Beyond bridging: How the FDA is changing the way it looks at foreign clinical data and diversity

Many sponsors of new drugs and biologics find themselves asking how they win FDA approval using data from foreign countries. But while the FDA has long been accepting of certain types of foreign data, it’s becoming increasingly concerned about certain practices, including one coined last week by the FDA’s Richard Pazdur, the director of its Oncology Center of Excellence: “tunneling.”

How do things work now?

To obtain FDA approval for a new drug or biologic requires data from clinical trials, but not exclusively from patients or trials in the United States. In fact, many clinical trials rely upon data obtained from sites located outside of the United States. For example, the ClinicalTrials.gov database maintained by the National Institutes of Health (NIH) continually updates high-level data on its listings. As of March 8, 2024, the site listed 490,272 studies with locations in all 50 States and in 223 countries. Of those, 30% of studies were registered in the US alone, while 54% of studies were registered in non-US countries only, and another 5% were registered both in the US and other countries (11% of listings did not include this information).
Among the most important questions regulators must ask when assessing clinical data is: To what extent is the clinical data reflective of the population in which the product will be used? Put simply, regulators must feel confident that if a drug is used in a population, it will be as safe and as effective as it was observed to be in the clinical study in which the product was tested. If a drug was tested in a non-representative population – such as a drug tested only in men, or only in a certain racial group – the risks of the drug having unknown or differential effects increases.
But ensuring the representativeness of a study population relative to the U.S. population can be made more difficult when the data were collected (in whole or in part) from a foreign country. The U.S. is more racially and ethnically heterogenous as compared to many other countries, which can make obtaining representative data from outside of the country difficult. But even beyond just the ethnic or racial makeup of the local population, foreign countries also may have different standards of clinical care which may affect the results of a clinical trial.
Traditionally, regulators have endorsed a strategy of “bridging” study data from one population to another. As its name implies, bridging refers to finding ways to connect one population of clinical evidence to a second population. For example, the International Council for Harmonization (ICH), a global regulatory harmonization group of which the FDA is a founding member, maintains a harmonized guideline known as the E5 guideline (which was expanded upon in a 2006 Q&A document) intended to address ethnic factors in foreign clinical data. As the Q&A document explains: “E5 says, in general, that if the data developed in one region satisfy the requirements for evidence in a new region, but there is a concern about possible intrinsic or extrinsic ethnic differences between the two regions, then it should be possible to extrapolate the data to the new region with a single bridging study.” A bridging study could involve a second clinical trial. Alternatively, it might only involve a pharmacodynamic study or a dose-response study meant to assess more limited concerns.
To avoid needing to conduct a series of bridging studies (or to minimize their use), companies may choose to make use of a multi-regional clinical trial (MRCT), which reduces the need for bridging by providing more “robust evidence,” according to another ICH standard known as E17. MRCTs are generally defined as a clinical trial conducted in multiple countries or regions under the same protocol and intended to support a future submission for approval to a regulator.

What’s the context?

If the FDA feels that a trial is not sufficiently representative, it may choose not to approve a product. One recent and notable example of this is sintilimab, an anti-PD-1 antibody developed by Eli Lilly and Innovent Biologics, which was under FDA review in 2021-2022. The drug’s application proposed an indication for “first-line treatment of patients with Stage IIIB, IIIC, or Stage IV non-squamous non-small cell lung cancer (NSCLC) with no epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinase genomic tumor aberrations in combination with pemetrexed and platinum-based chemotherapy.”
The clinical trial supporting the sintilimab application was conducted in 48 hospitals, all located in China. The FDA offered its perspective on the application and its data in a 2022 commentary in Lancet Oncology. In the commentary, entitled “Importing oncology trials from China: a bridge over troubled waters?”, FDA Oncology Center of Excellence (OCE) Director RICHARD PAZDUR and OCE Division Director HARPREET SINGH called for a balance between the need for innovation and generalizability of foreign clinical trial data. “The degree of regulatory flexibility in establishing the acceptability of data from a single country and its generalizability to a new population should be balanced against the drug’s innovation,” they wrote, indicating that regulators may not accept this data without a without a compelling reason and, further, that bridging studies may be insufficient in some cases. Further, the paper stressed the need to ensure established data reliability and accuracy, citing a 2016 paper which concluded that “ 80% of China’s clinical trial data are fraudulent.” [ Read AgencyIQ’s full analysis of the paper here.].
A February 2022 meeting of the FDA’s Oncologic Drugs Advisory Committee (ODAC) convened to review sintilimab’s application reached similar conclusions. Committee members at the meeting largely concurred with the agency’s assessment of the clinical study supporting sintilimab, with Pazdur asserting at the meeting that single-country trials exhibit “lack of diversity by design.” Ultimately, Eli Lilly reported in March 2022 that the application resulted in a Complete Response Letter (CRL) – an FDA response indicating that the application does not yet support approval. According to the press release, the letter recommended the firms conduct, “…an additional clinical study, specifically a multiregional clinical trial comparing standard of care therapy for first line metastatic NSCLC to sintilimab with chemotherapy utilizing a non-inferiority design with an overall survival endpoint.” [ Read AgencyIQ’s full analysis of the meeting here]

What’s new?

Now Pazdur, the FDA’s top oncology products regulator, is expressing concern that companies are engaged in a practice called “tunneling” that risks compromising FDA’s trust in clinical trials. In remarks made during a “fireside” discussion with AgencyIQ at the American Association for Cancer Research (AACR) inaugural Oncology Industry Partnering event on April 5, 2024, Pazdur explained that bridging studies originally functioned as “innovative processes” that worked when “moving from a heterogeneous population in the West to a homogeneous population,” such as one in Japan. In his view, the process does not work in the opposite direction – when going from a homogenous population to a heterogenous one. This “tunneling” practice involves companies that “take drugs that were already approved and that indication was approved in the United States, and repeating those drug trials in China and then bringing that data back to the United States,” Pazdur explained; this was the situation with sintilimab. [ Read the full transcript of AgencyIQ’s interview with Rrichard Pazdur here].
“The reason why we call it ‘tunneling’ is because when you tunnel, you could actually erode confidence,” Pazdur continued. He also argued that this practice fails to reward risk or innovation, erodes trust in the data, and raises ethical concerns of patients receiving suboptimal therapies in other countries. “[O]ne of the foundations of a company doing a trial and not knowing the results is that they are being rewarded for risk,” he said. “Well, those companies that were doing these trials in China were not taking any risks. They actually knew the results, basically. So that would undermine, really, companies really taking that risk in the future.”
A push for diversity and equity from the FDA means that regulators are trying to determine how to equitably apply these standards as well. Pazdur acknowledged that pushing for increased diversity while approving drugs based on data only from China “would be somewhat disingenuous. […] We can’t be speaking about diversity and then be approving drugs with no U.S. patients in it.” The OCE had published a policy in April 2022 on “ Diversity Plans to Improve Enrollment of Participants From Underrepresented Racial and Ethnic Populations in Clinical Trials,” in which it recommended that sponsors develop a Race and Ethnicity Diversity Plan to “enroll representative numbers of participants from under-represented racial and ethnic populations in the United States.”
According to Pazdur, multi-regional clinical trials (MRCTs) are perhaps the easiest solution to issues related to the use of foreign, “tunneled” data – and also help boost FDA confidence in foreign data. “Having a multi-regional trial brings in countries that we don’t have regulatory experience with. For example, we don’t have a lot of regulatory experience with clinical sites in China. So bringing them into a multi-regional trial, we will then take a look at this data and interrogate it, because we get the raw data,” Pazdur explained, noting that they’re looking for “consistency, not necessarily differences” to build confidence.
Pazdur also emphasized one important aspect of the sintilimab experience related to the clinical standards used in some countries. For sintilimab, the drug’s sponsors tested the product against a control group including a placebo paired with chemotherapy – the standard of care in China, where certain FDA-approved drugs are not legally marketed. Pazdur faulted the companies for “taking a population and exposing them to a drug, a control arm, when you know that was suboptimal.” Pazdur said these sorts of practices of “giving sub-optimal therapies to patients under the disguise that it is not approved in their country” is only serving to “erode confidence in the clinical trial system.”

What are the implications?

It’s worth noting that a lack of diversity isn’t an absolute barrier to FDA approval. Rather, data need to be understood in the context of the drug’s intended use and population of use.
A few examples pulled from the 51 new molecular entities the FDA approved in Fiscal Year 2023 illustrate this point. The majority of trials supporting these approvals enrolled patients in the U.S.; however, this was not always the case. AgencyIQ analyzed information in approval letters and review packages posted to the Drugs@FDA database to develop the following analysis.
FDA approved a handful of novel drugs last year based on non-US data due to small disease populations. For example, Lamzede (velmanase alfa-tycv, Chiesi), indicated to treat non-central nervous system manifestations of alpha-mannosidosis (AM), included zero U.S. participants in its development program of 25 patients. According to the review, “The review team assessed that using this European data to support velmanase approval in the US was acceptable. The etiology of AM, a deficiency of alphamannosidase, is the same across all populations worldwide. And the patient ages included in the trial (6 to 35 years old) is representative of the AM population in the US.”
One application was approved with limited U.S. participation – but more diverse regional representation: Xacduro (sulbactam, durlobactam, Entasis Therapeutics) was approved to treat hospital-acquired bacterial pneumonia and ventilator-associated bacterial pneumonia caused by susceptible isolates of Acinetobacter baumannii-calcoaceticus complex. The clinical trial included just one U.S. patient out of 180 participants. At an FDA advisory committee meeting on the application, the sponsor cited both COVID-19 pandemic condition complications, but but emphasized, “The challenges with enrolling in this trial were mainly focused on the underlying condition, which you can’t predict. You don’t know when you’re going to have a case of multidrug resistant Acinetobacter.” [ Read full AgencyIQ analysis of that meeting here.].
Only one drug with a cancer indication was approved based on non-US data: Loqtorzi (toripalimab-tpzi, Coherus Biosciences). The drug, approved to treat recurrent or metastatic nasopharyngeal carcinoma, included zero U.S. participants in its development program of 461. The two Phase III studies supporting its approval were conducted exclusively in Asia, with one exclusively in China. The review materials include a memorandum signed by OCE’s Deputy Director PAUL KLUETZ offering additional insight into the agency’s decision-making. The agency acknowledges that “…the patient population in the trials is not reflective of the racial and ethnic diversity of patients with NPC in the United States; in addition, patients with NPC with keratinizing histology are significantly underrepresented relative to the expected incidence of this histology.” That being said, the memo states that “…a flexible regulatory approach is warranted given the high unmet need and lack of FDA-approved therapies for US patients with NPC.” In addition, the approval was accompanied by a post-marketing commitment (PMC) to characterize Loqtorzi’s safety and efficacy in a population “whose demographic and disease characteristics are more representative of the population of U.S. patients.” The commitment is currently listed as “pending” per the agency’s database, with a projected completion date of June 30, 2028.
What does all of this mean? While there’s an increased focus on the data used to support regulatory applications reflecting the U.S. population, that doesn’t necessarily mean that the data needs to be generated entirely in the U.S. For example, OCE Deputy Center Director MARC THEORET has previously pointed to “key opportunities [in] increasing enrollment in Sub Saharan Africa, for example, [and] in Latin America” to generate OUS data that may still be applicable or relevant to the U.S. population.
And in cases where FDA is willing to approve a product despite concerns about representation, it is willing to make use of post-approval requirements. In recent years, FDA has included diversity requirements in dozens of required post-approval studies, including with respect to age, sex, race and ethnicity. For example, a review of the FDA’s post-marketing requirements database – a catalogue of all required studies that companies must conduct as a condition of approval – includes at least 46 examples of companies asked to assess the effects of their products in racially representative groups of patients. For example, Janssen Biotech’s TECVAYLI, approved in October 2022, was directed to “enroll sufficient numbers of racial and ethnic minority patients and older patients (ages 65-74 and 75 and above) to enable an evaluation of teclistamab in a study population that better reflects the U.S. population of patients with multiple myeloma.”
In addition, we are likely to soon receive additional insights from FDA about its approach to clinical trial diversity and foreign data. As part of a federal law passed in late 2022 known as the Consolidated Appropriations Act, the FDA was required to revise or finalize a guidance document on clinical trial Diversity Action Plans (DAPs). While the FDA missed its statutory deadline for releasing this guidance by the end of 2023, the document was recently sent by the FDA to the White House’s Office of Information and Regulatory Affairs (OIRA) – the last step before the FDA can legally publish the guidance. A senior FDA official who agreed to speak off-the-record told AgencyIQ that the document is a revised draft guidance, which means that the public and industry would have an opportunity to comment on the document before it is made final.
Under the FDA’s Diversity Action Plan policy, which is generally described in the Consolidated Appropriations Act (Section 3602), sponsors of certain research programs for both drugs and medical devices would need to justify their approach to recruiting (and retaining) a population in their research studies that reflect the expected U.S. patient population specifically. While the agency has been accepting these plans on a voluntary basis under a draft guidance document from 2022, the new revised draft guidance document is expected to lay out the parameters of when these plans are required. It is expected to also describe the conditions under which companies would be able to seek a waiver of federal requirements. So far, the FDA’s early experience with the voluntary plans (and from a limited subset of submissions) show that, so far, the FDA often has had feedback for sponsors related to how they’re setting and justifying their enrollment targets.
For sponsors of clinical trials, the FDA’s more pointed focus on trial diversity does run some short-term risks. Because trials can take years to plan and execute, it’s possible that there are in-progress clinical trials now which may be insufficient to support the approval of a future product. For those companies, they may be asked to run additional clinical studies, just as Eli Lilly was asked to for sintilimab, or additional post-marketing obligations.

To contact the authors of this item, please email Amanda Conti ( [email protected]), Laura DiAngelo ( [email protected]) and Alexander Gaffney ( [email protected])
To contact the editors of this item, please email Kari Oakes ( [email protected]) and Alexander Gaffney ( [email protected])

Key Documents

Transcript: AgencyIQ interview of OCE Head Richard Pazdur at AACR