Extensively drug resistant Pseudomonas aeruginosa (XDR-PA) is a growing concern due to its increasing incidence, limited therapeutic options, limited data on the optimal treatment, and high mortality rates. Being able to survive in many ecological settings, P. aeruginosa is highly flexible and can adapt to a wide array of environmental pressures. Having access to numerous metabolic pathways and a wide range of virulence and resistance factors thanks to its genetic plasticity, make it one of the most successful bacteria in Medical Microbiology (Behzadi et al., 2022a,b).

Indeed, although in the late 20th century the incidence of P. aeruginosa bloodstream infections was declining, by 2010 it had increased again to 6.5 per 10,000 (Werth et al., 2015). Moreover, an increase in the prevalence of multidrug resistant (MDR) and XDR-PA was evident, with XDR-PA prevalences ranging from 2.6 to 11.2% being described (Walkty et al., 2017; Sader et al., 2018a). Furthermore, according to the European Center for Disease Prevention and Control (ECDC), in 2021 most countries in Europe reported rates of resistance higher than 10% for the majority of the antimicrobial classes under surveillance and only two countries reported less than 5% resistance to carbapenems (European Centre for Disease Prevention and Control and World Health Organization, 2023). In Portugal, in 2021, 12.7% of all isolates were resistant to at least three classes. The highest resistance was seen to fluoroquinolones (18.1%) and piperacillin-tazobactam (16.4%). The most susceptible of the five surveilled classes was aminoglycosides (6.3% resistant) (European Centre for Disease Prevention and Control and World Health Organization, 2023). Moreover, carbapenem-resistant P. aeruginosa is of critical priority in the World Health Organization Priority Pathogens List (Behzadi et al., 2022b).

Even though P. aeruginosa is often a common cause of severe bloodstream infection, ventilator-associated pneumonia and other hospital-acquired infections (Bassetti et al., 2018b), XDR-PA carries an increased mortality. A meta-analysis by Matos et al. describes a mortality of 44.6% among patients infected with MDR P. aeruginosa versus 24.8% in those with non-MDR infections (de Matos et al., 2018). Recio et al. (2018) compared patients with bacteremia due XDR-PA versus susceptible strains and also found a difference in mortality (62.5% versus 30%).

Even though it is an opportunistic pathogen present in the environment, P. aeruginosa is seldom found in the microbiota of healthy humans (Silby et al., 2011; Estepa et al., 2014). Nevertheless, in patients at risk, such as those with a large exposure to the healthcare setting or with certain chronic illnesses (like cystic fibrosis), the colonization rate can reach up to 80% (Gómez-Zorrilla et al., 2014; Ciofu et al., 2015), while those submitted to antibiotic therapy have a higher risk of attaining MDR strains (Gómez-Zorrilla et al., 2014). It also has intrinsic resistance to many antimicrobial drugs and a high capacity of attaining resistance mutations and mobile genetic elements (Bassetti et al., 2018a,b; Horcajada et al., 2019). Although there are many factors contributing to the emergence of antimicrobial resistance, the misuse or overuse of antibiotics is paramount (Algammal et al., 2023). Indeed, despite the new antimicrobials or β-lactam inhibitors that have become commercially available in the last 5 years (Karvouniaris et al., 2023), the emergence of antimicrobial resistance remains one of the greatest threats to global health (Mendes et al., 2022).

Despite the existence of several articles depicting the molecular epidemiology and resistance mechanisms of P. aeruginosa, to the best of the authors’ knowledge there is a gap in the clinical characterization of these patients in Portugal, a country with one of the highest antimicrobial resistance rates in Europe (Pereira et al., 2013, 2015; Hernández-García et al., 2021a,b,c). The authors aim to characterize the population, the outcome and the microbiological characteristics of XDR-PA identified in a tertiary care university hospital center in Portugal.

Over the 3 years considered in this study (2019-2021) a total of 6,514 isolates of P. aeruginosa pertaining to 3,181 patients were identified. Considering XDR-PA, a total of 243 isolates pertaining to 178 episodes concerning 130 patients were found. The prevalence of XDR-PA identification among all P. aeruginosa isolates was 3.7%, corresponding to 4.1% of all patients with identification of this microorganism. A new episode was defined as a positive XDR-PA isolate when obtained more than 14 days after the last one, in the same patient. Of these 178 episodes, 114 (64.0%) were from infections and 64 (36.0%) were from colonizations. Moreover, 69 episodes (48 patients) happened in 2019, 61 (46 patients) occurred in 2020 and 48 (36 patients) in 2021. It is noteworthy that only six patients were co-infected with SARS-CoV-2, of which four of them died.

Extensively drug resistant Pseudomonas aeruginosa (XDR-PA) is a growing concern due to its increasing incidence, limited therapeutic options, limited data on the optimal treatment, and high mortality rates. Even though the criteria used to define XDR-PA did not match exactly those defined by Magiorakos et al. (2012), the prevalence reported in the present study is lower than the one displayed in 2017 by a large-scale Spanish multicenter study (17%) (Del Barrio-Tofinõ et al., 2019). Similarly, a trial including patients with ventilator-associated pneumonia in Spain, Greece and Italy showed a rate of 35.8% XDR isolates, mostly from Greece (Pérez et al., 2019). Other countries around the world also reported higher XDR-PA prevalence, such as 22.1% out of 447 P. aeruginosa isolates in Nepal (Mahto et al., 2021) or 15.5% out of 3248 isolates in Iran (Mirzaei et al., 2020). The prevalence reported in the present study is more similar to that described in Canada and in USA. McCracken et al. (2019) showed 4.5% XDR-PA among all 3864 Canadian isolates between 2007 and 2016, while Sader et al. (2017) depict 9.4% XDR-PA among 7452 American isolates from 2012 to 2015. Although there is a lack in data regarding the prevalence of XDR-PA in Portuguese isolates, ECDC reports a resistance rate to carbapenems of 14.1 and of 12.7% to any combination of at least three of the following antibiotics: piperacillin-tazobactam, ceftazidime, carbapenems, aminoglycosides and fluoroquinolones (Antimicrobial resistance surveillance in Europe 2023 - 2021 data. Stockholm: European Centre for Disease Prevention and Control and World Health Organization, 2023).

While P. aeruginosa is rarely found in the microbiota of healthy humans (Silby et al., 2011; Estepa et al., 2014), it can colonize up to 80% of patients with risk factors, such as a large exposure to the healthcare setting or certain chronic illnesses (such as cystic fibrosis, and solid or hematologic malignancies) (Gómez-Zorrilla et al., 2014; Ciofu et al., 2015). The presence of foreign devices, such as venous or urinary catheters, tracheostomy (especially in children), open abdominal surgery, diabetes, chronic hepatic disorder and end-stage renal disease also increase the risk for P. aeruginosa infection (Varaiya et al., 2008; Willmann et al., 2014; Miller et al., 2016; Bassetti et al., 2018a; Russell et al., 2019; Jean et al., 2020; Körpinar, 2021). Also prior antibiotic therapy (especially fluoroquinolones and carbapenems), so frequent in the healthcare setting and in patients with these comorbidities, is a major risk factor for attaining MDR strains (Falagas et al., 2006; Peña et al., 2012; Gómez-Zorrilla et al., 2014; Bassetti et al., 2018a; Jean et al., 2020). In fact, prior use of fluoroquinolones or carbapenems has been depicted as an independent risk factor for XDR-PA infections (Palavutitotai et al., 2018).

In P. aeruginosa, resistance mechanisms such as decreased permeability, expression of efflux pumps, target modifications and production of inactivating enzymes have all been described (Mesaros et al., 2007). In particular, production of extended-spectrum beta-lactamases and carbapenemases have been described, including metallo-beta-lactamases (Mesaros et al., 2007; Sacha et al., 2008; Ríos et al., 2018; Muddassir et al., 2021). Since these enzymes use zinc, which seems to be a valuable cofactor to hydrolyze beta-lactams, instead of serine in their active sites, they confer resistance to all beta-lactam antibiotics except aztreonam, while not being degraded by currently available beta-lactamase inhibitors (Sacha et al., 2008; Behzadi et al., 2020). In this study there was one VIM-producing XDR-PA that was susceptible only to colistin. Despite VIM not being able to degrade aztreonam, P. aeruginosa is capable of producing simultaneously additional inactivating enzymes, conferring resistance also to this drug (Mesaros et al., 2007; Ríos et al., 2018). So, the combination of aztreonam-avibactam (or ceftazidime-avibactam plus aztreonam if the former is not available) is pertinent as it combines the ability of avibactam to inactivate all serine-beta-lactamases, thus allowing aztreonam to be effective (Marshall et al., 2017; Mischnik et al., 2017). However, due to mechanisms of resistance to aztreonam independent of beta-lactamases, the activity of this combination has to be confirmed with synergy testing (Karakonstantis et al., 2020).

One drug that is often protected by the cross-resistance of other anti-pseudomonal antibiotics is colistin (Mesaros et al., 2007). And while pan-drug resistant isolates have been described, most often XDR-PA remains susceptible to colistin (Souli et al., 2008; Viedma et al., 2009; Giani et al., 2018). In this study all isolates but one were susceptible to colistin. However, while the prevalence of resistance to ceftazidime-avibactam and to ceftolozane-tazobactam appears to be overall low in Europe and the United States, the authors report approximately one quarter of non-susceptible isolates. This might be explained by the geographical variation of the predominance of different mechanisms of resistance (Flamm et al., 2014; Del Barrio-Tofiño et al., 2017; Giani et al., 2018; Livermore et al., 2018; Sader et al., 2018b; Evans et al., 2019), although when considering the cross-resistance of ceftolozane-tazobactam and ceftazidime-avibactam in this study, the authors believe the mechanism not to be predominantly related to carbapenemases. Regarding aztreonam and fosfomycin, the authors report high rates of non-susceptibility, as displayed by other centers and other Portuguese data (Sader et al., 2017; Safaei et al., 2017; Evans et al., 2019; Hernández-García et al., 2021a).

Regarding the analysis of the resistance patterns in XDR-PA isolates, it is not unexpected that the most frequent resistotype is resistance to both aztreonam and fosfomycin. It is noteworthy, however, that resistance patterns that included non-susceptibility to both aztreonam and ceftazidime-avibactam were frequent (40.0%, resistotypes V, VIII, X, and XII). Unfortunately, despite the knowledge that ceftazidime-avibactam resistance is emerging in K. pneumoniae in Portugal since its approval by the national regulatory authority in 2019 (Mendes et al., 2022, 2023), no data was available on synergy testing of ceftazidime-avibactam plus aztreonam or further study of resistance mechanisms in P. aeruginosa.

As XDR-PA is undoubtedly a growing threat with increasingly limited therapeutic options, it is essential to define new and innovative therapeutic strategies, either by developing new systemic drugs, new drug combinations, different methods of drug administration or alternative therapies (Mesaros et al., 2007; Marshall et al., 2017; Mischnik et al., 2017; Bassetti et al., 2018b; Horcajada et al., 2019; Ekkelenkamp et al., 2020; Gaurav et al., 2020; Yao et al., 2021). For example, the synergistic combination colistin-mefloquine has been suggested as a potential future therapeutic option against colistin-resistant strains due to its anti-biofilm activity (Behzadi et al., 2022a). Inhaled therapy has long since been used in cystic fibrosis and it can be a powerful adjunct in the treatment of XDR-PA infections. Since drug concentration in the lung is much higher with inhaled therapy than with the intravenous route, it can be effective even against in vitro resistant strains (Horcajada et al., 2019).

The authors report high mortality rates (35.1%), as observed in previous studies (Samonis et al., 2014; de Matos et al., 2018; Palavutitotai et al., 2018). Patients with hematologic malignancy have been described as being at an increased risk of a poor outcome (Samonis et al., 2014; Tofas et al., 2017; de Matos et al., 2018), perhaps due to their immunosuppressed status with frequent and recurrent infections and antimicrobial use. The authors report a 50.0% (p < 0.05) mortality among patients with hematologic malignancy.

Patients with bacteremia had a worse prognosis (mortality 44.4%, p < 0.05), similar to the 18.0% reported by Dantas et al. (2014) in their comparison between bacteremia due to susceptible and MDR strains. Having been established as an independent risk factor for mortality (Dantas et al., 2014), it surely is a marker for severe disease. Thus, it is interesting to note that in patients presenting with primary bacteremia the mortality was only 25.0%.

Pseudomonas aeruginosa is also a frequent cause of healthcare-associated respiratory and urinary tract infections (Quartin et al., 2013; Lamas Ferreiro et al., 2017; Jean et al., 2020), the most frequent sources of infection as reported in this study and consistent with the most frequent types of Healthcare Associated Infections, as reported by the ECDC (Suetens et al., 2023). Having been established that infections with resistant strains have a greater risk than susceptible strains (de Matos et al., 2018; Recio et al., 2018; Jean et al., 2020) it is not unexpected to find high mortality rates in this study. Nevertheless, 47.2% fatalities among patients with respiratory infections was higher than what had been previously described (Peña et al., 2013; Giaccari et al., 2021). The 26.5% mortality described among patients with urinary tract infections is similar to that described by Lamas Ferreiro et al. (2017).

Adjusting for the antimicrobial used, the authors report a higher mortality rate among colistin-treated patients (39.0%, p < 0.05). Although these results did not achieve statistical significance, it appears that the mortality among patients treated with ceftazidime-avibactam (28.2%) or ceftolozane-tazobactam (18.8%) was lower. Considering exclusively monotherapy, these results are consistent, with 28.6% of patients treated with colistin dying, against 15.8 and 15.4% of patients treated with ceftazidime-avibactam and ceftolozane-tazobactam, respectively. This seems paradoxical since colistin has been described as one of the most active antipseudomonal drugs (Del Barrio-Tofinõ et al., 2019; López Montesinos et al., 2021; Pinilla-Rello et al., 2021). The authors believe that this result might be related to both a better efficacy and safety profile of beta-lactams and a poorer penetration of colistin in the lung, which is supported by the 60.0% mortality when considering only patients with respiratory infection treated with colistin versus 32.3% when considering only other sources of infection. This findings support the last IDSA Guidance, where beta-lactams are put at the forefront of the treatment of MDR P. aeruginosa infections (Tamma et al., 2022). Patients that were submitted to combined directed therapy seemed to have a higher mortality rate than those treated with only one active drug. These results may be related to a possibly increased severity of disease among patients submitted to combined therapy.

The present study has several limitations. First, by using a narrower definition of XDR-PA than previously published, it is possible that other isolates were not considered. Therefore, comparisons with other studies must take this limitation into consideration. Second, several important variables, such as disease severity, exposure to venous catheterization or mechanical ventilation, recent or recurrent history of antibiotic therapy or healthcare exposure, and prior P. aeruginosa colonization status were most often not possible to ascertain.

The results of this study are relevant as they provide a first insight into the clinical outcome of Portuguese patients infected with XDR-PA, allowing for better care between relevant centers. However, in the future, it is essential to improve our understanding and to further characterize these patients, particularly in high-risk environments (such as focusing on Intensive Care Units and Hematology wards). Moreover, there is a crucial need for a current description of the molecular epidemiology and characterization of the resistance mechanisms of P. aeruginosa in Portugal. Furthermore, the integration of clinical data with molecular surveillance and the analysis of genetic resistance determinants may be relevant to improve the quality of care, allowing for a greater accent on research-based recommendations. Finally, with the potential impact of the COVID-19 pandemic on hospital epidemiology, it is paramount to evaluate this issue in the light of new therapeutic options.

In conclusion, patients infected with extensively drug resistant Pseudomonas aeruginosa are difficult to treat, with very limited therapeutic options. XDR-PA is a pathogen responsible for potentially severe disease that can have high mortality. Although the most frequent sources of infection were the respiratory and urinary sites, patients with bacteremia or hematologic malignancy had a higher risk. In this study, colistin was the most susceptible antibiotic in vitro. However, patients treated with ceftazidime-avibactam or ceftolozane-tazobactam appeared to have a better prognosis than those treated with colistin, especially considering those with respiratory infection. To the best of the authors’ knowledge, this is the first study to provide a clinical characterization of patients infected with XDR-PA strains in Portugal.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethical approval was not required for the study involving humans in accordance with the local legislation and institutional requirements. Written informed consent to participate in this study was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and the institutional requirements.

DM: Conceptualization, Formal analysis, Methodology, Writing – original draft. SP: Conceptualization, Methodology, Writing – review & editing. CS: Writing – review & editing. AF: Writing – review & editing. JM: Writing – review & editing. ÁP: Writing – review & editing. CC: Supervision, Writing – review & editing.