Clinical Features

The sudden onset of a stiff neck with fever in a child who appears ill usually indicates the presence of meningitis (Box 80-1). In these cases, a lumbar puncture is a crucial part of the initial evaluation.

Box 80-1 Signs and Symptoms of Bacterial Meningitis

Fever, acute course

Vomiting, anorexia

Nuchal rigidity

Meningismus

Kernig’s sign

Brudzinski’s sign

Tripod and knee-kissing sign

Depression of consciousness, coma

Irritability, photophobia, headache

Seizures, focal neurologic deficits

Lumbar Puncture

The diagnosis of bacterial meningitis is based on examination of CSF [Feigin et al., 1992]. Minimizing sequelae of bacterial meningitis depends on the prompt initiation of effective antibiotic therapy. A lumbar puncture should therefore be performed whenever there is even slight clinical evidence of meningitis. Rarely, other sites may be used to obtain CSF, including ventricles or shunts installed to divert CSF.

Lumbar puncture is a safe procedure for most patients. Minor discomfort may occur, but serious consequences are rare. Headache 1–6 hours after lumbar puncture is a troublesome but benign complication. When present, the headache is frontal and is exaggerated by movement from a lying to a sitting or standing position. Headache after lumbar puncture seems to be caused by continued leak of CSF from the site of the lumbar puncture, with resulting low CSF pressure and traction on pain-sensitive structures. The problem can be mitigated by removal of a minimal amount of fluid and use of a small-bore needle [Lynch et al., 1992; Lavi et al., 2010]. Pain in the low back or legs occurs occasionally after lumbar puncture; this pain is of uncertain etiology and resolves spontaneously in several days.

Herniation of the brainstem and cerebellar tonsils into the foramen magnum is extremely rare in children [Marton and Gean, 1986]. A risk of herniation is clearly related to substantially and focally increased intracranial pressure [Richards and Towu-Aghantse, 1986]. The risk of producing meningitis by performing a lumbar puncture in a child with septicemia is insignificant and is not a contraindication to the test [Shapiro et al., 1986].

With the widespread introduction of highly effective bacterial conjugate vaccines against H. influenzae type b and S. pneumoniae, there has been a significant decrease in the incidence of bacterial meningitis in children. This has reduced the probability that a child with CSF pleocytosis will have bacterial meningitis. Several studies were able to validate a previously developed clinical prediction rule, the so-called Bacterial Meningitis Score, for identifying children with CSF pleocytosis at very low risk of bacterial meningitis [Nigrovic et al., 2007]. Patients are classified as very low risk if they lack all of the following criteria: positive CSF Gram stain, CSF absolute neutrophil count of at least 1000 cells/μL, CSF protein of at least 80 mg/dL, peripheral blood absolute neutrophil count of at least 10,000 cells/μL, and a history of seizure before or at the time of presentation.

Repeated lumbar puncture is indicated only when delayed CSF sterilization or relapse or recurrence of meningitis is clinically suspected. Lumbar puncture should not be repeated, unless the expected findings would make a difference in management. Most patients with bacterial meningitis require just one initial diagnostic lumbar puncture.

Laboratory tests

The initial diagnosis of meningitis is based on examination of CSF. Bacterial infection is suggested by increased opening and closing pressure; a cloudy (i.e., not clear) CSF; or a CSF pleocytosis with polymorphonuclear predominance, increased protein and lactate concentration, or decreased glucose concentration. Details and normal values have been described previously (Table 80-1) [Ahmed et al., 1996; Kim, 2010].

Table 80-1 Cerebrospinal Fluid in Neurologic Infections

Confirmation of the diagnosis requires the detection in the CSF of a specific bacterial pathogen by microscopy or a positive culture, polymerase chain reaction, or rapid antigen-detection test of CSF. Smears of CSF should be examined by Gram stain (acid-fast stain if tuberculosis is suspected). Gram-stained smears of the CSF sediment reveal bacteria in 70–90 percent of untreated cases of meningitis caused by S. pneumoniae or N. meningitidis, but only in one-third of cases caused by L. monocytogenes [Kim, 2010]. The CSF should be cultured on a blood agar plate, on a chocolate agar plate (or on Fildes or Leventhal medium), and in broth. Special cultures for mycobacteria, fungi, or other microorganisms should be undertaken if clinically indicated. In an area of increasing antimicrobial resistance, adequate in vitro susceptibility testing must be routine for any cultured neonatal pathogen. Various antigen-detection tests are available (e.g., countercurrent immunoelectrophoresis, latex particle agglutination, enzyme-linked immunosorbent assay) and may be helpful in establishing an etiologic diagnosis rapidly.

Polymerase chain reaction and other sequence-based microbial detection methods are being used increasingly in many clinical microbiology laboratories. These tests improve detection of pathogens, particularly in patients pretreated with antibiotics [Fredricks and Relman, 1999; Chiba et al., 2009]. Careful standardization and validation of these tests is necessary.

Nonspecific parameters have been studied in the CSF to establish or exclude a bacterial etiology rapidly, including various indices of inflammation (e.g., cytokines, C-reactive protein, and lactate) and measurement of various enzyme activities [Gray et al., 1986; Lutsar et al., 1994]. With the exception of lactate, these parameters are not routinely used to assess patients with meningitis.

Other Tests

Most pediatric patients with bacterial meningitis are initially bacteremic, as might be expected based on the hematogenous pathogenesis of the disease. Blood cultures taken at hospital admission are valuable and are positive in 80–90 percent of cases of bacterial meningitis [Feigin et al., 1992].

Cultured bacteria from skin or mucosal surfaces, including throat and nose, are neither sensitive (organisms causing the meningitis may not be present) nor specific (other organisms are often present), and are not helpful. The culture of appropriate sites of primary or concomitant infected foci, such as cellulitis (e.g., periorbital and buccal), middle ear effusions, sinusitis, and mastoiditis, and urine in young infants may yield the meningitis-causing pathogen. Petechiae often represent microemboli, with bacteria (e.g., meningococci) present in the lesion. Gram-stained smears made from petechiae may provide immediate clues as to the identity of the invading microorganisms.

Tests of inflammatory responses, such as a total and differential peripheral white blood cell count, erythrocyte sedimentation rates, and measurement of C-reactive protein concentration, may be helpful diagnostically and during the treatment of invasive bacterial disease. Such tests are not useful, however, in making or excluding a diagnosis of bacterial meningitis.

Among new markers, serum procalcitonin level seems to be one of the most sensitive and specific predictors for discriminating between bacterial and aseptic infections. Preliminary experiences indicate that a serum procalcitonin level greater than 0.5 ng/mL predicts bacterial rather than aseptic meningitis in children [Dubos et al., 2008].

Brain Imaging

Neuroimaging studies, such as CT, MRI, and cranial sonography (in young infants with access through open fontanel), should be used in patients with signs of increased intracranial pressure (e.g., depressed consciousness, sluggishly reactive or dilated pupils, ophthalmoplegia, retinal changes, abnormal respiratory patterns, and cardiovascular instability). Increased intracranial pressure usually is caused by cerebral edema or other intracranial complications, such as subdural collection, thrombosis, infarction, brain abscess, and hydrocephalus [Ashwal, 1995; Tunkel and Scheld, 1995].

MRI has improved resolution compared with CT, provides additional insights into the pathogenesis of neurologic complications, and eventually may replace CT in most situations [Smith and Caldemeyer, 1999]. Cortical enhancement may be identified and is presumed to indicate cerebritis. Repeat imaging is useful in following the evolution of abnormalities [Bodino et al., 1982] and in determining the need for subsequent intervention. Abnormalities on imaging are associated with an unfavorable outcome [Packer et al., 1982]. Contrast MRI, although not usually considered a diagnostic test for meningitis, can show meningeal enhancement (Figure 80-1) [Kioumehr et al., 1995; Runge et al., 1995].

Fig. 80-1 Postcontrast coronal magnetic resonance image of a 6-month-old infant with Streptococcus pneumoniae meningitis.

Meningeal enhancement is visible (arrowheads).

(Courtesy of Dr. Blaine Hart.)

Pathophysiologic Changes

Seizures

Seizures complicate 20–30 percent of cases of childhood bacterial meningitis. Patients with severe illness are more likely to have seizures. They occur most frequently 2–3 days into the illness and cease 1–3 days later. Seizures may be focal or generalized, and the electroencephalogram (EEG) may be normal or may show generalized or focal epileptiform abnormalities, focal slowing, or only background slowing. The functional and structural alterations of the brain resulting from the meningitic process that may induce seizures include vasculitis and infarction, fever, electrolyte imbalances, and metabolic changes. Chronic seizures after recovery from meningitis are approximately 3-fold more common in patients than in age-matched control children [Bedford et al., 2001].

Deafness and Cranial Nerve Damage

Deafness is the most frequent neurologic sequela of bacterial meningitis. Of patients with bacterial meningitis, 1–30 percent develop hearing loss (see Table 80-2), and the risk is highest when meningitis is caused by S. pneumoniae. Low CSF glucose concentrations are another risk factor for hearing loss [Eisenhut et al., 2003]. Hearing impairment after meningitis can improve for several months, but many children have severe, permanent, and bilateral hearing loss [Wellman et al., 2003]. Typically, sensorineural hearing loss in both ears is found; rarely, a conductive hearing loss associated with a middle ear infection may be present. Hearing should be assessed in children with bacterial meningitis before discharge. Audiometry is an appropriate test in older children. In young infants, brainstem auditory-evoked responses or transient otoacoustic emissions can be used to detect hearing loss early [Cohen et al., 1988; Francois et al., 1997]. When hearing loss occurs, early intervention by a team that is knowledgeable in the management of this problem in infants and children should be encouraged. Cochlear implants are a consideration in the most severe cases.

Experimental studies suggest that bacteria reach the cochlea primarily via the cochlear aqueduct [Bhatt et al., 1993]. The deafness is caused by infection and inflammation in the cochlea [Klein et al., 2003]. Nitric oxide, other inflammatory mediators, and bacterial products, specifically the pneumococcal toxin pneumolysin, contribute to the loss of hairy cells in the cochlea [Amaee et al., 1995].

Other cranial nerve abnormalities develop in bacterial meningitis because of the involvement of the cranial nerves by the inflammatory process in the subarachnoid space. Increased intracranial pressure also can cause cranial nerve dysfunction by pressure or stretching. Cranial nerves VI and III are involved relatively frequently. Ocular palsies are usually reversible.

Neuronal Damage

The neurologic sequelae after meningitis, in addition to hearing loss (see previously), include focal sensorimotor deficits, mental retardation, seizure disorders, and cortical blindness (see Table 80-2). Detailed studies of the neurologic outcome after bacterial meningitis have led to an appreciation of the significant long-term functional impairment resulting from the disease [Grimwood et al., 1995; Bedford et al., 2001; Oostenbrink et al., 2002]. In a cohort of 5-year-old children who had survived bacterial meningitis during their first year of life, there was a 10-fold increase in the risk of moderate to severe disabilities, with 7.5 percent demonstrating learning difficulties; 8.1 percent, neuromotor disabilities; 7.3 percent, seizure disorders; 25.8 percent, hearing problems; and 11.9 percent, behavioral problems [Bedford et al., 2001]. Neurologic sequelae of childhood meningitis can persist at least into young adulthood, impairing learning during the entire time that children go to school [Grimwood et al., 2000]. S. pneumoniae consistently was associated with higher rates of disabilities than infection with H. influenzae or N. meningitidis [Grimwood et al., 1996].

Meningitis causes brain damage of cortical and subcortical structures (Figure 80-2). For much of the brain damage, particularly in the cortex, ischemia represents an important mechanism of injury (Figure 80-3) [Pfister et al., 2000]. In keeping with this notion, inhibition of excitatory amino acids and preservation of cerebral blood flow through inhibition of reactive oxygen radicals or endothelins are protective in experimental meningitis [Meli et al., 2002]. Matrix metalloproteinases, in addition to their effect on the blood–brain barrier and on cytokine release, also may contribute directly to ischemic changes in the brain.

Fig. 80-2 Axial T2-weighted magnetic resonance image of a 13-year-old child with Neisseria meningitidis meningitis.

The meningitis is complicated by multiple areas of cerebritis (arrowheads and other locations).

(Courtesy of Dr. Blaine Hart.)

Fig. 80-3 Computed tomography scan of a 16-day-old neonate with group B streptococcal meningitis.

There is massive infarction of the anterior circulation.

(Courtesy of Dr. Blaine Hart.)

A structure frequently damaged during meningitis is the dentate gyrus of the hippocampus. This structure is important for learning and memory acquisition, and represents one of the few areas of the brain where neurogenesis is on-going throughout life. Within this structure, pathology studies in humans dying from meningitis and animal models of meningitis consistently have documented damage to neurons [Gianinazzi et al., 2003; Nau et al., 1999]. Two forms of injury seem to affect the dentate gyrus neurons [Bifrare et al., 2003]. One is mediated by activation of caspase-3 and corresponds to classic apoptosis. This form of injury preferentially affects immature, postmitotic neurons in the subgranular zone, is associated with learning deficits in experimental animals surviving meningitis, and may represent an anatomic substrate for cognitive impairment and learning disabilities after meningitis [Bifrare et al., 2003; Leib et al., 2003]. The other form of dentate gyrus injury exhibits no obvious selectivity for a subset of neuronal cells and is associated with translocation of apoptosis-inducing factor. For both forms of dentate gyrus injury, the exact molecular triggers are unknown, but ischemia does not seem to be important. Potential mediators include pneumococcal products (pneumolysin and hydrogen peroxide), inflammatory mediators (e.g., tumor necrosis factor-α), matrix metalloproteinases, excitatory amino acids, reactive oxygen species, and others [Leib et al., 1996; Meli et al., 2002].

Hydrocephalus

Meningitis leads to disturbances of CSF homeostasis, with increased CSF production and reduced CSF absorption across the sagittal sinus – arachnoid villi system [Scheld et al., 1980]. In the acute phase, these disturbances can lead to ventriculomegaly associated with intracranial hypertension. This form of ventriculomegaly is frequently transient. Chronic hydrocephalus occurs as a long-term sequela of bacterial meningitis in infants and children (see Table 80-2). Its pathogenesis is either obstructive, secondary to inflammatory changes of CSF drainage pathways, or ex vacuo, secondary to loss of brain parenchyma. Chronic internal shunting with internal drainage is necessary in patients with persistent obstructive hydrocephalus.

Septic Shock and Disseminated Intravascular Coagulation

Bacterial meningitis can be complicated by sepsis and septic shock, independent of the infecting organism. All patients with bacterial meningitis need to undergo close cardiovascular monitoring until they are clinically stabilized with antimicrobial and supportive therapy. The association of sepsis and extensive disseminated intravascular coagulation suggests infection caused by N. meningitidis [Yung and McDonald, 2003]. A similarly fulminant and devastating course can be seen with S. pneumoniae in patients without a functional spleen. These patients present with rapidly progressive signs of severe illness, rigors, severe muscular pain, and skin lesions that progress from an early maculopapular rash to petechiae to enlarging purpuric lesions [Yung and McDonald, 2003]. The organism sometimes can be visualized on Gram stain of scrapings from purpuric lesions. Signs and symptoms of meningitis often are not prominent in patients with disseminated meningococcal sepsis. Laboratory abnormalities include thrombocytopenia, hypofibrinogenemia, and massively elevated D-dimers. Overt adrenal insufficiency seems to be relatively rare in children with meningococcal disease [Riordan et al., 1999]. The most fulminant form of sepsis and disseminated intravascular coagulation is the Waterhouse–Friderichsen syndrome, which leads to sudden cardiovascular collapse associated with adrenal hemorrhage.

Extra-Axial Fluid Collections

The subdural space is a potential intracranial space situated between the arachnoid and dura. Fluid can collect in the subdural space and in the subarachnoid space. The collection in the subarachnoid space may represent loculated CSF, widening of the subarachnoid space in response to inflammation, or loss of brain parenchyma. Because imaging often cannot distinguish between a fluid collection and a widened subarachnoid space, the fluid collection as seen on imaging is referred to as an extra-axial fluid collection. Extra-axial fluid collections are found in 50 percent of children with bacterial meningitis who have cranial CT performed [Syrogiannopoulos et al., 1986].

Extra-axial fluid collections occur more commonly in bacterial meningitis during the first year of life and often are recognized during the first week of illness. The extra-axial fluid is preferentially situated over the convexities, may accumulate bilaterally, is serosanguineous with granulocytes and high protein concentrations, and is usually sterile (Figure 80-4) [Snedeker et al., 1990]. Extra-axial fluid collections often are detected on imaging studies performed because of persistent fever, clinical signs of meningitis, seizures, or focal neurologic deficits. Often, they represent a coincidental finding, but they may also be the cause of persistent increase in intracranial pressure or focal neurologic deficits. Extra-axial fluid collections, as discussed here, must be distinguished from subdural empyema, which frequently is a complication of sinusitis or otitis media and needs emergent treatment. Subdural empyema will be discussed later in this chapter.

Fig. 80-4 Coronal T2-weighted magnetic resonance image of a 6-month-old infant with Streptococcus pneumoniae meningitis.

Massive extra-axial fluid collections are visible (arrowheads).

(Courtesy of Dr. Blaine Hart.)

Antibiotics

Effective eradication of the infecting pathogen requires that antibiotics display bactericidal activity in the CSF. Experimental studies have documented the fact that CSF concentrations that exceed the in vitro minimal inhibitory concentration severalfold (>10-fold for most antibiotics) are required to achieve maximal bacterial killing rates. CSF concentrations of antibiotics, even with inflamed meninges, are only a fraction of simultaneous serum concentration (usually between 5 and 25 percent), and antibiotics must be dosed higher for the treatment of meningitis than for most other infections [Täuber and Sande, 1990].

Empiric therapy should be instituted rapidly in all children with suspected bacterial meningitis. The progressive pathophysiologic alterations during meningitis, which rapidly lead to brain damage and death in most untreated patients, strongly suggest that any delay in antibiotic treatment may be harmful and should be avoided [Aronin et al., 1998, Koster-Rasmussen et al., 2008]. In practice, the appropriate sequence of obtaining CSF and blood for cultures and other tests, instituting antibiotics, and obtaining imaging studies to exclude mass lesions and other pathologies needs to be adapted to the individual patient. The sicker the patient and the more rapidly progressive the disease, the more urgent is the start of antibiotic therapy. Antibiotic therapy should not be delayed by more than approximately 15 minutes in critically ill patients for an attempt to obtain CSF or blood, despite the reduced yield of bacterial cultures if the samples are obtained after initiation of antibiotics. A CT scan is required before the lumbar puncture in all patients in whom there is suspicion of important CNS alterations, particularly if a mass lesion is suspected, and the patient has signs of increased intracranial pressure. In adults, criteria have been established to identify patients with suspected meningitis in whom a lumbar puncture can be performed safely [Hasbun et al., 2001]. Whether these criteria are valid in children has not been established. The contraindications for lumbar puncture were discussed earlier.

The choice of empiric antibiotics depends on the suspected pathogens and their antibiotic susceptibilities, which may display regional differences. All pathogens commonly encountered in a particular patient population should be covered empirically, and rare pathogens should be empirically targeted in patients with corresponding risk factors (Table 80-3 and Table 80-4). Recommended empiric treatment regimens adequate for patients without additional risk factors are shown in Table 80-5. When the causative pathogen has been isolated, and antibiotic sensitivities are known, therapy should be adjusted to cover the respective pathogen optimally (see later).

Table 80-3 Common Causes of Acute Bacterial Meningitis According to Age

Table 80-4 Rare Causes of Bacterial Meningitis in Infants and Children

* Rare past the neonatal period.

Table 80-5 Recommendations for Empiric Antibiotic Therapy of Bacterial Meningitis

The duration of therapy is not strictly defined and should take into account the clinical response to treatment. Meningitis caused by N. meningitidis can be treated for 4 days [Crosswell et al., 2006], and for meningitis caused by S. pneumoniae or H. influenzae, 4–7 days treatment were not inferior to 7–14 days in a recent meta-analysis [Karageorgopoulos et al., 2009]. Meningitis caused by Enterobacteriaceae, L. monocytogenes, other unusual pathogens, or organisms with reduced antibiotic sensitivity may require longer treatment courses. Antibiotic therapy should not be prolonged past the recommended durations because of a persistent fever in a child who has otherwise displayed a favorable clinical response. Other causes of fever, such as drug fever, intravenous line infections, other undiagnosed infections, and extra-axial fluid collections, should be considered.