Therapeutic trials in ALS.

Abstract

A number of facts about ALS are unquestionable. We need more effective drugs to treat the disease. The disease is remarkably variable in its site of involvement and rate of progression from patienttopatient.Ifadrugarrestedprogressionofthedisease,or even reversed it, we would only need one patient to prove this. Unfortunately, current expectations are for the discovery of drugs that only slow the progression. Until now, clinical trials of a drug that has laboratory features suggesting a likely effect on motor neuron death have proceeded along the classical path: phase 1 clinical trials designed to show that the drug is safe; phase 2 trials in ALS patients who show some possible clinical benefit; and phase 3 double blind, randomized, placebo controlledtrialsthatinvolve300patientspertreatmentarm,oreven more, and cost at least $20 million, the cost depending on the amount ofdata collected.Unfortunately,although todate many trials have been undertaken, only those involving Riluzole have unequivocally proved the positive effect of the drug. 1‐3 There has been much research into ways of speeding the process of identifying new drugs for ALS. Recent work has been devoted to trials in murine models of ALS, and in the rapid screening of potential drugs in motor neuron-like cultures. The question of the use of placebo controlled trials in ALS and in other lethaldiseases continues tobe a topicfordiscussion. From thepatient’spoint ofview, if thereisa drugthatmight producea benefit, he wants it now and does not want a trial involving a placebo. This attitude is counter-intuitive in that an unstudied drug is of untried efficacy, and cannot be classed as a therapy. However, patients with ALS are not interested in scientific proof that a drug works, but rather whether they can live longer. This is as true in ALS as it is for cancer and AIDS. Physicians looking after these dire diseases are among the most caring of our profession, and often have the same psychological drive to treat the patient to the exclusion of finding out if the drug actually helps or even hurts the patient. The latter is always possible. For instance, in the Regeneron high dose CNTF trial, patients treated with theactivedrugdeteriorated fasterthanthosetreated with the placebo. 4 For a number of years after that trial was published, the clamor for avoidance of placebo-controlled trials was muted, though we are now back in an era when ALS patients try to obtain every drug that is even suggested to be worthy of trial. Munsat and colleagues showed that Z-scores of data from a populationofALSpatientsdecreasedlinearly,andthattherewas a similar trend in individual patients.5 The implication of their work was that therapeutic trials in ALS could be designed to compare an active treatment group with a database from patients in an observational, natural history study. This would greatly reduce the cost the trial, and avoid some patients having to take a placebo. Unfortunately, this has not proved to be the case. Comparison of the placebo groups in the various controlled ALS trials in the last decade has shown significant differences, presumably derived from different inclusion and exclusion criteria, and the demographic details of the patients. Alternative trial designs allow the patient to act as his own control.Across-overdesigninALSiscomplicatedbythefactthat the patient is more advanced by the time of the cross-over. This might be avoided by collecting data during a natural history phase preceding an active treatment phase. However, this presupposesthatdiseaseprogressionislinearinbothphases,witha similar slope ofdeterioration on thecriticalmeasure,andso still needs a placebo arm if the placebo response is to be avoided. In this issue of the Journal, Bryan et al. revisit the question of the elimination of a placebo arm in ALS trials. 6 Based on data fromtheplacebogroupinthenegativelowdoseCNTFtrial, 7 the authors compare the rate of deterioration during the 3 month ‘‘lead-in’’ or natural history phase with that during the 6-month treatment phase. The authors report that the rate of decrease in FVC was significantly faster in the placebo phase than in the preceding natural history phase, casting doubt on the underlying theory of this type of study. Moreover, there was no correlation between the two phases of the study for the slopes of the leg and FVC megascores. There was a small but significant correlation between the deterioration in the two phases in the arm scores, indicating a smaller variation in the arm data, a fact previously reported. 8 The authors suggest that comparison with natural history database controls might allow exploratory trials to rule out ineffective treatments. They recommend the use of arm strength measurements in such trials. However, they indicate that double-blind placebo controlled trials are still needed to prove drug efficacy and to study safety of potential drugs for ALS, a point I have previously advocated.9 Until we have more effective drugs for the treatment of ALS, trial design will remain an important area of discussion. 3,10‐11