Introduction

The COVID-19 pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), continues to rage-on in waves, with over 1.5 billion cases and 3.1 million deaths recorded globally from December 2019 to May 2021 [1]. Although predominantly a respiratory illness, COVID-19 has increasingly been recognised to cause disorders of the nervous system including encephalitis, encephalopathy, acute disseminated encephalomyelitis, myelitis, Guillain–Barre syndrome, stroke, cerebral venous thrombosis and taste and smell dysfunction [2,3,4]. Neurological manifestations of SARS-CoV-2 can arise as a result of direct invasion, para-infectious or postinfectious immune mechanisms. Encephalitis related to SARS-CoV-2 is rare and is reported to occur usually within the first 2 weeks following infection [2, 3]. We report a patient manifesting encephalitis as a delayed complication of COVID-19. To our knowledge, this is the first authenticated case of SARS-CoV-2-induced encephalitis reported as late as 28 days after the initial infection.

Case presentation

A 47-year-old Sri Lankan woman with a past history of uncomplicated type 2 diabetes mellitus presented with lower respiratory symptoms in early January 2021, which was confirmed to be due to COVID-19 based on positive viral polymerase chain reaction (PCR) tests on nasal swabs and exclusion of other causes. She made an unremarkable recovery from the respiratory infection and returned home after 14 days. Twenty-eight days later, she developed gradual onset confusion and abnormal behaviour. There was no fever, headache or other significant neurological symptoms. Over the next 4 days, her symptoms worsened and she developed generalised seizures. On admission to hospital, she progressed into status epilepticus requiring assisted mechanical ventilation and admission to the intensive care unit (ICU). She did not have meningism. Initial blood investigations including her haemoglobin, white cell counts, random blood glucose, liver function tests, creatinine and electrolytes were within normal limits while inflammatory markers (ESR, CRP) were mildly elevated. Non-contrast-enhanced CT imaging of the brain showed bi-frontal white matter oedema. Cerebrospinal fluid (CSF) analysis revealed 10 lymphocytes/cumm, 5 polymorphs/cumm, normal protein and glucose concentrations. CSF viral PCR screening for Herpes simplex virus 1 and 2, Japanese encephalitis, Varicella zoster and SARS-CoV-2 were negative. SARS-CoV-2 IgM and IgG antibodies were detected in CSF using a chemiluminescent microparticle immunoassay (Abbott). However, NMDAR, LGI1, CASPR2, AMPA1/2R and GABARA/B antibodies were not detected in her CSF. Magnetic resonance imaging (MRI) of the brain on the fifth day of admission revealed confluent T2 fluid-attenuated inversion recovery (FLAIR) hyperintensities in the periventricular white matter, mainly clustered around frontal and occipital horns. There was no restricted diffusion, contrast enhancement or haemorrhagic changes in these lesions (Fig. 1A). FLAIR hyperintensities were also noted in the splenium, basal ganglia and in the ventral pons (Fig. 1B and C). The electroencephalogram showed generalised slow wave discharges consistent with encephalitis.

Fig. 1
figure 1

Axial MRI of the brain showing confluent T2-FLAIR hyperintensities in the periventricular white matter mainly clustered around frontal and occipital horns (A) and (B), in the basal ganglia and splenium of the corpus callosum (B) and a focal hyperintensity in the pons (C)

Full size image

She was initially treated with intravenous aciclovir until CSF test results were available. Intravenous methylprednisolone 1 g daily for 3 days and intravenous immunoglobulins (IVIg) 0.4 g/kg/day for 5 days were administered when it was evident that she had a postinfectious encephalitis. Her seizures were treated with intravenous midazolam and levetiracetam. She made a remarkable recovery following the immunotherapy and antiseizure medication. She was weaned off the ventilator 5 days after completion of IVIg and discharged home 2 weeks later with only minor residual cognitive deficits.

Discussion

Neurological complications have been observed in up to 36% of hospitalised COVID-19 patients [5]. Encephalitis is a rare complication of SARS-CoV-2 and has been reported mostly as a para-infectious manifestation occurring within days to 2 weeks of the initial infection [2,3,4]. Our patient presented 28 days after PCR confirmed COVID-19 pulmonary disease with a gradual onset encephalitis that rapidly progressed to status epilepticus requiring intensive care management. Inflammatory changes on brain MRI, delta rhythm on electroencephalogram and lymphocytic pleocytosis on CSF analysis were consistent with the clinical diagnosis of encephalitis. Her recent COVID-19 pulmonary disease suggested SARS-CoV-2 as a likely aetiology and the detection of SARS-CoV-2 IgM and IgG antibodies in her CSF in the absence of SARS-CoV-2 RNA coupled with a good response to immunotherapy strongly suggested the encephalitis in our patient to be a postinfectious immune-mediated manifestation of COVID-19.

The potential of coronavirus to cause delayed immune-mediated encephalitis has been demonstrated in murine models almost 40 years ago [6]. In humans, establishing SARS-CoV-2 as a cause of postinfectious encephalitis would always prove challenging, and more so, when it occurs as a delayed complication as in our patient. Indeed, in two previously reported cases, an immune-mediated mechanism was postulated without definitive evidence of intrathecal viral RNA or specific antibodies [7, 8]. Exclusion of other common aetiologies of encephalitis and evidence of SARS-CoV-2-specific intrathecal IgG and IgM antibodies in our patient authenticated SARS-CoV-2 as the cause of the postinfectious encephalitis fulfilling criteria for a definitive diagnosis [2]. Brain MRI abnormalities noted in our patient were consistent with imaging abnormalities previously reported in COVID-19 [9]. Herpes simplex virus encephalitis has been shown to trigger subsequent autoimmune encephalitis mediated by neuroglial surface antibodies [10] and CASPR2 antibodies have been reported in relation to COVID-19 [11]. None of the common encephalitogenic neuroglial surface binding antibodies were detected in the CSF of our patient.

Of the immune-mediated neuroinflammatory manifestations of COVID-19, acute disseminated encephalomyelitis appears to be the commonest and manifests at an average of 9 days post-infection [3]. Thus far, our patient is the presumptive first case of COVID-19 postinfectious encephalitis presenting with the longest delay of 4 weeks after the initial infection. Given that the mechanism of disease is immune-mediated, the viral genome was not detected in CSF. However, the PCR assay has not been validated for detection of SARS-CoV-2 in CSF and studies have shown that the viral genome is frequently undetectable in CSF [11].

As the pandemic continues to spread unabated and more and more people are infected worldwide, our case highlights the need to consider postinfectious encephalitis even in patients presenting after a protracted interval following COVID-19, the utility of detecting SARS-CoV-2-specific antibodies in the CSF in establishing diagnosis and the favourable response to immunotherapy in such patients. It would be prudent to maintain a high index of suspicion of this treatment-responsive neurological complication given that many patients may have had asymptomatic or mildly symptomatic COVID-19.