API Particle Size Distribution: Do You Need a Limit?

API particle size distribution is the share of an API powder found in each size band. D10, D50 and D90 are the three sizes that summarize it. D10 is the size below which 10 percent of the sample lies. D50 is the median. D90 is the size below which 90 percent lies. USP <429> defines a percentile size as the particle size below which m percent of the distribution lies, and results are on a volume or mass basis unless stated otherwise.

Three numbers are needed because one hides the tails. D50 shows where the middle sits. D10 shows the fine fraction and D90 the coarse fraction. Two lots can share a D50 and differ widely at D90.

As a quick width check we use span, which is (D90 minus D10) divided by D50. This is our working convention, not a pharmacopoeial limit. A lot with D10 of 2.5 micrometers (µm), D50 of 10 µm and D90 of 20 µm has a span of (20 − 2.5) ÷ 10 = 1.75. That lot comes from the FDA example below. For the basics of what counts as an API, see what an API is in pharma.

An API needs a PSD limit when the drug product is a solid, or a liquid containing undissolved drug, and particle size is critical to at least one of five attributes. ICH Q6A Decision Tree #3 lists them:

1. Dissolution, solubility or bioavailability.
2. Drug product processability.
3. Drug product stability.
4. Drug product content uniformity.
5. Maintaining product appearance.

A yes to any one means an acceptance criterion should be set. A no to all five means none is required. For solution dosage forms, no drug substance particle size criterion is required at all. ICH Q6A, section 3.1(b) adds that for some drug substances in solid or suspension products, particle size "can have a significant effect on dissolution rates, bioavailability, and/or stability", and that testing and acceptance criteria should then be provided.

Our recommendation is to answer question 1 with data. Make the finished product from API of two or three different particle sizes and run dissolution testing. If the profiles differ, you need a limit. The API here is the drug substance, the term regulators use for it.

FDA's example set the limits from dissolution data on product made with different API particle sizes. The FDA example Quality Overall Summary (QOS) describes an immediate-release 2 mg tablet in a generic drug application. The drug is the fictional "Ersatzine". It is poorly soluble under the Biopharmaceutics Classification System, which groups drugs by solubility and permeability. The aqueous granulating fluid carries suspended drug, so particle size affects the product.

The specification uses laser diffraction:

The room under the limit is our subtraction of the lot result from the limit. The justification in the QOS is for D90. The limits "were set based on product development studies that evaluated the dissolution" of product made with different drug substance particle sizes, and material with a D90 below 30 µm "provided an acceptable dissolution profile." The study compared D90 values of 30, 60 and 120 µm.

The same document shows why smaller is not automatically better. It says the poor flow of the micronized drug substance, and its tendency to form agglomerates and adhere to equipment, led the developers to wet granulation rather than direct compression.

Our reading of the table: all three limits are one-sided maximums. D10 has no lower bound, so excess fines are not capped. Add a lower D10 limit if fines cause flow, dust or agglomeration problems in your process. The example is an FDA illustration for a fictional drug. Its numbers are not targets for your API.

USP <429> sets repeatability limits for one laboratory's replicate preparations of the same batch (at least three): a coefficient of variation of less than 10 percent for D50 and no more than 15 percent for D10 and D90, both doubled below 10 µm. The coefficient of variation is the standard deviation divided by the mean.

Applied to lot 15531, this is our arithmetic, not a USP figure. It describes scatter among replicate preparations of one batch in one laboratory:

Our reading of the doubling: a D50 of exactly 10 µm is not below 10 µm, so the 10 percent limit applies. A lot with D50 under 10 µm gets 20 percent.

The linked USP chapter sets no limit for agreement between two laboratories. The clause must therefore fix the method and dispersion conditions, and our recommendation is to agree a tolerance between supplier and buyer in the same clause. USP says method development must show that the dispersant (the liquid or air that separates particles) breaks up agglomerates without breaking primary particles. Wet versus dry dispersion, dispersant and dispersion energy therefore belong in the specification, not left to each lab.

A particle size clause should state the method, the dispersion conditions, all three percentiles, the number of preparations and the evidence behind the limits. Here is a template for a quote request or quality agreement. NLT means not less than.

Test: Particle size distribution, drug substance
Method: Laser diffraction per USP <429>, volume basis
Dispersion: [wet or dry]; dispersant [name]; dispersion energy [setting];
            obscuration range [x to y %]
Limits: D10 [NLT ___ / NMT ___ µm]
        D50 [___ to ___ µm]
        D90 NMT [___ µm]
Replicates: at least 3 representative preparations per batch;
            report mean and CV for D10, D50 and D90
Basis for limits: [dissolution or content uniformity study reference]
Change control: notify buyer before any change of site, mill type
            or milling process; compare PSD of 3 consecutive
            new batches with 3 or more earlier batches

Obscuration is how much of the laser beam the sample blocks. Fill the limits from your own dissolution data, as the FDA example does. The change-control line is explained below. It belongs in a pharmaceutical quality agreement.

Check each COA line against the clause and mark it pass or fail. Ask for the COA and specification of the API you are evaluating, and read the PSD result like this:

A fail on the first row means the number cannot be compared with yours, whatever its value. Our guide to API quality testing and COA review covers the identity and purity lines of a COA that sit around the PSD result.

A change of source or milling process is when PSD is most likely to move, so repeat the PSD comparison on the new material. The FDA draft guidance on postapproval changes to drug substances (September 2018, nonbinding) calls solid-state form and particle size "critical for evaluation when a manufacturing change is made." It cites changing particle size reduction from dry milling to fluidized bed jet milling as an equipment change that needs attention.

The draft states that proving PSD equivalence is hard. Particle shape affects different measurement techniques, and there are no standard benchmarks for equivalence. So PSD data alone may not be enough for some changes. For one of its options, it says data should generally be obtained for three consecutive post-change batches and compared with three or more consecutive, representative pre-change batches. Another option adds drug product testing.

The draft says content uniformity is more at risk when the drug substance is a small share of the formulation, is a solid, and goes through a size-sensitive process such as dry granulation.

Our recommendation is to copy that three-versus-three comparison into the change-control line of the clause. Evaluate a trial lot against the clause before committing to commercial volume.

Send the clause as your specification with the quote request. We ask for the API name or CAS number, quantity, destination, grade or purity, documents needed, payment terms and Incoterms (International Commercial Terms, which set shipping responsibilities). A CAS number is the unique registry number for a chemical substance. We then check availability and sourcing options for that API and issue a commercial quotation.

We support sample and evaluation quantities for qualified business buyers. Sample quantity, pricing and documentation are confirmed product by product, subject to the API, availability and applicable regulatory requirements. Available documents include the COA, specifications and the safety data sheet (SDS/MSDS) where applicable. Test the evaluation lot against your clause. After successful evaluation and approval, we support the commercial and bulk requirement.

Prices are not published. We quote on the current rate once the API name and CAS number are provided. Send your clause through the contact form, by email to sunil@aarisepharma.com, or by WhatsApp on +91 9899783330.