Incidence of Carbapenem-Resistant Gram-Negative Bacterial Infections in Critically Ill Patients with COVID-19 as Compared to Non-COVID-19 Patients: A Prospective Case-Control Study

Introduction Critically ill COVID-19 patients hospitalized in intensive care units (ICU) are immunosuppressed due to SARSCoV-2-related immunological effects and are administered immunomodulatory drugs. This study aimed to determine whether these patients carry an increased risk of multi-drug resistant (MDR) and especially carbapenem-resistant Gram-negative (CRGN) bacterial infections compared to other critically ill patients without COVID-19. Materials and Methods A prospective case-control study was conducted between January 2022 and August 2023. The ICU patients were divided into two groups (COVID-19 and non-COVID-19). Differences in the incidence of CRGN infections from Klebsiella pneumoniae, Acinetobacter spp., and Pseudomonas aeruginosa were investigated. In addition, an indicator of the infection rate of the patients during their ICU stay was calculated. Factors independently related to mortality risk were studied. Results Forty-two COVID-19 and 36 non-COVID-19 patients were analyzed. There was no statistically significant difference in the incidence of CRGN between COVID-19 and non-COVID-19 patients. The infection rate was similar in the two groups. Regarding the aetiological agents of CRGN infections, Pseudomonas aeruginosa was significantly more common in non-COVID-19 patients (p=0.007). COVID-19 patients had longer hospitalisation before ICU admission (p=0.003) and shorter ICU length of stay (LOS) (p=0.005). ICU COVID-19 patients had significantly higher mortality (p < 0.001) and sequential organ failure assessment (SOFA) score (p < 0.001) compared to non-COVID-19 patients. Μortality secondary to CRGN infections was also higher in COVID-19 patients compared to non-COVID-19 patients (p=0.033). Male gender, age, ICU LOS, and hospital LOS before ICU admission were independent risk factors for developing CRGN infections. Independent risk factors for patients' mortality were COVID-19 infection, obesity, SOFA score, total number of comorbidities, WBC count, and CRP, but not infection from CRGN pathogens. Conclusions The incidence of CRGN infections in critically ill COVID-19 patients is not different from that of non-COVID-19 ICU patients. The higher mortality of COVID-19 patients in the ICU is associated with higher disease severity scores, a higher incidence of obesity, and multiple underlying comorbidities, but not with CRGN infections.

Introduction.Critically ill COVID-19 patients hospitalized in intensive care units (ICU) are immunosuppressed due to SARSCoV-2-related immunological efects and are administered immunomodulatory drugs.Tis study aimed to determine whether these patients carry an increased risk of multi-drug resistant (MDR) and especially carbapenem-resistant Gram-negative (CRGN) bacterial infections compared to other critically ill patients without COVID-19.Materials and Methods.A prospective casecontrol study was conducted between January 2022 and August 2023.Te ICU patients were divided into two groups (COVID-19 and non-COVID -19).Diferences in the incidence of CRGN infections from Klebsiella pneumoniae, Acinetobacter spp., and Pseudomonas aeruginosa were investigated.In addition, an indicator of the infection rate of the patients during their ICU stay was calculated.Factors independently related to mortality risk were studied.Results.Forty-two COVID-19 and 36 non-COVID-19 patients were analyzed.Tere was no statistically signifcant diference in the incidence of CRGN between COVID-19 and non-COVID-19 patients.Te infection rate was similar in the two groups.Regarding the aetiological agents of CRGN infections, Pseudomonas aeruginosa was signifcantly more common in non-COVID-19 patients (p � 0.007).COVID-19 patients had longer hospitalisation before ICU admission (p � 0.003) and shorter ICU length of stay (LOS) (p � 0.005).ICU COVID-19 patients had signifcantly higher mortality (p < 0.001) and sequential organ failure assessment (SOFA) score (p < 0.001) compared to non-COVID-19 patients.Μortality secondary to CRGN infections was also higher in COVID-19 patients compared to non-COVID-19 patients (p � 0.033).Male gender, age, ICU LOS, and hospital LOS before ICU admission were independent risk factors for developing CRGN infections.Independent risk factors for patients' mortality were COVID-19 infection, obesity, SOFA score, total number of comorbidities, WBC count, and CRP, but not infection from CRGN pathogens.Conclusions.Te incidence of CRGN infections in critically ill COVID-19 patients is not diferent from that of non-COVID-19 ICU patients.Te higher mortality of COVID-19 patients in the ICU is associated with higher disease severity scores, a higher incidence of obesity, and multiple underlying comorbidities, but not with CRGN infections.

Introduction
A minority of patients with COVID-19, usually older patients, with underlying comorbidities or immunosuppression, might develop a severe clinical course associated with acute respiratory distress syndrome (ARDS) and multiple organ failure.Tese critically ill patients are supported with mechanical ventilation in intensive care units (ICU).Pathophysiologically, the severe or critical illness from COVID-19 is characterized by a "cytokine storm" with excessive release of proinfammatory mediators [1].Terefore, therapeutically dexamethasone and biological factors, such as tocilizumab (an IL-6 inhibitor), baricitinib (a Janus kinase inhibitor), or anakinra (an IL-1 inhibitor) are used [2,3].However, these agents have been associated with an increased risk of severe bacterial infections [4,5].
Critically ill patients are at high risk of developing hospital-acquired infections, mainly ventilator-associated pneumonia (VAP) and bloodstream infections [6,7].Prior colonization with resistant pathogens, surgery during current admission, prior dialysis with end stage renal disease, use of carbapenems, and stay in the ICU for more than 5 days have been identifed as independent risk factors for multi-drug resistant (MDR), especially carbapenemresistant Gram-negative (CRGN), bacterial infections [8].Tis risk of nosocomial ICU-acquired infections is increased in COVID-19 patients due to frequent, signifcant, and prolonged lymphopenia and the use of glucocorticosteroids or biological factors that promote immunosuppression [9].Amongst these infections developed in critically ill COVID-19 patients, a proportion ranging between 32% and 50% is caused by MDR pathogens [10][11][12][13].Enterobacteriaceae, Pseudomonas aeruginosa, and Acinetobacter spp.are common Gram-negative resistant pathogens that account for bacterial superinfections in COVID-19 patients [9,11,14,15].However, most previous studies estimating the risk of MDR infections in COVID-19 patients did not include a comparative non-COVID-19 group of patients to counterbalance the contribution of the ICU epidemiology, which is especially necessary in ICUs with a high prevalence of MDR isolates [16].
Tis study aimed to determine whether critically ill ICU patients sufering from COVID-19 infection carry an increased risk of CRGN bacterial infections compared to critically ill patients without COVID-19.Tis risk was also estimated relative to the time spent in the ICU for each patient.Te study's secondary objectives were to identify diferences between the two groups of patients in baseline characteristics, disease severity and outcome, and identify factors related to CRGN infection and mortality risk.

2.1.
Patients.Tis prospective case-control study was conducted at the ICU of Patras General University Hospital, a tertiary, academic, 750-bed hospital, between January 2022 and August 2023.Te study protocol was approved by the Hospital Research Ethics Committee (PN: 10408), and the need for informed consent was waived.During the pandemic, informed consent was waived for all observational studies, according to our hospital's Research Ethics Committee, in order to limit the spread of the disease.However, prior to the patient's enrollment in the study, there was a phone call to the legal representative, and permission to collect data was requested.
Patients' primary clinical and demographic data were retrospectively collected from the electronic clinical records on the day of ICU admission.From that day onwards, the data of interest were collected prospectively, and survival was assessed at ICU discharge.Te patients were divided into two groups (COVID-19 and non-COVID-19).
Te criteria for inclusion in the study were patients aged >18 years who were hospitalized in the ICU during the study and who developed after 48 h of ICU admission a nosocomial infection, which was microbiologically investigated by cultures of relevant samples and at least one set of blood cultures.All the patients that were included in the COVID-19 group sufered from pneumonia and mild or severe ARDS.Te exclusion criteria were patients missing clinical or microbiological data and patients from whom blood culture samples were not taken or lost.All patients with positive cultures for CRGN pathogens were seen by an experienced infectious diseases specialist, and cases considered colonization were excluded.Te following characteristics of patients were recorded: age, gender, type of CRGN infection (K.pneumoniae, Acinetobacter spp., and P. aeruginosa), comorbidities (total number and type), sequential organ failure assessment (SOFA) score, hospital and ICU length of stay (LOS), hospital LOS before ICU admission, and patients' outcome (death or discharge from the ICU).Infammation and cell damage markers were also recorded: white blood cell count (WBC), C-reactive protein (CRP), D-dimer, and lactate dehydrogenase (LDH).Te SOFA score and infammation and cell damage markers were evaluated after positive cultures of CRGN were confrmed.Te criteria for patients' ICU discharge were exactly the same for both groups of patients (COVID-19 and non-COVID-19) by following the hospital's intensive care strategy (not to be needed interventions received in the ICU such as mechanical ventilation, vasoactive medications, renal replacement therapy, and invasive monitoring).In addition, for ICU discharge, patients should not have had clinical and laboratory signs of infection.During their hospitalisation in the ICU, patients with positive CRGN blood cultures who died while presenting clinical and laboratory signs of infection (fever, need for vasoactive medications, increased WBCs, etc.) were considered to die of CRGN infection.

Bacterial Identifcation and Antimicrobial Susceptibility
Testing.Identifcation of Gram-negative pathogens was performed by VITEK ® 2 Gram-negative identifcation cards (bioMérieux, Marcy-l'Etoile, France).Susceptibility to the following antimicrobial agents was studied: amikacin, gentamicin, tobramycin, ciprofoxacin, levofoxacin, imipenem, meropenem, tigecycline, trimethoprim/sulfamethoxazole, and colistin.MICs to ceftazidime, cefepime, ampicillin/sulbactam, piperacillin/tazobactam, and minocycline were evaluated from 2020 to 2023.Antimicrobial agents were selected according to the European Committee for Antimicrobial Susceptibility Testing (EUCAST) guidelines, as well as according to the European Centre for Disease Prevention and Control (ECDC) and the Centers for Disease Control and Prevention (CDC) suggestions [17].Results were evaluated, and isolates were defned as susceptible (including susceptible and susceptible-increased exposure) and resistant based on EUCAST guidelines [17,18].MIC to 2 Critical Care Research and Practice colistin was determined by the broth microdilution method (SensiTest ™ Colistin, Lioflchem, Roseto degli Abruzzi, Italy), as recommended [17].

MDR/XDR/PDR Defnitions and Infection Rate.
Te terms MDR (multidrug-resistant), XDR (extensively drugresistant), and PDR (pandrug-resistant) were used as described.Specifcally, MDR is defned as acquired nonsusceptibility to at least one agent in three or more antimicrobial categories among aminoglycosides, antipseudomonal carbapenems, antipseudomonal fuoroquinolones, antipseudomonal penicillins plus betalactamase inhibitors, extended-spectrum cephalosporins, folate pathway inhibitors, penicillin plus beta-lactamase inhibitors, polymyxins, and tetracyclines.XDR was defned as nonsusceptible to at least one agent in all but two or fewer antimicrobial categories (i.e., bacterial isolates remain susceptible to only one or two categories), and PDR was defned as nonsusceptible to all agents in all antimicrobial categories [17].Isolates that were not eligible for categorization as MDR, XDR, and PDR were assigned as susceptible (S).
In addition, an indicator of the infection rate of the patients during their ICU stay was calculated, with the numerator the number of patients' infections and the denominator the ICU LOS for all patients in both groups.Tree diferent analyses were conducted based on a predefned analysis plan.Te frst analysis aimed to determine the factors that difer among COVID-19 and non-COVID-19 patients.Te second analysis aimed to detect predictors of MDR/XDR infections from Gram (− ) microorganisms and predictors of patient mortality.In the third analysis, Kaplan-Meier curves and the logrank test were used to compare the overall survival and the survival secondary to CRGN infections of the two groups.To identify predictors of infections from Gram (− ) MDR/XDR microorganisms and predictors of patient mortality, a multivariate regression analysis was conducted using logistic regression or cox regression (for mortality) with backward stepwise elimination.Variables with p-values<0.1 in the univariate regression were included in the multivariable model, while the choice of variables was based on scientifc knowledge and considered potential collinearity.Multicollinearity issues were assessed using the variance infation factor (VIF).
Independent factors contributing to CRGN infections are presented as the Odds ratio (O.R.) and 95% confdence interval (CI), while independent factors contributing to patients' mortality are presented as the hazard ratio (H.R.) and 95% CI, after checking for the proportionality of hazards.In all cases, p-values <0.05 were considered signifcant.Data were analyzed using the Statistical Package for the Social Sciences (SPSS) (version 28.0; IBM, Armonk, USA) and GraphPad Prism Software (La Jolla, CA, version 10.1.0).

Diferences between COVID-19 and Non-COVID-19
Patients.A total of 173 patients were enrolled, of whom 78 completed the study (42 COVID-19 and 36 non-COVID-19 patients).A modifed CONSORT fow diagram, following the STROBE Statement-Checklist, of our study is presented in Figure 1.Diferences between the two groups of patients (COVID-19 vs. non-COVID-19) are presented in Table 1.Te two groups of patients did not signifcantly difer in terms of gender and age.Regarding underlying comorbidities, the incidence of obesity was higher in the COVID-19 patients (p � 0.034), while chronic obstructive pulmonary disease (COPD) was signifcantly more common in the non-COVID-19 patient group (p � 0.028).Concerning immunosuppression agents, COVID-19 patients statistically more often received glucocorticosteroids or biological factors compared to non-COVID-19 patients (p < 0.001).Tere were no diferences in the incidence of cancer (solid tumors and hematological malignancies) between the two groups (p � 0.06).
Tere were no statistically signifcant diferences in the incidence of infections between COVID- 19   Regarding infammation and cell damage markers, the two groups did not difer statistically signifcantly in WBCs, CRP, and D-dimers.However, patients with COVID-19 infection had statistically signifcantly higher LDH values than patients without COVID-19 infection (p � 0.005).

Predictors of Developing CRGN Infection.
Logistic regression with univariate and multivariate analyses showed that male gender, age, ICU LOS, and hospital LOS before ICU admission were independent risk factors for developing CRGN infections (Table 2).More specifcally, men presented 28.7% more infections by CRGN microorganisms than women.Also, the probability of developing CRGN microorganism infection increased by 3.7% (95% CI: 0.3-7%) for 1 year of age, while the probability of developing an infection during ICU hospitalisation increased by 43% (95% CI: 30-98%) with increasing patient age of 10 years.
Te probability of developing an infection by CRGN microorganism for each additional day of ICU LOS increased by 1.6% (95% CI: 0.1-3.2%).For every 10 days of increased ICU LOS, there was a 17% (95% CI: 1-37%) increased risk of developing infection.
Te probability of developing a CRGN infection for each additional day of hospital LOS before ICU admission increased by 4.6% (95% CI: 0.2-9.1%).For every 10 days of increased hospital LOS before ICU admission, there was a 56% (95% CI: 2-138%) increased risk of developing infection.Te infection rate was not an independent risk factor for developing CRGN infections (p � 0.37, p � 0.129, and p � 0.81 for Acinetobacter spp., Pseudomonas aur., and Klebsiella pn., respectively).

Predictors of Patient Mortality.
Univariate and multivariate cox regression analysis of data revealed that COVID-19 infection (p � 0.001), obesity (p � 0.03), SOFA score (p � 0.012), the total number of comorbidities (p � 0.026), absolute WBC count (p � 0.015), and CRP (p � 0.015) were independent risk factors for increasing patient mortality.Infection from CRGN pathogens is not an independent risk factor for increased mortality.Te results of the univariate and multivariate analyses are presented in Table 3. Te Kaplan-Mayer curve and the logrank test (Figure 2) showed that patients with COVID-19 infection had poorer survival rates than patients without COVID-19 infection (logrank test � 20.96 and p < 0.0001, Figure 2(a)), while mortality secondary to CRGN infections was also signifcantly higher in COVID-19 patients compared to non COVID-19 patients (logrank test � 10.18 and p < 0.001, Figure 2(b)).

Discussion
According to the literature, infection with the COVID-19 virus is an independent risk factor for developing immunosuppression and cell damage that increases the host's susceptibility to infections [19,20].Regarding the development of bacterial infections in patients with COVID-19 infection, one of the hypotheses is that viral infection causes direct damage to the epithelial cells of the lower airways associated with impaired clearance of mucus, which facilitates the binding of bacteria to cell surfaces [21].In addition, bacterial superinfection risk in COVID-19 patients is increased due to administration of glucocorticosteroids or biological factors that promote immunosuppression [9].Previous studies have shown that length of hospital or ICU stay is an important risk factor for acquiring a nosocomial MDRGN infection [22,23].Moreover, infections by CRGN pathogens are associated with an increased risk of mortality [24][25][26][27].Terefore, the present study was designed to address the question of contribution of CRGN bacterial infections in the mortality of COVID-19 critically ill patients.Most importantly, to discriminate whether nosocomialacquired CRGN infections in critically ill COVID-19 patients are associated with the underlying condition of virusor drug-induced immunosuppression, a comparative group of non-COVID critically ill patients, exposed to the same troubled epidemiological environment of the Greek ICUs, was included.
Our study results showed that the incidence of CRGN infections did not difer signifcantly in COVID-19 patients compared to non-COVID-19 patients.It should be considered that in the present study, COVID-19 patients remained hospitalized for a longer period before ICU admission, while non-COVID-19 patients had longer ICU LOS.In line with previous studies, our results showed that both hospital (non-ICU) and ICU LOS represent Critical Care Research and Practice independent risk factors for developing CRGN infections, and therefore, their impact on nosocomial infection acquisition seems to be counterbalanced in the two patient groups [22,23].
Administration of immunosuppressants was signifcantly higher in COVID-19 ICU patients.However, despite this fact, the incidence of CRGN infections in COVID-19 patients has not increased.On the contrary, regarding infections from resistant Pseudomonas aeruginosa, more cases were observed in non-COVID-19 patients.A recent study showed a decreased risk of colonization and/or infection with MDR in immunocompromised COVID-19 patients compared with immunocompetent subjects [28].Te authors attributed this fnding to the importance of contact precautions measures to prevent cross-transmission of MDR bacteria, which are more frequently applied in immunocompromised patients [29].In addition, previous studies with critically ill mechanically ventilated COVID-19 patients   [29,30].Except for solid tumors and hematological malignancies, which were similar between the two groups of patients, many other causes of immunosuppression could drive superinfections and mortality in ICU patients, although very recent data have challenged that common assumption [31].In our analysis,  In a recent review of published studies on the impact of COVID-19 infection in the incidence of ICU-acquired MDR colonization/infection, the authors concluded that MDR bacterial superinfection or colonization is similar or potentially lower in COVID-19 patients vs. controls [16].Terefore, according to our results, the higher mortality rate observed in critically ill COVID-19 patients was associated with higher disease severity, as assessed with the SOFA score, and the higher presence of comorbidities in this population and not with CRGN bacterial infections.
Tis study has some important limitations: First, this is a monocentric study with a low number of patients, which might have infuenced the potential to reveal diferences in the incidence of CRGN bacterial infections.Second, identifcation of CRGN pathogens was based on conventional microbiological techniques (cultures), whose sensitivity is dropped in patients receiving antibiotic therapy.However, the present study has some advantages as well: it has a prospective design and includes a control group of non-COVID-19 patients to balance the increased risk of CRGN infection acquisition in the troubled epidemiological environment of Greek ICUs.

Conclusions
Te present prospective study has shown that the incidence of infections with CRGN microorganisms did not difer signifcantly in critically ill ICU COVID-19 patients compared to non-COVID-19 patients.Increased mortality in the COVID-19 patients was not associated with higher CRGN infections but with higher disease severity and more underlying comorbidities.Tis study was conducted in a European country with a high incidence of infections from CRGN microorganisms.Terefore, its results should not be generalised to the global population of COVID-19 patients.To establish the contribution of critical COVID-19 infection in the incidence of CRGN infections more safely, prospective studies with larger numbers of patients are needed.
SOFA: sequential organ failure assessment score, WBC: white blood cells, LDH: lactate dehydrogenase, CRP: C-reactive protein, ICU: intensive care unit, LOS: length of stay.Statistically signifcant diferences are presented in bold.

Table 2 :
Univariate and multivariate regression analyses of factors associated with developing infection by CRGN microorganisms.: intensive care unit, LOS: length of stay.Statistically signifcant values are presented in bold. ICU

Table 3 :
Univariate and multivariate cox regression analyses of factors associated with patients' mortality.