Doxycycline for Creutzfeldt-Jakob disease: a failure, but a step in the right direction

Comment Doxycycline for Creutzfeldt-Jakob disease: a failure, but a step in the right direction Published Online January 8, 2014 http://dx.doi.org/10.1016/ S1474-4422(14)70001-8 See Articles page 150 No treatment trial so far has stopped or slowed the progression of prion diseases—the most rapidly progressive forms of neurodegenerative disease. 1–3 The most common form of prion disease—sporadic Creutzfeldt-Jakob disease—typically progresses rapidly, with most patients dying in less than a year from onset, often in less than 6 months. 4,5 Results of the trial reported in The Lancet Neurology by Stephane Haik and colleagues 6 that used oral doxycycline to treat Creutzfeldt-Jakob disease are no exception to previous study outcomes, with the drug showing no survival benefi t. Nonetheless, it is an important study because it showed that multinational collaborations, rapid enrolment, and rigorous study methods can be achieved for treatment of a rare, rapidly progressive dementia. Several aspects of this study deserve commendation. That two countries with diff erent medical systems were able to combine forces in a treatment trial is remarkable. Despite major diff erences in how patients were enrolled, such as exclusion on the basis of disease duration in Italy and involvement of many local study sites in France, the trials were suffi ciently similar to allow data to be combined for primary and secondary analyses. Additionally, the way in which the French cohort was enrolled—with rapid administrative procedures to enable local referring centres to become study sites— could serve as a model for future trials, because this method enabled a greater number of patients to be enrolled in a fairly short timeframe. Furthermore, in Italy, the investigators were able to use rapid genetic analyses to stratify patients for randomisation. The investigators also showed that patients with Creutzfeldt-Jakob disease can be enrolled in a randomised, placebo-controlled trial without off ering the possibility that all patients might eventually receive active drug. In our sporadic Creutzfeldt-Jakob disease trial of quinacrine, 2 we off ered open-label treatment to patients returning for their 2 month study visit because we were concerned that patients would not enrol if they were not eventually off ered active treatment. This turned out to be a weakness of our trial design, because the 2 month randomised phase might have been too short. 2 The results of this doxycycline randomised trial strongly contradict observational data from Italy and Germany suggesting that doxycycline prolonged survival in patients with Creutzfeldt-Jakob disease. 7,8 An important lesson from the trial, as the investigators allude, is that observational or historical series data should not be relied upon to determine whether treatments are effi cacious; all too often, the biases of observational studies become apparent when their fi ndings are tested with appropriate scientifi c methods in randomised controlled trials. Although this trial off ers much to emulate for future trials, it also had shortcomings, including the use of a drug without compelling preclinical effi cacy. Many complex factors need to be considered when choosing drugs for prion trials. Many compounds, including doxycycline, are eff ective in vitro or in in-vivo animal models when given before, with, or immediately after prion inoculation, but are ineff ective when given later in the disease course. 9–11 This observation is not analogous to Creutzfeldt-Jakob disease, in which patients are already symptomatic. Two compounds have shown signifi cant effi cacy in animal models when given at midpoint of incubation or at or around symptom onset, 12,13 but these drugs were either too toxic or failed even in observational studies. 13–16 Also, the route of administration might be important. The only doxycycline animal study that showed even minimal benefi t once animals had begun to show symptoms used intraventricular liposomal delivery, 17 not oral delivery, which was used in this trial. Additionally, drugs reported to be eff ective against one prion strain or in one animal genetic background are often ineff ective against other strains or in other animal backgrounds. 9,18,19 Many drugs, including doxycycline, are tested against only one or a few strains of prion and in animals from one or a small number of genetic backgrounds. 10,17 Recent research also suggests the development of drug-resistant prion strains with monotherapy. 18 Because of drug resistance and the number of human prion strains, some have suggested that future treatment might require multidrug therapy. 18 Testing of compounds at a more meaningful point, such as midway through the incubation period www.thelancet.com/neurology Vol 13 February 2014