Neuroglucopenia and Metabolic Distress in Two Patients with Viral Meningoencephalitis: A Microdialysis Study

A 38-year-old, otherwise healthy, woman presented to an external hospital with drowsiness and aphasia after a week of recurrent episodes of fever (>39 °C) accompanied by muscle weakness and an exanthema of trunk, arms, and legs. Travel- and exposure-history was uneventful and routine vaccination was performed. Initial computed tomography (CT) of the head was normal, and cerebrospinal fluid (CSF) analysis revealed a mild pleocytosis (3 lymphocytes and 1 granulocyte per µl), elevated protein of 311 mg/dl (normal range 15–40 mg/dl), and a normal glucose concentration (>60 % of serum glucose). Neurologic examination revealed a patient with altered level of consciousness, mixed aphasia, and neck stiffness. Serology for Tick-borne encephalitis virus (TBEV) and Borrelia burgdorferi, as well as polymerase chain reaction (PCR) of herpes simplex virus (HSV), varicella-zoster virus (VZV), Enterovirus, and Mycobacterium tuberculosis remained negative in serum and CSF samples. Serum C-reactive protein (CRP) was 4.21 mg/dl (normal range <0.85 mg/dl). Neuroimaging showed meningeal enhancement, supporting the clinically and laboratory suspected diagnosis of meningoencephalitis. Intravenous acyclovir (750 mg TID), ceftriaxone (2 g BID), and moxifloxacin (400 mg SID) were started. Twelve hours later, the patient developed repeated seizures, necessitating mechanical ventilation. Follow-up CT of the brain showed diffuse brain swelling (Fig. 1a), and she was transferred to our unit. Brain magnetic resonance imaging (MRI) revealed signal enhancement of the caput nuclei caudati as well as bilateral hyperintense white matter changes of the perivascular space in fluid-attenuated inversion recovery (FLAIR) sequence, suggestive for viral encephalitis. West-Nile virus (WNV) IgM-antibodies were detected in serum specimens. HIV, Hepatitis B and C virus, Borrelia burgdorferi, Listeria monocytogenes, TBEV, Hantavirus, Chikungunya virus, and Sandfly virus serology were all negative, so was PCR screening for HSV, VZV, Human cytomegalovirus (CMV), Epstein-Barr virus (EBV), Human herpesvirus 6 (HHV6) and enterovirus. Due to the patient’s clinical and radiologic deterioration, multimodal neuromonitoring probes were inserted for the assessment of ICP (NEUROVENT-P-TEMP, Raumedic^®, Helmbrechts, Germany), P_btO₂ (Licox^®, Integra LifeSciences, Saint Priest, France), and brain metabolism (71 High Cut-Off Brain Microdialysis Catheter, M Dialysis AB, Stockholm, Sweden). Probe location in normal-appearing white matter was confirmed by CT of the brain.

During the observation time, ICP remained <20 mmHg in 99 % of monitoring time, P_btO₂ dropped to 16 mmHg during the initial hours and increased to normal levels after immediate augmentation of cerebral perfusion pressure (CPP) to >70 mmHg. All values remained stable during the neuromonitoring time, except for fluctuations of CPP without decrease of P_btO₂ during days five and six (Fig. 2a). Parenteral amino acid supplementation (Aminoven^® 3.5 %, Fresenius Kabi Austria, Graz, Austria) was started at 10 h, enteral nutrition (EN) (Fresubin^® Original Neutral, Fresenius Kabi Austria) at 27 h after admission. Intravenous amino acid infusions are given until 70 % of the calculated energy demand are reached with EN. Intravenous insulin was started shortly after EN at a rate of 2 U/h when serum glucose concentration reached 10 mM/l (180 mg/dl). Daily median brain interstitial glucose concentration ranged from 0.35 mM/l (IQR 0.2–0.65, day two) to 2.38 mM/l (IQR 2.2–2.5, day eight). In 23 % of monitoring time CMD-glucose was <0.7 mM/l, in 3 % of time <0.2 mM/l. Our standard serum glucose protocol aims at maintaining systemic glucose levels between 6.1 and 8.3 mM/l (110–150 mg/dl). When neuroglucopenia was detected, we initiated a more liberal glucose regimen, allowing values up to 10 mM/l (180 mg/dl).

Serum glucose measurements were above 6 mM/l in 79 % of analyses. Cerebral glucose levels were influenced by modifications in insulin therapy, nutrition, and systemic glucose infusion, as detailed in Fig. 3c. Briefly, insulin application or dose augmentation was associated with a decrease in CMD-glucose and episodes of neuroglucopenia in the early phase of neuromonitoring. Stopping insulin infusion due to neuroglucopenia preceded a rapid increase in both systemic and cerebral glucose. On day 4, when the patient received combined EN and parenteral amino acid supplementation, brain glucose levels remained above 0.7 mM/l. When the caloric requirements were reached with EN, Aminoven^® was stopped. This was associated with a decrease in cerebral glucose to <0.7 mM/l. Bolus systemic glucose infusion (20 %) was associated with an intermediate increase in systemic and brain glucose levels. Overall, however, we did not find a significant correlation (Spearman’s rho) between serum and CMD-glucose (p = 0.38).

CMD-lactate and lactate-to-pyruvate-ratio (CMD-LPR) was constantly elevated (>4 mM/l in 98 % of measurements and >30 mM/l in 99 % of measurements), and CMD-pyruvate was within normal range (>70 µM/l in 99 % of measurements). The brain metabolic profile suggested non-ischemic metabolic distress, confirmed by P_btO₂ levels >20 mmHg, recently defined as mitochondrial dysfunction (Fig. 2b) [15].

From day six on, the patient developed short episodes of raised ICP without significant decrease in P_btO₂, which were successfully treated with hypertonic saline (10 %, 100 ml).

Clinical condition improved and the patient was extubated on day ten. At that time she followed commands, neurologic examination did not reveal any focal motor deficit, but the patient still suffered from encephalopathy with disorientation to time and person. Her neuropsychological status improved over the next weeks, and she could be discharged to an external hospital without neurologic deficit.

A 17-year-old, otherwise healthy, male patient presented with an upper respiratory tract infection and was treated with amoxicillin (dose not available). He was admitted to a peripheral hospital when he developed skin rash and muscle pain. Physical examination revealed tonsillitis and cervical lymphadenopathy; neurological examination was normal at this time. Differential blood count showed monocytosis (18 %, normal range 1–12 %); CRP was elevated (1.21 mg/dl, normal range <0.5 mg/dl).

Within 24 h, the patient developed drowsiness and neurologic exam showed neck stiffness. CSF analysis revealed mild pleocytosis (7 cells/µl, normal range 0–4/µl) and normal glucose levels (CSF/serum glucose ratio of 0.66, normal range >0.6). Antiviral therapy with acyclovir (500 mg TID) and antimicrobial chemotherapy with ceftriaxone (2 g BID) was initiated. The patient further deteriorated and was mechanically ventilated and transferred to our unit. On arrival, neurologic exam revealed bilateral non-reactive dilated pupils, which most likely developed during the 30 min of helicopter transfer. Head CT scan demonstrated global cerebral edema (Fig. 1b) with transtentorial herniation. The patient immediately underwent bilateral craniectomy. Neuromonitoring probes were placed into the right frontal white matter to measure ICP, brain metabolism, CBF, and brain water content (BWC) (HEMEDEX QFlow500™ Perfusion Probe, HEMEDEX™, Cambridge, MA). A broad diagnostic workup revealed Epstein-Barr virus (EBV)-associated encephalitis confirmed by serology (IgM/IgG antibodies in CSF and serum) and PCR (EBV-DNA in CSF). Intravenous ganciclovir (300 mg BID) and doxycycline (100 mg BID) treatment was initiated.

After craniectomy, ICP was still elevated (episodes of ICP >30 mmHg) and decreased only after osmotherapy, therapeutic hypothermia (33 °C), as well as thiopental (1.7 mg/kg/h) and propofol infusion (1 %, 10 ml/h), aiming to achieve a burst-suppression pattern on surface electroencephalogram (EEG). CBF was low (median 16 ml/100 g/min, IQR 13.7–18.1) during the first 24 h of monitoring and increased thereafter. BWC started at 86 % and decreased to 82 % during repeated osmotherapy concomitant to radiologic edema regression (significantly lower BWC values around 70 % have been described in normal-appearing brain tissue on head CT scans in acutely brain-injured patients) [16]. As ICP normalized on day four, intravenous thiopental and propofol infusions could be gradually reduced; slow rewarming was initiated, and the patient’s body temperature reached 36 °C on day nine of monitoring (Fig. 3a).

During the time of hypothermia, CMD-glucose levels were low (64 % of measurements <0.7 mM/l, 4 % <0.2 mM/l), despite a liberal systemic glucose management. Overall, 80.7 % of systemic glucose measurements were >6 mM/l. Insulin application (3 U/h) was started together with parenteral amino acid supplementation after 11 h, EN (Isosource^® Standard Neutral Smartflex^®, Nestlé Austria, Vienna, Austria) was started after 22 h. CMD-lactate, CMD-pyruvate, and CMD-LPR were within normal range [median of 1.9 mM/l (IQR 1.7–2.2), 84 µM/l (IQR 69.3–93.7) and 23 (IQR 21.4–25.5), respectively].

After the patient’s body temperature reached 36 °C, CMD-glucose levels remained low (54 % <0.7 mM/l, 11 % <0.2 mM/l), CMD-lactate and CMD-LPR exceeded pathologic thresholds for extended periods of time (32.4 % >4 mM/l and 46.3 % >30, respectively). As CMD-pyruvate levels were at a median of 92 µM/l (IQR 71–149) and 94 % >70 µM/l, several episodes (32.7 % of measurements) of mitochondrial dysfunction were observed (Fig. 3b).

When neuroglucopenia was detected, ICP was within normal range (<20 mmHg) in 80.2 %, CPP was >60 mmHg in 97.7 %, and CBF was >17 ml/100 g/min in 69.2 % of time. An association between systemic and CMD-glucose levels is shown in Fig. 3c. Serum- and CMD-glucose levels were significantly correlated (r = 0.47, p < 0.0001); insulin-dosage was negatively associated with CMD-glucose levels, and intermittent pauses of EN and PN were associated with a decrease in CMD-glucose.

The patient gradually improved and could be extubated after 23 ventilator days. He showed an atactic-ballistic movement disorder, discrete paresis of the lower limb (especially dorsal flexors), and transient psychosis. He could be discharged to an acute rehabilitation facility after 7 weeks. Fortunately, he improved within 3 months to a normal functional level with only mild neuropsychological deficits (modified Rankin Scale Score of 1, respectively).