Abstract
Background
Chemotherapy doses are usually prescribed on the basis of body surface area but dose banding is emerging as an efficient alternative. Dose banding presents the possibility of in-advance preparation in a Centralized Intravenous Admixture Service.
Aim of the study
To evaluate the long-term stability of 5-fluorouracil at banded doses (700 mg and 800 mg) in polyolefin bags.
Materials and methods
Ten polyolefin bags were prepared under aseptic conditions and stored at 23 ± 2°C for 24 days. Five of them were composed of 14 mL 5-fluorocuracil (700 g) in 100 mL 0.9% sodium chloride solution and the five other of 16 mL 5-fluorouracil (800 mg) in 100 mL 0.9% sodium chloride solution. At defined times, physical stability parameters were assessed: optical densities, pH measurements, visual and microscopical inspections. Solutions concentrations were measured using high-performance liquid chromatography coupled with a photodiode array detector.
Results
No change was observed on pH and optical density measurements during the study period. Visual and microscopical inspections remained free of colour change, precipitate, microagregate or crystal. The concentrations of 5-Fluorouracil in 800 mg bags remained stable for 24 days while the concentration in 700 mg bags showed a stability of at least 17 days.
Conclusion
Five-fluorouracil at banded doses of 700 and 800 mg in polyolefin bags is physicochemically stable for at least 17 days at 23 ± 2°C. These results support the possibility of in advance centralised preparation.
Introduction
Five-fluorouracil (5-FU) is a chemotherapeutic agent commonly used by oncologists. It is the standard therapy for advanced colorectal cancer but it is also used in case of head or breast cancer. This drug is a pyrimidine antagonist that inhibits DNA synthesising process. 1 The tumour cells usually consume more uracil than pyrimidine compared to healthy cells and the use of 5-FU by tumour cells stops their DNA synthesising process.
The ideal dose of chemotherapeutic agent for a specific patient is ‘the highest dose associated with an acceptable toxicity’. 2 Body surface area (BSA) is the most historically used parameter to determine individual antineoplastic drug dose. But there is poor evidence of correlation between the BSA, the dose delivered into the tumour, the clinical activity and the apparition of side effects. 3
The dose banding happens to be another approach to prescribe anticancer treatment. As Plumridge and Sewell defined it, dose banding has to be understood as a ‘system whereby, through agreement between prescribers and pharmacists, doses of intravenous cytotoxic drugs, calculated on an individualized basis that are within defined ranges or bands are rounded up or down to pre-determined standard doses. The maximum variation of the adjustment between the standard dose and the doses constituting each band is 5% or less. A range of pre-filled syringes of infusions, manufactured by pharmacy staff or purchased from commercial sources, can then be used to administer the standard dose’. 4 Dose banding presents the possibility of advance preparation for chemotherapeutic molecules and enables to reduce waiting time of patients in day care units as well as drug wastage. 5
The implementation of dose banding for that molecule drives us to confirm its long-term stability at selected standardised rounded doses (SRD) in polyolefin bags in order to support a centralised production. The choice of determined bands for 5-FU relies on a previous implementation study. 6 In this study, anticancer drugs were selected based on three criteria: the existence of a percentage of standardisable doses greater than or equal to 60%, a maximum of 5 SRD for a molecule and a minimum of one dispensation per week. A major part (61%) of the prepared doses of 5-FU were covered by bands from 700 mg to 800 mg.
The literature provides a lot of references about the stability of 5-FU.7–9 The parameters influencing the stability of the molecule are various (container, solvent, concentration, brand, storage temperature, exposure to light and study period). The specific conditions of bands of 700 and 800 mg stored in polyolefin bags at 23 ± 2°C, classically used in clinics, have never been studied. In the literature, there are references about the stability of the drug at concentrations varying from 0.1 to 50 mg/mL9–13 and fluorouracil has been tested in various containers including glass,14,15 polyethylene, 14 polyvinyl chloride (PVC) bags, 13 PVC or ethylene vinyl (EVA) reservoirs 12 and medical-grade silicon pump11,16 but they are poor references about polyolefin. Finally, the storage temperatures studied cover a range from −20°C to 37°C.13,17,18 Studies show that 5-FU can precipitate at 4°C 19 and that an evaporation may occur at 33°C in reservoirs. 12
Considering the lack of data about the stability of 5-FU in polyolefin bags and the risk of crystallisation at 4°C, the stability of the two selected bands in polyolefin bags at 23 ± 2°C will be assessed in this study.
Objective of the study
The objective of the study is to assess the stability of 5-FU at SRD in polyolefin bags conserved at 23 ± 2°C in order to centralise production.
Materials and methods
Chemical stability by chromatographic method
Solutions preparation
Ten polyolefin bags were prepared under aseptic conditions and stored at 23 ± 2°C for 24 days. Five bags of 700 mg were composed of 14 mL 5-FU (50 mg/mL, Teva Pharma, lot 3490715) in 100 mL 0.9% sodium chloride solution (Macopharma, lot 15L03A) for a final concentration of 6.1 mg/mL; the five other bags of 800 mg were composed of 16 mL 5-FU (50 mg/mL, Teva Pharma, lot 3490715) in 100 mL 0.9% sodium chloride solution (Macopharma, lot 15L03A) for a final concentration of 6.9 mg/mL.
All bags were kept at 23 ± 2°C for 24 days. Two aliquots from each solution were withdrawn at each time point (days 0, 2, 4, 7, 9, 11, 15, 17, 22 and 24). The first went through physical stability testing while the second was frozen. All samples were defrosted at the same time, just before chromatographic analyses in order to reduce the technical variability. 13
Standard solutions
Five levels of standard solutions were prepared in triplicate using the commercial solution of 5-FU 50 mg/mL diluted in distilled water (12 mg/mL, 8 mg/mL, 6 mg/mL, 4 mg/mL and 2 mg/mL) to determine the calibration curve considering the final concentration of the studied solutions.
Quality control solutions
Three quality control solutions were prepared using the commercial solution of 5-FU 50 mg/ml diluted in distilled water (10 mg/mL, 7 mg/mL and 4 mg/mL).
Chromatographic conditions
A high-performance liquid chromatography (HPLC) (Alliance, model 2695, Waters Association, Milford Massachusetts) using a reversed-phase column C18 (Hyperclone™ 3 µm ODS 100 × 4.6 mm, Phenomenex, H16-097915) and its associated pre-column (C18, 4 × 3 mm, Phenomenex, PRD 060439) maintained at 35°C were employed. The isocratic mobile phase was consisting of 5% methanol and 95% phosphate buffer 0.01M (pH 7.5) at a flow rate of 1.0 mL/min. The separation module was coupled to a photodiode array (model 996, Waters Association, Milford Massachusetts) with a wavelength set at 300 nm. The processing module was Empower 3 Software (Waters Association).
All standards, controls and samples were injected into the HPLC system after a 100-fold dilution.
Validation of the HPLC method
The chromatographic method was validated following the international conference on harmonisation (ICH) Q2(R1) guidelines. 20
The three quality control solutions were used to calculate within- (n = 10) and between (n = 9)-day reproducibility.
The linearity was evaluated by twofold serial dilutions (n = 10) in purified water from a 5-FU solution of 50 mg/mL.
Limits of detection (LOD) and quantification (LOQ) were determined using 10 blanks (mobile phase) measurements. LOD and LOQ were calculated as follows:
LOD = mean + 3* SD.
LOQ = mean + 10* SD.
A degradation study was performed to assess the stability-indicating capability of the method. Vials of 5-FU (50 mg/mL) were prepared in neutral, acidic (HCl 0.2 M), alkaline (NaOH 0.2 M) and oxidative (H2O2, 3%) conditions. Acidic and alkaline solutions were neutralised before injection. Solutions were injected immediately after preparation and after 2 and 4 days of preservation at room temperature and at 60°C.
Statistical analysis
As defined in ICH guidelines, 20 the shelf life of a product was defined as ‘the earliest time at which the 95% confidence limit for the mean intersects the proposed acceptance criterion’. Therefore, a unilateral 95% confidence interval on the mean was used to determine the earliest time at which the product concentration falls under 90% of the initial concentration or 95% of the initial concentration when any signs of physical instability exist 21
However, as stated by the Shelf-Life Working Group of the Product Quality Research Institute, the definition of the shelf life should be based on ‘an acceptably small proportion of product exceeding an acceptance criterion’. 22 The 95% unilateral prediction interval was used in addition to the confidence interval to cover, on average, 95% of the distribution.
Physical stability
At each time, optical densities were measured (Genesys 10 UV, Spectronic Unican) at 350, 410 and 550 nm to watch over the apparition of turbidity, as a control of the apparition of subvisible particles . 23 The pH of the solution was also monitored with a glass electrode pH meter (inoLab, WTW GmbH, Weilheim, Germany).
Additionally, solutions were visually inspected in front of black and white backgrounds to identify colour changes, appearance of opacity or particles. A spun aliquot (5 min at 2150 g, Heraeus Multifuge 1S, Thermo Scientific, USA) was also examined under the microscope (Jenamed, Carl Zeiss, Germany) with a 12.5x objective to detect crystals.
Results and discussion
Validation of the method
The within- and between-day reproducibility remained below 1% and 6%, respectively. The concentrations were linear over the range of 0.098 mg/mL–50 mg/mL with a calculated determination coefficient (r2) of 0.999. LOD and LOQ were both < 0.001 mg/mL. The degradation test showed a diminution of the main peak (± 20%), demonstrating the stability-indicating capability of the method 24 (Figure 1).
Forced degradation chromatogram of 5-fluorouracil (FU) after 4 days at room temperature: natural t0 (black), natural t4 (green), acid t4 (dark blue), alkaline t4 (red) and oxidant (light blue).
Physical stability
The optical densities remained stable over the 24 days (Table 1) as well as the pH measurements (bags of 700 mg pH = 8.79 ± 0.05 and bags of 800 mg pH = 8.77 ± 0.05).
The inspection of the solutions in front of white and black background did not show particles, change of colour or opacity. The microscopic examination of the solution after centrifugation did not highlight crystals.
Chemical stability
The concentrations of 5-FU are stable in the 800 mg condition bag for at least 24 days of storage.
The stability of 5-FU in the 700 mg condition bag is different in relation to the statistical method applied (Table 2 and Figure 2). After 17 days of storage, we expect more than 5% of the 700 mg bags to be non-compliant based on the prediction interval. But according to the confidence interval, the concentrations of 5-FU are stable for at least 24 days of storage.
Evolution of the Optical densities during storage.
Evolution of the relative concentration of 5-FU during storage at 25°C. Observed concentrations are represented with their mean ± standard deviation. Fitted values correspond to the regression line and are expressed as a percentage of the baseline fitted value that is 577 mg/100 mL and 618 mg/100 mL for 700 mg and 800 mg doses, respectively.
LL95CI, lower limit of the unilateral 95% confidence interval on the mean; LL95PI, lower limit of the unilateral 95% prediction interval.
Evolution of the relative concentration of 5-fluorouracil (5-FU) during the storage. The thick line represents the fitted mean or regression line while the thin lines represent the lower limit of the unilateral 95% confidence interval on the mean (continuous line) or the lower limit of the unilateral 95% prediction interval (dotted line).
The peak of interest remained free from interference during the study period.
Discussion
The dose banding combined with the centralised preparation of anticancer drugs offers the possibility to rationalise the use of those expensive molecules (reuse and reduction of wastages) while reducing the waiting time of patients in day care units. The aim of this study was the evaluation of the long-term stability of two concentrations of 5-FU in order to validate a dose banding system for this chemotherapeutic agent.
Two different statistical methods, the confidence and prediction intervals, were used to perform the analysis of the results. Although both methods match ICH guidelines, the use of the confidence interval shows an average stability for all solutions studied over the study period while the prediction interval on the mean ensures that each solution studied keeps 90% of its initial content for the study period. If following the confidence interval, 700 and 800 mg condition bags are stable for at least 24 days. The prediction interval is more stringent and shows that 700 mg condition bags start to decrease under 90% between day 18 and day 22. In the context of centralisation, the use of the prediction interval is preferred to ensure the preservation of 90% of the initial content of each bag.
The chromatographic method was validated according to the ICH guidelines. 20 The results of the validation show a between-day reproducibility of 6% which is possibly linked to the 100-fold dilution of the solutions before injection in the system. This coefficient of variation remains under the limit of 10% 25 but could influence the 95% confidence interval on the mean which is affected by the variability of the concentration of the bags and by the variability of the measurements.
The long-term stability of therapeutic solutions of 5-FU could be influenced by several parameters as the concentration, the temperature of storage and the containers. The stability of 5-FU has already been tested in dextrose 5% and sodium chloride 0.9% at different concentrations from 0.5 to 50 mg/mL. The higher concentrations (50 mg/mL, undiluted) are more likely to precipitate over time, especially at lower temperatures ( + 4°C). When lower concentrations are studied (0.5 to 8 mg/mL) at + 4°C, there is no evidence of precipitate.12,13 The storage of 5-FU at room temperature (21–25°C) protects solutions from such precipitate while the storage at higher temperatures (33°C) can have consequences on water loss giving falsely high readings. 12 The containers have also an influence on the long-term stability of the drug given that the containers may interact with the cytotoxic drug and reduce its concentration by phenomenon of adsorption, absorption, permeation or leaching. One hypothesis to explain the shorter stability period for the bands of 700 mg in this study could be an adsorption phenomenon 26 due to the side of the bag. As the surface of the bags is equal for both concentrations, the smaller concentration could be more affected by this phenomenon and show a restricted stability.
Martel et al., 12 Farhang-Ashani et al., 27 Roberts and Sewell, 11 Benaji et al. 28 and Galanti et al. 13 studied those parameters but the specific combination of 5-FU at selected SRD of 700 mg and 800 mg stored at 23 ± 2°C in sodium chloride polyolefin bags for 28 days had never been studied. 29
Conclusion
In our study, 5-FU in 0.9% sodium chloride solutions at selected SRD of 700 mg/114 mL and 800 mg/116 mL are compatible with and chemically and physically stable in polyolefin bags at 23 ± 2°C for at least 17 days.
This study supports a centralised production of 5-FU in accordance with the studied conditions.
Footnotes
Author’s Contribution
C.M., O.S., S.L., H.J.-D., L.G.: Conceptualisation. C.M., O.S., G.N.: Data curation. C.M., O.S., B.B.: Formal analysis. Q.M., O.S., G.N.: Investigation. C.M., C.M.-L., L.G.: Methodology. C.M.-L., J.J., H.J.-D., O.P., L.G.: Supervision. C.M., L.G.: Drafting. All authors reviewed and approved the final version of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
