Japan investigational drug GMP vs Annex 13 and phase 1

The standard is a notice under GCP, addressed to the sponsor

Japan's investigational drug GMP is not an ordinance. It is a Director-General notice, Yakushokuhatsu No. 0709002 of July 9, 2008, sitting under the GCP Ordinance.

The notice says so in its opening: it sets the manufacturing control, quality control and structural facility matters required under GCP Ordinance Article 17, paragraph 1 and Article 26-3. Section 2.1 says the standard sets what the sponsor must do under Article 17, paragraph 1 and what a person conducting a trial on their own initiative must do under Article 26-3. The 2008 text replaced two 1997 notices and covers investigational drugs used in trial plans notified on or after August 1, 2008 and manufactured on or after that date.

Two scope clauses matter to a site abroad. Section 2.2 states that the standard applies even where the manufacturing facility is overseas. Section 2.3 says it applies to manufacture for trials covered by those GCP articles whatever happens later: whether the GMP Ministerial Ordinance will apply after approval does not change the answer.

The form of the instrument is the part people get wrong. MHLW's material for the review panel that led to the 2008 rewrite put the two regimes side by side and labelled them: investigational drug GMP as administrative guidance through a notice, commercial drug GMP as something confirmed at approval and required for the manufacturing licence. That material described the 1997 version, so read it for the architecture only. There is also no certificate you can hold. Under MHLW's administrative notice of July 27, 2026, an investigational drug GMP certificate is issued only where the manufacturing facility is in Japan, and only where a trial in Japan is under way or planned, or the importing country asks for a certificate from the Japanese government. A site abroad cannot apply.

Two things you might expect here are not in the notice at all. It never mentions accreditation of foreign manufacturers, which belongs to the marketing approval that comes later, and it says almost nothing about the label: the Japanese "for clinical trial use" statement and the sponsor's or caretaker's address live in GCP Article 16, covered in the clinical trial notification guide.

Treat the notice as the sponsor's obligation and your contract as the transmission, because it is administrative guidance rather than a condition of any licence your site holds.

Section 20 moves most of the standard onto you, and keeps two duties back

Section 20 is the clause a contract manufacturer reads first. It sets out what the agreement must contain, then reads the rest of the standard across.

The eight agreed items are the scope of the outsourcing, the technical conditions, the sponsor's confirmation that the work is properly carried out at your facility, the instructions the sponsor may give you, confirmation of how you implemented any improvement it instructed, quality control during transport and handover, retention of documents and records, and anything else needed to secure proper control. Section 20.4 requires those instructions in writing, and 20.3 lets your own documents cover only the processes you perform.

Then comes 20.2. Sections 2 through 19 apply to the contract manufacturer, with "sponsor" read as "contract manufacturer" everywhere except in 2.4, and except for the reference sample retention in 10.1.8 and the record retention in 19.1.3. Those two are explicitly excluded, so the Japanese sponsor keeps them.

Expect to be looked at. Section 10.1.5 makes the sponsor's quality department confirm that control at your facility is properly carried out, and 20.6 has a person the sponsor designates perform the 20.1.3 and 20.1.5 confirmations and report in writing to the sponsor's own quality department. In the other direction, 20.5 makes you report in writing to the sponsor that your quality department evaluated the results and shipped. Ask the sponsor, before signing, which sections it is reading across and who will physically hold the reference samples.

Write the storage of reference samples and records into the agreement, because section 20.2 lifts both duties off your site and leaves them with the sponsor.

No qualified person: a designated person in the quality department releases

Japan names no qualified person for investigational drugs. Shipment is decided by a person the investigational drug quality department has designated in advance.

The structure around that person is simple. Section 5.1 puts a production department and a quality department at each investigational drug manufacturing facility, and adds that the sponsor must keep departments with the necessary functions even when it outsources the whole process. Section 5.2 makes quality independent of production. Section 6.1 then has the designated person evaluate the results for each item and decide whether shipment from the facility is allowed, where an item means the investigational drugs used in the trials for one marketing application. Section 6.2 asks only that this person has had sufficient training on the trial and on manufacturing and quality control, and has the knowledge and experience. The word pharmacist appears nowhere in the notice.

Compare what the same batch needs in Europe, and check which document your SOP means by Annex 13: in EudraLex Volume 4 that slot now holds the Commission's detailed guidelines of 8 December 2017, not the 2010 text of the same name. As those guidelines put it, release should not occur until the qualified person has certified under Article 62(1) of Regulation (EU) No 536/2014 that the requirements of Article 63(1) and (3) and of Article 12 of Delegated Regulation (EU) 2017/1569 are met. For product imported directly from a third country, the guidelines say that in the absence of a mutual recognition agreement, the qualified person should determine that equivalent standards of good manufacturing practice apply through knowledge of your quality system, normally acquired by auditing it, and may then certify on the basis of your documentation and document the rationale.

The United States splits the question by phase instead. Under 21 CFR 210.2(c), production of an investigational drug for a phase 1 study is exempt from part 211 while remaining subject to the statutory requirement in 21 U.S.C. 351(a)(2)(B), and the exemption falls away once the drug has been made available for a phase 2 or phase 3 study or lawfully marketed. FDA's July 2008 guidance covers how to apply statutory CGMP to most phase 1 investigational drugs. Japan has no phase-keyed exemption: the same notice covers phase 1 and phase 3.

Name the person who decides shipment under 6.1 and keep their training file current, because Japan places the decision on an individual rather than on a licensed role.

Phase-appropriate is written in, and verification is its named alternative

Japan does not set a separate requirement per phase, and says so. Section 3.1 states that requirements should move with the progress of development, that what is required can differ even within the same clinical stage, and that a uniform rule is therefore difficult to write.

The same clause tells you what to do instead: apply the standard flexibly, considering the stepwise situation that comes with development and the risk, and that it is desirable to use it as part of quality management with the product lifecycle in view.

The lever the notice gives you is vocabulary. Section 4 defines validation, verification and qualification separately. Validation is normally done once methods are established and repetition for reproducibility is needed. Verification, in section 4.8, is confirming from procedures, plans, records and reports that the expected quality was obtained, normally for limited situations and limited lots. Section 12.1.1 then lets you implement "necessary validation or verification" considering the stage of development, the objective of the trial and the risk. Section 21.3 extends the logic to hardware: special equipment used early in development, and single-use devices, do not necessarily have to meet the drug facility requirements referenced elsewhere in section 21, provided verification shows their function is effective and quality is secured.

Some things do not flex. Section 3.2 requires all manufacturing and quality control records to be retained so that later confirmation is possible. Section 9.2 says that because toxicity knowledge may be incomplete at the development stage, the production department must take measures specific to investigational drugs, cross-contamination prevention among them. Section 10.1.7 requires quality to be assured per lot until dosing ends in the planned trial, and for a drug with extremely poor stability, reproducibility to be examined with the time until administration in mind.

Section 21.1 points at the GMP Ministerial Ordinance and the Regulations for Buildings and Facilities for Pharmacies (MHW Ordinance No. 2 of 1961) as references, to be applied scientifically according to the properties of the drug, and 21.2 adds that other laws apply where relevant, naming the Cartagena Act on genetically modified organisms and the radiation hazard prevention act. Section 3.4 is the one to show your engineering team: design with awareness of the facility requirements that the marketing approval and the manufacturing licence will demand at the final stage of development. The EU reaches the same place from the other direction, since the current Annex 13 says procedures need to be flexible as process knowledge grows and appropriate to the stage of development. What Japan adds is a defined word for the lighter evidence.

Record for each early-phase batch whether you relied on validation or on verification under 4.8, and why the stage, the objective and the risk justified the choice.

Reference samples and records run on a Japanese clock

Japan ties retention to the approval, not to a count of years from the batch. That one difference can leave samples in your freezer long after an EU clock would have expired.

Section 10.1.8 sets both the quantity and the period. The quantity is at least twice the amount needed for the prescribed tests, after allowing for the amount that comparative testing on a change will consume. The period runs to the later of two dates: the day marketing approval is received for the drug, or where the sponsor has given notification under GCP Article 24, paragraph 3, the day three years after that notification; and the day three years after the trial is discontinued or completed. Drugs that do not form lots, and those whose nature makes storage extremely difficult, are outside the requirement. Section 19.1.3 puts documents and records on the same clock.

The European periods are stated as years. As the Commission's 2017 guidelines quote Article 8(3) of Delegated Regulation (EU) 2017/1569, documents forming part of the product specification file are kept at least five years, and batch manufacturing records at least five years after completion or formal discontinuation of the last clinical trial in which the batch was used. The sponsor separately keeps the clinical trial master file at least 25 years after the end of the trial. On sample size the two regions nearly agree: the EU asks for enough reference sample to run all critical quality attribute tests on at least two occasions, Japan for twice what the prescribed tests need.

Japan has a counterpart to that file. The document on the investigational drug in section 7.1 carries the ingredients and quantities, the specifications and test methods, the manufacturing procedures and an outline of the trial, is approved by the quality department, and under 7.2 is revised as development progresses. Its revision history is part of what Japan later asks to see.

Why it matters is in the purpose clause. Section 1.3 says investigational drug GMP exists, among other reasons, to assure consistency between the investigational drug and the marketed product once the drug has been narrowed down as a development candidate, and equivalence once the manufacturing and test methods are established. Section 3.3 turns that into a records duty: from the candidate stage, control, document and retain every change during development as the basis for showing that consistency and equivalence. Section 13.1 opens change control by noting that changes happen frequently in development, 13.1.1 asks for the possible quality impact to be evaluated by scientific and objective means such as small-scale experiments and approved by the quality department before the change is made, and 13.1.3 asks you to keep traceability for the day someone has to connect the trial material to the future commercial product.

Keep the change history as one unbroken chain from the first clinical batch, because Japan will read it as the evidence that your trial material and the marketed product are the same thing.

Steps before your first batch for a Japanese trial

Run these in order. Steps 1 and 2 belong in the contract, before any material moves.

1. Get the notice itself, Yakushokuhatsu No. 0709002 of July 9, 2008, with a translation, and ask the sponsor which sections it is reading across to you under 20.1 and 20.2.
2. Put the eight items of 20.1 into the quality agreement, and settle who holds the reference samples under 10.1.8 and the records under 19.1.3, since 20.2 leaves both with the sponsor. Add the sponsor's confirmations under 20.1.3 and 20.1.5, and your written shipment report under 20.5.
3. Name the person your quality department designates to decide shipment under 6.1, keep the training record 6.2 asks for, and check that your quality department is independent of production as 5.2 requires.
4. Build the document on the investigational drug under 7.1 for each item, have your quality department approve it, and set the trigger that makes you revise it when the process or the specification moves.
5. Compare the procedures listed in 8.4 against your own index: shipment control, validation and verification, change control, deviations, quality information and defects, recalls, self-inspection, training, and document and record control, plus the catch-all for anything else needed in 8.4.10.
6. For each batch, record whether you relied on validation or on verification under 4.8 and why, under 12.1.1. Take the measures 9.2 asks for, cross-contamination prevention among them, from the first batch, because 9.2 is written for the case where the drug's toxicity is not yet known.
7. If the same material also goes to an EU trial, ask your qualified person early which certification route applies, since in the absence of a mutual recognition agreement the certification rests on knowledge of your quality system, normally gained by audit.
8. If an outside laboratory does any of the testing, agree the five items in 11.1 with it: scope, technical conditions, your confirmation that testing is properly done, how sample transport and handover stay reliable, and anything else reliability needs.

For getting the material into Japan, see the Yakkan Shomei guide; for how PMDA looks at a site once the product is approved, the PMDA inspection guide. Japanese terms sit in the glossary.