ICH Q7: GMP Guide for Active Pharmaceutical Ingredients
Understand API manufacturing boundaries, connect quality responsibilities to production and laboratory controls, and organize the records needed for informed batch decisions.

On this page · 11 sections
- 1ICH Q7: Definition, purpose and essential terminology
- 2Regulatory mapping: Guidance, binding law and product classification
- 3Define the manufacturing boundary before assigning downstream controls
- 4Map controls across processing, storage and API distribution
- 5Quality-unit independence, review responsibilities and release decisions
- 6Material specifications, production instructions and batch calculations
- 7Laboratory controls, equipment qualification, cleaning and validation
- 8Change assessment and recovery from calibration exceptions
- 9Clinical-trial APIs: Development-stage controls under Section 19
- 10Connected evidence, interface recovery and limited software support
- 11Implementation checklist and recurring quality-review priorities
How does ICH Q7 apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01ICH Q7: Definition, purpose and essential terminology
ICH Q7 is the Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients. It provides guidance on API manufacturing under a system for managing quality. Its Section 1.1 treats good manufacturing practice and current good manufacturing practice as equivalent terms. The vocabulary below separates the guide itself from quality activities and manufacturing operations.
Scope is narrower than “all APIs”. Q7 1.3 applies to the manufacture of APIs for use in human drug (medicinal) products, and to sterile APIs only up to the point immediately prior to the APIs being rendered sterile; sterilization and aseptic processing of sterile APIs fall under GMP guidelines for drug (medicinal) products as defined by local authorities. It covers APIs made by chemical synthesis, extraction, cell culture/fermentation, recovery from natural sources or a combination, with cell culture/fermentation detail in Section 18. It excludes vaccines, whole cells, whole blood and plasma, blood and plasma derivatives (plasma fractionation) and gene therapy APIs, while including APIs produced using blood or plasma as raw materials. It also does not apply to medical gases, bulk-packaged drug (medicinal) products, or manufacturing/control aspects specific to radiopharmaceuticals.
| Term | Meaning | Operational distinction |
|---|---|---|
| ICH Q7 | GMP guidance for active pharmaceutical ingredients. | The reference framework, not a manufacturing operation. |
| Quality assurance | Organized arrangements supporting APIs’ intended quality and maintained quality systems. | The broader system perspective; see the quality assurance process. |
| Quality control | Checking or testing that specifications are met. | Narrower than QA: focuses on conformance to specifications. |
| Manufacture | All operations: receipt, production, packaging and repackaging, labeling and relabeling, quality control, release, storage and distribution, plus related controls. | The complete API manufacturing lifecycle, not just processing. |
| Production | API preparation from material receipt through processing and packaging. | The preparation sequence within manufacture, rather than the whole lifecycle. |
In practice, use these distinctions when naming procedures and assigning record ownership: a production workflow should not be mistaken for the complete manufacturing control system.
02Regulatory mapping: Guidance, binding law and product classification
Map the product and activity before assigning obligations. FDA Q7 guidance contains nonbinding recommendations; it is not the text of a binding regulation. Keep its API recommendations distinct from statutory or regulatory duties and from the procedures a manufacturer chooses to implement them.
| Jurisdiction or context | Product or activity | Instrument | Practical interpretation |
|---|---|---|---|
| International API guidance | API manufacturing | ICH Q7 | Internationally developed guidance; distinguish its recommendations from each market’s legal obligations. |
| United States: API guidance | API manufacturing and related controls | FDA Q7 guidance | Use as API guidance, not as a standalone binding regulation. |
| United States: drug-product controls | Quality-control decisions | 21 CFR 211.22(a) | Binding-law comparison: the quality-control authority described below is not a substitute for Q7’s API recommendations. |
| EU, UK or another market | Manufacture, use or supply of an API | Current local law and relevant Q7 implementation text | Check applicability before assigning legal obligations; do not assume identical implementation. |
| Non-API product classifications | Supplements, foods, devices and non-pharmaceutical chemicals | Applicable controls identified through classification assessment | Identify intended use, product category, activity and destination market before selecting controls. |
21 CFR 211.22(a) requires a quality control unit with responsibility and authority to approve or reject components, drug product containers, closures, in-process materials, packaging material, labeling and drug products. Read that authority in its drug-product regulatory context, rather than treating it as interchangeable with Q7.
For implementation, label each control’s basis: binding law, guidance recommendation or industry practice. A site’s cGMP procedures can translate those bases into assigned tasks. Treat software support separately as an implementation choice, not a source of legal authority or proof of compliance; human review and release decisions remain with authorized personnel.
03Define the manufacturing boundary before assigning downstream controls
The Q7 glossary definition of intermediate describes a material produced during API processing that undergoes further molecular change or purification before becoming an API. Its parenthetical scope note limits the guidance’s intermediate coverage to intermediates produced after the point the company has defined as the start of API production.
Q7 1.3 defines an “API Starting Material” as a raw material, intermediate or API that is used in the production of an API and is incorporated as a significant structural fragment into the structure of the API. It can be an article of commerce, purchased from one or more suppliers under contract or commercial agreement, or produced in-house, and normally has defined chemical properties and structure. The company should designate and document the rationale for the point at which API production begins: for synthetic processes, that is the point at which API Starting Materials enter the process; for other processes such as fermentation, extraction or purification, the rationale is established case by case, with Q7 Table 1 giving guidance on where the starting material is normally introduced. From that point, appropriate GMP from the Guide applies, including validation of critical process steps determined to impact API quality.
Q7 Section 1.3 recommends that the company designate that point and document its rationale. The practical task is therefore to connect a named process boundary to an assessable explanation, rather than merely label a purchased material or a convenient production step.
Build a boundary assessment record
A useful practice is to record:
- Route identifier: the process and revision being assessed.
- Material identity: the specific material selected for introduction.
- Introduction step: the operation marking the proposed start.
- Rationale owner: the person responsible for maintaining the assessment.
- Supporting process knowledge: the transformations, purification steps and material attributes supporting the reasoning.
- Approval reference: the documented decision and its applicable route version.
Recording a selected boundary does not establish that it is appropriate. Check the current Q7 scope text when assessing the selection. Link material and route identities to a backward genealogy trace so reviewers can follow the actual manufacturing history.
Worked case: X to Y to API Z
Assume a chemical route X → Y → API Z, with introduction of X designated as the manufacturing start for this exercise. Y requires further purification before becoming Z. Y is the post-boundary intermediate; the downstream map follows Y’s purification into Z and subsequent API handling. Attach relevant in-process checks to the mapped operations. Whether introduction of X is a justified start remains a separate assessment. This example involves no quantities or calculation.
If Y is purchased instead, treat that as a classification and rationale question—not an automatic shortcut for moving the GMP boundary downstream.
04Map controls across processing, storage and API distribution
Using the manufacture-versus-production distinction established above, map production stages within the wider manufacturing lifecycle. Connect intermediate processing and final API processing to packaging, storage and distribution, rather than ending the evidence trail at the last processing operation.
As guidance, Q7 6.41 recommends master production instructions covering process sequences, parameter ranges, sampling instructions and in-process controls with acceptance criteria where appropriate. It also includes storage instructions addressing suitability for use, labelling and packaging materials. Translate these recommendations into a stage-specific control strategy, not an identical checklist for every operation.
The following matrix is an implementation aid, not a prescribed sequence or universal ranking of control intensity. Select evidence according to the actual process; attach sampling points and acceptance criteria to the relevant in-process check.
| Stage | Illustrative evidence |
|---|---|
| Material entry | Material identity linked to the incoming lot |
| Intermediate processing | Conversion parameters and applicable sampling results |
| Final API processing | Isolation results against applicable acceptance criteria |
| Packaging | API lot, container and label identity |
| Storage | Specified conditions linked to storage records |
| Distribution | API lot linked to shipment and destination references |
Use the connections between records to follow one lot across handoffs, including intermediate holds, without assuming that each handoff needs the same tests.
Illustrative six-step workflow
Material entry → Intermediate processing → Final API processing → Packaging → Storage → Distribution
- Map boundaryIdentify the route, material entry point and downstream stages.
- Prepare controlsLink material specifications, production instructions and equipment status.
- Execute productionCapture actual quantities, step results and material genealogy.
- Evaluate resultsConnect samples, laboratory results and relevant investigations.
- Resolve exceptionsAssess affected batches and document approved follow-up actions.
- Review dispositionAssemble connected evidence for the independent quality unit.
05Quality-unit independence, review responsibilities and release decisions
ICH Q7 treats quality-unit independence as an organizational principle, not a requirement for a large department. Under Q7 2.13, the guidance recommends a quality unit independent of production that fulfills both quality assurance and quality control responsibilities. Separate QA and QC units, one group, or a single individual can serve this function, depending on organizational size and structure. That flexibility changes the staffing model, not the independence principle.
Q7 2.22 states that the independent quality unit’s main responsibilities should not be delegated. In practical terms, distinguish contributions to the work from ownership of the decision: production personnel and laboratory analysts can generate evidence and contribute technical assessments without taking over the listed quality-unit responsibilities.
Those responsibilities include releasing or rejecting all APIs, and releasing or rejecting intermediates for use outside the control of the manufacturing company (Q7 2.22(1)). They also include reviewing completed production and laboratory control records for critical process steps before API release for distribution (Q7 2.22(3)), and making sure critical deviations are investigated and resolved (Q7 2.22(4)). Consequently, a passing laboratory result alone does not cover the review supporting batch release.
The following implementation matrix separates supporting activities from accountable decisions; it is a practical allocation model, not additional Q7 text.
| Activity | Supporting contribution | Quality-unit responsibility |
|---|---|---|
| Production execution | Operators record steps and observations. | Review critical-step production records. |
| Laboratory testing | Analysts generate results and technical assessments. | Review critical-step laboratory control records. |
| Investigation | Relevant specialists establish facts and assess causes. | Make sure critical deviations are investigated and resolved. |
| Disposition | Staff assemble the review record. | Decide API release or rejection. |
Use the batch record to connect these contributions to the disposition decision. Software can support the record, but it does not replace human approval or release. A useful practical check is whether each supporting assessment identifies its author and whether the final decision clearly identifies the accountable quality-unit reviewer.
06Material specifications, production instructions and batch calculations
Separate material specifications from process criteria
Q7 6.17 recommends established, documented specifications for raw materials, intermediates where necessary, APIs, and labelling and packaging materials. The qualification “where necessary” belongs to intermediates; it should not disappear when translating the guidance into a site specification register.
The same provision separately recommends established, documented acceptance criteria for in-process controls. Keep these distinct in the control structure: material specifications address the material being evaluated, while in-process criteria provide the comparison points for checks during processing.
Make instructions calculable and execution traceable
Q7 6.41 recommends that master production instructions state accurately the quantity or ratio of each raw material or intermediate, including its unit of measure. Where quantities are not fixed, the instructions should include the calculation for each batch size or production rate. It also recommends expected yield ranges at appropriate processing phases or times—not simply a final yield number.
A practical implementation is to connect each execution to its instruction version, then link the actual material lots, quantities, units and processing observations. This is a suggested traceability pattern, not a prescribed electronic schema. In a batch record, reviewers can then compare the planned charge with actual use rather than reconstructing the calculation from disconnected entries.
For material consumption posting, preserve the relationship between each lot used and its recorded quantity. Keep planned quantities distinguishable from actual quantities so a correct formula does not conceal a different execution.
Worked case 2: Commercial dry API yield
Assume a commercial process with a theoretical dry API yield of 125.0 kg and actual dry isolated API of 112.5 kg, both on the same calculation basis. The illustrative site-defined expected range is 88.0–94.0%; it is not a Q7 limit.
Theoretical versus actual yield = 112.5 kg ÷ 125.0 kg × 100 = 90.0%. The result is within the illustrative range. It is one input to the quality review described above, not a release decision. Recording both masses and their common dry basis makes the percentage interpretable and reproducible.
07Laboratory controls, equipment qualification, cleaning and validation
Q7 addresses several distinct forms of control evidence. Laboratory procedures govern how materials are sampled, tested and dispositioned; qualification, cleaning, calibration and validation address different questions. The recommendations below are guidance, while the checklist is a practical review structure—not an additional regulatory requirement.
Keep the control questions separate
Q7 11.11 recommends documented procedures covering sampling, testing, approval or rejection of materials, and recording and storage of laboratory data. A test result is therefore only one element of the laboratory-control evidence.
The FDA Q7 glossary entry for qualification distinguishes equipment or ancillary-system qualification from process validation: qualification is part of validation, but individual qualification steps alone do not constitute process validation. Keep those conclusions separate when reviewing a validation package.
For cleaning, Q7 5.25 recommends defining and justifying residue acceptance criteria, cleaning procedures and cleaning agents. It does not assign a universal numerical residue limit. A practical cleaning validation review should connect the selected criteria to their justification rather than merely confirm that a cleaning instruction exists.
Q7 5.30 recommends calibration under written procedures and an established schedule for control, weighing, measuring, monitoring and testing equipment critical to intermediate or API quality. Under Q7 5.33–5.34, current calibration status should be known and verifiable, and instruments failing calibration criteria should not be used.
Set validation scope and review the evidence
Q7 5.40 recommends validating GMP-related computerized systems, with depth and scope depending on the application's diversity, complexity and criticality. A CSV traceability matrix can organize the practical connection between intended functions and test evidence; it is not itself a validation conclusion. Q7 2.22(10) assigns review and approval of validation protocols and reports to the quality unit.
Use this evidence checklist to keep each review focused. One category should not be treated as proof of another.
| Evidence category | Practical review question |
|---|---|
| Laboratory procedures | Are sampling, testing, material decisions and laboratory-data handling covered? |
| Qualification records | Which equipment or ancillary system was assessed, and what conclusion was reached? |
| Cleaning criteria | Are residue criteria, procedures and agents defined and justified together? |
| Calibration records | Can the instrument's current status and acceptability for use be established? |
| Validation protocols and reports | Does the scope match the subject being validated, with quality-unit review and approval? |
08Change assessment and recovery from calibration exceptions
A planned change and an equipment exception raise different questions: what could the proposed change affect, and what might already have been affected? Q7 provides guidance for both; the investigation record should keep those questions distinct.
Change approval and downstream notification
Q7 2.22(9) assigns approval of changes potentially affecting intermediate or API quality to the independent quality unit. Use the human decision responsibilities covered above when routing change control, rather than treating implementation as approval.
Q7 13.17 recommends notifying current dosage-form manufacturers of changes from established production and process-control procedures that can impact API quality. This is a specific downstream communication recommendation, not a statement that every administrative change needs customer notification.
Worked case 3: Look backward from calibration failure
Q7 5.35 recommends investigating deviations from approved calibration standards on critical instruments for possible quality effects on intermediates or APIs manufactured using that equipment since its last successful calibration.
Assume a critical balance passed calibration at elapsed hour 0 and failed at elapsed hour 72. B101, B102 and B103 each used it within that interval. The investigation window is 72 − 0 = 72 hours; all three batches enter impact assessment, not automatic rejection.
Examine each batch’s balance uses, recorded measurements and relevant acceptance margins before deciding disposition. As-found/as-left calibration helps distinguish the observed failure from the condition after adjustment. A later passing check does not, by itself, resolve the earlier batch question.
Practical exception chain
- Identify the trigger: proposed change or observed calibration deviation.
- Establish affected records: equipment history, uses and batch measurements.
- Assess quality impact against the relevant process and acceptance margins.
- Document the decision and its supporting reasoning.
- Determine follow-up actions, including corrective work and applicable downstream communication.
09Clinical-trial APIs: Development-stage controls under Section 19
Section 19 frames clinical-trial API controls around development maturity. Under Q7 19.11, controls should be consistent with the development stage of the drug product incorporating the API. The reference point is therefore the drug product’s development stage, not simply the size of the API batch.
The quality-unit discussion above still applies to clinical batch release: Q7 19.21 recommends a quality unit independent from production to approve or reject each clinical-trial API batch. Development-stage flexibility does not remove that independent decision.
Q7 19.60 describes process validation as normally inappropriate where only a single API batch is produced, or where process changes during API development make batch replication difficult or inexact. Those conditions matter: this is a qualified position for clinical development, not an unrestricted validation exemption for anything labelled investigational.
Yield expectations also differ from the commercial worked example above. Q7 19.51 recognises that expected clinical yields can be more variable and less defined than commercial yields, and states that investigations into yield variations are not expected.
Read that yield statement alongside Q7 19.24: process and quality problems should be evaluated. The distinction is between investigating yield variation as such and evaluating a process or quality problem; the former statement does not cancel the latter recommendation.
10Connected evidence, interface recovery and limited software support
A practical API evidence model connects consumed material lots to production batches, batches to equipment-use events and samples, and samples to laboratory results. Quality decisions should reference the records reviewed, including relevant exceptions, rather than sit as an isolated status. These are suggested design relationships, not regulatory database requirements.
Q7 5.43 recommends controls against unauthorized data access or changes and against data omissions, including loss of capture when a system is turned off. It also recommends recording each data change, the previous entry, who changed it and when. These guidance recommendations inform data integrity by design; they do not prescribe a particular relational schema.
For a generic production-to-laboratory interface, use shared batch and sample identifiers, source-system record references and an explicit reconciliation status. An equipment-data interface can additionally associate readings with equipment-use events. For MES to LIMS integration, distinguish receipt of a message from successful matching of its result to the intended sample.
After an outage, compare expected records with received records over the affected interval. Detect duplicates before accepting replayed messages, preserve correction history and resolve unmatched results before resuming review. Record the reconciliation outcome so reviewers can distinguish recovered data from unresolved gaps. These are suggested integration and recovery controls, not V5 Ultimate capability claims.
Suggested data-model diagram MaterialLot(lot_id) → BatchMaterial(batch_id, lot_id) → ProductionBatch(batch_id) ProductionBatch → EquipmentUse(batch_id, equipment_id) ProductionBatch → Sample(sample_id, batch_id) → LabResult(result_id, sample_id) ProductionBatch → QualityDecision(decision_id, batch_id) → ReviewedEvidence(decision_id, source_record_ref) Field names are implementation examples, not a prescribed Q7 schema or a description of a particular application's database.
Material Lot
Connect material identity, supplier reference and applicable specification.
- material_lot_id
- material_code
- supplier_lot_number
- specification_version
- boundary_role
Production Batch
Connect approved instructions, consumed lots and yield calculations.
- batch_id
- master_instruction_version
- input_lot_ids
- theoretical_yield_kg
- actual_yield_kg
Equipment Use
Connect each batch to equipment status when used.
- equipment_use_id
- batch_id
- equipment_id
- calibration_status_at_use
- cleaning_status_at_use
Laboratory Result
Connect samples and measured results to assessment criteria.
- result_id
- batch_id
- sample_id
- specification_version
- result_value
- result_unit
Quality Decision
Connect disposition to relevant events and accountable approval.
- decision_id
- batch_id
- deviation_ids
- change_control_ids
- disposition
- approved_by
11Implementation checklist and recurring quality-review priorities
Use this practical checklist alongside the boundary, production, laboratory and change-control sections above. It tests connections between controls and records; it is neither an exhaustive inspection standard nor a compliance certification.
| Review question | Evidence to examine | Action if unresolved |
|---|---|---|
| Is the manufacturing boundary justified? | Boundary rationale and process map | Resolve unclear scope. |
| Are specifications and instructions current? | Approved versions matched to execution dates | Assess version discrepancies. |
| Can materials be traced to batches? | Lot identities and batch records | Reconcile missing links. |
| Was equipment status suitable at use? | Equipment identifiers and dated status records | Assess affected operations. |
| Do laboratory records support reported results? | Sample links, raw data, calculations and reviews | Investigate gaps. |
| Are open quality events accounted for? | Event register and affected-batch links | Assign impact assessment. |
| Is disposition documented? | Decision, rationale and approval record | Obtain the responsible quality decision. |
Q7 guidance lists ensuring that internal audits are performed among the independent quality unit’s responsibilities in Q7 2.22(7). It also lists performing product quality reviews in Q7 2.22(15), referring to Section 2.5.
Close the review on reviewable connections, not document volume: each unresolved issue should have an identifiable owner, affected records and a recorded next action. Track those actions without treating checklist completion as a prediction of inspection outcomes.
Frequently asked questions
Q.Does FDA publication of ICH Q7 make every recommendation a binding U.S. legal requirement?+
No. FDA’s Q7 publication presents nonbinding recommendations, not a conversion of every recommendation into a legal command. For example, its quality-unit responsibilities in section 2.22 are expressed as guidance. When assessing a finding, distinguish the Q7 recommendation from any separately applicable legal obligation. A practical quality assurance process should record that distinction so an internal checklist does not mislabel every departure from guidance as a statutory violation.
Q.Can a purchased intermediate automatically become the API starting material?+
Do not use purchase status alone to make that designation. As a practical assessment, map the material’s identity, supplier processing, subsequent transformations and quality risks before proposing where API GMP controls begin. A purchase order describes a commercial transaction; it does not provide a technical justification for the boundary. Use lot pedigree and relevant critical material attributes to organize the assessment, and document the proposed boundary for quality and regulatory review.
Q.Can QA and QC be one group while remaining independent of production?+
Organizationally, a combined QA/QC group can be designed with separate authority from production; the department name is not the decisive control. Q7’s section 2.22 identifies independent quality-unit responsibilities and recommends that its main responsibilities not be delegated. As a practical governance check, document who controls quality decisions, who can reject material and how production disputes are escalated. Make those authorities explicit in the QA disposition step, rather than relying on an organization chart alone.
Q.Does an API that passes its specification still need review of critical batch records?+
Yes. Q7 section 2.22, item 3 recommends quality-unit review of completed batch production and laboratory control records for critical process steps before API distribution release. Passing analytical results therefore does not replace that recommended review. In practice, the batch record review examines how the batch was made, not just its final test values. Keep the reviewer’s conclusion and unresolved questions visible in the batch release decision rather than treating a passing certificate as the complete assessment.
Q.How should a failed calibration affect batches made since the last successful check?+
First stop using the instrument: Q7 section 5.34 recommends against using instruments that fail calibration criteria. For critical instruments, section 5.35 recommends investigating possible quality effects on intermediates or APIs manufactured since the last successful calibration. Practically, identify affected batches and compare the observed error with the measurements and decisions it could have influenced. An as-found/as-left calibration assessment helps distinguish the historical error from the instrument’s condition after adjustment.
Q.Must a one-batch clinical API process follow the same validation approach as a commercial process?+
No. Q7 19.60 says process validation for clinical-trial API production is normally inappropriate where a single API batch is produced or where process changes make batch replication difficult or inexact; the combination of controls, calibration and, where appropriate, equipment qualification assures API quality in that phase. Q7 19.61 adds that validation under Section 12 applies when batches are produced for commercial use, even at pilot or small scale. As a practical planning approach, distinguish development uncertainty from controls needed to protect batch quality. Document the process knowledge available, critical operations, proposed acceptance criteria and evidence the batch can generate. Do not turn a commercial validation template into an unsupported fixed batch-count rule for a clinical batch.
Q.What should be reconciled before batch review resumes after an interface outage?+
As an operational recovery practice, reconcile source and destination records for affected batches: identifiers, event counts, timestamps, units, results, status changes and duplicate or missing messages. Check manual fallback entries against replayed transactions so recovery does not double-post an event. Record unresolved discrepancies and obtain human authorization to resume review. For MES–LIMS integration, prioritize result-to-sample-to-batch associations. Q7 section 5.43 recommends controls against data omissions; reconciliation is a practical data integrity response, not a prescribed Q7 outage checklist.
Q.What should supplement, food, device and chemical manufacturers check before applying ICH Q7?+
Start with the material’s intended use, product classification, manufacturing activity, destination market and contractual commitments. Do not infer applicability merely because a site handles powders, fermentation or chemical synthesis. As a practical assessment, separate legally applicable requirements from customer expectations and voluntarily adopted API practices. Use a product-specific cGMP assessment to establish the governing framework, then select useful Q7 concepts for the control strategy without describing voluntary adoption as proof of compliance.
Primary sources
Explore this topic
ICH Q7 sits inside this topic cluster in our glossary. Every neighbour is one click away.
Drug-product cGMP rules, ICH Q-series, and the regulators that enforce them.
Want to see how ICH Q7 could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.
