Introduction

Severe viral meningoencephalitis has been a real diagnostic and therapeutic challenge for clinicians for decades. The diagnostic as well as treatment possibilities in these patients are often limited. In some patients the disease course might be even brutal with rapid onset of irreversible neuronal damage. Despite the advances in critical care, poor outcome in a significant proportion of patients urges for serious management revision and introduction of new treatment methods [1]. Of particular interest is therapeutic hypothermia (TH). It has well-documented neuroprotective effects and may have a potential use in selected patients with inflammatory central nervous system (CNS) diseases [2].

Although TH has already been used for neuroprotection in a variety of diseases, this method has been rarely applied in the treatment of CNS infections [3, 4]. The most important neuroprotective effects are reduced production of reactive oxygen and nitrogen species, inhibition of the neuroexcitatory cascade, a reduction in proinflammatory cytokine level, maintenance of blood–brain barrier integrity, decreased intracranial pressure (ICP) and finally decreased neuronal apoptosis [5]. Therefore, hypothermia may target the pathophysiological mechanisms that are in effect during viral meningoencephalitis as well as in bacterial meningitis [5, 6]. Adverse events of mild hypothermia are typically minor [2].

We present the case series of the patients with severe form of viral meningoencephalitis treated with TH and describe in detail the methods and the rationale for its use.

Patients and Methods

In the period between February 2009 and October 2010, 11 patients suffering from severe viral meningoencephalitis were treated with TH. All the patients were mechanically ventilated, deeply sedated, and relaxed using midazolam and vecuronium, respectively. Included patients were hospitalized with encephalopathy (depressed or altered level of consciousness ≥24 h, lethargy or change in personality) or ataxia, and had one or more of the following: fever (T ≥ 38C), seizure(s), focal neurologic findings, CSF pleocytosis, abnormal EEG or neuroimaging study (1). The major indication for TH in these patients was impaired carbon dioxide reactivity (CO2R) assessed by Transcranial Doppler (TCD). Because of strong correlation of breath-holding index (BHI) with the severity, particularly with Glasgow coma scale score (GCS) and the outcome of the disease, we defined the BHI level of ≤0.86 as the breakpoint for TH (unpublished data). In patients without temporal acoustic window, minor criteria [optic nerve sheath diameter (ONSD) ≥6.0 mm plus GCS ≤ 8] were required. The following treatment protocol was applied in all the patients: mild hypothermia (32–34°C) accompanied with daily assessment of cerebrovascular CO2R and ONSD. Jugular vein bulb oximetry was available in five patients. Initial antiviral treatment consisted of acyclovir.

TCD Ultrasound

The TCD measurement of CO2R was performed by using a Multidop 4 X (DWL, Sipplingen, Germany) with two 2-MHz pulsed wave probes 1.7 cm in diameter. The software used was TCD-8 for MDX (Version 8.0, Aaslid Rune).

The left and right middle cerebral arteries (MCA) were insonated simultaneously through the temporal bone windows at a depth of 50–55 mm. The probes were secured to the head of the patient with a specially designed spectacle frame that permitted a constant angle of insonation. The mean blood flow velocities (MBFV) were continuously recorded during normal ventilation and during interventions (induced hypercapnia, norepinephrine or urapidil infusion, and hyperventilation). CO2R was assessed using the breath-holding method (disconnection from the ventilator for 30 s in a deep sedated and relaxed patient). BHI was calculated by dividing the percentage of MBFV increase during breath holding by the time (in seconds) of apnea. The normal range of BHI is 1.03–1.65 [7].

ONSD

The ONSD measurements were made using a B-scan ultrasound with a 10 MHz linear probe (Accuson CV70, Siemens Medical Solutions Inc., WA, USA) before and during the induced hypothermia. The ONSD has been shown to be a very reliable measure of ICP. In adults ONSD greater than 5.8 mm correlated with a mean cerebrospinal fluid (CSF) pressure of more than 20 mmHg [8].

TH

We used an internal protocol designed to achieve mild hypothermia (rectal temperature of 32–34°C). Hypothermia was induced by intravenous infusion of cold (+4 to +8°C) isotonic saline (2000 ml/1 h) and maintained with continuous veno-venous hemofiltration (CVVHF) by using a Prismaflex (Gambro Dasco S.p.A, Medolla, Italy) machine for 72–120 h [9]. The blood flow rate was set to 150 ml/min, ultrafiltration rate (UFR) to 0 ml/h and the replacement solution rate was set to 2000 ml/h. Enoxaparin was used for anticoagulation of the circuit.

The hospital Ethics Committee approved the treatment protocol and informed consent was obtained from the relatives of all the patients.

Results

The overall mortality rate was 9% (1/11). Despite TH, one patient died 96 h from admission because of refractory intracranial hypertension. Deceased patient suffered from HSV encephalitis and had a delay in antiviral treatment that commenced on the fifth day of the disease. In the surviving ten patients, the ICU stay ranged from 10 to 44 days. Of these survivors, the outcome was favorable in five patients [Glasgow Outcome Scale score (GOS) 4–5] while the remaining five patients had severe residual neurological deficit (GOS 3). Median Acute Physiology and Chronic Health Evaluation (APACHE II) score in our patients at admission was 24 with predicted death rate of 49.7%. Median GCS at admission was 8, and at ICU-discharge in surviving ten patients it was 15. Demographic and clinical data of the patients are summarized in Table 1.

Table 1 Demographic and clinical data of patients with viral meningoencephalitis treated with hypothermia
Full size table

In five patients with jugular vein bulb catheter (BVJ) a significant improvement of cerebral perfusion during the first 24 h of TH was demonstrated owing to the lactate–oxygen index (LOI). The LOI was derived from arterio-jugular venous oxygen content difference (AjVDO2) and arterio-jugular venous lactate concentration difference (AVDL). The measurements were performed before the induction of hypothermia and afterward intermittently twice daily during the whole period of hypothermia. The LOI was not measured in six patients because of difficult internal jugular vein cannulation. Apart from LOI measurements, we followed the same treatment protocol in these patients with regular ONSD and TCD assessment (CO2R) on daily basis.

The ONSDs gradually decreased despite obvious clinical improvement after hypothermia and sedation ceased. This might implicate that ONSD has a low time-resolution in such setting.

Discussion

The available treatment exists for several types of viral meningoencephalitis [HIV, Herpes simplex virus (HSV), Varicella-zoster virus (VZV), Cytomegalovirus (CMV), Human herpesvirus 6 (HHV 6), B virus, and Influenza virus]. The most used antiviral drugs are acyclovir, valacyclovir, ganciclovir, foscarnet, oseltamivir, and finally Highly Active Antiretroviral Therapy (HAART) [10]. However, aside from acyclovir, the majority of listed antiviral drugs possess humble effectiveness against viruses in the CNS. In addition, in a significant proportion of the patients viral meningoencephalitis etiology remains elusive, particularly during the first several days [1]. Finally, a specific therapy for the majority of viral meningoencephalitis pathogens has not been developed so far (e.g., Tick-borne encephalitis). Thus, the cornerstone of treatment in the majority of viral meningoencephalitis cases is symptomatic treatment only, with the exception of HSV meningoencephalitis. Symptomatic treatment has remained unchanged during the last decades and relies on the administration of hyperosmolar solutions, hyperventilation, barbiturate-induced coma and finally steroids in selected cases [11]. The rationale for their use relies on the belief of beneficial effect of osmotic diuresis and hyperventilation to the brain edema. The neuroprotective effect of decreased brain metabolic demands induced by thiopental infusion accompanied with anticonvulsive action has also been presumed as highly beneficial [12]. However, neither brain edema and severe intracranial hypertension nor increased brain metabolic rate are obligatory in these patients. The brain oxygen and glucose metabolic rates are highly variable and will be rather decreased in the patients with CNS infections [13]. Therefore, an estimation of the cerebral blood flow (CBF) and metabolic coupling without appropriate measurement can only be inaccurate [13]. It should be stressed that ICP reduction cannot be achieved using the hyperventilation or thiopental in all the patients with viral meningoencephalitis. Cerebrovascular dysregulation due to perivascular inflammation makes these therapeutic measures completely useless [12]. Furthermore, the disruption of tight junctions between endothelial cells (which can be roughly estimated by the increased CSF protein concentration) should warn of a possibility of paradoxical brain edema aggravation after mannitol infusion [14]. In our patients, chemoregulation was severely impaired or completely absent.

Therefore, in patients with non-HSV meningoencephalitis and with CO2R loss we are almost helpless. Aside of direct neuronal damage by viruses, the severe alteration in CBF reflected in multisegmental hypoperfusion or hyperemia makes additional harm to the brain. Thus, to decrease ICP (if needed) and preserve adequate cerebral perfusion pressure, cerebrovascular reactivity-independent method should be employed. The appropriate assessment of ICP and intracranial hemodynamic status is mandatory. We used non-invasive and completely harmless methods (ONSD and TCD) for indirect measurement of ICP accompanied with intermittent calculation of LOI as a reliable marker of global cerebral perfusion.

Appreciating the CBF chemoregulation loss and in attempt to optimize CPP we used epinephrine or urapidil infusions in several patients. The target of these interventions was normal values of LOI and/or normal MBFV spectra on TCD. Soon after hypothermia was attained, only mild corrections in mean arterial blood pressure (MAP) were required to achieve the optimal LOI and MBFV, respectively.

We noted neither severe cardiac arrhythmias and coagulopathy nor increment in infection rate during TH. The risk of infection during such a short duration of treatment is maybe even overestimated [15]. Nevertheless, daily blood surveillance cultures as well as low threshold for antibiotic treatment in any suspicion of infection were included in our internal treatment protocol.

The intravascular cooling systems are preferred when available, because their use is much more convenient in comparison with other methods. We used an alternative method (CVVHF) which allows the maintenance of a stable body temperature and gradual rewarming of the patient upon recovery [9].

While there is evidence from animal studies that TH may be beneficial in meningitis and while scarce clinical experience suggests a role for TH, large scale randomized clinical trials are needed. Inclusion criteria in these trials should regard cerebral vasoreactivity status if the selection of the patients and the results aspire to be valid.

The use of TH during viral meningoencephalitis is complex, but it likely has a physiologic role in treating it. Adequate and complete treatment balancing between specific therapy, hypothermia, deep sedation, and controlled CPP early in the course of disease is of paramount importance.

The recovery of CO2R cannot be expected before the fourth day of treatment according to our experience. Therefore, we recommend the use of TH (if indicated) as soon as possible and at least during the first 3 days after presentation. In this case series of 11 patients only one died despite the fact that a predicted death rate measured by the median APACHE II score at admission was 49.7%. Furthermore, median GCS at admission was 8 and in the patients with CNS infections presenting with this value of GCS mortality is at least 33% [16]. In the patients with viral meningoencephalitis treated in our hospital before the implementation of the TH protocol, mortality was 29% if the GCS ≤ 10. According to our findings, TH in carefully selected patients with viral meningoencephalitis holds promise similar to the patients with bacterial meningitis [17]. However, the use of TH in this indication requires further investigation before definite conclusions can be drawn.