Outbreaks Caused by New Variants of Vibrio cholerae O1 El Tor, India

To the Editor: Vibrio cholerae O1, the causative agent of cholera, has 2 biotypes (classical and El Tor), which have traditionally been distinguished by phenotypic tests and by genetic differences in the major toxin-coregulated pilus (TCP) gene, the tcpA allele of the TCP cluster (1), the rstR region (regulatory region for phage lysogeny) of CTX phages (2), the type of cholera toxin (CT) produced, and the infection pattern of the disease they cause. However, 3 variants of the El Tor biotype have been described recently: Matlab (a place in Bangladesh) variants in 2002 (3), which could not be biotyped because they have a mixture of both classical and El Tor (4), Mozambique variant in 2004–2005, which has a typical El Tor genome but a tandem repeat of the classical CTX prophage in the small chromosome (5), and the altered El Tor type (a typical El Tor biotype and an El Tor CTX prophage that produces CT of the classical type) predominant in Bangladesh since 2001 (6). Hybrid vibrios have also been described in other regions of Asia and Africa (7). 
 
CT, encoded by the ctxA and ctxB genes, is the principal toxin produced by V. cholerae O1 and O139. Methods for differentiating the biotype-specific CT-B subunit of V. cholerae O1 include sequencing the ctxB gene, performing an ELISA wth a monoclonal antibody specific to the classical or El Tor CT, or by using a mismatch amplification mutation assay (MAMA)–PCR to distinguish between 2 kinds of ctxB genes. This assay detects sequence polymorphisms based on nt position 203 of the ctxB gene (8). 
 
In Punjab and Haryana states of northern India, during July–September 2007, 6 clusters of cholera outbreak were identified. A total of 745 case-patients were admitted to local government hospitals; the cholera attack rate was 183/1,000 population. Four deaths were reported (case-fatality rate 0.5%). The number of cases per cluster varied from 15 to 400, and adults were primarily affected (74%); 20% of patients had severe dehydration. V. cholerae O1 Ogawa was confirmed from stool cultures by using standard isolation, biochemical, and serotyping methods. Twenty-six isolates were phenotypically and genotypically characterized according to biotype. Phenotypic characterization included the Voges-Proskauer reaction, polymyxin B (50 U) susceptibility, chick cell agglutination, and sheep erythrocyte hemolysis; all isolates were confirmed as El Tor biotype. Genotypic characterization included PCR assays for ctxA and tcpA (2), rstR (3), and ctxB (8). All strains were toxigenic because each carried the ctxA gene. All strains also carried El Tor–specific tcpA (472 bp) and rstR genes (500 bp). MAMA-PCR showed the ctxB gene of both El Tor and the classical type in 21 (80%) of 26 isolates tested. 
 
Similarly, we also tested 20 available isolates from the 2002 outbreak and 4 and 19 sporadic isolates from 2003 and 2004, respectively; all were phenotypically and genotypically confirmed as El Tor and had only ctxB of the El Tor type. Of 53 water samples tested during the 2007 outbreak, 4 grew V. cholerae. Three samples were confirmed to be non-O1, non-O139 strains. Only 1 isolate was V. cholerae O1, which was positive for tcp, ctxA, and ctxB of both classical and El Tor types (Table). 
 
 
 
Table 
 
Phenotypic and genotypic traits of Vibrio cholerae O1 clinical strains isolated from northern India, 2002–2007* 
 
 
 
During the cholera outbreak of 2002 in Chandigarh (9), only 1 death was reported (case-fatality rate, <0.01%); the attack rate was 20/1,000, 58.6% were children, and only 10% had severe dehydration. Before the most recent outbreak, the affected regions of Panjab and Haryana (Ambala, Nurpur, Kurali, Mohali, Panchkula, and Raili) had been free of cholera outbreaks since 1994, though sporadic cases had been reported. The 4 deaths from cholera in 2007, along with adult preponderance, high attack rate, more severe illness, and 6 different clusters, point towards a change in the disease’s epidemiology. This change may be related to circulation of the hybrid vibrios in this region. In Bangladesh, all strains of V. cholerae O1 examined since 2001 belong to the altered El Tor type (6), which produces CT of the classical type. This altered type has replaced the seventh pandemic strain of the El Tor biotype that produced CT of the El Tor type, which indicates that a cryptic change has occurred in the seventh pandemic El Tor biotype strains of V. cholerae O1. 
 
Newly emerged variants from Bangladesh (8) have the genetic makeup of El Tor with ctxB gene of only classical, whereas our strains are unique in having ctxB of both the classical and El Tor biotypes. Our strains appear to be different from the Mozambique variant V. cholerae O1 (10), which has rstR of the classical type, in that our strains have rstR of only the El Tor type. Of 5 Matlab variants analyzed with MAMA-PCR, 3 had classical ctxB and 2 had El Tor type. Our study highlights the different genetic recombinations possible in V. cholerae and the epidemiologic role of these recombinations.


Outbreaks Caused by New Variants of Vibrio cholerae O1
El Tor, India To the Editor: Vibrio cholerae O1, the causative agent of cholera, has 2 biotypes (classical and El Tor), which have traditionally been distinguished by phenotypic tests and by genetic differences in the major toxincoregulated pilus (TCP) gene, the tcpA allele of the TCP cluster (1), the rstR region (regulatory region for phage lysogeny) of CTX phages (2), the type of cholera toxin (CT) produced, and the infection pattern of the disease they cause.However, 3 variants of the El Tor biotype have been described recently: Matlab (a place in Bangladesh) variants in 2002 (3), which could not be biotyped because they have a mixture of both classical and El Tor (4), Mozambique variant in 2004-2005, which has a typical El Tor genome but a tandem repeat of the classical CTX prophage in the small chromosome (5), and the altered El Tor type (a typical El Tor biotype and an El Tor CTX prophage that produces CT of the classical type) predominant in Bangladesh since 2001 (6).Hybrid vibrios have also been described in other regions of Asia and Africa (7).
CT, encoded by the ctxA and ctxB genes, is the principal toxin produced by V. cholerae O1 and O139.Methods for differentiating the biotype-specifi c CT-B subunit of V. cholerae O1 include sequencing the ctxB gene, performing an ELISA wth a monoclonal antibody specifi c to the classical or El Tor CT, or by using a mismatch amplifi cation mutation assay (MAMA)-PCR to distinguish between 2 kinds of ctxB genes.This assay detects sequence polymorphisms based on nt position 203 of the ctxB gene (8).
In case-patients were admitted to local government hospitals; the cholera attack rate was 183/1,000 population.Four deaths were reported (case-fatality rate 0.5%).The number of cases per cluster varied from 15 to 400, and adults were primarily affected (74%); 20% of patients had severe dehydration.V. cholerae O1 Ogawa was confi rmed from stool cultures by using standard isolation, biochemical, and serotyping methods.Twentysix isolates were phenotypically and genotypically characterized according to biotype.Phenotypic characterization included the Voges-Proskauer reaction, polymyxin B (50 U) susceptibility, chick cell agglutination, and sheep erythrocyte hemolysis; all isolates were confi rmed as El Tor biotype.Genotypic characterization included PCR assays for ctxA and tcpA (2), rstR (3), and ctxB (8).All strains were toxigenic because each carried the ctxA gene.All strains also carried El Tor-specifi c tcpA (472 bp) and rstR genes (500 bp).MAMA-PCR showed the ctxB gene of both El Tor and the classical type in 21 (80%) of 26 isolates tested.
Similarly, we also tested 20 available isolates from the 2002 outbreak and 4 and 19 sporadic isolates from 2003 and 2004, respectively; all were phenotypically and genotypically confi rmed as El Tor and had only ctxB of the El Tor type.Of 53 water samples tested during the 2007 outbreak, 4 grew V. cholerae.Three samples were confi rmed to be non-O1, non-O139 strains.Only 1 isolate was V. cholerae O1, which was positive for tcp, ctxA, and ctxB of both classical and El Tor types (Table ).
During the cholera outbreak of 2002 in Chandigarh (9), only 1 death was reported (case-fatality rate, <0.01%); the attack rate was 20/1,000, 58.6% were children, and only 10% had severe dehydration.Before the most recent outbreak, the affected regions of Panjab and Haryana (Ambala, Nurpur, Kurali, Mohali, Panchkula, and Raili) had been free of cholera outbreaks since 1994, though sporadic cases had been reported.The 4 deaths from cholera in 2007, along with adult preponderance, high attack rate, more severe illness, and 6 different clusters, point towards a change in the disease's epidemiology.This change may be related to circulation of the hybrid vibrios in this region.In Bangladesh, all strains of V. cholerae O1 examined since 2001 belong to the altered El Tor type (6), which produces CT of the classical type.This altered type has replaced the seventh pandemic strain of the El Tor biotype that produced CT of the El Tor type, which indicates that a cryptic change has occurred in the seventh pandemic El Tor biotype strains of V. cholerae O1.
Newly emerged variants from Bangladesh (8) have the genetic makeup of El Tor with ctxB gene of only classical, whereas our strains are unique in having ctxB of both the clas-sical and El Tor biotypes.Our strains appear to be different from the Mozambique variant V. cholerae O1 (10), which has rstR of the classical type, in that our strains have rstR of only the El Tor type.Of 5 Matlab variants analyzed with MAMA-PCR, 3 had classical ctxB and 2 had El Tor type.Our study highlights the different genetic recombinations possible in V. cholerae and the epidemiologic role of these recombinations.To the Editor: The genus Atopobium (1) accommodates species formerly designated Lactobacillus minutus, L. rimae, and Streptococcus parvulus (2).Use of 16S rDNA sequence analysis showed these species to be closely related and to form a distinct line of descent within the lactic acid bacteria (3).Atopobium spp.usually have been recognized as part of the human gingival oral fl ora; some species, including A. rimae and A. parvulum, have been identifi ed as agents of chronic periodontitis (4,5).A. rimae, formerly known as L. rimae (1), forms short, gram-positive, strictly anaerobic, elliptical bacteria with low G+C content (4). A. rimae produces large amounts of lactic acid and has been recovered previously from normal human gingival fl ora (4,5).Apart from periodontitis, it has not been implicated in other types of infection.We report an unusual case of A. rimae bacteremia.
In May 2007, a 77-year-old woman with a history of right thoracotomy for pneumothorax 2 years earlier was hospitalized for inhalation pneumonia caused by paralysis of the right vocal cord.During hospitalization, septic shock and a fever of 38°C developed in the patient, complicated by acute respiratory failure and stroke.She was transferred to an intensive care unit with a PaO 2 /FiO 2 >300 mm Hg, and a tracheotomy was performed.Three anaerobic blood specimens, drawn at entrance into the intensive care unit, yielded gram-positive cocci after 24-h incubation of the fi rst bottle and grampositive bacilli after 48-h incubation of the 2 other bottles.The gram-positive cocci were identifi ed as Streptococcus gordonii using API STREP (bioMérieux, Marcy l'Etoile, France).The gram-positive bacilli were catalase negative and oxidase positive but remained unidentifi ed with use of API ANA strip (bioMérieux).Minimum inhibitory concentrations of antibiotics were determined for the gram-positive bacilli using E-test assay (AB BIODISK, Solna, Sweden) on Columbia agar supplemented with 5% sheep blood.Minimum inhibitory concentrations were 0.064 μg/mL for penicillin G, 0.023 μg/mL for ampicillin, 0.012 μg/mL for amoxicillin-clavulanic acid, 0.032 μg/mL for imipenem, <0.016 μg/mL for azithromycin, <0.016 μg/mL for erythromycin, 0.06 μg/mL for ciprofl oxacin, and 1.25 μg/mL for vancomycin.DNA was extracted from 1 colony by using a QIAamp tissue kit (QIAGEN, Hilden, Germany) as described by the manufacturer.The 1,454-bp 16S rDNA sequence obtained using the fD1 5′-AGAGTTTGATCCTGGCTCAG-3′ and rP2 5′-ACGGCTACCTTGTTAC GACTT-3′ primer pair (6,7) showed 99% sequence similarity with the 16S rDNA sequence of A. rimae (GenBank accession no.AF292371) by use of BLAST version 2.2.9 software (National Center for Biotechnology Information).A phylogenetic neighbor-joining tree based on the Atopobium spp.16S rDNA sequences made with the MEGA software confi rmed that the isolate belonged to A. rimae (Figure).Initial treatment by intravenous tazocilline-amikacin was changed to intravenous amoxicillin-clavulanic acid (2 g/200 mg).The fever resolved, and the patient's condition improved.The treatment was stopped after 7 days, and the patient remained apyretic.
In this case, phenotypic identifi cation of gram-positive bacillus isolated from 2 blood cultures failed because the defi nite bacterial species A. rimae was not included in the API database used for the phenotypic identifi cation.Final identifi cation was achieved within 2 days by comparison of the almost complete 16S rDNA sequence with homologous sequences deposited in Genbank.This comparison yielded a 99% sequence similarity, regarded