Annual Product Quality Review (APQR/PQR): what to include

EU GMP Chapter 1 lists the minimum scope of a PQR, and the FDA expects similar content. A practical structure that covers both:

1. Product and scope: product, strengths, pack sizes, sites, review period and the batches manufactured, released, rejected and reprocessed.
2. Starting and packaging materials: review of materials used, especially those from new sources, and supply-chain traceability for active substances.
3. Critical in-process controls and finished-product results: trends against specifications, with process capability where the data supports it.
4. Batches failing specification: every failure and its investigation.
5. Deviations and non-conformances: significant deviations, their investigations and the effectiveness of resulting CAPAs.
6. Changes: process and analytical method changes made during the period, with their impact.
7. Marketing authorisation variations: variations submitted, granted or refused, including for third-country dossiers.
8. Stability programme: results and any adverse trends.
9. Returns, complaints and recalls: quality-related events and their investigations.
10. Previous actions: the adequacy of corrective actions from earlier reviews and investigations.
11. Post-marketing commitments, where applicable.
12. Qualification status: relevant equipment and utilities such as HVAC, water and compressed gases.
13. Technical agreements: confirmation that contracts with outsourced activities are up to date.

For sterile products, add environmental monitoring excursions, sterilisation and filter integrity failures and aseptic process simulation results — they are often the most informative trends in the review.

The most frequent criticism of APQRs is that they report data without evaluating it. Each section should answer a question, not just fill a table.

Look at trends, not totals

Twelve deviations means little on its own. Twelve deviations of which eight share a root cause in the second half of the year is a finding. Group events by cause, equipment, area and shift, and compare with previous periods.

Use statistics where the data supports it

Control charts and process capability indices (Cp, Cpk) for critical quality attributes show drift long before a specification failure. Be careful with small batch numbers — a capability index from a handful of batches is not meaningful, and saying so is better than over-interpreting it.

Link to the CAPA and change history

For each significant issue, show what was done and whether it worked. A recurring problem with a “closed” CAPA is exactly the sort of signal the review is meant to surface.

Write explicit conclusions

• Is the process in a state of control? State it, with the evidence.
• Are specifications still appropriate, or do they need tightening or revalidation?
• What actions are required, who owns them and by when?
• Is any revalidation needed because of cumulative changes?

Most of the effort in an APQR is collecting data from different places: batch records, the lab system, the deviation and CAPA logs, change control, stability, complaints, supplier files and engineering. A few habits make the job much lighter:

• Maintain a schedule that staggers products across the year, so QA is not reviewing every product in the same month.
• Agree data owners and cut-off dates for each section before the period ends.
• Freeze the dataset for the period, so the figures in the approved report can always be reproduced.
• Use rolling reviews where your procedure allows — updating trends quarterly makes the annual write-up a summary rather than an excavation.
• Group similar products where the regulations allow, with a documented scientific justification.

EU GMP expects the manufacturer and, where different, the marketing authorisation holder to evaluate the review results and decide whether corrective and preventive action or revalidation is needed. Actions should be completed in a timely and effective way, and management should review their progress. In practice:

1. The review is prepared by QA with input from production, QC, engineering and regulatory affairs.
2. It is reviewed and approved by people who are independent of the preparation.
3. Actions are raised in the CAPA or change control system — not left in the report’s final paragraph.
4. The next review starts by checking those actions were completed and effective.

Inspectors often pick a product and read its last two reviews side by side. If the same issues and the same unfulfilled actions appear in both, expect questions.

Product quality reviews are a favourite starting point for inspectors because they summarise a year of a product’s history in one place. The weaknesses they find are predictable:

• Late reviews: the report for a period is approved many months after the period ended, so any trend it reveals is already old news.
• Data without analysis: tables of results with no comment on trends, capability or comparison with earlier periods.
• Missing inputs: starting-material supplier changes, technical agreements, equipment qualification status or marketing authorisation variations simply absent.
• Inconsistent figures: batch or deviation counts in the review that do not match the source systems, because the data was compiled by hand from several places.
• No link to actions: issues identified in the review never appear in the CAPA or change control system.
• Copy-forward conclusions: “the process remains in a state of control” repeated year after year, even when the data in the same report suggests otherwise.

Each of these is avoidable with a schedule, a frozen and reproducible dataset, and a reviewer who is expected to write a genuine conclusion rather than confirm a template.

Swat iQua Suite includes a Product Quality Review module that helps with the hardest part — assembling the period’s quality figures from one source of truth. It is a newer module, and it does the groundwork rather than the analysis:

• For a chosen product and review period, it collects counts of batches, deviations, CAPAs, OOS investigations, change controls, environmental monitoring excursions and lab quality events. Areas you have not licensed are shown as “n/a”.
• The figures are frozen as a snapshot that cannot be changed, so the approved review can always be reproduced.
• The review moves from draft to prepared to approved and closed, with an electronic signature and segregation of duties — the approver cannot be the preparer.
• A controlled A4 PDF report is produced for the file.

The interpretation, trend analysis and conclusions remain your reviewers’ work. Related modules help them: stability studies provide ICH Q1E regression and shelf-life reporting, continued process verification calculates Cp and Cpk from the data points you enter, deviations and CAPAs are linked to their sources, and every record carries a full audit trail. Because the quality, lab and manufacturing records live on one platform, the inputs to the review are a few clicks away rather than spread across systems.