Introduction

Shigella, a group of Gram-negative, non-spore forming and rod-shaped bacteria, are the causative agents of shigellosis (or bacillary dysentery). The bacteria are considered to be highly contagious because of their low infectious inoculum size (10–100 organisms) [1]. Shigella is primarily transmitted through the fecal-oral route; therefore, it is still a major global public health threat, particularly in developing countries with poor sanitation conditions. One hundred sixty-five million episodes of shigellosis are estimated to occur worldwide annually, among which 163 million occur in resource-poor countries [2]. Moreover, a recent surveillance study in six Asian countries suggested that the global burden of Shigella infection might be much higher than these estimates [3]. Nonetheless, the number of Shigella-related deaths in Asia has substantially decreased due to current nonspecific interventions, including measles vaccination, vitamin A supplementation, and improved nutrition [4]. However, the treatment of shigellosis has recently become more difficult due to the emergence of multidrug-resistant strains, which has narrowed the clinical choice of antimicrobial agents [3].

Shigella bacteria are serologically grouped into four species (S. dysenteriae, S. flexneri, S. boydii, and S. sonnei), which are further subdivided into more than 40 serotypes based on the O-antigen structures of the membrane-associated lipopolysaccharide [5]. Though these species share similar pathogenic properties, they exhibit unique epidemiological characteristics. For instance, S. flexneri is known to be predominant in developing countries (median 60%), while S. sonnei is the most common species found in the industrialized world (median 77%) [2]. However, recent studies have indicated that S. sonnei is overtaking S. flexneri in newly industrialized nations such as Thailand, South Korea and Taiwan [3], [6], [7]. As initial Shigella infection produces only homologous, not heterologous, serotype protection [8], a comprehensive and current understanding of the prevalent Shigella species and serotypes is essential for the development of an effective vaccine [9]. Currently, though several strategies have been used to develop vaccines targeting shigellosis, only one live bivalent S. flexneri 2a and S. sonnei vaccine is licensed in China [8], [10]. Updated information on the circulating strains of Shigella is also important to enable assessment of the impacts of vaccination.

A retrospective review indicated that S. flexneri was responsible for 86% of shigellosis episodes in China during 1991–2000, and the predominant serotype was identified as S. flexneri 2a (80%) [11]. However, concurrent with the rapid economic development in China during the past decade, both morbidity and mortality from shigellosis have decreased steadily every year, based on annual reports from China’s Center for Disease Control and Prevention. A recent study using the national surveillance data from 2009 showed that the annual shigellosis morbidity rate was 20.28 cases per 100,000 people in mainland China, with S. flexneri (67.3%) and S. sonnei (32.7%) as the two major causative species [12]. Nevertheless, no comprehensive data are currently available regarding the causative species and serotypes of shigellosis in mainland China during the past decade. In this study, we performed a systematic review and meta-analysis of the most recently published data on Shigella species and serotypes circulating in mainland China during 2001–2010.

Methods

Results

We identified a total of 1,843 articles by literature search using different combinations of key terms from the databases, as shown in Figure 1. After exclusion based on title and abstract evaluation, 303 articles (9 in English and 294 in Chinese) were retrieved for detailed full-text review. Of these, 172 were further excluded according to the exclusion criteria. Detailed information on study identification is available in Table S1. In total, 131 articles (136 studies) were analyzed for the prevalence of Shigella species and/or S. flexneri circulating serotypes. Figure 1 shows the detailed selection process and the number of articles included/excluded in each phase.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Flow diagram of study identification.

https://doi.org/10.1371/journal.pone.0052515.g001

As shown in Figure 2 and Table S2, the included 136 studies were conducted in 29 of 31 provinces in mainland China. However, only 63 studies from 19 provinces provided detailed information on the prevalent S. flexneri serotypes (Table S2). Sample sizes of the studies ranged from 50 to 1,466, and a total of 30,959 Shigella isolates were analyzed in the 136 studies. More studies were performed in East (61) and North (24) China, as compared to the other geographic regions. The number of studies from Zhejiang, Shanghai and Beijing were 20, 19 and 15, respectively, whereas only one study was available for each of the other eight provinces (Table S2). In addition, 61.8% of the studies (84/136) were conducted in the 10 provinces with an average GRP per capita that was higher than the national GDP per capita, during 2001–2010 (abbreviated as “top 10 economic provinces” thereafter).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Geographic demarcation of mainland China, color coded as follows: red, East; orange, North; yellow, Middle; green, South; blue, Northwest; cyan, Southwest; violet, Northeast.

Two provinces without available data are marked with light grids. The top 10 economic provinces are highlighted by stripes.

https://doi.org/10.1371/journal.pone.0052515.g002

The meta-analyses of the prevalence of Shigella species circulating throughout mainland China during the past decade are shown in Table 1. S. flexneri and S. sonnei were found to be the two predominant species, with summarized prevalences of 76.2% (95% CI, 73.7%–78.5%) and 21.3% (95% CI, 19.0%–23.7%), respectively. However, evident heterogeneity was observed for both species. Moreover, the prevalence of the two species showed clear diversity in stratified analyses by geographic area, regional economic status, study period, and medical treatment type. For example, S. flexneri and S. sonnei individually accounted for 64.6% (95% CI, 57.2%–71.4%) and 34.9% (95% CI, 28.2%–42.3%) of shigellosis cases in North China, respectively, but the proportions were 85.9% (95% CI, 79.4%–90.6%) and 9.4% (95% CI, 5.8%–15.0%), respectively, in Northwest China. In general, a lower rate of S. flexneri cases and a higher rate of S. sonnei cases were observed in studies from East, North and Northeast China, as compared to the other regions (Table 1). In addition, a similar proportion of differences in the prevalent Shigella species also existed for the top 10 economic provinces, as compared to the other 21 provinces in mainland China. Interestingly, though Beijing and Shanghai are the two biggest modern cities in China and have very similar economic statuses, significant differences in the proportions of dominant Shigella species were observed (p<0.001); S. flexneri accounted for only 56.0% of shigellosis cases in Beijing and 70.2% in Shanghai, whereas S. sonnei accounted for 43.3% of shigellosis cases in Beijing and only 29.6% in Shanghai (Table 1). Moreover, the prevalence of S. flexneri in 2006–2010 was significantly lower than that in 2001–2005; in contrast, the situation for S. sonnei was just the opposite (p<0.001 for both). Similarly, S. flexneri infections constituted 67.1% of shigellosis cases in outpatients and 88.4% in inpatients, whereas S. sonnei infections constituted 29.9% of shigellosis cases in outpatients and only 7.4% in inpatients (Table 1). Figure S1 shows forest plots of meta-analyses on the prevalence of S. flexneri and S. sonnei.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Prevalence of causative species of shigellosis in mainland China, 2001–2010.

https://doi.org/10.1371/journal.pone.0052515.t001

S. flexneri is subdivided into 15 serotypes, whereas S. sonnei is clonal. Therefore, we further investigated the distribution of the causative serotypes of S. flexneri episodes (Table 2). S. flexneri 2a (21.5%, 95% CI, 16.7%–27.4%) was still the predominant serotype among S. flexneri, followed by S. flexneri 4c (12.9%, 95% CI 9.8%–16.9%), S. flexneri 1a (11.2%, 95% CI, 9.4%–13.3%), S. flexneri 2b (7.6%, 95% CI, 5.9%–9.7%) and S. flexneri x (4.7%, 95% CI, 3.4%–6.4%). However, high heterogeneity was observed in the results of most serotypes. For example, S. flexneri 2a presented as the dominant serotype in only 21 of the 63 studies included. Instead, S. flexneri 4c was the found to replace S. flexneri 2a as the predominant serotype in 19 studies. Furthermore, S. flexneri 4c was completely not detected in more than half of the included studies (Table S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of S. flexneri serotypes in mainland China, 2001–2010 (63 studies included).

https://doi.org/10.1371/journal.pone.0052515.t002

Discussion

This review addressed the status of the prevalent species and serotypes of shigellosis in mainland China during 2001–2010. One hundred thirty-one relevant articles, 3 in English and 128 in Chinese, were identified and summarized by meta-analysis. Our estimates showed that S. flexneri (76.2%, 95% CI, 73.7%–78.5%) and S. sonnei (21.3%, 95% CI, 19.0%–23.7%) were the two major causative species of Shigella circulating in mainland China during the past decade. Our results are generally in agreement with previous estimates from the national surveillance data from 2009 [12].

Mainland China consists of 31 provinces with over 1 billion people living in diverse areas with varying geographies, environments, economies, customs, and even climates. Therefore, geographic heterogeneity can be expected in the proportions of the aforementioned two major Shigella species. Indeed, significantly higher rates of S. sonnei cases and lower rates of S. flexneri cases were observed in East, North and Northeast China, as compared to those in Middle, South, Northwest and Southwest China (Table 1). Using the decade average of GRP per capita as an indicator, we ranked the economic status of each province in mainland China during the past decade (data not shown). Similar to the geographic differences discussed above, significantly higher rates of S. sonnei cases and lower rates of S. flexneri cases were also observed in the top 10 economic provinces, as compared to those in the other 21 provinces (Table 1). Furthermore, these results are internally consistent given that among the top 10 economic provinces in China, five are from the East, three are from the North, and one is from the Northeast, whereas only one province is from South China (Figure 2). Hence, the observed geographic differences in the proportions of the two causative Shigella species in mainland China are likely associated with the overall economic status of the different regions.

Historically, S. flexneri is known to be predominant in developing countries. In this study we observed that more than 70% of shigellosis was caused by S. flexneri in most geographic regions of mainland China (Table 1). Interestingly, the overall proportion of S. flexneri decreased by ∼10%, whereas that of S. sonnei increased by ∼10% during the past decade, as compared to the data from 10 years ago (1991–2000) [11]. Nevertheless, the previous estimates were produced from a systematic review rather than a meta-analysis as performed in this study; thus, direct comparison of the two values might not be valid. Our stratified analysis based on study duration indicates the same trends of causative species proportion changes of shigellosis in the past decade: S. flexneri cases decreased from 82.0% (95% CI, 79.3%–84.4%) in 2001–2005 to 67.3% (95% CI, 62.1%–72.1%) in 2006–2010, while S. sonnei cases increased from 15.7% (95% CI, 13.3%–18.3%) in 2001–2005 to 29.9% (95% CI, 25.3%–35.0%) in 2006–2010. Therefore, an undeniable tendency for S. flexneri cases to decease and S. sonnei cases to increase over time persisted in mainland China during the past decade. Cases of S. sonnei are also reported to have surpassed S. flexneri in newly industrialized countries, such as Thailand, South Korea and Taiwan [3], [6], [7]. Therefore, the observed causative species transition of shigellosis in mainland China might be linked to improvements in economic level, considering that China has experienced rapid economic growth during the past decade.

It is noteworthy that both geographic and temporal analyses implied that the prevalent species of Shigella are well associated with regional economic status. However, significant differences in the proportions of S. flexneri and S. sonnei were observed between Beijing and Shanghai, two modern cities with similar economic levels. These findings suggest that additional factors are likely contributing to the proportions of predominant Shigella species. Nonetheless, the essential factors that are driving the phenomenon are not yet known. Determination of the key factor(s) that influence the regional dominant Shigella species is critical for the development of better global shigellosis control and prevention practices in the future. As the Shigella species transition trend currently observed in China is likely to remain in the near future, the current and follow-up statuses of the populations of the bacteria, as well as their hosts in China, might be good targets for further investigations.

S. flexneri serotype 2a is known to have been responsible for 50–70% of endemic and pandemic shigellosis in mainland China in the last century [13]. Previous review also revealed that serotype 2a accounted for 80% of S. flexneri episodes during 1991–2000 [11]. However, our results indicate that S. flexneri 2a remained the dominant serotype in only one third of the studies, and the overall proportion has dramatically decreased to about 20% during the past decade. The decrease of S. flexneri 2a cases might be associated with the bivalent S. flexneri 2a and S. sonnei vaccine licensed in China in 1997 [8], [10], as the vaccination campaign may represent a new selective pressure on the pathogen [14]. Nevertheless, the current Shigella vaccine in China is an optional rather than obligatory vaccine in the national immunization program, and the information on vaccination rates is unavailable publicly. Therefore, further evaluations of the impact of the vaccine on Shigella serotype replacement in China will be an interesting focus in future studies.

Notably, the newly emerged S. flexneri 4c overtook S. flexneri 2a as the predominant serotype in approximately one third of the studies. S. flexneri 4c was first identified in Romania in the 1960s and was observed to differ biochemically from S. flexneri 4a and 4b representatives [15]. Moreover, S. flexneri 4c was found to account for as much as 17.2% of all Shigella isolates in the former Union of Soviet Socialist Republics (USSR) in the 1980s [15]. A recent study reported that fluoroquinolone-resistant S. flexneri 4c clones were isolated from East China [16]. So further investigations on S. flexneri serotype 4c circulated in China are urgently required. In addition, shigellosis cases caused by several other serotypes of S. flexneri, including serotypes 1a (11.2%), 2b (7.6%) and x (4.7%), are also significantly increased during the past decade as compared with previous results [11]. Therefore, to achieve better immune protections in China these additional causative S. flexneri serotypes should be incorporated into further Shigella vaccine developments.

The etiology distribution variation in each medical treatment type (clinic and hospital) may serve as a surrogate of disease severity of different Shigella species. Our results showed that the prevalence of S. flexneri in inpatients was significantly higher than that in outpatients; in contrast, the situation for S. sonnei was just the opposite (p<0.001 for both) (Table 1). This is in good agreement with the previous genetic results that S. sonnei is generally less virulent than S. flexneri from the genomic viewpoint [17]. Nevertheless, the threats from S. sonnei in mainland China should not be underestimated. Among the 29 studies addressing outbreaks that were excluded from the meta-analysis, 27 studies performed serological analyses, and S. flexneri and S. sonnei were responsible for 16 (59.3%) and 11 (40.7%) of the outbreaks, respectively. In addition, instead of a single or several predominant serotypes as might be expected, eight different serotypes of S. flexneri were confirmed to be associated with the 11 outbreaks that done detailed serotyping (data not shown). The results further highlight the aforementioned necessity to improve the valency coverage of currently conceived Shigella vaccines.

A recent multicenter study in six Asian countries found that a surprising 92% of shigellosis in China was caused by S. flexneri [3]. However, the surveillance was conducted in a rural area of China in 2002, which is far from representative of the current situation of Shigella prevalence in mainland China. The meta-analysis presented here combines data across studies to estimate the most current status of Shigella infections in China with more precision than is possible in a single study. Nevertheless, some limitations of this review should be considered. First, our results may not be fully representative of mainland China. The species information was not available for two provinces, both of which are in west of China (Figure 2). In addition, S. flexneri serotype information was not available, or did not meet the inclusion criteria, for 12 provinces in this analysis. Even for the provinces included by this study, evident discord was observed in a number of studies included for each province. However, we tried to ensure that the most valid studies were included through restrictive inclusion criteria. Second, several helpful stratify analyses were restricted by the limited information available from the original publications. For example, the primary targets of Shigella vaccines being developed are children under 5 years old, so the epidemiology of species/serotypes by meaningful age category would be very informative. However, only 15 of the 136 studies provided information on age of patients, and these studies employed different grouping criteria, which makes the stratify analysis using the basic age category can only include three studies (Table S2). Such data is obviously far from representing meaningful results. Third, several potential biases should be kept in mind for results interpretation. Only 3 of the 136 studies included were population-based; thus, the selection of subjects might have made the results prone to potential selection bias. Though the species/serotypes identification in all studies included was confirmed by both biochemical reactions and serological results, the sensitivity and coverage of the antisera sets used by each study may variant, which might have biased the results. The commercial Shigella antisera are usually available in different sets that offer diverse coverage of species/serotypes to accommodate the requirements of different users. But the details of the antisera set used in each study is generally absent from the publications, leading to the exclusion or evaluation of the potential bias impossible. In addition, historical improvements of the sensitivity in serological testing might have led to more accurate serotyping in recent studies than in previous work. Forth, the antimicrobial resistance patterns of Shigella isolates could not be analyzed due to the limited information available from the original publications.

To our knowledge, this is the first study to systematically review the status of circulating strains of Shigella in mainland China during the past decade. Our results reveal significant changes in the causative species and serotypes of shigellosis in recent years and emphasize the necessity to focus more attention on S. sonnei and S. flexneri serotype 4c episodes in future shigellosis prevention and control measures, as well as in basic research. In addition, our study highlights the limited information about Shigella epidemiology in many provinces of China with relatively low economic status. To achieve a precise estimate, further large-scale population-based surveillance studies using standardized methods are required.

Supporting Information

Acknowledgments

We are grateful to Lei Gao for inspirational discussion and assistance in meta-analyses.