Transtentorial herniation syndrome from meningococcal meningitis in a young woman: the case for neurocritical care

Abstract

We report a case of a previously healthy early adolescent female who presented with meningococcal meningitis. While in hospital, she had marked neurologic deterioration with clinical herniation from malignant cerebral oedema. She was transferred to a neurocritical care centre where she underwent invasive intracranial pressure (ICP) and brain tissue oxygen (P_btO₂) monitoring. Early in her course, she demonstrated a compete absence of autoregulation, with pressure passive cerebral blood flow. As a result, maintaining a mean arterial pressure between 50 mm Hg and 60 mm Hg, which ensured adequate cerebral oxygenation, while avoiding increases in ICP. Although her course was initially complicated by bilateral optic neuropathy, she has subsequently made a full neurologic recovery and is now undertaking postsecondary education. This case highlights that access to specialist neurocritical care, guided by neurophysiologic monitoring of ICP and P_btO₂, may help improve outcomes, even among those patients with catastrophic cerebral oedema from bacterial meningitis.

Background

Neisseria meningitidis is a common cause of community-acquired bacterial meningitis. Even with prompt recognition and treatment, the case-fatality rate is 15%,¹ and 28% of patients are left with neurologic complications.² Cerebral spinal fluid (CSF) infection results in cytokine release and marked inflammatory response, which leads to disruption of the blood–brain barrier.³ Blood–brain barrier breakdown results in the loss of cerebral autoregulation, cerebral oedema and increased intracranial pressure (ICP).⁴ Given these pathophysiologic changes, several studies have advocated for specialised neurocritical care, including the use of invasive multimodal neuromonitoring, in patients with severe bacterial meningitis.⁵

Case presentation

A previously healthy early adolescent female was brought to the emergency room in a community hospital reporting of a headache, shoulder pain, nausea, vomiting, diarrhoea, confusion and photophobia. On examination, she had a temperature of 36.6°C, blood pressure of 100/55 mm Hg, heart rate of 115 bpm, respiratory rate of 20 bpm and oxygen saturation of 96% on room air. Neurologically, she was confused and not able to follow commands. Her Glasgow Coma Scale was initially 13 (eyes—4, motor—5, verbal—4). She had no meningismus. The remainder of her examination was normal. Laboratory examination revealed a white blood cell (WBC) of 22.5×10⁹/L. Urinalysis and β-choriogonadotropin were negative. An initial CT of her brain was normal (figure 1A, B). She was started on cefotaxime, vancomycin and acyclovir. She did not received dexamethasone. Lumbar puncture demonstrated marked leucocytosis with a WBC count of 17 710×10⁹/L. CSF glucose was normal at 2.6 mmol/L and protein was elevated at 9.4 g/L. The CSF gram stain showed 2+ gram-negative diplococci and the PCR was positive for Neisseria meningitidis. Her level of consciousness subsequently deteriorated, and she was endotracheally intubated for transfer to a regional medical centre on the first day of admission. MRI of the brain was performed on admission day 2, which demonstrated extensive restricted diffusion and leptomeningeal enhancement in both hemispheres and cerebellum consistent with meningoencephalitis. On interruption of sedation on day 2, she would localise briskly with all four extremities. Early the next morning (admission day 3), her pupils became dilated to 9 mm and unresponsive to light. She had absent corneal reflexes bilaterally and an absent gag reflex. She had a strong cough. She had flexure posturing to pain in all four extremities. CT brain was performed and demonstrated marked cerebral oedema and obliteration of her basal cisterns (figure 1C, D). On consultation with neurocritical care services at the provincial quaternary intensive care unit (ICU), the following immediate resuscitative measures were instituted: hyperventilation (targeting a partial pressure of carbon dioxide of 35–40 mm Hg), therapeutic hypothermia (target temperature 34°C–35°C), sedation and pharmacologic paralysis and administration of hypertonic saline and mannitol. She was then transferred to the neurocritical care unit for consideration of invasive multimodal neuromonitoring.

Figure 1.

CT of her brain was normal on admission (A, B). On day 2, she had marked cerebral oedema with obliteration of the basal cisterns (C,D).

Treatment

On arrival to the neurocritical care unit, an intraparenchymal ICP catheter and brain tissue oxygen monitor were inserted into the right frontal lobe through a dual lumen cranial access bolt. ICP, brain tissue oxygen tension (P_btO₂) and mean arterial pressure (MAP) values were captured using ICM+Brain Monitoring Software (Cambridge Enterprise, Cambridge, United Kington). Observing the responses of P_btO₂ and ICP in response to fluctuations in MAP allowed us to target the optimal MAP to ensure the balance between adequate cerebral oxygen delivery and ICP. Furthermore, ICM+calculates the Pressure-Reactivity Index (PRx), which is a moving Pearson correlation coefficient between 30-consecutive, 10 s averaged values of MAP and corresponding ICP.⁶ A PRx above or below 0.3 indicates dysfunctional or preserved autoregulation, respectively. PRx is used to determine the optimal MAP where PRx is the most negative over a 4-hour window.⁷

During the first 4 days of multimodal neuromonitoring, she displayed marked intracranial hypertension and complete absence of autoregulation (PRx>0.3). This is evident in figure 2 as her MAP was positively associated with her ICP. When her MAP was above 65 mm Hg, her ICP went above 20 mm Hg. As such, during this time, we maintained a target MAP between 50 and 60 mm Hg in order to ensure a P_btO₂ threshold of greater than 15 mm Hg, while avoiding episodes of elevated ICP. In addition, she remained sedated using intravenous infusions of midazolam and fentanyl. Her temperature was maintained between 34°C and 35°C using an intravascular cooling catheter. Intermittent intravenous boluses of 5% hypertonic saline were used to maintain a serum sodium concentration of between 150 mEq/L and 155 mEq/L. Intermittent boluses of 0.5 g/kg of 20% mannitol were used to target an ICP of <25 mm Hg. Based on sensitivities, her antibiotic regimen was changed to cefotaxime for a total of 4 weeks.

Figure 2.

Screenshot from ICM+Brain Monitoring Software showing the collected physiologic parameters over approximately 20 min. This recording was taken early in the course of her invasive multimodal neuromonitoring. (A) is mean arterial pressure (MAP, yellow) and cerebral perfusion pressure (CPP, red). (B) is intracranial pressure (ICP), (C) is the pressure-reactivity index (PRx) and (D) is the brain tissue oxygen (P_btO₂). This figure demonstrates the complete loss of cerebral autoregulation as MAP is positively correlated with ICP, resulting in marked elevation of ICP when MAP was above 65 mm Hg. Figure created by DG.

Over time, there was marked reduction in cerebral oedema on CT Brain (figure 3A, B) and her ICP and autoregulation normalised (figure 4). Given these improvements in her cerebral physiology, we were able to normalise her temperature and sodium targets. We began to wean her intravenous sedation 11 days following insertion of her intraparenchymal neuromonitoring (14 days after admission) and she began to obey commands by day 17. The intraparenchymal neuromonitors were removed on day 18 of her admission and she was extubated on day 22, then discharged from the ICU on day 25. Following hospital discharge, she spent approximately 2 weeks at in-patient rehabilitation and was discharged home 52 days after her admission.

Figure 3.

Following treatment guided by invasive multimodal neuromonitoring, she had resolution of cerebral oedema (A, B).

Figure 4.

Screenshot from ICM+Brain Monitoring Software showing the collected physiologic parameters over approximately 24 min. This recording was taken near the end of the course of invasive multimodal neuromonitoring. (A) is mean arterial pressure (MAP, yellow) and cerebral perfusion pressure (CPP, red). (B) is intracranial pressure (ICP), (C) is the pressure-reactivity index (PRx) and (D) is the brain tissue oxygen (P_btO₂). This figure resolution of cerebral autoregulation as MAP and ICP are now inversely related as evident by a mostly negative PRx. Both ICP and P_btO₂ have normalised. Figure created by DG.

Outcome and follow-up

Her course was complicated by bilateral optic neuritis secondary to the N. meningitidis for which she received 5 days of intravenous methylprednisolone followed by prednisone. She subsequently made a full neurological recovery with respect to motor and sensory function, cognition as well as activities of independent and daily living. She is currently attending college.

Discussion

We present a case of meningococcal meningitis in a young woman who developed malignant cerebral oedema with transtentorial herniation as evident both radiographically and clinically with bilaterally fixed and dilated pupils. Fortunately, owing to aggressive neurocritical care resuscitation guided by invasive multimodal neuromonitoring, she made a remarkable recovery. There are several learnings that we can take from this case. First, access to specialist neurocritical care, guided by neurophysiologic monitoring of ICP and P_btO₂, can help improve outcomes among severely ill patients with bacterial meningitis. A historical cohort study in Sweden matched 52 patients in a neurocritical care unit to 53 patients who were managed in the conventional ICU.⁸ All patients in the neurocritical care group received ICP monitoring, 35 of 52 patients had episodes of significant intracranial hypertension (ICP>20 mm Hg for >5 min), and 48 of 52 patients received ICP-targeted treatment. Despite being well-matched at baseline, mortality in patients admitted to the neurocritical care unit was 10% compared with 30% admitted to the conventional ICU. However, care unit allocation in this study was not random, and, thus, the results are subjected to strong confounding by indication. Improved outcomes among neurocritically ill patients who received specialised care are also observed in other types of brain injuries.⁹ It must be stressed that it is not the monitoring itself that leads to improved outcomes, rather the access to expertise and improved processes of care seen in specialised care units. Another important observation was complete abolition of cerebral autoregulation observed in this case, a phenomenon that has been observed experimentally in bacterial meningitis.¹⁰ In our case, this pressure-passive physiology resulted in the ICP going above 20 mm Hg when the MAP went above 65 mm Hg. As such, using a traditional MAP threshold of >65 mm Hg for patients with septic shock would have resulted in marked intracranial hypertension and risk of further herniation. Thus, multimodal neuromonitoring allowed us to lower the threshold MAP, while maintaining adequate cerebral oxygen delivery as measured by P_btO₂. In addition, we were able to tolerate a lower P_btO₂ by employing strategies to decrease cerebral metabolic rate of oxygen utilisation, including: therapeutic hypothermia, sedation and use of barbiturates. As the patient improved, autoregulation was restored.

Learning points.

• Bacterial meningitis is a potentially catastrophic disease that carries a high mortality risk of neurologic injury even with prompt treatment. Clinicals must maintain an extremely high index of suspicion and initiate prompt treatment if this diagnosis is suspected.

• In appropriate patients with bacterial meningitis, access to neurocritical care, including the use of invasive multimodal neuromonitoring, may help clinicians titrate therapies, including mean arterial pressure, aimed to optimise the balance between cerebral oxygen delivery and metabolic rate of oxygen utilisation.

• Even in the face of transtentorial herniation, aggressive neurocritical care resuscitation may help salvage a potentially catastrophic outcome.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s)