The Lack of Survival Differences in Randomised Trials in CLL May Be Related to the Effect of Second Line Therapies. A Report from the LRF CLL4 Trial.

On behalf of the UK NCRI Haematological Oncology Clinical Studies Group and NCRI CLL Working Group The LRF CLL4 trial randomised 777 patients with previously untreated CLL who required therapy from January 1999 to October 2004 between fludarabine as single agent (n=194), fludarabine plus cyclophosphamide (FC; 196) and chlorambucil (387). Results presented at ASH 2005 and now with an extended follow-up show statistically significant better complete (CR) and overall (ORR) responses with FC (39% CR and 94% ORR) compared to fludarabine (15% and 80%) and chlorambucil (7% and 72%). Progression-free survival (PFS) at 5 years was also significantly better with FC (33%) than fludarabine (14%) or chlorambucil (9%). Better response rates and PFS for FC occurred in all age groups including those over 70, and all prognostic groups defined by VH mutations and cytogenetics. Despite these results with FC, also seen by others (Eichhorst et al, Blood 107, 885, 2006), no survival differences have emerged. The projected 5-year survival with FC and fludarabine is currently 54% and with chlorambucil 61% (p=ns). This is puzzling as survival in CLL has previously been demonstrated to correlate with the quality of response. In CLL4 this was confirmed and we have also shown that good responders to second line treatment survive better. We therefore analysed data from 180 patients who received second-line treatments: 125 originally in the chlorambucil arm, 44 in the fludarabine arm and 11 in the FC arm. Patients in the chlorambucil arm received, as second-line therapy, mostly fludarabine alone or in combinations including FC+/−Rituximab (R) (70% of cases), CHOP, alemtuzumab, high dose methylprednisolone and others. Patients from the fludarabine arm had CHOP (31%), FC+/−R (24%), or others. Results are summarised in the Table. In the chlorambucil arm the proportion of patients responding to second-line treatment was higher than the proportion responding to first-line chlorambucil. Median survival after progression in 496 patients was also better in the chlorambucil (42 months) and fludarabine arms (35 months) than FC (8 months). These figures were respectively 33, 49 and 7 months for 147 non-responders to first-line treatment, and 52, 37 and 27 months for 259 responders who had relapsed. We conclude from this analysis that the likely reason for the lack of survival differences in CLL4 (and in other CLL trials) relates directly to the better responses and improved survival rates after second line treatment in those receiving the less effective therapy first, i.e. chlorambucil in CLL4. This observation is also relevant for the analysis of results of salvage protocols, which need to take into account the quality of the initial treatments and the mechanisms underlying resistance. Our findings support the view that PFS and quality of life should be used when assessing new treatment modalities in CLL, while continuing to evaluate survival differences to ensure that there is no adverse effect.