Evaluation of procedures for typing of group B Streptococcus: a retrospective study

Introduction

Streptococcus agalactiae (group B streptococcus, GBS) is a well-known pathogen primarily causing infections in newborns and the elderly (Brigtsen et al., 2015; Ballard et al., 2016). The disease in neonates is generally described as occurring in two different varieties (Bulkowstein et al., 2016). Early-onset disease (EOD) occurs in the neonate during the first six days of life, while late-onset disease (LOD) occurs later than the seventh day of life and can develop up to three months of age (Le Doare & Heath, 2013; Vinnemeier et al., 2015). Possible clinical manifestations of GBS infection in neonates are sepis, meningitis and pneumonia (Schrag & Verani, 2013). Among adults, GBS may also be associated with invasive infections, particularly in elderly persons with underlying medical conditions (Le Doare & Heath, 2013). Since introduction of screening programmes for pregnant women (Ballard et al., 2016) in some parts of the developed world, particularly early onset GBS infections in neonates has been reduced, and for many years the GBS disease incidence has been low (Heath, 2016). In contrast, early onset GBS infection is still a major problem in the developing world and presumably an underestimated problem (Heath, 2016). Also in recent years, the developed world has seen an increasing interest in incidence of invasive GBS infections, in particular among the elderly (Sheppard et al., 2016). Surveillance and identification of GBS in humans are therefore increasingly essential (Ballard et al., 2016; Sheppard et al., 2016).

The GBS are currently divided into ten serotypes based on type specific capsular antigens and are designated as Ia, Ib, II, III, IV, V, VI, VII, VIII, and IX (Slotved et al., 2007; Le Doare & Heath, 2013). For decades, the precipitation test also known as the Lancefield precipitation test (LP test) has been considered the standard method for GBS serotype determination (Slotved, Sauer & Konradsen, 2002). However, the method is time-consuming and therefore not suited for typing large numbers of isolates (Slotved et al., 2003). At present, GBS isolates are in general serotyped by the phenotypical method latex agglutination test (latex test) (Afshar et al., 2011), for which several kits are commercially available. Increasingly simpler and affordable molecular techniques for genotyping of GBS isolates, predominantly based on PCR assays, have been introduced and are now commonly used (Brigtsen et al., 2015; Sheppard et al., 2016).

In recent years, all GBS isolates received at the Statens Serum Institut (SSI) have been typed using both a LP test and a latex test (Slotved, Sauer & Konradsen, 2002; Lambertsen et al., 2010). Furthermore, some of the GBS NT isolates have been tested for genotype using the PCR assay described by Imperi et al. (2010) and Poyart et al. (2007).

By using all our GBS typing data from the period 2010 to 2014, we evaluated the phenotypical typing procedure for GBS isolates, based on the comparison of the commercial latex test and the LP test. We will furthermore show the genotype distribution of phenotypically NT GBS isolates.

Methods

This is a retrospective study based on typing data obtained in the period from 2010 to 2014 at the national Neisseria and Streptococcus Reference Laboratory (NSR), SSI.

The Danish hospitals are serviced by regional departments of clinical microbiology, all of which are public. On a voluntary basis, they submit isolates of beta-hemolytic streptococci to the NSR for national surveillance (Lambertsen et al., 2010).

Results

Discussion

In general, laboratories use latex tests, and to a minor extent LP test for GBS serotyping (Sheppard et al., 2016; Afshar et al., 2011). In this study, we found the latex test able to serotype 83.8% (95% CI [80.7–86.6]), while the LP test was able to serotype 87.5% (95% CI [84.6–90.0]) of the isolates (Table 1). Both assays provided serotype identification to some isolates that were non-typeable with the other method, although to a varying degree (Table 1). We did not find any conflicting test results using the two assays except for the NT isolates. A serotyping identification rate between 80% and 90% is common (Brigtsen et al., 2015), and even lower serotyping rates have been observed (Slotved et al., 2003; Yao et al., 2013). Because the latex test is much easier to perform than the LP test, there is no question on which phenotypical method to use for routine serotyping (Brigtsen et al., 2015; Sheppard et al., 2016). At the NSR laboratory (SSI) we have chosen the following procedure for phenotypical serotyping (Fig. 1): we start with the latex test, if this method provide serotype identification, then the identification is accepted, and no further serotype testing is performed. If the latex test shows either cross-reactions or non-typeability, then we proceed to test the isolate using the LP test, by first testing for 0.2N HCl and then if necessary proceed to 0.1N HCl. If the LP test provides a specific serotype, then this is accepted, or else the isolate is defined as non-typeable. Using this procedure provides a serotyping percentage of 89.3% (95% CI [86.6–91.6]) (Fig. 1).

In recent years several studies (Brigtsen et al., 2015; Sheppard et al., 2016) have presented molecular based methods for typing of GBS isolates, and a standard GBS PCR method has been described (Imperi et al., 2010). The most recent molecular GBS typing method described is the Whole-Genome Sequencing, which has the advantage, that besides providing information on the capsular genes, it also can provide information on multilocus sequence type, analyses of relatedness to other sequenced isolates, and detailed phylogenetic analyses (Sheppard et al., 2016).

Our PCR assay provided a genotype for all the 41 phenotypically NT isolates (Table 2). According to other studies, nearly 100% of GBS isolates can be genotyped by the use of molecular based methods (Brigtsen et al., 2015; Yao et al., 2013; Sheppard et al., 2016). This typing rate is much higher than the approximately 90% rate obtained by phenotypical assays (Brigtsen et al., 2015) (Table 1).

The two GBS vaccines currently under Phase 2 trials are based on capsular polysaccharide conjugated vaccines, while another vaccine based on GBS surface proteins is under Phase 1 trial (Heath, 2016). The capsular polysaccharide based vaccines cover either serotype III or serotype Ia, Ib and IIII (Heath, 2016). Evaluating the predicted vaccine coverage for Danish invasive GBS isolates from 2010 and 2011 in this study, we found that genotyping suggested an apparent increase in predicted vaccine coverage of 6.9% in 2010 and 4.4% in 2011. However, as these isolates were only typeable by molecular methods, the type identification represents lack of phenotypical expression and therefore possible lack of vaccine coverage.

In conclusion, in this study we found that the latex test and the LP test showed similar identification percentages. Because of the greater workload with LP test, we recommend this method to be used only for latex test NT isolates (Fig. 1).

Molecular typing methods are advantageous for the surveillance of GBS infections in terms of evaluating transmission chains and possible description of, e.g., early onset neonatal infections (Bergseng et al., 2009). In contrast, phenotypical methods must be applied when evaluating possible vaccine coverage or vaccine failures as well as planning of future capsular polysaccharide vaccines. Therefore, appropriate typing methods must be chosen according to the purpose of surveillance. Based on our findings, we suggest that general surveillance can be performed either by using the phenotypical procedure shown in Fig. 1 or by molecular techniques such as those described by Sheppard et al. (2016), depending on the laboratory capacity and cost. Coverage studies to provide data for possible future polysaccharide based GBS vaccines will require a phenotypical procedure. In case a polysaccharide GBS vaccine will be implemented in the future, phenotypical procedures will be necessary when evaluating vaccine failure in patients with infections caused by vaccine serotypes. Molecular techniques including Multilocus sequence typing (MLST) are necessary if information on clonal relation is needed, e.g., for identification of transmission chains in case of outbreaks or clustered infections.

Supplemental Information