Characteristics of Multidrug Resistant Shigella and Vibrio cholerae O1 Infections in Patients Treated at an Urban and a Rural Hospital in Bangladesh

We determined the frequency of multidrug resistant (MDR) infections with Shigella spp. and Vibrio cholerae O1 at an urban (Dhaka) and rural (Matlab) hospital in Bangladesh. We also compared sociodemographic and clinical features of patients with MDR infections to those with antibiotic-susceptible infections at both sites. Analyses were conducted using surveillance data from the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b), for the years 2000–2012. Compared to patients with antibiotic-susceptible for Shigella infections, those in Dhaka with MDR shigellosis were more likely to experience diarrhea for >24 hours, while, in Matlab, they were more likely to stay inhospital >24 hours. For MDR shigellosis, Dhaka patients were more likely than those in Matlab to have dehydration, stool frequency >10/day, and diarrheal duration >24 hours. Patients with MDR Vibrio cholerae O1 infections in Dhaka were more likely than those in Matlab to experience dehydration and stool frequency >10/day. Thus, patients with MDR shigellosis and Vibrio cholerae O1 infection exhibited features suggesting more severe illness than those with antibiotic-susceptible infections. Moreover, Dhaka patients with MDR shigellosis and Vibrio cholerae O1 infections exhibited features indicating more severe illness than patients in Matlab.


Introduction
Shigella and Vibrio cholerae O1 are widely recognized causes of dysentery and acute watery diarrhea, respectively [1,2]. Both have been responsible for producing epidemics [3] and often require antibiotic treatment to mitigate the severity of disease [4][5][6]. For Shigella infections, in particular, increasing antibiotic resistance has led to fewer antibiotics capable of producing bacteriostatic or bactericidal minimum inhibitory concentrations (MICs) [5,7]. The challenge is compounded by the fact that while the regional prevalence of infection may be similar, rates of antibiotic resistance may differ substantially from one nation to another [5,8].
Since 1979 and 2000, the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b), has tested a systematic random sample of patients seeking care at an urban Dhaka Hospital and rural Matlab Treatment Centre for a spectrum of diarrhea-causing pathogens, including Shigella and Vibrio cholerae O1 [9]. Antimicrobial susceptibility patterns for these two pathogens are determined to inform clinicians about appropriate antibiotic treatment options [10]. Underscoring the need for antimicrobial sensitivity testing is the emergence of multidrug resistance (MDR), defined here as isolates resistant to ≥3 drugs [11][12][13]. Infections due to MDR strains are important not only because they are more difficult to treat but also because they may lead to higher fatality rates [14,15] Figure 1 illustrates the sampling frame for the study.

Laboratory Methodology.
Fresh whole stool specimens collected from patients were examined at either the central icddr,b laboratory in Dhaka or at the Matlab clinical laboratory. Using standard laboratory methods described elsewhere [16,17], each specimen was screened for common enteric pathogens, including Shigella spp. and Vibrio cholerae O1. Bacterial susceptibility to antimicrobial agents was determined by the disk diffusion method as recommended by the Clinical Laboratory Standards Institute (CLSI 2010, June update) with commercial antimicrobial discs (Oxoid, Basingstoke, UK) [18]. The antibiotic discs used in this study for Shigella spp. included ampicillin (10 g), . We compared differences in proportions using the Chi-square test. A probability value (P value) of <0.05 was considered to confer statistical significance. Magnitudes of association were determined by estimating odds ratios (OR) and 95% confidence intervals (CI). To identify MDR strains, we determined the frequency with which isolates exhibited resistance to ≥3 antibiotics. We first compared sociodemographic features of patients with MDR infections versus those with susceptible infections. Next, focusing solely on MDR infections, we determined whether patients treated in Dhaka differed from those in Matlab in terms of sociodemographic or clinical features. All statistically significant differences ascertained by univariate analysis were entered into a logistic regression model to calculate adjusted odds ratios, P values, and 95% confidence intervals.

Shigella spp.
From 2000 to 2012, the number of patients who yielded MDR Shigella spp. in Dhaka and Matlab was 371/1,232 (30%) and 552/1,728 (32%), respectively ( Figure 1). S. flexneri was recovered from a greater proportion of patients in Matlab compared to Dhaka, whereas S. boydii was recovered from a greater proportion of patients in Dhaka versus Matlab (Table 1). Resistance to ampicillin + nalidixic acid + trimethoprim-sulfamethoxazole was the most common resistance pattern for patients treated at both Dhaka and Matlab (Table 2). In Dhaka, the percentages of isolates resistant to individual antibiotics were trimethoprimsulfamethoxazole (94%), ampicillin (85%), nalidixic acid (91%), mecillinam (25%), and ciprofloxacin (31%). In Matlab, the percentages of isolates resistant to individual antibiotics were nalidixic acid (96%), trimethoprim-sulfamethoxazole (92%), ampicillin (87%), ciprofloxacin (32%), and mecillinam (12%). In Dhaka, compared to patients with antibioticsusceptible Shigella infections, those with MDR infections were more likely be male and to experience diarrhea for >24 hours (Table 3). In Matlab, patients with MDR Shigella infections were more likely than those with susceptible infections to stay in hospital >24 hours and to report having used antimicrobials at home prior to coming to hospital ( Table 4). For MDR Shigella infections, patients in Dhaka were more likely than those in Matlab to be male and to have dehydration, frequency of stools >10/day, and duration of diarrhea >24 hours (Table 5).
Patients with MDR Vibrio cholerae O1 infections in Dhaka were more likely than those in Matlab to experience a frequency of stools of >10/day, dehydration, and the presence in stools of 1-10 macrophages per high power field (Table 6).

Discussion
To our knowledge, few studies have assessed differences in sociodemographic and clinical features of multidrug resistant versus susceptible infections due to Shigella spp. within a single nation or to compare these features for multidrug resistant infections due to Shigella spp. and Vibrio cholerae O1 at an urban versus a rural treatment center [19]. Although multidrug resistant infections are a problem globally, the ease of availability of antibiotics to the public at large in nations such as Bangladesh may expedite the rate at which resistance develops [20]. In the present study, the proportion of multidrug resistant Shigella and Vibrio cholerae O1 isolates recovered from patients in urban Dhaka was similar to that in rural Matlab. Of concern, in both hospital settings, there was evidence of resistance to mecillinam and ciprofloxacin, drugs commonly used for the treatment of shigellosis and, in the case of ciprofloxacin, cholera. In general, the emergence  (7) -   of resistance poses a number of challenges by leading, in some instances, to an increase in morbidity and mortality and longer hospital stays, as a result of inadequate initial therapy or increased virulence [15]. Moreover, antibiotic resistance reduces choices for therapy and can cause health care costs to rise due to the need to use antimicrobial agents that are more expensive than those in current use. While antibiotic resistance in both Shigella spp. and Vibrio cholerae O1 has serious ramifications, for Shigella spp. the phenomenon carries added weight given that susceptibility rarely returns after resistant strains have become endemic in a region [12,21]. Although S. dysenteriae 1 frequently develops resistance to new antibiotics initially, resistance is often acquired subsequently in the other Shigella species. In contrast, Vibrio cholerae O1 strains often revert to antibiotic susceptibility [12]. This phenomenon was observed during the study period when, abruptly in late 2004, Vibrio cholerae O1 isolates at both Matlab and Dhaka demonstrated tetracycline resistance; however, two years later, in 2006, tetracycline susceptibility reappeared in large part [22].
We observed evidence that multidrug resistant infections due to Shigella spp. were associated with more severe illness compared to non-MDR strains. For example, patients in Dhaka from whom multiresistant Shigella spp. were recovered significantly were more likely to experience diarrhea for >24 hours compared to patients from whom susceptible isolates were recovered. Similarly, in Matlab, patients from whom multidrug resistant Shigella spp. was recovered are more likely than those with susceptible infections to stay in hospital >24 hours and to report having used antimicrobials at home prior to coming to hospital. We also found evidence of differing degrees of severity of multidrug resistant infections due to Shigella spp. and Vibrio cholerae O1 infections depending on whether patients resided in urban or rural areas 6 ISRN Microbiology  In the meantime, we hypothesize that differences in sociodemographic, nutritional, and economic characteristics of urban versus rural populations may, in part, explain our observations, as may differences in sources of drinking water and water-sanitation practices.
The present study has several limitations. Hospital-based data of the type incorporated in the Diarrheal Disease Surveillance System may not adequately represent the ill population at large. For example, a segment of the population in Bangladesh, particularly in rural regions, is known to use medicinal plants and traditional healers as a first-line of health care to cure gastrointestinal disorders [23]. Consequently, it is possible that sociodemographic and clinical features of patients with shigellosis and cholera described here may not be completely representative of all infections that occurred in the study area and study period. In addition, the results of antibiotic resistance testing we used were largely qualitative-presence, absence-as opposed to a quantitative nature as provided by minimum inhibitory concentrations. Thus, we were unable to ascertain pathogen-antibiotic relationships, wherein resistance levels may have approached limits known to confer resistance. Finally, our results did not incorporate the role of virulence factors or genetic typing that could have added valuable insight into outcomes of interest. Notwithstanding these limitations, we believe that our results justify further research to determine, first, whether multidrug resistant strains of Shigella and Vibrio cholerae O1 are associated with more severe infections than antibioticsusceptible strains, and, if so, what factors increase the severity of infections, and second, whether multidrug resistant infections caused by these two pathogens produce more severe infections in urban versus rural regions. Answers to these questions may help in efforts to prevent infections due to Shigella spp. and Vibrio cholerae O1, both of which are often leading bacterial causes of diarrhea in developing nations.