Validation of an automated method for cyclophosphamide reconstitution and comparison with the manual process

Introduction
Injectable cyclophosphamide is a chemotherapy drug available in powder form. Its slow, time-consuming dissolution requires anticipation. In our hospital, the established method consists of reconstitution of campaigns of 30 vials by hospital pharmacy technicians (HPTs) in an isolator, followed by 30 minutes of automatic agitation. The introduction of the Apothecachemo^® robotic system prompts the reconsideration of automating these reconstitutions. The objective of this study is to define and validate a robotized reconstitution method, and to compare it to the current method.

Materials and methods
Four agitation durations are tested on the robotic system to determine the optimal dissolution time. This duration is validated on three batches of four vials by three HPTs, via visual inspection. A reconstitution test with fluorescein is performed to ensure the absence of chemical contamination, and compare to a positive control. All elements are examined under UV light. The validated robotized method is compared with the manual method through a comparative analysis of campaign reconstitution durations; drudgery and musculoskeletal disorders risks are evaluated and compared by questionnaires (from 0 to 11 points with increasing drudgery) based on the Santé Publique France “Eval-Risk TMS” questionnaire. The accuracy of each method is compared using statistical analysis of deviations from expected vial weight, via a Student t-test.

Results
The shortest dissolution duration of 5 minutes, recommended by the manufacturer, is not sufficient. The validated duration, leaving no visible particles, is 7,5 minutes. The fluorescein test shows no visible traces on any observed elements : vial surface and septum, wipes used to clean the agitator and septum, or the preparation area. The positive control is conclusive. The average duration of the reconstitution per vial are 4,5 minutes and 10 minutes for the manual and the automated methods respectively. The questionnaires evaluating the manual method shows an average score of 7.8/11 across the ten HPTs who responded while the automated method one shows an average score of 1/11 among the four trained HPTs. All HPTs agree on the reduced task strain of the automated reconstitution. The average deviations from expected vial weight are 1.05% (+/-0.66) and 0.58% (+/-0.21) for the manual and automated method respectively, showing a significantly higher accuracy of the automated reconstitution (p<0.05).

Discussion
This work enabled the validation of a less physically demanding, more accurate automated reconstitution method, free of chemical contamination. Further optimization of campaign duration is required. A particle counter has been acquired to improve the detection of non-visible particles in solutions.

See the poster