Blend Uniformity Evaluation (BUE)
Blend Uniformity Evaluation (BUE) is the statistical demonstration that a powder blend's active ingredient is distributed evenly enough — at the dosage-unit scale — to deliver content uniformity in finished product. It is the bridge between the blender and the tablet press: pass BUE and downstream content uniformity is plausible; fail BUE and the batch is dead before compression.
On this page · 6 sections
How does Blend Uniformity Evaluation (BUE) apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01What BUE is
Blend Uniformity Evaluation samples the powder blend at multiple defined locations in the blender (or discharge stream) and assays each sample for active. The statistic that matters is the relative standard deviation (%RSD) of the assay results and how each location's potency compares to the target. Acceptable BUE is a prerequisite for blend release into compression, encapsulation or filling.
- Sample size targets the dose mass (1×–3× the dosage unit mass) — testing larger samples obscures segregation.
- Locations are stratified across the blender geometry — top/middle/bottom, corners, dead zones.
- Acceptance criteria are typically mean within 90.0–110.0% of label claim and %RSD ≤ 5.0%, tightened during PPQ/CPV.
- BUE is in-process — distinct from finished-dosage-unit content uniformity per USP <905>.
02The acceptance numbers
| Stage | Sample plan | Typical acceptance |
|---|---|---|
| Development/validation | 10 locations × 3 replicates | Mean 90.0–110.0%, %RSD ≤ 5.0%, individuals 75–125% |
| Routine in-process (CPV) | 10 locations × 1 (stratified) | Mean 95.0–105.0%, %RSD ≤ 5.0% |
| Tight-risk product (HBEL/low dose) | 10 locations × 3 | %RSD ≤ 3.0%, individuals 90–110% |
| Investigation / OOS follow-up | Re-sample full plan + dosage-unit content uniformity | Per protocol |
03Sampling mechanics — the thief problem
Powder-blend sampling is notoriously biased. The traditional sampling thief disturbs the powder bed as it enters, preferentially captures coarser or finer fractions depending on geometry, and can show false non-uniformity that disappears with stream sampling. FDA's 2003 draft guidance explicitly accepts stratified sampling at the press as an alternative when thief sampling is unrepresentative.
- Side-sampling thieves disturb less than end-loading thieves but still bias in cohesive blends.
- Stream sampling (drum/IBC discharge) collects the actual material going to the press — closer to reality.
- Wireless or NIR in-blender monitoring (PAT) removes the thief problem entirely but requires method validation under ICH Q14.
04Common mistakes
- Sampling 10× the dosage mass — masks small-scale segregation and gives false confidence.
- All samples from the top of the blender — misses bottom dead zones.
- Treating thief-sampling bias as a real result — leading to over-blending and unnecessary segregation.
- Skipping stratified press sampling for content uniformity — losing the regulator-preferred predictor.
- Using <905> limits as BUE limits — too loose; an in-process test must be tighter than the release test.
- Not investigating individual-location failures even when the mean passes — segregation hides in the spread.
05Cross-industry examples
- Solid oral pharma — BUE is mandatory at PPQ and routine in-process per FDA expectation.
- Dietary supplements (21 CFR 111) — equivalent uniformity testing required at the blend step per 111.110.
- Veterinary medicated feed — Type B/C medicated articles require uniformity per CVM GFI #135.
- Biopharma drug-substance powder fill — pre-fill blend uniformity supports vial-to-vial dose accuracy.
- Powder-fill cosmetics — bulk uniformity supports MoCRA stability and labelled potency claims.
06How V5 Ultimate handles BUE
Frequently asked questions
Q.How many locations must I sample?+
FDA's 2003 draft is the de facto standard: 10 locations, 1–3 replicates per location, sample size 1–3× the dosage unit. Some sites go to 20 locations for high-risk or low-dose products.
Q.Can I rely on NIR PAT instead of thief sampling?+
Yes, once the NIR method is validated per ICH Q2/Q14 and demonstrated equivalent or better than the reference assay. Many sites run both during PPQ and reduce to PAT-only in CPV with periodic verification.
Q.What if BUE fails?+
The blend is on hold pending investigation. Common causes: insufficient mix time, segregation during transfer, sticky-API agglomeration, sampling artefact. Re-blending is sometimes permitted with prior validation; otherwise reject.
Q.Is BUE required for capsules?+
Yes. Capsule fill weight uniformity is not a substitute for active uniformity in the blend.
Q.How tight should %RSD be for low-dose actives?+
Tighter — typically ≤3.0% for actives under 5 mg/dose. Segregation has more impact when the API mass fraction is low.
Primary sources
Further reading
Want to see how Blend Uniformity Evaluation (BUE) could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.