China Investigator-Initiated Trial (IIT) Spotlight Part 3: What does the future hold?

A readout of what we learned 100 days after 818, and what we expect for the future

Almost all news coverage so far has focused on the pre-818 era, but the landscape is changing. So we should talk more about what comes next.

Let’s start with what we already know.

Order 818 puts IITs under a standardised, regulated framework

Before May, hospital-led research in China sat in an institutional gap; CDE/NMPA had no jurisdiction over hospitals, and the NHC had no equivalent technical review capacity. Guidelines existed, but many areas depended on individualised interpretation and experience. That gap has now closed.

While hospitals are still the main reviewers, the study must be filed nationally with the NHC. There is a clear, standardised framework, with the same principles as IND under FDA/EMA/NMPA. There is a named sponsoring institution that must stand behind the non-clinical evidence. Academic and ethics reviews happen separately, before anything is filed. Non-clinical proof of safety is a precondition of dosing rather than a document assembled afterwards. Participants cannot be charged. Records must be kept for thirty years, and falsification or concealment carries real consequences. Every study that clears review receives a number and appears on a published list.

And it makes the choice of partner, site and PI matter more than it used to

The institutions that qualify under Order 818 are a narrower group than the hospitals accredited for registration trials, and the requirements are more demanding; dedicated clinical research management and quality assurance departments, a designated quality authorised person meeting criteria set by technology class, in principle national-level centre, laboratory or specialty status for the lead institution in multicentre work.

But hospitals alone cannot navigate the new rules. The formal screening pass rate was <6% in May, with the majority of failures due to an experience gap: writing the wrong dossiers, preparing the wrong evidence, having the wrong operations. Regulators have rolled out training across seven key regions, walking hospital teams through areas such as dossier preparation. That improved rates to ~20% in July, and it held there, further indicating the bar is permanently raised. And this is just one of the many pitfalls to navigate.

A partner with the right capabilities, an investigator with the right experience, and a site with the right infrastructure are, in our experience, the entry ticket for participation.

Given the lessons from the past and Order 818, safety will be a key focus for future review

For gene therapies, regulators may expect a 2–4-fold safety window (in some cases 5-fold or more) below the animal safety boundary. Simple allometric conversion would likely be rejected. Intrathecal doses require conversion on CSF volume and brain weight; intravenous doses must account for species differences in tissue tropism and organ and blood-volume ratios relative to body weight. Any case of integrating vectors, DNA editing, first clinical use of a novel delivery vector, weak species relevance, paediatric or critically ill patients, or high-dose exposure of CNS, heart or liver will earn greater scrutiny. For DNA editing products, off-target assessment is expected to be multi-layered and to include whole-genome sequencing, with the FDA's April 2026 NGS guidance cited as the reference. In-hospital preparation is still accepted but will be examined more closely (vs CDMO) given the significant variance across hospitals.

Selected examples of failed attempts shared by regulators (and, just in case you are wondering, none of these are trials supported by Vivara):

Preclinical

• Immunogenicity and immunotoxicity panels too narrow for systemic assessment — incomplete cytokine coverage, complement activation and immunophenotyping omitted.
• A biodistribution finding that flagged a pharmacological safety concern, then was not followed up.
• Efficacy not demonstrated in models, or endpoints inconsistent with the claimed clinical response.
• The same dataset used to support more than one clinical study.

Investigational material

• GMP declaration with no cleanroom layout and no environmental classification report, or a declaring entity different from the preparation facility.
• Release testing missing for the technology class — viral titre for gene therapy, mycoplasma for somatic cell work; incomplete residual limits; method validation absent.
• No storage or transport stability data.
• Third-party verification covering only part of what the manufacturer self-tested, or a third-party contract test report submitted where a verification report was required.
• Key process parameters and raw-material information redacted

Design, consent and recruitment

• Multiple indications bundled into one exploratory study; exploratory work not framed around safety as the primary objective.
• Dose selection insufficiently justified; follow-up too short; weak boundaries on paediatric or mixed-population enrolment.
• Risk assessment and adverse-event handling described in general terms rather than as procedures.
• Failing to disclose medium- and long-term risk, or to state clearly that participation is free of charge.
• Recruitment advertising claiming provincial drug-regulator approval or issued without ethics approval.

IITs serve a strong strategic purpose, but only if they are run correctly. And the bar has just risen considerably.

And with that, we end our Spotlight Series on China IIT. We hope you have enjoyed reading, and as always, we would love for you to reach out and share your thoughts!

Start the conversation