Hospital-based Surveillance Provides Insights Into the Etiology of Pediatric Bacterial Meningitis in Yaoundé, Cameroon, in the Post-Vaccine Era

Abstract

Background: Meningitis is endemic to regions of Cameroon outside the meningitis belt including the capital city, Yaoundé. Through surveillance, we studied the etiology and molecular epidemiology of pediatric bacterial meningitis in Yaoundé from 2010 to 2016.

Methods: Lumbar puncture was performed on 5958 suspected meningitis cases; 765 specimens were further tested by culture, latex agglutination, and/or polymerase chain reaction (PCR). Serotyping/grouping, antimicrobial susceptibility testing, and/or whole genome sequencing were performed where applicable.

Results: The leading pathogens detected among the 126 confirmed cases were Streptococcus pneumoniae (93 [73.8%]), Haemophilus influenzae (18 [14.3%]), and Neisseria meningitidis (15 [11.9%]). We identified more vaccine serotypes (19 [61%]) than nonvaccine serotypes (12 [39%]); however, in the latter years non-pneumococcal conjugate vaccine serotypes were more common. Whole genome data on 29 S. pneumoniae isolates identified related strains (<30 single-nucleotide polymorphism difference). All but 1 of the genomes harbored a resistance genotype to at least 1 antibiotic, and vaccine serotypes harbored more resistance genes than nonvaccine serotypes (P < .05). Of 9 cases of H. influenzae, 8 were type b (Hib) and 1 was type f. However, the cases of Hib were either in unvaccinated individuals or children who had not yet received all 3 doses. We were unable to serogroup the N. meningitidis cases by PCR.

Conclusions: Streptococcus pneumoniae remains a leading cause of pediatric bacterial meningitis, and nonvaccine serotypes may play a bigger role in disease etiology in the postvaccine era. There is evidence of Hib disease among children in Cameroon, which warrants further investigation.

Keywords: conjugate vaccine; genotyping; pediatric bacterial meningitis.

Figure 1.

Flowchart summarizing the number of patients, how many had specimens tested by each of the 3 bacteriological methods, and the outcome of the tests. Abbreviation: PCR, polymerase chain reaction.

Figure 2.

Bar chart showing the proportion of cases due to the 3 vaccine-preventable pathogens detected by a combination of culture, rapid tests, and polymerase chain reaction. Secondary axis shows the total number of patients recruited in each year. The Haemophilus influenzae type B vaccine was introduced in 2009 and the 13-valent pneumococcal conjugate vaccine was introduced in 2011.

Figure 3.

Distribution of the 13-valent pneumococcal conjugate vaccine (PCV13) and non-PCV13 serotypes over the surveillance period. A, Boxplot comparing the overall case counts of PCV13 and non–pneumococcal conjugate vaccine (PCV) serotypes based on the per annum counts. B, Stacked bar chart comparing the case counts per annum for PCV13 serotypes and non-PCV13 serotypes. C, Line graph of the proportion of PCV serotypes over time (years) with a nonlinear regression line of best fit.

Figure 4.

A, Maximum likelihood whole genome phylogenetic tree of Streptococcus pneumoniae isolates recovered from cerebrospinal fluid. B, Column graph comparing the number of resistance genotypes per genome among vaccine serotypes (VT) and nonvaccine serotypes (NVT). In the phylogeny, the serotypes in blue are nonvaccine types and black are vaccine types; earlier years are shaded in a brighter shade of green; resistance/intermediate resistance to chloramphenicol (C), trimethoprim-sulfamethoxazole (SXT), tetracycline (TE), oxacillin (OX), and erythromycin (E) is shown by a red box; and the presence of the antibiotic resistance genes catQ (chloramphenicol), folA/ folP (trimethoprim), tetM (tetracycline), penicillin-binding proteins (PBP; penicillin), and mef (erythromycin) is shown by a black circle. Blanks mean no data or not tested. Reference genome is shown on the tree branch with dashed line.