When qualifying a manual visual inspection (MVI) process or an automated visual inspection (AVI) system, pharmaceutical manufacturers routinely face a question that appears straightforward: which particle sizes should be included in the defect kits?
The range of answers encountered in practice is striking. Some facilities select a single particle size — typically 200 µm — and apply a stringent acceptance criterion. Others build test sets with twelve or more incremental sizes. Some choose 250 µm as their lower bound with no documented rationale. This variation is not random; it reflects a genuine gap between what the guidance says and what happens in practice.
What USP ⟨1790⟩ Actually Says
USP ⟨1790⟩ is explicit: visible particle detection is probabilistic. The probability of detection (PoD) is not a fixed property of a particle — it is a function of particle size, shape, color, reflectivity, solution clarity, container characteristics, and lighting conditions.
The guidance establishes that for clear solutions in transparent 10-mL glass vials under diffuse illumination of 2000–3750 lux:
- Particles below 100 µm are generally not detectable
- Routine reliable detection (≥70% PoD) is typically achieved near 150 µm for non-fibrous particles
- PoD reaches approximately 95% at 200 µm for non-fibrous particles
- Reliable detection for fibrous particles generally begins near 500 µm
The chapter also recommends that threshold studies use particle sizes in an approximate range of 100–500 µm, with increments of 100 µm. Detection Threshold studies are the first step in evaluating the performance of any new inspection method.
The Rationale for the Smallest Particle Size
The smallest particle in a test set defines the sensitivity floor of the inspection method. Based on USP ⟨1790⟩, 150 µm is a scientifically justified starting point for non-fibrous particles. Some facilities choose 200 µm instead, which is also defensible and provides a somewhat more robust challenge level. Both are sound, provided the rationale is documented with explicit reference to PoD data.
What is not defensible is selecting 250 µm, 300 µm, or higher as the smallest particle without rationale. Doing so excludes the near-threshold challenge level entirely and qualifies an inspection process against particles that are already comfortably detectable — which tells you very little about the actual sensitivity of the method.
For fibrous particles, the same logic applies. Since reliable detection begins near 500 µm, this is the appropriate lower bound. Seeding a fiber standard below this produces a near-zero PoD unit that is not a meaningful challenge to any inspection system.
Why a Single Particle Size Is Not Sufficient
A test set built around a single particle size — even if correctly chosen — misses the purpose of qualification. The objective is not just to confirm that inspectors can detect a threshold particle. It is to characterize detection capability across the range of challenge levels that reflect the real manufacturing environment: glass fragments, rubber particles, metal particulates, and fibers, each with different optical properties and detection profiles.
Multiple particle sizes across this range serve three purposes:
- They evaluate detection capability at different challenge levels (threshold, mid-range, and high-confidence)
- They provide a more robust statistical basis for qualification
- They allow meaningful comparison between inspectors, shifts, or inspection systems over time
Using only the smallest particle with a high acceptance criterion (e.g., 80–100%) creates a fragile qualification. Because borderline particles are inherently variable in their PoD — subtle differences in morphology, orientation, or buoyancy between nominally identical particles can shift detection probability significantly — applying a tight acceptance criterion to near-threshold particles leads to inconsistent results that are difficult to interpret and defend.
The Opposite Problem: Too Many Size Steps
On the other end of the spectrum, test sets with excessive size increments (e.g., 200, 300, 400, 500, 600, 700 µm and beyond) add handling complexity, calibration burden, and audit documentation overhead without proportionate qualification value. The goal is a scientifically bracketed set — not an exhaustive size ladder.
A Practical Reference Configuration
The following configuration reflects a well-structured approach that balances scientific rigor with operational practicality, aligned with USP ⟨1790⟩ guidance:
This configuration starts at the USP-justified threshold, uses meaningful increment to differentiate challenge levels, and avoids the noise of excessive intermediate steps. It covers the full range from near-threshold to high-confidence detection and generates qualification data that is interpretable, reproducible, and audit-ready.
Conclusion
Particle size selection for visual inspection qualification is a technical decision with direct regulatory and quality consequences. Selecting too few sizes, the wrong lower bound, or an arbitrary range produces a qualification that may pass on paper but fails to genuinely characterize the inspection system. USP ⟨1790⟩ provides the PoD-based framework to make this decision correctly — and documenting the rationale explicitly is what separates an audit-ready program from one that invites questions.
Pritam Pany
Chief Business Officer,
SSI Visual Inspection
SSI Visual Inspection works with 70+ injectable drug manufacturers worldwide for their Visual Inspection solutions and compliance requirements.
Learn more about our Visual Inspection qualification kits →