What's new?

Our study showed estimated cumulative failure of empirical treatment of urinary tract infections based on the latest guidelines in one‑quarter to nearly one‑half of patients (outpatients vs inpatients, respectively). Furthermore, resistance to cotrimoxazole exceeded the 20% accepted in the guidelines as a safe local level. Due to the unavailability of nitrofurantoin on the Polish market and the lack of labeling for the newly introduced drug pivmecillinam, it is necessary, based on this study, to adapt the guidelines to local epidemiologic conditions considering further newly acquired data. Currently, in practice, none of the first- or second‑line drugs proposed in the guidelines demonstrate an adequate level of evidence‑based efficiency.

Introduction

Urinary tract infections (UTIs) are a common condition that pose a significant health burden for both patients and health care systems worldwide. The absolute number of UTI cases increased by 60%, from 252.25 million in 1990 to 404.61 million in 2019. Global deaths due to UTIs rose by 140%, from 98 590 in 1990 to 236 790 in 2019.¹

UTIs can affect various structures of the urinary system, including the bladder (cystitis), urethra (ureteritis) and kidneys (pyelonephritis). Clinical consequences range from discomfort to serious complications, such as sepsis or kidney damage, often resulting in costly medical interventions. The 2025 Infectious Diseases Society of America (IDSA) guidelines introduced a new definition of a complicated UTI (cUTI) that significantly updates previous classifications. They now classify cUTIs as infections above the bladder in both women and men, as well as pyelonephritis, febrile or bacteremic UTI, catheter‑associated UTI, or prostatitis.² This classification is based on clinical presentation, risk factor stratification, and the availability of appropriate antimicrobial therapy.³ Importantly, underlying conditions, such as diabetes, immunocompromise, or benign prostatic hyperplasia do not automatically classify a UTI as complicated unless accompanied by the above features. Therefore, according to the new IDSA definition, the vast majority of UTIs are uncomplicated.²

In community‑acquired UTIs (CA‑UTIs), Escherichia coli predominates. In health care–acquired UTIs (HA‑UTIs), E. coli remains the most common pathogen, but with a higher proportion of nonfermenting strains, enterococci, and polymicrobial infections. In studies conducted in southern Italy, 95% of CA‑UTIs were caused by gram‑negative bacteria, with E. coli accounting for nearly 60% and Klebsiella pneumoniae for 25%. HA‑UTIs showed a higher share of enterococci and Pseudomonas spp. (approximately 10% each), as well as an increased proportion of K. pneumoniae (up to 30%), while the proportion of E. coli was lower (39%), as compared with CA‑UTIs.⁴

UTIs also exhibit one of the most pronounced sex disparities among infectious diseases, with premenopausal women being 20–40 times more likely to develop a UTI than men of the same age.⁵ Between 50% and 60% of adult women will experience at least 1 UTI in their lifetime, and nearly 10% of postmenopausal women report having had a UTI in the previous year.⁶

UTIs present differently in pre- and postmenopausal women due to physiological and microbiological factors. Premenopausal women with UTIs more frequently report severe symptoms, such as malaise, pain during urination, and dysuria, whereas postmenopausal women tend to have milder symptoms. In premenopausal women, UTIs are most often caused by E. coli, while postmenopausal women exhibit a greater variety of pathogens and often more resistant infections, likely due to increased hospitalization and higher cumulative antibiotic exposure.^7,8 The incidence of UTIs is by approximately 50% lower among older men than women; however, in individuals over 80 years, the proportion of women to men with UTIs is nearly equal.^9-11

In 2017, the World Health Organization (WHO) proposed the Access, Watch, Reserve (AWaRe) classification as an antibiotic management strategy.¹² This framework categorizes antibiotics according to their resistance potential into 3 groups. The Access group (A) includes agents with a low potential for resistance development, and should be used as the first‑line options. The Watch group comprises antibiotics that require strict monitoring, and should only be used when there is no alternative to A group agents. The Reserve group comprises agents that should be preserved as the last‑resort treatments, and includes newly developed drugs requiring special protection as well as older, more toxic antibiotics that should only be used when no alternatives are available. The AWaRe classification is updated every 2 years.¹² According to the WHO AWaRe Book, the recommended first‑line options for the treatment of lower urinary tract infections in primary health care are nitrofurantoin (A), sulfamethoxazole + trimethoprim, trimethoprim (A), and amoxicillin + clavulanic acid (A).¹³

However, in the face of increasing antimicrobial resistance, an evidence‑based approach and the use of local antibiotic guidelines are crucial when selecting optimal treatment options. The most effective strategy is to align empirical therapy with international recommendations, such as those of the IDSA or the WHO AWaRe framework, while also incorporating data from local cumulative antibiograms. These antibiograms provide a comprehensive overview of antimicrobial susceptibility patterns of bacterial isolates within a specific hospital or region, thereby reflecting the local epidemiology of resistance. This is particularly important, as resistance profiles can vary significantly between geographic locations and health care settings.

In clinical decision‑making, especially when data on optimal susceptibility thresholds are limited, it is generally recommended to select agents with 90%–95% susceptibility when the risk of morbidity or mortality is high. Antibiotics with 80%–85% susceptibility may be considered acceptable for patients with a lower risk of adverse outcomes within the next 24–48 hours.^14,15 Agents with susceptibility below 80% should be avoided as empirical treatment and reserved for situations when no alternatives are available. Nevertheless, interpretation of antibiograms should not occur in isolation; clinical and epidemiologic factors must also be considered to ensure optimal therapeutic outcomes.¹⁶

According to the latest guidelines of the European Association of Urology (EAU)¹⁷ and the Polish Urological Association,¹⁸ the first‑line treatment for uncomplicated UTIs in women includes fosfomycin, nitrofurantoin, or pivmecillinam. The Polish guidelines additionally recommend furazidine. As second‑line therapy for women, but first‑line for men, cephalosporins or trimethoprim‑sulfamethoxazole are advised.^17,18 However, the availability of certain antibiotics and resistance rates differ between Poland and other European countries; for example, nitrofurantoin is not available in Poland. Moreover, no prospective studies have validated the guideline‑recommended empirical treatments in the European settings. Given the limited representation of specific populations in the studies underlying these recommendations, countries such as Poland may face a higher risk of the empirical therapy ineffectiveness. Therefore, this study aimed to analyze the bacterial etiology of UTIs and their resistance patterns in the context of empirical regimens recommended by the European and national guidelines for for CA‑UTIs and HA‑UTIs. The analysis also assessed the safety of UTI treatments using the WHO AWaRe classification.

Patients and methods

Statistical analysis

Absolute values and percentages were used to describe the study population. The χ² test was applied for group comparisons. Statistical analysis was performed using Microsoft Excel version 16.102.3 (Redmond, Washington, United States).

The potential risk of available empirical therapy ineffectiveness was estimated as weighted average based on the following equation:

where S% is a susceptibility rate.

A P value below 0.05 was considered significant.

Results

Antibiotic susceptibility testing

First- and second‑line antibiotics

In the entire study population, only 2 first‑line antibiotics were tested frequently enough to meet the inclusion criterion: cotrimoxazole and nitrofurantoin (eg, no susceptibility tests were performed for oral fosfomycin and pivmecillinam).

The inpatients showed 73% susceptibility to cotrimoxazole across all patient groups, while susceptibility was lower in the outpatients. Among the outpatient women, E. coli isolates demonstrated 91.85% susceptibility to nitrofurantoin, with even higher rates observed for E. faecalis.

In the inpatient setting, E. coli isolates from premenopausal women showed high susceptibility to all recommended cephalosporins (≥86.99%), whereas in men and postmenopausal women, susceptibility exceeded 76.58%. For K. pneumoniae, regardless of sex or age, the only second‑line treatment with high susceptibility was cefoperazone / sulbactam (86.25%), while all other agents did not reach 60%.

In the outpatient setting, E. coli isolates remained highly susceptible to cephalosporins (>86%), whereas K. pneumoniae showed lower susceptibility (<⁠69%).

With the exception of outpatient treatment for E. faecalis, the data were insufficient to assess the recommended treatment for Enterococcus spp. (Table 2).

Table 2. Antibiotic susceptibility of common urinary tract infection pathogens in inpatients and outpatients in southern Poland, 2022–2023 (n = 20 479)

Inpatients	Premenopausal women				Others
Antimicrobial agent	Escherichia coli	Klebsiella pneumoniae	Enterococcus faecalis	Enterococcus faecium	Escherichia coli	Klebsiella pneumoniae	Enterococcus faecalis	Enterococcus faecium
a I scores – susceptible, increased exposure (or susceptible dose‑dependent) Abbreviations: iv, intravenous; n/a, not available; ND, no data, not routinely examined; po, per os; uUTI, uncomplicated urinary tract infection
First‑line treatment, %
Fosfomycin, po	ND	ND	n/a	n/a	ND	ND	n/a	n/a
Fosfomycin, iv	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a
Nitrofurantoin (uUTI)	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a
Pivmecillinam	ND	ND	n/a	n/a	ND	ND	n/a	n/a
Cotrimoxazole	75.8	60	n/a	n/a	67.5	40.5	n/a	n/a
Second‑line treatment, %
Cefepime	91.9	59.5a	n/a	n/a	86.5	42.2	n/a	n/a
Cefoperazone / sulbactam	100	97.2	n/a	n/a	98.8	86.3	n/a	n/a
Cefotaxime	87.7	55	n/a	n/a	78.2	30.9	n/a	n/a
Ceftazidime	89.5	52.8	n/a	n/a	82.7	33.3	n/a	n/a
Cefuroxime, iv	87a	50	n/a	n/a	76.6	29	n/a	n/a
Outpatients	Women				Men
First‑line treatment, %
Fosfomycin, po	ND	ND	n/a	n/a	ND	ND	n/a	n/a
Fosfomycin, iv	ND	ND	n/a	n/a	ND	ND	n/a	n/a
Nitrofurantoin (uUTI)	91.9	ND	98.9	n/a	n/a	ND	n/a	n/a
Pivmecillinam	ND	ND	n/a	n/a	ND	ND	n/a	n/a
Cotrimoxazole	72.4	69.2	n/a	n/a	71.5	52.1	n/a	n/a
Second‑line treatment, %
Cefotaxime	91.9	68.3	n/a	n/a	88.9	47.1	n/a	n/a
Cefuroxime, iv	90.5a	67.8a	n/a	n/a	86.9a	46.9a	n/a	n/a

Other antibiotics: Access group

Among the first‑line antibiotics listed in the WHO AWaRe Book, only nitrofurantoin from the A group showed sufficient susceptibility in selected outpatient populations. Among the other A group antibiotics, only gentamicin was potentially useful against the most common isolates in our study, with susceptibility ranging from 92% for E. coli to 38% for E. faecium.

An over 80% susceptibility threshold was reached only by aminoglycosides for E. coli and by amoxicillin and ampicillin for E. faecalis.

In the 60%–80% range, aminoglycosides for K. pneumoniae, as well as ampicillin / sulbactam and tetracycline for E. coli, met the threshold. All other routinely assessed A group antibiotics showed susceptibility below 60% (Figures 1 and 2).

Other antibiotics: Watch and Reserve groups

Susceptibility exceeding 80% was observed for carbapenems and piperacillin / tazobactam in E. coli, and for carbapenems in K. pneumoniae. In E. faecalis, all recommended EAU agents (linezolid, teicoplanin), except vancomycin, reached the above 80% threshold, whereas in E. faecium only linezolid met this criterion.

Within the 60%–80% range, only piperacillin / tazobactam in K. pneumoniae and vancomycin in E. faecalis met the threshold. All other Watch / Reserve group antibiotics demonstrated susceptibility below 60% (Figures 1 and 2).

Discussion

Gram‑negative bacilli were the dominant isolates in both inpatients and outpatients, accounting for 60% and 70% of cases, respectively. This was expected based on epidemiologic data; however, the proportion of Enterococcus spp., including multidrug‑resistant strains, was alarmingly high, accounting for about 15% of cases.²³ Antibiotic susceptibility data differed markedly from the EAU expectations. Resistance to cephalosporins reached up to 70% depending on the microorganism, while cotrimoxazole resistance was as high as 60%. Fluoroquinolone efficacy did not exceed 70% in the overall population. The prevalence of HLAR in the cultures positive for E. faecium exceeded 55%, and ESBL production in K. pneumoniae reached 41%. ESBL‑producing bacteria were the most common resistance phenotype.

Additionally, there was a high share of bacteria classified as “susceptible, increased exposure” (formerly “susceptible‑dose dependent”), particularly to cephalosporins. This indicates that standard doses or dosing frequencies may be insufficient to achieve therapeutic efficacy.

The latest 2025 IDSA guidelines¹² recommend cephalosporins for cUTIs. However, our results show that empirical use of cephalosporins in this patient population is questionable. Alternative options include carbapenems or piperacillin‑tazobactam, which were effective against E. coli and K. pneumoniae in our cohort but are unsuitable for oral outpatient therapy. Fluoroquinolones remain an oral option but susceptibility rates in Poland are unacceptably low.²³ Importantly, in severely ill patients with sepsis, antibiotic efficacy should exceed 80% at the population level.²

In Poland, as in other European countries, current guidelines recommend several safe options for empirical treatment of UTIs, with pivmecillinam and fosfomycin among the primary choices.¹⁸ However, to the best of our knowledge, susceptibility testing for pivmecillinam and oral fosfomycin is not routinely performed in Poland. Limited availability of reagents for manual testing, together with the absence of these antibiotics in most automated AST systems, further hinder their routine assessment in clinical microbiology laboratories. Consequently, based on data from southern Poland, where the susceptibility of Enterobacterales strains to pivmecillinam and fosfomycin is unknown, and susceptibility to cotrimoxazole is approximately 60%, there is a potential risk of therapeutic failure with their use as empirical therapy.

Nitrofurantoin is not available on the Polish pharmaceutical market, but it is routinely included in Polish AST. Both nitrofurantoin and furazidine are approved for use within the European Union; however, furazidine (commonly used in Poland) is not routinely tested because of the lack of standardized protocols and interpretation criteria. At present, EUCAST does not provide clinical breakpoints for furazidine and does not recommend susceptibility testing for nitrofuran derivatives other than nitrofurantoin.²⁴ Furthermore, furazidine is the only antibiotic available in Poland without prescription, and no evidence supports extrapolating of nitrofurantoin susceptibility results to furazidine. This creates significant challenges in assessing the potential effectiveness of empirical treatments recommended by the EAU.¹⁸

Additionally, cephalosporins recommended as the first‑line alternatives showed reduced efficacy, with resistance rates up to 55% in outpatients and over 70% in inpatients, consistent with the high share of ESBL producers in positive cultures. Aththanayaka et al²⁵ reported similar findings, with 70% resistance to cotrimoxazole despite ESBL prevalence of less than 19%. In their study, the highest efficacy was observed for mecillinam and nitrofurantoin, which were not assessed in Poland. Findings from other Polish studies also support our hypothesis that current guidelines may have limited applicability to the Polish population.²⁶ Among E. coli isolates, reported susceptibility was 85% to nitrofurantoin, 82% to cefuroxime, just over 60% to fluoroquinolones, and approximately 55% to cotrimoxazole.²⁶

In a large population‑based study from Scandinavian countries, where pivmecillinam has been widely used for more than a decade, resistance rates remained between 4%–6%, which was unusual given the substantial increase in its consumption during this period.²⁷ In our study, susceptibility to cephalosporins in premenopausal women was below 90% which, given the favorable safety profile of β-lactam antibiotics, may significantly limit available treatment options. Pivmecillinam may therefore represent a valuable therapeutic alternative for the Polish market.

The differences resulting from patient sex and age are primarily attributable to anatomical and hormonal factors, the latter mainly affecting the premenopausal period due to higher estrogen levels.⁵ Gu et al²⁸ highlighted the different distribution of pathogens by sex, which affects AST results. In their study, higher susceptibility to cephalosporins, trimethoprim / sulfamethoxazole, and penicillins was observed in premenopausal women, as compared with men.²⁸ In our cohort, sex‑related differences were mainly associated with variations in causative organisms rather than antimicrobial susceptibility. However, as shown in Figure 1, treatment of UTIs or bacteriuria in premenopausal women, particularly during pregnancy, may be challenging when K. pneumoniae is involved.

In a large United States (US) study conducted by Ku et al,²⁹ UTI treatment was primarily based on cephalosporins and fluoroquinolones, with resistance rates of around 10% during the first infectious episode. In our study, similar cephalosporin resistance patterns were observed only for E. coli, which was also the dominant pathogen in the US study. However, the susceptibility of E. coli to fluoroquinolones in our cohort was too low to justify empirical treatment (64.72%–68.47% vs 91.2% in the US).²⁹

The European Medicines Agency (EMA) issued warnings in 2019 regarding the use of fluoroquinolones due to rare but extremely serious adverse effects.³⁰ Consequently, the EAU guidelines recommend reserving fluoroquinolones as the last‑line option, to be used only in the absence of alternatives. The EMA Pharmacovigilance Risk Assessment Committee reaffirmed these recommendations in 2023, emphasizing that fluoroquinolones should not be used in nonsevere infections that may resolve spontaneously or in recurrent lower UITs.³¹ Our findings support the recommendation against fluoroquinolone use for UTIs in Poland, particularly for empirical therapy, given the high resistance rates among the most common bacterial uropathogens.

A comparable challenge in evaluating the EAU‑recommended empirical regimens involves Enterococcus spp., for which susceptibility testing to the first‑line antibiotics (fosfomycin, nitrofurantoin, pivmecillinam, and cotrimoxazole) is routinely performed. Moreover, we observed high resistance rates to aminoglycosides (>30%) and glycopeptides (>20% for vancomycin), particularly in hospitalized patients. El‑Mahdy et al³² reported similar HLAR prevalence in a study from Egypt, although in their cohort all isolates were resistant to penicillins but remained fully susceptible to glycopeptides. A large Hungarian study analyzing over 40 000 urine samples found HLAR prevalence of approximately 45%, with an increasing trend over several years, although high susceptibility to ampicillin was preserved, consistent with our findings.³³ Therefore, considering the recommended empirical options and the lack of determination of enterococcal susceptibility to the most used antibiotics (lack of routine AST), the risk of potential ineffectiveness of empirical therapy in this group may be considerable.

Although amoxicillin demonstrates high susceptibility rates against the most common isolates, it is not recommended for treatment, and its efficacy against E. coli has not been routinely assessed.

Considering the WHO AWaRe classification, our findings show that among the antibiotics recommended for the most frequently isolated pathogens, only aminoglycosides—despite their known toxicity—and penicillin derivatives for enterococci demonstrate acceptable susceptibility rates. Overall, the data suggest that empirical treatment based on current EAU guidelines carries a 50%–70% risk of potential empirical therapy ineffectiveness. This underscores the need to adapt international recommendations to local antibiotic susceptibility patterns, particularly within the Polish health care context. Our results highlight the substantial limitations of empirical therapy guided solely by international standards and support the use of β-lactam–based regimens as the most rational first‑line option in Poland, with aminoglycosides reserved for severe or complicated cases. Unfortunately, the limited effectiveness of the A group antibiotics and the overreliance on Watch and Reserve category drugs further emphasize the urgent need for revising national treatment guidelines to reflect local resistance trends.