Bacteremic versus Non-Bacteremic Urinary Tract Infections: Predictors of Poor Clinical Outcome

Urinary tract infections (UTIs) are among the most common infections in clinical practice, representing a significant burden in healthcare systems [1, 2]. They are strongly associated with sepsis, hospitalization, longer length of stay (LOS), and mortality in patients presenting to the emergency department [3]. Complicated UTIs are particularly challenging due to increasing resistance rates of uropathogens including E.coli [4, 5, 6] Contributing factors include misuse and overuse of broad-spectrum antibiotics, prolonged treatment [7] and inappropriate treatment of asymptomatic bacteriuria despite existing guidelines [8]. Antimicrobial resistance (AMR) not only complicates treatment options but also increases the risk of recurrent episodes and future resistant infections [9].

UTIs appear to be the most frequent source of secondary bacteremia [10], observed in up to 31% of acute pyelonephritis cases [11] and in 38% to 69% of instances involving severe sepsis or septic shock [12]. They account for approximately 24% of nosocomial secondary bloodstream infections [13]. Gram-negative bacteria, especially Enterobacterales, are the most prevalent pathogens in these cases [13], posing an increasing challenge due to emerging antibiotic resistance, affecting both nosocomial and community-acquired UTIs [14, 15].

Although previous research has explored the impact and economic burden of UTIs and bloodstream infections separately [16, 17], data evaluating the impact of bacteremia on the outcomes of UTIs remain scarce [18, 19]. This study aims to provide further insights into the influence of secondary bacteremia on the clinical outcomes of UTIs, identify potential predictive factors for the presence of bacteremia, and assess predictors of poor outcomes.

This study compared UTIs with and without bacteremia, revealing that patients with bacteremia exhibited more severe illness, poorer prognosis, higher severity scores, and more abnormal laboratory findings. Gram-negative pathogens were the predominant causative agents. Secondary bacteremia was associated with increased mortality and longer hospital stays. MDR status independently predicted for poorer outcomes as this reflected for need for antibiotic change, surgical intervention or death, despite similar resistance patterns across groups.

From our perspective, patients in the bacteremia group presented in a more deteriorated clinical condition than those without bacteremia, as reflected by higher qSOFA, APACHE, and SAP scores. Moreover, inflammatory markers (CRP) were significantly elevated in the bacteremia group, while surrogate indices of severe disease including Hematocrit (Ht) and PLT level were more severely affected. This comes as no surprise as bacteremia in sepsis patients is associated with higher mortality rates and longer Intensive Care Unit (ICU) stays compared to non-bacteremic infections [20].

Patients with bacteremia exhibit a more severe systemic inflammatory response, meeting more systemic inflammatory response syndrome (SIRS) criteria and showing elevated levels of inflammatory markers, SOFA scores and a greater need for intensive care [21]. The inflammatory mediator profile differs between sepsis patients with and without bacteremia, with significant differences in plasma levels of cytokines, adhesion molecules, and tissue inhibitors of metalloproteinases [22]. Given the significantly worse clinical scores and laboratory parameters observed in the bacteremia group, our findings clearly support the notion that bacteremia in UTIs reflects greater disease severity.

As already shown, we confirmed the presence of E.coli as the most commonly isolated pathogen [23, 24], followed by K. pneumoniae and P.aeruginosa. However, the presence of bacteria such as S.aureus and Enterococcus species, particularly in the bacteremia group, highlights the need to consider antibiotics effective against Gram-positive strains in patients with relevant risk factors.

In our study, the presence of an MDR pathogen was observed with a frequency of 23.1% in the non-bacteremia group and 25% in the bacteremia group. The absence of significant differences in resistance patterns between bacteremic and non-bacteremic UTIs in our study could potentially be explained by the overall high rates of antimicrobial resistance in Greece. Given the endemic presence of MDR pathogens in the region, resistance patterns may be similarly distributed across different patient groups, regardless of bacteremia status. Drug resistance is a major concern phenomenon with significant mortality forcing the identification of newer antibiotics and the global guidelines to rapidly update their strategies for the treatment of antimicrobial-resistant Gram-negative infections [25]. MDR prevalence in Greece is reported alarmingly high, as evidenced by previous and current studies [26] and a recent report by the European Centre for Disease Prevention and Control (ECDC) [27]. Since the late 2000s, Greece has faced an endemic challenge of MDR pathogens in hospitals, primarily due to carbapenem-resistant Gram-negative bacilli, consistently recording some of the highest antimicrobial resistance rates in Europe [27]. Our study, conducted on patient files before the COVID-19 pandemic, suggested that resistance trends will likely continue to escalate, consistent with local epidemiological patterns [28]. The increase of MDR pathogens is also confirmed by the UTILY study, which reported a 39.1% prevalence of MDR pathogens in patients admitted to emergency departments for UTIs, including a 21% prevalence among those without any known risk factors [29]. In special populations such as individuals with diabetes mellitus the prevalence of MDR pathogens can rise up to 58.4% [30]. Importantly, our analysis showed that MDR pathogens remained a significant independent factor for worse outcomes, underscoring MDR status as a key predictor of poor prognosis. The fact that MDR pathogens remain an independent predictor for adverse outcome in our study with an OR of 7.792 aligns with the results of other studies such as the SERPENS study where patients with MDR pathogens are at higher risk of poor outcomes [31]. However, the reason for the increased mortality risk associated with resistant pathogens remains under discussion since, many authors primarily attribute it to the initial use of inappropriate antibiotic therapy, rather than increased virulence of the resistant organisms themselves [32, 33, 34].

Significant differences in antibiotic regimens were found between groups. The use of carbapenems (38.7% vs. 14.3%, p=0.002), amikacin (25.8% vs. 9%, p=0.01), and vancomycin (12.9% vs. 1.5%, p=0.002) was significantly higher in bacteremic patients, while no significant differences were observed for ceftriaxone, piperacillin/tazobactam, quinolones, ampicillin/sulbactam, or other broad spectrum antibiotics. This increased use of broad-spectrum antibiotics was statistically significant and likely driven by the severity of infections. Meropenem remained the preferred option for the treatment of complicated urinary tract infections, as it exhibits a high test-of-cure rate of 91.4%, demonstrating superior efficacy among treatment options, according to a systematic review by Sivanandy et al [35]. Attending physicians seemed to prefer the use of carbapenems and aminoglycosides in bacteremia patients, often presenting in more severe condition, in line with their bacteriocidal vs bacteriostatic properties of other regimens e.g. quinolones [36]. This, in combination with the fact that MDR profile has been associated with prior use of antibiotics [37] underscores the importance of implementing antimicrobial stewardship programs (ASPs) in hospital wards, along with strategies like de-escalation in microbiologically documented infections, to minimize unnecessary antibiotic exposure and curb the emergence of multidrug-resistant pathogens [38].

Complicated UTIs are associated with a varied range of bacterial pathogens and an elevated risk of clinical complications [39]. Specific conditions classified under complicated UTIs, such as urolithiasis, serve as significant risk factors for the development of septic shock in older patients [40], which is strongly associated with increased mortality [41]. However, in our study, the prevalence of complicated infections did not statistically differ between the bacteremia and non-bacteremia groups. Moreover, complicated UTIs were not identified as a factor significantly correlated with poor outcomes in our analysis, suggesting that the impact of complicated UTIs on prognosis may depend on additional factors beyond their classification alone. Of note. on this parameter the authors do acknowledge the limitation of missing file data.

Our study has several limitations. Being a single-center study limits generalizability to other hospitals with different patient demographics, resources, or infection control protocols. However, managing UTIs complicated by bacteremia remains a global concern, and our findings are relevant to similar healthcare settings, especially those with high antibiotic resistance. Although the data were collected between 2016 and 2018, the evaluated antimicrobial regimens remain relevant, as current guidelines continue to recommend these agents. However, resistance patterns and microbiological epidemiology may have changed substantially over time, even within the same institution, which could limit the generalizability of our findings to the current clinical context.. While molecular sequencing methods were not available at the time, comparing historical and current data helps identify patterns, assess progress, and highlight persistent gaps. Fundamental patient characteristics such as age, sex, and comorbidities continue to influence outcomes despite evolving resistance patterns. As a retrospective study, some clinical and laboratory data were missing, potentially affecting accuracy and the strength of associations in our multivariate analysis. Our limited sample size restricted the number of variables we could examine, and wide confidence intervals, particularly for hospital-acquired infections and MDR-status as predictors, indicate model instability. Moreover, in our study, the primary outcome was defined as a composite of death, need for surgical intervention, or modification of antibiotic therapy, rather than mortality alone. As a result, the percentages and statistical analyses reported for outcomes differ from those that would be obtained if mortality were analyzed as a separate endpoint. This heterogeneity among outcome components, which vary considerably in clinical severity, likely contributed to a greater variance in the results and may have diluted the strength of associations observed. As we performed an analysis based on mortality as the primary outcome (death vs. survival), the small proportion of deaths in our cohort may have contributed to instability in the regression analysis. Consequently, some odds ratios showed wide confidence intervals, reflecting limited statistical power and preventing firm conclusions regarding independent predictors of mortality. Our study was limited by the retrospective nature of data collection, which did not allow us to capture information on pre-hospital antibiotic exposure. As a result, we cannot determine whether some negative cultures were influenced by prior antimicrobial therapy, potentially leading to an overestimation of culture-negative cases. This missing information represents an important limitation, as pre-admission antibiotics can alter pathogen detection and affect the assessment of resistance patterns. Despite these limitations, our study offers valuable insights into secondary bacteremia in UTIs, supporting clinical decision-making and underscoring the need for larger, multicenter research.

Bacteremia in UTIs leads to worse outcomes Early detection through blood cultures, urinalysis, and biomarkers is critical. Prompt, appropriate antimicrobial therapy—especially for MDR pathogens—guided by Infectious Disease specialists, along with careful monitoring and stewardship practices, improves outcomes, reduces complications, and optimizes patient care.