Outcomes in clinical trials

A recent paper in the New England Journal of Medicine has given us all something to think about. It did not concern CLL; rather it was about a similar disease, multiple myeloma.

Myeloma, like CLL, is an incurable disease of B cells. It has a rather worse prognosis than CLL, but like CLL it has a forme fruste Monoclonal Gammopathy of Undetermined Significance (MGUS) that fits alongside Monoclonal B cell Lymphocytosis.

For those who can take it, standard treatment is an autologous stem cell transplant. This treatment has been pioneered at Little Rock, and where one is good two may be better. Little Rock now propounds the virtue of two successive autografts. There is evidence that, if you can stand it, two are better than one, although some of the evidence is conflicting.

When the patient relapses what is there to do? It turns out that thalidomide will rescue more than a third of patients. If thalidomide is so useful why not add it in at the beginning and continue as long as possible? This was the trial reported in NEJM. A comparison between thalidomide all the way through or only later. Sure enough, patients receiving thalidomide had a greater number of complete remisions and longer lasting remissions, but unfortunately the overall survival was no better. The only difference was that they had to suffer the side effects of thalidomide throughout their illness: tiredness, peripheral neuropathy, constipation and venous thrombosis.

In chronic diseases like myeloma and CLL, trialists have been reluctant to look at overall survival as the end point of trials because it takes so long to get an answer. Instead, surrogate endpoints like remission rate and progression free survival are chosen. It suits pharmaceutical companies because if they wait 15 years for overall survival their patent will have run out before they are allowed to sell the drug.

In the recent CLL4 trial in the UK comparing Fludarabine with Fludarabine + cyclophosphamide (FC) with chlorambucil, FC gave the highest response rate and the longest progression free survival; but the overall survival for the three arms is so far not different. The same is true of the German F v FC trial.

If you look at the chlorambucil arm pf CLL4, which has remained the same in UK trials since the 1970s, there has been a progressive improvement in 5-year survivals over that time, from 46% to 62%. Part of this might be ascribed to better supportive care, antibiotics, growth factors, platelet transfusiohns and the like, but the most important factor has been the availability of a different drug to use when the patient relapses, namely fludarabine. It has been suggested that there will never be a difference in overall survival in CLL4 because it will be possible for the chlorambucil patients to crossover to FC or even FCR.

The immediate answer is why not? We are not playing a game that has to be fair for both sides. We are seeking the best treatment strategy for patients. There will be some patients who will be fine on chlorambucil, who may never need to be put at the risk of fludarabine immunosuppression. For the rest, if the risk of relapse is allayed by certainty of a response from FC or FCR why not have two bites of the cherry. Of course , the best trials would compare treatment sequences.