Surveillance Data on Outbreaks ofClostridiumdifficileInfection in Ontario, Canada, in 2008–2009

To the Editor—We read with great interest the Canadian Nosocomial Infection Surveillance Program (CNISP) data presented by Miller et al [1], who reported that the prevalence of the North American pulsed-field type 1 (NAP1) strain of Clostridium difficile was ∼20% in Ontario, Canada. During the period from 2008 through 2009, our public health laboratory conducted surveillance of all outbreak-related C. difficile identified in the province of Ontario as part of government legislation to document patient safety. Hospitals in Ontario are now required to report their incidence of C. difficile infection on a monthly basis to the Ministry of Health and Long-Term Care [2]. An outbreak was defined as ⩾6 cases of C. difficile infection per ward per month (Table 1). Isolates were analyzed to determine C. difficile pulsotypes and antibiotic susceptibility, for detection of toxin genes, and for genotyping of antibiotic resistance markers. A total of 155 C. difficile isolates from 16 distinct institutional outbreaks of C. difficile infection were identified. In contrast to the CNISP data presented by Miller et al [1], we found that the predominant outbreak isolate was NAP1 strain, with a prevalence of 53% (Figure 1). Furthermore, our data contrasted with the data collected from passive surveillance conducted in Ontario and presented by Martin et al [3], who reported that the NAP2 strain was identified as the most common strain in Ontario, with NAP1 being the second most common strain. NAP1 was the outbreak strain for all institutional outbreaks investigated during this surveillance. A large number of sporadic pulsotypes were given an “arbitrary” classification because their pattern did not conform to previously described NAP strains. The NAP2 strain was completely absent from outbreaks. Therefore, we believe that the NAP1 strain is more transmissible or may have a fitness advantage resulting in outbreaks, compared with other pulsotypes that may be causing sporadic cases of C. difficile infection [4]. NAP1 isolates and non-NAP1 isolates were susceptible to metronidazole; however, compared with the non-NAP1 isolates, the NAP1 isolates uniformly elaborated binary toxin, were more likely to be resistant to moxifloxacin (P < .05, χ2 test), and were less likely to be susceptible to ampicillin (P<.05 χ2 test). Sequencing of the quinolone resistance-determining region of DNA gyrase demonstrated that the gyrA mutation T82I (P < .05, χ2 test) and the gyrBmutationD426N(P <.05, χ2 test) were significantly associated with newergeneration fluoroquinolone resistance. In conclusion, although differences in NAP1 prevalence may be explained by selection criteria and temporal differences in sample collection, we believe that the NAP1 strain is present in a far greater number of cases of C. difficile infection than was reported by Miller et al [1] in Ontario and is still single-handedly responsible for driving institutional outbreaks of C. difficile infection [4]. We thank the clinical and research staff members of the hospital-acquired infection unit of the Public Health Laboratory (Toronto), the Public Health units, and infection control staff at the institutions for coordinating the samples collected in this surveillance. Potential conflicts of interest. All authors: no conflicts. Clostridium difficile pulsed-field gel electrophoresis (PFGE) types (155 isolates) identified in 16 institutional outbreaks in 2008 and 2009 in Ontario, Canada. Arbitrary implies that the PFGE pattern did not correspond to previously described North American pulsotypes (NAPs). NoGr, no growth; Untyp, untypeable. Demographics of Patients Whose Specimens Were Analyzed during Clostridium difficile Infection Outbreaks in Ontario, Canada, in 2008 and 2009