Article Text
PDF
Abstract
We describe a 64-year-old woman with relapsing encephalopathy. She initially presented with 5 days of psychomotor agitation, progressing to mania, psychosis and seizures that mimicked autoimmune limbic encephalitis. During her first hospital admission, extensive investigation failed to establish the underlying cause, and she improved with antiseizure medication alone. After a month at home, she relapsed with identical symptoms, and only then did we recognise that both episodes had been provoked by clarithromycin, prescribed for Helicobacter pylori eradication. Clarithromycin-induced neurotoxicity is rarely reported but likely to be under-recognised. It usually manifests within days of starting treatment, with delirium, mania, psychosis or visual hallucinations, sometimes termed ‘antibiomania’. Seizures and status epilepticus appear to be less frequent. A full recovery is expected on stopping the medication.
- NEUROTOXICOLOGY
- EEG
- NEUROPSYCHIATRY
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Statistics from Altmetric.com
Manuscript
A 64-year-old right-handed woman presented to the emergency department with a 5-day history of acute behavioural change. After listening to a BBC Radio 4 interview about childhood experiences of foster care, she had started reflecting aloud on her own childhood in care (not something that she usually discussed). She became increasingly animated and irritable. On day 5, she developed recurrent, brief periods of unresponsiveness, followed by intense agitation. There was no prior history of psychiatric illness, and no foreign travel, alcohol or drug misuse. She was married and had retired from a career in finance. She took famotidine for gastro-oesophageal reflux disease and occasional simple analgesics for osteoarthritis.
The frequent unresponsive spells continued during the initial days of admission, comprising 10–30 s of abrupt behavioural arrest without motor activity. Between events, she was alert and broadly orientated, but with marked psychomotor agitation, perseverative speech and variable attention. There were no signs of systemic illness or of infection, and no focal neurological deficits. Her initial electroencephalogram (EEG) (figure 1) showed intermittent runs of synchronised bilateral sharp-and-slow-wave complexes on a disorganised, slow background, indicating severe generalised encephalopathy with an increased susceptibility to seizures.
EEG in longitudinal bipolar montage on day 4 of admission, while she was alert and responsive. There is a disorganised slow background of mixed delta and theta activity, plus intermittent runs of bilateral synchronous sharp-and-slow-wave complexes, most prominent in the centroparietal and occipital regions (circled). EEG, electroencephalogram.
At this point, considering the acute onset, short psychiatric prodrome and symmetric EEG abnormalities, we suspected either an autoimmune or paraneoplastic limbic encephalitis or a toxic–metabolic encephalopathy. However, investigation along these lines was unrevealing: her blood tests on admission were normal, including full blood count, renal, liver, thyroid and bone profiles, C reactive protein, magnesium, ammonia and immunoglobulin concentrations, HIV and treponemal serology. Her MR brain scan and routine cerebrospinal fluid (CSF) studies were normal (cell count, protein, glucose ratio, oligoclonal bands and PCR panel for viral pathogens). Whole-body CT scan found no evidence of malignancy. In-house screening for antibodies to extractable nuclear antigens, antineutrophil cytoplasmic antibody, antidouble-stranded DNA antibody, antiglutamic acid decarboxylase antibody and paraneoplastic intraneuronal antibodies was negative. Using cell-based assays by the Oxford Neuroimmunology Group, serum tested negative for antibodies for neuronal surface antigens (LGI1, CASPR2, NMDAR, AMPA-1, AMPA-2, GABAA, GABAB and glycine receptor), and a small sample of CSF tested negative for NMDAR antibodies.
Following the abnormal EEG on day 4 of admission, we prescribed levetiracetam. The next day the suspected seizure activity stopped completely and she was less agitated. However, she remained in a hypomanic state that gradually improved over 2 weeks. Repeat EEG on day 18 (figure 2) was normal. She was discharged home on day 30.
Repeat EEG on day 18 shows return of a well-organised alpha rhythm posteriorly with normal faster activity anteriorly. The epileptiform abnormalities have disappeared. EEG, electroencephalogram.
After 25 days at home, during which time her family felt that she was well, she returned to the emergency department with a very similar presentation. For 3 days, she had become excitable, argumentative and paranoid. She then developed abrupt vacant spells followed by intense agitation, suggesting focal seizures. Again, there were no signs of systemic illness or infection, no focal neurological deficits and her baseline blood tests were unremarkable.
When we enquired about potential triggers, her husband highlighted that she had started Helicobacter pylori eradication therapy 5 days before her first presentation, following a positive urease breath test. This medication had been stopped on hospital admission and had been disregarded. Her general practitioner had then represcribed it 6 days before her second presentation, after she had reported ongoing dyspepsia. H. pylori ‘triple therapy’ comprises a 14-day course of oral amoxicillin, clarithromycin and omeprazole. She had taken amoxicillin previously, but these were her first exposures to clarithromycin and omeprazole. She also had past exposure to lansoprazole, which had been stopped due to gastrointestinal upset, rather than neuropsychiatric effects.
We therefore considered the possibility of clarithromycin-induced neurotoxicity and stopped this medication, leaving the levetiracetam dose unchanged. The focal seizures ceased rapidly. Repeat MR brain scan on day 3 and EEG on day 5 were both normal. We again observed a gradual improvement in her psychotic and manic symptoms. We switched levetiracetam to lamotrigine for its mood-stabilising benefits and discharged her home on day 24.
Subsequently, she had two further emergency department presentations with suspected focal seizures, after 3 and 5 months. There had been no further clarithromycin use, and these seizures were not accompanied by a psychiatric prodrome or by prolonged behavioural disturbance. The first occurred in the context of hyponatraemia (serum sodium 119 mmol/L), attributed to syndrome of inappropriate antidiuretic hormone secretion. She was admitted briefly under general medicine for oral fluid restriction and observation. The second seizure was unprovoked, and she was discharged from the emergency department. Specialist neurology input was not sought on either occasion. She has otherwise remained well.
Discussion
We describe a case of relapsing encephalopathy with prominent psychiatric features and seizures that initially mimicked autoimmune limbic encephalitis. Because we were unfamiliar with the potential neurotoxic effects of H. pylori triple therapy, we had not considered neurotoxicity until the patient re-presented having received the medication a second time. Clarithromycin, a frequently prescribed macrolide antibiotic, was the likely culprit due to the patient’s previous exposure to amoxicillin and proton pump inhibitors, as well as the fact that her presentation was consistent with published literature on clarithromycin-induced neurotoxicity. Her abnormal EEG provided objective evidence that her presentation was due to a neurological disorder of cerebral function rather than a primary psychiatric disorder, and extensive investigation for alternative causes, particularly autoimmunity, was negative. The Naranjo criteria provide a scoring system for estimating the probability of adverse drug reactions.1 This patient had a Naranjo score of 10 (table 1); scores above 9 indicate definite drug reactions.
The Naranjo adverse drug reaction probability scale, adapted from Naranjo et al 1
The fact that this patient subsequently re-presented with seizures suggests a predisposition to focal seizures, unmasked by clarithromycin. However, during clarithromycin use, we observed more frequent seizures and interictal epileptiform activity on EEG, as well as the prodrome of psychiatric disturbance, which persisted for several days after the seizures had stopped.
Bhattacharyya et al comprehensively reviewed the literature on antibiotic-associated encephalopathy (AAE) in 2016 and classified the reactions into three types (table 2). They identified 44 reports attributed to clarithromycin, plus 10 reports for other macrolides.2 Onset was within days of starting clarithromycin and psychosis occurred in 59% of cases. No seizures were reported. Just 6/44 patients underwent EEG, of which 2 showed encephalopathic abnormalities and none showed epileptiform discharges. This ‘psychosis phenotype’ (observed with macrolides, fluoroquinolones and sulphonamides) was categorised as type 2 AAE. In contrast, type 1 AAE, seen with penicillin and cephalosporins, is characterised by frequent seizures, myoclonus and epileptiform EEG abnormalities. Type 3 AAE refers to metronidazole-induced cerebellar ataxia, which typically occurs after weeks of treatment, and is associated with characteristic MRI abnormalities.2
Classification of antibiotic-associated encephalopathy, adapted from Bhattacharyya et al 2
A more evocative term used in many case reports for clarithromycin-induced neurotoxicity is ‘antibiomania’, emphasising the prominent psychomotor agitation, mania and psychosis.3–5 Visual hallucinations may also occur, with one patient describing a ‘constantly evolving landscape of sharks, priests, red lines and other technicolour’.6 Seizures and epileptiform EEG abnormalities appear to be much less common but are reported, including non-convulsive status epilepticus.7 8 Regardless of presentation, clarithromycin-induced neurotoxicity has an excellent prognosis, with symptoms resolving completely in all cases by 30 days, but usually more rapidly.2
Clarithromycin-induced neurotoxicity affects men and women equally, and most have no prior psychiatric history, or significant renal or hepatic impairment.2 7 Most cases have occurred at daily doses of 1000 mg or less.7 Those taking other drugs metabolised by cytochrome P450 isoenzymes of the CYP3A family, of which clarithromycin is a potent inhibitor, may have a higher risk of neurotoxicity.7 Such drugs include carbamazepine, benzodiazepines, calcium channel blockers, statins, ciclosporin, tacrolimus, azole antifungals and some antiretroviral drugs. Neurologists should be particularly aware of the risk of carbamazepine toxicity from clarithromycin-mediated enzyme inhibition, which necessitates monitoring and dose adjustments.9 10 In this case, aside from triple therapy, the patient was taking only famotidine, which does not interact significantly with hepatic enzyme metabolism.11
A large self-controlled case series study in Hong Kong (n=66 559) examined the risk of presenting with a first neuropsychiatric event during outpatient treatment with clarithromycin-containing H. pylori triple therapy.12 The age-adjusted incident rate ratio (IRR) quadrupled (IRR 4.12, 95% CI 2.94 to 5.76) during the 14-day treatment period and rapidly fell back to baseline thereafter. The crude absolute risk of composite neuropsychiatric events during treatment was 0.45 per 1000 prescriptions, equating to one event for every 2222 prescriptions.12 These figures are striking. Clarithromycin is also widely prescribed for respiratory, sinus, ear and soft tissue infections, suggesting that neurotoxicity is probably under-recognised and under-reported.
The mechanism of clarithromycin-induced neurotoxicity is not well understood. As a highly lipophilic molecule, it penetrates well into tissues, but because of its relatively high molecular mass, it reaches only low CSF concentrations in the absence of meningeal inflammation.13 Postulated mechanisms include direct toxicity of the active metabolite 14-hydroxyclarithromycin, altered cortisol and prostaglandin metabolism, and inhibition of glutaminergic transmission.7 14 Recent in vitro and ex vivo studies have shown that clarithromycin can increase neuronal excitability by inhibiting GABA-A receptor activation.15 16 Another study found that clarithromycin modulates the synthesis and metabolism of glycerophospholipids in the cerebral cortex of mice.17
It is important for neurologists to be familiar with the neurotoxic effects of clarithromycin since it is frequently prescribed in hospital and primary care. This case more broadly highlights the need to consider toxic encephalopathy in patients presenting with suspected autoimmune limbic encephalitis and to take a comprehensive drug history in all patients. In this case, our moment of clarity came the second time around.
Key points
An exhaustive drug history is essential when assessing patients with acute encephalopathy.
Recognising antibiotic-associated encephalopathy allows the offending drug to be stopped, prevents overinvestigation and avoids harm from its repeated administration.
Clarithromycin can cause psychomotor agitation, mania, psychosis, visual hallucinations and less often seizures.
A large population study found a fourfold increased risk of first neuropsychiatric during the 14-day treatment with clarithromycin-containing Helicobacter pylori eradication, suggesting that its neurotoxicity is under-recognised.
Further reading
Bhattacharyya S, Darby RR, Raibagkar P, Gonzalez Castro LN, Berkowitz AL. Antibiotic-associated encephalopathy.Neurology 2016; 86(10):963–71.
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Ethics statements
Patient consent for publication
References
Footnotes
Contributors DW drafted the manuscript and subsequent revisions. All authors contributed to clinical care of the patient, critically appraised and revised the manuscript for important intellectual content, and approved the final version of the article.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed by Aaron Berkowitz, San Francisco, USA.