Antimicrobial resistance phenotypes, disinfectant susceptibility, and resistance-associated gene detection in carbapenem-resistant Pseudomonas aeruginosa from two tertiary hospitals in Urumqi, China.

OBJECTIVE: To characterise antimicrobial resistance phenotypes, in vitro susceptibility to four commonly used disinfectants, and the distribution of resistance-associated genes among clinical carbapenem-resistant Pseudomonas aeruginosa (CRPA) isolates from two tertiary hospitals in Urumqi, China, in order to provide local surveillance evidence to inform antimicrobial stewardship and infection prevention and control.

METHODS: Eighty nonduplicate CRPA isolates were consecutively recovered from clinical specimens at two tertiary hospitals in Urumqi between January 2025 and March 2026. Antimicrobial susceptibility was determined using the VITEK 2 system, with disk diffusion performed when confirmatory testing was required, and interpreted according to CLSI M100-S31 (2021). Carbapenemase production was assessed by the modified carbapenem inactivation method (mCIM). A broth microdilution assay adapted from antimicrobial susceptibility testing, and not a standardised biocide efficacy test, was used to compare minimum inhibitory concentrations (MICs) and minimum bactericidal concentrations (MBCs) of benzalkonium bromide, ethanol, povidone-iodine, and a chlorine-containing disinfectant between isolate groups. Polymerase chain reaction (PCR) was performed to detect carbapenem resistance-associated genes (blaKPC, blaNDM, oprD, blaOXA-10, blaOXA-48, blaVIM, blaIMP) and disinfectant resistance genes (aceI, emrE, merA, qacEΔ1-sul1, tehA, oqxA). Outcomes were organised into one primary and three exploratory families, with Holm and Benjamini-Hochberg correction applied within families respectively.

RESULTS: Of the 80 CRPA isolates, 24 (30.00%) were extensively drug-resistant (XDR) and 56 (70.00%) were non-XDR. Resistance rates to imipenem and meropenem were 83.70% and 68.70% respectively, and no isolate was susceptible to either agent. Overall resistance rates to aztreonam, piperacillin/tazobactam, levofloxacin, ceftazidime, cefepime, amikacin, and tobramycin were 37.50%, 30.00%, 26.30%, 25.00%, 21.30%, 12.50%, and 6.30% respectively. The expected oprD amplicon failed to amplify in 55 isolates (68.75%). mCIM was positive in 23 isolates (28.75%), with 95.0% agreement with PCR-detected carbapenemase gene carriage (Cohen's κ 0.871, 95% CI 0.748-0.994). Detection rates for blaKPC, blaNDM, blaVIM, and blaIMP were 10.00%, 6.25%, 5.00%, and 2.50% respectively; neither blaOXA-10 nor blaOXA-48 was detected. qacEΔ1-sul1, merA, tehA, and oqxA were detected in 36.25%, 16.25%, 7.50%, and 2.50% of isolates respectively, whereas aceI and emrE were not detected. For ethanol the MIC50 and MIC90 were 75% (v/v) in both groups; for the chlorine-containing disinfectant and povidone-iodine the MIC50/MIC90 were 512/1024 mg/L in both groups, and neither MICs nor MBCs differed between groups. Benzalkonium bromide MICs ranged from 32 to 256 mg/L in XDR and 16 to 256 mg/L in non-XDR isolates, and MBCs from 64 to 512 and 64 to 256 mg/L respectively; the MIC difference was the primary outcome and remained significant after Holm correction (adjusted P < 0.001), whereas the MBC difference did not survive correction when considered alongside the other disinfectant endpoints (Benjamini-Hochberg q=0.144). qacEΔ1-sul1 was more frequent among XDR isolates in unadjusted analysis (58.3% versus 26.8%; OR 3.83, 95% CI 1.40-10.45; unadjusted P = 0.011) but did not retain significance after correction within the eight-test gene family (q = 0.088). Resistance to aztreonam and meropenem was nominally more frequent among qacEΔ1-sul1-positive isolates (55.20% versus 27.50% and 86.2% versus 58.80%) but neither association survived correction (q=0.072 for both; Holm-adjusted P = 0.111 and 0.100). Direct comparison of benzalkonium bromide endpoints by qacEΔ1-sul1 status showed no difference (median MIC 64 mg/L in both groups, P = 0.097; median MBC 128 mg/L in both groups, P = 0.697).

CONCLUSIONS: CRPA isolates from these two hospitals showed uniform nonsusceptibility to carbapenems, with failure to amplify the expected oprD target the most frequent molecular observation. XDR isolates showed higher benzalkonium bromide MICs in vitro compared with non-XDR isolates. qacEΔ1-sul1 was the predominant disinfectant resistance gene detected, but its association with the XDR phenotype did not survive correction for multiple comparisons, and carriage was not accompanied by any measurable difference in benzalkonium bromide susceptibility. These findings provide baseline epidemiological data for Urumqi, though whole-genome sequencing, molecular typing, expression analysis, and standardised biocide efficacy testing would be required to establish mechanistic determinants.