Clinical Application of Reverse-Transcription Polymerase Chain Reaction and Intravenous Immunoglobulin for Enterovirus Encephalitis

Jpn. J. Infect. Dis., 61, 18-24, 2008 Original Article Clinical Application of Reverse-Transcription Polymerase Chain Reaction and Intravenous Immunoglobulin for Enterovirus Encephalitis Ming-Fang Cheng 1,4, Bao-Chen Chen 2,4, Tsi-Shu Huang 2,4, Kai-Sheng Hsieh 1,4, Shu-Nuan Chen 2,4 and Yung-Ching Liu 2,3,4 * 1 Department of Pediatrics, 2 Department of Microbiology, and 3 Section of Infectious Diseases, Department of Internal Medicine, Veterans General Hospital-Kaohsiung, Kaohsiung, and 4 National Yang-Ming University, Taipei, Taiwan (Received June 5, 2007. Accepted October 19, 2007) SUMMARY: Although polymerase chain reaction (PCR) is a highly sensitive procedure for the diagnosis of enteroviruses, it has never been systemically applied to the treatment of enteroviral encephalitis using intravenous immunoglobulin (IVIg). We conducted a 2-year randomized, controlled comparison of reverse transcription (RT)-PCR of cerebrospinal fluid (CSF) with traditional viral isolation to guide IVIg treatment. Seventy-five patients were enrolled and classified into three groups: one group with clinical manifestations of enteroviral infections and two without. The latter two groups were separated on the basis of whether IVIg treatment was guided by RT-PCR or virus culture assay. CSF specimens from the 18 confirmed cases of enteroviral encephalitis were RT-PCR positive for enterovirus in all but one case. Of the remaining 57 cases of nonenteroviral encephalitis, only 4 were positive for enterovirus RT-PCR. One patient in the group of IVIg treatment guided by viral isolation subsequently displayed a sequel of epilepsy. No patients in the IVIg treatment groups guided by RT-PCR had any neurological sequelae. In conclusion, the use of RT-PCR allowed rapid, sensitive, and specific detection of enteroviral RNA in CSF. When used to guide IVIg treatment, RT-PCR may shorten hospitalization and improve outcomes of patients with enteroviral encephalitis. INTRODUCTION An enterovirus (EV) epidemic in Taiwan in 1998, predominantly caused by EV 71 in 42% of cases, coxsackievirus A16 in 18%, and other EVs in 40%, resulted in a total of 129,106 cases reported to have the primary clinical marker of enteroviral infection--herpangina and hand-foot-mouthdisease (HFMD), with 405 severe cases and 78 deaths (1-3). This epidemic prompted the use of intravenous immunoglobulin (IVIg) for the treatment of severe enteroviral infections, including EV-mediated encephalitis in Taiwan (4). This treatment policy has been hampered by the need to prove the presence of enteroviral infections for those patients without characteristic herpangina or HFMD, since current tissue culture-based methods are laborious, time-consuming, and frequently unsuccessful exercises (5). In contrast, detection of EVs by reverse transcription-polymerase chain reaction (RT-PCR) is more rapid and sensitive (6). The objective of the present study was to assess the impact of high-dose IVIg adjuvant therapy in patients with enteroviral encephalitis diagnosed by the RT-PCR assay. EVs are a leading cause of meningoencephalitis (7). They reside in a genus within the family of Picornaviridae, and are single-stranded RNA viruses that can be subdivided into five species including human enterovirus (HEV) A to D species and poliovirus type species (8). Specific antiviral treatment is not currently available. The clinical disease observed during central nervous system enteroviral infections varies *Corresponding author: Mailing address: Section of Infectious Diseases, Department of Internal Medicine, Veterans General Hospital-Kaohsiung, 386 Ta-Chung 1st RD, Kaohsiung, Taiwan 81346. Tel: +886-7-3422121 ext. 2029, Fax: +886-7-3468067, E-mail: mulstin@gmail.com with the host s age and immune status. Virulence determinants of the circulating virus may also be involved (9,10); antibodies play a pivotal role in resolving infection. Indeed, the use of immune globulin has quelled the incidence of enteroviral infections in the nursery (11-13) and in individually ill newborns (14-16). However, the clinical efficacy of IVIg in enteroviral encephalitis remains controversial (9). Patients with encephalitis who present characteristic herpangina or HFMD are easily diagnosed as enteroviral infections. However, many patients with enteroviral encephalitis do not display these characteristic manifestations; instead, diagnosis relies on the direct detection of the virus from cerebrospinal fluid (CSF), throat swab, or stool. These detection methods suffer from a slow turnaround time (17) and relative insensitivity (17,18). Brain biopsy, perhaps the only existing gold standard in diagnosing viral encephalitis, is rarely justified because of its invasive nature. RT-PCR more rapidly and sensitively detects EV genomic RNA in specimens as varied as muscle biopsy, CSF, throat swabs, serum, and stool (18,19). We used a commercially available PCR assay (AMPLICOR EV Test; Roche Molecular Systems, Branchburg, N.J., USA) that combined sufficient sensitivity and maximum specificity (17,18,20) for the rapid detection of EV RNA in clinical specimens by taking advantage of the new real-time LightCycler PCR technology (6) using the TaqMan format (21). Presently, we conducted a detailed prospective study of the clinical application of RT-PCR and the effects of treatment with IVIg on symptomatic enteroviral encephalitis. MATERIALS AND METHODS Study design: The subjects (n = 75; all <15 years of age) enrolled in the prospective study were admitted to Veterans 18

General Hospital-Kaohsiung, Taiwan, presenting with indications of encephalitis from January 1, 2003 to December 31, 2004. The clinical criteria for encephalitis included pleocytosis (leukocytes 25/mm 3 ), an absence of bacteria upon culture of CSF, and some or all of the following symptoms: fever; sensorium changes (such as confusion, drowsiness, stupor, coma, combativeness, convulsions, abnormal behavior, ataxia, limb paralysis, and clumsiness), hemiplegia, specific cranial nerve dysfunction, and brain dysautonomia. At the time of the study the use of IVIg therapy for enteroviral meningitis was not prompted in Taiwan. Hence, patients presenting with symptoms or signs of meningitis such as headache, stiff neck, nausea, and vomiting, but without the definable signs of encephalitis summarized above, were excluded. Encephalitis caused by nonviral pathogens cultured in CSF (such as bacteria, mycobacterium, fungus, parasites) and conscious changes caused by factors other than infection (such as hemorrhage, infarction, hypoglycemia, or other metabolic disorders) were also excluded in this study. Standard laboratory analyses conducted upon admission included viral cultures of the throat, rectal or stool specimens, CSF, and urine. Among the enrolled encephalitis patients, those who had characteristic manifestations of herpangina or HFMD were defined as clinically diagnosed enteroviral encephalitis, and those who had positive cultures of nonpolio EVs were defined as virus-culture confirmed enteroviral encephalitis. Patients who did not have positive cultures of EVs nor characteristic manifestations of herpangina or HFMD were defined as nonenteroviral encephalitis. The patients were divided into three groups. Clinically diagnosed patients with enteroviral encephalitis were assigned to group A. The CSF of these patients was evaluated by RT- PCR within 24 h of admission, and IVIg (1 g/kg body weight; the same dose was used for groups B and C) was administered via an intravenous route (groups B and C as well) after withdrawing the CSF sample. Patients with encephalitis without herpangina or HFMD were randomized into groups B and C. CSF specimens of the group B patients were sent for RT-PCR within 24 h of admission, and IVIg was administered to those patients upon a positive result. The CSF of the group C patients was not analyzed by RT-PCR until near the completion of the study. Additionally, group C patients were not administered a placebo, and IVIg was not administered until the isolation of EV. All patients received standard supportive care as determined by their primary care providers. With the exception of IVIg treatment, all treatment regimens and management decisions, including duration of hospitalization, were determined by the primary caregivers and were unaffected by the study protocol among the different groups. Patients were reevaluated during follow-up visits at 7 and 30 days, 3 months, and every 3 months after discharge until the final visit at the end of 2006. Viral isolation and serotyping: CSF, throat swabs, urine, stools, and/or rectal swabs from the patients were used for virus isolation. Samples were inoculated onto RD, A549, LLC MK2, Vero, and HEL cells. A culture positive for nonpolio EVs from any source was accepted as evidence of enteroviral infection. Urine was added to Eagle s minimum essential medium (EMEM) and passed through a 0.2 m pore size filter prior to the culture. Cultures were maintained in EMEM with 2% FCS at 37 C in an atmosphere of 5% CO 2 and were monitored daily for cytopathic effects. Samples that did not display cytopathic effects on day 7 were passed and observed for an additional 14 days. Once a cytopathic effect involved over 50% of the cell monolayer, the cells were scraped loose and indirect fluorescent antibody staining with a panenteroviral monoclonal antibody (Chemicon International, Inc., Temecula, Calif., USA) was performed to identify the EV antigens. All isolates were subsequently serotyped by immunofluorescence with type-specific monoclonal antibodies (all from Chemicon International) to poliovirus 1 to 3; coxsackievirus A2, A6 to A16, A24, and B1 to B6; echovirus 3, 4, 6, 9, 11, and 30; EV 70; and EV 71. In addition, EV 71 VPI-IgM Rapid tests (Oncoprobe Biotech, Taipei, Taiwan) were carried out to identify the EV 71 infection for the serum samples from all of the enrolled cases. RNA extraction and reverse transcription: RNA was extracted from 140 l of CSF or viral culture supernatant using the QIAamp viral RNA extraction kit (Qiagen, Hilden, Germany) according to the manufacturer s instructions. RNA was recovered in 60 l of nuclease-free water and either used immediately or stored at 80 C until further analysis. cdnas were obtained by RT using the Reverse Transcription System (Promega, Madison, Wis., USA). LightCycler PCR: A master mix (LightCycler FastStart DNA Master Hybridization Probes; Roche Diagnostics GmbH, Penzberg, Germany) was optimized for the LightCycler. The mix was composed of 0.2 mm concentrations of each of the deoxyribonucleoside triphosphates (50 mm KCl, 10 mm Tris-Cl [ph 8.3]), 3 mm MgCl2, 0.5 M concentrations of the primers (forward primer: 5 -GTA ACG GGC AAC TCT GCA GC-3 ; reverse primer: 5 -ATT GTC ACC ATA AGC AGC CA-3 ), and 0.2 M of the fluorescein probe (5-6FAM-CAD GGA CAC CCA AAG TAG TCG GTT CC TAMRA-TP-3 ). Five microliters of cdna was added to 15 l of PCR mixture in each reaction capillary. The reaction mixture was centrifuged in the capillary to facilitate mixing. All capillaries were then sealed and amplified using the following protocol: 95 C for 2 min for one cycle, followed by 5 s of denaturation at 95 C, 10 s of annealing at 65 C, and 10 s of primer extension at 72 C for 40 cycles. The melting curve analysis was performed in one cycle of 95 C for 10 s and 50 C for 60 s, each with a temperature transition rate of 20 C/s, and then ramping to 85 C at 0.1 C/s. Negative controls without virus cdna were included in every PCR. RESULTS Seventy-five cases with aseptic encephalitis were enrolled during the 2-year study. Of these, 29 (39%) occurred in the summer (between May and July); the seasonal distribution among the three groups was not significantly different (P = 0.43). Eighteen cases were caused by EVs, and the other 57 cases were nonenteroviral. The serotype distribution and the specimens used to isolate EVs and nonenteroviral viruses in groups A, B, and C during the entire course of the study are summarized in Table 1. Of the 18 CSF specimens acquired from the patients having virus-culture confirmed enteroviral encephalitis, 17 were EV positive by RT-PCR; the remaining specimen was found to contain coxsackievirus A9. Of the CSF specimens obtained from patients with nonenteroviral encephalitis, 4 were EV-positive and 53 were EV-negative by RT-PCR. RT-PCR was performed within 24 h of admission for all 10 patients in group A (i.e., patients with herpangina or HFMD) (Fig. 1). All CSF specimens from this group were positive for EV by RT-PCR, and at least one specimen was 19

Table 1. Distribution of virus isolates and serotypes in the different groups Groups/patients Results of CSF EV RT-PCR and culture from various specimens with culture Serotype 1) RT-PCR Cultures confirmed encephalitis CSF CSF Throat swab Rectal swab Stool Urine Group A Patient 1 Echovirus 11 + + NA + Patient 2 Echovirus 11 + + + + NA + Patient 3 Echovirus 11 + + NA Patient 4 Echovirus 11 + + + NA Patient 5 Echovirus 11 + + + NA Patient 6 Echovirus 9 + + NA Patient 7 Echovirus 9 + + NA + Patient 8 Coxsackievirus A9 + + NA Patient 9 Coxsackievirus B3 + + NA Patient 10 Nontypable + + NA Group B Patient 1 Echovirus 11 + + + NA Patient 2 Echovirus 11 + + NA + Patient 3 Coxsackievirus A9 + + NA Patient 4 Nontypable + + + NA Patient 5 Cytomegalovirus + NA + Patient 6 Cytomegalovirus NA + Patient 7 Adenovirus + + NA Patient 8 Influenza A virus + NA Group C Patient 1 2) Echovirus 11 + + NA Patient 2 Echovirus 11 + + NA + Patient 3 Coxsackievirus A9 NA + Patient 4 Coxsackievirus B3 + + + NA Patient 5 Cytomegalovirus + NA Patient 6 Adenovirus + NA Patient 7 Influenza B virus + NA Patient 8 HSV type 2 + + + 1) : EV-serotyping was done for all of the specimens culture-positive for virus. 2) : The patient with symptoms of epilepsy in the post-discharge follow-up. CSF, cerebrospinal fluid; EV, enterovirus; HSV, herpes simplex virus; +, positive;, negative; NA, not available. culture-positive for EV. The 10 isolates included echoviruses 11 and 9 (n = 5, 2, respectively), and coxsackieviruses A9 (n = 1), and B3 (n = 1), and untypable EV (n = 1). All 10 patients received IVIg soon after admission. The other 65 patients without herpangina or HFMD were randomly assigned to group B (RT-PCR conducted within 24 h of admission; n = 32) and group C (RT-PCR performed later in the study; n = 33). Five group B patients were PCR-positive for EV and immediately received IVIg. Four of these 5 patients were subsequently culture-positive for echovirus 11 (n = 2), coxsackievirus A9 (n = 1), and an untypable EV (n = 1). The 27 patients who were PCR-negative for EV were subsequently negative on viral culture. Of the group C patients, 4 were culture-positive for EV including echovirus 11 (n = 2), coxsackievirus A9 (n = 1), and coxsackievirus B3 (n = 1). The patients infected with coxsackievirus A9 were subsequently PCR-negative for EV, and the other 3 patients culturepositive for EV were subsequently PCR-positive. The patient infected with coxsackievirus B3 and one of the patients infected with echovirus 11 had frequent seizure episodes; IVIg was administered upon identification of the particular EV. The other 29 group C patients whose enteroviral isolation was negative did not receive IVIg during hospitalization. However, 3 patients subsequently proved to be viruspositive in retrospective RT-PCR analysis. A patient infected with echovirus 11 who did not receive IVIg until detection by viral isolation displayed symptoms of epilepsy at the postdischarge follow-up visit. The EV 71 VPI-IgM Rapid test to detect serum EV 71 IgM and the results of indirect immunofluorescence staining with monoclonal antibodies (mabs) against EV 71 (mab 979, 3323, and 3324) were negative in all 75 cases. There were no significant differences between the three groups in baseline demographic, clinical, and laboratory characteristics (Table 2). Infusions of IVIg (1 g/kg body weight) in the 17 patients who were so-treated were generally welltolerated. Skin rash, fever, tachycardia, and tachypnea were observed in 2 patients; these potential but unconfirmed side effects did not prompt discontinuation of the infusion. No association of IVIg and clinically significant changes in renal function, neutropenia, or thrombocytopenia were observed in the 2-year-long study. No fatalities occurred during the study among those with aseptic encephalitis. During postdischarge follow-up, only a single group C patient infected with echovirus 11 displayed neurological sequelae (epilepsy). The two most common EV serotypes were echovirus 11 and 9, and the two most common nonenteroviruses were cytomegalovirus and adenovirus. Bacterial cultures of blood, CSF, and urine were negative for all of the 75 patients, except for 4 urine cultures subsequently determined to be the 20

Fig. 1. Flow chart of study enrollment and outcomes. Table 2. Baseline characteristics of the patients in the different groups 1) Characteristic Group A Group B Group C (n = 10) (n = 32) (n = 33) Age (y) 1.62 ± 2.11 1.59 ± 3.25 1.62 ± 2.97 No. of male (%) 6 (60) 15 (47) 17 (52) White cell count per mm 3 6,725 ± 4,126 7,510 ± 7,531 6,980 ± 5,968 C-reactive protein, mg/dl 2.3 ± 3.9 2.9 ± 4.6 2.7 ± 5.1 Blood sugar, mg/dl 92.5 ± 20.3 89.3 ± 45.2 96.2 ± 35.6 Cerebrospinal fluid White cell count per mm 3 54 ± 126 46 ± 154 60 ± 127 Red-cell count per mm 3 8 ± 25 12 ± 35 10 ± 28 Glucose, mg/dl 73.2 ± 32.0 68.0 ± 28.0 65.9 ± 27.2 Protein, mg/dl 40.0 ± 18.2 42.2 ± 20.9 35.7 ± 15.8 1) : Plus-minus values are mean ± SD. There were no significant differences among the groups. result of specimen contamination. All the patients with positive cultures of influenza A or B virus, herpes simplex virus, and cytomegalovirus were treated with specific antiviral drugs such as oseltamivir, acyclovir, and gancyclovir, respectively, no matter which group they belonged to. The median time to discharge was 10, 9, and 14 days in groups A, B, and C, respectively (P = 0.08). There was no statistically significant difference in the duration of hospitalization between the three groups, though the patients in group C stayed in the hospital 4 and 5 days longer than those in groups A and B, respectively. DISCUSSION During the epidemic of enteroviral infection in 1998, the viruses isolated at two major diagnostic laboratories that processed samples from inpatients and outpatients, suggested that although approximately half of the EVs isolated were EV 71, other EVs were also active. Coxsackievirus A16 was known to cause 18% of the cases, and 40% of the cases were caused by other EVs (1-3). After this severe epidemic, IVIg was introduced for the treatment of severe enteroviral infections, including enteroviral encephalitis. The indications of 21

IVIg treatment are not only limited to severe EV 71 infection. The criteria for the indications include patients with clinical manifestations of HFMD or herpangina or epidemiological evidence of enteroviral infection, and the presence of at least one severe clinical syndrome (4). This policy for IVIg administration has been hampered by the need to prove the presence of enteroviral infections for those patients without characteristic herpangina or HFMD, since current tissue culture-based methods are laborious, timeconsuming, and frequently unsuccessful exercises (5). Patients having enteroviral encephalitis without the characteristic occurrences of herpangina or HFMD (8/18 patients, 44%, in the present study) are typically diagnosed by the isolation of the causative virus from clinical specimens. While conclusive, this route of diagnosis is long and laborious, and so is entirely impractical for rapid diagnosis and prompt treatment. In the present study, biochemistry and cell counts in CSF, white cell counts, and serum C-reactive protein levels were indistinguishable between enteroviral encephalitis and nonenteroviral encephalitis (Table 2). Thus, differentiation of enteroviral from nonenteroviral encephalitis on the basis of hematology and biochemical analysis is virtually impossible. In the present study, the sensitivity, specificity, positive predictive value, and negative predictive value were 94% (17/ 18), 93% (53/57), 81% (17/21), and 98% (53/54), respectively. These values are very similar to other reported study results (17,18,20), indicating the suitability of PCR for rapid diagnosis, which allows for a more rapid initiation of antiviral therapy to improve patient outcomes. In the 65 patients without characteristic manifestations of herpangina or HFMD, viral isolation was positive in only 8 patients (12.3%), and viral RNA was detected by PCR in only 11 (16.9%). Antibodies play a key role in the host defense against EVs. IVIg is prepared from pools of plasma samples obtained from at least 1,000 healthy donors by the Cohn alcohol fractionation method; thus, IVIg covers numerous arrays of antibodies for variable viruses that comprise a broad range of immune antibodies directed against pathogens (13-16,22,23). Patients with agammaglobulinemia are at risk of developing severe and/or chronic enteroviral infections, and their condition may be improved with administration of IVIg (24,25). In newborns, the absence of type-specific antibodies is a risk factor for the development of symptomatic enteroviral infections. Since commercial IVIg products contain neutralizing antibodies to commonly circulating enteroviral serotypes (12,24-27), IVIg has been used for prophylaxis during nursery outbreaks (11-13,24). In addition to increased viral clearance due to antibody-dependent neutralization, the efficacy of high-dose IVIg has been demonstrated in a wide range of additional autoimmune diseases, including Guillain-Barré syndrome, myasthenia gravis, Kawasaki disease, systemic vasculitis, and systemic lupus erythematosus (22,28-31). Although a complete definition of the mechanism of IVIg action is still lacking, extensive research suggests that IVIg may achieve its therapeutic effects through multiple mechanisms of immunomodulation (22). For example, several proposed mechanisms dependent on: (i) the antigen-binding (Fab) regions only (e.g., immunomodulation mediated by anti-idiotypic Fab within IVIg) (32); (ii) the Fc regions only (e.g., competitive blockade of Fc receptors mediated by Fc within IVIg) (33); (iii) Fab and Fc regions (e.g., inhibition of autoantibody production by crosslinking the B-cell receptor and Fc receptor IIB) (34); or (iv) IVIg contaminants (e.g., immunosuppression mediated by soluble cytokine receptors Fig. 2. Distribution of serotypes in patients of virus-culture confirmed enteroviral encephalitis from 1999 through 2006. present as contaminants within IVIg) (35). Wang et al. demonstrated the modulation of cytokine production by IVIg in patients with EV 71-associated brainstem encephalitis. Plasma levels of inferferon (IFN)-, interleukin (IL)-6, IL-8, IL-10, and IL-13 levels significantly decreased in patients with pulmonary edema after administration of IVIg. Additionally, plasma levels of IL-6 and IL-8 were significantly decreased in patients with autonomic nervous system dysregulation after administration of IVIg. These findings suggest that IVIg may play a therapeutic role in EV 71-associated brainstem encephalitis and indicate the possible mechanisms of IVIg therapy in such kinds of severe enteroviral infection (36). The role of IVIg in the treatment of viral encephalitis caused by the West Nile virus, herpes simplex virus, and Japanese encephalitis virus has been reported for clinical application, based on the theory of increased viral clearance due to antibody-dependent neutralization and/or modulation for the inflammatory conditions (23,37,38). However, the clinical efficacy of IVIg in enteroviral encephalitis remains controversial (9). As shown in Table 1, of the 18 virus-culture confirmed enteroviral encephalitis cases, the virus was most frequently obtained from stool and rectal swabs, but less from throat swabs and CSF, and least from urine. No EV 71 infection was identified in any of the specimens. In addition, EV 71 VPI-IgM Rapid tests were negative in all of the patients. The above tests might support the absence of EV 71 infections in this study. However, physicians should be aware that RT-PCR can detect EV71 at a higher rate in throat swabs, stools, and/ or rectal swabs than in CSF specimens (39). Consequently, we strongly recommend collecting other clinical specimens (such as throat swabs, stool swabs, and rectal swabs) in addition to CSF and their simultaneous testing by RT-PCR and virus culture to increase the detection rate for enteroviral encephalitis. HEV B species are the most common cause of aseptic meningitis. They include all echoviruses, six serotypes of coxsackie B virus, coxsackievirus A9, and a number of newer EVs (8,40). The predominant EV types vary from year to year, with echovirus 30, echovirus 13, and echovirus 18 being the most frequent in Europe and the United States over the past few years (40). Figure 2 summarizes the distribution of EV species in patients with culture-confirmed enteroviral encephalitis in our hospital from 1999 through 2006, and it suggests that the main cause of viral encephalitis after the 1998 EV 71 epidemic was HEV B species. Interestingly, we have also found that serotypes varied during these years in Taiwan, as in Europe and the United States. Coxsackievirus B1, B2, B3, and B5 were found primarily in 1999. Coxsackievirus A9, B1, and B3 and echovirus 4, 9, 24, and 30 were found to co-circulate in 2000. Echovirus 30 and 11 22

caused local outbreaks in 2001 and 2003, respectively, which is compatible with the report by Chen et al. of a disease outbreak due to echovirus 30 with mosaic genome structure in Taiwan in 2001 (41). Outbreaks typically peak during the summer and early fall. The outbreak of echovirus 11 in 2003 explains why the majority of patients with enteroviral encephalitis in the present study were infected with HEV B species instead of HEV A species, which is the main cause of HFMD and herpangina. Since encephalitis was diagnosed in all patients in the present study, it is not surprising to find that HEV B was the main cause of this disease, even in group A patients with clinical manifestations of HFMD or herpangina. Epilepsy as a sequel of encephalitis was found in one group C patient, whose evaluation by RT-PCR (establishing echovirus 11 as the cause) was not undertaken until late in the study. The epilepsy in this case might have been prevented by earlier administration of IVIg, and conducting RT-PCR testing soon after admission might have resulted in earlier treatment. In Taiwan, IVIg is the treatment indicated for patients with severe enteroviral infections including encephalitis (4). This strategy is controversial (9). The present results, however, support the use of this strategy, and suggest that IVIg teamed with rapid diagnosis using RT-PCR may shorten hospital stays and improve outcomes. In the present study, IVIg was administered as indicated for enteroviral infection by the CDC of Taiwan (4). Additionally, the consensus for IVIg therapy of other viral encephalitides remains unclear; therefore, patients who are culturepositive for viruses other than EVs are not treated with IVIg. However, patients infected with influenza A or B virus, herpes simplex virus, and cytomegalovirus are treated with oseltamivir, acyclovir, and gancyclovir, respectively; all of these patients achieve good outcomes without any sequelae. Specific antiviral drugs for EVs are not currently available, which emphasizes the importance of effective adjuvant treatment for severe enteroviral infections. In addition, IVIg, steroid pulse therapy has also been reported to treat acute viral encephalitis (42,43). However, the effect remains uncertain (44,45) and is not currently recommended in the guidelines for treatment of severe enteroviral infection in Taiwan, and thus was not considered in the study design. Though the present study supports the strategy of using IVIg teamed with rapid diagnosis by RT-PCR, double-blind, randomized, placebo-controlled trials using larger patient populations will be needed to confirm the efficacy of this strategy for treating enteroviral encephalitis at an earlier stage. ACKNOWLEDGMENTS This work was supported by a grant from the Veterans General Hospital- Kaohsiung, Taiwan (VGHKS93-74). REFERENCES 1. Ho, M., Chen, E.R., Hsu, K.H., et al. (1999): An epidemic of enterovirus 71 infection in Taiwan. N. Engl. J. Med., 341, 929-935. 2. Lin, T.Y., Twu, S.J., Ho, M.S., et al. (2003): Enterovirus 71 outbreaks, Taiwan: occurrence and recognition. Emerg. Infect. Dis., 9, 291-293. 3. Wang, J.R., Tsai, H.P., Chen, P.F., et al. (2000): An outbreak of enterovirus 71 infection in Taiwan, 1998. II. Laboratory diagnosis and genetic analysis. J. Clin. Virol., 17, 91-99. 4. Center for Disease Control, Taiwan (2006): Guidelines for treatment of severe enteroviral infection. p. 11. Clinical management and indication of IVIg usage for severe enterovirus infection (in Chinese). 5. Rotbart, H.A. and Romero, J.R. (1995): Laboratory diagnosis of enteroviral infections. p. 401-410. In H.A. Rotbart (ed.), Human Enterovirus Infections. ASM Press, Washington, D.C. 6. Holger, F.R., Alexandra, C., Gerhard, M., et al. (2002): Rapid detection of enterovirus infection by automated RNA extraction and real-time fluorescence PCR. J. Clin. Virol., 25, 155-164. 7. Rotbart, H.A. (1995): Meningitis and encephalitis. p. 271-289. In H. A. Rotbart (ed.), Human Enterovirus Infections. ASM Press, Washington, D.C. 8. Stanway, G. (2005): Family Picornaviridae. p. 757-758. In C.M. Fauquet, M.A. Mayo, J. Maniloff, et al. (ed.), Virus Taxonomy. Eighth Report of the International Committee on Taxonomy of Viruses. Elsevier/Academic Press, London. 9. Jochem, M.D., Maria, T.E., Gerco, H.J., et al. (1997): Antibodies against enteroviruses in intravenous Ig preparations: great variation in titres and poor correlation with the incidence of circulating serotypes. J. Med. Virol., 53, 273-276. 10. Melnick, J.L. (1996): Enteroviruses: polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses. p. 655-712. In B.N. Fields, D.M. Knipe, P.M. Howley (ed.), Virology. Lippincott-Raven Press, Philadelphia. 11. Nagington, J. (1982): Echovirus 11 infection and prophylactic antiserum [letter]. Lancet, 1, 446. 12. Carolane, D.J., Long, A.M., McKeever, P.A., et al. (1985): Prevention of spread of echovirus 6 in a special care baby unit. Arch. Dis. Child., 60, 674-676. 13. Nagington, J., Gandy, G., Walker, J., et al. (1983): Use of normal immunoglobulin in an echovirus 11 outbreak in a special-care baby unit. Lancet, 2, 443-446. 14. Black, S. (1983): Treatment of overwhelming neonatal ECHO 5 virus infection with intravenous gamma globulin (abstract no. 314). p. 140. In S. Black (ed.), Program and Abstracts of the 23rd Interscience Conference on Antimicrobial Agents and Chemotherapy (Las Vegas). ASM Press, Washington, D.C. 15. Johnston, J.M. and Overall, J.C., Jr. (1989): Intravenous immunoglobulin in disseminated neonatal echovirus 11 infection. Pediatr. Infect. Dis. J., 8, 638-641. 16. Valduss, D., Murray, D.L., Karna, P., et al. (1993): Use of intravenous immunoglobulin in twin neonates with disseminated coxsackie B1 infection. Clin. Pediatr. (Phila.), 32, 561-563. 17. Romero, J.R. (1999): Reverse-transcription polymerase chain reaction detection of the enteroviruses. Arch. Pathol. Lab. Med., 123, 1161-1169. 18. Rotbart, H.A. (1997): Reproducibility of AMPLICOR enterovirus PCR test results. J. Clin. Microbiol., 35, 3301-3302. 19. Pozo, F., Casas, I., Tenorio, A., et al. (1998): Evaluation of a commercially available reverse transcription-pcr assay for a diagnosis of enteroviral infection in archival and prospectively collected cerebrospinal fluid specimens. J. Clin. Microbiol., 36, 1741-1745. 20. Lina, B., Pozzetto, B., Andreoletti, L., et al. (1996): Multicenter evaluation of a commercially available PCR assay for diagnosing enterovirus infection in a panel of cerebrospinal fluid specimens. J. Clin. Microbiol., 34, 3002-3006. 21. Nitsche, A., Steuer, N., Schmidt, C.A., et al. (1999): Different real-time PCR formats compared for the quantitative detection of cytomegalovirus DNA. Clin. Chem., 45, 1932-1937. 22. Jin, F. and Balthasar, J.P. (2005): Mechanisms of intravenous immunoglobulin action in immune thrombocytopenic purpura. Hum. Immunol., 66, 403-410. 23. Caramello, P., Canta, F., Balbiano, R., et al. (2006): Role of intravenous immunoglobulin administration in Japanese encephalitis. Clin. Infect. Dis., 43, 1620-1621. 24. Abzug, M.J., Keyserling, H.L., Lee, M.L., et al. (1995): Neonatal enterovirus infection: virology, serology, and effects on intravenous immune globulin. Clin. Infect. Dis., 20, 1201-1206. 25. McKinney, R.E., Jr., Katz, S.L. and Wilfert, C.M. (1987): Chronic enteroviral meningoencephalitis in agammaglobulinemic patients. Rev. Infect. Dis., 9, 334-356. 26. Dagan, R., Prather, S.L., Powell, K.R., et al. (1983): Neutralizing antibodies to non-polio enteroviruses in human immune serum globulin. Pediatr. Infect. Dis., 2, 454-456. 27. Keyseling, H.L. and Torfason, E.G. (1986): Comparison of intravenous gamma globulin preparations in neutralization assays against enteroviruses (abstract no. 328). p. 156. In Program and Abstracts of the 26th Interscience Conference on Antimicrobial Agents and Chemotherapy (New Orleans). ASM Press, Washington, D.C. 28. Arsura, E. (1989): Experience with intravenous immunoglobulin in myasthenia gravis. Clin. Immunol. Immunopathol., 53, S170-179. 29. Jordan, S.C. (1989): Intravenous gamma-globulin therapy in systemic lupus erythematosus and immune complex disease. Clin. Immunol. Immunopathol., 53, S164-169. 30. Jayne, D.R., Davies, M.J., Fox, C.J., et al. (1991): Treatment of systemic vasculitis with pooled intravenous immunoglobulin. Lancet, 337, 1137-23

1139. 31. Newburger, J.W., Takahashi, M., Burns, J.C., et al. (1986): The treatment of Kawasaki syndrome with intravenous gamma globulin. N. Engl. J. Med., 315, 341-347. 32. Berchtold, P., Dale, G.L., Tani, P., et al. (1989): Inhibition of autoantibody binding to platelet glycoprotein IIb/IIIa by anti-idiotypic antibodies in intravenous gammaglobulin. Blood, 74, 2414-2417. 33. Bussel, J.B. (2002): Another interaction of the FcR system with IVIG. Thromb. Haemost., 88, 890-891. 34. Ballow, M. (1997): Mechanisms of action of intravenous immune serum globulin in autoimmune and inflammatory diseases. J. Allergy Clin. Immunol., 100, 151-157. 35. Sewell, W.A. and Jolles, S. (2002): Immunomodulatory action of intravenous immunoglobulin. Immunology, 107, 387-393. 36. Wang, S.M., Lei, H.Y., Huang, M.C., et al. (2006): Modulation of cytokine production by intravenous immunoglobulin in patients with enterovirus 71-associated brainstem encephalitis. J. Clin. Virol., 37, 47-52. 37. Haley, M., Retter, A.S., Fowler, D., et al. (2003): The role for intravenous immunoglobulin in the treatment of West Nile virus encephalitis. Clin. Infect. Dis., 37, 88-90. 38. Yoshidome, Y., Hayashi, S. and Maruyama, Y. (2005): A case of brainstem encephalitis caused by herpes simplex virus type 1 with possible infection via trigeminal nerve. Clin. Neurol., 45, 293-297 (in Japanese). 39. Tsao, L.Y., Lin, C.Y., Yu, Y.Y., et al. (2006): Microchip, reverse transcription-polymerase chain reaction and culture methods to detect enterovirus infection in pediatric patients. Pediatr. Int., 48, 5-10. 40. Thoelen, I., Moes, E., Lemey, P., et al. (2004): Analysis of the serotype and genotype correlation of VP1 and the 5 noncoding region in an epidemiological survey of the human enterovirus B species. J. Clin. Microbiol., 42, 963-971. 41. Chen, G.W., Huang, J.H., Lo, Y.L., et al. (2007): Mosaic genome structure of echovirus type 30 that circulated in Taiwan in 2001. Arch. Virol., [Epub ahead of print]. 42. Nakano, A., Yamasaki, R., Miyazaki, S., et al. (2003): Beneficial effect of steroid pulse therapy on acute viral encephalitis. Eur. Neurol., 50, 225-229. 43. Koyama, S., Morita, K., Yamaguchi, S., et al. (2000): An adult case of mumps brainstem encephalitis. Intern. Med., 39, 499-502. 44. Johnson, R.T., Intralawan, P. and Puapanwatton, S. (1986): Japanese encephalitis: identification of inflammatory cells in cerebrospinal fluid. Ann. Neurol., 20, 691-695. 45. Matsui, M. (1999): Infection in the central nervous system and corticosteroid therapy. Clin. Neurol., 39, 26-28 (in Japanese). 24