3PC-022 Performances’ evaluation of a new quality control system: the spectrophotometer druglog

BackgroundThe good preparation practices recommend a control of chemotherapy compound to minimise the risk of mistakes before patient’s administration.We tried to put in place an automated analytic method taking into account financial and technical criteria to control IV chemotherapy preparations.PurposeThe purpose was to evaluate the identification and qualification performances of a new quality control system: the spectrophotometer (Druglog, pharmacolog, Uppsala).Material and methodsThe system is based on absorption spectroscopy in the ultraviolet and visual spectral range calculated according to the Beer-lambert law.Eight cytotoxic drugs were initially chosen based on level of clinical use.The statistical approach used for validation referred to the International Conference Harmonisation.1 Each drug calibration was made in triplicate the same day. These operations were repeated for three different sets and three different days. A mathematical model of linear regression representing the relationship between absorbance and concentration of the molecule has been applied. We were able to determine statistically the best calibration curve for each drug through the correlation coefficient’s (R²).Each validation standard was then analysed by the spectrophotometer to determine the relative error of the concentrations measured.Finally, a number of tests were performed on cytotoxic infusion kits.ResultsAll calibration curves present a linear profile: R² average and maximal are higher than 0.98 except irinotecan with R²=0.96.The system has specifically identified each validation standard. The results show that the system can measure all compounds with a relative error less than 12%.The tests in production are presented in the following table.Abstract 3PC-022 Table 1 Drug Mean error (%) Paclitaxel 22.14 5FU 9.6 Carboplatin 17.51 Cisplatin 5.72 Epirubicin 1.86 Gemcitabin 11.4 Irinotecan 20.52 Oxaliplatin 5.2 All drugs were correctly identified.ConclusionWe can doubt the linearity of the irinotecan calibration and enquire about any matrix effect, or even a stability problem with NaCl diluent.During production tests, no kits were found with erroneous drugs. The dosing error rates were higher than those of the validation standards. We have considered a problem with the homogenisation kits.These first results are promising but we will therefore continue the collaboration with Pharmacolog and Uppsala to perform our analytical method of control.Reference and/or Acknowledg