Introduction

Streptococcus pneumoniae remains a leading cause of morbidity and mortality worldwide, particularly among children under five years of age, despite substantial advances in prevention and treatment strategies [1]. The human nasopharynx serves as the primary reservoir for S. pneumoniae, playing a critical role in both disease development and transmission within the community [2–4]. Pneumococcal carriage is influenced by several host and environmental factors, including age, daycare attendance, number of siblings, and vaccination status [5,6]. As a result, isolates colonizing the nasopharynx often reflect those circulating strains in the community, some of which may also cause invasive infections.

Over recent decades, the development and global implementation of pneumococcal conjugate vaccines (PCVs) have significantly reduced the burden of invasive pneumococcal disease (IPD) related with the specific serotypes included in the vaccine. These vaccines are highly effective not only in preventing IPD caused by vaccine serotypes but also in reducing nasopharyngeal (NP) carriage of these serotypes [7,8]. Currently, five PCVs are available worldwide for the prevention of IPD in children. These include the 10-valent vaccine (PCV10, Synflorix®, GSK), which targets the serotypes: 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F, and 23F; the 13-valent vaccine (PCV13, Prevenar13®, Pfizer), which includes all PCV10-GSK serotypes plus serotypes 3, 6A, and 19A; the 10-valent Serum Institute of India vaccine (PCV10-SII PNEUMOSIL®), covering the serotypes:1, 5, 6A, 6B, 7F, 9V, 14, 19A, 19F, and 23F; the 15-valent vaccine (PCV15, Vaxneuvance®, Merck Sharp & Dohme Corp.), which adds 22F and 33F to the PCV13 formulation; and the 20-valent vaccine (PCV20, Prevenar 20, Pfizer Inc.), which expands PVC15 with serotypes 8, 10A, 11A, 12F, and 15B [9,10].

Brazil introduced PCV10-GSK (Synflorix®) into its national childhood immunization program in March 2010. The current schedule recommends a two-dose primary series at 2 and 4 months of age, followed by a booster dose at 12 months. As of 2023, vaccination coverage among children under one year of age reached 83.4% [11]. Several studies have documented the vaccine’s impact on reducing IPD caused by PCV10-GSK serotypes, both in the short and long term. Data from Brazil’s national laboratory-based surveillance system showed an 85.6% reduction in IPD caused by PCV10-GSK serotypes among children aged two months to five years within five years of vaccine introduction [12]. Another study, over a nine-year post-vaccine period, reported an 84.7% decline in overall IPD cases and a 98.0% reduction in IPD due to PCV10-GSK serotypes [13].

Despite this success, data on the impact of PCV10-GSK vaccine on NP carriage remain limited, especially outside Brazil’s Southeast region. Studies from that region have demonstrated a marked decline in vaccine-type carriage, alongside an increase in non-vaccine serotypes, particularly serotypes 6C and 19A, which are frequently associated with resistance to penicillin and ceftriaxone [14–16].

Updated data on pneumococcal carriage in the context of a mature PCV10-GSK immunization program are crucial to inform policy decisions regarding the adoption of newer or higher-valency PCVs. This study aimed to estimate the prevalence of pneumococcal NP carriage among children under five years of age in Salvador, the fifth most populous city in Brazil, located in the country’s Northeast region [17], and to identify factors associated with overall pneumococcal carriage and given the near elimination of PCV10-GSK serotypes, carriage of serotypes not covered by PCV20, which currently predominate in the study population. These findings will contribute to evidence-based policymaking regarding the potential incorporation of expanded-valency pneumococcal conjugate vaccines in Brazil.

Materials and methods

Results

Between August and November 2023, all 10 pre-selected schools were visited, and a total of 400 eligible children were enrolled in the study. Participant characteristics are summarized in Table 1. Children aged two to three years represented 16.8% of the total, while those aged three to four and four to five years accounted for 42.8% and 40.5%, respectively. Overall, 50.8% of the participants were male. Regarding race, 16.5% identified as white, 33.3% as black, 50.3% as mixed race. Respiratory tract infections in the preceding week were reported for 58.0% of children, and 22.2% had received antimicrobial treatment in the past month. Vaccination records indicated that 92.0% (368 out of 400) had completed PCV10-GSK Brazilian immunization schedule of two doses plus a booster, while 8.0% had incomplete vaccination. Additionally, 1.5% had received PCV13.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and clinical characteristics of the study population according to Streptococcus pneumoniae nasopharyngeal carriage status (n = 400).

https://doi.org/10.1371/journal.pone.0357903.t001

The overall NP carriage rate was 39.5% (158 out of 400), yielding 162 pneumococcal isolates due to co-colonization detected in four children. Recent antibiotic use, number of household contacts, household size, exposure to tobacco smoke, contact with other children under five years of age, and complete PCV10-GSK vaccination was not significantly associated with pneumococcal carriage. White race was independently associated with lower odds of pneumococcal carriage compared with mixed race in the main multivariable model (adjusted OR = 0.47, 95% CI: 0.25–0.87, p = 0.02), and this association remained similar in the sensitivity analysis using a mixed-effects logistic regression model including school as a random effect (S1 Table). Living with elderly individuals was also associated with lower odds of carriage in the main model (adjusted OR = 0.48, 95% CI: 0.24–0.90, p = 0.03), although this association was attenuated when school was included as a random effect.

The number of enrolled children was relatively balanced across the 10 participating schools, ranging from 32 to 56 children per school (S2 Table). Pneumococcal carriage prevalence varied across schools, ranging from 18.2% to 64.3%. In a sensitivity analysis, the inverse association between white race and pneumococcal carriage remained similar, whereas the association with living with elderly individuals was attenuated (S1 Table).

Among the 158 children colonized with S. pneumoniae, 64 (40.5%) carried at least one PCV20 serotype, whereas 94 (59.5%) carried only non-PCV20 serotypes and/or non-typeable isolates (Table 2). In both the univariate and multivariable analyses, no demographic, clinical, household, or vaccination-related factor was significantly associated with carriage of non-PCV20 serotypes. In the adjusted model, children aged 4–5 years had higher odds of carrying non-PCV20 serotypes than those aged 2–3 years, although this association did not reach statistical significance (adjusted OR = 2.54, 95% CI: 0.92–7.22; p = 0.07). Similarly, living with another child younger than five years of age was positively associated with carriage of non-PCV20 serotypes, but the association was not statistically significant (adjusted OR = 1.82, 95% CI: 0.90–3.80; p = 0.10).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Factors associated with carriage of non-PCV20 serotypes among children colonized with Streptococcus pneumoniae (n = 158).

https://doi.org/10.1371/journal.pone.0357903.t002

The most common S. pneumoniae serotypes identified among carriers were 6C (17.9%, 29/162), 19A (13.0%, 21/162), 11A (9.3%, 15/162), 15B (8.6%, 14/162), 23A (8.6%, 14/162), and 15A (7.4%, 12/162) (Fig 1). Among the PCV10-GSK vaccine serotypes, 14 (n = 3) and 19F (n = 2) were detected, corresponding to 3.2% of colonized children. Overall, 17.3% (28/162) of circulating serotypes were represented in PCV13, PCV15, or PCV10-SII, whereas 39.5% (64/162) corresponded to serotypes included in PCV20. Co-colonization, presence of more than one pneumococcus serotype, were identified in 2.5% (4/158) of the pneumococcal carriers. The dual-serotype combinations detected were 19A/6C, 19A/20, 14/23A, and 23B/6C.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Serotype distribution of Streptococcus pneumoniae in the nasopharynx of children aged 2 to 5 years in Salvador, Brazil.

PCV10:10-valent vaccine (PCV10, Synflorix®, GSK), which targets the serotypes: 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F, and 23F; PCV13: the 13-valent vaccine (PCV13, Prevenar13®, Pfizer), which includes all PCV10-GSK serotypes plus serotypes 3, 6A, and 19A; PCV10-SII: the 10-valent Serum Institute of India vaccine (PCV10-SII PNEUMOSIL®), covering the serotypes:1, 5, 6A, 6B, 7F, 9V, 14, 19A, 19F, and 23F; PCV15: the 15-valent vaccine (PCV15, Vaxneuvance®, Merck Sharp & Dohme Corp.), which adds 22F and 33F to the PCV13 formulation; PCV20: the 20-valent vaccine (PCV20, Prevenar 20, Pfizer Inc.), which expands PVC15 with serotypes 8, 10A, 11A, 12F, and 15B.

https://doi.org/10.1371/journal.pone.0357903.g001

Antimicrobial susceptibility was available for 160/162 isolates. Oxacillin disk diffusion screening test identified 68 isolates (42.5%) as non-susceptible to penicillin (zone diameter < 20 mm). Among these, seven isolates (4.4%, 7/160) exhibited penicillin resistance (MIC > 2.0 μg/mL), and 28 isolates (17.5%) showed intermediate resistance (MIC 0.125 – < 2.0 μg/mL). Additionally, two isolates (1.3%, 2/160) had ceftriaxone MIC values of 2.0 μg/mL (Table 3).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Antimicrobial susceptibility of Streptococcus pneumoniae isolated from nasopharyngeal carriage in children aged 2 - 5 years old (n = 160).

https://doi.org/10.1371/journal.pone.0357903.t003

Trimethoprim-sulfamethoxazole (SXT) non-susceptibility was common, with 8.1% (13/160) of isolates showing intermediate susceptibility and 56.3% (90/160) exhibiting full resistance. Resistance to erythromycin and clindamycin were observed in 33.8% (54/160) and 28.8% (46/160) of isolates, respectively. All isolates were susceptible to levofloxacin and vancomycin.

Non-susceptibility of individual pneumococcal serotypes to penicillin or ceftriaxone is shown in Table 4. Six serotype 19A isolates and one serotype 11A isolate had a penicillin MIC > 2.0 μg/mL. Among the 28 isolates with a penicillin MIC value between 0.125 and < 2.0 μg/mL, the most frequent serotypes were 19A (n = 9), 6C (n = 6), 23A (n = 5), 14 (n = 3), 15A (n = 2), 10A (n = 2), and 15B (n = 1). Overall, 38 isolates (23.8%) were classified as MDR, with serotypes 19A (n = 15), 6C (n = 8), and 23A (n = 6) accounting for the majority of MDR cases.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Penicillin and ceftriaxone MIC values according to Streptococcus pneumoniae serotypes isolated from children aged 2–5 years.

https://doi.org/10.1371/journal.pone.0357903.t004

Discussion

Pneumococcal vaccine coverage in Brazil for the PCV10-GSK was historically high (>90%) within the Unified Health System (SUS) but declined after 2015, because of multiple factors, including declining vaccine confidence, operational challenges in immunization services, and disruptions to routine childhood vaccination during the COVID-19 pandemic, reaching approximately 62% among children under two years of age in 2021. Recovery efforts implemented in 2023–2024, including the National Vaccination Movement, resulted in a partial rebound, with national coverage estimates rising to about 83.1%, although regional disparities persist [11].

In this study involving 400 preschool-aged children enrolled between August and November 2023 in Salvador, Brazil, we observed a pneumococcal carriage rate of 39.5%. Carriage prevalence varied across the participating schools, ranging from 18.2% to 64.3%. This heterogeneity may reflect differences in classroom crowding, socioeconomic conditions, household composition, exposure to respiratory infections, antibiotic use, or local transmission dynamics. In addition, because only one school was sampled per administrative region, the observed variation may also reflect geographic heterogeneity across Salvador.

Notably, only 3.2% of carriage isolates corresponded to serotypes represented in the PCV10-GSK vaccine, underscoring the continued circulation of S. pneumoniae among young children in a context of incomplete and recently recovering vaccine coverage and the predominance of non-PCV10-GSK serotypes. Comparable overall pneumococcal nasopharyngeal carriage rates in the post-PCV10-GSK era have been reported in Latin America and the Caribbean, as documented in a recent systematic review and meta-analysis [21]. However, considerable heterogeneity exists across Brazilian regions: studies from Niterói (Southeastern Brazil) reported overall carriage rates of 17.4% [22] and 22.6% [16,21], whereas a recent study from southern Brazil documented an overall carriage prevalence of 63.7% in a similar age group [23]. Together, these findings indicate that nasopharyngeal colonization remains frequent among vaccinated children and continues to represent an important reservoir for pneumococcal transmission and genetic exchange.

Serotype distribution was dominated by non-PCV10-GSK types, particularly serotypes 6C (17.9%), 19A (13.0%), 11A (9.3%), and 15B/15A (16.0% combined). These findings align with previous Brazilian studies, where serotype 6C has emerged as a leading colonizer [15,16,22]. An increasing trend in 6C prevalence has also been reported in countries such as Iceland and Belgium following the replacement of PCV13 with PCV10-GSK, possibly due to the removal of serotypes 6A and consequent loss of cross-protection against 6C that may have been indirectly provided by PCV13 [24,25]. In Brazil, serotype 6C was recently identified as the third most frequent serotype among isolates related with invasive pneumococcal disease in children under five years of age and in the general population [26]. Similar trends have been observed across other South American countries using PCV10-GSK [27]. The emergence of MDR serotype 6C isolates raises additional concerns for both colonization and disease [16].

Environmental and behavioral factors such as household crowding, exposure to tobacco, contact with other young children, and recent antimicrobial use were not significantly associated with overall pneumococcal carriage in the multivariable analysis. Although close contact with young children is a well-established risk factor for pneumococcal carriage, all participants in our study attended preschool, where frequent contact with peers likely represented the predominant source of pneumococcal exposure [28]. Consequently, additional household contact with young children may have contributed little incremental risk. The white race was associated with lower odds of carriage compared with mixed race. Living with elderly individuals was associated with lower odds of carriage in the primary multivariable model. However, this association was attenuated after adjustment for school, suggesting that the initial finding was likely influenced by school-level or geographic confounding rather than representing a true protective effect.

Among children colonized with S. pneumoniae, we did not identify demographic, clinical, household, or vaccination-related factors independently associated with carriage of non-PCV20 serotypes. Although older age and living with another child under five years of age showed positive trends in the adjusted analysis, these associations did not reach statistical significance. The absence of clear predictors may reflect the widespread circulation of non-PCV20 serotypes in this highly vaccinated population, suggesting that exposure to these serotypes is no longer restricted to specific demographic or behavioral subgroups. These findings support the hypothesis that non-PCV20 serotypes have become established in the community following the decline of PCV10-GSK vaccine serotypes, reinforcing the importance of continued surveillance to monitor the emergence and dissemination of replacement serotypes.

The high prevalence of serotype 19A, an invasive, antimicrobial resistant type not included in PCV10-GSK vaccine was particularly concerning [29]. This serotype is included in PCV13 and newer, higher-valency vaccines. Consistently, serotype 19A has been detected with increasing frequency in colonization and disease in several Brazilian regions and other countries that have adopted PCV10-GSK [15,24,25,30].

According to national laboratory surveillance data, 19A was the most common serotype associated with IPD in Brazil in 2023, responsible for nearly half of IPD cases in children under five years old and remained the leading cause of IPD across all age groups (26). Similarly, in other Latin American countries that adopted the PCV10-GSK, serotype 19A has also emerged as a leading cause of IPD [27]. These findings emphasize the limited serotype coverage of PCV10-GSK and highlight the potential benefits of introducing broader-valency formulation. In our study, only 17.3% of the identified serotypes were covered by PCV13, PCV15, or PCV10-SII, whereas 39.5% were included in PCV20. At the child level, 40.5% of colonized children carried at least one PCV20 vaccine-type serotype, while 59.5% carried only non-PCV20-targeted serotypes and/or non-typeable isolates.

On the other hand, serotype 3, highly invasive and frequently isolated in IPD was not detected in this carriage study. Although rarely found in healthy children, serotype 3 is a major cause of bacteremia and meningitis in Brazil [29,31,32]. Its low colonization capacity, paired with high disease potential, supports the inclusion of this serotype in future vaccine strategies and the development of formulations capable of eliciting stronger protection against it [33].

The antimicrobial resistance patterns observed in our study was concerning. High levels of resistance to trimethoprim-sulfamethoxazole (56.3%) were consistent with its historical overuse in Brazil and other settings [34]. Resistance to erythromycin (33.8%) and clindamycin (28.8%) was also notable, limiting the effectiveness of empirical therapy. Importantly, 42.5% of isolates were nonsusceptible to penicillin by oxacillin screening, with 4.4% demonstrating high-level penicillin resistance (MIC > 2.0 µg/mL). These findings align with national antimicrobial resistance surveillance data, particularly among serotype 19A and 11A [15,29].

Co-colonization with multiple serotypes was identified in 2.5% of children, including combinations such as 19A/6C and 14/23A. Although infrequent, such events have important epidemiological implications, as they may facilitate horizontal gene transfer and promote the emergence of recombinant isolates [35,36], including those with enhanced virulence or resistance characteristics.

Despite the valuable insights provided, this study has some limitations. First, the cross-sectional design precludes establishing causal relationships between risk factors and pneumococcal carriage. Second, the potential underreporting of antibiotic use by caregivers may have led to misclassification and biased the observed associations. Another limitation is that only one school was sampled within each administrative region. Although this approach ensured geographic representation across Salvador, pneumococcal carriage prevalence varied substantially between schools, suggesting that unmeasured school-level characteristics may have influenced transmission dynamics. A sensitivity analysis including school as a random effect was performed for the multivariable model of overall carriage; however, residual confounding by unmeasured school-level factors cannot be excluded. Therefore, the findings may not fully capture the variability in carriage patterns among all preschools in the city.

Overall, our findings reinforce the importance of ongoing pneumococcal surveillance in pediatric populations, particularly in the context of widespread vaccine use, shifting serotype dynamics, and rising antimicrobial resistance. The observed serotype distribution, predominance of non-PCV10-GSK and non-PCV20-targeted carriage among colonized children, and resistance patterns suggest that higher-valency vaccines could offer improved protection in these settings. The data offer valuable insights into current carriage dynamics and resistance patterns in a highly vaccinated pediatric population.

Supporting information

Acknowledgments

The authors sincerely thank the Municipal Department of Education of Salvador, Bahia, for its institutional support and collaboration throughout this study. We are especially grateful to the principals, teachers, staff, children, and their families from all participating schools for their cooperation, hospitality, and willingness to participate in this research. Their support was essential for the successful conduct of the nasopharyngeal carriage investigation and greatly contributed to the completion of this study. ChatGPT 4 was used to review grammar, synonyms, and punctuation.