Measuring Carbapenem-Resistant Enterobacteriaceae in the United States
The Journal of the American Medical Association October 05, 2015
A Critical Step for Control
Mary K. Hayden, MD
Rush University Medical Center, Chicago, Illinois
Carbapenem-resistant Enterobacteriaceae (CRE) may be the most concerning contemporary antibiotic resistance threat. Enterobacteriaceae comprise a large group of bacteria, including Escherichia coli and Klebsiella pneumoniae, and are common causes of health care–associated and community-acquired infections. Carbapenems, such as imipenem, meropenem, ertapenem, and doripenem are among the broadest-spectrum and most potent β-lactam antibiotics. For decades, carbapenems were reliably active against Enterobacteriaceae, and thus were frequently selected for empirical treatment of suspected serious gram-negative infections.
Significant carbapenem resistance in Enterobacteriaceae was first described in 2001 in a US strain of K pneumoniae that produced an enzyme, termed “K pneumoniae carbapenemase” that was able to destroy all β-lactam antibiotics, including carbapenems. Infections due to K pneumoniae carbapenemase–producing CRE are overwhelming health care–associated infections, and mortality related to bloodstream infections is substantial, with an estimated case fatality rate of more than 50%.1,2 This appears to be due more to delayed or inadequate antibiotic therapy for these multidrug-resistant organisms than to increased bacterial virulence.
Significant carbapenem resistance in Enterobacteriaceae was first described in 2001 in a US strain of K pneumoniae that produced an enzyme, termed “K pneumoniae carbapenemase” that was able to destroy all β-lactam antibiotics, including carbapenems.1 Infections due to K pneumoniae carbapenemase–producing CRE are overwhelming health care–associated infections, and mortality related to bloodstream infections is substantial, with an estimated case fatality rate of more than 50%.1,2 This appears to be due more to delayed or inadequate antibiotic therapy for these multidrug-resistant organisms than to increased bacterial virulence.2,3
Reports of K pneumoniae carbapenemase increased rapidly after 2005, and several new carbapenemases were identified, most notably New Delhi metallo-β-lactamase and OXA-48-like oxacillinase.1 Carbapenemase-producing CRE have since spread globally, including into the community in developing countries,1 although long-term reduction in incidence has been achieved in regions in which aggressive multimodal interventions were introduced early after CRE emergence.4 Carbapenem resistance can be due to mechanisms other than carbapenemase production; however, CRE that do not produce a carbapenemase are of lesser epidemiological importance because these organisms have not demonstrated the same potential for rapid transmission and regional spread.1
In this issue of JAMA, Guh and colleagues5 report on a critical step toward control of CRE in the United States. The investigators conducted active, laboratory-based surveillance through the US Centers for Disease Control and Prevention (CDC) Emerging Infections Program (EIP) Multi-site Gram-negative Surveillance Initiative. Three metropolitan EIP sites (in Georgia, Minnesota, and Oregon) participated in 2012, and 4 additional sites (in Colorado, Maryland, New Mexico, and New York) joined in 2013, covering a total population of 13.2 million. All incident cases underwent medical record review. Standardized incidence ratios adjusted for age and race were calculated for each catchment area to facilitate comparisons between sites. The result is the most comprehensive evaluation to date of the burden of CRE in the United States.
As Guh et al5 report, during 2012-2013, 599 incident CRE cases were identified in 481 individuals across the 7 EIP sites, with CRE detected in urine in 520 cases and in blood in 68 cases. Overall, 51 patients (9.0%) with incident CRE infection died. These findings suggest both bad news and perhaps some good news. The bad news is that CRE were identified in every surveillance site in the Multi-site Gram-negative Surveillance Initiative network (range of standardized incidence ratios per 100.000 population, 0.35 [95% CI, 0.14-0.74] in 2012 in Oregon to 4.80 [95% CI, 3.89-5.85] in 2013 in Maryland), with incidence rates in some regions high enough to suggest that CRE are endemic. The good news is that even in the region with the highest number of cases, the estimated crude incidence rate of CRE was relatively low (4.68 per 100.000 population in Georgia). The overall crude incidence rate of CRE in the network was estimated at 2.93 per 100.000. As noted by the authors, this rate is much lower than population-based estimates for established health care–associated pathogens, such as methicillin-resistant Staphylococcus aureus (25.1 per 100.000) or Clostridium difficile (147.2 per 100.000), and suggests that interventions implemented now to control CRE could have a sizeable effect.4,6
The study by Guh et al relied on data generated by local clinical laboratories, which resulted in some limitations. For instance, because knowledge of the mechanism of carbapenem resistance is not needed for treatment decisions, clinical laboratories usually do not characterize the mechanism of carbapenem resistance in CRE, thus, it was not known for most isolates. Susceptibility testing methods likely differed among laboratories, and the accuracy of laboratory results could not be guaranteed. Not all commercial laboratories in catchment areas participated in surveillance, which could have resulted in incomplete case ascertainment.
In about one-third of cases, isolates were available and analyzed in a CDC laboratory. The only carbapenemase detected was K pneumoniae carbapenemase (90/188 isolates; 47.9% [95% CI, 40.6%-55.1%]), but this percentage may not be representative because the sample of isolates was not random. Ideally, so that the most effective infection prevention efforts can be instituted, the specific mechanism of carbapenem resistance in CRE should be determined, particularly when CRE are first observed in a health care facility or region, and in outbreaks. Yet, most clinical laboratories and many state public health laboratories lack this capability.
Another limitation of the study5 was that isolates that were resistant only to ertapenem were excluded from the case definition, likely reducing the sensitivity of case finding.7 In January 2015, the CDC modified the CRE surveillance definition to include ertapenem resistance.7 Although the new definition improves sensitivity of CRE detection, the ability to discriminate carbapenemase-producing CRE is poor, with false-positive rates ranging from 27.1% to 76.9% in different US regions.7 The increasing number of New Delhi metallo-β-lactamase and OXA-48-like oxacillinase reports pose another challenge for accurate CRE surveillance8 because little information is available on the performance characteristics of either surveillance definition for CRE that express these or other less common resistance mechanisms.7 The surveillance definition change and continued emergence of new mechanisms of carbapenem-resistance may limit the usefulness of the report by Guh et al5 as a baseline measure against which to compare future CRE incidence.
Guh et al identified most CRE cases (253/386; 65.5%) in an outpatient setting or emergency department; however, more than 90% had multiple comorbid conditions or had been exposed to an inpatient health care facility during the previous year. Exposure to a short-stay acute care hospital (399/531; 75.1%) was more common than to a long-term acute care hospital (42/318; 13.2%) or a long-term care facility (259/531; 48.8%). Other investigators have reported high incidence rates of CRE in postacute care settings, particularly long-term acute care hospitals.9,10 Although the study by Guh et al may appear at odds with those other reports, the discrepancy might reflect differences in regional epidemiology that require distinct prevention strategies. Alternatively, the results may be explained by more temporally distant acquisition of CRE by patients; ie, carriage of CRE for more than 3 years after hospital discharge has been reported.11 Molecular subtyping of CRE isolates using a highly discriminatory method such as whole genome sequencing could clarify the epidemiology, as well as more precisely identify and locate sources of CRE acquisition and determine whether hot spots for incident colonization or infection exist; prevention efforts in such facilities may yield benefit.6 Subtyping might also help to ascertain whether the approximately 8% of case-patients who did not have recent health care exposure acquired CRE in the community, which would be an alarming finding. More than half of hospitalized patients in the study5 were discharged to a long-term care facility, underscoring the importance of interfacility communication of CRE status for regional control.6
The study by Guh et al represents an important step forward for CRE control in the United States. Expansion of surveillance to more geographic regions, including rural settings and metropolitan areas known to have high prevalence of CRE, would provide a more complete picture of the US burden. Molecular characterization of isolates would also inform prevention efforts. Whether the resources needed for this work will be made available is unclear. The 2014 presidential executive order on combating antibiotic resistance contained actions to strengthen national surveillance efforts for resistant bacteria, including the establishment of regional public health laboratories with advanced molecular diagnostic capabilities.12 These actions were not approved for funding in FY 2015; however, an appropriation to support the initiative currently awaits congressional approval of the FY 2016 federal budget. In the meantime, physicians, infection control practitioners, and public health workers will continue to rely on the Multi-site Gram-negative Surveillance Initiative and other surveillance networks10,13- 15 to measure the extent of CRE and estimate the effects of prevention efforts.