Preparation errors detection using gravimetric control on finished product, despite compliant analytical control or in absence of analytical control possibilities

Introduction
For more than 20 years, gravimetric control of finished product (FP) has been implemented on all chemotherapy preparations in our establishment. Preparation theoretical weight is calculated using a weight database of final preparation material and components (container, solvent volume, tubing, packaging, etc.), taking into account the anticancer agent volume added (densities close to 1) and the operating procedure specific features. This weight is compared with the real weight, taking into account a variability of +/- 2 standard deviations of the infusion bags average weight.
This process, included in the FP control measures necessary for products release, complements in-process controls and, since 2009, analytical controls.
The aim of this work is to show the interest of gravimetric control on FP, by analyzing over a period, non-conformities detected by weighing (WNC).

Material and method
Analysis of WNCs on FP recorded from 01/07/2022 to 30/06/2023 and comparison with the analytical result when an analytical control of preparation was possible.

Result
Over the study period, 16 WNCs were identified, i.e. 0.044% of production, 44% of which related to doseable preparations and 56% to preparations for which analytical controls were not possible.
The WNCs concerning the 7 doseable preparations were distributed as follows: 4 drug reconstitution errors (2 carfilzomib, 1 bendamustine, 1 dacarbarzine), 2 solvent withdrawal volume errors (2 carfilzomib) and 1 anticancer withdrawal volume error (carfilzomib). Analytical control was compliant for all these preparations (-9.9% to +8.6% of theoretical concentration). The 4 preparations produced with a reconstitution error were destroyed, while the 3 others were released after corrective action. The WNCs for preparations for which analytical controls were not possible, corresponded to 8 5-FU in elastomeric pumps and 1 blinatumomab. In all cases, an error was made in volume of anticancer agent added, leading to the destruction of the preparation.

Discussion/conclusion
Gravimetric control on FP therefore makes it possible to detect preparation errors where other controls have failed. It can be applied to all our production, unlike analytical control, which cannot be used for all preparations (limit of detectability, product incompatibility with components of dosing equipment, no validated dosing method (new molecule, clinical trials), packaging in syringes or elastomeric pumps).
It’s easy to implement, very quick for an experienced pharmacist, has a low installation cost and is more robust than in-process control alone. In the event of dosing machine failure, inaccessibility to camera controls or cyber-attack, this process provides additional security for pharmaceutical controls.