Outcome of and Prognostic Factors for Herpes Simplex Encephalitis in Adult Patients: Results of a Multicenter Study

Transcription

1 MAJOR ARTICLE Outcome of and Prognostic Factors for Herpes Simplex Encephalitis in Adult Patients: Results of a Multicenter Study Franck Raschilas, 1,2 Michel Wolff, 2 Frédérique Delatour, 3 Cendrine Chaffaut, 4 Thomas De Broucker, 5 Sylvie Chevret, 4 Pierre Lebon, 1 Philippe Canton, 6 and Flore Rozenberg, 1 for the French Herpes Simplex Encephalitis Study Group a 1 Laboratoire de Virologie, Hôpital Saint-Vincent-de-Paul, 2 Service de Réanimation des Maladies Infectieuses and 3 INSERM U13, Hôpital Bichat Claude Bernard, and 4 Département de Biostatistiques et Informatique Médicale, Hôpital Saint-Louis, Paris; 5 Service de Neurologie, Hopital Delafontaine, Saint-Denis; and 6 Service des Maladies Infectieuses, Centre Hospitalier Universitaire Brabois, Nancy, France Management of herpes simplex encephalitis (HSE) has been considerably improved by the availability of acyclovir therapy and rapid polymerase chain reaction (PCR) based diagnostic assays. Prognostic factors for this rare affliction are, however, misestimated. We conducted a large retrospective multicenter study that included 93 adult patients in whom HSE was diagnosed by PCR from 1991 through 1998 and who were treated with intravenous acyclovir. Among the 85 patients assessed at 6 months, 30 (35%) had a poor outcome, which led to death in 13 patients (15%) and severe disability in 17 (20%). The outcome was favorable for 55 patients (65%). A multivariate analysis identified 2 factors that were found to be independently associated with poor outcome: a Simplified Acute Physiology Score II 27 at admission and a delay of 12 days between admission to the hospital and initiation of acyclovir therapy. Early administration of antiviral therapy is the only parameter that can be modified to improve the prognosis of patients with HSE. Herpes simplex encephalitis (HSE) is the most common sporadic necrotizing encephalitis in the Western world [1]. The spontaneous mortality rate associated with HSE is 70%, and, before the advent of antiviral therapy, most survivors had severe neurological impairment [2]. Two major advances have considerably improved the management of HSE. First, 2 large randomized trials performed in the mid-1980s showed that use of intravenous acyclovir reduced the 6-month mortality rate to 20% and significantly decreased morbidity (i.e., minor or no Received 8 November 2001; revised 8 March 2002; electronically published 10 July Financial support: Glaxo-Wellcome (grant ). a Members of the study group are listed after the text. Reprints or correspondence: Dr. Flore Rozenberg, Laboratoire de Virologie, Hôpital Saint-Vincent-de-Paul, AP-HP and Faculté Cochin, 82 Avenue Denfert- Rochereau, Paris Cedex 14, France (flore.rozenberg@svp.ap-hop-paris.fr). Clinical Infectious Diseases 2002; 35: by the Infectious Diseases Society of America. All rights reserved /2002/ $15.00 neurological impairment was observed in 37.5% and 55.5% of treated patients, respectively) [3, 4]. Second, herpes simplex virus (HSV) DNA amplification by PCR analysis of CSF has been the reference standard for early diagnosis of HSE since the early 1990s [5 7], thus greatly improving early therapeutic decisions [8, 9]. Despite this marked progress, recent reports have suggested that CNS sequelae are still frequent [10, 11]. The age of the patient and the level of consciousness at initiation of therapy have been identified as major determinants of prognosis [4]. More recently, the delay between admission of the patient to the hospital and initiation of acyclovir was identified as a main predictor of outcome [12]. However, the rarity of the disease has hampered the design of studies focused on the long-term outcome of survivors of HSE or aimed at identifying early prognostic factors. To further identify parameters independently associated with poor prognosis, we conducted a large, multicenter, retrospective study to assess the outcome of HSE for acyclovir-treated adult patients. 254 CID 2002:35 (1 August) Raschilas et al.

2 METHODS Patients. Patients were traced from records kept by the Virology Department of Saint-Vincent-de-Paul Hospital, where CSF samples from patients with infections of the CNS were prospectively assayed by PCR according to a procedure published elsewhere [7]. From January 1991 through June 1998, CSF samples from 241 adult patients tested positive for HSV DNA by PCR (180 with HSV type 1 and 61 with HSV type 2). After the exclusion of patients with HSV type 2 associated meningitis, the study was limited to 98 patients who originated from 2 geographical centers, Paris and Nancy, France. Procedures. Demographic data and information about previous health status, neurological signs, and symptoms at hospital admission were retrieved from medical records. The severity of underlying disease was assessed by the MacCabe [13] and Knaus [14] scores. The following parameters were determined within the first 24 h of admission: the Simplified Acute Physiology Score (SAPS II) [15], the Glasgow Coma Scale (GCS) score [16], and the need for mechanical ventilation. The following biological data were obtained: serum sodium concentration, CSF leukocyte count, CSF glucose and protein levels, and IFN-a dosage. CT scan or MRI profiles were reviewed by a neurologist (T.d.B.). The delay between the onset of symptoms and initiation of treatment, the delay between hospital admission and initiation of acyclovir, the dose and duration of antiviral treatment, and the use of drugs to reduce cerebral edema and/or to control seizures were recorded. Outcome assessments. Neurological impairment and quality of life were assessed at 6 and 12 months by review of patients medical charts and by questioning of the patients, their relatives, or their general practitioners about specific items: the patient s professional activity and ability to drive before and after the onset of HSE; autonomy in daily life, including the ability to wash him or herself, to cook, to clean the home, and to shop; the need for permanent home care; impairment in speaking, fluency, judgment, or memory; the need for antiepileptic treatment and occurrence of seizures after the onset of HSE; and the need for speech, motor, or any other rehabilitation after the onset of HSE. Ten patients were first evaluated by a member of the team to assess the reliability of the evaluation. All other patients were then evaluated either by members of the team or by their general practitioners after detailed information and the objectives of the study had been provided to the general practitioner. Handicap and quality of life were graded according to a 5-grade scale derived from the Glasgow Outcome Scale [17]: I, good recovery, allowing independent life without any neurological impairment; II, mild disability, defined by the presence of minimal cognitive alterations (speech disturbances, memory, or attention impairment) and/or seizures (partially controlled with antiepileptic drugs), without consequences for socioprofessional life; III, moderate disability, defined by criteria identical to those of group II but with consequences for socioprofessional life; IV, severe disability, defined by loss of autonomy requiring institutionalization or constant life aid; and V, death. To identify HSE prognostic factors, the patients were assigned to 1 of 2 categories, according to outcome at 6 months: favorable outcome, for patients with good recovery or mild or moderate disability (grades I III), or poor outcome, for patients with severe disability or who died (grades IV and V). Statistical methods. For quantitative variables, data were given as mean SD; for qualitative variables, data were given as n/n (%). For the analysis of prognostic factors, outcome at 6 months was the main end point and was considered to be binary (favorable outcome vs. poor outcome). Univariate prognostic analyses were calculated using Fisher s exact test (for qualitative variables) and the nonparametric Wilcoxon test (for quantitative variables). Multivariate analysis was then conducted that included the covariates that showed statistical association with the outcome in univariate analysis. A logistic regression model was used in which the covariates were introduced as binary, using, for each continuous covariate, the median value computed in the whole sample as the cutoff point. Estimated odds ratios (with 95% CIs) were computed. All statistical tests were 2-sided, and P.05 was considered to be statistically significant. Statistical analysis was performed using the SAS version 8.1 software package (SAS, Inc.). RESULTS Study population. Five of the 98 patients enrolled in the study were excluded from the analyses because incomplete data were available. The 93 remaining patients originated from 36 hospitals (34 in the Paris area and 2 in Nancy). HSE was associated with HSV type 1 in 92 patients and with HSV type 2 in 1 patient. Clinical characteristics. The mean patient age was 53.5 years, and there was a slight predominance of male patients. The first clinical evaluation was performed at the emergency department for 32 patients (34%), in an intensive care unit (ICU) for 10 patients (11%), in a neurology department for 10 patients (11%), and in an infectious diseases department for 4 patients (4%). Thirty-seven patients (40%) were admitted to other medical departments. Overall, 66 (71%) of 93 patients were admitted to the ICU, either directly or from the ward after a mean delay of days. At admission, the mean temperature was 38.9 C, but 8 patients (8.5%) were not febrile. The mean GCS score was , although the score was!8 for 9 patients, and the median value was 14. Neurological signs and symptoms at admission mainly consisted of disorientation (71 patients Herpes Simplex Encephalitis CID 2002:35 (1 August) 255

3 [76%]), behavioral changes (38 [41%]), speech disturbances (55 [59%]), and seizures (31 [33%]). The first neurological examination had no abnormal results for 5 (5.5%) of 93 patients. At admission to the ICU, the mean SAPS II was , and the mean GCS score was (table 1). Laboratory findings. Most of the initial CSF samples exhibited monocytic pleocytosis, but polymorphonuclear cells were predominant in samples from 2 (2%) of the 93 patients. The mean protein level was g/l in the initial CSF samples and reached g/l in the second CSF samples, obtained days later. In 3 patients (3%), the leukocyte count in the initial CSF samples, obtained at admission, was normal (!5 cells/mm 3 ); protein levels were normal (!0.50 g/l) in 2 patients and were slightly elevated (0.69 g/l) in 1 patient. The initial CSF samples were not analyzed with HSV PCR, but results of PCR were positive for samples drawn 2 7 ( mean SD, ) days later, and those samples had high protein levels and WBC counts. IFN-a levels were high in the initial CSF sample from 43 patients. The mean level was IU/mL (normal level,!2 IU/mL). Hyponatremia (serum sodium level,!135 mm) was present in 75 (81%) of 93 patients (table 1). Brain imaging findings. The mean delay between hospital admission and the first neuroimaging study was days (range, 0 4 days). A CT scan was performed for 91 patients, and MRI was done for 2 patients (table 2). No abnormalities were seen on the first CT scan in 19 (21%) of 91 patients. No other imaging studies were done for 2 (2%) of these 91 patients. CNS lesions were obvious on a second CT scan obtained days later for all 17 other patients. Brain imaging revealed temporal involvement in 83 (89%) of the 93 patients, and this was associated with frontal involvement in 34 (36%) of 93 patients. In 8 (9%) of 93 patients, radiologic examination showed only occipital or parietal lesions. Therapy. All patients received intravenous acyclovir. The mean delay between hospital admission and the initiation of acyclovir therapy was days. In 82 (88%) of 93 patients, the mean delay between the reported onset of symptoms and initiation of treatment was days. The intervals were as follows: 1 3 ( n p 20), 4 6 ( n p 40), 7 9 ( n p 14), ( n p 5), and days ( n p 3). In 11 patients, this infor- mation could not be obtained. The mean deduced delay between onset of symptoms and admission was therefore 3.5 days. The mean dosage of acyclovir was mg/kg 3 times daily for a mean duration of days. Seventy-three patients took antiepileptic drugs, and 28 received antiedematous therapy, either with corticosteroids ( n p 12) or with hypertonic mannitol ( n p 16). Overall, 66 patients (71%) received care in the ICU, and 42 patients (45%) required mechanical ventilation, for a mean duration of days (range, 1 56 days). Outcomes. Eighty-five (91%) of 93 patients were available for evaluation 6 months after the onset of HSE, and 8 were lost to follow-up. Among these 85 patients, 13 (15%) died, and 8 of those died within the first month of hospitalization. All deaths that occurred during the first 6 months of follow-up were due to HSV disease, either as a direct result (7 patients) or as a consequence of complications (6 patients), mainly nosocomial infections. Only 12 (14%) of 85 patients completely recovered; 19 (23%) of the 85 patients lived with mild disability, 24 (28%) had moderate disability, and 17 (20%) had severe disability. Thus, the outcome at 6 months was favorable for 55 (65%) of 85 patients and poor for 30 (35%) of 85 patients. A relapse of HSE (confirmed by HSV PCR analysis of CSF) occurred in 1 patient 2 months after the end of a first, 24-day course of acyclovir. This patient had a poor recovery. Fiftythree of 83 patients were assessed 1 year after the onset of HSE. Among these, 15 (28%) died, 9 (17%) had complete recovery, 12 (23%) had mild disability, 10 (19%) had moderate disability, and 7 (13%) had severe disability. Two patients living in a longterm care facility died before the end of the first year after the onset of HSE, 1 patient at 8 months after nosocomial infection occurred and the other at 10 months after the onset of neurological sequelae of HSE. Prognostic factors. By univariate analysis, no significant Table 1. Demographic characteristics and baseline clinical and laboratory findings for 93 patients with herpes simplex encephalitis. Characteristic Value Age, mean years SD (range) (16 88) Male sex, no. (%) of patients 56 (60) MacCabe score, mean SD (range) (0 2) Knaus score, mean SD (range) (1 3) Temperature, mean C SD (range) ( ) SAPS II, mean SD (range) (8 58) GCS score, mean SD (range) (3 15) Serum sodium concentration, mean mm SD (range) ( ) Neurological features, no. (%) of patients Seizures 31 (33) Focal neurological deficit 22 (24) CSF parameters, mean SD (range) Leukocyte count, cells/ml (1 3900) Mononuclear cell proportion, % (13 100) Polymorphonuclear cell proportion, % (0 69) Protein level, g/l ( ) IFN-a level, a IU (2 200) NOTE. GCS, Glasgow Coma Scale; SAPS, Simplified Acute Physiology Score. a Data for IFN-a levels in CSF were available for only 43 patients. 256 CID 2002:35 (1 August) Raschilas et al.

4 Table 2. Findings of CT (91 patients) and MRI (2 patients) for 93 patients with herpes simplex encephalitis. Brain imaging results No. (%) of patients Localization of the lesion Laterality Unilateral involvement 53 (57) Right hemisphere 18 Left hemisphere 35 Bilateral involvement 38 (41) Brain lobe involved Temporal 49 (53) Temporal and frontal 34 (36) Other a 8(9) No abnormal results 2 (2) a Other involved sites were parietal ( n p 6) and occipital ( n p 2). differences between patients with favorable and poor outcome were found for the following factors: age, MacCabe score, focal neurological deficit or seizures, need for mechanical ventilation, serum sodium concentration, and CSF parameters. In contrast, patients who experienced poor outcome had significantly higher Knaus and significantly lower GCS scores at admission. The delay between admission and initiation of acyclovir therapy was!2 days in 41 (75%) of 55 patients with favorable outcome, compared with 9 (30%) of 30 patients with poor outcome ( P p.00008). Nosocomial infection occurred significantly more frequently in the latter group (table 3). On multivariate analysis, 2 factors were found to be independently associated with poor outcome of HSE at 6 months: a delay of 12 days between admission and initiation of acyclovir therapy and a SAPS II 127 (table 4). DISCUSSION Although the prognosis for patients with HSE has been dramatically improved by the availability of specific antiviral therapy, sequelae in surviving patients may include severe neurological deficits, seizures, and/or neuropsychological dysfunctions that greatly impair quality of life [10 12]. Therefore, the identification of early factors that are predictive of outcome might contribute to better management of the disease. Patient age and level of consciousness at onset of therapy have been identified elsewhere as major determinants of prognosis [4]. However, in that study, one-half of the patients were between the ages of 6 months and 20 years. In addition, HSE was diagnosed by cerebral biopsy, which could have slightly modified the evolution of CNS lesions and also delayed the administration of antiviral treatment, as indicated by the duration of disease before therapy. In a more recent study, the delay between hospital admission and the initiation of acyclovir therapy was twice as long for patients with poor outcomes as for those with favorable outcomes [12]. Both studies included pediatric patients, for whom the prognosis might differ from that for adults [18]. Moreover, patients were included starting in 1983, and diagnosis was confirmed by heterogeneous methods. Finally, no multivariate analysis was performed to identify the parameters independently associated with poor outcome. In the present study, we retrospectively analyzed 93 adult patients in whom HSE was diagnosed by PCR, which is the reference standard for diagnosis [8], and all patients received acyclovir therapy. To date, this is the largest series of adults with HSE in whom prognoses have been evaluated using a multivariate analysis. The disease of patients in the present study was highly representative of classical adult HSE. The mean age (53.5 years) was similar to that of patients included in other epidemiological studies [19]. Mononuclear pleocytosis and mildly elevated protein level were common findings of analysis of CSF samples, although several CSF samples were acellular or pleocytic, with polymorphonuclear predominance at the onset of the disease. However, no correlation was found between CSF abnormalities and outcome, as reported elsewhere [20]. Neuroimaging showed temporal lobe involvement in most cases. However, less typical lesions, such as occipital or parietal lobe involvement, were observed in 11% of patients. Such localizations of HSE lesions have been recently reported in a detailed radiological study [11]. These previously misdiagnosed cases of HSE are presently better recognized by PCR [21]. Remarkably, the consciousness level of patients at hospital admission differed from that usually reported in association with adult HSE. Indeed, the mean GCS score at admission for our patients was moderately altered, compared with the high percentages of comatose patients reported in 2 large studies elsewhere [3, 4]. However, more recently, the initial GCS score of patients with HSE was found to be 12 in most cases [22]. Several explanations may account for this high GCS score at admission: patients may now be referred more rapidly to the hospital and/or HSE may be suspected earlier, as was suggested by the short delay (3.5 days) observed in our study between onset of symptoms and hospital admission. Early use of PCR may also allow recognition of milder or moderate forms of HSE [21, 23]. However, despite having a high GCS score at admission, most patients in the present study exhibited rapid neurological deterioration, which led to ICU admission in 71% of the patients. Surprisingly, despite the use of rapid diagnostic procedures and effective antiviral therapy, HSE was associated with significant mortality and morbidity in the present series. The 6- month fatality rate was only slightly lower than those for acyclovir-treated patients from 3 studies published elsewhere (19%, Herpes Simplex Encephalitis CID 2002:35 (1 August) 257

5 Table 3. Univariate analysis of factors associated with outcomes at 6 months for 85 patients with herpes simplex encephalitis. Characteristic Patients with favorable outcome (n p 55) Patients with poor outcome (n p 30) P Age, mean years SD MacCabe score, mean SD Knaus score, mean SD GCS score, mean SD SAPS II, mean SD SAPS II 127, no. (%) of patients 9 (16) 17 (57).0001 Seizures, no. (%) of patients 18 (33) 9 (30).79 Focal neurological deficit, no. (%) of patients 11 (20) 9 (30).3 Serum sodium level, mean mm SD CSF parameters, mean SD Leukocyte count, cells/ml Protein level, g/l Less than 2 days between hospital admission and initiation of acyclovir therapy, no. (%) of patients 41 (75) 9 (30) Mechanical ventilation, no. (%) of patients 23 (42) 18 (60).97 Hospital-acquired infection, no. (%) of patients 17 (31) 20 (67).001 NOTE. GCS, Glasgow Coma Scale; SAPS, Simplified Acute Physiology Score. 28%, and 19%), which involved a total of 101 patients [3, 4, 12]. Moreover, the 6-month morbidity assessment revealed that only 37% of our patients returned to conditions of life that were nearly the same as those preceding the illness; 14% completely recovered and 23% had mild impairment. This percentage was even lower at the 1-year follow-up. These figures are comparable to those reported in the first trials of acyclovir treatment, in which the percentages of patients with minor or no neurological impairment among the entire population of treated patients were 37.5% and 55.5% [3, 4]. To identify prognostic factors for HSE, patients were assigned to 1 of 2 categories according to outcome at 6-months (favorable and poor). This categorization was performed before the statistical analysis of the data set, to distinguish a clear-cut and clinically relevant group of patients that included those with very severe neurological disability and those who had died as result of HSE ( poor outcome ). A similar categorization has been used in 2 recent studies that included patients with bacterial meningitis [24, 25]. Two prognostic factors independently associated with a poor 6-month outcome were identified for adult patients with HSE: a SAPS II 127 at admission and a delay of 12 days between admission and initiation of acyclovir therapy. Our observation that a high SAPS II is associated with poor outcome is not unexpected, because this score takes into account many parameters, including age and neurological status, that have been identified elsewhere as prognostic factors [4]. The second prognostic factor identified in this study, that the length of the delay between hospital admission and the initiation of acyclovir therapy affects outcome, was also reported in a retrospective study by McGrath et al. [12] that included 42 patients; the study showed that this delay had been longer for patients with a poor outcome (4 days) than for those with a good outcome (1.8 days). Because of the rarity of HSE, the heterogeneous origin of the patients, and the retrospective design of the study, our results carry some limitations. First, the precise duration of symptoms before admission was not taken into account in the statistical analysis, because it could not be accurately evaluated in all patients. However, both the high GCS score observed at hospital admission in patients included in the present study and the mean delay observed between the onset of symptoms and initiation of treatment suggest that evolution of the disease before admission was shorter than that reported by studies elsewhere [21, 23]. Second, the neurological examination evaluated handicap and quality of life, but neuropsychological Table 4. Multivariate analysis of factors associated with poor outcome at 6 months for 85 patients with herpes simplex encephalitis. Parameter OR (95% CI) P SAPS II 127 at hospital admission 3.7 ( ).014 More than 2 days between hospital admission and initiation of acyclovir therapy 3.1 ( ).037 NOTE. SAPS, Simplified Acute Physiology Score. 258 CID 2002:35 (1 August) Raschilas et al.

6 status was not precisely assessed. In a study by McGrath et al. [12], all but 1 of 34 surviving patients had neurological symptoms and abnormal results of neurological examination, but nearly one-half of patients were able to perform everyday activities as before HSE. Thus, the handicap and quality-of-life scores defined in our study may have underevaluated the occurrence of neurological sequelae in patients with HSE. In conclusion, although the availability of PCR has greatly facilitated early diagnosis of HSE, a number of patients die or do not recover completely, despite administration of early acyclovir therapy, and have severe sequelae. Our results underline the fact that HSE prognosis is correlated with the delay between hospital admission and initiation of acyclovir. In the future, further reduction of mortality and morbidity due to HSE might be achieved with use of more-effective antiviral agents. However, because the early administration of antiviral therapy is, at present, the only parameter that can be modified to improve the prognosis for patients with HSE, acyclovir treatment should be initiated as soon as HSE is suspected. FRENCH HERPES SIMPLEX ENCEPHALITIS STUDY GROUP MEMBERS Prof. Agid, Dr. Bolgert, Prof. Bricaire, Prof. Lyon-Caen, and Prof. Pierrot-Deseilligny (Hôpital Pitié-Salpétriêre, Paris); Dr. Akrouf and Dr. Meignan (Hôpital Sainte-Anne, Paris); Prof. Bousser (Hôpital Lariboisière, Paris); Prof. Degos (Fondation- Hôpital Saint-Joseph, Paris); Prof. Degos (Hôpital Henri Mondor, Paris); Prof. Dhainaut (Hôpital Cochin, Paris); Prof. Offenstadt (Hôpital Saint-Antoine, Paris); Prof. Roullet (Hôpital Tenon, Paris); Prof. Safar (Hôpital Broussais, Paris); Prof. Valcke (Hôpital Boucicaut, Paris); Prof. Bleichner and Dr. Davous (Hôpital Victor Dupouy, Argenteuil); Dr. Fouet, Dr. Hilpert, and Dr. Lanoe (Hôpital Robert Ballanger, Aulnay-sous- Bois); Dr. Hoang The Duan and Prof. Robineau (Hôpital Avicenne, Bobigny); Prof. Jardin (Hôpital Ambroise Paré, Boulogne-Billancourt); Prof. Chaput and Prof. Galanaud (Hôpital Antoine Béclère, Clamart); Prof. Dreyfuss (Hôpital Louis Mourier, Colombes); Prof. Lemaire (Hôpital Henri Mondor, Créteil); Dr. Tenaillon (Hôpital Louise Michel, Evry); Dr. Clair and Prof. Gajdos (Hôpital Raymond Poincarré, Garches); Prof. Nouhaillat (Hôpital Léon Touladjian, Mantes-la-Jolie); Dr. Loriferne (Centre Hospitalier, Montfermeil); Prof. Canton, Prof. Gérard, and Prof. May (Hôpital Brabois, Nancy); and Prof. Weber (Hôpital Central, Nancy); Dr. Cambon and Dr. Ricome (Centre Hospitalier Intercommunal, Poissy-Saint-Germain-en- Laye); Dr. Trouillet (Hôpital René Dubos, Pontoise); Dr. Fraisse (Hôpital Delafontaine, Saint-Denis); Dr. Graveleau, Dr. Loirat, and Dr. Truelle (Hôpital Foch, Suresnes); Dr. Vuong (Clinique du Vert Galant, Tremblay-en-France); Dr. Blin (Hôpital de Gonesse); Dr. Caen (Hôpital de Corbeil); Dr. Casciani and Dr. Patey (Hôpital de Villeneuve-Saint-Georges); and Dr. Coudray (Centre Hospitalier Général de Longjumeau). Acknowledgment We thank Glaxo-Wellcome for supporting this study. References 1. Whitley RJ. Viral encephalitis. N Engl J Med 1990; 323: Whitley RJ, Soong SJ, Dolin R, Galasso GJ, Ch ien LT, Alford CA. Adenine arabinoside therapy of biopsy-proved herpes simplex encephalitis. National Institute of Allergy and Infectious Diseases collaborative antiviral study. N Engl J Med 1977; 297: Skoldenberg B, Forsgren M, Alestig K, et al. Acyclovir versus vidarabine in herpes simplex encephalitis: randomised multicentre study in consecutive Swedish patients. Lancet 1984; 2: Whitley RJ, Alford CA, Hirsch MS, et al. Vidarabine versus acyclovir therapy in herpes simplex encephalitis. N Engl J Med 1986; 314: Powell KF, Anderson NE, Frith RW, Croxson MC. Non-invasive diagnosis of herpes simplex encephalitis. Lancet 1990; 335: Aurelius E, Johansson B, Skoldenberg B, Staland A, Forsgren M. Rapid diagnosis of herpes simplex encephalitis by nested polymerase chain reaction assay of cerebrospinal fluid. Lancet 1991; 337: Rozenberg F, Lebon P. Amplification and characterization of herpesvirus DNA in cerebrospinal fluid from patients with acute encephalitis. J Clin Microbiol 1991; 29: Lakeman FD, Whitley RJ. Diagnosis of herpes simplex encephalitis: application of polymerase chain reaction to cerebrospinal fluid from brain-biopsied patients and correlation with disease. National Institute of Allergy and Infectious Diseases Collaborative Antiviral Study Group. J Infect Dis 1995; 171: Cinque P, Cleator GM, Weber T, Monteyne P, Sindic CJ, van Loon AM. The role of laboratory investigation in the diagnosis and management of patients with suspected herpes simplex encephalitis: a consensus report. The EU Concerted Action on Virus Meningitis and Encephalitis. J Neurol Neurosurg Psychiatry 1996; 61: Gordon B, Selnes OA, Hart J, Hanley DF, Whitley RJ. Long-term cognitive sequelae of acyclovir-treated herpes simplex encephalitis. Arch Neurol 1990; 47: Kapur N, Barker S, Burrows EH, et al. Herpes simplex encephalitis: long term magnetic resonance imaging and neuropsychological profile. J Neurol Neurosurg Psychiatry 1994; 57: McGrath N, Anderson NE, Croxson MC, Powell KF. Herpes simplex encephalitis treated with acyclovir: diagnosis and long term outcome. J Neurol Neurosurg Psychiatry 1997; 63: MacCabe WR, Jackson GG. Gram negative bacteriemia: etiology and ecology. Arch Intern Med 1962; 110: Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE acute physiology and chronic health evaluation: a physiologically based classification system. Crit Care Med 1981; 9: Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270: Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet 1974; 2: Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet 1975; 1: Lahat E, Barr J, Barkai G, Paret G, Brand N, Barzilai A. Long term neurological outcome of herpes encephalitis. Arch Dis Child 1999; 80: Koskiniemi M, Piiparinen H, Mannonen L, Rantalaiho T, Vaheri A. Herpes encephalitis is a disease of middle aged and elderly people: polymerase chain reaction for detection of herpes simplex virus in the Herpes Simplex Encephalitis CID 2002:35 (1 August) 259

7 CSF of 516 patients with encephalitis. The Study Group. J Neurol Neurosurg Psychiatry 1996; 60: Koskiniemi M, Vaheri A, Taskinen E. Cerebrospinal fluid alterations in herpes simplex virus encephalitis. Rev Infect Dis 1984; 6: Domingues RB, Tsanaclis AM, Pannuti CS, Mayo MS, Lakeman FD. Evaluation of the range of clinical presentations of herpes simplex encephalitis by using polymerase chain reaction assay of cerebrospinal fluid samples. Clin Infect Dis 1997; 25: Domingues RB, Fink MC, Tsanaclis AM, et al. Diagnosis of herpes simplex encephalitis by magnetic resonance imaging and polymerase chain reaction assay of cerebrospinal fluid. J Neurol Sci 1998; 157: Fodor PA, Levin MJ, Weinberg A, Sandberg E, Sylman J, Tyler KL. Atypical herpes simplex virus encephalitis diagnosed by PCR amplification of viral DNA from CSF. Neurology 1998; 51: Aronin SI, Peduzzi P, Quagliarello VJ. Community-acquired bacterial meningitis: risk stratification for adverse clinical outcome and effect of antibiotic timing. Ann Intern Med 1998; 129: Auburtin M, Porcher R, Bruneel F, et al. Pneumococcal meningitis in the intensive care unit: prognostic factors of clinical outcome in a series of 80 cases. Am J Respir Crit Care Med 2002; 165: CID 2002:35 (1 August) Raschilas et al.