My colleagues, Barb Geiger and Chris Wang, recently held a webinar, “Demystifying Asia: Best Practices For Conducting Multinational Clinical Trials,” which took a deep dive into the why, what, and how of enrolling patients in Asia.
After the webinar, they were asked: What percentage of patients can be enrolled from Asia for oncology studies where the primary focus is for registration of the product in the US? Is there a precedent in a multinational study where data from Asian patients have contributed to US product registration? What percentage of patients was registered in the Asia?
According to Barb, she has seen US product registrations where more than 20% of the patients came from Asia. In fact, she was involved in one study, which received FDA approval for the compound, that only had 3 out of 700 total patients enrolled from the US. To her understanding, there is no limit on the number of patients that come from any one particular global region for the registration of a product in the US.
I agree with Barb; I am not aware that the FDA has a specific percentage of acceptable foreign data for a pivotal study. In fact, technically, they can accept all foreign data for an NDA/BLA. The relevant guideline for this topic is ICH E5 Ethnic Factors in the Acceptability of Foreign Clinical Data, which explains that the sponsor will have to demonstrate:
- The enrolled population is representative of the US patient population.
- The medical practice is representative of US medical practice (diagnosis criteria, other treatments received such as surgery or first line treatment, concomitant medications, etc.).
While not tied directly to whether the data are from the US or foreign sites, the statistical analysis plan (SAP) pre-defines what result is considered a “win” (meets the endpoints) in a clinical study. The SAP for a clinical study is part of the critical negotiations with the FDA, especially for a pivotal study. Parameters for the SAP include: sample size, effect size, superiority/non-inferiority, stats model, p-value, and statistical power. This plan is submitted to the FDA prior to the study becoming unblinded.
The SAP will usually group all subjects who received the same dose/treatment together for analysis irrespective of country of origin of the data. However, you also have to test the rigor of the data to prove there was not a special situation or specific groupings that affected the results more than they should.
For example, you’d need to perform analysis by foreign data vs. US data, results per country, by ethnicity, by sites with the most enrolled patients, by most protocol deviations, best and worst results, most adverse events, ITT vs. per protocol population, and so forth. This is all a part of the medical and statistical justification that the data from foreign sites may be included in the study and NDA—that the enrolled population is representative of the US patient population, and that the medical practice is representative of US medical practice.
It is required for all foreign sites and US sites to conduct studies according to good clinical practice (GCP) and to the study protocol. The FDA will inspect pivotal studies, and in particular, they will inspect most (if not all) foreign sites that contributed data to pivotal studies. During inspections, the FDA ensures all data from patient files match the database submitted in the NDA, and double-checks that the site followed all regulations, including GCP, IRB/Ethics Committee, Informed Consent, and US requirements. It is up to the sponsor and the CRO to monitor the studies correctly and prepare the sites to be inspection ready.
The FDA will not accept the data or the interpretation of the data until they have verified that there is no reason to question any of it. If they find reason to doubt, they have been known to throw out data from certain patients or from entire sites.
Because of previous experience, the FDA may look at foreign sites, and specifically sites in certain countries, with greater scrutiny. I’ve heard lots of stories about the FDA throwing out foreign data from a pivotal study, but the fact of the matter is those instances had nothing to do with the site being foreign. In those instances, it was an issue of bad data, poor site management, lack of trained site personnel, poor monitoring, etc. Sponsors and CROs should be aware of the potential for greater scrutiny and manage sites accordingly.
For more information on managing multinational clinical trials in Asia, check out the recorded webcast and be on the lookout for the related eBook and infographic. If you’d like to be alerted when they become available, please leave your contact information in the comments section.
Clara Li is vice president of regulatory affairs and strategic development at Clinipace.