FDA Blend Uniformity Guidance: Sampling, ASTM E2810, and OOS Handling

FDA blend uniformity guidance sits in an unusual place: the underlying requirement is codified in 21 CFR 211.110, but the agency has withdrawn two successive draft guidances that would have spelled out sampling and acceptance criteria, leaving manufacturers to work from a risk-based framework built around stratified sampling, a relative standard deviation of no more than 5.0 percent with individual results between 90.0 and 110.0 percent of target, and ASTM E2810 statistical methodology. If you are designing a program or responding to an inspection finding, the practical rules below are what FDA reviewers actually apply, even though no single finalized document consolidates them.

The Regulation That Creates the Obligation

The requirement to prove a powder blend is uniform comes from 21 CFR 211.110(a). It mandates written procedures describing in-process controls, tests, and examinations for each batch, and subsection (a)(3) specifically calls for monitoring the “adequacy of mixing to assure uniformity and homogeneity.”¹ That phrase is the regulatory hook for the entire blend uniformity program.

Two neighboring provisions tighten the obligation. Section 211.100 requires written production and process control procedures designed to assure that drug products have the identity, strength, quality, and purity they claim, with any deviation recorded and justified.² Section 211.110(b) adds that in-process specifications must be consistent with final product specifications and, where possible, derived from historical process data using suitable statistical methods.¹ Read together, they mean a documented, statistically grounded program is required before a blend moves to compression or encapsulation.

Why There Is No Finalized Guidance

FDA has tried twice to publish specific blend uniformity guidance and withdrawn both attempts. The 1999 draft, “ANDAs: Blend Uniformity Analysis,” drew industry pushback on sampling and statistics and was withdrawn on May 17, 2002.³ A second draft, built on Product Quality Research Institute work and titled “Powder Blends and Finished Dosage Units — Stratified In-Process Dosage Unit Sampling and Assessment,” was also withdrawn after FDA concluded that its USP <905>-based criteria provided “only limited statistical assurance that batches of drug products met appropriate specifications and statistical quality control criteria.”⁴

Between 2020 and 2022, FDA’s Office of Pharmaceutical Manufacturing Assessment (OPMA) developed internal review guides for assessing blend and content uniformity of solid oral dosage forms.⁵ These are not published as formal industry guidance. Reviewers instead evaluate each program on its own facts, weighing the product, process, and data against current statistical expectations, particularly ASTM E2810.

When Blend Uniformity Analysis Applies

Blend uniformity analysis matters most for products that require content uniformity testing under the United States Pharmacopeia. USP requires content uniformity testing when the active ingredient is less than 50 milligrams per dosage unit or makes up less than 50 percent of the dosage unit by weight. The 1999 draft used these same thresholds to define when BUA should be performed, and the principle still governs practice.

Blend uniformity testing is broadly expected as a CGMP control for any solid oral dosage form. Products above the USP thresholds still benefit from blend monitoring during process validation. Low-dose and high-potency formulations demand more rigorous sampling than standard protocols, because small absolute shifts in drug concentration translate into large percentage deviations from target.

How to Sample the Blend

Assessing blend uniformity requires pulling samples from multiple locations throughout the blender. The standard approach is stratified sampling: predetermined locations distributed across the three-dimensional space of the blend to capture potential segregation zones, dead spots, and concentration gradients. A common protocol calls for sampling from at least 10 locations, including areas near the top, middle, bottom, and sides of the container.⁵ The sample weight at each location should approximate the weight of a single dosage unit, so results reflect what an individual tablet or capsule would contain.

Most manufacturers use a sampling thief, a hollow probe inserted into the blend at various depths and positions. The analytical method used to quantify the active ingredient in each sample must be validated for accuracy, specificity, and precision before results carry regulatory weight.

The Known Weaknesses of Thief Sampling

Thief sampling is the default, and both industry and regulators recognize its limits. Inserting the thief into a static powder bed can disturb the blend and cause localized segregation at the point of sampling, so the sample pulled may not represent what was there before the probe went in. Active ingredient can adhere to the thief’s metal surfaces, producing consistently low results that reflect the instrument rather than the blend. Achieving consistent depth and location across operators and batches is difficult, adding variability that has nothing to do with mixing quality. These weaknesses are a major reason the industry has moved toward real-time monitoring for validation and development work.

Acceptance Criteria

The core statistical measure for blend uniformity is the Relative Standard Deviation (RSD), which expresses how tightly individual sample results cluster around their mean. Expectations carried forward from both withdrawn draft guidances and current practice set the threshold at an RSD of no more than 5.0 percent.⁶ An RSD above 5.0 percent signals meaningful pockets of higher or lower drug concentration.

Each individual sample result should also fall within 90.0 percent to 110.0 percent of the target active ingredient concentration.⁶ A blend can have a low RSD overall and still fail if one or two locations show concentrations well outside that range. The dual requirement catches both systemic mixing problems and isolated dead spots.

ASTM E2810 Is Now the Working Standard

FDA reviewers expect manufacturers to use ASTM E2810 for evaluating content uniformity data, and the same statistical framework increasingly applies to blend uniformity. ASTM E2810 provides a procedure for computing a lower bound on the probability that a future sample from the lot will pass the USP Uniformity of Dosage Units test, at a prescribed confidence level.⁷ In a published case study, FDA recommended ASTM E2810 sampling plan 2, with 90 percent confidence and 95 percent coverage, for evaluating stratified content uniformity of exhibit, validation, and commercial batches. The same case study flagged that a manufacturer’s failure to properly understand the ASTM E2709/E2810 methodology led to critical deficiencies in its application.⁸ Familiarity with these standards is essential, not optional.

Connecting Blend Results to Finished Dosage Testing

Blend uniformity testing feeds into finished dosage unit testing under USP General Chapter <905>, Uniformity of Dosage Units, and should correlate with it. FDA expects manufacturers to demonstrate that a uniform blend translates to uniform finished units. When a scientifically sound Process Analytical Technology (PAT) monitoring strategy is in place, it can help establish the potential correlation between powder blend samples and finished dosage units.⁴

A blend that passes the 5.0 percent RSD threshold but produces dosage units with high variability points to a downstream problem, in the feed frame, hopper, or compression equipment, that blend testing alone will not catch. The reverse is also informative: consistently uniform dosage units from a blend with borderline results may justify reduced blend sampling in routine commercial production.

Real-Time Monitoring With PAT

Thief sampling gives you a snapshot at a single moment. Process Analytical Technology offers continuous monitoring of the blend as it mixes. FDA’s 2004 PAT framework encourages manufacturers to develop and implement innovative tools for real-time process understanding and control.⁹

Near-infrared (NIR) spectroscopy is the most widely used PAT tool for blend uniformity. NIR probes mounted inside the blender collect spectral data throughout the mixing cycle and track homogeneity in real time. This identifies the actual blending endpoint, the point at which further mixing produces no meaningful improvement, rather than relying on a fixed mixing time validated months earlier. NIR also eliminates the sampling bias inherent in thief probes because it measures the blend in situ.

The barrier is validation. Quantitative NIR methods require significant upfront work to build calibration models, and maintaining those models across formulation changes and equipment transfers is time-consuming, which has slowed adoption at smaller manufacturers. Qualitative NIR approaches, which detect whether the blend has reached a stable, uniform state without predicting exact concentration, are emerging as a simpler option for endpoint determination and development troubleshooting.

Handling an Out-of-Specification Blend Result

When blend uniformity testing produces an out-of-specification (OOS) result, meaning the RSD exceeds the acceptance limit or an individual sample falls outside the acceptable range, 21 CFR 211.192 requires a thorough investigation. Any unexplained discrepancy or failure of a batch to meet specifications must be investigated whether or not the batch has already been distributed, and a written record of the investigation, including conclusions and follow-up, is mandatory.¹⁰

Rule Out Laboratory Error First

The first step is determining whether the OOS reflects a real manufacturing problem or a laboratory error. FDA’s 2022 OOS guidance instructs the analyst to begin assessing the accuracy of results immediately upon obtaining unexpected data, with an objective and timely review by the laboratory supervisor. The assessment covers instrument calibration, sample preparation, analytical method execution, and calculation accuracy. If a clear laboratory error is identified, the original result can be invalidated, and the firm should determine the source of that error and take corrective action to prevent recurrence.¹¹

Then Investigate the Process

When no laboratory error explains the result, the investigation escalates to the manufacturing process. FDA expects these investigations to receive the “highest priority.” The objective is to identify the root cause, which could involve inadequate mixing time, incorrect blender loading sequence, poor raw material particle size distribution, equipment wear, or environmental conditions during processing. FDA sets no fixed number of days for completion, but the standard is “thorough, timely, unbiased, well-documented, and scientifically sound.”¹¹

A confirmed OOS result means the batch does not meet established standards and should be rejected under 21 CFR 211.165(f).¹¹ Rejecting the batch does not end the obligation. The investigation must extend to other batches of the same product and other products that may be associated with the same failure.¹⁰ Corrective and preventive actions must be implemented and documented.

What FDA Actually Cites

Blend uniformity deficiencies appear regularly in FDA warning letters, and the citations cluster around two failures: inadequate sampling plans and incomplete OOS investigations. In a 2024 warning letter, FDA cited a manufacturer for a blend uniformity sampling plan that failed to account for the depth of the mixing bowl and lacked justification for the number of samples collected, the size of each sample, and how samples were prepared for analysis. The same letter cited failures to thoroughly investigate discrepancies, noting that investigations did not consistently identify root causes or establish scientifically justified preventive actions.¹²

Inspectors are not simply checking whether you perform blend uniformity testing. They evaluate whether your sampling plan is scientifically justified for your specific product and equipment, whether your acceptance criteria reflect current statistical expectations, and whether your investigation process is genuinely diagnostic when results go wrong. A sampling plan copied from a template without product-specific justification is a common trigger for findings, as is an investigation that concludes “no root cause identified” without demonstrating the rigor FDA expects.

1
eCFR. 21 CFR 211.110 – Sampling and Testing of In-Process Materials and Drug Products
2
eCFR. 21 CFR 211.100 – Written Procedures; Deviations
3
Federal Register. ANDAs: Blend Uniformity Analysis; Withdrawal of Draft Guidance
4
U.S. Food and Drug Administration. Questions and Answers on Current Good Manufacturing Practice Regulations – Production and Process Controls
5
U.S. Food and Drug Administration. Risk-Based Blend and Content Uniformity Assessment: A Case Study
6
Product Quality Research Institute / Food and Drug Administration. Powder Blends and Finished Dosage Units — Stratified In-Process Dosage Unit Sampling and Assessment Draft Guidance
7
ASTM International. ASTM E2810 – Standard Practice for Demonstrating Capability to Comply With the Uniformity of Dosage Units Test
8
FDA. Risk-Based Blend and Content Uniformity Assessment: A Case Study
9
U.S. Food and Drug Administration. PAT — A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance
10
eCFR. 21 CFR 211.192 – Production Record Review
11
Food and Drug Administration. Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production – Guidance for Industry
12
U.S. Food and Drug Administration. EyePoint Pharmaceuticals, Inc. – Warning Letter 679747