Emerging Streptococcus pneumoniae Strains Colonizing the Nasopharynx in Children after 13-Valent (PCV13) Pneumococcal Conjugate Vaccination in Comparison to the 7-Valent (PCV7) Era, 2006-2015

Abstract

Background

After introduction of the 7-valent pneumococcal conjugate vaccine (PCV7) in the US in 2000, emergence of replacement serotypes occurred leading to the introduction of a 13-valent pneumococcal conjugate vaccine (PCV13) in 2010 that contained all PCV7 serotypes plus six additional serotypes (1, 3, 5, 6A, 7F & 19A). Here, we describe a 9 year prospective, longitudinal study characterizing S. pneumoniae (Spn) strains colonizing the nasopharynx (NP) of young children based on serotype, sequence type and antibiotic susceptibility during the PCV7 and PCV13 eras.

Methods

NP samples were obtained for pneumococcal identification from prospectively followed children at 6, 9, 12, 15, 18, 24 & 30 months of age. A total of 1072 visits during the PCV7 era (Jun 2006-Sep 2010) and 2044 visits during the PCV13 era (Oct 2010-Sep 2015) were included from 665 children. Serotyping and MLST (multi-locus sequence typing) types of Spn isolates were evaluated along with their antibiotic resistance pattern.

Results

A total of 1045 Spn were isolated; 350 during the PCV7 era and 685 during the PCV13 era. The most common serotypes identified during the PCV7 era were 19A and 23B compared with 35B, 23B and 21 in PCV13 era. Serotypes 15A/B/C emerged in equal proportion during the PCV13 era. Serotypes 16 and 20 were only observed in the PCV13 era. NP carriage of 19A persisted 5 years after PCV13 introduction (5% of all isolates). MLST sequence type (ST) 199 remained a dominant ST during both the PCV7 and PCV13 eras and ST558 and ST62 emerged following PCV13. Antibiotic resistance to penicillin, ceftriaxone, cefotaxime, erythromycin, tetracycline, and trimethoprim/sulfamethoxazole significantly decreased from the PCV7 to the PCV13 era.

Conclusion

Serotypes 35B, 23B, 21, and 15A/B/C rapidly emerged as NP colonizers in the early PCV13 era. Genetically divergent strains with ST558 and ST62 emerged. Resistance to common antibiotics declined following the introduction of PCV13.

Introduction

Streptococcus pneumoniae (Spn) Nasopharyngeal (NP) colonization occurs commonly in the first two years of life and the prevalence of carriage in children under 2 years of age varies from 30% to 60% (1, 2). There are 94 Spn serotypes and their frequency, in both NP colonization and invasive disease, differs by age, geographic region, and socioeconomic conditions (3-5). NP colonization is the first step in disease pathogenesis of Spn, leading to non-invasive (e.g. acute otitis media (AOM), sinusitis and non-bacteremic pneumonia) and invasive infections (e.g. bacteremic pneumonia, bacteremia and meningitis) (6). The introduction of the heptovalent pneumococcal conjugate vaccine (PCV7) represented an important step towards the prevention and control of invasive pneumococcal diseases (IPD) and non-invasive diseases associated with vaccine serotypes (7, 8).

Following the introduction of PCV7, there was an initial decrease in the incidence of NP colonization, IPD and AOM by vaccine serotypes; however, within 2 years there was an increase in NP colonization, IPD and AOM rates by non-PCV7 serotypes (9-12). Despite the change in the serotype distribution, the proportion of Spn NP colonization remained unchanged from the PCV7 into the PCV13 eras (13). Serotype 19A became the predominate Spn isolate in both NP colonization and disease, was frequently resistant to penicillin, and was associated mainly with sequence types (ST) 320 and ST199 indicate clonal expansion of serotype 19A (14-16).

In 2010, the PCV13 vaccine, including serotype 19A and five additional serotypes (1, 3, 5, 6A, 7F), was introduced in the US and several other countries. The present study reports on the changes in NP colonization in US children ages 6-30 months old following the introduction of PCV13. We describe the overall and serotype-specific trends in Spn colonization and compare them with the late PCV7 era in Rochester, NY, US. The antibiotic susceptibility pattern and molecular sequence typing patterns of the Spn serotypes to observe if particular clones are emerging, among Spn strains are compared.

Methods

Study population and sample processing

Subjects participated in our 9-year, prospective, longitudinal study of NP colonization and AOM in young children funded by the National Institutes of Deafness and Communication Disorders (R0108671) from Jun 2006 to Sep 2015. Children were enrolled from a middle-class, suburban socio-demographic population in Rochester, New York at 6 months of age and prospectively followed to 30 months of age. The details of the study design have been previously described (17).

Children received PCV7 vaccine from the introduction of study in 2006 until April 2010 and PCV13 vaccine after April 2010. Children received 3 doses of PCV vaccine according to the U.S. schedule at 2, 4, and 6 months of age; either PCV7 or PCV13 depending on the date of enrollment period, a booster dose was given at 15 months of age. Written informed consent was obtained from parents prior to enrollment in the study and the IRBs of the University of Rochester and the Rochester Regional Health System approved this study. In this study data obtained from children is compared between the PCV13 (Oct2010-Sep 2015) and the late PCV7 (2006-sept 2010).

Demographic data collected included daycare attendance, breast-feeding, sibling information, family history of AOM, and tobacco smoke exposure. NP wash and NP swab samples were collected at the well child visits at 6, 9, 12, 15, 18, 24, and 30 month of age and processed as previously described (16). Occasionally a child presented with AOM at the time of the well child visit; these visits were excluded in this analysis. Standard microbiology processing and identification techniques were used to process and identify the organism nasal wash cultures. The NP samples were inoculated on chocolate and blood agar plates, and Spn was identified first by colony morphology, α-hemolysis of blood agar plate, and optochin disc differentiation. Rough strains that demonstrate α -hemolysis and positive optochin zone were selected for further characterization. Serotypes were determined by Quellung reaction using Latex pool and serotype-specific pneumococcal antisera (Serum Staten Institute, Denmark) on pure cultures of Spn isolates.

Antibiotic Susceptibility

Oxacillin susceptibility of each isolate was determined using Sensi-Disc from Becton and Dickinson. The antibiotic susceptibility of Spn isolates to 16 different antibiotics including penicillin, ceftriaxone, trimethoprim-sulfamethoxazole, chloramphenicol, tetracycline, and erythromycin were determined with the VITEK 2 Gram Positive Susceptibility Card-AST-GP68 (BioMerieux, Inc.) in the clinical laboratories of Rochester General Hospital. Spn strain ATCC49619 was used as a control for each batch of antibiotic susceptibility testing. The numerical values for each antibiotic were expressed in μg/ml and Spn isolates were classified as susceptible, intermediate, or resistant based on current Clinical and Laboratory Standard Institute (CLSI) breakpoints for non-meningeal isolates (18).

Multilocus Sequence Typing (MLST)

Sequence types (STs) of some randomly picked Spn isolates from our population were determined using the MLST technique as described previously (19, 20). An eBURST analysis was performed to compare ST of Spn in the PCV7 vs. PCV13 eras as described previously (21, 22).

Statistics

The statistical analysis was performed on GraphPad Prism version 6. The statistical significance was assessed using Fisher's exact test with a significance level of p value 0.05.

Results

Study Population and Microbiology

Six hundred sixty-five subjects were enrolled and had study visits from June 2006 to September 2015. 52% of the children were male. Table 1 shows the demographic data of our study population.

Table 1.

Demographic information of Children at Enrollment

Number of Children	665
Sex
Male	52%
Female	48%
Race/Ethnicity
Caucasian	70.9%
African American	11.8%
Hispanic	3.9%
Asian	1.4%
Other	12%
Family history of AOM	49.6%
Exposed to tobacco smoke	12.7%
Day care attendance	43.9%
Breast-feeding >6 mo	27.2%

There were a total of 1072 well child visits during PCV7 period (Jul 2006-Sept 2010) and 2044 well child visits during the PCV13 period (Oct 2010-Sept 2015). Data during the transition period of vaccines from April-September 2010 were included in the PCV7 era. Table 2 shows the percentage of the three otopathogens isolated from NP samples of healthy children during PCV7 and PCV13 eras. Overall Spn colonization rates remained stable during the two time periods. There was a significant decrease in nontypeable H. influenzae (NTHi) NP colonization while NP colonization by Moraxella catarrhalis (Mcat) remained stable following the introduction of PCV13.

Table 2.

Comparison of otopathogens causing Nasopharyngeal Colonization in Healthy children during PCV7 vs. PCV13 eras

Bacterial Respiratory	PCV7 (n=1072)	PCV13 (n=2044)
S. pneumoniae	350 (32.7%)	685 (33.5%)
Oxacillin resistance	115 (32.8%)	177 (25.8%)*
Nontypeable H. influenzae	181(16.9%)	252(12.3%)***
β-lactamase positive	60 (33.2%)	87 (34.5%)
M. catarrhalis	403 (37.6%)	737 (36.1%)
β-lactamase positive	403 (100%)	737 (100%)

^*Significant decrease (p<0.05) in oxacillin susceptibility among strains between PCV7 vs. PCV13.

^***Significant decrease (p<0.001) in proportion difference among NTHi strains between PCV7 vs. PCV13.

S. pneumoniae Serotypes

Figure 1 shows the percentage distribution of the most commonly isolated Spn serotypes during the PCV7 and PCV13 eras. There was a significant decrease in vaccine serotypes 19A and 6A in the PCV13 era, p < 0.0001 and p< 0.001 respectively. Colonization of Spn serotype 3 did not differ between the PCV7 and PCV13 eras. Non-vaccine serotypes 35B, 35 and 16 increased significantly, p< 0.0001, p< 0.001 and p<0.05 respectively. Within serogroup 15 there were equal proportions of serotypes 15A, 15B and 15C. Serotypes 16 and 20 (>6 isolates each) were only observed in the post-PCV13 era.

Figure 1.

Comparison of most common serotypes of S. pneumoniae isolated from nasopharynx of healthy children in the PCV7 versus PCV13 eras.

****p<0.0001, *** p< 0.001 and * p<0.05.

Notes: Dotted line represent to separate vaccine types from non-vaccine types.

Y-axis is split at 15 to zoom in to show proportion of below 15% serotypes on the same graph.

Sequence types

A total of 289 Spn isolates from the PCV7 era and 503 isolates from the PCV13 era were characterized using MLST which represented typing of 59% of the isolates and 73% of the isolates from the PCV7 and the PCV13 eras, respectively. Table 3 shows the most common STs, and the associated serotypes observed in the PCV7 and PCV13 eras. ST 199, 558 and 62 were seen most often following PCV13 introduction. There was a significant increase (p<0.0001) in ST 558 associated with emergence of serotype 35B. Despite the significant decrease in serotype 19A, ST 199 remained unchanged due to the increase in serotypes 15A, B, and C. ST 320, associated with serotype 19A, significantly decreased following the introduction of PCV13 (p<0.0001). ST 36, 235, 448, 1840, 3280 and 3811 were only observed following introduction of PCV13.

Table 3.

Comparison of the most commonly observed sequence types (STs) and associated serotypes of Streptococcus pneumoniae in the PCV7 vs. PCV13 eras

Common Sequence types	# (%) of Spn MLST typed in PCV7 period (n=289)	# (%) of Spn MLST typed in PCV13 period (n=503)	Main Serotypes associated with ST
199	29 (10.0%)	42 (8.3%)	19A, 15B, 15A, 15C
558†	12 (4.2%)	66 (13.1%)†	35B, 35 (not B)
62	19 (6.6%)	44 (8.8%)	11A, 11B, 11C, 6A, 35, 9
432	7 (2.4%)	32 (6.4%)	21
439	6 (2.1%)	23 (4.6%)	23A, 23B
320†	21 (7.3%)	5 (1.0%)	19A
338	8 (2.8%)	20 (4.0%)	23A, 23B, 21
433	13 (4.5%)	16 (3.2%)	2211C
1262	6 (2.1%)	13 (2.6%)	15A, 15B, 15C
816	5 (1.7%)	7 (1.4%)	10, 39
695	7 (2.4%)	6 (1.2%)	19A
1847	10 (3.5%)	2 (0.4%)	23B, 23A
63	5 (1.7%)	6 (1.2%)	15A
2733	3 (1.0%)	8 (1.6%)	19A
473	3 (1.0%)	6 (1.2%)	6A, 6C, 21
1292	5 (1.7%)	3 (0.6%)	6A, 6C, 11C, 21
7927	4 (1.4%)	4 (0.8%)	21, 12
448*	0	7 (1.4%)	NT
498	2 (0.7%)	9 (1.8%)	35 (not B)
36*	0	5 (1.0%)	23B
1379	5 (1.7%)	1 (0.2%)	6A, 6C
3280*	0	5 (1.0%)	35B, 15B, 15C
1840*	0	6(1.2%)	23B, 15B
235*	0	5(1.0%)	20
3811*	0	13(2.5%)	15A

Note

^*represent STs only observed in PCV13 period.

^†represent significant difference (p<0.0001) between PCV7 and PCV13.

Figure 2 shows an eBURST analysis comparison of STs observed during the PCV7 and PCV13 eras. Each ST is represented by a point and its size is determined by the number of isolates. Groups of related STs are referred to as clonal complexes (CC). STs differing at a single allele are linked by straight line. Those STs, which cannot be linked to any other in the sample, are termed singletons. STs shown in pink were found in both the PCV7 and PCV13 eras, in black when found from the PCV7 era only and in green when found only during the PCV13 era.

Figure 2.

A Comparative eBURST diagram showing the combined NP colonized S. pneumoniae isolated from children during PCV7 and PCV13 period. The Sequence Type numbers are colored according to whether they are found only during PCV13 (green), during PCV7 (black) or in both (pink). Size of the point that represents each sequence type is determined by the number of isolates it contains. Sequence types differing at a single allele are linked by straight line. Those sequence types, which cannot be linked to any other in the population, are termed singletons.

Antibiotic susceptibility

There was a higher percentage of Spn isolates that were oxacillin resistant in PCV7 era and the percentage of β-lactamase producing NTHi and Mcat strains remained stable (Table 2). Table 4 shows the antibiotic susceptibility to 16 different antibiotics among 210 Spn isolates during the PCV7 and 438 Spn isolates during the PCV13 era. Among Spn isolates the prevalence of penicillin-nonsusceptibile strains decreased (33% vs. 26%, p<0.05) and the prevalence of penicillin-resistant strains decreased (13% vs. 4%, p<0.001) in the PCV13 era. Conversely, 56% of Spn isolates were penicillin susceptible (MIC value <=0.06 μg/ml) in the PCV7 era compared to 69% in the PCV13 era (p<0.001). Antibiotic resistance to Cefotaxime, Ceftriaxone, Erythromycin, Tetracycline and Trimethoprim/Sulfamethoxazole significantly decreased in the PCV13 era. Spn strains carrying multiple drug resistance significantly decreased (p<0.0001) in the PCV13 era; 10.5% Spn isolates exhibited resistance to ≥ 3 antibiotics among strains isolated in the PCV7 era compared to 2.8% in the PCV13 era (p<0.0001).

Table 4.

Comparison of antibiotic susceptibilities of the Streptococcus pneumoniae nasopharyngeal isolates during PCV7 versus PCV13 eras.

		PCV7			PCV13
Antibiotics	Antibiotics sensitivity range in μg/ml	Percentage of Spn isolates (n=210)			Prcentage Spn isolates (n=438)			p-value of change in Resistance
	(S; I; R)	S	I	R	S	I	R
Benzylpenicillin	<=.06;.12-1;>=2	55.7	31.4	12.9	68.5	27.2	4.3	<0.001
Amoxacillin /Calvulanic Acid	<=2;I4;R>=8	100.0	0.0	0.0	98.4	1.1	0.5	ns
Cefotaxime	<=.5;1;>=2	84.3	10.0	5.7	93.4	0.5	6.1	ns/0.01*
Ceftriaxone	<=1;2;>=4	84.7	8.6	6.7	96.8	2.3	0.9	<0.0001
Ertapenem	<=1;2;>=4	98.6	1.4	0.0	98.1	0.9	0.0	ns
Meropenem	<=.25;.5;>=1	84.8	12.9	2.4	87.8	10.1	2.1	ns
Levofloxacin	<=2;4;>=8	100.0	0.0	0.0	100.0	0.0	0.0	ns
Moxifloxacin	<=1;2;>=4	100.0	0.0	0.0	100.0	0.0	0.0	ns
Ofloxacin	<=2;4;>=8	100.0	0.0	0.0	99.6	0.2	0.2	ns
Erythromycin	<=.25;.5;>=1	70.0	0.0	30.0	84.2	0.0	15.8	<0.0001
Telithromycin	<=1;2;>=4	100.0	0.0	0.0	100.0	0.0	0.0	ns
Linezolid	<=2	100.0	-	0.0	100.0	-	0.0	ns
Vancomycin	<=1	100.0	-	0.0	98.6	-	1.4	ns
Tetracycline	<=2;4;>=8	80.5	0.5	19.0	94.1	0.9	5.0	<0.0001
Chloramphenicol	<=4S;>=8R	98.6	-	1.4	100.0	-	0.0	0.01
Trimethoprim/Sulfamethoxazole	<=9.5;19-38;>=76	73.3	4.3	22.4	87.2	6.1	6.6	<0.0001

S: sensitive; I: intermediate; R: resistant; ns: not significant

^*significant decrease in Cefotaxime Intermediate resistance level among strains between PCV7 vs. PCV13

Discussion

Pneumococcal conjugate vaccines containing 7, 10 and 13 serotypes are in use in multiple countries resulting in a decline in colonization and both invasive and non-invasive pneumococcal infections with vaccine serotypes. However, the emergence of replacement serotypes occurred (15, 23, 24) and continues to occur as shown from the results presented here. Our group has been prospectively monitoring the serotype distribution, molecular type diversity and antibiotic susceptibility of Spn strains since 1995 (14, 16, 25). Here we provided analysis of 9 years of results, 2006 – 2015, and showed the adaptability of the pneumococcus to respond to vaccine pressure.

A consistent observation among most studies [19-22], including ours, but not all studies [23, 24] is that PCVs do not result in an overall reduction in the frequency of Spn NP colonization. PCVs effectively eliminate all or nearly all of the strains expressing capsules corresponding to the polysaccharide-protein antigens in the vaccines. However, new serotypes quickly emerge to take the place of the strains eliminated.

We found a reduction in the NP colonization of vaccine serotype 19A after introduction of PCV13. Consistent with our results, Kaplan et al (26) and Bruce et al (9) have previously reported a decline in the prevalence of serotype 19A among children hospitalized with IPD after introduction of PCV13. However, we note that there is still a substantial frequency of Spn 19A colonization (5% of total Spn strains) in our population 5 years after PCV13 introduction. Olarte et al (27) found that 19A was still the most common serotype causing Spn meningitis in children during 2011-2013.

From an international perspective, another important serotype is 3. We did not observe a change in the frequency of serotype 3 NP colonization since introduction of PCV13. However, isolation of serotype 3 is relatively uncommon in our population so it is not possible to draw a conclusion of vaccine effectiveness on the reduction in NP colonization. Similarly, a study in France by Cohen et al (28) found no change in NP colonization by serotype 3.

We observed only 1 isolate of 7F and have not observed serotypes 1 and 5 in our population. Low detection rates of the serotypes 1, 3, 5 and 7F are expected irrespective of vaccination because they are rarely found in the NP, probably due to low density and short time of colonization (29).

Non-PCV13 serotypes accounted for 91% of the isolates in the PCV13 era, with 35B, 21 and 23B being the most commonly isolated (9-15% of total Spn isolates each). Other emergent serotypes of potential importance were non-vaccine serotypes 15A, 15B, 15C, 23A and 11A. Martin et al (30) recently reported the emergence of serogroup 15 and 35 but they did not subtype the strains. This is highly relevant as the serotypes 15A, 15B and 15C do not cross protect (31) and are observed in equal proportion in our population. The clinical relevance of serotypes 15B, 15C, 23A and 11 has been noted recently in invasive diseases (26). Among serogroup 35, 13% were 35B and 7% were either serotypes 35A/C/or F. Serotype 6C, which is not included in PCV13, accounted for ~2% of the total Spn isolates in our study during both the PCV7 and PCV13 eras. Some authorities hoped that inclusion of 6A in PCV13 might induce cross functional anti-6C opsonophagocytic responses and reduce 6C colonization (32) but this is not what we observed.

We found that the frequency of isolation of other potential respiratory bacterial pathogens, NTHi and Mcat remained stable across the PCV7 and PCV13 eras indicating that strains expressing non-vaccine Spn serotypes occupy the niche emptied by vaccine serotypes. The impact of age on Spn carriage and multicarriage analysis is not provided here and has been described previously (33).

Antibiotic resistance in the pnumococccal isolates decreased in the PCV13 era. Because serotype 19A accounted for the majority of antibiotic resistant strains during the PCV 7 era (15, 34) the decline in prevalence of serotype 19A was the major contributor to the significant decrease in nonsusceptible Spn strains during the PCV13 era. Our results differ from those of Lee et al (35), among children from Boston where they did not find a substantial change in the pattern of antibiotic nonsusceptibility in 2011, despite a modest decline in the rate of Spn serotype 19A carriage. Dagan et al (36) had similar findings among children in Israel.

Previous reports from our lab and other groups showed a predominance of ST320 and ST199 during the PCV7 era (14, 37). ST320 emerged as a predominant 19A clone following the introduction of PCV7 but was nearly eliminated following the change to PCV13. The Spn 19A serotypes we isolated during the PCV13 era were mainly ST199 and ST695. It is unclear why vaccination pressure by PCV13 appeared to eliminate clone ST320 but not ST199 or ST695. A newly predominant ST558 emerged in the PCV13 era associated with serotype 35B. ST558 was the main clone observed (13% of all isolates) in our population followed by ST199 (8%) and ST62 (9%) in the PCV13 era. This is the first report of the emergence of ST558 in the PCV13 era. Special note is taken of the emergence of serogroup 15 strains that share a ST199 origin. ST199 strains expressing a 19A capsule emerged in the PCV7 era to predominate and acquire antibiotic resistance. Additional studies are needed to see if strains expressing capsules of serogroup 15 become the next serotype 19A to cause invasive diseases.

Our study population is derived from a predominantly middle class, suburban population of children in upstate NY and may not be representative of other types of populations in the US or those in other countries. We showed that four years after the introduction of PCV13, Spn NP colonization rates have remained stable, there has been a reduction in vaccine serotypes 19A and 6A and an emergence in non-vaccine serotypes, 35B, 23B, 21, 15A, 15B and 15C. As occurred after introduction of PCV7, there was a decrease in antibiotic non-susceptible Spn strains with the elimination of 19A and 6A. The elimination of NP colonization of vaccine serotypes following PCV13 vaccination led to herd immunity and herd immunity extends the impact of PCV13 beyond vaccinated children to unvaccinated children and adults (38).