RESUMEN
Heteroresistance (HR) to colistin is especially concerning in settings where multi-drug-resistant (MDR) K. pneumoniae are prevalent and empiric use of colistin might lead to treatment failures. This study aimed to assess the frequency of occurrence of colistin HR (CHR) among (MDR) K. pneumoniae (n = 676) isolated from patients hospitalized in 13 intensive care units (ICUs) in six European countries in a clinical trial assessing the impact of decolonization strategies. All isolates were whole-genome-sequenced and studied for in vitro colistin susceptibility. The majority were colistin-susceptible (CS) (n = 597, MIC ≤ 2 µg/mL), and 79 were fully colistin-resistant (CR) (MIC > 2 µg/mL). A total of 288 CS isolates were randomly selected for population analysis profiling (PAP) to assess CHR prevalence. CHR was detected in 108/288 CS K. pneumoniae. No significant association was found between the occurrence of CHR and country, MIC-value, K-antigen type, and O-antigen type. Overall, 92% (617/671) of the K. pneumoniae were MDR with high prevalence among CS (91%, 539/592) and CR (98.7%, 78/79) isolates. In contrast, the proportion of carbapenemase-producing K. pneumoniae (CP-Kpn) was higher among CR (72.2%, 57/79) than CS isolates (29.3%, 174/594). The proportions of MDR and CP-Kpn were similar among CHR (MDR: 85%, 91/107; CP-Kpn: 29.9%, 32/107) and selected CS isolates (MDR: 84.7%, 244/288; CP-Kpn: 28.1%, 80/285). WGS analysis of PAP isolates showed diverse insertion elements in mgrB or even among technical replicates underscoring the stochasticity of the CHR phenotype. CHR isolates showed high sequence type (ST) diversity (Simpson's diversity index, SDI: 0.97, in 52 of the 85 STs tested). CR (SDI: 0.85) isolates were highly associated with specific STs (ST101, ST147, ST258/ST512, p ≤ 0.003). The widespread nature of CHR among MDR K. pneumoniae in our study urge the development of rapid HR detection methods to inform on the need for combination regimens.
RESUMEN
Colistin heteroresistance has been identified in several bacterial species, including Escherichia coli and Klebsiella pneumoniae, and may underlie antibiotic therapy failures since it most often goes undetected by conventional antimicrobial susceptibility tests. This study utilizes population analysis profiling (PAP) and time-kill assay for the detection of heteroresistance in K. pneumoniae and for evaluating the association between in vitro regrowth and heteroresistance. The mechanisms of colistin resistance and the ability of combination therapies to suppress resistance selection were also analysed. In total, 3 (18%) of the 16 colistin-susceptible strains (MIC ≤ 2 mg/L) were confirmed to be heteroresistant to colistin by PAP assay. In contrast to the colistin-susceptible control strains, all three heteroresistant strains showed regrowth when exposed to colistin after 24 h following a rapid bactericidal action. Colistin resistance in all the resistant subpopulations was due to the disruption of the mgrB gene by various insertion elements such as ISKpn14 of the IS1 family and IS903B of the IS5 family. Colistin combined with carbapenems (imipenem, meropenem), aminoglycosides (amikacin, gentamicin) or tigecycline was found to elicit in vitro synergistic effects against these colistin heteroresistant strains. Our experimental results showcase the potential of combination therapies for treatment of K. pneumoniae infections associated with colistin heteroresistance.