Epidemiology and resistant profile of bacterial pathogens in a tertiary health care hospital, Medan City: a retrospective study

Study design and ethics statement

This is a retrospective cross-sectional study conducted on patients’ medical records from a clinical microbiology laboratory in a private hospital in Medan, North Sumatra Province, Indonesia. Among the inclusion criteria for the data collection were: (1) all patients from all age groups who were admitted to the hospital from January 2022 to December 2023; (2) patients suspected of having an infection and having the microbiological analysis results of clinical specimens taken from the suspected focus of infection (blood, wound, sputum, and/or urine); and (3) patients admitted to the emergency department, wards (inpatient), and the hospital clinics (outpatient).

Clinical specimens other than blood, wound, sputum, and urine were excluded. Repeated cultures with microbial growth (positive) results from one patient for less than a 48-h interval were also excluded. This study was approved by the Ethics Committee of the Universitas Sumatera Utara in agreement with the Nuremberg Code and Declaration of Helsinki, as registered in letter number 1000/KEPK/USU/2023. This study does not require informed consent as it used anonymous patients’ medical records. The hospital’s ethics committee approved the consent waiver.

Collection of clinical specimens

Specimen collection and transport were processed according to the hospital protocol by following the literature (Church, 2016). At least 10 mL of blood was drawn from each adult patient. Blood sample volumes from neonates and pediatric patients complied with the body weight rule (Church, 2016). Pus, exudates, and swabs from wounds were categorized as wound specimens. Urine specimens were mostly collected by clean-voided midstream or urinary catheter methods. Urine specimens collected through suprapubic needle aspiration directly into the bladder were also accepted. Gram staining was used to determine the number of squamous epithelial cells (SEC) per low-power field (LPF) in wound (especially swab), sputum, and urine specimens that were eligible for culture. Sputum specimens containing 25 polymorphonuclear neutrophils and <10 SEC per LPF were accepted for culture.

Microbial isolation, identification, and antimicrobial susceptibility tests

Specimen culture for the isolations of bacteria and yeast was done according to the protocol of processing, isolation, detection, and interpretation of aerobic bacteriology cultures (Church, 2016). Blood specimens were cultured using the BACTEC™ automated system (Becton Dickinson, United States) and incubated at 35 °C for up to 3–5 days. Gram staining was performed on a positive blood culture, which was then sub-cultured into a solid medium. Skin microbiota (including coagulase-negative staphylococci (CoNS) and viridans streptococci) recovered from the culture were not reported as contaminants because we only collected one blood specimen (10 mL) per patient. Instead, we advised clinicians to match culture results to the patient’s infectious status before initiating antibiotic therapy. However, spore-producing Gram-positive rods grown in aerobic sheep blood agar (SBA) plates were most likely not reported.

Gram staining results determined the solid media used for wound and sputum culture. SBA plates were inoculated with 1 µL of clean midstream and urinary catheter specimens using a calibrated loop for colony count, then incubated in ambient air overnight at 35–37 °C. Gram-positive cocci were subcultured on SBA and mannitol salt agar plates. Gram-negative rods were subcultured on SBA, chocolate, MacConkey (MAC), and Eosin-Methylene Blue (EMB) agar plates (Oxoid™, Waltham, MA, UK). The cultures were incubated aerobically at 35–37 °C for 24–48 h. To recover yeast, specimens were inoculated on Sabouraud dextrose agar plate (Oxoid, Basingstoke, Hampshire, UK) and aerobically incubated at 37 °C for 48 h. We did not analyze sputum specimens for Mycobacterium tuberculosis because they were identified separately in an external reference laboratory (Adam Malik Hospital Clinical Microbiology Sub-installation laboratory).

Isolates of pathogenic bacteria and yeast grown in the culture media were subjects for species identification using the Vitek^® 2 Compact commercial system with Gram-negative (GN), Gram-positive (GP), and yeast cards (bioMérieux, Marcy-l’Étoile, France). Antimicrobial susceptibility tests (ASTs) for bacterial isolates were done using the Vitek^® 2 automated susceptibility system (bioMérieux, Marcy-l’Étoile, France) and Kirby-Bauer methods using antimicrobial susceptibility discs (Oxoid™, Waltham, MA, UK). Disk diffusion susceptibility testing for aerobic bacteria was performed by following the Clinical & Laboratory Standards Institute (CLSI) guidelines (CLSI, 2023).

Data analysis

The data were analyzed by descriptive statistics using Prism 10 (GraphPad Software, LLC., San Diego, CA, USA) and then presented in graphs and tables. Analysis of variance (ANOVA) was performed at a 95% confidence interval and p ≤ 0.05 using Prism 10. The antimicrobial susceptibility test (AST) results were shown only from the most isolated organisms, and the percentage of resistant organisms was reported in a table.

Distribution of patients and clinical specimens

A total of 2,501 patients’ data were collected from January 2022 to December 2023, consisting of 1,292 (51.66%) males and 1,209 (48.34%) females. The distribution of patients based on age groups is presented in Table 1. Children were categorized in the age group below 18 (<18) as mentioned elsewhere (Ogbonna et al., 2019). The number of patients hospitalized in wards/inpatient [1,726 (69.01%)] outnumbered those admitted in the emergency department [522 (20.87%)] and the hospital clinics/outpatient [253 (10.12%)]. However, there are no significant differences in the number of patients between the age groups (one-way ANOVA, p = 0.175) and between the hospital admissions (one-way ANOVA, p = 0.071). During the 2 years, we collected 714 (28.54%) blood, 655 (26.18%) wound, 545 (21.79%) sputum, and 587 (23.49%) urine specimens. A total of 1,190 (47.58%) microbial isolates were recovered, consisting of 98 (13.73%), 454 (69.32%), 394 (72.29%), and 244 (41.57%) from blood, wound, sputum, and urine, respectively (Fig. 1).

Distribution of microbial isolates

Following the microbial recovery from the aerobic culture, we identified bacteria and yeast, presented in Fig. 2. Gram-negatives (GNs) are the primary isolates [69/98 (70.41%)] recovered from blood specimens, pursued by the Gram-positives (GPs) [26/98 (26.53%)], and yeast Candida spp. [3/98 (3.06%)]. Escherichia coli was the most prevalent isolate identified from blood [27/98 (27.55%)] specimens. Previous studies reported E. coli as the most frequent GNs in patients with bloodstream infection, followed closely by Klebsiella pneumoniae (Sormani et al., 2023; McGuinness et al., 2024; Moon et al., 2024; Suganthini, Suvintheran & Nor Zanariah, 2024). E. coli from blood cultures indicates bacteremia, whether it is catheter-related (McGuinness et al., 2024) or the non-catheter related infections (Irigoyen-von-Sierakowski et al., 2024). Most patients with E. coli bacteremia in this study were from the age group of ≥ 55 years [17/27 (63%)] and thus associated with a higher risk of case fatality (MacKinnon et al., 2021). Coagulase-negative staphylococci (CoNS) [20/98 (20.41%)] were prevalent after E. coli in blood specimens, followed by Klebsiella spp. [16/98 (16.33%)], which was dominated by K. pneumoniae [15/16 (93.75%)]. Staphylococcus hominis [7/20 (35%)], S. epidermidis [3/20 (15%)], and S. haemolyticus [3/20 (15%)] were the three most common CoNS found in blood specimens.

S. aureus was highly predominant [103/454 (22.69%)] in wound specimens after E. coli [107/454 (23.57%)], which was in accordance with former studies (Rajput, Johri & Goyal, 2023; Sheeba, Prathyusha & Anila, 2024). Frequent identification of E. coli and S. aureus in this study could be a clue of fecal contamination in wounds and nosocomial infection risk as an indication for further investigations regarding the proper management of wound infections. Similar to findings in blood and wounds, E. coli was also the most prevalent isolate in urine specimens [107/244 (43.85%)], followed by K. pneumoniae [26/244 (10.65%)]. E. coli isolates were found in 83 (77.57%) females and 24 (22.43%) males, while K. pneumoniae was found in 17 (65.38%) females and nine (34.61%) males with significant bacteriuria. Our results are in line with other studies (Khan et al., 2023; Hossain et al., 2024; McGuinness et al., 2024) confirming the role of E. coli and K. pneumoniae among uropathogenic bacteria.

K. pneumoniae [68/394 (17.26%)] predominated the microbial isolates from sputum, followed by Acinetobacter spp. [57/394 (14.47%)]. The latter was dominated by Acinetobacter baumannii [52/57 (91.22%)]. K. pneumoniae is consistently identified as the leading isolate in sputum specimens (Goel et al., 2022; Algurairy, 2024; Rafia et al., 2024), yet it is also saprophytic in the airway. Thus, K. pneumoniae’s presence in sputum samples does not always indicate a true pathogenic role and requires further examination to diagnose a lower respiratory tract infection.

Besides monomicrobial, we found polymicrobial isolates from 4/98 (4.08%) blood, 41/454 (9.03%) wound, 52/394 (13.2%) sputum, and 9/244 (3.69%) urine specimens. Pairs of GNs dominated the polymicrobial isolates from blood [2/4 (50%)], yet there was no GP pair. GP and GN pairs were prevalent in polymicrobial wound specimens [18/41 (43.9%)], while bacteria and yeast pairs dominated the sputum’s polymicrobial [26/52 (50%)]. GP pairs did not exist in urine specimens, yet the other pairs were found equally (Fig. 3A). The most common species pairs in polymicrobial isolates were K. pneumoniae and S. aureus [6/106 (5.7%)].

The polymicrobial pairs observed in this study were dominated by non-multidrug-resistant (non-MDR) strains (Fig. 3B). Pairs of non-MDR and MDR strains were found in 1/4 (25%) blood, 9/41 (21.95%) wound, 17/52 (32.69%) sputum, and 3/9 (33.33%) urine. At a lower percentage, pairs of MDRs were found in 4/41 (9.76%) wounds, 1/52 (1.92%) sputum, and 1/9 (11.11%) urine specimens. We found the highest percentage of MDR polymicrobial isolates in urine specimens, lower than the estimated prevalence in clinical settings (Harris et al., 2023). The most common MDR polymicrobial isolate in urine was extended-spectrum beta-lactamase (ESBL)-producing E. coli [4/5 (80%)]. A high prevalence of MDR strains was also found in polymicrobial isolates of sputum specimens, which were dominated by the Gram-positive CoNS [8/19 (42.10%)] and the Gram-negative ESBL-producing K. pneumoniae [4/19 (21.05%)].

Antimicrobial resistance profile of the bacterial isolates

Among Gram-positives, CoNS isolates from sputum showed the lowest sensitivity against most antibiotics tested (Table 2), followed by Enterococci from urine and wound specimens. S. haemolyticus was the dominant CoNS [17/28 (60.71%)] and the most common MDR CoNS isolate from sputum [15/28 (53.57%)]. S. epidermidis was the other common MDR CoNS in sputum [10/28 (35.71%)], which was dominated by methicillin-resistant S. epidermidis (MRSE) [6/10 (60%)]. E. faecalis was the most prevalent isolate [10/15 (66.67%)] among the uropathogenic Enterococci.

Resistance to cefoxitin and oxacillin was the most common among Gram-positives, dominated by CoNS (Table 2). Resistance to clindamycin and ampicillin came after, and it was also dominated by CoNS. Enterococci resistance to quinupristin-dalfopristin was high, as the majority of the isolates were E. faecalis [9/18 (50%) and 10/15 (66.67%) enterococci from wound and urine samples, respectively], which are mostly resistant to this antibiotic. Streptococcus pyogenes and Streptococcus pneumoniae, on the other hand, showed zero to minimal resistance to the antibiotics tested.

A. baumannii in wound specimens were the highly resistant isolates among the Gram-negatives (Table 3). Enterobacter cloacae was the sole species from Enterobacter found in urine [7/7 (100%)] and also showed a high level of resistance towards the majority of antimicrobials tested, followed by Pseudomonas aeruginosa isolates in wound and sputum specimens. Overall, the lowest sensitivity of Gram-negative isolates was shown towards ampicillin, followed by cefazolin, and ampicillin-sulbactam. K. pneumoniae isolates from urine and blood demonstrated increasing resistance to ceftazidime, a third-generation cephalosporin, most likely due to the beta-lactamase production such as ESBL and carbapenemase. Likewise, increasing resistance of E. coli isolates to ceftazidime indicated a significant proportion of the ESBL-producing strains.

ESBL-producing E. coli was the most prevalent MDR found in patients with bacteremia, with equal distribution [5/49 (10.2%)] per year in 2022 and 2023 (Fig. 4). In 2022, ESBL-producing K. pneumoniae was less identified [1/49 (2.04%)] than methicillin-resistant S. aureus (MRSA) [2/49 (4.08%)]. However, in 2023, there was a fourfold increase in ESBL K. pneumoniae found in blood specimens [4/49 (8.16%)], surpassing MRSA [2/49 (4.08%)]. MRSA and ESBL E. coli both dominated the MDR isolates of wound specimens [9/206 (4.37%) each] in 2022, but MRSA increased [24/248 (9.68%)] more than ESBL E. coli [20/248 (8.06%)] in 2023. ESBL K. pneumoniae was the most common MDR isolate in sputum in 2022 [4/161 (2.48%)] and 2023 [12/233 (5.15%)]. MRSA was detected in sputum [7/233 (3.0%)] only in 2023. ESBL E. coli dominated the MDR isolates of uropathogens in 2022 [11/119 (9.24%)], and increased twofold in 2023 [22/125 (17.60%)].