Centralized preparation of injectable cytotoxic drugs in hospital pharmacy is implemented to protect the final product and to protect handlers against exposure to hazardous drugs. Due to this centralization more than 50 drugs are used simultaneously in a contained environment (laminar airflow hoods or isolators). The risk of cross contamination is thus theoretically important and the potential risk for patients is varying according to specific toxicity of drugs. Some of them, in particular the monoclonal antibodies, have a lower toxicity and are not included in the NIOSH or IARC lists (1, 2). The risk for patients be that they could be then to be exposed to low levels of drugs more toxic than the drugs that they should receive.
Several studies have shown that external contamination cannot be excluded; for example, some studies indicated that vials delivered from pharmaceutical companies may be contaminated on the outside surface with the cytotoxic drug (3–5). Other studies have shown that aseptic manipulation using a syringe and needle technique resulted in environmental contamination which can be minimized by closed-transfer systems. Nevertheless, the internal contamination may be not significant due to the use of closed transfer method for the preparation of hazardous drugs (6–10).
However, human errors during handling process cannot be excluded and could lead to cross contamination. The objective of the study was therefore to assess the risk of cross-contamination between different cytotoxic preparations by simulating the potential errors of preparation with tracer solutions. We considered in our work the most commonly worldwide used process which is manual aseptic preparation and considered that the facility where the preparations may be performed (i. e. biosafety cabinet or isolator) will hardly impact on the improper reuse of devices by operators.
Materials and methods
Tracers solutions was produced from retinol (Vitamine A Nepalm 100,000 UI / 2 ml, Haupt pharma livron, Livron sur Drôme, France) diluted in Water For Injection (WFI) at the final concentration of 6 mg/ml and from thiamine chlorhydrate (Cooper, Melun, France) diluted in WFI at the final concentration of 50 mg/ml.
The handling material used were syringes (Becton Dickinson, Drogheda, Ireland), needles (Becton Dickinson, Drogheda, Ireland), pads (Tetra, Annonay, France) and spikes (ICU medical, Hannover, Germany).
The HPLC system consisted of a pump (L-2130, VWR-Hitachi, Tokyo, Japan) and a diode array detector (L2455, VWR-Hitachi, Tokyo, Japan). The chromatographic peaks were analyzed using the software EZChrom elite (VWR-Hitachi, Tokyo, Japan).
For retinol, the column used was a Hypersil C8 (250 mm × 4.6 mm, id=5 µm) (Shandon, Eastmore, United kingdom). The mobile phase consisted of methanol (Sigma-Aldrich, Steinheim, Germany)/dichloromethane (Merck, Darmstadt, Germany) (95/5 v/v). The volume of injection was 50 µl. The flow rate was 1 ml/min. The peaks were analyzed at a wavelength of 325 nm.
For thiamine, the column used was a C18 Nucleodur Gravity (Macherey-Nagel, Düren, Germany) (250 mm× 4.6 mm, id=5 µm). The mobile phase consisted of a 0.05 M phosphate buffer (sodium phosphate dibasic, Sigma-Aldrich, Steinheim, Germany) pH=2.9/Acetonitrile (Sigma-Aldrich, Steinheim, Germany) (99/1 v/v). The volume of injection was 50 µl. The flow rate was 1 ml/min. The peaks were analyzed at a wavelength of 254 nm.
Choice of tracers
Tracers which simulate the physicochemical characteristic of cytotoxics were used in place of cytotoxic drugs to minimize the risk of toxicity.
The choice of tracers was based on a review of the physicochemical properties of most cytotoxic drugs. This review was conducted using the databases Toxnet (11), Drugbank (12) and the European pharmacopeia (13). The characteristics selected were the log P and the solubility in water. In a second step tracers which have similar physicochemical characteristics were investigated. The selected tracers exhibiting less toxicity were retinol and thiamine.
Usual handling process
The pad is used in the process to limit the dispersion in the environment of droplets at the time of connection or disconnection of the different transfer devices (syringe, needle and spike). The syringe is used for transfer and withdrawal of the drug from vial to final content (e. g. bag). To ensure homogenization of the drug in the final bag, five flushes with the same syringe is applied. The needle and the spike are used for withdrawing liquids from the vials and transferring to final contents.
Selected handling errors
The whole team involved in the production of chemotherapy (pharmacists, operators and control technicians) considered possible sources of cross-contamination coming from human errors.
Four types of handling errors were identified. Errors associated with the re-use of a disposable transfer device i. e. syringe, spike and needle, and errors associated with surface contamination such as a contaminated pad. Thus a total of six operators have carried out four types of simulations repeated three times. Each preparation was performed with two tracers simulating hydrophilic and lipophilic drugs.
The process performed to quantify the rate of cross-contamination related to the re-use of a disposable transfer device consisted of preparing a 5 ml syringe of tracer (thiamine or retinol) and then preparing a 5 ml placebo syringe of WFI with the same disposable transfer device (syringe, needle or spike). The process performed to quantify the level of cross-contamination due to a surface contamination has consisted in the contamination of a pad with a volume of 2 ml of tracers (retinol or thiamine) and then by the preparation of a syringe of WFI using the contaminated pad. The same process without the use of a tracer served as a control. The tracers were quantified in the placebo syringes of water. All operators involved in the simulation were previously trained and qualified for cytotoxic aseptic compounding.
The simulated area used was uncontrolled areas but was an easily cleanable working room.
Quantification of tracers
Tracers were quantified separately by HPLC. The validation of methods for the determination of retinol and thiamine was validated according to the criteria of SFSTP (14) and ICH (15). For thiamine linearity was validated to 1–2,000 µg/ml. The coefficient of variation (CV) of repeatability was 2.12 %. The reproducibility CV was 2.75 %. The accuracy of the method was 96.2 % (IC : 91.2 –101 %). The limit of detection (LD) was 0.2 µg/ml. The limit of quantification (LQ) was 1 µg/ml. For the retinol linearity was validated to 1–200 µg/ml. The CV of repeatability was 2.55 %. The reproducibility CV was 4.95 %. The accuracy of the method was 94.9 % (IC: 88.6–101 %). The LD was 0.3 µg/ml. The LQ was 1 µg/ml.
Statistical comparisons were performed between operators, tracers and critical cases using ANOVA statistical test.
Among all cytotoxic drugs, it is possible to differentiate them by their lipophilic affinity and solubility in water. Some drugs are hydrophilic and soluble in water like cisplatin, 5-Fluorouracile, or cyclophosphamide, others are lipophilic and poorly soluble in water like carmustine, docetaxel or paclitaxel (Table 1).
In order to simulate the real cases, it was necessary to both select a lipophilic tracer and an hydrophilic one which can be diluted in aqueous solution.
We selected thiamine, hydrophilic and freely soluble in water and retinol, lipophilic and practically insoluble in water. In addition, retinol has the advantage of sharing the same adjuvant as paclitaxel i. e. polyoxylated castor oil.
Cross contamination assessment
Tables 2 and 3 give the results of cross contamination levels obtained in the placebo syringes for the four simulation processes of handling errors with thiamine and retinol. To facilitate the comparison between the results, they are expressed by volume (µl) taking into account the concentrations of the solutions of thiamine (50 mg/ml) and retinol (6 mg/ml). The conversion in volume was carried out following two steps. First, by multiplying the concentration measured by HPLC by the volume of 5 ml of the placebo syringe using the formula C=m/V where C is concentration (µg/ml), m is the mass (µg) and V is the volume (ml). This allows to calculate an amount (µg) of tracer in the placebo syringe. This amount is subsequently transformed by volume taking into account the respective concentrations of thiamine or retinol using the same formula C=m/V. Whether by thiamine or retinol the results show rates of contamination above the limit of quantification with the re-use of a syringe, a needle or a spike (Figure 1). In contrast, the use of contaminated pad failed to detect contamination above the threshold of quantification.
The results of the statistical analysis show that whatever the tracer and whatever the critical case there was no statistically significant difference between operators (Table 4). Thus all operators were merged into a single homogeneous group.
For the tracers, there was no statistically significant difference with the re-use of a needle (p=0.98), and pad. In contrast, a statistically significant difference was found with the syringe and the spike between retinol and thiamine (p < 0.001).
Finally, it has been shown that the four critical cases are not equivalent between themselves (statistically significant difference p < 0.001). Whatever the tracer, the risk of cross-contamination was in decreasing order spike > needle > syringe.
Using the experimental results obtained with the tracers, it can be extrapolated cross-contamination quantity for cytotoxic drugs using the following equation (Figure 2). For example, in case of inappropriate re-use of a disposable medical device such as spike, the rate of cross-contamination may exceed 1 mg. The extrapolated rates calculated for the most common cytotoxic drugs are given in Table 5.
Our study showed that handling errors due to inappropriate re-use of single use devices i. e. syringe, needle spike are potential sources of cross-contamination. In contrast, surface chemical contamination of the working environment would be at low risk for cross-contamination. Despite the large surface contamination with cytotoxic drugs found on working surfaces (16–20), the manufacturing process using only enclosed vials and transfer systems remaining in place throughout the duration of transfer according to good manufacturing guidelines (21) would dramatically limit transfer of contaminant from the outside to the inside of a preparation. This is confirmed by our study which was using a worst case surface contamination of 4 mg/cm2 and 0.24 mg/cm2 for thiamine and retinol respectively in comparison to real surface contamination usually found around 100 ng/cm2. We can safely conclude that surface contamination is unlikely involved in potential cross-contamination.
On the other hand, inappropriate re-use of a single use device cannot be excluded due to the human management of the preparation. Few studies investigated the error rate of injectable preparation when performed in hospital pharmacies. One study, in five hospital pharmacies in USA estimated a residual mean error rate of 9 % (6–10 %) (22). More recently, a wide study performed in the UK over 4 years on centralized injectable preparation units (43 pharmacies) found a 0.49 % error rate (23). The main cause can be related to the human individual error.
Because of the complexity of the manual process, more than 50 different drugs simultaneously prepared each day in the same controlled area, with multiple handling actions (i. e. injection, transfer, dilution), the human errors are utmost possible. Furthermore, the inappropriate re-use of single use device is very difficult to detect during the process of preparation and may be enough frequent to be problematic. In our study we did not investigate “closed transfer devices” (10) in that case, the risk of reuse of the withdrawal device is theoretical. In fact, withdrawal device used in place of needle or spike is permanently connected to the vial and cannot be reused, however one can expect that the improper reuse of the syringe is still possible and will lead to the same risk.
The cross-contamination rate found with the tracers when extrapolated to cytotoxic drugs can give significant rate of contamination.
We clearly noticed a difference of rate of contamination between lipophilic tracer and the hydrophilic one. The difference may be explained by the possible interactions of the lipophilic drug with the surface of plastic devices as previously described for carmustine (24), or possible adsorption of cytotoxics on the filter of spikes (25). Moreover, the process can introduce variability of the rate of cross-contamination depending on the lipophilic characteristic of the drug. With the usual process, syringes are systematically flushed five times to ensure homogenization of the preparation. To assess the impact of flushing syringes on cross-contamination, a complementary experiment was carried out without flushing. The lack of flushing was showing a significant impact (p < 0.001) on cross-contamination in comparison to flushing condition with a much higher contamination rate of 30.6 µl versus 2.4 µl and 17.3 µl versus 1.1 µl for retinol and thiamin respectively. Moreover, for both tracers, the inappropriate re-use of the spike was higher than needle, which can be easily explained by the higher dead volume in spike in comparison to needle. In summary, the process, the device and the chemical characteristic of the drug would affect the rate of cross-contamination.
Then the question is to know if the contamination observed may have consequences for the patient. According to the GMP for the cleaning validation (26) and applying the formulae used in industry published by Fourman and Mullen (27) with the criterion limit of 10 ppm, it can be found that the maximum admissible residue may be exceeded.
However, these limits do not take into account the available pharmacological and toxicological data. According to current regulatory practice it is assumed that genotoxic compounds have the potential to damage DNA at any level of exposure and that such damage may lead to tumour development (28, 29). For genotoxic active substances for which there are no discernible thresholds, it is considered that any level of exposure carries a risk. However, a pre-defined level of acceptable risk for non-threshold related genotoxics has been established in the form of the Threshold of Toxicological Concern (TTC) of 1.5 µg/person/day. With these thresholds, the risk exists for almost all preparations.
Thus the guidelines (21, 30) which recommended to separate the preparation of cytotoxic drugs from others injectable treatments such as monoclonal antibodies or at a minimum to work using segregated preparation by campaign seems to be fully justified.
In conclusion, external surface contamination seems to have little impact on the risk of cross-contamination but in contrast, the manual process can lead to errors of re-use of medical devices. Such errors present low or no detectability and can potentially have consequences for the patient. It is therefore important to find ways to minimize this risk, in particular separated areas for cytotoxic drug preparation. Automation and batch production would be interesting perspectives to limit the human error risk.
3. Nygren O, Gustavsson B, Ström L, Friberg A. Cisplatin contamination on the outside of drugs vials. Ann Occup Hyg 2002 May;46:555–7. Google Scholar
4. Connor TH, Sessink PJ, Harrison BR, Pretty JR, Peters BG, Alfaro RM, Bilos A, Beckmann G, Bing MR, Anderson LM, Dechristoforo R. Surface contamination of chemotherapy drug vials, and evaluation of new vial-cleaning techniques: Results of three studies. Am J Health Syst Pharm 2005 Mar;62:475–84.Google Scholar
5. Fleury Souverin S, Nussbaumer S, Mattiuzzo M, Bonnabry P. Determination of the external contamination and cross-contamination by cytotoxic drugs on the surfaces of vials available on the Swiss market. J Oncol Pharm Pract 2014 Apr;20:100–11. CrossrefWeb of ScienceGoogle Scholar
6. Harrisson BR, Peters BG, Bing MR. Comparison of surface contamination with cyclophosphamide and fluorouracil using a closed-system drug transfer device versus standard preparation techniques. Am J Health Syst Pharm 2006 Sep 15;63:1736–44. CrossrefGoogle Scholar
7. Nygren O, Gustavsson B, Strom L, Eriksson R, Jarneborn L, Friberg A. Exposure to anti-cancer drugs during preparation and administration. Investigations of an open and closed system. J Environ Monit 2002 Oct;4:739–42.CrossrefGoogle Scholar
8. Spivey S, Connor TH. Determining sources of workplace contamination with antineoplastic drugs and comparing conventional IV drug preparation with a closed system. Hosp Pharm 2003 Feb;38:135–9.Google Scholar
9. Tans B, Willems L. Comparative contamination study with cyclophosphamide, fluorouracil and ifosfamide: standard technique versus a proprietary closed-system. J Oncol Pharm Pract 2004 Dec;10:217–23.CrossrefGoogle Scholar
10. Sessink PJ, Rolf M-AE, Ryden NS. Evaluation of the PhaSeal hazardous drug containment system. Hosp Pharm 1999; 34:1311–17. Google Scholar
12. Wishart DS, Knox C, Guo AC, Shrivastava S, Hassanali M, Stothard P, Chang Z, Woolsey J. DrugBank: a comprehensive resource for in silico drug discovery and exploration. Nucleic Acids Res 2006 Jan 1;34(Database issue):D668–72. CrossrefGoogle Scholar
13. EDQM (European Directorate for the Quality of Medicines and Healthcare). European pharmacopoeia, 8th edn. Strasbourg: Council of Europe; 2014. Google Scholar
14. Caporal J, Nivet JM, Algranti P, Guilloteau M, Histe M, Lallier M, N’Guyen-Huu, Russotto R. Guide de validation analytique. Rapport d’une commisson SFSTP I méthodologie. STP Pharma Pratiques 1992 2:205–26. Google Scholar
15. ICH Guideline. Q2(R1) Validation of analytical procedures: text and methodology. [Internet]. 1994 [cited 2015 Nov 30] Available from: http://www.ich.org/products/guidelines/quality/quality-single/article/validation-of-analytical-procedures-text-and-methodology.html.
16. Schmaus G, Schierl R, Funck S. Monitoring surface contamination by antineoplastic drugs using gas chromatography-mass spectrometry and voltammetry. Am J Health Syst Pharm 2002 may 15;59: 956–61. Google Scholar
17. Connor TH, Anderson RW, Sessink PJ, Broadfield L, Power LA. Surface contamination with antineoplastic agents in six cancer treatment centers in Canada, and United States. Am J Health Syst Pharm 1999 Jul 15;56:1427–32.Google Scholar
18. Schulz H, Bigelow S, Dobish R, Chambers C. Antineoplastic agent workplace contamination study: The Alberta cancer board pharmacy perspective. J Oncol Pharm Pract 2005 Sep;11:101–9.CrossrefGoogle Scholar
19. Crauste-Manciet S, Sessink PJ, Ferrari S, Jomier JY, Brossard D. Environmental contamination with cytotoxic drugs in healthcare using air pressure isolators. Ann Occup Hyg 2005 Oct; 49:619–28.Google Scholar
20. Nussbaumer S, Geiser L, Sadeghipour F, Hochstrasser D, Bonnabry P, Veuthey JL, Fleury-Souverain S. Wipe sampling procedure coupled to LC-MS/MS analysis for the simultaneous determination of 10 cytotoxic drugs on different surfaces. Anal Bioanal Chem 2012 Mar;402:2499–509. CrossrefGoogle Scholar
21. ANSM (Agence national de sécurité du médicament et des produits de santé). Bonnes pratiques de préparation. [Internet]. 2007 [cited 2015 Nov 30] Available from: http://ansm.sante.fr/Activites/Elaboration-de-bonnes-pratiques/Bonnes-pratiques-de-preparation/(offset)/4.
22. Flynn EA, Pearson RE, Barker KN. Observational study of accuracy in compounding i.v. admixtures at five hospitals. Am J Health-System Pharm 1997 Apr 15;54: 904–12. Google Scholar
23. Bateman R, Donyai P. Errors associated with the preparation of aseptic products in UK hospital pharmacies: lessons from the national aseptic error reporting scheme. Qual Saf Health Care 2010 Oct;19:e29. doi:. CrossrefGoogle Scholar
24. Beitz C, Bertsch T, Hannak D, Schrammel W, Einberger C, Wehling M. Compatibility of plastics with cytotoxic drug solutions – comparison of polyethylene with other container materials. Int J Pharm 1999 Aug;185:113–21. CrossrefGoogle Scholar
27. Fourman GL, Mullen MV. Determining cleaning validation acceptance limits for pharmaceutical manufacturing operations. Pharm Technol 1993 Apr;17:54–60. Google Scholar
28. European Medicines Agency. Guideline on setting health based exposure limits for use in risk identification in the manufacture of different medicinal products in shared facilities [Internet]. 2014 [cited 2015 Nov 30] Available from: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2014/11/WC500177735.pdf.
29. European Medicines Agency. Guideline on the limits of genotoxic impurities. [Internet]. 2006 [cited 2015 Nov 30] Available from: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500002903.pdf.
30. PIC/S guide to good practices for the preparation of medicinal products in Healthcare Establishments. Pharmaceutical Inspection convention. Pharmaceutical co-operation scheme [Internet]. 2014 [cited 2015 Nov 30] Available from: http://www.picscheme.org/bo/commun/upload/document/pe-010-4-guide-to-good-practices-for-the-preparation-of-medicinal-products-in-healthcare-establishments-1.pd.
About the article
Raphaël Vazquez is a hospital pharmacist in the sector of drug manufacturing and control at Poissy – Saint Germain en Laye hospital’s pharmacy, France. He completed is PharmD degree at University of Paris Sud (Châtenay-Malabry) with a thesis on Simultaneous quantification of water-soluble and fat-soluble vitamins in parenteral nutrition by HPLC-UV-MS/MS. He also completed his master’s degree in innovative therapies in immunology and a university diploma of hospital pharmaceutical productions. His work resulted in participation at more than 20 conference presentation and publications.
Kevin Boubet is a pharmacist. He also completed his Master degree’s in Quality Control of drugs and medical devices. He was resident in the sector of drug manufacturing and control at Poissy/Saint Germain hospital. He is currently consultant in the field of quality control of health product in pharmaceutical companies. His fields of expertise include flow management in quality control laboratorie especially the substance standard and reagents management.
Marie-Noelle Guerrault-Moro is a hospital pharmacist (pharmD) at Poissy-Saint-Germain-en-Laye hospital, France. Since 2006, she was senior pharmacist involved in the sterile preparations (cytotoxic and parenteral nutrition) and not sterile preparations and she became head of the pharmaceutical technology department in 2014.
Sylvie Crauste-Manciet is currently Professor in Pharmaceutical Technology at Bordeaux University and she is the Head of the pharmaceutical technology department of Bordeaux university hospital (CHU de Bordeaux). She obtained her PharmD degree in 1993, her PhD in 1997 and accreditation to supervise research in 2013 in Paris University. Her academic research field is on design and characterization of nanovesicular systems for therapeutic and imaging. Her hospital research is on sterile compounding, controlled areas such as isolators, on the chemical risks for operators involved in the preparation of cytotoxic drugs and development of pharmaceutical technologies for sterile hospital preparations (i.e. robotics, dose standardization). Since 1998, she is the president of Group for Evaluation and Research on Protection in Areas under Control (GERPAC).
Published Online: 2016-06-02
Published in Print: 2016-09-01
Conflict of interest statement: Authors state no conflict of interest. All authors have read the journal’s publication ethics and publication malpractice statement available at the journal’s website and hereby confirm that they comply with all its parts applicable to the present scientific work.