Prolonged neuropsychiatric effects following management of chloroquine intoxication with psychotropic polypharmacy.

Susceptibility to quinoline antimalarial intoxication may reflect individual genetic and drug-induced variation in neuropharmacokinetics. In this report, we describe a case of chloroquine intoxication that appeared to be prolonged by subsequent use of multiple psychotropic medications. This case highlights important new considerations for the management of quinoline antimalarial intoxication.

Introduction

Currently available synthetic quinoline antimalarials, including chloroquine, exhibit idiosyncratic neuropsychiatric effects when administered at low doses used for chemoprophylaxis. Idiosyncratic cases of acute intoxication from quinoline antimalarials have been well characterized in the literature since the first widespread use of the now-outmoded synthetic quinoline quinacrine (also known as atabrine) in the mid-1940s 1,2. With subsequent widespread use of chloroquine, a drug developed as a less toxic replacement for quinacrine 3, similar idiosyncratic cases of intoxication 4–6 were soon reported. More recently, similar intoxicating effects from mefloquine, itself developed as a replacement for chloroquine 7, have become well characterized in the literature 8. Although the naturally occurring quinoline antimalarials have traditionally been considered free of such effects, cases of a related syndrome of intoxication have also occasionally been reported after relatively low doses of quinine 9,10 and quinidine 11,12, suggesting a common intoxicating class effect.

Descriptions of chloroquine intoxication in the literature, commonly characterized as toxic psychosis, are similar to those associated with other quinolines and typically feature insomnia 13,14, mania 15,16, paranoia and persecutory delusions 17, and auditory and visual hallucinations 18. Symptoms of intoxication may resemble those of brief psychotic disorder but are more likely to include prominent visual hallucinations, anxiety, restlessness, agitation, and derealization 19. Suicidality is not uncommon reported 20, and reports of completed suicide 21,22 attest to the potentially life-threatening nature of such intoxication. Other symptoms of intoxication may include impulsivity, over-talkativeness with flight of ideas, unprovoked laughing or crying 20, personality change, grandiosity, depersonalization 6, and delusional misidentification 23.

Occasionally, intoxication may also induce a delirium 24 or encephalopathy whose symptoms may include confusion 25, difficulties with concentration 26, memory loss and amnesia 27, and catatonia 28.

Although theories have been proposed to explain quinoline intoxication 29,30, the precise pathophysiological basis of these effects remain unclear despite numerous molecular mechanisms having been identified experimentally. Antimalarial quinolines typically cross the blood–brain barrier to concentrate within the central nervous system (CNS) 31, where many have been demonstrated to directly affect chemical synaptic communication 32, including at serotonin 32,33, adenosine 34, gamma-amino butyric acid (GABA) 35, and opioid 36 receptors. Various quinolines have also been identified as potent inhibitors of interneuronal electrical communication 37 and are suspected of disrupting transmission between networks of GABA inhibitory interneurons 38,39. Such drug-induced inhibitory dysfunction may be sufficient to induce a toxic encephalopathy that mimics in particular the limbic effects of anti-N-methyl-D-aspartate receptor (NMDA) receptor hypofunction 40.

As with other causes of intoxication delirium 41, which may induce neurological abnormalities 24, chloroquine intoxication has been associated with certain extrapyramidal-like movement disorders including stereotypical repetitive movements 42, ataxia 43, facial dystonia and torticollis 44,45, dysphonias 45, oculogyric dystonia 46, and other neurological effects 47 including tinnitus, vestibulopathy, and accommodative dysfunction 48. While many symptoms of quinoline intoxication delirium are seemingly reversible with elimination of the drug, use of various quinolines has also been associated with the development of chronic neurological disorders including temporal lobe epilepsy 49, dysautonomia 50, and central vestibulopathy 40. In addition to reflecting the direct sequelae of encephalopathy, drug-induced brain and brainstem neurotoxicity may underlie at least some prolonged neuropsychiatric effects 51. For example, in animal models many drugs of the class induce highly focal neurotoxic lesions throughout multiple brain and brainstem areas including the medulla, pons, striatum, and limbic system 51. Chloroquine may also induce histopathological alterations within multiple neuron subtypes in CNS including within the brainstem and limbic system 52,53. In this regard, similarities in the chronic neuropsychiatric sequelae from the antimalarial quinolines shared to a degree with the effects of even the minimally substituted antiparasitic quinoline clioquinol 49,54 may suggest either neurotoxicity from an as-yet uncharacterized common metabolite, or a direct dose-dependent neurotoxic effect shared among drugs of the class.

Similarly, as neuropsychiatric effects that occur with low doses of quinoline antimalarials during chemoprophylaxis mimic effects more reliably produced at higher doses or dose rates with use in treatment, susceptibility to intoxication and subsequent neurotoxicity during chemoprophylaxis may reflect multifactorial genetic and drug-induced variations in neuropharmacokinetics which may serve to locally increase such metabolite or drug concentrations in brain 51 despite normal serum levels.

In this report, we describe a case of quinoline intoxication that developed while taking low dose chloroquine as chemoprophylaxis that was further exacerbated by the addition of various psychotropic agents used to manage neurobehavioral symptoms associated with the acute intoxication. Pharmacogenetic testing was consistent with absent enzymatic activity of cytochrome P450 (CYP) 2D6, as well as potentially altered activities of CYP1A1, CYP1A2, CYP2C8, CYP2C9, CYP2C19, and P-glycoprotein (P-gp, or ABCB1). In this report, the possible role of these polymorphisms in contributing to susceptibility to chloroquine intoxication, as well as in prolonging its neuropsychiatric sequelae in the context of psychotropic polypharmacy are discussed together with new considerations for appropriate clinical management of this condition.

Case History

While volunteering in Honduras, a 16-year-old female with an unremarkable medical history became ill during her ninth week of chloroquine prophylaxis (250 mg weekly). She had begun malaria prophylaxis 2 weeks prior to travel, and by her seventh week in Honduras had developed insomnia, paranoia, and confusion that progressed in severity over the next 2 days. Following a dose of diphenhydramine to aid sleep, her symptoms progressed to include hallucinations, irrational guilt and self-persecution, suicidal ideation, and catatonic features. She was taken to a local neurologist where initial workup revealed a normal head CT, unremarkable lumbar puncture, normal complete blood count, negative infectious disease studies, and negative urine toxicology screen.

Upon repatriation to the United States 2 days later, she was admitted to the emergency room of a local hospital where olanzapine and lorazepam were administered for acute management of behavioral symptoms. Upon hospital admission, risperidone and fluoxetine were initiated with lorazepam continued on an as-needed basis for anxiety. On the second day, benztropine was added while lorazepam was replaced with clonazepam. Her mental status improved briefly by the evening of the second hospital day and she was able to converse with her family and recall many of her experiences over the prior week. However, by the morning of the third day, she had slipped back into a state of marked confusion, self-persecution, paranoia, suicidal ideation, visual and auditory hallucinations, delusional thinking, and catatonia. These progressed in a waxing and waning manner over the next 14 days.

Differential Diagnosis, Investigations, and Treatment

Although chloroquine intoxication was considered in the differential diagnosis, owing to the severity of symptoms and their similarity to those of NMDA receptor (NMDAR) encephalitis 55, cerebrospinal fluid (CSF) collected in Honduras was sent for NMDAR antibody titers, but these were found to be negative. Two additional lumbar punctures were performed over the course of the first 17 hospital days, but these repeat NMDAR antibody titers were also negative. As an additional precaution, an abdominal ultrasound was performed which ruled out ovarian tumors often found in NMDAR encephalitis 56. CSF samples from each of these collections were also sent to the California Encephalitis Project 57 for additional viral studies but were also negative. A head MRI was normal. A video-EEG was attempted but the patient would not cooperate. Beta-human chorionic gonadotropin (β-HCG) was negative, serum copper was normal, and serum porphyrins were low. Urine toxicology, porphyrins, and heavy metals were all normal. Infectious disease studies revealed a positive serum IgM but negative IgG for Brucella and a positive serum IgM for Mycoplasma pneumoniae, for which she was treated with a standard dose of azithromycin daily for 5 days. Rheumatologic disease studies were remarkable for a mild elevation in erythrocyte sedimentation rate and thyroglobulin antibody, but normal C-reactive protein. A neurologic exam on hospital day 17 revealed extreme mydriasis, amaurosis, patellar areflexia, dysphonia, and facial palsies.

Despite worsening neurological findings and mental status, the absence of evidence of a treatable organic cause and the need to manage continued symptoms led to transfer to a psychiatric hospital 3 days later. At the time of transfer, she exhibited near-catatonic features. The following day, risperidone, benztropine, and fluoxetine were replaced with olanzapine and lorazepam. Over the next 3 days, she improved remarkably. However, after the fourth day she progressively declined with increasing visual illusions, extreme paranoia with magical and delusional thoughts (e.g., each of the staff were interpreted as celebrities, and she believed a bird outside the window was videotaping her and that she was imprisoned in a foreign country), suicidal ideation and anxiety. She demonstrated a return of facial palsies and abnormal phonation as her dose of olanzapine was increased from 2.5 mg to 17.5 mg daily. The correlation of medication dosage with symptom severity raised suspicion for an exacerbation of toxicity from psychotropic polypharmacy. By hospital day 58, olanzapine was tapered off, after which she rapidly improved and was discharged on hospital day 66 on a single tapering dose of lorazepam only. Outpatient neuro-ophthalmologic exam revealed findings of extreme mydriasis, despite the patient not having been on an anticholinergic medication for over 45 days. This led to suspicion that abnormal medication metabolism was contributing to central anticholinergic toxicity.

Pharmacogenetic testing was then obtained using the Drug Metabolizing Enzymes and Transporters (DMET) DNA Microarray GeneChip® 58, with selected polymorphisms confirmed by direct analysis at the Mayo Clinic Laboratories. Results (Table1) were consistent with absent CYP2D6 activity 59, and potentially altered CYP1A1, CYP1A2, CYP2C8, CYP2C9, CYP2C19, and ABCB1 activities. Samples of urine, serum, and CSF that were stored over the course of hospitalization at the California Encephalitis Project were later retrieved and analyzed for the presence of chloroquine and its metabolite. Their presence was confirmed and concentrations of chloroquine in serum were found to be within the expected range 60.

Table 1

Selected Positive Cytochrome P450 (CYP) and ATP-binding cassette (ABC) Single Nucleotide Polymorphisms (SNP)

Gene	SNP	Reference SNP (rs) ID	Observed Alleles	Notes
ABCB1	-25G>T	rs2235015	T/T	1
	24G>A	rs2235040	A/A	1
CYP1A1	2452C>A	rs1799814	C/A	1
CYP1A2	-163C>A	rs762551	A/A	1,2
CYP2C8	30411A>G	rs10509681	A/G	1
CYP2C9	430C>T	rs1799853	C/T	2
CYP2C19	2608C>T	rs10509681	C/T	1
	-806C>T	rs12248560	C/T	1,2
CYP2D6	1846G>A	rs3892097	A/A	1,2
	100C>T	rs1065852	T/T	1,2
	1661G>C	rs1058164	C/C	1
	4180G>C	rs1135840	C/C	1

1. From DMET DNA Microarray GeneChip®.

2. From direct polymorphism analysis.

Outcome and Follow-up

The patient convalesced at home off all medications over the next 4 months. She underwent extensive neuropsychological testing but no localizing deficits could be identified. She then returned to high school where she was noted by family and friends to have experienced further behavioral changes (e.g., baseline introversion gave way to extroversion with excessive use of cellphone and social media and attention-seeking behavior), attention deficit-like symptoms (e.g., diminished attention, concentration) and uncharacteristic learning deficits. During her final exams, she decompensated in apparent exhaustion. Thereafter, she began to demonstrate increasingly bizarre behavior including: periods of catatonic-like features (e.g., often sitting for hours and staring at herself in the mirror), punctuated by hysterical fits of inappropriate laughter. She exhibited ritualistic semiconscious behavior that seemed to appear each evening, with associated insomnia. She reported vague but intense pains in the throat and genitalia. Workups aimed at identifying organic causes of these pains were negative. Attempts at providing an ideal environment for her to rest and sleep failed as her insomnia and declining behavior progressed.

The patient's prolonged neuropsychiatric symptoms were suspected to represent sequelae of chloroquine intoxication encephalopathy. Specialists in quinoline toxicity and psychopharmacology were consulted, and in addition to limbic encephalitis, suspicion of quinoline-induced limbic 49, and brainstem 8 neurotoxicity with possible associated epilepsy or nonconvulsive status epilepticus 61 and neuralgias were also raised. Certain organic causes were again ruled out following a repeat MRI and lumbar puncture that were normal, and CSF sent for a broad panel of midbrain receptor autoantibody titers returning negative. A 48-hour, continuous video-EEG monitoring was also negative for cortical seizures.

Guided by the newly obtained pharmacogenetic testing, a new medication regimen directed at managing her neuropsychiatric symptoms and central cholinergic hypoactivity, and to empirically manage a possible occult seizure disorder was designed. To these effects, valproic acid, paliperidone, and rivastigmine were sequentially initiated. On the third day of valproic acid, an abrupt improvement in mental status occurred whereby the bizarre behavior subsided and the patient was able to engage in detailed conversations on a wide range of topics. The sequential additions of paliperidone followed by transdermal rivastigmine appeared to add small incremental improvements in her cognition, mental status, and behavior. None of these medications initially appeared to lead to any adverse effects.

After a few weeks, the patient's behavior had significantly improved and she was weaned off paliperidone. After a few months, she was weaned off valproic acid while continuing on transdermal rivastigmine only. She convalesced at home for 4 months and returned to finish high school without any behavioral issues but with emerging symptoms of extreme fatiguability, orthostatic intolerance, and attentional deficit all suggestive of the autonomic dysfunction following brain injury 62. Concerns for rivastigmine contributing to these lingering symptoms led to its substitution with a subsequent gradual taper of an oral formulation that continued through the end of reported follow-up, 36 months after the initial intoxication.

Discussion

While this patient's initial symptoms were remarkably consistent with those in other reports of quinoline intoxication and specifically in prior reports of chloroquine intoxication 6,14,15,17,20,25,42,63, the extended nature of certain neuropsychiatric effects has not previously been reported with this drug. In this case, pharmacogenetic testing revealed evidence of absent CYP2D6 activity, alterations in the activity of at least five other enzymes including CYP1A1, CYP1A2, CYP2C8, CYP2C9, and CYP2C19 involved in the metabolism either of chloroquine or the psychotropic drugs used to manage her symptoms, and evidence of altered P-gp activity. While these pharmacogenetic findings do not provide definitive insight into the etiology of the patient's original intoxication, they do suggest that it may have been a direct effect of an inability to metabolize chloroquine. Likewise, these pharmacogenetic findings may provide pharmacokinetic explanations for the extended duration of at least some of her symptoms. The plausible pharmacogenetic correlates in this case therefore raise important new considerations for the management of quinoline antimalarial intoxication.

Etiology

In this case, the patient exhibited characteristic symptoms of quinoline intoxication, including insomnia, paranoia, and confusion well prior to use of any other psychotropic drugs. This suggests that the etiology of this intoxication could reflect a primary genetic susceptibility to these effects. For example, genetic susceptibility to mefloquine intoxication has been postulated to primarily reflect polymorphisms in ABCB1 64,65. However, although chloroquine is a potent P-gp inhibitor 66, there is yet little direct evidence that the drug is a significant P-gp substrate in vitro 66. Therefore, whatever alteration in activity might be associated with the homozygous variation in ABCB1 (rs2235015 and rs2235040) observed in this case may not have directly contributed to risk of initial intoxication. Conversely, among other enzymes with polymorphisms suggesting altered activity, CYP1A1, CYP1A2, CYP2C8, and CYP2C19 are all believed to be involved in chloroquine metabolism 67,68. Our patient was found to have heterozygous variation in CYP1A1 (rs1799814), CYP2C8 (rs10509681), and CYP2C19 (rs12248560), and homozygous variation in CYP1A2 (rs762551), but the clinical significance of these in the context of chloroquine metabolism is not clear based on existing studies.

The only enzyme in which a well-characterized alteration in activity was identified in our patient was CYP2D6. Although prior reports have suggested that chloroquine is not significantly metabolized in vitro by CYP2D6 67, it is plausible that brain-specific variations in CYP2D6 expression might play a relevant role in the etiology of chloroquine intoxication in vivo. Of the major isoforms thus far found responsible for chloroquine metabolism, only CYP2D6 is known to be relatively more expressed than these isoforms in brain relative to liver 69. Interestingly in this respect, the proportion of subjects who are found to lack CYP2D6 activity 70, typically 10% or fewer, is similar to the proportion of subjects in studies of chloroquine prophylaxis who report prodromal psychiatric symptoms such as depression, irritability, and confusion 71, which may predict risk of more serious intoxication with continued dosing.

In this case, it may be hypothesized that if brain-specific expression of chloroquine-metabolizing CYP isoforms is necessary to mitigate the risk of intoxication, the absence of CYP2D6 activity might in part explain the observed susceptibility. According to this hypothesis, the addition of fluoxetine, a potent CYP2D6 inhibitor 72, administered during the first 20 days of hospitalization, may have potentially further reduced whatever minimal chloroquine metabolic activity may have been initially present in brain tissue.

Whether or not the hypothesis is correct that the absence of CYP2D6 mediated metabolism of chloroquine in brain may have been primarily responsible in this case for the initial intoxication and its extended duration, it is tempting to speculate that the absence of CYP2D6 activity may also have contributed through other mechanisms to the extended duration of illness. CYP2D6 provides a major pathway for the metabolic inactivation of many of the drugs to which the patient was subsequently exposed, including diphenhydramine 73 used to manage initial symptoms of insomnia, as well as the risperidone 74, fluoxetine 75, and benztropine 76 employed during subsequent hospitalization. In the absence of metabolic inactivation, their potential accumulation to toxic levels may have further complicated clinical presentation. For example, among those with absent CYP2D6 activity and treated with risperidone, a case report has described prolonged neurologic side effects including extrapyramidal movement disorders 74, whereas among those treated with fluoxetine, another case report has described prolonged neurologic side effects including incoordination, ataxia, and seizure 75. Certain of these side effects might have readily confounded the presentation of chloroquine intoxication had they indeed occurred in our patient.

Clinical application

While pharmacogenetic testing allowed the treatment team to recognize the potential contribution of polypharmacy to the patient's original intoxication, this may have remained of little more than academic interest had it not been for the latent relapse necessitating further treatment. Once symptoms returned, an understanding of the potential contributions of pharmacogenetics to the etiology of her illness aided the treatment team in selecting appropriate therapy.

Antiepileptic agents have been successfully employed in the management of multiple neurological sequelae of brainstem 77 and limbic encephalitis, which may serve as a reasonable pathophysiological model for the brain and brainstem injury 8,51 that may follow quinoline intoxication encephalopathy 40. With knowledge of the possibility of a nonconvulsive status as causative or contributory to the waxing and waning presentation, valproic acid was chosen to manage symptoms demonstrated by the patient. Other reasons for its selection included its neuroprotective properties 78,79, and its efficacy in managing agitation and psychotic symptoms associated with acute confusional states 80,81. Whether or not this presumed etiology was correct, the patient seemed to have a dramatic response in mental status by the third dose of valproic acid. Unlike certain other antiepileptic agents, valproic acid is not significantly metabolized by CYP2D6 82, making it a reasonable choice given the absent activity of this enzyme.

Additionally, paliperidone was also selected to manage the patient's psychotic symptomatology, given its lack of associated CYP metabolism, absence of anticholinergic activity, and minimal sedation and extrapyramidal symptoms 83. Finally, given evidence suggestive of a dysautonomia and central anticholinergic state, rivastigmine 84 was selected in an attempt to enhance central cholinergic activity.

In addition to informing management of individual future cases of intoxication, this case also highlights the important emerging role of pharmacogenetic testing in the routine prescribing of antimalarial quinolines. Recently, metabolism by CYP2D6 has been shown to be necessary in ensuring reliable antimalarial effects from the related antimalarial primaquine 85. As standard pharmacogenetic testing becomes ever more widely available, the assessment of CYP2D6 activity may soon help to inform consideration of alternatives to primaquine in the chemoprophylaxis of malaria disease 86. However, should the increased risk of neuropsychiatric toxicity among those with reduced CYP2D6 activity hypothesized in this case be found to be shared among other quinolines, this would preclude increasing the dosing of these and related drugs 87 as a safe strategy for overcoming reduced antimalarial effects.

Conclusions

This case suggests important new considerations for the appropriate clinical management of quinoline antimalarial intoxication. Given the generally self-limiting nature of most acute symptoms of quinoline intoxication, clinicians managing future cases may wish to avoid psychotropic polypharmacy. This may potentially avoid prolonging the duration of acute symptoms and may reduce the risk of chronic neuropsychiatric sequelae; both of which may be plausibly linked to individual genetic and drug-induced variation in neuropharmacokinetics.

This case also points to a need for quinoline antimalarial drug development, including prelicensing trials assessing safety and efficacy, to be informed by relevant pharmacogenetic testing among enrolled subjects.

Lastly, this case also emphasizes the critical need for those reporting neuropsychiatric adverse events from antimalarial quinolines to include, wherever possible, the results of standard pharmacogenetic testing panels that include the enzymatic pathways plausibly involved in the neuropharmacokinetics of these drugs.