This site is intended for health professionals only

Formulating IV drugs for neonatal intensive care


William Griffiths
MRPharmS SwissFedDip FPH
Pharmacist R&D
Department of Pharmacy
University Hospitals of Geneva
E:[email protected]
The author would like to thank Dr P Bonnabry PD, Chief Pharmacist, Mr H Ing, Pharmacist, and Dr J Favet, Microbiologist, in the HUG Pharmacy and Dr R Pfister of the NICU for making this study possible

Medical errors and microbiological contamination associated with intravenous (IV) manipulations carried out in clinical areas can lead to potentially life-threatening effects in the neonatal intensive care unit (NICU). For these reasons, the pharmacy of the University Hospitals of Geneva (HUG) decided to develop “almost” ready-to-use (ARTU) formulations of high-risk drugs beginning with vancomycin and midazolam.

Medical errors
More and more reports of medical errors are being made. They can harm the patient and can increase morbidity, length of stay and ultimately hospital costs.

The preparation of drugs in syringes for IV administration is a complex manual process, especially since a final product check is rarely possible. A study carried out in the HUG in an experimental environment showed quite a high error frequency (6.5%) when calculations and preparation of medicines in syringes were performed by anaesthetists.(1) The possible consequences were wrong dose, wrong patient and wrong quantity. Other studies carried out in the field gave similar results.(2,3) A recent study carried out by the School of Pharmacy at the University of London revealed that almost half of all IV doses made up on hospital wards are associated with errors, especially in the case of multistep preparations.(4)

The majority of paediatric prescribing errors occur because most of the drugs require weight-based dosing, which often leads to incorrect calculations and misplaced decimal points.(5) This is particularly relevant in the case of neonates. Mistakes produced during paediatric antibiotic infusion can lead to serious, even fatal dilution errors.(6)

Microbiological contamination
Recent research using a simulation model has shown a high risk of contamination due to standard preparation of syringes for continuous IV administration in intensive care units (ICUs).(7) These risks were drastically reduced when the syringes were prepared by pharmacy technicians working under aseptic conditions. Air sampling during ward-based aseptic preparation showed that the environment is completely uncontrolled, and the failure of a validation test carried out in the ward situation by experienced pharmacy technicians confirmed this.(8)

The Geneva approach
The first mandates obtained from the NICU by the HUG pharmacy were to develop IV preparations of vancomycin and midazolam that would reduce errors, microbiological contamination and reduce nurse time.

The vancomycin example
Vancomycin is the drug of choice in NICUs when severe neonatal infection is suspected after 3–4 days of life.(9) Initial empirical therapy in this rapidly life-threatening condition often includes an aminoglycoside to broaden the spectrum of activity. When sepsis is suspected in the neonate, antibiotic treatment has to be initiated immediately. Commercially available vancomycin targeted primarily for adults needs to be reconditioned for neonatal use. Before IV infusion, 500mg of sterile dry powder are reconstituted to produce a concentration of 100mg/ml. Based on dosage regimens of 10–15mg/kg in the very low birth weight infant (<1500g),(10) further dilutions are necessary to obtain a final strength of 5mg/ml. The material required in the NICU for this multiple dilution procedure is shown in Figure 1a.

  • Stage 1: reconstitution of the lyophilised vancomycin (5 or 10ml syringe).
  • Stage 2: withdrawal of the required quantity into a 1ml syringe.
  • Stage 3: further dilute into a 10ml syringe.
  • Stage 4: attach the syringe to the pump driver.
  • Stage 5: choose the desired volume after purging the line.


These dilution procedures are carried out by clinical staff who may be preoccupied with other aspects of the patient’s care. Medical errors increase in such stressful situations and can lead to dangerous under- or overdosage.(11) Vancomycin has been associated with lasting ototoxicity.(12)

The pharmacy solution
Most of the NICUs in the UK are provided by the pharmacy aseptic unit with syringes containing diluted vancomycin. A dilution of 50mg/ml in 1ml syringes is that most often supplied. A final dilution has then to be carried out in the NICU, which does not fully remove potential problems of error and/or contamination.

In Geneva, it was decided to prepare 10ml syringes of an ARTU solution at a concentration of 5mg/ml, in NaCl 0.9% or glucose 5%. The syringe contents would cover body weights from 500g to 3kg at a dosage of 15mg/kg. The solutions were prepared, filtered (0.2mm) and filled under aseptic conditions into 10ml polypropylene syringes (Plastipak(®), Becton Dickinson)(*) and subsequently stored at 4C and 25C.

*These syringes were obligatory, as they were the
standard model used on HUG pump drivers.

Shelf life
The majority of articles concerning chemical stability gave shelf lives varying from 14 to 94 days in different kinds of packaging including bottles, plastic bags, syringes, elastomeric devices and medication cassettes, and stored at various temperatures.(13–15) Only one study was found using microbiological potency as a stability indicator.(16) Due to the wide variations reported, it was decided to carry out inhouse stability tests by comparing the results obtained from the chemical HPLC assay,(17) with the microbiological potency according to the European Pharmacopoeia 2002. The Swiss Pharmacopoeia 2002 was used for the statistical methods. The results are shown in Figure 2 and can be obtained from the HUG website.(18) The two solutions were found to be stable for five months at 4C. An article on these stability results will be published shortly.


Preparation for delivery
After manufacture, each batch (50 syringes) is placed in quarantine, analysed and a sterility test is carried out. Full manufacturing procedures have been developed to ensure trace back. The Geneva directives state that all vials on the wards must be used for only one dose whatever the volume and then discarded. In the case of these syringes substantial cost savings are thus obtained as only five vials of vancomycin 500mg are used instead of 50, even though there is wastage of solution due to body weight variations. The syringe is shown in Figure 1b and has been used in the NICU for four months. It has only to be fixed to the pump driver after the desired volume has been chosen and after purging the line.

The midazolam example
Midazolam is a short-acting, water soluble benzodiazepine that reduces seizure activity. It is used as firstline sedation in paediatric intensive care units. IV administration can cause respiratory depression and therefore it should be used with caution in the NICU because it can accumulate in neonates due to inadequate clearance by their renal and hepatic systems. The dosages employed are very small – 0.05mg/kg for body weights varying from 500g to 3kg.

The IV midazolam available on the market is dosed at 5mg/ml. In the NICU the following dilution steps are carried out:

  • Stage 1: 1ml of the commercial product is added to 4 ml of NaCl 0.9% or glucose 5%.
  • Stage 2: 1ml of the above diluted to 10ml (giving a concentration of 0.1mg/ml).
  • Stage 3: the dose required for the neonate is determined and placed on the pump driver.

The risk is similar to that encountered with vancomycin.

The pharmacy solution
As in the case of vancomycin, the problem was to obtain a concentration permitting the administration of a convenient volume that wasn’t too large in order to avoid hydric overload in the neonate. Table 1 shows the chosen weight/dose/volume range. This permits a minimal dose of 1ml (0.02mg/ml) for a neonate weighing 400g.


Preparation and shelf life
A solution of midazolam at a concentration of 0.02mg/ml is prepared in either NaCl 0.9% or glucose 5% and filled aseptically to a volume of 7.5ml in a 10ml syringe in the same way as the vancomycin. The solution is stored in light-proof sachets at 4C and 25C. Stability tests are ongoing and the results will be published shortly.

Future perspectives
Other possibilities for pharmacy preparation are under discussion with the NICU staff, including several other antibiotics, heparin, phenobarbital and other high-risk products.

Considering the wide range of doses to be administered to neonates, we cannot really speak of a ready-to-use solution but an “almost” ready-to-use solution. This means that the syringes are prepared containing a fixed concentration of the product and it is the volume chosen by the user that is the determining factor. The syringes are used on a one injection/one patient basis and the surplus is discarded. The wastage is compensated by the “minuscule” quantity of drug used to prepare the dilution and by the fact that batch production is used. While errors can still be produced by injecting the wrong volume, tenfold to hundredfold errors are impossible. It is imperative to use clear and concise labelling without ambiguity and to make sure dosage regimens are well defined for the nursing staff.

The following advantages have been observed in the HUG following the introduction of this kind of IV infusion:

  • No more multiple aseptic manipulation carried out for ARTU solutions in the NICU, thus avoiding bacterial contamination.
  • Fewer calculations, reducing medical errors.
  • No dilution procedures that could lead to dramatic under- and overdoses.
  • A product available for onsite use in emergency situations without delaying the onset of the treatment.
  • Reduction of nurse time being spent on IV duties (available for other tasks).
  • Stability data sufficient for batch production (analytical control and sterility testing can be performed).
  • Cost reductions due to better use of raw materials and to batch production.

The vancomycin study was presented as a poster at the 8th congress of the European Association of Hospital Pharmacists in February 2003 in Florence, Italy.


  1. Meier B, Garnerin Ph, Chopard P Bonnabry P. 70th Assembly of the Swiss Society of Internal Medicine 2001. 25–27 April, Geneva. Available from URL:
  2. Flynn EA, Pearson RE, Barker KN. Am J Health-Syst Pharm 1997;54:904-12.
  3. Ferner RE, Langford NJ, Anton C, et al. Br J Clin Pharmacol 2001;52:573-7.
  4. Taxis K, Barber N. BMJ 2003;326: 684-7.
  5. Neonatal and Paediatric Pharmacists Group. Pharmaceut J 2002;270:60-1.
  6. Leuroueil-Le-Verger M, Dufay E, et al. J Pharm Clin 1995;14(3):195-201.
  7. Van Grafhorst P, Foudraine NA, Nooteboom F, et al. Crit Care Med  2002;30:832-6.
  8. Beaney AM, Goode J. Poster. Guild of Healthcare Pharmacists National conference 2002. Available from URL: News 2002;3:April.
  9. Rubin LG, Sanchez PJ, Siegel J, et al. Pediatrics 2002;110(4):e42.
  10. Tan WH, Brown N, Kelsall AW, McClure RJ. ADC 2002;87:F214-16.
  11. Sexton JB, Thomas EJ, Helmreich RL. BMJ 2000;320:745-9.
  12. Brumett RE. Otolaryngol Clin North Am 1993;26(5):821-8.
  13. Allen Jr LV, Stiles ML, Prince SJ, Smeeding J. Am J Health-Syst Pharm 1996;53:2740-3.
  14. Wood MJ, Lund R, Beavan M. J Clin Pharm Ther 1995;20:319-25.
  15. Allen LV, Stiles L. IJPC 1997;1(2): 123-4.
  16. Mann JM, Coleman DL, Boylan JC. Am J Hosp Pharm 1971; 28:760-4.
  17. Galanti LM, Hecq JD, Vanbeckbergen D, Jamart J. J Clin Pharm Ther 1997;22:353-6
  18. Griffiths W, Favet J, Ing H, et al. Poster at the 8th congress of the European Association of Hospital Pharmacists 2003,26-28 February; Florence, Italy. Available from URL:

Latest Issue

Be in the know
Subscribe to Hospital Pharmacy Europe newsletter and magazine