An Evidence-Based Multidisciplinary Approach Focused on Creating Algorithms for Targeted Therapy of Infection-Related Ventilator-Associated Complications (IVACs) Caused by Pseudomonas aeruginosa and Acinetobacter baumannii in Critically Ill Adult Patients

(1) Background: To develop evidence-based algorithms for targeted antibiotic therapy of infection-related ventilator-associated complications (IVACs) caused by non-fermenting Gram-negative pathogens. (2) Methods: A multidisciplinary team of four experts had several rounds of assessments for developing algorithms devoted to targeted antimicrobial therapy of IVACs caused by two non-fermenting Gram-negative pathogens. A literature search was performed on PubMed-MEDLINE (until September 2021) to provide evidence for supporting therapeutic choices. Quality and strength of evidence was established according to a hierarchical scale of the study design. Six different algorithms with associated recommendations in terms of therapeutic choice and dosing optimization were suggested according to the susceptibility pattern of two non-fermenting Gram-negative pathogens: multi-susceptible Pseudomonas aeruginosa (PA), multidrug-resistant (MDR) metallo-beta-lactamase (MBL)-negative-PA, MBL-positive-PA, carbapenem-susceptible Acinetobacter baumannii (AB), and carbapenem-resistant AB. (3) Results: Piperacillin–tazobactam or fourth-generation cephalosporins represent the first therapeutic choice in IVACs caused by multi-susceptible PA. A carbapenem-sparing approach favouring the administration of novel beta-lactam/beta-lactamase inhibitors should be pursued in the management of MDR-MBL-negative PA infections. Cefiderocol should be used as first-line therapy for the management of IVACs caused by MBL-producing-PA or carbapenem-resistant AB. Fosfomycin-based combination therapy, as well as inhaled colistin, could be considered as a reasonable alternative for the management of IVACs due to MDR-PA and carbapenem-resistant AB. (4) Conclusions: The implementation of algorithms focused on prompt revision of antibiotic regimens guided by results of conventional and rapid diagnostic methodologies, appropriate place in therapy of novel beta-lactams, implementation of strategies for sparing the broadest-spectrum antibiotics, and pharmacokinetic/pharmacodynamic optimization of antibiotic dosing regimens is strongly suggested.


Introduction
Infection-related ventilator-associated complications (IVACs) represent the most prevalent infective events in patients admitted to the intensive care unit (ICU) and requiring mechanical ventilation [1], accounting approximately for one-third of hospital-acquired pneumonia (HAP) cases [2]. IVACs are associated with high mortality rate (over 50%) and with a remarkable impact on length of ICU stay, antibiotic use, and overall health care costs [2][3][4][5]. Gram-negative pathogens are responsible for the majority of HAP and ventilator-associated pneumonia (VAP), and among these, the non-fermenting Gram-negative pathogens (especially Pseudomonas aeruginosa and Acinetobacter baumannii) are responsible for a remarkable amount of IVACs in critically ill patients, second only to Staphylococcus aureus in terms of prevalence [6,7].
Pseudomonas aeruginosa and Acinetobacter baumannii are both characterized by innate resistance mechanisms against multiple antimicrobials. Furthermore, these pathogens may easily acquire new resistances by different mobile elements, thus making extremely challenging the choice of appropriate antibiotic therapy in this setting [6][7][8][9]. Prompt initiation of empirical broad-spectrum antibiotic treatment is necessary in critically ill patients with suspected IVAC, including agents with activity against non-fermenting Gram-negative pathogens according to the existence of specific risk factors or in case of epidemiological settings characterized by high prevalence. In the last report on antimicrobial surveillance issued by the European Center for Disease Prevention and Control (ECDC) [10], the proportion of invasive isolates of multidrug resistant (MDR) Pseudomonas aeruginosa and carbapenem-resistant Acinetobacter baumannii accounted for 12.1% and 32.6%, respectively. In the Italian context, the proportion of the invasive strains of MDR Pseudomonas aeruginosa is in line with European data (13.1%), while that of carbapenem-resistant Acinetobacter baumannii isolates accounts for as much as 80% [10]. However, once that the causative pathogen has been identified and its susceptibility pattern has been defined, therapy should be revised and targeted, as recommended by the Surviving Sepsis Campaign guidelines [11].
Microbiological confirmation of IVAC represents a crucial step for enabling targeted antibiotic therapy in critically ill patients. Unfortunately, traditional culture-based methods are time-consuming as they often require at least 1-2 days for pathogen isolation and an additional day for determination of the antibiotic susceptibility pattern. Furthermore, culture-based methods exhibit low sensitivity, as they may provide pathogen identification in less than 40% of patients with clinically diagnosed IVACs [12]. This has traditionally made challenging and often difficult a rapid implementation of targeted antibiotic therapy [10]. In recent years, the development of rapid molecular tests has revolutionized microbiological diagnosis in the pneumonia setting. These tests, based on syndromic panels, may provide fast identification of the respiratory pathogens coupled with detection of relevant genotypic markers of resistance. This approach is expected to have a huge impact in daily clinical practice, by making more and more applicable both targeted antibiotic therapy and antimicrobial stewardship strategies [12,13].
Appropriate targeted antimicrobial therapy, coupled with antibiotic dose optimization and implementation of antimicrobial stewardship programs, could simultaneously maximize efficacy, reduce antibiotic overconsumption, and minimize the development of resistance in ICU patients [14]. This could be achieved by means of the coordinated approach of a multidisciplinary task force, composed by the intensive care physician, the infectious disease consultant, the clinical microbiologist, and the MD or PharmD clinical pharmacologist. Early diagnosis with prompt identification of the causative pathogen of IVACs coupled with targeted antibiotic therapy based on antimicrobial susceptibility testing and on dosing optimization according to the pharmacokinetic/pharmacodynamic (PK/PD) concepts and the therapeutic drug monitoring (TDM)-guided approach may be fundamental for this purpose.
This multidisciplinary opinion article aims to develop evidence-based algorithms for targeted antibiotic therapy of IVACs caused by two non-fermenting Gram-negative pathogens in critically ill adult patients. The objective is to provide a useful guidance for intensive care physicians either in appropriately placing novel antimicrobial agents in lack of definitive evidence or in considering antimicrobial stewardship strategies for possibly sparing the broadest-spectrum antibiotics.

Targeted Treatment of IVACs Caused by Pseudomonas aeruginosa in Critically Ill Adult Patients
Therapeutic algorithm for targeted treatment of IVACs caused by Pseudomonas aeruginosa in adult ICU patients is shown in Figure 1.

Multi-Susceptible Pseudomonas aeruginosa
Continuous infusion piperacillin-tazobactam (18 g/day after 6.75-9 g loading dose [LD]) is recommended as target therapy for the management of IVACs caused by multisusceptible Pseudomonas aeruginosa. Ceftazidime (6-8 g/day after 2 g LD) or cefepime (2 g LD followed by 6 g/day) in continuous infusion should be reserved as second-line alternatives in case of clinical or microbiological failure (Figure 1, panel A.1). ATS/IDSA guidelines [15] recommended the use of piperacillin-tazobactam, ceftazidime or cefepime as definitive therapy of HAP or VAP requiring a coverage on Pseudomonas aeruginosa with demonstrated susceptibility, suggesting the administration in EI or CI to maximize lung exposure. A summary of the studies evaluating the efficacy of piperacillin-tazobactam or third/fourth-generation cephalosporins in patients affected by IVACs caused by multisusceptible Pseudomonas aeruginosa is provided in Table 1. Table 1. Summary of the studies investigating the treatment of multi-susceptible Pseudomonas aeruginosa infection-related ventilator-associated complications (IVACs) with piperacillin-tazobactam or fourth-generation cephalosporins. Jaccard   Two different RCTs investigated the efficacy of piperacillin-tazobactam in the management of IVACs caused by Pseudomonas aeruginosa [16,17]. Jaccard et al. [16] randomized 371 patients affected by nosocomial infections (49.2% HAP, of which 28% caused by Pseudomonas aeruginosa) to piperacillin-tazobactam or imipenem-cilastatin. Although no difference in clinical failure rate (17.0% vs. 29.0%; p = 0.09) and mortality rate (8% vs. 9%; p = 0.78) was found between the two agents in HAP subgroup, a significantly lower clinical failure rate was found in patients affected by HAP due to Pseudomonas aeruginosa receiving piperacillin-tazobactam compared to imipenem-cilastatin (10.0% vs. 50.0%; p = 0.004). Notably, a significantly higher rate of resistance development was reported with imipenem-cilastatin compared to piperacillin-tazobactam in HAP due to Pseudomonas aeruginosa. Joshi et al. [17] randomized 300 patients (87% HAP, of which 7.7% due to Pseudomonas aeruginosa), reporting a significantly higher clinical cure rate in 155 patients receiving piperacillin-tazobactam associated with tobramycin compared to 145 patients treated with ceftazidime plus tobramycin (74.2% vs. 57.9%; p = 0.004). Furthermore, a trend to higher microbiological eradication was found in Pseudomonas aeruginosa subgroup with piperacillin-tazobactam (67% vs. 30%; p = 0.19). Babich et al. [18] retrospectively evaluated clinical outcome in 767 patients receiving definitive monotherapy for the treatment of BSI due to Pseudomonas aeruginosa. No difference in 30-day mortality rate emerged between ceftazidime (OR 1.14; 95% CI 0.52-2.46) or piperacillin-tazobactam (OR 1.30; 95% CI 0.67-2.51) compared to carbapenems (meropenem or imipenem) at propensity score analysis, while a higher development of new resistance in Pseudomonas aeruginosa isolates occurred in patients treated with carbapenems compared to ceftazidime or piperacillin-tazobactam (17.5% vs. 12.4% vs. 8.4%; p = 0.007). However, no subgroup analysis according to IVACs was performed.
Several observational studies recently investigated the role of ceftazidime and cefepime in the management of Pseudomonas aeruginosa BSIs [18][19][20]. Rate of IVACs ranged from 14.7% to 30% and ICU admission was reported in up to 32.2% of cases. Conflicting findings emerged in retrospective studies performing a subgroup analysis according to MIC of Pseudomonas aeruginosa for ceftazidime or cefepime [19,20]. Su et al. [19] found that a MIC for cefepime ≥ 4 mg/L may predict an unfavourable outcome among patients with serious infections due to Pseudomonas aeruginosa, including 30% of HAP/VAP. Conversely, Ratliff et al. [20] found no difference in mortality rate between patients exhibiting a MIC for ceftazidime or cefepime ≤ 2 mg/L compared to those with a MIC > 2 mg/L in Pseudomonas aeruginosa BSIs (17.2% vs. 27.6%; p = 0.34), although subgroup analysis in subjects affected by IVACs was not performed.
In regard to recommended dosages, some well-established evidence may support the use of CI over intermittent infusion in administering traditional antipseudomonal beta-lactams in critically ill patients [21]. Additionally, we recently showed in a large cohort of critically ill patients having documented Gram-negative infections treated with CI traditional beta-lactams that early achievement of an aggressive PK/PD target of Css/MIC > 5 within the first 72 h was significantly associated with both microbiological eradication and prevention of resistance development [22]. In the same study, Pseudomonas aeruginosa-related infections were independently associated with higher risk of microbiological failure [22]. Accordingly, we consider that when treating IVACs caused by multi-susceptible Pseudomonas aeruginosa isolates, the use of CI piperacillin-tazobactam and/or ceftazidime, and/or cefepime after loading may represent a valuable approach for rapidly achieving and maintaining an aggressive PK/PD target helpful at achieving microbiological eradication.

Pseudomonas aeruginosa
The emergence of MDR and extensively drug-resistant (XDR) Pseudomonas aeruginosa is a major clinical concern. The underlying mechanisms of the MDR/XDR phenotype are heterogeneous and can be mediated by the selection of mutations in the chromosomal genes or by horizontal acquisition of resistance determinants, including beta-lactamase and carbapenemase genes [23]. A major distinction regards MDR/XDR Pseudomonas aeruginosa isolates producing or not metallo-beta-lactamases. In case of MDR/XDR metallo-betalactamases (MBL)-negative isolates, antibiotics should be chosen according to antimicrobial susceptibility tests. Among the different beta-lactamases responsible for MDR/XDR phenotype, GES enzymes, belonging to class A carbapenemases, could play a remarkable role in the selection of appropriate targeted anti-pseudomonal agents. GES enzymes represents a major resistance mechanism of MDR/XDR high-risk clones (e.g., ST-235 highly virulent clone), exhibiting higher virulence compared to other clones (e.g., ST111 or ST175) and resulting in more severe acute infections with significant mortality [23]. Furthermore, production of GES enzymes confers resistance to ceftolozane-tazobactam [23]. In case of IVACs caused by MDR/XDR MBL-negative GES-negative isolates (Figure 1, Panel A.2), prolonged infusion of ceftolozane-tazobactam (3 g q8h CI after 3 g LD) is recommended as first-line therapy. Cefiderocol (2 g LD followed by 2 g q8h in CI) should be reserved as second-line alternative in case of clinical or microbiological failure (Figure 1, Panels A.2-A.3). A summary of the studies evaluating the efficacy of ceftolozane-tazobactam or cefiderocol in patients affected by IVACs caused by MDR GES-negative Pseudomonas aeruginosa is provided in Table 2.
Two different phase III RCTs [41,42] supported the efficacy of cefiderocol for the treatment of IVACs caused MDR Pseudomonas aeruginosa, reporting no significant difference in mortality or clinical cure rate compared to meropenem or best available therapy (including combination of aminoglycoside, carbapenems, colistin, fosfomycin or tigecycline) in case of carbapenem-resistant isolates. However, the proportion of carbapenemase-producer isolates in these studies was only 8-15%. Several in vitro studies [43][44][45] demonstrated the activity of cefiderocol against MDR (carbapenemase-negative meropenem non-susceptible) Pseudomonas aeruginosa (MIC range 0.002-8; MIC50 0.12). Notably, high susceptibility (90%) against GES-positive isolates was reported.    In regard to recommended dosages, the use of CI for ceftolozane-tazobactam is supported by the findings of a prospective study showing that among 72 patients affected by MDR Pseudomonas aeruginosa infections (mainly VAP), CI was associated with a higher probability of target attainment of the aggressive PK/PD target of 100% f T > 4 × MIC [46]. Additionally, a case-control study evaluating 28 patients affected by MDR Pseudomonas aeruginosa infections (67.9% pneumonia) showed that prolonging infusion up to 3 h allowed significantly lower resistance development compared to intermittent infusion (0.0% vs. 58.3%; p = 0.04) [47]. Consequently, we suggest CI of ceftolozane-tazobactam after loading for rapidly achieving and then maintaining aggressive PK/PD targets in IVACs caused by MDR MBL-negative GES-negative Pseudomonas aeruginosa isolates.
Continuous infusion of ceftazidime-avibactam (7.5 g/day after 2.5 LD) or cefiderocol (2 g LD followed by 2 g q8h in CI) in monotherapy or in association with fosfomycin (6-8 g LD followed by 16 g/day CI) represents the first-line therapy for the management of IVACs caused by MDR GES-positive Pseudomonas aeruginosa (Figure 1, Panel A.3). A summary of the evidence assessing the efficacy of ceftazidime-avibactam alone or with fosfomycin in patients affected by IVACs caused by MDR GES-positive Pseudomonas aeruginosa is provided in Table 3.
In a phase III RCT [48], Torres et al. reported no significant difference in clinical cure rate (64.3% vs. 77.1%; p = NS) between ceftazidime-avibactam and meropenem in subgroup of critically ill patients affected by HAP or VAP caused by MDR Pseudomonas aeruginosa. However, Pseudomonas aeruginosa accounted only for 30% of overall isolates, and no further analysis was provided to identify resistance mechanisms of Pseudomonas aeruginosa isolates (including carbapenemases/GES-production). Several observational studies and case series/reports [49][50][51][52][53] demonstrated the efficacy of ceftazidime-avibactam in critically ill patients (ICU admission ranging from 41.5% to 100%) affected by HAP/VAP in most cases. Particularly, Vena et al. [50] reported a clinical success of 87.8% in Pseudomonas aeruginosa infections (48.8% HAP/VAP) in 41 patients treated with ceftazidime-avibactam (prolonged infusion in 36.6% of cases). Notably, an in vitro analysis of a retrospective study including 24 patients affected by XDR Pseudomonas aeruginosa infections (33.3% nosocomial pneumonia) found an overall susceptibility to ceftazidime-avibactam of 100% in GES-5-positive strains belonging to ST235 clone, showing a MIC 90 of 6 mg/L [54].
Only preclinical evidence supported the association therapy between ceftazidimeavibactam and fosfomycin against MDR Pseudomonas aeruginosa [55][56][57]. Synergism between the two agents was retrieved in 25-61.9% of isolates and was also confirmed in a murine model of infection [55].
Cefiderocol could represent a valuable alternative to ceftazidime-avibactam in the treatment of IVACs caused by MBL-negative MDR GES-positive Pseudomonas aeruginosa, but caution is required due to the limited clinical experience in this setting and the quite low number of GES-positive strains that have been tested in vitro for susceptibility to this antibiotic. Anyway, considering that in most clinical studies MBL-negative MDR GESpositive Pseudomonas aeruginosa isolates were rarely characterized, caution is recommended in choosing any agent, including ceftazidime-avibactam.
In regard to recommended dosages, evidence supporting the use of CI for ceftazidimeavibactam is stemmed from a large observational study among 577 patients with KPCproducing Klebsiella pneumoniae treated with ceftazidime-avibactam, in which prolonged and/or CI was independently associated with higher survival rate compared to intermittent infusion [58]. In regard to cefiderocol, we recently showed in a descriptive case series of PK/PD target attainment and microbiological outcome in critically ill patients with documented severe XDR Acinetobacter baumannii BSI and/or VAP treated with cefiderocol that the standard 3 h infusion was ineffective in achieving the aggressive PK/PD of 100%fT > 4 × MIC in more than 50% of included patients. This resulted in a remarkable high rate of microbiological failure, especially among VAP cases [59].  Overall susceptibility rate to CTV in GES-5-positive strains:100%; MIC 90  Preclinical studymurine model infection The association between CTV and FOS significantly reduced the P. aeruginosa (CFUs), by approximately 2 and 5 logs, compared with stasis and in the vehicle-treated control, respectively. Administration of ceftazidime-avibactam and fosfomycin separately significantly increased CFUs, by approximately 3 logs and 1 log, respectively, compared with the number at stasis, and only reduced CFUs by approximately 1 log and 2 logs, respectively, compared with the number in the vehicle-treated control. The combination of CTV + FOS was superior to either drug alone and has the potential to offer infected patients with high bacterial burdens a valid therapeutic choice against infection with MDR-PA that lack metallo-beta-lactamases. Consequently, we recommend CI of ceftazidime-avibactam and/or of cefiderocol after loading for rapidly achieving and then maintaining aggressive PK/PD targets in IVACs caused by MDR Pseudomonas aeruginosa, also taking into account the limited ELF penetration rate of these agents [60][61][62].

Multidrug-Resistant (MDR) Metallo-Beta-Lactamase (MBL)-Positive Pseudomonas aeruginosa
Prolonged infusion of cefiderocol (2 g LD followed by 2 g q8h in EI or CI) in combination therapy with inhaled colistin (2 MU q8h) is recommended as first-line therapy for the management of IVACs caused by MDR metallo-beta-lactamase-producers (VIM, IMP, or NDM) Pseudomonas aeruginosa. Combination therapy including high-dose prolonged infusion meropenem (1-1.5 g q6h after 2 g LD), fosfomycin (6-8 g LD followed by 16 g/day CI), and inhaled colistin (2 MU q8h) should be reserved as second-line alternative (Figure 1,  panel A.4). A summary of the evidence assessing the efficacy of these antibiotic regimens in patients affected by IVACs caused by MDR metallo-beta-lactamase-positive Pseudomonas aeruginosa is provided in Table 4.
Although no clinical evidence for the use of meropenem in combination with fosfomycin for the management of IVACs caused by MBL-producing Pseudomonas aeruginosa currently exists, Albiero et al. [64] reported in vitro the synergic effect of the combination regimen in ten MBL-positive Pseudomonas aeruginosa. Synergism was found in 100% of isolates, resulting in a median decrease of MIC50 and MIC90 by eight-fold. PK/PD simulation showed that high-dose fosfomycin (6-8 g q8h) or meropenem (1.5 g q6h in 3 h EI) achieved the probability of target attainment of ≥ 90% respectively at an MIC of 32 mg/L and 16 mg/L. Additionally, combination therapy resulted in a significantly increase in the cumulative fraction rate against MBL-positive Pseudomonas aeruginosa compared to monotherapy with meropenem (32% vs. 68%) or fosfomycin (0% vs. 74%) [64].
A systematic review including patients affected by HAP or VAP due to MDR Pseudomonas aeruginosa reported respectively a clinical success and microbiological eradication in 70.4% and 71.3% of cases with inhaled colistin monotherapy [65]. No evidence for inhaled colistin in association with cefiderocol or meropenem/fosfomycin combination therapy currently exists. Table 4. Summary of the studies investigating the treatment of multidrug-resistant (MDR) metallo-beta-lactamase-positive Pseudomonas aeruginosa infection-related ventilator-associated complications (IVACs) with cefiderocol in association with inhaled colistin or combination therapy between meropenem, fosfomycin and inhaled colistin.

Meropenem + Fosfomycin + inhaled colistin
Albiero et al., 2019 [64] In vitro study 19 10 MBL+. Synergism was found in 100% of isolates with a FICI ≤ 0.5. Median reduction in MIC50 and MIC90 by 8-fold. PK/PD simulation showed that 6-8 g q8h FOS achieved the probability of target attainment of ≥90% at an MIC of 32 mg/L. 1.5 g q6h MER in 3 h EI achieved the probability of target attainment of ≥90% at an MIC of 16 mg/L. Combination therapy significantly increase the cumulative fraction rate against MBL-PA compared to monotherapy with MER (32% vs. 68%) or FOS (0% vs. 74%).

Vardakas et al., 2018 [65]
Systematic review and meta-analysis In regard to recommended dosages, evidence supporting the use of high-doses CI meropenem after 2 g loading stemmed from a recent study that assessed the impact of maximizing Css/MIC ratio on efficacy of CI meropenem against documented Gram-negative infections in critically ill patients [68]. In that study, it was shown that a Css/MIC ratio ≥4.63 was significantly associated with favourable clinical outcome among 74 ICU patients [68]. Monte Carlo simulation showed that higher meropenem dosages by CI should be recommended for dealing tackling appropriately Pseudomonas aeruginosa and Acinetobacter baumannii infections in critically ill patients with preserved renal function [68]. Likewise, a similar aggressive PK/PD target of Css/MIC > 5 within the first 72 h was significantly associated with both microbiological eradication and prevention of resistance development in a large cohort of critically ill patients having documented Gram-negative infections treated with CI traditional beta-lactams [22]. In regard to cefiderocol, we recently showed in a descriptive case series of PK/PD target attainment and microbiological outcome in critically ill patients with documented severe XDR Acinetobacter baumannii BSI and/or VAP that treatment with cefiderocol at the standard 3 #h infusion was ineffective in achieving the aggressive PK/PD of 100%fT > 4 × MIC in more than 50% of included patients. This resulted in a remarkable high rate of microbiological failure, especially among VAP cases [59]. Consequently, we suggest CI of cefiderocol or of high dose meropenem after loading for promptly achieving and then maintaining an aggressive PK/PD target in critically ill patients affected by IVACs due to MDR Pseudomonas aeruginosa.

Targeted Treatment of IVACs Caused by Acinetobacter baumannii in Critically Ill Adult Patients
Therapeutic algorithm for targeted treatment of IVACs caused by Acinetobacter baumannii in adult ICU patients is shown in Figure 2.

Carbapenem-Susceptible Acinetobacter baumannii
Continuous infusion of high-dose meropenem (1-1.5 g q6h after 2 g LD) or imipenem (500 mg q4-6h after 1 g LD) are recommended as targeted therapy for the management of IVACs caused by carbapenem-susceptible Acinetobacter baumannii (Figure 2, panel B.1). Several clinical and in vitro evidence [69][70][71][72][73][74][75] demonstrated the efficacy of carbapenems (particularly meropenem and imipenem) for the treatment of carbapenem-susceptible Acinetobacter baumannii infections (Table 5). Two different observational studies [69,70] compared imipenem with intravenous colistin for the management of critically ill patients affected by VAP due to MDR Acinetobacter baumannii, reporting no significant difference in clinical cure rate, mortality rate, resistance development, and adverse events between the two agents. Wang [71] retrospectively evaluated 30 critically ill patients affected by HAP/VAP due to MDR carbapenem-susceptible Acinetobacter baumannii, reporting a clinical cure rate of 100.0%. Furthermore, the administration of meropenem in EI was a cost-effective approach in this setting, although showing an equal clinical efficacy compared to intermittent infusion. Several in vitro studies [72][73][74][75] reported a variable susceptibility to carbapenems of Acinetobacter baumannii strains in different ecological settings, ranging from 41% to 75.8% and from 32.2% to 77.8% for meropenem and imipenem, respectively.
In regard to recommended dosages, evidence supporting the use of high-doses CI meropenem after 2 g LD stemmed from a PK/PD analysis carried out among 74 ICU patients affected by documented Gram-negative infections (of which pneumonia accounted for half of cases), in which a Css/MIC ratio ≥4.63 was significantly associated with favourable clinical outcome [68]. Monte Carlo simulation showed that, according to cumulative fraction of response, higher meropenem dosages by CI should be recommended for the management of Acinetobacter baumannii related infections among patients with preserved renal function [68]. Likewise, a similar aggressive PK/PD target of Css/MIC > 5 within the first 72 h was significantly associated with both microbiological eradication and prevention of resistance development in a large cohort of critically ill patients having documented Gram-negative infections treated with CI traditional beta-lactams [22].

Carbapenem-Resistant Acinetobacter baumannii
Prolonged infusion of cefiderocol (2 g q8h EI or CI after 2 g LD) represents the first-line therapy in the management of IVACs caused by MDR Acinetobacter baumannii. Combination therapy including fosfomycin (6-8 g LD followed by 16-24 g/day CI), high-dose ampicillinsulbactam (6 g/3g q8h CI after 6-8 g/3-4 g LD), and inhaled colistin (2 MU q8h) could be considered as second-line therapeutic alternative (Figure 2, panel B.2). A summary of the evidence assessing the efficacy of these antibiotic regimens in patients affected by IVACs caused by MDR Acinetobacter baumannii is provided in Table 6. Table 6. Summary of the studies investigating the treatment of carbapenem-resistant Acinetobacter baumannii infection-related ventilator-associated complications (IVACs) with cefiderocol or combination therapy between fosfomycin and ampicillin/sulbactam.    Although in vitro studies [44,45,63,79] reported high susceptibility rate of carbapenem non-susceptible Acinetobacter baumannii isolates against cefiderocol (MIC 50 ranging from 0.12 mg/L to 0.25 mg/L, and susceptibility above the 85% in all different OXA-producing strains), clinical evidence are currently limited. In the CREDIBLE-CR RCT [41], a significant higher mortality rate in IVACs caused by Acinetobacter baumannii (accounting for 65% of HAP/VAP) was reported in patients treated with cefiderocol compared to best available therapy (49% vs. 18%; p = 0.04). In a case series including 13 patients, Gatti et al. [59] found a microbiological failure rate of 54% in critically ill patients affected by XDR Acinetobacter baumannii infections (84.6% VAP). Notably, microbiological failure occurred in 80% of patients with suboptimal f C min /MIC < 1 compared to 29% of those achieving optimal (f C min /MIC > 4) or quasi-optimal (f C min /MIC = 1-4) PK/PD target. In a case series, Falcone et al. [77] described two critically ill COVID-19 patients developing VAP caused by XDR Acinetobacter baumannii and treated with cefiderocol. Clinical failure was reported in 50% of cases. Conversely, Trecarichi et al. [78] reported a successful case of bacteraemic VAP caused by XDR Acinetobacter baumannii treated with cefiderocol as targeted therapy.

Author, Year and Reference
Different RCTs and observational studies [80,81,83] demonstrated the efficacy of highdose sulbactam (up to 12 g/day) in monotherapy or combination therapy in IVACs caused by MDR Acinetobacter baumannii, showing no difference in clinical cure rate compared to colistin-based regimens [81]. However, evidence supporting combination therapy with fosfomycin are currently scanty. A case report [82] demonstrated the efficacy of combination therapy between high-dose sulbactam (9 g/day) and fosfomycin (16 g/day) in a critical care patients affected by meningitis due to MDR Acinetobacter baumannii. Notably, a synergism between sulbactam and fosfomycin was reported in 74% of MDR Acinetobacter baumannii isolates, resulting in a median MIC 50 and MIC 90 decrease up to eight-fold [83].
A systematic review including patients affected by HAP or VAP due to MDR Acinetobacter baumannii reported respectively a clinical success and microbiological eradication in 70.4% and 71.3% of cases with inhaled colistin monotherapy [65]. A retrospective casecontrol study [84] assessing 78 patients affected by respiratory infection or colonization due to MDR Acinetobacter baumannii found that the use of inhaled colistin was the only independent factor associated with microbiological eradication within 14 days after the index day (OR 266.33; 95% CI 11.26-6302.18, p < 0.001), and shortened the duration of MDR Acinetobacter baumannii recovery from the respiratory tract by 13.3 days.
In regard to recommended dosages, no evidence for administering cefiderocol by CI still exists Currently. However, it was shown in a descriptive case series of PK/PD target attainment and microbiological outcome in critically ill patients with documented severe XDR Acinetobacter baumannii BSI and/or VAP that treatment with cefiderocol with the standard 3 h infusion was ineffective in achieving the aggressive PK/PD of 100%fT > 4 × MIC in more than 50% of included patients. This resulted in a remarkable high rate of microbiological failure, especially among VAP cases [59].
Consequently, we recommend CI of cefiderocol after loading for rapidly achieving and then maintaining aggressive PK/PD targets in IVACs caused by Acinetobacter baumannii, also taking into account the limited ELF penetration rate of these agents [60][61][62].

Overview of Recommendations
Non-fermenting Gram-negative pathogens represent a leading cause of HAP or VAP in critically ill patients requiring mechanical ventilation [2]. Particularly, Pseudomonas aeruginosa represents one of the leading causative pathogens in critically ill patients affected by HAP or VAP [85]. The widespread diffusion of MDR and XDR isolates coupled with the emergence of high-risk clones (ST111, ST175, and ST235) calls for the prompt administration of adequate antibiotic therapy and optimization of lung exposure [23,85].
Two main cornerstones should be pursued in the management of IVACs caused by Pseudomonas aeruginosa: (1) the implementation of carbapenem-sparing regimens both in case of multi-susceptible isolates (considering the higher risk of relapse or resistance development [16,18]), preferring the use of piperacillin-tazobactam or third/fourth-generation cephalosporins (ceftazidime or cefepime), and of MDR/XDR isolates, favouring the administration of novel BL/BLIs (ceftolozane-tazobactam or ceftazidime-avibactam) or cefiderocol, according to the resistance mechanism exhibited by the specific strain); (2) the adoption of altered dosing strategies (high-doses and/or prolonged infusion) in order to achieve optimal PK/PD target at the site of infection, considering the limited pulmonary penetration of piperacillin-tazobactam, ceftazidime, cefepime, ceftazidime-avibactam, and cefiderocol (below 20%) [60][61][62][86][87][88]. Notably, in a prospective observational study including 72 patients affected by MDR Pseudomonas aeruginosa infections (79% ICU admission; 66.7% HAP/VAP), Plimis et al. [46] reported that intermittent infusion of ceftolozane-tazobactam was inadequate to achieve optimal PK/PD target for MIC ≥ 4 mg/L compared to continuous infusion. Similarly, ATS/IDSA guidelines recommended the use of prolonged infusion of piperacillin-tazobactam, ceftazidime, or cefepime in HAP or VAP due to Pseudomonas aeruginosa to maximize lung exposure [15].
The treatment of IVACs caused by MBL-producing Pseudomonas aeruginosa is challenging. In this scenario, cefiderocol could play a major role, although evidence is currently limited to only in vitro studies. The administration of inhaled colistin could represent a valuable therapeutic strategy in association with cefiderocol or meropenem and fosfomycin, providing a targeted antibiotic delivery in respiratory tract infections and resulting in lower systemic toxicity compared to intravenous colistin [89,90]. It should not be overlooked that after administration of a single 2 MU of inhaled colistimethate, colistin concentrations in the ELF ranged between 9.53 and 1137 mg/L, which are values much higher than that achievable after the administration of the same dose by the intravenous route (1.48-28.9 mg/L) [91]. The best dosing regimen of inhaled colistimethate has still to be defined, as quite variable dosages have been proposed in the literature, ranging from 1 MU q12h up to 5 MU q8h [91]. However, we believe that the dosage of 2 MU q8h should be preferred Currently considering that is the one supported by the major clinical evidence [65,91].
MDR or XDR Acinetobacter baumannii is a leading causative pathogen in critically ill patients affected by VAP, characterized by mortality rate up to 60% [92]. Although colistinor polymyxin-based regimens were widely used in clinical practice for the management of severe Acinetobacter baumannii infections, relevant toxicity (mainly nephrotoxicity) and PK/PD issues (low ELF exposure, high occurrence of breakthrough infections) strongly affect their efficacy [92].
Notably, CREDIBLE-CR trial found [41] a remarkable mortality rate amongst patients affected by Acinetobacter baumannii infections, thus mitigating the initial expectations for cefiderocol [93]. Recently, it has been suggested to limit the use of cefiderocol to situations when intolerance or resistance to other generally active drugs has been shown [93]. However, the favourable safety profile, the high in vitro activity, and the potential maximization of lung exposure through the implementation of CI, make cefiderocol the first-line choice for targeted therapy of IVACs due to MDR Acinetobacter baumannii in critically ill patients.
By virtue of their lung penetration [94], fosfomycin could be a valid alternative for combination therapy with high-dose sulbactam (9-12 g/day) as second-line strategy for the management of MDR Acinetobacter baumannii. A recent observational study [95] identified fosfomycin-containing regimen as an independent predictor for 30-day survival in severe pneumonia caused by MDR Acinetobacter baumannii. However, none of these combination therapies included sulbactam.
Overall, in the management of IVACs caused by non-fermenting Gram-negative pathogens, the implementation of beta-lactams altered dosing strategies consisting in high-dose CI administration is strongly recommended for attaining aggressive PK/PD target of 100% fT > 4-8 × MIC and maximizing lung exposure [87,96,97]. This approach may both maximize clinical efficacy and prevent the development of resistance [87,98]. This aggressive strategy should be pursued also in the treatment of infections caused by wild-type strains. Optimal exposure into the ELF is difficult to be achieved and maintained with intermittent infusion of piperacillin-tazobactam, ceftazidime, and/or cefepime due to the limited penetration rates of these hydrophilic agents, and conversely, Pseudomonas aeruginosa was found to be an independent predictor of microbiological failure in critically ill patients affected by documented Gram-negative infections [22]. Clinicians must be aware that the antibiotic dosing regimens that we recommended throughout the manuscript are focused only on treatment of patients with normal renal function. It should not be overlooked that the pharmacokinetics of hydrophilic antimicrobial agents, namely beta-lactams and fosfomycin, may be affected among critically ill patients by several pathophysiological conditions that may alter volume of distribution and/or renal clearance [97,99]. Consequently, dose adjustments are needed in critically ill renal patients, especially among those with transient acute kidney injury, augmented renal clearance (ARC), and/or undergoing renal replacement therapy (RRT) [97,100]. In this scenario, adaptative daily therapeutic drug monitoring (TDM) may represent a valuable tool in addressing these issues. Currently, routinely TDM of BLs is strongly recommended for optimizing treatment among critically ill patients with Gram-negative infections [101], and an aggressive PK/PD target is considered useful among patients undergoing CRRT or having ARC [97]. Interestingly, the achievement of aggressive PK/PD target of Css/MIC ratio around 5 within the first 72 h of treatment with CI traditional beta-lactams was recently shown to be associated with significantly higher probability of both clinical outcome and microbiological eradication among critically ill patients with documented Gram-negative infections [22,68]. This suggests that early optimization of drug exposure by means of real-time TDM may be extremely helpful in maximizing treatment with beta lactams among the critically ill patients. However, it should be recognized that the extensive use of a real-time TDM-guided clinical pharmacologist advice approach is still burdened by many barriers [97,[102][103][104]. To mention some of these, measurement of unbound concentrations, daily TDM assessment, timely turnaround time, appropriate interpretation of TDM data performed by well-trained MD or PharmD clinical pharmacologists, and implementation of user-friendly methods for novel beta-lactams still represent critical issues [97,[102][103][104].
Furthermore, alternative agents should be considered for the treatment of patients with well-documented life-threatening beta-lactam allergies. Fluoroquinolones, aminoglycosides and colistin could be helpful in these cases depending on the susceptibility pattern of Pseudomonas aeruginosa or Acinetobacter baumannii. Besides, in case of life-threatening infections caused by MDR non-fermentative Gram-negatives with limited therapeutic options, desensitization protocols should be taken into consideration.
Finally, it should be mentioned that the current COVID-19 pandemic has led to a worldwide rise in antimicrobial resistance due to the massive disruption of infection control and antimicrobial stewardship measures in COVID ICUs. This caused a remarkable proportion of MDR bacterial co-infections and super-infections (including Pseudomonas aeruginosa and Acinetobacter baumannii) in severe COVID-19 patients [105]. In this scenario of more and more growing antimicrobial resistance, educational programs for improving the culture of knowledge of appropriate antibiotic use and the pivotal role of antimicrobial stewardship should be planned and delivered to young physicians [106].
In recent years, several leading guidelines for the management of Gram-negative infections have been issued by the most important Infectious Disease Societies worldwide [107][108][109]. Unfortunately, none of these focused on the treatment of IVACs and/or provided extensive recommendations for appropriate place in therapy of novel agents and/or for dosing optimization according to the PK/PD principles. Consequently, the implementation of developed therapeutic algorithms based on susceptibility pattern of non-fermenting Gram-negative isolated pathogens, coupled with the administration of altered dosing strategies of beta-lactams [97], could provide intensive care physicians a useful guidance for maximizing antibiotic treatment in ICU patients affected by IVACs, in order to address three main purposes: (a) to provide a personalized and targeted antibiotic therapy in each critically ill patient affected by HAP/VAP due to Pseudomonas aeruginosa or Acinetobacter baumannii; (b) to appropriately place novel antimicrobial agents in lack of definitive evidence; (c) to consider antimicrobial stewardship strategies for sparing the broadest-spectrum antibiotics (namely carbapenems). It is supposed that this strategy could maximize clinical outcome while minimizing resistance development in a challenging scenario as the management of IVACs in ICU patients.

Materials and Methods
A multidisciplinary task force, composed by one intensive care physician (B.V.), one infectious disease consultant (P.V.), one clinical microbiologist (G.M.R.), and one MD clinical pharmacologist (F.P.) met virtually on several occasions with the intent of developing algorithms for targeted antimicrobial therapy of IVACs caused by Pseudomonas aeruginosa and Acinetobacter baumannii in ICU critically ill patients. IVACs were defined as the presence of a ventilator-associated condition (consisting in a 48 h stable or decreasing daily minimum positive end-expiratory pressure [PEEP] or FiO 2 followed by a rise in PEEP of 3 cm H 2 O or a rise in FiO 2 of 0.2 sustained for 48 h) coupled with the occurrence of body temperature <36 • C or >38 • C and the start of at least one antibiotic agent continued for over 96 h. VAP was considered a subgroup of IVACs, consisting in the presence of at least 25 neutrophils/field coupled with positive semi-quantitative/quantitative culture for pathogenic organisms at bronchoalveolar lavage [110]. The definitive agreement for each therapeutic algorithm was reached by the multidisciplinary team after thoroughly discussion based on specific long-standing experience and on the specific expertise of each single member in terms of management of critically ill patients affected by IVACs, in appropriately placing in therapy of the old and novel antimicrobial agents, in implementing appropriate target antibiotic therapy and antimicrobial stewardship strategies in challenging scenarios, in applying traditional and novel microbiological methods and in interpreting microbiological findings and susceptibility test according to the specific clinical scenarios, and in optimizing and individualizing antibiotic dosing regimens according to the specific pathophysiological alterations. Each therapeutic algorithm was designed after that unanimous agreement among the four members of the multidisciplinary team was achieved. Six different scenarios were identified based on the resistance genotype of the pathogens and/or on the antibiotic susceptibility pattern (namely multi-susceptible Pseudomonas aeruginosa, MDR metallo-beta-lactamase-negative Pseudomonas aeruginosa, MDR metallo-beta-lactamase-positive Pseudomonas aeruginosa, carbapenem-susceptible Acinetobacter baumannii, and carbapenem-resistant Acinetobacter baumannii). MDR Pseudomonas aeruginosa isolates were defined according to the classification proposed by Magiorakos et al. [111]. A hierarchical scale was established whenever agreement on multiple therapeutic options was achieved in one specific scenario. Optimized antibiotic dosing schedules were provided as well.
A researcher (M.G.) retrieved the scientific evidence needed for supporting the specific choices included in the algorithms by means of a PubMed-MEDLINE literature search (until October 2021). Key terms were selected antibiotics, HAP, VAP, IVACs, and bacterial pathogens with genotype of resistance and/or antibiotic susceptibility pattern. Quality of evidence was established according to a hierarchical scale of the study design, as reported in the evidence pyramid [112]: randomized controlled trials (RCTs), prospective observational studies, retrospective observational studies, case series, case reports, and in vitro studies. International guidelines issued by the Infectious Disease Society of America and/or by the European Society of Clinical Microbiology and Infectious Diseases, systematic reviews and meta-analyses were also consulted. Consistence between retrieved studies was also considered, by assessing the concordance in clinical outcome of the included studies at each level of the evidence pyramid. Only articles published in English were considered, with a focus mainly on studies published in the last ten years.

Conclusions
In an era characterized by widespread diffusion of MDR Gram-negative pathogens and continuous increase in antibiotic resistance, the implementation of a multidisciplinary taskforce focusing on targeted therapy in critically ill patients has become a real need. Our approach is focused on prompt revision of inappropriate/unnecessary antibiotic therapy, implementation of "carbapenem-sparing" strategies, and PK/PD optimization of antibiotic exposure hopefully guided by real-time TDM whenever feasible. Rational use of broadspectrum antibiotics, especially carbapenems, could represent a powerful strategy for tackling resistance spread in the ICU setting [14]. We believe that this strategy and these algorithms could be helpful in improving clinical outcomes and in avoiding resistance spread in the IVAC setting. The availability of molecular diagnostic tests that can rapidly provide information about the nature of the infecting pathogens and the presence of some relevant resistance determinants will be instrumental to improve antimicrobial stewardship practices based on the proposed algorithms.