Abstract
Accurate detection of carbapenemase-producing Gram-negative bacilli is of utmost importance for the control of nosocomial spread and the initiation of appropriate antimicrobial therapy. The modified Hodge test (MHT), a carbapenem inactivation assay, has shown poor sensitivity in detecting the worldwide spread of New Delhi metallo-β-lactamase (NDM). Recent studies demonstrated that NDM is a lipoprotein anchored to the outer membrane in Gram-negative bacteria, unlike all other known carbapenemases. Here we report that membrane anchoring of β-lactamases precludes detection of carbapenemase activity by the MHT. We also show that this limitation can be overcome by the addition of Triton X-100 during the test, which allows detection of NDM. We propose an improved version of the assay, called the Triton Hodge test (THT), which allows detection of membrane-bound carbapenemases with the addition of this nonionic surfactant. This test was challenged with a panel of 185 clinical isolates (145 carrying known carbapenemase-encoding genes and 40 carbapenemase nonproducers). The THT displayed test sensitivity of >90% against NDM-producing clinical isolates, while improving performance against other carbapenemases. Ertapenem provided the highest sensitivity (97 to 100%, depending on the type of carbapenemase), followed by meropenem (92.5 to 100%). Test specificity was not affected by the addition of Triton (87.5% and 92.5% with ertapenem and meropenem, respectively). This simple inexpensive test confers a large improvement to the sensitivity of the MHT for the detection of NDM and other carbapenemases.