Escherichia coli is the most common Gram-negative bacillary organism that causes meningitis, and is also the most prevalent and best-studied infection in the neonatal population, with 80% of cases caused by the K1 serotype of E. coli, which expresses a capsular polysaccharide virulence factor [1,2]. In adults, E. coli meningitis accounts for just 3%–5% of all cases, but has a significant mortality rate of 36% [3-5]. The Biofire FilmArray Meningitis/Encephalitis panel uses polymerase chain reaction (PCR) amplification of the K1 gene to screen for E. coli [6].

Despite antibiotic treatment, cerebrovascular complications such as cerebral vasospasm have been reported to occur with infectious meningitis [7,8]. Advanced age, altered consciousness on admission, and seizures during hospitalization, have all been identified as independent risk factors for cerebrovascular complications [9]. To date, only two previous case reports have described cerebral vasospasm associated with E. coli meningitis: Klein et al. [8] reported transcranial Doppler (TCD) measurements indicative of cerebral vasospasm, while Gogia et al. [10] deduced cerebral vasospasm based on an infarct pattern. However, neither report specified whether the K1 virulence factor was present.

A 63-year-old man with Parkinson disease presented with new-onset seizures that did not return to baseline. He had no relevant neurosurgical history or immunocompromising conditions, but did have chronic unexplained thrombocytopenia. The patient was reported to be well in the morning; in the evening, his wife found him in bed able to sit up; however, he could not speak and looked unwell. The patient was transported to the hospital by ambulance. On arrival, he experienced multiple generalized tonic-clonic seizures that were suppressed by 6 mg of lorazepam and 3 g of levetiracetam. The only notable abnormality on assessment of initial vital signs was hypertension: blood pressure was 160/88 mm Hg, pulse was 96 beats per minute, temperature was 36 °C (96.8 °F), and respiratory rate was 18 breaths per minute, with an oxygen saturation of 99% on room air.

His postictal neurological examination revealed stupor, rightward gaze deviation, and weak withdrawal from noxious stimuli in all limbs. Nuchal rigidity was not observed. Labs on presentation were notable for a venous pH of 6.88, lactate >17 mEq/L, white blood cell count of 10,600 cells/μL, platelet count of 62 × 10⁹/L, and alanine transaminase/aspartate aminotransferase of 10⁶/221 IU/L, respectively. Contrast-enhanced computed tomography (CT) of his brain, chest, abdomen, and pelvis revealed no intracranial abnormalities, but revealed multiple non-obstructive renal calculi. Urinalysis revealed 3+ urine leukocyte esterase, with 25 white blood cells per high-power field, which were sent for culture. He was subsequently treated with ceftriaxone (1 g), and admitted with the suspicion that his seizures were caused by a neurodegenerative disorder in the setting of a urinary tract infection.

The next morning, his white blood cell count had doubled to 21,400 cells/μL. His neurological exam results also worsened; the patient was found to be comatose with a newly fixed-dilated left pupil. He was subsequently intubated for airway protection; no coughing or gag reflexes were noted during intubation. CT angiography of the brain and neck revealed no intracranial artery occlusion. An urgent magnetic resonance image (MRI) of the brain with contrast indicated meningoencephalitis with purulence around the brainstem (Fig. 1). The patient was subsequently administered empiric intravenous antibiotic therapy with ceftriaxone (2 g every 12 hours), vancomycin (15 mg/kg every 12 hours), and ampicillin (2 g every 6 hours). A lumbar puncture was performed, which revealed cloudy cerebrospinal fluid (CSF) with glucose concentrations below the limit of measurement, protein concentrations above the limit of measurement, and 4,050 nucleated cells with 94% neutrophilic predominance. Gram-negative rods were observed during initial CSF examination. The treatment was modified to monotherapy with meropenem 2 g every 8 hours, to enhance the coverage of anaerobic Gram-negative pathogens. Continuous electroencephalographic monitoring revealed diffuse continuous delta range slowing, consistent with severe encephalopathy.

The Biofire FilmArray Meningitis/Encephalitis panel was performed on the CSF, with no organism detected, while CSF cultures did not reveal any growth. However, because the initial CSF examination revealed Gram-negative rods, broad-range PCR for the 16S ribosomal ribonucleic acid (rRNA) gene and next-generation sequencing were performed, indicating the presence of Enterobacteriaceae, consistent with the presence of either E. coli or Shigella. His urine culture showed pan-sensitive E. coli, correlating with the CSF analysis results, supporting the diagnosis of E. coli without the K1 virulence factor as the cause of meningitis, with likely hematogenous spread from a urinary source.

By hospitalization day 4, the patient regained his cough and gag reflexes, and was noted to have withdrawn from the pain in his upper extremities. However, on hospitalization day 8, he developed tachypnea, tachycardia, and decreased responsiveness to noxious stimuli. Brain CT revealed bilateral evolving infarcts in a pattern typical of diffuse vasospasm. Concurrent CT of his neck with contrast, which was ordered to screen for other sources of infection, incidentally demonstrated a new vasospasm in the major cerebral arteries (Fig. 2). Shortly thereafter, the patient’s neurological status worsened with the development of malignant cerebral edema and herniation. On hospitalization day 10, the patient was declared brain dead.

This case presents a challenge for neurological localization and diagnosis. Although the patients seizures without fever or nuchal rigidity were initially attributed to the progression of his neurodegenerative disease, his rapid progression to coma with impaired gag reflex on intubation and unilateral fixed-dilated pupil indicated an acute brainstem injury, which was further supported by the purulence observed in his cisterna magna on MRI. This case demonstrates how a CSF pathogen can be identified despite early antibiotics rendering the pathogen non-viable, and despite the pathogen being absent from a screening PCR panel. When 16S rRNA gene sequencing becomes available, this technique could provide further reassurance to clinicians that antibiotic administration should not be delayed before obtaining a CSF sample.

This case further highlights the need to consider utilizing neuromonitoring tools such as serial TCD examinations in high-risk patients with meningitis, particularly when it is challenging to monitor neurological examinations due to a poor neurological baseline or the presence of confounders, such as sedative medications. We did not use TCD as it is not yet part of the standard of care for the management of meningitis; however, research has demonstrated it to be a reliable modality for diagnosing vasculopathy associated with bacterial meningitis [11]. A prospective multicenter cohort study with 506 patients identified altered mental status (defined as a Glasgow Coma Score <14) on presentation and seizures within 48 hours of presentation as independent risk factors for cerebrovascular complications, with odds ratios of 2.23 (95% CI, 1.21–4.10) and 1.90 (95% CI, 1.01–3.52) respectively [9].

The management of post-infectious vasospasm is challenging, given that there is currently no well-established approach or a dearth of studies on this topic. Treatments that have been reported with varying degrees of success are derived from treatment protocols for aneurysmal subarachnoid hemorrhage-related vasospasms. These include induced hypertension, intraarterial verapamil, milrinone, and balloon angioplasty [12]. In our case, experimental treatment attempts for the cerebral vasospasm were not made because of the severe irreversible neurological injury that was already present.

Our case is the first report of arterial vasospasm related to non-K1 E. coli meningitis, and the first to present with CT imaging evidence of vasospasm following meningitis in all strains of E. coli. One weakness of this report is that his arterial vasospasm was incidentally detected on a contrast-enhanced CT scan of the neck, which confirmed his infarct pattern. A dedicated CT angiogram for the occurrence of vasospasm would provide more definitive evidence. The patient’s widow shared that her faith was important as she coped with the tragedy, and she expressed gratitude to the medical team that had cared for him, stating: “It is late for our family but I hope and pray that no other families go through what we went through.”