The PDU Medical College/Hospital is a tertiary care and teaching hospital that provides a full range of health care services, including medical, surgical, and superspecialty services, to patients in and around Rajkot district. In a 100-km radius of Rajkot, there is only one multispeciality government hospital. Every day, more than 400 patients load from Rajkot city as well as rural areas of Rajkot district. Also, people from Amreli, Jamnagar, Junagadh, Kachchh, Morbi, Porbandar, Surendranagar, and Veraval visit PDU Medical College/Hospital Rajkot for treatments. PDU Medical College/Hospital is located in the center of the Saurashtra region in Rajkot, Gujarat.

Based on the data of the past 6 months, we selected the six most commonly reported pathogens from hospitalized and outpatient specimens, including Pseudomonas aeruginosa, Escherichia coli, Enterobacter spp., Klebsiella pneumoniae, Staphylococcus spp., and Acinetobacter spp. The top six most frequently observed bacterial strains were selected for the present study. Samples were collected during March to June 2022 from the in-patients and out-patients at PDU Medical College/Hospital in Rajkot, Gujarat. Samples were collected in sterile containers according to Cheesbrough (2005) in various wards by the assigned clinicians, and they were processed further at the bacteriology lab immediately. Preservation and storage of specimens varied from place to place and time to time; for example, blood, urine, and sputum were stored until the analysis and discussion with the assigned doctor. While for precious specimens like postoperative samples, body fluids were stored for 7–10 days, etc. Gender, age, ward, collection date, specimens, and other information were noted along with the sample collection. Ten samples were randomly selected daily for this survey, conducted for 3 months. PEEKSA (Pseudomonas aeruginosa, Escherichia coli, Enterobacter spp., Klebsiella pneumoniae, Staphylococcus spp. and Acinetobacter spp) were isolated from the collected samples (CLSI, 2018). Blood samples collected from adult patients (10–20 mL) and pediatric patients (5–10 mL) were mixed in brain heart infusion broth bottles and analyzed for bacterial growth up to 7 days using automated blood culture system BD-BACTEC FX40 (United States) (Procop et al., 2020). Absence of turbidity after 7 days was considered negative. Positive samples were subcultured on blood agar, nutrient agar, and MacConkey agar plates. Urine samples (20–30 mL) were collected in a sterile plastic container (50 mL capacity) and streaked on MacConkey agar, blood agar, cysteine lactose electrolyte deficient agar, and nutrient agar plates. Pus/swab samples were streaked on nutrient agar, blood agar, and MacConkey agar plates. Sputum samples (2–5 mL) were collected in a sterile plastic container (50 mL capacity) and streaked on chocolate agar, blood agar, nutrient agar, and MacConkey agar plates. The samples except blood were processed the same day; streaked plates were incubated at 35°C for 24–72 h and subcultured on nutrient agar plates. Bacterial identification was done based on colony morphology, biochemical tests, and the Gram reaction (Collee et al., 1996; Procop et al., 2020).

Gram-positive and-negative bacterial isolates were evaluated for antibiotic susceptibility employing Kirby-Bauer disk-diffusion method (CLSI, 2018). A single colony was picked and suspended in sterile normal saline (0.85% NaCl) to generate the equivalent of 0.5 McFarland standard solution. 1 mL of bacterial suspension was mixed with 19 mL of sterile Muller-Hinton soft agar (45°C) and poured in Petri plates, incubated at 35°C for 24 h after the transfer of antibiotics disks by disk dispenser. PEEKSA isolates were classified as resistant, intermediate, and sensitive to antibiotics on the basis of the size of the zone of inhibition according to the Clinical Laboratory Standard Institute (CLSI) guidelines. Antibiotics used for antibiotic susceptibility test of Enterobacter spp., Klebsiella pneumoniae, and Escherichia coli isolated from pus, swabs, sputum, and blood samples were ciprofloxacin (CIP), levofloxacin (LVX), gentamicin (GM), amikacin (AN), meropenem (MEM), cefuroxime (CXM), cefotaxime (CTX), ceftazidime (CAZ), ceftazidime-clavulanate (CAC), cefepime (FEP), piperacillin-tazobactam (TZP), trimethoprim-sulfamethoxazole (SXT), tetracycline (TE), and ampicillin-sulbactam (SAM). Strains isolated from urine samples were also evaluated for sensitivity against the antibiotics TR, NX, and NIT along with the above-mentioned antibiotics. Imipenem-EDTA (IE), imipenem (IPM), aztreonam (ATM), CIP, GM, LVX, AN, and FEP were employed for the evaluation of the antibiotic sensitivity of Pseudomonas aeruginosa. Staphylococcus spp. were against erythromycin (E), linezolid (LZD), CIP, rifamycin (RIF), clindamycin (CM), TE, vancomycin (VA), SXT, GM, chloramphenicol (C), meropenem (MEM), cefoxitin (FOX), and penicillin (P). Antibiotics class, abbreviation, and concentration in Supplementary Table S1.

According to the CLSI recommendations, bacterial pathogens were classified as sensitive, intermediate, and resistant based on antibiotic susceptibility values (CLSI, 2022). Descriptive statistics such as relative abundance, percentage of categorical variation, and frequency were calculated. The chi-square test was used to compare the abundance of bacterial isolates with patient specimens. A Tukey post-hoc test was conducted for multiple comparisons between the mean values of the number of resistant antibiotics. p values less than 0.05 were considered statistically significant. Antibiotic resistance index (R_I) was derived by n/N where, “n” is number of resisrant isolates and “N” is the total number of isolates tested. Pearson’s correlation analysis was studied between isolation source and its antibiotic resistance pattern (antibiotics that are tested more than 90% were used for the correlation study). Natural selection was determined by (n/N) × 100, where “n” is the number of sensitive, intermediate, or resistant phenotypes expressed by each isolate and “N” is the total number of bacterial isolates.

475 clinical specimens in the present study exhibited bacterial growth; the majority (34%) of the 475 isolates originated from pus, followed by blood (24%), urine (20%), and sputum (13%), but other isolation sources were common for some pathogens (Figure 1A). The results of the chi-square test revealed that bacterial abundance in specimens was significantly different (p value <0.05 Pearson chi square test). The relative abundance of Staphylococcus spp. (148; 31%) were isolated most commonly from the collected specimen samples, i.e., 95 of the 148 isolates were observed to be prevalent in blood. Klebsiella pneumoniae (135; 28%) was the second most common isolate; that was most prevalent in pus and sputum specimens, 67 and 34, respectively. Escherichia coli (74; 15%) was prevalent in urine specimens 35. Pseudomonas aeruginosa (65; 14%) pus specimens 51 were most prevalent, followed by Enterobacter spp. (28; 6%), equally distributed in pus and sputum, and the majority of Acinetobacter spp. (25; 5%) from pus (Figure 1B). The most commonly tested (frequency > 0.9) antibiotics against all the studied bacterial pathogens were aminoglycosides, carbapenems, cephalosporins, and fluoroquinolones. Other classes of antibiotics were often tested for diverse infections; antifolates and tetracyclines were routinely tested against Staphylococcus spp., K. pneumoniae, E. coli, Enterobacter spp., Acinetobacter spp. except P. aeruginosa. Monobactam and lipopeptide-class antibiotics were tested only against P. aeruginosa, whereas macrolide, anisomycin, lincosamide, glycopeptide, etc. were tested against Staphylococcus spp. (Figure 1F). Patients were classified into three age groups, viz., 0–15 years (19%), 16–35 years (10%), >35 years (23%), and unknown age (49%) (Figure 1C; Table 1).

The penicillin resistance index (R_I) of Staphylococcus spp. was proportionately 1 (100%)and against ciprofloxacin, cefoxitin, and erythromycin was >0.8. Chloramphenicol was the most effective antibiotic against Staphylococcus spp. (Supplementary Figure S1A). Pseudomonas aeruginosa exhibited a 0.6 R_I to imipenem, ceftazidime, cefepime, and piperacillin-tazobactam. Imipenem-EDTA, and amikacin were most effective against P. aeruginosa (Supplementary Figure S1B). Acinetobacter spp. exhibited R_I > 0.8 against ciprofloxacin, gentamycin, ceftazidime, cefotaxime, piperacillin-tazobactam, ampicillin-sulbactam, and trimethoprim-sulfamethoxazole; however, meropenem resistance was <0.4 (Supplementary Figure S1C). Klebsiella pneumoniae exhibited high resistance against all antibiotics tested except tetracyclines (Supplementary Figure S1D). Enterobacter spp. exhibited a R_I of 0.5 to tetracyclines and amikacin but were more resistant (>0.6) to the other antibiotics (Supplementary Figure S1E). Escherichia coli was more sensitive to amikacin (<0.3), and highly resistant to other antibiotics (>0.7) (Supplementary Figure S1F). Similar report by Gupta et al. (2014) stated that uropathogenic E. coli, Enterococcus faecalis, K. pneumoniae, Staphylococcus aureus, P. aeruginosa, and Proteus mirabilis revealed ampicillin resistance up to 94–100%. Uropathogenic E. coli strains prevalent in Rajasthan, India, exhibited 95% resistance to nalidixic acid and 80% resistance to ampicillin and amoxiclav antibiotics (Sood and Gupta, 2012).

Klebsiella pneumoniae spp. isolated from patients of medical and surgical wards exhibited a strong antibiotic-resistance correlation 0.85 (Table 2). Similarly, K. pneumoniae strains obtained from patients in medical ward and intensive care unit (ICU) exhibited antibiotic-resistance correlation 0.79. Escherichia coli isolates showed strong antibiotic-resistance correlation 0.96, 0.87, and 0.82 between the out-patient department (OPD) and medical ward, OPD and surgical ward, and surgical-medical wards, respectively. Pseudomonas aeruginosa isolated from skin-burns, skin-surgical, and burns-surgical wards, showed substantial antibiotic-resistance correlation of 0.97, 0.93, and 0.91, respectively. Staphylococcus spp. exhibited a strong antibiotic-resistance correlation 0.98 between medical and surgical wards, but a moderate antibiotic-resistance correlation 0.59 between Emergency and tuberculosis chest diseases (TBCD) wards. These findings strongly suggest that the antibiotic resistance patterns of the above-mentioned isolates were similar with respect to isolation sources. However, differences were also observed in the antibiotic resistance patterns of bacterial pathogens isolated from the patients of other wards; e.g., K. pneumoniae isolates revealed a weak antibiotic-resistance correlation 0.29 between burns and ICU wards. Similarly, E. coli exhibited an antibiotic-resistance correlation 0.30 between the ICU and medical ward and P. aeruginosa isolates showed antibiotic-resistance correlation 0.14 and 0.22 between the TBCD-ear nose throat (ENT) and burns-ENT wards, respectively (Kelch and Lee, 1978). There is ample published literature describing the antibiotic sensitivity phenotype of non-clinical P. aeruginosa, E. coli, Enterobacter spp., K. pneumoniae, and Staphylococcus spp. (Supplementary Table S2).

Klebsiella pneumoniae (102) and Staphylococcus (87) isolates were resistant to 6–10 tested antibiotics; E. coli (43) strains were resistant to >12 antibiotics; and P. aeruginosa (36) isolates were resistant to 4–6 antibiotics (Figure 2A). Enterobacter spp. and E. coli show 12 and 13 mean values for the number of resistant antibiotics, respectively. Whereas P. aeruginosa and Staphylococcus spp. only have four and six mean values for the number of antibiotic resistant. The mean values of number of resistant antibiotics (9) were not significant for K. pneumoniae and Acinetobacter spp. (p = 0.80). The comparison of Enterobacter spp. with other bacteria was significant (p < 0.05), except for E. coli and Acinetobacter spp. (p = 0.167) (Figure 2B). In the current study, we classified pathogens into three selection groups viz. directional selection, disruptive selection, and stabilizing selection. Klebsiella pneumoniae isolates were observed to undergo directional selection toward resistant phenotypes in gaining resistance to antibiotics CIP, LVX, FEP, TZP, and CXM; while it showed evolutionary disruptive selection toward antibiotics SXT, CAZ, CTX, GM, AN, TE, SAM, and CAC; and appeared to adopt stabilizing selection to antibiotic MEM in (Supplementary Figure S2A). Escherichia coli showed directional selection toward resistant phenotypes to antibiotics CIP, LVX, CAZ, FEP, TZP, CTX, and CXM, disruptive selection to antibiotics SXT, GM, TE, SAM, MEM, and CAC, and stabilizing selection to antibiotic AN (Supplementary Figure S2B). Pseudomonas aeruginosa isolates exhibited directional selection toward sensitive phenotypes to antibiotics ATM and IE, while toward resistant phenotype to antibiotic CIP, and disruptive selection to antibiotics GM, AN, CAZ, FEP, TZP, and IPM (Supplementary Figure S2C). Staphylococcus spp. were observed to adopt directional selection toward resistant phenotype to antibiotics CIP and E, and disruptive selection to antibiotics SXT, GM, TE, CM, C, and RIF (Supplementary Figure S2D). The selection patterns for Enterobacter and Acinetobacter spp. isolates are drawn in (Supplementary Figures S2E,F).

Non-clinical P. aeruginosa isolated from caterpillar carcasses was more sensitive compared to clinical isolates against fluoroquinolones LVX, CIP, and NX, third generation cephalosporins cefpodoxime (CPD), ceftriaxone (CRO), and CTX, amikacin and gentamicin. The zone of inhibition of non-clinical P. aeruginosa was 1.1–1.5 times bigger than those of clinical P. aeruginosa against CIP and LVX, indicating higher resistance of clinical than non-clinical P. aeruginosa (Figures 3C,D). Similarly, clinical K. pneumoniae isolates tended to be resistant to LVX and CIP and second and third generation cephalosporins, while most being resistant also to fourth generation cephalosporins. Non-clinical K. pneumoniae strains were sensitive to CIP, LVX, and NX, third generation cephalosporins CPD, CTX, and CRO, with the exception of cefixime (CFM), and to AN and GM. More than 60% of clinical K. pneumoniae strains were resistant to the antibiotics AN and GM (Figures 3A,B).