Repurposing a drug for treating a novel disease is difficult, particularly with an injectable therapy. While hospital pharmacies are capable of masking approved tablets to look the same as a placebo tablet, this is not possible with an injectable device. So what does one do? How can one start to take such therapies to clinic with the hope of seeing an effect that merits further investigation in the absence of a large, expensive, properly powered double-blind, placebo-controlled trial? Here, the investigators studied a parallel group of patients that were matched to those in the treatment arm, but did not receive any injectable therapy (1). This approach obviously lacks the rigor of a placebo-controlled trial, but does allow one to draw conclusions about how the intervention fares against best medical therapy in patients with PD at a similar stage of illness.
The standard approach with any therapy is to do a first-into-human phase I study and then move to phase II trials to ascertain whether the agent is well tolerated and exhibits some efficacy. An approved drug may enter the regulatory pathway later in the process. In this trial, Aviles-Olmos et al. demonstrated that exenatide is well tolerated in PD (1); given the large body of safety data already gathered from its use in the diabetes population, exenatide may now perhaps be ready for the phase III, pivotal trial stage of testing. Phase II testing also assesses efficacy, and in the present study, the investigators were challenged to explore whether exenatide might modify the slowly evolving progressive course of PD, rather than simply have some direct symptomatic effect (1). In an effort to reduce investigator bias, Aviles-Olmos et al. used an established procedure in which participants are videotaped and assessed by a blinded third party (16).
In the absence of a placebo group or any reliable peripheral measure that tracks the pathology of PD, how could the investigators at least get some indication of whether exenatide might have a disease-modifying effect? Imaging the dopaminergic pathway using f-dopa PET is extremely expensive (approximately $9,000 per scan), but reliably measures signals over a short period of time. However, functional imaging is not always straightforward, and interpreting changes in f-dopa PET can be difficult (15, 17, 18). Using123I-FP-CIT SPECT scanning may offer a cheaper, more feasible approach, but it lacks the sensitivity to detect gradual changes in the network once the disease process has begun (19).
Aviles-Olmos and coworkers chose to treat participants with exenatide for twelve months, and then reassess them after a washout of two months, with the assumption that any symptomatic effects would have been lost by the time the patient was reevaluated. The study team saw a sustained benefit two months after the exenatide was discontinued, including in a number of motor measures as well as nonmotor measures. Notably, cognitive scores were improved by exenatide (1). Although a sustained symptomatic effect potentially could explain the improvement after exenatide was discontinued, the observations suggest that exenatide may truly have slowed disease progression. However, without a placebo injectable study arm, one cannot entirely exclude that there was a protracted placebo effect after cessation of the drug therapy. Indeed, protracted placebo effects have been seen in some gene therapy trials while patients have still been blinded (20, 21). In the present study, one would predict that if a placebo effect played a major role, it ought to have waned during the final two months, when the patients were aware that they were no longer receiving injections of the active drug.