Data without borders: Assessing China-only evidence for EMA approval – insights from ex-EMA and seasoned subject matter experts

As Chinese biopharmaceutical innovation accelerates, a critical question confronts global sponsors: Can China-only or China-heavy clinical data truly support approval in Europe? 

While the European Medicines Agency (EMA) has demonstrated openness to foreign data, the threshold for acceptance is evolving and far from guaranteed. Expectations around data quality, global applicability, standard of care, and regulatory strategy are becoming increasingly stringent.

On this expert-led webinar facilitated by Dr. Sinan B. Sarac, former Committee for Medicinal Products for Human Use (CHMP) members from the EMA, Dr. Sinan Sarac, Dr. Aaron Sosa and Dr. Simona Stankeviciute, joined Dr. Matthias Kruse to explore when China-generated data can be sufficient, when it falls short, and how to design development programs that meet European requirements from the outset. Drawing on recent approvals, regulatory precedents, and real-world case studies, the panel of experts provided practical insights to help companies de-risk their global strategy and avoid costly rework.

Regulatory precedents: EMA approvals based on China-only datasets

Dr. Aaron Sosa presented a comprehensive timeline of EMA approvals based on China-only datasets, revealing that while precedent exists, success is far from guaranteed. Since 2022, approximately 15 approvals have been granted, predominantly in oncology, particularly for PD-1/PD-L1 inhibitors. Notable successes include Cejemly (sugemalimab), Tevimbra (tislelizumab), Loqtorzi (toripalimab) and Hetronifly (serplulimab), which replicated study designs and safety profiles consistent with established checkpoint inhibitors already approved in the EU.

However, not all applications succeeded. Sevsury (surufatinib), a tyrosine kinase inhibitor intended for patients with progressive neuroendocrine tumours, was withdrawn following several rounds of review. Fruzaqla (fruquintinib), also a tyrosine kinase inhibitor, was approved for metastatic colorectal cancer but only in the fifth line setting, rather than the requested fourth line indication. These outcomes contrast with the generally more favorable reception of PD-1/PD-L1 inhibitors and underscore the CHMP’s more cautious assessment of products that lack well-established regulatory precedents in Europe. 

Key takeaways: EMA is not lowering evidentiary standards but showing greater flexibility regarding geographic representativeness in selected cases. Success depends on the totality of evidence, including data quality, GCP compliance, clinical benefit, unmet need, and CHMP familiarity with the product class. Products supported by development programmes that align with established EU regulatory precedents generally face less uncertainty than those employing novel mechanisms or unfamiliar development paradigms.

Usage of solely foreign data for non-malignant indications: Can it be done?

Dr. Simona Stankeviciute emphasized that oncology's success cannot be directly extrapolated to non-malignant indications due to fundamental differences in endpoint objectivity and population sensitivity. While objective endpoints like disease cure for example in infectious diseases or weight loss are more readily accepted from foreign data, more subjective measures, clinical outcome assessments and patient-reported outcomes, introduce significant challenges.

Extrinsic factors largely influence acceptability: medical practice differences, concomitant treatments, disease definitions, standard of care, and even dietary factors can bias the results. For example, hypertension definitions differ between the EU, US, and China; psychiatric treatment resistance is defined differently across regions; and traditional Chinese medicine use may influence neurology outcomes in ways that cannot be fully quantified.

Through three real-world scientific advice cases, Dr. Stankeviciute illustrated CHMP's hesitancy with China-only data for non-malignant indications. In one neurology case, certain symptoms were not assessed in the Chinese dataset, preventing CHMP from confirming population equivalence. CHMP suggested including at least a small European cohort to verify comparable treatment effects. Notably, CHMP also requested bridging Pharmacokinetics (PK) studies for US-only data, indicating geographic flexibility is not limited to China but applies broadly to any foreign data.

Key takeaways: A robust clinical study design is essential for success regardless of geographic location. It is important to ensure that inclusion/exclusion criteria, diagnostic criteria, disease staging, endpoints, comparators, and concomitant treatments are similar to or extrapolatable to EU populations. Engaging with the CHMP early in the development process is critical.

From China to Europe: Can clinical pharmacology close the evidence gap?

Dr. Matthias Kruse provided the clinical pharmacology perspective, grounding the discussion in International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) E5 guidance^1,2 on ethnic sensitivity assessment. EMA has followed the principles of ICH E5 guidance ever since and accepted foreign clinical data, though rarely as the only data source while European data remain preferred. The framework distinguishes between ethnically insensitive drugs (which may bridge without additional studies) and those requiring bridging information due to differences in medical practice, background therapies, study design, or drug class novelty.

The Aumseqa (Aumolertinib) case exemplified this complexity. Although the Phase 3 trial was China-only with limited US patients in earlier phases, CHMP required a dedicated PK bridging study in European Non-Small Cell Lung Cancer (NSCLC) patients to compare exposure between Asian and non-Asian patient populations. The study revealed Cmax was approximately 15% higher in Europeans, but other parameters were comparable and clinically irrelevant. This additional PK data, combined with literature extrapolation for the different comparator CHMP expected, ultimately supported approval at the original dose and allowed the applicant to extrapolate the safety and efficacy from one region to the other without additional clinical efficacy trials. However, whether a single PK study can bridge from a different geography cannot necessarily be generalized from this example and remains a case-by-case decision. In addition, adequate pre-planning is required and complete data be submitted with an application to be successful in future. In this case the data were only provided after a major objection was raised by the CHMP.

Dr. Kruse emphasized that PK studies in healthy volunteers may be insufficient; patient population PK data is what agency reviewers would look for. Additionally, Pharmacokinetics/Pharmacodynamics (PK/PD) and exposure-response relationships must be fully characterized for the region where the data has been collected, and data quality, completeness, and clarity are non-negotiable.

Key takeaways: You should conduct thorough intrinsic and extrinsic factor analysis early with a pre-defined plan to address the requirements set out in ICH E5 guidance. It is recommended to include non-Asian patients in first-in-human trials if possible, as these trials are less expensive than larger Phase 2 or Phase 3 trials but still can provide essential information on PK, PD, safety and efficacy in a foreign population. You should present a strong bridging strategy in the dossier, including PK/PD analysis and population PK. Clarify expectations with CHMP is necessary before submission. While a small PK or PD study may suffice, success remains case-by-case.

Panel discussion highlights

The panel addressed critical scenarios: China-only confirmatory trials carry high regulatory risk and typically require global randomized controlled trials for conditional approvals. Rare and ultra-rare diseases in the oncology and non-oncology setting are granted significantly greater flexibility, particularly when endpoints are agreed by the CHMP and treatments are not ethnically sensitive nor background therapy significantly differs between regions. In the case of a more fluid endpoint, lesser unmet need therapeutic area, less flexibility will likely be applied. Bridging PK studies are a helpful tool to de-risk programs but may not be sufficient alone, with patient population PK data preferred over healthy volunteer studies. Cell and gene therapies developed solely in China face significant scrutiny due to higher expectations for manufacturing consistency, comparability, and long-term safety.

Conclusion

The webinar revealed a nuanced regulatory landscape: The EMA has shown greater flexibility toward China-only data in selected oncology settings, but approval remains highly case specific. Success depends on the totality of evidence, including data quality, clinical benefit, GCP compliance, unmet need, and alignment with established regulatory precedents. Therapies that deviate from familiar European development models, including novel mechanisms, non-malignant indications, and Advanced Therapy Medicinal Products (ATMPs), are likely to face greater regulatory scrutiny and would benefit from early engagement with regulators.

Resources

1. ICH E5: https://database.ich.org/sites/default/files/E5_R1__Guideline.pdf
2. ICH E5 Q&A: https://database.ich.org/sites/default/files/E5_Q%26As__R1_Q%26As.pdf

Return to Insights Center