KEY WORDS Stem cell transplant Bloodstream infection Vancomycin-resistant enterococcus Allogeneic Screening

Transcription

1 Biology of Blood and Marrow Transplantation 13: (2007) 2007 American Society for Blood and Marrow Transplantation /07/ $32.00/0 doi: /j.bbmt Colonization, Bloodstream Infection, and Mortality Caused by Vancomycin-Resistant Enterococcus Early after Allogeneic Hematopoietic Stem Cell Transplant David M. Weinstock, 1,2 Mary Conlon, 3 Christine Iovino, 3 Tanya Aubrey, 2 Carlota Gudiol, 2 Elyn Riedel, 4 James W. Young, 1 Timothy E. Kiehn, 5 Gianna Zuccotti 2 1 Adult Allogeneic Bone Marrow Transplant Service and 2 Infectious Diseases Service, 3 Department of Medicine, Department of Nursing, 4 Department of Biostatistics, and 5 Department of Clinical Laboratories, Memorial Sloan- Kettering Cancer Center, New York, New York Correspondence and reprint requests: David M. Weinstock, MD, Department of Medicine, Memorial Sloan- Kettering Cancer Center, 1275 York Avenue, Box 199, New York, NY ( weinstod@mskcc.org). Received November 27, 2006; accepted January 21, 2007 ABSTRACT Bloodstream infection caused by vancomycin-resistant enterococcus (VRE) is associated with very high mortality among allogeneic hematopoietic stem cell transplant (allohsct) recipients. However, it remains unclear whether VRE bloodstream infection directly causes mortality in the early posttransplant period or is simply a marker of poor outcome. To determine the risk factors for VRE bloodstream infection and its effect on outcome, we followed 92 patients screened for stool colonization by VRE upon admission for allohsct. Patient records were reviewed to determine outcomes, including mortality and microbiologic failure. Colonization by VRE was extremely common, occurring in 40.2% of patients. VRE bloodstream infection developed in 34.2% of colonized patients by day 35, compared to 1.8% without VRE colonization (P <.01). VRE bloodstream infection was associated with a significant decrement in survival and frequent microbiologic failure, despite treatment with linezolid and/or daptomycin. Five (35.7%) of 14 patients with VRE bloodstream infection had attributable mortality or contributing mortality from the infection. Strain typing by pulsed-field gel electrophoresis identified 9 different VRE strains among the 37 colonized patients and 5 patients with different strains recovered from the stool and the blood. In conclusion, stool screening effectively identified patients at extremely high risk for VRE bloodstream infection. The high mortality of VRE in the early posttransplant period supports the use of empiric antibiotics with activity against VRE during periods of fever and neutropenia in colonized patients American Society for Blood and Marrow Transplantation KEY WORDS Stem cell transplant Bloodstream infection Vancomycin-resistant enterococcus Allogeneic Screening INTRODUCTION Enterococci are commensal flora of the human gastrointestinal tract. Although capable of causing gastrointestinal, genitourinary, and endovascular infections, enterococci were previously considered to have limited pathogenicity [1]. The unique nature of vancomycin-resistant enterococci (VRE) became apparent soon after the first VRE isolates were described in By the mid-1990s, over 25% of enterococci causing bloodstream infections in the United States were resistant to vancomycin [2] and VRE are now an important cause of hospital-acquired infections globally [1,3,4]. The vast majority of VRE infections in the United States are caused by Enterococcus faecium, with a smaller fraction caused by E. faecalis, E. avium, and other species [5]. The enhanced virulence of vancomycin-resistant E. faecium, compared to antibioticsusceptible enterococci, is likely to be mediated by a pathogenicity island that includes the esp, hyl, and acm genes [4,6]. In a recent report, hospital-derived vancomycin-resistant E. faecium from 5 continents were found to be primarily derived from a single clonal lineage, arguing that this organism is highly adapted 615

2 616 D. M. Weinstock et al. for success in hospital environments and has undergone a remarkably rapid, global dissemination [4]. In addition, a meta-analysis of enterococcal infections recently demonstrated that bloodstream infection by vancomycin-resistant isolates is associated with increased mortality, compared to infection with vancomycin-susceptible isolates [7,8]. Immunocompromised patients are at especially high risk for VRE bloodstream infection [9-12]. Colonization of the gastrointestinal tract with VRE may precede bloodstream infection in these patients, either through nosocomial acquisition or as a result of extensive exposure to broad-spectrum antibacterial antibiotics [13,14]. Among patients colonized with VRE, mucosal damage from high-dose chemotherapy may permit enteric VRE to seed the portal and systemic circulations. In several reports, VRE bloodstream infection was associated with very high mortality among hematopoietic stem cell transplant (HSCT) recipients [15-18]. Avery et al [16] reported 100% mortality among 12 allogeneic HSCT (allohsct) recipients who developed VRE bloodstream infection between 1997 and In a study from Memorial Sloan-Kettering Cancer Center (MSKCC) between 1993 and 1996 [17], 17 (85%) of 20 patients with VRE bloodstream infection died. Despite the high mortality, some authors have questioned whether VRE infection is a true cause of adverse outcome among immunocompromised patients or simply a marker of poor prognosis [16,19]. It also remains unclear whether the use of newer agents, such as linezolid and daptomycin, will reduce the mortality from VRE bloodstream infection in allohsct recipients. We sought to determine whether routine screening of allohsct recipients for stool colonization with VRE effectively identifies patients at high risk for VRE bloodstream infection in the early posttransplant period, offering the potential for targeted empiric or prophylactic strategies. In addition, we reviewed the outcomes of allohsct recipients with VRE bloodstream infection within 35 days after transplant, to determine the efficacy of newer antimicrobials in this population. PATIENTS AND METHODS Patient records and the allohsct database were reviewed to identify patients admitted for an initial allohsct between August 2004 and February Routine isolation for all patients undergoing allohsct at MSKCC includes a private room, gloves, and mask. Gowns are added to precautions for patients colonized by VRE. Since August 2004, patients admitted for allohsct underwent screening for VRE colonization by stool culture. Further routine stool cultures were not performed, although subsequent stool cultures were obtained in patients with diarrhea or upon transfer to the intensive care unit. Antibacterial prophylaxis was not given for the first 19 months of the study period. Because of an unexpectedly high rate of viridans Streptococci infections among allohsct recipients at our center [20], empiric vancomycin was given to patients undergoing myeloablative allohsct during the final 3 months of the study period. No enteral or topical antibiotics were routinely administered as prophylaxis. T cell depletion was performed by either soybean lectin agglutination or CD34 selection (Isolex 300i; Nexell Therapeutics, Irvine, CA), in each case followed by rosetting with sheep erythrocytes [21]. Vancomycin and linezolid susceptibility were determined using the MicroScan Dried Gram-Positive panel (Dade Behring). Daptomycin susceptibility was performed by Etest (AB Biodisk, Piscataway, NJ). Pulse-field gel electrophoresis (PFGE) was performed after SmaI digest, as previously described [22]. Associations between demographic and treatment factors and (1) VRE colonization determined at transplant, and (2) VRE bloodstream infection among colonized patients were examined using Fisher s exact test for categoric variables and Wilcoxon rank sum test for continuous variables. Examined demographic and treatment factors included patient age, sex, underlying disease, time from diagnosis to transplant, previous cycles of chemotherapy, myeloablative versus nonmyeloablative preparative regimen, use of total body irradiation 2 Gy in the preparative regimen, allograft T cell depletion, bone marrow versus peripheral blood stem cell allograft, and matched sibling versus alternative donor. Survival was compared using a Cox proportional hazards model with time-dependent covariates for colonization and infection. RESULTS Frequent Stool Colonization Prior to allohsct During the study period, 92 patients (Table 1) underwent screening for stool colonization. Thirtyseven (40.2%) of the 92 patients had VRE stool colonization, including 2 with colonization identified before admission, 25 with VRE isolated from the initial stool culture after admission, and 10 who had negative initial stool cultures then had VRE isolated from subsequent cultures. Patients with acute leukemia or refractory anemia with excess blasts (RAEB) had an increased risk for VRE colonization, compared with other underlying diseases (P.03). No other demographic, disease, or treatment-related factors were associated with VRE colonization. Bloodstream Infections There were 34 bloodstream infections among the 92 patients within the first 35 days after allohsct

3 VRE among Stem Cell Transplant Recipients 617 Table 1. A Comparison of All Adult Allogeneic Hematopoietic Stem Cell Transplant Recipients at Memorial Sloan-Kettering Cancer Center between August 2004 and February 2006 Who Underwent Stool Screening for Vancomycin-Resistant Enterococcus (VRE) Colonization, Those Who Were Found to Have VRE Stool Colonization and Those Who Developed a VRE Bloodstream Infection All Patients (%) VRE Colonized (%) VRE Bloodstream Infection (%) Patients Median age (range) 50 (23-71) 48 (23-68) 51 (29-68) Male 52 (56.5%) 18 (48.6%) 6 (42.9%) Primary disease Acute lymphoblastic leukemia 14 (15.2%) Acute myelogenous leukemia 37 (40.2%) Chronic lymphocytic leukemia 1 (1.1%) Chronic myelogenous leukemia 2 (2.2%) Hodgkin Lymphoma 3 (3.3%) Myelodysplastic syndrome 20 (21.7%) Multiple myeloma 1 (1.1%) Non-Hodgkin lymphoma 13 (14.1%) Prolymphocytic leukemia 1 (1.1%) High-risk primary disease* 57 (62.0%) 19 (51.4%) 7 (50%) Acute leukemia or RAEB 62 (67.4%) 32 (86.5%) 10 (71.4%) Transplant characteristics Myeloablative conditioning regimen 80 (87.0%) 36 (97.3%) 13 (92.9%) Total body irradiation >2 Gy 41 (44.6%) 18 (48.6%) 7 (50%) Allograft T cell depletion 57 (62.0%) 25 (67.6%) 11 (78.6%) Peripheral blood stem cell allograft 86 (93.5%) 34 (91.9%) 13 (92.9%) HLA-matched sibling donor 43 (46.7%) 16 (43.2%) 8 (57.1%) VRE bloodstream infection 14 (15.2%) 13 (34.2%) 14 (100%) Median follow-up for survivors (months) 11 (3-22) 13 (3-21) 9 (3-21) *Patients with acute myelogenous leukemia in first or second complete remission (CR), acute lymphoblastic leukemia in first CR, and chronic myelogenous leukemia in first chronic phase were classified as standard risk. All other patients were considered high risk. p.05 compared to patients with any disease other than acute leukemia or refractory anemia with excess blasts (RAEB). Patients who did not have sibling donors received allografts from HLA-mismatched or unrelated donors. (Table 2). Three bloodstream infections were polymicrobial, including 1 patient with coagulase-negative Staphylococcus and Lactobacillus, 1 patient with coagulase-negative Staphylococcus and Fusobacterium, and 1 patient with Lactobacillus and Klebsiella pneumoniae. Seven patients had blood cultures that were thought to be contaminated by coagulase-negative Table 2. Thirty-Seven Bacterial and Fungal Organisms That Caused 34 Bloodstream Infections among Adult Allogeneic Hematopoietic Stem Cell Transplant (allohsct) Recipients within the First 35 Days after Transplant Organism Isolates (%) Day(s) after allohsct Enterococcus faecium 13 (35.1%) 0,3,5,5,5,5,6,6,7,8,12,20,33 Klebsiella pneumoniae 4 (10.8%) 0,1,3,9 Fusobacterium 3 (8.1%) 3,9 Coagulase-negative Staphylococcus 3 (8.1%) 3,3,5 Streptococcus mitis 3 (8.1%) 3,5,5 Enterobacter cloacae 2 (5.4%) 4,4 Lactobacillus 2 (5.4%) 3,3 Acinetobacter baumanii 1 (2.7%) 32 Bacillus 1 (2.7%) 1 Candida kruseii 1 (2.7%) 2 Candida lusitaniae 1 (2.7%) 5 Enterococcus avium 1 (2.7%) 5 Escherichia coli 1 (2.7%) 5 Streptococcus oralis 1 (2.7%) 5 VRE isolates are in light face. Staphylococcus. VRE caused 14 (41.1%) of the 34 bloodstream infections, making it the most common organism by over 3-fold (Table 2). The mean number of days from hospitalization to bloodstream infection did not differ between patients infected by VRE and patients infected by other organisms (16.6 versus 13.4 days, respectively; p.25). Among the 37 patients colonized with VRE, 13 (34.2%) developed a VRE bloodstream infection by day 35 after allohsct (Table 3), including 10 of 27 (37.0%) who had VRE isolated from the first culture after admission and 3 of 10 (30.0%) found to be colonized on subsequent culture (P 1). In comparison, only 1 (1.8%) of 55 patients who did not have VRE on stool culture developed a VRE bloodstream infection (P.001 compared with VRE colonized). No other demographic, disease, or treatment-related factors were associated with an increased risk for VRE bloodstream infection. None of the 14 patients with VRE bloodstream infection had graft-versus-host disease (GVHD) at the time of their bloodstream infection. None of the 92 study patients had 1 bloodstream infection. Twenty (25.6%) of the 78 patients who did not have a VRE bloodstream infection had a bloodstream infection caused by an organism other than VRE, compared to 0 (0.0%) of 14 who had a VRE bloodstream infection (P.03). Bloodstream infec-

4 618 D. M. Weinstock et al. Table 3. A Description of the 14 Allogeneic Hematopoietic Stem Cell Transplant Recipients Who Developed Bloodstream Infection Caused by Vancomycin-Resistant Enterococci (VRE) between August 2004 and February 2006 at Memorial Sloan-Kettering Cancer Center Patient Days of Bacteremia ANC on First Day of BSI VRE Species Treatment/Days <200/ L E. faecium Linezolid/7 Daptomycin Gentamicin/ <200/ L E. faecium Daptomycin/3 Linezolid/13 Total Days of Treatment Days to CVC Removal Alive Cause of Death 29 1, 7 Yes 16 2 No EBV posttransplant lymphoproliferative disorder <200/ L E. faecium Linezolid/2 2 Not removed No VRE BSI <200/ L E. faecium Linezolid/ Yes Daptomycin/ <200/ L E. faecium Linezolid/ Yes <200/ L E. faecium Linezolid/ Yes <200/ L E. faecium Linezolid/ Yes <200/ L E. faecium Linezolid/ Yes Daptomycin/ / L E. faecium Daptomycin/2 Linezolid/5 7 0 No Diffuse alveolar hemorrhage VRE BSI (contributing) <200/ L E. faecium Linezolid/ No Leukemia relapse Daptomycin/ <200/ L E. faecium Linezolid/ Yes Veno-occlusive disease Daptomycin/ <200/ L E. avium Linezolid/3 Ampicillin Gentamicin/29 Dapto/3 (added) 32 1 No VRE BSI (Episode #1) 9 (Episode #2) (Episode #1) 1 (Episode #2) 600/ L E. faecium Daptomycin/34 Linezolid/15 (added) <200/ L E. faecium Linezolid/5 Daptomycin/ Not removed No No Pneumonia VRE BSI (contributing) Drug reaction VRE BSI (contributing) Mortality was attritutable to VRE bloodstream infection in patients 103 and 112 and contributed to mortality in patients 109, 113, and 114. Patients 113 and 114 had second episodes of BSI during daptomycin therapy. Patient 101 had 2 central venous catheters. ANC indicates absolute neutrophil count; CVC, central venous catheter; EBV, Epstein-Barr virus; VRE, vancomycin-resistant enterococcus. tion caused by organisms other than VRE were also less common among the 24 patients who were colonized with VRE but did not have a VRE bloodstream infection (12.5%) than among the 55 patients who were not colonized with VRE (30.9%; P.05). Outcomes Mortality was directly attributable to VRE bloodstream infection in 2 (14.3%) of the 14 patients (Table 3). VRE bloodstream infection contributed to mortality in 3 (21.4%) additional patients. Overall survival at 1 year after allohsct (Figure 1A) did not differ between patients who were not colonized with VRE and patients who were colonized but did not develop bloodstream infection (hazard ratio [HR], 0.8; 95% confidence interval [CI], ; P.55). Of note, patients who had VRE bloodstream infection had significantly worse survival compared to patients without VRE bloodstream infection (HR, 5.1; 95% CI, 1-26; P.05). Four (28.5%) of 14 patients with VRE bloodstream infection in our series had microbiologic failure while receiving daptomycin (n 3) or linezolid (n 1) (Table 3). In patient 109, culture of a femoral venous catheter tip grew VRE on day 24 and the patient was started on daptomycin. Blood cultures at that time were negative. Nine days later, while still on daptomycin, blood cultures grew VRE (daptomycin MIC 4 g/ml). Patient 113 developed a bloodstream infection caused by VRE (daptomycin MIC 4 g/ml) on day 20, received daptomycin, and cleared the bacteremia within 72 hours. Ten days later, while still on daptomycin, blood cultures again grew VRE (daptomycin MIC 4 g/ml), which persisted for 8 days until linezolid was added. For patient 114, VRE (daptomycin MIC 2 g/ml) was isolated from urine and blood cultures on day 3. Linezolid was started but changed 5 days later to daptomycin resulting from pancytopenia. On the 14th day of daptomycin treatment, VRE (daptomycin MIC 3 g/ml) was again isolated from blood cultures. In all 3 cases, daptomycin was dosed at 6 mg/kg daily, or the equivalent dose in the setting of renal insufficiency. All 3 patients died, including 2 within a week after the second episode of bacteremia.

5 VRE among Stem Cell Transplant Recipients 619 Figure 1. (A). Survival curve for patients after allogeneic hematopoietic stem cell transplant, divided into those who were not colonized with vancomycin-resistant enterococcus (VRE), those who were colonized but did not develop a VRE bloodstream infection, and those who developed a VRE bloodstream infection. The graph does not take into account time-dependent nature of the covariates. Survival at 1 year was significantly worse for patients with VRE bloodstream infection compared to the other 2 cohorts (P.05). (B) Pulsed-field gel electrophoresis for blood (B) and stool (S) isolates of VRE. Of these 7 patients, #103, #105, #107, and #108 had different clones identified from blood and stool isolates. indicates a DNA standard. In the 1 case of linezolid failure (patient 101), blood cultures were initially positive (linezolid MIC 1 g/ml) on day 5 after allohsct. A temporary venous catheter was removed and linezolid was started, but a tunneled venous catheter was left in place. Blood cultures remained positive for 7 days (highest linezolid MIC 4 g/ml), until linezolid was changed to daptomycin and gentamicin and the tunneled venous catheter was removed. Microbiology Strain typing performed by PFGE (Figure 1B) identified 9 distinct strains among the 37 patients colonized with VRE, with no strain present in 6 patients. Among the 13 patients with blood and stool typing performed, 5 (35.7%) had different strains identified from stool and blood. DISCUSSION Preemptive and prophylactic approaches are widely used in immunocompromised patients to prevent infections by opportunistic pathogens, including candida, cytomegalovirus, and pneumocystis [23]. In contrast, approaches to prevent or mitigate VRE infection are lacking. We identified an extraordinarily high risk (incidence, 142/10,000 patient days) for VRE bloodstream infection by day 35 after allohsct among patients colonized with VRE. VRE bloodstream infection was associated with a significant reduction in overall survival (OS) and directly caused or contributed to death in over 1/3 of cases. Vancomycin-resistant enterococcus is known to cause nosocomial outbreaks [1,5,9,13]. However, strain typing among the 37 patients colonized with VRE in our series revealed significant heterogeneity, arguing that acquisition did not occur through an outbreak. In addition, 5 (35.7%) of 14 patients with bloodstream infections had different strains identified from the stool and the blood, supporting the possibility of polyclonal colonization. The risk for colonization was highest among patients with acute leukemia or high-risk MDS. Importantly, the treatment of these diseases prior to allohsct typically requires 1 or more hospitalizations for high-dose chemotherapy and broad-spectrum antibiotics for fever and neutropenia, presumably increasing the risk for acquiring VRE colonization. A higher rate of VRE colonization was also reported among hospitalized patients with leukemia (5.5%) at the M.D. Anderson Cancer Center, compared to patients with lymphoma (2.2%) [12]. These rates are markedly lower than in our study. However, 28% of HSC transplant recipients in the M.D. Anderson series who were colonized with VRE developed a VRE bloodstream infection, very similar to the 34.2% rate we report. In fact, several studies have reported rates of VRE bloodstream infection between 25% and 35% among colonized patients, including adult and pediatric stem cell transplant recipients, liver transplant recipients, and patients with hematologic malignancies receiving chemotherapy [12,24-26], despite significant differences in the rates of colonization. Thus, the local epidemiology of VRE colonization appears to be less important in predicting VRE bloodstream infection than the pathogenicity of VRE, including its ability to colonize and persist within intravascular catheters. It remains unclear whether the esp

6 620 D. M. Weinstock et al. gene, a member of the pathogenicity island found within nosocomial VRE isolates, fosters the ability of VRE to form and colonize biofilm [6,27]. The appropriate frequency of stool screening for VRE has not been clearly defined. We elected to screen all patients upon admission and to perform subsequent screening only in patients with diarrhea or upon transfer to the intensive care unit. This approach identified VRE colonization prior to bloodstream infection in all but 1 case and required significantly fewer stool cultures than routine weekly surveillance. Stool culture is 100% sensitive for identifying VRE colonization [28]. Thus, the 11 patients who acquired VRE during the allohsct hospitalization (10 with positive stool cultures and 1 with bloodstream infection without positive stool cultures) may have actually been colonized with VRE upon admission, which subsequently overgrew in the presence of broad-spectrum antibiotics. Some groups have reported high rates of VRE bloodstream infection among allohsct recipients, but concluded that VRE infection is simply a marker, rather than the cause, of poor outcome [16,19]. In these and other studies [15-19], patients who developed VRE bloodstream infection in the early posttransplant period were analyzed together with patients who developed infection much later, frequently in the setting of GVHD. We chose to focus solely on the early posttransplant period, when VRE bloodstream infection is closely associated with neutropenia and mucosal damage resulting from the preparative regimen. If VRE bloodstream infection in this setting were simply a marker of higher infection risk, we would expect the patients with VRE bloodstream infections to have high rates of bloodstream infection by other organisms. Instead, bloodstream infection resulting from organisms other than VRE was more common among patients who did not have VRE infections (25.6%) than among patients with VRE infections (0.0%). The lower rate argues against VRE as a marker of greater risk for infection and possibly for a protective effect from VRE colonization. The latter is supported by the low rate (12.5%) of bloodstream infection from organisms other than VRE among the 24 patients who were colonized by VRE but did not develop a VRE infection. In 5 (35.7%) of the 14 patients with VRE bloodstream infection in our series, mortality was directly attributable to VRE infection (2 patients) or VRE infection contributed to mortality (3 patients). Importantly, day 45 survival among patients with VRE bloodstream infection was 71.4%, compared to 96.2% among patients who did not have a VRE bloodstream infection (absolute difference, 24.8%). If VRE bloodstream infection in the early posttransplant period were simply a marker for higher mortality, we would expect patients with VRE infections to continue to die at higher rates than the patients who did not have VRE bloodstream infections. Instead, the absolute difference of 25% in survival by day 45 remains relatively constant over the subsequent months of follow-up, with comparable late mortality among both groups, presumably from factors like disease relapse and GVHD. Thus, patients who survive their VRE bloodstream infection appear to be at comparable risk for mortality from other complications as patients who did not have a VRE bloodstream infection. Linezolid and daptomycin are the agents used most frequently to treat VRE bloodstream infection. However, their broad spectra of activity and potential side effects, including myelosuppression from linezolid, have contributed to a general reluctance to use them for prophylaxis. Importantly, 4 of 14 patients with VRE bloodstream infection in our series had microbiologic failure while receiving 1 of these agents. In all 3 cases of daptomycin failure, the VRE isolated during therapy remained susceptible in vitro to daptomycin, arguing against a clear correlation between daptomycin MIC and efficacy in highly immunocompromised patients. Our study has several shortcomings. First, the lack of a control arm prevents any conclusions about the direct effects of VRE colonization or VRE bloodstream infection on survival. A randomized study is needed to determine whether an approach that includes: (1) routine stool screening of allohsct recipients, (2) early isolation of colonized patients, and (3) empiric treatment of fever during neutropenia with daptomycin or linezolid in the VRE colonized patients can improve overall mortality. The second shortcoming is the variation in approach between patients, including differences in the day that the stool isolate was obtained and the antimicrobial therapy for VRE bloodstream infection. Third, several patients received care for their underlying malignancy at other centers before transferring their care to our institution for allohsct. Thus, we were unable to analyze the role of certain factors on the risk for VRE colonization or bloodstream infection, including lifetime exposure to antibiotics and lifetime days of hospitalization. Fourth, vancomycin prophylaxis was used in 15 patients, although their rates of VRE colonization (40%) and bloodstream infection among colonized patients (33.3%) did not differ from the group that did not receive vancomycin prophylaxis. Finally, the frequent use of allograft T cell depletion may affect the applicability of our findings to other transplant centers. In conclusion, we find that colonization in the setting of allohsct confers an extraordinarily high risk for VRE bloodstream infection in the early posttransplant period, with a considerable decrement in survival. Approximately 1/3 of patients in our study were VRE colonized, 1/3 of those developed VRE blood-

7 VRE among Stem Cell Transplant Recipients 621 stream infection, and 1/3 of those succumbed to the infection. As a result of these findings, we routinely use daptomycin or linezolid for 48 hours as empiric treatment of fever and neutropenia in allohsct recipients who are colonized by VRE. Although these agents were associated with a high rate of microbiologic failure in our patients, the high mortality of VRE bloodstream infection in the early posttransplant period supports the aggressive use of empiric therapy until culture results are available. Studies of new approaches to prevent colonization, minimize the burden of VRE growth in the gastrointestinal tract, and inhibit seeding of vascular access devices are promptly needed. REFERENCES 1. Bonten MJ, Willems R, Weinstein RA. Vancomycin-resistant enterococci: why are they here, and where do they come from? Lancet Infect Dis. 2001;1: National Nosocomial Infections Surveillance (NNIS) System Report. Data summary from January 1992-June 2001, issued August Am J Infect Control. 2001;29: Stampone L, Del Grosso M, Boccia D, et al. Clonal spread of a vancomycin-resistant Enterococcus faecium strain among bloodstream-infecting isolates in Italy. J Clin Microbiol. 2005;43: Willems RJ, Top J, van Santen M, et al. Global spread of vancomycin-resistant Enterococcus faecium from distinct nosocomial genetic complex. Emerg Infect Dis. 2005;11: Zirakzadeh A, Patel R. Vancomycin-resistant enterococci: colonization, infection, detection, and treatment. Mayo Clin Proc. 2006;81: Leavis H, Top J, Shankar N, et al. A novel putative enterococcal pathogenicity island linked to the esp virulence gene of Enterococcus faecium and associated with epidemicity. J Bacteriol. 2004;186: DiazGranados CA, Jernigan JA. Impact of vancomycin resistance on mortality among patients with neutropenia and enterococcal bloodstream infection. J Infect Dis. 2005;191: DiazGranados CA, Zimmer SM, Klein M, et al. Comparison of mortality associated with vancomycin-resistant and vancomycin-susceptible enterococcal bloodstream infections: a metaanalysis. Clin Infect Dis. 2005;41: Bhavnani SM, Drake JA, Forrest A, et al. A nationwide, multicenter, case-control study comparing risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia. Diagn Microbiol Infect Dis. 2000;36: Bakir M, Bova JL, Newell KA, et al. Epidemiology and clinical consequences of vancomycin-resistant enterococci in liver transplant patients. Transplantation. 2001;72: Zaas AK, Song X, Tucker P, et al. Risk factors for development of vancomycin-resistant enterococcal bloodstream infection in patients with cancer who are colonized with vancomycin-resistant enterococci. Clin Infect Dis. 2002;35: Matar MJ, Safdar A, Rolston KV. Relationship of colonization with vancomycin-resistant enterococci and risk of systemic infection in patients with cancer. Clin Infect Dis. 2006;42: Muto CA, Jernigan JA, Ostrowsky BE, et al. SHEA guideline for preventing nosocomial transmission of multidrug-resistant strains of Staphylococcus aureus and enterococcus. Infect Control Hosp Epidemiol. 2003;24: Recommendations for preventing the spread of vancomycin resistance. Recommendations of the Hospital Infection Control Practices Advisory Committee (HICPAC). MMWR Recomm Rep. 1995;44: Koc Y, Snydman DR, Schenkein DS, et al. Vancomycin-resistant enterococcal infections in bone marrow transplant recipients. Bone Marrow Transplant. 1998;22: Avery R, Kalaycio M, Pohlman B, et al. Early vancomycinresistant enterococcus (VRE) bacteremia after allogeneic bone marrow transplantation is associated with a rapidly deteriorating clinical course. Bone Marrow Transplant. 2005;35: Gregory JJ, Small TN, Papadopulos E, et al. Influence of vancomycin-resistant enterococcus (VRE) on the outcome of recipients of bone marrow transplants. Am Soc Clin Oncol Meeting, Kirkpatrick BD, Harrington SM, Smith D, et al. An outbreak of vancomycin-dependent Enterococcus faecium in a bone marrow transplant unit. Clin Infect Dis. 1999;29: Dubberke ER, Hollands JM, Georgantopoulos P, et al. Vancomycin-resistant enterococcal bloodstream infections on a hematopoietic stem cell transplant unit: are the sick getting sicker? Bone Marrow Transplant. 2006;38: Jaffe D, Jakubowski A, Sepkowitz K, et al. Prevention of peritransplantation viridans streptococcal bacteremia with early vancomycin administration: a single-center observational cohort study. Clin Infect Dis. 2004;39: O Reilly RJ, Collins N, Dinsmore R, et al. Transplantation of HLA-mismatched marrow depleted of T-cells by lectin agglutination and E-rosette depletion. Tokai J Exp Clin Med. 1985; 10: Goossens H, Jabes D, Rossi R, et al. European survey of vancomycin-resistant enterococci in at-risk hospital wards and in vitro susceptibility testing of ramoplanin against these isolates. J Antimicrob Chemother. 2003;51(Suppl 3):iii5-iii Zuccotti G, Strasfeld L, Weinstock DM. New agents for the prevention of opportunistic infections in haematopoietic stem cell transplant recipients. Expert Opin Pharmacother. 2005;6: Kapur D, Dorsky D, Feingold JM, et al. Incidence and outcome of vancomycin-resistant enterococcal bacteremia following autologous peripheral blood stem cell transplantation. Bone Marrow Transplant. 2000;25: McNeil SA, Malani PN, Chenoweth CE, et al. Vancomycinresistant enterococcal colonization and infection in liver transplant candidates and recipients: a prospective surveillance study. Clin Infect Dis. 2006;42: Tsiatis AC, Manes B, Calder C, et al. Incidence and clinical complications of vancomycin-resistant enterococcus in pediatric stem cell transplant patients. Bone Marrow Transplant. 2004; 33: Raad II, Hanna HA, Boktour M, et al. Vancomycin-resistant Enterococcus faecium: catheter colonization, esp gene, and decreased susceptibility to antibiotics in biofilm. Antimicrob Agents Chemother. 2005;49: Shigei J, Tan G, Shiao A, et al. Comparison of two commercially available selective media to screen for vancomycin-resistant enterococci. Am J Clin Pathol. 2002;117: