease. The real incidence of the relapse rate may be underestimated unless bone marrow biopsy is performed. The results of our study indicate that a 2-h infusion of 2-CdA in HCL patients is at least as effective as a 24-h infusion but more convenient to the patients, and may be given on an outpatient basis.