Patients with PA BSI who have underlying or immunosuppressive diseases have poor clinical outcomes such as increased mortality, prolonged hospitalization, etc. A previous study showed that 41.8% and 25.6% of patients were admitted to the ICU because of a BSI caused by PA or another pathogen within 90 days prior to BSI onset, respectively[6]. In this study, we found that 19.5% of the patients originated from the ICU. Most PA isolates came from the ICU, with possible reasons for the high risk of infection due to the long length of stay in the ICU, the prolonged use of antibiotics in the bowel, and the fact that patients admitted to the ICU are more likely to receive invasive therapies[16]. Hospitalized patients usually have a history of antibiotic exposure (in our study, 50.0% of the patients had a history of antibiotic use within 30 days prior to BSI onset). Previous exposure to antibiotics can lead to drug resistance. Many factors may contribute to the increase in antibiotic resistance, such as the overexpression of efflux pumps, the reduced permeability of the bacterial outer membrane, and the production of drug inactivating enzymes[17],[18]. We found that the majority of PA-infected patients were male, probably due to the large proportion of males who smoke, as tobacco smoke can cause adverse immune responses that can lead to lung lesions and various infectious diseases[19]. In this case, the patient's immune function declines, leading to an increased chance of bacterial invasion in the blood. In previous studies of PA BSI, patients with BSI caused by drug-resistant PA, especially imipenem-resistance, had a higher mortality rate than patients infected with susceptible strains[9]. This was not observed in our study, but the resistance and mortality rates of patients in this study were low, possibly because of the following reasons: First, few patients were infected with carbapenemase-producing PA. Second, a review of the mechanism of resistance to PA indicated that the decreased sensitivity of PA to carbapenems was associated with the loss of the outer membrane, OprD[20]. PA resistance to carbapenems (imipenem) is mainly OprD-mediated. This mechanism is different than that of carbapenem resistance in K. pneumoniae, which is resistant to all β-Lactam antibiotics after it is produced[21],[22]. This finding suggests that PA may be susceptible to antibiotics other than imipenem, with a larger variety of available antibiotics, better treatment outcomes, and a low mortality rate. Finally, the sample size included in this study was small. However, both MDR and carbapenem resistance were higher in the group with high SOFA scores than in the low SOFA score group. Our findings indicate that patients with higher SOFA scores had more severe symptoms, may require longer hospital stays, had prolonged antibiotic use, and were more likely to develop resistance.
In our study, SOFA score and previous antibiotic exposure were identified as risk factors for 30-day mortality. At the same time, the 14-day treatment failure and 30-day mortality rate of infected patients with SOFA score ≥ 4 were higher than those with SOFA score <4. There is a good relationship between increased SOFA score and mortality[23]. SOFA score is an indicator of multiple organ dysfunction and is considered an important tool for defining the clinical status of individual patients and treatment response in clinical trials. Similarly, a study showed that SOFA score (≥ 10) is an independent risk factor associated with failure of the first definitive treatment for PA BSI[24].
Combination therapy for PA BSI did not reduce patient mortality, improve clinical failure rates, or affect development of drug resistance[25]. However, another study demonstrated lower mortality in patients treated with ciprofloxacin combined with definitive therapy compared to those treated with monotherapy[26]. Scott et al. suggested that inappropriate initial empirical antimicrobial therapy may increase mortality in PA BSI patients[27]. The association between inappropriate initial empirical therapy and mortality from BSI has not been examined at this time, however, we found that empirical imipenem treatment increased the 14-day treatment failure rate of patients regardless of SOFA score. Empirical imipenem treatment was a risk factor for 14-day treatment failure in both SOFA score groups. A possible reason for this is that antimicrobials binding to serum proteins can affect their antibacterial effects; patients can show different pathophysiological characteristics, while clinical use is usually recommended according to the standard dosage instead of individualized treatment, resulting in a possible low drug concentration in some patients and clinical treatment failure[28]. In addition, patients under this treatment modality are prone to develop resistance and, thus, clinical treatment failure. Bricheux et al. found that lower plasma imipenem concentrations were associated with a higher clinical failure rate despite its therapeutic effect[29].
A multicenter retrospective study showed that there was no difference in PA mortality using ceftazidime, carbapenems, and piperacillin tazobactam as definitive treatments, and ceftazidime or piperacillin tazobactam is recommended for susceptible infections because of the high rate of resistance to carbapenems in patients receiving carbapenems[30]. Another meta-analysis showed no difference in sepsis mortality between carbapenems and BLICs treatment[31]. We assessed 30-day mortality with single antibiotic regimens versus either BLICs or carbapenems, and among the 174 patients, 119 were treated only with BLICs or carbapenems, including 48 treated with carbapenems and 71 treated with BLICs. We found no significant difference in the clinical outcomes of patients treated with BLICs and carbapenems (P=0.874). The European Society of Clinical Microbiology and Infectious Diseases (ESCMID) suggested the use of old antibiotics with in vitro activity in patients with low-risk carbapenem resistance Pseudomonas aeruginosa[32]. The IDSA recommends the use of traditional non-carbapenem, β-lactam drugs when PA isolates are susceptible to them[33]. In this study, PA showed the highest susceptibility to aminoglycosides, followed by third-generation cephalosporins, and quinolones. We recommend the use of aminoglycosides, third-generation cephalosporins, and quinolones for the treatment of PA BSIs. In critically ill patients, we recommend the use of β-lactamase inhibitors. Given the nephrotoxicity of aminoglycosides, their use should be avoided in immunosuppressed patients, patients with renal dysfunction, and pregnant women[34].
We also found that, although the median Pitt score was equal between the two groups divided by SOFA score (≥ 4 and < 4), the group with SOFA scores ≥ 4 had a higher incidence of mortality or 14-day treatment failure compared to the other group (P < 0.001). This suggests that a high SOFA score may be an important factor that affects the clinical outcome of patients with PA BSI.
This was a single-center retrospective study, and the results may be influenced by sample size and local epidemiology. This study suggests that the change in SOFA score may be crucial for patients hospitalized with infection and requires increased attention from clinicians. There is currently debate regarding the optimal treatment for PA BSI, and the emergence and spread of antibiotic resistance are associated with many epidemiological risk factors. Therefore, it is essential to detect pathogens and investigate their resistance mechanisms, epidemiology, and alternative treatments. Clinicians should consider their potential toxicity and the emergence of resistance when using antimicrobial therapy empirically and appropriate individualized treatment.