Skip to content
Entourage

How do you replace the manual visual inspection of your parenterals with an automated visual inspection that holds up to the notified body and to inspection?

We support pharma, biotech and MedTech manufacturers in the selection, qualification and validation of AVI systems for 100 percent final inspection: from the defect library and method validation per EU GMP Guide Annex 1 to computer system validation per GAMP 5. The critical spot is rarely the camera, but the evidence: an AVI may only replace manual inspection once it demonstrates at least equivalent detection accuracy in a planned detection study. Anyone who plans this comparison only after the line conversion is validating against a reference standard that no longer exists.

  • Pharma
  • Biotech
  • MedTech
  • IVD

Overview

What requirements does Annex 1 place on visual inspection?

Support from the defect library to the validated method · EU GMP Guide Annex 1 (2022), GAMP 5

Last updated: 2026-06-13

The EU GMP Guide Annex 1 requires a 100 percent inspection for visible particles and defects for parenteral products. Manual visual inspection is subjective, fatigue-dependent and barely reproducible at high unit volumes. Four points determine whether an automated inspection holds up regulatorily:

  • 100 percent final inspection: Annex 1 requires the complete visual inspection of every container; the method, whether manual, semi-automated or automated, must be validated and tailored to the specific defect spectrum.
  • Method validation against a reference standard: when manual inspection is replaced by an AVI, its detection capability must be demonstrated against a defined reference standard and shown to be at least equivalent to the established method.
  • Qualified defect sets: detection limits and acceptance criteria are based on test samples with known, graded defects (particles, cracks, fill level, cosmetic defects) rather than on nominal specifications.
  • Computer system validation and data integrity: the AVI system must be qualified as a computerized system per GAMP 5 and EU GMP Guide Annex 11 and integrated into change control and deviation management.

Services

How we support you

Requirements analysis & defect library

Analysis of product type, defect spectrum and throughput target, plus the build-up of a documented defect library with graded test samples as the basis for detection limits and acceptance criteria. The outcome is an approved requirements catalogue (User Requirements Specification).

System selection & supplier qualification

Technology comparison and supplier selection based on objective, weighted criteria, plus supplier qualification. The deliverable is a traceable selection matrix with a documented rationale for the award decision.

Method validation (Probability of Detection)

Design and evaluation of a planned detection study to determine the Probability of Detection and to compare the AVI against the manual reference standard per Annex 1. The deliverable is a validation report with detection rates per defect class.

Computer system validation & QMS integration

Qualification of the AVI system per GAMP 5 and Annex 15 (IQ, OQ, PQ) and connection to Annex 11 including change control and deviation handling for OOS signals. The deliverable is a validated system state with a traceability matrix.

Learn more

Training & performance monitoring

Training of operators and QA, plus the build-up of ongoing monitoring of the false-reject rate, the detection rate and system availability with a defined revalidation trigger. The deliverable is a documented monitoring program with metrics and thresholds.

What it comes down to

The EU GMP Guide Annex 1 turns visual inspection into a question not of the camera, but of evidence. Three strands have to fit together, and in this order: first comes the defect library with graded test samples, because it defines what the system must be able to find in the first place. On this rests the method validation with a probability-of-detection study that pits the AVI against manual inspection as the reference standard. Only after that does the computer system validation per GAMP 5 carry the system state in a compliant manner. Anyone who moves the line conversion ahead of the detection study loses exactly the reference standard against which equivalence would have to be demonstrated.

This is precisely where we come in: the requirements and defect analysis makes visible early on which defect spectrum actually has to be inspected, before technology is selected and the line is converted. This shifts the effort forward, to where defect sets and acceptance criteria can still be corrected inexpensively, and away from operation, where an overly optimistic detection rate comes back as a complaint. The ongoing monitoring of the false-reject and detection rate closes the loop, because it detects drift in optics and algorithm before it undermines the validated performance.

Our approach

Our approach

01

Requirements & defect analysis

Approved requirements catalogue and documented defect library with graded test samples.

02

System selection

Evaluated selection matrix, qualified supplier, defined inspection concept.

03

Method validation

Detection study with probability-of-detection evaluation and demonstrated equivalence to the manual reference.

04

Computer system validation

IQ/OQ/PQ completed, system brought into the validated state per GAMP 5 and Annex 11.

05

QMS integration

Change control, deviation and OOS handling for the AVI anchored in the QMS.

06

Operation & monitoring

Ongoing performance monitoring with metrics, thresholds and a revalidation trigger in operation.

Common pitfalls

Where projects commonly fail

The detection study is planned only after the line conversion.

Annex 1 requires the demonstration against the established manual inspection as the reference standard; if the manual line has already been dismantled, the benchmark is missing and the AVI can no longer be cleanly demonstrated as equivalent.

The defect library does not reflect the real defect spectrum.

Test samples with too few or non-graded defects produce a Probability of Detection that does not hold in operation; borderline particles and cosmetic defects are then systematically missed.

The AVI is treated as a machine rather than as a computerized system.

Without qualification per GAMP 5 and connection to Annex 11, meaning access rights, audit trail and change control, a data integrity gap arises that surfaces during inspection.

The false-reject rate is set too aggressively.

A threshold that is too tight drives up scrap of good units and rework and tempts unqualified manual re-inspections that undermine the validated state of the line.

Monitoring after validation is dropped.

Drift in optics, lighting or algorithm gradually changes detection performance; without ongoing monitoring of detection and reject rates with a revalidation trigger, the loss of performance goes unnoticed until the next complaint.

FAQ

Frequently asked questions

The EU GMP Guide Annex 1 requires a 100 percent inspection for visible particles and defects for parenteral products, as well as a validated inspection method. When replacing manual inspection, the AVI's detection capability must be demonstrated against a reference standard. In addition, Ph. Eur. 2.9.20 and USP <790> apply to visible particles, as does the FDA Guidance for Industry on the inspection of injectable products for visible particulates.

Sources
  • EU GMP Guide Annex 1 (Manufacture of Sterile Medicinal Products, 2022 version), primary text
  • GAMP 5 (ISPE Good Automated Manufacturing Practice) and EU GMP Guide Annex 11/15
  • FDA Guidance for Industry: Inspection of Injectable Products for Visible Particulates; Ph. Eur. 2.9.20; USP <790>
  • Writer material: automatisierte-visuelle-inspektion (Quality Management & Operational Excellence)
  • https://theentourage.de/quality-management-operational-excellence/automatisierte-visuelle-inspektion/ (existing page content)

Life Science Journal

Regulatory updates, straight to your inbox.

New requirements, authority decisions and practice notes. Once a month, unsubscribe any time.

Regulations & standards considered

  • EU GMP Guide Annex 1 (Manufacture of Sterile Medicinal Products, 2022 version)
  • GAMP 5 (Good Automated Manufacturing Practice, ISPE)
  • EU GMP Guide Annex 11 (Computerised Systems)
  • EU GMP Guide Annex 15 (Qualification and Validation)
  • FDA Guidance for Industry: Inspection of Injectable Products for Visible Particulates
  • Ph. Eur. 2.9.20 (Particulate Contamination: Visible Particles)
  • USP <790> (Visible Particulates in Injections)

Have a concrete project?

Briefly outline your situation. We'll respond with an initial assessment, usually within one business day.

Prefer direct? +49 89 4161170-0
info@theentourage.de

  • Reply usually within one working day
  • 4 offices: DE · CH · IT · US
  • 100% life sciences