Can Duration of Antibiotic Therapy for Bacterial Meningitis in Infants Be Shortened?

In this issue of Pediatrics, Van Hentenryck et al¹  report on their systematic review of publications assessing the relationship between the duration of the administration of parenteral antibiotics and clinical outcomes of infants with bacterial meningitis who are younger than 3 months of age. They found that the number and quality of publications related to their aim were very limited and thus they could not perform a standard quantitative meta-analysis. The authors describe 2 studies, including 1 randomized control trial, in detail and concluded that “longer courses were not associated with improved outcomes.” However, both studies were very problematic in that the majority of the patients did not have culture-proven bacterial meningitis. In the end, Van Hentenryck et al concluded there was a “pressing need for comparative effectiveness research to establish the shortest safe duration of antibiotic therapy in this population.”

The standard recommendations for the duration of antibiotic treatment of bacterial meningitis in children endorsed by experts in the field have been the same for more than 40 years.^2–4  In all cases, the duration is ultimately based on how the child has responded to therapy, with a normalization in body temperature for several days as well as overall improvement in neurologic status. The duration of fever in some children may be quite prolonged, leading to a longer treatment duration. In many instances, the cause of the prolonged fever is not determined and is ultimately thought to be associated with differences in host response. In our multicenter study of 173 children with pneumococcal meningitis, the median duration of fever after admission was 2 to 3 days with an interquartile range of 1 to 5 days, meaning that 25% of patients had fever for more than 5 days.⁵  In a multicenter study of 159 episodes of meningococcal infections in children (70% meningitis), the mean duration of fever after administration of antibiotics was 1.9 days with a range of 0 to 14 days.⁶  Furthermore, the development of a complication such as subdural empyema demonstrated on neuroimaging will prolong antibiotic treatment.

In the patient who has an uncomplicated course, experts recommend 5 to 7 days for meningococcal meningitis, 10 to 14 days for pneumococcal meningitis and 7 to 10 days for meningitis resulting from Haemophilus influenzae type b.^3,4,7  For the neonate and younger infant, 14 days of antibiotics are recommended for treatment of meningitis resulting from Group B Streptococcus (GBS).⁷ 

The recommended duration of antibiotic administration for treating meningitis resulting from Gram-negative enteric bacteria in the neonate and younger infant is at least 3 weeks or 2 weeks beyond the first sterile repeat cerebrospinal fluid (CSF) culture, whichever is longer. The leading expert on the antibiotic treatment of Gram-negative neonatal meningitis, George McCracken Jr, MD, put forth this particular recommendation in 1974,⁸  and it is still the standard recommendation in official guidelines.⁴  I asked Dr McCracken how he came up with these recommendations and he responded, “If I recall correctly, we had several failures with 14 days of therapy with Gram-negative meningitis in neonates and chose 21 days, which resulted in 100% bacteriologic cure. That was before the availability of more active drugs like cefotaxime, etc. I am not sure I would shorten the duration very much since outcome with the newer drugs can also be severe.” Indeed, in 1972, Dr McCracken published his experience with the time to sterilization of repeat CSF cultures in a group of neonates with Gram-negative meningitis.⁹  Of 19 infants with Gram-negative meningitis and follow-up CSF cultures (majority with Escherichia coli), only 50% had sterile CSF cultures after 7 days and 75% after 11 days of therapy.

Michael Radetsky outlined the history of the recommendations for duration of treatment of bacterial meningitis in children in a very well-documented and well-crafted review published in 1990.¹⁰  He concluded that “the standards for duration of treatment in meningitis have been more the distillations of clinical experiences than the fruit of scientific study. Therefore as such they ought to constitute general guidelines rather than absolute standards. Because it is historically valid to say that patient response has remained the ultimate bioassay of the therapy, there is no demanding reason to treat meningitis for any predetermined length of time. What emerges instead is the concept of a range of treatment durations. This range has not been established by experimentation.”

It will be interesting to determine if comparative effectiveness studies can actually be conducted. In the United States, the incidence of early-onset (0-6 days) and late-onset (7-89 days) GBS infection is 0.25 and 0.28 cases per 1000 live births, respectively; up to 10% of early-onset GBS cases and up to 30% of late-onset GBS cases are meningitis, respectively.¹¹  In 2021, there were almost 3.7 million live births. Thus, for bacterial meningitis caused by GBS, there are approximately 400 cases a year in the United States in the first 3 months of life. So, over several years using national inpatient databases, a large number of uncomplicated GBS meningitis cases likely could be included in a comparative effectiveness study of the duration of antibiotic treatment and outcome. This assumes that uncomplicated cases can be identified from these databases and that a substantial number of physicians do not follow the 14-day treatment duration recommended by the 2021 through 2024 Red Book or other authoritative sources. For meningitis resulting from Streptococcus pneumoniae in this age group, the numbers are likely to be substantially fewer than for GBS meningitis. However, because the recommendations for treating uncomplicated pneumococcal meningitis generally are for 10 to 14 days, an analysis of outcome based on duration may be possible if enough cases can be identified. Over the years 2015 to 2021, 31 cases of pneumococcal meningitis in infants 0 through 90 days of age were seen at 8 United States children’s hospitals in the multicenter pediatric pneumococcal surveillance study (K.G. Hulten, PhD, personal communication, 2022). For Gram-negative meningitis in this age group, even if enough cases can be identified, many challenges exist in trying to analyze various durations of treatment considering the many different enteric species that can cause meningitis, varying gestational and postnatal age of the infant, increasing antibiotic resistance of the isolates, and results of neuroimaging, and assuming an adequate number of these infants are not treated for the recommended minimum of 21 days. Treatment in these infants beyond 21 days is often the result of significant abnormalities on neuroimaging.

It may be that the optimal population for these investigators to assess is the group of infants they are perhaps most concerned about: febrile infants with unsuccessful or uninterpretable CSF results and negative blood and CSF cultures and, I would add, negative viral testing. It is likely there is much greater heterogeneity across institutions in the duration of antibiotics administered to these infants than there is in young infants with culture proven bacterial meningitis.

CSF

cerebrospinal fluid

GBS

group B Streptococcus