Back

Evaluation of Six Commercial and Non-Commercial Colistin Resistance Diagnostics

Leshaba, T. M. S.; Mmatli, M.; Skosana, L.; Mbelle, N.; Osei Sekyere, J.

2024-03-22 pathology
10.1101/2024.03.21.24304472 medRxiv
Show abstract

BackgroundResistance to colistin, a last-reserve antibiotic used for treating drug-resistant infections, is increasing globally. This study evaluated six diagnostic tests designed to detect colistin-resistant pathogens. MethodsPCR and broth microdilution assays (BMD) were used to respectively characterize the molecular mechanisms and phenotypic colistin resistance of 142 Gram-negative bacterial isolates and controls. The sensitivity, specificity, positive- and negative-predictive values, major (ME) and very major errors (VME), categorical and essential agreements (EA) of ComASP Colistin, CHROMagar COL-APSE, Rapid NP Test, Sensititre, MicroScan, and Vitek 2 were determined with these isolates; the BMD was used as gold standard. ResultsThe Vitek 2, Sensititre, and ComASP tests were more efficient, albeit with concerning ME and VMEs and low EAs. Sensititre was 100% specific with 0% ME and 3.61% VME; Vitek 2 had the least VME (1.25% and 0%) and a low EA (57.50%). ComASP had an EA of 75.35%. MicroScan was highly sensitive (96.55%) but less specific (87.50%), with very below-accetable EAs (48.11%). The CHROMAgar COL-APSE efficiently identified the species with their unique colours but was the least specific (67.80%), with the highest ME (32.20%) and high VME (7.23%). The Rapid NP test had the highest VME (7.84%), producing results within 4 hours with 92.16% sensitivity and 96.08% specificity. ConclusionVitek 2, MicroScan, ComASP colistin, and Sensititre are good for determining colistin resistance; the latter two tests are recommendable for low-resourced laboratories. The in-house Rapid NP test has short turnaround time with high efficiency for initial resistance screening.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.