FUSOBACTERIUM NECROPHORUM MENINGITIS ASSOCIATED WITH CEREBRAL VESSEL THROMBOSIS

Abstract

Anaerobic meningitis is rare in immunocompetent children unless there is a predisposing condition such as a neurenteric fistula.1 The exception to this rule is Fusobacterium necrophorum, a Gram-negative rod that is part of the normal upper respiratory tract flora. This organism can cause meningitis in previously healthy children with significant morbidity and neurologic sequelae.2 One of the major complications of infection with this organism is thrombosis.3 We report a case of an infant with fulminant meningitis in which thrombosis of cerebral vessels both clinically and pathologically played an important role in the brain injury that lead to his demise. Case report. The patient was a 9-month-old previously healthy Caucasian boy who was one of fraternal twins. After a 2-day history of fever, cough, congestion and irritability he became less responsive and had a episode of his eyes rolling up and staring off so he was taken to the local emergency room. His temperature was 105°F rectally, his neck was stiff and he was lethargic. A lumbar puncture revealed purulent cerebrospinal fluid with the following values: glucose, <10 mg/dl; protein, >600 mg/dl; white blood cell count, 36 540/cm3 with 79% polymorphonuclear leukocytes, 20% lymphocytes and 1% monocytes; a stained smear showing many polymorphonuclear leukocytes and Gram-negative rods. The cerebrospinal fluid culture grew a Gram-negative rod in the thioglycolate broth only which was subsequently identified biochemically using the RapANA system (Innovative Diagnostic Systems, Norcross, GA) as F. necrophorum. This identification was confirmed by the Centers for Disease Control and Prevention, Atlanta, GA, by gas-liquid chromatography. A head computerized tomography scan showed mild enlargement of the lateral ventricles but was otherwise unremarkable. He was treated intravenously with ampicillin and cefotaxime after which he was given 1.5 mg of dexamethasone. Shortly after his initial doses of antibiotics he had a brief seizure and received one iv dose of lorazepam, after which he was described as more awake, cooing and kicking in the crib. Within 3 h his neurologic status deteriorated dramatically, resulting in his transfer to our hospital. On arrival he was breathing on his own but had no spontaneous movements or response to voice. The anterior fontanel was concave with no suture diastasis. The pupils were unequal and unresponsive to light. Doll's eyes reflex was absent. There was no facial grimace or gag response. With stimulation the upper extremities assumed the decorticate position and the lower extremities withdrew. He had bilateral hyperreflexia, bilateral ankle clonus and bilateral Babinski signs. An electroencephalogram showed diffuse slowing for age without focal features or epileptiform activity. A head computerized tomography scan done 6 h after his acute neurologic deterioration showed areas of decreased density consistent with infarction in the posterior half of the right thalamus, the left caudate, internal capsule, lentiform nuclei and portions of the left thalamus. The lateral ventricles were slightly larger then they were on the scan done 12 h earlier. There were no areas of enhancement and cortical perfusion appeared normal. Because of his acute deterioration and with the ventricles being somewhat larger on the follow-up scan, a ventricular tap with pressure measurement was performed. A ventriculostomy was placed and ventricular pressure was measured in the 10- to 17-mm Hg range. Analysis of the ventricular cerebrospinal fluid showed: glucose, 70 mg/dl; protein, 147 mg/dl; white blood cell count, 161 cm3 with 76% polymorphonuclear leukocytes, 2% lymphocytes and 22% monocytes. Stained smear revealed Gram-negative rods but the culture was sterile. A trial of mannitol therapy had no benefit and metronidazole was added to the antibiotic regimen. The next morning he developed diabetes insipidus still without elevated ventricular pressure. That evening (now 48 h after presentation) the patient had two brief seizures and the ventricular pressure increased to 20 to 25 mm Hg associated with cardiovascular instability. During the next 48 h intracranial pressure increased to the 30- to 40 mm Hg range and the patient's neurologic examination became consistent with brain death. An electroencephalogram showed absence of electrocerebral activity and perfusion brain scan showed no cerebral perfusion. On postmortem examination there was extensive thick grayish fibropurulent exudate extending from the dorsal convexities of the brain to the cauda equina. This firm exudate filled the subarachnoid space, encasing the cranial nerves and blood vessels and the pituitary within the sella turcica. There was no evidence of communication between the sinuses, middle ear or oropharyngeal cavity and the intracranial vault or of a defect at the spinal cord level. Thrombi were noted in multiple vessels with ischemic infarcts noted including areas of the cortex, basal ganglia, thalamus, hypothalamus, pons, cerebellum, dorsal medulla and spinal cord. The brain and spinal cord were edematous but there was no evidence of herniation. Discussion. In children with anaerobic meningitis without brain abscess, two clinical entities can be distinguished.2 The first is the immunocompromised host where the most common organism is Bacteroides spp., especially Bacteroides fragilis. In contrast anaerobic meningitis in a previously healthy child, often after an upper respiratory infection or otitis media, is most frequently caused by F. necrophorum. These children can have a rapid fulminant course with a poor outcome.2 Historically human F. necrophorum infection has been characterized by a sore throat, sepsis and metastatic abscesses of the lungs or joints.4 This was first described by Lemierre and is now known as necrobacillosis.5 One of the well-known complications of necrobacillosis is septic thrombophlebitis of the internal jugular vein.3 There are also reports documenting thrombotic complications of intracranial vessels in F. necrophorum infections. These have been in the form of cortical vein, venous sinus and arterial thromboses.6-8 The early rapid neurologic deterioration in our patient suggested a primary vascular ischemic event followed later by cerebral edema and increased intracranial pressure as a result of the ischemia. Neuroimaging studies and autopsy findings support this interpretation. There are two cases in the literature that also report infarction as a complication of meningitis caused by this organism.9, 10 The first case was a 23-month-old child whose computerized tomography scan findings showed left parietal lobe, caudate nucleus and bifrontal infarctions.9 The second case was a 5-year-old girl whose autopsy showed extensive purulent leptomeningitis with infarctions in the left internal capsule, mesencephalon, pons and medulla,10 findings very similar to our case. F. necrophorum produces endotoxin and exotoxins such as leukocidin, hemolysin, lipase and cytoplasmic toxin.10 It also produces coagulase that encourages clot formation.7 These toxins may be the cause of the clinical complication of thrombosis that is associated with the organism.11 Treatment with dexamethasone as well as antibiotics did not prevent the vascular complications that resulted in our patient's death. Paul D. Larsen, M.D. Stephen A. Chartrand, M.D. Edward D. Adickes, D.O. Departments of Neurology (PDL, EDA), Pediatrics (PDL, SAC) and Pathology (EDA); Creighton University Medical Center; Omaha, NE