Herpes Zoster Encephalitis

2 Case Reports and Review of Literature

Abstract

Herpes zoster encephalitis (HZE) is a rare complication of varicella-zoster virus infection. Because of its nonspecific clinical presentation, HZE is underrecognized. The advent of polymerase chain reaction for detecting viral particles in cerebrospinal fluid provides a rapid and accurate means of diagnosis. Two patients with HZE diagnosed by a positive cerebrospinal fluid-polymerase chain reaction and neurological changes are presented, and the literature is reviewed. We provide a practical approach to the diagnosis and treatment of HZE and conclude that HZE is underdiagnosed and should be considered in all patients with encephalitis even in the absence of a herpes zoster rash.

Varicella-zoster virus (VZV) causes a primary illness known as varicella or chicken pox that usually occurs in childhood, becomes latent in the cranial nerve and sensory nerve ganglia, and can later reactivate to cause herpes zoster or shingles. In rare occasions, zoster sine herpete occurs, which is herpes zoster without the characteristic rash.¹

It has been estimated that 0.01% to 0.25% of patients with varicella develop overt neurological complications such as cerebellar ataxia, encephalitis, transverse myelitis, aseptic meningitis, polyneuritis, cranial neuropathies, and Reye syndrome.^2,3 Neurological sequelae are more common in the herpes zoster phase.⁴ These sequelae include-in order of decreasing incidence-postherpetic neuralgia, cranial nerve palsies, peripheral motor neuropathy, myelitis, encephalitis, thrombotic cerebral vasculopathy, acute ascending polyradiculitis, and aseptic meningitis.^2,5

Herpes zoster encephalitis (HZE) is an uncommon complication of herpes zoster.^6-8 Immunosuppression is the principal risk factor for the development of HZE.^9-11 Historically, the diagnosis of HZE depended on the presence of the characteristic rash along with the temporal development of clinical encephalitis.³ With the advent of the polymerase chain reaction (PCR) technique for identifying VZV in the cerebrospinal fluid (CSF), HZE can be definitively diagnosed, thus allowing for directed therapy.^12,13

CASE 1

A 57-year-old man presented with a 6-day history of localized zoster on the right thigh. Three days after the outbreak of the rash, he developed headache, neck pain, and fever, followed by confusion, memory impairment, and drowsiness with episodes of agitation and restlessness. He had a history of type 2 diabetes mellitus and stage I B-cell chronic lymphocytic leukemia. He had no recent travel, animal exposure, corticosteroid intake, chemotherapy, or radiation therapy. Physical examination showed a blood pressure of 151/66, pulse rate of 75/min, respiratory rate of 20/min, temperature of 99.5°F, nuchal rigidity, and an erythematous bandlike vesiculopapular rash over the right third lumbar dermatome. Neurological examination revealed an altered sensorium and the presence of asterixis and bilateral ankle clonus. Otherwise, result of the neurological examination was normal. Results of the following were normal: serum electrolytes, renal and hepatic profiles, urinalysis, electrocardiography, chest roentgenography, and head computed tomography (CT). Complete blood cell count was significant for leukocytosis (22,600 cells/uL), which consisted of 75% lymphocytes, 20% neutrophils, 3% monocytes, and 2% eosinophils. Lumbar tap showed a cloudy CSF. Cerebrospinal fluid analysis revealed 2110 white blood cell per microliter with 96% lymphocytes. Cerebrospinal fluid glucose was 52 mg/dL (reference range, 40-80 mg/dL), which was 44% of the serum glucose of 117 mg/dL (normal, >40%), and protein was elevated at 320 mg/dL (reference range, 15-60 mg/dL). Intravenous acyclovir (1000 mg every 8 hours) and intravenous ceftriaxone (2 g daily) were initiated. The patient continued to exhibit a generalized cerebral impairment with slow speech, disorientation, personality changes, poor insight and judgment, impaired recent memory, and visual hallucinations. He did not develop seizures or focal neurological deficits. Bacterial antigens in CSF for Haemophilus influenzae, Streptococcus pneumoniae, group B Streptococcus, and Neisseria meningitidis were not detected. Levels of vitamin B12, folate, and thyroid-stimulating hormone were all within normal limits. The finding of a serum fungal immunodiffusion panel for Aspergillus, Blastomyces, Candida, Coccidioides, and Histoplasma was negative; the finding of a serum cryptococcal antigen was also negative. Ceftriaxone was discontinued. Cerebrospinal fluid cytology was performed and was consistent with reactive pleocytosis from a viral infection. No virus was isolated from culture of the skin lesion done on the sixth hospital day. The result of a purified protein derivative skin test was negative. Results of CSF bacterial, mycobacterial, and fungal cultures were negative. Results of CSF herpes simplex virus (HSV) DNA and Enterovirus RNA by PCR were negative. Varicella-zoster virus DNA PCR was positive. A magnetic resonance imaging (MRI) of the brain was remarkable only for the presence of a left choroidal detachment. Ophthalmologic and slit lamp examinations revealed glaucoma that was attributed to acute scleritis from the choroidal detachment possibly caused by the patient's recent VZV infection. There were no external ophthalmic lesions. The patient was treated with a 3-week course of intravenous acyclovir and was transferred to an extended care facility for rehabilitation. One month later, there was considerable improvement in his neurological function but then developed another decline in mental status. The CT and MRI of the head were unremarkable except for a finding of acute mastoiditis on the right side. A lumbar puncture now revealed only 67 nucleated cells with 92% lymphocytes, a normal glucose (88 mg/dL), and an elevated protein level (178 mg/dL). The results of bacterial, fungal, and mycobacterial cultures of CSF were again negative. Results of PCR testing for HSV and VZV were negative. Cerebrospinal fluid cytology showed no malignant cells. The patient has since made little neurological progress. An electroencephalography (EEG) 1 month after revealed a diffuse severe encephalopathy without focal abnormalities or seizure activity.

CASE 2

A 58-year-old man was admitted with a 2-day history of bizarre behavior and confusion associated with nausea, vomiting, and generalized weakness. Five days before, the patient was evaluated for left axillary chest pain and nausea. The finding on the initial workup was negative. His medical history included type 2 diabetes mellitus complicated by peripheral neuropathy. Physical examination was remarkable for a blood pressure of 118/80, pulse rate of 75/min, respiratory rate of 18/min, temperature of 100.8°F, and a vesicular rash with an erythematous base distributed over the left third and fourth thoracic dermatomal levels. Neurological examination was significant for drowsiness, disorientation, and confusion without focal deficits. Complete blood cell count, comprehensive drug screen, chest roentgenography, and urinalysis were unremarkable. A basic metabolic panel revealed acute renal failure attributed to dehydration. The patient received 2 g of intravenous cefotaxime, 1 g of intravenous vancomycin, and 800 mg of intravenous acyclovir and underwent lumbar puncture with CSF analysis showing a glucose of 136 mg/dL, a protein of 61 mg/dL, 6 red blood cell per microliter, and 127 white blood cell per microliter consisting of 1% neutrophils, 92% lymphocytes, and 4% monocytes. Results of Gram stain and bacterial culture were negative. Vancomycin and cefotaxime were thus discontinued. Intravenous acyclovir (800 mg) was given continuously every 8 hours. By the next hospital day, there was dramatic improvement in the patient's sensorium. The acute renal failure resolved with hydration. An MRI of the brain showed mild atrophy with remote small vessel ischemic changes. Results of blood and urine cultures were negative. Cerebrospinal fluid PCR for VZV was positive. The patient received intravenous acyclovir for 1 week and was discharged home with significant improvement in his symptoms.

DISCUSSION

The overall age-adjusted annual incidence rate of herpes zoster has been reported to be 1.3 per 1000 person-years.^2,14 It occurs approximately in 1% of the general population and in up to 25% of those with underlying neoplasms.¹⁵ However, the incidence of HZE in the general population is unknown. Among patients infected with VZV, HZE has been found to occur in approximately 0.1% to 0.2%.¹⁶ Reports have shown that neurological complications of herpes zoster are infrequent except for postherpetic neuralgia. Herpes zoster accounts for 8% to 13% of the cases of herpes zoster-associated neurological complications excluding postherpetic neuralgia.^17-19 A recent prospective study identified VZV as the most common etiology in adult patients with acute encephalitis.²⁰

Herpes zoster encephalitis has greater prevalence in such immunocompromised states as acquired immune deficiency syndrome, transplantation, malignancy, and advanced age.^1,5,10,11 Patients with leukemia or lymphoma seem to be at particular risk for VZV reactivation in general and HZE specifically.²¹ In a retrospective study of 1125 patients with systemic cancer, HZE occurred in almost 1% of the population.¹¹ In another cohort study comprising 962 patients diagnosed with chronic lymphocytic leukemia, 2 patients developed herpes zoster-related encephalitis.¹⁴ In a Finnish study, HZE was reported to be the leading cause of encephalitis in patients older than 65 years.²²

The severity and location of herpes zoster involvement affect the risk for development of HZE. In 1 study, 10 of 32 patients with disseminated herpes zoster had encephalitis.²³ Disseminated herpes zoster increases the risk of developing encephalitis by 30%.² Herpes zoster encephalitis also seemed to be more common after trigeminal distribution of shingles compared with other sites.^11,15 The presence of 2 or more prior episodes of herpes zoster and cervical nerve involvement has also been found to predispose to the development of HZE.² In both patients presented previously, diabetes mellitus was present and may have increased the susceptibility to develope HZE. Diabetes has been implicated as a predisposing factor in the development of herpes zoster-associated neurological disease.¹⁸

Varicella-zoster virus can affect any area of the central or peripheral nervous system through a vasculopathic process.^10,12 Herpes zoster encephalitis-affected patients may present with such diverse findings as multiple ischemic and hemorrhagic infarctions, demyelinating disease, seizures, peripheral neuropathy, and mental status changes.^23-25 Generalized cerebral impairment is common and may include a decreased level of consciousness, behavioral and personality changes, cognitive decline, and memory impairment. These deficits are suggestive of a subcortical type of cognitive impairment similar to that reported in 1997 in a study of 40 patients with HZE.^23,26

In HZE, the occurrence of a recent or concomitant episode of herpes zoster or shingles is important in distinguishing this entity from other causes of encephalitis. The onset of central nervous system (CNS) symptoms usually occurs days to weeks, sometimes up to months after the herpes zoster eruption.^2,5,25 In a small number of cases, neurological manifestations appear before the appearance of a rash or, even rarer, in the absence of a rash.^5,24,27 Interestingly, Koskiniemi et al¹⁷ reported that 40% of 174 patients with HZE had zoster sine herpete. Seizures and ataxia are infrequent sequelae of HZE.^23,28 Fever is usual at the onset of the rash and CNS symptoms but then resolves with improvement in neurological status.²³

After the development of herpes zoster, the virus can spread to the spinal cord and brain, leading to CNS complications.¹ Herpes zoster encephalitis exists in any or a combination of 3 pathological patterns-large vessel vasculopathy, small vessel vasculopathy, and ventriculitis/meningitis, thus explaining the variability of the presentation of the disease.¹⁰ In all cases, the main pathological finding is ganglionic hemorrhage and inflammation.²⁹ Large-vessel encephalitis predominantly affects elderly immunocompetent patients and is characterized by acute focal neurological deficits such as hemiplegia or blindness.^1,25 Small-vessel encephalitis is more commonly seen in immunocompromised patients who present with fever, headache, mental status changes, seizures, and focal deficits. Imaging shows multiple deep-seated ischemic or demyelinating lesions in neural white matter. Ventriculitis, the least common form of HZE, occurs with VZV infection of ependymal cells and may produce ataxia and hydrocephalus.¹

In the past, diagnosis of HZE was inferred by the coexistence of an encephalitic state with the herpes zoster rash. Electroencephalography abnormalities, compatible CSF results, and exclusion of other causes of encephalitis then supported the diagnosis.²³ Cerebrospinal fluid viral culture and antibody detection are conclusive, but the delay in obtaining results compromises timely initiation of therapy.³⁰ Cerebrospinal fluid analysis typically shows a lymphocytic pleocytosis with high normal-to-elevated protein levels and normal glucose levels. Such findings support the diagnosis of HZE but are not specific, occurring in other forms of CNS infection and in many as half the cases of uncomplicated herpes zoster.^9,15,31 Rare cases of HZE with low CSF glucose have been reported.^32,33 Cell counts and protein levels do not seem to correlate with disease severity.²³

Brain computed tomographic scans are generally unremarkable in patients with HZE and are mainly used to exclude other diagnoses such as tumor or hemorrhage.²⁸ Magnetic resonance imaging is more sensitive and specific than CT for evaluating viral encephalitis.¹² Magnetic resonance imaging findings that are suggestive of encephalitis from herpes zoster infection include discrete subcortical nonenhancing spherical lesions that eventually coalesce, develop enhancement, and spread to the gray matter.⁸ Autopsy reveals these spherical lesions to be the areas of demyelination and hemorrhagic infarction.⁸ However, MRI abnormalities have also been observed in patients with uncomplicated herpes zoster, with the CNS lesions corresponding to the homuncular distribution of the rash.⁹ Another technique that has been found to be helpful in the diagnosis of acute encephalitis is single photon emission computed tomography. In combination with the clinical presentation, CSF studies, and radiographic results, the finding of unilateral hyperperfusion on single photon emission computed tomography has been found to significantly increase the diagnostic yield for viral encephalitis, including HZE.³⁴ Most patients with HZE demonstrate mild diffuse slowing of brain activity without focal abnormalities on EEG.^15,23 This abnormality also lacks specificity for HZE because nearly 31% of patients with herpes zoster without neurological manifestations have abnormal EEG findings.²⁸ Diagnosis of VZV CNS infection has also been done using electron microscopy of CSF through identification of the characteristic viral cytopathic effect, rate of progression of this effect, and cell susceptibility. However, this tool does not provide a rapid means of diagnosis but may be valuable in confirming viral dissemination into spinal fluid and in evaluating the efficacy of antiviral therapy.³⁵

The diagnostic armamentarium has been improved with the recent introduction of VZV PCR.³⁶ Detection of viral DNA using PCR identifies recent or ongoing CNS infection.^12,37 Polymerase chain reaction, in combination with detection of intrathecal specific immunoglobulin G antibody, represents the most accurate method of diagnosis of neurological infections including HZE.³⁸ A ratio of antibody levels in serum to CSF of 20 or less indicates intrathecal antibody production, provided no other antibodies are present in the CSF.¹³ In 1 study, VZV antibodies can be measured by indirect membrane immunofluorescence in up to 94% of patients with HZE.² Detection of VZV by PCR provides a quick, minimally invasive, and accurate diagnostic marker for HZE especially when there is no concomitant rash.³⁹ Polymerase chain reaction for viral DNA has been observed to be present as early as day 3 after the appearance of vesicles with results available within 1 day, unlike viral culture and serology which take more than a week to obtain results.^30,37 Negative PCR results do not rule out the diagnosis. False-negative results may be due to insufficient DNA in CSF or variation in viral genome with advanced disease.³⁷ Because VZV DNA may be undetectable in patients with HZE, antibody testing of CSF for VZV has been recommended to confirm the diagnosis.^16,17 Cerebrospinal fluid PCR retains its sensitivity after initiation of antiviral therapy and thus permits empirical treatment before obtaining CSF.¹² Polymerase chain reaction is also very useful in cases where serological techniques may be equivocal because of the presence of cross-reactive HSV and VZV antibodies.³⁹ One disadvantage of PCR is the fall in sensitivity in protracted cases. It was observed that CSF VZV PCR was positive only in patients with overt neurological deficits during the acute phase-initial 1 to 2 weeks of the disease- when the amount of replicating virus is maximal.¹² It has been shown that PCR become negative over time because the replicating virus and viral DNA do not persist indefinitely in the CSF.¹² Cerebrospinal fluid antibody testing is particularly useful in this group of patients with neurological symptoms manifesting weeks to months after the onset of the rash.³⁷ Multiplex PCR, which can detect several viral DNA simultaneously, is cost effective, although sensitivity and specificity still need to be determined.³⁰ This test may be most useful in patients in whom viral encephalitis is considered, but a specific agent is not clinically evident.

Because of the rarity of HZE, randomized control trials have not been performed to assess medical treatment. Moreover, the significant morbidity of HZE coupled with the benign side-effect profile of antiherpetic therapy make it unlikely that a placebo-controlled study could be ethically performed. Numerous case reports and case series have illustrated the effectiveness of acyclovir in HZE.⁴⁰ However, other case reports suggest that acyclovir may be ineffective in HZE.¹⁸ Acyclovir administration is standard therapy despite limited data because of its relatively benign side-effect profile and the significant morbidity associated with HZE.³ Although not proven to eradicate VZV, treatment with acyclovir prohibits active infection and presumably keeps the virus in its latent state.²⁴ Doses used for VZV encephalitis have been similar to that proven to be effective for herpes simplex encephalitis-10 to 15 mg/kg intravenously every 8 hours for a duration of 10 to 14 days.^13,41 The efficacy of therapy is implied by the disappearance of VZV DNA with antiviral treatment. One study revealed undetectable VZV DNA by PCR in patients treated with acyclovir for at least 2 days but persistence of viral DNA in a third of patients with less than 2 days of treatment or those without treatment.⁹ Cerebrospinal fluid PCR may thus be useful in monitoring therapeutic response and determining prognosis in patients with HZE.^12,29 Clinical failure after treatment with acyclovir may be due to persistence of the virus in the CSF, inadequate CNS drug concentration, or development of acyclovir-resistant strains.⁴¹ Alternative approaches such as the use of drug combinations or other drugs including famciclovir or sorivudine have been successfully used for VZV infection, but no information is available for cases with associated neurological complications.⁴¹ Another antiviral agent, ganciclovir, has been shown to be effective in some patients with varicella zoster encephalitis.⁴⁰ In 1 case of HZE where acyclovir resistance occurred, vidarabine was found to be effective.⁴²

The use of corticosteroids in HZE is controversial, perhaps best suited for immunocompetent patients with severe encephalitis or evolving cerebral edema.¹³ It is unknown if antiviral therapy prevents the development of HZE in patients with shingles. A recent multicenter trial proved the efficacy of the zoster vaccine, a vaccine 14 times the potency of the current live varicella virus vaccine (VARIVAX), in markedly reducing the incidence of herpes zoster (by 51.3%) in older adults.⁴³ This vaccine should thus potentially reduce the incidence of HZE, especially in susceptible immunocompetent patients who can receive the vaccine.

In a 1983 study evaluating 25 patients with HZE, the average length of neurological symptoms was approximately 16 days, with note of longer duration in cases of disseminated herpes zoster and immunocompromised states.²³ In another report, the mean duration of CNS symptoms was 3 and 7 days, respectively, for acyclovir-treated and nontreated patients.²⁸ The clinical outcome in patients with HZE has been variable. In a study by Jemsek et al,²³ patients consistently returned to their preencephalitic mental status regardless of severity of disease. Normal cognitive function resulted in 90% of cases.²³ This was comparable with other studies wherein most patients with HZE recovered completely with²⁶ or without⁴⁴ antiviral therapy. A slow but eventual return to baseline cognitive state was the trend in 12 cases of HZE.²³A separate study reported long-term neurological sequelae of HZE to be present in up to 30% of patients.⁵ Other studies have found the rate of HZE mortality to be as high as 20% to 50%.^5,15 Deaths from HZE have been mainly attributed to nonneurological complications such as pneumonia.⁵

CONCLUSIONS

The clinical manifestations of HZE are variable, nonspecific, and often indistinguishable from other forms of encephalitis. Older and immunosuppressed individuals are more prone to acquiring the disease. The finding of an antecedent or concurrent episode of herpes zoster rash is a key in the diagnosis but is not always present. The CSF VZV PCR, a rapid and highly sensitive test, is most valuable in proving the diagnosis of HZE and thus should be used when HZE is being considered. Timely treatment with acyclovir may be beneficial and should be instituted empirically in suspected cases.

ACKNOWLEDGMENT

The authors thank Dr Joseph Myers for his valuable comments and suggestions.