Atypical herpes simplex virus encephalitis diagnosed by PCR amplification of viral DNA from CSF

Herpes simplex virus encephalitis (HSVE) is the most commonly recognized cause of acute sporadic encephalitis in the Western world.^1,2 Early initiation of antiviral therapy with acyclovir dramatically reduces mortality and morbidity.³ The availability of safe and effective antiviral therapy, and the evidence that early treatment enhances its efficacy,^3,4 make rapid and accurate diagnosis of HSVE essential. Until recently, definitive diagnosis of HSVE depended on direct demonstration of virus or viral antigens in brain tissue obtained by biopsy or at autopsy. Several studies have shown that amplification of HSV DNA from CSF using PCR techniques^5-7 is diagnostic for HSVE with a sensitivity and specificity equaling or exceeding that of brain biopsy.⁸

The classic clinical syndrome of HSVE has been defined almost exclusively on the basis of clinical and laboratory features exhibited by patients whose diagnosis has been established at autopsy^9,10 or by brain biopsy.^3,11 Enrollment in clinical trials of treatment for HSVE sponsored by the Collaborative Antiviral Study Group of the National Institute on Allergy and Infectious Diseases(CASG-NIAID) required that patients have an acute febrile encephalopathy with disordered mentation, focal cerebral signs, evidence of localization by diagnostic procedures (e.g., EEG, arteriography, brain scan, cranial CT), and CSF findings compatible with viral infection.^3,11 The clinical picture of HSVE derived from the CASG-NIAID trials and autopsy-based data has inherent ascertainment bias for features associated with severe HSVE and under recognizes mild or otherwise atypical illnesses.

It has long been recognized that mild or atypical presentations of HSVE occur,^12-14 but it is impossible to determine their frequency on the basis of isolated case reports in the literature. This problem is further compounded by varying standards used for diagnosis or by using diagnostic criteria of low specificity.¹⁵ Increasingly widespread use of CSF PCR has demonstrated the role played by HSV in a variety of CNS infections, including recurrent meningitis,¹⁶ brainstem encephalitis,¹⁷ and myelitis.¹⁸ Widespread use of CSF PCR might also improve recognition of mild or otherwise atypical cases of HSVE, as demonstrated by a case report describing a case of mild HSVE in a child.¹⁹ A range of clinical presentations of HSVE was observed in a Brazilian study in which 17% of patients with PCR-proven HSVE had mild disease, and up to 25% of a selected population of patients with mild encephalitis had HSV infection documented by PCR.²⁰

In our institution, CSF PCR to detect HSV DNA is the diagnostic procedure of choice. In order to determine the frequency of mild cases of HSVE among patients whose CSF specimens were submitted for HSV DNA PCR, we analyzed prospectively collected clinical data, performed a retrospective review of medical records, and interviewed referring physicians to obtain information concerning the clinical features of all patients with a positive HSV CSF PCR obtained during a 4-year period in our diagnostic virology laboratory.

Methods. All CSF specimens sent to the Diagnostic Virology Laboratory at the University of Colorado Health Sciences Center (UCHSC) between January 1, 1993, and December 31, 1996, for detection of HSV DNA by PCR were reviewed. Clinical information was prospectively collected on all patients with positive HSV CSF PCR results and supplemented by retrospective review of medical records and direct contact with referring physicians.

All PCR reactions were performed in accordance with established guidelines for avoiding contamination with exogenous DNA of either CSF samples or DNA amplification reactions (see ^{reference 16}). PCR for HSV was performed as previously described.¹⁶ Briefly, DNA was extracted from CSF specimens using the InstaGene DNA purification matrix(BioRad Laboratories, Hercules, CA) according to the manufacturer's instructions. PCR was carried out in 50-µL mixtures containing 20µL of extracted DNA, 1 unit of PfuI (Stratagene, La Jolla, CA), 100 µmol of each of the four deoxynucleoside triphosphates, and 1µmol of each primer (HSVPo1A1 and HSVPo1A2) in PfuI buffer(Stratagene). The primers selected have complete sequence homology to a region within the DNA polymerase gene and produce a 224-base pair amplification product. Extracted DNA was amplified in duplicate for 45 cycles, separated according to molecular weight using agarose gel electrophoresis, and transferred to nylon membranes. The identity of the DNA bands was confirmed by hybridization with a digoxigenin-labeled internal probe for the DNA polymerase gene, followed by anti-digoxigenin antibody conjugated to alkaline phosphatase(Boehringer Mannheim, Indianapolis, IN) and CSPD chemiluminescent substrate(Tropix, Bedford, MA). Each assay included two negative water controls and two positive sensitivity controls. The tests were considered valid if the controls yielded the expected results and if the results obtained in the replicate patient CSF specimens were in agreement. This viral HSV PCR method has a sensitivity of two viral copies per reaction.

The primers we used for HSV PCR do not distinguish between the DNA polymerase genes of HSV-1 and HSV-2. CSF HSV PCR-positive specimens from patients with atypical HSVE were serotyped as previously described^16,17 using restriction enzyme analysis of the amplified DNA based on the presence of Alu I and Bal I restriction enzyme polymorphisms between the amplified DNA polymerase genes of HSV-1 and HSV-2.

Clinical information was classified as follows: meningitis, recurrent meningitis, classic encephalitis, atypical encephalitis, or neonatal encephalitis. Criteria for diagnosis of HSV meningitis and recurrent meningitis have been previously described.^16,21 Patients were classified as having classic encephalitis if they met CASG-NIAID criteria for entry into clinical trials of antiviral therapy for HSVE.^3,11 Patients were considered to have atypical encephalitis if they had an acute encephalopathy with mild alteration in mental status or level of consciousness in the absence of focal findings on neurologic examination. The presence of focal findings on EEG or MRI did not exclude the diagnosis of atypical encephalitis (see Results). Patients were considered to have neonatal HSV infection if they developed signs or symptoms within 3 weeks of birth.^2,22

Results. A total of 1,224 CSF specimens were sent to the Diagnostic Virology Laboratory at UCHSC for determination of HSV DNA by PCR during the 4-year study period (table 1). A total of 7.6% (n = 93) were positive for HSV DNA. The percentage of positive specimens obtained in each individual year of the study remained relatively constant(mean 8.2%, range 6.1 to 11.1%). Patients with meningitis or recurrent meningitis accounted for 52% (n = 48) of the positive cases. A total of 58%(28 of 48) of the patients with meningitis met criteria for benign recurrent lymphocytic meningitis.¹⁶ Adult patients with encephalitis accounted for 26% (n = 24) of the positive CSF PCR results. Of this group 17% (n = 4) of patients were classified as having atypical encephalitis and 83% (n = 20) as having classic encephalitis. A total of 17%(n = 16) of the positive HSV PCR results were obtained from patients with neonatal HSV infections. Clinical information was not available for 5% (n = 5) of patients with positive CSF PCR results. The clinical and laboratory features of the patients with atypical HSVE are summarized below.

Patient 1. Y.F., a 55-year-old right-handed woman with steroid-dependent (prednisone 40 mg/d) asthma, presented with the onset of recurrent focal motor seizures involving the left face, arm, and leg with subsequent secondary generalization. At presentation to the hospital she was febrile (38.5 °C) but had no nuchal rigidity. Her neurologic examination was normal except for the mental status examination. She was somnolent but could be aroused and had diminished verbal responsiveness as well as decreased attention and concentration. Glasgow Coma Scale (GCS) score was 14. Initial CSF examination on the first day of illness was normal, although lymphocytic pleocytosis was present on repeat lumbar puncture (LP) at day 10(table 2). This CSF specimen was positive for HSV DNA by PCR, and restriction enzyme polymorphism typing indicated the DNA was from HSV-1. CT, obtained without contrast, was normal on day 1. T2-weighted MRI(displayed as axial and coronal sections) obtained on the fourth day of illness demonstrated increased signal involving both gray and white matter within a diffusely swollen right temporal lobe (figure 1). Acyclovir (30 mg/kg/d) was started on the fifth day of illness and continued for 14 days. The patient made a complete recovery. Neurologic examination including mental status was entirely normal when evaluated 3 months after hospital discharge.

Comment. The lack of CSF pleocytosis, normal CT, and absence of focal features caused the treating physicians to initially suspect the patient had status epilepticus perhaps secondary to a cerebrovascular accident or vasculitis. HSVE was not suspected until MRI was obtained on the fourth day of illness and was subsequently confirmed by a positive CSF PCR for HSV DNA. Because the diagnosis of HSVE was not initially suspected, institution of antiviral therapy was delayed until the fifth day of illness.

Patient 2. A.S., a 61-year-old right-handed man with chronic obstructive pulmonary disease (COPD) and a history of alcohol abuse, was hospitalized for suspected pneumonia. Treatment included antibiotics, bronchodilators, and steroids. Steroid therapy included 3 days of IV methylprednisolone (100 mg IV qd), 2 days of prednisone (60 mg po dq), and 2 days of dexamethasone (16 mg IV qd). On his seventh hospital day he had several tonic-clonic seizures. He was febrile (39 °C), with mild nuchal rigidity. His neurologic examination was normal except for the mental status examination. He was somnolent, with mild confusion, disorientation to place, inattention, and difficulty concentrating. GCS score was 14. Results of CSF examination are shown in table 2. A noncontrast-enhanced cranial CT obtained on day 2 after onset of neurologic symptoms showed subtle hypodensity in the right temporal lobe including parts of the insular cortex, extreme capsule, and external capsule. EEG obtained on the fourth day after onset of neurologic symptoms showed periodic lateralized epileptiform discharges (PLEDs) localized over the right hemisphere. MRI with both T2 weighting and fluid attenuated inversion recovery (FLAIR) sequences^23,24 obtained on the fifth day after onset of neurologic symptoms demonstrated increased signal in the right anterior inferior temporal lobe with associated involvement of the adjacent insular cortex as well as the hippocampal and parahippocampal gyri(figure 2). His neurologic condition appeared to be stable or improving despite the absence of specific therapy. He was treated with acyclovir (30 mg/kg/d for 14 days) beginning on the eighth day of illness after his CSF HSV PCR was reported positive for HSV DNA. Restriction enzyme polymorphism typing of the amplified DNA identified it as derived from HSV-1. At discharge his neurologic examination was nonfocal. He continued to have mild residual abnormalities in concentration (e.g., poor performance on the serial subtraction subtest of the Mini-Mental Status Examination). Family members felt that his intellectual capacities had returned to his premorbid baseline. He died 5 months after hospitalization from causes unrelated to his encephalitis. Postmortem examination was not performed.

Comment. This patient was initially suspected of having seizures secondary to viral meningitis because of the absence of focal neurologic features. After neuroimaging studies indicated the presence of focal parenchymal abnormalities and EEG showed PLEDs, he was thought to have viral encephalitis, but the diagnosis of HSVE was considered unlikely because of the relatively mild and nonprogressive nature of his symptoms despite the absence of antiviral therapy. Antiviral therapy was instituted after CSF PCR confirmed the diagnosis of HSVE.

Patient 3. J.B., a 47-year-old man, developed the acute onset of severe global headache, photophobia, nausea, and vomiting. At the emergency room of a local community hospital on the third day of illness he was afebrile (36.4 °C) but had nuchal rigidity. His neurologic examination was normal. Cranial CT after IV contrast administration obtained on the third day of illness was normal. CSF examination revealed lymphocytic pleocytosis (see table 2). He was hospitalized for 1 week and discharged with a diagnosis of viral meningitis. Over the next several days his headache worsened, and he became increasingly confused. His wife noted episodic "twitching and jerking" movements in his hands but no generalized seizures. Upon admission to the hospital he was febrile, with mild nuchal rigidity. His neurologic examination was normal except for his mental status. He was somnolent, disoriented to time, and confused. GCS score was 14. Results of his repeat CSF examination (day 18) are shown in table 2. HSV PCR on this CSF was positive, and restriction enzyme polymorphism typing indicated the DNA was derived from HSV-2. Cranial CT (obtained on the 18th day after onset of his initial symptoms) showed loss of definition of cortical sulci thought to be consistent with edema. EEG showed generalized slowing without periodic or paroxysmal activity. He was treated with acyclovir (30 mg/kg/d for 12 days) beginning on the 18th day of illness. Upon completion of therapy his neurologic examination was normal.

Comment. The absence of focal neurologic features and the presence of a normal CT lead to an initial diagnosis of viral meningitis. Although J.B.'s condition clearly worsened after his initial hospital discharge, the diagnosis of HSVE was not suspected until he developed obvious mental status changes, apparent focal seizures, and had evidence suggestive of generalized cerebral edema on CT. The diagnosis was confirmed by a positive CSF PCR for HSV DNA. Despite the delay in initiation of antiviral therapy (started on day 18 of illness), his clinical course progressed very slowly, and he made a complete recovery.

Patient 4. M.C., a 35-year-old man with AIDS and a CD4 count of<200 cells/mm³, was admitted to the hospital with recurrence of Pneumocystis carinii pneumonia. Medications included prednisone(40 mg bid), pentamidine, and zidovudine. He was febrile (37.2 °C). Three days into his hospitalization he became acutely confused and had a generalized tonic-clonic seizure. His neurologic examination was normal except for his confused mental state. GCS score was 14. Results of CSF examination (day 1) are shown in table 2. The HSV PCR on this CSF was positive, and restriction enzyme polymorphism typing indicated the DNA was from HSV-2. Cranial CT was unremarkable. He died secondary to continued deterioration in his pulmonary status and before the institution of antiviral therapy.

Comment. The diagnosis of HSVE was not considered antemortem because of the lack of CSF pleocytosis, normal CT, and the absence of focal features on neurologic examination.

Discussion. The CSF specimens that form the basis for this study were sent to the UCHSC Diagnostic Virology Laboratory with a request that HSV PCR be performed. It can be inferred that the referring physicians suscepted the possibility of HSV CNS infection. A total of 7.6% of submitted specimens were positive for HSV DNA by PCR (see table 1). This figure is higher than the 1.5% incidence of HSV PCR-positive CSF specimens from 2,233 consecutive nonselected samples sent during an overlapping period to a diagnostic virology laboratory in England.²⁵ However, the CSF samples in that study were tested based on clinical suspicion of a viral CNS infection and not specifically because of suspected HSV infection. Conversely, the incidence of HSV PCR-positive specimens found in our study was considerably lower than the 18% positive rate reported in a group of patients with "possible HSV encephalitis."⁶ Among the 93 patients with positive HSV CSF PCR in our study were 24 with encephalitis. Of this group, 17% (4 of 24) were classified as having atypical encephalitis. This figure is identical to that reported in a series of 49 Brazilian patients with HSVE.²⁰ In another study with clinical data on 14 HSV CSF PCR-positive patients with encephalitis, 14% (2 of 14) did not have focal features on neurologic examination and would have met the diagnostic criteria for atypical encephalitis used in this study and in that of Domingues et al.²⁰ Taken together, these results indicate that atypical HSVE is not uncommon. However, contrary to some prior predictions,¹⁴ most cases of HSVE do not fall into this category.

There are a number of clinical features that may help explain why the four patients identified in this study had atypical encephalitis. Two of the four had encephalitis caused by HSV-1 and two by HSV-2. Previous studies have shown that HSV-2 accounts for most cases of both monophasic² and recurrent¹⁶ HSV meningitis but for only 6 to 15% of cases of adult HSVE.^2,26,27 Although most reports suggest that the clinical features of HSV-1and HSV-2 encephalitis in adults are identical (see^{reference 26}), there are also examples of HSV-2 producing mild encephalitis.^26,28

Three of the four patients were immunocompromised: one was receiving steroid treatment for asthma (Y.F.), another for an exacerbation of COPD(A.S.), and the third (M.C.) had AIDS. It has been suggested that immunocompromised patients may have atypical presentations of HSVE.^29,30 In a carefully studied case of this type it was noted that an anergic patient had markedly decreased CNS inflammation and necrosis compared with that expected in an immunocompetent individual.³⁰ Another pathologic study of HSVE in a patient with AIDS suggested that the inflammatory response to HSV infection was diminished, although the amount of necrosis appeared to be comparable with that found in immunocompetent individuals.³¹

Two of the three immunocompromised patients (Patient M.C. with AIDS and Patient Y.F. with steroid-dependent asthma) did not have CSF pleocytosis on their initial LP. In one patient (Y.F.), pleocytosis was present on repeat LP. The second patient (M.C.) died without having a repeat LP. Because CSF pleocytosis is an almost invariant feature of HSVE in immunocompetent adults,^2,32,33 the absence of an initial pleocytosis in these two patients suggests that they failed to mount an inflammatory response within the CNS to HSV infection or that this was delayed. The presence of immune compromise in three of our four patients with atypical HSVE suggests that a vigorous inflammatory response contributes to CNS tissue injury and the consequent nature and severity of clinical disease.

Cranial MRI is the most sensitive test for defining the anatomic distribution of tissue injury in HSVE.^34-36 Cranial MRIs in two of our patients demonstrated their disease was primarily localized to the nondominant (right) temporal lobe, including the insular, hippocampal, and parahippoccampal gyri. In one patient(A.S.), FLAIR images demonstrated superior definition of the extent of temporal lobe abnormalities than did standard T1- and T2-weighted MR images, a finding that has also been noted by others.^23,24

Patients with severe neuropsychological impairment after HSVE typically have bilateral damage that involves the diencephalon, temporal lobe, hippocampus, and parahippocampus.³⁷ Among patients with predominantly unilateral disease, those with involvement of the dominant(left) temporal lobe often have sequelae that include dysnomia and impairment in new learning, memory, and verbal IQ.³⁸ Conversely, those with unilateral damage to the non-dominant temporal lobe appear to have better clinical and neuropsychological outcomes.³⁹ Sequelae in these patients may include decreased visual memory and diminished performance IQ.³⁸ The atypical encephalitis seen in two of our patients may reflect the fact that they had predominantly unilateral injury that involved the nondominant temporal lobe and adjacent structures. Patients with nonviral temporal lobe pathology show a similar hierarchy of clinical deficits, with the most pronounced impairment seen in patients with bilateral temporal lobe and hippocampal lesions and the least severe deficits seen in those with unilateral nondominant temporal lobe and hippocampal lesions.⁴⁰

In contradistinction to the prognosis in classic HSVE, in which less than 5% of survivors regain completely normal function,² three of our four patients with atypical encephalitis made a complete recovery. The fourth patient died of respiratory complications of AIDS.

We found that atypical forms of HSVE occurred in approximately 20% of patients. Clinically these patients presented with focal or generalized seizures and neurologic abnormalities limited to the mental status examination. None had focal neurologic deficits. This suggests that in patients with encephalopathy and lymphocytic CSF pleocytosis, CSF should be tested for the presence of HSV DNA by PCR even when focal features are not present on neurologic examination and CT in normal. Neuroimaging studies are exceedingly helpful in identifying patients with encephalitis. However, two patients in our study had HSVE despite the presence of both a non-focal neurologic examination and normal CT. MRIs were abnormal in the two patients in our series on whom they were performed. This suggests that MRI abnormalities occur even in patients with mild or atypical forms of HSVE. We did not find any examples of patients with encephalitis and lymphocytic CSF pleocytosis who had a positive HSV CSF PCR in the presence of a normal MRI. It is also important to recognize that in immunocompromised patients the signs and symptoms of HSVE may be atypical and may occur in the absence of initial CSF pleocytosis. Clinicians should consider performing HSV CSF PCR on this subset of patients when they present with encephalopathy, even in the absence of CSF pleocytosis or focal neurologic findings.