The classic triad of fever, neck stiffness, and a change in mental status is present in less than 50% of pediatric meningitis cases.
Streptococcus species (particularly viridans , pneumoniae , and milleri ) and Staphylococcus aureus are the most common pathogens identified in pediatric subdural empyemas, although up to 30% demonstrate no bacterial growth.
Viral causes of encephalitis are very common in children, with enteroviruses, herpesviruses, and arboviruses causing the greatest burden of disease in the northern hemisphere.
Central nervous system (CNS) infections and the need for neurologic intensive care represent an important patient population in the pediatric intensive care unit (PICU). These infections can progress rapidly and lead to devastating and permanent neurologic sequelae. The diversity of infectious agents causing CNS disease includes bacteria, viruses, fungi, parasites, and amoebas. In addition, inflammatory postinfectious diseases such as acute disseminated encephalomyelitis and autoimmune processes such as acute hemorrhagic leukoencephalitis and anti- N -methyl- d -aspartate (NMDA) receptor-mediated encephalitis feature diverse presentations and require a high index of suspicion. Care of these patients requires a knowledge of epidemiology, signs and symptoms, predisposing factors, and appropriate laboratory and radiologic diagnostic tools. Understanding the importance of timely diagnosis and treatment of these conditions could help reduce both morbidity and mortality.
Acute bacterial meningitis is less common than in the past but remains a potentially severe infection in childhood. The majority of children with bacterial meningitis are admitted to a PICU for the initial phases of their care. Bacterial meningitis occurs when bacteria invade the cerebrospinal fluid (CSF) within the pia-arachnoid spaces surrounding the brain and spinal cord. This is most often the result of hematogenous dissemination of microorganisms arising from colonization or infection of the upper respiratory tract or other distant sites. Bacterial meningitis also may be associated with trauma, surgery, and placement of neurosurgical intracranial devices.
The incidence of bacterial meningitis in children has significantly decreased over the past 4 decades in countries that have introduced capsular polysaccharide-protein conjugate vaccines against the most common bacterial etiologies. Incidence of bacterial meningitis remains highest in the first year of life, with greatest risk among infants younger than 2 months. In 2009, the overall incidence of bacterial meningitis by age group in the United States was about 65 per 100,000 in the first year of life and between 2.0 and 2.5 per 100,000 in children 1 to 17 years old.
Epidemiology of common organisms causing pediatric meningitis is summarized on ExpertConsult.com .
Epidemiology of common organisms causing pediatric meningitis
Many bacterial species can cause meningitis. The most common etiologies historically have been Haemophilus influenzae type b (Hib), Streptococcus pneumoniae , Neisseria meningitidis, Streptocococus agalactiae (Group B streptococcus; GBS), Escherichia coli , and Listeria monocytogenes , , , with the latter three causing disease primarily, but not exclusively, in the first month of life. Staphylococcus aureus meningitis usually occurs within the context of neurosurgical procedures or head trauma, but newly detected strains can cause meningitis in the absence of other risk factors.
In the United States, the incidence of Hib, which was the most common cause of bacterial meningitis, has decreased from 22 to 24 per 100,000 children younger than 5 years in the mid-1980s to fewer than 1 per 100,000. This success has been attributed to vaccine induction of concentrations of serum immunoglobulin G antibodies against the Hib capsular polysaccharide that were sufficient to eradicate nasopharyngeal colonization with this microbe. Meningitis due to other capsular types of H. influenzae occur sporadically worldwide but have not emerged to the level of Hib in the prevaccine era.
Rates of all types of infection due to the more than 90 serotypes of S. pneumoniae have decreased since the introduction of a 7-valent vaccine in 2000, followed by a 13-valent vaccine in 2010. , The majority of pneumococcal meningitis cases now are caused by nonvaccine serotypes, although cases due to the vaccine serotype 19a remain proportionately common. The frequency of pneumococcal isolates with reduced susceptibility to penicillin and ceftriaxone has declined to date since the introduction of the 13-valent vaccine. There remains potential for nonvaccine serotypes to emerge with subsequent increases in incidence of invasive pneumococcal infections, including meningitis, toward prevaccine levels.
N. meningitidis has been the second most common cause of sepsis and meningitis beyond the neonatal period. Following the introduction of conjugated meningococcal vaccines against capsular serogroups A, C, Y, and W-135, the incidence of N. meningitidis meningitis decreased from 0.7 per 100,000 people in 1997 to 0.1 per 100,000 people in 2010. Globally, N. meningitidis causes around 500,000 cases of meningitis and septic shock every year, although incidence rates vary from fewer than 1 per 100,000 per year in North America to 10 to 1000 per 100,000 per year in sub-Saharan Africa. The peak incidence of invasive meningococcal disease occurs in children between 6 months and 2 years of age, with a second smaller peak in adolescents and young adults.
In N. meningitidis infections, capsular group B is the major disease-causing group in most countries, causing the highest rates of disease in North America and Europe. Serogroup Y strains have become more common in the United States in recent years. Serogroup A strains are more common in parts of Africa and Asia. Vaccines using various serogroup B protein antigens are now available but not yet widely used.
Among infants younger than 90 days, GBS and E. coli remain the most common etiologies of bacterial meningitis, each accounting for about one-third of cases, especially among neonates. From 1997 to 2017, early-onset neonatal GBS infections in the United States declined from about 0.60 to 0.22 per 1000 live births due to implementation of guidelines for maternal screening and intrapartum antimicrobial prophylaxis. Rates of late-onset GBS cases have remained relatively stable at about 0.3 per 1000 live births over this time period. , Klebsiella , other Enterobacteriaceae species, L. monocytogenes , and S. aureus can also cause meningitis in this age group. , Viral meningitis, as in older age groups, is more common than bacterial meningitis in young infants.
Lymphocytic meningitis can be a manifestation of early Lyme disease. This entity should be considered for patients in endemic areas or those who have traveled to such areas. Presentation usually involves cranial nerve deficits, especially Bell palsy, and papilledema, as well as the clinical finding of erythema migrans. Most patients with Lyme meningitis are not as acutely ill as with other bacterial etiologies of CNS infection.
Mycobacterium tuberculosis (TB) is a rare but still important cause of pediatric meningitis in the United States. TB meningitis disproportionately affects young children and is the most common form of childhood extrapulmonary tuberculosis after scrofula. Before antituberculosis drugs, tuberculous meningitis invariably caused death, usually within weeks. A meta-analysis in 2014 showed that risk of death from tuberculous meningitis was 19.3%, with risk of neurologic sequelae among survivors greater than 50%. Use of effective antimicrobials and adjunctive corticosteroids, along with advances in management of intracranial hypertension, has improved outcomes. Longer duration of symptoms (days to weeks) before diagnosis is associated with poorer outcomes. Early diagnosis remains a substantial challenge because the earliest stage of TB meningitis often manifests with nonspecific signs and symptoms.
Tissue injury and subsequent morbidity in acute bacterial meningitis is caused collectively by multiple factors, including bacterial proliferation in the subarachnoid space, release of bacterial products, local immune response from CNS tissue, and leukocyte migration into infected tissue. These factors then can lead to injury to the brain parenchyma via (1) ischemia from loss of autoregulation of cerebral blood flow, thrombosis of arteries, veins, or dural venous sinuses secondary to vasculitis, microvascular thrombosis, and/or increased intracranial pressure (ICP); and (2) direct cytotoxicity from bacterial products or the host immune response that may damage neurons and other cell types in the brain.
Relatively recent colonization of the nasopharynx by the causative microbe, often in conjunction with a viral respiratory tract infection, is frequently the antecedent event in bacterial meningitis. After entrance into the bloodstream, microbes reach the CSF most likely via the blood-CSF barrier of the choroid plexus, and possibly other sites, including the intraparenchymal cerebral capillaries of the blood-brain barrier. Magnitude (i.e., higher CFU/mL) and duration of bacteremia may be factors in development of meningitis. Experimental data show that various pathogens pass between or through cerebral capillary endothelial cells or potentially traverse the blood-CSF barrier within phagocytic cells. This ingestion causes a nonfatal phagocytosis, often labeled a “Trojan horse.” ,
Bacterial strains that cause hematogenous meningitis usually have one or more traits that facilitate their invasion into the CSF. Examples include the following.
E. coli: Type 1 fimbriae, outer membrane protein A, and flagella
Group B streptococcus (GBS): Laminin-binding protein, fibrinogen-binding protein A, and pilus protein A
S. pneumoniae: Phosphorylcholine and pneumococcal surface protein C
N. meningitidis: Type IV pili and neisserial adhesin A
These and other identified microbial structures interact with various host cell receptors that connect with host-signaling pathway molecules. A recently identified mutation in pneumococcal penicillin-binding protein, pbp1bA641C , was associated with increased risk of meningitis via increased microbe tolerance of antimicrobial therapy.
Host factors also are important in pathogenesis. There is a general lack of immune defenses in the subarachnoid space, which can allow for rapid multiplication of bacteria within the CSF. Children with asplenia, congenital immunodeficiencies, cochlear implants, human immunodeficiency virus (HIV), and CSF leakage secondary to trauma are at increased risk for pneumococcal meningitis. Patients with asplenia or complement deficiencies are at increased risk for meningococcal meningitis. On rare occasions, acute otitis media has been associated with Streptococcus pyogenes meningitis. Infections by S. aureus and coagulase-negative staphylococci are generally associated with neurosurgery or head trauma.
Human genetic determinants of infectious diseases are increasingly recognized risk factors for infection. In invasive meningococcal disease, several mutations and polymorphisms have been associated with a higher susceptibility and severity of illness. To date, the genetic markers with the strongest supportive evidence are specific polymorphisms in the coding genes of plasminogen activator inhibitor 1 (PAI-1) and receptors of the Fc fraction of immunoglobulins. Deficiencies in the interleukin-1 (IL-1) receptor–associated kinase 4 (IRAK-4) and MyD88, both components of toll-like receptor (TLR) pathways, appear to be associated with increased risk of pneumococcal meningitis. Analysis of 11 single-nucleotide polymorphisms (SNPs) in seven genes involving host immune response suggests a potential cumulative risk when multiple such SNPs are present. These genes included complement factor H, mannose-binding lectin, and toll/interleukin-1 receptor domain containing adaptor protein (TIRAP). In another study, nine candidate SNPs in TLRs 2, 4, and 9 were not associated with increased risk of pneumococcal or meningococcal meningitis.
No constellation of clinical signs and symptoms is 100% predictive of meningitis. The classic clinical presentation of meningitis is characterized by fever, headache, nuchal rigidity, and altered mental status. Clinical symptoms can be ambiguous in the pediatric patient such that the triad of fever, neck stiffness, and a change in mental status is present in less than 50% of pediatric cases. Most but not all children with meningitis have fever. Nuchal rigidity is frequently absent, and changes in mental status such as irritability can often be misattributed to other causes, especially in young infants. On admission, less than 25% will have focal neurologic signs (often cranial nerve involvement), and ≈15% will be obtunded or comatose. The classic signs of Kernig and Brudzinski frequently go unrecognized or are absent and appear to have limited utility in the diagnosis of bacterial meningitis in children. , , The presence of petechiae and purpura may suggest meningococcus as the causative agent, but similar findings can occur when S. pneumoniae and Haemophilus influenzae type b (Hib) are etiologies. Poorly reactive pupils, a bulging fontanel, diplopia, papilledema, and uncontrollable vomiting can be present and are indications of increased ICP.
Acute bacterial meningitis patients can present with seizure activity. A multiyear retrospective study from South America evaluated risk factors in pediatric patients with confirmed bacterial meningitis for the subsequent development of seizure activity. Risk factors for seizure development included age less than 2 years, pneumococcal etiology, altered mental status at admission, and CSF leukocyte count less than 1000 cells. Generalized seizures were present in 20% to 30% of patients within the first 3 days of illness. The presence of focal seizures and onset past the third day of illness was associated with long-term neurologic sequelae. Development of seizure activity in these patients also was associated with a ninefold increase in mortality.
The microbiology laboratory plays a critical role in early identification of the causative bacterium in meningitis. When clinically feasible, the diagnosis of meningitis should include examination of CSF. In most cases of bacterial meningitis, the diagnosis is made on the basis of isolation of pathogenic bacteria from CSF. CSF white blood cell (WBC) count and differential, protein, glucose, Gram stain, and culture are essential in confirming the diagnosis. In the past few years, multiplex polymerase chain reaction (PCR) panels for the most common bacterial, viral, and fungal etiologies of meningitis have become available. Evidence is growing to support the clinical utility of rapid identification of the etiology of meningitis using these technologies. Molecular tests using specific, universal (e.g., 16S ribosomal ribonucleic acid [rRNA] sequence analysis), or multiplex PCR tests may provide evidence of the causative microbes, especially when antibiotics have been given before obtaining material for culture.
Many meningitis patients will have concurrent bacteremia, but the use of blood culture alone as an initial screening test is not appropriate. Prior oral antimicrobial therapy will not significantly alter CSF cell counts, differential, and chemistries in children with bacterial meningitis. However, sensitivity of the CSF Gram stain and culture are greatly reduced. Bacterial antigen tests do not have sufficient diagnostic validity for routine use. Measurement of opening pressure when performing lumbar puncture may help guide decision-making regarding need for acute neurosurgical interventions in some cases. Lumbar puncture should be postponed in children with an ongoing coagulopathy, elevated ICP (such as altered mental status or hydrocephalus seen on head computed tomography [CT] scan or brain magnetic resonance imaging [MRI]), cardiorespiratory difficulty, and patients with suspected mass-occupying lesions. Head CT imaging is indicated before lumbar puncture in those with focal neurologic findings or signs of elevated ICP. It is important to remember that even when head CT scans are reported as normal, herniation may still occur in both children and adults with meningitis. ,
CSF with pleocytosis, usually greater than 1000/μL, and a predominance of polymorphonuclear (PMN) leukocytes are highly suggestive of bacterial meningitis. High protein and low glucose (usually less than half serum value) concentrations also support the diagnosis. A predominance of lymphocytes and monocytes suggests a viral pathogen. This can also be observed in patients with rickettsial and tuberculous meningitis. Pleocytosis also can occur in Kawasaki disease and various autoimmune disorders. CSF eosinophils are commonly observed in patients with CSF shunt devices, but also can be found in parasitic CNS infections. CSF pleocytosis is also observed in adverse reactions to numerous drugs, including nonsteroidal antiinflammatory agents, trimethoprim-sulfamethoxazole, and intravenous (IV) gammaglobulin.
Meningitis caused by enterovirus and related strains can be associated with PMN predominance early in the clinical course, but these patients usually have normal CSF protein and glucose concentrations. Interestingly, hypoglycorrhachia can be observed with certain viral pathogens, such as enterovirus and mumps virus. On occasion, a small percentage of patients may have an initial normal-appearing CSF analysis, especially with N. meningitidis . Gram stains demonstrate organisms in 50% to 75% of patients with bacterial meningitis. The clinician must be cautioned not to make significant changes in antimicrobial therapy solely based on a Gram stain result, as differences in CSF staining during slide preparation sometimes occur.
Interpretation of CSF WBC count when there is significant blood contamination of the specimen (the “bloody tap”) is often difficult. Rules such as allowance of 1 WBC for every 500 to 1000 red blood cells (RBCs) in the specimen or determination of expected versus observed CSF WBCs using concurrent ratio of peripheral blood WBCs to RBCs can be applied, but none of these approaches are well validated. Comparison of the CSF to peripheral blood WBC differential proportions may be useful at times. Reasonable data exist to adjust the CSF protein concentration for blood contamination, with an allowance of 1.1 mg/dL for every 1000 RBCs.
Serum procalcitonin and C-reactive protein assays have been studied as a means of estimating the likelihood of invasive bacterial infections and differentiating between bacterial and aseptic (usually viral) meningitis. To date, the utility of these serum tests for these purposes is unclear. Numerous CSF biomarkers also have been evaluated for ability to distinguish between bacterial and viral meningitis. Sufficient data to provide clinical confidence regarding predictive values for any specific marker, including CSF lactate, are still lacking.
Repeated lumbar punctures during antibiotic therapy are no longer routinely recommended. However, repeat lumbar puncture after 48 hours of therapy should be considered in cases caused by S. pneumoniae when the isolate is (1) nonsusceptible to third-generation cephalosporins; (2) the condition of the patient is not improving; or (3) dexamethasone has been administered, which may affect interpretability of the clinical response since this agent can suppress fever. Repeat lumbar puncture also should be considered in infants with meningitis caused by gram-negative enteric bacteria until sterility of the CSF is documented.
Once meningitis is suspected, antimicrobial therapy must be initiated promptly. Antibiotic therapy and other supportive measures must be started early when meningitis is suspected and should not be delayed while awaiting CSF results when clinical suspicion is high. Delays in antibiotic administration have been repeatedly shown to increase mortality, with delays as short as 3 hours associated with unfavorable outcomes. The risk for poor outcome increases by up to 30% per hour of antibiotic treatment delay.
Initial therapies for bacterial meningitis in children are based on achievable CSF concentrations of various antimicrobial agents and the likelihood of antimicrobial resistance among the most likely pathogens based on age and other clinical circumstances. Hospital antibiograms can serve as a partial guide to antimicrobial resistance considerations. Critical care clinicians should be familiar with recommendations for antibiotic management in bacterial meningitis.
eTables 67.1 to 67.3 list the most frequent meningeal pathogens and recommended antimicrobial regimens along with dosage information. The duration of antimicrobial therapy varies according to causative pathogen. Children with gram-negative bacterial meningitis require a minimum of 3 weeks of therapy. Classically, infants with GBS meningitis require 14 to 21 days; with S. pneumoniae , 10 to 14 days; with N. meningitidis , 7 days; and with H. influenzae , 7 to 10 days of therapy. For treatment of tuberculosis (TB) meningitis in children, up-to-date guidance from the Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) should be reviewed due to changing antimicrobial sensitivities.
|Age and/or Predisposing Condition||Pathogen||Antimicrobial Therapy|
|<1 mo||Escherichia coli, Streptococcus agalactiae, Klebsiella species, Listeria monocytogenes||Ampicillin plus cefotaxime a ± aminoglycoside|
|1–2 mo||S. pneumoniae, Neisseria meningitidis, S. agalactiae, Haemophilus influenzae, E. coli||Vancomycin plus third-generation cephalosporin b|
|2 mo–5 y||S. pneumoniae, N. meningitidis, H. influenzae||Vancomycin plus third-generation cephalosporin b|
|>5 y||S. pneumoniae, N. meningitidis||Vancomycin plus third-generation cephalosporin b|
|Humoral immunodeficiency, HIV, asplenia||S. pneumoniae, N. meningitidis, H. influenzae, Salmonella spp.||Vancomycin plus third-generation cephalosporin b|
|Complement deficiencies||N. meningitidis, S. pneumoniae||Vancomycin plus third-generation cephalosporin b|
|Basilar skull fractures||S. pneumoniae, N. meningitidis, Streptococcus pyogenes||Vancomycin plus third-generation cephalosporin b|
|Cerebrospinal fluid shunt–related||Coagulase-negative staphylococci, Staphylococcus aureus (methicillin-susceptible, methicillin-resistant), aerobic gram-negative bacilli (including Pseudomonas aeruginosa ), Propionibacterium acnes||Vancomycin plus ceftazidime or vancomycin plus cefepime or vancomycin plus meropenem c|
|Postneurosurgery||Coagulase-negative staphylococci, Staphylococcus aureus (methicillin-susceptible, methicillin-resistant), aerobic gram-negative bacilli (including Pseudomonas aeruginosa ), Propionibacterium acnes||Vancomycin plus ceftazidime or vancomycin plus cefepime or vancomycin plus meropenem c|
|Cochlear implants||S. pneumoniae||Vancomycin plus third-generation cephalosporin b|
|Streptococcus agalactiae||Penicillin G ± gentamicin or ampicillin ± gentamicin|
|Haemophilus influenzae , β-lactamase-negative||Ampicillin|
|H. influenzae , β-lactamase-positive||Third-generation cephalosporin a|
|Streptococcus pneumoniae , penicillin-susceptible||Penicillin G or ampicillin b|
|S. pneumoniae , penicillin-resistant, cephalosporin-susceptible||Third-generation cephalosporin a|
|S. pneumoniae , drug-resistant (multiply resistant)||Vancomycin plus third-generation cephalosporin a ± rifampin|
|Neisseria meningitidis||Third-generation cephalosporin a|
|N. meningitidis , penicillin-susceptible||Penicillin G or ampicillin b|
|Escherichia coli , other aerobic enteric gram-negative bacilli (not including Pseudomonas aeruginosa )||Cefotaxime or ceftriaxone (if >1 mo old) plus aminoglycoside c |
E. coli may produce extended-spectrum β-lactamases that require use of carbapenems (i.e., meropenem)
|Pseudomonas aeruginosa||Meropenem plus aminoglycoside or cefepime plus aminoglycoside or ceftazidime plus aminoglycoside d|
|Coagulase-negative staphylococci, methicillin-resistant Staphylococcus aureus||Vancomycin ± rifampin|
|Methicillin-susceptible S. aureus||Nafcillin|
|Antibiotic (Total Daily Dose)||Neonates 0–7 d, >2000 g a||Neonates >7 d, >2000 g||Infants and Children||Maximum Daily Adult Dose|
|Ampicillin (mg/kg/day)||300||300||300–400||12 g|
|Cefepime (mg/kg/day)||100||100||150||6 g|
|Cefotaxime (mg/kg/day)||150||200||200–300||12 g|
|Ceftriaxone (mg/kg/day)||—||—||100||4 g|
|Gentamicin (mg/kg/day)||5||7.5||7.5||5 mg/kg|
|Meropenem (mg/kg/day)||120||120||120||6 g|
|Penicillin G (U/kg/day)||300,000||400,000||250,000–400,000||24 million U|
|Rifampin (mg/kg/day)||—||10–20||10–20||600 mg|
|Vancomycin (mg/kg/day)||30||45||60||45 mg/kg|
In neonates, empiric therapy using a combination of ampicillin with either gentamicin or cefotaxime has been a longstanding recommendation. Currently, cefepime is often substituted for cefotaxime owing to a manufacturing-related shortage of the latter. Some have recently proposed that a third- or fourth-generation cephalosporin alone, without ampicillin, may be sufficient in this age group given the rarity of meningitis due to L. monocytogenes and enterococcal species in recent case series. , If more frequent expression of extended-spectrum β-lactamases emerges among E. coli and other gram-negative enteric bacteria at the population level over time, this may require consideration of inclusion of alternative agents such as carbapenems in empiric regimens for neonatal meningitis.
Because of the ongoing potential for reduced susceptibility to penicillin and cephalosporins among pneumococcal strains, empiric therapy beyond the neonatal period (or if S. pneumoniae is suspected or confirmed as the etiology in the neonate) should include vancomycin and a third-generation cephalosporin (i.e., ceftriaxone or cefotaxime). If the isolate is shown to be susceptible to penicillin, therapy may be completed with penicillin or a third-generation cephalosporin. If the isolate is nonsusceptible to penicillin but susceptible to third-generation cephalosporin, the latter should be continued. If the isolate is nonsusceptible both to penicillin and third-generation cephalosporins, vancomycin plus either ceftriaxone or cefotaxime should be continued. If the patient is not improving after 24 to 48 hours of therapy, follow-up CSF culture remains positive, or the microbe is fully resistant to third-generation cephalosporins, addition of rifampin should be considered if the microbe is susceptible to this agent.
Additional empiric agents could be added according to epidemiologic exposures, underlying conditions, and the presence of resistant organisms in the community or hospital. Once a specific pathogen has been identified and susceptibility data are available, antimicrobial therapy could be narrowed to specifically target the offending pathogen. Carbapenems may offer a treatment option in antibiotic-resistant strains causing meningitis or in cases of hypersensitivity to penicillins or cephalosporins. Meropenem is as effective as ceftriaxone or cefotaxime as empiric therapy for bacterial meningitis. It can be considered for use in nosocomial meningitis either as a single agent or in combination with vancomycin. Linezolid, a member of the oxazolidinone class, readily enters the CSF and is active against gram-positive organisms.
Accumulating evidence suggests that shorter durations may be effective and safe in some cases. A 2009 meta-analysis on duration of therapy in pediatric meningitis that focused on the three most common pathogens and compared short-course (4–7 days) versus long-course (7–14 days) therapy found no significant difference in clinical success or long-term neurologic complications. Subsequently, a multinational trial enrolled over 1000 children with meningitis caused by Hib, S. pneumonia , or N. meningitidis who were stable after 5 days of IV ceftriaxone therapy and randomized them to receive placebo or an additional 5 days of ceftriaxone. Again, there were no significant differences in bacteriologic failures, clinical failures, or clinical sequelae in survivors. Patients with persistent seizures, bacteremia, brain abscesses, other infections, or who were judged to be severely ill at the 5-day mark were excluded. A 2017 study of 263 children and adults with meningococcal meningitis found that a 4-day course was noninferior in all outcomes compared with 7 days.
The correction of hyponatremia and hypotension (key to maintaining adequate cerebral perfusion) when present and the treatment of cerebral edema may be just as important as the selection of an effective antimicrobial regimen. In patients with altered mental status, airway protection and assisted ventilation may be required. Patients who present in septic shock may require vasoactive-inotropic support. Anticonvulsant therapy is indicated in patients with seizures. Rapid placement of an external ventriculostomy may be indicated in patients with radiographic or clinical evidence of obstructive hydrocephalus with risk for impending herniation. Consultation with infectious disease and pharmacy colleagues is often valuable.
When a child with suspected bacterial meningitis is admitted, droplet precautions should be implemented immediately in addition to standard precautions until 24 hours of effective antimicrobial therapy against N. meningitis has been administered. , Droplet precautions are not required for confirmed pneumococcal infections.
In patients who present in shock or dehydration, administration of IV fluids is critical for the maintenance of normal blood pressure, adequate cerebral perfusion, and reducing the risk of central venous thrombosis. In the absence of shock and dehydration, there has been disagreement regarding IV fluid management in patients with meningitis. There are potential risks from giving too much fluid (especially brain swelling) as well as too little fluid (especially shock). In a recent Cochrane review, the authors evaluated differing volumes of fluid given in the initial management of bacterial meningitis in three pediatric trials, only one of which was considered high quality. They concluded that low-quality evidence showed no difference between maintenance versus restricted fluid regimens regarding death or severe acute neurologic complications. However, there was low-quality evidence supporting maintenance therapy over fluid restriction to reduce severe neurologic events at 3 months following infection.
The search for therapeutic adjuncts to improve outcomes in bacterial meningitis often has been hampered by the difficulty in conducting randomized clinical trials due to the acuity of the disease, narrow time frame in which adjuncts can influence outcome, and the increasingly small number of cases in the postvaccine era. Immunomodulation, hypothermia, free radical scavengers, C5 monoclonal antibodies, neutrophil proapoptosis agents, and TLR antagonists have all shown benefit in experimental models of meningitis, but no large clinical trials have been conducted to date.
The use of dexamethasone (DXM) in meningitis has been studied for over 3 decades. In early studies in children with Hib meningitis, DXM-treated children had a lower incidence of moderate-to-severe hearing deficits when compared with placebo. However, one of the study antimicrobial agents, cefuroxime, was later shown to be associated with increased neurologic sequelae. In adults with pneumococcal meningitis, the early administration of DXM, 15 to 20 minutes before or with first dose of antibiotic, was associated with a reduction in mortality and unfavorable outcomes. , There is no evidence that corticosteroid use is beneficial in neonatal bacterial meningitis.
A Cochrane database review of the effect of adjuvant corticosteroid therapy on meningitis found no effect on overall mortality. This study pooled 25 randomized clinical trials from adults and children. Subgroup analyses for causative organisms showed that corticosteroids reduced mortality in S. pneumoniae meningitis (a finding potentially driven by adult data) but not in meningitis caused by Hib or N. meningitidis . The rate of hearing loss was reduced in children with Hib meningitis but not when caused by other bacteria.
Patients with TB meningitis treated with steroids have improved survival compared with nonrecipients owing to reduction in cerebral vasculitis associated with this infection. However, the severity of neurologic disability observed with TB meningitis is not altered.
When DXM is administered, clinicians should be cognizant that fever associated with the infection may be suppressed and then rebound when the agent is discontinued. Such rebounds may occur in time frames when complications such as subdural effusions or empyema begin to emerge, which also can generate new or higher temperature elevations. Thus, beyond the apparent reduction in hearing loss associated with Hib meningitis when DXM is administered shortly before antimicrobial therapy, studies of adjuvant corticosteroid therapy in children have not unequivocally demonstrated beneficial effects. The decision to use DXM thus remains an individualized, case-by-case decision that involves balancing potential benefits and risks that are often difficult to quantify.
With aggressive support and timely and appropriate antimicrobial therapy, the overall mortality rate of bacterial meningitis remains low, at about 2% to 7% overall. Unfortunately, neurologic sequelae still occur in about 50% of survivors. , , Prognosis can vary by age (worse in young infants), progression of illness before effective antimicrobial therapy is initiated, causative microbe, bacterial burden in the CSF at the time of diagnosis, and host factors that may impair the immune response. Infants younger than 90 days may have sequelae in more than two-thirds of cases.
Overall, hearing loss is the most common long-term neurologic deficit associated with bacterial meningitis in children. Hearing loss is often associated with fibrosis and even ossification of the cochlea with destruction of the organ of Corti. Seizure disorders; motor deficits, including hemiparesis and hypertonicity; cognitive impairments of various types, including disorders of speech and language development; communicating and noncommunicating hydrocephalus; visual disturbances and other cranial nerve deficits; and ataxia can occur as a result of brain injuries from bacterial meningitis. Children without apparent permanent neurologic sequelae at discharge are at low risk for future epilepsy related to their meningitis.
When compared with other pathogens, pneumococcal, TB, and gram-negative enteric meningitis are most commonly associated with neurologic sequelae. Among the most common bacterial etiologies, S. pneumoniae is associated with highest mortality and with hearing loss rates of 20% to 30%. Antimicrobial resistance in pneumococcal meningitis does not appear to be associated with increased mortality or neurologic sequelae compared with strains susceptible to penicillin or ceftriaxone, though this may be due to routine use of vancomycin as part of initial empiric therapy.
Individual outcomes upon initial admission to the PICU are difficult to predict, and no single factor is entirely predictive. In a retrospective study of 15 PICUs in the United States, patients with both bacterial and nonbacterial meningitis had an unadjusted mortality rate of 7%. Nonsurvivors had a significantly higher Pediatric Risk of Mortality (PRISM) III score. Mortality was threefold higher in patients with bacterial-confirmed meningitis. The presence of coma upon initial presentation to the PICU was associated with 10-fold increased mortality. In another study evaluating only pneumococcal meningitis, the presence of shock, hyponatremia, or coma on admission to the ICU was associated with a higher use of invasive medical devices and higher mortality. A third pediatric study of all-cause meningitis reported risk factors associated with increased mortality, including low Glasgow Coma Scale score, respiratory distress at presentation, seizures at any point during admission, and—in resource-limited countries—malnutrition. Among laboratory findings, the presence of low WBC and platelet counts was associated with increased mortality. ,
The hippocampus, which is important to information processing, is frequently affected in bacterial meningitis. As such, meningitis survivors may demonstrate neuropsychological problems and cognitive issues later in life. An educational outcomes evaluation showed that children recovering from CNS infections significantly underperformed on neuropsychological evaluations in areas of memory function and teacher-rated academic performance in comparison with healthy controls.
The best means of prevention of bacterial meningitis is receipt of recommended vaccinations in childhood and adolescence, especially the recommended capsular polysaccharide-protein conjugate vaccines against Hib, pneumococci, and meningococci. Two protein antigen-based vaccines effective against many N. meningitidis serogroup B strains are available but are considered optional based on patient preference at this time. Vaccines against GBS have been developed; future maternal vaccination against GBS could further reduce early-onset and potentially impact late-onset GBS infections of all types.
Prevention also involves family members. Individuals (family or hospital staff) with significant, prolonged, or close exposures to children with N. meningitidis meningitis should receive antimicrobial prophylaxis with ceftriaxone, ciprofloxacin, or rifampin. Rifampin also is recommended for prophylaxis of close contacts of patients with Hib (or H. influenzae type a) who meet selected criteria.
A subdural empyema (SDE) is a purulent fluid collection outside the brain parenchyma contained between the dura and the arachnoid layers of the CNS. SDEs can occur anywhere in the subdural space, but the majority are in the supratentorial compartment. They are usually encapsulated and often loculated. Multiple predisposing etiologies are responsible, including extension of local infections (i.e., sinuses, otitis, mastoiditis) and hematologic spread from distant sites. In younger patients, an SDE frequently accompanies acute bacterial meningitis wherein the infection extends through the arachnoid and into the subdural space. In the pediatric population, extraaxial CNS infections are most common in two distinct age groups, corresponding to known etiologies. Subdural empyemas following meningitis are more common in infants, while sinus-related infections are most common in those older than 6 years.
In addition to meningitis, otorhinolaryngeal infections are an important predisposing factor for development of an SDE in older children. In a retrospective study conducted over 24 years at a single children’s hospital, 70 children developed SDEs. Sinusitis was the most common etiology, followed by bacterial meningitis. In this study, all patients were older than 7 years at diagnosis. However, another study that included infants demonstrated that only 10% of SDEs were related to otorhinolaryngeal infection. The drop in SDEs from sinus infections is presumed to be due to increased antibiotic use. In this study, meningitis, head trauma, and recent neurosurgery were identified as common predisposing factors for SDE development. A more recent pediatric prevalence study for SDE reported that neurosurgical procedures were the most commonly associated factor.
Initial symptoms may be vague, but a history of sinus infections is common. As an example, Pott puffy tumors are found frequently among patients with prior sinusitis and can have associated empyema development ( Fig. 67.1 ). Fever, headaches, eye pain, and altered level of consciousness are common presenting features in SDE. Prolonged fever (90%), seizures (70%), and focal neurologic signs (60%) are the most common clinical signs noted in children. Though often a source of fever, these fluid collections can be sterile. Focal neurologic deficits are frequently observed. Seizures and cerebral edema from cortical vein thrombosis are common complications among patients presenting with SDEs.
Before recent changes in vaccination strategies, SDE was often secondary to Hib. Currently, Streptococcus species (particularly viridans , pneumoniae , or milleri ) and S. aureus are the most common pathogens identified. , , The most common pathogen in infants less than 4 months old is GBS. Abscess fluid cultures are positive in almost two-thirds of cases, with ≈15% demonstrating polymicrobial growth.
The diagnosis of SDE is often made with radiologic assistance. Patients with sinus infections, progressive headaches, and new neurologic deficits should be evaluated with CT or MRI. It is often difficult to distinguish an SDE from a sterile reactive subdural effusion (RSE). SDEs can be differentiated from epidural empyemas, as they do not cross the midline and tend to layer along the convexity of the brain in a crescent shape. SDEs appear as isodense to low-dense extraaxial collections with rim enhancement on CT. MRI is superior to CT in the demonstration of both the extraaxial fluid and enhancement of the inner membrane with contrast medium ( Fig. 67.2 ). Diffusion-weighted imaging (DWI) is increasingly used to distinguish SDE from nonpurulent RSE. While SDEs appear bright with a low apparent diffusion coefficient (ADC), SDEs have a low signal and ADC values similar to CSF. Serial cranial ultrasounds may be considered for the evaluation of response to antibiotics in infants when repeated imaging is indicated.
The goal of treatment is evacuation of purulence and eradication of the source of infection. Medical management alone may be adequate if the empyema is small, but most cases require surgical evacuation of the purulent material. Craniotomy is the preferred procedure in most cases. Drainage through a burr hole may be considered in the unstable patient. In a retrospective review of 699 patients treated for SDE, performance of a craniotomy was associated with a mortality of 8% compared with 23% for patients who received burr hole drainage. In young infants, a subdural tap through the anterior fontanel may be considered for intracranial purulent exudate evacuation because of ease of access. Many of these patients required repeated surgical aspirations owing to empyema recurrence.
Following surgical drainage, broad-spectrum antibiotics should be initiated. A third- or fourth-generation cephalosporin plus vancomycin is a reasonable initial antibiotic option. Metronidazole may be added to cover anaerobic organisms, especially if an infected sinus is suspected as the source of the SDE. Duration of IV antibiotics varies from 2 to 8 weeks, with longer courses recommended for patients with bone involvement.
The keys to an optimal outcome are early accurate diagnosis, timely intervention, and appropriate antibiotic therapy. Among adults, reported mortality rates of SDE are ≈10% following meningitis and up to 20% following untreated sinus infections. Germiller and Sparano reported a 4% mortality rate in a pediatric series of intracranial suppurative infections. Survivors often had significant long-term morbidity with persistent seizures, hemiparesis, or residual neurologic deficits. Age, initial level of consciousness, initiation of treatment, and the rapidity of disease progression all influence outcome. The sequelae of SDE may be more severe in young infants than in any other age group.
The brain parenchyma itself is remarkably resistant to microbial infections. Despite the relative frequency of occult bacteremia in the pediatric patient, cerebral abscess formation is rare. Brain abscesses typically begin with a localized area of cerebritis, evolve through various stages, and finally organize into a collection of encapsulated purulent material. Among intracranial masses, the incidence of brain abscess is ≈8% in developing countries and ≈1% to 2% in Europe and North America. Most abscesses occur in the first 2 decades of life, but this associative factor is changing with improvement in treatment of sinus and otologic infections. Incidence of brain abscesses in children is 4 cases/1,000,000 with a peak age of presentation between 4 and 7 years. Certain medical conditions predispose patients for brain abscess formation. These include a history of congenital heart disease with intracardiac or intrapulmonary shunting, otologic infections, immunodeficiencies, prolonged corticosteroid use, diabetes mellitus, and alcoholism. At a single children’s hospital over a 19-year period, congenital heart disease and sinus/otologic infections were the most common predisposing factors to development of cerebral abscesses.
CNS abscesses can occur from a number of different etiologies and in different locations. In a large series examining brain abscesses in adults and children, almost 90% were in supratentorial locations. Associated infections or factors included otorhinogenic (38.6%), traumatic (32.8%), pulmonary (7.0%), cryptogenic (4.6%), postsurgical (3.2%), meningitis (2.8%), cardiac (2.7%), and “other” (8.6%). In children, the most common origin of brain abscesses is direct or indirect extension from the middle ear, paranasal sinus, or dental infections. Brain abscesses are sequelae in 6% to 8% of untreated sinusitis cases and up to 10% of mastoiditis cases. Abscesses in the temporal lobe or cerebellum are typically linked to ear or mastoid air cell points of entry and tend to be solitary. Frontal lobe abscesses are often due to ethmoid sinus or dental infections. Sphenoid sinusitis is associated with abscesses in the temporal lobes as well as the pituitary gland.
Abscesses caused by hematogenous origin are often noted in specific areas of the brain, especially in the middle cerebral artery distributions, including the parietal and occipital regions. These abscesses tend to be multiple and of varying sizes. The etiology of these abscesses is often from distant sources, such as endocarditis or pulmonary infections.
Not surprisingly, cyanotic heart disease increases the risk of brain abscess formation. , In one study, over 30% of patients with brain abscesses were found to have an underlying heart defect. Tetralogy of Fallot, transposition of the great vessels, atrial septal defects, ventricular septal defects, and pulmonary arteriovenous fistulas are all reported to predispose children to brain abscesses. This increased predisposition may be due to the presence of areas of brain ischemia caused by decreased arterial oxygen saturation and increased viscosity from an elevated hematocrit. Right-to-left shunting likely also predisposes to brain abscess formation, as the removal of organisms from the systemic circulation by the pulmonary capillaries is bypassed. Hematogenous spread can also occur from venous drainage into the cavernous sinus, resulting in frontal lobe abscesses that correlate with infections of the facial tissues or ethmoidal sinuses.
Extensions from cranial osteomyelitis, scalp infections, endocarditis, and meningitis are other known causes of CNS abscess formation. Several studies examined the etiology of brain abscesses in noncardiac patients. Antecedents include endocarditis in 10% of patients, bacteremia in 8%, immunodeficiency in 12%, pulmonary anomalies in 5%, and skin folliculitis in 3%. , Infants and toddlers are more susceptible than other age groups to brain abscesses arising from sequelae of bacterial meningitis or bacteremia.
Penetrating head trauma and neurosurgery represent a small proportion of the predisposing causes of CNS abscess, but the proportion is increasing, possibly as other predisposing factors such as middle ear infection become less important. In roughly one-quarter of brain abscess cases, no identifiable route or predisposing factors are identified.
CNS abscess formation can occur in any area of the brain. The most frequent areas are frontal, temporal, and frontal-parietal. Occipital lobes are least frequently involved. Histologically, brain abscesses begin as regions of cerebritis ( Fig. 67.3 ). They evolve into discrete collections of encapsulated pus over a series of stages (early cerebritis, late cerebritis, early capsular, and late capsular). This evolution takes 4 to 9 days as the center of the lesion undergoes liquefaction necrosis. By 10 to 14 days, a well-vascularized collagenous capsule with peripheral gliosis or fibrosis is typically present ( Fig. 67.4 ).
The clinical presentation of brain abscesses varies depending on the size, multiplicity, and location of the lesion. The initial presentation can be relatively nonspecific and occur due to focal mass expansion, intracranial hypertension, or destruction of the CNS parenchyma. Most patients are symptomatic during the early cerebritis phase. Symptoms include headache (70%), fever (60%), vomiting (50%), focal neurologic deficits (45%), and seizures (40%). , Clinical symptoms occur less commonly in pediatric cases. The triad of fever, headache, and focal deficits occurs in roughly one-third of patients, while meningeal signs occur in less than 25%. Over half of cases will have papilledema. The sudden worsening of a preexisting headache can indicate rupture of the brain abscess into the ventricular space or impending herniation from the lesion’s mass effect. Significant alteration in mental status is an ominous clinical finding. Abscesses located within their brainstem typically present with fever, headaches, hemiparesis, and focal cranial nerve findings involving CN III, CN VI, and CN VII.
Since the clinical presentation of brain abscess is often nonspecific, the diagnosis is often made by CT. Without imaging, it may be difficult to differentiate between brain abscess and other intracerebral pathologic processes. The increased use of CT has altered the prognosis of brain abscess by improving early diagnosis. Serial CT scanning may provide additional information about response to treatment. CT with IV contrast can be used for determining the size and number of abscesses but cannot consistently discriminate between metastases or primary brain tumors and brain abscesses. MRI is also a useful diagnostic tool and may have significant advantages in differentiating brain abscess from primary, cystic, or necrotic tumors using DWI and ADC images. MRI may also have a diagnostic advantage over CT by better identification of edema from liquefaction necrosis and greater sensitivity for early detection of satellite lesions ( Figs. 67.5 and 67.6 ). A 2014 meta-analysis of 11 studies demonstrated the value of DWI in brain abscess discrimination. DWI imaging provided both sensitivity and specificity of 95% in identification or brain abscesses.