Article Text

Download PDFPDF

Accuracy of the concentration of morphine infusions prepared for patients in a neonatal intensive care unit
  1. Virginia Aguado-Lorenzo1,
  2. Kevin Weeks2,
  3. Paul Tunstell3,
  4. Karen Turnock4,
  5. Timothy Watts4,
  6. Sara Arenas-Lopez5
  1. 1Department of Pharmacy, Guy's and St. Thomas’ NHS Foundation Trust, London, UK
  2. 2Pharmacy QC Department, Guy's and St. Thomas’ NHS Foundation Trust, London, UK
  3. 3Pharmacy Aseptic Services, Guy's and St. Thomas’ NHS Foundation Trust, London, UK
  4. 4Neonatal intensive care Unit, Evelina Children's Hospital, Guy's and St. Thomas’ NHS Foundation Trust, London, UK
  5. 5Pharmacy Evelina Children's Hospital, Guy's and St. Thomas’ NHS Foundation Trust, London, UK
  1. Correspondence to Dr Virginia Aguado-Lorenzo, Pharmacy, St Thomas' Hospital, Westminster Bridge Road, London SE1 7EH, UK; Virginia.Aguado{at}


Objective To investigate the accuracy of morphine infusions prepared for neonates in relation to the label strength and to identify the differences in deviation between infusions made in neonatal intensive care unit (NICU) and those dispensed ready-to-use from pharmacy.

Methods Unused portions of morphine solution for infusion were collected over a 6-weeks period and used to determine the concentration of the drug by high-performance liquid chromatography (HPLC).

Results A total of 19.2% of infusions prepared by nurses in the ward and 7.8% prepared in the pharmacy were outside the limit required by the British Pharmacopoeia (±7.5%). Moreover, a deviation in concentration of more than 20% was found in ward-prepared infusions, although this was caused by volume discrepancies of less than 0.2 mL. The frequency and magnitude of deviations found in infusions prepared in pharmacy was lower than in those prepared by NICU. The latter showed significantly higher number of out-of-specification samples (p=0.015); however, deviations from intended concentration occurred in both settings. Significant differences between pharmacy and NICU for volumes of less than 0.5 mL or for less than 1 mL were not identified probably due to small sample size, but statistical data show a trend for differences.

Conclusions Current practice of preparation of infusions from strengths intended for older children and adults involves dilution of small volumes in a syringe and leads to inaccuracy in the final concentration of infusions for neonatal use. We propose the implementation of standard concentrations for this patient group to effectively eliminate these errors.

  • accuracy
  • neonates
  • standard concentrations
  • morphine
  • infusion
View Full Text

Statistics from

What is already known on this topic

  • Errors occur during preparation of intravenous drugs in paediatric and neonatal clinical areas.

  • The intravenous products available are frequently inadequate for administration of doses required in children.

  • The use of standard concentrations has been recommended to reduce risk.

What this study adds

  • Incorrect concentrations occur both in ward and pharmacy prepared infusions.

  • The use of small drug volumes to prepare infusions correlates to an increased frequency and magnitude of deviation from the intended concentrations.


Errors with intravenous medications are common, and drug preparation has been identified as one of the steps of the process where errors are more frequent.1 Multiple-step preparation processes, in particular, have been associated with an increased occurrence of error.2

The preparation of morphine infusion for neonates is a complex procedure that involves using an open system during the dilution process, which requires syringe-to-syringe transfer of the drug. Morphine sulfate continuous infusion is prepared on individualised doses, both in pharmacy and at ward level, prescribed using a calculation tool in an attempt to reduce errors during preparation and administration.3–5 The concentration of morphine sulfate infusion prepared using this method varies from patient to patient as it is determined by weight. Routine practice in neonatal units is to withdraw the required amount of morphine sulfate to a syringe of the appropriate size and then transfer to a second syringe to dilute to a larger volume of 50 mL with glucose 5%.6

Morphine infusions were chosen for this project because it is a high-risk drug widely prescribed for neonates as first-line analgesic, doses frequently require use of small volumes and there is strict controlled documentation of this drug. Previous studies have reported that two-thirds of morphine infusions prepared in clinical areas were outside pharmaceutical standards.7

The preparation of intravenous therapy by a pharmacy-run centralised intravenous admixture service (CIVAS) is intended to reduce errors and microbiological risk of preparation of injectable drugs in clinical areas. However, a significant number of the treatments continue to be prepared by nursing staff on the wards8 ,9 mainly due to limited capacity of pharmacy services, leading to impractical turnaround time for first doses and lack of an out-of-hours service.

The objective of this study was to investigate the accuracy of morphine infusions prepared for neonates in relation to the label strength and to identify the differences in deviation between infusions made in neonatal intensive care unit (NICU) and those dispensed ready-to-use from pharmacy.


Morphine syringes containing the unused portion of morphine solution for infusion administered to patients in NICU (typically neonate with a weight range between 0.5 and 4 kg) were collected over a 6-weeks period and used to determine the concentration of the drug. Hospital policy is to run intravenous infusions for a maximum of 24 h. Nurses were asked to retrieve syringes containing unused portion of solution from all morphine infusions administered in NICU during the study period. The syringes were stored in a refrigerator prior to analysis for a maximum of 10 days. Little or no degradation would be expected as morphine sulfate has been shown to be stable, less than 3% loss over 4 months stored in syringes in previous studies performed in-house (unpublished work).

The concentration of morphine in the syringes was determined by high-performance liquid chromatography (HPLC) using Hewlett Packard HP 1100/1200 HPLC System attached to a HP Computer system with Hypersil ODS 5 μm 100 mm×4.5 mm id+20 mm guard column (particle size 5 μm) at 20°C. Mobile phase was 0.005 M dioctyl sodium sulphosuccinate+0.01 M sodium acetate (pH=5) and methanol (40 : 60); flow rate 1.5 mL/min and detection by ultraviolet at 285 nm. The limit of detection for the method was determined to be 14.4 ng/μL, and the limit of quantitation for the method was determined to be 48.0 ng/μL.

Morphine sulfate B.P. was used to prepare the solution standard for the analysis. The method has been validated for stability testing by means of forced degradation using acid alkali, heat, light and oxidation (unpublished work). External standards were used for all assays. Linearity was shown over the range of concentrations injected as part of method validation performed in-house (r2=0.9994). Triplicate readings were obtained for each of the samples analysed, and the average was used as the measured concentration of morphine sulfate.

Data analysis

Using as reference the British Pharmacopoeia (BP) concentration limits for morphine sulfate injection, a maximum deviation of ±7.5% of the concentration in product label was considered acceptable.10 For interpretation of results in this study, we consider that a deviation has occurred when an infusion fails BP quality standards (out of specification).

The clinical significance of the deviations identified on the concentration of the infusions was not the purpose of the study, and, therefore, the results obtained were not linked to individual patients or staff.

Statistical analysis

A Fisher's test was carried out to determine if there was a significant difference between errors found in infusions prepared on the ward and CIVAS.


A total of 214 samples of morphine infusions were collected. They were prepared either in pharmacy CIVAS (n=115, 54%) or by nurses in the ward (n=99, 46%). Morphine 10 mg in 1 mL was the starting material used in all cases to prepare the infusions, as confirmed by Controlled Drug (CD) registers.

The concentrations recorded of morphine infusions prepared for NICU patients during the study period ranged from 0.5 mg to 52 mg in 50 mL, median concentration was 8.65 mg in 50 mL (interquantile range 4–17 mg in 50 mL).

Theoretical volumes of morphine sulfate 10 mg/mL used during preparation of the infusions were as small as 0.05 mL. In 60% of the infusions, withdrawal of volumes of less than or equal to 1 mL of morphine sulfate was required, being less than or equal to 0.5 mL in 37% of the cases.

Accuracy of morphine sulfate concentration

The concentration of 19.2% of infusions prepared by nurses on the ward and 7.8% of infusions prepared in the pharmacy were outside the ±7.5% BP limit. The difference between preparation in both settings was significant (p=0.015, OR 2.79), indicating that the probability of out of specification of ward infusions is nearly three times higher.

The rate of errors in relation to volumes of morphine sulfate withdrawn to prepare the solutions is shown in table 1.

Table 1

Distribution of errors with respect to theoretical volume of morphine withdrawn

Infusions prepared using volumes of the starting material less than or equal to 1 mL of morphine accounted for 93% (26/28) of results out of specification and 71% (20/28) when using volumes of less than 0.5 mL.

No significant differences were found on frequency of out-of-specification results in CIVAS and NICU for less than 1 mL and 0.5 mL (p=0.184 and 0.301, respectively). However, a trend was observed indicating that out-of-specification infusions are more likely to occur on the ward for subgroups (ORs 1.497 and 1.921, respectively).

Concentration accuracy in relation to strength in label

The deviation of morphine volumes withdrawn to prepare the infusions was calculated from the difference between measured concentration of the solution in syringe and concentration in label.

Table 2 shows the calculated volume deviations for each of the 28 samples found to be outside BP limits. The calculated deviation was 0.01–0.08 mL for infusions prepared by CIVAS and up to 0.19 mL for infusions prepared on the ward. Furthermore, a deviation of more than 20% in relation to the strength on the label was found in 3/99 infusions prepared on the ward, including one infusion in which morphine concentration was 66.5% more than expected.

Table 2

Morphine infusion samples with a concentration outside BP limits

A similar volume deviation was observed in other infusions that were compliant with BP limit (figure 1). All the three infusions showing a volume deviation of 0.1 mL or greater were prepared on the ward.

Figure 1

Relationship between concentration deviations observed and volume of morphine withdrawn. *Outlier result 66.5% deviation not shown in chart (volume of morphine 0.24 mL). BP, British Pharmacopoeia.


A difference in accuracy of syringes in relation to strength in label, outside the BP limits, was found in syringes prepared for the neonatal unit. The frequency of error found in syringes made on the ward was significantly higher than in pharmacy (p=0.015), but deviations were found in both settings.

Variation of the magnitudes described may result in morphine delivery that is significantly higher or lower than that prescribed. It is difficult to predict the clinical effects on neonates following these deviations as this population presents an additional challenge with the pharmacokinetic and pharmacodynamic morphine profile compared with term neonates and older children.

These neonates are more vulnerable to the effects of morphine due to a decreased glucuronidation capacity and excretion. Therefore, increases of more than 10% higher concentrations than intended could potentially put the baby at increased risk of side effects such as respiratory depression.11

However, the main problem is that individualised syringes made by weight either by nurses or in pharmacy cannot reliably deliver the intended prescribed concentrations (in up to 20% of the cases in our study). This makes dose adjustments and interpretation of clinical effects challenging as they rely on the label concentration and may not correlate the observed effect due to the deviation in concentration. The variation between syringes prepared daily could compromise clinical response in neonates.

There are no international guidelines as to safe rounding of doses in paediatrics; therefore, having the correct concentration according to pharmacopoeial standards as a starting point is paramount.

Previous authors have reported incidence of error higher than those observed in this study, even when limits are based on US Pharmacopeia (±10%),7 ,12 or preparation was taken place in non-clinical settings for the study.13 In this study, calculation errors or wrong volume measurements may have been prevented by independent double check by two registered nurses during preparation and administration of the morphine infusions, documented on CD records. The standard procedures for selection of materials and manipulations, documented supervision of each preparation step and automated validated systems in CIVAS contribute to eliminate certain sources of error, thereby minimising inaccuracies during preparation. Wheeler et al12 reported fourfold to fivefold errors attributable to calculation mistakes. The magnitude of the deviations observed in this study, with only three volume errors over 0.1 mL and none over 0.2 mL, suggests those were due to inaccuracies during preparation of the infusion rather than due to calculations or volume measurement errors.

There is evidence that most errors occur during the preparation process and are associated to multiple steps and the use of small volumes.14 ,15 Although in our study most volumes measured were greater than 0.1 mL compared with previous studies,15 small deviations in a volume measured up to 0.5 mL or even 1 mL caused a significant concentration error.

Our results showed that even when preparation is carried out by experienced supervised staff there seems to be a residual number of infusions that still lay outside pharmacopoeial limits. Deviations in volume were less than 0.1 mL in 86% out-of-specification morphine infusions, and they were as low as 0.01 mL in some of them. Statistical analysis showed that these were twice as likely to occur in the ward as in CIVAS for volumes up to 0.5 mL and 1.5 times more likely for volumes up to 1 mL.

Discrepancies in volume could be attributable to accumulated factors: syringe accuracy,16 volume in syringe dead space (0.07 mL in BD 1 mL and 3 mL syringes used)17 ,18 or mixing of the final product as the content of drug may not be uniformly distributed in the syringe.13 In addition, the choice of syringe size can also have an impact,19 with smallest syringes providing more accurate and reproducible results.20 ,21

These findings can also be applied to other infusions such as inotropes, protaglandines, etc., using a similar preparation method and a range of volumes in this patient group.22 We propose the implementation of batch manufactured standard concentrations of morphine infusions. This would guarantee compliance with pharmacopoeia limits and reduce microbiological risks.23

The use of standardised concentrations is recommended on numerous safety alerts24–26 by the US Government in 200827 and has demonstrated reduced administration errors.28 ,29 However, in order to implement this change in paediatric practice safely, careful considerations such as agreement on clinically appropriate concentrations and diluents, introduction of a change in prescribing and administration practice, labelling of different strengths to avoid mis-selection and safety of storage are required. The main risk envisaged with the implementation of this new system is the mis-selection of the prefilled syringes. Therefore, the use of barcoding and smart pump technology need to be considered for a successful implementation of standardised concentrations.30


Current practice of preparation of infusions from strengths intended for older children and adults involves dilution of small volumes in a syringe and leads to inaccuracy of final concentration of infusions in neonatal use. We propose the implementation of standard concentrations for this patient group to effectively eliminate these errors.


The authors would like to thank Saloni Chandaria, School of Pharmacy (University of London), and the Pharmacy Quality Control staff at Guy's and St. Thomas’ Foundation Trust for their assistance with the analysis of the samples.


View Abstract


  • Contributors VA-L and SA-L were the lead authors, providing design of the study, supervision data collection, analysis and interpretation of the data and wrote manuscript. KW and PT contributed with data collection and analysis of samples and review of the manuscript. KT and TW were responsible for the interpretation of the data from the clinical perspective and review of manuscript.

  • Competing interests None.

  • Ethics approval This study was registered as an audit following trust procedures. No patient data were retrieved from any sample, and therefore no ethics permission was required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.