Capability measurement of size-exclusion chromatography with a light-scattering detection method in a stability study of bevacizumab using the process capability indices

• Application of the capability indices to the SEC/LS/RI method using BAS as model.
• The X- and moving range chart confirmed that the method is in-control and stable.
• The method meets the capability requirement and run under conditions preset by the analyst.
• Method was used in stability study of bevacizumab pharmaceutical preparations.
• Method does not comply with the specification limits and a precision control is required.

In this study, we investigated if the size-exclusion chromatography coupled with light-scattering and refractive index detection (SEC/LS/RI) method is fitted for its intended purpose and checked if the analytical method is able to provide enough conforming results. For this, the process capability indices Cp, Cpk, and Cpm were computed. The traditional X-chart and moving range (MR) chart were used by the same analyst to monitor the equipment in the laboratory over a 1-year period. For this, a bovine serum albumin (BSA) sample (0.3 mg mL−1) with a nominal Mw of 66.4 kDa was analyzed each working day. The results confirmed that the analytical method is in-control and stable. To determine whether the given process meets the present capability requirement and runs under the desired quality conditions, the Pearn and Shu (2003) method based on the lower confidence bound C on Cpm was used. The estimator Cpm was 1.81, and the lower confidence bound C was 1.40. We therefore conclude that the true value of the method capability Cpm is no less than 1.40 with a 95% level of confidence. This result indicates that the method is satisfactory and no stringent precision control is required. The usefulness of this method was applied in the characterization of bevacizumab commercial pharmaceutical preparations stored under different conditions that lead to aggregation. In this case, the computed Cpm index was 0.98 (0.70, 1.26), which indicates that the method does not comply with the specification limits and needs to be revised. The quality improvement effort should: (1) reduce the uncertainty in the absolute Mw determination; (2) either move the process mean closer to the target value or reduce the process variation, i.e. improve the method accuracy (μ–T) and precision (σ2). On this point, the Bayesian posterior distribution of the mean and standard deviation pointed out the need to control the precision but specially accuracy in order to reduce the overall uncertainty of analytical method and thus, the method is capable.