Carbapenem-producing Enterobacteriaceae (CPE) are on the rise worldwide. National guidelines for the prevention and control of CPE recommend screening for the detection of asymptomatic carriers on admission.

To evaluate the benefit of serial screens for detecting the carriage of CPE and other antibiotic-resistant Gram-negative bacteria following hospital admission.

All CPE screens, which were cultured on chromogenic media and the presence of a carbapenemase confirmed by polymerase chain reaction, were analysed for a six-month period. National guidelines in England recommend three serial screens for CPE separated by 48 h for admission screening for ‘at-risk’ patients, during which the patient is isolated. Two screening scenarios were tested. In scenario A, patients received three screens at the specified timepoints, in line with English national guidelines; in scenario B, patients received three consecutive screens, but not necessarily within the specified timepoints, during one admission. General linear models or conditional logistic regression were used to detect any significant change in the rate of carriage.

There was no significant increase in the detected carriage rate of CPE across any of the three timepoints in the scenarios tested. However, there was a significant increase in the detected rate of carriage of Gram-negative bacteria, Enterobacteriaceae, and resistant Enterobacteriaceae (excluding CPE) in scenario B.

Three serial screens were not useful for the detection of CPE carriage on admission. The increase in the carriage rate of other Gram-negative bacteria may be explained by ‘unmasking’ of pre-existing carriage, or acquisition. This argues for regular screening of long-stay patients.