Antimicrobial resistance of Klebsiella pneumoniae strains isolated from urine in hospital patients and outpatients

Conclusion: Strains of K. pneumoniae isolated from the urine of hospital patients and outpatients have developed signifi cant resistance against all tested antibiotic substances. A higher occurrence of ESBL strains was observed in hospital patients than in outpatients. ESBL strains were resistant to all penicillins and almost all cephalosporins. Highly eff ective antimicrobials were amikacin, colistine, carbapenem and fosfomycin. The best therapeutic results were achieved when patients were treated with fosfomycin and imipenem.


Introduction
pneumoniae is a species that often causes infections in humans. As a primary pathogen, it most commonly causes pneumonia and urinary tract infection [1]. The bacterium lives as a commensal mainly in the digestive system of humans and animals, but is also widespread in the environment. Infections occur most frequently in hospitalised or immunocompromised patients, and are routinely treated with ß lactam antibiotics and other antibiotics against enterobacteria [2]. Due to its rapid spread, Klebsiella infection is the third most common hospital infection in the United States (9.9%), after Clostridium dif icile and Staphylococcus aureus. K. pneumoniae causes serious infections including pneumonia, urinary tract and circulatory infections [3]. Outbreaks of Klebsiella infections are often caused by newly emerged resistant strains with expanded spectrum beta lactamase production (ESBL) [4], and their frequency has been increasing for years. The increasing trend of the occurrence of Klebsiella infection is particularly pronounced in Eastern Europe [5]. The environment is an important source and reservoir of K. pneumonia, and it is frequently isolated from water, sewage, soil and plant surfaces. In a study of surface water samples, 123 Klebsiella strains were isolated from 53% of samples with K. pneumoniae as the dominant species [6]. Recent studies have also shown that highly pathogenic K. pneumonia strains, containing resistance and virulence genes similar to those detected in human strains, are often found in pets such as dogs and cats. High similarity has been demonstrated between strains isolated from pets and humans with urinary infection and highly pathogenic clones of K. pneumoniae (> 86.7%). Hence, owners and pet handlers must be cautious to prevent the spreading of infection [7].
One of the most common modes of transmission of K. pneumoniae in the hospital environment is direct contact between healthcare professionals and patients. Contaminated surfaces of instruments have also been detected as sources of infection [8]. Transmission of K. pneumoniae in the community is considered to be lower than in hospitals. The colonisation of the nasopharynx is higher in adults than in children (15% vs. 7%). The main means of transmission is poor food and water hygiene [9]. The gastrointestinal system is a signi icant reservoir of K. pneumoniae and source of infection, since it is often colonised [10]. Urinary tract infections are among the most common human infections and represent an important public health problem with high economic implications. Due to the widespread use of antibiotics to treat urinary infections, antibacterial resistance of Enterobacteriaceae, in particular the major uropathogens Escherichia coli and K. pneumoniae, has increased signi icantly worldwide. The increased prevalence of resistance of Enterobacteriaceae toward the most commonly used antibiotics limits the availability of treatment for infections caused by these bacteria [11]. Klebsiella producing ESBL are resistant to many antibiotics, including aminoglycosides, luoroquinolone, tetracyclines, chloramphenicol, trimethoprim + sulfametoxazole, penicillins and cephalosporins [12]. Infections caused by ESBL strains can only be treated by carbapenem, such as imipenem, meropenem, ertapenem and others [13], though carbapenems should be avoided whenever possible, as the last line of defence in case of infection with a highly resistant strain if no other effective antibiotic against infection is available. Such cases have been reported in India, Europe and across the world [14].
The main goal of this study was to determine the resistance pattern of K. pneumoniae isolates from hospital patients and outpatients, to determine the extent of resistance and sensitivity of K. pneumonia among isolates, and to determine associations of resistance patterns with patient gender and age. Finally, the aim was to express the proportion of resistant ESBL strains of K. pneumoniae in hospital patients and outpatients.

Materials and methods
The study was carried out from January to December 2018 at the Institute of Microbiology and Molecular Diagnosis of Mostar University Clinical Hospital (Bosnia and Herzegovina). Hospital patients were admitted to the Mostar University Clinical Hospital and samples were taken during their stay. Urine samples from outpatients were delivered to the Institute of Microbiology and Molecular Diagnosis, Mostar University Clinical Hospital based on a physician's instructions from primary health care.
A total of 430 K. pneumoniae strains isolated from urine samples were analysed. During the study period, 153 strains were isolated from hospital patients and 277 from outpatients.

Isolation and identifi cation of isolated bacteria
To sample urine, 10 mL of the middle stream was taken. In exceptional circumstances urine was taken by catheter in hospitalised patients. Urine samples were inoculated on nonselective (5% blood agar) and selective McConkey agar (Oxoid, UK) and incubated for 24 hours in a thermostat at 37 °C. Gram negative bacteria were identi ied by standard microbiological methods.
Phenotypic and biochemical properties of grown colonies were evaluated in order to identify Klebsiella. The colonies were cultivated on McConkey agar and blood agar. The colonies on McConkey agar were up to 3 mm in diameter, mucous and lactose positive. Colonies on blood agar appeared as grey coloured, shiny colonies up to 3 mm in diameter. Commercially available substrates (Oxoid, UK) were used to avaluate biochemical properties of suspected Klebsiella colonies such as Kligler agar (containing two sugars), SIM medium (sulphide, indole, motility) and citrate agar. After 24 hours incubation at 37 °C Klebsiella ferments both glucose and lactose producing gas during the growth. Bacterium is non-motile, does not produce H 2 S, it is indole negative and ferments citrate. The inal identi ication was carried out using the automated closed system for the identi ication and testing of susceptibility VITEK 2 Compact (BioMérieux, France) according to the manufacturer's instructions [1,15].
After the identi ication, isolates were submitted to the susceptibility test toward antimicrobials using Kirby Bauer disc diffusion method. Five colonies of a freshly cultured bacterial strain (overnight culture) were picked up using inoculation loop and diluted in 2 ml of sterile saline. Suspension was vortexed in order to create a smooth suspension and turbidity was adjusted to a 0.5 according to McFarland. Suspension was inoculated on surface of Miller Hinton agar within 15 minutes after preparation with a sterile cotton swab dipped into inoculum tube. Excess luid was removed by rotating the swab against the side of the tube (above the luid level) using irm pressure. The surface of a MH agar plate was inoculated by streaking the swab three times over the entire agar surface;
Infection was more frequently con irmed in female (244% or 56.74%) than male patients (186% or 43.26%). Assuming that infection occurs with the same frequency in both sexes, the observed difference in the distribution by gender was statistically signi icant (χ 2 = 7.82; df = 1; p < 0.05).
The age of patients ranged from 15 to 95 years, with a mean age of 66.14 years (SD = 18.07). The mean age of male patients was 69.02 years (SD = 15.87) and female of 63.93 years (SD = 19.33), and the observed difference in patient age by gender was statistically signi icant (t = 2.992; df = 424.4; p = 0.003).
Analysis of the susceptibility pattern of K. pneumoniae isolated from hospital and non-hospital patients showed that strains isolated from hospital patients were more resistant to each of the tested antibiotic, with the exception of fosfomycin (Table 1).
Of the 430 analysed isolates of K. pneumoniae, ESBL production was detected in 169 (39.30%) analysed strains. Of the ESBL positive strains, 92 (60.13%) came from hospital patients (of the 153 analysed), and 77 (27.8%) originated from outpatients. Strains able to synthesise ESBL were detected in 111 (66.07%) males and 58 (34.53%) females. The observed difference of the distribution of ESBL strains between genders was statistically signi icant. Male patients were in more frequently infected with strains able to produce ESBL (χ 2 = 16.095; DF = 1; p < 0.05). Table 2 shows the sensitivity of ESBL producing strains toward different antibiotics in the two patient groups.

Discussion
Urinary tract infections are among the most common of all bacterial infections and are often a reason for visiting health institutions. Most of these infections are treated in primary health care facilities. However, resistance to antimicrobials used to treat urinary infections is on the rise [20], and the number of hospitalised patients increases with the increase of antimicrobial resistance. In recent decades, hospital treatments of urinary tract infections have increased dramatically, and resistance to standard antibiotics has become very common [21].
In this study, the resistance pattern of K. pneumoniae isolated from hospital patients and outpatients with urinary infection to certain antibiotics is presented. These data primarily provide insight into the possibility of successful treatment of these infections. Resistance to penicillin is very widespread, therefore penicillin cannot be considered rotating the plate approximately 60 degrees each time to ensure an even distribution of the inoculum. Inoculated plate was left several minutes to dry. Disks were applied irmly on the agar surface within 15 min of inoculation of the plates. Within 15 min of application of antimicrobial disks, the plates were inverted and incubated at 35 ± 1 °C for 16-20 h. After incubation, inhibition zones were read at the point where no obvious growth was detected by the unaided eye when the plate is held about 30 cm from the eye. The inhibition zone diameters are measured to the nearest millimetre with a ruler. The procedure is standardized and the results are interpreted according to EUCAST [16]. We tested bacterial isolates for the following substances: amikacin (AKN) 30

Testing of bacterial strains K. pneumoniae for ESBL production
The production of ESLB was determined by the double disc synergy test (DDST) method. Screening and con irmatory tests for detection of ESBL strains were performed according to valid standards (Clinical and Laboratory Standard Institute) and EUCAST [17][18][19]. Discs containing ceftazimidin (CAZ 30), amoxicillin + clavulanic acid (AMC10/10) and ceftriaxone (CRO30) were placed on the surface of Mueller Hinton's agar inoculated with bacterial suspension of proper density (0.5 according to McFarland). The distance between the centres of discs was 20 mm. Plates were incubated at 37 °C for 18-24 hours. After incubation, synergism between discs was evaluated. The smallest sign of synergism between discs was considered a positive test result, i.e., the strain had the ability to produce expanded spectrum ß lactamase (ESBL).

Statistical analysis
The IBS SPSS Statistics for Mac OSX statistical package (version 23, SPSS Inc, Chicago, Illinois, USA) was used for data analysis.
The Chi-square test (χ 2 ) was used to compare the differences of infection occurrence between hospital patients and outpatients by gender. The Chi-square (χ 2 ) and Fisher exact tests were used to compare the differences of bacterial susceptibility against antimicrobials isolated from hospital patients and outpatients. Student's t -test was used to compare age differences between hospital patients and outpatients. Probability less than 0.05 was considered statistically signi icant. https://doi.org/10.29328/journal.abb.1001021 a reliable antibiotic in combatting infections caused by K. pneumoniae. Acceptable results were achieved using cephalosporins, luoroquinolones or gentamicin, a satisfactory therapeutic effect was achieved with nitrofurantoin and trimethoprim + sulfomethoxazole. The best therapeutic success was expected with carbapenems and colistine as the last line defence antibiotics, and with amikacin and fosfomycin. Fosfomycin was particularly successful, since only 2 of 142 isolates expressed resistance toward it, and its ef icacy was equal to imipenem and better than ertapenem and meropenem. A study conducted in the USA from 1998 to 2010 showed a constant trend of increased resistance of K. pneumoniae against antibiotics. Resistance against sulfametoxazole + trimethoprim reached 10.9% in 1998 and increased to 19.3% in 2010, while in the same period, resistance to cipro loxacin increased from 5.5% to 16.8%. Our study showed 40.5% resistant K. pneumonia isolates against sulfametoxazole + trimethoprim from hospital patients and 26.4% from outpatients, while resistance to cipro loxacin was 53.7% in isolates from hospitalised patients and 34.3% from outpatients. The resistance in the broader region is even more pronounced due to the inappropriate and uncontrolled use of antibiotics. Resistance against amikacin increased from 0.7% to 4.5% in the US study, while in this study it reached 4.6%. Resistance toward imipenem was not detected in the US at the beginning of the study (1998) but eventually reached 4.3% [22]. In this study, the resistance of isolates from hospital patients to imipenem was 2.8%. These indings frequency of indings in some earlier surveys was 34% in Portugal, 37% in Italy, 44% in the USA, 30% -60% in Latin American countries and 58% in Turkey [28]. The choice of antibiotics is particularly important in seriously affected patients with infections caused by ESBL producing K. pneumoniae. This study has con irmed that penicillins and cephalosporins are unsuccessful in the treatment of infection caused by ESBL strains and that luoroquinolones can cure less than 30% of patients. Nitrofurantoin is effective in approximately 50% of patients. Good results, with a success rate of over 95%, were observed in treatment with carbapenem, colistin, amicacin and fosfomycin. Administration of cabazenem, primarily imipenem, prevented mortality from infection with ESBL strains compared with the use of other antibiotics [29]. The therapeutic ef icacy of carbapenem has been con irmed in India, the United States, Israel, Greece, and in many European countries, including Croatia. Almost all alternative antibiotics are inef icient against infections with ESBL producing K. pneumonia. Resistance is observed even against antimicrobials still under development. Hence, there is a growing awareness about the possible inability to cure life-threatening infections caused by carbonate-insensitive strains in the future [30,31]. The highest frequency of ESBL strains in relation to patient age was determined in the age group above 65 years. Regarding gender, the affection rate of the pathogen and the body site from which the isolates originated, our indings are in accordance with similar studies, con irming females as the more frequently infected gender and the urinary tract as a commonly affected body system [32,33].
Based on these results, it can be concluded that the con irm that new strains have appeared. These strains are also resistant against carbapenem (CRKPA) and could pose a serious problem in the future. In 2000, less than 1% of K. pneumoniae isolates were resistant to carbapenem, while in 2007 this grew to 8% [22,23]. A study in Bangladesh showed a 100% resistance rate of K. pneumonia against ampicillin, 90% against amoxicillin, 45% against ceftriaxone, 40% against cipro loxacin, 25% against gentamicin and 35% against tetracycline. The indings of that study are much closer to the results presented here, and far from desirable values [24].
Strains secreting ß-lactamase expanded spectrum (ESBL) appeared for the irst time in Germany in the 1980s, and have since spread worldwide, posing a major problem. ESBL production was detected in gram-negative rod-shaped bacteria, with the majority detected in the Enterobacteriaceae family [25]. Prevalence of infection caused by ESBL strains is most common among K. pneumoniae isolates in the hospital environment, reaching about 25%, although in some studies this varied from 30% to 80% [25,26]. In this study, ESBL strains of K. pneumoniae were detected in 169 (39.3%) patients, with 92 isolates (60.13%) isolated from hospital patients and 77 isolates (27.8%) originated from outpatients. According to the literature, the rate of strains able to produce ESBL in Europe varies from less than 5% in Scandinavian countries and the Netherlands, to more than 50% in Greece and Poland. The ability of K. pneumoniae strains isolated in Croatia to produce ESBL depends on the origin and reaches about 56% among invasive isolates and 34% among clinical isolates, with signi icant variations among health institutions [27]. The frequency of ESBL producing bacteria may vary within a geographical area or even within a country. The frequency of infection with isolated K. pneumoniae species was determined in 35.6% of hospital patients and in 64.4% of outpatients. The infection occurs more frequently in female patients and in patients over 65 years of age. K. pneumoniae developed signi icant resistance against each tested antibiotic, except fosfomycin, carbapenem, amikacin and choline. The ef icacy of cephalosporins, luoroquinolone, aminoglycosides and trimethoprima + sulfametaxel was satisfactory, and the most effective medicines to be kept for dif icult situations in the treatment of infections are carbapenems, amicacin, colistine and fosfomycin. There was a high occurrence of ESBL strains (39.35 cases), which was more common in hospitalised (60.13%) than in non-hospitalised patients (27.8%). ESBL strains are resistant against all penicillins and almost all cephalosporins, even against third-generation cephalosporin and in less than 2% of cases lead to a cure. In less than half of cases, infections can be cured using luoroquinolones, gentamicin and trimethoprim + sulfometoxazole. Nitrofurantoin was effective in approximately 50% of cases. Amicacin, colistine, carbapenem and fosfomycin were highly effective, while the best results were achieved when fosfomycin and imipenem were used to treat infection.