LABORATORIO DE URGENCIAS

Risk factors and outcomes associated with persistent vancomycin resistant Enterococcal Bacteremia

                                                                    Emily Fox, David Ha, Mark Bounthavong , Lina Meng, Emily Mui,                                                                                  Marisa Holubar, Stanley Deresinski and William Alegria

                                                                    BMC Infectious Diseases (2022) 22:855 https://doi.org/10.1186/s12879-022-07864-8

BACKGROUND

Enterococcal bloodstream infections (BSIs) are associ[1]ated with signifcant mortality, with some estimates as high as 50% [1–6]. Vancomycin resistance has also been identifed as an independent predictor of all-cause and infection-related mortality in patients with enterococcal BSIs [3, 4]. Although no consensus defnition of “persis[1]tent” vancomycin resistant enterococcus (VRE) bacte[1]remia exists, multiple studies have demonstrated that BSIs that take four days or more to clear are indepen[1]dently associated with mortality [7–11]. A recently pub[1]lished prospective observational study similarly reported that failure of VRE bacteremia to clear within 4 days of the index culture was the strongest predictor of poor outcomes [8]. While  associated risk factors, similar studies are not available for VRE [12]. In addition to the lack of data to identify those at high[1]est risk of developing persistent VRE bacteremia, optimal management remains controversial [13–16]. Enterococci express intrinsic and acquired resistance mechanisms that signifcantly limit the number of treatment options available. Clinical uncertainty is further exacerbated when VRE bacteremia persists despite directed ther[1]apy. Recognizing these challenges, both the Centers for Disease Control and Prevention and the World Health Organization have designated VRE as a high-priority multi-drug resistant organism that poses a threat to pub[1]lic health. Given the association with increased mortality in patients with persistent VRE BSIs and the limited num[1]ber of treatment options available, further insight into risk factors associated with persistence is warranted. We aimed to identify risk factors associated with persistent VRE BSIs, including those related to patient manage[1]ment, such as antimicrobial selection and source control.

METHODS
Study design and population
We conducted a single center, retrospective, case-control study. Adult patients (≥18 years old) admitted to Stan[1]ford Hospital (Stanford, California, USA) between 2016 and 2020 with ≥1 blood culture positive for vancomy[1]cin resistant Enterococcus spp. were screened for inclu[1]sion. Case patients included those who had an episode of persistent bacteremia, defned as bacteremia for ≥4 days despite VRE-directed therapy. Control patients had non[1]persistent bacteremia, defned as documented clearance of blood cultures within 4 days of active treatment initia[1]tion. Patients were excluded if they had incomplete data records, did not have repeat blood cultures drawn after the initial index culture, did not receive VRE-directed therapy, or expired within 4 days of treatment initiation. Tis study was reviewed and deemed to be non-human subjects research by the Stanford University School of Medicine Panel on Human Subjects in Medical Research.
Data collection
Patient data was extracted from the medical record by a single trained reviewer, using a structured data collection form within REDCap (Research Electronic Data Cap[1]ture, Stanford University) [17, 18]. Patient demographics, including comorbidities, presence or absence of a central line or prosthetic device at the time of index blood cul[1]ture, infection site, antimicrobial management and out[1]comes were collected. Te Charlson comorbidity index (CCI) was used to provide a composite score of comorbid conditions. Te severity of bacteremia at the time of the initial positive VRE blood culture was assessed using the Pitt bacteremia score, with most signifcant values within 48 h of index culture.
Defnitions
Duration of bacteremia was defned as the number of days between the frst positive and last positive blood culture. Recurrent VRE BSI was defned as a new positive blood culture result for VRE following at least 14 days of negative blood cultures after receiving active ther[1]apy. Breakthrough VRE BSI was defned as initial clear[1]ance of bacteremia followed by a subsequent positive blood culture within 14 days of the initial index culture. Severe neutropenia was defned as an absolute neutrophil count<500 cells/mm3 . Source of VRE bacteremia and necessity for source control was determined via review of treating physicians’ notes and the available clinical and diagnostic data, including vital signs, laboratory and microbiology parameters, and imaging. Patients with identifed sources of infection were further evaluated to determine if a source control intervention was required. Source was considered uncontrolled if an intervention was necessary and not performed> < 500 cells/mm3 . Source of VRE bacteremia and necessity for source control was determined via review of treating physicians’ notes and the available clinical and diagnostic data, including vital signs, laboratory and microbiology parameters, and imaging. Patients with identifed sources of infection were further evaluated to determine if a source control intervention was required. Source was considered uncontrolled if an intervention was necessary and not performed.
Microbiologic data
Standard blood culture collection policy was to obtain two sets of blood cultures: one bottle of aerobic and one bottle of anaerobic culture medium in set one, and two bottles of aerobic culture medium in set two. Te frst positive blood isolate from each patient was used for microbiologic and molecular assessments. Screening for vancomycin resistance was performed by polymerase chain reaction detection of the vanA gene. Antimicrobial susceptibilities were subsequently determined using an automated system (MicroScan WalkAway Plus System (Beckman Coulter, Brea, CA)) and interpreted accord[1]ing to Clinical and Laboratory Standards Institute (CLSI) guidance at the time of the index blood culture. Dapto[1]mycin minimum inhibitory concentrations (MICs) of ≥4 mg/L were confrmed via E test.
Outcomes
Te primary outcome was risk factors associated with persistent VRE BSI. Secondary outcomes included in[1]hospital mortality and rates of breakthrough and recur[1]rent bacteremia.
Statistical analysis
Clinical characteristics were compared using inde[1]pendent t tests for continuous data and chi square tests for discrete data. A logistic regression model was con[1]structed to evaluate the association between specifc fac[1]tors for persistent bacteremia. Risk factors chosen inclusion in the multivariable logistic regression model were determined by plausibility of clinical signifcance as perceived by the researchers. Risk factors included in the regression model were age, gender, CCI, cirrhosis, his[1]tory of malignancy (hematologic and solid tumor), solid organ transplant history, severe neutropenia, Pitt bacte[1]remia score, ICU stay at index culture, 4 out of 4 positive blood cultures, lack of source control, and receipt of an oxazolidinone within the initial 72 h of treatment. Results were presented as odds ratios (OR) with corresponding 95% confdence intervals (CI). Model ft was assessed using the Hosmer-Lemeshow goodness-of-ft test, which evaluated the null hypothesis that the observed event rates were no diferent from the expected event rates [19]. Subgroup analyses were performed among patients who received daptomycin within the initial 72 h of treat[1]ment and source control interventions. Statistical signif[1]cance was set at a two-sided alpha<0.05. All statistical analyses were performed using Stata SE version 15 (Stata Corp., College Station, TX)> <0.05. All statistical analyses were performed using Stata SE version 15 (Stata Corp., College Station, TX)

RESULTS
Study population
During the study period, 138 patients had a blood culture positive for VRE. A total of 30 patients were excluded. Te most common reason for exclusion was death within four days of VRE index culture (Fig.  1). Of the patients that met inclusion criteria (n=108), 24 (22.2%) had per[1]sistent and 84 (77.8%) had non-persistent VRE bacte[1]remia. Te median (IQR) duration of bacteremia was 7 (6–11) days versus 1 (1–2) day in the persistent and non[1]persistent group, respectively.
Primary outcome
No signifcant diference in baseline demographics were identifed except for a higher number of patients on intermittent hemodialysis in the persistent group (33.3% versus 14.3%; p=0.034; Table 1). All patients in the per[1]sistent group and 96% (81/84) in the non-persistent group had blood cultures positive with Enterococcus fae[1]cium. More patients in the persistent group had 4 out of 4 blood culture bottles positive (54.2% versus 21.4%; p=0.002), and more patients in the non-persistent group had 1 out of 4 bottles positive (31.0% versus 8.3%; p=0.026; Table 1). Te only signifcant risk factor for persistence pertain[1]ing to management identifed in the bivariate analysis was an uncontrolled source of infection.Te most common source of infection in both groups was intra-abdominal, accounting for 50% (12/24) and 52.4% (44/84) of patients in the persistent and non-persistent groups, respectively. Other identifed sources of infection included central line-related (4.2% [1/24] in the persistent group and 9.5% [8/84] in the non-persistent group), undiferentiated (20.8% [5/24] in the persistent group and 26.2% [22/84] in the non-persistent group), and other (25% [6/24] in the persistent group and 11.9% [10/84] in the non-persistent group). Of those in the persistent group, 67% (16/24) of patients required a source control intervention versus 39% (33/84) of patients in the non-persistent VRE BSI group. In the subgroup analysis of those patients, 56% (9/16) underwent a source control intervention in the persistent group versus 91% (30/33) in the non-persistent group. Additionally, the median (IQR) time to source control intervention was 9 (2.5–17.5) days in the persis[1]tent group and 3 (1–4) days in the non-persistent group (p=0.0127; Additional fle  1: Table  S1). No diferences were observed between patients initially treated with daptomycin, oxazolidinones, or tetracyclines. Addition[1]ally, in the subgroup analysis of patients treated with dap[1]tomycin (n=62), no diferences were observed based on daptomycin MIC or the initial daptomycin weight-based dose utilized (Additional fle 1: Table S2). In the logistic regression analysis (Fig. 2), patients with severe neutropenia at initial index culture had a 12.12 higher odds of having persistent bacteremia compared to patients without severe neutropenia (95% CI 1.47, 100.26) while controlling for the other covariates. Four out of 4 positive index blood culture bottles were associ[1]ated with an 11.29 higher odds of having persistent bac[1]teremia (95% CI 2.42, 52.68). Lack of source control was associated with an 11.88 higher odds of having persis[1]tent bacteremia compared to those who obtained source control or had an infection that did not require a source control intervention (95% CI 2.48, 56.91). Te Hosmer[1]Lemeshow goodness-of-ft test p-value was 0.13, indi[1]cating that in our model the observed event rates did not deviate from the expected event rates, thus the the regression model ft the data well.
Secondary outcomes
The results of the unadjusted second[1]ary outcome analysis. In-hospital mortality was observed in 58% versus 40% of patients in the persistent and non[1]persistent groups (p=0.121). Tere were no statistically signifcant diferences in recurrent bacteremia (17% ver[1]sus 6%; p=0.090) or breakthrough bacteremia (13% ver[1]sus 7%; p=0.402) between groups. Discussion Prior studies that have identifed risk factors for persis[1]tent BSIs have almost exclusively focused on Staphylo[1]coccus aureus. However, persistent bacteremia due to VRE is also often encountered in clinical practice. In our cohort, 22% of patients experienced persistent bactere[1]mia for four days or more despite VRE-directed therapy, an incidence almost identical to that reported in a recent prospective multicenter study [10]. While other publica[1]tions have shown that persistent VRE bacteremia is an independent predictor of mortality, none have sought to elucidate risk factors for persistence [7–11]. As such, our fndings fll a much-needed gap in identifying those at high risk of persistent VRE BSI who may beneft from more aggressive management strategies. Delayed source control has been well-described as a risk factor for persistent Staphylococcus aureus bac[1]teremia [12]. Our fndings indicated similar results for VRE, as patients in the persistent group were more likely to have an infection source that required a source con[1]trol intervention, and the corresponding median time to intervention was also longer. Tis coincides with the breakdown of primary infection source between groups. Te non-persistent group had a higher incidence of unknown infection sources. In the absence of an obvious source, many of these were presumably secondary to gas[1]trointestinal translocation, where source control inter[1]ventions may not be indicated. Severe neutropenia was not a statistically signifcant risk factor in the unadjusted analysis, but it was sig[1]nifcant in the adjusted analysis when controlling for confounders. To our knowledge, neutropenia has not been previously described as a risk factor for persistent Staphylococcus aureus bacteremia. However, it has been identifed as a strong risk factor for VRE colonization and infection [1, 4]. Additionally, in a retrospective cohort analysis, Bhavnani et al. similarly found that patients with multiple positive blood cultures for E. faecalis or E. fae[1]cium had higher proportions of hematologic malignant neoplasms or neutropenia [20]. Our fndings expand upon these results and support the assertion that neutro[1]penia should be considered an important risk factor for persistent VRE bacteremia. To our knowledge, 4 out of 4 positive index blood cul[1]tures has never been identifed as a risk factor for persis[1]tent bacteremia. We collected number of positive blood culture bottles as a surrogate for overall bacterial burden. While acknowledging this is not a perfect marker, it can be easily applied in real-world clinical practice. In our study, 31.0% of patients in the non-persistent VRE BSI group had a single positive blood culture versus just 8.3% in the persistent group. Some clinicians may assert that a single positive blood culture for VRE may be more con[1]sistent with contamination rather than infection. Sexton et  al. previously suggested that the National Healthcare Safety Network (NHSN) defnition of primary Enterococ[1]cal BSI should require at least 2 positive blood cultures for this reason [21, 22]. While the signifcance of a single positive blood culture with Enterococcus may warrant additional study, our results do indicate that 4 out of 4 positive index blood cultures is associated with a higher risk of persistent VRE bacteremia.
Another interesting fnding was the lack of therapy implications on persistent VRE bacteremia. Within the subgroup of patients that received linezolid or tedizolid for management within the frst 72 h, there was no difer[1]ence in persistent bacteremia, suggesting that initial treat[1]ment with a “bacteriostatic” antimicrobial is not a risk factor for persistence. Studies comparing treatment with linezolid to daptomycin for VRE bacteremia have yielded conficting results, so these fndings were not unexpected [13–15]. In contrast, one management recommendation well-supported by existing literature, prompting a CLSI breakpoint revision in 2019, is that daptomycin doses of ≥8 mg/kg/day are required for optimal treatment of VRE (particularly when the MIC is 3–4  mg/L) [13].  Surpris[1]ingly, in our subgroup analysis of patients treated with daptomycin, there was not a signifcant diference in ini[1]tial dosages or daptomycin MIC between the persistent and non-persistent groups. Tis may be explained by low numbers of patients in the daptomycin subgroup analysis (particularly in the persistent VRE BSI group).
We observed a numerically higher incidence of mortal[1]ity, recurrent bacteremia, and breakthrough bacteremia in the persistent VRE group, but this did not meet statistical signifcance. Tat said, the 2009 Infectious Disease Soci[1]ety of America guideline for intravascular catheter-related infections state that Enterococcal bacteremia persisting>4 days is associated with mortality. [9]. Additionally, multi[1]ple prior studies have found that persistent Enterococcal BSI is an independent predictor of poor outcome. Tis was most recently confrmed in the study by Contreras et al., in which microbiologic failure (at ≥4 days from index cul[1]ture) had a HR of 5.03 for in-hospital mortality [7–11]. Te lack of a statistically signifcant diference in mortality between groups observed in our study may be due to Type II error owing to limitations of our sample size.
Tis study has several limitations. First, this was a retro[1]spective single-center study with a relatively small sample size and therefore may not be generalizable to all patient populations. Only a limited number of variables could be included in the multivariable analysis to avoid over-ftting.
Because of this, only variables considered to have the high[1]est plausibility of clinical signifcance were included. Sec[1]ond, 25% of VRE BSI cases had an unknown source, as it was difcult in some cases to determine retrospectively. Tis is a common limitation given the frequency of VRE BSI cases secondary to gastrointestinal translocation, and other retrospective Enterococcal BSI studies have reported a similar incidence of ~10–36% of patients with VRE BSI secondary to unknown source [5, 8]. Tird, blood culture practices were heterogeneous throughout the study period. Repeat blood cultures were not always drawn at the same intervals, which could afect the duration of bacteremia reported. Additionally, some patients only had one set of blood cultures drawn, and not all institutions may utilize our same blood culture collection approach, in which one set may constitute two aerobic bottles or one anaerobic and one aerobic bottle. Fourth, while we evaluated the number of positive blood culture bottles, we did not capture the time from specimen collection to blood culture positivity, which has been associated with severity of infection. Future studies should consider evaluating time to positivity as a marker for persistent VRE bacteremia. Finally, “persistent” VRE BSI has not been formally defned, and the defnition used in this study may be controversial amongst practi[1]tioners. Tis, however, would not be unique to VRE, as the defnition of persistent Staphylococcus aureus BSI has been heavily debated and studied extensively, and still remains variable in the literature ranging anywhere from 2 to 7 days [23]. Nonetheless, the defnition used in our study was sup[1]ported by the mortality impact of ≥4 days of persistent VRE BSI noted in existing studies.

CONCLUSION
Our results indicated that 22% of patients with VRE BSI, particularly those with severe neutropenia, 4 out of 4 positive index blood culture bottles, and an uncon[1]trolled source had persistent bacteremia despite 4 days of appropriate antibiotic treatment. We noted a numeri[1]cally higher incidence of mortality, recurrent bacteremia, and breakthrough bacteremia in the persistent VRE BSI group. Larger prospective studies should be conducted to confrm the mortality impact and explore manage[1]ment strategies for persistent VRE BSI. While awaiting further data, clinicians may consider aggressive VRE BSI management, such as early source control interventions, for neutropenic patients or those with 4 out of 4 positive index blood culture bottles who are at risk for persistent VRE bacteremia

NOTE: figures and tables in original article.

REFERENCES

1. Vydra J, et al. Enterococcal bacteremia is associated with increased risk of mortality in recipients of allogeneic hematopoietic stem cell transplantation.ClinInfect Dis.2012;55:764–70.
   2. Noskin GA, Peterson LR, Warren JR. Enterococcus faecium and Enterococcus faecalis Bacteremia: acquisition and outcome. Clin Infect Dis. 1995;20:296–301.
   3. Thomas P, Peggy L, Vincent HT, Rybak MJ. Clinical outcomes for patients with bacteremia caused by vancomycin-resistant enterococcus in a level 1 trauma center. Clin Infect Dis. 2002;34:922–9.
   4. Billington EO, Incidence, et al. Risk factors, and outcomes for Enterococcus spp. blood stream infections: a population-based study. Int J Infect Dis. 2014;26:76–82.
   5. Zasowski EJ, Claeys KC, Lagnf AM, Davis SL, Rybak MJ. Time is of the essence: the impact of delayed antibiotic therapy on patient outcomes in hospital[1]onset enterococcal bloodstream infections. Clin Infect Dis. 2016;62:1242–50.
   6. Ye J-J, et al. Clinical characteristics and treatment outcomes of vancomycin[1]resistant Enterococcus faecium bacteremia. J Microbiol Immunol Infect. 2018;51:705–16.
   7. DiazGranados CA, Zimmer SM, Klein M, Jernigan JA. Comparison of mortal[1]ity associated with vancomycin-resistant and vancomycin-susceptible enterococcal bloodstream infections: a meta-analysis. Clin Infect Dis. 2005;41:327–33.
   8. Contreras GA, et al. Contemporary clinical and molecular epidemiology of vancomycin-resistant enterococcal bacteremia: a prospective multicenter cohort study (VENOUS I). Open Forum Infect Dis. 2021. https://doi.org/10. 1093/ofd/ofab616.
   9. Mermel LA, et al. Clinical practice guidelines for the diagnosis and manage[1]ment of intravascular catheter-related infection: 2009 update by the Infec[1]tious Diseases Society of America. Clin Infect Dis. 2009;49:1–45.
   10. Chuang Y-C, et al. A retrospective clinical comparison of daptomycin vs daptomycin and a beta-lactam antibiotic for treating vancomycin-resistant Enterococcus faecium bloodstream infections. Sci Rep. 2018;8:1632.
   11. Bhavnani SM, et al. A nationwide, multicenter, case-control study com[1]paring risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia. Diagn Microbiol Infect Dis. 2000;36:145–58.
   12. Chong YP, et al. Persistent Staphylococcus aureus bacteremia: a prospective analysis of risk factors, outcomes, and microbiologic and genotypic charac[1]teristics of isolates. Med (Baltim). 2013;92:98–108. vanco[1]mycin-resistant enterococcal bloodstream infections?*. Crit Care Med. 2018;46:1700–3.
   13. Kanjilal S, Kalil AC, Klompas M. What is the best treatment for vanco[1]mycin-resistant enterococcal bloodstream infections?*. Crit Care Med. 2018;46:1700–3.
   14. Patel R, Gallagher JC. Vancomycin-resistant enterococcal bacteremia phar[1]macotherapy. Ann Pharmacother. 2015;49:69–85.
   15. Barber KE, King ST, Stover KR, Pogue JM. Therapeutic options for vancomy[1]cin-resistant enterococcal bacteremia. Expert Rev Anti-infective Therapy. 2015;13:363–77.
   16. Miller WR, Murray BE, Rice LB, Arias CA. Vancomycin-resistant enterococci: therapeutic challenges in the 21st century. Infect Dis Clin North Am. 2016;30:415–39.
   17. Harris PA, et al. The REDCap consortium: building an international commu[1]nity of software platform partners. J Biomed Inform. 2019;95:103208.
   18. Harris PA, et al. Research electronic data capture (REDCap)—a metadata[1]driven methodology and workfow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–81.
   19. Hosmer DW, Lemesbow S. Goodness of ft tests for the multiple logistic regression model. Commun Stat Theory Methods. 1980;9:1043–69.
   20. Bhavnani SM, et al. A nationwide, multicenter, case-control study compar[1]ing risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia☆. Diagn Microbiol Infect Dis. 2000;36:145–58.
   21. Freeman JT, Chen LF, Sexton DJ, Anderson DJ. Blood culture contamination with Enterococci and skin organisms: implications for surveillance defni[1]tions of primary bloodstream infections. Am J Infect Control. 2011;39:436–8.
   22. Sexton DJ, Chen LF, Anderson DJ. Current defnitions of central line–associ[1]ated bloodstream infection is the emperor wearing clothes? Infect Control Hosp Epidemiol. 2010;31:1286–9.
   23. Kuehl R, et al. Defning persistent Staphylococcus aureus bacteraemia: secondary analysis of a prospective cohort study. Lancet Infect Dis. 2020;20:1409–17.