Chapter 16. Meningitis

Meningitis is defined as an inflammation of the leptomeninges of any cause. Bacteria, which cause meningitis by invading and replicating in the subarachnoid space, are associated with significant morbidity and mortality. Viral infections may also cause meningitis, most commonly enteroviruses, but few children with viral meningitis suffer any long-term sequelae. Therefore, the focus of this chapter will be on bacterial meningitis. Figure 16–1 displays the age and organism-specific rates of bacterial meningitis in the United States prior to the introduction of currently used conjugate vaccines (note the y-axis is a log scale). As can be seen, the greatest risk period for bacterial meningitis is in the first 6 months of life.

Overall, there has been a remarkable decline in the rate of bacterial meningitis in the developed world over the last two decades with the introduction of the Haemophilus influenzae type B conjugate vaccines, the Streptococcus pneumoniae conjugate vaccines and greater use of meningococcal vaccines. H. influenzae type B was once the leading cause of bacterial meningitis in children but has been virtually eliminated in countries utilizing the conjugate vaccine. In the first 2 months of life Enterobacteriaceae (e.g., Escherichia coli, Klebsiella spp.), group B streptococci, and occasionally Listeria monocytogenes, Salmonella spp. or enterococci will cause bacterial meningitis. Infections caused by S. pneumoniae occur with increasing in frequency over the second month to become the most likely cause of bacterial meningitis and continue to increase in frequency until 4 or 5 months of age when they begin to decline. Neisseria meningitidis is the most common cause of bacterial meningitis by 1 year of age. S. pneumoniae remains the second most common cause after 1 year of age and all other pathogens trail behind considerably. These two pathogens occur more commonly in the winter months presumably in association with common respiratory viruses, which disrupt mucosal barriers thereby allowing these colonizing pathogens to move from the nasopharynx to the bloodstream more easily. Research is ongoing to develop vaccines that will be effective against a greater number of pneumococcal serotypes and improved meningococcal vaccines that may work for younger children and against group B strains. Table 16–1 reviews microbial causes of meningitis.

Bacteria most often gain access to the central nervous system via the bloodstream. Pathogens gain access to the bloodstream as a result of nasopharyngeal colonization or local infections. When these specific bacterial strains are relatively new to the host, and the child has no existing circulating antibody against that strain, the bacteria may at least temporarily evade other host defenses and cause a transient or sustained bacteremia. Some of these bacteria may traverse the blood–brain barrier and replicate in the subarachnoid space. Once bacteria begin to replicate within the central nervous system, the human host has no adequate mechanism to recover without medical intervention.

The risk for sequelae and mortality vary significantly by age of the child, pathogen, and underlying host compromise at baseline and at presentation with meningitis. Patients may have impaired neurologic function as a result of brain inflammation secondary to the bacteria and the host response to infection as well as because of cerebral hypoperfusion related to septic shock with hypotension, raised intracranial pressure, and/or disorders in local circulatory regulation including microvascular obstruction. Brain injury centers prominently on the cortex, hippocampus, and the inner ear, and the injuries are not simply the result of bacterial invasion but also occur as a result of a symphony of inflammatory mediators whose roles and control remain the focus of intense evaluation and give the promise of potential for future clinical intervention.1

Meningitis may also develop as the result of direct extension of pathogens from sites, such as the sinuses, mastoids, dermoid sinuses, or the skull. Recurrent episodes of bacterial meningitis should prompt immunologic evaluation as well as careful anatomic evaluation of contiguous structures (e.g., for congenital or traumatic bony, vascular, or dural defects).

In practice, patients present in one of two manners. The first is easier to recognize quickly but carries a worse prognosis. Approximately 25% of patients have a short clinical course of only a few hours and present with fever, mental status changes (irritability, lethargy, confusion) and the diagnosis is not elusive. The other 75% of patients with bacterial meningitis have a more insidious clinical presentation with simple fever and nonspecific symptoms that progress over 1, 2, or more days to cause signs and symptoms typical of meningitis.

The most reliably present finding in bacterial meningitis is fever (98%) whereas complaints of headache, photophobia, or neck pain will not be heard until the child is verbal and can communicate these symptoms (around 2 years of age). Since bacterial meningitis is most common before this age, it is important to watch for other clues. Infants may initially present with irritability that is difficult for the parents to alleviate, poor feeding, sleepiness or lethargy, and possibly vomiting. Later in the illness a bulging fontanelle, nuchal rigidity, and coma may be seen.

For an older child, where neck pain can be more easily assessed, the neck pain seen with meningitis is worsened by neck flexion. The typical child can easily touch the chin to the chest without the need to open the mouth and reluctance to do so is worrisome. Patients with meningismus will often open the mouth in an attempt to reach the chest with less neck flexion. Flexion typically causes pain at the back of the neck and moving down to the upper mid-back area. It should be noted that ibuprofen and other analgesics may temporarily resolve the neck pain despite the presence of meningitis.

Seizures occur as part of the presentation in 15–30% of patients but will only rarely occur as simple (brief, nonfocal) seizures without other signs or symptoms worrisome for meningitis.2 On the other hand, the risk of meningitis is increased significantly among febrile patients with convulsive status epilepticus (a seizure or series of seizures without recovery of consciousness between seizures lasting at least 30 minutes).3 Therefore, if faced with a febrile patient with focal, prolonged, or recurrent seizures, or a febrile infant younger than 6 months with any type of seizure, a lumbar puncture for cerebrospinal fluid (CSF) examination should be strongly considered.

Bacterial meningitis should be considered a medical emergency and the usual attention to airway, breathing, and circulation should be applied as depressed mental status and associated septic shock may complicate management. The rapid institution of supportive measures and prompt administration of antibimicrobials is the goal.

Severe pharyngitis, parapharyngeal abscess or adenitis, musculoskeletal strain, upper lobe pneumonia, cervical spine osteomyelitis, epidural abscess, and subarachnoid hemorrhage may all produce meningismus. Depressed mental status may occur as a result of encephalitis, ingestions, mass lesions of the CNS, or simple febrile seizure.

The possibility of another primary focus for infection, which might also require specific therapy or follow-up, should be considered. This site may be distant (e.g., pyelonephritis, pleural empyema, osteomyelitis) but by causing bacteremia may have caused meningitis. In addition, a contiguous focus may also be the cause. Sinusitis, mastoiditis, osteomyelitis of the skull, or even a brain abscess may require additional intervention.

While the diagnosis is simple when patients come with classic symptoms of photophobia, nuchal rigidity, severe headache, or mental status changes, this is not generally the case at the initial evaluation. There is a continuum from a complete absence of clinical signs or symptoms at the time the first bacteria cross the blood–brain barrier to severe symptoms.

The point in that continuum of symptoms when the child is seen and the rate at which the disease is progressing will determine how difficult it will be for the clinician to recognize the child as having meningitis. It is for this reason that approximately half of all patients ultimately diagnosed with bacterial meningitis will visit a clinician during the course of their illness and be sent home from the encounter without meningitis being considered.4

The highest risk for meningitis during childhood is in the first month or two of life and at this age the clinical signs are few. Therefore, a lumbar puncture is generally warranted in the evaluation of the febrile infant younger than 2 months. After 2 months of age the diagnosis remains a challenge but more clinical cues are available to the clinician experienced in evaluating young children.

Children with prior antibiotic therapy during the illness can be more difficult to diagnose. Pretreated patients are less likely to have fever at examination, less likely to have altered mental status, and will have longer duration of symptoms at diagnosis.5Table 16–2 summarizes the pertinent history and physical examination in the evaluation of patients suspected of having meningitis. Figure 16–2 is a flow diagram for the diagnostic approach for patients being evaluated for meningitis.

Cerebrospinal Fluid Studies

The CSF should be sent for cell count, glucose, protein, Gram stain, and bacterial culture. A CSF fungal culture should also be performed in premature infants and in children with other immunocompromising conditions. Most patients with bacterial meningitis will have an elevated CSF white blood count (WBC) count with 80% or more polymorphonuclear cells, the CSF glucose will be low in about 50% and the CSF protein concentration is generally elevated. The CSF culture is positive in about 80% of patients and the culture is positive within 24–48 hours unless antimicrobials were given prior to lumbar puncture.

For the infant under 1 month of age with CSF pleocytosis, no prediction rule can reliably exclude bacterial meningitis and these infants should be treated as possibly having bacterial meningitis until the results of CSF culture are known. Further, not even normal CSF parameters exclude bacterial meningitis as 10% of neonates with bacterial meningitis will not have a CSF pleocytosis6–8 and a majority of very low birth weight (≤1.5 kg) neonates with bacterial meningitis have no CSF pleocytosis at diagnosis (by positive CSF culture).6

For older infants and children with CSF pleocytosis, a multicenter review of pediatric cases of bacterial meningitis clearly demonstrated that children older than 1 month with CSF pleocytosis (and no antibiotic pretreatment) can be considered at very low risk of bacterial meningitis if the Gram stain has no organisms, the CSF absolute neutrophil count is ≤1000 cells/mm3, the CSF protein is ≤80 mg/dL, the peripheral blood absolute neutrophil count is ≤10,000 cells/mm3, and there is no history of seizure before or at the time of presentation.7 Nonetheless, even with these serially applied low-risk criteria, 2 of 121 children with bacterial meningitis had none of these risk factors for a sensitivity of 98% (95% confidence intervals: 94–100%). Both of these children were younger than 2 months of age.

Among patients beyond the first months of life, the serum procalcitonin (≥0.5 ng/mL), C-reactive protein (≥20 mg/L), CSF lactate, and the CSF glucose or CSF glucose to serum glucose ratio can all be helpful in distinguishing bacterial from nonbacterial causes of acute meningitis.8–11 While not routinely available the CSF levels of interleukin 6 (IL-6) and tumor necrosis factor are also much higher in patients with bacterial versus aseptic meningitis.12

Patients with bacterial meningitis have a predominance of polymorphonuclear cells before treatment. Approximately half of patients with aseptic or demonstrated enteroviral meningitis will also have a CSF polymorphonuclear cell predominance. Symptom duration does reliably result in a change in the CSF white blood profile, so serial lumbar punctures for this purpose are not recommended.13,14

Latex agglutination studies of the CSF for bacterial antigens do not typically provide any benefit and are not routinely recommended.15 Polymerase chain reaction (PCR) tests for bacteria are not readily available and have not shown much benefit beyond culture but some available viral PCR tests are faster and more sensitive than viral culture and should be considered for specific etiologies. Because enteroviruses are common causes of meningitis and enterovirus PCR tests are now readily available, it is suggested that if the turn around time for this test at your institution is usually ≤48 hours this test should be routinely considered. Previous studies have demonstrated a decrease in hospital length of stay when enterovirus PCR testing is routinely performed during enteroviral season.16,17 Because it is slower than most routine bacterial culture or PCR tests, the routine use of viral culture is not recommended. Where herpes simplex meningoencephalitis is a possible concern a herpes simplex PCR should be sent.

A positive Gram stain or growth of a pathogen in CSF culture definitively identifies bacterial meningitis. Unfortunately the diagnosis is more difficult to establish or exclude in the patient who has received antibiotic treatment prior to obtaining CSF for examination and culture. Pretreated patients are less likely to have a positive CSF Gram stain or culture.5

While bacteria can be recovered in some circumstances for one or more days after initiation of antimicrobial therapy, the CSF is sterilized in 2 hours or less with meningococcal infections, in 4 to 12 hours with pneumococcal infections, and as early as 8 hours with group B streptococcal infections.18Table 16–3 includes suggested studies to be ordered in the evaluation of the patient with suspected meningitis.

It should be noted that it may not be possible to safely perform a lumbar puncture and the diagnosis will need to be a presumptive one in some patients. The lumbar puncture may need to be deferred if there are concerns of clinical instability, significantly elevated intracranial pressure (bradycardia, hypertension, irregular respirations, papilledema), or concern for intracranial lesion. The presence of thrombocytopenia (≤50,000 platelets/mm3) should prompt consideration for platelet transfusion prior to lumbar puncture. The presence of skin infection in the area overlying lumbar puncture should also be considered a relative contraindication because of the concern for inoculating bacteria from the skin site to the CSF.

Other Studies

The decision to perform a lumbar puncture should not rest on the results of blood testing. Among infants, a very low or high peripheral WBC does increase the risk for bacterial meningitis but 41% will have peripheral WBC values in the normal range.19 A negative peripheral blood culture does not rule out bacterial meningitis. Among patients with confirmed bacterial meningitis, no bacterial growth is seen in blood cultures in approximately one-third of cases.13,14,18–20 Despite these limitations, blood should be routinely obtained for culture from patients evaluated for suspected meningitis. A positive blood culture will influence the choice of antimicrobial therapy in cases where the CSF does not yield growth of a pathogen. Serologic tests may be helpful and Lyme serology should be routinely ordered in the proper epidemiologic setting (e.g., areas where infection is common, opportunity for exposure).

Neuroimaging does not need to be routinely ordered but should be performed during the course of therapy when patients experience unexplained changes in mental status, focal neurologic deficits, or there is other reason for concern of intracranial abscess, empyema, hemorrhage, infarct, or thrombosis.

Optimize supportive care to ensure adequate oxygen delivery including consideration of adequate cerebral perfusion pressure. Patients with septic shock will require appropriate resuscitation to maintain adequate oxygen and glucose delivery to the brain. In patients without the need for additional fluids to support the circulation, intravenous fluids should provide for “maintenance” needs but cautiously to avoid worsening the hyponatremia of the syndrome of inappropriate antidiuretic hormone secretion (SIADH) should it occur. Baseline and daily weight measurements and the evaluation of electrolytes may be helpful in monitoring for the development of SIADH.

When there is a high clinical suspicion for bacterial meningitis, the lumbar puncture should be performed as soon as it is clinically safe and feasible and the administration of antibiotic(s) should occur without awaiting the results of CSF studies or computed tomography (which itself is not routinely indicated unless there are focal signs or symptoms). Delays in antibiotic administration should be avoided and are associated with worsened outcome in adults.21

Until an organism is isolated, patients should be treated with empiric antibiotic therapy that covers the range of potentially causative pathogens. Empiric antimicrobial coverage should rarely be narrowed based on the reading of the Gram stain alone as errors in interpretation are common. Conversely if the Gram stain suggests an organism not well covered by empiric therapy the coverage should be broadened awaiting the results of culture. It is important to consider Listeria as a potential pathogen since optimal therapy includes the use of ampicillin, which is no longer routinely included in many empiric treatment strategies.

The initial empiric antibiotic therapy for most children is listed in Table 16–4. Dosing recommendations for these antimicrobials are listed in Table 16–5. Once the specific etiologic agent causing the meningitis is known, the treatment can be focused on that organism. Recommendations by organism (and the organism antibiotic susceptibility) are listed in Table 16–6. Duration of antibiotic therapy varies depending on the causative organism, the time to CSF sterilization, the extent of CSF inflammation, and the patient's clinical response. In uncomplicated cases, parenteral antibiotic courses are typically 7 days for N. meningitidis, 7–10 days for H. influenzae type B, 10–14 days for S. pneumoniae, 14–21 days for group B streptococci, and 21 days for gram-negative organisms.

Vancomycin is a necessary component in the empiric treatment of patients with possible penicillin-resistant pneumococcus. However, one retrospective study has suggested that the use of vancomycin in the first 2 hours of treatment may be associated with an increased risk for hearing loss in pneumococcal meningitis.22 Further corroboration is needed before changes in practice can be recommended but the priority should be to administer the dexamethasone and cefotaxime as quickly as is feasible and then administer vancomycin.

Corticosteroid Therapy

Dexamethasone when given 15 minutes prior or simultaneously with antibiotics appear to reduce mortality and sequelae among patients with H. influenzae type B meningitis.23 Dexamethasone use in children with meningitis caused by S. pneumoniae is more controversial.24 It is not known whether dexamethasone therapy has an effect on cognitive outcome. There is no demonstrated efficacy for the use of dexamethasone after antibiotics have been administered. There has been no demonstrated benefit to the use of steroids in meningococcal or neonatal meningitis.25–27 The key steps in the evaluation and management are summarized in Table 16–3. Most authors reserve the use of dexamethasone to patients with clinical or laboratory findings highly suggestive of bacterial meningitis thought most likely to be associated with H. influenzae type B or S. pneumoniae. Dexamethasone is not routinely recommended for the child with suspected aseptic meningitis receiving antimicrobial therapy pending the results of bacterial culture. Dexamethasone has not been adequately studied or demonstrated beneficial in the treatment of neonates with bacterial meningitis.

Additional Considerations

A repeat lumbar puncture for a follow-up cerebrospinal fluid culture is not routinely recommended but should be considered when patients fail to improve within the first 24–48 hours or when the identified pathogen can be anticipated to be difficult to eliminate with conventional antibiotic therapy based on susceptibility testing or prior experience (e.g., nosocomial gram-negative infections, resistant pneumococcus).

Consultation with an infectious diseases specialist is recommended for most cases of gram-negative meningitis, fungal meningitis, meningitis in a patient with reduced immunity or indwelling intracranial hardware, and in the treatment of organisms with reduced susceptibility to antibiotics.

Meropenem is preferred to a third-generation cephalosporin for the treatment of specific gram-negative infections caused by organisms, such as Citrobacter-, Enterobacter-, or Serratia-related cases which may possess the ability to express extended spectrum beta-lactamases (ESBL) that may be induced by cephalosporin therapy.

The mortality is associated with many factors but overall is about 5% for H. influenzae and meningococcal meningitis and in the vicinity of 20% for pneumococcal and Listeria meningitis in developed countries. The most common complications are listed in Table 16–7.

The risks for mortality and sequelae are related to the severity of disease. Not surprisingly, mortality (approximately 30%) and the rate of sequelae among survivors (approximately 70%) are very high among patients requiring mechanical ventilation during therapy of acute bacterial meningitis and are highly correlated with severity of illness at admission.28 Mortality also increases if there is a significant neurologic event during the acute phase of the illness. These may be caused by general cerebral edema, hypoperfusion, loss of microcirculatory control, and vascular thromboses, which cause secondary effects such as herniation and infarction.

The time course of presentation is also associated with severity of disease. Children who present and are diagnosed as having bacterial meningitis within the first 24 hours of illness are much more likely to have severe disease than those with a more insidious presentation that may develop over 2 days or more before diagnosis.29

Subdural effusions are commonly noted during the first week of therapy for bacterial meningitis of infants (approximately 30%) but do not typically require intervention.30 Subdural empyema is a rare but important complication of meningitis which can usually be distinguished from effusion with diagnostic imaging.31

Monitoring of urine output, daily weights, and serum electrolytes is prudent in the early phase in order to quickly identify hyponatremia and manage patients developing the SIADH secretion. Less common complications include acute spinal cord dysfunction—myelitis.

Long-Term Sequelae

Hearing loss occurs in a large proportion (10–30% depending on the pathogen) of children and is associated with severity of infection (lower CSF glucose levels, raised intracranial pressure, nuchal rigidity, Strep. pneumoniae as the causative agent) with profound hearing loss in 5%.32 Hearing loss, when it occurs, will be noted to some degree with the earliest testing.33 While many children will experience some improvement in hearing over time, others may have further progression of hearing loss over months to years.34 After bacterial meningitis, children are more likely to have cognitive impairments, such as poor linguistic and executive functions, and behavioral problems than their peers.35 As a result, even after apparent complete recovery from bacterial meningitis, children benefit from formal neuropsychiatric evaluation in addition to routine tests of hearing as the findings may be subtle. Three percent to 5% of children will have seizures, mental retardation, and/or some degree of spasticity or paresis.32

Special Circumstances

Patients with immunocompromise, intracranial implants including cochlear implants, ventriculoperitoneal shunts, and infants in the intensive care setting are at increased risk of meningitis, including meningitis caused by bacteria or fungi not typical of other children. Therefore, they should be considered separately with regard to primary prevention and optimal initial empiric therapy.

Citrobacter meningitis in the neonate, and Citrobacter koseri in particular, is associated with a very high rate of mortality (30%) and sequelae (75%) and is the most common cause of neonatal abscess (Figure 16–3). Infants with Citrobacter meningitis should have brain imaging during the course of therapy to evaluate for this complication.

Meningitis with cranial nerve palsies, appropriate local epidemiology, or protracted symptoms, should raise suspicion for Lyme disease, and depending on the cranial nerves involved, a basilar meningitis as can be seen with tuberculous or some fungal meningitis should be considered.36 Causes of chronic meningitis are also quite varied but pathogens such as Brucella, Lyme, syphilis, cryptococci, and other fungi may cause a more indolent and chronic meningitis. Noninfectious causes include malignancies, sarcoid, and autoimmune disease.

Discussion of eosinophilic meningitis (≥10 eosinophils/mm3 of CSF) is beyond the scope of this chapter but may be seen in relation to indwelling foreign body, certain parasitic diseases, malignancies, in response to medications, and with coccidioidal meningitis.

Chemoprophylaxis is recommended for close contacts of patients with disease caused by N. meningitides or H. influenzae type B. Active immunization against H. influenzae type B and Strep. pneumoniae is now part of routine childhood vaccination schedules and immunization for N. meningitidis is typically required prior to college entry in the United States.

Intrapartum ampicillin administered to women at high risk of transmitting group B streptococci to their infants is effective in reducing the risk of early onset but not late-onset group B streptococcal disease. Women with penicillin allergy can receive clindamycin, erythromycin, or in the absence of immediate-type hypersensitivity, a first-generation cephalosporin. Up to 20% of pregnant women require intrapartum antibiotic prophylaxis. Recurrent group B streptococcal infections occur; most of the isolates from recurrent infections are identical to the isolates from the initial infection. Recurrent infection likely occurs as a consequence of persistent mucosal group B streptococcal colonization. Attempts to eradicate mucosal colonization with oral rifampin have met with varied success. Consultation with a pediatric infectious diseases or immunology specialist should be considered in infants with recurrent group B streptococcal infections.