Multilocus sequence types associated with neonatal group B streptococcal sepsis and meningitis in Canada

Abstract

Group B streptococci (GBS), a leading cause of neonatal sepsis and meningitis, are transferred to neonates from colonized mothers during childbirth. Prior studies using multilocus sequence typing (MLST) have found specific GBS clones (e.g., sequence type 17 [ST-17]) to be associated with neonatal disease in several geographic locations. Few population-based studies, however, have been conducted to determine the frequency of disease caused by specific GBS clones. MLST was used to assess the genetic diversity of 192 GBS strains from neonates and young children identified by population-based surveillance in Alberta, Canada, from 1993 to 2002. Comparisons were made to 232 GBS strains collected from colonized pregnant women, and all strains were characterized for one of nine capsule (cps) genotypes. A total of 47 STs were identified, and more than 80% of GBS strains were represented by 7 STs that have been shown to predominate in other populations. ST-17 and ST-19 were more prevalent in strains causing early onset disease (EOD) and late onset disease (LOD) than from pregnant women, whereas STs 1, 12, and 23 were more common in pregnant women. In addition, ST-17 strains and close relatives more frequently caused meningitis than sepsis and LOD versus EOD in this population of neonates. Further research is required to better understand why strains belonging to the ST-17 phylogenetic lineage are more likely to cause both LOD and meningitis and may provide clues into the pathogenesis of these conditions.

FIG. 1.

Phylogenetic relationships among MLSTs from 424 GBS strains recovered from either neonates with invasive disease, colonized pregnant women, or both. The consensus tree was constructed using the neighbor-joining algorithm based on the distance matrix of pairwise differences between STs. Only those relationships with node percentages of >70% bootstrap confidence values based on 1,000 replicates are shown. Each CC, with the exception of CC-19 (64%), comprises STs that cluster with a 70% or greater bootstrap confidence value. STs formerly published (22) as variants (v) were assigned new ST designations: ST-1v is ST-448; ST-19v is ST-447; and the six ST-23 variants are STs 441, 442, 443, 444, 445, and 446.

FIG. 2.

Phylogenetic network applied to 45 PI sites out of the 3,457 total nucleotide sites using the neighbor-net algorithm for 424 strains of GBS representing 47 STs. The colored circles mark the CCs identified in the neighbor-joining phylogeny. STs formerly published (22) as variants (v) were assigned new ST designations: ST-1v is now ST-448; ST-19v is now ST-447; and the six ST-23 variants are STs 441, 442, 443, 444, 445, and 446.

FIG. 3.

Distribution of GBS CCs as determined by MLST and frequency of cps genotypes among all 424 GBS strains from neonates with invasive disease (n = 192) and women colonized during pregnancy (n = 232). The CCs are ranked in order of decreasing overall frequency, and the frequency of cps genotypes is presented for both invasive and maternal colonizing strains. Singletons (S) refer to the STs that were not associated with a CC or GBS cluster. Footnote symbol a indicates that, together, the CC-19 and CC-17 invasive strains more frequently had cps3 than colonizing strains (χ², 360.71; df, 1; P < 0.0001). Footnote symbol b indicates that colonizing CC-1 strains more commonly had cps5 than invasive strains (χ², 305.81; df, 1; P < 0.0001).

FIG. 4.

Frequency of disease stratified by GBS CCs among invasive strains recovered from Alberta, Canada. The CCs are ranked in order of decreasing overall frequency, and the singleton STs represent clones not associated with any CCs. (A) Percentage of EOD and LOD cases by CC isolated from 181 of the 192 neonates with invasive disease. The 10 strains isolated from stillbirths (n = 6) and neonatal tissue (n = 4) were included in this analysis because the disease onset was known. However, the seven strains from children and the four strains identified retrospectively between 1993 and 1994 were omitted. (B) Disease severity by CC for 171 neonates with neonatal disease and GBS isolation from either blood (n = 152) or CSF (n = 19). In addition to the 7 strains from children and 4 strains from cases identified retrospectively, the 10 strains isolated from neonatal tissue and stillbirths were excluded to examine associations between disease type and CC.