Risk Factors for 30-Day Mortality in Nosocomial Enterococcal Bloodstream Infections

Enterococci commonly cause nosocomial bloodstream infections (BSIs), and the global incidence of vancomycin-resistant enterococci (VRE) BSIs is rising. This study aimed to assess the risk factors for enterococcal BSIs and 30-day mortality, stratified by Enterococcus species, vancomycin resistance, and treatment appropriateness. We conducted a retrospective cohort study (2014–2021) including all hospitalized adult patients with at least one blood culture positive for Enterococcus faecalis or Enterococcus faecium. We included 584 patients with enterococcal BSI: 93 were attributed to vancomycin-resistant E. faecium. The overall 30-day mortality was 27.5%; higher in cases of BSI due to vancomycin-resistant E. faecium (36.6%) and vancomycin-sensitive E. faecium (31.8%) compared to E. faecalis BSIs (23.2%) (p = 0.016). This result was confirmed by multivariable Cox analysis. Independent predictors of increased mortality included the PITT score, complicated bacteremia, and age (HR = 1.269, p < 0.001; HR = 1.818, p < 0.001; HR = 1.022, p = 0.005, respectively). Conversely, male gender, consultation with infectious disease (ID) specialists, and appropriate treatment were associated with reduced mortality (HR = 0.666, p = 0.014; HR = 0.504, p < 0.001; HR = 0.682, p = 0.026, respectively). In conclusion, vancomycin-resistant E. faecium bacteremia is independently associated with a higher risk of 30-day mortality.


Introduction
The Enterococcus species are a common cause of nosocomial bacteremia.In the United States, enterococci are the first most common bacteria causing central line-associated bloodstream infections (BSIs) in long-term acute-care hospitals [1].In 2019, the Enterococcus species ranked second among pathogens responsible for intensive care unit (ICU)-acquired BSIs in Europe [2].
Two species cause most of the enterococcal infections in humans: Enterococcus faecalis and Enterococcus faecium [3].Other species involved in human infections are as follows: E. casseliflavus, E. gallinarum, E. raffinosus, E. avium, and E. durans [4,5].Enterococci are not highly virulent bacteria.Typical enterococcal virulence factors are cytolysin, pili, gelatinase, aggregation substance, and extracellular surface proteins.These virulence factors contribute to the ability of enterococci to form biofilms [6].Most of them are absent in E. faecium.Moreover, E. faecalis and E. faecium have different patterns of acquired antimicrobial resistance, and these are more frequently observed in E. faecium, such as with penicillins and glycopeptides [3].Most E. faecium isolates exhibit ampicillin resistance, which is mainly due to the production of low-affinity penicillin-binding proteins (PBPs), especially PBP5.In contrast, high-level penicillin resistance in E. faecalis is much less common.While glycopeptide resistance can be found in both E. faecalis and E. faecium, it is more frequently associated with E. faecium [3,7,8].
Antimicrobial resistance (AMR) is known to be a leading cause of death around the world [9].Recently, the European Antimicrobial Resistance Collaborators conducted a cross-country systematic analysis about bacterial antimicrobial resistance in the World Health Organization (WHO) European region in 2019.They estimated that 541,000 deaths were associated with bacterial AMR, and 47,200 of these were attributable to bloodstream infections.E. faecium was one of the seven pathogens responsible for most deaths associated with AMR [10].AMR is also a serious threat to public health and national health systems.In this regard, nosocomial vancomycin-resistant enterococci (VRE) colonization [11] and infections significantly increase hospital costs [12].
A retrospective multicenter study, conducted in Italy from 2011 to 2017, found a progressive increase in the incidence of enterococcal bacteremia, and particularly those caused by vancomycin-resistant (VR) E. faecium.Resistance to ampicillin was detected in 6.8% and 89.1% of E. faecalis and E. faecium bacteremia cases, respectively.Resistance to vancomycin was detected in 1.3% and 14.1% of E. faecalis and E. faecium bacteremia cases, respectively.Resistance to tigecycline and linezolid was rarely observed [13].
Sources of enterococcal bacteremia usually are urinary and found in the gastrointestinal tract for community-acquired BSIs and intravascular and urinary catheters for hospital-acquired bacteremia [14].In almost 20% of cases, the source of infection is not identified [15].Most cases of enterococcal bacteremia are caused by E. faecalis, followed by E. faecium [15].Polymicrobial bacteremia is often observed, ranging from 25% to 50% of enterococcal BSIs, depending on the study considered, and is usually associated with abdominal sources of infections [15,16].The main risk factors for enterococcal BSIs are advanced age, immunosuppression, nosocomial infection under broad-spectrum antibiotics, prior enterococcal infections or colonization, recent surgery (mainly urinary or intra-abdominal), comorbidities related to urogenital and intra-abdominal organs, and presence of intravascular devices and/or indwelling urinary catheters [15].
Enterococcal BSIs are associated with high mortality rates, from 20% to 40% [15,17].Before the approval of effective drugs for VRE strains, such as daptomycin and linezolid, two systematic reviews compared the outcomes of VRE versus vancomycin-sensitive Enterococcus (VSE) bacteremia.Both studies found an increased risk of mortality for VRE bacteremia (relative risk [RR], 2.38; 95% confidence interval [CI] 2.13-2.66;odds ratio [OR], 2.52; 95% CI, 1.87-3.39)[18,19].In 2016, Prematunge et al. conducted a meta-analysis of 11 studies comparing VRE versus VSE bacteremia, confirming that VRE bacteremia is associated with an increased risk of in-hospital mortality and length of stay (LOS) [20].Kramer et al., in 2018, in a retrospective cohort study on patients with enterococcal bacteremia, reported that in-hospital mortality and infection-attributed hospital stay are not influenced by vancomycin resistance but by the Enterococcus species (E.faecium is an independent risk factor for in-hospital mortality) [21].More recently, a systematic review and meta-analysis found out a higher mortality for VR E. faecium bacteremia compared with vancomycinsensitive (VS) E. faecium BSI (RR 1.46; 95% CI 1.17-1.82),while no difference was observed when comparing VR E. faecium vs. VR E. faecalis BSI (RR 1.00; 95% CI 0.52-1.93)[22].Thus, according to the available studies, we cannot draw definitive conclusions about the outcome of enterococcal bacteremia.
The main aim of the present study is to investigate the risk factors for 30-day mortality for enterococcal BSIs, according to the Enterococcus species, resistance to vancomycin and appropriate treatment.

Results
During the study period, a total of 618 patients with enterococcal bacteremia were considered for inclusion.We excluded the following: 2 episodes for missing antimicrobial susceptibility test, 24 episodes because of species other than E. faecalis or E. faecium (10 Enterococcus casseliflavus, 5 Enterococcus gallinarum, 4 Enterococcus avium, 3 Enterococcus durans, 1 Enterococcus hirae, and 1 Enterococcus raffinosus), 5 episodes for concomitant E. faecalis and E. faecium, and 3 for missing species type.
Eventually, we included 584 patients with enterococcal bacteremia.Eleven patients had two separate episodes of enterococcal bacteremia (ten patients with the same species of Enterococcus, one patient had one E. faecalis BSI, and another one caused by E. faecium).

VRE Annual Prevalence
Only three E. faecalis BSIs, one observed in 2019 and two in 2020, were resistant to vancomycin.We excluded these patients from uni-and multivariate analysis due to the low number of strains isolated.A total of 93 vancomycin-resistant E. faecium were identified over the study period.The proportion of vancomycin-resistant E. faecium increased from 6.45% (n = 2/31) in 2014 to 51.06% (n = 24/47) in 2021 (Table 1).

Population Analysis According to Species and Vancomycin Susceptibility
Demographic and clinical characteristics of the included patients according to Enterococcus species and resistance to vancomycin are described in Table 2.
For the majority of cases of E. faecalis BSI, the source of infection was urologic, while an intra-abdominal focus was predominant in E. faecium BSIs (p < 0.001).
Hospital LOS before BSI diagnosis was significantly longer for vancomycin-resistant E. faecium, compared with other groups (22.3 ± 24.2 days vs. 13.3 ± 22.7 for VS E. faecalis and 16.6 ± 17.9 for VR E. faecium, p = 0.002).Previous chemotherapy and chronic immunosuppressive therapy were reported in 16.2% (p < 0.001) and 14.9% (p = 0.006) of VS E. faecium BSIs, respectively.Among patients receiving chronic immunosuppressive therapy (n = 59), seventeen were solid organ transplant recipients, and two were hematopoietic stem cell transplantation (HSCT) recipients.Only two patients had HIV infection.Between 2020 and 2021, nine patients with moderate/severe COVID-19 were diagnosed with enterococcal BSI.All of them received steroids during their hospital stay.Three patients received tocilizumab.
No statistically significant difference was found in complicated BSI rates according to enterococcal species and vancomycin resistance (p = 0.055).Approximately 59.8% of females had complicated bacteremia compared to 53.9% of males.Infective endocarditis was largely caused by E. faecalis (n = 28, p = 0.014).
We observed only nine cases of Clostridioides difficile infection within 60 days of discontinuing therapy.Relapse of bacteremia within 60 days of discontinuing therapy occurred in 17 patients (for seven of them, source control was not performed).

Population Analysis According to Mortality
Demographic and clinical characteristics of patients and bacteria according to 30 daymortality after BSI diagnosis are described in Table 3. Regarding the ward upon BSI diagnosis, 23.6% of patients were hospitalized in the ICU.Only 54% of them survived, while 30-day mortality rates in medical and surgical wards were lower (24.7% and 16.3%, respectively; p < 0.001).
Complicated BSIs were observed in 65.4% of deceased patients compared to 52.4% of surviving patients (p = 0.004).
Appropriate antimicrobial treatment was started in 41.4% of deceased patients compared to 56.4% of surviving patients (p < 0.001).ID specialist consultation was performed in 273 patients (47% of the total).Approximately 79.9% of surviving patients received an ID consultation, while the survival rate decreased in the group that did not receive it (p < 0.001).
The 30-day mortality was analyzed through a multivariable Cox model (Table 4).Cox regression analysis confirmed an adjusted higher 30-day mortality rate for vancomycinresistant E. faecium bacteremia compared to vancomycin-sensitive E. faecium BSIs and vancomycin-sensitive E. faecalis BSIs (Figure 1).The risk of death was higher for patients with complicated BSIs, higher PITT scores, and older age, while male gender, ID consultation, and appropriate antimicrobial treatment were predictive of lower mortality rates (Table 4).
For patients receiving an appropriate antimicrobial treatment, the sensitivity analysis confirmed a lowered risk for 30-day mortality only for patients with E. faecalis BSI (Figure 2).4).For patients receiving an appropriate antimicrobial treatment, the sensitivity analysis confirmed a lowered risk for 30-day mortality only for patients with E. faecalis BSI (Figure 2).

Discussion
The burden of enterococcal BSIs is increasing worldwide [1].Our country observed an increasing incidence of vancomycin-resistant E. faecium bacteremia over the last few years [13].We calculated the annual prevalence of vancomycin-resistant strains on the total enterococcal BSIs reported at our Institution between 2014 and 2021.We observed 93 vancomycin-resistant E. faecium BSIs, and only 3 vancomycin-resistant E. faecalis BSIs.The

Discussion
The burden of enterococcal BSIs is increasing worldwide [1].Our country observed an increasing incidence of vancomycin-resistant E. faecium bacteremia over the last few years [13].We calculated the annual prevalence of vancomycin-resistant strains on the total enterococcal BSIs reported at our Institution between 2014 and 2021.We observed 93 vancomycin-resistant E. faecium BSIs, and only 3 vancomycin-resistant E. faecalis BSIs.The prevalence of VR E. faecium bacteremia has increased from 6.45% in 2014 to 51.06% in 2021.However, most cases of BSIs were caused by E. faecalis (343 over a total of 584 bacteremia, 58.73%).
The origin of enterococcal bacteremia was identified mainly as urologic or intraabdominal for E. faecalis and intra-abdominal for E. faecium (both for VRE and VSE strains).The source of infection has not been identified in 19.4% of overall cases.These observations are in line with other cohort studies [23,24].Hospital LOS before BSI diagnosis was significantly longer for patients with vancomycin-resistant E. faecium BSIs.This reflects the current evidence of a major involvement of E. faecium in nosocomial infections rather than the community-acquired ones [20,25].
Another interesting point concerns polymicrobial bacteremia.Usually 25-50% of enterococcal bacteremia are polymicrobial and have an abdominal origin [15,16].In our study, we identified 187 cases of polymicrobial bacteremia, accounting for 31.9% of the total cases.In 99 instances, in addition to Enterococcus, bacteria from the Enterobacteriaceae family were also detected.The mortality rates for both polymicrobial and monomicrobial enterococcal bacteremia cases in our cohort were similar.As our data collection focused solely on therapies targeting Enterococcus species, we could not evaluate the appropriateness of antimicrobial treatment for the other isolated bacteria.However, our results are consistent with those reported by Lagnf et al. [16], in the only study known to us conducted with a focus on the outcome of polymicrobial enterococcal bacteremia.
Among major risk factors for enterococcal BSIs we have to consider advanced age, immunosuppression, and recent abdominal surgery [15].Considering our cohort, older patients had a higher 30-day mortality rate (still significant in multivariable analysis).Chronic immunosuppressive therapy and previous chemotherapy were significantly associated with E. faecium bacteremia (both vancomycin-sensitive and resistant), as highlighted previously [26].Recent abdominal surgery was not identified as a risk factor for enterococcal BSI and for 30-day mortality.
The risk factors for VRE bacteremia are mainly prior vancomycin use and VRE colonization [25].Glycopeptide exposure before BSI diagnosis was not identified as a risk factor for enterococcal VR E. faecium BSI and for 30-day mortality in our cohort.At our institution, rectal swabs for the detection of VRE colonization are not routinely carried out in every ward.Thus, we did not analyze this variable.VR Enterococcus faecium colonization is also a risk factor for C. difficile infections in particular populations, such as HSCT recipients [27].We reported only nine cases of CDI in our cohort.Consequently, we cannot stratify these data according to the Enterococcus species or vancomycin susceptibility.
Our study confirmed that the 30-day mortality rate was higher for vancomycinresistant E. faecium BSIs (30-day mortality rate: 36.6%)compared with vancomycin-sensitive E. faecium BSIs (30-day mortality rate: 31.8%) and vancomycin-sensitive E. faecalis BSIs (30-day mortality rate: 23.2%).The multivariate analysis confirmed these observations.In particular, the risk of death was 1.5 and 2.0 times higher for vancomycin-sensitive and vancomycin-resistant E. faecium BSIs, respectively.The 30-day mortality rate in the group of patients who received an ID specialist dropped by 50% compared with the group who did not receive it (HR = 0.504, p < 0.001).These data are not new, but they reinforce the importance of bundles for the management of enterococcal BSI [28].Male gender was associated with lower 30-day mortality (HR = 0.666, p = 0.014).Recently, a meta-analysis found a male/female ratio in VRE BSIs of 1.4 [29].To the best of our knowledge, there are no other studies that have observed the protective role of being male in enterococcal BSIs.This result could be due to the higher proportion of complicated bacteremia in females compared to that in males.In the multivariate analysis, complicated bacteremia was associated with a higher 30-day mortality rate.There is not a consensus for the definition of enterococcal complicated bacteremia.We must mention two potential biases when considering our definition of complicated BSI.Firstly, given the retrospective design of our study, we were not able to properly assess positive follow-up blood cultures because they are not routinely performed by clinicians.Secondly, we considered all primary bacteremia and all bacteremia without control of infections' sources as complicated.
While E. faecium is generally considered less virulent than E. faecalis [3], infections caused by E. faecium are associated with higher mortality rates and longer lengths of hospital stay [20], a trend corroborated by our study.Several potential explanations, though not conclusive, can be proposed.Firstly, the peculiar intrinsic and acquired antimicrobial resistance profile of E. faecium presents challenges in treatment [3].Secondly, patients with E. faecium infections tend to be more medically fragile and have a higher burden of comorbidities [20].Thirdly, given the relatively few therapeutic options for VRE BSIs, it is easier to miss the right empiric treatment compared to E. faecalis BSIs.Additionally, other less explored factors may contribute, such as disparities in biofilm formation between E. faecalis and E. faecium [30], as well as variations in host immune responses.
The multivariate analysis evidenced how an appropriate treatment was associated with a 30% reduction in the adjusted risk of death; however, this association was confirmed only for E. faecalis BSIs in the subgroup analysis.Inappropriate antibiotic therapy has already been identified as an independent risk factor for mortality in enterococcal bacteremia [31].Recently, Russo et al. observed that starting an appropriate therapy for VRE bacteremia within 48 h from blood culture collection was independently associated with improved survival [32].In this study, we did not assess the impact of different timing for appropriate treatment on 30-day mortality.
Our work has some limitations that deserve to be considered when interpreting the results.First, this is a retrospective study.Second, this study was conducted in only one hospital.Third, we did not consider active therapy against other isolated bacteria (for polymicrobial bacteremia) or stratify the analysis for different antimicrobial regimens.Finally, the definition we have chosen for appropriate therapy is arbitrary and not standardized.The strengths of this study are the long study period (from 2014 to 2021), the large sample (584 enrolled patients), and the fact that we made comparisons not only between VRE and VSE BSIs but also between E. faecalis and E. faecium BSIs.
Our future objectives are to enroll more patients through a multicentric study including enterococcal bacteremia from other Italian and European Hospitals, focus also on Enterococcus species other than E. faecalis or E. faecium, and assess the impact of different timing for appropriate treatment on 30-day mortality, according to different antimicrobial regimens.

Objectives of This Study
The aims of this study were as follows: (1) to calculate the annual prevalence of vancomycin-resistant strains on total enterococcal BSIs reported at our institution between 2014 and 2021; (2) to investigate risk factors for enterococcal BSIs, according to the Enterococcus species and resistance to vancomycin; and (3) to investigate risk factors for 30-day mortality, according to bacteria characteristics (Enterococcus species and resistance to vancomycin) and appropriate treatment.

Study Design and Population
We conducted a retrospective cohort study at Trieste University Hospital, Italy.All adult patients (aged > 18 years) hospitalized at our institution with at least one BC positive for Enterococcus faecalis or Enterococcus faecium during hospital stay were included.The study period ranges from 1 January 2014 to 31 December 2021.
Exclusion criteria were as follows: (1) pregnancy; (2) BCs positive for species other than E. faecalis and E. faecium; (3) lack of antimicrobial susceptibility test of isolated Enterococcus; (4) lack of Enterococcus typing; and (5) BCs positive for both E. faecalis and E. faecium at the same time.Additionally, duplicate BCs (up to 60 days following the last positive culture for the same Enterococcus spp.) from the same patient were excluded.

Data Collection and Definitions
The following data were retrospectively collected from hospital electronic medical records: demographics (age and gender); comorbidities; chronic therapy; previous exposure to glycopeptides (vancomycin and teicoplanin); date of hospital admission; date of first positive BC; result of the in vitro susceptibility testing; source of infection; ICU admission and PITT score [33] at BSI diagnosis; ID specialist consultation; treatment prescribed for enterococcal BSI; evidence of polymicrobial bacteremia and complicated bacteremia; date of hospital discharge and date of death; relapse of bacteremia and Clostridioides difficile infection within 60 days of discontinuing therapy.We defined a new BSI caused by the same organism within 60 days of clinical and microbiological resolution of a previously treated BSI as a relapse.After 60 days, we considered the new BSI as a separate episode.
According to the Centers for Diseases Control and Prevention (CDC), possible sources of infection were classified as catheter-related, urologic, intra-abdominal, heart/cardiovascular devices, bone/skin/soft tissue, and unknown.In this last case, the bacteremia was defined as primary [34].
Enterococcal bacteremia was defined complicated when at least one of the following features were present: (a) infective endocarditis, (b) device-associated infection, (c) metastatic infection, (d) source control not done, (e) positive follow up BCs after 48-72 h, and (f) persistency of fever after 48-72 h from first positive BC.In the case of primary bacteremia, the source control was automatically defined as not documented and, therefore, not done.Enterococcal bacteremia was defined as polymicrobial when at least one non-enterococcal bacterial species was isolated from the same blood culture as Enterococcus spp.and met the CDC criteria for bloodstream infection [34].
Appropriate antibiotic therapy was defined as an active therapy against isolated bacteria started within 24 h from BSI diagnosis and continued for at least five days.Antibiotic therapy was defined as active in accordance with in vitro isolate susceptibility.
The treatment prescribed for enterococcal BSI was documented as follows: time to empiric therapy; time to pathogen-specific therapy; duration of pathogen-specific therapy.
Antibiotic regimens that we considered appropriate are listed in Supplementary Materials S1.
All isolates were identified by MALDI-TOF mass spectrometry (bioMérieux, Marcy-l'Etoile, France), while antimicrobial susceptibility was assessed with the Vitek2 system (bioMérieux, Marcy-l'Etoile, France).Resistance to vancomycin was defined when a minimal inhibitory concentration > 4 was detected, according to the EUCAST criteria.
Mortality was defined as death of any cause within the 30 days following the first positive BC for E. faecalis or E. faecium.
All data were pseudonymized via a web-based central, password-protected clinical database management system.

Statistics
Continuous variables were presented as means ± standard deviations (SD).The between-group comparisons were analyzed via Student's t test for independent samples after determining whether or not equal variance could be attributed to the subgroups as per Levene's test.Nominal variables were shown as a number and percentage, and the respective contingency tables were analyzed using χ test or Fisher's exact test, as appropriate.
The prevalence of vancomycin resistance among E. faecium was calculated as the number of resistant strains over the total number of E. faecium isolates.
The 30-day mortality was analyzed according to bacteria characteristics (Enterococcus species and resistance to vancomycin) and appropriate treatment through multivariable Cox proportional hazards models with forward stepwise selection.The results were presented as an adjusted proportional hazard ratio (HR) and 95% confidence intervals (CIs).Aiming at examining the potential impact of survival bias among the patients receiving or not receiving an appropriate antimicrobial therapy, 30-day mortality was computed separately among bacteria subgroups for sensitivity analysis.
A p-value < 0.05 was set for statistical significance.All statistical analyses were performed using the software IBM SPSS Statistics, version 24.0 (New York, NY, USA: IBM Corp.).

Conclusions
The mortality rate of enterococcal bacteremia is high.Our study confirms that vancomycin-resistant E. faecium bacteremia is independently associated with a higher risk of 30-day mortality, and delayed appropriate antimicrobial treatment is associated with a higher mortality rate.However, the employment of ID specialist consultation and appropriate antimicrobial therapy, along with patients' male gender, were associated with significant lower mortality rates.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics13070601/s1.File S1.Appropriate Therapy.Informed Consent Statement: Informed consent was obtained from all subjects involved in this study.

Author Contributions:
Conceptualization, V.Z., and S.D.B.; methodology, V.Z., G.S., and S.D.B.; investigation, V.Z., R.P., D.A.S., E.P., V.T., J.M., M.B., L.S.C., D.A.T., and R.L.; writing, review and editing, V.Z., G.S., J.M., L.S.C., D.A.T., R.L., and S.D.B.; supervision, G.S., R.L., and S.D.B.All authors have read and agreed to the published version of the manuscript.Funding: The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.Institutional Review Board Statement: This study was performed in line with the principles of the Declaration of Helsinki.Approval was granted by the Ethics Committee of University of Trieste (Date 28 June 2022/N • V123).

Table 1 .
Annual prevalence of vancomycin-resistant strains on total enterococcal BSIs.

Table 2 .
Characteristics of patients according to bacterial species and vancomycin susceptibility.

Table 3 .
Demographic and clinical characteristics of patients and bacteria according to 30-day mortality.

Table 4 .
Results of Cox regression of 30-day mortality on study variables.

Table 4 .
Results of Cox regression of 30-day mortality on study variables.