Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

VRE and VSE Bacteremia Outcomes in the Era of Effective VRE Therapy: A Systematic Review and Meta-analysis

VRE and VSE Bacteremia Outcomes in the Era of Effective VRE Therapy: A Systematic Review and... infection control & hospital epidemiology january 2016, vol. 37, no. 1 original article VRE and VSE Bacteremia Outcomes in the Era of Effective VRE Therapy: A Systematic Review and Meta-analysis 1 1 1,2,3 1 Chatura Prematunge, MSc; Colin MacDougall, MSc; Jennie Johnstone, MD, PhD; Kwaku Adomako, MSc; 1 1 1,3,4,5 Freda Lam, MPH; Jennifer Robertson, PhD; Gary Garber, MD background. Prior data suggest that vancomycin-resistant Enterococcus (VRE) bacteremia is associated with worse outcomes than vancomycin-sensitive Enterococcus (VSE) bacteremia. However, many studies evaluating such outcomes were conducted prior to the availability of effective VRE therapies. objective. To systematically review VRE and VSE bacteremia outcomes among hospital patients in the era of effective VRE therapy. methods. Electronic databases and grey literature published between January 1997 and December 2014 were searched to identify all primary research studies comparing outcomes of VRE and VSE bacteremias among hospital patients, following the availability of effective VRE therapies. The primary outcome was all-cause, in-hospital mortality, while total hospital length of stay (LOS) was a secondary outcome. All meta-analyses were conducted in Review Manager 5.3 using random-effects, inverse variance modeling. results. Among all the studies reviewed, 12 cohort studies and 1 case control study met inclusion criteria. Similar study designs were combined in meta-analyses for mortality and LOS. VRE bacteremia was associated with increased mortality compared with VSE bacteremia among cohort studies (odds ratio [OR], 1.80; 95% confidence interval [CI], 1.38–2.35; I = 0%; n= 11); the case-control study estimate was similar, but not significant (OR, 1.93; 95% CI, 0.97–3.82). LOS was greater for VRE bacteremia patients than for VSE bacteremia patients (mean difference, 5.01 days; 95% CI, 0.58–9.44]; I = 0%; n= 5). conclusions. Despite the availability of effective VRE therapy, VRE bacteremia remains associated with an increased risk of in-hospital mortality and LOS when compared to VSE bacteremia. Infect. Control Hosp. Epidemiol. 2016;37(1):26 –35 Enterococcus spp. are typically commensal organisms, Whether outcomes associated with VRE bacteremia are 1,2 common in the human gastrointestinal tract, but in some worse than those associated with vancomycin-sensitive circumstances can cause serious infections including bacteremia, enterococci (VSE) bacteremia remains unclear. Two prior particularly among hospitalized patients with underlying systematic reviews have compared outcomes of VRE 1,2 comorbid conditions. Since its discovery in 1988, bacteremia VSE bacteremia; both found VRE bacteremia to be vancomycin-resistant enterococci (VRE) have emerged as associated with an increased risk of mortality when compared important nosocomial pathogens and are occurring with to VSE bacteremia (relative risk [RR], 2.38; 95% confidence increasing frequency due to widespread use of antibiotics, interval [CI], 2.13–2.66; odds ratio [OR], 2.52; 95% CI, prolonged hospitalizations, and increased intensive care unit 1.87–3.39 ). However, both of these systematic reviews (ICU) admissions, especially among patients with malignant included studies conducted prior to the availability of effective 1–3 5,6 health conditions. In Canada, the incidence of VRE infections VRE therapies. Since late 1999, a number of antibiotic drugs has risen to 0.5 infections per 1,000 admissions, a 6-fold have been licensed as treatment for VRE bacteremia by the increase in recent years. Similarly in the United States, US Food and Drug Administration (FDA), Health Canada, 1,2,7 hospitalizations with VRE infection reached 0.6 per 1,000 and other national approval agencies. Quinupristin- admissions by 2006. dalfopristin was approved in 1999, followed by linezolid Affiliations: 1. Infection Prevention and Control, Public Health Ontario, Toronto, Ontario, Canada; 2. St. Joseph’s Health Centre, Toronto, Ontario, Canada; 3. Department of Medicine, University of Toronto, Toronto, Ontario, Canada; 4. Department of Medicine, University of Ottawa, Ottawa, Ontario, Canada; 5. Ottawa Hospital Research Institute, Ottawa, Ontario, Canada. PREVIOUS PRESENTATION. The preliminary findings of this systematic review and meta-analysis were presented at Association of Medical Microbiology and Immunology (AMMI) Annual Conference, Quebec, Canada, in April 2015 as an oral presentation. Received April 30, 2015; accepted August 30, 2015; electronically published October 5, 2015 © 2015 by The Society for Healthcare Epidemiology of America. All rights reserved. This is an Open Access article, distributed under the terms of the Creative Commons Attribution license (http://creativecommons.org/licenses/ by/3.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited. All rights reserved. 0899-823X/2016/3701-0006. DOI: 10.1017/ice.2015.228 vreand vsebacteremia outcomes after vre therapy 27 7 1,2,7,8 in 2000. In 2003, daptomycin was formally licensed for telavancin for treating any part of the illness. Penicillin, complicated skin and soft tissue VRE infections, but it is ampicillin, amikacin, streptomycin, chloramphenicol, doxycy- frequently used as an off-label therapy for VRE bacteremia. cline, rifampin, imipenem-cilastatin, and nitrofurantoin were Thus, understanding whether VRE bacteremia-associated not considered effective VRE treatments. outcomes are different from those of VSE bacteremia, since the To capture standard, off-label, and compassionate study emergence of effective VRE therapy, is critically important to use of effective VRE treatment(s), literature published after help inform future VRE infection control recommendations. January 1997 was considered. Studies analyzing data collected To this end, we performed a systematic review and between January 1997 and January 2000 were excluded if the meta-analysis of studies comparing outcomes of patients antibiotics used for the treatment of VRE bacteremia patients with either VRE or VSE bacteremia, when patients with VRE were not reported or could not be obtained by contacting bacteremia were treated with effective VRE therapy. study authors. Studies conducted after January 2000 were assumed to have administered effective VRE treatment(s) and were included in the review. methods Narrative reviews, case series, case reports, and commentaries All methods including literature searches, study selection, data were excluded. Only the most recent peer-reviewed collection, and quantitative analysis processes were developed a publication was included when multiple reports using the priori and were reported according to the Preferred Reporting same study data existed. We limited our review to English Items for Systematic Reviews and Meta-Analyses (PRISMA) language articles. guidelines and the Cochrane Handbook for Systematic Reviews of 9,10 Intervention. Study Selection and Data Extraction Titles and abstracts of articles captured by literature searches Search Methodology and Data Sources were independently screened in duplicate by two reviewers The Public Health Ontario (PHO) Library Services department (CP and CM). Articles flagged for full-text review by either assisted with the development and implementation of search reviewer were included in the full-text review, and the full-text strategies for electronic databases, as well as with the retrieval of review process was duplicated and independently completed full-text articles. Medline, Embase, CINAHL, ProQuest by the same reviewers. Inter-rater reliability following full-text dissertations and theses, and the Cochrane Central Register of review was calculated using Cohens Kappa statistic and any Controlled Trials (CENTRAL) databases were searched from disagreements on study inclusion were resolved via arbitration January 1997 to December 2014. A sample search strategy is by a third reviewer (JJ). provided in Supplemental Table 1. Websites of infection control authorities and proceedings from infection control conferences Quality Assessment held within the most recent 5 years (ie, January 1, 2010, to Data extraction and quality assessment for included studies were January 1, 2015) were searched as outlined in Supplemental performed in duplicate (by CP and CM). An electronic data Table 2. Conference proceedings prior to 2010 were not extraction template was developed, pilot tested, and refined considered because we assumed that valuable data contained prior to the initiation of data extraction. The extracted data within such abstracts had become available in peer-reviewed elements included study design, sample size, study period, literature. Additionally, the reference lists of all relevant study setting, study population, patient type, study location, publications were hand searched to identify additional citations. Enterococcus spp., VRE/VSE bacteremia definition, VRE therapy administered, and number of VRE and VSE bacteremia patients Study Inclusion Criteria with the above stated outcome(s) of interest along with The study inclusion criteria for the review were randomized associated effect estimates and confidence intervals. Whenever controlled trials (RCTs), cohort studies, case-control studies, required information was not reported, attempts were made and cross-sectional studies, sampling adult (≥18 years of age) to contact the first and/or corresponding authors to and/or pediatric (<18 years of age) hospital patients, diag- obtain missing information; after 2 attempts, authors were nosed with VRE bacteremia and treated with effective VRE considered unresponsive. Data requests were limited to therapy, alongside VSE bacteremia patient comparators, and missing information on administered VRE treatment type(s), reporting on various mortality and morbidity outcomes. primary outcomes, and any secondary outcomes reported The primary outcome of interest was all-cause in-hospital within the primary study. mortality. Secondary outcomes were bacteremia-attributable Study quality was assessed using the Newcastle-Ottawa Scale mortality, total hospital length of stay (LOS), total intensive (NOS) scale or Cochrane risk of bias tool. The NOS was used care unit (ICU) LOS, post-VRE/VSE bacteremia diagnosis to establish quality of evidence within non-randomized cohort hospital LOS, and post-VRE/VSE bacteremia diagnosis ICU or case control studies, via a 9-star system. A study awarded a LOS. Effective VRE therapies were defined as quinupristin- greater number of stars is considered to be of higher dalfopristin, linezolid, daptomycin, tigecycline, teicoplanin, and methodological study quality. Although we did not 28 infection control & hospital epidemiology january 2016, vol. 37, no. 1 anticipate finding any RCTs, the Cochrane risk of bias tool December 2008. However, these authors confirmed that all was assigned to assess RCT study quality in the event an patients with VRE bacteremia were diagnosed after January 10 16 RCT meeting inclusion criteria was discovered. 2000. Billington et al sampled all residents within a Canadian health zone who developed enterococcal bloodstream infec- tions. We contacted these authors to obtain mortality Data Analysis and LOS information for study participants. In addition, Outcome effect measures for each study were calculated using 8 studies exclusively sampled adult patients within tertiary care 15,17–19,21,22,24,26 numbers of patients with VRE and VSE bacteremia with the hospital settings, and 4 of these studies were 19,22,25,26 outcome(s) of interest. We pooled studies of the same study limited to immunocompromised patient populations design via inverse variance method and random effects such as hematopoietic stem cell transplant patients or modeling in Review Manager 5.3; summary effect measures chemotherapy recipients. are reported as odds ratios (OR) and 95% confidence intervals All included studies defined patients with at least 1 VRE- or 15–27 (CI) for mortality, and mean difference and standard deviation VSE-positive blood culture to be cases of bacteremia. Both (SD) are reported for continuous LOS outcomes. When the E. faecalis and E. faecium were captured in 12 study 15–25,27 26 median and interquartile ranges (IQR) were reported, the samples, but the study by Yoo et al only included median was assumed to reflect the mean, and IQR was E. faecium infections. Outcome data for 2,575 bacteremias, assumed to be 1.35 SD. Statistical heterogeneity was assessed specifically 1,863 VSE and 712 VRE bacteremia cases, were using the I statistic, and VRE and VSE bacteremia outcomes identified in our literature review. were further explored via planned subgroup analyses of the following patient populations: (1) adult versus pediatric Outcomes patients, (2) immunocompromised versus varied immune status patients, ICU versus non-ICU admissions, (3) multi- Of the reviewed studies, 12 studies were cohort studies and 1 was center versus single study site, and low versus moderate- a case control study. When in-hospital mortality from the to-high study quality for included cohort studies. cohort studies were combined using unadjusted analysis, VRE Publication bias was examined via the visual interpretation of bacteremia was associated with an increased risk of in-hospital funnel plot symmetry and limited to the mortality outcomes. death when compared to VSE bacteremia with no heterogeneity (OR, 1.80; 95% CI, 1.40–2.32; I = 0%; n= 12) (Figure 2). The Role of the Funding Source single case-control study did not report a statistically significant increase in risk of VRE bacteremia death when compared with The design, conduct, and reporting for this systematic review VSE bacteremia in an unadjusted analysis (OR 1.93; 95% CI, and meta-analysis was funded by the Ontario Agency for 0.97–3.82) ; adjusted analyses were not reported. Health Protection and Promotion (Public Health Ontario). Of the 12 cohort studies, 5 reported adjusted analyses for 15,16,18,19,21 in-hospital mortality risk, and 2 studies found VRE results 18,19 bacteremia to be associated with adjusted mortality. Cheah Literature Search et al adjusted for prior ICU admission, comorbidities measured by the Charlson Comorbidity Index, Enterococcus sp., The literature searches identified 4,878 citations; among these, additional non-enterococcal infections, time to effective therapy, 155 citations were chosen for full-text review, and 20 studies and VRE bacteremia (OR, 1.21; 95% CI, 0.53–2.79]) via logistic were determined to meet our inclusion criteria (Figure 1). Of regression analysis. Cho et al adjusted for severity of these, 1 study did not indicate a study period and 5 studies illness using Simplified Acute Physiology Index, length of reported study periods between January 1997 and January hospitalization, and vancomycin resistance (hazard ratio [HR], 2000 and required confirmation of VRE therapies within 0.75; 95% CI, 0.24–2.36) via Cox proportional hazards each study. Corresponding authors were contacted, but modeling. VRE bacteremia was not included in the final models administered VRE therapy information could not be obtained 15,16,21 of the remaining 3 studies reporting adjusted mortality. and all 7 studies were excluded from the review. Excluded The study by Cho et al was the only study to report on VRE/ study details are provided in Supplemental Table 3. Therefore, VSE bacteremia-attributable mortality, which was defined 13 studies were included in the systematic review. as death within 7 days of bacteremia when no other cause could be identified. There was no significant difference in Description of Studies attributable mortality risk between VRE and VSE bacteremia The study characteristics of all included studies are outlined in patients in the unadjusted analysis (6 of 24 patients with VRE Table 1. All were observational and retrospective studies, 12 of bacteremia vs 15 of 67 patients with VSE bacteremia; OR, 1.15; which were conducted between January 2000 and December 95% CI, 0.39–3.43). 2011, following the formal regulatory approval of the first Total hospital LOS data were reported within 6 studies. Data 15–26 27 17 18 effective VRE therapy. The study by da Silva et al reported by Butler et al and Cheah et al, and data obtained 16 27 reported a study period between September 1998 and by contacting authors of Billington et al, da Silva et al, and vre and vse bacteremia outcomes after vre therapy 29 figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flowchart of the literature search and study selection. table 1. Characteristics of Studies Included in Systematic Review and Meta-Analysis Sample Size, No. Study Study Period Location Patient Population VRE VSE Cohort studies Bar et al, 2006 Nov 2000–Dec 2002 Richmond, VA USA Adult 17 33 16 a Billington et al, 2014 2000–2008 Calgary, Canada Mixed 27 640 Butler et al, 2010 Jan 2002–Dec 2003 St Louis, MO USA Adult, Non-surgical, >2 days LOS 94 182 Cheah et al, 2013 Jan 2002–March 2010 Victoria, Australia Adult, >2 days LOS 116 116 Cho et al, 2013 July 2009–Dec 2011 Seoul, Korea Adult, neutropenia post CHEMO or SCT 24 67 27 c c da Silva et al, 2014 Sep 1998–Dec 2008 Sao Jose do Rio Preto, Brazil Mixed 30 273 20 a Haas et al, 2010 2001–2006 Philadelphia, PA USA Pediatrics 39 300 Marschall et al, 2013 Jan 2006–Dec 2006 St. Louis, MO USA Adult, CVC associated bacteremias 67 39 22 a Mikulska et al, 2012 2004–2011 Genoa, Italy Adult, allogeneic HSCT 9 58 23 b Mohr et al, 2009 Jan 2000–Dec 2009 58 sites, USA Mixed, dap Tx. 151 211 Vydra et al, 2012 Jan 2004–Dec 2008 Minneapolis, MN USA Mixed, HSCT 50 43 26 D Yoo et al, 2004 Jan 2000–Dec 2001 Seoul, Korea Adult, HSCT or cytotoxic CHEMO 19 8 Case control study Peel et al, 2011 Jan 2000–Dec 2009 Victoria, Australia Adult 80 360 NOTE. LOS, length of stay; HSCT, hematopoietic stem cell transplantation; CHEMO, chemotherapy; dap Tx, daptomycin treatment; CVC, central venous catheter; SCT, stem cell transplantation. Months not reported. Assumed to be mixed, unconfirmed due to demographics being reported as ≤ 30 years of age. Data obtained by contacting study authors. A total of 8 VRE patients received VRE therapies and were included in the review. 30 infection control & hospital epidemiology january 2016, vol. 37, no. 1 figure 2. VRE and VSE bacteremia unadjusted in-hospital mortality risk by study design. Results of included studies for VRE and VSE bacteremia unadjusted in-hospital mortality risk stratified by study design. Abbreviations: 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. figure 3. VRE and VSE bacteremia total hospital LOS mean difference. Results of studies reporting on VRE and VSE bacteremia total hospital LOS. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. figure 4. VRE and VSE post-bacteremia total hospital LOS mean difference. Results of studies reporting on VRE and VSE post-bacteremia hospital LOS. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. Haas et al were pooled; VRE bacteremia was associated development of clinically significant bacteremia, which is with a longer LOS than VSE bacteremia (mean difference, different from the LOS definition used in our review. 2 26 5.01; 95% CI, 0.58–9.44; I = 0%; n = 5) (Figure 3). Data Data from Yoo et al were excluded because their LOS from Cho et al were excluded because they defined LOS estimates combined patients treated with effective and as the number of days from hospital admission to the noneffective VRE therapy. vre and vse bacteremia outcomes after vre therapy 31 Post-bacteremia LOS data reported by Cheah et al and Haas bacteremia compared with VSE bacteremia. The mortality et al were also pooled via a meta-analysis. There was no significant summary estimate demonstrated lack of heterogeneity across difference in LOS following a VRE bacteremia compared with studies and no significant influence on the point estimate by VSE bacteremia (mean difference, 0.53 [95% CI –8.98, 10.04]; age, immune status, study site(s), or study quality. VRE I = 26%; n= 2) (Figure 4). Yoo et al also reported on bacteremia was also associated with increased total hospital post-bacteremia LOS, but data were omitted because estimates LOS and post-bacteremia LOS with no heterogeneity. The combined patients treated with both effective and noneffective post-bacteremia LOS estimate was not statistically significant, VRE therapy. which may have been due to lack of statistical power influenced by the small number of studies reporting on post Subgroup Analyses bacteremia LOS outcomes. Our finding, that there is an increased risk of mortality and No significant interactions were detected between any of the LOS associated with VRE bacteremia when compared to VSE subgroups we had planned to analyze for in-hospital 5,6 bacteremia, is consistent with 2 previous systematic reviews. mortality including age (pediatric patients [OR, 1.62; 95% In the systematic review by Salgado et al, the authors CI, 1.18–2.22] vs adult [OR, 1.93; 95% CI, 0.89–4.18]; speculated that the increased morbidity and mortality could be interaction P = .68), immune status (immunocompromised because patients with VRE bacteremia were more likely to patients [OR, 1.24; 95% CI, 0.65–2.35] vs varied immune 5,6 receive ineffective therapy. However, our findings suggest status [OR, 1.93; 95% CI, 1.47–2.54]; interaction P = .21), that a lack of effective therapy is not the explanation. It should study site (single center studies [OR, 1.85; 95% CI, 1.37–2.50] be noted that our systematic review was unable to capture time vs multicenter studies [OR, 1.70; 95% CI, 1.12–2.58]; to effective therapy. Thus, it is possible that patients with VSE interaction P= .75) and study quality (low-quality studies bacteremia received effective therapy sooner than patients with [OR, 0.36; 95% CI, 0.04–3.11] vs moderate- to high-quality VRE bacteremia because VRE may be less likely to be covered studies [OR, 1.84; 95% CI, 1.43–2.37]; interaction P= .14) by empiric therapy, and effective therapy may only have (Figure 5). The planned subgroup analysis for ICU stay was been administered following a VRE-positive microbiological not performed due to unavailable data. 6,15,18,19 culture result. Age was not found to significantly influence total hospital An alternative explanation for the observed increase in LOS by subgroup analysis (pediatric patients [OR −13.00; 95% mortality risk and LOS could be differences in illness severity CI, −39.90–13.90] vs adult [OR, 6.12; 95% CI, 0.82–11.42]; or comorbidities between patients with VRE and VSE interaction P= .17) (Figure 6). The remaining LOS subgroup bacteremia, particularly because patients with VRE bacteremia analyses could not be performed due to a lack of studies in each 18,24,28 may have more comorbidities. Due to limited reporting companion subgroup. No significant interaction was detected of adjusted mortality and morbidity risks among included in the subgroup analysis of post-bacteremia LOS by age studies, we were unable to calculate adjusted summary (pediatric patients [OR, −9.0; 95% CI, −28.13–10.13] vs adult estimates in this systematic review. Thus, the effect of [OR, 3.0; 95% CI −3.37–9.37]; interaction P = .24) (Figure 4). confounding factors on our unadjusted mortality and LOS summary estimates remains unclear. We hypothesize that not Study Quality adjusting for potential confounders (ie, comorbid conditions Study quality ratings based on NOS criteria are presented in and severity of illness) may lead to overestimates of our Table 2. Of the 13 studies reviewed, 12 were of moderate to associations of interest because patients colonized with VRE high study quality, with the most frequent number of stars tend to have more comorbid conditions and more severe awarded per study being 6 or 7. Among all studies, patients illness than patients not colonized with VRE. with VRE and VSE bacteremia were selected from the same However, the earlier systematic review by DiazGranados hospital population, and bacteremia diagnosis was confirmed et al, which only considered studies controlling for underlying by patient chart reviews or microbiology reports. severity of illness, found VRE bacteremia adjusted mortality risk to be greater in comparison to VSE bacteremia. The worse outcomes associated with VRE bacteremia Publication Bias compared to VSE bacteremia may also be linked to differences The asymmetrical funnel plot indicates that the review’s in the causative species as there may have been proportionately in-hospital mortality estimates may be subject to publication more patients with E. faecium than E. faecalis in the bias (Figure 7). VRE bacteremia group when compared to the VSE bacteremia 1,2 group. Our results should be interpreted recognizing the systematic discussion review’s limitations. First, studies included within each In this systematic review and meta-analysis, we found that meta-analysis were non-randomized observational studies, since the advent of effective VRE therapy, there remains an and accordingly, our results reflect association and not increased risk of in-hospital mortality associated with VRE causation. Second, as discussed above, only a small number of 32 infection control & hospital epidemiology january 2016, vol. 37, no. 1 figure 5. Subgroup analysis of VRE and VSE bacteremia un-adjusted in-hospital mortality risk by age, immune status, study site(s), and study quality, for each included cohort study reporting these data. Abbreviations: 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. vreand vsebacteremia outcomes after vre therapy 33 figure 6. Subgroup analysis of VRE and VSE bacteremia hospital LOS by age, for each included cohort study reporting these data. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. table 2. Assessment of Study Quality, Based on the Newcastle-Ottawa Scale (NOS) Star System Study Selection Comparability Outcome/Exposure Total Stars Cohort studies Bar et al, 2006 **** *** 7 Billington et al, 2014 **** *** 7 Butler et al, 2010 **** *** 7 Cheah et al, 2013 **** ** *** 9 Cho et al, 2013 *** ** *** 8 da Silva et al, 2014 **** *** 7 Haas et al, 2010 *** *** 6 Marschall et al, 2013 *** *** 6 Mikulska et al, 2012 ** ** 4 Mohr et al, 2009 **** *** 7 Vydra et al, 2012 *** *** 6 Yoo et al, 2004 *** *** 6 Case control study Peel et al, 2011 *** *** 6 Illness severity and comorbid conditions were selected as the most important factors when assessing comparability. studies adjusted for potential confounders, and thus confounding may have influenced the investigated associations. Third, our results may be limited by the exclusive review of English language reports published after January 1997, but it is unlikely such language restrictions biased our findings. Fourth, the majority of studies sampled immunocompromised hospital patient populations, which limited our ability to generalize our findings to all healthcare settings. Last, our funnel plot suggests that there may be publication bias. However, the 2 studies that contributed to this asymmetry had high standard error and odds ratios close to 1. Thus, if the asymmetry in the funnel plot is due to publication bias, it would bias the results towards the null hypothesis. We conclude that using the best available evidence, VRE bacteremia remains associated with increased risk of morbidity figure 7. Asymmetrical funnel plot of VRE and VSE bacteremia in-hospital mortality effect estimates of all included studies. and mortality when compared with VSE bacteremia in the era of 34 infection control & hospital epidemiology january 2016, vol. 37, no. 1 effective VRE therapy. Future research is needed to determine vancomycin-resistant enterococcal bacteremia. Antimicrob Agents Chemother 2014;58:734–739. whether these results are related to unadjusted differences in the 9. MoherD,LiberatiA,TetzlaffJ,AltmanDG, PRISMA Group. patient populations, differences in treatment effectiveness, or Preferred reporting items for systematic reviews and meta-analyses: differences in proportions of patients with E. faecalis and the PRISMA statement. BMJ 2009;339:b2535. E. faecium comprising the VRE and VSE bacteremias. 10. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 ed. The Cochrane Collaboration; 2011. 11. Linden PK. Treatment options for vancomycin-resistant acknowledgments enterococcal infections. Drugs 2002;62:425–441. The authors would like to acknowledge Public Health Ontario, Library Services 12. Moher D, Pham B, Lawson ML, Klassen TP. The inclusion of for their contributions to the development and execution of literature searches reports of randomised trials published in languages other than referenced in this work. English in systematic reviews. Health Technol Assess 2003; Financial support: Financial support for this work was provided by 7:1–90. Ontario Agency for Health Protection and Promotion (Public Health 13. Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Ontario). Potential conflicts of interest: All authors have no potential conflicts of Scale (NOS) for assessing the quality of nonrandomized studies in interest to report. The work under consideration for publication was solely meta-analyses. Ottawa hospital Research Institute Website. supported by funds from Ontario Agency for Health Protection and Promo- http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. tion (Public Health Ontario) and did not receive third-party funds. All listed Published 2014. Accessed April, 20, 2015. authors have no financial relationships with entities in the biomedical arena 14. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and that could be perceived to influence, or that give the appearance of potentially variance from the median, range, and the size of a sample. influencing, the submitted work. Finally, all listed authors have no other BMC Med Res Methodol 2005;5:13. relationships or activities that readers could perceive to have influenced, or that 15. Bar K, Wisplinghoff H, Wenzel RP, Bearman GM, Edmond MB. give the appearance of potentially influencing, what is written in the Systemic inflammatory response syndrome in adult patients with submitted work. nosocomial bloodstream infections due to enterococci. BMC Address correspondence to Jennie Johnstone, MD, PhD, Public Infect Dis 2006;6:145. Health Ontario, 480 University Avenue, Suite 300, Toronto ON M5G 1V2 16. Billington EO, Phang SH, Gregson DB, et al. Incidence, risk (jennie.johnstone@oahpp.ca) factors, and outcomes for Enterococcus spp. blood stream infections: a population-based study. Int J Infect Dis 2014;26: supplementary material 76–82. 17. Butler AM, Olsen MA, Merz LR, et al. Attributable costs of To view supplementary material for this article, please visit enterococcal bloodstream infections in a nonsurgical hospital http://dx.doi.org/10.1017/ice.2015.228 cohort. Infect Control Hosp Epidemiol 2010;31:28–35. 18. Cheah AL, Spelman T, Liew D, et al. Enterococcal bacteraemia: references factors influencing mortality, length of stay and costs of hospitalization. Clin Microbiol Infect 2013;19:E181–E189. 1. Cetinkaya Y, Falk P, Mayhall CG. Vancomycin-resistant enter- 19. Cho SY, Lee DG, Choi SM, et al. Impact of vancomycin resistance ococci. Clin Microbiol Rev 2000;13:686–707. on mortality in neutropenic patients with enterococcal 2. Murray BE. Vancomycin-resistant enterococcal infections. bloodstream infection: a retrospective study. BMC Infect Dis N Engl J Med 2000;342:710–721. 2013;13:504. 3. Ramsey AM, Zilberberg MD. Secular trends of hospitalization 20. Haas EJ, Zaoutis TE, Prasad P, Li M, Coffin SE. Risk factors and with vancomycin-resistant Enterococcus infection in the outcomes for vancomycin-resistant Enterococcus bloodstream United States, 2000–2006. Infect Control Hosp Epidemiol 2009; infection in children. Infect Control Hosp Epidemiol 2010;31: 30:184–186. 1038–1042. 4. Gravel D, Archibald CP, Pelude L, Mulvey M, Golding G. 21. Marschall J, Piccirillo ML, Fraser VJ, Doherty JA, Warren DK. Antimicrobial resistance surveillance in Canadian hospitals, Catheter removal versus retention in the management of 2007–2012. Canada Communicable Disease Report CCDR 2014, catheter-associated enterococcal bloodstream infections. Can J Volume 40:November 21, 2014. 40 S-40 2; Accessed November Infect Dis Med Microbiol 2013;24:e83–e87. 21, 2014. 22. Mikulska M, Del Bono V, Raiola AM, et al. Enterococcal bloodstream 5. Salgado CD, Farr BM. Outcomes associated with vancomycin- infection after hematopoietic stem cell transplant: experience of a resistant enterococci: a meta-analysis. Infect Control Hosp Epidemiol center with a low prevalence of vancomycine8resistant enterococci. 2003;24:690–698. Clin Infect Dis 2012;55:1744. 6. DiazGranados CA, Jernigan JA. Impact of vancomycin resistance 23. Mohr JF, Friedrich LV, Yankelev S, Lamp KC. Daptomycin on mortality among patients with neutropenia and enterococcal for the treatment of enterococcal bacteraemia: results from bloodstream infection. J Infect Dis 2005;191:588–595. the Cubicin Outcomes Registry and Experience (CORE). Int J 7. Erlandson KM, Sun J, Iwen PC, Rupp ME. Impact of the Antimicrob Agents 2009;33:543–548. more-potent antibiotics quinupristin-dalfopristin and linezolid 24. Peel T, Cheng AC, Spelman T, Huysmans M, Spelman D. on outcome measure of patients with vancomycin-resistant Differing risk factors for vancomycin-resistant and vancomycin- Enterococcus bacteremia. Clin Infect Dis 2008;46:30–36. 8. Balli EP, Venetis CA, Miyakis S. Systematic review and sensitive enterococcal bacteraemia. Clin Microbiol Infect 2012; meta-analysis of linezolid versus daptomycin for treatment of 18:388–394. vreand vsebacteremia outcomes after vre therapy 35 25. Vydra J, Shanley RM, George I, et al. Enterococcal bacteremia is 27. da Silva NS, Muniz VD, Estofolete CF, Furtado GH, Rubio FG. associated with increased risk of mortality in recipients of Identification of temporal clusters and risk factors of bacteremia allogeneic hematopoietic stem cell transplantation. Clin Infect Dis by nosocomial vancomycin-resistant enterococci. Am J Infect 2012;55:764–770. Control 2014;42:389–392. 26. Yoo JH, Lee DG, Choi SM, et al. Vancomycin-resistant enterococcal 28. Cheah AL, Peel T, Howden BP, et al. Case-case-control study bacteremia in a hematology unit: molecular epidemiology on factors associated with vanB vancomycin-resistant and and analysis of clinical course. J Korean Med Sci 2005;20: vancomycin-susceptible enterococcal bacteraemia. BMC Infect 169–176. Dis 2014;14:353. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Infection Control and Hospital Epidemiology Pubmed Central

VRE and VSE Bacteremia Outcomes in the Era of Effective VRE Therapy: A Systematic Review and Meta-analysis

Infection Control and Hospital Epidemiology , Volume 37 (1) – Oct 5, 2015

Loading next page...
 
/lp/pubmed-central/vre-and-vse-bacteremia-outcomes-in-the-era-of-effective-vre-therapy-a-lUVljsEuqA

References (33)

Publisher
Pubmed Central
Copyright
© © 2015 by The Society for Healthcare Epidemiology of America 2015
ISSN
0899-823X
eISSN
1559-6834
DOI
10.1017/ice.2015.228
Publisher site
See Article on Publisher Site

Abstract

infection control & hospital epidemiology january 2016, vol. 37, no. 1 original article VRE and VSE Bacteremia Outcomes in the Era of Effective VRE Therapy: A Systematic Review and Meta-analysis 1 1 1,2,3 1 Chatura Prematunge, MSc; Colin MacDougall, MSc; Jennie Johnstone, MD, PhD; Kwaku Adomako, MSc; 1 1 1,3,4,5 Freda Lam, MPH; Jennifer Robertson, PhD; Gary Garber, MD background. Prior data suggest that vancomycin-resistant Enterococcus (VRE) bacteremia is associated with worse outcomes than vancomycin-sensitive Enterococcus (VSE) bacteremia. However, many studies evaluating such outcomes were conducted prior to the availability of effective VRE therapies. objective. To systematically review VRE and VSE bacteremia outcomes among hospital patients in the era of effective VRE therapy. methods. Electronic databases and grey literature published between January 1997 and December 2014 were searched to identify all primary research studies comparing outcomes of VRE and VSE bacteremias among hospital patients, following the availability of effective VRE therapies. The primary outcome was all-cause, in-hospital mortality, while total hospital length of stay (LOS) was a secondary outcome. All meta-analyses were conducted in Review Manager 5.3 using random-effects, inverse variance modeling. results. Among all the studies reviewed, 12 cohort studies and 1 case control study met inclusion criteria. Similar study designs were combined in meta-analyses for mortality and LOS. VRE bacteremia was associated with increased mortality compared with VSE bacteremia among cohort studies (odds ratio [OR], 1.80; 95% confidence interval [CI], 1.38–2.35; I = 0%; n= 11); the case-control study estimate was similar, but not significant (OR, 1.93; 95% CI, 0.97–3.82). LOS was greater for VRE bacteremia patients than for VSE bacteremia patients (mean difference, 5.01 days; 95% CI, 0.58–9.44]; I = 0%; n= 5). conclusions. Despite the availability of effective VRE therapy, VRE bacteremia remains associated with an increased risk of in-hospital mortality and LOS when compared to VSE bacteremia. Infect. Control Hosp. Epidemiol. 2016;37(1):26 –35 Enterococcus spp. are typically commensal organisms, Whether outcomes associated with VRE bacteremia are 1,2 common in the human gastrointestinal tract, but in some worse than those associated with vancomycin-sensitive circumstances can cause serious infections including bacteremia, enterococci (VSE) bacteremia remains unclear. Two prior particularly among hospitalized patients with underlying systematic reviews have compared outcomes of VRE 1,2 comorbid conditions. Since its discovery in 1988, bacteremia VSE bacteremia; both found VRE bacteremia to be vancomycin-resistant enterococci (VRE) have emerged as associated with an increased risk of mortality when compared important nosocomial pathogens and are occurring with to VSE bacteremia (relative risk [RR], 2.38; 95% confidence increasing frequency due to widespread use of antibiotics, interval [CI], 2.13–2.66; odds ratio [OR], 2.52; 95% CI, prolonged hospitalizations, and increased intensive care unit 1.87–3.39 ). However, both of these systematic reviews (ICU) admissions, especially among patients with malignant included studies conducted prior to the availability of effective 1–3 5,6 health conditions. In Canada, the incidence of VRE infections VRE therapies. Since late 1999, a number of antibiotic drugs has risen to 0.5 infections per 1,000 admissions, a 6-fold have been licensed as treatment for VRE bacteremia by the increase in recent years. Similarly in the United States, US Food and Drug Administration (FDA), Health Canada, 1,2,7 hospitalizations with VRE infection reached 0.6 per 1,000 and other national approval agencies. Quinupristin- admissions by 2006. dalfopristin was approved in 1999, followed by linezolid Affiliations: 1. Infection Prevention and Control, Public Health Ontario, Toronto, Ontario, Canada; 2. St. Joseph’s Health Centre, Toronto, Ontario, Canada; 3. Department of Medicine, University of Toronto, Toronto, Ontario, Canada; 4. Department of Medicine, University of Ottawa, Ottawa, Ontario, Canada; 5. Ottawa Hospital Research Institute, Ottawa, Ontario, Canada. PREVIOUS PRESENTATION. The preliminary findings of this systematic review and meta-analysis were presented at Association of Medical Microbiology and Immunology (AMMI) Annual Conference, Quebec, Canada, in April 2015 as an oral presentation. Received April 30, 2015; accepted August 30, 2015; electronically published October 5, 2015 © 2015 by The Society for Healthcare Epidemiology of America. All rights reserved. This is an Open Access article, distributed under the terms of the Creative Commons Attribution license (http://creativecommons.org/licenses/ by/3.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited. All rights reserved. 0899-823X/2016/3701-0006. DOI: 10.1017/ice.2015.228 vreand vsebacteremia outcomes after vre therapy 27 7 1,2,7,8 in 2000. In 2003, daptomycin was formally licensed for telavancin for treating any part of the illness. Penicillin, complicated skin and soft tissue VRE infections, but it is ampicillin, amikacin, streptomycin, chloramphenicol, doxycy- frequently used as an off-label therapy for VRE bacteremia. cline, rifampin, imipenem-cilastatin, and nitrofurantoin were Thus, understanding whether VRE bacteremia-associated not considered effective VRE treatments. outcomes are different from those of VSE bacteremia, since the To capture standard, off-label, and compassionate study emergence of effective VRE therapy, is critically important to use of effective VRE treatment(s), literature published after help inform future VRE infection control recommendations. January 1997 was considered. Studies analyzing data collected To this end, we performed a systematic review and between January 1997 and January 2000 were excluded if the meta-analysis of studies comparing outcomes of patients antibiotics used for the treatment of VRE bacteremia patients with either VRE or VSE bacteremia, when patients with VRE were not reported or could not be obtained by contacting bacteremia were treated with effective VRE therapy. study authors. Studies conducted after January 2000 were assumed to have administered effective VRE treatment(s) and were included in the review. methods Narrative reviews, case series, case reports, and commentaries All methods including literature searches, study selection, data were excluded. Only the most recent peer-reviewed collection, and quantitative analysis processes were developed a publication was included when multiple reports using the priori and were reported according to the Preferred Reporting same study data existed. We limited our review to English Items for Systematic Reviews and Meta-Analyses (PRISMA) language articles. guidelines and the Cochrane Handbook for Systematic Reviews of 9,10 Intervention. Study Selection and Data Extraction Titles and abstracts of articles captured by literature searches Search Methodology and Data Sources were independently screened in duplicate by two reviewers The Public Health Ontario (PHO) Library Services department (CP and CM). Articles flagged for full-text review by either assisted with the development and implementation of search reviewer were included in the full-text review, and the full-text strategies for electronic databases, as well as with the retrieval of review process was duplicated and independently completed full-text articles. Medline, Embase, CINAHL, ProQuest by the same reviewers. Inter-rater reliability following full-text dissertations and theses, and the Cochrane Central Register of review was calculated using Cohens Kappa statistic and any Controlled Trials (CENTRAL) databases were searched from disagreements on study inclusion were resolved via arbitration January 1997 to December 2014. A sample search strategy is by a third reviewer (JJ). provided in Supplemental Table 1. Websites of infection control authorities and proceedings from infection control conferences Quality Assessment held within the most recent 5 years (ie, January 1, 2010, to Data extraction and quality assessment for included studies were January 1, 2015) were searched as outlined in Supplemental performed in duplicate (by CP and CM). An electronic data Table 2. Conference proceedings prior to 2010 were not extraction template was developed, pilot tested, and refined considered because we assumed that valuable data contained prior to the initiation of data extraction. The extracted data within such abstracts had become available in peer-reviewed elements included study design, sample size, study period, literature. Additionally, the reference lists of all relevant study setting, study population, patient type, study location, publications were hand searched to identify additional citations. Enterococcus spp., VRE/VSE bacteremia definition, VRE therapy administered, and number of VRE and VSE bacteremia patients Study Inclusion Criteria with the above stated outcome(s) of interest along with The study inclusion criteria for the review were randomized associated effect estimates and confidence intervals. Whenever controlled trials (RCTs), cohort studies, case-control studies, required information was not reported, attempts were made and cross-sectional studies, sampling adult (≥18 years of age) to contact the first and/or corresponding authors to and/or pediatric (<18 years of age) hospital patients, diag- obtain missing information; after 2 attempts, authors were nosed with VRE bacteremia and treated with effective VRE considered unresponsive. Data requests were limited to therapy, alongside VSE bacteremia patient comparators, and missing information on administered VRE treatment type(s), reporting on various mortality and morbidity outcomes. primary outcomes, and any secondary outcomes reported The primary outcome of interest was all-cause in-hospital within the primary study. mortality. Secondary outcomes were bacteremia-attributable Study quality was assessed using the Newcastle-Ottawa Scale mortality, total hospital length of stay (LOS), total intensive (NOS) scale or Cochrane risk of bias tool. The NOS was used care unit (ICU) LOS, post-VRE/VSE bacteremia diagnosis to establish quality of evidence within non-randomized cohort hospital LOS, and post-VRE/VSE bacteremia diagnosis ICU or case control studies, via a 9-star system. A study awarded a LOS. Effective VRE therapies were defined as quinupristin- greater number of stars is considered to be of higher dalfopristin, linezolid, daptomycin, tigecycline, teicoplanin, and methodological study quality. Although we did not 28 infection control & hospital epidemiology january 2016, vol. 37, no. 1 anticipate finding any RCTs, the Cochrane risk of bias tool December 2008. However, these authors confirmed that all was assigned to assess RCT study quality in the event an patients with VRE bacteremia were diagnosed after January 10 16 RCT meeting inclusion criteria was discovered. 2000. Billington et al sampled all residents within a Canadian health zone who developed enterococcal bloodstream infec- tions. We contacted these authors to obtain mortality Data Analysis and LOS information for study participants. In addition, Outcome effect measures for each study were calculated using 8 studies exclusively sampled adult patients within tertiary care 15,17–19,21,22,24,26 numbers of patients with VRE and VSE bacteremia with the hospital settings, and 4 of these studies were 19,22,25,26 outcome(s) of interest. We pooled studies of the same study limited to immunocompromised patient populations design via inverse variance method and random effects such as hematopoietic stem cell transplant patients or modeling in Review Manager 5.3; summary effect measures chemotherapy recipients. are reported as odds ratios (OR) and 95% confidence intervals All included studies defined patients with at least 1 VRE- or 15–27 (CI) for mortality, and mean difference and standard deviation VSE-positive blood culture to be cases of bacteremia. Both (SD) are reported for continuous LOS outcomes. When the E. faecalis and E. faecium were captured in 12 study 15–25,27 26 median and interquartile ranges (IQR) were reported, the samples, but the study by Yoo et al only included median was assumed to reflect the mean, and IQR was E. faecium infections. Outcome data for 2,575 bacteremias, assumed to be 1.35 SD. Statistical heterogeneity was assessed specifically 1,863 VSE and 712 VRE bacteremia cases, were using the I statistic, and VRE and VSE bacteremia outcomes identified in our literature review. were further explored via planned subgroup analyses of the following patient populations: (1) adult versus pediatric Outcomes patients, (2) immunocompromised versus varied immune status patients, ICU versus non-ICU admissions, (3) multi- Of the reviewed studies, 12 studies were cohort studies and 1 was center versus single study site, and low versus moderate- a case control study. When in-hospital mortality from the to-high study quality for included cohort studies. cohort studies were combined using unadjusted analysis, VRE Publication bias was examined via the visual interpretation of bacteremia was associated with an increased risk of in-hospital funnel plot symmetry and limited to the mortality outcomes. death when compared to VSE bacteremia with no heterogeneity (OR, 1.80; 95% CI, 1.40–2.32; I = 0%; n= 12) (Figure 2). The Role of the Funding Source single case-control study did not report a statistically significant increase in risk of VRE bacteremia death when compared with The design, conduct, and reporting for this systematic review VSE bacteremia in an unadjusted analysis (OR 1.93; 95% CI, and meta-analysis was funded by the Ontario Agency for 0.97–3.82) ; adjusted analyses were not reported. Health Protection and Promotion (Public Health Ontario). Of the 12 cohort studies, 5 reported adjusted analyses for 15,16,18,19,21 in-hospital mortality risk, and 2 studies found VRE results 18,19 bacteremia to be associated with adjusted mortality. Cheah Literature Search et al adjusted for prior ICU admission, comorbidities measured by the Charlson Comorbidity Index, Enterococcus sp., The literature searches identified 4,878 citations; among these, additional non-enterococcal infections, time to effective therapy, 155 citations were chosen for full-text review, and 20 studies and VRE bacteremia (OR, 1.21; 95% CI, 0.53–2.79]) via logistic were determined to meet our inclusion criteria (Figure 1). Of regression analysis. Cho et al adjusted for severity of these, 1 study did not indicate a study period and 5 studies illness using Simplified Acute Physiology Index, length of reported study periods between January 1997 and January hospitalization, and vancomycin resistance (hazard ratio [HR], 2000 and required confirmation of VRE therapies within 0.75; 95% CI, 0.24–2.36) via Cox proportional hazards each study. Corresponding authors were contacted, but modeling. VRE bacteremia was not included in the final models administered VRE therapy information could not be obtained 15,16,21 of the remaining 3 studies reporting adjusted mortality. and all 7 studies were excluded from the review. Excluded The study by Cho et al was the only study to report on VRE/ study details are provided in Supplemental Table 3. Therefore, VSE bacteremia-attributable mortality, which was defined 13 studies were included in the systematic review. as death within 7 days of bacteremia when no other cause could be identified. There was no significant difference in Description of Studies attributable mortality risk between VRE and VSE bacteremia The study characteristics of all included studies are outlined in patients in the unadjusted analysis (6 of 24 patients with VRE Table 1. All were observational and retrospective studies, 12 of bacteremia vs 15 of 67 patients with VSE bacteremia; OR, 1.15; which were conducted between January 2000 and December 95% CI, 0.39–3.43). 2011, following the formal regulatory approval of the first Total hospital LOS data were reported within 6 studies. Data 15–26 27 17 18 effective VRE therapy. The study by da Silva et al reported by Butler et al and Cheah et al, and data obtained 16 27 reported a study period between September 1998 and by contacting authors of Billington et al, da Silva et al, and vre and vse bacteremia outcomes after vre therapy 29 figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flowchart of the literature search and study selection. table 1. Characteristics of Studies Included in Systematic Review and Meta-Analysis Sample Size, No. Study Study Period Location Patient Population VRE VSE Cohort studies Bar et al, 2006 Nov 2000–Dec 2002 Richmond, VA USA Adult 17 33 16 a Billington et al, 2014 2000–2008 Calgary, Canada Mixed 27 640 Butler et al, 2010 Jan 2002–Dec 2003 St Louis, MO USA Adult, Non-surgical, >2 days LOS 94 182 Cheah et al, 2013 Jan 2002–March 2010 Victoria, Australia Adult, >2 days LOS 116 116 Cho et al, 2013 July 2009–Dec 2011 Seoul, Korea Adult, neutropenia post CHEMO or SCT 24 67 27 c c da Silva et al, 2014 Sep 1998–Dec 2008 Sao Jose do Rio Preto, Brazil Mixed 30 273 20 a Haas et al, 2010 2001–2006 Philadelphia, PA USA Pediatrics 39 300 Marschall et al, 2013 Jan 2006–Dec 2006 St. Louis, MO USA Adult, CVC associated bacteremias 67 39 22 a Mikulska et al, 2012 2004–2011 Genoa, Italy Adult, allogeneic HSCT 9 58 23 b Mohr et al, 2009 Jan 2000–Dec 2009 58 sites, USA Mixed, dap Tx. 151 211 Vydra et al, 2012 Jan 2004–Dec 2008 Minneapolis, MN USA Mixed, HSCT 50 43 26 D Yoo et al, 2004 Jan 2000–Dec 2001 Seoul, Korea Adult, HSCT or cytotoxic CHEMO 19 8 Case control study Peel et al, 2011 Jan 2000–Dec 2009 Victoria, Australia Adult 80 360 NOTE. LOS, length of stay; HSCT, hematopoietic stem cell transplantation; CHEMO, chemotherapy; dap Tx, daptomycin treatment; CVC, central venous catheter; SCT, stem cell transplantation. Months not reported. Assumed to be mixed, unconfirmed due to demographics being reported as ≤ 30 years of age. Data obtained by contacting study authors. A total of 8 VRE patients received VRE therapies and were included in the review. 30 infection control & hospital epidemiology january 2016, vol. 37, no. 1 figure 2. VRE and VSE bacteremia unadjusted in-hospital mortality risk by study design. Results of included studies for VRE and VSE bacteremia unadjusted in-hospital mortality risk stratified by study design. Abbreviations: 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. figure 3. VRE and VSE bacteremia total hospital LOS mean difference. Results of studies reporting on VRE and VSE bacteremia total hospital LOS. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. figure 4. VRE and VSE post-bacteremia total hospital LOS mean difference. Results of studies reporting on VRE and VSE post-bacteremia hospital LOS. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. Haas et al were pooled; VRE bacteremia was associated development of clinically significant bacteremia, which is with a longer LOS than VSE bacteremia (mean difference, different from the LOS definition used in our review. 2 26 5.01; 95% CI, 0.58–9.44; I = 0%; n = 5) (Figure 3). Data Data from Yoo et al were excluded because their LOS from Cho et al were excluded because they defined LOS estimates combined patients treated with effective and as the number of days from hospital admission to the noneffective VRE therapy. vre and vse bacteremia outcomes after vre therapy 31 Post-bacteremia LOS data reported by Cheah et al and Haas bacteremia compared with VSE bacteremia. The mortality et al were also pooled via a meta-analysis. There was no significant summary estimate demonstrated lack of heterogeneity across difference in LOS following a VRE bacteremia compared with studies and no significant influence on the point estimate by VSE bacteremia (mean difference, 0.53 [95% CI –8.98, 10.04]; age, immune status, study site(s), or study quality. VRE I = 26%; n= 2) (Figure 4). Yoo et al also reported on bacteremia was also associated with increased total hospital post-bacteremia LOS, but data were omitted because estimates LOS and post-bacteremia LOS with no heterogeneity. The combined patients treated with both effective and noneffective post-bacteremia LOS estimate was not statistically significant, VRE therapy. which may have been due to lack of statistical power influenced by the small number of studies reporting on post Subgroup Analyses bacteremia LOS outcomes. Our finding, that there is an increased risk of mortality and No significant interactions were detected between any of the LOS associated with VRE bacteremia when compared to VSE subgroups we had planned to analyze for in-hospital 5,6 bacteremia, is consistent with 2 previous systematic reviews. mortality including age (pediatric patients [OR, 1.62; 95% In the systematic review by Salgado et al, the authors CI, 1.18–2.22] vs adult [OR, 1.93; 95% CI, 0.89–4.18]; speculated that the increased morbidity and mortality could be interaction P = .68), immune status (immunocompromised because patients with VRE bacteremia were more likely to patients [OR, 1.24; 95% CI, 0.65–2.35] vs varied immune 5,6 receive ineffective therapy. However, our findings suggest status [OR, 1.93; 95% CI, 1.47–2.54]; interaction P = .21), that a lack of effective therapy is not the explanation. It should study site (single center studies [OR, 1.85; 95% CI, 1.37–2.50] be noted that our systematic review was unable to capture time vs multicenter studies [OR, 1.70; 95% CI, 1.12–2.58]; to effective therapy. Thus, it is possible that patients with VSE interaction P= .75) and study quality (low-quality studies bacteremia received effective therapy sooner than patients with [OR, 0.36; 95% CI, 0.04–3.11] vs moderate- to high-quality VRE bacteremia because VRE may be less likely to be covered studies [OR, 1.84; 95% CI, 1.43–2.37]; interaction P= .14) by empiric therapy, and effective therapy may only have (Figure 5). The planned subgroup analysis for ICU stay was been administered following a VRE-positive microbiological not performed due to unavailable data. 6,15,18,19 culture result. Age was not found to significantly influence total hospital An alternative explanation for the observed increase in LOS by subgroup analysis (pediatric patients [OR −13.00; 95% mortality risk and LOS could be differences in illness severity CI, −39.90–13.90] vs adult [OR, 6.12; 95% CI, 0.82–11.42]; or comorbidities between patients with VRE and VSE interaction P= .17) (Figure 6). The remaining LOS subgroup bacteremia, particularly because patients with VRE bacteremia analyses could not be performed due to a lack of studies in each 18,24,28 may have more comorbidities. Due to limited reporting companion subgroup. No significant interaction was detected of adjusted mortality and morbidity risks among included in the subgroup analysis of post-bacteremia LOS by age studies, we were unable to calculate adjusted summary (pediatric patients [OR, −9.0; 95% CI, −28.13–10.13] vs adult estimates in this systematic review. Thus, the effect of [OR, 3.0; 95% CI −3.37–9.37]; interaction P = .24) (Figure 4). confounding factors on our unadjusted mortality and LOS summary estimates remains unclear. We hypothesize that not Study Quality adjusting for potential confounders (ie, comorbid conditions Study quality ratings based on NOS criteria are presented in and severity of illness) may lead to overestimates of our Table 2. Of the 13 studies reviewed, 12 were of moderate to associations of interest because patients colonized with VRE high study quality, with the most frequent number of stars tend to have more comorbid conditions and more severe awarded per study being 6 or 7. Among all studies, patients illness than patients not colonized with VRE. with VRE and VSE bacteremia were selected from the same However, the earlier systematic review by DiazGranados hospital population, and bacteremia diagnosis was confirmed et al, which only considered studies controlling for underlying by patient chart reviews or microbiology reports. severity of illness, found VRE bacteremia adjusted mortality risk to be greater in comparison to VSE bacteremia. The worse outcomes associated with VRE bacteremia Publication Bias compared to VSE bacteremia may also be linked to differences The asymmetrical funnel plot indicates that the review’s in the causative species as there may have been proportionately in-hospital mortality estimates may be subject to publication more patients with E. faecium than E. faecalis in the bias (Figure 7). VRE bacteremia group when compared to the VSE bacteremia 1,2 group. Our results should be interpreted recognizing the systematic discussion review’s limitations. First, studies included within each In this systematic review and meta-analysis, we found that meta-analysis were non-randomized observational studies, since the advent of effective VRE therapy, there remains an and accordingly, our results reflect association and not increased risk of in-hospital mortality associated with VRE causation. Second, as discussed above, only a small number of 32 infection control & hospital epidemiology january 2016, vol. 37, no. 1 figure 5. Subgroup analysis of VRE and VSE bacteremia un-adjusted in-hospital mortality risk by age, immune status, study site(s), and study quality, for each included cohort study reporting these data. Abbreviations: 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. vreand vsebacteremia outcomes after vre therapy 33 figure 6. Subgroup analysis of VRE and VSE bacteremia hospital LOS by age, for each included cohort study reporting these data. Abbreviations: LOS, length of stay; 95% CI, 95% confidence interval; SE, standard error; IV, random, inverse-variance, random-effects method. table 2. Assessment of Study Quality, Based on the Newcastle-Ottawa Scale (NOS) Star System Study Selection Comparability Outcome/Exposure Total Stars Cohort studies Bar et al, 2006 **** *** 7 Billington et al, 2014 **** *** 7 Butler et al, 2010 **** *** 7 Cheah et al, 2013 **** ** *** 9 Cho et al, 2013 *** ** *** 8 da Silva et al, 2014 **** *** 7 Haas et al, 2010 *** *** 6 Marschall et al, 2013 *** *** 6 Mikulska et al, 2012 ** ** 4 Mohr et al, 2009 **** *** 7 Vydra et al, 2012 *** *** 6 Yoo et al, 2004 *** *** 6 Case control study Peel et al, 2011 *** *** 6 Illness severity and comorbid conditions were selected as the most important factors when assessing comparability. studies adjusted for potential confounders, and thus confounding may have influenced the investigated associations. Third, our results may be limited by the exclusive review of English language reports published after January 1997, but it is unlikely such language restrictions biased our findings. Fourth, the majority of studies sampled immunocompromised hospital patient populations, which limited our ability to generalize our findings to all healthcare settings. Last, our funnel plot suggests that there may be publication bias. However, the 2 studies that contributed to this asymmetry had high standard error and odds ratios close to 1. Thus, if the asymmetry in the funnel plot is due to publication bias, it would bias the results towards the null hypothesis. We conclude that using the best available evidence, VRE bacteremia remains associated with increased risk of morbidity figure 7. Asymmetrical funnel plot of VRE and VSE bacteremia in-hospital mortality effect estimates of all included studies. and mortality when compared with VSE bacteremia in the era of 34 infection control & hospital epidemiology january 2016, vol. 37, no. 1 effective VRE therapy. Future research is needed to determine vancomycin-resistant enterococcal bacteremia. Antimicrob Agents Chemother 2014;58:734–739. whether these results are related to unadjusted differences in the 9. MoherD,LiberatiA,TetzlaffJ,AltmanDG, PRISMA Group. patient populations, differences in treatment effectiveness, or Preferred reporting items for systematic reviews and meta-analyses: differences in proportions of patients with E. faecalis and the PRISMA statement. BMJ 2009;339:b2535. E. faecium comprising the VRE and VSE bacteremias. 10. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 ed. The Cochrane Collaboration; 2011. 11. Linden PK. Treatment options for vancomycin-resistant acknowledgments enterococcal infections. Drugs 2002;62:425–441. The authors would like to acknowledge Public Health Ontario, Library Services 12. Moher D, Pham B, Lawson ML, Klassen TP. The inclusion of for their contributions to the development and execution of literature searches reports of randomised trials published in languages other than referenced in this work. English in systematic reviews. Health Technol Assess 2003; Financial support: Financial support for this work was provided by 7:1–90. Ontario Agency for Health Protection and Promotion (Public Health 13. Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Ontario). Potential conflicts of interest: All authors have no potential conflicts of Scale (NOS) for assessing the quality of nonrandomized studies in interest to report. The work under consideration for publication was solely meta-analyses. Ottawa hospital Research Institute Website. supported by funds from Ontario Agency for Health Protection and Promo- http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. tion (Public Health Ontario) and did not receive third-party funds. All listed Published 2014. Accessed April, 20, 2015. authors have no financial relationships with entities in the biomedical arena 14. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and that could be perceived to influence, or that give the appearance of potentially variance from the median, range, and the size of a sample. influencing, the submitted work. Finally, all listed authors have no other BMC Med Res Methodol 2005;5:13. relationships or activities that readers could perceive to have influenced, or that 15. Bar K, Wisplinghoff H, Wenzel RP, Bearman GM, Edmond MB. give the appearance of potentially influencing, what is written in the Systemic inflammatory response syndrome in adult patients with submitted work. nosocomial bloodstream infections due to enterococci. BMC Address correspondence to Jennie Johnstone, MD, PhD, Public Infect Dis 2006;6:145. Health Ontario, 480 University Avenue, Suite 300, Toronto ON M5G 1V2 16. Billington EO, Phang SH, Gregson DB, et al. Incidence, risk (jennie.johnstone@oahpp.ca) factors, and outcomes for Enterococcus spp. blood stream infections: a population-based study. Int J Infect Dis 2014;26: supplementary material 76–82. 17. Butler AM, Olsen MA, Merz LR, et al. Attributable costs of To view supplementary material for this article, please visit enterococcal bloodstream infections in a nonsurgical hospital http://dx.doi.org/10.1017/ice.2015.228 cohort. Infect Control Hosp Epidemiol 2010;31:28–35. 18. Cheah AL, Spelman T, Liew D, et al. Enterococcal bacteraemia: references factors influencing mortality, length of stay and costs of hospitalization. Clin Microbiol Infect 2013;19:E181–E189. 1. Cetinkaya Y, Falk P, Mayhall CG. Vancomycin-resistant enter- 19. Cho SY, Lee DG, Choi SM, et al. Impact of vancomycin resistance ococci. Clin Microbiol Rev 2000;13:686–707. on mortality in neutropenic patients with enterococcal 2. Murray BE. Vancomycin-resistant enterococcal infections. bloodstream infection: a retrospective study. BMC Infect Dis N Engl J Med 2000;342:710–721. 2013;13:504. 3. Ramsey AM, Zilberberg MD. Secular trends of hospitalization 20. Haas EJ, Zaoutis TE, Prasad P, Li M, Coffin SE. Risk factors and with vancomycin-resistant Enterococcus infection in the outcomes for vancomycin-resistant Enterococcus bloodstream United States, 2000–2006. Infect Control Hosp Epidemiol 2009; infection in children. Infect Control Hosp Epidemiol 2010;31: 30:184–186. 1038–1042. 4. Gravel D, Archibald CP, Pelude L, Mulvey M, Golding G. 21. Marschall J, Piccirillo ML, Fraser VJ, Doherty JA, Warren DK. Antimicrobial resistance surveillance in Canadian hospitals, Catheter removal versus retention in the management of 2007–2012. Canada Communicable Disease Report CCDR 2014, catheter-associated enterococcal bloodstream infections. Can J Volume 40:November 21, 2014. 40 S-40 2; Accessed November Infect Dis Med Microbiol 2013;24:e83–e87. 21, 2014. 22. Mikulska M, Del Bono V, Raiola AM, et al. Enterococcal bloodstream 5. Salgado CD, Farr BM. Outcomes associated with vancomycin- infection after hematopoietic stem cell transplant: experience of a resistant enterococci: a meta-analysis. Infect Control Hosp Epidemiol center with a low prevalence of vancomycine8resistant enterococci. 2003;24:690–698. Clin Infect Dis 2012;55:1744. 6. DiazGranados CA, Jernigan JA. Impact of vancomycin resistance 23. Mohr JF, Friedrich LV, Yankelev S, Lamp KC. Daptomycin on mortality among patients with neutropenia and enterococcal for the treatment of enterococcal bacteraemia: results from bloodstream infection. J Infect Dis 2005;191:588–595. the Cubicin Outcomes Registry and Experience (CORE). Int J 7. Erlandson KM, Sun J, Iwen PC, Rupp ME. Impact of the Antimicrob Agents 2009;33:543–548. more-potent antibiotics quinupristin-dalfopristin and linezolid 24. Peel T, Cheng AC, Spelman T, Huysmans M, Spelman D. on outcome measure of patients with vancomycin-resistant Differing risk factors for vancomycin-resistant and vancomycin- Enterococcus bacteremia. Clin Infect Dis 2008;46:30–36. 8. Balli EP, Venetis CA, Miyakis S. Systematic review and sensitive enterococcal bacteraemia. Clin Microbiol Infect 2012; meta-analysis of linezolid versus daptomycin for treatment of 18:388–394. vreand vsebacteremia outcomes after vre therapy 35 25. Vydra J, Shanley RM, George I, et al. Enterococcal bacteremia is 27. da Silva NS, Muniz VD, Estofolete CF, Furtado GH, Rubio FG. associated with increased risk of mortality in recipients of Identification of temporal clusters and risk factors of bacteremia allogeneic hematopoietic stem cell transplantation. Clin Infect Dis by nosocomial vancomycin-resistant enterococci. Am J Infect 2012;55:764–770. Control 2014;42:389–392. 26. Yoo JH, Lee DG, Choi SM, et al. Vancomycin-resistant enterococcal 28. Cheah AL, Peel T, Howden BP, et al. Case-case-control study bacteremia in a hematology unit: molecular epidemiology on factors associated with vanB vancomycin-resistant and and analysis of clinical course. J Korean Med Sci 2005;20: vancomycin-susceptible enterococcal bacteraemia. BMC Infect 169–176. Dis 2014;14:353.

Journal

Infection Control and Hospital EpidemiologyPubmed Central

Published: Oct 5, 2015

There are no references for this article.