Clinical trials in CNS indications are notoriously difficult to recruit patients for. Screen failure rates greater than 100% are becoming more and more common. For example in the publication of the phase III trials in Bapinizumab [i] clinical trials, there were 3,855 subjects screened to randomise 1,121 subjects in the first trial and for the second trial, 2768 subjects were screened to randomise 1,331 subjects. And this was after an initial screen had rejected 746 prior to genotyping. So in total 7,369 subjects were screened to enrol a total of 2,452, a staggering screen failure percentage of over 300%. Why is this and why is it more difficult than for other indications? Here are a few top tips to consider if you are a sponsor with a pivotal study in a CNS indication.
Top tips for recruitment:
The protocol must match the patient population. Although this should really go without saying, there is a tendency to design study entry criteria that do not match the patient population. Detailed feasibility, involving the study sites, is a must to ensure that overly narrow medical parameters will not exclude patients unnecessarily. This is particularly important for studies using an elderly population such as an Alzheimer’s disease trial that could have many comorbidities. A selection of the issues we have encountered includes a restrictive vital sign, creatinine clearance and QT limits.
Think about how you will achieve patient recruitment. You could rush to get patients by opening lots of sites, which only contribute 1 or 2 patients, and then risk drowning a positive result in site variability. On many levels, this is an expensive way to get your recruitment. Our preferred strategy is to aim for a smaller number of carefully selected sites that have a clear patient recruitment and retention strategy, with additional referral or funnelling of patients to these sites. Close monitoring and active management of these sites help a positive result to surface.
Often it is the less well-known sites that are highly motivated to deliver the patient recruitment. This is true even in western countries such as the UK and US. But a high number of screen failures are demoralising for the site staff, so we recommend listening to the sites and giving early consideration to adjustments to the protocol entry criteria to match the ‘real’ study population, rather than persevering with a trial design that would suggest it is only going to capture a smaller portion of patients.
It often goes without saying but the patient population for CNS indications can be very ill with additional complications, for example, cognition in Parkinson’s disease, behavioural issues in dementia, mobility in Huntington’s disease and mental health issues in schizophrenia and bipolar disorder, so this patient population is very vulnerable. In order to assist with recruitment and in particular, retention of subjects (missing data is another problem area that we’ll save for another time), the study schedule and level and type of assessments should take account of the needs of the patient, rather than thinking of them in the same way as for paid Phase-I subjects.
The Golden Rules
In summary, in CNS clinical trials there are some golden rules that must be obeyed:
Feasibility should include comments on the study design. Inclusion/exclusion criteria and the schedule of assessments being the main areas to get real feedback from experienced sites.
Only set up sites that will recruit. Site selection is absolutely key. These may be sites that are not well known in terms of publications.
Make sure all sites contribute as many patients as possible. Recruiting a small number of subjects at each site just adds variability to the data.
Ensure there is an overall study patient recruitment and retention plan which is then translated into a detailed site level plan that is agreed with each study site. Make sure the plan is not only set up but monitored and acted upon throughout the study.
[ Supplement to: Salloway S, Sperling R, Fox NC, et al. Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer’s disease. N Engl J Med 2014;370:322-33. DOI: 10.1056/NEJMoa1304839 http://www.nejm.org/doi/suppl/10.1056/NEJMoa1304839/suppl_file/nejmoa1304839_appendix.pdf