To our knowledge this study provides the most in-depth insight of the serotype epidemiology of colonizing isolates in mother-newborn dyads and invasive GBS isolates in an industrializing country setting, and particularly from Africa. Although there are extensive data on maternal colonizing serotype distribution from industrialized countries 
, there are limited data comparing serotype distribution from colonized mothers and their newborns from industrializing countries 
. The findings from our study are similar to those reported elsewhere, 
and confirm that maternal GBS vaginal-colonization is commonly (52.5%) associated with infant colonization. The vertical transmission of GBS from mother to the newborn was corroborated by the high concordance of matching serotypes in the colonized mother-newborn dyads. Our study, however, also identified GBS in 113 (28.1%) of 402 colonized newborns born to mothers in whom GBS was not identified by vaginal swabbing.
The latter, as well as the incomplete, albeit high, concordance in serotypes distribution associated with colonization in mother-newborn dyads in our study may be due to the limitations of the study in the sensitivity of identifying maternal colonization, limitations in detecting concurrent multiple colonizing serotypes, or possibly acquisition of GBS from non-maternal sources during or soon after birth 
. The exclusion of rectal swabs to detect GBS colonization may have compromised the sensitivity in detecting GBS colonization 
. Previous studies have reported that the sensitivity of detecting GBS colonization in pregnant women by undertaking both rectal and vaginal swabs is 18.5%–51% higher compared to taking vaginal swabs alone 
and suggests that maternal rectal colonization may be important source for acquisition of GBS in newborns. Additionally, the yield of GBS on vaginal swabbing undertaken after the onset of labor may have been affected by draining liquor following the rupture of the placental membranes. Consequently, our study provides a minimal estimate as to the prevalence of GBS colonization in pregnant women during labor, as well as the proportion of colonized newborns who have acquired GBS from their mothers.
The serotype distribution identified in invasive isolates in our study are consistent with other smaller studies on invasive isolates elsewhere from Sub-Saharan Africa such as in Malawi 
, Zimbabwe 
, and an earlier study at the same centre as this study from 1997–1998. 
) The serotype data of invasive isolates from this study are nevertheless important as it confirms the absence of temporal variation in serotype distribution of invasive isolates over more than a decade in the study-setting. 
This has positive implications for the design of serotype-specific polysaccharide vaccines for prevention of newborn GBS sepsis.
Serotype III was responsible for 69.4% of invasive GBS disease among infants under the age of 90 days, which is in keeping with other studies from Sub Saharan Africa (56.0–58.3%) 
. Serotype Ia was the second most frequently identified invasive serotype in infants under 90 days of age in our study (18.8%), consistent with previous studies from South Africa (28.7%) and Malawi (21%) 
. These findings are also consistent with serotype distribution data of invasive isolates in young infants from industrialized countries (44% to 65% for serotype III and 33% to 15% for serotype Ia) 
. Overall, 85.4% and 98.2% of invasive isolates in children less than 7 days and those aged 7 to 90 days of age, respectively belonged to serotypes Ia, Ib or III, which are currently being considered as serotypes to be included in a polysaccharide-protein conjugate vaccine against GBS 
. Similarly a trivalent serotype-specific GBS vaccine including these serotypes would provide potential cover against at least 80% of invasive serotypes in young infants in USA and European countries 
Our data indicated that serotype III (OR
2.29) was the most invasive of all studied serotypes based on a method of analysis described by Brueggemann et al
. These data are also consistent with findings from Portugal (OR
, Sweden (OR
, and Israel (OR
where the same method was used. Due to this consistency we benchmarked invasive potential of other serotypes with type III as a referent serotype 
as summarized in . The data consistently reports that serotype Ib, II and V, are less invasive than serotype III. However, the invasive potential of serotype Ia is variable across different sites. In South Africa, Taiwan and The Netherlands 
, serotype Ia was less invasive than type III, although more invasive in Portugal 
, Israel 
and Sweden 
. The reasons for variation in invasive potential of serotype Ia reported at different sites are unclear, however data from Portugal suggests that the high invasive index of this serotype can be attributed to a dominant clone (ST-23 and ST-24), suggesting that the underlying genotype can influence the invasive potential.
Serotype V was identified in a low proportion of invasive isolates (5.6%) in our study, which is consistent with earlier studies from South Africa (5.6%) and Malawi (4.0%) 
. In the United States, serotype V is responsible for a high proportion of the invasive cases among infants (28%) 
. Similarly, a higher prevalence of serotype V has also been reported in studies from England and Wales (13%) 
.Recent identification of these historically uncommon circulating strains in both colonization and invasive disease isolates raises questions as to whether introduction of a vaccine composed of the most common invasive serotypes could cause a shift in serotype distribution in colonization, and thus disease. This warrants the need for ongoing international surveillance of invasive GBS serotypes in order to optimize GBS vaccine formulations.
Our study demonstrated differences between serotype epidemiology of isolates associated with invasive disease in newborns compared to maternal colonizing isolates. These findings corroborate the finding of other studies from industrialized country settings 
. Our study, however, clarified that the difference in serotype distribution between invasive and colonizing isolates is related to an increased invasive potential of certain GBS serotypes (i.e. III and Ia) rather than due to increased risk of acquisition of these by the newborn from the mother.
Despite the success of intrapartum antibiotic prophylaxis in reducing GBS associated neonatal sepsis in industrialized countries, the high costs and resource intensiveness is prohibitive in industrializing countries 
. Further limitations to this intervention strategy include concerns about the emergence of antimicrobial resistance in GBS, and a lack of efficacy in reducing GBS sepsis in infants older than 7 days of age 
. As a result, neonatal morbidity due to GBS remains a public-health problem. In Sub-Saharan Africa GBS neonatal disease occurs in 3.06 per 1,000 live births in South Africa 
, and 1.81 per 1,000 live births in Malawi 
. Consequently, the agenda of maternal immunization with GBS vaccines aimed at either reducing maternal recto-anal colonization, or preventing disease in the newborns through transplacental antibody transfer, needs to be explored in settings with a high burden of GBS disease.