The challenge of diagnosing and successfully treating anti-NMDA receptor encephalitis in a toddler

Abstract

Anti-N-methyl-D-aspartate (anti-NMDA) receptor encephalitis is an immune-mediated syndrome that is still under-recognised, with grave consequences if not treated early. A multidisciplinary team approach is required in the process of diagnosis and management of this potentially treatable and reversible disorder. We report on a 26-month-old Sudanese girl who presented with focal seizures associated with fever (temperature = 38.9°C) and history of trivial head trauma a day before. Viral encephalitis was suspected, and she was started on acyclovir and ceftriaxone. Cranial computed tomography revealed small high density in the right frontal lobe, and magnetic resonance imaging showed the features of cortical haemorrhagic lesion at the right frontoparietal lobe. Polymerase chain reaction for herpes simplex virus 1 and 2 revealed negative results. Her condition worsened over the course of 1 week, with recurrent seizures, insomnia, violent chorea and orofacial dyskinesia. Electroencephalography showed diffuse slow activity and the presence of ‘extreme delta brush’ pattern, a specific abnormality seen in anti-NMDA receptor (NMDAR) encephalitis. Cerebrospinal fluid was positive for anti-NMDAR antibodies (titre = 1:100). She was treated with intravenous (IV) corticosteroids, IV immune globulin, plasma exchange and rituximab. Her condition improved gradually, with full recovery when last seen 19 months after the onset of the disease.

INTRODUCTION

Antibody-associated encephalitis, causing acquired neurological deficits in children, is increasingly being recognised. They are hypothesised to result from autoantibodies triggered by pathogens or directed against tumoural antigens [1]. Of these, anti-N-methyl-D-aspartate (NMDA) receptor encephalitis results from antibodies directed against the NMDA receptor (NMDAR), which has a vital role in synaptic transmission and brain plasticity, resulting in cortical and subcortical manifestations. Although the estimated incidence is very low (0.85 children/million/year) [2], it constitutes the most common identifiable cause of encephalitis in children [1,3]. In affected females over 18 years of age, ovarian teratomas are frequently seen (>50%), whereas a low rate of associated neoplasm (0%-9%) is seen in prepubertal girls and male patients [1,4]. Herpes simplex virus (HSV) is an important trigger of anti-NMDAR encephalitis [5,6].

Reports on infants and toddlers (aged <36 months) with anti-NMDA encephalitis are rare, and reports in children from the Middle East and North Africa countries are extremely rare [6,7]. This may be due to the misrecognition of symptoms and/or the unavailability of diagnostic testing in many of these countries.

We describe a challenging case of a 26-month-old Sudanese girl with anti-NMDA encephalitis who initially presented with focal seizures, associated with fever and preceded by trivial head trauma.

CASE REPORT

A 26-month-old Sudanese girl presented with a history of focal seizure in the form of facial twitching and tonic-clonic movement of the left upper and lower limbs associated with fever (temperature = 38.9°C). There was no history of previous seizures but a history of trivial head trauma few days earlier. She was treated empirically with ceftriaxone and intravenous (IV) acyclovir for possible herpetic encephalitis. The conducted investigations included cerebrospinal fluid (CSF) analysis which showed white blood cell count of 27/μl (all lymphocytes), protein of 0.39 g/l, glucose of 3.7 mmol/l (blood glucose of 5 mmol/l) and red blood cells of 0 cells/mm³. Gram stain and bacterial culture were negative, and polymerase chain reaction (PCR) viral workup for Enteroviruses and HSV 1 and 2 were also negative. Cranial computed tomography (CT) revealed a small high-density lesion, noted in the right frontal lobe, reported as suggestive of haemorrhagic contusion (Figure 1).

Figure 1.

Cranial CT revealing a small high-density lesion, noted in the right frontal lobe (arrow), reported as suggestive of haemorrhagic contusion.

The patient was stabilised, fever subsided and clinically improved. Two days after admission, she spiked a fever (temperature of 37.8°C), manifested a decreased level of conscious associated with brief focal seizures and aborted with phenytoin loading, and she was started on levetiracetam. She was seen by neurosurgery team who suspected intracranial bleeding and requested brain magnetic resonance imaging (MRI) which showed right frontoparietal lobe cortical swelling with cortical and subcortical patchy haemorrhage, evaluated as keeping with cortical haemorrhagic contusion (Figure 2). The visualised orbital cavities and their structures, as well as the upper cervical spinal cord, were grossly unremarkable. Magnetic resonance angiography (MRA) showed grossly maintained normal signal voids of the visualised vascular structures.

Figure 2.

Brain MRI showing right frontoparietal lobe cortical swelling with cortical and subcortical patchy congruent high signal intensities in T2-weighted image (A, B and C) and foci of blooming artefacts in diffusion restriction (D) in keeping with cortical haemorrhage (arrows).

The Glasgow Coma Scale was 11/15, and a few days later, she started to have abnormal movements in wakefulness which disappear during sleep. Initially, seizure was suspected, and a loading dose of phenytoin was given. These choreiform movements became worse, and electroencephalography (EEG) showed diffuse slowing with superimposed rhythmic beta frequency activity (extreme delta brush) and no epileptiform discharges (Figure 3).

Figure 3.

Electroencephalogram (EEG) showing generalised rhythmic delta frequency activity with superimposed rhythmic beta frequency activity (extreme delta brush) and no epileptiform discharges.

Clinically, the level of conscious decreased (Glasgow Coma Scale 8/15), and the choreiform movement worsened with remarkable orofacial dyskinesia. Video EEG was done, and the recording was supportive of chorea with no epileptiform activity. CSF analysis was repeated and sent for autoantibodies. The result came positive for anti-NMDAR antibodies (titre =1:100). She was treated with IV corticosteroids, IV immune globulin and plasma exchange (five cycles). She was also maintained on acyclovir and received ceftriaxone for 14 days.

The patient improved clinically, the Glasgow Coma Scale was 13/15, chorea persisted but with less severity and orofacial dyskinesia remained the same. She also had difficulty in feeding and lack of sleep. Following a multidisciplinary meeting (neurology, infectious diseases and rheumatology units), rituximab (two doses) was started, and acyclovir was continued. Valproic acid and haloperidol were used to control chorea and chloral hydrate to improve her sleep. Improvement was observed after 2 weeks, with less chorea and better sleep. The second brain MRI (Figure 4), conducted 3 weeks after the first one (Figure 2), showed a reduction of the previous haemorrhage and the surrounding oedema. MRA and magnetic resonance venography were unremarkable.

Figure 4.

Brain MRI, done 3 weeks after the first one (Figure 2). There is a reduction of the previous haemorrhage with evidence of blooming artefact and reduction at the surrounding oedema (arrows in A and B, T2-weighted images). (C) There is mild linear high signal intensity seen on axial fluid-attenuated inversion recovery (FLAIR) image surrounding haemorrhagic component with brain gliosis (arrow).

The patient stayed 3 months in hospital and was discharged with mild orofacial dyskinesia and distal chorea. Ultrasound abdomen and pelvis revealed normal study, and the swallowing test was normal. Brain CT angiography, done before discharge, showed normal intracranial major arteries, without any significant stenosis or occlusion. Dural venous sinuses were visible without any filling defect. She was referred to rehabilitation centre for physiotherapy and functional therapy.

Three months after rehabilitation, she came to the clinic for follow-up. There was no chorea or oral dyskinesia, but she had mild drooling. Her neurological examination was normal, apart from fine motor abnormality. She was unable to feed herself and could hold a spoon but had difficulty reaching her mouth, with mild tremor. She could walk without assistance, started to say single words, hold a pen and make a scratch.

The patient travelled back home (Sudan), where she continued on physiotherapy and was off medications. A video taken 18 months after the onset of the disease showed her communicating and walking normally. She was assessed a month later at the Paediatric Neurology Outpatient Clinic, Soba University Hospital, Khartoum, Sudan. She joined a kindergarten and had coherent speech, and neurological examination revealed no abnormality (Figure 5).

Figure 5.

Photo of the patient taken at the Pediatric Neurology Outpatient Clinic, Soba University Hospital, Khartoum, Sudan, 19 months after the onset of the disease, revealing normal cognition.

DISCUSSION

Encephalitis is a pathological process referring to inflammation of the brain parenchyma with significant morbidity and mortality if not promptly diagnosed and treated [8]. It can result from direct infectious causes or follow immune-mediated processes [1]. On the other hand, encephalopathy is a general term that describes a state, where alteration of awareness occurs. It can result from toxic, metabolic or vascular aetiologies, in addition to infectious causes or immune-mediated inflammation [1,8]. In a retrospective single-centre cohort study in children, the estimated frequency of immune-mediated encephalitis (approximately 34%) surpassed that of infectious aetiologies [9].

To confirm a specific diagnosis in encephalitis is challenging since the clinical presentation as well as the requested special investigations are often non-specific [10]. Clinically, both infectious and inflammatory processes may present with focal neurological deficits or manifest diffuse neurological abnormalities such as alteration of awareness or seizures [1]. A possible inflammatory aetiology is suggested by the presence of psychiatric manifestations (psychosis or mood alteration) in addition to abnormal movements including chorea, athetosis and dystonia [1].

Common symptoms in the early stage of anti-NMDAR encephalitis include seizures (in up to 75% of children), speech dysfunction and abnormal behaviour or cognition [1]. Movement disorder, such as chorea and facial dyskinesias (>80% of children) and dystonia, is a prominent feature. In the late stage of the disease, there is a decreased level of consciousness and autonomic instability or central hypoventilation [1]. Although the reported patient presented with fever and seizures following trivial head trauma, the evolution of symptoms, mainly the chorea and facial dyskinesia, suggested anti-NMDAR encephalitis to be the underlying cause. The diagnosis was further supported by EEG which showed generalised slowing and superimposed rhythmic beta frequency activity ‘extreme delta brush’, characteristically found in anti-NMDAR encephalitis) [1]. ‘Extreme delta brush’ pattern has been reported in up to 53% of affected children [1,11]. The CSF analysis showed lymphocytic pleocytosis and was positive for anti-NMDAR antibodies.

In up to 65% of patients with anti-NMDAR encephalitis, the MRI brain may be normal. Diffuse or patchy non-enhancing cortical abnormalities may be seen in the remainder. These often involve the medial temporal lobe and insular region [1]. MRI of the patient revealed hyperintense signal on T2-weighted images and restricted diffusion abnormalities (on diffusion-weighted images) corresponding to oedematous changes in the right frontoparietal lobe. It also revealed cortical haemorrhage. The serial radiology features and evolution of the frontoparietal lesion are compatible with herpes simplex encephalitis as an underlying cause rather than brain contusion following the trivial head trauma, preceding the fever and seizures [6,12]. This is despite a negative result of PCR for HSV in CSF taken on admission. It has been documented that for the first 3 days after the onset of encephalitis, a HSV PCR is often negative [10]. The patient was treated empirically with IV acyclovir for HSV, and a follow-up brain MRI (Figure 4) showed a reduction of the previous haemorrhage and the surrounding oedema.

The patient received the first-line immunotherapy for anti-NMDAR encephalitis [1] namely, IV corticosteroids, IV immune globulin and plasma exchange. Due to her persistent symptoms, she received rituximab. Both rituximab and cyclophosphamide have been used in an open-label fashion to treat children with persistent symptoms following the use of first-line therapies [1]. The use of other drugs (tacrolimus, methotrexate, azathioprine, mycophenolate mofetil or methotrexate) has also been reported [1]. Surgical removal of the associated tumour is required in paraneoplastic cases. However, ultrasound abdomen and pelvis revealed a normal study in the patient.

In anti-NMDAR encephalitis, symptomatic treatment of movement disorders is challenging, and medications have been used to treat the symptoms with variable responses [6]. In the present patient, valproic acid and the neuroleptic agent haloperidol were used to control chorea. Neuroleptics are the most widely used agents in the treatment of chorea and are less likely to interfere with volitional movement [13]. Another option in the treatment of chorea is dopamine-depleting agents (e.g., tetrabenazine and deutetrabenazine). However, in susceptible individuals, high doses can interfere with volitional movement [13]. Valproic acid, oxcarbazepine and benzodiazepines (clonazepam) can also be helpful. Some cases of florid chorea will respond to levetiracetam which can be loaded intravenously following admission [13].

Full or substantial recovery occurs in nearly 80% of children with anti-NMDAR encephalitis, and this may be hastened by the use of immunotherapy; one or more relapses can occur in +25%-33% of children [1,4]. A prompt diagnosis and early treatment should be the aim of therapy if full recovery is to be achieved [2,6].

CONFLICTS OF INTEREST

The authors declare that there are no conflicts of interest.

FUNDING

None.

ETHICAL APPROVAL

Signed informed consent for participation and publication of medical details and photographs were obtained from the parents of the child. The authors declare that ethics committee approval was not required for this case report.