Active Surveillance and Isolation of Asymptomatic Carriers of Clostridium difficile at Hospital Admission: Containing What Lies Under the Waterline.

Abstract

During the 2000s, the incidence of Clostridium difficile infection (CDI) increased dramatically, in large part due to the emergence of a hypervirulent strain, BI/NAP1/027, responsible for many hospital outbreaks in the United States. In 2011, C difficile accounted for 12% of all US health care–associated infections, surpassing Staphylococcus aureus as the most common cause of such infections.1 Hospitalized patients with CDI are a recognized source of health care–associated (HA) transmission, and a primary control measure is to limit the spread of the organism from symptomatic patients. However, increasing molecular evidence, based on genomic sequence–based methods, indicates that asymptomatic patients colonized with C difficile also contribute to transmission.2,3 A 2013 study3 found that incident CDI cases in a hospital were as frequently linked to transmission from asymptomatic carriers as to symptomatic patients. Despite the potential for patients with asymptomatic colonization to serve as a reservoir for CDI, no data currently exist to determine whether interventions targeting asymptomatically colonized patients could be effective in reducing HA-CDI. In this issue of JAMA Internal Medicine, Longtin et al4 report findings from a quasi-experimental controlled study using time series analysis to determine the effect of active surveillance and isolation of asymptomatic carriers on the incidence of HA-CDI. The authors conducted rectal sampling of all patients admitted through the emergency department of a tertiary acute care hospital over an approximate 17-month period. Clostridium difficile testing was performed using a polymerase chain reaction (PCR) assay that targeted the tcdB gene and has been demonstrated to detect the subset of carriers with heavier organism loads who were more likely to contaminate the skin or environment.5 All identified asymptomatic carriers were placed under isolation precautions until discharge. Among 7599 of 8218 patients screened, 4.8% were identified as asymptomatic carriers, which is similar to a previous study6 that included additional Quebec hospitals. However, other researchers have reported a higher prevalence of carriage on admission, ranging from 7% to 18%.7 Longtin et al4 did not provide the proportion of all patients admitted who were either transferred from other hospitals or directly admitted from long-term care facilities. Although these individuals are likely to be higher-risk patients, they were excluded from screening due to logistical restraints, and their exclusion could have contributed to the low prevalence of asymptomatic carriage. The limited sensitivity of direct screening using a commercial PCR (ie, insufficient for detecting carriers with lower organism burden) might be another reason for the lower prevalence. Nonetheless, Longtin et al4 found that the incidence of HA-CDI decreased by 7% per 4-week period during the interventionperiod, resulting in anoverall decreaseof 7.2 casesper 10000 patient-days. The interventionwas estimated to have prevented 63 of 101 expected cases. This meant that, for every single HA-CDI case prevented, 121 patients had to be screened, and 6 asymptomatic carriers had to be isolated. No concomitant decrease in the incidence of HA-CDI was detected among other hospitals in Quebec City and throughout Quebecprovince. It is conceivable thatmore caseswouldhave beenprevented ifperiodic screeningwasperformedamongpatientsafter admission, identifyingadditional carrierswhowere missed during admission screening. In addition, a greater reductionofHA-CDI incidencemighthavebeendetected if ithad not been for the previously mentioned limitations (ie, exclusion of direct admissions and transfers and use of a less sensitive screening method). Several strengths of this study should be noted. Although the study design was not as rigorous as a cluster randomized approach,multiple statisticalmethodswere used to measure the effect of the intervention, including segmented regression analysis and autoregressive integratedmoving averagemodeling, and these analyses produced similar results, while accounting for seasonality andchanges indiagnostic assays. Multiple control hospitals were also included for interhospital comparison.Other important confounders thatwere assessed includedhandhygienecomplianceandantibiotic and proton pump inhibitor use. Hand hygiene compliance increasedduring the interventionphase, but almost all handhygiene was performed with an alcohol-based hand sanitizer, which is not effective against C difficile spores. The results of this study are promising for reducing HA-CDI.Additional informationonpatient-specific factors that couldhave affected the incidence ofHA-CDI, such aswhether asymptomatic carrierswere less likely than noncarriers to receive antibiotics and therefore todevelopCDI,wouldbehelpful. Adherence to isolation precautionswas also not assessed to ensure that the intervention was adequately implemented. In addition, isolation precautions were modified to allow asymptomatic carriers to share a room with noncarriers as longas the curtain separating the2bedswasdrawn.The effectiveness of this approach in preventing C difficile transmission is unknown, andnodatawere available regarding the proportion of noncarrier roommates who subsequently Related article page 796 Effect of Detecting and Isolating Clostridium difficile Carriers Original Investigation Research