Antibiotic Resistances of Enterobacteriaceae with Chromosomal Ampc in Urine Cultures: Review and Experience of a Spanish Hospital

The Enterobacteriaceae Citrobacter freundii, Enterobacter cloacae, Klebsiella aerogenes, Morganella morganii, Providencia stuartii, and Serratia marcescens (CESPM group) produce numerous urinary tract infections (UTIs) which are difficult to treat due to their high multiresistance rate. The objectives of this study were to carry out a systematic review of antibiotic resistances by UTIs and to determine changes over time in urine cultures from a reference hospital in southern Spain. The literature was searched for European data on the resistance rates of each microorganism, and a retrospective cross-sectional descriptive study was performed in samples with suspicion of UTI from patients in Virgen de las Nieves University Hospital (Granada, Spain) between 2016 and the first half of 2021. Among 21,838 positive urine cultures, 1.85% were caused by E. cloacae, 0.77% by M. Morganii, 0.65% by K. aerogenes, 0.46% by C. freundii, 0.29% by P stuartii, and 0.25% by S. marcescens. The lowest resistance rates by microorganism were: E. cloacae to amikacin (3.47%) and imipenem (5.28%); M. morganii to piperacillin–tazobactam (1.79%), cefepime (4.76%), and tobramycin (7.74%); K. aerogenes to tobramycin (3.55%), gentamicin (4.25%), trimethoprim–sulfamethoxazole (4.96%), imipenem (5.75%), and cefepime (6.43%); C. freundii to imipenem (no resistance), nitrofurantoin (1.96%), fosfomycin (2.80%), and ertapenem (6.12%); P. stuartii to cefepime (3.28%) and ceftazidime (3.28%); and S. marcescens to gentamicin (1.8%), ciprofloxacin (3.64%), cefepime (3.70%), piperacillin–tazobactam (3.70%), and trimethoprim–sulfamethoxazole (5.45%). In our setting, CESMP Enterobacteriaceae showed the lowest resistance to piperacillin–tazobactam, cefepime, imipenem, gentamicin, and colistin, which can therefore be recommended for the empirical treatment of UTIs. The COVID-19 pandemic may have had a clinical impact in relation to the increased resistance of E. cloacae and M. morgani to some antibiotics.


Introduction
Urinary tract infections (UTIs) are highly frequent [1], particularly among sexually active women of childbearing age [1,2]. It has been estimated that one-third of Primary Care visits are for infections, 10% of which are UTIs [3], and this does not take into account the numerous UTIs treated by self-medication or in hospital emergency departments, where 22% of treated infections are UTIs, the second most frequent after respiratory infections [1]. UTIs are the most common infection in hospitals, in which 80% are related to vesical catheters or other devices [4]. oratory between 1 January 2016 and 30 June 2021. No exclusion criteria were applied. Sample processing always followed the standard laboratory protocol of our hospital [15]. Matrix-assisted laser desorption ionization (MALDI) Biotyper (Bruker Daltonics, Billerica, MA, USA) or MicroScan (Beckman Coulter, Barcelona, Spain) systems were employed to identify microorganisms grown in culture, and MicroScan was used to evaluate their antibiotic susceptibility. The minimum inhibitory concentration (MIC) was recorded for each antibiotic. Isolates were categorized as susceptible, intermediate, or resistant to antibiotics in accordance with Clinical and Laboratory Standards Institute (CLSI) recommendations for each year until 2019, and thereafter, in accordance with recommendations of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for each year.
Information on urine sample origin, microorganism, and patient age was collected from the MODULAB ® system used by the Public Health System of Andalusia to support electronic clinical records. Data on UTI episodes were stratified by sex, age (≤14 years, 15-64 years, and ≥65 years), and origin (hospitalized vs. community). No clinical information was gathered to analyze clinical factors related to the presence of microorganisms.
The percentage resistance to the different antibiotics was compared by sex, age, and origin using Pearson's chi-square test, applying Fisher's exact test when chi-square test conditions were not met (≤20% of expected frequencies <5). R 4.4.1 software was used for data analyses, and p < 0.05 was considered significant.

Ethical Considerations Ethical Approval
The study protocol was carried out in accordance with the Declaration of Helsinki [16]. This was a non-interventional study with no additional investigation to routine procedures. Biological material was only used for standard infection diagnostics ordered by the attending physician. There was no additional sampling or modification of the routine sampling protocol of the laboratory. Data analyses were based on an anonymous database. For these reasons, ethics committee approval was considered unnecessary according to national guidelines. The Clinical Microbiology Clinical Management Unit of the University Hospital Virgen de las Nieves of Granada (Spain) granted permission to access and use the data.

Informed Consent
The study protocol was carried out in accordance with the Helsinki Declaration [16]. Data analyses were performed using an anonymous database. Therefore, approval was considered unnecessary according to the guidelines of our country (Law on Data Protection -Organic Law 15/1999 of 13 December on the protection of data of a personal nature, available online: https://www.boe.es/buscar/doc.php?id=BOE-A-1999-23750 (accessed on 30 June 2021)).

Global Prevalence
The HUVN microbiology laboratory processed 74,106 urine samples for suspicion of UTI between 1 January 2016 and 30 June 2021, with 21,838 (29.47%) testing positive. Table 1 displays the number of clinical isolates and the percentage of positive urine cultures per microorganism. E. cloacae was isolated in 405 patients, representing 1.85% of positive urine cultures, while S. marcescens was isolated in 55 patients, representing 0.25% of positive urine cultures.  Table 2 lists the number and percentage of clinical isolates per microorganism according to patient sex and age and sample origin and type. E. cloacae (p = 0.001) and S. marcescens (p = 0.019) were more frequently detected in males, while C. freundii (p = 0.006) was more frequently isolated in females. All microorganisms were more frequent in the hospital setting, except for P. stuartii, which was more prevalent in community samples. Although some antibiotics yielded resistance rates <10% (imipenem-relebactam, meropenem, doripenem, levofloxacin, and colistin), many were not effective in vitro against >50% of clinical isolates.

HUVN Laboratory Study
This study gathered 405 clinical isolates of Enterobacter cloacae with antibiogram. Table S3 exhibits the annualized general resistance rates, and Tables S4-S15 (Supplementary Material) show the results by category.
For piperacillin-tazobactam and ciprofloxacin, the respective cutoff points of MIC 16 and 0.5 mg/dL correspond to areas of technical uncertainty (ATUs) according to EUCAST 2022, and these were observed for piperacillin-tazobactam (9.45%) and ciprofloxacin (4.45%) in the present sample. Figure 1 depicts the upward trend over the years, especially between 2019 and 2021, in the percentage resistance of the antibiotics most frequently prescribed to treat UTIs caused by E. cloacae.

HUVN Laboratory Study
The laboratory identified 168 clinical isolates of M. morganii during the study period. Table S17 displays the annualized general resistance rates, and Tables S18-S29 (Supplementary Material) show the results by category.
Resistance rates to imipenem were higher in adults than in the elderly or children (p = 0.016), higher in the elderly than in adults, and higher in adults than in children against nalidixic acid (p < 0.001), ciprofloxacin (p = 0.016), and trimethoprim-sulfamethoxazole (p = 0.040), respectively.

HUVN Laboratory Study
The laboratory identified 141 isolates of Klebsiella aerogenes in samples with suspicion of UTI received during the study period. Table S31 lists the annualized general resistance rates, and Tables S32-S43 (Supplementary Material) show the results by category.
ATUs were observed for piperacillin-tazobactam (13.47%) and ciprofloxacin (2.84%). Figure 3 depicts the downward trend over the years, especially between 2020 and 2021, in the percentage resistance of the antibiotics most frequently prescribed to treat UTI caused by K. aerogenes.

HUVN Laboratory Study
The laboratory identified 107 clinical isolates of C. freundii during the study period. Table S45 exhibits the annualized general resistance rates, and Tables S46-S57 (Supplementary Material) show the results by category.
No significant differences in resistance rates were found among years or age groups. The resistance rates were higher in females versus males against cefixime (p = 0.013) and in hospital versus community samples against cefuroxime (p = 0.014), cefotaxime (p=0.016), and ceftazidime (p = 0.005).
ATUs were observed for piperacillin-tazobactam (8.05%) and ciprofloxacin (3.74%). Figure 4 depicts the upward trend in percentage resistance to ciprofloxacin, nitrofurantoin, and fosfomycin and the downward trend in resistance to trimethoprim-sulfamethoxazole, gentamicin, and cefepime between 2020 and 2021.

Systematic Review
No study was traced on the resistance of P. stuartii in urine cultures in Europe.

HUVN Laboratory Study
The laboratory identified 64 isolates of P. stuartii during the study period. Table S58  displays the annualized general resistance rates, and Tables S59-S70 (Supplementary Material) show the results by category. No significant differences were found by age, sex, sample, or year. No resistance was observed to ertapenem or piperacillin-tazobactam, while the resistance rates were low against cefepime (3.28%) and ceftazidime (3.28%) but much higher against imipenem (19.64%).

Serratia marcescens
No study was traced on the resistance of P. stuartii in urine cultures in Europe.

HUVN Laboratory Study
The laboratory identified 64 isolates of P. stuartii during the study period. Table S58  displays the annualized general resistance rates, and Tables S59-S70 (Supplementary  Material) show the results by category. No significant differences were found by age, sex, sample, or year. No resistance was observed to ertapenem or piperacillin-tazobactam, while the resistance rates were low against cefepime (3.28%) and ceftazidime (3.28%) but much higher against imipenem (19.64%).
ATUs were observed for ciprofloxacin (17.46%) but not for piperacillin-tazobactam. Table S71 (Supplementary Material) exhibits the two studies selected for review, which reported resistance rates that were relatively low, observing a rate of >30% against colistin alone (98.03%).

HUVN Laboratory Study
The laboratory identified 55 clinical isolates of Serratia marcescens during the study period. Table S72 lists the annualized general resistance rates, and Tables S73-S84 (Supplementary Material) show the results by category.

Antibiotic Resistances of Chromosomal AmpC-Producing Enterobacteriaceae E. cloacae
Cutoff points for E. cloacae followed EUCAST [17,18] guidelines in two studies and recommendations of the CLSI [19] and the Comité de l'antibiogramme de la Société Française de Microbiologie (CA-SFM) [20] in one study each. Resistance rates for E. cloacae were lower in our hospital than in the systematic review against all antibiotics except for cefepime, tobramycin, and colistin, which showed slightly higher resistance rates in our setting. Resistance rates were higher in one study than in the others, possibly because it only included the elderly [20]. An increase in resistance rates against numerous antibiotics has been detected over the past few years, which may be related to a wider prescription of antibiotics during the COVID-19 pandemic. Comparisons with data from the same laboratory in previous years [19] reveal an increase in resistance to fosfomycin (from 28 to 32.51%) and cefepime (20 to 26.48%). One of the largest reductions in the resistance rate was against gentamicin (18 to 10.62%) and imipenem (8 to 5.28%), while no major differences were observed for the other antibiotics. According to these findings, the lowest resistance rates (<10%) were against amikacin (3.47%) and imipenem (5.28%), which may therefore be the best choice for the empirical treatment of E. cloacae, with colistin (12.67%) being another possible option.

M. morganii
Cutoff points for M. morganii followed CLSI guidelines in all studies [19,[21][22][23] except for two that followed EUCAST recommendations [17,24]. In comparison to the studies in the review, isolates detected in the HUVN laboratory had higher resistance rates against cefotaxime, imipenem, gentamicin, fosfomycin, ciprofloxacin, and trimethoprimsulfamethoxazole but lower rates against ceftazidime, cefepime, piperacillin-tazobactam, and tobramycin. The resistance rate against imipenem was higher in adults than in the elderly or children but showed a general trend towards a reduction (29.41 to 9.09%) over the past few years. There was an increase in the resistance to cefuroxime, cefotaxime, ceftazidime, cefepime, tobramycin, and nitrofurantoin in 2020 and 2021, possibly attributable to a greater exposure to at-home and oral versus hospital and intravenous treatments. Resistance rates to fluroquinolones and trimethoprim-sulfamethoxazole were higher in the elderly than in adults or children. A comparison with data from the same laboratory in previous years [19] revealed a marked increase in the resistance of M. morganii to gentamicin, fosfomycin, nalidixic acid, ciprofloxacin, imipenem, and trimethoprim-sulfamethoxazole, especially to fosfomycin (from 4 to 80.61%), nalidixic acid (18 to 46.95%), and trimethoprimsulfamethoxazole (14 to 33.33%). In contrast, resistance rates decreased against tobramycin (22 to 7.74%), piperacillin-tazobactam (19 to 1.79%), and cefepime (10 to 4.76%).
Antibiotics with in vitro resistance rates <10% were observed against piperacillintazobactam, cefepime, and tobramycin, which may be appropriate empirical treatments of M. morganii, whereas fosfomycin cannot be recommended.

K. aerogenes
Cutoff points for K. aerogenes followed CLSI guidelines in two studies [18,19] and EUCAST guidelines in the other [17]. In comparison to the reviewed studies, the resistance rates of K. aerogenes isolates detected in the HUVN laboratory were higher against cefepime, ertapenem, fosfomycin, and colistin but lower against the other antibiotics under study.
Resistance rates to third-and fourth-generation cephalosporins and piperacillintazobactam were higher in males versus females and in hospital versus community samples. The comparison with previous data from the same laboratory [19] showed a decrease in resistances against gentamicin, tobramycin, nitrofurantoin, imipenem, piperacillintazobactam, and trimethoprim-sulfamethoxazole, most markedly against nitrofurantoin (47 to 16.15%), imipenem (18 to 10.64%), and piperacillin-tazobactam (18 to 10.64%), but an increase in resistance rate to fosfomycin (13 to 18.44%). The resistance patterns of this bacterium in our setting indicate numerous antibiotics had a resistance rates <10% against K. aerogenes in our setting, and cefepime, imipenem, piperacillin-tazobactam, gentamicin, tobramycin, ciprofloxacin, and trimethoprim-sulfamethoxazole may all be appropriate empirical treatments.

C. freundii
Three of the reviewed studies reported on 218 clinical isolates of C. freundii in Europe. Cutoff points followed CLSI guidelines [18,19] in two of the studies, while EUCAST guidelines were followed by other. In comparison to the reviewed studies, the resistance rates of C. freundii isolates detected in the HUVN laboratory were higher against cefepime, ertapenem, amikacin, tobramycin, levofloxacin, and colistin and lower against ceftazidime, imipenem, piperacillin-tazobactam, gentamicin, and nitrofurantoin. Differences were almost nonexistent. Very low resistance rates (<1.5%) were observed against nalidixic acid, fosfomycin, ciprofloxacin, and trimethoprim-sulfamethoxazole. Significantly higher resistance rates to cefixime, cefotaxime, and ceftazidime were found in females versus males. The comparison with previous data from the same laboratory [19] showed an absence of resistances to imipenem; a reduction in resistance rates to gentamicin (16 to 8.41%), tobramycin (15 to 9.3%), fosfomycin (7 to 2.8%), and nitrofurantoin (7 to 1.96%); and an increase in resistance to cefepime (4 to 14.95%). The antibiotics with resistance rates <10% that can be recommended for empirical treatment are gentamicin, tobramycin, amikacin, fosfomycin, nitrofurantoin, imipenem, ertapenem, piperacillin-tazobactam, and colistin.

P. stuartii
An insufficient number of cases of P. stuartii have been reported in the literature to reveal any trends in antibiotic resistances. The information obtained in our hospital indicates resistance rates <10% to ceftazidime and cefepime and no resistance to ertapenem or piperacillin-tazobactam.

S. marcescens
The systematic review retrieved two European studies with a total sample of 152 isolates of S. marcescens. Cutoff points in CLSI [18] guidelines were applied in one study, and those recommended by EUCAST [17] were applied in the other. In comparison to the systematic review, S. marcescens had lower resistance rates in the HUVN samples to all tested antibiotics except for ceftazidime, which were almost the same, and tobramycin, which demonstrated much higher rates in our setting (100 vs. 12.2%). The resistance rates to cefepime, imipenem, piperacillin-tazobactam, gentamicin, ciprofloxacin, and colistin were <10%.
Piperacillin-tazobactam appears appropriate for the empirical treatment of chromosomal AmpC-producing Enterobacteriaceae in our setting (in vitro resistance <10%), except for E. cloacae. Prospective studies reveal a low risk of clinical failure due to emerging resistance [25]. A Canadian study of 2394 urinary tract isolates between 2007 and 2009 observed resistance rates to piperacillin-tazobactam of 7.4 % for E. cloacae, 10.3% for Citrobacter spp., and 4.2% for M. morganii [26]. Piperacillin-tazobactam has been found to preserve its activity against M. morganii even when high levels of its AmpC enzyme are expressed [27]. An in vitro study in an animal model found that piperacillin-tazobactam did not select resistant mutants of E. cloacae as effectively as cephalosporin [28].
In the hospital laboratory series, ATUs observed for piperacillin-tazobactam occurred in 9.45% of E. cloacae isolates, 13.47% of K. aerogenes isolates, and 8.05% of C. freundii isolates. These cases require an additional test for confirmation or a change in clinical category [29].

Possible Impact of the COVID-19 Pandemic on Antibiotic Resistances
Recommendations by the Spanish Agency of Medicines and Medical Devices for optimal antibiotic prescriptions during the COVID-19 pandemic (2020 and 2021) emphasized that excessive or inappropriate prescriptions could favor the emergence of resistant bacteria, compromising the effectiveness of treatments [30].
In the present hospital series, increased resistance rates were observed in 2020 and 2021 for E. cloacae against cefuroxime, ceftazidime, cefepime, piperacillin-tazobactam, tobramycin, nalidixic acid, ciprofloxacin, levofloxacin, and trimethoprim-sulfamethoxazole and for M. morganii against cefuroxime, cefotaxime, ceftazidime, cefepime, tobramycin, and nitrofurantoin. No significant trend attributable to antibiotic use during the pandemic was observed for K. aerogenes, C. freundii, P. stuartii, or S. marcescens.

Limitations
No data were gathered on the antibiotic therapy applied in clinical practice, preventing the direct correlation of susceptibility/resistance results in vitro with therapeutic success or failure in vivo.
Information was collected in 2020 and the first half of 2021, i.e., during the COVID-19 pandemic, but data from the whole of 2021 should be studied to fully elucidate its effect on the development of antibiotic resistance. Furthermore, the study of UTIs due to P. stuartii and S. marcescens was limited by small sample sizes, preventing the evaluation of the effects of patient age and sex or sample origin and type.
The search of the literature traced relatively few articles (maximum of six) on each microorganism, and no European studies on P. stuartii were found.

Conclusions
In comparison to the data gathered in our systematic review of European studies, the overall resistance rates in our hospital were higher against cefepime, tobramycin, fosfomycin, colistin, and ertapenem. This is likely attributable to the greater utilization of these antibiotics against UTIs treated in our hospital.
In patients with suspicion of UTI caused by CEMPS microorganisms, clinicians should initially consider empirical antibiotic therapy administered via the parenteral route. This is because none of the antibiotics that showed the lowest resistance rates in vitro (piperacillintazobactam, cefepime, imipenem, and gentamycin), and may therefore be most useful as empirical therapy, can be administered orally.
The clinical impact of the COVID-19 pandemic influenced an upward trend in the antibiotic resistance of UTIs produced by some microorganisms of the CEMPS group. This especially affected empirical treatments with second-and third-generation cephalosporins and fluoroquinolones, which we would not recommend as first choices for empirical antibiotic therapy.