Herpes simplex virus encephalitis in a patient receiving ustekinumab associated with extensive cerebral oedema and brainshift successfully treated by immunosuppression with dexamethasone

Abstract

Herpes simplex virus (HSV) encephalitis affects 2–4 people per million/year. Immunocompomised patients can have atypical presentations of HSV encephalitis, including a lack of cerebrospinal fluid (CSF) pleocytosis. We present the case of a patient who was receiving ustekinumab therapy for psoriasis which inhibits interleukin (IL)-12 and IL-23 signalling pathways. The initial presentation was suggestive of encephalitis, but he was discharged prior to the reporting of HSV positivity due to the lack of CSF pleocytosis. On representation, he had worsening symptoms and imaging showed midline shift, indicating cerebral oedema despite the immunosupressant effects of ustekinumab. He required intensive care unit support and treatment with high dose aciclovir and dexamethasone; after a month of treatment he made a good recovery. This case is the first to report a link between ustekinumab and HSV encephalitis, and also emphasises that imunocompromised patients can lack CSF pleocytosis and develop significant cerebral oedema which responds to immune suppression.

Background

Herpes simplex virus (HSV) encephalitis is a condition of brain inflammation which has an estimated incidence of 2–4 cases per million per year worldwide.¹ It is known that 10%–15% of patients with encephalitis can have normal cerebrospinal fluid (CSF) at presentation, a proportion that increases in the immunocompromised.² Immunomodulatory therapies have expanded significantly in recent times, particularly for autoimmune conditions affecting the vascular, dermatological and neurological systems. A recent development in the treatment of psoriasis includes ustekinumab, which inhibits the cytokines, interleukin (IL)-12 and IL-23, binding to their cognate receptors.³ In this paper, we report the first case of HSV encephalitis in a patient receiving ustekinumab to the authors’ knowledge. Broad-spectrum immune-suppressive agents, including steroids and methotrexate, have been reported to reduce cerebral oedema in HSV encephalitis.^{4 5} However, this case demonstrates that new, more targeted, immunomodulatory therapies can be associated with marked cerebral oedema and brainshift, which was successfully treated by immunosuppression with dexamethasone.^{4 5}

Case presentation

A 39-year-old man presented to the accident and emergency department with a 2-day history of severe holocephalic headaches, neck stiffness and fever on a background of headaches, irritability and alterations in personality and behaviour for the last month. His past medical history included psoriasis, for which his only recent therapy was ustekinumab, an anti-IL-12/anti-IL-23 therapy, which he had been receiving for several years prior to presentation. Some years earlier he had been treated with methotrexate. Prior to a lumbar puncture (LP), he was treated with intravenous antibiotics, but continued to have intermittent fevers throughout the week of his admission. He was discharged following an LP from which the only CSF abnormality on initial testing was a raised protein; the virological PCR results were still pending at this time. An unenhanced CT brain scan at this time was reported as normal. The initial differential diagnosis was that of a non-central nervous system (CNS) viral infection, as the LP had shown no evidence of CSF pleocytosis. He re-presented the day following his discharge with ongoing severe headaches and vomiting and had a reduced level of consciousness. On the day of his re-admission, 1 week after his initial presentation, a second LP was performed, and he was started on intravenous aciclovir along with broad spectrum antibiotics. Over the following 2 days he became more confused, his level of consciousness continued to decline, and he eventually required intubation.

Investigations

During the initial presentation, the CSF showed 1 white cell per mm³, 0.65 g/L protein and 4.1 mmol/L glucose (ratio unavailable). Following the patient’s discharge, PCR for HSV was reported to be positive, and pneumococcus, meningococcus and cryptococcus PCR was negative.

CSF analysis on the second presentation to hospital 1 week after discharge showed 556 white cells per mm³ (95% lymphocytes), 1.99 g/L protein and 3.8 mmol/L glucose (ratio 52%), and PCR for HSV remained positive. CSF oligoclonal bands, IgG index and immunoglobulin levels were not tested for. The imaging during the second presentation is shown in figure 1. The CT brain scan (1A and 1B) showed an area of low density in the right medial temporal lobe and evidence of midline shift. These abnormalities were further elucidated on MRI, with medial temporal lobe T2-weighted hyperintensity, midline shift and uncal herniation (1 C and 1D).

Figure 1.

Imaging on re-presentation to hospital. A and B: CT brain. C and D MRI brain (coronal FLAIR and Axial T2-weigted sequences).

There was initial deterioration with a drop in consciousness level, and following transfer to the tertiary neurology centre, he underwent further MRI (figure 2) which showed more pronounced oedema within the right temporal lobe associated with significant midline shift and uncal herniation.

Figure 2.

Imaging taken following deterioration at the tertiary neurology centre The left MRI is a coronal slice of T2 FLAIR sequence, and right shows an axial slice of T2 sequence. Both images show worsening of the marked right medial temporal lobe T2 hyperintensity and worsening midline shift compared with figure 1.

Treatment

During the second admission to hospital, intravenous aciclovir, ceftriaxone and amoxicillin were given to cover a broad spectrum of CNS infections pending microbiology and virology investigation, although viral encephalitis was considered the likely diagnosis. When the patient developed a reducing level of consciousness and imaging confirmed significant midline shift, a neurosurgical review for consideration of a decompressive craniectomy was under-taken and it was concluded that there was insufficient evidence to perform the procedure. Therefore, intravenous dexamethasone was commenced at 8 mg three times/day. On review by a neurologist at the tertiary neurology centre, dexamethasone was increased to 10 mg three times/day and, given the history of immune suppression, the dose of intravenous aciclovir was increased to 15 mg/kg with ongoing monitoring of renal function.

Following 3 weeks of intravenous aciclovir, a further LP was performed to see if HSV could still be detected by PCR. There were 370 white cells per mm³ (100% lymphocytes), protein was 0.34 g/L and glucose was 5.2 mmol/L (ratio 78%); HSV PCR remained weakly positive and intravenous aciclovir was continued for a further 7 days. The fourth LP, after 33 days of aciclovir, was performed and HSV could not be detected by PCR; therefore, the intravenous aciclovir was stopped. Steroids were weaned gradually to discontinuation.

In parallel with neurological management, this patient was supported in the intensive care unit (ICU) for over 3 weeks. He required ICU supportive care, including intubation and ventilation, primarily to prevent worsening of intracranial hypertension in the context of an agitated and low conscious level state. The patient received ventilator support for over 3 weeks, necessitating a tracheostomy, and during this time he was also treated for a lower respiratory tract infection.

Outcome and follow-up

With neurological improvement, the patient received rehabilitation support. Currently, several months after initial presentation, the patient has had no receptive or expressive dysphasia, or clinical evidence of a relapse or autoimmune ‘pseudo-relapse’. However, his family has noticed personality changes, headaches, dizziness and fatigue. He will receive community-based neuropsychological rehabilitation.

Discussion

HSV infects humans through the mucous membrane. Once the virus has entered the sensory nerves, it travels to the ganglia where it enters the latent phase. The mechanisms by which HSV enters the CNS and causes encephalitis are incompletely understood. However, the innate immune system and CD4+ and CD8+ T cells have all been shown to be involved.⁶ Ustekinumab prevents IL-12 and IL-23 from interacting with their cognate receptors, thus disrupting the production of IFN-γ that is induced by IL-12, IL-17 and IL-22 which are in turn induced by IL-23. IL-12 and IL-23 signalling pathways are initiated by cells of the innate immune system, such as dendritic cells and tissue-resident macrophages, and effect T_H cell differentiation.⁷ This is the first reported case of HSV encephalitis in a patient taking ustekinumab, and suggests that IL-12 and IL-23 signalling pathways are involved in the immunological response to HSV in CNS.

Immunocompromise can be caused by treatments used in a wide range of clinical scenarios, such as cancer, transplantation and inflammatory conditions.^{8 9} There have been many case reports linking specific drugs to cases of HSV encephalitis, particularly methotrexate,⁵ ciclosporine,¹⁰ fingolimod,¹¹ antiTNFα class of drugs,¹² temozolomide¹³ and ipilimumab.¹⁴ As these are either singular case reports or case series, we cannot characterise the differences in clinical presentations of HSV encephalitis caused by each drug. However, some similarities can be seen within these cases, including the one we present; they can have a very subtle presentation and CSF parameters and imaging can be initially normal, making HSV encephalitis more difficult to diagnose in this population. CNS infection was not suspected in this patient because the CSF white cell count and glucose were within normal range, with only mildly increased protein. However, immunocompromise of various causes has been shown to lead to atypical presentations of HSV encephalitis, particularly subacute presentations and lack of CSF pleocytosis.^15–17 Despite the lack of pleocytosis in CSF, the patient’s initial HSV PCR was positive, in keeping with the pattern of HSV encephalitis in immunocompromised individuals.¹⁷ This case reiterates the importance of both suspecting encephalitis in patients who are on immunomodulatory therapies and awaiting negative PCR result prior to discontinuation of the antiviral therapy. National UK guidelines recommend that HSV encephalitis can be excluded if two separate LPs, separated by 48 hours, fail to identify HSV by PCR and an MRI brain scan is normal, if performed after at least 72 hours of symptoms.²

These same guidelines suggest performing an urgent LP on presentation, unless there are clear clinical contraindications, and intravenous aciclovir at 10 mg/kg started within 6 hours of admission.² On the second presentation, these guidelines were followed. However, the dose of aciclovir was increased to 15 mg/kg due to the history of immunocompromise and the high dose of dexamethasone used under specialist supervision. The use of corticosteroids in encephalitis is controversial, and a randomised control trial is currently being performed to evaluate its place in the management of HSV encephalitis.² This patient was not eligible for the trial as dexamethasone had already been commenced prior to consideration for randomisation due to the declining consciousness level secondary to extensive cerebral oedema and midline shift. In this case, decompressive hemicraniectomy was considered but not performed, as the risks potentially outweighed the benefits in the absence of any evidence to support this operation. There have been a small number of anecdotal case reports of successful recovery following decompressive hemicraniectomy in HSV encephalitis but evidence for large cohorts or controlled trials is lacking.¹⁸

While the patient had a good clinical outcome following the treatment outlined, we cannot exclude the possibility that he may have had a similar outcome had intravenous dexamethasone not been prescribed. Nevertheless, the progressive decline in consciousness was associated with significant cerebral oedema, midline shift and uncal herniation despite aciclovir; these features of raised intracranial pressure are typically associated with a poor prognosis, suggesting that broad immunosuppression with dexamethasone may have contributed to his positive outcome.¹⁹ Indeed, in observational studies of adjunctive dexamethasone, improved outcomes have been reported and in a murine model of HSV encephalitis, dexamethasone was associated with markedly reduced cerebral oedema in response to the treatment.^{20 21}While IL-12 and IL-23 have important roles in T_H1 and T_H17 differentiation respectively as well as natural killer responses, the more broad effects of dexamethasone may reduce cerebral oedema through multiple mechanisms, including stabilisation of blood-brain barrier permeability and reduced neutrophil transmigration, as has been demonstrated in an in vitro model of viral encephalitis.²²

Learning points.

• Immunocompromised patients with encephalitis can have relatively normal cerebrospinal fluid (CSF) appearances. When encephalitis is suspected, viral PCR for herpes viruses is mandatory, even where there is no CSF pleocytosis, and required to be proven negative before withdrawing antiviral aciclovir treatment. National UK guidelines recommend that two lumbar punctures should be negative for herpes simplex virus (HSV), and an MRI normal, before HSV is excluded in a patient with suspected encephalitis.

• There are increasing immunomodulatory therapies for ever increasing ranges of medical conditions and the presence of opportunistic infections must be considered when patients on therapies such as ustekinumab present with clinical features suggestive of possible central nervous system infections.

• Decompressive hemicraniectomy has been described in HSV encephalitis, but there is no strong evidence to support its use. It is not possible in individual anecdotal cases to determine if such treatment approach is worthwhile and higher-level evidence for the management of intracranial pressure in encephalitis is required.

• Despite receiving a targeted immunosuppressive agent, this patient developed marked cerebral oedema with midline shift and uncal herniation. Dexamethasone was associated with a reduction in cerebral oedema and a good clinical outcome.