– Pharma Safety Solutions, LLC, Noblesville, IN, United States
Background: The safety and effectiveness profile of biologic products could be affected if changes in their manufacturing and purification processes ensued, which are often needed for insulin products.
Objectives: We describe batch-specific safety monitoring program for insulin lispro following the introduction of a manufacturing change in the purification process that was applied in October 2013.
Methods: Multi-item Gamma Poisson Shrinker disproportionality analysis was performed on the manufacturer-maintained database of adverse event reports for insulin lispro. Batch-level analyses included insulin lispro batches for reports submitted between October 2013-October 2018 (post-change batches); insulin lispro batches for reports between January 2008-October 2013 (historic, pre-change batches); and lots for reports between October 2013-October 2018 (concurrent, old and new batches). MedDRA Preferred Terms of interest included changes in insulin lispro effect and hypersensitivity (including immunogenicity, angioedema, and injection site reactions). Empirical Bayes Geometric Mean (EBGM) and 95% confidence interval (EB05-EB95) values were calculated as disproporitonality metrics, with values of EB05≥2.0 are considered potential signals that warrant further review.
Results: About 60% of submitted reports included batch numbers. Of those with recorded numbers, a total of 35,075 reports for post-change insulin lispro batches; 11,302 reports for historic batches; and 32,737 reports for concurrent batches were identified. Disproportionality analysis results were: lack of drug effect [post-change batches, EBGM=1.15 (EB05-EB95=1.12-1.17); historic batches, EBGM=1.06 (EB05-EB95=1.02-1.10); and concurrent batches, EBGM=0.97 (EB05-EB95=0.95-0.99)]; increased drug effect [post-change batches, EBGM=0.73 (EB05-EB95=0.71-0.76); historic batches, EBGM=0.68 (EB05-EB95=0.63-0.73); and concurrent batches, EBGM=0.63 (EB05-EB95=0.61-0.66)]; and hypersensitivity reactions [post-change batches, EBGM=0.72 (EB05-EB95=0.65-0.79); historic batches, EBGM=0.60 (EB05-EB95=0.50-0.70); and concurrent batches, EBGM=0.75 (EB05-EB95=0.68-0.83)].
Conclusions: Insulin lispro is not associated with worsening of glycemic control or hypersensitivity reactions following purification process change. Batch-specific safety monitoring of biologics provides an objective assessment of product comparability before and after changes in the manufacturing process.
