Prevalence and serotype distribution of nasopharyngeal carriage of Streptococcus pneumoniae in China: a meta-analysis

Background To explore the overall prevalence and serotype distribution of nasopharyngeal carriage of Streptococcus pneumoniae(S. pneumoniae) among healthy children. Methods A search for pneumococcal nasopharyngeal carriage studies including children published up to July 31th, 2016 was conducted to describe carriage in China. The review also describes antibiotic resistance in and serotypes of S. pneumoniae and assesses the impact of vaccination on carriage in this region. Summary measures for overall prevalence, antibiotic resistance, and serotype distributions extracted from the analyzed data were determined with 95% confidence intervals (CIs) using random-effects models. Heterogeneity was assessed using I 2 test statistics. Results Thirty-seven studies were included in this review, and the majority of studies (64.9%) were located in the pre-introduction period of 7-valent pneumococcal conjugate vaccine (PCV7) in China. The pooled prevalence of S. pneumoniae nasopharyngeal carriage was 21.4% (95% CI: 18.3–24.4%). Carriage was highest in children attending kindergartens [24.5%, (19.7–29.3%)] and decreased with increasing age. Before the introduction of PCV7 into China, the prevalence of S. pneumoniae nasopharyngeal carriage was 25.8% (20.7–30.9%), the pooled carriage of S. pneumoniae sharply dropped into the 14.1% (11.3–16.9%) by PCV7 vaccination period (P < 0.001). Before the pneumococcal conjugate vaccine (PCV) was introduced in China, the penicillin resistance rate in S. pneumoniae isolated from healthy children was 31.9% (21.2–42.6%); however, this rate sharply decreased after the introduction of PCV7 in China [21.6%, (7.4–35.9%)], and the difference between the rates during these two time periods was statistically significant (P value <0.05). Serotypes 19F, 6A and 23F were the most commonly isolated. Meta-analysis of data from young children showed a pooled rate estimate of 46.6% (38.8–54.4%) for PCV7 vaccine coverage and 66.2% (58.6–73.8%) for PCV13 vaccine coverage. Conclusions The prevalence of nasopharyngeal carriage among children was high in China. PCV7 immunization was found to be associated with reduction of nasopharyngeal colonization of S. pneumoniae. Conjugate vaccination coverage was slightly affected by the introduction of PCV7 into China because of low vaccination rate. The government should implement timely adjusted conjugate vaccination strategies based on our findings. Electronic supplementary material The online version of this article (10.1186/s12879-017-2816-8) contains supplementary material, which is available to authorized users.


Background
Streptococcus pneumoniae (S. pneumoniae) is a major pathogen that can cause invasive pneumococcal disease (IPD) and respiratory tract infections and result in high morbidity and mortality. The World Health Organization has reported that nearly 500,000 children under 5 years of age are infected by S. pneumoniae annually, and the vast majority of these infections occur in developing countries [1]. Asymptomatic nasopharyngeal carriage of S. pneumoniae is an essential element of the transmission of pneumococcal disease [2], a prerequisite for the occurrence of invasive pneumococcal disease, and a known risk factor for subsequent acute and recurrent otitis media [3,4].
The prevalence of nasopharyngeal pneumococcal carriage has been found to vary in different countries and regions [5]. Because S. pneumoniae carriage is more common than the S. pneumoniae disease, it is important to investigate carriage status to evaluate the effect of new pneumococcal vaccines [6]. When the 7-valent pneumococcal vaccine was introduced in mainland China, the invasive pneumococcal disease burden decreased sharply, especially disease caused by the vaccine type (VT) serotypes; this decrease was accompanied by an increase in non-vaccine type (NVT) serotype, particularly serotype 19A, as previously seen in Europe [7,8].
This systematic review was conducted to describe the nasopharyngeal carriage status of S. pneumoniae in healthy children, describe the major serotypes of S. pneumoniae, and evaluate the impact of pneumococcal vaccination on the coverage of PCV7.

Inclusion and exclusion criteria
Studies were required to meet the following criteria for inclusion in this meta-analysis: (1) subjects were healthy children, (2) samples were collected from nasopharyngeal or oropharyngeal swabs, (3) studies focused on nonvaccination group and (4) sufficient information was provided to compute positive carriage rates and their 95% confidence intervals (CIs). Exclusion criteria were as follows: (1) if a study included both adults and children, only children data were enrolled, (2) studies reporting clinical infectious diseases caused by S. pneumoniae, (3) if studies included both vaccinated and non-vaccinated children, only non-vaccinated data were enrolled, (4)studies with a lack of sufficient baseline information to compute carriage rates and their 95CIs, (5) review studies, or conference studies or newspaper articles, (6) studies determining antibiotic resistance rates without carriage data, or studies were referred to infections rather than colonization, and (7) duplicate reports.

Data extraction
Two reviewers (LW and JF) independently identified and extracted the following data: first authors, sample year, study location, study population, number of participants, number of participants with pneumococcal carriage, pre/post vaccination period, vaccination history, type of swabs, immediately incubated into plates or not, transportation period, culture plates, culture into the 5% CO 2 or not, identification methods, serotyping methods, storage medium, rates  of antibiotic resistance, and prevalence of S. pneumoniae serotypes and their corresponding 95% CIs.

Quality assessment
The quality of included studies was assessed in accordance with the STROBE statement [9], studies with scores <8 were excluded from the systematic review.

Statistical analysis
STATA version 10.0 was used to perform the statistical analyses. DerSimonian and Laird random-effects models (REM) were used to pool the data. Funnel plots were  used to examine publication bias, which was further assessed using Egger's test, with P < 0.10 indicating potential bias [10]. Stratified analyses were carried out to assess the heterogeneity across subgroup defined by age and PCV7 vaccination period.

Characteristic of included studies
The flow chart in Fig. 1 depicts the selection process for the included studies. Overall, 614 studies were written in Chinese, and 21 studies were written in English. By  (Fig. 2).
Identification and confirmation of S. pneumoniae with different methods Table 2 summarizes the methods used to identify and confirm the S. pneumoniae strains. Three different methods, including PCR, optochin disk with bile solubility and latex agglutination were used. There was no impact on the prevalence of S. pneumoniae when using three different identification methods, see Fig. 3. Nasopharyngeal carriage of S. pneumoniae by age Figure 4 summarizes the prevalence of nasopharyngeal carriage of S. pneumoniae in healthy children in different age groups. Six studies [11,14,17,34,39,42] reported the prevalence of nasopharyngeal carriage of S. pneumoniae among children younger than 2 years of age.
Among the 6656 healthy children in this age group, a total of 847 were identified to be positive for nasopharyngeal carriage of Streptococcus pneumoniae; thus, the pooled prevalence was 11.7% (9.1-14.2%). Twentyseven studies [12-16, 18, 21-33, 36-38, 41, 44-47] including 10,480 kindergarten children (2-5 years of age) investigated the prevalence of nasopharyngeal carriage of S. pneumoniae. Within these studies, a total of 2437 children were identified to be positive for S. pneumoniae carriage, and the pooled prevalence was 24.5% (19.7-29.3%). Among the 1122 healthy children who were older than 5 years of age [15,19,37], 104 were identified as S. pneumoniae carriers; therefore, the prevalence of nasopharyngeal carriage was 8.8% (6.0-11.5%) in this age group. The prevalence of nasopharyngeal carriage of S. pneumoniae varied between the three age groups, with the highest rate reported in kindergarten children (P = 0.002). The 7-valent pneumococcal conjugate vaccine was introduced to China in October 2008, but it has not yet been included in the Chinese Expanded Program on Immunizations (EPI) [48]. Unlike the vaccination in Chinsed EPI schedule, the PCV7 vaccine was not free to the public and the coverage was estimated as 9.91% [49]. Before the PCV7 was introduced in mainland China, 24 studies [12, 16, 18-23, 25-31, 33, 36, 40, 41, 43-45, 47] had reported the prevalence of nasopharyngeal carriage of S. pneumoniae; within these studies, the pooled prevalence was 25.8% (20.7-30.9%), Fig. 5. The prevalence of nasopharyngeal carriage sharply declined following the introduction of PCV7, with a pooled prevalence of 14.1% (11.3-16.9%) identified in studies conducted post-PCV7 introduction [11, 13-15, 17, 24, 32, 34, 35, 37, 39, 42, 46]. There was a highly significance differences in the prevalence between these two time periods (P < 0.001). In kindergarten children, before the

Overall heterogeneity and publication bias
Stratified analyses were carried out to assess the heterogeneity across subgroups defined by age, PCV7 introduction period and PCV7 introduction period within kindergarten children groups. The sensitivity analysis indicated that the pooled prevalence of S. pneumoniae carriage had only slight variations by stratified studies into pre/post vaccination period when individual studies were omitted one by one. The prevalence estimates ranged from 13.4% (10.6, 16.1%) to 14.8% (12.5, 17.1%) in post vaccination period and from 25.2% (20.7, 31.1%) to 26.3% (21.1, 31.6%) in prevaccination period, suggesting that the results were stable.
Slight publication bias was noted from the statistical tests (Egger's test, P = 0.011; Begg's test, P = 0.01). After stratified the pooled prevalence of S. pneumoniae by PCV7 vaccination period, the potential publication bias was adjusted as no significant (Egger's test, P = 0.134; Begg's test, P = 0.602) in pre-vaccination period and (Egger's test, P = 0.353; Begg's test, P = 0.125) in post vaccination period.

Discussion
This systematic review analyzed the prevalence and serotype distributions of nasopharyngeal carriage of S. pneumoniae, antibiotic resistant rates in S. pneumoniae, and the rates corresponding the serotype coverage provided by PCV7 and PCV13.
Since the serotypes distribution of and antibiotic resistance in S. pneumoniae isolates have been found to vary from region to region, the prevalence of S. pneumoniae has also been found to vary in different populations. The prevalence of nasopharyngeal carriage of S. pneumoniae was found to be 60% in infants under 2 years of age in Greenland [49], while the prevalence of nasal carriage was only identified as 9.8% in elderly populations in Italy [50]. In Hong Kong, the prevalence of nasopharyngeal carriage S. pneumoniae was identified as 13.5% in children younger than 5 years of age who had never received any pneumococcal vaccines, 14.1% in children who received at least one dose of PCV13, and 15.3% in children who received at least 3 doses of the PCV13 vaccine [51]. In Taiwan, the prevalence of nasopharyngeal carriage of S. pneumoniae identified in children younger than 5 years of age was 14.1%, similar to that identified Hong Kong [52]. However, data collected in mainland China have differed from data collected in Taiwan and Hong Kong. The pooled prevalence of nasopharyngeal carriage of S. pneumoniae was determined to be 21.4% (18.3-24.4%) among children in China.
A variety of studies have confirmed that colonization by S. pneumoniae begins in infanthood and early childhood. It has been reported that carriage of this pathogen is acquired within the first 6 months of life and, the prevalence of the epidemic appeared to peak in children of pre-school age [53]. A study conducted by Ueno M [53] showed that prevalence of nasopharyngeal carriage of S. pneumoniae increased with age within pediatric age groups, with rates of 19 and 23% identified in infants younger than 1 years-old and children 2 to 3 years old, respectively. The highest prevalence has been identified during the pre-school period. Our data were consistent with the findings of Ueno M [53], suggesting that carriage trends differed with age. The prevalence was 12.8% (10.0-15.6%) in children younger than 2 years old; the prevalence increased with age and reached a peak at 24.7% (19.7-29.7%) in children aged 2 to 5 years and then decreased to 8.8% (6.0-11.5%) in children aged 5 years and older. It is well known that attending kindergarten has been identified as a risk factor [52,53] for colonization by opportunistic pathogens, such as S. pneumoniae, due to poor hygiene, confined physical environmental conditions and frequent interaction with other children. Nasopharyngeal carriage of S. pneumoniae in kindergarten children results in this population serving as an asymptomatic reservoir that spreads this pathogen into community. Since the PCV7 was introduced into China in October 2008, the studies conducted between 2009 to 2012 in age 2 to 5 years-old children were the coverage and the active population of getting shot by PCV7 vaccine, which leads to a reduction of prevalence of nasopharyngeal carriage of S. pneumoniae.
Unlike the GAVI Alliance [54] in the world and EPI in China, the PCV7 is available at immunization clinics for a fee during 2008-2015, these clinics designated as "point of vaccination" centers, children at 2, 4, 6 months will get shot of one dose of PCV7 and at 1 years old will get the fourth shot of does to enhance the immunity after purchase the vaccine [54]. Because of the high price of PCV7, the PCV7 coverage level was not as many other countries [8,9]. According to a survey of children age 1 to 2 years selected from 31 provinces throughout China conducted in 2012, 9.9% of children had received one dose of PCV7 [49]. Another study from Shanghai reported a similar PCV7 coverage level at 11.4% [55]. We observed a slightly change of PCV7 coverage level from 43.9% (34.1, 53.6%) to 52.1% (37.3, 66.9%) between pre/post vaccination period because of the limited herd immunity from low vaccine rate of pneumococcal conjugate vaccination.
High antibiotic resistance rates in S. pneumoniae may facilitate transmission of this pathogen among young children. Crowding and barriers to maintaining quality hygiene facilities could accelerate the transmission of highly antibiotic resistant S. pneumoniae in the kindergarten environment [56]. Our pooled data indicated that the rates of erythromycin, clindamycin, trimethoprimsulfamethoxazole and tetracycline resistance among isolates were all more than 60%. High-level resistance to the aforementioned antibiotics has also been identified in previous studies [57]. Macrolides and lincosamides have been reported to be the first-line empirical antibiotic therapy for pneumococcal infections in China, and the use of these agents has led to a high rate of antibiotic resistance in S. pneumoniae [42,57]. Previous studies have demonstrated that the penicillin-nonsusceptible pneumococci (PNSP) rate varied in different regions. The prevalence of nasopharyngeal carriage of S. pneumoniae in Brazilian and Korean children who attended day care centers were identified as 26.0 and 31.3%, respectively [58,59]. A marked modification in pneumococcal antibiotic susceptibility rates was observed after the introduction of pneumococcal conjugate vaccines. The PNSP rate was 47.1% before the introduction of PCV13 in France, and this rate rapidly decreased to 39% 3 years after PCV13 was introduced [60]. The pooled data in this study were consistent with results identified in France. The proportion of pneumococcal isolates resistant to penicillin identified in this study decreased from 31.9% (21.2-42.6%) to 21.6% (7.4-35.9%) after the introduction of PCV7.
A remarkable decrease in the incidence and mortality of invasive pneumococcal disease has been observed following the introduction of pneumococcal conjugate vaccines into pediatric immunization programs [61]. With the introduction of these PCVs and further reductions in the prevalence of nasopharyngeal carriage of S. pneumoniae in pediatric groups. Our data demonstrated that the prevalence of nasopharyngeal carriage of S. pneumoniae was 25.8% (20.7-30.9%) among healthy children before the introduction of PCV7. The prevalence dropped sharply to 14.1% (11.3-16.9%) following the introduction of PCV7 in China, indicating that the impact of PCV7 introduction on disease prevalence can be determined by assessing the nasopharyngeal carriage of S. pneumoniae in healthy children.

Conclusions
Pneumococcal carriage was identified to occur at generally high prevalence among children in China. PCV7 immunization was associated with a reduction in the rate of penicillin resistance among nasopharyngeal carriage isolates of S. pneumoniae. The distribution of serotypes identified in the nasopharynx was only slightly modified following the introduction of the PCV7 vaccination because of the low PCV7 immunization rates. The Centers for Disease Control and Prevention should timely adjust PCV vaccination strategies based on these findings to reduce the incidence and morbidity of pneumococcal invasive disease in pediatric populations.