Rapid phenotypic testing for detection of carbapenemase- or extended-spectrum ß-lactamase-producing Enterobacterales directly from blood cultures: a systematic review and meta-analysis.

BACKGROUND: Early identification of extended-spectrum ß-lactamase (ESBL) and carbapenemase-producing Enterobacterales (CP-CRE) is critical for timely therapy. Rapid phenotypic tests identifying these resistance mechanisms from pure bacterial colonies have been developed. OBJECTIVES: To determine the operating characteristics of available rapid phenotypic tests when applied directly to positive blood cultures. METHODS OF DATA SYNTHESIS: Bivariate random effects models were used unless convergence was not achieved where we used separate univariate models for sensitivity and specificity. DATA SOURCES: MEDLINE, CENTRAL, Embase, BIOSIS, and Scopus from inception to 16 March 2021. STUDY ELIGIBILITY CRITERIA: Studies using any rapid phenotypic assay for detection of ESBL or CP-CRE directly from blood cultures positive for Enterobacterales, including those utilizing spiked blood cultures. Case reports/series, posters, abstracts, review articles, those with ≤5 resistant isolates, and studies lacking data or without full text were excluded. PARTICIPANTS: Consecutive patient samples (main analysis) or spiked blood cultures (sensitivity analysis). TESTS: Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry assays (MALDI-TOF) and commercially available chromogenic or immunogenic assays. REFERENCE STANDARD: Conventional laboratory methods and/or polymerase chain reaction (PCR). ASSESSMENT OF RISK OF BIAS: Quality Assessment of Diagnostic Accuracy Studies Version 2 (QUADAS-2). RESULTS: For detection of the ESBL phenotype the respective pooled sensitivities and specificities for consecutive clinical samples were as follows: 94% (95% CI 93-99%) and 97% (95% CI 95-100%) for MALDI-TOF/mass spectrometry (n = 1); and 98% (95% CI 92-100%) and 100% (95% CI 96-100%) for chromogenic assays (n = 7). For the CP-CRE phenotype the respective pooled sensitivity and specificities for consecutive clinical samples were as follows: 100% (95% CI 99-100%) and 100% (95% CI 100-100%) for MALDI-TOF (n = 2); 96% (95% CI 77-99%) and 100% (95% CI 81-100%) for chromogenic assays (n = 4); and 98% (95% CI 96-100%) and 100% (95% CI 100-100%) for immunogenic testing (n = 2). CONCLUSIONS: Rapid phenotypic assays that can be directly applied to positive blood cultures to detect ESBL and carbapenemase production from Enterobacterales exist and, although clinical studies are limited, they appear to have high sensitivity and specificity. Their potential to facilitate patient care through timely identification of bacterial resistance should be further explored.