Dantrolene is a hydantoin derivative originally developed as an antibiotic. Its ability to reduce muscular tone by interfering with calcium efflux from the sarcoplasmic reticulum was discovered soon thereafter. It has since become the definitive treatment for malignant hyperthermia (MH), a disease that is triggered by volatile anesthetics and succinylcholine in susceptible individuals. Because the incidence of MH crises is between 1 in 15,000 and 1 in 150,000 (1,2), pharmacokinetic (PK) and pharmacodynamic data from patients being treated for acute MH are rare. No compartmental PK models have been published for this drug.
Based on available evidence, prophylactic plasma concentrations of dantrolene in humans are between 2.8 and 4.2 mg/L and are able to depress the response of a single muscle twitch by 70%–75% (3). Peak dantrolene plasma concentrations of 4.2 mg/L were obtained 3 h after administering bolus injections of 0.1 mg/kg until a twitch depression plateau was reached (2.2- to 2.5-mg/kg cumulative dose). Plasma elimination half-life t1/2el is 10–13 h in adults and 10 ± 2.6 h in children (3,4). Quantifying the half-life and appropriate treatment regimen is important because MH symptoms reoccur in up to 25% of patients within 48 h (1), which might be caused by plasma concentrations decreasing to less than the therapeutic threshold.
The aim of this study was to determine the compartmental PK parameters in adults including covariate analysis. Based on computer simulations, we also evaluated the plasma concentration time course according to the guidelines recommended by European Sources (ES) and the Malignant Hyperthermia Association of the United States (MHAUS) and developed an “ideal” dosing regimen, maintaining a constant plasma concentration, based on the individualized plasma concentrations required to control MH.
The data presented in this report were obtained during a previously published study investigating the effect of dantrolene on the thresholds of sweating, vasoconstriction, and shivering (5). Although plasma concentrations were measured, no PK analysis was performed. Briefly, nine ASA physical status I volunteers were enrolled. All were between 23 and 36 yr old. Volunteers were excluded from the study if they were taking any medications or had a history of neuromuscular disease, Raynaud syndrome, dysautonomia, or thyroid disease.
Having obtained IRB approval of the institutional human investigation committee and written informed consent, a 14-gauge catheter was inserted into a right antecubital vein for blood sampling, and an 18-gauge catheter was inserted into the left forearm vein for dantrolene administration. Dantrolene sodium was administered as a continuous IV infusion (initial dose, 5 mg/kg for 0.5 h; maintenance thereafter for 5 h, 0.05 mg · kg−1 · h−1). Venous blood samples were drawn during and after dantrolene administration for up to 60 h (at minutes 0, 2, 4, 6, 10, 15, 20, 30, 35, 40, 50, 60, 90, 120, 150, 180, 210, 270, 330, 335, 340, 345, 350, 360, 390, 420, 450, 1440, 1800, 2160, 2880, 3240, and 3600 and at the sweating, vasoconstriction, and shivering thresholds). Plasma was separated by centrifuge and stored at −30°C until assayed. Thermoregulatory manipulations (forced air warming and a circulating water mattress until sweating) took place between approximately 1 and 5 h after starting the dantrolene infusion.
Plasma concentrations of dantrolene were determined by reverse phase, high-performance liquid chromatography (HPLC), as previously described (6). Plasma aliquots were precipitated and centrifuged with acetonitrile for plasma protein removal. Reverse HPLC was performed in columns (mobile phase was acetonitrile; aqueous was glycine 35/45 vol/vol) with a flow rate of 2 mL/min. The effluent from the HPLC column flowed through an ultraviolet absorption spectrometer for detection at 405 nm. Results were normalized to external standards prepared in human plasma. The coefficient of variation (CV) was <5%, and the limit of quantification (smallest concentration of the standard probe) was 0.5 mg/L.
One-, two-, and three-compartment mammillary models were fitted to the dantrolene plasma concentration data and compared using the Akaike information criterion (7). The models were parameterized in terms of volumes of distribution, elimination, and distributional clearance (Clel and Cldist).
Interindividual variability was characterized by a multiplicative model: θ(n,i) = θ(n,m) · (1 +η(i)), where θ(n,i) refers to the individual value of the respective PK parameter, θ(n,m) is the population mean of the parameter, and η(i) varies randomly among individuals with a mean of zero and a diagonal variance-covariance matrix ω2.
A multiplicative error model was chosen to model residual variability: cobs = cexp(1+ε), where cobs refers to the observed concentration, and cexp to the concentration predicted based on dose, time, and the individual PK parameters. ε is normally distributed with a mean of zero and variance σ2.
Covariates were age, weight, sex, height, body mass index, lean body mass, and body surface area. The parameters were plotted against these covariates for visual inspection. Covariates were added one by one and kept in the model if they improved the goodness of fit, judged by the likelihood ratio criterion. For age and weight, the influence of covariates was expressed as deviation per unit of the covariate from the median value in the study population: θ(n,i) = θ(n,m)[1 + θ(d)(Cov(i) − Cov(median))], where θ(n,i) refers to the value of the respective PK parameter for the patient, θ(n,m) is the population mean of the parameter, θ(d) is the deviation from the population mean for one unit of the covariate, Cov(i) is the individual value of that covariate, and Cov(median) is the median value of the covariate in the study population. Therefore, the population mean of the parameters equals the value for the median patient. We tested for model misspecification by plotting the ratio of the measured and the predicted dantrolene plasma concentrations against observation time. For all calculations, NONMEM version V (NONMEM, San Francisco, CA) with the first-order method for all model fits, and empirical bayesian estimation of the individual parameters was used (8). Simulations were performed with Stanpump/Stangraph, freely available from Steven L. Shafer, MD (http://anesthesia.stanford.edu/pkpd).
Among the nine volunteers, five were men and four were women. They were (mean ± sd) 27 ± 5 yr old (23–36 yr), weighed 70 ± 12 kg (55–89 kg), and were 174 ± 9 cm tall (163–188 cm). A total of 306 blood samples were drawn from these volunteers, including 9 samples before the drug administration. The plasma concentration time course best fitted a two-compartment model (Fig. 1). All parameters correlated linearly with body weight (Table 1, the objective function value decreased from 45.3 to 21.2). The goodness-of-fit plot for the weight adjusted two-compartment model shows that the bayesian prediction for each volunteer matched the observations (measured plasma concentrations) well over the entire concentration range (Fig. 2).
The PK parameters of dantrolene are shown in Table 1. An unexpectedly large difference between the elimination (mean ± se) and distributional clearance (Clel = 0.03 ± 0.003 L/min and Cldist = 1.24 ± 0.22 L/min; factor 40) was observed. The PK of dantrolene can therefore be approximated by assuming one single compartment. The plasma elimination half-life of IV administered dantrolene was 10.3 h.
The compartmental PK parameter set was used to simulate the plasma concentration time course resulting from current dosing recommendations (ES and MHAUS guidelines). The simulated plasma concentration time curve based on the ES is shown in Figure 3A. Based on the severity and persistence of MH symptoms, these guidelines recommend 1 to 4 bolus(es) of IV dantrolene: 2.5 mg/kg over 4 min every 5 min, followed by a continuous infusion of 10 mg · kg−1 · 24h−1 (2,9,10) for at least 24 h, regardless of the initial dose. After the bolus(es), simulated peak plasma concentrations reached 28 to 42 mg/L. If the infusion was continued, plasma concentrations would reach a plateau of approximately 16.5 mg/L after approximately 52 h (note that Fig. 3A shows only the first 24 h). After 1 or 2 bolus injections, the plasma concentration gradually increased towards the steady-state concentration during the maintenance infusion, and after 3 or 4 boluses, the plasma concentration slowly decreased until the steady-state was reached.
MHAUS-dosing guidelines for dantrolene suggest giving 1 to 4 IV bolus(es) of 2.5 mg/kg every 5 min, followed by 1 mg/kg every 6 h for up to 48 h. As shown in Figure 3B, simulated peak plasma concentrations were equal to those described above for the ES protocol (Fig. 3A), but the maintenance bolus every 6 h resulted in fluctuating plasma concentration between 4.5 and 25 mg/L. As with the ES protocol, plasma concentrations gradually increased after 1 or 2 boluses and slightly decreased after either 3 or 4 boluses until trough concentrations were between 5 and 8 mg/L after 52 h. The trough-simulated plasma concentrations according to the MHAUS recommendations for each successive bolus injection of dantrolene were 4.2, 8.6, 13.3, and 17.3 mg/L.
As shown in Table 2, the MHAUS guidelines can be adapted to maintain a dantrolene plasma concentration of 4.5, 9.0, 13.5, or 18.0 mg/L by starting a continuous infusion approximately 5 h after the initial bolus(es) (infusion rate dependent on the initially administered dose).
The PK of dantrolene, which is used worldwide in the treatment and prophylaxis of MH and neuroleptic malignant syndrome, were evaluated, and a weight-corrected mammillary two-compartment model was found to adequately describe the PK. The calculated plasma elimination half-life (t1/2β) during this study was 10.3 hours. This agrees well with previously reported terminal elimination half-lives (t1/2β) for dantrolene (12 ± 2 h in adults (3), 10 ± 3 h in children (4), and 16 ± 6 h in MH-susceptible patients (6)), supporting the validity of our analysis.
As shown in Figure 3A, simulation of the ES dantrolene dosing scheme for treatment of acute MH using our two-compartment model yielded a wide range of plasma dantrolene concentrations (7–25 mg/L after the initial peak), which converged to a steady-state concentration of 16.5 mg/L over time. Simulating the MHAUS recommendations yielded with undulating concentrations (trough concentrations between 5 and 11 mg/L). After the initial treatment the simulated MHAUS plasma concentrations, not including the short bolus peak every six hours, were always smaller than the simulated ES values.
The undulations associated with repeated bolus administration in the current MHAUS guidelines can be avoided by administering a continuous infusion starting five hours after the initial bolus(es). Although no data about concentration-related adverse effects of dantrolene are available, we speculate that maintaining constant concentrations more than the individualized therapeutic concentration might be superior to the acceptance of large fluctuations of plasma concentrations. The current initial bolus regime was not altered. It has proved effective in emergency situations and is used to determine the maintenance rate (individualized therapeutic concentrations).
Known susceptibility to MH no longer mandates prophylactic use of dantrolene (1,9), because the syndrome is rare when susceptible patients are given trigger-free anesthesia (11). Common side effects of dantrolene, such as muscle weakness, dizziness, light-headedness, drowsiness, tachyarrhythmia, nausea, vomiting, and allergic reactions, limit its administration to MH patients detected during surgery.
Because the major aim of the original clinical investigation was to determine the effect of dantrolene on thermoregulation, forced core temperature changes (36.6°C–37.0°C) were performed during the first five hours of the study, possibly altering dantrolene’s PK. Because of the small temperature range, we believe this effect to be negligible. A far more relevant problem is the effect of possible wide fluctuations of carbon dioxide production, muscle blood flow, and cardiac output during an acute MH crisis and their effects on the PK of dantrolene. Although we cannot exclude profound changes of distributional phenomena, the small extraction ratio of this hepatically eliminated drug precludes relevant changes of the CLel. Although age was not identified as a covariate, this might very well be caused by the small age range of our study population (23–36 years). In Fig. 2, we plotted measured versus predicted plasma concentrations (predictions based on individual doses and typical parameter values [= population predictions] and individual doses and parameter values [= bayesian predictions]). The bayesian predictions match the observations well over the entire concentration range. The fact that several measured concentrations deviate profoundly from the population predictions and that these data points are observed during the infusion time becomes obvious when assessing the interindividual variabilities of parameters determining the drug concentrations in this phase before steady-state (Table 1). Both the central volume of distribution and the distribution clearance display the largest interindividual variability (CV 60% and 43%, respectively). After the termination of the infusion, the decrease of plasma concentrations is predominantly related to the CLel, which displays only modest interindividual variability (CV = 29%).
In conclusion, there has been immense improvement in MH therapy over the last few decades, mainly because of dantrolene use. However, dosing recommendations could not be based on compartmental PK analysis. We hope that our current findings fill this void and will encourage implementation of a continuous infusion with an individualized infusion rate dependent on the number of initial bolus(es) starting 5 h after initial emergency treatment. Such a protocol will achieve individualized stable plasma concentrations, according to the initial response of the patient.