An assessment of the variation in the concentration of acetylcysteine in infusions for the treatment of paracetamol overdose

Abstract

Background

Intravenous acetylcysteine is the treatment of choice for paracetamol poisoning. A previous UK study in 2001 found that 39% of measured acetylcysteine infusion concentrations differed by >20% from anticipated concentrations. In 2012, the UK Commission on Human Medicines made recommendations for the management of paracetamol overdose, including provision of weight‐based acetylcysteine dosing tables. The aim of this study was to assess variation in acetylcysteine concentrations in administered infusions following the introduction of this guidance.

Methods

A 6‐month single‐centre prospective study was undertaken at a UK teaching hospital. After preparation, 5‐ml samples were taken from the first, second and third/any subsequent acetylcysteine infusions. Acetylcysteine was measured in diluted (1:50) samples by high‐performance liquid chromatography. Comparisons between measured and expected concentrations based on prescribed weight‐based dose and volume were made for each infusion.

Results

Ninety samples were collected. There was a variation of ≤10% in measured compared to expected concentration for 45 (50%) infusions, of 10–20% for 27 (30%) infusions, 20.1–50% for 14 (16%) infusions and >50% for four (4%) infusions. There was a median (interquartile range) variation in measured compared to expected concentration of −3.6 mg ml⁻¹ (−6.7 to −2.3) for the first infusion, +0.2 mg ml⁻¹ (−0.9 to +0.4) for the second infusion and −0.3 mg ml⁻¹ (−0.6 to +0.2) for third and fourth infusions.

Conclusion

There has been a moderate improvement in the variation in acetylcysteine dose administered by infusion. Further work is required to understand the continuing variation and consideration should be given to simplification of acetylcysteine regimes to decrease the risk of administration errors.

What is Already Known about this Subject

• A previous UK study in 2001 found large variations in measured acetylcysteine concentrations in intravenous infusions for the treatment of paracetamol overdose compared to anticipated concentrations.

• In 2012 the UK Commission on Human Medicines made a number of recommendations for the management of paracetamol overdose, including provision of weight‐based acetylcysteine dosing tables.

What this Study Adds

• There appears to have been a moderate improvement in the variation in acetylcysteine dose administered by infusion since the introduction of weight‐based dosing charts.

• Significant variation in infusion concentrations is still present and has the potential to be associated with a risk of either treatment failure, or an increase in adverse effects.

Table of Links

LIGANDS
Paracetamol

This Table lists key ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 1.

Introduction

Paracetamol is the most common substance used by patients presenting with deliberate self‐poisoning in the UK, accounting for between 42–44% of cases 2, 3, 4. In a 2014 UK multicentre study, half of the patients presenting to Emergency Departments following paracetamol overdose required treatment with acetylcysteine 5. In the UK, paracetamol overdose is treated with a regime of three infusions of intravenous acetylcysteine over 21 hours. Each infusion has a different acetylcysteine dose in a different volume of infusion fluid. The dose is calculated for each patient based on body weight and there is the potential for error at many stages of the process including prescription, drug preparation and administration.

Ferner et al. compared measured concentrations of acetylcysteine in infusion bags to the expected concentration and found large variations 6. These included 39% of measured infusion concentrations differing by >20% from the anticipated concentration 6. A US retrospective chart review of intravenous acetylcysteine use found medication errors for 33% of patients including 1.4% being prescribed an incorrect dose and an incorrect infusion rate in 5% of cases 7. In a Malaysian audit of intravenous acetylcysteine treatments there were prescription errors for 5.9% of patients, incorrect infusion rates for 37.3% and treatment interruptions for 8.5% 8.

There are reports of deaths caused by acetylcysteine dosing errors including a 53‐year‐old man who died from a myocardial infarction following administration of 10 times the recommended dose 9 and a 2‐year‐old child dying after developing status epilepticus following a prescribing error 10. A 23‐year‐old patient died from haemolytic uraemic syndrome 12 days after being administered 10 times the intended loading dose of acetylcysteine and developing haemolysis 11. A young child was given a two‐fold overdose of acetylcysteine and died despite the treatment not having been required 12. There are also reports of a patient developing intractable seizures and cerebral oedema with resulting permanent neurological disability 13, another developing haemolysis and haemolytic uraemic syndrome after a five‐fold overdose 14 and a 10‐fold overdose with no adverse consequences related to incorrect entry of the patient's weight on a computer prescribing system 15. Two cases of 10‐fold under‐dosing errors are reported: both patients developed significant hepatotoxicity but made a full recovery 16.

In 2012, the UK Commission on Human Medicines (CHM) made a number of recommendations for the management of paracetamol overdose with the aim of reducing the risk of acetylcysteine toxicity and inter alia of reducing the risk of acetylcysteine administration errors. This included changing the time over which the first acetylcysteine infusion is given from 15 min to 1 h, the provision of weight‐based dosing tables for acetylcysteine for both adults and children, and including a technical information leaflet regarding preparation of the infusions in every pack of acetylcysteine 17. Other changes were also made to the assessment of patients with paracetamol overdose that have resulted in a lowering of the threshold for treatment and a consequent increase in the number of patients being given acetylcysteine 5.

Selvan et al. assessed prescribing and volume calculations by doctors and nurses with the introduction of a weight‐based dosing chart for acetylcysteine 18. Using manual calculation methods, errors were made in 26% of cases, but no errors were made using the dosing chart. In an Australian pre‐ and post‐intervention study, McIntyre et al. found that introducing a weight‐based dosing chart reduced the overall rate of acetylcysteine prescription errors but did not significantly reduce major prescription error rates 19. Their chart included the acetylcysteine dose in milligrams but not the ampoule volume.

The aim of this study was to assess variation in acetylcysteine concentrations in administered infusions following the introduction of the new guidance from the CHM.

Methods

A 6‐month single‐centre prospective study was undertaken at a London teaching hospital with a clinical toxicology service from 1^st October 2014 to 31^st March 2015. Patients aged 18 years and over, receiving treatment with intravenous acetylcysteine for paracetamol overdose were included. Patients weighing <40 kg were excluded.

Acetylcysteine infusions made up with 5% glucose or 0.9% sodium chloride solution (supplied by Baxter–Viaflo) were prepared by nursing staff using the appropriate UK weight‐based acetylcysteine dosing tables. The dosing tables are included as part of the prescription charts at the study hospital (see Table 1). No changes were made to routine clinical practice.

Table 1.

Table within the acetylcysteine infusion prescription chart, showing the weight‐based dose bands and infusion volumes

Prescriber to initial next to patient weigh selected	Regimen of infusion	First infusion		Second infusion		Third infusion
	Infusion fluid	200 ml glucose 5% or sodium chloride 0.9%		500 ml glucose 5% or sodium chloride 0.9%		1000 ml glucose 5% or sodium chloride 0.9%
	Duration	1 hour		4 hours		16 hours
	Drug dose	150 mg kg −1 acetylcysteine		50 mg kg −1 acetylcysteine		100 mg kg −1 acetylcysteine
	Patient weight	Ampoule volume	Infusion rate	Ampoule volume	Infusion rate	Ampoule volume	Infusion rate
	(kg)	(ml)	(ml h −1 )	(ml)	(ml h −1 )	(ml)	(ml h −1 )
	40–49	34	234	12	128	23	64
	50–59	42	242	14	129	28	64
	60–69	49	249	17	129	33	65
	70–79	57	257	19	130	38	65
	80–89	64	264	22	131	43	65
	90–99	72	272	24	131	48	66
	100–109	79	279	27	132	53	66
	≥110	83	283	28	132	55	66
Administration signatures
Date and time		Start time	Stop time	Start time	Stop time	Start time	Stop time

After preparation, infusion bags were hand‐mixed and a 5‐ml sample taken from the injection port of the first (150 mg kg⁻¹ over 1 h), second (50 mg kg⁻¹ over 4 hours) and third and any subsequent (100 mg kg⁻¹ over 16 h) acetylcysteine infusions using separate syringes. During working hours (Monday to Friday 09:00–17:00 h) samples were taken by a member of the clinical toxicology team. Out of hours samples were taken by nursing staff and stored at 4°C in the Emergency Department or ward refrigerator until the next working day. The infusion number and weight of the patient were recorded. Samples were stored at −20°C in plastic screw‐top universal containers until being transferred in a single batch for analysis.

No patient information other than body weight was recorded. The study was approved by the Trust's audit committee. Patient consent was not required because the sampling did not significantly alter care.

Acetylcysteine was measured by high‐performance liquid chromatography with UV detection (215 nm) 20. Samples (100 μl) were diluted in duplicate 1 + 49 with 0.02 mol l⁻¹ orthophosphoric acid solution (4.9 ml) before being mixed with internal standard solution (0.4 mg ml⁻¹ tyrosine in dilute orthophosphoric acid, 100 μl). Portions (20 μl) of the resulting solutions were analysed (column: Hypersil Gold, 5 μm aps, 150 × 4.6 mm i.d., ThermoFisher Scientific, eluent: 0.05% m/v aqueous sodium perchlorate, pH 3.0, eluent flow‐rate: 1.0 ml min⁻¹, column oven temperature: 25°C). Acetylcysteine concentrations were calculated by comparison to the results obtained by analysis of standard acetylcysteine solutions (calibration range: 0.125–2 mg ml⁻¹) prepared in 5% (w/v) aqueous D‐glucose. This calibration was verified by analysis of the standard solutions at an independent laboratory. The initial sample analysis was followed by re‐analysis after a minimum of 13 days storage at 4°C. Mean (standard deviation) results (mg ml⁻¹) were: (infusion 1) 36.15 (7.56) and 35.27 (7.22), r ² = 0.91; (infusion 2) 7.54 (5.32) and 7.45 (4.60), r ² = 0.99; and (infusion 3) 6.91 (6.19) and 7.00 (6.30), r ² = 1.00. The mean results of the separate analyses were reported. Solutions containing acetylcysteine in 5% (w/v) aqueous D‐glucose were analysed to assess intra‐assay precision and accuracy. %CV was ≤6.1% and bias ≤5.1% for all solutions across the measuring range of the assay.

Comparisons between measured and expected concentrations based on prescribed weight‐based dose and volume were made for each infusion. The results were analysed to assess the degree of variation between the measured and expected concentrations. The variations in measured compared to expected concentrations were expressed as a percentage.

Expected acetylcysteine concentrations in the infusion bags were calculated accounting for overfill of infusion bags using the manufacturer's average target fill volumes of 271 ml for the 250 ml bag, 530 ml for the 500 ml bag and 1047 ml for the 1000 ml bag. These were confirmed with the measurement of volumes of five randomly selected bags of each volume. The variation between the measured and average target overfill volumes ranged from 0% to 1.8%. For the first infusion 50 ml fluid is removed from the bag prior to addition of acetylcysteine, for subsequent infusions nothing is removed prior to addition of acetylcysteine 17. The different volumes of fluid in acetylcysteine ampoules added to infusions based on the weight bands were also accounted for in calculating the expected concentration.

Results

There were 90 samples (39 first, 24 second, and 27 third or fourth acetylcysteine infusion samples) collected from infusions for 54 patients. The variation in measured as compared to expected acetylcysteine concentrations is shown in Figure 1. There was a variation of ≤10% in measured compared to expected concentration for 45 (50%) infusions, of 10–20% for 27 (30%) infusions, 20.1–50% for 14 (16%) infusions and >50% for four (4%) infusions.

Figure 1.

The percentage variation in measured concentrations of acetylcysteine compared to expected concentrations based on the prescribed dose and volume of infusion bags

Measured concentrations were higher than expected for 23 (26%) samples and lower than expected for 67 (74%) samples. There were two samples that varied by more than +100% from the expected concentration including one second‐infusion sample that varied by +333% and one third‐infusion sample that varied by +334% from the expected concentration. The sample containing the lowest percentage of measured acetylcysteine concentration was a second infusion sample that contained −38% of the expected concentration.

Comparing the different volume infusions, there was a variation of ≤10% in measured compared to expected concentration for 18/39 (46%) first‐infusion samples, 11/24 (46%) second‐infusion samples and 16/27 (59%) third‐ or fourth‐infusion samples. There was a median (interquartile range [IQR]) variation in measured compared to expected concentration in mg ml⁻¹ of −3.6 mg ml⁻¹ (−6.7 to −2.3 mg ml⁻¹) for the first infusion, +0.2 mg ml⁻¹ (−0.9 to +0.4 mg ml⁻¹) for the second infusion and −0.3 mg ml⁻¹ (−0.6 to +0.2 mg ml⁻¹) for third and fourth infusions.

The total dose (mg kg⁻¹) of acetylcysteine administered in each infusion sampled is shown in Figure 2. This was calculated based on the laboratory measured concentrations and dose weight band and compared to the expected doses of acetylcysteine in infusions. The median (IQR) total dose was 135 mg kg⁻¹ (125–141 mg kg⁻¹) for the first infusion (prescribed dose = 150 mg kg⁻¹), 48 mg kg⁻¹ (43–52 mg kg⁻¹) for the second infusion (prescribed dose 50 mg kg⁻¹) and 94 mg kg⁻¹ (90–103 mg kg⁻¹) for third and fourth infusions (prescribed dose 100 mg kg⁻¹).

Figure 2.

Acetylcysteine dose (mg kg⁻¹ h⁻¹) administered per hour in infusions based on laboratory measured concentration and weight‐based dose band compared to the prescribed (expected) doses. Prescribed doses marked by black lines

The relationship between patient weight and percentage variation in acetylcysteine concentration was analysed using Spearman's rank correlation coefficient: ρ = −0.0268 (P = 0.79).

Discussion

There was less variation in our study compared to that reported by Ferner et al. 6, who studied variation in acetylcysteine concentration using the previous UK regime prior to the use of weight‐based tables. They found that 37% of the 184 samples analysed contained acetylcysteine within 10% of the anticipated dose and 61% within 20% of the anticipated dose 6.

We considered a variation in measured compared to expected concentration of up to 10% as an acceptable level; this was the case for 50% of samples in our study. This suggests that although there has been a modest improvement in the accuracy of preparation of acetylcysteine infusions since the introduction of the new guidance from the CHM together with weight‐based dosing tables, a significant proportion of infusion bags still have variation that may be of potential clinical significance.

The dosing tables are included on the prescription charts in our Trust with the intention of simplifying the calculations required for preparation of acetylcysteine infusions. The charts clearly provide the volume of acetylcysteine (based on patient weight) that is required to be added to each infusion bag. Prior to the 2012 introduction of weight‐based dosing tables, acetylcysteine infusions would be written on drug charts and would include the number of milligrams of acetylcysteine but would not routinely state the ampoule volume. The effect of removing 5 ml from each infusion bag on delivered doses was much lower than the potential differences caused by variations between actual patient weight and the weight‐based dose category calculated doses.

In our study there was a tendency for samples to contain lower than expected concentrations; this could be associated with treatment failure and risk paracetamol‐related hepatotoxicity. However, markers of hepatotoxicity were not assessed in this study and the impact of infusions containing lower acetylcysteine concentrations than prescribed is not certain. The reason for a higher proportion of patients being relatively under‐dosed was not clear. For the first infusion, it might be explained in some cases if 50 ml was not removed (removal of 50 ml is recommended) from infusion bags prior to addition of acetylcysteine but this is not relevant to the second and subsequent infusions.

Some patients received higher than prescribed doses. This risks patients being exposed to more frequent or more severe adverse drug reactions. The occurrence of anaphylactoid reactions mostly coincides with peak acetylcysteine concentrations 21 and higher rates of anaphylactoid reactions (11/15 patients) were reported for a series of patients with acetylcysteine overdoses given in error 22. However, the adverse effects are not related to acetylcysteine concentrations alone 23. Lower paracetamol concentrations are associated with a higher incidence of anaphylactoid reactions to acetylcysteine 24. In vitro, paracetamol has been found to modify histamine secretion 25 and, in a study of 22 patients, higher plasma histamine concentrations were correlated with more severe adverse reactions to acetylcysteine 23.

Errors resulting in higher doses for the first and second infusions have greater implications regarding adverse drug reactions because of the higher doses administered per hour compared to the third and subsequent infusions. Correct dosing for the first‐infusion is more important regarding adverse effects and likely antidotal effect. There were no major outliers in the first‐infusion group in this study.

Infusion preparation requires use of the correct volume infusion bag and insertion of the correct volume of acetylcysteine. Medication errors are common. A UK disguised observational study found that at least one drug error occurred for 212/430 (49%) intravenous drug doses, there were preparation errors for 7% of doses and administration errors for 36% of doses 26. A Canadian observational study on preparation of morphine infusions in nonclinical practice found concentrations errors of ≥10% for 35% of infusions 27. This rate of error was similar to that found in our study for acetylcysteine.

Selvan and colleagues demonstrated that the introduction of a weight‐based dosing chart for acetylcysteine reduced calculation errors 18. However, this was a written calculation exercise only and was not in a clinical setting and did not involve infusion preparation. Our results suggest that simplification of prescription charts may have resulted in a moderate reduction in the variability of acetylcysteine concentrations in infusions.

There is the potential that inadequate mixing of infusions prior to sampling or that concentrations not being homogenous throughout the infusion bags could in part have contributed to the variations found. All of our samples were taken after preparation and mixing but prior to administration. Ferner and colleagues measured 69.6% of their samples both before and after infusion and found a median difference of 0% (IQR −5.2 to +14.6%). This level of discrepancy between pre‐ and postinfusion samples would not fully account for the variation found in both studies but it may have influenced the results.

Further work is required to ascertain the impact of mixing and homogeneity of the infusions. If variable concentrations of acetylcysteine are being delivered throughout infusions this may influence the efficacy of the treatment (even if the overall delivered dose is correct) or influence the rate of adverse effects. This is particularly important in view of the high rate of adverse drug reactions reported for intravenous acetylcysteine (rates of 12–77%) 4, 23, 28.

The expected concentrations were calculated using the average manufacturer target volumes for infusion bags. Variation in the volumes of the infusion bags could have affected the results. However, the magnitude of this effect would not account for the discrepancies found. There was a weak negative correlation between patient weight and percentage variation in acetylcysteine concentration but this was not statistically significant.

A reduced incidence and severity of adverse effects have been demonstrated with modified two infusion acetylcysteine regimens in both the UK and Australia 29, 30. However, these still involve two different infusions, with different acetylcysteine doses, volumes and rates and are therefore likely to be prone to similar errors as those identified in this study.

The level of variation in acetylcysteine doses observed was higher than the variation for doses recommended for individual patients of different weights, based on the weight bands. Therefore, the use of a single dose for all patients weighing above 40 kg could be a possible change that would further simplify the acetylcysteine regimen.

Pharmacy preprepared acetylcysteine infusions is another option that could reduce errors but this would lead to high costs and is unlikely to be practical in view of the multiple weight‐based dose bands and requirements for starting infusions out‐of‐hours.

Limitations

This was a single centre study at a teaching hospital with a clinical toxicology service. In most cases fewer than three infusion samples were collected for each patient; this meant that the total doses administered could not be calculated. This was in part due to the logistical difficulties of acetylcysteine being started in different locations at varying times and being prepared by many different members of staff. The rates of administration were not analysed and there was no assessment of patient outcomes, recording of the rate of adverse effects, paracetamol concentrations or markers of hepatotoxicity.

Samples were taken preinfusion but not postinfusion to limit the concentration removed from infusions and to negate the need for patient consent. However, the lack of a comparison between pre‐ and postinfusion samples limited the assessment of the impact of possible inadequate mixing.

The Trust prescription chart permits the use of either 5% glucose or 0.9% saline as the infusion fluid with a preference for 5% glucose. The type of fluid used for the samples was not recorded.

The use of percentage difference between measured and expected concentration to assess variation implies a greater impact for any variation per mg ml⁻¹ for the second, third and fourth infusions compared to the first due to the higher concentrations in the first infusion.

In comparing the results to a previous study in different centres, a historical control was used that may not have reflected previous results at the study hospital.

Conclusion

There has been a moderate improvement in the variation in acetylcysteine dose administered by infusion since the introduction of weight‐based dosing charts. However, significant variation is still present which has the potential to be associated with a risk of either treatment failure, or an increase in adverse effects. Further work is required to understand the reasons for this variation and consideration should be given to simplification of acetylcysteine regimes to decrease the risk of administration errors.

Competing Interests

There are no competing interests to declare.

Contributors

P.D. conceived and supervised the study. G.B. organised the study protocol and sample collection and analysed the data. E.R. analysed the samples in the laboratory, supported by R.F. All authors contributed to and approved the final manuscript.

Bailey, G. P. , Wood, D. M. , Archer, J. R. H. , Rab, E. , Flanagan, R. J. , and Dargan, P. I. (2017) An assessment of the variation in the concentration of acetylcysteine in infusions for the treatment of paracetamol overdose. Br J Clin Pharmacol, 83: 393–399. doi: 10.1111/bcp.13099.