Antimicrobial Resistance and Biofilm Production in Uropathogens from Renal Disease Patients Admitted to Tribhuvan University Teaching Hospital, Nepal

Background . Various antibiotics are prescribed empirically by physicians to cope with infections in renal disease patients. A urinary tract infection (UTI) is often caused by bioflm-forming multidrug-resistant (MDR) uropathogens. Tis study aimed to analyze the antibiogram of UTI strains from renal disease patients and the bioflm-forming ability of those strains. Methods . 102 patients clinically diagnosed with a UTI and renal disease were recruited into the study from August 2017 to January 2018. Clean-catch midstream urine samples were processed for the isolation and identifcation of the bacteria following standard methodologies. Te antibiogram of the isolates ( n � 106) was produced by the Kirby–Bauer disc difusion method. Detection of bioflm formation was performed in tissue culture plates. Results . Te incidence of a UTI in renal disease was 19.1%. Most patients were diagnosed with chronic kidney disease (18.63%), nephrotic syndrome (16.67%), and nephrolithiasis (14.71%). Te commonest uropathogens were Escherichia coli (52.8%), Klebsiella pneumoniae (16%), and Enterococcus spp. (15.0%). Ceftriaxone was the most common antibiotic prescribed empirically (37%), whereas nitrofurantoin was the most prescribed antibiotic as adjusted therapy (36.1%). Among the frst-and second-line antibiotics, most Gram-negative bacteria were sensitive to amikacin (70.7%), meropenem (70.7%), cefoperazone-sulbactam (70.0%), piperacillin-tazobactam (67.2%), gentamicin (66.7%), and nitrofurantoin (66.7%). Most Gram-positive bacteria were sensitive to doxycycline (90.0%), nitrofurantoin (72.2%), gentamicin (66.7%), and tetracycline (62.5%). All MDR Gram-negative uropathogens were susceptible to colistin sulfate and polymyxin B. Among the 106 isolates, 74.5% produced bioflms and 70.8% were MDR. In 67.0% of cases, including both MDR and bioflm-producing bacteria, the empirical therapy needed adjustment. Conclusions . Aminoglycoside, carbapenem, beta-lactam combination agents, and nitrofuran group of antibiotics may be the optimal frst-line empirical therapies for uropathogens in hospitalized renal disease patients. Regular surveillance of resistance patterns and the study of bioflm formation in uropathogens must be performed to ensure efective management of the patients.


Introduction
Antimicrobials remain the mainstay of infectious disease treatment; however, the undiscerning use of antibiotics in many countries has resulted in the emergence of multidrug-resistant (MDR) microorganisms [1,2]. Antimicrobial resistance (AMR) is a serious public health threat recognized by the World Health Organization (WHO) [3]. Te spread of antimicrobial resistance and selection of MDR pathogens have most likely been caused by combinations of failure to adherence to proper infection control techniques, irrational use of antibiotics, increased use of antibiotics in animals and plants, availability of antibiotics without a prescription, and counterfeit products of dubious quality [4]. Antibiotic resistance leads to higher medical expenses, prolonged hospital stays, and increased mortality rates [5]. Tere is an urgent need for change in the way antibiotics are prescribed and consumed. Even if new antibiotics are developed, without behavior change, antibiotic resistance will continue to be a major threat [6]. In the meantime, information about the antibiotic spectrum of activity against existing MDR strains may help reduce the rate of emergence and spread of antimicrobial resistance [7].
An additional factor contributing to antibacterial resistance is bioflm production by bacteria [8]. Bioflm is an association of microorganisms in which microbial cells adhere to each other on living or nonliving surfaces within a self-produced matrix of extracellular polymeric substances [9,10]. Within bioflms, microbes are 10-100 times more resistant to antimicrobial agents [11]. Reasons for this increase in resistance can be due to the growth patterns of bacteria in bioflms, extracellular substances retarding antibiotic difusion, and the upregulation of certain genes in bacteria in bioflms [11][12][13][14]. Te National Institutes of Health (NIH) has estimated that approximately 65% of all microbial infections and 80% of all chronic infections are associated with bioflms [15,16]. Human diseases in which bioflms have been associated include urinary tract infections (UTIs), catheter infections, middle ear infections, contact lens-associated infections, and less common but more lethal infections such as endocarditis and cystic fbrosis [17][18][19].
UTIs are commonly encountered by clinicians with an estimated annual global incidence of at least 250 million [20]. Furthermore, the prevalence of UTI increases in patients with preexisting renal disease [21]. Tis study examined the antibiogram patterns of common uropathogens in renal disease patients and the ability of the strains to form bioflms. Tis study also aimed to provide evidence for the rational use of antibiotics in patients with renal disease having UTIs.

Materials and Methods
A descriptive, hospital-based, cross-sectional study was conducted among renal disease patients diagnosed with UTI and admitted to the 750-bedded Tribhuvan University Teaching Hospital (TUTH), Nepal, from August 2017 to January 2018. Te type of uropathogens and their antimicrobial resistance patterns along with bioflm-forming abilities were examined. Te "universal sampling" technique was used to determine the sample size for this study. We had taken urine samples of 102 patients, but of them, 4 of the samples had polymicrobes, i.e., 2 uropathogens from each of 4 urine samples. So we treated them like individual isolates and performed antimicrobial susceptibility testing and bioflm analysis. Tis explains why the sample size varied between patient size (102) and their sociodemographic analysis and uropathogens isolated (106) and their further analysis. Clean catch mid-stream urine specimens (n � 102) were collected aseptically from the patients. Specimen collection, culture, and identifcation were performed according to standard guidelines [22][23][24].
Te study conformed to the tenets of the Declaration of Helsinki. Ethical approval was obtained from the Institutional Review Board of the Institute of Medicine (IOM), Tribhuvan University (Ref. 256(6-11-6)2/074/075). Renal disease patients diagnosed with UTIs provided their approval by signing patient consent forms if ≥ 18 years or parents signing consent forms for younger subjects. Cases of any age, both males and females, were included in the study. Patients diagnosed with renal disease but with no signs of UTIs or who did not agree to sign the patient consent form were excluded from the study. Relevant clinical and epidemiological information was recorded from the patients. Patients with a history of recurrent UTI, kidney transplant patients, or patients under immuno-suppressing drugs were excluded from the study. Hemodialysis patients were also excluded from the study. We did take samples from catheterized patients too. A semistructured data sheet was pretested in the same study area, and pretesting bias was avoided. In the questionnaire, patients of age >16 years were considered legible for enquiring about marital status (children ≤16 years are not included in this variable) and occupation. Children and infants below primary education were also not included in the education level variable. Data entry, data checking, compiling, and editing were performed manually.

Antibiotic Susceptibility Testing (AST).
All the isolates were subjected to antibiotic susceptibility testing (AST) by the Kirby-Bauer disc difusion method on the Mueller-Hinton agar (HiMedia, India). Te frst-, second-, and third-line antibiotics (HiMedia, India) were chosen based on the Clinical and Laboratory Standards Institute (CLSI) 2017 guidelines [22] with some modifcations based on the antibiotic testing policy in the microbiology laboratory of TUTH, Nepal. Te results were reported as sensitive, intermediate, or resistant as described by CLSI [22]. Initially, all the isolates were subjected to frst-line AST of the appropriate antibiotics for Gram-positive or Gram-negative bacteria. If the uropathogen was found to be resistant to more than 3 diferent classes of antibiotics, then it was tested with the appropriate second-line antibiotics. If any Gramnegative uropathogen was found to be resistant to meropenem or if the isolate was susceptible to only one antibiotic among the battery of second-line antimicrobials, it was further subjected to third-line antibiotic sensitivity testing. If an isolate showed resistance to ≥1 antibiotic from at least 3 diferent structural classes, it was considered to be multidrug-resistant (MDR) [25,26].
Bioflm production: Te bioflm production assay was performed using the tissue culture plate method [27]. A wellisolated colony of the organism isolated from the clinical specimen was inoculated in 2 mL of brain heart infusion (BHI) broth (HiMedia, India). Te broth was incubated at 37°C for 24 h. Te cultures were then diluted 1 : 100 with fresh medium (BHI broth supplemented with 1% glucose) in the individual wells of the sterile well of fat bottom microtiter plates so that the fnal volume in each well was 200 μl. Te plates were incubated at 37°C for 24 h. Ten, the contents of each well were removed by gentle tapping. Te wells were washed with phosphate-bufered saline (pH 7.2) three times and then were stained with 0.1% safranin. After drying the wells, the adhered dye was then dissolved by 100% ethanol. Finally, absorbance of released safranin from each well was measured using OD 490nm . Quantifcation was performed according to the criteria described by Stepanovic et al., as shown in Table 1 [28].
Here, OD is the average optical density of each isolate and OD c is the cutof OD for the microtiter-plate test as three standard deviations above the average optical density of the negative control. Previously identifed bioflmproducing in-house clinical isolates were used as the positive control for bioflm production [27].

Data
Analysis. Data analysis was performed using the 17.0 version of Statistical Package for Social Sciences (SPSS) software. Major variables analyzed in this study were sociodemographic data, the result of AST, bioflm production, and days of hospital stay. A descriptive type of analysis was performed to generate frequency and percentage. Chi-square/cross-tabulation was performed to test signifcance.

Results
During the study period, a total of 10,404 urine samples were received by the microbiology laboratory. Among them, 534 samples were obtained from renal disease patients. A total of 102 renal disease patients who had a UTI were recruited to the study. Te prevalence of a UTI in renal disease was 19.1%. A total of 106 uropathogens were isolated from the UTI-confrmed urines. A UTI was most seen in renal disease patients of age group 55-64 years (17.6%) followed by 35-44 years (14.7%) and 15-24 years (13.7%) ( Table 2).

Patient Type and Diferent Wards to Which Patients Were
Admitted. Of the 102 patients in the study, the majority were inpatients (95, 93.1%). Most of them were admitted to the nephrology (34, 33.3%) or pediatric (16,15.7%) wards, whereas the lowest number of the patients was from neurology and surgery (4.9%). Te majority of patients were diagnosed with chronic kidney disease (18.6%), followed by nephrotic syndrome (16.7%), nephrolithiasis (14.7%), and acute kidney disease (AKD) (11.8%), as shown in Table 3.

Types of Uropathogens.
A total of 10 diferent bacterial species accounting for 106 diferent isolates from 102 urinary samples were isolated and identifed. Te majority were Gram-negative bacteria (87, 82.1%), and the remaining isolates were Gram-positive. Te most common bacteria were Escherichia coli (52.8%), followed by Klebsiella pneumoniae (16.0%) ( Table 4). Enterococci were the most common Gram-positive bacteria (Table 4).

Bioflm Production and MDR.
Among the 106 isolates, 70.8% were MDR, while 74.5% were bioflm producers ( Table 7). Te majority of MDR bacteria (54.3%) were found to produce bioflm but of diferent levels, and the relationship with multidrug resistance was statistically insignifcant (p value >0.05).

Duration of Hospital Stay with Respect to MDR and
Bioflm Production. Patients infected with MDR strains had a slightly higher duration of hospital stays (9.01 days; 95% CI (7.71, 10.32)) than patients with non-MDR (6.67 days; 95% CI (5.11, 8.22)). In the case of bioflm production, patients with strong bioflm-producing bacterial isolates had longer hospital stay (10 days) followed by patients with moderate bioflm producers (9 days). Patients having nonbioflm producers had the least days of hospital stay (Figures 1  and 2).

Te Change in Antibiotic Use after Sensitivity Assessment.
Te antibiotic use pattern for empirical therapy and adjusted treatment therapy was evaluated. Tird-generation cephalosporins (cefxime, ceftriaxone, and ceftazidime) were the most frequently prescribed empirical antibiotics (49.0%), followed by fuoroquinolones which were prescribed in 28% of patients. Once the results of the culture and sensitivity were obtained, the most commonly prescribed antibiotic was nitrofurantoin (36.1%) followed by fuoroquinolones (24.9%) ( Table 8).

Comparison of Empirical Terapy with MDR and Bioflm
Production. In 67% of patients, the antibiotic used as an empirical therapy did not match the culture sensitivity results and so had to be prescribed alternative antibiotics. Most of the patients in which the empirically used antibiotic was changed were infected by MDR strains (n � 53) and bioflmproducing bacteria (n � 50) (Table 9).

Discussion
Te prevalence of renal disease patients is increasing, and it has become a serious threat to the health sector of Nepal. Antimicrobial resistance (AMR) has further aggravated the prognosis of renal disease patients developing UTI. Our study highlights the alarming AMR threat to renal disease patients and why strict antibiotic stewardship practice should be enforced [29]. In this study, the prevalence of a UTI in renal disease was 19.1%. A previous study found a similar prevalence (17%) of a UTI during the frst 6 months after renal transplantation [30,31]. In Nepal, the prevalence of renal disease has been reported to approximately double in males compared to females with a ratio of 1.8 : 1 [32], although the study recruited approximately equal numbers of males and females. Te estimated glomerular fltration rate (eGFR) declines in parallel with age [33], and this coincides with increasing trends of CKD prevalence from 7.4% for 18-39 years to 24.2% for 60-70 years [34]. Tis pattern is also refected in the current study, with the highest number of renal diseases being seen in the age group of 55-64 years followed by 25-44 years. Te current study found that a UTI was commonly caused by Gram-negative bacteria, with E. coli being the most common bacteria followed by K. pneumoniae then Enterococcus species. Tis is a similar bacteriological profle that has been reported by many others in CKD patients as well as studies of community acquired UTI [35][36][37]. Antimicrobial resistance is a looming threat to humankind, though its rate varies from place to place, and a common trend shows its rate is increasing. Similar to that, in our study, the resistance rate is high with 70.8% of MDR bacteria which is higher than fndings from those in previously conducted studies. Baral found only 41.1% of MDR bacteria in their study in 2012 [38], but a more recent study conducted by Parajuli et al. [39] has reported 64.9% of MDR bacteria causing UTIs and in a more recent study by Shilpakar et al. [40] who have reported that more than 90% of Gram-negative bacteria were MDR. Tis highlights the difculties that are likely to be encountered with the treatment of UTI patients in Nepal.
Resistance to amoxycillin, ceftriaxone, cotrimoxazole, and ciprofoxacin among the frst-line antibiotics was ≈70% of all isolates [41]. Fortunately, no strains were resistant to the last-line drugs such as polymyxin B and colistin sulfate. Similar fndings were reported by the other studies [42]. However, with the known problems of renal toxicity during therapy with polymyxin B and colistin sulfate, careful optimization of the polymyxin dose and drug monitoring is needed [43,44]. Bioflm production was seen in 74.5% of isolates with the majority being E. coli (51.9%), followed by K. pneumoniae (20.2%) and P. aeruginosa (10.1%). Tese results agree with, although slightly higher than, those of another study from Nepal that reported bioflm formation in  [27]. Te current study found that bioflm-producing bacteria also tended to be MDR strains. Studies from Egypt and Iran [45,46] found that the prevalence of MDR and XDR was higher in bioflm-producing strains. Of the bioflmproducing isolates from ventilator-associated pneumonia, 42.2% were MDR, but the relationship was statistically insignifcant [47]. 67% of patients had their empirical therapy adjusted after the AST result, and this adjustment was higher in infections with MDR and bioflm-forming isolates. It would be useful in future studies to test the resistance of strains in bioflms. Te numbers of ciprofoxacin-resistant Pseudomonas aeruginosa and Staphylococcus aureus strains in bioflms could not be reduced even when four times the minimum inhibitory concentration of ciprofoxacin was used, whereas the numbers of bacteria in bioflms of ciprofoxacin-sensitive strains could be reduced by ≥ 60% by ciprofoxacin at its minimum inhibitory concentration [48,49]. Infection with MDR bacteria results in longer days in the hospital [50], and the current study found that on average patients with MDR UTI remained in the hospital 2 days longer. Similarly, UTI patients infected with strong bioflm producers remained in the hospital for 3 days longer than patients infected with weak or nonbioflm producers. Tis correlation between days of hospital stays with MDR and bioflm formation may be interconnected. It is possible that the greater number of days in the hospital had a selective pressure for the generation of MDR and/or strong bioflmproducing isolates. Regardless of the relationship between MDR and bioflm, both can add signifcant levels of burden to healthcare providers as well as patients, as well as costs to the hospital system.

Conclusion and Recommendations
Overall, the antibiotic resistance patterns in the current study showed most of the empirically used antibiotics were inefective and over 50% of patients required therapy adjustment. A high percentage of uropathogens were found to be MDR and bioflm producers. Te study refects the crisis in resource-starved large hospital settings in developing countries without signifcant antibiotic stewardship strategies and the threat posed by increasing drug-resistant UTIs in renal disease. Strategies to inhibit or disperse bioflm formation by bacteria in vivo in renal disease patients should also be considered.
Tere is a need for the development of a protocol for rational use of antibiotics, and physicians and pharmacists must be aware of the rational use of antibiotics. Tere should be the use of a narrow spectrum of antibiotics when supported by clinical situations and culture reports. Furthermore, more thorough and routine studies are necessary to monitor for future changes in resistivity patterns. Tere must be provision for availability of trained pharmacists and microbiologists so that appropriate use of medicines and patient adherence to the treatment can be enhanced.

Data Availability
Data are available from corresponding author upon request.