- Email: [email protected]
Acute central nervous system complications in varicella zoster virus infections
Journal of Clinical Virology 25 (2002) 293– 301 www.elsevier.com/locate/jcv
Acute central nervous system complications in varicella zoster virus infections Marjaleena Koskiniemi a,b,*, Heli Piiparinen b, Timo Rantalaiho a, Pekka Era¨nko¨ c, Markus Fa¨rkkila¨ b, Kirsti Ra¨iha¨ b, Eeva-Marjatta Salonen b, Pentti Ukkonen b, Antti Vaheri a,b, The Study Group 1 a
Department of Virology, Haartman Institute, Helsinki Uni6ersity, POB 21, FIN 00014 Helsinki, Finland b Helsinki Uni6ersity Central Hospital, Uni6ersity of Helsinki, Helsinki, Finland c Glaxo Wellcome Pharmaceutical Company, Helsinki, Finland
Abstract Background: In a previous multicenter study on central nervous system (CNS) viral infections varicella zoster virus (VZV) appeared the most frequent etiologic agent and appeared often without rash. Objecti6e: To evaluate the appearance and diagnostics of VZV in CNS more thoroughly, we studied the cases systematically by using sensitive and specific methods to learn the best diagnostic approach in order to start specific therapy. Study design: We analyzed all serum and cerebrospinal fluid samples of 174 patients, 88 females and 86 males, with acute CNS symptoms associated with VZV infection diagnosed in the multicenter study on viral CNS infections. Results: About 38 patients (22%) had chickenpox, 59 (34%) had shingles, and 77 (44%) had no cutaneous symptoms at all. The mean age of chickenpox patients was 8.6 years, of the others 46.6 and 41.4 years. VZV-specific nucleic acid was detected in the CSF in one fourth of the patients in all groups, primarily during the first week of illness. In serum specimens,
Abbre6iations: CNS, central nervous system; CSF, cerebrospinal fluid; HSV, herpes simplex virus; VZV, varicella zoster virus; PCR, polymerase chain reaction. * Corresponding author. Tel.: + 358-919-126-579; fax: + 358-919-126-491 E-mail address: [email protected] (M. Koskiniemi). 1 Study Group: Asko Ja¨rvinen, Esko Kinnunen, Suvi Koskiniemi, Laura Mannonen, Marketta Muttilainen, Kimmo Linnavuori, Jukka Porras, Mirja Puolakkainen, Ville Valtonen, Helsinki University Central Hospital; Marjo-Riitta Aine, Matti Junes, Central Hospital of Lapland, Rovaniemi; Vilho Myllyla¨, Heikki Rantala, Anne Saari, Oulu University Hospital, Oulu; Roger Byring, Lauri Hakamies, Harri Kivela¨, Bjarne Udd, Vaasa Central Hospital, Vaasa; Merja Kaislakoski, Kirsi Mustonen, Annamari Juntunen, Central Hospital of North Karelia, Joensuu; Heidi Fagerstro¨m, Dag Nyman, A, land Central Hospital, Mariehamn; Niilo-Pekka Huttunen, Taneli Sarasoja, Central Hospital of Central Finland, Jyva¨skyla¨; Eila Herrga˚rd, Matti Korppi, Juho Nuutinen, Kuopio University Hospital, Kuopio; Juhani Gro¨nlund, Ulla Kaski, Ritva Norja, Central Hospital of South Pohjanmaa, Seina¨joki; Christine Hedman, Ulla-Maija Ritanen-Mohammed, Central Hospital of Pa¨ija¨t-Ha¨me, Lahti; Helena Pesonen, Ari Saarinen, Kajaani Central Hospital, Kajaani; Ilkka Pieninkeroinen, Tiina Wallden, Central Hospital of Kymenlaakso, Kotka; Pentti Lautala, Raili Saari, Jussi Valpas, Central Hospital of South Karelia, Lappeenranta; Juhani Sankala, Hannu Kilpela¨inen, Savonlinna Central Hospital, Savonlinna; Pirkko Karttunen, Matti Pere, Anja Selonen, Jaakko Taalas, Mikkeli Central Hospital, Mikkeli; Seppo Mannila, Kari Uotila, Central Hospital of Kanta-Ha¨me, Ha¨meenlinna; Seppo Tuisku, Markus Va¨re, Central Hospital of Central Pohjanmaa, Kokkola; Finland. 1386-6532/02/$ - see front matter. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 2 ) 0 0 0 2 0 - 3
294
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
specific IgM was present in two thirds of the patients with chickenpox, whereas in the others in one third of the cases. In CSF, specific IgM was present in 15 – 17% of patients with skin manifestations, compared with 6% of those without rash. Conclusions: The role of VZV infections in CNS complications seems remarkable, often presenting without rash. Even these cases should be promptly recognized in order to conduct proper antiviral therapy. In children, a combination of PCR and IgM tests is the best approach. In adults, PCR, together with the measurement of intrathecal antibody production yields best results. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. Keywords: Central nervous system; Chickenpox; Shingles; Zoster; Varicella-zoster virus
1. Introduction
2. Patients and methods
Conventional childhood infections have been commonly attributed to central nervous system (CNS) diseases. Many microbes have been largely eradicated by vaccination programs (Beghi et al., 1984; Koskiniemi and Vaheri, 1989; Koskiniemi et al., 1991; Peltola et al., 1994; Heisler and Richmond, 1994). Concomitantly, CNS infections associated with varicella zoster virus (VZV) have been increasingly recognized among adults (Gershon et al., 1980; Weller, 1983; Echevarria et al., 1997; Gilden et al., 2000). It is important to identify the infections associated with VZV because the patients could benefit from antiviral therapy (Whitley, 1992; Meszner et al., 1993; Wood et al., 1994). Furthermore, the new diagnostic methods, the polymerase chain reaction (PCR) and sensitive and specific serological tests have remarkably widened knowledge on VZV infections (Puchhammer et al., 1991; Echevarria et al., 1994; Gilden et al., 2000). Several cases with CNS involvement but without classical skin manifestations have been described (Bale, 1993; Echevarria et al., 1994; Gilden et al., 1994b; Bergstro¨ m, 1995). Their impact in CNS diseases as a whole has not been defined. In a previous multicenter collaborative study on CNS infections in Finland, we have shown VZV to be today the microbe most frequently associated with various neurological complications (Koskiniemi et al., 2001). The present study of 174 patients with VZV-associated CNS diseases emphasizes the significance of a broad diagnostic approach.
2.1. Patients The study was approved by the Research Ethics Committee of Helsinki University Hospital, and by all the participating centers. During the years 1995–1996, we studied 3231 patients with neurologic symptoms of suspected viral etiology treated at neurologic, pediatric or infectious disease wards or intensive care units of 18 tertiary care hospitals covering 77% of population in Finland and of whom serum and cerebrospinal fluid (CSF) specimens were sent for viral studies to the Department of Virology, University of Helsinki (Koskiniemi et al., 2001). From 174 patients with VZV associated confirmed or probable infection serum and CSF samples were available for systematic studies for VZV. VZV-associated infection was regarded as confirmed if VZV-specific nucleic acid was detected in the CSF or VZV-specific intrathecal antibody production (the ratio of serum/CSF VZV-specific antibodies was 5 20 by end-point titers or B 10 in EIUs, whereas other antibodies were negative in CSF), or VZV-specific IgM was detected in CSF. The diagnosis was probable if seroconversion or a ] 40 EIUs increase in VZV-specific antibody level occurred, VZV-specific IgM antibody was observed in serum, or a characteristic rash appeared within 4 weeks of the CNS symptoms. Based on symptoms on admission, the clinical diagnoses distributed into six mutually exclusive categories as follows: encephalitis (symptoms and signs indicative of involvement of the brain parenchyma, such as focal or generalized seizures,
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
motor or sensory pareses, coma, mental function impairement, depression of consciousness), meningitis (meningeal signs without signs of the brain parenchyma involvement), myelitis (signs indicative of medullary dysfunction), Guillain Barre´ polyradiculitis (flaccid paresis evaluated by a neurologist), facial paresis (Bell’s palsy), and discrete symptoms without clinical disease entity (only a single sign/symptom was present such as dizziness, ataxia, convulsions). Polyradiculitis and facial paresis were included to the study because the viral diagnostic approach is similar to the CNS infections, and occasionally, CNS involvement in these diseases can be seen. In all, CSF and serum were obtained on clinical basis for suspected viral neurological infection, usually on admission to hospital, and repeated, if clinically relevant.
295
antibodies were negative in the CSF, the serum and CSF specimens were retested to determine the end-point titers, and the ratio of serum and CSF antibodies was calculated. The ratio of the endpoint titers B 1:20 and of EIU titers B 10 were regarded. HSV- and VZV-specific IgMs were determined by using a commercial kit for testing indirect IgM immunofluorescence (Gull Laboratories, Salt Lake City, UT) from serum (dilution 1:10 and 1:20) and from CSF (dilution 1:2). Positive samples were retested after removal of IgG (Gullsorp absorption).
2.3. Nucleic acid detection PCR assays for HSV-1, HSV-2 (on 148 patients), and VZV (166 patients) were performed on all CSF samples available. The primers were selected from the DNA polymerase genes. The PCR product was detected using luminometric microplate hybridization (Piiparinen and Vaheri, 1991 Koskiniemi et al., 1997, 2001).
2.2. Serologic tests Antibodies to herpes simplex virus-1 (HSV-1), HSV-2, VZV, RSV (commercial antigens, Virion®, Wu¨ rzburg, Germany), adeno, influenza A and B, rota, coxsackie B5, nontyped entero and parainfluenza 1 (antigens grown and purified as described), EBV, HHV-6, TBE, Nephropathia epidemica, Chlamydia pneumoniae, and Mycoplasma pneumoniae were measured from serum and CSF as described (Koskiniemi et al., 2001). If specific antibodies were detected and all other
3. Results
3.1. Chickenpox Chickenpox or characteristic rash 5 4 weeks before or soon after the acute CNS symptomatology was observed in 38 patients, in most the
Table 1 Different VZV-associated CNS diseases and symptoms in relation to skin manifestations Admission diagnosis
Encephalitis Meningitis Myelitis Guillain barre´ Facial paresis Discrete symptoms 2661 All
No rash
Shingles
Chickenpox
All
N
%
N
%
N
%
N
2552 1762 4 2 31 65 77
27.2 65.4 100 66.7 33.3 11 44.3
37 7 0 1 3 27.5 59
40.2 26.9 0 33.3 33.3 3 33.9
30 2 0 0 3 7.5 38
32.6 7.7 0 0 33.3 40 21.9
92 26 4 3 9 174
In the patient group without skin manifestations, the PCR positivity to VZV is indicated by the number in the right superscript, and the CSF-VZV IgM positivity by the number in the right subscript.
296
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
Fig. 1. (a – d) Age distribution of the patients with VZV infection in relation to skin manifestations (no rash, shingles or chickenpox). Shadowed area indicates VZV-PCR positivity. In patients with no rash, VZV-specific nucleic acid is especially detected in middle aged and elderly patients. In shingles, the PCR positivity is distributed evenly, and in patients with chickenpox, VZV-PCR is positive only in the age group B10 years.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
297
Fig. 2. The distribution of VZV-specific diagnostic findings (CSF-VZV-PCR, CSF-VZV-IgM, serum VZV-IgM, intrathecal antibody production) in relation to skin manifestations (no rash, shingles, chickenpox). IT-ab =intrathecal antibody production/antibodies to VZV in CSF. All samples are negative for HSV-1 and HSV-2 specific nucleic acid. Usually only one test per patient was positive. In children with chickenpox, serum VZV-IgM was positive in two thirds of all children, and CSF VZV-IgM was present in 15%, whereas the specific intrathecal IgG antibody production is only occasionally seen. In the other patient groups, intrathecal antibody production is diagnostic in 50 –80% of patients.
admission diagnosis was encephalitis (Table 1). The mean age was 8.6 years with a range of B 1–40 years (Fig. 1). In the 10 VZV-PCR positive patients the mean age was clearly less, 4.1 years. In addition to PCR, the diagnosis was based on VZV-IgM in CSF in six patients (four had positive PCR concomitantly) (Fig. 2). Serum VZV-IgM was positive in 20 patients but in only two of the PCR-positive group. Intrathecal VZV-IgG antibody production was detected in only two (5%).
3.2. Shingles/zoster Shingles was observed in 59 patients altogether, in 37 of them with encephalitis (Table 1). In all patients the zoster was dermatomal. None of the cases showed disseminated zoster. The mean age was 46.6 years (range of 1– 80 years) (Fig. 1). CSF VZV-PCR was positive in 12 and VZV-IgM was detected in CSF in nine patients (four had positive PCR concomitantly) (Fig. 2). The diagnosis was based on intrathecal antibody production to VZV in 33 patients
(56%). VZV-IgM was positive in serum in 17 of the 47 studied patients.
3.3. No skin manifestations In 77 patients no skin manifestations were observed during the disease or within 4 weeks of onset of various CNS complications (Table 1). The mean age of these patients was 41.4 years with a range from 1 to 77 years (Fig. 1). VZVPCR was positive in 18 patients, VZV-IgM in CSF was present in four, and in serum in 19 of the 57 studied patients. Intrathecal antibody production appeared in 81%. The spectrum of symptoms and diseases associated with this type of VZV infection was wide, mostly categorized into discrete symptoms. Remarkably, in three previously healthy patients, development of multiple sclerosis was suspected based on characteristic plaques in the MRI of the brain. In addition, the disease seemed to be more severe in patients without rash, about 18% needed intensive care on admission compared with 12% in shingles and 3% in chickenpox.
298
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
3.4. Impact of diagnostic methods VZV-specific nucleic acid was detected in 40 patients of 151 studied (26.5%), in about the same proportion in all three groups (Fig. 3). All specimens were tested negative for HSV-1 and HSV-2 DNA. Clearly, in chickenpox and shingles the nucleic acid findings were obtained during the first week of symptom onset, while in patients without rash even samples from the second week appeared occasionally positive. In meningitis patients without rash, the PCR positivity appeared more frequently than in the other disease groups. CSF VZV-IgM was positive in 15– 17% of the patients with rash and in the others the percentage was 6 (Fig. 2). In up to 80% of adult patients, intrathecal antibody production against VZV was observed and thus it was the most powerful diagnostic tool. In three children with chickenpox, adeno- or echovirus was concomitantly found in a throat swab. In nine patients (5%), intrathecal antibody production appeared to HSV besides VZV.
4. Discussion We used PCR analysis and antibody testing from CSF samples to confirm the role of VZV in producing disorders of the CNS. The number of CNS complications in 44% of patients without skin manifestations was conspicuous. Similar cases have been reported in the literature (Gilden et al., 1994b; Bergstro¨ m, 1995; Echevarria et al., 1997), but the big proportion of ‘sine herpete’ cases is a new finding. The microbiological diagnosis was confirmed by CSF-PCR, intrathecal antibody production to VZV or specific IgM in the CSF. The specificity of the finding was confirmed and the damage of blood– brain barrier was excluded by using a large panel of antibodies. The concurrent presence of HSV and VZV antibodies
in CSF in a few patients may be a coreactivation, possibly even a dual infection. In general, the severe nature of VZV may easily run unnoticed and the diagnosis may be missed in cases without rash, if not especially searched for (Fleisher et al., 1981; Weller, 1983; Preblud et al., 1984; Law et al., 2000). Our data present an exceptional spectrum of neurological complications associated with VZV infections. Encephalitis was most frequent, but besides it myelitis, polyradiculitis, facial paresis and various single symptoms were observed. These features have been described earlier in chickenpox and zoster associated CNS complications, but in regard to VZV without rash frequent complications are a novel finding (Appelbaum et al., 1953; Jemsek et al., 1983; Gilden et al., 1994b). Histopathologically, small and large-vessel alterations have been described with necrosis and demyelination, which may explain the large spectrum (Gershon et al., 1980; Horten et al., 1981; Gilden et al., 1994a; Amlie-Lefond et al., 1995; Gilden et al., 1996, 2000). Remarkably, three 26–34-year-old, previously healthy patients with no skin manifestations, but a confirmed VZV infection, had signs of demyelination in MRI of the brain suggesting the possibility of multiple sclerosis. Our series shows the relevance of measuring intrathecal antibody production, specific antibodies from CSF in parallel with serum antibodies. The specificity of intrathecal antibody production was confirmed by using a large panel of antigens. The benefit of detecting intrathecal antibody production and specific nucleic acid has clearly been documented since the 1980s and 1990s (Gershon et al., 1980; Puchhammer et al., 1991; Gilden et al., 1998), but the combined use is seldom applied. In our series, PCR gave one fourth of diagnoses independently of the presence of rash. Measurement of intrathecal antibody production was especially useful in adult patients without rash, and
Fig. 3. The distribution of patients in different groups (no rash, shingles, chickenpox) shown as days after the onset of symptoms and by CSF-VZV PCR findings. Shadowed area indicates VZV-PCR positivity and white area negative PCR findings. The PCR positivity accumulates to the first week after the onset of symptoms, except in patients without rash, VZV-specific nucleic acid can be detected after the first week as well.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
Fig. 3.
299
300
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
even in those with shingles intrathecal antibody production appeared in half of the cases. In chickenpox, the presence of CSF IgM was a valuable tool in combination with PCR (Echevarria et al., 1994; Gilden et al., 2000). The mutual relationship between PCR, intrathecal antibody production and IgM may be complicated. They do not exclude each other, but their appearance may be sequential. All these methods are needed in comprehensive diagnostics. Especially in children, IgM is a valuable addition to PCR. CNS involvement in association with chickenpox is difficult to evaluate. Traditionally, it has been regarded as a postinfectious immunological disease. By using new sensitive tests the CNS complications have often been proved to be viral meaning a direct invasion of the virus to the CNS. The frequent presence of IgM in serum has been reported earlier, especially after the fourth day of illness, both in primary and reactivated VZV infections. However, it does not confirm the CNS disease (Sundqvist, 1982; Weller, 1983), but in association with the characteristic rash it clearly supports the VZV etiology. The CNS symptoms with confirmed VZV infection without rash emerged, was a conspicuous finding. Nearly half of all cases had no cutaneous symptoms and most of them were previously healthy immunocompetent people. This suggests frequent activation of VZV without skin manifestations even in healthy persons and supports the findings reported before (Echevarria et al., 1994; Gilden et al., 1994b; Bergstro¨ m, 1995). In studying CNS complications this possibility should be taken into account. By using sensitive antibody tests and nucleic acid detection the diagnosis can be obtained, and proper antiviral treatment started. The high proportion of VZV infections without rash gives a new perspective to vaccination programs as well (Law et al., 2000). It is time to realize that VZV infection is more than chickenpox or zoster. Specific therapy is available to VZV. It underlines the importance of correct diagnosis (Whitley, 1992; Wood et al., 1994). A combination of sensitive tests, as recently recommended by (Gilden et al., 2000) is a reliable way to obtain viral diagnosis even on admission to hospital and specific therapy can be readily started.
Acknowledgements This study was supported by the Helsinki University Hospital Research Funds and the GlaxoWellcome Pharmaceutical Company.
References Amlie-Lefond C, Kleinschmidt-DeMasters BK, Mahalingam R, Davis LE, Gilden DH. The vasculopathy of varicella-zoster virus encephalitis. Ann Neurol 1995;37:784 – 90. Appelbaum E, Rachelson MH, Dolgopol VB. Varicella encephalitis. Am J Med 1953;5:223 – 30. Bale JF. Viral encephalitis. Med Clin North Am 1993;77:25 – 42. Beghi E, Nicolosi A, Kurland LT, et al. Encephalitis and aseptic meningitis, Olmsted County, Minnesota, 1950 – 1981: I. Epidemiology. Ann Neurol 1984;6:283 – 94. Bergstro¨ m T. Polymerase chain reaction for diagnosis of varicella zoster virus central nervous system infections without skin manifestations. Scand J Infect Dis Suppl 1995;100:41 – 5. Echevarria JM, Casas I, Tenorio A, et al. Detection of varicellazoster virus-specific DNA sequences in cerebrospinal fluid from patients with acute aseptic meningitis and no cutaneous lesions. J Med Virol 1994;43:331 – 5. Echevarria JM, Casas I, Martinez-Martin P. Infections of the nervous system caused by varicella-zoster virus: a review. Intervirology 1997;40:72 – 84. Fleisher G, Henry W, McSorley M, et al. Life threatening complications of varicella. Am J Dis Child 1981;135:896 – 9. Gershon A, Steinberg S, Greenberg S, Taber L. Varicella-zosterassociated encephalitis: detection of specific antibody in cerebrospinal fluid. J Clin Microbiol 1980;12:764 – 7. Gilden DH, Beinlich BR, Rubinstein EM, et al. Varicella-zoster virus myelitis; an expanding spectrum. Neurology 1994a;44:1818 – 23. Gilden DH, Wright RR, Schneck SA, et al. Zoster sine herpete, a clinical variant. Ann Neurol 1994b;35:530 – 3. Gilden DH, Kleinschmidt-DeMasters BK, Wellish M, et al. Varicella-zoster virus, a cause of waxing and waning vasculitis. Neurology 1996;47:1441 – 6 The New England Journal of Medicine case 5-1995 revisited. Gilden DH, Bennett JL, Kleinschmidt-DeMasters BK, et al. The value of cerebrospinal fluid antiviral antibody in the diagnosis of neurologic disease produced by varicella-zoster virus. J Neurol Sci 1998;59:140 – 4. Gilden DH, Kleinschmidt-DeMasters BK, LaGuardia JJ, et al. Neurologic complications of the reactivation of varicellazoster virus. New Engl J Med 2000;342:635 – 45. Heisler MB, Richmond JB. Lessons from Finlands successful immunization program. New Engl J Med 1994;331:1446 – 7. Horten B, Price RW, Jimenez D. Multifocal varicella-zoster virus leukoencephalitis temporally remote from herpes zoster. Ann Neurol 1981;9:251 – 66.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301 Jemsek J, Greenberg SB, Taber L, et al. Herpes zoster-associated encephalitis: clinicopathologic report of 12 cases and review of the literature. Medicine (Baltimore) 1983;62:81 – 97. Koskiniemi M, Vaheri A. Effect of measles, mumps, rubella vaccination on pattern of encephalitis in children. Lancet 1989;1:31– 4. Koskiniemi M, Rautonen J, Lehtokoski-Lehtiniemi E, Vaheri A. Epidemiology of encephalitis in children: a 20-year survey. Ann Neurol 1991;29:492 –7. Koskiniemi M, Mannonnen L, Kallio A, Vaheri A. Luminometric microplate hybridization for detection of Varicellazoster virus PCR product from cerebrospinal fluid. J Virol Methods 1997;63:71 –9. Koskiniemi M, Rantalaiho T, Piiparinen H, et al. Infections of the central nervous system of suspected viral origin: a collaborative study from Finland. J Neurovirol 2001;7:400– 8. Law B, MacDonald N, Jalpern S, et al. The immunization monitoring program active (IMPACT) prospective five year study of Canadian children hospitalized for chickenpox or associated complications. Pediatr Infect Dis J 2000;19:053– 1059. Meszner Z, Nyerges G, Bell AR. Oral acyclovir to prevent dissemination of varicella in immunocomprimised children. J Infect 1993;26:9 – 15.
301
Peltola H, Heinonen OP, Valle M, et al. The elimination of indigenous measles, mumps, and rubella from Finland by a 12-year, two-dose vaccination program. New Engl J Med 1994;31:1397 – 402. Piiparinen H, Vaheri A. Genotyping of herpes viruses by polymerase chain reaction. Arch Virol 1991;119:275 – 83. Preblud SR, Orenstein WA, Bart KJ. Varicella: clinical manifestations, epidemiology and health impact in children. Pediatr Infect Dis J 1984;3:505 – 9. Puchhammer E, Popow-Kraupp T, Heinz F, et al. Detection of varicella-zoster virus DNA by polymerase chain reaction in the cerebrospinal fluid of patients suffering from neurological complications associated with chicken pox or herpes zoster. J Clin Microbiol 1991;29:1513 – 6. Sundqvist VA. Frequency and specificity of varicella zoster virus IgM response. J Virol Methods 1982;5:219 – 27. Weller TH. Varicella and herpes zoster. Changing concepts of the natural history, control, and importance of a Not-SoBenign virus (second of two parts). New Engl J Med 1983;309:1362 – 8. Whitley RJ. Therapeutic approaches to varicella-zoster virus infections. J Infect Dis 1992;166(Suppl. 1):551 – 7. Wood MJ, Johnson RW, McKendrick MW, et al. A randomized trial of acyclovir for 7 or 21 days with and without prednisolone for treatment of acute herpes zoster. New Engl J Med 1994;330:896 – 900.
Acute central nervous system complications in varicella zoster virus infections Marjaleena Koskiniemi a,b,*, Heli Piiparinen b, Timo Rantalaiho a, Pekka Era¨nko¨ c, Markus Fa¨rkkila¨ b, Kirsti Ra¨iha¨ b, Eeva-Marjatta Salonen b, Pentti Ukkonen b, Antti Vaheri a,b, The Study Group 1 a
Department of Virology, Haartman Institute, Helsinki Uni6ersity, POB 21, FIN 00014 Helsinki, Finland b Helsinki Uni6ersity Central Hospital, Uni6ersity of Helsinki, Helsinki, Finland c Glaxo Wellcome Pharmaceutical Company, Helsinki, Finland
Abstract Background: In a previous multicenter study on central nervous system (CNS) viral infections varicella zoster virus (VZV) appeared the most frequent etiologic agent and appeared often without rash. Objecti6e: To evaluate the appearance and diagnostics of VZV in CNS more thoroughly, we studied the cases systematically by using sensitive and specific methods to learn the best diagnostic approach in order to start specific therapy. Study design: We analyzed all serum and cerebrospinal fluid samples of 174 patients, 88 females and 86 males, with acute CNS symptoms associated with VZV infection diagnosed in the multicenter study on viral CNS infections. Results: About 38 patients (22%) had chickenpox, 59 (34%) had shingles, and 77 (44%) had no cutaneous symptoms at all. The mean age of chickenpox patients was 8.6 years, of the others 46.6 and 41.4 years. VZV-specific nucleic acid was detected in the CSF in one fourth of the patients in all groups, primarily during the first week of illness. In serum specimens,
Abbre6iations: CNS, central nervous system; CSF, cerebrospinal fluid; HSV, herpes simplex virus; VZV, varicella zoster virus; PCR, polymerase chain reaction. * Corresponding author. Tel.: + 358-919-126-579; fax: + 358-919-126-491 E-mail address: [email protected] (M. Koskiniemi). 1 Study Group: Asko Ja¨rvinen, Esko Kinnunen, Suvi Koskiniemi, Laura Mannonen, Marketta Muttilainen, Kimmo Linnavuori, Jukka Porras, Mirja Puolakkainen, Ville Valtonen, Helsinki University Central Hospital; Marjo-Riitta Aine, Matti Junes, Central Hospital of Lapland, Rovaniemi; Vilho Myllyla¨, Heikki Rantala, Anne Saari, Oulu University Hospital, Oulu; Roger Byring, Lauri Hakamies, Harri Kivela¨, Bjarne Udd, Vaasa Central Hospital, Vaasa; Merja Kaislakoski, Kirsi Mustonen, Annamari Juntunen, Central Hospital of North Karelia, Joensuu; Heidi Fagerstro¨m, Dag Nyman, A, land Central Hospital, Mariehamn; Niilo-Pekka Huttunen, Taneli Sarasoja, Central Hospital of Central Finland, Jyva¨skyla¨; Eila Herrga˚rd, Matti Korppi, Juho Nuutinen, Kuopio University Hospital, Kuopio; Juhani Gro¨nlund, Ulla Kaski, Ritva Norja, Central Hospital of South Pohjanmaa, Seina¨joki; Christine Hedman, Ulla-Maija Ritanen-Mohammed, Central Hospital of Pa¨ija¨t-Ha¨me, Lahti; Helena Pesonen, Ari Saarinen, Kajaani Central Hospital, Kajaani; Ilkka Pieninkeroinen, Tiina Wallden, Central Hospital of Kymenlaakso, Kotka; Pentti Lautala, Raili Saari, Jussi Valpas, Central Hospital of South Karelia, Lappeenranta; Juhani Sankala, Hannu Kilpela¨inen, Savonlinna Central Hospital, Savonlinna; Pirkko Karttunen, Matti Pere, Anja Selonen, Jaakko Taalas, Mikkeli Central Hospital, Mikkeli; Seppo Mannila, Kari Uotila, Central Hospital of Kanta-Ha¨me, Ha¨meenlinna; Seppo Tuisku, Markus Va¨re, Central Hospital of Central Pohjanmaa, Kokkola; Finland. 1386-6532/02/$ - see front matter. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 2 ) 0 0 0 2 0 - 3
294
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
specific IgM was present in two thirds of the patients with chickenpox, whereas in the others in one third of the cases. In CSF, specific IgM was present in 15 – 17% of patients with skin manifestations, compared with 6% of those without rash. Conclusions: The role of VZV infections in CNS complications seems remarkable, often presenting without rash. Even these cases should be promptly recognized in order to conduct proper antiviral therapy. In children, a combination of PCR and IgM tests is the best approach. In adults, PCR, together with the measurement of intrathecal antibody production yields best results. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. Keywords: Central nervous system; Chickenpox; Shingles; Zoster; Varicella-zoster virus
1. Introduction
2. Patients and methods
Conventional childhood infections have been commonly attributed to central nervous system (CNS) diseases. Many microbes have been largely eradicated by vaccination programs (Beghi et al., 1984; Koskiniemi and Vaheri, 1989; Koskiniemi et al., 1991; Peltola et al., 1994; Heisler and Richmond, 1994). Concomitantly, CNS infections associated with varicella zoster virus (VZV) have been increasingly recognized among adults (Gershon et al., 1980; Weller, 1983; Echevarria et al., 1997; Gilden et al., 2000). It is important to identify the infections associated with VZV because the patients could benefit from antiviral therapy (Whitley, 1992; Meszner et al., 1993; Wood et al., 1994). Furthermore, the new diagnostic methods, the polymerase chain reaction (PCR) and sensitive and specific serological tests have remarkably widened knowledge on VZV infections (Puchhammer et al., 1991; Echevarria et al., 1994; Gilden et al., 2000). Several cases with CNS involvement but without classical skin manifestations have been described (Bale, 1993; Echevarria et al., 1994; Gilden et al., 1994b; Bergstro¨ m, 1995). Their impact in CNS diseases as a whole has not been defined. In a previous multicenter collaborative study on CNS infections in Finland, we have shown VZV to be today the microbe most frequently associated with various neurological complications (Koskiniemi et al., 2001). The present study of 174 patients with VZV-associated CNS diseases emphasizes the significance of a broad diagnostic approach.
2.1. Patients The study was approved by the Research Ethics Committee of Helsinki University Hospital, and by all the participating centers. During the years 1995–1996, we studied 3231 patients with neurologic symptoms of suspected viral etiology treated at neurologic, pediatric or infectious disease wards or intensive care units of 18 tertiary care hospitals covering 77% of population in Finland and of whom serum and cerebrospinal fluid (CSF) specimens were sent for viral studies to the Department of Virology, University of Helsinki (Koskiniemi et al., 2001). From 174 patients with VZV associated confirmed or probable infection serum and CSF samples were available for systematic studies for VZV. VZV-associated infection was regarded as confirmed if VZV-specific nucleic acid was detected in the CSF or VZV-specific intrathecal antibody production (the ratio of serum/CSF VZV-specific antibodies was 5 20 by end-point titers or B 10 in EIUs, whereas other antibodies were negative in CSF), or VZV-specific IgM was detected in CSF. The diagnosis was probable if seroconversion or a ] 40 EIUs increase in VZV-specific antibody level occurred, VZV-specific IgM antibody was observed in serum, or a characteristic rash appeared within 4 weeks of the CNS symptoms. Based on symptoms on admission, the clinical diagnoses distributed into six mutually exclusive categories as follows: encephalitis (symptoms and signs indicative of involvement of the brain parenchyma, such as focal or generalized seizures,
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
motor or sensory pareses, coma, mental function impairement, depression of consciousness), meningitis (meningeal signs without signs of the brain parenchyma involvement), myelitis (signs indicative of medullary dysfunction), Guillain Barre´ polyradiculitis (flaccid paresis evaluated by a neurologist), facial paresis (Bell’s palsy), and discrete symptoms without clinical disease entity (only a single sign/symptom was present such as dizziness, ataxia, convulsions). Polyradiculitis and facial paresis were included to the study because the viral diagnostic approach is similar to the CNS infections, and occasionally, CNS involvement in these diseases can be seen. In all, CSF and serum were obtained on clinical basis for suspected viral neurological infection, usually on admission to hospital, and repeated, if clinically relevant.
295
antibodies were negative in the CSF, the serum and CSF specimens were retested to determine the end-point titers, and the ratio of serum and CSF antibodies was calculated. The ratio of the endpoint titers B 1:20 and of EIU titers B 10 were regarded. HSV- and VZV-specific IgMs were determined by using a commercial kit for testing indirect IgM immunofluorescence (Gull Laboratories, Salt Lake City, UT) from serum (dilution 1:10 and 1:20) and from CSF (dilution 1:2). Positive samples were retested after removal of IgG (Gullsorp absorption).
2.3. Nucleic acid detection PCR assays for HSV-1, HSV-2 (on 148 patients), and VZV (166 patients) were performed on all CSF samples available. The primers were selected from the DNA polymerase genes. The PCR product was detected using luminometric microplate hybridization (Piiparinen and Vaheri, 1991 Koskiniemi et al., 1997, 2001).
2.2. Serologic tests Antibodies to herpes simplex virus-1 (HSV-1), HSV-2, VZV, RSV (commercial antigens, Virion®, Wu¨ rzburg, Germany), adeno, influenza A and B, rota, coxsackie B5, nontyped entero and parainfluenza 1 (antigens grown and purified as described), EBV, HHV-6, TBE, Nephropathia epidemica, Chlamydia pneumoniae, and Mycoplasma pneumoniae were measured from serum and CSF as described (Koskiniemi et al., 2001). If specific antibodies were detected and all other
3. Results
3.1. Chickenpox Chickenpox or characteristic rash 5 4 weeks before or soon after the acute CNS symptomatology was observed in 38 patients, in most the
Table 1 Different VZV-associated CNS diseases and symptoms in relation to skin manifestations Admission diagnosis
Encephalitis Meningitis Myelitis Guillain barre´ Facial paresis Discrete symptoms 2661 All
No rash
Shingles
Chickenpox
All
N
%
N
%
N
%
N
2552 1762 4 2 31 65 77
27.2 65.4 100 66.7 33.3 11 44.3
37 7 0 1 3 27.5 59
40.2 26.9 0 33.3 33.3 3 33.9
30 2 0 0 3 7.5 38
32.6 7.7 0 0 33.3 40 21.9
92 26 4 3 9 174
In the patient group without skin manifestations, the PCR positivity to VZV is indicated by the number in the right superscript, and the CSF-VZV IgM positivity by the number in the right subscript.
296
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
Fig. 1. (a – d) Age distribution of the patients with VZV infection in relation to skin manifestations (no rash, shingles or chickenpox). Shadowed area indicates VZV-PCR positivity. In patients with no rash, VZV-specific nucleic acid is especially detected in middle aged and elderly patients. In shingles, the PCR positivity is distributed evenly, and in patients with chickenpox, VZV-PCR is positive only in the age group B10 years.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
297
Fig. 2. The distribution of VZV-specific diagnostic findings (CSF-VZV-PCR, CSF-VZV-IgM, serum VZV-IgM, intrathecal antibody production) in relation to skin manifestations (no rash, shingles, chickenpox). IT-ab =intrathecal antibody production/antibodies to VZV in CSF. All samples are negative for HSV-1 and HSV-2 specific nucleic acid. Usually only one test per patient was positive. In children with chickenpox, serum VZV-IgM was positive in two thirds of all children, and CSF VZV-IgM was present in 15%, whereas the specific intrathecal IgG antibody production is only occasionally seen. In the other patient groups, intrathecal antibody production is diagnostic in 50 –80% of patients.
admission diagnosis was encephalitis (Table 1). The mean age was 8.6 years with a range of B 1–40 years (Fig. 1). In the 10 VZV-PCR positive patients the mean age was clearly less, 4.1 years. In addition to PCR, the diagnosis was based on VZV-IgM in CSF in six patients (four had positive PCR concomitantly) (Fig. 2). Serum VZV-IgM was positive in 20 patients but in only two of the PCR-positive group. Intrathecal VZV-IgG antibody production was detected in only two (5%).
3.2. Shingles/zoster Shingles was observed in 59 patients altogether, in 37 of them with encephalitis (Table 1). In all patients the zoster was dermatomal. None of the cases showed disseminated zoster. The mean age was 46.6 years (range of 1– 80 years) (Fig. 1). CSF VZV-PCR was positive in 12 and VZV-IgM was detected in CSF in nine patients (four had positive PCR concomitantly) (Fig. 2). The diagnosis was based on intrathecal antibody production to VZV in 33 patients
(56%). VZV-IgM was positive in serum in 17 of the 47 studied patients.
3.3. No skin manifestations In 77 patients no skin manifestations were observed during the disease or within 4 weeks of onset of various CNS complications (Table 1). The mean age of these patients was 41.4 years with a range from 1 to 77 years (Fig. 1). VZVPCR was positive in 18 patients, VZV-IgM in CSF was present in four, and in serum in 19 of the 57 studied patients. Intrathecal antibody production appeared in 81%. The spectrum of symptoms and diseases associated with this type of VZV infection was wide, mostly categorized into discrete symptoms. Remarkably, in three previously healthy patients, development of multiple sclerosis was suspected based on characteristic plaques in the MRI of the brain. In addition, the disease seemed to be more severe in patients without rash, about 18% needed intensive care on admission compared with 12% in shingles and 3% in chickenpox.
298
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
3.4. Impact of diagnostic methods VZV-specific nucleic acid was detected in 40 patients of 151 studied (26.5%), in about the same proportion in all three groups (Fig. 3). All specimens were tested negative for HSV-1 and HSV-2 DNA. Clearly, in chickenpox and shingles the nucleic acid findings were obtained during the first week of symptom onset, while in patients without rash even samples from the second week appeared occasionally positive. In meningitis patients without rash, the PCR positivity appeared more frequently than in the other disease groups. CSF VZV-IgM was positive in 15– 17% of the patients with rash and in the others the percentage was 6 (Fig. 2). In up to 80% of adult patients, intrathecal antibody production against VZV was observed and thus it was the most powerful diagnostic tool. In three children with chickenpox, adeno- or echovirus was concomitantly found in a throat swab. In nine patients (5%), intrathecal antibody production appeared to HSV besides VZV.
4. Discussion We used PCR analysis and antibody testing from CSF samples to confirm the role of VZV in producing disorders of the CNS. The number of CNS complications in 44% of patients without skin manifestations was conspicuous. Similar cases have been reported in the literature (Gilden et al., 1994b; Bergstro¨ m, 1995; Echevarria et al., 1997), but the big proportion of ‘sine herpete’ cases is a new finding. The microbiological diagnosis was confirmed by CSF-PCR, intrathecal antibody production to VZV or specific IgM in the CSF. The specificity of the finding was confirmed and the damage of blood– brain barrier was excluded by using a large panel of antibodies. The concurrent presence of HSV and VZV antibodies
in CSF in a few patients may be a coreactivation, possibly even a dual infection. In general, the severe nature of VZV may easily run unnoticed and the diagnosis may be missed in cases without rash, if not especially searched for (Fleisher et al., 1981; Weller, 1983; Preblud et al., 1984; Law et al., 2000). Our data present an exceptional spectrum of neurological complications associated with VZV infections. Encephalitis was most frequent, but besides it myelitis, polyradiculitis, facial paresis and various single symptoms were observed. These features have been described earlier in chickenpox and zoster associated CNS complications, but in regard to VZV without rash frequent complications are a novel finding (Appelbaum et al., 1953; Jemsek et al., 1983; Gilden et al., 1994b). Histopathologically, small and large-vessel alterations have been described with necrosis and demyelination, which may explain the large spectrum (Gershon et al., 1980; Horten et al., 1981; Gilden et al., 1994a; Amlie-Lefond et al., 1995; Gilden et al., 1996, 2000). Remarkably, three 26–34-year-old, previously healthy patients with no skin manifestations, but a confirmed VZV infection, had signs of demyelination in MRI of the brain suggesting the possibility of multiple sclerosis. Our series shows the relevance of measuring intrathecal antibody production, specific antibodies from CSF in parallel with serum antibodies. The specificity of intrathecal antibody production was confirmed by using a large panel of antigens. The benefit of detecting intrathecal antibody production and specific nucleic acid has clearly been documented since the 1980s and 1990s (Gershon et al., 1980; Puchhammer et al., 1991; Gilden et al., 1998), but the combined use is seldom applied. In our series, PCR gave one fourth of diagnoses independently of the presence of rash. Measurement of intrathecal antibody production was especially useful in adult patients without rash, and
Fig. 3. The distribution of patients in different groups (no rash, shingles, chickenpox) shown as days after the onset of symptoms and by CSF-VZV PCR findings. Shadowed area indicates VZV-PCR positivity and white area negative PCR findings. The PCR positivity accumulates to the first week after the onset of symptoms, except in patients without rash, VZV-specific nucleic acid can be detected after the first week as well.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
Fig. 3.
299
300
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301
even in those with shingles intrathecal antibody production appeared in half of the cases. In chickenpox, the presence of CSF IgM was a valuable tool in combination with PCR (Echevarria et al., 1994; Gilden et al., 2000). The mutual relationship between PCR, intrathecal antibody production and IgM may be complicated. They do not exclude each other, but their appearance may be sequential. All these methods are needed in comprehensive diagnostics. Especially in children, IgM is a valuable addition to PCR. CNS involvement in association with chickenpox is difficult to evaluate. Traditionally, it has been regarded as a postinfectious immunological disease. By using new sensitive tests the CNS complications have often been proved to be viral meaning a direct invasion of the virus to the CNS. The frequent presence of IgM in serum has been reported earlier, especially after the fourth day of illness, both in primary and reactivated VZV infections. However, it does not confirm the CNS disease (Sundqvist, 1982; Weller, 1983), but in association with the characteristic rash it clearly supports the VZV etiology. The CNS symptoms with confirmed VZV infection without rash emerged, was a conspicuous finding. Nearly half of all cases had no cutaneous symptoms and most of them were previously healthy immunocompetent people. This suggests frequent activation of VZV without skin manifestations even in healthy persons and supports the findings reported before (Echevarria et al., 1994; Gilden et al., 1994b; Bergstro¨ m, 1995). In studying CNS complications this possibility should be taken into account. By using sensitive antibody tests and nucleic acid detection the diagnosis can be obtained, and proper antiviral treatment started. The high proportion of VZV infections without rash gives a new perspective to vaccination programs as well (Law et al., 2000). It is time to realize that VZV infection is more than chickenpox or zoster. Specific therapy is available to VZV. It underlines the importance of correct diagnosis (Whitley, 1992; Wood et al., 1994). A combination of sensitive tests, as recently recommended by (Gilden et al., 2000) is a reliable way to obtain viral diagnosis even on admission to hospital and specific therapy can be readily started.
Acknowledgements This study was supported by the Helsinki University Hospital Research Funds and the GlaxoWellcome Pharmaceutical Company.
References Amlie-Lefond C, Kleinschmidt-DeMasters BK, Mahalingam R, Davis LE, Gilden DH. The vasculopathy of varicella-zoster virus encephalitis. Ann Neurol 1995;37:784 – 90. Appelbaum E, Rachelson MH, Dolgopol VB. Varicella encephalitis. Am J Med 1953;5:223 – 30. Bale JF. Viral encephalitis. Med Clin North Am 1993;77:25 – 42. Beghi E, Nicolosi A, Kurland LT, et al. Encephalitis and aseptic meningitis, Olmsted County, Minnesota, 1950 – 1981: I. Epidemiology. Ann Neurol 1984;6:283 – 94. Bergstro¨ m T. Polymerase chain reaction for diagnosis of varicella zoster virus central nervous system infections without skin manifestations. Scand J Infect Dis Suppl 1995;100:41 – 5. Echevarria JM, Casas I, Tenorio A, et al. Detection of varicellazoster virus-specific DNA sequences in cerebrospinal fluid from patients with acute aseptic meningitis and no cutaneous lesions. J Med Virol 1994;43:331 – 5. Echevarria JM, Casas I, Martinez-Martin P. Infections of the nervous system caused by varicella-zoster virus: a review. Intervirology 1997;40:72 – 84. Fleisher G, Henry W, McSorley M, et al. Life threatening complications of varicella. Am J Dis Child 1981;135:896 – 9. Gershon A, Steinberg S, Greenberg S, Taber L. Varicella-zosterassociated encephalitis: detection of specific antibody in cerebrospinal fluid. J Clin Microbiol 1980;12:764 – 7. Gilden DH, Beinlich BR, Rubinstein EM, et al. Varicella-zoster virus myelitis; an expanding spectrum. Neurology 1994a;44:1818 – 23. Gilden DH, Wright RR, Schneck SA, et al. Zoster sine herpete, a clinical variant. Ann Neurol 1994b;35:530 – 3. Gilden DH, Kleinschmidt-DeMasters BK, Wellish M, et al. Varicella-zoster virus, a cause of waxing and waning vasculitis. Neurology 1996;47:1441 – 6 The New England Journal of Medicine case 5-1995 revisited. Gilden DH, Bennett JL, Kleinschmidt-DeMasters BK, et al. The value of cerebrospinal fluid antiviral antibody in the diagnosis of neurologic disease produced by varicella-zoster virus. J Neurol Sci 1998;59:140 – 4. Gilden DH, Kleinschmidt-DeMasters BK, LaGuardia JJ, et al. Neurologic complications of the reactivation of varicellazoster virus. New Engl J Med 2000;342:635 – 45. Heisler MB, Richmond JB. Lessons from Finlands successful immunization program. New Engl J Med 1994;331:1446 – 7. Horten B, Price RW, Jimenez D. Multifocal varicella-zoster virus leukoencephalitis temporally remote from herpes zoster. Ann Neurol 1981;9:251 – 66.
M. Koskiniemi et al. / Journal of Clinical Virology 25 (2002) 293–301 Jemsek J, Greenberg SB, Taber L, et al. Herpes zoster-associated encephalitis: clinicopathologic report of 12 cases and review of the literature. Medicine (Baltimore) 1983;62:81 – 97. Koskiniemi M, Vaheri A. Effect of measles, mumps, rubella vaccination on pattern of encephalitis in children. Lancet 1989;1:31– 4. Koskiniemi M, Rautonen J, Lehtokoski-Lehtiniemi E, Vaheri A. Epidemiology of encephalitis in children: a 20-year survey. Ann Neurol 1991;29:492 –7. Koskiniemi M, Mannonnen L, Kallio A, Vaheri A. Luminometric microplate hybridization for detection of Varicellazoster virus PCR product from cerebrospinal fluid. J Virol Methods 1997;63:71 –9. Koskiniemi M, Rantalaiho T, Piiparinen H, et al. Infections of the central nervous system of suspected viral origin: a collaborative study from Finland. J Neurovirol 2001;7:400– 8. Law B, MacDonald N, Jalpern S, et al. The immunization monitoring program active (IMPACT) prospective five year study of Canadian children hospitalized for chickenpox or associated complications. Pediatr Infect Dis J 2000;19:053– 1059. Meszner Z, Nyerges G, Bell AR. Oral acyclovir to prevent dissemination of varicella in immunocomprimised children. J Infect 1993;26:9 – 15.
301
Peltola H, Heinonen OP, Valle M, et al. The elimination of indigenous measles, mumps, and rubella from Finland by a 12-year, two-dose vaccination program. New Engl J Med 1994;31:1397 – 402. Piiparinen H, Vaheri A. Genotyping of herpes viruses by polymerase chain reaction. Arch Virol 1991;119:275 – 83. Preblud SR, Orenstein WA, Bart KJ. Varicella: clinical manifestations, epidemiology and health impact in children. Pediatr Infect Dis J 1984;3:505 – 9. Puchhammer E, Popow-Kraupp T, Heinz F, et al. Detection of varicella-zoster virus DNA by polymerase chain reaction in the cerebrospinal fluid of patients suffering from neurological complications associated with chicken pox or herpes zoster. J Clin Microbiol 1991;29:1513 – 6. Sundqvist VA. Frequency and specificity of varicella zoster virus IgM response. J Virol Methods 1982;5:219 – 27. Weller TH. Varicella and herpes zoster. Changing concepts of the natural history, control, and importance of a Not-SoBenign virus (second of two parts). New Engl J Med 1983;309:1362 – 8. Whitley RJ. Therapeutic approaches to varicella-zoster virus infections. J Infect Dis 1992;166(Suppl. 1):551 – 7. Wood MJ, Johnson RW, McKendrick MW, et al. A randomized trial of acyclovir for 7 or 21 days with and without prednisolone for treatment of acute herpes zoster. New Engl J Med 1994;330:896 – 900.