Antibiotic resistance and genotyping of gram-negative bacteria causing hospital-acquired infection in patients referred to Children’s Medical Center

Introduction

Antimicrobial resistance among gram-negative bacteria (GNB) is growing worldwide.¹ It is a main public health problem causing both significant morbidity and mortality among hospitalized patients. A direct correlation between resistance of GNB and patient mortality, cost of patient care, and length of hospital stay has been shown.^2,3 Lots of gram-negative organisms are responsible for hospital-acquired infections (HAIs), among which the Enterobacteriaceae family is the most commonly recognized group.

Multidrug-resistant organisms such as Pseudomonas aeruginosa, Acinetobacter baumannii, and ESBL-producing or carbapenemase-producing Enterobacteriaceae, are increasingly being reported worldwide.⁴

It is essential to understand pathogen distribution and relatedness in order to determine the epidemiology of HAIs.⁵ There have been various technical approaches to bacterial source tracking,^6,7 but a consensual single method for field application has not been recognized until now.⁸ Genomic fingerprinting methods are now regarded as the most accurate methods for the typing of microorganisms.⁹ These methods comprise pulsed field gel electrophoresis, ribotyping, and PCR-based fingerprinting methods.¹⁰ The aim of our study was to evaluate antibiotic resistance of GNB and genotyping of common GNB causing HAI in patients referred to Children’s Medical Center during 6 months by random amplified polymorphic DNA (RAPD) PCR.

Material and methods

The current study was carried out in the referral hospital of Children’s Medical Center, Tehran, Iran between July 2017 and January 2018. The whole GNB strains isolated from clinical specimens of blood, wounds, and sterile fluids of patients hospitalized in this center were used in study. Ethical approval was obtained from the Ethical Committee of Tehran University of Medical Sciences, Tehran, Iran. GNB isolates were not specifically isolated for this research and were part of the routine hospital laboratory procedure. The BACTEC 9120 Blood Culture System (BD, Franklin Lakes, NJ, USA) was used for quick detection of microorganisms in blood samples. Gram staining and subculture of the samples were performed on MacConkey agar, chocolate agar, and blood agar plates. Identification of microorganisms was carried out using conventional biochemical methods. The tests used were a Kligler iron agar slant, catalase and oxidase tests, sugar fermentation, Simmons’ citrate agar slant, urea hydrolysis slant, methyl red/Voges–Proskauer test, and motility test.

Antimicrobial resistance of each organism was assessed using standard techniques according to Clinical and Laboratory Standards Institute criteria.¹¹ The evaluated antibiotics include piperacillin-tazobactam, imipenem, meropenem, gentamicin, amikacin, cefepime, ceftazidime, colistin, ciprofloxacin, cefotaxime, and trimethoprim-sulfamethoxazole.

In order to accomplish the genotyping of the common hospital-acquired GNB (including A. baumannii, Klebsiella pneumoniae, P. aeruginosa, Escherichia coli, and Enterobacter spp.), bacterial genomic DNA extraction was performed using DNA extraction kit (Bioneer, Daejeon, South Korea) according to manufacturer’s guidelines. RAPD-PCR was done as described previously.⁵ Briefly, DNA amplification was accomplished on a thermo cycler (Bio-Rad Laboratories Inc., Hercules, CA, USA) in a final volume of 25 mL containing 1× PCR buffer, 50 mM MgCl2, 10 mM dNTP mix, 50 mM primer 272 (3‘-AGCGGGCCAA-5‘),12 1 unit Taq polymerase (Qiagen NV, Venlo, the Netherlands), doubledistilled water, and genomic DNA equivalent to 40 ng.⁵

The RAPD-PCR was carried out under the following conditions: initial four cycles of denaturation at 94°C for 5 minutes, annealing at 36°C for 5 minutes, extension at 72°C for 5 minutes followed by 30 more cycles of denaturation at 94°C for 1 minute, annealing at 36°C for 1 minute, and extension at 72°C for 1 minute. The RAPD-PCR products were loaded on a 1% (w/vol) agarose gel containing 1/10,000 gel red (Biotium, Fremont, CA, USA), and were analyzed by gel electrophoresis and banding patterns were visualized and photographed in Geldocumentation system (Uvitec, Cambridge, UK). A 100 bp DNA ladder (Thermo Fisher Scientific, Waltham, MA, USA) was used as a molecular size standard. SPSS software (version 13; SPSS Inc., Chicago, IL, USA) was used for data analysis and RAPD-PCR fingerprinting profile was analyzed with Gelcompar II software, version 6.5 (Applied Maths, Sint-Martens-Latem, Belgium). Similarity analysis of results was calculated using the Dice coefficient/unweighted pair-group method with arithmetic mean. The criteria for related clones were taken as profiles with 80% or more similar bands. Isolates with 100% similarity were considered as the same RAPD-PCR type.

Results

In the current study, 142 (2.2%) GNB strains were isolated from 6,524 blood, wound, and sterile body fluid cultures. The GNB were isolated from 84 boys (59.2%) and 58 girls (40.8). The median age of the patients was 11.4 months (IQR: 1.9 months-4 years) (Table 1). Ninety-one patients (64%) had a history of previous hospitalization and 13 patients (9%) had died by the end of the study.

Table 1 Characteristics of studied patients

As shown in Table 1, the most isolates were collected from surgical unit and intensive care units (ICUs) (particularly EICU and Neonatal Intensive Care Unit). Forty-four patients (31%) utilized ventilators, 80 patients (56%) used catheters, and 12 of them (8.5%) had positive central venous catheter tip culture during their hospitalization.

The most common isolated GNB was E. coli (n=32, 22/5%), followed by K. pneumoniae (n=29, 20%) and Pseudomonas spp. (n=17, 12%) (Table 2). A total of 142 organisms were isolated from blood (63%), wounds (13.7%), ascites (8.7%), cerebrospinal fluid (6%), central venous catheter tip (5%), dialysis fluid (1%), pleural fluid (0.7%), and the other sterile cultures (1.4%). Fourteen strains were separated from peritoneal fluid in which E. coli with a number of nine strains (64%) was regarded as the most common.

Table 2 Frequency of occurrence of GNB isolated from blood, wounds, and sterile body fluid cultures

Among all the gram-negative organisms, 100 isolates met the criteria of nosocomial infection. All of the S. marcescens, A. baumannii, Enterobacter spp., and P. aeruginosa isolates met the criteria of HAI.

Forty-one strains were from ICUs, most of which were S. marcescens (N=13, 22%), K. pneumoniae (N=12, 21.3%), E. coli (N=9, 15.2%), and A. baumannii (N=8, 13.5%). While the most prevalent isolates in the surgical unit were E. coli (N=11, 57.9%). HAI was reported in all the patients with positive culture of central venous catheter tip, 79 patients (99%) with catheters, and 43 patients (98%) who had used a ventilator. Interestingly, 99% of children who met the criteria of HAI had a history of previous hospitalization.

Antibiotic resistance

Antibiotic resistance frequencies of evaluated GNB were depicted in Table 3. High frequency of antibiotic resistance to cefotaxime (81%) and trimethoprim-sulfamethoxazole (75%) was seen in E. coli strains, whereas the most sensitive antibiotics were amikacin (84%) and gentamycin (81%). K. pneumoniae strains were highly resistant to cefotaxime (96.5%) and cefepime (86%) and sensitive to imipenem (65%), amikacin (64%), and gentamycin (57%). Likewise, in Klebsiella oxytoca, high resistance to the same antibiotics (75%) was seen. High sensitivity to gentamycin and trimethoprim-sulfamethoxazole was reported (75%) and all of them were sensitive to imipenem. Besides, the low sensitivity of A. baumannii to common antibiotics and high resistance of Enterobacter spp. to cefotaxime (100%) was reported (Table 3).

Table 3 In vitro antimicrobial resistance of evaluated gram-negative bacteria

Genotyping

The genotyping of A. baumannii causing HAI showed the presence of one clone in which two clusters with high similarity (95%) were present (Figure 1A). Among ten A. baumannii strains, five strains had 100% similarity and it might be due to the development of one strain in the Children’s Medical Center. The genotyping of E. coli and K. pneumoniae (Figure 1B and C) demonstrated the presence of one and two clones, respectively with ≥80% genetic affinity. However, the trend varied completely in P. aeruginosa strains and Enterobacter spp. (Figure 1D and E). Among 15 isolates of P. aeruginosa, just two isolates had 100% genetic similarity and four strains showed ≥80% similarity.

Figure 1 Genotyping of isolated Acinetobacter baumannii (A), Klebsiella pneumoniae (B), Escherichia coli (C), Pseudomonas aeruginosa (D), and Enterobacter spp. (E) Strains that met hospital-acquired infection criteria.

Discussion

We assessed the antimicrobial resistance of GNB recovered from blood, wounds, and sterile fluids from patients who were referred to the Children’s Medical Center and the genotyping of the isolated strains from the children who had HAI.

In the current study, the most frequent GNB were E. coli (n=32, 22/5%), K. pneumoniae (n=29, 20%), Pseudomonas spp. (n=17, 12%), S. marcescens (n=15, 11%), S. maltophilia (n=11, 8%), A. baumannii (n=10, 7%), and Enterobacter spp. (n=6, 4%). This report is similar to our previous study13 with the exception of the higher prevalence of K. pneumoniae compared to E. coli. In another study performed by Mahmoudi et al,¹⁴ K. pneumoniae (n=263, 27.5%) was reported as the most frequent GNB.

The most isolated bacteria were collected from bloodstream (63%), followed by wounds (13%). Primary bloodstream infections (28%), pneumonia (21%), and urinary tract infections (15%) were found to be most frequent in Richards et al’s study.¹⁵ It is obvious from the reported data that all the children with a history of hospitalization and almost all of the patients who used an invasive device met the criteria of nosocomial infection. K. pneumoniae was the most common organism isolated from blood culture (21%). However, this organism turned out to be Enterobacter spp. in Richards et al’s study¹⁵ and P. Aeruginosa in Murni et al’s study.¹⁶

The greatest proportion of isolated strains was from ICUs (particularly EICU and Neonatal Intensive Care Unit) (51%). Likewise, in Folgori et al’s study, the most GNB in Europe were isolated from Pediatric Intensive Care Unit/Cardiac Intensive Care Unit (22%).¹⁷

A. baumannii strains were significantly resistant to almost all tested antibiotics, which is similar to previous studies,^14,18 but different from Jafari et al’s study¹⁹ in which resistance rate to imipenem (41%), ciprofloxacin (78%), and meropenem (60%) was lower than ours. Whereas, in Ballot et al’s study,²⁰ resistance rates to ceftazidime, ciprofloxacin, and gentamycin (15%, 11%, and 4%) were lower than we reported. A. baumannii strains were sensitive to colistin and ampicillin-sulbactam. The resistance rate of P. aeruginosa in the current study, compared to the previous reports in the Children’s Medical Center,^13,14,21 was substantially lower and showed quite high sensitivity to the tested common antibiotics. The K. pneumoniae strains depicted high resistance rate to cefotaxime (96.5%), which was compatible with the results reported by Molana et al (90%).²² Various studies show that many resistant strains of bacteria that cause nosocomial infections are caused by an initial strain and expanded subsequently in the hospital environment between patients, or even transmitted by patients from different hospitals to each other.

Typing of bacteria is vital for monitoring newly emerging pathogens and for examining local outbreaks.²³ In this report, we used RAPD-PCR followed by agarose gel electrophoresis techniques for the molecular typing of common GNB collected during 6 months from the patients in the Children’s Medical Center who had HAI. The standardization of the reaction conditions has contributed to making RAPD typing a trustworthy and reproducible technique,²⁴ with even more discriminatory power than DNA macro-restriction analysis by PFGE for bacterial typing.²⁵ However, random amplification could simply be applied for the surveillance and prevention of nosocomial infections by clinical microbiology laboratories by enhancing the resolution of the electrophoretic separation and the sensitivity of the staining.²⁵ The dendrogram of genotyping of A. baumannii strains showed the presence of two clones with ≥80% genetic similarity. Besides, the low genetic similarity in Enterobacter spp. and P. aeruginosa could be plausible based on the high records of previous hospitalization of the patients. In previous studies carried out in the same center, the possibility of incidence of nosocomial infection was reported in presence of one bacterial clone.^18,26

In previous studies which were conducted in our hospital, the analysis of genotyping of A. baumannii, K. pneumoniae, and P. aeruginosa strains showed the presence of one clone in the Children’s Medical Center. Ninety-three percent of isolated Klebsiella spp. and all the A. baumannii strains that met the criteria of nosocomial infection were determined in one and three clusters, respectively with the high genetic similarity of ≥80%. In addition, 90% of P. aeruginosa strains were identified in one clone.^18,26

In Wasfi et al’s study,²⁷ RAPD-PCR showed 18 distinct patterns of K. pneumonia isolates with ≥80% similarity. In addition, the genotyping of E. coli, K. pneumonia, and P. aeruginosa demonstrated seven RAPD patterns for E. coli, five for each of K. pneumonia and P. aeruginosa isolates.²⁸ Likewise, according to Sachse et al’s study²⁹ in 2014, RAPD is probably functional as a rapid screening method for the intra-hospital surveillance of K. pneumoniae, allowing discrimination of clonally related strains. In another study, Asadollahi et al³⁰ reported three predominant genotypes and ten unrelated genotypes of A. baumannii.

Conclusion

The results of this study showed the presence of clones with ≥80% similarity in three common GNB, E. coli, A. baumannii, and K. pneumoniae, indicating occurrence of a possible outbreak in the Children’s Medical Center. Since there is a possibility of transmission of an infection from one patient to another, much attention should be paid to the basic methods of preventing infection (standard precautions), disinfection of medical equipment, and personnel hygiene. The length of hospitalization and the use of antibiotics in patients, especially during long-term hospitalization, could have impacts on the reduction of antibiotic resistance and the occurrence of HAI.