The Importance of Sampling Plans for Uniformity of Dosage Testing

Numerous health authorities have cited companies for inadequate sampling plans of finished drug products. A recent 483 issued to a finished drug‑product firm specifically mentions the firm’s lack of a scientifically sound statistical sampling plan during production of all strengths of its drug product.  The FDA also mentions that this firm uses content uniformity results as part of its product’s release, but the firm has not established acceptance limits that would ensure that each lot will pass the tests outlined in USP Chapter <905>, Uniformity of Dosage Units.

The (quite reasonable) expectation of regulators is that a company will devise internal testing requirements to provide assurance that the batch will routinely meet the quality requirements, as specified within the USP with a defined level of confidence, and not rely on applying the USP <905> test at the time of release.  This is critical for process validation and demonstrating the capability of that process to afford a uniform product.  This forms the basis of the USP Stimuli article “Batch Release Testing for Ensuring Compliance with Uniformity of Dosage Units (UDU)” where considerations are provided in relation to both sample size and criteria to provide such assurance with internal test procedures.  The USP Stimuli article states that the key is the definition of a Quality Goal, which for UDU is described in terms of confidence of meeting the USP<905> quality requirements and then to design an internal sampling plan and criterion accordingly.  Such an internal sampling plan is expected to be larger than what is defined within USP <905> and, with criterion more stringent, to provide (for example) 95% confidence in meeting the USP <905> requirements.  It is strongly advised that, when devising such sampling plans and setting such release criteria, statistical expertise is sought so that the sampling plan is statistically appropriate in terms of parent batch representativeness, and the risk of releasing a “bad batch” (customer risk) is acceptable.  The USP Stimuli article provides approaches that can be taken for devising such a sampling plan, such as applying the ASTM E2709 and E2810 standards.

If you have any questions in relation to batch release sampling plans, please contact Paul Mason, Ph.D. at