Volume 56, Issue 4 p. 441-450
Free Access

Moclobemide poisoning: toxicokinetics and occurrence of serotonin toxicity

Geoffrey K. Isbister

Corresponding Author

Geoffrey K. Isbister

Discipline of Clinical Pharmacology, University Of Newcastle,

Department of Clinical Toxicology and Pharmacology, Newcastle Mater Misericordiae Hospital, Newcastle,

Dr Geoffrey K. Isbister, Discipline of Clinical Pharmacology, Level 5, Clinical Sciences Building, Newcastle Mater Misericordiae Hospital, Waratah NSW 2298, Australia. Tel.: + 61 2 4921 1293; Fax: + 61 2 4960 2088; E-mail: [email protected]Search for more papers by this author
L. P. Hackett

L. P. Hackett

Clinical Pharmacology and Toxicology, The Western Australian Centre for Pathology and Medical Research, Perth, WA, Australia

Search for more papers by this author
Andrew H. Dawson

Andrew H. Dawson

Discipline of Clinical Pharmacology, University Of Newcastle,

Department of Clinical Toxicology and Pharmacology, Newcastle Mater Misericordiae Hospital, Newcastle,

Search for more papers by this author
Ian M. Whyte

Ian M. Whyte

Discipline of Clinical Pharmacology, University Of Newcastle,

Department of Clinical Toxicology and Pharmacology, Newcastle Mater Misericordiae Hospital, Newcastle,

Search for more papers by this author
Anthony J. Smith

Anthony J. Smith

Discipline of Clinical Pharmacology, University Of Newcastle,

Search for more papers by this author
First published: 12 September 2003
Citations: 57

Abstract

Aims To investigate the spectrum of toxicity of moclobemide overdose, the occurrence of serotonin toxicity, and to estimate toxicokinetic parameters.

Methods All moclobemide overdoses presenting over a 10-year period to the Hunter Area Toxicology Service were reviewed. Clinical features, complications, length of stay (LOS) and intensive care (ICU) admission rate were extracted from a standardized, prospectively collected database. Comparisons were made between moclobemide alone and moclobemide with a serotonergic coingestant poisoning. Serotonin toxicity was defined by a combination of Sternbach's criteria and a clinical toxicologist's diagnosis. In five patients serial moclobemide concentrations were measured. Time to maximal plasma concentration (Tmax), peak plasma concentration (Cmax) and terminal elimination half-lives were estimated.

Results Of 106 included patients, 33 ingested moclobemide alone, 21 ingested moclobemide with another serotonergic agent (in some cases in therapeutic doses) and 52 ingested moclobemide with a nonserotonergic agent. Eleven (55%) of 21 patients coingesting a serotonergic drug developed serotonin toxicity, which was significantly more than one (3%) of 33 moclobemide-alone overdoses (odds ratio 35, 95% confidence inteval 4, 307; P < 0.0001). In six of these 21 cases severe serotonin toxicity developed with temperature> 38.5 °C and muscle rigidity requiring intubation and paralysis. The 21 patients had a significantly increased LOS (34 h) compared with moclobemide alone overdoses (12 h) (P < 0.0001) and a significantly increased ICU admission rate of 57% vs. 3% (P < 0.0001). Time to peak plasma concentration was delayed in two patients where prepeak samples were obtained. Cmax increased slightly with dose, but all three patients ingesting ≥ 6 g vomited or had charcoal. The mean elimination half-life of moclobemide in the five patients in whom serial moclobemide concentrations were measured was 6.3 h and elimination was first order in all cases. There was no evidence of a dose-dependent increase in half-life.

Conclusions The effects of moclobemide alone in overdose are minor, even with massive ingestions. However, moclobemide overdose in combination with a serotonergic agent (even in normal therapeutic doses) can cause severe serotonin toxicity. The elimination half-life is prolonged by two to four times in overdose, compared with that found in healthy volunteers given therapeutic doses. This may be a result of wide interindividual variation in overall elimination, also seen with therapeutic doses, but appears not to be due to saturation of normal elimination pathways.

Introduction

Moclobemide is a reversible and selective inhibitor of monoamine oxidase (MAO) type A [1, 2]. Controlled clinical trials have demonstrated it to be effective in the treatment of depressive disorders. It, and other newer antidepressants, such as the selective serotonin reuptake inhibitors (SSRIs), have increasingly been used for the treatment of depression, instead of the older tricyclic antidepressants (TCAs) [1, 2].

There is conflicting opinion in the literature regarding the safety of moclobemide in overdose based on case reports and small case series. Some authors have concluded that the drug does not cause major toxicity and is far safer than tricyclic antidepressants [3, 4]. However, there are reports of fatalities from moclobemide alone [5, 6] and in combination with clomipramine [6–10], citalopram [8] and paroxetine [11]. Severe effects have been reported with the combination of moclobemide and venlafaxine [12] and moclobemide and paroxetine in overdose [13], all attributed to serotonin toxicity. It appears that most severe or fatal cases of moclobemide poisoning occur with coingestion of serotonergic agents [8–10, 12, 13], increasing the likelihood that the effects are the result of serotonin toxicity.

Two early series of moclobemide poisonings were reported prior to the widespread recognition of serotonin toxicity [4, 14]. One is from initial reports of 18 cases to the manufacturer [4] and the other is a small series of reports to a poisons information centre [14]. Both conclude that moclobemide is relatively nontoxic unless ingested with other agents. However, they do not identify cases of serotonin toxicity, unlike many subsequent case reports since, although many of the features described in the poison centre series are consistent with the diagnosis [14].

The kinetics of drugs in overdose are poorly defined, and moclobemide is no exception. Moclobemide is rapidly and extensively metabolized, although only two metabolites are present to a large extent (Figure 1). Neither of the two metabolites have major inhibitory activity on MAO-A, but their kinetics are helpful in differentiating poor and extensive metabolizers. There are two reports that measure moclobemide and its metabolites following overdose [9, 15] and a number of deaths reported in which drug concentrations were measured [6, 11]. It is important to determine the toxicokinetics of drugs to predict the time course of their adverse effects so that safe observation, admission and discharge policies can be developed.

Details are in the caption following the image

Diagram of moclobemide and its four major metabolic pathways with the important plasma and urinary metabolites illustrated. Not all 19 metabolites are included and those included are either marked by manufacturer's numbers where available or the numbering system as per Jauch et al. [35]. Full structures of the parent and two major plasma metabolites are included. Circled percentages indicate the approximate percentage dose recovered in urine. M5 has some monoamine oxidase (MAO)-A-inhibitory activity, and M2 and M4 have been found to inhibit MAO-B in rat liver. Importantly, M15, the major plasma metabolite, is inactive. Modified from Jauch et al. 1990 [35].

We present a cohort of moclobemide overdoses presenting to a tertiary hospital toxicology unit over a period of 10 years. Serial drug concentrations were determined in five cases associated with moderate to severe effects, and were modelled to estimate some toxicokinetic parameters.

Methods

The Hunter Area Toxicology Service (HATS) is a regional toxicology unit situated at the Newcastle Mater Misericordiae Hospital that serves a population of about 350 000 people and is a tertiary referral centre for a further 150 000 [16]. All toxicological presentations to emergency departments in the region are either admitted to this unit or notified to HATS and entered prospectively into a clinical database. A preformatted admission sheet is used by medical staff to collect data on admission [17] and this and additional information from the medical record is entered into the database by two trained personnel blinded to any study hypotheses. Detailed demographic and clinical information is recorded [18]. From the database, cases of moclobemide poisoning presenting between January 1992 and June 2002 were identified.

Moclobemide overdose was defined as ingestion of> 600 mg (maximum recommended daily dose of 300 mg twice daily). Out of all admissions for moclobemide poisoning only a patient's first admission was included. Second and subsequent admissions were excluded to remove the potential bias of an individual susceptibility to moclobemide.

The data reviewed included patient demography (sex, age), details of the moclobemide ingestion (estimated time of ingestion, estimation of amount, coingestion), clinical features [heart rate (HR), blood pressure (BP), maximum temperature, Glasgow coma score (GCS)], investigations, outcomes [mortality, seizures, coma (GCS < 8), arrhythmias, hypotension, serotonin toxicity, length of stay (LOS) and intensive care (ICU) admission], and treatment.

Cases of serotonin toxicity were defined as those fulfilling Sternbach's criteria where the attending clinical toxicologist also diagnosed serotonin toxicity. Sternbach's criteria are the presence of at least three of the following clinical features: mental status changes (confusion, hypomania), agitation, myoclonus, hyperreflexia, diaphoresis, shivering, tremor, diarrhoea, incoordination, fever in a patient commencing or having an increased dose of a serotonergic agent. In addition, other medical causes must be excluded, and a neuroleptic not commenced or increased in dose prior to the commencement of symptoms. The use of the additional inclusion criterion for serotonin toxicity rather than Sternbach's criteria alone [19] is because Sternbach's criteria are not sensitive or specific enough for the diagnosis of serotonin toxicity in our unit's experience [20]. Severe serotonin toxicity was defined as a temperature> 38.5 °C or severe hypertonia and/or spontaneous clonus necessitating muscle paralysis. Serotonergic coingestants included all selective serotonin reuptake inhibitors, venlafaxine, all tricyclic antidepressants, other monoamine oxidase inhibitors, lithium and amphetamines.

Toxicokinetics

In five cases with moderate to severe poisoning serial moclobemide concentrations were measured. Plasma concentrations of moclobemide and three of its metabolites (Ro 12-8095, Ro 12-5637 and Ro 16-3177) were quantified by reverse phase high-performance liquid chromatography (HPLC) as described previously by Geschke et al. with some minor modifications [21]. The limit of detection for moclobemide and the three metabolites was 5.0 µg l−1.

The maximum plasma concentration (Cmax) and the time to Cmax (Tmax) were read directly from the plasma concentration–time curves if possible, otherwise an estimated Cmax was used which was taken to be the first and highest plasma concentration.

The terminal elimination half-lives for the parent drug were derived from plasma concentration vs. time data using a first-order elimination, one-compartment disposition model. Data points prior to peak plasma concentration were ignored and an exponential decay was fitted to the plasma concentration–time data and half-life was calculated. Half-lives for two of the three metabolites were also determined. Analysis and curve fitting were performed using GraphPad Prism version 3.02 for Windows (GraphPad Software, San Diego, CA, USA).

Statistical analysis

For descriptive statistics, means and standard deviations (SD) are quoted for normally distributed data, while medians and interquartile ranges (IQR) are used for nonparametric data. For comparison of two groups, unpaired t-test was used for normal populations and the Mann–Whitney test for nonparametric populations. For comparison of proportions of two groups compared with either control group, Fisher's exact test was used. All statistical analysis was done using GraphPad InStat (version 3.02 for Windows 95; GraphPad Software).

Results

There were 127 admissions with moclobemide poisoning to HATS between January 1992 and June 2002. Twelve were excluded because they ingested 600 mg or less. Only first admissions were included, so another nine admissions were excluded, leaving 106 cases for analysis. Of these patients, 33 ingested moclobemide alone. There were 42 males and 64 females with a mean age of 33 years (SD 12 years).

The clinical effects of all moclobemide poisonings are summarized in Table 1. There were no deaths, complications or major toxic effects in the moclobemide alone group despite a median ingested dose of 6 g (IQR 3.6–7.5 g; range 0.9–18 g). The patient who ingested 18 g developed hyperreflexia, tachycardia and a temperature of 37.4 °C, but no other significant effects.

Table 1. Clinical effects of 33 moclobemide overdose admissions where moclobemide was the only drug ingested (column A) compared with all other moclobemide overdose admissions where coingestants were taken.
Moclobemide alone
overdoses (33)
Moclobemide overdoses
with coingestants (73)
Baseline characteristics
Age (years) 27 (IQR 23–34) 34 (IQR 26–44)***
Gender (% males) 8, 24% 34**,47%
Dose (g) 6 (IQR 3.6–7.5) 4 (IQR 2.3–6.7)*
Presentation time (h) 1.5 (IQR 1.0–3.3) 1.9 (IQR 0.8–4.0) NS
Clinical features
Neurological
 Coma 0, 0% 4,1 5.5%
 Seizures 0, 0% 6,2 8.2%
Cardiovascular
 Tachycardia 7, 21% 15, 21%
 Arrhythmia 0, 0% 2,3 3%
 Length of stay (median, IQR) 12 (9–20) 20 (13–38)***
 ICU admission 1, 3% 16,**** 22%
 Serotonin toxicity 1, 3% 13,NS 18%
  • * P = 0.028;
  • ** P = 0.034;
  • *** P < 0.01;
  • **** P = 0.020. NS, Not significant;
  • 1 , all four had coingested a tricyclic antidepressant;
  • 2 2 , all had coingested proconvulsant drugs: dothiepin, doxepin, venlafaxine (3) and thioridazine;
  • 3 3 , asystole occurred in one patient who coingested metoprolol and ventricular tachycardia occurred in another who coingested dothiepin.

All major complications in the 106 patients, including seizures and coma, occurred in patients taking coingestants and could be accounted for by the coingested drug (Table 2). The median dose of moclobemide in the 73 patients coingesting other drugs was 4 g (IQR 2.3–6.7), which was significantly less than the moclobemide-alone overdoses (P = 0.028).

Table 2. Comparison of serotonergic effects and complications between moclobemide alone overdoses and moclobemide overdoses where a serotonergic coingestant was taken.
Moclobemide alone
overdoses (33)
Moclobemide overdoses
with serotonergic
coingestants (21)
Moclobemide overdoses
with nonserotonergic
coingestants (52)
Baseline characteristics
Age (year) 27 (IQR 23–34) 30 (IQR 25–41)NS 36 (IQR 26–44)
Gender (% male) 8, 24% 10, 48% 24, 46%
Dose (g) 6 (IQR 3.6–7.5) 3 (IQR 1.7–7.9)NS 4.2 (IQR 2.7–6.3)
Presentation time (h) 1.5 (IQR 1.0–3.3) 1.5 (IQR 1–3.9) NS 2 (IQR 0.8–3.9)
Clinical feature
Serotonin toxicity 1, 3% 11,* 52% 2, 4%
Severe serotonin toxicity 0, 0% 6, 29% 0, 0%
Features of serotonin toxicity
Clonus 0, 0% 11, 52% 2, 4%
Hyperreflexia 5, 15% 12, 57% 7, 13%
Rigidity 0, 0% 6, 29% 0, 0%
Temp.> 37.2 °C 0, 0% 8, 19% 1, 2%
Temp.> 38.5 °C 0, 0% 4, 14% 1, 2%
Length of stay 12 (9–20) 39 (20–79)** 18 (12–26)
ICU admission 1, 3% 12,* 55% 4, 8%
  • The third column contains all other overdoses where a nonserotonergic agent was taken. Comparison is between the first two columns only. NS, Not significant. *P < 0.01; **P < 0.0001.

The 33 cases of moclobemide-alone overdoses are compared with the 21 cases where a serotonergic agent was taken as a coingestant in Table 2. Eleven (52%) of the 21 patients coingesting a serotonergic drug developed serotonin toxicity, which was significantly more than for moclobemide alone overdoses, one of 33 (3%), [odds ratio (OR) 35, 95% confidence interval (CI) 4, 307; P < 0.0001]. In six of the 11 cases the features were consistent with severe serotonin toxicity necessitating sedation and intubation of the patient. There was a significantly increased LOS for serotonergic coingestant overdoses of 39 h (IQR 20–79 h) vs. moclobemide alone overdoses 12 h (IQR 9–20 h) and a significantly increased rate of ICU admission of 11 in 22 (52%) vs. one in 33 (3%) (OR 35, 95% CI 4, 307; P < 0.0001). The coingested agent in these 11 patients was an SSRI in five patients, venlafaxine in four patients, doxepin and tranylcypromine in one patient each. No patient ingested more than 30 defined daily doses (DDD) of the coingestant, one patient coingested only 150 mg of venlafaxine and another 750 mg of venlafaxine. The patient coingesting 150 mg venlafaxine developed severe serotonin toxicity requiring intubation and paralysis (patient B).

Toxicokinetics

In five patients who developed significant serotonin toxicity serial plasma concentrations of moclobemide and three metabolites were measured and plotted as plasma concentration–time curves. Figure 2 shows the plasma concentration–time curves for the five patients. Information on coingestants taken in these five cases was based on the history, and was confirmed in each case by HPLC (Table 3).

Details are in the caption following the image

Plasma concentration–time profile for moclobemide (▪) (Ro 11-1163) and three of its metabolites, the N-oxide metabolite (Ro 12-5637 (□)), the oxo-metabolite (Ro 12-8095 (○)) and the ring-opened metabolite (Ro 16-3177 (◆)) for five of the patients.

Table 3. Toxicokinetic information on five patients where serial plasma concentrations were measured.
Patient A B C D E
Estimated dose (g)  6 12  9 3  1.8
Coingested drugs Venlafaxine Venlafaxine Fluvoxamine Sertraline, amitriptyline Citalopram
C max (mg l–1) 28 50 30 8 25
Terminal half-life (h) 10.2  7.5  4.7 4.3  4.9
Metabolites
Ro 12-5637 half-life (h)  8.6  6.8  6.5 4.3  4.6
Ro 12-8095 half-life (h) 10.9 10.3 1 4.3  6.2
  • Information on Ro 16-3177 not included. 1, Too few points for analysis.

The time to peak plasma concentration was delayed in at least two patients, occurring at 13 h in patient B and 4.5 h in patient C. In patient B the delayed peak corresponded with a delayed onset of serotonin toxicity and a generalized tonic clonic seizure 12 h after ingestion. For three of the patients the first plasma concentration was taken after the peak.

A plot of Cmaxvs. time is shown in Figure 3. Included in the figure are data from pharmacokinetic studies and a linear regression of these data [22], as well as two other overdose cases where data were available [15]. Cmax appears to increase with increasing dose, but the points in our study fall below the expected values based on data from therapeutic doses (Figure 3).

Details are in the caption following the image

Maximum concentration (Cmax) vs. estimated dose plot (top) and plasma elimination half-life vs. dose (bottom). PK data are taken from Mayersohn et al. [22] and the other overdose data are from Hackett et al. [15]. Patient A had gastric lavage and activated charcoal approximately 4 h after ingestion. Patient B who took 12 g was given activated charcoal 75 min after ingestion and patient C who took 9 g spontaneously vomited some tablet material within 60 min. In patients D and E, the Cmax may be a significant underestimate, so has been marked with an open box. The drug ingestion history for patient D taking 3 g was very unreliable. Cmax (▪); estimated Cmax (□); PK Data (▾); and other overdoses (○).

The mean terminal elimination half-life for the five patients for the parent drug was 6.3 h (SD 2.5 h) and individual values are listed in Table 3. A plot of plasma elimination half-life vs. time shows that there is no significant correlation between dose and elimination half-life (Figure 3). The elimination of moclobemide was first order in all cases with no evidence of dose-dependent kinetics.

Two of the metabolites with first-order elimination also had half-lives estimated (Table 3). Ro 12-5637, the N-oxide metabolite, had a similar terminal elimination half-life to the parent drug, declining in parallel, suggesting that its kinetics was formation limited. In contrast, the plasma concentration curve of Ro 12-8095, the oxo-metabolite, declined with a longer half-life (Table 3), which suggests that the metabolite was elimination limited. Ro 16-3177 was found in very low concentrations in the plasma, so no kinetic analysis was possible.

Discussion

Clinical effects of moclobemide poisoning

This study shows the spectrum of toxicity of moclobemide poisoning when taken alone and when combined with serotonergic agents. Single-agent poisoning with moclobemide causes minimal toxicity even when massive amounts of drug are ingested. However, the combination of moclobemide with other serotonergic drugs causes serotonin toxicity in over half of cases and severe serotonin toxicity in 29% of cases necessitating intubation, paralysis and sedation. The odds of developing serotonin toxicity in moclobemide plus a serotonergic drug overdose were 35 times those of moclobemide-alone overdoses.

There was no evidence in this study to suggest that the serotonin toxicity occurring in the coingestant group was solely due to the overdose of the serotonergic agent (e.g. SSRI), and therefore, unrelated to the moclobemide poisoning. The amounts of the serotonergic agent ingested in patient who developed serotonin toxicity were not massive (all < 30 DDD), and in two patients were therapeutic, one of which had severe serotonin toxicity. This is also supported by the fact that in SSRI overdoses with no other serotonergic coingestants, serotonin toxicity only occurred in 15% and severe serotonin toxicity did not occur in 320 of our patients (unpublished data).

There continues to be disagreement over the definition of serotonin toxicity [23–25], which appears to be a continuum of toxicity rather than a discrete syndrome [23, 25, 26]. This study provides some support for this hypothesis. There were patients with definite serotonin toxicity who did not require any major interventions and were often treated with oral cyproheptadine alone. However, there was a group arbitrarily defined as severe serotonin toxicity, who had potentially life-threatening effects and required major interventions in a critical care setting.

The specific treatment of serotonin toxicity is continually evolving. There is increasing use of 5-HT2A receptor antagonists, such as cyproheptadine and chlorpromazine, which appear to be effective based on clinical experience [24, 27] and animal studies [28, 29]. However, all six severe cases in this series required intubation, sedation and most importantly muscle paralysis. In the severe cases hypertonia and clonus caused both respiratory compromise due to chest wall rigidity and increased muscle activity contributing to hyperthermia. In severe cases this muscle activity can result in rhabdomyolysis [9]. Patients coingesting moclobemide and another serotonergic agent must be managed aggressively in a critical care area to prevent severe serotonin toxicity. 5-HT2A receptor antagonists may reverse these effects in isolated cases [24, 27, 30] and may be useful if administered early in the course of the overdose, but there remain no controlled trials to support the use of these agents. However, in severe cases supportive care must take priority to prevent life-threatening complications. A rapidly rising temperature, muscle rigidity and ventilatory failure, with a rising pCO2, are indicators of life-threatening serotonin toxicity and patients with these signs should be intubated and paralysed.

Plasma moclobemide drug concentrations were only determined for five patients in this study. Although this is the best way to confirm ingestion of the drug, it was not possible for all patients. However, all overdose patients admitted to HATS have the drug ingested confirmed by history from the patient, and this is confirmed with collateral history from family, friends and ambulance officers, as well as evidence of empty drug containers. Previous studies of overdose have demonstrated that this corresponds to definite ingestion in the majority of cases [31, 32].

Toxicokinetics of moclobemide

The kinetics of drugs in overdose is a far less precise science than the kinetics of drugs at therapeutic doses because it is opportunistic, rather than planned. There are inaccuracies in the time of the ingestion, the stated dose and often the first plasma sample is not collected until 1–4 h after ingestion. In addition, the patient often spontaneously vomits or receives decontamination with activated charcoal, which decreases the ingested dose by an unknown amount. This means that the absorption phase of the plasma concentration–time graph is either missing or incomplete and it is usually difficult or impossible accurately to estimate the AUC, Tmax or Cmax.

After a single therapeutic dose moclobemide is rapidly and completely absorbed with a Tmax of 1–1.5 h [22]. In only two of our patients was it possible to estimate the Tmax, and in both cases it exceeded 1.5 h. A prolonged absorption phase may occur following overdose for a variety of reasons, including delayed gastric emptying, formation of tablet aggregates and saturation of the usual mechanisms of absorption, due to the excessive amount of drug in the gastrointestinal tract.

Moclobemide has an oral bioavailability ranging from 43% to 59% because of significant variable first-pass metabolism [33], which further increases to 86% with multiple therapeutic doses, due to either saturation or autoinhibition [33]. Peak concentration (Cmax) increases linearly with dose (although there is substantial interindividual difference [22]), and almost doubles at steady state after 10 days of multiple doses [34].

An apparent Cmax in our cases was estimated from either the actual peak concentration (available in two patients) or the first plasma concentration in the other three patients. The Cmax appears to increase slightly with dose, but falls below those predicted from the kinetics derived after therapeutic dose (Figure 3). This may be explained partly by the fact that the patients A, B and C either spontaneously vomited or were decontaminated early.

Following oral administration of 50 mg, a mean of 95% of a dose of moclobemide is excreted in urine within 96 h [35]. The parent drug is almost completely metabolized, with at most 0.4% excreted unchanged. There are at least 19 metabolites accounting for 64% of the dose [35]. There are two major metabolic cascades, morpholine C-oxidation and morpholine N-oxidation (Figure 1). Morpholine C-oxidation is the most important metabolic pathway producing the major plasma metabolite, Ro 12-8095, which is detectable in plasma in amounts equivalent to the parent drug. It undergoes extensive metabolism and is not detectable in urine (Figure 1) [21, 35, 36]. It has no inhibitory activity on MAO-A. The second important pathway is N-oxidation producing the other major metabolite detectable in plasma, Ro 12-5637 (M5) (Figure 1). It has some MAO-A inhibitory activity, but occurs in plasma at much lower concentrations than the parent compound. It is detectable in trace amounts in urine. Aromatic hydroxylation and deamination also occur but are quantitatively less important pathways (Figure 1).

The involvement of the different families of cytochrome P450 enzymes in the metabolism of moclobemide is unresolved. Gram et al. demonstrated that the formation of the Ro 12-8095 by C-oxidation was mainly due to the cytochrome P450 2C19 enzyme [37]. In poor metabolizers (PM) of mephenytoin (2C19 phenotype) the elimination half-life of moclobemide was prolonged at 4 h and the plasma concentration decline of Ro 12-8095 was in parallel (i.e. formation-limited kinetics). In extensive metabolizers (EM), the mean half-life of moclobemide was 1.8 h, and the decline of Ro 12-8095 was longer than the parent, indicating elimination-limited kinetics of the metabolite [37]. At steady state the difference remained and earlier observations [36] were confirmed that moclobemide had nonlinear kinetics, with a reduced clearance at steady state [37].

In the five overdose patients the pattern of elimination of moclobemide and its metabolites was almost identical to the elimination of moclobemide seen in EM of mephenytoin at steady state [37], with formation-limited elimination of Ro 12-5637, but not Ro 12-8095, and the relatively similar concentrations of the parent and Ro 12-8095. Thus, all five patients are likely to be EM of mephenytoin (2C19) [38] and genetic polymorphism of P450 2C19 is unlikely to explain the increased terminal elimination half-lives.

The apparent oral clearance of moclobemide varies significantly between individuals, as expected for a drug with intermediate to high hepatic extraction [22], but decreases overall with increasing dose. This has been attributed to high first-pass metabolism which appears to saturate [22]. Clearance is also reduced at steady state, suggesting that moclobemide may reduce its own metabolism [22]. However, the plasma concentration–time curves remain log-linear before and after multiple dose administration, indicating that there is inhibition of metabolism (possibly metabolite inhibition), rather than overall saturation of metabolism [22, 34, 36, 39].

Moclobemide has an elimination half-life of 1.6–2.1 h in healthy volunteers [2, 33, 36, 39], which does not increase with age or renal impairment. The half-life is prolonged in patients with hepatic cirrhosis [33], and prolonged with multiple doses. A terminal plasma elimination half-life was estimated in all five of our patients, but there was no relationship between dose and plasma elimination half-life (Figure 3). The plasma concentration–time curves in our five patients were similar to those from single-dose and multiple-dose pharmacokinetics studies, with first-order elimination and no evidence of saturation [21, 34, 39]. Our study thus provides evidence that overall moclobemide metabolism and elimination is not saturable, even with extreme doses and very high plasma concentrations.

There are only two previous studies on the toxicokinetics of moclobemide [9, 15]. The first by Hackett et al. reported two patients with delayed absorption and prolonged elimination half-lives [15]. The elimination half-lives in these patients were 5.6 h and 12.5 h. Both had been taking the drug therapeutically for months, which may partly explain the increase in half-life, but the authors were not able to account for the much longer half-life in one patient [15]. In a later case report, Power et al. described a fatal case where the initial elimination half-life was about 5 h [9]. However, the elimination of moclobemide was significantly compromised about 30 h after ingestion due to multiorgan failure and most probably a dramatic reduction in hepatic clearance [9]. The estimated terminal elimination half-lives in these studies are consistent with the prolonged half-lives determined in our study.

Patients C, D, and E had plasma elimination half-lives that were about twice that predicted from a single therapeutic dose. None was taking moclobemide therapeutically, so the increased half-lives were not a result of multiple dosing. All three coingested antidepressants (Table 3), but this is unlikely to account for the longer half-lives because only fluvoxamine is a potent inhibitor of 2C19 [40, 41].

Patient A and B, who both coingested venlafaxine, had plasma elimination half-lives more than twice that of the other three patients. There is no evidence that venlafaxine or its metabolites inhibit 2C19 [40], so this is unlikely to account for the prolonged plasma elimination half-life in these two patients. Neither was unwell or had a history of renal or hepatic disease. It is thus likely that the explanation for their prolonged terminal elimination half-lives in patients A and B is unrelated to venlafaxine. Interindividual differences in clearance between the five patients may be the only reason for the large variation in half-lives in overdose, which is consistent with the large interindividual clearance seen with therapeutic doses [22].

The prolonged absorption and elimination of moclobemide is consistent with the time course of the clinical effects, when they were apparent. In patients with serotonin toxicity the median length of stay was> 36 h at which time there were still significant plasma concentrations of moclobemide. The major plasma metabolite Ro 12-8095 was inactive, and the plasma concentration of the only active metabolites, Ro 12-5637 and Ro 16-3177 were much lower in the five patients. Thus, the effects of moclobemide in overdose would be expected to correlate mainly with plasma concentrations of the parent drug.

Although this study suffers from being a retrospective review of a clinical database, the information in the database had been entered prospectively in a standardized blinded manner [17, 18]. The routine collection of all cases allows a better characterization of the full spectrum of effects and the importance of coingestants. A further problem with the study was the use of Sternbach's criteria to define serotonin toxicity. The use of the clinical toxicologist's diagnosis of serotonin toxicity was added to improve the definition, but improved objective criteria are needed to clarify this diagnosis.

Conclusions

This study is the first to attempt to characterize more fully the toxicokinetics of acute overdose with moclobemide. It illustrates the difficulties in obtaining data, the inaccuracies of the data and the estimates that need to be made. We have shown that in overdose both the absorption and the terminal elimination half-life of moclobemide are prolonged. There is no evidence of saturation of hepatic metabolism, and the reason for the reduced elimination is not clear. The time course of overdose is much longer than that predicted by therapeutic dose pharmacokinetic studies.

Major toxicity associated with moclobemide occurred only in the presence of coingestion of a serotonergic drug, and the effect appeared to be an interaction between the serotonergic coingestant and moclobemide, rather than attributable to the toxicity of the coingestant alone. The effects of moclobemide-alone overdose are minor, even with massive ingestions, and do not need to be managed in an ICU or require prolonged monitoring.

We would like to thank Stuart Allen for extracting the data from the database, and Deb Whyte and Toni Nash for entering the data into the database.