1. Introduction
Pseudomonas aeruginosa is the third most commonly identified cause among gram-negative microorganisms causing bloodstream infection (BSI) and carries a very high mortality, higher than that by other gram-negative pathogens and also higher than that of
Staphylococcus aureus, and is estimated to be at 30% after 30 days from the BSI episode [
1,
2,
3,
4]. BSI by
Pseudomonas spp. is a universal problem and is more commonly hospital-acquired [
5,
6,
7,
8,
9]. In specific settings, such as in intensive care units (ICU), the rate of
Pseudomonas spp. isolation as pathogens causing BSIs is even higher. For example, in a study about ICU-acquired BSI,
Pseudomonas aeruginosa was the second most common cause of BSI among gram-negative pathogens [
10]. In another study regarding patients with liver transplantation, 35% of BSIs were due to
P. aeruginosa and had an attributed mortality of 30% [
11]. Commonly described risk factors for bacteremia by
P. aeruginosa include recent antimicrobial use (most commonly involving the last three months), recent hospitalization, severe burns, pancreatobiliary disease, neutropenia or other immunodeficiency (such as hematologic malignancy or transplantation), and presence of an indwelling central venous or urinary catheter [
6,
8,
12,
13,
14,
15,
16].
The coronavirus-disease 2019 (COVID-19) pandemic has caused pressure in healthcare systems, has led to increased rates of hospitalization, increased antimicrobial consumption, and has also led to a higher prevalence of multi-drug-resistant (MDR) pathogens [
17,
18]. This could be associated with increased rates of hospitalization, and at least to some extent, unnecessary antimicrobial use, healthcare personnel exhaustion, and possibly, a relative reduction in the implementation of antimicrobial stewardship practices [
19]. Thus, due to the high mortality associated with
Pseudomonas BSIs and the changing landscape in antimicrobial resistance and incidence of these infections in the post-COVID-19 era, studies evaluating the characteristics of these infections are warranted.
The aim of the present study was to assess the epidemiological and microbiological characteristics of patients with BSI by Pseudomonas spp. in a tertiary hospital, characterize the resistance rates of different Pseudomonas strains to the most clinically relevant antimicrobials, estimate the mortality rate, and identify factors independently associated with mortality.
4. Discussion
The present study investigated the microbiology and antimicrobial resistance of Pseudomonas spp. BSI in a tertiary hospital in Crete, Greece. The most common species was P. aeruginosa, which was also the most resistant to antimicrobials. Isolation of Pseudomonas spp. was stable throughout the study period. Mortality at 30-days after occurrence of BSI was high and patients who died had a higher age, were more likely to have hospital-acquired BSI, and more specifically, were more likely to have acquired the BSI later in their hospitalization, were less likely to be hospitalized in a surgical ward and more likely to be hospitalized in the ICU, were more likely to have a longer duration of hospitalization, and were also more likely to have a BSI by P. aeruginosa as well as increased antimicrobial resistance. Increased age, ICU-acquisition of BSI, and more days in the hospital when positive blood culture was drawn were independently positively associated with 30-day mortality.
BSIs by
Pseudomonas spp. and more specifically, by
P. aeruginosa, are commonly encountered in clinical practice, especially in patients with risk factors for acquisition of this pathogens, such as advanced age, neutropenia or other immunodeficiency (e.g., malignancy or transplantation), indwelling central venous or urinary catheter, severe burns, pancreatobiliary disease, previous receipt of antimicrobials, and previous or current hospitalization [
5,
6,
7,
8,
9,
12,
16,
23]. The clinical presentation includes the typical features noted for all patients with gram-negative sepsis, such as fever, tachycardia and tachypnea, hypotension, respiratory failure, and disorientation, even though these findings may also occur in cases of sepsis by other organisms as well, and they are not specific exclusively for gram-negative sepsis. The most common sources of
P. aeruginosa sepsis and bacteremia include the respiratory tract, infected intravascular catheters, the gastrointestinal and hepatobiliary tract, the urinary tract, and the skin and soft tissues. However, in up to 40% of cases, the source of bacteremia is not evident [
7,
24,
25].
In the present study, the most common species, as anticipated, was
P. aeruginosa, which is the most clinically relevant species causing infection, and more specifically BSIs, in humans. Indeed, most studies on BSIs refer exclusively to
P. aeruginosa so that it may be hard to compare data on non-
aeruginosa species to others in the literature. A study referring to non-
aeruginosa Pseudomonas species in respiratory samples of patients with cystic fibrosis identified
P. fluorescens,
P. putida, and
P. stutzeri as the most common species with 33, 18, and 6 isolated strains, respectively, followed by
P. alcaligenes,
P. fragi,
P. mendocina,
P. nitroreducens,
P. oleovorans,
P, oryzihabitans, and
P, veronii, each with one isolated stain [
26]. In the same study, the clinical importance of the non-
aeruginosa Pseudomonas species is discussed, suggesting that some of them may be of low pathogenicity. In the present study, the most commonly isolated species beyond
P. aeruginosa were
P. putida,
P. oryzihabitans, and
P. stutzeri. An older study evaluating non-
aeruginosa Pseudomonas species as causes of BSI identified a small number of such species and discusses the clinical relevance their isolation may have, since most patients with such species did not receive antimicrobial treatment, with no mortality noted [
27]. In the present study, even though there were no data regarding antimicrobial treatment of the patients, a statistically significant association between isolation of
P. aeruginosa and 30-day mortality was noted, even though this was not identified as an independent factor associated with 30-day mortality in a multivariate logistic-regression analysis model.
Even though the number of
Pseudomonas isolates identified in blood cultures remained stable through the years, there was a statistically significant reduction of
P. aeruginosa strains in the post-COVID-19 era. Infection by non-
aeruginosa Pseudomonas species is relatively uncommon and it mostly occurs in patients with underlying immunocompromise, or in association with an infected medical device [
28,
29,
30].
Antimicrobial resistance among gram-negative microorganisms is a problem of increasing magnitude in medicine, associated with increased morbidity, mortality, and health costs [
31,
32]. Currently, for many pathogens, available antimicrobial treatments are limited, and choices include revived antibiotics or combinations of antimicrobials [
33,
34,
35,
36,
37].
P. aeruginosa has many different mechanisms of antimicrobial resistance that may include production of antibiotic-inactivating enzymes, efflux systems, and modification of the permeability of the outer membrane [
38]. In the present study, regarding antimicrobial resistance,
P. aeruginosa has the highest rates of resistance to clinically relevant antimicrobials with anti-pseudomonal activity, such as piperacillin, ciprofloxacin, and cefepime. However, antimicrobial resistance to old antibiotics such as colistin is still low, meaning that even in cases of BSI by
P. aeruginosa with multiple resistance mechanisms, there are still available options in terms of treatment. Nowadays, newer antimicrobials and antimicrobial combinations, such as the combination of ceftolozane with tazobactam, the combination of ceftazidime with avibactam, cefiderocol, the combination of imipenem with cilastatin and relebactam, or the combination of aztreonam with avibactam have emerged as important therapeutic options that should be preferred for the treatment of BSI by sensitive microorganisms such as
Pseudomonas species [
21,
39,
40,
41]. In the present study, these newer antimicrobials were not tested; thus, no result can be drawn regarding their susceptibility in the isolated strains. Interestingly, other species such as
P. putida also have high rates of antimicrobial resistance to the same antimicrobials. This is in line with studies showing that, even though non-
aeruginosa Pseudomonas species may have been previously considered to have low pathogenicity, some of them, such as
P. putida, are causing significant mortality in cases of BSI, and also have important antimicrobial resistance [
30,
42]. For example,
P. putida has been identified as a nosocomial cause of infection with multi-drug-resistance and production of VIM-2 metallo-beta-lactamase, probably by independent horizontal transfer of resistance genes [
43,
44].
In the present study, antimicrobial resistance of
P. aeruginosa seemed to have remained stable throughout the study period when studying each antimicrobial individually, with the exception of gentamicin and tobramycin, as, in the post-COVID-19 era,
P. aeruginosa had less resistance against these two aminoglycosides. Interestingly, rates of MDR, XDR, and DTR were lower in patients with BSI by
P. aeruginosa in the post-COVID-19 era. However, it is of note that, during the study period, a carbapenem-focused antimicrobial stewardship program was implemented and led to a reduction of use of carbapenems through unsolicited consultations of infectious disease physicians; this may have affected the time-course of development of antimicrobial resistance, by halting a possibly increasing trend [
45,
46]. This is of utmost importance since it increases the common understanding that implementation of antimicrobial stewardship interventions may lead to reduction of antimicrobial use, reduction of hospital costs, and reduction of antimicrobial resistance without an increase in mortality [
47]. In some other studies, antimicrobial resistance of
P. aeruginosa strains also remained stable during the COVID-19 pandemic [
48,
49,
50]. More specifically, in some studies, an increase in infections by carbapenem-susceptible
P. aeruginosa was noted, which could theoretically be associated with a higher consumption of third generation cephalosporins during the first wave of the COVID-19 pandemic [
49]. In other studies, however, trends for antimicrobial resistance of
P. aeruginosa was rising during the COVID-19 pandemic [
51,
52].
Interestingly, for all antimicrobials tested, with the single exception of colistin, there was a statistically significant association between antimicrobial resistance and 30-day mortality, while MDR, XDR, and DTR
P. aeruginosa was more frequently isolated among patients with BSI who died within 30-days from the positive blood culture. However, this could have been a result of the higher likelihood of ICU-acquisition of
Pseudomonas BSI among patients who died, as there was a non-statistically significant trend for higher antimicrobial resistance of
P. aeruginosa strains isolated in the ICU. To that end, a multivariate logistic-regression analysis identified ICU-acquisition of
Pseudomonas BSI to be independently associated with 30-day mortality, while antimicrobial resistance was not. Notably, the fact that diagnosis of the BSI in the ICU is associated with higher mortality is probably an epiphenomenon of the more severe clinical condition of the patients diagnosed in this setting. Indeed, patients admitted to the ICU have higher severity of illness and higher mortality when diagnosed with BSI [
53]. However, since the present study is based mostly on microbiological data, a direct association of BSI mortality with the severity of the underlying illness addressed with clinical severity scores such as SOFA or APACHE could not have been performed.
The multivariate logistic-regression analysis identified also increased age and increased length of hospitalization before drawing the positive blood culture for
Pseudomonas spp. to be independently associated with 30-day mortality. Increased age is well known to be associated with increased mortality in studies of hospitalized patients in general; thus, this finding was not a surprise [
54]. On the other hand, length of hospitalization before the positive blood culture could also be an epiphenomenon, since length of hospitalization may also be a factor that could be theoretically associated with the clinical condition of the patient, as patients with more complex medical or surgical problems may have a need for more prolonged hospitalization and may have higher mortality, especially in the case of BSI by drug-resistant gram-negative pathogens [
55]. However, the present study did not focus on the clinical characteristics of the patients; thus, this presumed association of length of hospitalization before drawing a blood culture in patients with BSI by
Pseudomonas spp. and the severity of their clinical condition has not been confirmed in the present study. In other studies, though, prolonged hospitalization in general has also been associated with mortality [
54].
This study has some limitations that should be acknowledged. First, it is a single-center study; thus, the results should be read cautiously as they represent the microbiology and antimicrobial resistance patterns of a specific area. Furthermore, as this study mostly included microbiological data, there are no data regarding patients’ clinical characteristics or treatment. For example, the criterion of hospital-acquired bacteremia was whether the blood culture was drawn after 48 h from admission, which may not necessarily coincide with the most appropriate definition of development of symptoms of infection after 48 h from the admission. Moreover, since clinical data were not collected, whether BSI episodes were primary or secondary was not addressed in the present study. On the other hand, clinical and laboratory data allowing estimation of the severity of the underlying disease through calculation of SOFA or APACHE scores was not possible in the present study. Moreover, the antimicrobials tested are not all that are available; thus, newer antimicrobial combinations, such as that of ceftazidime/avibactam or ceftolozane/tazobactam have not been included in the analysis. Finally, some of the Pseudomonas species were very few; thus, the data regarding antimicrobial resistance may not be reliable enough to draw safe conclusions.