A comprehensive, intensive patient surveillance program for carbapenemase-producing bacteria at the NIH Clinical Center

Background: Carbapenemase-producing organisms (CPO) are important nosocomial pathogens that spread from colonized patients. Following a 2011 outbreak of CPO at the NIH Clinical Center, the hospital prioritized early identification and isolation of CPO carriers. Admission perirectal surveillance cultures were added in 9/13 to an existing surveillance program. We describe our experience during the first 2.5 years of CPO admission surveillance.

Methods: From 9/13 to 4/16, perirectal swabs were automatically ordered for all patients on admission to non-behavioral health wards. Manual orders were placed for swabs twice weekly in the ICU, weekly in other high-risk wards, and monthly for point prevalence throughout the hospital. Patients hospitalized in the U.S. in the previous week or abroad in the previous 6 months met CDC criteria for higher risk of colonization, and were empirically isolated pending 2 perirectal swabs.

Swabs (n=18,487) were cultured onto chromogenic media. All Gram negative bacilli (GNB) were identified by MALDI-TOF MS and screened by PCR for blaKPC/blaNDM (blaOXA48 was tested for specific cases). Selected GNB were subject to susceptibility testing. Whole-genome sequencing was used for epidemiological strain typing of CPOs. Some perirectal swabs were tested directly by bla_KPC PCR (n=770) in response to a known bla_KPC carrier or when selective media was unavailable.

Results: Among 26,495 orders for surveillance swabs, 19,257 swabs were collected from 6,440 patients, a 73% compliance rate. Out of 21 CPO-colonized patients identified from 9/13 through 4/16, 11 were identified via admission surveillance, 4 in point prevalence surveys, 4 from high-risk ward surveillance, and 2 from clinical cultures. Sequencing demonstrated no relatedness among CPO isolates. Although only 0.18% of patients sampled on admission were found to be colonized with CPO, those meeting CDC high-risk criteria were 38 times as likely to be colonized on admission.

Conclusion: At a hospital with a geographically broad referral base, more than half of CPO-colonized patients were identified on admission. Despite aggressive admission surveillance, some carriers were identified well after admission. Every CPO carrier who is identified and isolated early represents a potentially averted reservoir for nosocomial transmission.