Case report: quantification of methadone-induced respiratory depression using toxicokinetic/toxicodynamic relationships

Methadone has been the most widely delivered maintenance therapy for heroin users for almost 50 years. Prescribed with doses adequate to the actual needs of individual addicts, methadone efficiently reduces heroin use, overdose mortality, as well as criminal activities [1‐3]. The major pharmacological activity of methadone, a long-lasting opioid agonist, is related to its binding to the mu-opioid receptors. The available marketed methadone (6-dimethylamino-4,4-diphenyl-3-heptanone) is a racemic mixture of two stereoisomeric forms, the R- and S-enantiomers [4]. R-methadone is the main pharmacologically active isomer, believed to account for most if not all of the therapeutic effects of methadone maintenance treatment. However, although lacking strong opioid effects, S-methadone may play a significant role in the adverse responses to R,S-methadone [5‐7].

Opioid toxicity produces a classic syndrome characterised by a depressed level of consciousness, bradypnea, and miosis. Methadone overdose results in life-threatening respiratory depression, which may be reversed using naloxone, a mu-receptor competitive antagonist [8]. Regardless of whether initial doses of naloxone restore adequate respiration and further therapy is needed, repeated boluses or continuous infusion of naloxone can be used [9, 10]. In the setting of close monitoring of neurological and respiratory parameters, naloxone may obviate the need for tracheal intubation and mechanical ventilation.

To our knowledge, toxicokinetic/toxicodynamic (TK/TD) relationships in regard to respiratory depression have not been studied in methadone poisonings. Moreover, the relative kinetics of each methadone enantiomer after the ingestion of a toxic dose remains unclear. Here, we report a case of methadone poisoning with severe respiratory depression treated using naloxone. We performed a toxicokinetics study of both S- and R-methadone enantiomers and a TK/TD study to understand the relationships between the respiratory effects and the concentrations of the active R-enantiomer.

A 44-year-old man was found unconscious outdoors and admitted to the emergency department of a neighbouring hospital. On presentation, he was comatose (Glasgow Coma Score 3) with a blood pressure of 118/71 mm Hg, a heart rate of 117 beats per minute, a respiratory rate of 10 breaths per minute, and an SpO₂ (percutaneous oxygen saturation) of 75%. His temperature was 36.5°C. Neurological examination showed hypotonic and hyporeflexic coma with pinpoint pupils. Pulmonary auscultation was normal. Arterial blood gas analysis while breathing on room air showed a pH of 7.30, a PaO₂ (arterial oxygen pressure) of 56 mm Hg, PaCO₂ (arterial carbon dioxide pressure) of 48 mm Hg, and a serum bicarbonate concentration of 26 mmol/l. Other routine laboratory tests were unremarkable. Plasma lactate concentration was 2.1 mmol/l. Significant improvement in consciousness level (Glasgow Coma Score 15) without mental confusion was obtained after an intravenous bolus injection of 0.3 mg of naloxone followed by a continuous 0.3-mg/hour infusion. The patient rapidly awoke and declared that he had intentionally ingested 240 mg (3.14 mg/kg) of methadone and 2 mg of flunitrazepam to relax. The delay between ingestion and admission to the hospital was estimated to be 2.7 hours. He had a history of heroin addiction with a maintenance treatment of 70 mg/day of methadone (confirmed by a prescription from his maintenance program centre) and occasional use of 2 mg of flunitrazepam. He suffered from depression, and a recently diagnosed hyperthyroidism was treated with neomercazole. He was transferred for monitoring to our intensive care unit (ICU). The attending physicians managed the patient according to the standard treatment guidelines currently used in our department. Naloxone infusion rate was progressively adapted, aiming to keep the patient calm with a normal mental status (Glasgow Coma Score 15) and a respiratory rate of more than 12 breaths per minute. Reduction in naloxone dose was systematically attempted, and when evidence of toxicity occurred, a return to the previous dose was made. Routine toxicological screening of blood and urine, including benzodiazepines, opiates, buprenorphine, cocaine, and tetrahydrocannabinol, was negative, except for methadone. This case outcome was favourable with an uneventful 48-hour ICU stay, except for an aspiration pneumonia treated with amoxicillin-clavulanic acid for six days. The duration of naloxone intravenous infusion was 31 hours with a cumulative dose of 9.1 mg.

R-methadone accounts for most, if not all, of the effect of methadone [5‐7]. To this extent, we considered that TK/TD relationships in regard to methadone respiratory effects for S-enantiomer would probably just mirror those of R-enantiomer. Therefore, we performed a TK/TD analysis of the naloxone infusion rate needed to maintain the respiratory rate of more than 12 breaths per minute (as toxicodynamics parameter) versus plasma R,S- and R-methadone concentrations (as toxicokinetics parameter). Upon admission to the ICU, verbal informed consent was obtained from the completely awakened and non-confused patient. Venous blood samples were obtained at 2.7, 10.5, 14.5, 18.5, 20.5, 22.5, 26.5, and 30.5 hours after methadone ingestion. Plasma was separated and frozen until methadone concentration measurement. The rate of naloxone infusion was prospectively collected at each blood sampling time, and the nurses in charge were blinded to the results of the toxicological analyses. Both R- and S-methadone enantiomer concentrations were measured using a validated chiral high-performance liquid chromatography assay as previously described [11]. The plasma kinetics of each enantiomer were analysed using a noncompartmental method. Toxicokinetics parameters and TK/TD relationships were determined using a computerised curve-fitting program (WinNonlin Pro 3.0; Pharsight Corporation, Mountain View, CA, USA). The area under the curve from 0 to infinity (AUC _inf ) was calculated using the model-independent trapezoidal method.

The maximal observed plasma concentration of R,S-methadone was 1,204 ng/ml at 2.7 hours (Table 1). Decrease in plasma R,S-, R-, and S-methadone concentrations was of first-order pharmacokinetics (Figure 1). During the course of overdose after naloxone infusion, the respiratory rate was maintained at more than 12 breaths per minute (therapeutic objective). The Glasgow Coma Score was constantly at 15, and the haemodynamic parameters were within the normal range. TK/TD correlations between naloxone infusion rate (E) and R,S- and R-methadone concentrations (C) fitted well the sigmoidal E _max model, E = E _max C ^γ /[EC₅₀ ^γ + C ^γ ] + E₀, where E₀ is the baseline value of the measured rate of naloxone infusion, E _max represents the maximum possible infusion rate, EC₅₀ represents the concentration associated with the half-maximum infusion rate, and γ (or Hill coefficient) determines the steepness (slope) of the concentration-versus-response curve (Figure 2). Values of EC₅₀ and γ are given in Table 2.

Methadone pharmacokinetics and pharmacodynamics are characterised by a high inter-individual variability, even in toxic conditions [7, 12, 13]. There is up to 17-fold inter-individual variation of blood methadone concentrations for a given dosage [7]. Because methadone is administered as a racemic mixture, our objective was to study the relationships between the respiratory effects and the racemic as well as the active R-enantiomer concentrations.

R-methadone is believed to account for most, if not all, of the therapeutic effects of methadone maintenance treatment [5‐7, 14]. S-methadone may be a clinically important determinant of adverse responses to racemic methadone and may significantly contribute to the adverse but not the therapeutic effects of racemic methadone during maintenance treatment for opioid dependence. In healthy volunteers, 7.5 mg of oral S-methadone did not significantly differ from the placebo response in respiratory and pupillary effects, whereas 7.5 mg of R-methadone and 15 mg of R,S-methadone induced dramatic and sustained respiratory depression and miosis [15]. In the same volunteers, S-methadone doses between 50 and 100 mg slightly depressed ventilation. In dependent patients, S-methadone administered at dosages of 650 to 1,000 mg/day induced subjective morphine-like effects, partially suppressed abstinence from morphine, and created a mild degree of physical dependence [16].

At therapeutic doses, plasma concentrations of R,S-methadone decrease in a bi-exponential manner, with an elimination half-life estimated at 22 ± 7 hours (range 14 to 40) in opioid users after 26 days of methadone [7]. Trough plasma concentrations of 400 ng/ml for R,S-methadone or 250 ng/ml for R-methadone are usual target values when measurement of methadone concentration is performed to investigate a non-complete clinical response [17]. In our overdosed patient, the maximal observed R,S-methadone concentration was approximately three times higher than the reported peak levels in patients treated with a daily therapeutic dosage of more than 80 mg [18].

The almost mono-exponential first-order decline in plasma R- and S-methadone concentrations suggests the absence of any saturation of the distribution and the elimination processes, even at supra-therapeutic concentrations. Methadone toxicokinetics was in close agreement with pharmacokinetics, with an observed half-life in the lower range of the reported pharmacokinetic half-lives [7]. Methadone is metabolised mainly into EDDP (2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine) through cytochrome P450 (CYP) 3A4 and 2B6 pathways, for which genetic polymorphisms and high inter-individual variability of activity were reported, resulting in large inter-individual variability of methadone half-lives. In vitro metabolism studies using human liver microsomes showed apparent affinity constants (K_m) regarding CYP-mediated EDDP formation far above the maximal observed methadone concentration in this patient [19], suggesting that saturation of methadone metabolism did not occur in our patient. Moreover, no study of methadone dose-linearity was performed at such elevated concentrations. Thus, we may hypothesise that similar data (first-order kinetics) would have been observed with a larger data set of intoxicated patients presenting similar plasma methadone concentrations. Both enantiomer half-lives were rather short in comparison to previously reported values in eight patients undergoing methadone maintenance treatment (R-methadone 23 to 50 hours and S-methadone 16 to 37 hours) [20]. Thus, although no definitive explanation could be given, we may suggest that our patient was an extensive CYP 3A4 metaboliser, because no CYP 3A-inducing drugs were identified in his treatment. Moreover, as previously reported [7], the half-life of R-methadone was longer than that of S-methadone.

TK/TD relationships allow the descriptive and quantitative characterisation of the time course of the in vivo drug effect in relation to its corresponding drug concentrations within an individual [21]. There are several studies dealing with pharmacokinetic/pharmacodynamic relationships for opioids commonly used in anaesthesia or for the treatment of pain [22, 23]. However, only a limited number of studies addressed the pharmacokinetic/pharmacodynamic relationships of methadone in healthy volunteers, patients treated for chronic pain, or maintenance patients [5, 23‐25]. To our knowledge, there is no case of human methadone poisoning with a TK/TD analysis of respiratory effects. In our patient, we found sigmoidal TK/TD relationships. The maximal naloxone infusion rate was associated with an R, S-methadone concentration ranging from 400 to 1,200 ng/ml, suggesting a saturation of the mu-opioid receptors at these concentrations. Similarly, a study in healthy volunteers showed a sigmoidal response with minimal hysteresis between the iris/pupil ratio and its corresponding plasma R-methadone concentration (EC₅₀ 2.3 ng/ml and γ 9.0) [5]. In patients with chronic pain, pain relief was related to the R, S- methadone concentrations by a sigmoid E _max model defined with an EC₅₀ of 290 ng/ml and a steepness of 2.03 [25]. In our case, the high value of γ (10.0) clearly indicated that a small decrease in plasma concentrations near the EC₅₀ was associated with a dramatic decrease in the rate of naloxone infusion. R-methadone exhibited a lower γ than R,S-methadone, suggesting that S-methadone enantiomer could demonstrate a positive allosteric cooperation with R-methadone activity. However, in the present case, the confidence intervals regarding the γ values were largely overlapping, precluding any definitive conclusion. R,S- methadone EC₅₀ of 334 ng/ml was associated with significant respiratory effects in our patient. This concentration was less elevated than the target concentration of 400 ng/ml in patients treated with elevated doses, suggesting that our patient was probably equilibrated with lower methadone concentrations. However, a relationship if any between EC₅₀ and a therapeutic drug monitoring tool, as the trough plasma R,S-methadone concentration need to be established.

Moreover, although we may hypothesise that the EC₅₀ values regarding methadone respiratory effects in our patient are higher than values that could have been observed in opiate-naïve patients due to a pharmacodynamic tolerance, these EC₅₀ values were in the same range as those obtained for pain relief in chronically treated patients [25]. Increase of EC₅₀ could also be related in our patient to a pharmacogenetic variant of mu-receptor, as previously reported [26]. Thus, our data clearly support the tight safety index of methadone, previously illustrated by the occurrence of significant respiratory effects after intake at pharmacological dosages [27]. Consistently, during the first months of methadone maintenance, there is a continual alveolar hypoventilation due to the depression of both central and peripheral chemoreception [28]. Interestingly, alteration in ventilatory response to hypoxia persists after a prolonged treatment, although alveolar hypoventilation is abolished after the fully acquired tolerance within two months of the carbon dioxide-sensitive chemoreflex. However, this adaptation of ventilatory response to chronic use of methadone does not prevent the occurrence of alveolar hypoventilation in the case of acute methadone overdose.

Our toxicokinetics study has significant limitations. First, AUC _inf could be slightly underestimated due to the absence of any sample before 2.7 hours post-ingestion and between 2.7 and 10.5 hours, as the mean maximal observed concentration has been previously reported between 2.5 and 4.4 hours post-ingestion [29]. The bioavailability of methadone is also unknown in such an overdose situation. In addition, the AUC versus time was calculated between 0 and infinity as if our patient ingested a single dose of methadone, regardless of a possible long-term maintenance therapy, which was not certain in the absence of previous therapeutic drug monitoring. Thus, the apparent volume of distribution (V _z /F) and the total plasma clearance (Cl _t /F) values calculated using AUC _inf should be interpreted with caution. However, Cl _t /F and V _z /F values for both methadone enantiomers were comparable to those reported in patients undergoing methadone maintenance at therapeutic dosage [20], suggesting the absence of saturation of methadone pharmacokinetics in this case. The V _z /F values of R- and S-methadone (211 versus 170 l, respectively) were lower than those reported in treated patients (470 versus 318 l, respectively [20]) but similar to those reported in healthy volunteers (106 versus 262 l, respectively [7]). We have no definitive explanations for these differences, due to the imprecision of the AUC _inf calculation. However, methadone V _z /F is influenced by the acid α-1-glycoprotein levels with regard to its transfer to and from the central to the peripheral compartments and by extension to the site of methadone pharmacological action [7]. The Cl _t /F was similar for both methadone R- and S-enantiomers (9.0 versus 9.1 l/hour, respectively) and close to values reported in treated patients (10.7 versus 9.7 l/hour, respectively [20]). In contrast, in healthy adults, the pharmacologically active R-enantiomer has a lower Cl _t /F when compared with S-methadone (4.0 versus 20.4 l/hour, respectively) [7].

We considered the naloxone infusion rate as a surrogate marker of methadone-induced respiratory depression. However, because the elimination rate of naloxone varies with a half-life of 45 to 90 minutes, a continuous infusion runs the risk of naloxone accumulation, complicating the pharmacodynamic interactions. Otherwise, such a toxicodynamics surrogate marker requires a close concordance between real effects on respiration and the titrated infusion rate of naloxone. Moreover, serum and urine specimens were not tested for flunitrazepam in our patient. Consistently, one unusual feature in this case is the rapidity with which the patient appears to have suffered unconsciousness. Given that the peak effects of methadone usually occur approximately four to six hours after ingestion [7] and that flunitrazepam is a short-acting agent with respect to methadone, we think the initial presentation may have been influenced by the co-ingestion of flunitrazepam. However, several studies suggest that methadone peak plasma concentrations and effects may occur much earlier [30, 31], supporting (as in our patient) methadone-related steep concentration-effect relationships. Otherwise, we cannot rule out potential interactions between flunitrazepam and methadone that contributed to the respiratory depression. Both molecules are metabolised by the CYP 3A4, which may lead to clinically significant drug-drug interactions. However, the relatively short half-life of R, S-methadone does not support any inhibition of CYP 3A4-mediated methadone metabolism by flunitrazepam.

Sources of inter-individual variability of the response to methadone are numerous, including demographic and physiopathological characteristics of the methadone-maintained subjects [23, 32]. Thus, analysis of the TK/TD relationships in one patient precludes any definitive conclusion regarding the various possible situations of exposures to methadone in other patients. In addition, because plasma is not the effect site of methadone and time is needed before the drug distributes into the central nervous system, another significant concern in our study is the absence of characterisation of the plasma-versus-biophase distribution. For some opioids, there may be a delay between the time course of the plasma concentrations and the time course of the effects [5, 22, 25]. Moreover, the rate of plasma protein binding was not determined in this patient. However, because methadone is a highly lipophilic drug, the rate of cerebral distribution of the pharmacologically active molecules should not be a limiting step in this study of TK/TD relationships at toxic concentrations. Here, we described TK/TD relationships during the methadone elimination phase.

Plasma toxicokinetics of R- and S-methadone enantiomers were parallel. The study of TK/TD relationships appears to be helpful for quantifying the respiratory response to methadone during poisonings. Despite a possible pharmacodynamic tolerance in our patient, the EC₅₀ values regarding the methadone respiratory effects were in the same range as those previously reported for the analgesic effects in chronically treated patients, thus suggesting a limited safety index for methadone. A better understanding of the toxicokinetics and toxicodynamics of methadone and its enantiomers will be possible as more data accumulate in this regard.