Animal Study Requirements for Medical Device Registration in China

1. Key takeaways

Animal studies are one of the nine non-clinical study types in Article 28 of the Medical Device Registration and Filing Measures (SAMR Order No. 47, published 26 August 2021, effective 1 October 2021). Their position in the evidence set is supplementary, not a default requirement.

• The test for necessity is whether a safety or effectiveness risk cannot be adequately evaluated through physicochemical, electrical safety, radiation safety, software, biological characteristics, sterilisation or stability studies. Where other studies can cover it, the animal study should not be repeated.
• Article 29 sets two pillars, method validation and statistical analysis, which directly govern sample size and endpoint design. Article 31 requires samples to represent the product applied for, including sterilisation state, model and storage.
• Acceptance has clear boundaries: animal studies show mechanistic and local safety and performance, cannot replace clinical evaluation, and do not extrapolate automatically to humans.
• GRZAN’s assessment: problems surface at protocol stage rather than execution. A protocol that states why the study is done, what is observed and how passing is judged keeps execution deviations manageable; a vague protocol is unconvincing even when the study is completed.

2. Legal basis

Administrative regulation. The Regulation (State Council Order No. 739, published 9 February 2021, effective 1 June 2021) provides:

• Article 5: supervision follows risk management, whole-process control, scientific oversight and social co-governance.
• Article 14: seven categories of materials, including risk analysis, technical requirements, test report and clinical evaluation.
• Article 24: two situations exempt from clinical evaluation. Article 25: two routes for clinical evaluation.

Departmental rule. Order No. 47 specifies:

• Article 19: applicable mandatory standards met; inapplicable ones explained with materials.
• Article 24: risk management across the life cycle. Article 28: the nine study types include animal studies.
• Article 29: method validation required; where statistics apply, the statistical method, sample size basis and acceptance criteria must be specified.
• Article 31: test samples must represent the product applied for.
• Articles 33 and 34: clinical evaluation as the principle, with two exemptions. Article 36: the material requirements for the equivalence and clinical trial routes.
• Article 41: deficiencies notified once in full, answered once within one year, otherwise review terminates with a refusal.
• Article 121: IVDs follow the separate IVD measures.

Normative documents. Announcement No. 121 of 2021 sets the dossier catalogue; the QMS inspection guide (notice No. 50 of 2022) defines four inspection outcomes; No. 126 governs self-testing; No. 63 of 2025 clarifies review requirements.

Background data: the 2025 Annual Medical Device Registration Report (NMPA, published 24 February 2026) records 76 innovative devices approved in 2025, up 17 per cent year on year, and 25 priority review devices, up 212.5 per cent. High-risk devices in these channels typically face higher justification requirements for animal work.

Practical estimate: ethics review, animal quality and facility conditions, institution qualification and observation period length have no clause-level device rule; animal welfare legislation, current guidelines and review points govern. Verify qualification and capacity before selection and settle ethics within the protocol.

3. Expert insights

Supplementary evidence, tested through three gates. Article 28 anchors the decision in design requirements and risk management.

Gate two is most often skipped: companies schedule studies because comparable products ran them, while the risk file holds no matching item, so the study and risk analysis never connect.

The protocol answers three questions. Most challenges trace to three omissions: why this study, meaning which risk item it addresses and why other studies cannot cover it; which endpoints, meaning the primary and secondary endpoints and why they reflect the risk under evaluation; and how passing is judged, meaning the criteria per endpoint and their source in standards, comparator data or engineering argument. A protocol listing species, numbers, procedure and observation times cannot show why the study proves safety and effectiveness, however well executed.

Model selection is a comparability argument.

The fifth row is most often omitted and questioned; stating species limits strengthens credibility by showing the boundary of the evidence is understood.

Statistics as a checkable prior design. Article 29 requires the method, the sample size basis and the criteria. Four points apply: the sample size basis, including effect size, significance level and power, and how ethical minimums were balanced against statistical power with reasons recorded; randomisation and control type, whether negative, positive or comparator; blinding and bias control, including independent imaging or pathology review; and handling of deaths and dropouts, including whether they enter the analysis. Only protocol-stage decisions count; a sample size note added afterwards cannot repair the design.

Three boundaries of acceptance.

The third boundary deserves attention. Where a unit contains several sizes, animal work usually covers a representative size with the rest covered by worst-case argument; without a stated basis, the chain for other sizes has a gap.

Sample representativeness. Article 31 means the implanted or used sample is the marketed product. Four breaks recur: samples not finally sterilised, or sterilised by a different method; test sizes inconsistent with the sizes claimed, without a representativeness argument; samples at the start of shelf life with no statement of difference from the end; and samples from development batches with different processes or suppliers. A sample traceability matrix recording identifier, batch, sterilisation state, differences from production and typicality argument closes the gap.

4. Practical recommendations

• Run the three gates before committing, recording conclusions and any decision not to study.
• Open the protocol with the risk item addressed and why other studies cannot cover it.
• Separate primary and secondary endpoints, justifying each against the risk.
• Argue comparability across five dimensions, with species limits in their own section.
• Fix statistics in advance: sample size, randomisation, controls, blinding, dropouts (Article 29).
• Where only a representative size is tested, justify worst-case coverage.
• Verify institution qualification, facilities and ethics arrangements before starting.
• Keep conclusions within the model, period and sample, leaving the rest to clinical evaluation.

5. Verification checklist

• □ Applicability judged under Article 28, with reasons traceable to the risk file?
• □ A risk item requiring animal evaluation identified, with the protocol aligned to it?
• □ Confirmed that physicochemical, electrical, radiation, software, biological, sterilisation and stability work cannot cover the risk?
• □ Primary and secondary endpoints stated with their justification?
• □ Comparability argued across anatomy, physiology, procedure, period and species limits?
• □ Sample size basis, randomisation, controls, blinding and dropout rules in the protocol (Article 29)?
• □ Criteria per endpoint and their source stated?
• □ Samples consistent with production in sterilisation, size and batch (Article 31)?
• □ Worst-case basis stated where several sizes exist?
• □ Institution qualification, facilities and ethics review verified?
• □ Conclusions confined to model, period and sample, without clinical extrapolation?
• □ Protocol, raw records, statistical output and report archived?
• □ Cited standards current; inapplicable ones explained (Article 19)?
• □ Correction answered once within one year (Article 41)?

Regulatory information verified as at 29 October 2026. Legal texts follow the versions published by NMPA, SAMR and gov.cn and must be rechecked after amendment or repeal. Time limits, fees and catalogue scope reflect the official position on that date; content marked practical estimate has no legal effect.

How GRZAN can help. GRZAN provides medical device and IVD registration, quality system, authorised representative and local testing support for the Chinese market. Our team prepares NMPA-compliant dossiers, coordinates typetesting and clinical evaluation documents, and supports post-market vigilance obligations. Business enquiries: info@grzanglobal.com