USP <1790> Visual Inspection of Injections

USP ⟨1790⟩ is the how-to guide for visual inspection of injectables; USP ⟨790⟩ is the pass/fail standard (AQL 0.65%). Together they operationalize "essentially free from visible particulates", which means compliance with a validated 100% inspection process, not zero particles.

<1790> gives recommendations for the design, qualification/validation, and application of visual inspection programs across the product lifecycle.
It addresses prevention, 100% inspection, acceptance sampling, investigation, and trending.

It is complementary to USP <790> (compendial standard for visible particulates) and USP <1> (parenteral product quality tests).

Under USP <790>, “essentially free” means that, when inspection is performed as described, no more than the specified number of units may exhibit visible particulates (i.e., compliance is demonstrated via 100% inspection plus acceptance sampling). This clarifies that the standard is implemented through defined procedures and criteria, not a claim of absolute zero.

Lifecycle approach—prevention first, then detection and control. <1790> emphasizes building processes to prevent particle introduction (materials, components, equipment, environment), employing qualified inspection to detect/remove affected units, and using investigation/CAPA and trending to sustain control. FDA also frames visible-particle control within a lifecycle/CGMP context.

Manual Visual Inspection (MVI)—compendial reference method. MVI remains the baseline reference described by major compendia; inspectors and methods must be qualified and periodically requalified. Key conditions from USP <790> for manual inspection include:
• No magnification (other than corrective lenses as needed); inspect against both black and white backgrounds.
• Illumination at the inspection point: 2,000–3,750 lux.
• Container movement (gently swirl/invert) to mobilize potential particles.
• Inspection time ≈ 5 seconds per background (≈10 seconds total per unit).
These conditions are used for acceptance sampling and are widely adopted for MVI stations.

Automated Visual Inspection (AVI) qualification expectations. AVI may be used for 100% inspection in high-volume operations, but systems must be validated to demonstrate detection performance that meets or exceeds suitably qualified MVI under representative conditions (including container formats, product attributes, and relevant defect classes). Regulators expect documented capability studies and ongoing performance monitoring. AVI qualification typically includes probability-of-detection (POD) studies comparing automated results against qualified manual inspection.

100% inspection vs. AQL acceptance sampling (distinct, complementary steps).
• 100% inspection: Every container is inspected with a qualified method; all units with visible particulates are rejected.

• AQL acceptance sampling (per USP <790>): After 100% inspection, a statistical sample of accepted units is re-inspected manually under the <790> conditions, using ANSI/ASQ Z1.4 or ISO 2859 sampling plans. USP <790> specifies General Inspection Level II, single sampling, AQL = 0.65% for visible particulates, with alternative plans allowed if providing equal or better protection.

<1790> provides recommendations; <790> provides the compendial acceptance standard and inspection procedure (including backgrounds, illumination, inspection time, and AQL sampling). FDA’s draft guidance aligns with the “essentially free” principle and notes that, depending on clinical risk, stricter standards than the compendia may be expected to meet CGMP.