The time frame for conducting this study was from February 2023 to December 2023. The inclusion criteria of the patients were being clinically non-responding to antibiotics, and the exclusion criteria were not having nosocomial infections and being outdoor patients or staying at the hospital for less than 7 days. According to patient records, a total of 132 non-duplicate isolates were obtained from distinct samples (wound swab, urine, tracheal aspirate, pus, and sputum) of patients suffering from nosocomial infections (NIs) at Dhaka Medical College Hospital (DMCH), Bangladesh. Information was gathered about the patients’ age, sex, and place of origin for the samples. Every isolate was obtained from a separate patient, and the patients were not associated with epidemic outbreaks. To identify pure colonies of P. aeruginosa, the isolates were picked from the MacConkey agar plates, onto which the raw samples were cultured, and then streaked onto Cetrimide agar. Pseudomonas aeruginosa isolates were determined by their colony characteristics on Cetrimide agar medium followed by a series of biochemical tests (Gram staining, catalase, oxidase, sugar fermentation, nitrate reduction, motility, indole, urea, citrate utilization, methyl red, and Voges-Proskauer) and molecular identification based on 16S rRNA gene sequencing for further study.

The pathogens’ antibiogram was assessed using the disk diffusion method onto Mueller–Hinton agar in accordance with the CLSI (Clinical and Laboratory Standards) guidelines. The antimicrobials tested for resistance included ciprofloxacin (5 μg), co-trimoxazole (25 μg), amikacin (30 μg), ceftazidime (30 μg), piperacillin/tazobactam (100/10 μg), levofloxacin (5 μg), amoxicillin/clavulanic acid (20/10 μg), doxycycline (30 μg), aztreonam (30 μg), ceftriaxone (5 μg), meropenem (10 μg), netilmicin (10 μg), gentamicin (10 μg), and tigecycline (30 μg). The strains were classified as pandrug-resistant (PDR, non-susceptible to all antimicrobial classes), extensively drug-resistant (XDR, non-susceptible to all but two classes), multidrug-resistant (MDR, non-susceptible to ≥3 antimicrobial classes), and low-level drug-resistant (LDR).

Hemolysin and pigment production were evaluated on blood agar (8%) and cetrimide agar plates, respectively. A clear zone around the colonies elucidated the beta-hemolytic activity of the P. aeruginosa strains, and the lack of a clear halo was identified as a negative result. Pigments produced on the cetrimide agar were pyoverdine (green), pyocyanin (blue-green), and fluorescein (yellow).

Proteolytic enzyme (gelatinase) production was studied by stabbing in 4% gelatin media containing test tubes. Hydrolysis (partial or total) of gelatin indicated the production of gelatinase enzyme by the isolates, while a lack of hydrolysis indicated a negative result.

As three types of motility can be detected in P. aeruginosa strains, a motility assay was carried out on Luria–Bertani (LB) agar plates using different concentrations of agar to evaluate the following types of motility: twitching (1% agar), swarming (0.5% agar), and swimming (0.3% agar). The media were inoculated with the overnight cultured strains using sterile yellow tips on the surface of the agar to study swarming, by stabbing halfway down the thick agar media for evaluating swimming and for twitching the strains were introduced at the bottom of the agar. All culture plates were kept overnight at a 37°C incubator. To evaluate the twitching motility, the 1% agar plates were flooded with TM developer solution and left for up to 30 min. The colony on the agar surface was scraped away by a sterile loop, and the solution was decanted. The remaining interstitial colony attached to the Petri plate is indicative of the twitching motility of the isolates.

P. aeruginosa strains were inoculated in LB broth on a 96-well microtiter plate to quantify biofilm formation. As a negative control, a sterile broth was used to assure sterility. Biofilm producers were differentiated from non-biofilm producers by measuring their OD values. If the value is ≤ (average OD of negative control +3ơ), the strain was considered non-biofilm producer, and the isolates having OD value > (ANC + 3ơ) were interpreted as biofilm producer strains.

DNA extraction of clinical P. aeruginosa isolates was carried out by boiling DNA method. The presence of 6 virulence genes (plcH, lasB, exoS, algD, toxA, and aprA) and 10 MBL genes (bla_VIM-2, bla_SPM, bla_SIM, bla_DIM, bla_BIC, bla_OXA-48, bla_GIM, bla_SHV, bla_DHA, and bla_AIM) was evaluated by amplifying the genes with specific primers listed in Tables 1, 2, respectively. Previously well-characterized strains were used as controls in this study. PCR conditions were maintained as in the reference articles, respectively. Gel electrophoresis was conducted by running the amplified PCR products for 50 min at 100 V using a 1.5% agarose gel using 1X TBE buffer. The sizes of the amplicons were determined by comparing them with the mobility of the 100 bp and 1 kb + DNA ladder (NEB, UK). The amplicons were visualized, and the gels were photographed using the gel documentation system (Axygen, USA) having a UV trans-illuminator.

Using binary logistic regression analysis, with a 95% confidence interval at a 5% significance level, the relationship between MDR phenotypes and the virulence factors as well as epidemiological and clinical factors was assessed. The statistical program SPSS version 25 (IBM, New York, USA) was used for the analysis.

Initially collected 132 isolates taken from patients aged 25–68 years were undergone biochemical tests (Supplementary Table S1), and of them, 108 isolates were plausibly identified as Pseudomonas aeruginosa. Twenty-five selected isolates out of 108 were further identified as P. aeruginosa using molecular techniques based on 16S rRNA gene sequencing (Supplementary Table S2). Among the patients, 46% were male and 54% were female having low fever and UTI.

The prevalence rate of antibiotic resistance was found to be extremely high (53.7–98.1%) for all antimicrobial agents tested among the 108 clinical isolates of P. aeruginosa. In comparison with the other 13 antibiotics (doxycycline, meropenem, ceftazidime, tigecycline, levofloxacin, aztreonam, amoxiclav, netilmicin, gentamycin, co-trimoxazole, ceftriaxone, ciprofloxacin, and amikacin), P. aeruginosa strains exhibited significantly less resistance (53.7%) against piperacillin–tazobactam. Comparing the isolates to the other antibiotics revealed the highest level of resistance to aztreonam (98.1%) which belongs to the class monobactam (Figure 1). Amoxiclav (93.5%), ceftriaxone (94.4%), and doxycycline (95.4%) are the remaining drugs that came next in order of resistance (Figure 1).

Out of 108 isolates, only 1 was found to be resistant to each of the 14 antibiotics across 8 distinct classes. The other 107 strains were found to be resistant to at least one of the antibiotics tested. Of them, one strain was resistant to two antibiotics of different classes, while two isolates were found to be resistant to only one antibiotic. These strains of P. aeruginosa (2.7%) fall into the low-level drug-resistant (LDR) category. A total of 27.8% of the isolates were classified as MDR strains because they exhibited resistance to at least three classes of antibiotics. Since 58.3% showed no resistance to at least two classes of antibiotics, they were classified as extensively drug-resistant (XDR). A total of 10.2% of the clinical P. aeruginosa strains tested positive for every antibiotic and thus fell under the category of pandrug-resistant (PDR).

PCR and gel electrophoresis were performed to detect the presence of beta-lactamase genes in the P. aeruginosa isolates (Figure 2). The rates of different classes of beta-lactamase (Ambler classes A, B, C, and D) gene production among the clinical Pseudomonas aeruginosa isolates are shown in Figure 3. Among the 10 beta-lactamase genes, 6 are under class B (bla_VIM-2, bla_SPM, bla_SIM, bla_DIM, bla_GIM, and bla_AIM), 2 of them are classified as class A (bla_BIC and bla_SHV), 1 is from class D (bla_OXA-48), and another 1 is from class C (bla_DHA). Most of the isolates (74.1%) were detected positive for MBL genes, which are under class B beta-lactamases. Among them, 37.5% harbored the bla_VIM-2 gene, followed by bla_SPM (32.5%), bla_AIM (27.5%), and bla_GIM (2.5%) genes. Only one isolate was found to be positive for the bla_DIM gene, while the bla_SIM gene was not identified in any of the isolates. The highest rate of MBL gene production was found in strains isolated from wound swabs (50.1%).

Class D beta-lactamase (bla_OXA-48) gene was found in 23.2% of the P. aeruginosa isolates, followed by class A beta-lactamase genes (n = 3, 2.7%). No isolates were found to harbor the class C beta-lactamase gene.

All but two isolates of P. aeruginosa showed different types of motilities, and most of the strains (84.26%) were capable of swimming which requires rotating flagella, followed by twitching (80.56%) which indicates the existence of type IV pili and swarming (59.25%). The second most frequently found virulence factor synthesized by the P. aeruginosa isolates was pigment (88.89%). Among the pigment-producing strains, 70.83% produced fluorescein and subsequently pyoverdine (28.12%) and pyocyanin (9.37%). A total of 7.30% of the strains produced both pigments, fluorescein, and pyoverdine. Gelatinase production was marked in 84.26% of the P. aeruginosa strains followed by hemolysin (73.15%) and biofilm formation (67.59%) (Figure 4).

Among the six virulence genes, plcH and lasB were most frequent among the P. aeruginosa isolates (92.59 and 94.44%, respectively) (Figure 5). A total of 82.41% of the strains harbored the toxA gene, and the prevalence of the toxA gene among MDR P. aeruginosa strains was 90.27%. A total of 70.37% of the isolates were found to be positive for the algD gene. A total of 62.96 and 55.56% of the isolates were found positive for aprA and exoS genes, respectively. In addition, 18.52% of the strains were positive for all the virulence genes tested. The frequency of the virulence genes among different sample types is depicted in Table 3.

Biofilm formation, gelatinase activity, hemolysin, and pigment production, which are known virulence factors for P. aeruginosa, were found to have no statistically significant correlation with the MDR pattern of P. aeruginosa isolates. Among the three types of motilities, swimming and swarming had no association while twitching appeared to be significantly associated (p < 0.05) with the level of antibiotic resistance. MDR strains were statistically more likely to harbor the toxA gene than the other five when analyzing the virulence genes (Figure 6).

In the statistical model developed for determining the virulence factors as the risk factors for the MDR pattern, the strongest predictor of multidrug resistance was the toxA gene, recording an odds ratio (OR) of 8.45 (Table 4). This indicated that the P. aeruginosa isolates containing the toxA gene were 8.45 times more likely to be multidrug-resistant than the isolates lacking the toxA gene. In addition, the odds ratio for twitching motility was 3.31, which indicated that the existence of this type of motility, which is acquired by type IV pili, enhanced the risk of the strains to be MDR by 3.31 times, as found in the statistical analysis. Epidemiological clinical factors such as patient’s age, sex, and types of samples were analyzed and not found to be statistically related to antibiotic resistance.