Position statement
Posted: Oct 19, 2020
Nicole Le Saux; Canadian Paediatric Society. Updated by Nicole Le Saux, Infectious Diseases and Immunization Committee
The incidence of bacterial meningitis in infants and children has decreased since the routine use of conjugated vaccines targeting Haemophilus influenzae type b, Streptococcus pneumoniae, and Neisseria meningitidis. However, this infection continues to be associated with considerable mortality and morbidity if not treated effectively with empirical antimicrobial therapy. Diagnosis still rests on clinical signs and symptoms, and cerebrospinal fluid analysis. This position statement outlines the rationale for current recommended empirical therapy using a third-generation cephalosporin and vancomycin for suspected bacterial meningitis. It also provides new recommendations for the use of adjuvant corticosteroids in this setting. Once antibiotic susceptibilities of the pathogen are known, antimicrobials should be reviewed and modified accordingly. Recommendations for treatment duration as well as audiology testing are included. The present statement updates a Canadian Paediatric Society position statement on bacterial meningitis revised in 2008.
Keywords: Ampicillin; Antimicrobial resistance; Cephalosporin; Corticosteroids; Meningitis; PCV13; Vancomycin
The purpose of this statement is to review the current epidemiology of bacterial meningitis in children older than 2 months of age and provide guidelines for the empiric management of non-hospital-acquired suspected bacterial meningitis in previously healthy children in Canada. It does not address meningitis associated with cerebrospinal fluid (CSF) shunts or meningitis caused by organisms such as Escherichia coli and other enterobacteriaceae. Referral to other resources and, preferably, consultation with an infectious diseases specialist are recommended in such cases. Viral meningoencephalitis caused by herpes simplex or other viral pathogens is also beyond the scope of the present statement; however, this diagnosis should be considered in the proper clinical contexts. Importantly, meningitis caused by Mycobacterium tuberculosis (TB) can present with symptoms that are similar to other forms of bacterial meningitis, especially in areas with high TB rates. However, while diagnosis and management of TB meningitis are clinically distinct processes and morbidity remains high, this infection is also beyond the scope of this statement.
The epidemiology of meningitis in Canada has been influenced dramatically by universal immunization programs delivering conjugate vaccines for Haemophilus influenzae type b (Hib), Neisseria meningitidis, and Streptococcus pneumoniae [1][2]. The epidemiology of meningitis in the United States, and elsewhere where universal immunization programs are similar to those in Canada, is also evolving (see Figure 1 at www.nejm.org/doi/full/10.1056/NEJMoa1005384) [3]-[5]. The Netherlands reported a substantial drop in adult cases of community-acquired bacterial meningitis between 2006 and 2014 [6]. In the North American Arctic, including the Canadian Arctic, data from 2000 to 2010 indicated decreasing pneumococcal meningitis but increasing prevalence of H influenzae serogroup type a meningitis [7]. Despite limitations with reporting, there has also been an observed decrease in meningitis rates within the African meningitis belt due to improved access to vaccines [8].
In Canada, the Hib vaccine has been provided in public programs in all provinces and territories since 1998. Hib meningitis is now very rare and occurs primarily in unimmunized or partially immunized children, or in individuals who are immune-incompetent or immunosuppressed. It is worth noting that disease due to other serogroups (i.e., non-b) has been increasing in all parts of Canada, but particularly in Northern populations [9]-[11].
Publicly funded infant immunization programs using heptavalent conjugate vaccines against S pneumoniae (PCV7), which contained the capsular serotypes 4, 6B, 9V, 14, 18C, 19F, and 23F, were offered in all provinces and territories by 2005. The incidence of pneumococcal meningitis in the United States and Canada decreased significantly in all age groups following the introduction of PCV7 [12]-[16]. In Canada, the number of meningitis cases caused by S pneumoniae reported to Immunization Monitoring Program ACTive (IMPACT) hospitals decreased from 75 to 20 cases annually between 2000 and 2007, and there was an 87.5% decrease in cases of invasive pneumococcal disease (which includes meningitis and isolation of pneumococcus from normally sterile sites). However, the phenomenon of serotype replacement, with increases in the relative and absolute incidence of 19A, 15B, 6A, and other serotypes not present in PCV7, did occur here and elsewhere [1][3][13][15]-[18].
A 13-valent pneumococcal conjugate vaccine (PCV13) has now replaced the PCV7 vaccine. The PCV13 vaccine includes the seven serotypes in the PCV7 vaccine and an additional six serotypes (1, 3, 5, 6A, 7F, and 19A). In 2010, a significant proportion of invasive pneumococcal isolates from children <2 years of age were serotypes included in PCV13 and not in PCV7. As of 2011, all Canadian immunization programs had completed the conversion to PCV13. One recent study, using isolates collected from normally sterile clinical sites since 2010, determined that the PCV13 serotypes in Canada declined from 66% (224 of 339) to 41% (101 of 244; P<0.001) in children <5 years of age, and from 54% (1262 of 2360) to 43% (1006 of 2353; P<0.001) in children ≥5 years of age. Serotypes 19A, 7F, 3, and 22F were the most common serotypes in 2012, with 19A decreasing from 19% (521 of 2727) to 14% (364 of 2620; P<0.001) [19].
The incidence of meningococcal disease in children and adults has decreased significantly since the introduction of routine meningococcal serogroup C immunization programs [20]-[23]. The impact of the introduction of the quadrivalent conjugated A, C, Y, and W meningococcal vaccine for adolescents is not yet known because this vaccine is not part of publicly funded programs in all provinces and territories. Two vaccines that target serogroup B (Bexsero, (Novartis Canada) and Trumemba (Pfizer Canada)) are now licensed in Canada, though both are only publicly funded for persons at increased risk for meningococcal disease [24].
Meningitis caused by group B streptococcus (GBS, also referred to as Streptococcus agalactiae) remains an important cause of meningitis in infants up to 90 days old [25]. Although Listeria monocytogenes is an uncommon cause of meningitis beyond the neonatal period, it should be considered when specific host risk factors, such as immunosuppression, are present, or if brain stem infection is the initial presentation.
Meningitis susceptibility breakpoints should always be applied in the setting of presumed or confirmed meningitis, given the requirement for adequate drug levels in the central nervous system (CNS). S pneumoniae breakpoints have been specifically designed for interpretation in the context of meningitis. The organism is penicillin-susceptible if minimal inhibitory concentration (MIC) is ≤0.06 mcg/mL penicillin, and penicillin-resistant if MIC is ≥0.12 mcg/mL [26].
Using current criteria for antimicrobial susceptibility, the SAVE study analysed 6001 invasive isolates of S pneumoniae from adults and children in Canada [27]. There was a significant increasing prevalence of penicillin-susceptible isolates from 2011 to 2015. In 2015, using non-meningitis and meningitis breakpoints, 99% and 89.5% of isolates respectively were parenteral penicillin-susceptible, and 99.7% and 97.3% were susceptible to ceftriaxone, using similar infection site criteria [27]. The proportion of isolates represented in PCV10 or PCV13 vaccines, including 7F, and 19A, decreased over the same time period. Of serotypes, only 68.1% of 19A serogroup isolates were penicillin-susceptible using meningitis breakpoints. Of note, the rate of multidrug-resistant S pneumoniae (11.8%) in children 1-<2 years of age was higher than in other age groups, possibly due to selection pressure from higher rates of antibiotic use [28].
In the past several years, many countries, notably Belgium, Australia, and several countries in Latin America, have reported increasing prevalence (ranging from 30% to 80%) of N meningitidis with reduced susceptibility to penicillin [29]-[31]. In the United States, ciprofloxacin-resistant N meningitidis has also emerged [32]. One report from Ontario indicated that the percentage of strains with reduced susceptibility to penicillin between 2000 and 2006 was 21.7% [33]. Surveillance data of 408 Canadian isolates of N meningitidis analyzed at the National Microbiology Laboratory from 1996 to 2010 showed 18.6% with reduced susceptibility to penicillin, although no endemic isolates were resistant to ciprofloxacin (personal communication, Raymond Tsang, National Microbiology Laboratory (Winnipeg, Manitoba)). There is one report of an N meningitidis serogroup Y isolate in Canada with penicillin resistance [34].
While Hib is now an uncommon cause of meningitis in children, it as well as other H influenzae serotypes should still be considered in a child who is not fully immunized or unimmunized, or from an area in northern Canada with a higher incidence of invasive H influenzae. Increasingly, Hib and other typeable strains of H influenzae have shown increased beta-lactamase production, ranging from 4% to 42%, making these isolates resistant to ampicillin [35]. Because of this trend, ceftriaxone or cefotaxime should be used as empiric therapy, pending susceptibility testing.
Penicillin is currently the drug of choice for infection caused by group B streptococcus (S agalactiae). However, empirical coverage with cefotaxime or ceftriaxone in infants would be reasonable until culture results are available.
Current susceptibility data reaffirm the empiric management of suspected meningitis with ceftriaxone and vancomycin until susceptibility results are available.
Infants with meningitis often present with nonspecific findings of fever, poor feeding, lethargy (or decreased interaction with caregivers), vomiting, and irritability. They sometimes have a rash. Inconsolable crying, prolonged or worsening irritability, or progressive lethargy are also important clinical features that may indicate a CNS focus such as meningitis. Nuchal rigidity is uncommon in infants. Older children are more likely to have specific symptoms related to meningitis, such as headache, nuchal pain or rigidity, and impaired consciousness, as well as other nonspecific symptoms [36]. Patients should undergo a full examination, including respiratory status and detailed neurological examinations, to detect focal neurological signs, posturing, cranial nerve abnormalities, and assessment of level of consciousness.
A lumbar puncture (LP) for CSF analysis (cell count, glucose and protein levels, microbiological culture and molecular detection of bacterial DNA (if clinical suspicion is high and bacterial cultures are negative), and viral studies where appropriate, as well as consideration for specific testing for TB in high-risk children) is indispensable for the definitive diagnosis of meningitis. Multiplex polymerase chain reaction (PCR) testing is often helpful if cultures of blood or CSF are negative, and should be pursued with the microbiology laboratory when needed. An LP should always be attempted unless there are contraindications, such as coagulopathy, cutaneous lesions at the proposed puncture site, signs of herniation, or an unstable clinical status such as shock. If papilledema, new onset seizures, focal neurological deficits, or decreased level of consciousness or coma are present, an LP should be deferred until imaging (a contrast-enhanced computed tomography and/or magnetic resonance imaging of the head) is performed, and the risk of potential herniation is ruled out. Although there are no specific studies involving children, herniation following an LP in meningitis is rare in the absence of focal CNS lesions [37][38].
Because timely empiric antimicrobial therapy is critical to treatment, antimicrobial administration should not be delayed when imaging studies are not immediately available or an LP cannot be performed. Blood cultures using weight-based volumes should be obtained before starting antimicrobial therapy.
In general, in a setting of bacteremia, when there is either a clinical diagnosis of meningitis on presentation or the CSF parameters are consistent with meningitis (despite negative cultures), the patient should be managed as a bacterial meningitis case.
Other investigations, such as urine culture, pharyngeal culture, or chest radiograph, should be performed as clinically indicated.
Because the prognosis of meningitis depends on treating infection before clinically severe disease ensues, the timely administration of empirical antimicrobial therapy (Table 1) is critical. Antimicrobials should be administered without delay when meningitis is suspected or confirmed. Also, the careful, ongoing assessment and appropriate management of hemodynamic status is required. An LP should be performed to support the diagnosis, but if an LP is not possible, antimicrobials should be given empirically irrespective of the delay in obtaining an LP, and the patient should be transferred to a facility where an LP can be performed. One study involving adults showed that a delay in starting antimicrobial treatment was one of three independent variables associated with poor prognosis. The other two factors were the severity of clinical state at presentation and the isolation of non-penicillin-susceptible S pneumoniae [39][40].
Other factors to consider in the choice of antimicrobials are the child’s age and underlying diseases or risk factors such as immunodeficiency. For example, if there is an underlying immunodeficiency, then Listeria is a possible risk and ampicillin should be added to the empirical regimen. Management should also include monitoring for early complications associated with acute meningitis (e.g., syndrome of inappropriate antidiuretic hormone secretion and increased intracranial pressure).
The bacterial organisms most likely to cause community-acquired meningitis in healthy, immunized children >2 month of age are S pneumoniae and N meningitidis, but E coli and GBS should also be considered in infants up to 3 months of age. As mentioned previously, Hib is still occasionally observed in incompletely immunized patients, but other encapsulated H influenzae cases are being diagnosed with increasing frequency. In Canada, where penicillin-resistant S pneumoniae is known to occur, empiric therapy using a third-generation cephalosporin (ceftriaxone or cefotaxime) is recommended. In areas where there have not been cephalosporin-resistant S pneumoniae cases, this single drug may be adequate empiric therapy. However, pending culture results, most experts recommend adding vancomycin to the third-generation cephalosporin to protect against the possibility of a cephalosporin-resistant S pneumoniae, which has emerged in some parts of Canada. Third-generation cephalosporins also provide adequate empiric coverage for N meningitidis and H influenzae, because both organisms remain susceptible to these agents. If there are contraindications to third-generation cephalosporin use, other alternatives (such as meropenem) may be used and the early advice of an infectious disease expert should be requested.
Close contacts of patients with meningococcal disease should receive chemoprophylaxis. Hib chemoprophylaxis should be administered to all occupants of contact households with infants <12 months old (who have not completed the primary Hib immunization series), children <4 years old who are incompletely immunized, or immunocompromised children of any age. Any index case of Hib aged <2 years and not treated with cefotaxime or ceftriaxone should also receive chemoprophylaxis at the end of therapy. Public health should be consulted regarding management of possible contacts in child care and school.
In adults, multiple studies and meta-analysis have determined that adjuvant empiric steroids offered clinical benefit resulting in slightly lower mortality rates and reduction in hearing losses [41]-[43]. Studies in Sweden and Denmark, where adult patients who received steroids were compared with those who did not, concluded that there was likely some mortality benefit, although rapid diagnosis and treatment made historical comparisons difficult [42][43]. Based on the potential benefit and the low risk profile of 24 h to 48 h of initial dexamethasone therapy, the European guidelines recommend empiric dexamethasone for both adults and children with suspected or proven meningitis [44].
A landmark study in children who received empiric steroids in the management of acute bacterial meningitis due mainly to Hib showed that when they were administered just before or within 2 h of antimicrobials, there was a reduction in severe hearing loss (RR 0.34, 95% CI 0.20 to 0.59) [45]. Subsequent trials in children have focused mainly on pneumococcal meningitis, and all have been weakened by the heterogeneity of clinical severity at presentation, making conclusions regarding beneficial effects of steroid use in children much less definitive [41]. A descriptive study of children in the United States between 2011 and 2014 reported that only 8.1% of 6665 children received empiric steroids on admission with suspected meningitis or encephalitis [46].
Dexamethasone should be considered for infants and children with meningitis when CSF Gram stain testing shows Gram-negative coccobacilli consistent with H influenzae. If Hib is subsequently identified by molecular testing or cultured within 48 h, steroids should be continued for a total duration of 4 days. When Hib has not been positively identified within 48 h, steroids should be discontinued. Based on likely similar pathogenesis in adult meningitis, empiric steroids can be considered for infants and children with presumptive S pneumoniae meningitis. In this setting also, steroids should be discontinued if pneumococcus is not identified within 48 h.
The recommended dose of dexamethasone is 0.6 mg/kg/day in four divided doses administered every 6 h immediately before, concomitant with, or within 4 h of administering the first dose of antimicrobials. Clinical benefit is probably greater when steroids are received earlier within this 4 h window.
There is a higher risk for rebound of fever after steroids are discontinued, but if all other parameters indicate improvement and the clinical diagnosis continues to support bacterial meningitis alone, fever alone does not indicate need for additional testing.
There is insufficient information at the present time to recommend other types of adjuvant therapy.
When cultures and antimicrobial susceptibility data are available, therapy should be adjusted accordingly. As mentioned previously, an S pneumoniae isolate is considered susceptible to penicillin when the MIC is ≤0.06 mcg/mL. However, an isolate is considered susceptible to cefotaxime or ceftriaxone when the MIC is ≤0.5 mcg/mL, intermediately susceptible when MIC is 1.0 mcg/mL, and resistant when MIC is ≥2.0 mcg/mL [26]. Vancomycin is active against cefotaxime- or ceftriaxone-resistant strains. Treatment should be modified according to Table 1, depending on the results of the CSF culture and sensitivity. See Table 2 for dosage recommendations for antimicrobial agents.
Generally, repeat CSF sampling is not required in the context of common pathogens, unless a child does not clinically improve with initial therapy. When meningitis is caused by S pneumoniae, repeat CSF sampling at 48 h may be considered if the patient has received dexamethasone or if S pneumoniae is resistant to penicillin or cefotaxime/ceftriaxone. For meningitis due to GBS, some experts recommend documentation of CSF sterilization at 24 h to 48 h after initiation of therapy [47]. Although not discussed in this statement, repeat CSF culture at 24 h to 48 h is recommended for meningitis caused by Gram-negative enteric pathogens (e.g., E coli). CNS imaging is recommended when there is failure of sterilization of CSF, or if neurological symptoms or other specific complications develop during the course of treatment.
Treatment of bacterial meningitis should always be with intravenous antimicrobials to achieve high CSF levels. The recommended length of treatment varies with the pathogen and the clinical course of infection. Recommended length of therapy for uncomplicated meningitis due to S pneumoniae is 10 to 14 days; due to Hib, 7 to 10 days; and due to N meningitidis, 5 to 7 days. Recommended therapy for uncomplicated GBS meningitis is 14 to 21 days, and may be longer if cerebritis or ventriculitis is present.
Formal audiology assessment should be performed as soon as possible after diagnosis of meningitis for all children affected (and always before discharge from hospital) to optimize management in the event of hearing loss [48].
This statement has been reviewed by the Acute Care and Community Paediatrics Committees of the Canadian Paediatric Society.
Members: Michelle Barton-Forbes MD; Sean Bitnun MD; Natalie A Bridger MD (past member); Shalini Desai MD (past member); Michael Forrester MD; Ruth Grimes MD (Board Representative); Nicole Le Saux (past Chair); Jane C McDonald MD; Heather Onyett MD; Laura Sauve MD (Chair); Marina I Salvadori MD (past member); Karina Top MD
Consultant: Noni E MacDonald MD
Liaisons: Upton D Allen MBBS, Canadian Pediatric and Perinatal HIV/AIDS Research Group; Toby Audcent MD, Committee to Advise on Tropical Medicine and Travel (CATMAT), Public Health Agency of Canada; Carrie Byington MD, Committee on Infectious Diseases, American Academy of Pediatrics; Fahamia Koudra MD, College of Family Physicians of Canada; Marc Lebel MD, Immunization Monitoring Program, ACTive (IMPACT); Yvonne Maldonado MD, Committee on Infectious Diseases, American Academy of Pediatrics; Jane McDonald MD, Association of Medical Microbiology and Infectious Disease Canada; Dorothy L Moore MD, National Advisory Committee on Immunization (NACI); Howard Njoo MD, Public Health Agency of Canada
Principal author: Nicole Le Saux MD
Updated by: Nicole Le Saux MD
Disclaimer: The recommendations in this position statement do not indicate an exclusive course of treatment or procedure to be followed. Variations, taking into account individual circumstances, may be appropriate. Internet addresses are current at time of publication.
Last updated: May 27, 2021