USP expert answers questions regarding particulate matter testing

ECA Conference: Particles in Parenterals May 7-8, 2013 Vienna, Austria

Conference Questions answered by: Scott Aldrich, Ultramikro LLC, USP Expert Committee

1. Which analytical performance characteristics are required for the (light obscuration) method verification?

Both USP 788 methods provide a robust and reproducible particle content determination which many companies have collected over an extended history, and are appropriate for most product formulae as written.  However, verifying acceptability for specific product families is an important step.  Certainly, the elements of USP <1225> provide comprehensive guidance for validation just as the IST routines endorsed by the vendors.  Clear, low concentration formulations which mix easily often require no further method modifications for validation.  For those formulations of high concentration, or are colored, or viscous, or with specific gravity significantly different from water, effects of the formulation character will require further consideration for method acceptability. Products in very low or large volumes (SVP: <2mL or >50mL and any LVP) will require further consideration for sample mixing and proper sampling (recovery). Why? The size and count determination via light obscuration (LO) is derived from a fraction of each container; therefore, the product sample (pooling and sampling) may affect particle recovery and counting linearity and so should be standardized and verified.  Membrane microscopy (MM) samples the entire (drained) product fills and thus has no similar sampling issues.

There are certain requirements directed by USP for each method; for light obscuration, the instrument standardization tests (IST) found in USP 1788 must be performed for each instrument and found acceptable in a period of not more than 6-months. First is selection of a sensor suitable for the product and particle size range, then the IST routine will verify instrument performance. Note that a system suitability check using the USP particle count reference standard is one of these tests which I would recommend as an interim system check on a weekly or even daily basis. USP 1788 further describes the use of other commercial standards or comparison to a microscopic count for this purpose.  Two additional tests I recommend for LO beyond IST are a) use of a wide range of calibration standards, spanning 2µm - 100µm and b) verification of 80-120% recovery of a known bimodal bead spike in the final sample preparation.

For membrane microscopy, the USP graticule within the microscope system must be verified with a NIST-traceable linear scale to be within 2% of dimension calibration. Consult USP 1788 for discussion regarding both methods.  Overall, I recommend use of product release Targets which are a fraction of the USP limits (1/3 to 2/3) in order to accommodate assay error and as a sensitive monitor of batch variation.

2. What material should be used as a reference material for the method verification (light obscuration)? The problem here is, that no certified count standards are available (for spherical structured proteins)

NIST-traceable bead standards are typically used for light obscuration calibration. Non-spherical standards would be ideal for protein formulations, and are being developed by Dean Ripple, NIST (1) ( ) but are not yet commercially available. Note that for optical particle counting by light obscuration, the size of the particle is determined by correlation of the light blockage for each particle against that for the calibration profile established by truly spherical standards yielding a size of equivalent circular diameter.  As the aspect ratio of the particle deviates from equant (1:1:1 or spherical/cubic), the registered size will have a negative bias for diameter.  Other factors affecting determination of the size are diminishing relative contrast, from particle-medium refractive index similarity and effects of formulation color.  The count may be affected by a) high concentration of the particles (negative bias due to coincidence counting) and b) artifacts such as air bubbles and immiscible liquids such as silicone oil content (positive bias).

The USP PCRS is a mono-disperse 15µm bead suspension with a count range established through a collaborative study using many commercial labs.  Within a lab, especially one instrument, the count will be much more reproducible.  Use your internal data to set acceptable variance expectations. My own criteria would be ±5% inter-instrument, ±10% intra-instrument and ±20% across labs.  This also applies to individuals counting mono-dispersive bead standards by the microscopic method, e.g., ±5% individual, ±10% between individuals and ±20% across labs.

3. How to proceed, if the Light Obscuration test exceeds the limit and the microscopic particle count test does comply? Can the lot be released? Is an OOS procedure necessary?

The membrane microscopic method 2 is used as an orthogonal count to verify the particle population.  Since air and liquids have no effect on the count, and even soft or low refractive index particles will be evident (some may sink into the membrane), it is a reasonable check of the method 1 count.  A passing count by membrane indicates that artifacts likely increased the light obscuration counts.  A passing count by method 2 allows batch release; however, what else is evident in the isolated particles?  Are there many soft or amorphous particles and did you count them?  Is there an evidence of point-source, e.g., a singular particle type that may indicate a detrimental change?  What evidence is apparent to explain the elevated light obscuration count?  I would endorse an OOS if both methods fail.  If method 1 fails and method 2 shows passing results, the batch passes USP (and harmonized) 788 criteria.

4. The draft USP chapter <787> "Subvisible particulate matter in protein injections" states that the microscopic test is not equivalent to the Light Obscuration test. Could the microscopic test be used as a release test (in the case the LO test does exceed the limit)?

Draft USP <787>(2) minimizes the utility of the membrane method 2, since the purpose of the new chapter is size and count for protein and other biologic formulae for ?10µm particles with attention to sub-10µm content.  Further, new chapter USP <1787> (3) provides discussion and recommendations regarding the particle size and nature of the principal proteinaceous ingredient(s) using a variety of techniques to size, count and characterize particle load.  We are in an information-gathering stage, in which new approaches to elucidate the active ingredient character are being added to the standard particle determination methods (LO and MM) both of which certainly capture the presence of extraneous particles.  USP <787> provides refinement of <788> methods, specifically for the more sensitive protein therapeutic formulae.  Microscopic method 2 does not provide useful quantitation of the protein fraction, but does recognize the isolated material and allow enumeration of the other (extrinsic and intrinsic) particles retained.  Other alternate methods may be useful and illuminating, such as flow microscopy, Coulter (electrical sensing zone) and nephelometry.

5. What are the expectations for a statistical assessment with regard to USP chapter <788> "The number of test specimens must be adequate to provide a statistically sound assessment"?

Statistical sampling is recommended by USP, but the sampling methods and plan are not.  Note that the compendial methods and limits apply to the samples run, but do not directly indicate batch acceptability. Many companies use broad sampling during development, and once variation is known and the final product form is stabilized, decrease the number of samples for routine release. Common guidance such as ANSI Z1.4 (Mil Std 105E) and ISO 2859.1 provide direction for sampling and acceptance according to batch size and desired acceptable quality limits (AQL). A strategy using beginning-middle-end sampling with both pooled and individual samples should provide a clear determination of particle load and variance. What are the counts and sizes and what types of particles are isolated? It is expected that both USP 788 methods 1 and 2 will be performed in Development, since one may need method 2 to decide batch release and the qualitative to quantitative information provided by retention of particulate matter is an invaluable aid for process improvement activity. Sampling is also quite dependent upon batch performance, investigation of particle incidents (visible and subvisible), from field experience and after product changes.

References:
(1) Ripple DC, Wayment JR, Carrier MJ.  Standards for the optical detection of protein particles.  American Pharmaceutical Review. 2011, July Issue, available at: http://www.americanpharmaceuticalreview.com/Featured-Articles/36988-Standards-for-the-Optical-Detection-of-Protein-Particles/ visited January 10, 2013.

(2) USP <787> "Subvisible Particulate Matter in Therapeutic Protein Injections", Pharmacopeial Forum 39(2).

(3) USP <1787>, "Measurement of Subvisible Particulate Matter in Therapeutic Protein Injections", Pharmacopeial Forum 39(6).