Rising Rates of Carbapenem-Resistant Enterobacteriaceae
Our study is, to our knowledge, the first to use mixed methods to describe and compare epidemiologic trends and microbiologic practices in a large number of community hospitals. There are major and interrelated conclusions from our analysis: (1) the rate of CRE detection in community hospitals is rapidly and significantly increasing, (2) microbiology laboratories in community hospitals use widely varying techniques for detection and reporting of CRE, and (3) more preparation is needed to combat this emerging threat.
The rate of CRE detection increased more than fivefold in our network of community hospitals from 2008 to 2012. The higher rate of CRE infection likely stems from a combination of factors, including increased use of broad-spectrum carbapenems, the ease with which carbapenemase enzymes can be transmitted among bacteria, and increased transmission between patients in healthcare settings. For example, 94% of cases of CRE infection detected in our study were healthcare associated. Given the recent finding that 30% of patients in long-term acute care hospitals are colonized with carbapenemase-producing K. pneumoniae, there are likely to be reservoirs of CRE within acute care hospitals, skilled nursing facilities, and hemodialysis centers that are driving the increase in CRE transmission in healthcare settings.
The use of more sensitive detection methods by clinical microbiology laboratories also contributes to recent increases in CRE. Microbiology laboratories in our network used different automated MIC systems, carbapenem breakpoints, and methods to verify the presence of carbapenemases (if any verification was done at all). Rates of CRE detection varied on the basis of these different approaches. We believe the low adoption of the decreased carbapenem breakpoints from the 2010 CLSI guidelines results in a significant underestimation in the true rate of CRE detection in our network of community hospitals. In our cohort, the rate of CRE detection was over threefold higher in hospitals that adopted and implemented the 2010 CLSI guidelines. Interestingly, there was a statistically significant association between the use of the Vitek 2 automated MIC system and increased rate of CRE detection. Previous head-to-head studies have demonstrated that the Microscan is at least equivalent to if not superior to the Vitek 2 system in detecting CRE. Thus, the cause of this association is unclear. It is conceivable that hospitals using Vitek 2 in this study had other confounding factors that caused this observation (eg, higher colonization pressure or worse CRE prevention practices).
The relatively low rate of CRE detection among the community hospitals in this study punctuates the importance of using networks of hospitals to identify larger epidemiologic trends and to better understand the importance of microbiology laboratory practices in detecting these pathogens. The overall number of CRE detected in individual hospitals is generally not sufficient to make firm conclusions regarding trends. As a result, practitioners in these single institutions would find it difficult to detect the increased rate of CRE detection or assess how variability in laboratory practices influences CRE detection. Analysis of surveillance data from the DICON network of community hospitals, however, has demonstrated that multiple epidemiologically important bacteria—extended-spectrum β-lactamase–producing bacteria and now CRE—are present and increasing in number in community hospitals in the United States.
There are several limitations in this study. First, the CRE surveillance data were recorded by local hospital IPs after interpretation of microbiology data. There is inherent subjectivity in surveillance practices despite the use of standard protocols and database. Given the overall low number of CRE present in community hospitals, this variability could be significant and skew the data in unclear ways. We recently performed a retrospective analysis, however, that confirmed that the definition used for "multidrug-resistant" gram-negative phenotypes by IPs detected 100% of CRE that occurred at a single tertiary care medical center. This issue is the reality of infection surveillance in community hospitals and further highlights the nationwide challenge in detecting and controlling this emerging threat. Second, this study used qualitative data from microbiology laboratory directors regarding laboratory practices, which may be subject to recall bias. We used multiple surveys to gain a more complete data set. Third, we did not collect data on the presence of carbapenemases, and so we are unable to comment on the molecular epidemiology of genes important in carbapenem resistance. Despite these limitations, we believe that this study provides a practical description of the limited data available on incidence of CRE in a multicenter sample of community hospitals and highlights deficiencies in preparedness for this public health threat.
We believe community hospitals in our network have much more work to do to prepare for and respond to CRE. Specifically, 2 focus areas must be developed to prevent CRE transmission: infection control and improved laboratory detection. Hospitals must be vigilant to limit person-to-person transmission. The CDC has outlined basic strategies to decrease transmission, including hand hygiene, contact precautions, healthcare personnel education, limitation of medical device use, patient and staff cohorting, laboratory notification strategies, antimicrobial stewardship, and CRE active screening. Of note, hospitals in our network have had significant difficulty implementing these recommendations because of resource limitations. No community hospitals in our cohort have yet adopted active surveillance programs for CRE. Finally, microbiology laboratories in these hospitals must make every effort to switch to the new carbapenem breakpoints. For example, all of the hospitals that did not detect a single CRE also did not adopt the new breakpoints. Thus, we remain skeptical that CRE are indeed absent from these hospitals. The inability to detect and control CRE makes in-hospital transmission more likely and could further drive the increasing trend and lead to hospital outbreaks.
In conclusion, our data indicate that the rates of CRE, while still infrequent, are increasing dramatically in community hospitals, where the majority of Americans receive their healthcare. We believe this increase is attributable to growing reservoirs and transmission of CRE and improvement in detection. Overall, we believe the estimates from study hospitals are underestimates of the true incidence in these hospitals. This point underscores the fact that these organisms are increasingly important and relevant in all areas of healthcare, including small community hospitals. Greater adherence to the 2010 CLSI guidelines is necessary to better understand the true prevalence of CRE and to better define what public health measures must be undertaken to prevent further spread of this serious, emerging threat.
Discussion
Our study is, to our knowledge, the first to use mixed methods to describe and compare epidemiologic trends and microbiologic practices in a large number of community hospitals. There are major and interrelated conclusions from our analysis: (1) the rate of CRE detection in community hospitals is rapidly and significantly increasing, (2) microbiology laboratories in community hospitals use widely varying techniques for detection and reporting of CRE, and (3) more preparation is needed to combat this emerging threat.
The rate of CRE detection increased more than fivefold in our network of community hospitals from 2008 to 2012. The higher rate of CRE infection likely stems from a combination of factors, including increased use of broad-spectrum carbapenems, the ease with which carbapenemase enzymes can be transmitted among bacteria, and increased transmission between patients in healthcare settings. For example, 94% of cases of CRE infection detected in our study were healthcare associated. Given the recent finding that 30% of patients in long-term acute care hospitals are colonized with carbapenemase-producing K. pneumoniae, there are likely to be reservoirs of CRE within acute care hospitals, skilled nursing facilities, and hemodialysis centers that are driving the increase in CRE transmission in healthcare settings.
The use of more sensitive detection methods by clinical microbiology laboratories also contributes to recent increases in CRE. Microbiology laboratories in our network used different automated MIC systems, carbapenem breakpoints, and methods to verify the presence of carbapenemases (if any verification was done at all). Rates of CRE detection varied on the basis of these different approaches. We believe the low adoption of the decreased carbapenem breakpoints from the 2010 CLSI guidelines results in a significant underestimation in the true rate of CRE detection in our network of community hospitals. In our cohort, the rate of CRE detection was over threefold higher in hospitals that adopted and implemented the 2010 CLSI guidelines. Interestingly, there was a statistically significant association between the use of the Vitek 2 automated MIC system and increased rate of CRE detection. Previous head-to-head studies have demonstrated that the Microscan is at least equivalent to if not superior to the Vitek 2 system in detecting CRE. Thus, the cause of this association is unclear. It is conceivable that hospitals using Vitek 2 in this study had other confounding factors that caused this observation (eg, higher colonization pressure or worse CRE prevention practices).
The relatively low rate of CRE detection among the community hospitals in this study punctuates the importance of using networks of hospitals to identify larger epidemiologic trends and to better understand the importance of microbiology laboratory practices in detecting these pathogens. The overall number of CRE detected in individual hospitals is generally not sufficient to make firm conclusions regarding trends. As a result, practitioners in these single institutions would find it difficult to detect the increased rate of CRE detection or assess how variability in laboratory practices influences CRE detection. Analysis of surveillance data from the DICON network of community hospitals, however, has demonstrated that multiple epidemiologically important bacteria—extended-spectrum β-lactamase–producing bacteria and now CRE—are present and increasing in number in community hospitals in the United States.
There are several limitations in this study. First, the CRE surveillance data were recorded by local hospital IPs after interpretation of microbiology data. There is inherent subjectivity in surveillance practices despite the use of standard protocols and database. Given the overall low number of CRE present in community hospitals, this variability could be significant and skew the data in unclear ways. We recently performed a retrospective analysis, however, that confirmed that the definition used for "multidrug-resistant" gram-negative phenotypes by IPs detected 100% of CRE that occurred at a single tertiary care medical center. This issue is the reality of infection surveillance in community hospitals and further highlights the nationwide challenge in detecting and controlling this emerging threat. Second, this study used qualitative data from microbiology laboratory directors regarding laboratory practices, which may be subject to recall bias. We used multiple surveys to gain a more complete data set. Third, we did not collect data on the presence of carbapenemases, and so we are unable to comment on the molecular epidemiology of genes important in carbapenem resistance. Despite these limitations, we believe that this study provides a practical description of the limited data available on incidence of CRE in a multicenter sample of community hospitals and highlights deficiencies in preparedness for this public health threat.
We believe community hospitals in our network have much more work to do to prepare for and respond to CRE. Specifically, 2 focus areas must be developed to prevent CRE transmission: infection control and improved laboratory detection. Hospitals must be vigilant to limit person-to-person transmission. The CDC has outlined basic strategies to decrease transmission, including hand hygiene, contact precautions, healthcare personnel education, limitation of medical device use, patient and staff cohorting, laboratory notification strategies, antimicrobial stewardship, and CRE active screening. Of note, hospitals in our network have had significant difficulty implementing these recommendations because of resource limitations. No community hospitals in our cohort have yet adopted active surveillance programs for CRE. Finally, microbiology laboratories in these hospitals must make every effort to switch to the new carbapenem breakpoints. For example, all of the hospitals that did not detect a single CRE also did not adopt the new breakpoints. Thus, we remain skeptical that CRE are indeed absent from these hospitals. The inability to detect and control CRE makes in-hospital transmission more likely and could further drive the increasing trend and lead to hospital outbreaks.
In conclusion, our data indicate that the rates of CRE, while still infrequent, are increasing dramatically in community hospitals, where the majority of Americans receive their healthcare. We believe this increase is attributable to growing reservoirs and transmission of CRE and improvement in detection. Overall, we believe the estimates from study hospitals are underestimates of the true incidence in these hospitals. This point underscores the fact that these organisms are increasingly important and relevant in all areas of healthcare, including small community hospitals. Greater adherence to the 2010 CLSI guidelines is necessary to better understand the true prevalence of CRE and to better define what public health measures must be undertaken to prevent further spread of this serious, emerging threat.
SHARE
