Antimicrobial resistance among Salmonella and Shigella isolates in five Canadian provinces (1997 to 2000) résistance antimicrobienne dans les isolats

Antimicrobial resistance among Salmonella and Shigella isolates in five Canadian provinces (1997 to 2000). Can J Infect Dis Med Microbiol 2006;17(4):243-250. OBJECTIVE: To describe rates of antimicrobial resistance (AMR) among Salmonella and Shigella isolates reported in five Canadian provinces, focusing on clinically important antimicrobials. METHODS: The authors retrospectively investigated AMR rates among 6219 Salmonella and 1673 Shigella isolates submitted to provincial public health laboratories in Alberta, Newfoundland and Labrador, Ontario, Prince Edward Island and Saskatchewan from 1997 to 2000; these isolates were estimated to represent 41% of Salmonella cases and 72% of Shigella cases reported by the study provinces. RESULTS: Among Salmonella isolates, 27% (1704 of 6215) were resistant to ampicillin, 2.2% (135 of 6122) to trimethoprim/ sulfamethoxazole, 1.5% (14 of 938) to nalidixic acid, 1.2% (one of 84) to lomafloxacin and 0.08% (five of 6163) to ciprofloxacin. Among Shigella isolates, 70% (1144 of 1643) were resistant to trimethoprim/sulfamethoxazole, 65% (1079 of 1672) to ampicillin, 3.1% (eight of 262) to nalidixic acid, 0.49% (eight of 1636) to ciprofloxacin, 0.14% (one of 700) to ceftriaxone and 0.08% (one of 1292) to ceftazidime. CONCLUSIONS: Higher rates of resistance to clinically important antimicrobials (including ciprofloxacin) were observed among both Salmonella and Shigella isolates than has previously been reported. Current Canadian data on rates of AMR for these pathogens are required.


METHODS
AB, NL, ON, PE and SK provided available data on human Salmonella and Shigella isolates tested for AMR. Data available differed by province by time span of availability (for NL, data were only available between 1999 and 2000), antimicrobials tested and bacterial typing methods (Table 1). Laboratories used standard methods for identification of Salmonella and Shigella species. The University of Guelph (Guelph, Ontario) provided ethical approval of the study.

Data cleaning
The National Committee for Clinical Laboratory Standards (now the Clinical and Laboratory Standards Institute) guidelines (6) were used to interpret minimum inhibitory concentrations; isolates with minimum inhibitory concentrations in the intermediate range of resistance were classified as resistant. Isolates were eliminated if they were not submitted between 1997 and 2000, not tested for antimicrobial susceptibility or missing susceptibility results, known to be nonhuman in origin (either animal or environmental samples, although some not labelled as such may have remained in the data) or missing genus information. Laboratories may have tested a sample from a single case more than once, creating more than one test result per case. For provinces that provided laboratory or patient numbers, isolates were defined as duplicates if they had the same identification number, were the same organism (at the serotype level for Salmonella and serogroup level for Shigella), had the same antimicrobial resistance pattern, and were collected or received in the same calendar month and year. If the number of antimicrobials tested differed by duplicate isolate, the isolate tested for the greater number of antimicrobials was retained.

Investigating AMR
Resistance was investigated by antimicrobial class (stratified by the most prevalent Salmonella serotypes and Shigella serogroups) and by resistance to antimicrobials of particular clinical importance. Analyses were conducted using SAS version 9.1 (SAS Institute, USA).

Salmonella
After 48 duplicates were removed and exclusion criteria were applied, 6219 Salmonella isolates submitted by the five study provinces were included in the analyses (Table 1). One hundred thirty-one (2.1%) isolates were Salmonella enterica serovar Typhi (S typhi) and 19  ; specimen source was unknown or not provided for 2129 (34%) isolates. The median age for patients with Salmonella was 22.0 years (age range of less than one to 97 years) (n=5843); 51% of isolates were from female patients (n=6071).

Shigella
After 21 duplicates were removed and exclusion criteria were applied, data on 1673 Shigella isolates, submitted by all provinces except NL, were included in the analyses (Table 1)    Continued on next page unknown sources. The median age for patients with Shigella was 24.0 years (age range of less than one to 88 years) (n=1619); 54% of isolates were from female patients (n=1634). By antimicrobial class, S sonnei isolates had the highest rates of resistance to other beta-lactams (26%), sulfonamides and trimethoprim (76%), and quinolones (12 isolates [1.0%]). S flexneri isolates had the highest rates of resistance to tetracyclines (90%), extended-spectrum penicillins (73%) and chloramphenicol (71%). Although few S dysenteriae isolates were tested, this serogroup showed the highest rates of resistance to beta-lactam/beta-lactamase inhibitor combinations (five isolates [63%]) and aminoglycosides (six isolates [27%]); S sonnei and S flexneri showed lower rates of resistance to betalactam/beta-lactamase inhibitor combinations and aminoglycosides (S sonnei: 52% and 15%, respectively; S flexneri: 53% and 16%, respectively). S boydii isolates had the highest rate of resistance to nitrofurantoin (four isolates [18%]). Sixteen Shigella isolates were resistant to quinolones (12 S sonnei and four S flexneri isolates).
Among clinically important antimicrobials, high rates of resistance were observed for T/S (70%) and ampicillin (65%) ( Table 3). Most Shigella isolates (n=1636 [98%]) were tested for resistance to ciprofloxacin; eight (0.49%) were resistant (six S sonnei and two S flexneri isolates), all of which were additionally resistant to ampicillin and/or T/S (Table 4). Comparatively few isolates were tested for resistance to nalidixic acid (262 isolates [16%]); of these isolates, eight (3.1%) were resistant (six S sonnei and two S flexneri isolates) ( Table 3), one of which was also resistant to ciprofloxacin (Table 4). Two S boydii isolates were resistant to third-generation cephalosporins: one to ceftriaxone (0.14%) and one to ceftazidime (0.08%) ( Table 3). The ceftriaxone-resistant isolate originated from a two-year-old patient (sex unknown) with a history of travel to India (Table 5), and the ceftazidimeresistant isolate (also resistant to ampicillin, amoxicillin/ K clavulanate, chloramphenicol, cephalothin, imipenem, nitrofurantoin, piperacillin, tetracycline and T/S) originated from a 44-year-old man with no available information on travel history.

DISCUSSION
The present retrospective study describes AMR rates for Salmonella and Shigella isolates passively reported in five Canadian provinces between 1997 and 2000. These data do not include all cases reported in these provinces. A closer estimate, however, can be obtained from the National Notifiable Diseases database (NND). For the provinces and years included in our study (limited to 1999 and 2000 for NL), our data represent 41% (6219 of 14995) of Salmonella isolates (including Typhi and Paratyphi serotypes) and 72% (1673 of 2313) of Shigella isolates reported by the NND (1). Therefore, our results may be more representative for Shigella than for Salmonella.

Continued on next page
Salmonella When antimicrobial therapy for nontyphoidal Salmonella infections is recommended, fluoroquinolones, T/S, ampicillin and third-generation cephalosporins are considered drugs of choice (7). Multidrug-resistant S typhi and S paratyphi infections should be treated with fluoroquinolones, third-generation cephalosporins and azithromycin (8). Among the Salmonella isolates in our study, we observed higher rates of resistance to clinically important antimicrobials than previously reported. For example, in a Quebec study of Salmonella isolates from patients hospitalized between 1991 and 1995, Gaudreau and Turgeon (9) reported no resistance to ciprofloxacin, 4% of isolates resistant to ampicillin (five isolates) and 0.08% of isolates resistant to T/S (one isolate). In contrast, we observed 0.08% of Salmonella isolates resistant to ciprofloxacin, 27% to ampicillin and 2.2% to T/S. This variation could be the result of differences in study design (hospital isolates versus passive provincial laboratory surveillance), differences in study location, or to real increases in AMR rates over time. Our data suggest that resistance to T/S has increased from 1.7% in 1997 to 3.3% in 2000 (Table 2). More recent Canadian data are available from the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS), which was established to monitor the resistance of enteric pathogens, including Salmonella, that are isolated from humans, animals and animal-derived foods across the country (10). Compared with our study results, for Salmonella isolates tested in the five study provinces in 2003, CIPARS observed higher rates of resistance to nalidixic acid (1.5% versus 6.8%), similar rates of resistance to T/S (3.3% versus 3.8%) and ciprofloxacin (0.08% versus 0.06%), and lower rates of resistance to ampicillin (27% versus 19%), amoxicillin (14% versus 4.8%) and ceftriaxone (0.46% versus 0.06%) (10). However, the methodologies used by CIPARS and the present study were different. In CIPARS, all or a sample of Salmonella isolates from the provincial laboratories were systematically tested for resistance to the same antimicrobials. In our study, however, the antimicrobials tested varied among the isolates and provinces. This difference may account for the observed variations.
Based on our results, ampicillin appears to be an inappropriate choice for antimicrobial therapy for a number of Salmonella isolates tested in Canada over the study period. Low rates of resistance to T/S, nalidixic acid, ciprofloxacin and third-generation cephalosporins were also observed. In patients with Salmonella infections, susceptibility testing should be performed and reported as per current Clinical and Laboratory Standards Institute guidelines. In the present study, only 15% of Salmonella isolates were tested for nalidixic acid resistance. However, Crump et al (11) have emphasized the importance of testing for nalidixic acid resistance, citing evidence that inadequate clinical response to fluoroquinolones has occurred among cases infected with fluoroquinolonesusceptible, nalidixic-resistant Salmonella isolates.

Shigella
Among Shigella isolates, we observed high rates of resistance to T/S (70%) and ampicillin (65%), drugs that were once commonly used to treat shigellosis (12). Because 98.2% and 99.9% of isolates were tested for resistance to T/S and ampicillin, respectively, these resistance rates are representative of the Shigella isolates included in our study. In comparison, a 1990 study of 598 Shigella isolates (11) reported a lower rate of resistance to T/S (26.7% to 37.6%) than our study, as well as rates of resistance to ampicillin ranging from 39.3% to 66.5%. Compared with our study, Gaudreau and Turgeon (9) reported    (13), which reported a higher resistance rate for ampicillin (78%) but a lower rate for T/S (46%) than our study. Based on the results of our study, ampicillin and T/S appear to be inappropriate therapeutic choices for most Shigella infections reported in the study provinces. One S boydii isolate tested in 1998 from a patient with a history of travel to India was resistant to ceftriaxone. Extendedspectrum beta-lactamase-producing S sonnei and S flexneri isolates have been reported in several countries, including France, Argentina, Korea, Turkey, Bangladesh and Taiwan (14)(15)(16)(17)(18)(19). However, from 1999 to 2002, no ceftriaxone resistance among Shigella isolates was observed in the US (14), and to our knowledge, this is the first report of ceftriaxone resistance among Shigella isolates in Canada.
Eight (3.1%) Shigella isolates included in our study were resistant to nalidixic acid, which is higher than the numbers reported in a 1990 study (which found no resistance to nalidixic acid [20]) and by the National Antimicrobial Resistance Monitoring System (which found 1% of Shigella isolates resistant between 1999 and 2002 in the US [13]). In contrast, the resistance rate we observed was lower than that reported by a study conducted in England and Wales in 2002 (21), which found 13% of S sonnei isolates, and 10% of S dysenteriae, S flexneri and S boydii isolates resistant to nalidixic acid. Nalidixic acid resistance among Shigella isolates has been associated with decreased susceptibility to ciprofloxacin (19); therefore, it is important to monitor resistance to nalidixic acid to prevent possible fluoroquinolone treatment failures.
Eight (0.49%) Shigella isolates included in our study were resistant to ciprofloxacin, the therapy currently recommended by the World Health Organization to treat shigellosis (12). To our knowledge, few studies have previously reported resistance to ciprofloxacin among Shigella isolates tested in North America. In one example in 2001, a US study (22) reported the uncommon occurrence of an S flexneri isolate resistant to ciprofloxacin from a patient with a history of travel to China. In another example, a Canadian study (23) reported an S dysenteriae type 1 isolate resistant to ciprofloxacin and nalidixic acid from a 56-year-old man from AB in 2004 with a history of travel to India. Unfortunately, travel histories were unavailable for Shigella cases with ciprofloxacin-resistant infections in our study; therefore, we are uncertain whether these organisms were acquired domestically or internationally. Travel histories from Shigella cases should be included with specimens submitted to the laboratory to enable a more comprehensive understanding of the epidemiology of ciprofloxacin-resistant Shigella isolates reported in Canada. This information should indicate whether the case had recently travelled ('yes or no'), as well as the location. Regardless of the country of origin, Shigella is transmitted via person-to-person contact; thus, there is a potential risk for secondary transmission and infection. If these retrospective resistance rates are predictive of current rates, then ciprofloxacin -the currently recommended first-line therapymay not be effective for a small percentage of Shigella infections occurring in Canada.

Limitations
The present study has several limitations. First, we collected data retrospectively; therefore, methods were not uniform   across the provinces or over time, and provincial laboratories may have selectively tested isolates for resistance to certain antimicrobials. These differences could have affected observed rates of resistance; consequently, these data should not be taken as representative of all laboratory-confirmed Salmonella or Shigella isolates reported in the study provinces. Although the culture methods may have varied among the laboratories, each laboratory participated in external proficiency testing programs to maintain high standards for the testing of enteric pathogens. Second, travel histories were available for few cases. Therefore, for isolates from patients without information about travel history, we could not be sure whether the isolates tested originated in Canada or whether they were imported from another country. Third, these data were collected over five years ago; therefore, they should not be considered as characteristic of present Salmonella or Shigella resistance rates. However, these data provide baseline information, and context for future research and surveillance efforts. Despite the limitations of these data, these results represent, to our knowledge, the most comprehensive description of AMR rates for Salmonella during the study period and one of the only multiprovincial descriptions of AMR rates available for Shigella in Canada.

CONCLUSIONS
The rates of resistance we observed among Salmonella and Shigella isolates are concerning; they demonstrate that treatment options are more limited for these infections than previously reported. Current information on AMR rates for these pathogens is needed. CIPARS conducts national surveillance on AMR for Salmonella; however, given that we found Shigella isolates resistant to the recommended first-line therapy, resistance among Shigella isolates should also be monitored by CIPARS, along with the routine collection of patients' travel histories to differentiate between domestically and internationally acquired infections. Because humans are the key reservoir for Shigella, the resistance observed in these organisms is likely associated with human antimicrobial use. Therefore, prudent antimicrobial drug use is essential to maintain effective antimicrobial therapies for these infections.
ADDENDUM: Aspects of the data found in this study were presented in the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) 2002 Annual Report, which is available online at <http://www.phac-aspc.gc.ca/cipars-picra>.