|Home | About | Journals | Submit | Contact Us | Français|
Background.Since the introduction of live attenuated varicella zoster virus (VZV) vaccine in 1995 there has been a significant reduction in varicella incidence and its associated complications, but the impact on VZV-associated central nervous system (CNS) disease has not been assessed.
Methods.In this descriptive study we evaluated patients referred to the California Encephalitis Project from 1998 to 2009 with VZV PCR-positive cerebrospinal fluid (CSF). Epidemiological, clinical, and laboratory data were collected using a standardized case form. Specimens were genotyped using multi-single nucleotide polymorphism (SNP) analysis.
Results.Twenty-six specimens were genotyped from patients 12–85 years of age (median, 46 years). Clinical presentations included meningitis (50%), encephalitis (42%), and acute disseminated encephalomyelitis (ADEM) (8%). Only 11 patients (42%) had a concomitant herpes zoster rash. Genotype analysis identified 20 European Group (Clade1, Clade 3) strains; 4 Asian (Clade 2) strains, and 2 Mosaic Group (Clade 4, Clade VI) strains. One specimen was recognized as vaccine strain by identifying vaccine-associated SNPs.
Conclusions.VZV continues to be associated with CNS disease, with meningitis being the most frequent clinical presentation. CNS VZV disease often presented without accompanying zoster rash. Sequencing data revealed multiple genotypes, including 1 vaccine strain detected in the CSF of a young patient with meningitis.
Varicella vaccine was licensed in the United States in 1995 . According to the National Immunization Survey, single-dose varicella vaccine coverage in the United States reached 90% in 2007 . Although a single dose of varicella vaccine provided high levels of protection and greatly reduced the incidence of varicella, it was inadequate to prevent all varicella outbreaks and breakthrough disease among vaccinated children. To further reduce varicella disease incidence and its complications in the United States, a second dose of varicella vaccine was recommended in 2007 . The additional dose of vaccine is expected to confer improved protection for the 15% - –20% of children who responded inadequately to the first dose.
Even though complications involving the central nervous system (CNS) are rare following natural VZV infection, VZV was found to be the most common etiological agent associated with encephalitis, meningitis and myelitis prior to vaccine introduction . Significant reductions in varicella complications and deaths have been documented since the introduction of live attenuated VZV vaccine . The impact of varicella vaccination on CNS disease, however, has not been reported. Neurological complications of VZV infection are important to recognize due to the morbidity and mortality associated with them and potential benefit in some with antiviral treatment . While vaccine strain VZV can reactivate in rare instances to cause herpes zoster (HZ), only 6 cases of vaccine-associated VZV CNS disease have been reported in the literature [7–13]. Too few cases of vaccine-associated CNS diseases have been described to know whether the clinical presentation differs from disease caused by wild virus. Thus, genotyping and strain surveillance is essential in distinguishing wild type and vaccine strain infections.
The California Encephalitis Project (CEP) was initiated in 1998 to better understand the clinical and epidemiological characteristics of encephalitis in California. Referral of patients to the CEP is open to all hospitals and physicians in California. Although the original emphasis was on immunocompetent patients with encephalitis, specimens from immunocompromised patients and patients with a range of CNS disorders other than encephalitis are referred voluntarily by physicians throughout California because a standardized panel of diagnostic tests (including VZV PCR in CSF) is performed routinely by the State Laboratory.
The objectives of our study were to identify patients from the CEP whose cerebrospinal fluid (CSF) samples tested positive for VZV by PCR in order to (1) describe the clinical and laboratory features of VZV CNS disease in the post-vaccine era, (2) perform genotyping on all available samples to better understand the epidemiology of VZV CNS disease in California, and (3) identify VZV vaccine strain specimens and describe their clinical presentation.
Between July 1998 and July 2009, samples from 4290 patients were voluntarily referred to the CEP by physicians from 249 facilities throughout California. A core battery of tests was performed on specimens obtained from all patients referred. Details about CEP testing are described elsewhere . We retrospectively identified patients whose CSF samples tested positive for VZV: at least 1 CSF sample was received and tested from 94% of patients (n = 4021), and of these, 43 (1%) tested positive for VZV by PCR. Frozen aliquots of CSF were available on 33 of these patients and sent to the Centers for Disease Control and Prevention (CDC) for genotype analysis. The CDC was unable to genotype 2 specimens and 5 tested negative. Genotyping was successfully performed on the remaining 26 CSF specimens according to the following protocol: total DNA was extracted from CSF samples using the MagNa Pure LC automated DNA purification system with the Tissue DNA Purification Kit (Roche Diagnostics), and recovered in a final volume of 100 μL of elution buffer. To verify VZV positive specimens and to discriminate vaccine Oka from wild-type VZV, at least 4 FRET (Förster Resonance Energy Transfer)-based real-time polymerase chain reaction (PCR) protocols targeting 4 vaccine-associated single base polymorphisms in ORF38 (69348), ORF54 (94167), and ORF62 (106262, 107252, and 108111) were performed on the LightCycler platform (Roche Diagnostics) [15, 16]. Genotyping was performed using conventional PCR targeting polymorphic regions in ORF21, ORF22, and ORF50 as reported elsewhere [17, 18]. Automated DNA sequencing was performed with the ABI PRISM 377 Genetic Analyzer (Perkin Elmer). All base positions are based on the published complete genome sequence for Dumas strain (GenBank accession no. gi: 9625875). Clades of varicella are referred to using the newly revised nomenclature adopted in 2008 .
Clinical information was obtained from the CEP case history form. This form includes information on demographics, exposures, travel history, laboratory findings, and clinical characteristics. Additional clinical information was requested when data were missing. Immunization histories were requested for all pediatric patients. Neurologic syndromes were classified into 3 primary groups. Patients were considered to have encephalitis as per CEP case definition  if they were hospitalized with encephalopathy (defined by a depressed or altered level of consciousness lasting more than 24 h, lethargy, a personality change, or ataxia) with 2 or more of the following characteristics: fever, seizure, focal neurological findings, pleocytosis, or electroencephalography or neuroimaging findings consistent with encephalitis. Patients were classified as having meningitis or ADEM based on the Brighton Collaboration Case Definitions [20, 21]. Activities of the California Encephalitis Project were reviewed by the California Committee for the Protection of Human Subjects.
Demographic, clinical, laboratory, and imaging data for the 26 patients whose CSF specimens were confirmed positive for VZV and genotyped are summarized in Table 1. Individual patient characteristics are shown in Table 2. Patients ranged in age from 12 to 85 years old, with a median age of 46.5 years. Seven (27%) were <18 years of age, 10 (38%) were adults between the ages of 27–60, and 9 (35%) were over the age of 60. The majority of patients (53%) were white, non-Hispanic, followed by Hispanic (19%). Sixteen patients (62%) reported no underlying immune deficiency. However, 5 of those (19%) were >60 years and thus expected to have declining VZV-specific T cell-mediated immunity (immunosenescence) [22, 23]. Ten patients (38%) reported primary or secondary causes of immunodeficiency, including human immunodeficiency virus (HIV) (n = 5), immunosuppressant drugs (n = 4), and asplenia (n = 1).
Three different CNS presentations were observed: meningitis (50%), encephalitis (42%), and acute disseminated encephalomyelitis (ADEM) (8%). Meningitis was the most common CNS presentation, seen in 91% (10/11) of immunocompetent patients ≤60 years of age. Encephalitis was the most common CNS presentation, seen in 67% (10/15) of patients with immunodeficiency and/or immunosenescent. ADEM was only seen in 2 immunocompromised patients (HIV and immunosupressant drugs).
A complete CSF WBC differential was obtained in 23 of 26 patients (Table 1). Lymphocytic predominance was observed in 52% of patient cases, while monocytes predominated in 26% and neutrophils in 22%. One HIV positive patient had two CSF samples that tested positive for VZV 8 days apart despite acyclovir treatment.
Neuroimaging studies were obtained in 25 of 26 patients; 80% (20/25) were reported as normal (n = 17) or showing abnormalities consistent with old age (n = 3). Abnormal imaging findings (5/25) included 2 patients with diffuse meningeal enhancement; 1 patient with abnormal spinal cord signal as well as enhancement of the temporal lobe area (patient was HIV-positive and had positive PCR on CSF for cytomegalovirus [CMV], VZV, and herpes simplex virus [HSV]); 1 patient with increased signal in fronto-temporal region (CSF HSV PCR negative); and 1 patient with enhancing occipital lobe and thoracic/lumbar lesions as well as abnormal thickened nerve roots.
Eleven (42%) patients had a rash diagnosed by the referring clinicians as HZ, and 14 (54%) patients presented without any rash. One additional patient (4%) presented with a rash not diagnosed as HZ but described in a trigeminal distribution that we believe to be consistent with HZ. Of the 15 patients with immunodeficiency and/or immunosenescent, 53% (8/15) presented with HZ and 47% (7/15) presented without. Of the 11 patients reported to be immunocompetent and ≤60 years old, 36% (4/11) presented with a HZ rash and 64% (7/11) presented without.
Of the 7 pediatric patients (ages 12–16), 4 had a history of at least 1 dose of varicella vaccine. Two children had a history of varicella (10 months of age and 18 months of age) and thus were not immunized. The remaining child was unimmunized with no clinical history of varicella disease prior to this presentation.
VZV genotypes were defined using the SNP profiles illustrated in Figure 1. Genotype analysis identified 20 European Group strains (Clade 1, Clade 3); 4 Asian strains (Clade 2), 1 of which was Oka vaccine strain; and 2 Mosaic Group strains, 1 Clade 4 and 1 Clade VI. Of the 20 E strains, 9 were Clade 1 and 7 were Clade 3, but we were unable to amplify DNA from either ORF21 or ORF50 to discriminate Clade 1 from Clade 3 in 4 cases (Figure 2).
The Oka vaccine strain identified was isolated from a 12-year-old previously healthy girl who presented with HZ rash, fever, elevation in liver transaminases, and meningitis (Table 3). She was treated with 7 days of acyclovir and recovered uneventfully. Her vaccination history was remarkable for a single varicella vaccination at the age of 1 year. Identification of this Oka vaccine strain represents the seventh published, laboratory-confirmed association of varicella vaccine with later CNS disease, and the longest interval from vaccination to presentation (11 years) reported to date.
After implementation of routine universal varicella vaccination in early childhood in the United States, CNS disease caused by VZV still occurs and, albeit rarely, can result from the reactivation of vaccine virus. However, the epidemiology of CNS VZV disease appears to be changing. Studies performed in different regions of the world prior to licensure of the varicella vaccine found VZV to be a leading cause of childhood encephalitis [24–27]. Encephalitis was infrequently seen among our pediatric cohort (2 pediatric cases among the 17 without genotyping data presented with encephalitis). Even though acute cerebellar ataxia is the most common complication of VZV infection in children , no cases were reported in this cohort, most likely due to referral bias. All children with specimens genotyped in this study (n = 7) presented with meningitis. In contrast, VZV is now recognized as one of the leading causes of adult encephalitis. Indeed, in 2 recent encephalitis studies, VZV was the second most common etiology identified, second only to HSV [29, 30]. All patients older than 60 years (n = 9) with specimens genotyped in this study presented with encephalitis. Among our cases, VZV CNS disease frequently presented without the characteristic HZ rash. The clinical implications are important because VZV may not be considered in the differential diagnosis of a patient presenting without a rash with CNS disease as routinely as HSV, and antiviral medications thus may be prematurely discontinued once HSV is ruled out. In fact, VZV CNS disease can resemble HSV encephalitis clinically and on neuroimaging as illustrated by 2 of our patients who had temporal lobe abnormalities, a characteristic finding of HSV but not usually associated with VZV .
VZV isolates are classified based on their geographic distribution as European Group (Clade 1, Clade 3), Asian (Clade 2), and Mosaic Group (Clade 4, Clade 5, Clade VI, and Clade VII). There are currently at least 5 and possibly as many as 7 distinctive genotypes identified: Clades 1–5, and provisional genotypes Clade VI and Clade VII [18, 19, 32]. The vaccine strain belongs to Clade 2 and can be differentiated from wild-type VZV by identifying vaccine-associated single base pair polymorphisms as described in our methods. In a global survey of more than 300 VZV isolates, Clade 2 strains were identified only in Japan and Hawaii . Several more recent studies failed to detect Clade 2 isolates in multiple countries in Europe, but 13% of isolates evaluated in Australia were Clade 2 viruses [17, 18, 32, 33]. While the overall numbers of isolates evaluated worldwide remains relatively small, it is striking that Clade 2 viruses have never been observed in areas with predominantly European descent that have limited immigration and have not yet been shown to be circulating in the United States. The emerging premise is that countries with a history of European colonization that have received large numbers of Asian immigrants in recent decades are undergoing a shift in the distribution of VZV genotypes, with increasing numbers of Clade 2 strains. The observation of 3 wild-type Clade 2 isolates in this small study may support this notion. Of the 3 cases with wild-type clade 2 cases, all were young (likely reflecting more recent infection trends), one had no recorded ethnicity or travel history, and the other 2 cases, 1 Hispanic and 1 Asian/Pacific Islander, denied any international travel. More detailed information on these cases can be found in Table 2. Notably, the Clade 2 strains detected in this study as wild-type using ORF21/ORF22/ORF50 genotyping would have been identified as vaccine genotype using only ORF38/ORF54-based genotyping (BglI+, PstI−), a method that is still in use for the identification of vaccine strain among clinical isolates [34, 35]. Another important finding was the identification of Clade VI strain, previously not seen in the United States but found in Spain and France . Sequencing data from the 26 VZV isolates reported here show that all VZV genotypes can be involved in CNS disease.
Only 1 of 26 isolates genotyped in this study was identified as the Oka vaccine strain. Six additional cases of CNS disease associated with VZV vaccine strain have been reported in the literature (Table 3), with 2 additional cases of vaccine meningitis confirmed more recently, 1 in Japan (identified in 2008, unpublished data) and 1 in Minnesota (in press). The previously reported cases ranged in age from 15 months to 9 years old. Four children were previously healthy, and 2 were immunocompromised. Length of time from vaccine administration to presentation ranged from 3 months to 8 years. Our case is the oldest reported (12 years old) and had the longest interval to reactivation (11 years). Even though only 7 cases of vaccine-associated CNS diseases are described to date, it is interesting to note that most patients presented with meningitis, and all were preceded by HZ (median time of 4.5 days) prior to CNS symptoms. Based on this very limited number of cases, clinical presentation of vaccine strain reactivation does not appear to differ from wild-type VZV CNS disease, making the diagnosis of vaccine strain reactivation difficult based on clinical presentation alone. The small number of cases reported may also suggest that the vaccine strain is less virulent in the CNS than the wild-type virus. However, the actual number of cases due to VZV vaccine strain reactivation is most likely higher than the number of cases that have been laboratory-confirmed, making any conclusion premature. This highlights the importance of continuing genotyping and strain surveillance studies to distinguish wild-type and vaccine strain infections.
One of the important limitations of this study is that the CEP is not a population-based study. The CEP is more likely biased toward patients that are severely ill and that are diagnostically challenging . For example, cases that presented with a characteristic rash of VZV and CNS symptoms are likely underrepresented in this sample. Similarly immunocompromised hosts are likely underrepresented since the focus of the project is the immunocompetent host.
In summary, after implementation of routine universal varicella vaccination in the United States, all but 1 case of VZV associated CNS disease observed in this cohort study were due to reactivation of the wild-type virus. Sequencing data from the 26 VZV isolates reported here show that all VZV genotypes can be involved in CNS disease. Our report of a seventh patient with VZV vaccine strain reactivation is similar to cases previously reported, although the time to reactivation, causing HZ and meningitis in an otherwise healthy young girl, was 11 years. In contrast to wild-type VZV CNS illness, which most often causes encephalitis in children, VZV vaccine strain is more often associated with meningitis. This presentation cannot be differentiated from natural VZV CNS disease clinically, emphasizing the importance of sequencing and genotyping.
This study was funded through a subcontract with America's Health Insurance Plans (AHIP) under contract 200-2002-00732 from the Centers for Disease Control and Prevention (CDC) for the Clinical Immunization Safety Assessment Network and under CDC Emerging Infections Program (U50/CCU015546-09).
We thank all members of the CEP for their respective contributions to the project and Kay Radford at the CDC for performing PCR and genotyping.