Although in the latest British CLL4 trials fludarabine and the higher dose of chlorambucil were indistinguishable in efficacy, fludarabine was slightly more toxic. Neutrophil counts of less than 1000/cu mm occurred in 41% compared to 28%, hospitalization for more than a day occurred in 36% compared to 22% and diarrhea in 24% compared to 13%. Interestingly, haemolytic anemia was equally common in these two groups at 11% and 12%, but it tended to be more severe after fludarabine and there were some deaths among this group. Small wonder then, that NICE found that chlorambucil was actually superior to fludarabine as a single agent as first line treatment of CLL.
Why, you may ask, has the rest of the world preferred fludarabine? It was clearly on the basis of chlorambucil being given at a reduced dose in the CALGB trial. Remember that trial used chlorambucil at 57% of the dose used in the CLL4 trial, and even at the larger dose, chlorambucil was less toxic than fludarabine.
What is the proper dose of chlorambucil? For many years I have used the dose of 10mg daily for 14 days followed by 14 days rest. Few patients have not tolerated this dose. Expressed as mg/sq m/month and assuming an average surface are of 1.72 sq m, that works out as 81 mg/sq m. Clearly the LRF trial was not overdosing the chlorambucil at 70 mg/sq m.
Far be it for me to suggest that there was any irregularity in the choice of dose of chlorambucil in the CALGB trial, but the effect of this choice has been to maximize sales of fludarabine.
Another area of comparison between the two drugs has been in suppression of T cells. The results for CD4+ T cells after fludarabine are given in a paper by Keating et al in Blood 1998; 92:1165-71. They studied 127 patients whose median pre-treatment CD4+ count was 1562 per microliter. After 3 courses of fludarabine the median level was 172 and after 6 courses 163. It is difficult to discover what happens on follow up as no numbers are given, but there is a figure with a logarithmic scale. My reading of this figure is that the median CD4+ T cell count does not return to above 200 by 24 months follow-up. I should remind you that a CD4+ count of less than 200 in patients with HIV is what defines AIDS.
There are no precisely similar results for treatment with chlorambucil, but a paper appeared in Annals of Hematology last month (Lazlo et al, Ann Hematol 2006; 85:813-4). They studied the effect of chlorambucil 42 mg/sq m/week for 6 weeks and rituximab 350mg/sq m for 4 weeks followed by 4 additional cycles of chlorambucil 84mg/sq m/month on 27 patients with low grade lymphomas (some of which had CLL). Median counts at baseline were just over 500 and after the first 6 weeks they had fallen to a median of 246. After the 4 further months of chlorambucil they had fallen further to a median of 216. At nine months follow up the median count was back in the normal range.
While these studies are not completely comparable, they tend to confirm what has long been believed that fludarabine is particularly immunosuppressive compared with chlorambucil. Fludarabine is regularly used as an immunosuppressive in mini-stem cell transplants – nobody would use chlorambucil for that purpose.
Does this matter? The risks of T-cell depletion are mainly those of infection. The MDACC paper reported a risk of shingles in 13.8% of patients at risk and reactivation of herpes simplex in 3.8%. More important 8.8% developed Richter’s syndrome, which as we now know is frequently cause by failure of control of existing EBV infection. This is reminiscent of the experience of the Royal Marsden Hospital in London where they observed a 12% incidence of Richter’s transformation among CLL patients treated with fludarabine. (Thornton et al Leuk Res. 2005 ;29:389-95).
In the unpublished MRC CLL3NR trial patients failing to respond to chlorambucil or relapsing after chlorambucil were offered treatment with fludarabine second-line. The response rate was 80%. In LRF CLL4 the response rate to fludarabine given first line was 80%. In the UK fludarabine is licensed and recommended for use second-line in CLL.
As far as it goes the NICE decision not to recommend fludarabine as a single agent for first line use in CLL seems to me to be perfectly correct.