Accurate prediction of the chemical stability of the pancuronium and its degradation products in injectable solution.
6 October 2021C. Merienne1, R. Guery1, T. Bourdais1, B. Lapras1, S. Filali1, C. Paillet2, F. Pirot1
1 Unité de Préparation et de Contrôle des Médicaments, plateforme FRIPHARM, Pharmacie à usage intérieur, Groupement Hospitalier Edouard Herriot - Hospices Civils de Lyon
2 Pharmacie à usage intérieur, Groupement Hospitalier Edouard Herriot - Hospices Civils de Lyon
Against the threat of a curare shortage, our department has developed in April 2020 an injectable formulation of pancuronium (PC) 2 mg/mL. Its stability study was performed by semi-predictive and ICH methods. The main objective of this study was to compare the prediction of the chemical stability (CS) of the drug by Accelerated Stability Program with the results of ICH method.
Materials and methods
A batch of PC 2 mg/mL was compounded and packaged in colorless type I glasses with a bromobutyl rubber stopper. It was incubated at 5 °C (long term storage condition), 25 °C, 40 °C, and 60 °C (accelerated storage conditions). Chemical analysis was performed at D0, D30, D180, and D365 for the long-term storage and at D0, D15, D30, D45, D60, and D90 for the accelerated environments. PC content was quantified by ultra-high performance liquid chromatography tandem with mass spectrometry (UHPLC-MS). PC, 3-désacetylpancuronium (dPC1), 17 désacetylpancuronium (dPC2) and 3,17-didésacétylpancuronium (ddPC) were quantified at 286,2 ; 265,2 ; 265,2 and 244,2 m/z respectively. Modeling was performed with « R » software. The equations screened were the chemical kinetic models . The fits were evaluated by Bayesian and Akaike Information Criterions (BIC and AIC) and Root Mean Square Error (RMSE). The results of the predictions at D365 / 5°C were compared with experimental data. Then the maximal CS was predicted.
Results and discussion
The best-fitted mathematical models were a zero order reaction model for the PC (BIC = -237,4 ; AIC = -248,1 ; RMSE = 3,98), an autocatalytic model (Prout-Tompkins) for the first dPC1 (retention time (RT) = 1,33 minutes, BIC = 213,3 ; AIC = 202,4 ; RMSE = 0,82), a zero order reaction model for the second dPC2 (retention time (RT) = 1,48 minutes, BIC = 232,9; AIC = 220,0 ; RMSE = 0,96), and a nucleation model for the ddPC (BIC = 131,5 ; AIC = 120,6 ; RMSE = 0,56). Remaining PC content measured and predicted at D365 / 5°C were 96,38 ± 0,82 % and 95,62 ± 1,42 %. The percentage of dPC1, dPC2 and ddPC measured and predicted were 1,00 ± 0,02 % vs 1,35 ± 0,37 % ; 0,42 ± 0,02 % vs 0,34 ± 0,15 %; 0,015 ± 0,002 % vs 0,010 ± 0,004. Thus, the formulation maintain its specified quality during 2 years.
The CS of the PC was successfully predicted for 1 year and was used to predict 2 years of CS for the formulation.
References:  « Accelerated Predictive Stability (APS): Theory and fundamentals of accelerated predictive stability studies », Elsevier 2018. 1st Edition », G. Scrivens, D. Clancy et J. Gerst, Elsevier 2018.