Background: In a previous review of hairy cell leukemia (HCL) patients we found no difference in efficacy between cladribine and pentostatin [Else et al, Cancer 104:2442- 8; 2005]. With longer follow-up (median 15.4 years since diagnosis) we aimed to find whether efficacy remained equivalent and whether factors associated with relapse could be identified retrospectively.

Methods: We reviewed data retrospectively from 228 HCL patients, treated between 1986–2005 with either single-agent pentostatin (n=187) or cladribine (n=41), at a median of 15.4 (range 1–38) years from diagnosis.

Results: Patient demographics were similar for both treatment groups, apart from a male:female ratio of 4.5:1 (pentostatin) versus 2:1 (cladribine) (p=0.02). No statistically significant differences were found between treatment groups after first-line therapy. Complete response (CR) rates were 82% (pentostatin) and 76% (cladribine). After a median followup since treatment of 13 years (pentostatin) and 9 years (cladribine), 43% of pentostatin and 37% of cladribine patients had relapsed. Relapse rates at 5, 10 and 15 years were 25%, 42% and 47% (pentostatin) and 35%, 44% and 44% (cladribine). Overall, patients who reached 5 years remission had only a 30% risk of relapsing by 15 years. Median progression-free survival (PFS) was 10 years. After relapse (n=75) or non-response (n=5), 24 patients received single-agent pentostatin and 56 cladribine, with CR rates of 63% and 75% respectively (not significant). Eleven patients subsequently relapsed after pentostatin and 18 after cladribine. Ten years after second-line treatment the rate of relapse was 53% and 39% for the two agents respectively (not significant) and second-line PFS was marginally shorter after pentostatin than after cladribine (p=0.04). Relapse rates and PFS were no different between those switching treatment (n=51) and those re-treated with the same agent (n=29). There were 45 deaths altogether, of which only 8 were HCL-related. With non-HCL-related deaths censored, overall survival from first treatment was 95% (pentostatin) and 100% (cladribine) at both 10 and 15 years (not significant). Compared with all others, the 95 patients (42%) whose disease progressed (who failed to respond to treatment, relapsed, or died of HCL-related causes) were more likely to have haemoglobin <10g/dl (p=0.01) and platelets <100x109/l (p=0.002). The median first-line PFS of patients with haemoglobin <10g/dl and/or platelets <100x109/l was 8 years versus 16 years for all others (p<0.001).

Conclusions and Recommendations: The long-term outcome for HCL patients was similar, whether treated with pentostatin or cladribine. The risk of dying of HCL-related causes was less than 5% and, among those surviving 15 years after treatment, 52% had still not relapsed. As in chronic lymphocytic leukaemia Binet stage C, patients with a low haemoglobin and/or platelet count appeared to fare worse. Further research will elucidate the factors underlying this difference in long-term prognosis. The outcome for relapsed patients may improve with the combination of either pentostatin or cladribine with the monoclonal antibody rituximab (Else et al, Cancer 110:2240–7; 2007).

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author