INTRODUCTION
Diarrhoeal diseases constitute a major public health problem in the community of both developing and developed countries [Reference Steiner, Samie and Guerrant1]. Different diarrhoeagenic pathogens, including bacteria, viruses, and parasites, have been identified as the aetiological agents of diarrhoea with most of them being highly related to the consumption of contaminated food products [Reference Scallan2]. The features of diarrhoeagenic pathogens may vary from place to place depending on dietary habits, local meteorology, geography, and socioeconomic elements [Reference Yamashiro3, Reference Talan4]. Epidemiological characterization of these pathogens in the community is the initial step towards introduction of correct interventions for controlling the diseases. Several countries have conducted different prospective population-based or cross-sectional studies to estimate the overall number of foodborne illnesses [Reference Flint5]. Recently, powerful active surveillance networks for diarrhoeagenic pathogens have been established and improved, especially in countries with extensive health resources [6–Reference Siebenga8].
However, until now, only limited information has been available on the prevalence of the diarrhoeal aetiological spectrum and antimicrobial resistance patterns of major diarrhoeagenic bacterial pathogens in China as there is no requirement for hospitals to identify pathogens from stool samples of outpatients. Recently, China CDC has initiated a Diarrhoea Surveillance Network in several provinces, including Henan. The purpose of this study is to describe the prevalence of the main aetiological infectious agents of diarrhoea and the antimicrobial susceptibility of Salmonella and Shigella from outpatients in Henan province, China.
MATERIALS AND METHODS
Stool specimen collection
Twenty hospitals in Zhengzhou city, Shangqiu city and Suiyang county of Henan province were recruited in this study. These hospitals consisted of tertiary (n = 2), secondary (n = 6), and primary (n = 6) hospitals and community health centres (n = 6). These hospitals represented the most popular hospital categories in Henan province and were also selected as the sentinel hospitals in the Diarrhoea Surveillance Network of China CDC. The population of Henan province was 99·18 million in 2008, which ranked highest in China. There are 1192 hospitals and 2089 community health centres in this province. The population of Zhengzhou city or Shangqiu city is more than five million and the population of Suiyang county is >500 000. Faecal specimens from outpatients with diarrhoea (defined by World Health Organization guidelines as passage of >3 loose or liquid stools in a 24-h period) were collected from January to December in 2007. From June to September, ⩽30 faecal specimens were collected from each hospital or health centre every month and ⩽10 specimens were collected from each hospital or health center in the remaining months. Faecal specimens were collected from each outpatient before antimicrobial treatment. Each day, only the first eligible stool specimen from the outpatient who presented with diarrhoea at each recruiting hospital was collected. Faecal swabs (Cary–Blair transport swabs; Copan, Italy) from each outpatient were shipped to the local CDC laboratiry for bacterial pathogen isolation, and ten ∼30 g faecal specimen from each outpatient were shipped to the provincial CDC laboratories for virus analysis. Questionnaires regarding information of clinical symptoms as well as general data such as age, gender and place of residence were collected.
Pathogen analysis
Each outpatient contributed ⩽1 specimen, and each specimen was tested for 12 diarrhoeagenic pathogenic species, including Aeromonas hydrophila, Campylobacter spp., diarrhoeagenic Escherichia coli, Salmonella spp., Shigella spp., Vibrio spp., Yersinia enterocolitica, astrovirus, adenovirus, noroviruses, rotavirus and sapovirus. The presence of pathogenic bacteria was tested for by routine culture assays on appropriate selective media as follows: diarrhoeagenic E. coli, MacConkey agar (BD Diagnostics, China); Salmonella spp., BBL™ CHROMagar™ Salmonella (BD Diagnostics); Shigella spp., Salmonella-Shigella agar (BD Diagnostics); Campylobacter spp., CCDA agar (Oxoid, China); A. hydrophila and Vibro spp., thiosulfate-citrate-bile-sucrose agar (BD Diagnostics); Y. enterocolitica, Yersinia-selective agar (BD Diagnostics). Campylobacter plates were incubated microaerophilically at 42 ± 1 °C; Yersinia plates were incubated aerobically at 27 ± 1 °C; all other plates were incubated aerobically at 35 ± 1 °C. From each plate, ⩽3 suspected colonies were selected and confirmed by API strip tests (bioMérieux, France). Salmonella and Shigella isolates were further serotyped with commercially available antiserum (Statens Serum Institute, Denmark) through slide agglutination according to the manufacturer's instructions. Pathotypes of diarrhoeagenic E. coli were identified through a multiplex PCR method, including enteroaggregative E. coli (EAEC), enterohaemorrhagic E. coli (EHEC), enteroinvasive E. coli (EIEC), enteropathogenic E. coli (EPEC) and enterotoxigenic E. coli (ETEC) [Reference Muller9]. This multiplex PCR amplifies the following target genes: uidA, escV, bfpB, stx 1, stx 2, elt, estIa, estIb, invE, astA, aggR, and pic. Adenovirus, astrovirus, norovirus, rotavirus and sapovirus in faecal specimens were detected by reverse transcription multiplex PCR methods [Reference Yan10, Reference Yan11].
Antimicrobial susceptibility testing
Antimicrobial susceptibility of all Salmonella and Shigella isolates was determined via the disc diffusion method, and interpreted according to the Clinical and Laboratory Standards Institute (CLSI) interpretive standards [12]. Susceptibility to seven antimicrobials was measured for all isolates, including ampicillin (AMP, 10 μg), cefotaxime (CTX, 30 μg), chloramphenicol (CHL, 30 μg), ciprofloxacin (CIP, 5 μg), nalidixic acid (NAL, 30 μg), tetracycline (TET, 30 μg) and trimethoprim–sulfamethoxazole (SXT, 1·25/23·75 μg).
Statistical analysis
The differences of antimicrobial resistance frequencies were analysed by χ2 test and Mantel–Haenszel test using Epi Info v. 6 software (www.cdc.gov/epiinfo/).
RESULTS
From January to December 2007, 1526 stool specimens from individual outpatients with diarrhoea were collected and tested for bacterial and viral pathogens from Zhengzhou city (n = 574), Shangqiu city (n = 498) and Suiyang county (n = 454). The outpatients comprised 515 females and 1011 males that ranged in age from <1 month to 98 years (median age 3 years, mean age 16·4 years). Of 1526 outpatients, 682 subjects had watery diarrhoea, 574 subjects had abdominal pain, 389 subjects had fever (⩾38 °C) and 182 subjects had blood in their stool (bloody diarrhoea). A total of 1228 subjects exhibited at least one of the above symptoms, 500 outpatients exhibited more than one symptom, and 102 outpatients exhibited more than two symptoms (Table 1).
Pathogens of 12 diarrhoeagenic pathogenic species were detected in 432/1526 (28·3%) stool specimens. Bacterial pathogens of seven species were recovered from 279/1526 (18·3%) stool specimens and viral pathogens of five species were detected in 178/1526 (11·7%) stool specimens; more than two pathogenic species were found in 60/1526 (3·9%) stool specimens. Both bacterial and viral pathogens were found in 25/1526 (1·6%) stool specimens. The top five pathogenic species were diarrhoeagenic E. coli (n = 140, 9·2%, including EAEC, EHEC, EIEC, EPEC and ETEC), rotavirus (n = 79, 5·2%), Shigella spp. (n = 69, 4·5%), Salmonella spp. (n = 66, 4·3%) and norovirus (n = 56, 3·7%) (Tables 1 and 2). The median age of patients infected with a bacterial enteric pathogen was 8 years, compared to 1 year for those infected with a viral enteric pathogen and 4 years for outpatients with no pathogen detected.
EPEC, Enteropathogenic E. coli; EHEC, enterohaemorrhagic E. coli; ETEC, enterotoxigenic E. coli; EIEC, enteroinvasive E. coli; EAEC, enteroaggregative E. coli.
–, Not detected.
Values given are n (%).
Stool specimens collected in spring and summer were more likely to yield bacterial pathogens (19·6% and 21·6%, respectively), and those collected in autumn were more likely to yield viral pathogens (23·3%). Stool specimens collected in spring also showed higher prevalence for rotavirus (n = 10, 6·1%). The pathogen prevalence in three regions showed significant difference: 24·8% (113/454) of specimens from Suiyang county had a positive bacterial culture result, compared to 13·9% (69/498) and 16·9% (97/574) of those from Shangqiu city and Zhengzhou city (P < 0·01); 16·0% (92/574) of specimens from Zhengzhou city had a positive viral test result, compared to 7·0% (35/498) and 11·2% (51/454) of those from Shangqiu city and Suiyang county (P < 0·01).
The prevalence of clinical symptoms showed a correlation with the distribution of pathogen categories. Outpatients with a positive bacterial culture had a higher prevalence of abdominal pain (119/279, 42·7%) and bloody diarrhoea (49/279, 17·6%), compared to 13·5% (24/178) and 7·9% (14/178) of those with a positive viral test result (P < 0·01). Outpatients infected by Shigella spp. showed a significantly higher prevalence of bloody diarrhoea (28/69, 40·6%) compared to other diarrhoeagenic pathogens (P < 0·01). However, outpatients with a positive viral result had a higher prevalence of watery diarrhoea (115/178, 64·6%), compared to those infected with a bacterial enteric pathogen (99/279, 35·5%) or no pathogen detected (478/1094, 43·7%) (P < 0·01).
The prevalence of pathogen categories also showed a correlation with outpatient age. Stool specimens from outpatients aged <3 years showed similar prevalence of bacterial (113/749, 15·1%) and viral (136/749, 18·2%) pathogens, but a higher prevalence of bacterial pathogens was observed in other age groups (Table 3).
EPEC, Enteropathogenic E. coli; EHEC, enterohaemorrhagic E. coli; ETEC, enterotoxigenic E. coli; EIEC, enteroinvasive E. coli; EAEC, enteroaggregative E. coli.
–, Not detected.
Values given are n (%).
Salmonella serotyping and antimicrobial susceptibility
Of 66 Salmonella isolates, 13 distinct serotypes were identified, including two dominant serotypes: S. Typhimurium (n = 23, 34·8%) and S. Enteritidis (n = 16, 24·2%). Minor serotypes included Norwich (n = 7), London (n = 5), Derby (n = 5), Stanley (n = 2) and Bovismorbificans (n = 2). Only one isolate was recovered for each of the following serotypes: Agona, Meleagridis, Sanktmarx, Schwarzengrund, Senftenberg and Tennessee.
Of 66 Salmonella isolates, 15 (22·7%) were susceptible to all tested antimicrobials. Resistance to nalidixic acid was the most common (n = 48, 72·7%), followed by tetracycline (n = 34, 51·5%), ampicillin (n = 29, 43·9%), trimethoprim-sulfamethoxazole (n = 27, 40·9%) and chloramphenicol (n = 24, 36·4%). The antimicrobial-resistant profiles differed among serotypes. Of 23 S. Typhimurium isolates, 17 (73·9%) were resistant to at least four antimicrobials. Whereas, in 16 S. Enteritidis isolates, only three (18·9%) were resistant to four or more antimicrobials. Of 27 isolates of minor serotypes, 10 (37·0%) were susceptible to all tested antimicrobials and seven (25·9%) isolates were resistant to at least four antimicrobials. In total, 13 ciprofloxacin-resistant Salmonella isolates were recovered, including Typhimurium (n = 9), Norwich (n = 2), Enteritidis (n = 1) and Senftenberg (n = 1). All these 13 ciprofloxacin-resistant isolates were also resistant to four non-quinolone antimicrobials. Two isolates of different serotypes (Enteritidis and Senftenberg) were resistant to cefotaxime and the Senftenberg isolate also showed resistance to ciprofloxacin.
Shigella serotyping and antimicrobial susceptibility
Of 69 Shigella isolates, three dominant serotypes were identified, including S. flexneri serotype 2a (n = 21, 30·4%), S. flexneri serotype 4c (n = 13, 18·8%) and S. flexneri serotype 1a (n = 12, 17·4%). Minor serotypes included S. flexneri serotype fx (n = 7, 10·1%), S. sonnei (n = 7, 10·1%) and S. flexneri serotype 2b (n = 6, 8·7%). Only one isolate was recovered for each of the following serotypes: S. flexneri serotype f1b, S. flexneri serotype f4a and S. flexneri serotype fy.
Of 69 Shigella isolates, resistance to nalidixic acid was the most common (69, 100%), followed by tetracycline (67, 97·1%), ampicillin (64, 92·8%), chloramphenicol (60, 87%), trimethoprim–sulfamethoxazole (40, 58·0%) and ciprofloxacin (n = 25, 36·2%). All 69 Shigella isolates were resistant to at least two antimicrobials, and 65 (94·2%) isolates were resistant to at least four antimicrobials. Four dominant antimicrobial-resistant profiles were identified: AMP-CHL-NAL-TET (n = 18), AMP-CHL-NAL-SXT-TET (n = 14), AMP-CHL-CIP-NAL-SXT-TET (n = 14) and AMP-CHL-CIP-NAL-TET (n = 10). Twenty out of 21 S. flexneri serotype f2a isolates were resistant to five (n = 8) or six (n = 12) antimicrobials. All 25 ciprofloxacin-resistant isolates were also resistant to three or four non-quinolone antimicrobials and belonged to three S. flexneri serotypes: S. flexneri 2a (n = 20), S. flexneri 1a (n = 4) and S. flexneri 4c (n = 1). Six cefotaxime-resistant isolates were also resistant to three or four non-β-lactam antimicrobials and belonged to three serotypes: S. sonnei (n = 3), S. flexneri 1a (n = 2) and S. flexneri 2a (n = 1). The cefotaxime-resistant S. flexneri 2a isolate also showed resistance to ciprofloxacin and three other non-quinolone antimicrobials.
DISCUSSION
This study provides useful information regarding the aetiologies of diarrhoea and antimicrobial resistance profiles of major diarrhoeagenic bacteria in densely populated developing regions with inadequate medical resources. More than 25% of 1526 stool specimens were found to have known diarrhoeagenic pathogens that have been reported with variable prevalence in different countries [Reference Echeverria13–Reference Vu Nguyen16]. Two major diarrhoeagenic pathogens (Salmonella and Shigella), were characterized by serotyping and antimicrobial susceptibility. Our data showed multiple drug-resistant (MDR) Salmonella and Shigella isolates were widely disseminated in recruited outpatients. To the best of our knowledge, this is the first prospective diarrhoeagenic pathogen spectrum study in outpatients in mainland China.
Although stool specimens were collected from the first outpatient who presented with diarrhoea in each recruiting hospital, 66·3% (1011/1526) of subjects were male and 49·1% (749/1526) of subjects were aged <3 years. This demographic distribution profile was significantly different from other studies [Reference Klein14, Reference Olesen15]. This profile might be a unique demographic pattern for outpatients with diarrhoea in China. Since China implemented the one-child policy, parents take illness in young children very seriously, especially male children. There are now more opportunities for parents to bring their children (especially those aged <3 years) suffering from diarrhoea to see a doctor early in the morning if they have concerns. Further, younger children are more likely to be infected by diarrhoeagenic pathogens because of their dietary habit and immature immune system. Moreover, since most diarrhoeagenic pathogen infections are self-limiting, adults with diarrhoea may not seek medical help or visit a doctor in small clinics or obtain medicines over the counter. As a result, the data of this study might be biased because more diarrhoea cases of children aged <3 years were included.
In this study, 28·3% of 1526 stool specimens were found to have known diarrhoeagenic pathogens and the top five pathogenic species were diarrhoeagenic E. coli (n = 140, 9·2%), rotavirus (n = 79, 5·2%), Shigella spp. (n = 69, 4·5%), Salmonella spp. (n = 66, 4·3%) and norovirus (n = 56, 3·7%). The total diarrhoeagenic pathogen-positive ratio and ranking pattern were different from other studies on the aetiology of diarrhoea; in particular, Campylobacter and viral pathogens showed significantly lower prevalence in outpatients with diarrhoea than other studies [Reference Olesen15, Reference Scallan17]. The differences in the spectrum of diarrhoeal agents were critically dependent on the employed technology. Since a variety of bacteria species existed in the stool specimens, bacteria similar to the target pathogens may interfere with the recovery of these pathogens. Appropriate selective enrichment step adoption may help to change the ratio of diarrhoeagenic bacterial pathogens in the media and improve the recovery of target pathogens. Other selective methods, such as immunomagnetic beads and better viral RNA extraction methods may also help to improve pathogen detection. In future studies, new techniques should be applied to improve the positive rate of the target pathogens. Another reason might be differences in healthcare-seeking behaviours of the subjects. In this study, close to 50% of the subjects were aged <3 years and two thirds were male, which was different from other studies [Reference Talan4, Reference Olesen15, Reference Erdem18]. Moreover, our study population may differ from the population of other studies because of geographical and dietary habit variations in prevalence of pathogens.
Our data showed outpatients aged >50 years continued to have the highest incidence of bacterial infections. Studies have shown infected elderly subjects are at highest risk for hospitalization and death from most of these infections [Reference Schiller19]. MDR Salmonella and Shigella were especially widely disseminated in these regions which made prompt diagnosis and treatment, as well as careful attention to food safety, particularly important in this age group.
As expected, rotavirus was the most common pathogen (n = 79, 5·2%), especially for children in autumn. The appreciable detection rates of astroviruses and sapovirus, as reported in Australia, the USA, and France [Reference Palombo and Bishop20–Reference Rodriguez-Baez22], suggests that these viruses are also important causes of diarrhoea, especially in children and older subjects (Table 2).
Our data showed that Shigella and S. Typhimurium isolates had a strong capacity to accumulate antimicrobial resistance mechanisms. More than 90% of S. flexneri isolates and more than 70% of S. Typhimurium isolates in this study showed MDR (resistant to ⩾3 antimicrobial categories) profiles and were highly resistant to older generation antimicrobials, such as ampicillin, chloramphenicol, nalidixic acid and tetracycline. This rapid accumulation of resistance mechanisms in Shigella and S. Typhimurium has also been described by other studies which reported the acquisition of resistance mechanisms was through mobile genetic elements [Reference Von Seidlein23, Reference Ghosh24]. The precise causes of the observed capacity of resistance mechanism accumulation and high prevalence of MDR Shigella and S. Typhimurium remain unknown. In contrast to Salmonella and E. coli which are widely distributed, humans and a few primates are the only reservoirs of Shigella, and its circulation in humans is through person-to-person contact or consumption of contaminated food, which is facilitated by the low infective dose of Shigella (as few as 10 viable organisms). The prevalence of antimicrobial resistance in Shigella should be highly related to the volume and categories of antimicrobial consumption in the local community. Inappropriate prescription and easy access to antimicrobials by outpatients could facilitate the selection and spreading of MDR Shigella isolates in the local community. Studies have estimated that half of outpatient antimicrobials were prescribed for inappropriate indications, such as viral illness [Reference Gonzales, Steiner and Sande25]. In China, inappropriate prescriptions are more common, especially through intravenous infusion. The over-the-counter sale of antimicrobials without prescription might be another reason for the resistance dissemination. Since most epidemic Shigella isolates in China are MDR, as shown in our data, the abuse of antimicrobials in the community might be the main reason for the wide dissemination of MDR Salmonella and Shigella.
Fluoroquinolones and third-generation cephalosporins were two popular empirical options to treat severe gastrointestinal infections caused by pathogenic bacteria [Reference Gendrel and Cohen26, Reference Guerrant27]. However, more than 70% of Salmonella and 100% of Shigella isolates in this study exhibited resistance to nalidixic acid and decreased susceptibility to ciprofloxacin. These isolates had a high probability of becoming ciprofloxacin resistant, thereby causing treatment failure, as shown in Salmonella infection [Reference Helms28]. Besides these, more than 30% of S. Typhimurium and Shigella spp. isolates showed resistance to ciprofloxacin. Widespread quinolone resistance in Shigella and S. Typhimurium has also been documented as a unique characteristic in other studies of China and some Asian countries with easy access of antimicrobials [Reference Von Seidlein23, Reference Cui29]. However, in countries with strict antimicrobial controls, such as the USA, less than 1% of Shigella or S. Typhimurium isolates were resistant to nalidixic acid and few ciprofloxacin-resistant isolates were identified nationwide [Reference Sivapalasingam30, Reference Herikstad31]. These data further emphasize the necessity of strict antimicrobial application control in developing nations. Besides quinolone resistance, several cefotaxime-resistant Salmonella and Shigella isolates were also identified; even when only two isolates show co-resistance to both ciprofloxacin and cefotaxime the surveillance agencies should remain alert for the spreading of these isolates. The wide distribution of MDR Salmonella and Shigella isolates clearly raises the perennial issue of strict control of antimicrobial prescriptions and improved sanitation rather than new antibiotics, as a long-term solution to the plethora of salmonellosis and shigellosis.
Although this aetiological study involved 1526 subjects, it does have limitations. First, we did not include a case-control group as in other studies [Reference Olesen15, Reference Okeke32]. Without the control group, we could not rule out the carriers, as indicated in other studies, and obtain the true infection rate and aetiological agents of infection. Compared to other studies, a lower ratio of outpatients with diarrhoeal pathogens was found, this might be caused by the techniques we used and the proficiency of staff. The aetiology of high ratio of unexplained diarrhoea warrants further investigation. Future investigations of the aetiology of diarrhoea should include appropriate control populations to assess the frequencies with which pathogens are found in asymptomatic subjects, and they should also incorporate techniques to identify other pathogens not sought in the present study. However, taken together, these data we present demonstrate the need for – and opportunities in – expanded diarrhoeal aetiological investigations in China.
In summary, aetiological data on diarrhoeal diseases and susceptibility patterns of diarrhoeal pathogens are important information for clinical management and control strategic planning. These data on the aetiology of diarrhoea and antimicrobial resistance patterns in China will have significant effect on public health policies and resource prioritization practices.
ACKNOWLEDGEMENTS
This research was supported by a grant (5U2GGH 00018-02) from the China–US Collaborative Program on Emerging and Re-emerging Infectious Diseases and a grant (30972487) from the National Natural Science Foundation of China.
DECLARATION OF INTEREST
None.