Normal Spinal Fluid Results Hid a Reversible Brain Injury in a Comatose Sepsis Patient

A woman in her 50s arrived at an emergency department in profound coma, her body fighting a septic storm that no routine test could initially explain. Her temperature was 40.5 degrees Celsius, her heart rate raced at 127 beats per minute, and her blood pressure had fallen to 86/53 mm Hg despite oxygen delivered through a high-flow reservoir mask. On the Glasgow Coma Scale she scored E1V1M1, meaning she did not open her eyes, utter words, or move in response to stimulation. Her pupils were 3 millimeters on both sides with brisk light reflexes, but the depth of coma prevented any reliable assessment of focal or lateralizing motor deficits, and detailed cranial-nerve testing was impossible during active resuscitation. Notably, she had no stiff neck, no Kernig or Brudzinski signs, and no rash, findings that might otherwise have pointed immediately toward meningitis.

Laboratory studies painted a picture of overwhelming systemic inflammation paired with failing clotting and kidneys. C-reactive protein measured 8.85 mg/dL against a reference ceiling of 0.20, prothrombin time and international normalized ratio stood at 1.25, and fibrin and fibrinogen degradation products reached 26.58 µg/mL, far above the 5.00 threshold. Her platelet count had dropped to 13.5 × 10⁴/µL, blood urea nitrogen was mildly elevated at 23.3 mg/dL, and creatinine reached 1.44 mg/dL. Clinicians diagnosed septic shock complicated by disseminated intravascular coagulation and acute kidney injury. Contrast-enhanced computed tomography, including a head scan that ruled out any intracranial mass lesion, failed to reveal a source of infection. Because the patient carried a levonorgestrel-releasing intrauterine device, physicians performed transvaginal ultrasonography to search for a genital-tract origin and found no abscess; urinalysis was likewise unremarkable. To chase down the unexplained fever, they performed a lumbar puncture and simultaneously began fluid resuscitation, norepinephrine, vasopressin, meropenem, vancomycin, levofloxacin, and intravenous hydrocortisone. On day 2, blood cultures returned the answer: Neisseria meningitidis serogroup B, the bacterium behind meningococcal septic shock.

The cerebrospinal fluid drawn that first day looked deceptively clean. It was clear, contained only 1 leukocyte per microliter, carried just 26 mg/dL of protein, and showed a glucose level of 115 mg/dL against a paired serum glucose of 172 mg/dL. Gram stain and culture were negative, and a multiplex polymerase chain reaction panel designed to detect 14 meningitis and encephalitis pathogens came back negative across the board. By conventional standards, these results would end the neurological workup and support a diagnosis of sepsis-associated encephalopathy, a condition that occurs in up to 70 percent of patients with severe sepsis and is recognized only after primary central nervous system infection, metabolic causes, and drug effects have been excluded. Yet the treating team understood that routine cerebrospinal fluid can be falsely reassuring. Early meningoencephalitis may precede pleocytosis, fastidious or partially treated organisms can evade culture, viral loads may fall below PCR detection thresholds, and delayed processing can reduce diagnostic yield. When they measured interleukin-6 in the same fluid sample, the concentration was 6190 pg/mL, a level the authors describe as comparable to values reported in acute bacterial meningitis. A conventional noninflammatory profile, in other words, had not excluded inflammatory signaling inside the central nervous system.

Serial biomarker tracking sharpened the picture. Serum interleukin-6 was measured over 7 days, while cerebrospinal fluid interleukin-6, neuron-specific enolase, and myelin basic protein were measured on days 1 and 10. On admission, serum interleukin-6 reached a staggering 150,000 pg/mL. The CSF-to-serum albumin quotient, a standard marker of blood-brain barrier integrity, was 3.31 × 10⁻³ on day 1, comfortably within the age-adjusted normal range, whose upper limit sits near 7.5 × 10⁻³ by the Reiber formula. In meningococcal shock, lipooligosaccharide and outer-membrane vesicles released by the bacteria can drive such a cytokine surge, but experimental work shows that only a very small fraction of circulating interleukin-6 crosses the blood-brain barrier, so passive leakage alone cannot explain the cerebrospinal fluid level observed. Peripheral inflammation can instead signal to the brain through saturable transport across the barrier, through the circumventricular organs, and along vagal afferent pathways. These routes can activate microglia, the resident immune cells of the brain, which then produce interleukin-6 locally within the central nervous system. The combination of extreme serum elevation, markedly elevated cerebrospinal fluid interleukin-6, and a preserved albumin quotient therefore pointed most consistently toward active inflammatory signaling and possible intrathecal cytokine production rather than simple barrier leakage. The authors interpret this cautiously: the albumin quotient was normal only at symptom onset and had risen to 9.44 × 10⁻³ by day 10, so barrier integrity at the time of the later imaging abnormality remains unknown, particularly because it was not measured when the day 7 MRI change appeared. By day 10, cerebrospinal fluid interleukin-6 had normalized to 7.1 pg/mL, and serum interleukin-6 had collapsed from 150,000 pg/mL on day 1 to 57.8 pg/mL by day 2, remaining between 20 and 90 pg/mL through day 7.

Despite withdrawal of sedation, altered consciousness persisted. Brainstem reflexes stayed intact and pupils remained reactive throughout, while the motor component of the Glasgow Coma Scale improved only gradually. On day 7, brain magnetic resonance imaging revealed bilateral symmetric high signal in the thalami on diffusion-weighted imaging, with corresponding low signal on the apparent diffusion coefficient map, confirming true restricted diffusion. No other site of restricted diffusion, edema, or mass effect was present. Magnetic resonance angiography showed a patent basilar artery and patent bilateral posterior cerebral arteries, and no earlier MRI existed to date the lesion’s onset.

Bilateral symmetric thalamic restricted diffusion carries a broad differential diagnosis. Bilateral paramedian thalamic infarction from top-of-the-basilar syndrome or occlusion of the artery of Percheron was a key consideration, but the major arteries were patent and the lesions resolved completely by day 33, arguing against infarction. Deep cerebral venous thrombosis was judged unlikely because there was no hemorrhage, no vasogenic edema, and no visible venous thrombus, and complete reversibility would be atypical for venous infarction. Posterior reversible encephalopathy syndrome seemed improbable without hypertensive surges or posterior cortical-subcortical lesions, and carbon monoxide poisoning was excluded by an admission carboxyhemoglobin of 0.8 percent. Acute necrotizing encephalopathy was also deemed unlikely given the isolated thalamic involvement, the absence of hemorrhage, normal cerebrospinal fluid neuron-specific enolase and myelin basic protein, and complete radiographic resolution. The team instead favored a reversible inflammatory-metabolic process: normal structural-injury biomarkers together with complete MRI resolution pointed away from infarction, demyelination, or necrosis and toward transient dysfunction in a thalamus rendered vulnerable by high metabolic demand and systemic inflammatory stress.

That interpretation shaped a deliberately staged approach to prognosis. The clinicians deferred early prognostication on three grounds: the major arteries were patent, the cerebrospinal fluid structural-injury biomarkers were within normal limits, and serum interleukin-6 was falling rapidly, consistent with swift attenuation of systemic inflammatory activity. Early tracheostomy was withheld because the bilateral thalamic lesions raised the possibility of a poor neurological prognosis and the patient remained dialysis dependent without a clear route to recovery. When consciousness recovered on day 20, she was extubated the next day and tracheostomy proved unnecessary. Follow-up MRI on day 33 showed complete resolution of the thalamic restricted diffusion. Continuous electroencephalography with a headset device showed no epileptiform activity, though the authors acknowledge a limitation: because the recordings were not retained for independent review, background continuity and reactivity could not be reliably reassessed, and a documented exclusion of nonconvulsive status epilepticus was lacking. They also note that continuous renal replacement therapy for acute kidney injury ran from days 4 to 12, overlapping with the later biomarker normalization, so extracorporeal cytokine clearance cannot be excluded as a contributor, even though the steepest serum interleukin-6 decline on days 1 and 2 preceded renal replacement.

Recovery of arousal did not translate into full functional recovery. At discharge, severe flaccid tetraparesis had become apparent. Nerve conduction studies showed motor-predominant axonal neuropathy, and the Medical Research Council sum score was 24 out of 60, meeting criteria for severe intensive-care-unit-acquired weakness. The authors attribute this weakness to multiple concurrent factors: sepsis, prolonged immobilization, hyperglycemia, and vasopressors were all present, while corticosteroid exposure was limited to hydrocortisone 200 mg/day for 6 days and neuromuscular blockade consisted of a single intubating dose. They regard the central and peripheral injury as parallel consequences of severe systemic inflammation rather than a single proven pathway. Two years later, the patient could walk to the toilet with a walker and perform simple manual tasks, indicating recovered arousal but persistent peripheral disability.

The case yields several lessons for neurocritical care. Severe sepsis with disproportionate encephalopathy is a neurological problem, not only a hemodynamic and infectious one, and reversible but clinically important brain injury can hide behind normal routine studies unless the nervous system is actively evaluated from the start. A normal routine cerebrospinal fluid profile should not prematurely terminate the neurological evaluation of patients with severe sepsis and unexplained coma; when suspicion is high, cerebrospinal fluid cytokines such as interleukin-6 may provide complementary evidence of inflammatory activity when cell count, protein, culture, and multiplex PCR are unrevealing. Bilateral thalamic restricted diffusion in septic shock is not invariably irreversible, and with patent major vessels and normal structural-injury biomarkers, serial imaging may disclose a reversible process, so prognostication should be multimodal, combining vascular imaging, biomarkers, and electroencephalography rather than fixed at the first scan. Finally, as arousal returns, intensive-care-unit-acquired weakness may dominate functional recovery, and early bedside strength assessment with proactive rehabilitation targets the injury that routine neuroimaging does not show.

Subject of Research: Sepsis-associated encephalopathy diagnosed through cerebrospinal fluid cytokine measurement and reversible bilateral thalamic MRI changes in meningococcal septic shock

Article Title: Looking Beyond a Normal Cerebrospinal Fluid Profile: Lessons Learned About Sepsis-Associated Encephalopathy in Meningococcal Septic Shock

Article References: Tomita, Y., Suzuki, K., Inoue, G., Yagi, M., & Dohi, K. (2026). Looking Beyond a Normal Cerebrospinal Fluid Profile: Lessons Learned About Sepsis-Associated Encephalopathy in Meningococcal Septic Shock. Neurocritical Care. https://doi.org/10.1007/s12028-026-02708-8

Image Credits: AI Generated

DOI: 10.1007/s12028-026-02708-8

Keywords: sepsis-associated encephalopathy, meningococcal septic shock, cerebrospinal fluid, interleukin-6, blood-brain barrier, bilateral thalamic lesions, restricted diffusion, neuron-specific enolase, myelin basic protein, prognostication, ICU-acquired weakness, Neurocritical Care