This article discusses the challenges overcome during the development of a
blend-sampling technique and the successful validation of the blending oper
ation for a tablet dosage form containing 2% active ingredient. Content uni
formity results are discussed for three pilot-scale (15-kg) and seven comme
rcial-scale (150-kg) batches of tablets. Blend and core content uniformity
data from the pilot-scale batches were acceptable. For the initial commerci
al-scale batches, although the tablet core content uniformity data were acc
eptable, the blend uniformity results were poor. The blend data for these b
atches had very high mean values, but acceptable relative standard deviatio
ns (RSDs). This suggested that the drug was being preferentially sampled by
the thief but in a consistent, reproducible manner. Extensive testing was
performed on a commercial-scale development batch to identify potential cau
ses of sampling error. The results of this testing helped define the blend-
sampling technique and strategy used to validate the mixing operation.