|Fatal Drug-Drug Interaction As a Differential Consideration in Apparent Suicides
SHELDON H. PRESKORN, MD
Journal of Psychiatric Practice, July 2002, 233-238
This column is another in a series presenting real-life cases to illustrate clinical pharmacological principles.1 In contrast to earlier columns in this series, this one will raise more questions than it will answer. The topic is drug-drug interactions as a possible mechanism contributing to death in drug overdoses or even as the cause of a false positive diagnosis of suicide.
When DDIs are discussed, it is usually within the context of routine clinical care; however, DDIs may also occur in acute overdoses. Like all DDIs, those occurring in an overdose may be pharmacodynamically or pharmacokinetically mediated. An example of the former is found in Neuvonen et al.'s report of five fatal cases of the serotonin syndrome occurring as a result of an acute overdose of moclobemide and either citalopram or clomipramine.2 This column will review the possibility that a pharmacokinetic DDI may have contributed to the death of 13 people who presumably died of acute overdoses of fluvoxamine and other co-ingested drugs.3
It is even conceivable that one or more than one of these 13 people considered to have died from an acute overdose may have died simply due to a DDI mimicking an acute overdose. An example of this phenomenon was presented by Preskorn and Baker and concerned a patient whose death was initially attributed to suicide via an acute intentional overdose, but who in fact died as a result of a pharmacokinetic interaction between fluoxetine and amitriptyline which led to the gradual accumulation of toxic levels of amitriptyline.4
Criteria for establishing death due to an acute overdose
The level of proof for establishing death as a result of an acute overdose may range from "beyond a reasonable doubt" to the best explanation given a limited number of facts. Evidence meeting the "beyond a reasonable doubt" criteria can include a suicide note in the victim's handwriting or the statement of the victim to a reliable witness, the presence of an empty pill bottle that had recently been filled, the presence of a substantial quantity of pill fragments in the stomach, the ingestion of a drug with a narrow therapeutic index, high concentrations of that drug in bodily fluids (typically blood) and/or tissue, documented physical manifestations of severe drug intoxication (e.g., arrhythmia's following a tricyclic antidepressant [TCA] overdose), and the absence of other causes of death. The documentation of physical manifestations of an acute overdose generally occurs when the victim is found alive and treated by medical personnel prior to death.
However, the conclusion of death due to overdose may be more inferential. The victim may be found dead. There may be no suicide note and no one who observed the death. There may be a pill bottle but the number of pills gone and the reason they are missing may not be clear. Pill fragments may not be found in the stomach or the quantity may be low. However, there may be no other explanation for death except drug intoxication (e.g., no other cause found at autopsy). Typically, there is a search for a motive or risk for suicide, which might be recent or chronic significant life stressors or a history of having suffered from a depressive episode. The latter may be inferred from the fact that the victim is on an antidepressant. The pivotal finding is often the presence of a high concentration of a toxic drug in bodily fluids and/or tissue.
However, such a finding, and even the death itself, can be the result of a gradual accumulation of lethal concentrations of the toxic drug as a result of a pharmacokinetically mediated DDI, as happened in the case reported by Preskorn and Baker, in which the conclusion of death due to overdose was only inferred.4 This situation can occur when a substantial metabolic inhibitor is co-administered with a drug with a narrow therapeutic index whose clearance is dependent on the inhibited pathway for elimination. In such a case, concentrations of the latter drug may gradually build up until the patient suffers the fatal event, which may occur without warning (e.g., an arrhythmia due to a TCA). The body is then discovered some time after death. The autopsy may find no other cause for death beyond the high levels of the toxic drug. If the metabolic inhibitor or the inhibited drug is an antidepressant, that fact provides the motive or the risk for suicide.
A DDI may thus either contribute to a drug overdose death, by potentiating the effects of the ingestion, or may cause a false positive diagnosis of death due to overdose. It is from this perspective that this column will discuss cases of apparent acute overdoses of fluvoxamine reported in a publication by Garnier et al.3
Fluvoxamine was chosen for several reasons. First, a published article reviewing an extensive series of apparent overdoses with this antidepressant was available. Second, because fluvoxamine is an antidepressant, it may be used as a marker for suicide risk. Third, under usual dosing conditions fluvoxamine inhibits several cytochrome P450 (CYP) enzymes (1A2, 2C19, and 3A3/4) to a substantial degree5,6 and can thus be a perpetuator of a pharmacokinetically mediated DDI. Specifically, fluvoxamine can increase the amount of a co-ingested drug reaching the systemic circulation if that co-ingested drug normally undergoes first pass metabolism via these CYP enzymes during its absorption phase. The inhibition of the same enzyme can also delay the subsequent clearance of the drug. Both of these effects can increase the degree and the duration of the toxic effect of the co-ingested drug.
While fluvoxamine is the focus of this column, the same issues can be raised about any other drug capable of substantially inhibiting the clearance of a co-prescribed drug. This is particularly true if either the inhibitor or the drug that is inhibited is an antidepressant, because the use of such an agent raises the concern that the patient was at risk for a suicide attempt. Among marketed antidepressants, bupropion, fluoxetine, fluvoxamine, nefazodone, and paroxetine all cause substantial inhibition of one or more drug metabolizing enzymes under usual dosing conditions. TCAs, on the other hand, have a narrow therapeutic index and hence can be dangerous if the ingested dose is high relative to the individual's ability to clear it, whether that scenario occurs as a result of an acute ingestion or the gradual buildup of toxic levels due to inhibition of clearance.
The Article by Garnier et al.
In 1993, Garnier et al.3 reviewed 299 cases of acute overdoses of fluvoxamine; 221 of these cases of deliberate acute fluvoxamine overdoses came from reports to the Paris Poison Centre and 78 cases came from reports collected by the International Safety Department of Duphar BV, the manufacturer of fluvoxamine at the time of the publication.
The focus of the article was the safety of fluvoxamine when taken in an acute overdose. Garnier et al. pointed out that the acute toxicity attributable to fluvoxamine alone was rarely severe. At doses below 1,000 mg, symptoms were benign. At doses above 1,500 mg, there were only a few instances of seizures, but these were easily manageable and did not cause significant patient morbidity.
However, the authors also noted that fluvoxamine was taken alone in only 20% of the cases. In the vast majority of cases, the patient ingested multiple other drugs in addition to the fluvoxamine. Multiple drug ingestion in acute overdoses is not unique to fluvoxamine, but rather is the norm. However, that fact raises the possibility that a DDI could have contributed to the toxicity. Yet Gardiner et al. did not discuss this possibility in their article, perhaps due to the fact that CYP enzyme-mediated DDIs were just beginning to be understood at that time. This column will therefore review these cases from this perspective and will suggest that such DDIs may well have contributed significantly to one or more of 13 fatalities reviewed in that case series.
The case material
The outcome was not known in 50 of the 299 cases. While the majority of the remaining 249 cases recovered without sequelae, 13 of the patients (5%) died. In all of these latter cases, fluvoxamine was taken in combination with other drugs, raising the possibility that a DDI could have occurred.
In many cases, the bases for the conclusion that the victim committed suicide are not given in the Garnier et al paper. However, the authors did list all of the coingested drugs taken with fluvoxamine in these fatal cases. In each of the 13 cases, the patient co-ingested at least one drug whose metabolism is dependent on a CYP enzyme that is normally substantially inhibited by fluvoxamine under usual dosing conditions, and hence a DDI was likely to have occurred in all of these cases. However, such a DDI may have been an incidental rather than a critical element in terms of the fatal outcome. Thus, a more in-depth analysis of each case is required.
Five of the 13 fatal cases involved complicated overdoses of multiple drugs from multiple different classes. These cases were so complicated that it would be difficult to speculate on how much of a role a CYP-enzyme-mediated DDI might have played in the fatal outcome.
Another 4 of these cases involved the co-ingestion of a tertiary amine TCA. In 2 of these cases, the TCA (amitriptyline and clomipramine, respectively) was the only drug known to have been co-ingested with fluvoxamine. In the third case, the patient ingested fluvoxamine, dothiepin, and alcohol. In the fourth case, the patient ingested fluvoxamine, trimipramine, trihexyphenidyl, benzodiazepines, and alcohol. TCAs, of course, can be fatal if taken alone in sufficiently large quantities. In 1993, Härtter et al. reported that fluvoxamine markedly interfered with the metabolic fate of tertiary amine TCAs.7 As discussed above, such inhibition is comparable to ingesting a higher dose of the TCA and then slowing its clearance. Thus, the co-ingestion of fluvoxamine in these 4 cases would have contributed to the risk of a fatal outcome. Nevertheless, the overdose of the TCA alone might have been sufficient to cause the patient's death.
|Table 1 - Four cases in which fluvoxamine-induced inhibition of the metabolism of a co-ingested drug is likely to have contributed substantially to the fatal outcome |
|Patient sex |
and age (years)
ingested dose of
|Female, 35 ||unknown ||astemizole, |
|Male, 36 ||3,000 mg ||propoxyphene, |
|Female, 71 ||2,600 mg ||thioridazine |
|Male, 80 ||unknown ||flunitrazepam |
| In all four of these cases, the authors of the report concluded either that fluvoxamine probably was not directly responsible for the death or that its role could not be evaluated. |
The final 4 cases in this series raise the strongest possibility that a pharmacokinetically mediated DDI between fluvoxamine and a co-ingested drug may have been a meaningful contributor to the fatal outcome. These 4 cases are summarized in Table 1. In three of the cases, the patient died before medical assistance arrived so that a number of the observations used to convincingly prove suicide were likely missing. The fourth patient, an 80-year-old male, came to medical attention when he became deeply comatose with respiratory pauses after a suspected overdose of fluvoxamine and flunitrazepam. He died 3 days later from sudden cardiorespiratory arrest as a result of a premature extubation.
There are a number of reasons to consider that fluvoxamine-induced inhibition of drug metabolism may have contributed to the fatal outcome in these cases. First, each of these cases involved relatively simple ingestion of only one or two drugs in addition to fluvoxamine. Second, fluvoxamine is a known inhibitor of the metabolism of one or more of the drugs co-ingested in each of these cases. Third, the drugs co-ingested either have a large therapeutic index and thus rarely cause fatality when taken in overdose (e.g., flunitrazepam) or may be toxic when taken alone but are not often given in sufficient quantities to be toxic if taken as an overdose (e.g., thioridazine). Each of these four cases is reviewed in greater detail below.
Case of the 35-year-old female who died after taking fluvoxamine, astemizole, and prochlorperazine
Astemizole has been removed from the market because of fatalities that occurred when it was taken as prescribed in combination with a drug capable of substantially inhibiting its metabolism.8 Prior to its removal, the package insert for astemizole contained a warning about QT prolongation and the risk of ventricular arrhythmias.9 The mechanism responsible for this effect was concentration-dependent blockade of cardiac K+ channels, leading to delayed ventricular repolarization.10 The package insert contraindicated the concomitant administration of astemizole with erythromycin, ketoconazole, itraconazole, and related drugs capable of inhibiting CYP 3A. That is one of the CYP enzymes that is substantially inhibited by fluvoxamine. The basis for this contraindication included both case reports of ventricular polymorphic arrhythmia (i.e., torsades de pointes) and human pharmacokinetic data indicating that drugs such as ketoconazole significantly inhibited the metabolism of astemizole.
The co-ingestion of prochlorperazine was a complicating variable in this case. Prochlorperazine is a phenothiazine derivative. Some members of this drug class (e.g., thioridazine discussed below) have concentrationdependent effects on slowing intracardiac conduction and hence can be cardiotoxic at sufficiently high concentrations. Many phenothiazines have complicated metabolisms that can be inhibited by other drugs such as fluvoxamine.6,11 In addition, some phenothiazines can inhibit the metabolism of other drugs. Nevertheless, prochlorperazine is an older drug and there is minimal, if any, information available about its metabolism, the effect of metabolic inhibitors on its levels, its effects on the levels of other drugs, or its cardiotoxicity. For all of these reasons, its role in this patient's death and whether it was involved in a DDI with either astemizole or fluvoxamine, and if so, what was affecting what, are uncertain.
This patient died before medical attention arrived so that details concerning the nature of the apparent overdose were not available.
Case of the 36 year-old male who died after taking fluvoxamine, nitrazepam, and propoxyphene
The corpse of this patient was found 3 weeks after the presumed date of the overdose. This case is intriguing because of the complexity of the potential DDI. First, benzodiazepines and narcotics can interact pharmacodynamically to cause respiratory depression. Second, there is the possibility of at least two pharmacokinetically mediated DDIs in this case.
While there are no reports concerning the effects of fluvoxamine on the metabolism of nitrazepam, erythromycin has been shown to decrease the metabolism of nitrazepam.12 The effect of erythromycin described in this report was modest (25%), but it involved the administration of only a single dose of nitrazepam, which would tend to underestimate the effect with chronic dosing and would certainly underestimate the effect in an overdose.
The fact that erythromycin inhibited the metabolism of nitrazepam suggests that CYP 3A, which metabolizes a number of benzodiazepines that are structurally related to nitrazepam, was involved. Fluvoxamine also inhibit this CYP enzyme.
Propoxyphene has also been shown to interfere significantly with the metabolism of alprazolam, reducing its clearance by one-third and increasing its half-life from 12 to 18 hours.13 Although the mechanism for this effect was not established, alprazolam is principally metabolized by CYP 3A. Propoxyphene has also been shown to inhibit the metabolism of carbamazepine, another CYP 3A substrate, leading to a caution against using propoxyphene and carbamazepine together.14 Thus this patient was prescribed two drugs, fluvoxamine and propoxyphene, likely to inhibit the metabolism of nitrazepam, leading to its increased accumulation. A gradual accumulation of nitrazepam in the presence of propoxyphene could cause the patient to go into a coma and experience pulmonary arrest and death.
It would also be useful to know whether propoxyphene and fluvoxamine have any effect on each other's metabolism. Unfortunately, both of these are older drugs and the mechanisms underlying their clearance are not well established.15,16
Case of the 71-year-old female who died after taking fluvoxamine and thioridazine
After almost 40 years on the market, thioridazine received a black box warning in 2000 concerning the risk of sudden unexpected death due to its potential to cause ventricular polymorphic arrhythmias.17 Thioridazine has been shown to cause a concentrationdependent blockade of the K+ rectifier current and thus lengthens cardiac repolarization.18 Thioridazine use was recently shown to be associated with a statistically significant increased risk of sudden unexplained death in routine clinical practice (adjusted odds ratio = 5.3, p < 0.01).19
There have been well documented cases of lifethreatening overdoses on thioridazine alone,20 so that this 71-year-old patient in the Garnier et al. article may have died from thioridazine alone. However, there has been a case of sudden death in routine clinical practice in a patient taking thioridazine and fluvoxamine in which an overdose was unlikely (personal communication, State of Connecticut coroner's office, 1995). Consistent with that case, fluvoxamine at a relatively low dose of 50 mg/day has been shown to cause a significant 3-fold increase in the levels of thioridazine, its primary metabolite, mesoridazine, and another metabolite, sulforidazine.21
Thus, a DDI between thioridazine and fluvoxamine may have amplified the effects of any overdose or the DDI may have caused a false positive diagnosis of an overdose. This patient was found dead and no pathological or toxicological analysis was done. Therefore, the basis for the conclusion of death due to an acute overdose cannot be determined from the Garnier et al paper.
Case of the 80-year-old male who died after taking fluvoxamine and flunitrazepam
This case involved the presumed overdose ingestion of two drugs that are usually benign even when taken in appreciable amounts. The death apparently resulted from a cardiopulmonary arrest while the patient was in the hospital. The death was attributed to premature extubation. However, this death occurred after 3 days of hospitalization, suggesting prolonged intoxication with levels of flunitrazepam sufficiently high to suppress respiratory drive.
Since the publication of the Garnier et al paper, CYP 3A3/4 and CYP 2C19 have been identified as the major drug metabolizing enzymes for flunitrazepam.22 Desmethylflunitrazepam, 3-hydroxyflunitrazepam, and 7-aminoflunitrazepam are the major metabolites of flunitrazepam in humans. CYP 3A and CYP 2C19 mediate 40% and 60% of desmethylflunitrazepam formation, respectively. CYP 3A mediates 80% of the formation of 3-hydroxyflunitrazepam. Fluvoxamine at usual doses substantially inhibits both CYP 2C19 and 3A.
The prolonged course and fatal outcome in this case is compatible with slow clearance of flunitrazepam caused by fluvoxamine-induced inhibition of its metabolism. The basis for concluding that this patient took an overdose is not given in the report. Conceivably, this patient's coma could have resulted from a gradual accumulation of flunitrazepam in his system due to fluvoxamine-induced inhibition of its clearance.
Rather than providing definitive answers, the goal of this column was to raise questions and increase awareness about the possible role of a pharmacokinetically mediated DDI when managing drug overdoses or evaluating whether a death occurred as a result of an acute overdose. In some cases, the death may have resulted from a pharmacokinetically mediated DDI which led to the gradual accumulation of high concentrations of the toxic drug due to inhibition of its metabolism and/or clearance.
|Table 2. Masked ways in which pharmacokinetically mediated DDIs can present* |
|Toxicity 23 |
|Loss of or inadequate efficacy 24,25 |
|Drug withdrawal 26 |
|Worsening of underlying disease 27,28 |
|Emergence of a new disease 29 |
|Acute overdose 30 |
(also see cases described in this column)
| In all four of these cases, the authors of the report concluded either that fluvoxamine probably was not directly responsible for the death or that its role could not be evaluated. |
This issue is important for psychiatrists because their patients are at increased risk for suicide and so they must take these issues into consideration when selecting treatment regimens. Psychiatrists are also likely, over the course of their professional careers, to have patients for whom suicide might be an expected cause of death but whose death might instead have resulted from a DDI.
Another goal of this column was to again point out that "not seeing" a DDI is not the equivalent of a DDI "not occurring." That appears to have been the case in the Garnier et al report, in which the question of the possible role of DDIs was not raised. The absence of that consideration likely reflects the period when the report was written, in terms of the extent of knowledge of pharmacokinetic DDIs then available and the modest attention given to the problem of this type of DDI at that time.
The central point of this column is that, even when a DDI causes death, the outcome may be recognized but the cause not understood. This column thus provides another example of how DDIs may be missed in clinical practice because they can present in a myriad of masked ways, including as death due to an acute overdose (i.e., suicide) (Table 2). In subsequent columns, more definitive examples of DDIs presenting as acute overdoses will be described.
- Preskorn SH, The polypharmacology section. (collection of columns published in the Journal of Psychiatric Practice). Accessed July 10, 2002
- Neuvonen PJ, Pohjola-Sintonen S, Tacke U, Vuori E, Five fatal cases of serotonin syndrome after moclobemide-citalopram or moclobemide-clomipramine overdoses. Lancet. 1993;342;1419. Letter
- Garnier R, Azoyan P, Chataigner D, Taboulet P, Dellattre D, Efthymiou ML, Acute fluvoxamine poisoning. J Intern Med Res. 1993;21:197-208
- Preskorn SH, Baker B, Fatality associated with combined fluoxetine-amitriptyline therapy. JAMA. 1997;277:1682. Letter
- Preskorn SH, Clinically relevant pharmacology of selective serotonin reuptake inhibitors. An overview with emphasis on pharmacokinetics and effects on oxidative drug metabolism. Clin Pharmacokinet 1997; 32:1-21
- Sproule BA, Naranjo CA, Brenmer KE, et al, Selective serotonin reuptake inhibitors and CNS drug interactions: A critical review of the evidence. Clin Pharmacokinet 1997;33: 454-471
- Hartter S, Wetzel H, Hammes E, Hiemke C, Inhibition of antidepressant demethylation and hydroxylation by fluvoxamine in depressed patients. Psychopharmacol. 1993; 110:302-308
- Preskorn SH, Drug approvals and withdrawals over the last 60 years. Journal of Psychiatric Practice 2002;7:410
- Physicians Desk Reference, 53rd edition. Montvale, NJ: Medical Economics Company, 1999
- DuBuske LM, Second-generation antihistamines: The risk of ventricular arrhythmias. Clin Ther 1999;21:281-95
- Devane CL, Markowitz JS, Antipsychotics. In: Levy R, Thummel K, Trager W et al. (eds) Metabolic drug interactions. Philadelphia: Lippincott William & Wilkins 2000:245-59
- Luurila H, Olkkola KT, Neuvonen PJ, Interaction between erythromycin and nitrazepam in healthy volunteers. Pharmacol Toxicol 1995;76:255-8
- Abernethy DR, Greenblatt DJ, Morse DS, et al, Interaction of propoxyphene with diazepam, alprazolam and lorazepam. Br J Clin Pharmacol 1985;19:51-7
- Bergendal L, Friberg A, Schaffrath AM, et al, The clinical relevance of the interaction between carbamazepine and dextropropoxyphene in elderly patients in Gothenburg, Sweden. Eur J Clin Pharmacol 1997;53:203-6
- Kharasch ED, Opioid analgesics. In: Levy R, Thummel K, Trager W, et al. (eds) Metabolic drug interactions. Philadelphia: Lippincott William & Wilkins 2000:297-319
- Chiba K, Kobayashi K, Antidepressants. In: Levy R, Thummel K, Trager R; et al. (eds) Metabolic drug interactions. Philadelphia: Lippincott William & Wilkins 2000:233-243
- FDA website. Accessed July 9, 2002
- Drolet B, Vincent F, Rail J, et al, Thioridazine lengthens repolarization of cardiac ventricular myocytes by blocking the delayed rectifier potassium current. J Pharmacol Exp Ther 1999;288:1261
- Reilly JG, Ayis SA, Ferrier IN, et al, Thioridazine and sudden unexplained death in psychiatric in-patients. Br J Psychiatry 2002;180:515-22
- Schmidt W, Lang K, Life-threatening dysrhythmias in severe thioridazine poisoning treated with physotigmine and transient atrial pacing. Crit Care Med 1997;25:19250
- Carrillo JA, Ramos SI, Herraiz AG, et al, Pharmacokinetic interaction of fluvoxamine and thiroidazine in schizophrenic patients. J Clin Psychopharmacol. 1999;19:494-9
- Kilicarslan T, Li N-Y, Busto U, et al, Flunitrazepam: Polymorphic metabolism by CYP2C19 (abstract). Clin Pharmacol Ther 1998;63:218
- Preskorn SH, I dont see em. J Prac Psych Behav Hlth. 1997; 3:302-307
- Preskorn SH, Case 3: A 44-year old man at risk for seizures. In Outpatient management of depression. Caddo, OK Professional Communications; 1999:188-9
- Preskorn SH, Silkey B, Multiple medications, multiple considerations. Journal of Psychiatric Practice 2001;7:48-52
- Preskorn SH, Case 4: The 28-year-old man with abdominal cramps, chills, and rummy nose. In Outpatient management of depression. Caddo, OK Professional Communications; 1999:189-90
- Preskorn SH, A message from Titanic. J Prac Psych Behav Hlth. 1998;4:236-242
- Preskorn SH, Do you believe in magic? J Pract Psychiatry Behav Health 1997;3:99-103
- Preskorn SH, Case 2: Severe parkinsonism and confusion in a 78-year-old woman. In Outpatient management of depression. Caddo, OK Professional Communications; 1999:186-8
- Preskorn SH, Case 1. Sudden death in a 36-year-old man. In Outpatient management of depression. Caddo, OK Professional Communications; 1999:184-6