Carbapenem De-escalation in urinary tract infections: prevalence and outcomes among hospitalized patients

Abstract

Objective

To evaluate the prevalence and outcomes of carbapenem de-escalation among hospitalized urinary tract infection (UTI) patients at Jordan University Hospital from January 2022 to March 2024.

Methods

This retrospective study included adult patients who received carbapenems as empirical therapy and underwent urine culture testing. Patients who were discharged before culture results (n = 31) were excluded. The primary outcome was to assess the rate of effective de-escalation, which is the transition to a narrower-spectrum antibiotic without therapeutic failure. Secondary outcomes included the effect on successful de-escalation on patients’ hospital length of stay, and the relationship between urine culture results and the success of de-escalation. When determining whether to de-escalate, factors such as clinical stability, and urine culture results and susceptibility were considered. Chi-square test assessed associations between culture results and de-escalation success.

Results

A total of 205 patients who received carbapenems as empirical therapy and underwent urine culture testing were included in the study. Out of these, 116 individuals (56.6%) had negative culture results, indicating no infection, while 89 individuals (43.4%) tested positive, confirming the presence of an infection. Among these patients, 95.6% (196 out of 205) required de-escalation of their treatment, whereas 4.4% did not. The prevalence of successful de-escalation was 40.3% (79 out of 196). The analysis revealed that successful de-escalation is much more likely in cases with urine cultures showing growth (86.8%) compared to those with no growth (17.2%) (p = 0.001). The study also indicated no significant differences in the length of hospital stay between the successfully de-escalated group and those who failed or were incorrectly de-escalated (P > 0.05).

Conclusions

The study underscores the challenges of implementing effective antibiotic stewardship in UTI management, particularly regarding carbapenem de-escalation. Enhanced protocols and clinician education are essential to optimize de-escalation practices, especially in cases with negative microbiological results. Further research is needed to refine these strategies and address the growing issue of antibiotic resistance in UTIs.

Introduction

Urinary tract infections (UTIs) are among the most prevalent bacterial infections globally, affecting approximately 18 per 1,000 individuals annually [19]. Beyond their high incidence, UTIs pose significant challenges due to their impact on patient quality of life and the considerable clinical and financial burdens they impose on healthcare systems [27]. The rise of antibiotic resistance has exacerbated these challenges, with mortality rates exceeding 10% in cases complicated by UTI-related bacteremia [11].

Infections caused by antibiotic-resistant bacteria are typically more difficult to treat, leading to potential relapses and a substantial increase in morbidity and mortality (Christaki, Marcou and Tofarides [3]. Surveillance networks in Europe and Asia have recorded a noteworthy rise in antibiotic-resistant bacteria over the past decade, while the United States witnesses millions of people infected with antibiotic-resistant strains annually, resulting in thousands of deaths [3]. Globally, antimicrobial resistance claims approximately 700,000 lives each year, with the potential for this figure to reach 10 million deaths annually by 2050 [3].

Empirical antibiotic therapy for UTIs must strike a delicate balance between effectively treating infections and minimizing unnecessary antimicrobial exposure, which contributes to the development and spread of antimicrobial resistance [16]. Carbapenems, known for their broad-spectrum activity, are frequently chosen as empirical treatment options due to their effectiveness against common UTI pathogens, including multi-drug resistant strains [25].

Escherichia coli remains the predominant causative pathogen in UTIs, demonstrating notable resistance to antibiotics such as ampicillin, amoxicillin, and trimethoprim/sulfamethoxazole [6]. Monitoring antibiotic resistance trends is crucial not only for guiding clinical treatment decisions but also for informing public health policies aimed at combating resistance through targeted interventions [6].

In recent years, there has been a growing emphasis on antibiotic de-escalation as a strategic approach to optimizing UTI treatment outcomes [2]. Antibiotic de-escalation involves the process of switching from empiric, broad-spectrum medication to narrow-spectrum agents based on clinical response and microbiological data, with the aim of minimizing unnecessary antimicrobial exposure and reducing the emergence of resistance [13, 14]. This practice is widely recognized as safe and effective in various infectious diseases, including pneumonia (hospital-acquired or community-acquired), bacteremia, UTIs, sepsis with bloodstream infection, severe sepsis among neutropenic patients, and pneumococcal bacteremia [5]. This study aims to investigate the prevalence and outcomes of carbapenem de-escalation among hospitalized UTI patients at a tertiary teaching hospital in Jordan. By assessing de-escalation success, our findings contribute to improving antimicrobial stewardship efforts. And since carbapenems are broad-spectrum antibiotics, optimizing their use is considered to prevent resistance.

Methods

Study setting

This study was a retrospective single-center cohort study conducted at Jordan University Hospital (JUH) from January 2022 to March 2024, to study the prevalence of carbapenem de-escalation for patients with UTI. Established in 1971, JUH is the first university teaching hospital in Amman, Jordan, and a pioneer in the Arab World. JUH boasts a capacity of 600 beds across 64 specialty and subspecialty areas [23].

Patients

The study’s inclusion criteria targeted adults aged 18 years or older who were suspected of having a UTI based on their symptoms, as noted by the attending physician in the admission notes. These patients received carbapenem as empirical antibacterial therapy, and urine cultures were collected before the initiation of treatment. Patients were required to remain hospitalized until culture and susceptibility results were available. Exclusion criteria for this study encompassed patients with uncertain diagnoses, pregnant individuals, those presenting with infections other than UTI, or individuals with incomplete data.

Data collection

During the data collection phase, comprehensive information was sourced from the computerized medical records of eligible patients. A systematic extraction process using a preprepared data collection form involved retrieving key patient details, including age, gender, dates of admission and discharge, length of hospital stay, and admitting diagnosis.

Moreover, specific attention was directed towards the collection of antibiotic-related data. This encompassed careful documentation of empiric antibiotics prescribed, including details such as the number of empiric antibiotics and administration routes. Information about urine culture and susceptibility results was also collected to provide crucial insights into the microbial agents responsible for the UTIs and their susceptibility patterns. The date on which microbiological cultures were requested was also documented. Furthermore, information on specific antibiotics was systematically recorded, including the number, type and administration route. The precise timing of antibiotic de-escalation, if applicable, was also documented.

Primary outcome

Carbapenem de-escalation was the primary outcome of this study. The carbapenems available in our hospital’s formulary are imipenem-cilastatin, meropenem, and ertapenem. Carbapenem de-escalation involves changing the intravenous empiric broad-spectrum carbapenem to narrower spectrum antibiotics (either oral or intravenous) such as fluoroquinolones, cephalosporins, penicillins (ampicillin/sulbactam, amoxicillin/clavulanate), and trimethoprim/sulfamethoxazole, following the availability of final culture and sensitivity results [17]. If multiple antibiotics are prescribed, the number should be reduced to a single antibiotic whenever possible for positive culture results [26]. Additionally, there is a commitment to discontinue antimicrobial treatment if no infection is confirmed, which means negative culture results [18, 26]. Furthermore, clinical stabilization for urinary tract infections (UTI), when available in the medical record, was also considered in evaluating the success of de-escalation. Also, de-escalation is recommended to be performed within 48 h after reporting culture and sensitivity results [17, 22].

Following the availability of final culture and sensitivity results, patients were categorized into two groups: those necessitating de-escalation and those not requiring it. Patients not necessitating de-escalation were those with positive culture results for Pseudomonas, Acinetobacter, or extended spectrum β-lactamase (ESBL) -producing Enterobacterales [7, 28]. These patients should continue carbapenem antibiotics as part of empiric treatment only if the isolated microorganisms are susceptible to carbapenems [28]. However, in cases of carbapenem-resistant infections, such as those caused by Carbapenem-resistant Enterobacteriaceae (CRE), a different treatment strategy is needed due to limited therapeutic options and high mortality rates. Colistin-based regimens, including monotherapy or combination therapy with fosfomycin, have been scouted, but the optimal approach remains uncertain [8]. Any instances of de-escalation for them were considered incorrect. Patients requiring de-escalation were divided into three groups: those who were successfully de-escalated, those for whom de-escalation failed, and those who were incorrectly de-escalated.

Successful de-escalation means switching from a broad-spectrum carbapenem to a narrower-spectrum antibiotic if bacterial growth is present, discontinuing antibiotics if no bacterial growth is detected, or transitioning from multiple antibiotics to a single agent whenever possible [13, 14].

Failed de-escalated therapy was defined as the continuation of empirical antibiotics in cases of negative culture results, the continuation of broad-spectrum empirical antibiotics in cases of culture-positive results when a narrower spectrum would have been sufficient, and the continuation of combination empirical antibiotics in cases of culture-positive results when a single agent would have been sufficient.

Incorrectly de-escalating antibiotic therapy occurs when de-escalation is done improperly. In cases where culture results are negative, antibiotics have been de-escalated from broad spectrum to narrower spectrum options instead of being stopped altogether, or from multiple to single antibiotics. Conversely, in cases where culture results are positive, antibiotics have been stopped instead of being converted to narrower spectrum options, leaving patients untreated.

Secondary outcomes

The study’s secondary objectives included several features that were analyzed between patients who successfully de-escalated and those who failed or incorrectly de-escalated after excluding patients who did not require de-escalation and cases where de-escalation was inaccurately implemented, such as antibiotics being stopped instead of transitioning to a narrower spectrum following positive urine culture results. For these two included groups, we assessed the following outcomes: length of hospital stays (duration from admission to discharge), and ICU visits. Additionally, we compared the rates of successful carbapenem de-escalation between patients with positive results and those with negative results.

Statistical analysis

Data were analyzed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY, USA). The descriptive data were presented using frequencies and percentages for categorical variables and median with an interquartile range for continuous variables. The normality of the continuous variable (length of stay) was assessed using the Shapiro-Wilk test, which indicated that the data were not normally distributed. Therefore, the Mann-Whitney U test was used for comparisons between the two groups. The differences between the two groups (the successfully de-escalated group and the failed/incorrectly deescalated group) were assessed using Mann Whitney U test or Chi-square test. All reported p values were two-tailed. The threshold for statistical significance was set at P < 0.05.

Results

Demographic and medical characteristics of the study sample

During the study period, a total of 236 patients were admitted with various complaints. All of these patients underwent urine culture tests and were prescribed carbapenems as empirical antibiotics for suspected UTIs, rendering them eligible for inclusion in our study. Among them, 13.1% (n = 31) were discharged before the availability of culture results and thus were excluded from the study. The median age of the remaining 205 patients was 62.0 years (IQR = 19.0). Roughly half of the participants (n = 88, 42.9%) were elderly (≥ 65 years old), and more than two-thirds of them were males (n = 149, 72.7%). Further details regarding the demographic and medical characteristics of the study sample are provided in Table 1.

Table 1 Demographic and medical characteristics of the study sample (n = 205)

Admission characteristics, ICU transfers, and readmissions

The study reveals that the primary reasons for patient admissions (Table 2) included renal stone (n = 48, 23.4%), bladder-neck obstruction (n = 33, 16.1%), and benign prostatic hyperplasia (n = 15, 7.3%). Notably, no patients required transfer to the ICU. Within six months, only a minority of patients (n = 75, 36.6%) experienced readmissions, with the majority having one readmission (n = 59, 28.8%). Renal stone was the most prevalent reason for readmission (n = 19, 9.3%).

Table 2 Patient admission characteristics, ICU transfers, readmissions, and reasons for readmission (n = 205)

Antibiotics prescription patterns among the study sample

In the study sample (n = 205), all patients received carbapenem, predominantly imipenem-cilastatin IV (n = 184, 89.8%). The majority solely received carbapenem (n = 161, 78.5%), while a minority were prescribed other antibiotics alongside carbapenem (n = 44, 21.5%), most commonly levofloxacin IV (n = 16, 27.5%). The median number of prescribed empiric antibiotics was 1.0 (IQR = 0.0), totaling 263 prescribed empiric antibacterial agents, including carbapenem. Further details are available in Table 3.

Table 3 Carbapenem and empirical antimicrobial treatment details (n = 205)

Urine culture and susceptibility testing

Table 4 presents details regarding urine culture and susceptibility testing. The microbial culture results indicate that most cultures had negative results (n = 116, 56.6%), while positive results were less common (n = 89, 43.4%). Most culture results were available within 24 h of hospital admission (n = 72, 35.1%). Approximately less than half of the positive culture results underwent susceptibility testing for the isolated pathogen (n = 34 out of 89, 38.2%). Notably, most of the isolated pathogens exhibited susceptibility to carbapenem (n = 29 out of 34, 85.3%), while a few were resistant to carbapenem (n = 5 out of 34, 14.7%).

Table 4 Urine culture and susceptibility results for the study sample (n = 205)

Of the positive urine culture specimens, the majority revealed the presence of one microorganism (60 out of 89, 67.4%), while a smaller proportion showed two pathogens (25 out of 89, 28.1%). In total, 122 microorganisms were identified. The most frequently reported pathogens among the urine cultures were Escherichia coli (39 out of 122, 31.9%), Enterococcus spp. (20 out of 122, 16.4%), Staphylococcus spp. (15 out of 122, 12.2%), and Klebsiella spp. (12 out of 122, 10%). For further details, please refer to Fig. 1.

Fig. 1

Distribution of microorganisms identified in positive culture results (n = 122)

Assessing the appropriateness of carbapenem De-escalation

Among patients with suspected UTIs included in the study, 56.6% had no growth in their urine cultures (n = 116). These patients were categorized into three groups: incorrectly de-escalated (n = 83, 71.6%), failed de-escalation (n = 13, 11.2%), and successfully de-escalated (n = 20, 17.2%). Incorrect de-escalation involved a shift from broad-spectrum to narrow-spectrum antibiotics (n = 77, 66.4%) or from multiple antibiotics to a single one (n = 6, 5.2%). Failed de-escalation included persistent carbapenem use (n = 12, 10.3%) or increasing the number of antibiotics (n = 1, 0.9%).

For patients with positive urine cultures (n = 89), 89.9% required de-escalation (n = 80), with 13.5% incorrectly de-escalated (n = 12), 10.1% failed de-escalation (n = 9), and 66.3% successfully de-escalated (n = 59). Successful de-escalation included transitioning to narrower spectrum antibiotics (n = 55, 61.8%) or from multiple antibiotics to one (n = 4, 4.5%). Overall, 40.3% of patients needing de-escalation were successfully de-escalated (79 out of 196). Overall, the prevalence of successfully de-escalated patients for the two arms was 40.3% (79 out of 196 who needed de-escalation) (Fig. 2).

Fig. 2

Decision-making Scheme for De-escalation Process

Assessing the secondary objectives of the study

To refine the analysis, the initial sample of 205 patients was reduced by excluding those who did not require de-escalation (n = 9) and those with inaccurately implemented de-escalation (n = 12). This resulted in a final sample size of 184 patients, with 79 successfully de-escalated and 105 experiencing incorrect or failed de-escalation. This refined sample was used to evaluate the study’s secondary objectives.

In examining these objectives, the study found no statistically significant difference in the median length of hospital stay between patients who were successfully de-escalated and those who failed or were incorrectly de-escalated (Table 5). Additionally, no patients required ICU admission during their hospital stay, precluding analysis related to ICU visits. However, a notable association was found between urine cultures showing growth and successful de-escalation practices compared to cases with no growth (Table 6).

Table 5 Length of hospital stay and 6-month all-cause mortality among the study sample (n = 184)

Table 6 Association between culture results and de-escalation practice among the study sample (n = 184)

Discussion

In this retrospective study, we assessed the prevalence of successful carbapenem de-escalation in the management of UTIs among hospitalized patients. Our findings revealed a 40.3% success rate for de-escalation. These results emphasize the importance of timely and appropriate antibiotic adjustments to optimize patient outcomes and antimicrobial stewardship.

This study identified that Escherichia coli was the most frequently reported pathogen among patients with positive urine cultures, accounting for 31.9% of cases. Recognizing Escherichia coli as the predominant bacterium can facilitate more targeted and effective antibiotic therapy. Since Escherichia coli is a well-characterized organism with established UTI treatment protocols, this knowledge can improve the accuracy of empirical therapy and potentially enhance the outcomes of antibiotic de-escalation strategies [5].

Also, we identified an alarming percentage of patients (13.1%) excluded from our study because they were discharged too early before the availability of culture results. This early discharge poses significant challenges to effective antibiotic stewardship and optimal patient care [15]. Several factors may contribute to early discharge, including financial burdens as patients may face in covering the costs associated with ordering a urine culture, the need to free up hospital beds, reliance on clinical judgment, patients’ personal or work-related preferences, and hospital policies or insurance limitations favoring shorter stays. Additionally, the role of urinalysis in early diagnosis should not be overlooked, as it may help reduce the need for cultures and antibiotic therapy [1]. These issues underscore the need for a balanced approach considering both economic and clinical factors (Steiger, Comito and Nicolau [24], Kollef [10].

The negative consequences of early discharge before culture results may have impacted the results, as patients may have received initial treatments that were not fully effective or appropriate for their condition. This is because the culture results, which are crucial for accurate diagnosis and optimal antibiotic treatment, were unavailable at the time of discharge. As a result, patients may have been treated with suboptimal antibiotics, which could lead to persistent or worsening symptoms and a higher risk of readmission [20]. To improve management for early-discharge patients, they should have a follow-up appointment with their clinician at an outpatient clinic. This allows the clinician to assess the appropriateness of the prescribed antibiotics and ensure the resolution of the infection.

Our study aimed to assess the de-escalation of successful carbapenem de-escalation within 48 h among patients hospitalized with UTIs. We found that the prevalence of successful carbapenem de-escalation for both positive and negative urine cultures of patients who needed de-escalation was 40.3% (n = 79/196). It was slightly lower than that reported in previous studies [9, 12, 22]. A study by Sadyrbaeva et al. conducted in a Spanish hospital showed a successful de-escalation rate of 49.7% (n = 81) among patients with complicated UTIs within 72 h [22]. Another retrospective study investigated antibiotic de-escalation practices in 65 patients with bacteremic UTI at a community hospital. The initial antibiotic regimen was successfully de-escalated in 52.3% of cases [9]. Additionally, a single-center retrospective study spanning one year by Lew et al. reported a 68% rate of successful carbapenem de-escalation (n = 204) [12].

Several factors could explain these differences in de-escalation rates. One critical variable is the time frame for de-escalation. Our requirement of a 48-hour window for successful de-escalation may have contributed to the lower prevalence compared to studies with more extended windows of 72 h or five days. Additionally, local antimicrobial stewardship policies, healthcare practices, and the spectrum of pathogens in our patient population might influence de-escalation practices and success rates.

In assessing the hospital length of stay difference between patients who were successfully de-escalated and those who either failed or were incorrectly de-escalated, we found no statistically significant differences between the two groups. This finding suggests that while de-escalation is crucial for reducing antibiotic resistance [13], its direct impact on the length of hospital stay might be less pronounced, as initial broad-spectrum antibiotics effectively controlled infections. Comparing our findings with other studies, Sadyrbaeva et al. demonstrated a reduction in median hospital stay by five days associated with carbapenem de-escalation, suggesting that successful de-escalation may indeed lead to shorter hospital stays [22]. Another study by Alshareef et al. demonstrated that successful de-escalation was linked to a notably shorter median hospital length of stay compared to patients who did not undergo de-escalation. Specifically, patients who underwent successful de-escalation had a median hospital stay of 3 days, while those who did not experience de-escalation had a significantly longer median hospital stay of 10 days [2]. Another prospective study showed the length of hospital stay was 5 days shorter in the carbapenem-de-escalated group compared to the carbapenem-non-de-escalated group [21].

During the assessment of ICU admissions, we found that none of the patients, whether de-escalated or not, required ICU admission during their hospital stay. We observed the absence of ICU admissions is a notable finding, suggesting that the patients managed with the de-escalation protocol did not experience severe clinical deterioration necessitating intensive care. This observation could indicate the effectiveness of the de-escalation strategy in managing infections without progressing to critical illness requiring ICU intervention.

Our study found a significant association between microbial culture results and the success of de-escalation practice (p = 0.001). Interestingly, successful de-escalation is much more likely in cases with positive UTI results (86.8%) than negative UTI results (17.2%). This counter intuitive result suggests that de-escalation is more successful in positive UTI cases, potentially because these situations involve switching to a narrower spectrum antibiotic or reducing the number of antibiotics, which may follow clearer clinical guidelines. In contrast, de-escalation in negative culture cases, often involving the complete cessation of antibiotics, might be more challenging for clinicians. Physicians may be more cautious about stopping antibiotics entirely due to concerns about undetected infections or the risk of relapse, leading to a lower success rate in de-escalation when cultures are negative, and some patients may have already taken antibiotics at home before visiting the healthcare facility due to the pain caused by a UTI. This prior use of antibiotics can complicate the clinical decision-making process. A study by William et al. showed that de-escalation of antimicrobial therapy was rare in cases with negative microbiology results. Only one out of 40 reviewed episodes (~ 3%) resulted in a clear de-escalation of antimicrobials following the negative result. This indicates a reluctance or oversight to modify treatment based on negative microbiology findings [4].

These findings have significant clinical implications. The reluctance to de-escalate antimicrobial therapy despite negative microbiology results highlights a potential area for improvement in antibiotic stewardship. It emphasizes the need for enhanced education and more robust Antibiotic Stewardship Programs to guide clinicians in adjusting treatments based on microbiology results. Such improvements would not only enhance patient safety but also promote more sustainable antibiotic practices and reduce unnecessary use of broad-spectrum antibiotics.

Currently, our hospital does not have specific guidelines for the empirical treatment of UTIs; instead, treatment decisions rely on international guidelines. Furthermore, when it comes to carbapenem prescriptions, physicians are permitted to prescribe them for the full course without requiring an infectious disease consultation, as there is no specific protocol or antimicrobial stewardship team overseeing their use.

However, the study’s limitations include its reliance on data from a single tertiary center in Jordan, potentially limiting the generalizability of findings to other settings or populations nationwide and globally. The relatively small sample size may reduce statistical power and limit the ability to detect significant differences or generalize results broadly. Being retrospective, the study lacked patient interviews to ascertain pre-admission antibiotic use, which could influence negative culture results. One limitation of this study is the lack of data on the type of suspected UTI, which may affect the interpretation of the results. Additionally, uncontrolled variables such as illness severity were not fully accounted for, potentially introducing biases that affect study outcomes differently.

Conclusion

This study found that successful carbapenem de-escalation was more likely in patients with urine cultures showing growth compared to those with no growth. While a portion of patients who required de-escalation were successfully de-escalated, many experienced incorrect or failed de-escalation. Early discharge, without proper culture results, complicates appropriate de-escalation, and contributes to unnecessary use of broad-spectrum antibiotics. The results emphasize the importance of urine culture results in guiding antibiotic treatment and highlight the need for improved antibiotic stewardship to optimize de-escalation practices and reduce unnecessary antibiotic use. Future research should focus on developing strategies for better antibiotic de-escalation, minimizing early discharge risks, and improving adherence to stewardship guidelines, ultimately optimizing treatment outcomes and reducing resistance.