A male infant weighing 768 g was born at 25 weeks gestation by spontaneous vaginal delivery after prolonged premature rupture of membranes. He was treated with ampicillin and gentamicin for suspected sepsis. Compete blood count and C-reactive protein were normal, and blood and urine cultures were negative. A screening head ultrasound on day of life (DOL) 1 showed bilateral periventricular leukomalacia (PVL) with a unilateral grade IV intraventricular hemorrhage. He had an unremarkable clinical course until DOL 7 when he developed increasing abdominal distension and respiratory deterioration necessitating intubation. A second sepsis evaluation included a lumbar puncture (LP), which showed cerebral spinal fluid (CSF) with protein of 1350 mg/dL, glucose of 5 mg/dL, 1940 red blood cells (RBC)/mm3, and 263 white blood cells (WBC)/mm3 (differential of 89% neutrophils, 7% lymphocytes, 4% histiocytes). Gram stain of the CSF revealed many Gram-negative rods. Vancomycin and piperacillin/tazobactam were added, gentamicin was continued, and ampicillin was discontinued. CSF, blood, and urine cultures were negative. On DOL 7 the infant developed near continuous seizures, consisting of rhythmic jerking movements of the extremities. Head ultrasound revealed worsening ventricular dilatation and PVL. On DOL 10 the seizures subsided, and repeat LP showed CSF protein of 1116 mg/dL, glucose of 5 mg/dL, 14,400 RBC/mm3, and 251 WBC/mm3 (differential of 71% neutrophils, 21% lymphocytes, 1% histiocytes, and 7% other). Gram stain was negative, and cultures of CSF, blood, and urine remained negative.
On DOL 11, the Pediatric Infectious Disease Service recommended starting cefotaxime at a meningitic dose, continuing gentamicin, and discontinuing piperacillin/tazobactam and vancomycin. A video EEG on DOL 18 showed frequent, high amplitude, irregular spike wave bursts thought to represent “aborted” ictal discharges. On DOL 19, the infant had recurrent episodes of apnea and bradycardia, requiring an increase in ventilator support. A third CSF evaluation on DOL 20 showed rising protein of 1881 mg/dL, improved glucose of 65 mg/dL, RBC count of 31/mm3, and lower WBC count of 111/mm3. CSF Gram stain was negative. Vancomycin was added empirically. A head computed-tomography scan (CT) on DOL 22 revealed increased uptake in the ependyma and a possible periventricular cystic abscess. Piperacillin/tazobactam was restarted empirically.
Bacterial ventriculitis with noncommunicating hydrocephalus was suspected. A diagnostic test was performed.
CSF obtained on DOL 23 from a transfontanelle ventricular tap was cultured and had growth of Bacteroides fragilis after 48 hours. This was the first CSF specimen sent for anaerobic culture. The infant was diagnosed with anaerobic meningitis. The isolate was susceptible to metronidazole, piperacillin/tazobactam, meropenem, but resistant to penicillin with an MIC >256 mg/dL. Intravenous metronidazole (7.5 mg/kg IV daily) was started on DOL 26 and all other antibiotics were stopped. The infant received 28 days of metronidazole and remained clinically stable, although seizure activity on EEG prompted an increase in phenobarbital dose on DOL 29. Repeat CSF culture on DOL 31 was sterile. A subsequent head CT on DOL 41 showed stable massive hydrocephalus, which was noted on DOL 11 (Fig. 1), and evolution of bilateral germinal matrix hemorrhages with calcifications. He was discharged on DOL 79. At 4 months of age, the infant underwent placement of a ventriculoperitoneal shunt. At his most recent follow-up evaluation at age 11 months, he had right occipital plagiocephaly, increased lower extremity tone, and global developmental delay with gross motor skills at the 3 to 4 month level and fine motor skills at 5 to 6 months.
Anaerobic meningitis is rare. Only 20 pediatric cases with isolated anaerobic meningitis (versus with abscess) were described in a 1986 review spanning 40 years. Eight were due to B. fragilis.1 It appears that B. fragilis meningitis is primarily a pediatric disease, particularly affecting young infants. Of the 14 cases of B. fragilis meningitis reported to date, 10 have occurred in infants, 7 of whom were neonates.2 The bacteriology in isolated anaerobic meningitis in newborns is usually monobacterial, with B. fragilis being the only bacterium able to induce abscess formation as the sole infecting organism. Abscess formation has been linked to its complex polysaccharide capsule with a zwitterionic motif which provides docking sites for immunomodulatory molecules.3 Interestingly, abscess formation induced by B. fragilis is dependent on T-cell function, which is impaired in the neonate. This might explain the slow and minor abscess formation in our case even after 3 weeks.
Predisposing factors in isolated anaerobic meningitis in newborns are chorioamnionitis in the mother, gastrointestinal disease or gastrointestinal surgery, and congenital malformations (spinal-colonic fistula, anal stenosis, inguinal hernia, or myelomeningocele).1,2,4–10 The clinical presentation is indistinguishable from other causes of neonatal meningitis.
We speculate that our patient developed transient anaerobic bacteremia due to maternal chorioamnionitis, which resulted in preferential infection of the patient's intraventricular bleed. The propensity of B. fragilis to grow in blood clots has been observed in adult patients with portal vein thrombosis. Subclinical bowel perforation leading to dissemination is less likely.
Although safety and efficacy data in children are limited,10 metronidazole is the treatment of choice for Gram-negative anaerobic meningitis, due to its excellent penetration into the CNS and abscesses. This agent has been used with great success since the 1970s, with the parenteral form used in children since the late 1980s. Since Bacteroides sp. have the most antibiotic resistance mechanisms of all anaerobic pathogens and the highest resistance rates, some authors have suggested adding a beta lactam/ beta-lactamase inhibitor or a carbapenem. It should be noted, however, that piperacillin/tazobactam is not FDA-approved for the treatment of meningitis in the United States.11 There is some evidence that the usual dosage of tazobactam may not achieve adequate drug concentrations in CSF. This could explain why use of piperacillin/tazobactam led to treatment failure in our patient despite in vitro susceptibility. In rabbits with experimental Klebsiella meningitis receiving piperacillin/tazobactam, CSF concentrations of both compounds were less than 5 times the minimum bactericidal concentration with suboptimal bactericidal effect.12 Inadequate tazobactam concentrations in the CSF have also been described in hydrocephalic adults.13 Other reasons for treatment failure in our case might have been inadequate drug transport into the intraventricular clot where the infection likely began.
A minimum of 14 days of treatment is usually recommended, but the optimum treatment duration for B. fragilis meningitis is unknown. The role of surgery, aspiration, excision, and intraventricular antibiotic instillation is unclear. The outcome of patients with isolated anaerobic meningitis has historically been guarded with full recovery in only 30%, survival with sequelae in 45%, and mortality in 25%.1
A high index of suspicion is needed to diagnose anaerobic meningitis. In this case, the presence of Gram-negative rods in the Gram stain of the initial CSF with a negative CSF culture should have alerted the clinicians to the possibility of an anaerobic infection. In the absence of a positive CSF Gram stain, an anaerobic culture of the CSF should also be considered if a neonate with meningitis is not responding to standard therapy and aerobic cultures of the CSF, are negative.
1.Tärnvik A. Anaerobic meningitis
in children. Eur J Clin Microbiol
2.Cooke RW. Bacteroides fragilis
septicaemia and meningitis
in early infancy. Arch Dis Child
3.Dysart NK, Griswold WR, Schanberger JE, et al. Letter: meningitis
due to Bacteriodes fragilis
in a newborn infant. J Pediatr
4.Feldman WE. Bacteroides fragilis
ventriculitis and meningitis
. Report of two cases. Am J Dis Child
5.Berman BW, King FH Jr, Rubenstein DS, et al. Bacteroides fragilis meningitis
in a neonate successfully treated with metronidazole. J Pediatr
6.Law BJ, Marks MI. Excellent outcome of Bacteroides meningitis
in a newborn treated with metronidazole. Pediatrics
7.Feder HM Jr. Bacteroides fragilis meningitis
. Rev Infect Dis
8.Webber SA, Tuohy P. Bacteroides fragilis meningitis
in a premature
infant successfully treated with metronidazole. Pediatr Infect Dis J
9.Brook I. Meningitis
and shunt infection caused by anaerobic bacteria in children. Pediatr Neurol
10.American Academy of Pediatrics. In: Pickering LK, ed. 2006 Red Book: Report of the Committee on Infectious Diseases (Antimicrobial Agents, Chap 4:752).
27th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2006.
11.Leleu G, Kitzis MD, Vallois JM, et al. Different ratios of the piperacillin-tazobactam combination for treatment of experimental meningitis
due to Klebsiella pneumoniae producing the TEM-3 extended-spectrum beta-lactamase. Antimicrob Agents Chemother
12.Nau R, Kinzig-Schippers M, Sörgel F, et al. Kinetics of piperacillin and tazobactam in ventricular cerebrospinal fluid of hydrocephalic patients. Antimicrob Agents Chemother
13.Wexler HM. Bacteroides: the good, the bad, and the nitty-gritty. Clin Microbiol Rev