During 2002–2010, a total of 12,719 fecal specimens were collected, which represented every fifth hospitalized diarrhea patient at the Infectious Diseases Hospital in Kolkata and all children at the outpatient unit at B.C. Roy Memorial Hospital for Children in Kolkata. Fecal specimens were screened for V. cholerae and other enteric pathogens as described (8). NAGs were serotyped by using 206 polyclonal O antisera according to the protocol developed at the National Institute of Infectious Diseases (Tokyo, Japan) (2).

Antimicrobial drug susceptibility assays were performed by using the disk diffusion method and commercially available disks (Becton Dickinson, Sparks Glencoe, MD, USA), according to standards of the Clinical and Laboratory Standards Institute (9). Because these standards do not include interpretive criteria for V. cholerae, breakpoints for Enterobacteriaceae were adopted. Escherichia coli ATCC 25922 was used as a quality-control strain.

Simplex and multiplex PCRs were performed by using published methods specific for ctxA, tcpA, rtxA, rtxC, stn, and hlyA genes (classical/El Tor) and the TTSS-coding genes (7). Pulsed-field gel electrophoresis was performed according to the PulseNet standardized protocol for V. cholerae (www.pulsenetinternational.org/SiteCollectionDocuments/pfge/5.71_2009_PNetStandProtVcholerae.pdf. Gel Compare II software (Applied Maths NV, Sint-Martens-Latem, Belgium) was used for comparison of electrophoresis patterns. This software uses a Dice similarity index and contains an unweighted pair group with arithmetic mean method.

Of the 12,719 diarrhea feces specimens screened, 2,206 (17.3%) contained V. cholerae, which included V. cholerae O1 in 1,841 (83.4%), and O139 in 84 (3.8%). In the remaining 281 (12.7%) specimens, V. cholerae strains did not agglutinate with serogroups O1 or O139. This result was confirmed by species-specific ompW PCR, which included strains collected during 2003 (7). Among 281 strains, 175 (62.3%) NAGs were the only enteric pathogen found, and 106 (37.7%) of those NAGs were found with other enteric pathogens. The isolation frequency of NAGs ranged from 1.2% to 3.2% (Table 1).

Although V. cholerae O1 is highly prevalent in Kolkata, more NAGs were detected than in our previous study (10). A total of 224 (79.7%) strains were categorized into 80 serogroups; the remaining 57 (20.3%) were untypeable. Among typeable serogroups, 14 (6.2%) strains belonged to the O37 serogroup, and 11 (4.9%) each belonged to serogroups O6 and O34. Serogroups O97 (4.5%), O11 (3.6%), and O59 (3.1%) were also identified in this study. Although serogroups O11, O35, and O37 showed a lower prevalence, their prevalence was higher among patients with diarrhea in Kolkata (7,10). The prevalence rate of NAGs (≈2%) in Kolkata is similar to that in the Haizhu District of Guangzhou, China (11). However, the number of serogroups identified in China (26 serogroups) was less than in our study (80 serogroups).

Among age groups of patients, NAGs were detected mostly in patients >5 years of age (>72%) than in those <5 years of age (≈28%) (Table 2). NAGs produce a spectrum of gastrointestinal symptoms ranging from asymptomatic infection to severe cholera-like illness or bloody diarrhea. In this study, most (70%) patients had watery diarrhea, which was similar to that for patients with cholera and those exclusively infected by NAGs (Table 2). Other clinical symptoms, such as dehydration status (22.3%) and fever (41.1%), were also high in patients with a single infection. In patients with mixed infections, bloody diarrhea (23.6%) and abdominal pain (41.5%) were the 2 major symptoms (Table 2), perhaps because >1 pathogen was involved.

Most NAGs were resistant to nalidixic acid (57.6%), ampicillin (55.5%), furazolidone (36.6%), and streptomycin (32.4%) and highly susceptible to gentamicin (96%), tetracycline (80%), and chloramphenicol (80.4%). During 1992–1997, antimicrobial drug resistance was high among V. cholerae isolated in Kolkata (12). In this study, patterns of antimicrobial drug resistance in NAGs were different than those in previous reports (10,12). This finding might be caused by discontinuation of ineffective antimicrobial drugs, such as co-trimoxazole and furazolidone, rational use of fluoroquinolones, and introduction of azithromycin for treating diarrhea.

Unlike our previous study (7), in this study, factors involved in the virulence of NAGs were not comprehensively elucidated in the study region. Unlike V. cholerae O1/O139, the pathogenicity of NAGs has been associated with >1 virulence factor (7,10). All NAGs lacked ctxA and El Tor tcpA genes. However, 6 (2.1%) strains had the tcpA gene; in 5 of them it was the only virulence gene detected. Most (94%) strains had the gene encoding El Tor type hemolysin, followed by rtxA (91.4%) and rtxC (75%) genes; only 5 (1.8%) had the stn gene. The RTX family includes a group of protein toxins produced by gram-negative bacteria, including V. cholerae with hemolytic, leukotoxic, and actin cross-linking activities, which may play a role in virulence (13). In this study, prevalence of NAGs that have genes encoding TTSS as the only virulence factor (41 cases) was higher than in Bangladesh and Argentina (14,15).

Strains containing hlyA, rtxA, and rtxC genes were predominant, followed by strains containing hlyA-rtxA-rtxC-TTSS and hlyA-rtxA genes (Table 3). Some gene combinations, such as hlyA-rtxA-rtxC-TTSS, hlyA-rtxA-rtxC, and hlyA-rtxA, were detected predominantly NAGs as the only virulence factor genes (Table 3). However, there was no correlation between type of serogroup and prevalence of putative virulence gene(s). Pulsed-field electrophoresis profiles of 70 strains representing the predominant serogroups showed distinct patterns (overall similarity ≈70%) (Figure). There was no profile match among strains belonging to the same serogroup.

Prevalence of NAGs associated with severe traits of infection indicates the role of these pathogens in cholera. The pathogenic mechanism of NAGs is multifarious; there are several virulence factors in genetically distinct strains. On the basis of our results, current antimicrobial drug therapy in the clinical management of NAG-mediated diarrhea can be continued. Further epidemiologic studies are needed to determine the ecology, virulence factors, and public health role of NAGs.