Trials of orphan drugs often involve smaller sample sizes. How much smaller? Cut the typical phase 3 sample size by 75% and then, for many orphan trials, cut a lot more. The Evaluate Pharma Orphan Drug Report 2013 provides valuable data on sample size in phase 3 orphan trials. While the median size was >500 subjects, some orphan drugs received approval based on trials involving only 34 subjects (Factor VIIa), 36 (Von Willebrand factor ) or 65 (Factor VIIa). The most recent example is Juxtapid, a drug for homozygous familial hypercholesterolemia, which received approval in 2012 based on a phase 3 trial involving 54 subjects.
How big a sample size will you need for a pivotal trial of your orphan drug? There is no rule of thumb, no design you can pull off the shelf. Selection of an appropriate design and sample size for an orphan trial is case-by-case. Designs may use frequentist or Bayesian statistical methods. Traditional parallel group designs are sometimes possible, but other types of designs abound, especially in early phases, including crossover, N-of-1, delayed start, stepped wedge and randomized withdrawal. N-of-1 designs, such as an early-phase design testing a treatment of cystic fibrosis variants, can quickly demonstrate any effect that is going to happen. In one N-of-1 trial, Bayesian analysis of aggregate data provided an excellent basis for selecting target populations for subsequent studies.
With so much variety and uncertainty, you need access to relevant design and statistical expertise to identify a design that will work for your drug given its unique circumstances. Regulators are open to discussions of alternative approaches. Take them up on the offer and be sure you understand everything said. Sadly, many small and medium companies miss some of the import of regulatory input. EMA officials report:
“. . . large companies not only asked for scientific advice more frequently than medium-sized and small pharmaceutical companies but, importantly, they were significantly more compliant with the scientific advice given than their smaller peers.”
Orphan drug experience at the FDA is similar: Big companies fare better. Anecdotal reports suggest that small and medium companies simply lack the experience required for a full understanding of what the regulators are telling them about their programs. Regardless of the size of your company, don’t walk into regulatory meetings alone or unprepared. You need experienced advisors on your side of the table.
Also, be prepared for operational challenges in an orphan trial. Patient registries may make it easy to find subjects motivated by serious unmet medical needs, but that’s not the whole story. Geographic dispersion of a small population increases the risk of an intolerable lag between events in the field and awareness by the study team, especially if you are using a typical EDC system and CTMS. Every subject in an orphan drug trial is precious. You need rapid feedback to identify and address errors quickly and ensure collection of complete, valid data on every subject.
One of the exciting things about orphan trials is the opportunity for creativity in operations as well as design. For example, a CF trial followed patients in part by placing spirometers and other devices in their homes and transmitting data to the study database through a custom wireless hub. Remote data collection may also extend to patients and caregivers’ completing questionnaires and assessments. Remote data collection has the potential to reduce the number of patient visits required to conduct a trial. This is an exciting prospect, but reliance on remote data collection makes it even more important to know exactly what is going on in your trial every day.
What does all this mean for you and your orphan drug? It’s simple:
- Find people with the experience and expertise to design your studies
- Meet with regulators early and bring your experts along – level the playing field with big competitors
- Be sure you fully understand and comply with regulatory guidance
- Select a CRO with systems that enable early detection and correction of problems in the field.