Optimal dosing of lidocaine for seizure control in preterm and term neonates using population pharmacokinetic modelling and simulation
Marcel PH van den Broek (1), Alwin DR Huitema (2), Floris Groenendaal (3), Antoine CG Egberts (1,4), Carin MA Rademaker (1), Linda S de Vries (3)
(1) Department of Clinical Pharmacy, UMCU, Utrecht; (2) Department of Pharmacy & Pharmacology, Slotervaart Hospital/Netherlands Cancer Institute, Amsterdam; (3) Department of Neonatology, UMCU, Utrecht; (4) Division of Pharmacoepidemiology and Pharmacotherapy, Utrecht University, Utrecht
Objectives: Lidocaine is widely used around the world for different indications. It has anaesthetic, sedative, antiarrhythmic and anticonvulsive properties. The exact mechanism of its anticonvulsive effect is unknown. It is administered to neonates that are not responding to first-line anticonvulsant therapy, such as phenobarbitone. Since lidocaine also has antiarrhythmic properties, therapeutic drug monitoring (TDM) is considered useful to prevent the occurrence of cardiac arrhythmias, mainly in the form of bradycardia, although a therapeutic window for its anticonvulsant action has never been fully established. For term neonates a dosing regimen has been developed. However, this regimen was not evaluated for preterm neonates [1]. Lidocaine pharmacokinetics (PK) may be different in preterm neonates because of differences in maturation of metabolic enzymes and body size [2]. The objective of this study was to develop an optimal dosing strategy for lidocaine in preterm and term neonates using population PK modelling and simulation.
Methods: Pharmacokinetic data were available of term and preterm neonates admitted to the neonatal intensive care unit (NICU) that were treated with lidocaine intravenously. After completion of the loading dose and during the maintenance dose, blood samples were collected using an arterial line to determine efficacy and/or toxicity. Lidocaine plasma concentrations were measured in plasma using a fluorescence polarization immunoassay (FPIA). The PK-analysis was performed using NONMEM 6.2 using FOCE-I. The log-likelihood ratio test was used to discriminate between hierarchical models, based on the objective function value (OFV). Goodness-of-fit plots were used for diagnostic purposes. The influence of body size (body weight, WT) and maturation (gestational age, GA, and postnatal age, PNA) was assessed. The bootstrap re-sampling method (n=1000) was used for model evaluation. Distribution of the bootstrap parameter estimates were compared to parameter estimates of the original data set. Simulations were used to establish optimal lidocaine dosing strategies for preterm and term neonates. Several requirements for this dosing strategy were defined:
1. The dosing strategy should be easy to implement on a NICU and insensitive to calculation errors, therefore, dosing per kg WT was preferred; 2. Seizures require rapid intervention, therefore an initial bolus dose should be administered followed by an infusion during 4 hours; 3. Seizure control requires only a short duration of dosing, however, doses should be reduced slowly to decrease the occurrence of neurological withdrawal symptoms; 4. Although a therapeutic window for lidocaine for neonatal seizure control has not been established, neonatologists regularly use a target plasma concentration (at the end of the 4-hour infusion) of 6 to 7 mg/L. Lidocaine plasma concentration of >9 mg/L have been associated with increased risk for cardiac arrhythmias and should therefore be avoided. However, these findings are based on studies in a cardiological setting [3].
Results:
Model development
A total of 163 plasma concentrations from 48 neonates (WT 0.84-4.46 kg, GA 25.0-42.7 weeks) that received lidocaine were obtained. All neonates received lidocaine within 10 days after birth (PNA 0-10 days). 38% had a GA of less than 34 weeks (i.e. premature). A one-compartmental PK model was selected. As expected, body size (WT) and age (GA and PNA) were closely related. The effects of body size (allometry) and GA/PNA (maturation) on PK could therefore not be described independently. Both effects were captured by the significant relationship between WT and clearance (CL) and distribution volume (V) using: CL = θ1 * (WT/3)**alpha and V = θ2* (WT/3)**beta. Parameter estimates were θ1 = 1.41 L/h (RSE 7.87%) and θ2 = 8.95 L (RSE 4.13%). The allometric power coefficients (alpha and beta) were estimated at 1.32 (RSE 13.5%) and 1.13 (RSE 5.50%), respectively. Interindividual variability (IIV) on CL and V was 49.4% (CV14.2%) and 19.4% (CV 23.1%), respectively. Values obtained by bootstrapping were very close to the typical values and parameter precision was adequate for all PK-parameters.
Simulations
We developed a new infusion strategy based on simulations. Our strategy consisted of an initial bolus of 2 mg/kg (for all weight categories, to allow rapid administration within the NICU) in 10 minutes, followed by a body weight (WT) based infusion during 4 hours, with different doses for the different weight categories. After the 4-hour infusion, a first dose reduction is applied which is half of the loading infusion rate for 6 hours. Then a second dose reduction is applied. Again, this is half of the previous infusion rate, but now for 12 hours. The selected optimal dosing regimen is displayed in Table 1.
Table 1. Optimal dosing regimen based on simulations
Body weight (kg) | Initial bolus | Infusion (during 4h) | First dose reduction (during 6h) | Second dose reduction (during 12h) |
0.8 - 1.5 | 2 mg/kg in 10 minutes | 5 mg/kg/h | 2.5 mg/kg/h | 1.25 mg/kg/h |
1.6 - 2.5 | 6 mg/kg/h | 3 mg/kg/h | 1.5 mg/kg/h | |
2.6 - 3.5 | 6.5 mg/kg/h | 3.25 mg/kg/h | 1.625 mg/kg/h | |
3.6 - 4.5 | 7 mg/kg/h | 3.5 mg/kg/h | 1.75 mg/kg/h |
With this dosing regimen, the median concentration achieved at the end of the 4-hour infusion was 6.4 mg/L (IQR 5.5 - 7.3). At this moment only 3.8% of the simulated individuals had a concentration above 9.0 mg/L (median 9.5 mg/L, IQR 9.2-9.9). One hour later only 2.5% still had a concentration above 9.0 mg/L (median 9.4 mg/L). The initial bolus resulted in a median concentration of 0.68 mg/L (IQR 0.59 - 0.78). Results were comparable for the different body weight categories.
Conclusions: The effects of body size and maturation on the pharmacokinetics of lidocaine in this population could not be estimated separately. Therefore, bodysize (body weight) was the only significant covariate remaining in the PK model. Estimates of the allometric power coefficient were higher than 1, strongly suggesting an effect of both maturation and body size. Therefore, extrapolation with the current model beyond conditions on which the model was developed will need further validation. Based on the PK model, a dosing strategy for lidocaine for neonatal seizure control within the first 10 days after birth has been developed, which allows rapid and safe administration of lidocaine in this population. With this strategy routinely TDM would not be necessary anymore and would only be advised in case of (suspected) toxicity. This dosing strategy will be implemented on the NICU, which will allow prospective validation of this study
References:
[1] Malingré et al. (2006). Eur J Pediatr 165(9):589-604.
[2] Kearns et al. (2003). N Engl J Med 349(12): 1157-1167.
[3] Lie et al. (1974). N Engl J Med 291(25): 1324-1326.
