Geneva University Hospitals (HUG)
2School of Pharmaceutical Sciences
University of Geneva
University of Lausanne
At the Geneva University Hospitals (HUG), individualised parenteral nutrition (PN) are prepared daily using a MM12 MicroMacroR compounder (Baxa UK). An error in the concentration of electrolytes or glucose, due for example to an inversion of ingredients during the connection to the automatic filler, can induce an increased risk for the patient, especially for neonates. The objective of this study was to develop and validate quality control methods to analyse each PN before patient administration. The first method involved capillary electrophoresis with a capacitively coupled contactless conductivity detector (CE-C4D) for the quantification of potassium, sodium, calcium and magnesium in each PN. The second method, based on an enzymatic reaction, allowed the determination of glucose. The developed methods were found appropriate for controlling electrolytes and glucose in PN formulations and they were successfully applied in our daily quality control.
During the past few years, a lot of work has been carried out at HUG to improve the safety and the quality of drug use, from prescription to administration.1–4 Particular attention was focused on drugs aseptically compounded in the hospital pharmacy, considered as a high-risk process, especially when individual formulations are prepared, such as PN. PN are complex mixtures of almost 50 components made from more than 10 different solutions, and a high risk of error and microbiological contamination exists during the production process. As shown by the proactive risk analysis of Bonnabry et al, the use of an electronic prescription and an automatic compounder to prepare PN reduced the criticality of the process.5 This analysis also emphasised that new risks can appear during the re-engineering of a process, such as the inversion of compounds on the automatic filler, leading to a preparation error. Therefore, additional actions, such as quality control of PN before release, help to indicate the non-conforming PN before delivery, whatever the production process (automatic filling or not).
Choice and development of methods
Several points have to be taken into account during the development of quality control of PN before patient administration:
1. How much time has the laboratory to perform PN control?
2. Which critical parameters should be controlled?
3. Which methods are available in the laboratory?
How much time has the laboratory to perform PN control?
The analysis of PN has to be inserted into the process, from prescription to administration. In our hospital, electronic prescriptions are performed by physicians until 1pm and qualified operators produce PN under a laminar airflow hood (GMP Class A/ISO 4.8) in a cleanroom (GMP Class B/ISO 5) until 3pm with an automatic compounder (MM12 MicroMacroR compounder, Baxa UK). At 5pm, the PN should be transported to the wards, to allow their administration to the patient at 6pm. Consequently, the quality control laboratory has a two-hour period to analyse all PN (see Figure 1).
Which critical parameters should be controlled?
The concentration of the main electrolytes (potassium, sodium, calcium and magnesium) and glucose in PN are undoubtedly the most critical parameters. An error in the concentration of these compounds can induce serious clinical problems, especially in neonates. For these reasons, we decided to quantify the main four cations and glucose in order to ensure the adequacy of the production with the prescription.
The control of the sterility of PN also appears important because microbiological contamination of PN during compounding can have dramatic consequences for the patient. However, it cannot be performed in the time available before administration and other quality assurance methods (a posteriori sterility testing, media-fill tests) have to be implemented.
Which methods are available in the laboratory?
Nowadays, no sterility test, recognised by different pharmacopeias, on pharmaceutical fomulations can give a result in two hours. To bring confidence to the system, a full validation of the aseptic filling process has been performed during the implementation process. Moreover, a sterility test, carried out according to the European pharmacopoiea, is performed on ‘control’ PN formulations produced over the weekend, to ensure the maintenance of sterility during the aseptic compounding process.6 These PN contain standard concentrations of glucose, sodium chloride, potassium chloride, calcium chloride and water for injectables and are produced in the same way as PN for patients. Until now, results have shown no contaminated samples.
Quantification of the four main electrolytes
Concerning the chemical analysis of PN, it should be noted that the determination of four electrolytes in PN formulation is an extremely difficult task, given the multi-composition nature of these solutions. Usually, a rapid quantification of inorganic ions in aqueous solution is performed by atomic absorption. However, in the case of PN, an obstruction of the injection system is rapidly observed and a complete wash of this instrument part has to be achieved after a few injections even if diluted solutions of PN are used. Capillary electrophoresis (CE) coupled to a conductivity detection (capacitively coupled contactless conductivity detection, namely C4D), already and currently used in our laboratory in routine analysis, appeared to be an attractive strategy to perform the rapid analysis of inorganic ions in PN. Indeed, CE is particularly suitable for analysing small inorganic ions because the separation mechanism depends on the charge and the size of compounds. Moreover, the main advantages of CE are a high efficiency, low organic solvent consumption, low cost of capillaries, rapid method development and high versatility. The CE-C4D method developed for the rapid analysis of sodium, potassium, calcium and magnesium in PN is described elsewhere.7 In summary, the CE-C4D method is considered as:
- Rapid, with a complete separation of the four ions in less than four minutes. A typical electropherogram obtained for the analysis of a PN is reported in Figure 2. This short analysis time is compatible with the two-hour period available to achieve the analysis of c.a. 10–15 produced PN bags
- Easy to operate by the laboratory staff. Indeed, with the CE-C4D method, a diluted sample of PN is directly injected in the CE instrumentation
- Reliable, as demonstrated by the complete validation of the developed method. Accuracy values between 99.7 and 101.9% were obtained with repeatability and intermediate precision values less than 2% for all cations.
Nowadays, more than 700 individualised PNs, produced by the HUG pharmacy, have been analysed with the CE-C4D method. Only one PN was considered as non-conforming because the potassium level was more than 110% (120%) of the prescribed concentration. This formulation was destroyed and a new one was produced and released after analysis. A calibration of the compounding system corrected this problem.
Quantification of glucose
The hexokinase method developed for the analysis of glucose in human blood is used for the determination of glucose in PN. This approach is based on the formation of NADPH proportional to the glucose concentration and can be measured photometrically at 340nm. A complete description of the reaction is reported in the guidelines of the reagent Gluco-quant Glucose/HK commercialised by Roche.8 Like the CE-C4D method, the hexokinase method is:
- Rapid: the analysis of glucose in PN is performed in less than 10 minutes (including a reaction time of five minutes)
- Easy to operate: 100µL diluted sample of PN was added to 900µL of reagent. The solution is mixed and analysed by spectrophotometry after 5 minutes
- Reliable: as demonstrated by the validation of the hexokinase method. Accuracy values between 99.4 and 102.1% were obtained with repeatability and intermediate precision values less than 2%.
A single non-conformity with glucose concentration has been revealed on 700 produced PN and human error was detected. After investigation, it appeared that the empty bottle of glucose G70% was substituted by an identical bottle of aminoacids during the production process. The PN with the wrong glucose concentration was destroyed, and new PN was produced and released after analysis.
Data and non-conformity handling
All results obtained for the analysis of PN (cations and glucose determination and sterility test) are reported in a ‘home-made computerised system’ designed in collaboration with the medical informatics department of HUG. Production and analysis protocols for all prepared PN can be continually accessed by the pharmacy personnel. Thus, the traceability and the safety of PN flow are improved, as described in previously published work.5
Non-conforming PN are eliminated and new ones are produced and released after analysis. In this case, the physicians are informed about the delay and appropriate actions can be taken.
Quality control of PN formulations has been implemented at the HUG pharmacy. This included the quantification of sodium, potassium, calcium, magnesium and glucose for each PN before patient administration. The microbiological quality of the production process is checked by a complementary sterility test. Two simple and rapid analytical methods have been developed and validated. Both methods have been successfully applied in routine analysis for one year at the laboratory of the HUG pharmacy and permitted the detection of non-conforming PN. The implementation of a quality control of PN, in addition to an electronic prescription and an automatic compounding system, has contributed to improvements in the safety of the PN process in the hospital.
1. Garnerin P, et al. Eur J Clin Pharmacol 2007;63:769–76.
2. Spahni S, et al. Stud Health Technol Inform 2007;129:853–7.
3. De Giorgi I, et al. Int J Qual Health Care 2010;22:170–8.
4. Bonnabry P, et al. Hospital Pharmacy Europe 2009;47:67–8.
5. Bonnabry P, et al. Qual Saf Health Care 2005;14:93–8.
6. Ing H, et al. Le Pharmacien Hospitalier 2003;38:155–60.
7. Nussbaumer S, et al. J Pharm Biomed Anal 2010;53;130–6.
8. Gluco-quant glucose/HK. Roche Diagnostics GmbH: 68298 Mannheim, Germany;2008.