Serotype distribution and antimicrobial resistance of Streptococcus pneumoniae causing invasive diseases in China: a meta-analysis

Background To summarize information about invasive pneumococcal disease (IPD) among children in mainland China. Methods Sixteen eligible studies were included in this systematic review and the random effect model was used to estimate the pool prevalence of IPD. Results The most predominant serotypes circulating in children were 19F (27.7, 95% confidence interval (95% CI): 17.7–37.6%), 19A (21.2%, 16.4–26.1%), 14 (16.5%, 12.8–20.1%), 6B (8.6%, 5.2–10.8%) and 23F (7.3%, 5.2–9.5%). The serotype coverage of the available pneumococcal conjugate vaccines PCV7, PCV10, and PCV13 was 60.8% (52.5–69.4%), 65.1% (57.7–72.4%), and 90.0% (87.1–92.8%), respectively. The pooled antibiotic resistance rates of Streptococcus pneumoniae revealed a resistance to penicillin prevalence rate of 32.0% (12.1–51.9%). Approximately 94.4% (90.7–98.1%) and 92.3% (87.4–97.3%) of isolates were resistant to erythromycin and clindamycin. eBURST analysis revealed great diversity among isolates, with 102 sequence types (STs) for 365 isolates. The major predominant clonal complexes (CCs) were CC271 (43.6%, 159/365), CC876 (13.4%, 49/365), CC81 (5.2%, 19/365), and CC90 (4.1%, 15/365). Long-term and regional surveillance of S. pneumoniae is necessary. Conclusions Based on our pooled results showing that PCV13 coverage of the reported serotypes was 90% and that most serotypes contributed to the distribution of antibiotic-resistant isolates, implementation of PCV13 into the Chinese Expanded Program on Immunizations (EPI) would achieve health benefits in Chinese children.


Background
Streptococcus pneumonia (S. pneumoniae) is one of the most prominent pathogens, causing mild to lifethreatening invasive diseases due to its well-known capsule pathogenicity [1]. It has been reported that approximately 1 million people die of pneumococcal diseases annually, most of whom are children under 5 years old [2][3][4]. Pneumococcal conjugate vaccines (PCVs) targeting 7, 10, or 13 of more than 90 serotypes of S. pneumoniae have been successively introduced to reduce the burden of invasive pneumococcal disease (IPD) in vaccinated children in developed countries [5,6]. In fact, it has been reported that after licensing PCV13, pneumococcal diseases of any cause, but especially those caused by PCV13 minus PCV7 serotypes, were further decreased across each age group in the UK [7].
In China, S. pneumoniae is one of the most common pathogens and can cause infectious diseases, especially pneumonia, in children under 5 years of age [8]. PCV7 was first introduced in mainland China in 2008 [9] but due to its high price was not included as part of the national immunization schedule. Because of its high cost in the the private market [10], and the vaccine may not be available to some low-income families, especially in western China. But a model predicted that every year, 112,629 cases of pneumococcal-related disease could be prevented in Shanghai if the PCV7 vaccine could be introduced [11]. After employing PCV7, another multicenter study conducted in Shanghai in 2013 reported that the serotype coverage of PCV7, − 10, and − 13 was 58.6, 59.4 and 85.1%, respectively [12].
Both genetic background and capsular type contribute to the ability of S. pneumoniae to cause invasive diseases [13,14]. The aim of this study was to obtain a systematic estimate of the serotypes of S. pneumonia and their antibiotic resistances and to determine clonal types causing IPD in mainland China.

Literature search strategy
The design and construction of this systematic review was performed and completed according to the PRISMA checklist [13]. (see Additional file 1). The following databases were searched: PubMed, MEDLINE, EMBASE, WanFang Data (http://www.wanfangdara.com.cn), and the China National Knowledge Infrastructure (CNKI) database (http://cnki.net/). The time frame was January 2000 to September 2016. The keywords for the literature search were "Streptococcus pneumoniae OR invasive pneumococcal disease OR pneumococcal conjugate vaccine", "children OR school children OR 0-18 years", "serotype OR capsular type" and "China OR Chinese".

Eligibility criteria
Two reviewers (RS Yi and JC Chen) independently identified eligible studies and extracted data based on the following criteria: (1) the publication language was Chinese or English, (2) the study participants were generally Chinese children aged from 0 to 18 years, (3) the study provided invasive pneumococcal serotypes and/or specific patterns of antibiotic sensitivity, and (4) the serotyped pneumococcal isolates were obtained from normally sterile sites (e.g., blood, cerebrospinal fluid, and pleural fluid). Exclusion criteria were as follows: (1) the study data included adult patients, and specific infection data about children could not be extracted; (2) the data were not regarding invasive pneumococcal diseases; (3) the data did not provide specific serotypes and/or specific patterns of antibiotic sensitivity; (4) the studies were reviews, lectures, or editorials; or (5) the studies were duplicates published in both Chinese and English (in this case, only the English-language studies were included).

Data extraction
The data were screened and reviewed independently by two authors (RS Yi and JC Chen) according to the inclusion and exclusion criteria, and the following information was extracted: first author's name; study location; district; study period; publication year; age group; positive strains; serotyping method; serotypes; serotype coverage of PCV7, − 10, and − 13; and antibiotic resistance patterns of the positive cases and total cases.

Statistical analysis
Statistical analyses were performed using Stata statistical software version 13.0 (Stata corporation LP, College Station, Texas, USA). The pooled relevant serotypes and antibiotic resistance prevalence and their corresponding 95% confidence intervals (CIs) were evaluated using the DerSimonian and Laird random-effects model [15]. The Cochrane chi-square (x 2 ) test and quantification with the I 2 statistic were used to calculate the source of heterogeneity [16,17]. Begg's funnel plots and Egger's test (significant at P ≤ 0.1) were used to assess publication bias [18].
MLST data analysis performed using eBURST v3 grouped the isolates into 4 CCs. ST81, ST90, ST876 and ST271 were confirmed as probable founder genotypes (Fig. 2). Eleven single-locus variants (SLVs) were found in ST271. Figure 3 shows high genetic diversity of IPD  isolates from mainland China based on the global S. pneumoniae genetic background.

Discussion
Epidemiological and clinical studies have shown that invasive pneumococcal diseases caused by S. pneumoniae is a major public health burden. Based on a systematic review of the literature, Chen et al [35] reported that approximately 700,000 children are diagnosed with pneumococcal disease each year and that approximately 30,000 deaths are attributed to S. pneumoniae; moreover, over 50% of children with pneumonia in China died because of this pathogen. The high pneumococcal disease burden makes the investigation and evaluation of S. pneumoniae serotypes important for providing evidence and guidance for the use of vaccines and antibiotics targeting the pathogen.
Our study revealed that the predominant serotypes circulating among Chinese children were 19F, 19A, 14, 6B and 23F. For children under 5 years old, the predominant serotypes were 19F, 19A, 14, 23F and 6B. A global serotype project confirmed that serotype 14 was the most prevalent cause of IPD among children under five in every region, including Asia; however, this systematic review did not include studies from mainland China, which contains the largest population of children under 5 years of age in the world [36]. In that report, seven serotypes (1, 5, 6A, 6B, 14, 19F and 23F) accounted for over 50% of IPD in every region, including Asia, and our results demonstrated that 14, 19F, 6B, 23F and 19A accounted for 80.7 and 73.5% of all IPD in Chinese children 0-14 years and children under 5 years of age, respectively. The major serotype patterns of S. pneumoniae in mainland China were consistent with the ANSORP study, which revealed that 19F, 23F, 19A, 14 and 6B were the major serotypes in Asia [37]. Serotype 1 is reported as one of the 5 most common serotypes in many developing countries such as in South Africa Kenya and in Philippines [37][38][39], but this serotype was infrequency in our study, which may revealed that the serotype distribution of S.pneumoniae was varied geographically [40].
Pooled analysis of multiple surveillance sites revealed that vaccine-serotype and overall IPD rates declined consistently and significantly after PCV7 introduction, and this effect was still observed after 7 years [41]. Unfortunately, these promising effects may not represent the important public health implications in some countries that did not implement routine immunization with PCVs, such as China [10]. Although the PCV7 vaccine was imported into and licensed in China in 2008, this vaccine has not yet been included in the Chinese Expanded Program on Immunizations (EPI) [42]. The PCV7 vaccine was removed from the market in 2015, though new vaccines, such as PCV10 and PCV13, have not yet been employed in the Chinese market, making the situation even worse [10].
Our analyses revealed serotype coverage rates for PCV7, − 10, and − 13 of 60.8, 65.1, and 90.0%, respectively, for IPD among Chinese children. A cohort and case-control study conducted in Europe [43] revealed that after the 13-valent pneumococcal conjugate   Global multi-region studies have revealed that the highest antibiotic resistance to S. pneumoniae was consistently found in Asia [37], and the ANSORP reports demonstrated that S. pneumoniae isolates from mainland China had the highest antibiotic resistance rates [37]. Indeed, it was reported that S. pneumoniae isolated from Chinese children had the highest antibiotic resistance rates [31]; the most predominant resistance, 96.4%, was to the antibiotic erythromycin, followed by tetracycline and sulfamethoxazole. Our study revealed S. pneumoniae to be most frequently resistant to erythromycin, followed by clindamycin, tetracycline, and sulfamethoxazole, consistent with previous study and the ANSORP study [37].
Our findings indicated a pooled penicillin resistance rate to S. pneumoniae of 32.0% and a PNSP of 74.6%, which was much higher than the ANSORP study results but consistent with a study conducted in Beijing that revealed a total non-susceptibility rate of penicillin of 91.5% if based on an oral breakpoint of Clinical and Laboratory Standards Institute (CLSI) [31]. The limitation of our study is that we cannot estimate the resistance rate to penicillin based on revised CLSI breakpoints, as most of the included studies did not report the adaptation of the breakpoints. However, our results were partially consistent with the ANSORP study, which reported a resistance rate to penicillin of 60.0% for meningeal isolates [37].
The snapshot from eBURST showed a great diversity among IPD strains isolated in mainland China. The predominant clones were CC271, CC876, CC81 and CC90. CC271 was the most predominant CC, and the inclusion of 11 STs, such as ST271, ST320, ST236, and ST4664, accounted for 43.6% of all the included isolates from the literature. Several studies have confirmed that this clone became prevalent prior to the introduction of PCV7 and spread rapidly among both adults and children in China [44]. The widespread use of antibiotics has also been implicated in the emergence of serotype 19A and 19F isolates in both communities and hospitals. The two serotypes of the isolates were reported to be multidrug resistant and carriers of the ermB and mefA genes, making the situation even worse [47,48].
A significantly increased risk of CC271 strains, which caused IPD in our study, is anticipated among children and is in agreement with existing trends in the literature regarding transmission [49]. ST236, namely the original Taiwan 19F -14 clone, was the major international antibiotic-resistant strain. In addition, long-term monitoring data revealed that ST271 and ST320 evolved rapidly and replaced ST236 due to their higher fitness and that they were becoming well established in local regions of China [47]. It was also reported that CC271 is prevalent in Asian countries, such as Japan and Korea, due to the introduction of PCV7 [50,51]. After implementation of PCV7, non-vaccine-related serotypes, such as 19A, circulated among pediatric groups. This serotype was reported to be multidrug resistant due to PCV7 implementation and antibiotic selection pressure. The high prevalence of CC271 clones was consistent with the high penicillin non-susceptibility rate found in IPD isolates pooled in this study.
According to a multicenter study conducted in China, ST320, ST271, and ST876 were the prevalent types among IPD isolates collected from children [52]. Another study showed that the CC876 prevalence increased from 0% in 1997-2000 to 96.4% in 2010-2012 in China, and the most important findings of this study were that this clone had high non-susceptibility rates to β-lactam antibiotics [53].

Conclusions
This is the first study to pool data for invasive pneumococcal diseases among children. The study has several limitations. First, the studies pooled were highly heterogeneous. Second, the studies included did not report the specific breakpoint from CLSI, and therefore we could not estimate the actual resistance of meningeal isolates. Long-term monitoring is required to estimate antibiotic resistance and serotype distributions in an effort to provide a policy to introduce a new PCV vaccine for Chinese children.