Evaluation of phenotypic and genotypic methods for detecting KPC variants

Abstract

Klebsiella pneumoniae carbapenemases (KPCs) have spread and diversified extensively. To date, 242 clinical variants have been identified and harbor different hydrolytic capacities, thereby interfering with rapid diagnostic tests. The accurate detection of KPC variants is crucial to guide treatment and control measures in healthcare settings. We constructed KPC variants to assess the mutational impact on detection capacities of resistance-based tests. KPC variants (n = 45) were characterized phenotypically and used to measure the detection sensitivity of KPC detection methods (two lateral flow immunoassays [LFIAs], three hydrolysis tests, three selective culture media, and two PCR-based tests). We identified four antibiotic susceptibility patterns: "KPC-like" (23/45; 51%), "extended-spectrum beta-lactamase-like" (6/45; 13%), "ceftazidimase" (9/45; 20%), and outlier variants with "mixed-profiles" (5/45; 11%). These phenotypes had different impacts on the detection capabilities of hydrolysis tests (0%-100%), LFIA (44%-100%), and selective culture media (0%-100%), highlighting a risk of misdiagnosis for some KPC variants. All variants were detected with PCR-based tests. To detect the maximum of KPC variants, fecal carriage screening requires a combination of selective media targeting resistance to carbapenems, third-generation cephalosporins, and ceftazidime-avibactam. From antibiotic susceptibility testing, resistance to ceftazidime ± avibactam and specific phenotypic profiles should be used as warnings to track the presence of KPC variants. We recommend LFIA as a first-line test, owing to its high sensitivity in detecting KPC variants. Nevertheless, using a combination of tests may remain wise in some situations. The spread of KPC variants remains a significant concern, particularly as reversion to ancestral phenotype could restore carbapenem resistance and lead to therapeutic failure.

Keywords: KPC beta-lactamase; LFIA tests; clinical KPC variants; detection tests; hydrolysis-based tests.

Fig 1

PCA on the inhibition zone diameters of representative beta-lactams for the 45 strains studied. Each point is labeled with the strain indicated in Table 1. The clusters “ancestral KPC,” “KPC-like,” “ESBL-like,” and “ceftazidimase” are shown in yellow, green, pink, and blue, respectively, and the “mixed” cluster (21, 22, 46, 55, and 72) is shown in red. Arrows indicate higher susceptibility to antibiotics. The first two principal components explained 79.03% (dimension 1—Fig. 1a) and 8.5% (dimension 2—Fig. 1b) of the total variance.

Fig 2

Kiviat diagram showing the distribution of mean inhibition zone diameters for the principal antibiotic molecules tested, by major phenotypic profile: “ancestral KPC,” “KPC-like,” “ceftazidimase,” and “ESBL-like.” AMX, amoxicillin; TZP, piperacillin-tazobactam; CTX, cefotaxime; CAZ, ceftazidime; ETP, ertapenem; IMP, imipenem; MEM, meropenem. Diameters: millimeters; blue crosses: EUCAST breakpoints (11).