What is the Easier and More Reliable Dose Calculation for IvPhenytoin in Children at Risk of Developing Convulsive Status Epilepticus, 18 kg/kg or 20 mg/kg?

Convulsive status epilepticus (CSE) is defined as a continuous or recurrent convulsive seizure with loss of consciousness lasting 30 minutes or more, or a cluster of repeated convulsions during which consciousness is not regained, lasting 30 minutes or more [1]. CSE in childhood constitutes a medical emergency as it is a life threatening condition with serious risk of neurological squeal [2]. In addition, the longer the duration of the episode, the more difficult it is to terminate [3].


Introduction
With the convulsive status epilepticus in children guideline due for renewal, we wondered if a Phenytoin dose of '20 mg/kg' would be easier to calculate correctly and therefore less prone to error and so potentially safer than the previously recommended '18 mg/kg' dose.

Background
Convulsive status epilepticus (CSE) is defined as a continuous or recurrent convulsive seizure with loss of consciousness lasting 30 minutes or more, or a cluster of repeated convulsions during which consciousness is not regained, lasting 30 minutes or more [1]. CSE in childhood constitutes a medical emergency as it is a life threatening condition with serious risk of neurological squeal [2]. In addition, the longer the duration of the episode, the more difficult it is to terminate [3].
Data from epidemiological studies suggest that four to eight children per 1000 may be expected to experience an episode of CSE before the age of 15 years [4], and in children with first seizures, 12% present with CSE as their first unprovoked seizure [5]. CSE in children has a mortality of approximately 4 % [6].
The 2000 guideline by the 'Status Epilepticus Working Party' for treating and preventing status epilepticus by treating prolonged convulsive seizures (lasting more that 5 or 10 minutes) in children in the UK, advised infusion of 18 mg/kg of Phenytoin by slow intravenous (iv) infusion over 20 minutesas a third line treatment if other treatments (generally benzodiazepines) had failed to control the seizure [7]. However, there is little agreement between hospital protocols when treating CSE in children globally, and it is well known that many hospitals in the UK and in North America use 20 mg/kg dose [8].
The objective of this study was to test medical students, trainees and consultant doctors as part of an educational exercise in dose calculation, and see if it is easier and less prone to error to calculate a dose of 20 mg/kg rather than 18 mg/kg.

Methods
A standard question paper was prepared, comprising five clinical scenarios with children of varying ages and estimated body weights. Medical students, trainee doctors at registrar and senior house officer level, and consultant paediatricians were asked to complete the exercise, confidentially, anonymously, in private, as an educational exercise, by one of two medical students (SD, PS).
Calculations were done with and without a calculator, for 18 mg/kg and for 20 mg/kg in randomised order. Speed was recorded with a stop watch, and errors were determined. For our exercise, only calculation errors of greater than 10% different from the correct dose were counted as significant errors.
The whole exercise took 5-10 minutes of the student's or doctor's time.
The data analysis was performed using SPSS version 18 (SPSS Inc, Chicago, IL). Continuous variables were tested for normality using Kolmogrov-Smirnov test and Wilcoxson signed rank test was used for paired data when not normally distributed. One way ANOVA with post hoc analysis using Dunnett C test (unequal variance) was performed for normally distributed data when comparing effect of level of seniority on time taken to calculate. The categorical data was analysed using Fischer's exact test and p values below 0.05 were taken as significant.

Results
Data was collated from the 20 scenarios as completed by 15 consultant paediatricians, 15 registrars, 15 SHOs, and 15 medical students. All answered all 20 scenarios, giving a total of 300 answers per doctor / student group, and 300 answers per type of calculation ( Table 1).
The students 'and doctors' performances were similar, with respect to the significant error rate ( Figure 1). There was a significant effect of seniority on the time taken to calculate the dose, F (3,658) = p<0.05. There was a significant quadratic trend with the time taken decreasing with seniority at registrar level and then again increasing with seniority, F(3, 658) = p = 0.028. Post hoc analysis shows registrars calculations were significantly faster than medical students (p<0.001) and SHO (p=0.006) ( Figure 4).

Error rate > 10%
When comparing the 2 doses, the numbers of errors more than 10% were significantly less in 20 mg/kg dose (0.33%) as compared to the 18 mg/kg dose (9.3%) (p<0.0001, Fischer exact test) when not using the calculator. The odds ratio for making a significant error is 30.77 *Corresponding author: Dr. M Prasad, Department of Paediatric Neurology, Nottingham Children's Hospital, Nottingham University Hospitals NHS Trust, UK, E-mail: m.prasad@nhs.net with increased risk in 18 mg/kg calculation. There was no significant difference in the significant error rate when using the calculator.
When using the 18 mg/kg dose, using a calculator significantly decreased the significant error rate from 9.3% to 0.66% (p<0.001). The odds ratio of making an error while calculating the 18 mg/kg dose without a calculator is 15.38 as compared to when using a calculator.
For the 20 mg/kg doses, there was no difference in significant error rate with or without a calculator (1%vs. 0.3%), in fact using 20 mg/kg dose the significant error rate was less without a calculator.

Speed of calculation
When comparing the 2 doses, the time taken to perform the calculations was significantly decreased using the 20 mg/kg dose as compared to 18 mg/kg dose with (median 6 seconds and 8 seconds respectively, T = 102, p<0.001, r= 0. 21) or without (median 4 seconds and 18 seconds respectively, T = 39, p< 0.0001, r= 0.79) the calculator (Figure 3). A similar trend was seen irrespective of the category of the doctor or the medical student with the speed of calculation being fastest when using 20 mg/kg dose without calculator and slowest when using 18 mg/kg dose without calculator ( Figure 2).
For the 18 mg/kg doses, speed of calculation was better than halved by using a calculator (8.0 vs. 18.0 seconds) which is statistically significant (T=39, p<0.001, r= 0.75),however for the 20 mg/kg dose there was no significant difference in time taken with (median 6 seconds) or without (median 4 seconds) the calculator. In fact the speed was quicker without the calculator in the 20 mg/kg dose (Figure 2).

Discussion
Medication errors are considered to be the commonest type of medical error [9][10][11], and recent reviews have established that paediatric patients are at particularly high risk compared to adults [12,13]. It is estimated that the true incidence of paediatric dosing errors could be approximately 500,000 per year in England. There is, therefore, an urgent need to minimise such errors [14].
Phenytoin is one of the most effective drugs for treating acute convulsive seizures, whether primarily or secondarily generalised, and status epilepticus. The main advantage of Phenytoin is the relative lack of sedating effect. However, it is considered one of the medicines most commonly responsible for dosing errors in childhood by the Royal College Paediatrics and Child Health [18].
Historically, doses quoted for ivPhenytoin and Phenobarbitone range from 15-20 mg/kg. The guidelines already recommend 20 mg/ kg as the dose for ivPhenobarbitone [7]. The difference between 20 mg/ kg and 18 mg/kg, 2mg/kg is 11.1% of 18 mg/kg. This is relatively small. The 18 mg/kg dose was first published to our knowledge in the paper by David M. Treiman [19]. This report does not justify their choice of 18 mg/kg over 20 mg/kg. We contacted Pfizer pharmaceutical company which now owns Parke-Davis who initially marketed EPANUTIN® (phenytoin sodium). According to Pfizer the dose of 18 mg/kg quoted was derived from numerous clinical pharmacology studies (dose response studies) along with safety data from their phase 3 clinical programme.
The iv infusion of Phenytoin does sometimes cause adverse cardiovascular effects e.g. bradycardia; hence in children the infusion rate should not exceed 1 mg/kg/min and should be administered with cardiac monitoring.
It is well known that many hospitals' local guidelines advocate 20 18    For this educational exercise, only calculation errors greater than 10% different from the correct dose were counted as significant errors. The significant error rate was considerably lower and the speed much quicker for calculating 20 mg/kg dose when compared with 18 mg/kg dose, without a calculator, which is the current recommended dose [7].
The exercise demonstrated that doctors and medical students will make errors in simple dose calculations in at best 0.3-1% of calculations even with a calculator. These errors may be higher in the real life scenarios as managing status epilepticus is a medical emergency which can be stressful, making us more prone for errors. This underlines the importance of checking all dose calculations.
Often in emergency situations a calculator will not be readily to hand and checking will require another professional to use the calculator again, this all compounds the stress and anxiety of the situation.
We propose that new status epilepticus guidelines should make an attempt to minimise ivPhenytoin dose calculation errors and therefore recommend 20 mg/kg.  Recently in the latest edition of 'Advanced Paediatric Life Support', the status epilepticus algorithm has advised the use of 20 mg/kg iv Phenytoin dose replacing the previously recommended dose of 18 mg/ kg [20].