SEARCH

SEARCH BY CITATION

Keywords:

  • Encephalitis of unknown aetiology;
  • functional deficits;
  • long-term sequelae;
  • neurological sequelae;
  • outcome

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Clin Microbiol Infect 2011; 17: 621–626

Abstract

Encephalitis is caused by a variety of conditions, including infections of the brain by a wide range of pathogens. A substantial number of cases of encephalitis defy all attempts at identifying a specific cause. Little is known about the long-term prognosis in patients with encephalitis of unknown aetiology, which complicates their management during the acute illness. To learn more about the prognosis of patients with encephalitis of unknown aetiology, patients in whom no aetiology could be identified were examined in a large, single-centre encephalitis cohort. In addition to analysing the clinical data of the acute illness, surviving patients were assessed by telephone interview a minimum of 2 years after the acute illness by applying a standardized test battery. Of the patients with encephalitis who qualified for inclusion (n = 203), 39 patients (19.2%) had encephalitis of unknown aetiology. The case fatality in these patients was 12.8%. Among the survivors, 53% suffered from various neurological sequelae, most often attention and sensory deficits. Among the features at presentation that were associated with adverse outcome were older age, increased C-reactive protein, coma and a high percentage of polymorphonuclear cells in the cerebrospinal fluid. In conclusion, the outcome in an unselected cohort of patients with encephalitis of unknown aetiology was marked by substantial case fatality and by long-term neurological deficits in approximately one-half of the surviving patients. Certain features on admission predicted an unfavourable outcome.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Encephalitis is a clinical syndrome characterized by central nervous symptoms such as headache, mental status change, neurological deficits, seizures and fever [1]. The disease represents a significant burden in terms of morbidity, mortality and healthcare costs [2,3].

Encephalitis-like illnesses are caused by several pathogenically distinct mechanisms, including infections, post-infectious immune reactions, other inflammatory processes, paraneoplastic conditions and intoxications [4,5]. Despite considerable progress in imaging techniques and the detection of microbial pathogens, diagnostic efforts fail to identify an aetiological agent in many patients with encephalitis [6,7]. Consequently, clinicians commonly care for patients with encephalitis of unknown aetiology for prolonged periods of time.

Treatment options for patients with encephalitis are limited. Antiviral drugs that are typically used to empirically treat patients with encephalitis of unknown aetiology cover only a small spectrum of the potentially involved viruses (i.e. herpes viruses). Many patients are treated with other pharmacological agents, such as corticosteroids, in addition to anti-epileptic drugs indicated in patients with evidence of seizures [1]. The uncertainty regarding the aetiology and prognosis in these patients hampers the optimal management and represents a major burden for the patients themselves as well as those who care for them [8].

Little is known regarding the outcome and especially the long-term neurological outcome of patients with encephalitis, in whom no aetiological diagnosis can be established [9]. In the present study, we therefore assessed the long-term outcome in a retrospectively identified cohort of consecutive patients with encephalitis, in whom no aetiology could be identified at the time of hospitalization. Specifically, we assessed the health status, including the presence of permanent neurological sequelae and the daily performance of the surviving patients by structured telephone interviews for a minimum of 2 years after discharge from hospital. Furthermore, we examined which clinical and laboratory parameters at the time of admission were significantly associated with an unfavourable outcome (defined as death or long-term functional impairment). Finally, we assessed how many of the patients in our collective presented with one of the ten characteristic clinical profiles of encephalitis as recently proposed by the investigators of the California Encephalitis Project [5].

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

The study was approved by the institutional review board (KEK) of the Kanton of Bern, Switzerland. The study population consisted of patients who had a clinical diagnosis of encephalitis and in whom diagnostic studies had failed to identify an aetiological cause of their disease. Patients hospitalized between the years 1992 and 2004 in the Department of Neurology, University Hospital Inselspital, Bern, were retrospectively identified based on the Hospital records. In a first step, all patients were identified in whom the neurological diagnosis code of discharge during the study period indicated any form of encephalitis (n = 267). Insufficient information was available for 35 of these patients, leaving 232 patients for further study. Next, a review to confirm the aetiological diagnosis was performed in all patients who were discharged with a code indicating a specific aetiology or a clearly-defined clinical syndrome (n = 83). The charts of the remaining patients (n = 149) were examined in detail to determine their eligibility for the outcome study of patients with encephalitis of unknown aetiology.

To be included in the outcome study of a patient with encephalitis of unknown aetiology, the following inclusion criteria had to be met: signs and symptoms of encephalitis/encephalopathy with quantitative or qualitative impairment of level of consciousness for more than 24 h duration and at least one of the following signs or symptoms: body temperature of >38.0°C, seizure, focal neurological signs, cerebrospinal fluid (CSF) pleocytosis, or neuroradiological or neurophysiological signs of encephalitis. Exclusion criteria were: (i) age <16 years; (ii) start of symptoms >3 weeks before hospitalization; (iii) HIV infection; (iv) chronic inflammatory systemic disease or neoplasia; (v) abuse of intravenous drugs or alcohol; (vi) a diagnosis of migraine with aura; (vii) residence in Switzerland for <6 months before beginning of symptoms; and (viii) proven or probable infectious or non-infectious aetiology of encephalitis. With this selection procedure, a total of 39 patients with a diagnosis of acute encephalitis of undetermined aetiology were retained for further analysis. All relevant data were extracted from the charts and radiology and laboratory reports and entered into a database.

For the assessment of long-term outcome, patients were contacted by telephone between 2006 and 2007 and were asked to provide their informed consent for participation in the study. Using a standardized questionnaire, residual neurological symptoms and their impact on the patient’s daily activities were queried. In addition, four standardized tests were performed: (i) a six-item orientation memory concentration test [10]; (ii) a questionnaire for attention deficits [11]; (iii) the short version of the Geriatric Depression Scale [12]; and (iv) the Epworth Sleepiness Scale [13].

Outcome was defined as ‘healed’ if the patient was alive, there were no signs and symptoms of encephalitis, and the neuropsychological test results were normal. ‘Deficit’ was defined as any abnormal neuropsychological test result or overt neurological deficit. Death was categorized based on clinical judgement as either the result of encephalitis or unrelated causes.

Statistical analysis was performed with statview, version 5.0 (SAS Institute Inc., Cary, NC, USA). Proportions were compared with the chi-square test or Fisher’s exact test as appropriate. Differences between means were assessed by Student’s t-test. p ≤ 0.05 (two-tailed) was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Of the 149 patients with a discharge code that did not suggest an obvious aetiology, sixteen were excluded because they were <16 years old, and thirteen were excluded because the diagnosis of encephalitis could not be confirmed based on the detailed chart review. Of the remaining 120 patients, 39 were classified as encephalitis of unknown origin (Table 1) and constituted the study collective for this analysis. For the other patients (n = 164), the proven or likely aetiologies based either on discharge records (n = 83) or on detailed chart reviews (n = 81) are shown in Table 1.

Table 1.   Diagnosis in 203 patients with encephalitis
Encephalitis with unknown aetiology39 (19.2%)
Encephalitis with infectious aetiology107 (52.7%)
 Neuroborreliosis20
 Herpes simplex virus encephalitis16
 HIV-associated encephalopathy13
 Other viral encephalitis6
 Meningitis and brain abscess9
 Para- and post-infectious15
 Previous encephalitis9
 Other infectious19
Non-infectious causes57 (28.1%)
 Acute demyelinating disease, multiple sclerosis15
 Sarcoidosis of the central nervous system8
 Tumour-associated/paraneoplastic6
 Vascular/migraine5
 Various others23

The patients in the study collective presented with a range of symptoms typically seen in patients suffering from encephalitis (Table 2) [5]. The average age was 47 years and males predominated slightly (56%). Approximately one-half (53%) of patients had prodromal symptoms, most often fever (43%) or respiratory symptoms (25%). The mean delay from onset of neurological symptoms to referral was 9 days. Almost one-third of patients presented with seizures, but only 12% were comatose at presentation. All but two patients had elevated white blood cells in the CSF, with a mean cell count of 74 per mm3 and a clear predominance of mononuclear cells. CSF protein levels were elevated in two-thirds of patients (n = 30), whereas glucose level was mostly in the normal range (n = 37). Only one-fifth of patients had a C-reactive protein (CRP) level >20 mg/dL (normal value <5 mg/dL). Cases clustered in the months of January to June with 26 of 39 cases (74%).

Table 2.   Characteristics of 39 patients with acute encephalitis of unknown aetiology
  1. CSF, cerebrospinal fluid.

Total, n (%)39 (100)
Age (years), mean ± SD47.2 ± 16.6
Male gender, n (%)22 (56.4)
Swiss nationality, n (%)35 (89.7)
Travel abroad in last 4 weeks1, n (%)4 (10.2)
Days from onset of illness to hospital admission, mean ± sd8.8 ± 9.4
Admission to intensive care, n (%)9 (23.0)
Prodrome (any), n (%)21 (53.8)
 Fever, n (%)17 (43.5)
 Respiratory symptoms, n (%)10 (25.6)
 Gastrointestinal symptoms, n (%)5 (12.8)
 Myalgia/arthralgia, n (%)8 (20.5)
Rash, n (%)1 (2.5)
Seizure, n (%)11 (28.2)
Coma, n (%)5 (12.8)
Death, n (%)5 (12.8)
Days of hospital stay, mean ± SD18.8 ± 18.2
CSF white blood cell count per mm3, mean ± sd74.3 ± 74.4
CSF polymorphonuclear cells (%), mean ± SD15.4 ± 29.1
CSF protein level (mg/dL), mean ± sd1.0 ± 5.4
CSF glucose level (mmol/L), mean ± SD3.2 ± 0.7
Initial abnormal neuroimaging, n (%)19 (65.5)
Steroid treatment, n (%)10 (25.6)
 Days, mean ± sd11.3 ± 13.2
Antiviral treatment, n (%)29 (74.3)
 Days, mean ± SD8.0 ± 5.6
Antiepileptic treatment, n (%)14 (35.8)
White blood cell count per mm3, mean ± sd8.5 ± 3.9
Platelet count per mm3, mean ± SD282 ± 85
C-reactive protein >20 mg/dL, n (%)8 (20.5)

Clinical classification

Only 11 of 39 patients with encephalitis of unknown aetiology presented with one of the ten clinical profiles that have been described by Glaser et al. [5] in their large cohort of patients with encephalitis. Three patients presented with temporal lobe involvement (profile 1), two with cerebellar involvement (profile 3), three with primary seizures with rapid recovery (profile 7), two with psychosis (profile 8), and one with multifocal white matter lesions (profile 10). The clinical presentation of the other 28 patients did not clearly fit one of the ten profiles.

Outcome

Of the 39 patients, 34 (87.1%) were alive at the time of follow-up, although two of the 34 patients who were still alive could not be reached for interview. Five (12.8%) patients had died with severe encephalopathy, two of them in the acute phase of illness. Among the 32 patients who were available for follow-up interviews, 15 (46.8%) patients were considered to be healed and 17 (53.1%) had neurological deficits. A wide range of symptoms was reported by the patients, with the most frequent deficits being attention and sensory deficits (Fig. 1).

image

Figure 1.  Long-term outcome. For assessment of long-term outcome, patients were contacted by phone between 2006 and 2007. (a) Four standardized tests were performed: a six-item orientation memory concentration test [10], a questionnaire for attention deficits [11], the short version of the Geriatric Depression Scale [12], and the Epworth Sleepiness Scale [13]. (b) Using a standardized questionnaire, residual and neurological symptoms, as well as their impact on patients’ daily activities, were queried.

Download figure to PowerPoint

Factors that were associated with death were older age, comorbidity, coma at referral, low CSF white blood cell count, higher percentage of polymorphonuclear cells in the CSF, and a CRP value >20 mg/dL (Tables 3 and 4). Full recovery was associated with relative young age, lack of comorbidity, and presentation with minor focal neurological deficits but intact vigilance.

Table 3.   Clinical characteristics of 37 patients with encephalitis of unknown aetiology stratified according to outcome
 HealedDefectDeathp
  1. NS, nonsignificant.

  2. Two of 39 patients were excluded because they were unavailable for interview at follow-up.

  3. aChi-square or Fisher’s exact test used for the comparison of patients who died with those who survived.

  4. bChi-square test for trend.

  5. cChi-square test comparing all three groups with each other.

Total, n (%)15 (100)17 (100)5 (100) 
Age (years), mean ± sd42.2 ± 17.147.0 ± 13.862.4 ± 19.20.02a
Male gender, n (%)10 (66.6)8 (47.0)3 (60.0)NS
Comorbidity, any, n (%)4 (26.6)7 (41.1)4 (80.0)0.05b
Days to hospital admission, mean ± SD7.1 ± 6.510.4 ± 10.78.4 ± 13.1NS
Intensive care, n (%)1 (6.6)4 (23.5)3 (60.0)0.01b
Prodrome, any, n (%)10 (66.6)6 (35.2)3 (60.0)NS
 Fever, n (%)8 (53.3)6 (35.2)2 (40.0)NS
Minor focal deficits without mental status changes, n (%)11 (73.3)3 (17.6)3 (60.0)0.002c
Seizure, n (%)5 (33.3)4 (23.5)1 (20.0)NS
Coma, n (%)1 (6.6)1 (5.8)3 (60.0)0.01a
Glasgow coma scale, mean ± sd13.3 ± 2.613.9 ± 2.67.6 ± 4.4<0.001a
Days of hospital stay, mean ± SD12.0 ± 7.020.5 ± 16.945.5 ± 16.90.03a
Steroid treatment, n (%)3 (20.0)5 (29.4)2 (40.0)NS
Antiviral treatment, n (%)11 (73.3)13 (76.4)3 (60.0)NS
Antiepileptic treatment, n (%)5 (33.3)6 (35.2)2 (40.0)NS
Table 4.   Laboratory findings in 37 patients with encephalitis of unknown aetiology stratified according to outcome
 HealedDefectDeathp
  1. CSF, cerebrospinal fluid; CRP, C-reactive protein; NS, nonsignificant.

  2. Two of 39 patients were excluded, because they were unavailable for interview at follow-up.

  3. aChi-square or Fisher’s exact test used for the comparison of patients who died with those who survived.

  4. bChi-square test for trend.

Total, n (%)15 (100)17 (100)5 (100) 
CSF white blood cell count per mm3, mean ± sd85.8 ± 82.185.0 ± 77.223.0 ± 77.20.09a
CSF polymorphonuclear cells (%), mean ± SD10.9 ± 28.08.5 ± 14.347.0 ± 6.90.005a
CSF protein level (mg/dL), mean ± sd0.77 ± 0.6421.4 ± 82.01.28 ± 1.75NS
CSF glucose level (mmol/L), mean ± SD3.5 ± 0.823.1 ± 0.63.2 ± 0.5NS
Abnormal neuroimaging, n (%)6 (40.0)7 (41.1)3 (60.0)NS
White blood cell count per mm3, mean ± sd8.0 ± 3.58.8 ± 4.78.9 ± 2.3NS
Platelet count per mm3, mean ± SD257 ± 45292 ± 85307 ± 171NS
CRP >20 mg/dL, n (%)0 (0)2 (11.6)5 (100)0.02b

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

The present study focussed on patients with a clinical diagnosis of encephalitis, in whom no aetiology could be established. The rationale for this approach was based on the fact that the management of these patients is particularly challenging, not least because the prediction of outcome is more uncertain than in patients with a known aetiology of encephalitis. The results obtained show that, in a large, unselected cohort of patients with encephalitis, 19.2% were classified as having encephalitis of unknown origin. In these patients, mortality attributable to encephalitis was 13% and only one-half of the surviving patients interviewed at least 2 years after the acute illness were considered to be cured and free of long-term neurological impairment.

Compared to the other series, the percentage of patients in whom no aetiology could be identified was relatively low in the present study (19%). For comparison, in the large, prospective California Encephalitis Project (CEP), no aetiology could be identified in 63% of the patients [5,6]. In a large study from Finland, 41% of patients had no identified aetiology [8], whereas, in a smaller study from the same country, the aetiology could not be determined in 64% of 42 patients with encephalitis [4]. The corresponding figure was 48% in a similar study from Sweden [7]. Finally, a prospective study in Taiwan failed to identify the aetiology in 31% of 124 enrolled patients [14].

These differences likely reflect differences in the study design, in the population of patients that were studied, the local epidemiology of encephalitis, and the available methodology and stringency applied to the criteria for establishing the cause of disease. The present study was based on an unselected patient population in a large tertiary care hospital in Switzerland, which had included patients since the early 1990s and was retrospective. By contrast, the CEP study, which showed one of the highest percentages of patients with unknown aetiology, represents a prospective collection of CSF samples and patient data at a State reference laboratory. The selection of cases with this design likely favours the selection of difficult-to-diagnose cases. Furthermore, the very strict criteria required for diagnosing the aetiology of encephalitis in the present study contributed to the high rate of unknown causes [5,6].

By contrast to the widely diverse rates of unknown aetiologies, the spectrum of identified causes of encephalitis was quite similar between our study and previous studies [4–8,14,15]. As in most of these previous studies, the biggest group of identified aetiologies in the present study was viral infections (herpes simplex virus, varicella zoster virus, and other neuropathic viruses). A second important group consisted of infections with nonviral pathogens, such as bacteria (in particular Borrelia encephalitis), whereas a third group encompassed a wide range of non-infectious or para-infectious causes, in particular demyelinating disorders.

The focus of the present study was on outcome (i.e. mortality and long-term sequelae). For the most part, the available data regarding outcome in patients with encephalitis is derived from series of patients infected with one specific pathogen. The almost 100% mortality of rabies represents the extreme spectrum of the disease. Infections by herpes simplex virus, various flaviviruses (e.g., Japanese encephalitis) and paramyxoviruses (e.g. Nipah and Hendra virus) are among the aetiologies that are associated with mortality rates typically in excess of 10% [1,16–20]. Enterovirus infection, on the other hand, generally leads to a relatively mild illness with low mortality, even though some serotypes such as enterovirus 71 can cause severe and fatal illness [21]. Besides mortality, neurological sequelae are common after encephalitis. Herpes simplex encephalitis is well known for the often severe neurological deficits, and many other forms of encephalitis are also often associated with debilitating neurological deficits [19,22]. Deficits range from subtle neuropsychological impairment and memory deficits to severe sensomotor sequelae and chronic epileptic seizures.

The present study confirms the significant morbidity and mortality of encephalitis in a group of patients with unknown aetiology. Importantly, the mortality attributable to acute encephalitis was 13%, which is comparable to the current mortality for herpes simplex encephalitis [23]. Equally important, the surviving patients in the present study showed remarkably high rates of chronic deficits, in that only one-half of them were considered to be completely recovered years after the acute illness. Hokkanen and Launes [17] demonstrated that neuropsychological deficits tend to improve rather than worsen after acute encephalitis. Thus, our patients, who on average were re-examined several years after their acute infection, most likely had even more severe deficits in the early phase after encephalitis. Thus, the short- and long-term neurological outcome in patients with encephalitis of unknown aetiology is characterized by a high rate of persisting neurological sequelae.

We found that certain factors were prognostic for an unfavourable outcome. These factors included advanced age, increased CRP, a predominance of neutrophil granulocyte in the CSF, and coma. Of note, the presence of mild focal neurological deficits with preserved vigilance was predictive of favourable outcome. Little is known from other studies about the prognostic factors in encephalitis. In herpes simplex encephalitis, two predictors of adverse outcome were found to be a high clinical severity score on admission and a long delay of antiviral therapy [20]. In children, young age, seizures, coma and documented viral infection have been associated with poor outcome [24,25]. A recent study from France also confirmed age and coma as factors associated with death [26]. Underlying illnesses, sepsis, mechanical ventilation, and infection caused by herpes viruses, Mycobacterium tuberculosis or Listeria monocytogenes, were other adverse risk factors in this study [26].

Glaser et al. [5] have identified ten clinical profiles, into which approximately one-half of the 1570 patients in their study could be categorized. For most profiles, the mortality rate was approximately 10%, which is similar to the collective in the present study. However, two profiles were associated with much higher mortalities (profile 5 ‘diffuse cerebral oedema’ and profile 6 ‘intractable seizures’) [5], whereas two other profiles [e.g. ‘cerebellar disease’ (profile 3) and ‘seizures with rapid recovery’ (profile 7)] showed very low mortality rates (2% and 0%, respectively). We attempted to categorize our patients into the ten profiles. Only 11 of our 39 (28%) patients clearly fit one of the profiles, with five of them fitting one of the two low-mortality profiles (i.e. profiles 3 and 7), and none fitting the high mortality profiles. More prospective studies are needed to assess the value of the described profiles for the prediction of outcome in patients with encephalitis of unknown aetiology.

In summary, the present study shows that patients with encephalitis of unknown aetiology, as a group, suffer from significant mortality (approximately 13%) and long-term morbidity (approximately 50%) with a wide range of deficits. Some clinical parameter on admission, in particular older age, coma and elevated CRP, were associated with an adverse outcome.

Footnotes
  • 1

    Spain, Tunisia, Malta, Thailand

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

This study did not receive any financial sponsoring. All authors declare that there are no conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References