GFAP Astrocytopathy Can Look Like Severe Infectious Encephalitis
Two women arrived with the kind of inflammatory central nervous system syndrome that usually drives clinicians toward infection first: fever, headache, meningism, delirium, respiratory failure, and cerebrospinal fluid pleocytosis with elevated protein. In both cases, the microbiological work-up stayed negative while the inflammatory pattern persisted. The diagnosis that finally emerged was GFAP astrocytopathy, a rare autoimmune disorder in which astrocytic injury can closely resemble infectious meningoencephalitis at the bedside and in the laboratory.
The clinical trap is obvious. When a patient has abnormal CSF, systemic inflammation, and rapidly worsening neurologic status, it is rational to start antimicrobials immediately. Yet these cases show how a parallel autoimmune work-up can become decisive once repeated testing fails to identify a pathogen. GFAP-IgG detection, especially in CSF, provided the turning point in both women.
Persistent Aseptic Meningitis in a 36-Year-Old Woman
The first patient, a 36-year-old woman with no notable past history, presented with 5 days of low-grade fever and severe stabbing headache. Neurologic examination was largely normal except for marked meningeal signs. CT of the head was unrevealing, but CSF showed a strongly inflammatory profile: 183 cells/µL with 97.8% lymphocytes and protein of 83.3 mg/dL. PCR testing did not identify an infectious cause, and serologies for tick-borne infections, including Borrelia burgdorferi, were negative.
She was treated initially as having lymphocytic meningitis syndrome. Intravenous dexamethasone was given for 9 days, and repeat lumbar puncture showed partial improvement in CSF inflammation. After discharge, symptoms did not settle. Headache and photophobia continued, prompting two emergency visits, and she was readmitted within 3 days because of severe orthostatic headache, tremor of the upper limbs, dizziness, neck stiffness, and a positive Kernig sign. Brain MRI remained normal. Spinal MRI showed disc herniation at L3–L4 and L4–L5 with possible nerve root compression, but the CSF still pointed toward ongoing inflammatory disease: 178 cells/µL with 97.7% lymphocytes.
The autoimmune panel changed the diagnostic frame. Neuronal surface antibodies were negative, including NMDAR, AMPAR1/2, DPPX, LGI1, CASPR2, and GABABR. GFAP-IgG was detected in both CSF and serum by indirect immunofluorescence. The extensive infectious evaluation remained negative, including HIV, Borrelia, tick-borne encephalitis, Treponema pallidum, repeated multiplex CSF PCR, broad real-time PCR testing, and other targeted studies.
A short course of acyclovir was given because of lip lesions suspicious for herpes simplex infection, and doxycycline was also administered. Physiotherapy was introduced when hamstring contractures were thought to contribute to the apparent meningeal irritation. The bedside detail mattered: what initially looked like stubborn meningeal inflammation was partly mechanical limitation from muscle contracture. After 20 days, she improved enough for discharge, then started oral prednisone at 60 mg/day followed by 50 mg/day. A telephone follow-up about a month later showed partial clinical improvement, although upper-limb tremor and generalized weakness persisted.
Residual neurologic symptoms after treatment fit with a disease that can improve yet not fully disappear at first pass. Autoimmune inflammation may recede slowly, especially when recognition comes after repeated episodes and prolonged diagnostic uncertainty.
Fulminant Encephalopathy in the ICU in a 57-Year-Old Woman
The second patient, a 57-year-old woman with migraine with aura and depression, arrived in far more dramatic fashion. She became confused, agitated, delirious, and increasingly disorganized in speech and behavior, with her husband describing altered voice and abnormal movements. Stroke was suspected initially, but CT angiography and non-contrast head CT were normal. CRP was elevated at 34.5 mg/L, and lumbar puncture showed 99 cells/µL with 88% lymphocytes and protein of 139 mg/dL.
Her neurologic status worsened quickly. She required sedation, physical restraint, and empiric coverage with acyclovir, ceftriaxone, ampicillin, and vancomycin. Despite treatment, she developed acute respiratory failure, required intensive care and mechanical ventilation, and stayed in the ICU from 2 January to 21 January 2026. Brain MRI on day 6 showed slight segmental meningeal thickening with moderate contrast enhancement, but no other major abnormality. CRP rose further to 167 mg/L, which led to escalation to meropenem and metronidazole while acyclovir continued.
A repeat CSF sample on day 14 showed partial laboratory improvement, with pleocytosis falling to 81 cells/µL and protein to 60 mg/dL. Comprehensive microbiological and pathomorphological investigations excluded bacterial and viral CNS infection. Autoimmune testing was then pursued, including the same neuronal surface antibody panel used in the first case.
The clinical course gradually turned. Mechanical ventilation was stopped, oxygen support continued through the tracheostomy tube, and by day 19 she was conscious and able to follow commands, although generalized weakness remained. The record provided here stops before the final antibody result and longer-term neurologic outcome, but the diagnostic logic had already shifted: a syndrome that looked infectious from the first hour was behaving like autoimmune meningoencephalitis.
Why These Two Cases Matter at the Bedside
GFAP astrocytopathy sits in a dangerous diagnostic overlap zone. The syndrome can present with meningitis, encephalitis, or meningoencephalomyelitis, and CSF often shows the same lymphocytic pleocytosis and elevated protein that prompt infectious treatment. When MRI is normal or only mildly abnormal, the autoimmune diagnosis becomes easier to miss, especially if the clinical picture is noisy or the patient is critically ill.
The biology helps explain the mimicry. GFAP is an astrocytic intracellular protein, and the inflammatory process appears to involve cellular immune mechanisms rather than a classic surface-antibody syndrome alone. Astrocytic dysfunction disrupts CNS homeostasis, and the result can be headache, fever, meningeal signs, tremor, encephalopathy, and focal neurologic change. In practice, that means the pattern can look like viral encephalitis, bacterial meningitis, or mixed meningoencephalitis long before the autoimmune signal is considered.
Both women improved after corticosteroids, which fits the broader treatment pattern reported for this disorder. The first patient still had persistent tremor and weakness at follow-up, while the second improved after prolonged ICU-level support. Each case also demonstrates how much antibiotic and antiviral exposure can accumulate before the diagnosis is secured. In a patient with negative infectious studies and persistent inflammatory CSF, GFAP-IgG testing deserves a place early in the differential rather than at the end of it.
The final lesson is not subtle: infectious exclusion alone is not enough when the CSF stays inflammatory and the neurologic syndrome keeps evolving. The remaining uncertainty is practical and important — which presentation patterns should trigger autoimmune testing immediately, especially when neuroimaging is unrevealing and the patient is already receiving empiric antimicrobials.