Positive impact of selective outpatient management of high-risk acute myelogenous leukemia on the incidence of septicemia

Transcription

1 Annals of Oncology 18: , 2007 doi: /annonc/mdm112 Published online 17 April 2007 Positive impact of selective outpatient management of high-risk acute myelogenous leukemia on the incidence of septicemia T. Y. Halim 1, K. W. Song 2, M. J. Barnett 2, D. L. Forrest 2, D. E. Hogge 2, S. H. Nantel 2, T. J. Nevill 2, J. D. Shepherd 2, C. A. Smith 2, H. J. Sutherland 2, C. L. Toze 2 & J. C. Lavoie 2 1 Department of Immunology and Microbiology, University of British Columbia, Vancouver; 2 The Leukemia/Bone Marrow Transplant Program of British Columbia, Division of Hematology, Vancouver Hospital Health Sciences Centre, BC Cancer Agency, University of British Columbia, Vancouver, Canada Received 7 September 2006; revised 19 January 2007; accepted 26 February 2007 original article Background: Curative intent chemotherapy for acute myelogenous leukemia (AML) leads to prolonged severe neutropenia, during which patients are highly susceptible to infection. Traditionally these high-risk patients were treated as inpatients. Our center recently implemented a selective ambulatory management policy for AML patients undergoing chemotherapy. Materials and methods: A retrospective analysis was conducted to assess the occurrence of septicemia in AML patients treated over a 5 years period with curative intent chemotherapy. This review encompasses a change in policy from primarily inpatient care to selective outpatient management coupled with prophylactic antibiotic therapy. Results: A total of 294 patients, receiving 623 cycles of chemotherapy were identified. A significant decrease in septicemia was observed from the inpatient to outpatient cohort (22% to 13% P < 0.05), which correlated with the shift towards outpatient treatment of consolidation cycles. A shift from Gram-negative to Gram-positive organisms as the cause of septicemia was also detected in the outpatient cohort, likely due to the introduction of ciprofloxacin prophylaxis. No significant emerging resistance and no septicemia-related mortality were noted in the outpatient cohort. Conclusion: The observed decrease in the incidence of septicemia in the ambulatory cohort adds supportive evidence to the feasibility of selective outpatient management of AML patients with respect to infectious complications. Key words: acute myelogenous leukemia, ambulatory, neutropenia, outpatient, prophylaxis, septicemia introduction Curative intent chemotherapy for acute myelogenous leukemia (AML) leads to prolonged severe neutropenia, during which patients are highly susceptible to infection. Traditionally these patients have been hospitalized for the duration of the chemotherapy until count recovery. An increasing number of institutions are implementing selective early discharge and outpatient protocols of intensive chemotherapy in response to limited health care resources and to increase patient comfort. Ambulatory care has regularly been described in solid tumours, high-dose chemotherapy followed by autologous hematopoietic stem-cell transplantation (HSCT) and more recently in the allogeneic nonmyeloablative HSCT setting, despite earlier studies of protective isolation indicating an infection preventative benefit [1 3]. A small number of studies have reported outpatient management of select patients with AML *Correspondence to: Dr J. C. Lavoie, Vancouver General Hospital, Division of Hematology, Room 10149, 10th floor, 2775 Laurel Street, Vancouver, BC V5Z 1M9, Canada. Tel: ; Fax: ; jlavoie@bccancer.bc.ca [4 8]. These investigations have documented the feasibility of outpatient management in high-risk neutropenic AML patients [9, 10]. The principal concerns with ambulatory care are the reduced surveillance of patients for febrile neutropenia and bleeding, and the possible delay in medical intervention, which could compromise the safety of patients. Building on our previous experience with the outpatient management of AML patients treated with curative intent chemotherapy [11], we conducted a retrospective study of the incidence of septicemia in AML patients over a 5 years period. The time period of the review included before and after a change in policy from primarily inpatient management to selective outpatient management coupled with prophylactic antibiotic therapy, to allow a comparative retrospective analysis of both treatment periods and safety assessment. materials and methods institution, setting The Leukemia/Bone Marrow Transplant (L/BMT) Program of British Columbia (BC) is a referral adult provincial program for patients from BC ª 2007 European Society for Medical Oncology

2 Annals of Oncology and the Yukon. The L/BMT program is centered at the Vancouver General Hospital site and is associated with the BC Cancer Agency and the University of BC. The program currently sees >350 new patient referrals annually, including 80 cases of new AML per year. The Research Ethics Board of the University of BC approved this retrospective study. population The population of interest met the following inclusion criteria: all adult AML patients treated by the L/BMT program of BC from February 1999 to February 2004 with curative intent chemotherapy for de novo, secondary AML, and treatment-related AML. Some patients have been already described in a previous report [11]. Exclusion criteria removed AML patients not treated with the standard induction or consolidation chemotherapy regimen from the study, and patients treated with palliative intent. Salvage cycles of intensified chemotherapy such as VP16 cyclophosphamide or cyclophosphamide carboplatin given for refractory or relapsed AML were also excluded, as those are never done as outpatient in our centre. All patients signed informed consent. treatment protocol A variety of chemotherapy regimens were administered and are described in Table 1. Chemotherapy regimens were classified as or similar (group 7 + 3), or as high-dose cytarabine-containing regimens (group HIDAC). In our institution, patients <60 years of age were preferentially treated with HIDAC. Historically, patients have been hospitalized for the duration of chemotherapy and until count recovery. Since September 2001, our institution has implemented a selective ambulatory protocol, which allows the majority of consolidation cycles to be administered entirely as outpatient. In addition, some select patients receiving induction cycles are discharged early from the inpatient ward. Inpatient cycle (IptC) implied that patients received chemotherapy and supportive care as inpatients, and left the hospital after day+15 after chemotherapy, or after absolute neutrophil count (ANC) /l. Outpatient cycle (OptC) implied ambulatory administration of chemotherapy and supportive care. Early discharge cycle (EdcC) implied that patients received chemotherapy as inpatients, but left the hospital before day+15 after chemotherapy or before ANC /l. For all patients, antimicrobial prophylaxis consisted of acyclovir 600 mg p.o. q.d.s. or valacyclovir 500 mg p.o. o.d. (if herpes Table 1. Chemotherapy regimen Group Cytarabine 200 mg/m 2 on days 1 7, daunorubicin 45 mg/m 2 on days Cytarabine 200 mg/m 2 on days 1 5, daunorubicin 30 mg/m 2 on days (M3) Cytarabine 200 mg/m 2 on days 1 6, daunorubicin 60 mg/m 2 on days 1 3, ATRA 60 mg/m 2 on days 1 60 (induction only) (M3) Daunorubicin 50 mg/m 2 on days 1 3, ATRA 45 mg/m 2 on days 1 7 Group HIDAC HIDAC Cytarabine 3.0 gm/m 2 on days 1 6 HIDAC + DN Cytarabine 3.0 gm/m 2 on days 1 6, daunorubicin 45 mg/m 2 on days 1 3 M3, promyelocytic leukemia; ATRA, all-trans retinoic acid; HIDAC, high-dose cytarabine-containing regimen; DN, daunorubicin. original article simplex virus positive), and fluconazole mg p.o. o.d. or itraconazole 200 mg p.o. BID (if previously proven or probable Aspergillus infection) until ANC recovery. Inpatients did not receive prophylactic antibacterial prophylaxis. After 1 September 2001, ciprofloxacin 500 mg po BID was added as antibacterial prophylaxis to ambulatory patient, starting on the day following chemotherapy (OptC) or on the day of discharge (EdcC) until ANC recovery. Patients were not routinely treated with granulocyte colony-stimulating factor. Patients were transfused two units of packed red cells if the hemoglobin fell <90 gm/l and either five units of random donor platelets or one unit of single donor platelets if the platelet count was < /l and the patient was afebrile or if the platelet count was < /l and the patient was bleeding or febrile. All patients had indwelling Hickman-type catheters. ambulatory clinic To be candidates for the outpatient or early discharge chemotherapy program every AML patients must be clinically stable, had accommodations <1 h from our day care and, had a willing and able caregiver. Affordable housing was made available for out-of-town patients. For patients who could not meet the expense of outpatient treatment, a referral was made to social services to find extra resources and founding. The ambulatory clinic is open 7 days/week from 8 AM to 7 PM. An attending physician, two clinical associates or fellows, and four to seven nurses staff this clinic. Routine and walk-in assessments, laboratory, blood bank, and pharmacy support are available during clinic hours. AML patients in this analysis were seen in the L/BMT ambulatory clinic at least three times per week for routine vital signs checkup, bloodwork, and medical assessment. The care team discusses all patients on a daily basis. Further testing is ordered and therapy modified according to their current clinical situation. Outside clinic hours patients are instructed to contact the L/BMT physician on call if they develop a temperature >38 C or if they have other concerns. As required the physician can then chose to evaluate the patient further by having them present to the hospital s emergency within the hour. Empiric i.v. antibiotic therapy was immediately initiated in febrile neutropenic episodes, consisting of tobramycin 5 mg/kg i.v. once daily, ceftriaxone 1 2 g/kg i.v. once daily and vancomycin 20 mg/kg i.v. once daily [12]. Febrile neutropenic patients were reevaluated daily to determine their eligibility for continued ambulatory treatment based. The criteria for admission of patients with neutropenic fever included hemodynamic instability (hypotension unresponsive to fluid challenge or marked tachycardia), hypoxia (O 2 saturation <92% on room air), a temperature >39.5 C, fever unresponsive to therapy with antibiotics for 3 days, rigors, World Health Organization grade 3 bleeding, requirement of i.v. antibiotics more than o.d., neutropenic colitis, failure to thrive as an outpatient or if the caregiver was unable to adequately care for the patient. AML neutropenic patients who were requiring admission to the hospital were admitted on the same day or night. Direct admissions to the L/BMT unit were requested in priorities. If this was not possible, off-service bed or emergency room admission were done. The L/BMT attending were responsible for the offservice AML patients. microbial samples and definitions Blood samples for culture were drawn from indwelling catheters upon presentation with an oral temperature of 38.0 C on two or more occasions over a 12 h period, a single oral temperature of 38.3 C, or hemodynamic instability. Subsequent cultures were obtained 2 3 times a week in the event of prolonged febrile episodes. Each set consisted of an aerobic and anaerobic bottle with 10 ml blood in each, incubated at 35.0 C in the BACTEC 9240 continuous monitor system (Becton, Dickinson and Company, Franklin Lakes, NJ). Aspartate aminotransferase was carried out according to Clinical and Laboratory Institute (formerly National Committee on Clinical Laboratory Standards) using either disk diffusion Volume 18 No. 7 July 2007 doi: /annonc/mdm

3 (Kirby Bauer) or microbroth breakpoint technique (MicroScan, Dade Behring, Sacramento, CA) [13]. Due to the high-risk nature of this patient population, septicemia was diagnosed upon presentation of one positive blood culture bottle regardless of organism in a setting of clinical significance (presence of fever, chills or shock, or the presence of a likely source of infection). Septicemia episodes were defined as isolation from blood of one (monomicrobial) or more (polymicrobial) organisms per chemotherapy cycle. data collection, chart data entry Eligible patients were identified by a computer query of the patient database (BMT Serve 3.0) maintained and updated prospectively by the L/BMT program of BC. We found 294 eligible patients from all AML cases evaluated during the study period. A dataset of chemotherapy cycles administered to inpatient and outpatient AML patients were obtained. In addition to the electronic database, individual chart reviews were conducted to complete any missing information. Data collected included the following: patient age, sex, diagnosis, type of treatment (including chemotherapy regimen, date, chemotherapy cycle, outcome, and IptC\OptC\EdcC status), hospital length of stay, ANC recovery period, complications, current patient status, last day of contact, relapse date, HSCT date, length of follow-up, and the identification and antibiograms of bacterial isolates. Microbiological results and antibiograms of documented bloodstream infections were separately analyzed. statistical methods The primary outcome was the development of clinically significant positive blood cultures (septicemia). Stratification was carried out according to the type of cycle (induction or consolidation), chemotherapy regimen (group versus group HIDAC), and IptC versus OptC and EdcC. All data were analyzed using the SPSSÒ statistical software package (SPSS, Inc., Chicago, IL). Student unpaired t-tests, Fisher s exact tests, and chi-square tests were used for univariated analysis. Logistic regression was used to carry out multivariate analysis. Results were considered significant at P 0.05 (two-tailed for all tests). Table 2. Patient and cycle characteristics Patient No. of patients 294 Total no. of cycles 623 Gender, male (female) 156 (138) Age, median (range) 52 (17 76) years AML FAB Subtypes M0 (25), M1 (49), M2 (46), M3 (43), M4 (61), M5 (39), M6 (3), M7 (11), 2N AML (7), TX-AML (7), BI (3) ANC recovery (n = 516) Median (range) days 24 (8 122) Platelet recovery (n = 240) Median (range) days 19 (4 59) AML, acute myelogenous leukemia; FAB, French American British; 2N AML, secondary AML; TX AML, treatment-related AML; BI, Bi-phenotypic; ANC, absolute neutrophil count. Table 3. AML chemotherapy cycle distribution Annals of Oncology Cycle Induction Consolidation Total IpT OpT EdC IpT OpT EdC a (53%) HIDAC b (47%) Subtotal a 7 + 3, 5 + 3, 96-02, chemotherapy regimen. b high-dose cytarabine-containing regimen +/2 daunorubicin chemotherapy regimen. AML, acute myelogenous leukemia; IpT, inpatient; OpT, outpatient; EdC, early discharge patient. results patients characteristics From February 1999 to February 2004, a total of 623 cycles of chemotherapy that met our inclusion criteria were administered to 294 patients, with a mean of 2.1 cycles per patient (range 1 8) (Table 2). In all, 332 cycles of the group chemotherapy regimen and 291 cycles of the group HIDAC chemotherapy regimen were administered. Chemotherapy sequence data indicated that 328 cycles were given as induction chemotherapy, which includes reinductions with conventional chemotherapy (no intensification) after failed induction cycles or relapses; 295 cycles were given as consolidation chemotherapy. When comparing patient discharge status, we observed that 426 cycles were spent entirely as inpatient (IptC); 157 cycles were spent entirely as outpatient (OptC) and 40 cycles as early discharge (EdcC) (Table 3). septicemia One hundred and twenty-six episodes of microbiologically defined septicemia were recorded in the 623 cycles of chemotherapy (20%), of which 21 episodes yielded >1 identified organism. All polymicrobial episodes occurred in different patients. Age of patients with septicemia was 48.4 versus 53.5 years old for patients without septicemia (P = 0.01). When stratified by chemotherapy sequence, we noted an increase in septicemia incidence from 17% (55/328) for induction cycle chemotherapy to 24% (71/295) for consolidation cycle chemotherapy (P 0.05). Percent occurrence of septicemia stratified by chemotherapy regimen indicated 17% (57/332) for group and 24% (69/291) for group HIDAC (P 0.05). ambulatory program The incidences of septicemia episodes when stratified by discharge status were 22%, 13%, and 28% for IptC, OptC, and EdcC, respectively (P 0.05) (Table 4). During OptC, there were 27 admissions (17%) before day+30 of chemotherapy. Two patients were admitted to continue chemotherapy (day+1 and day+2) as inpatient which was considered to be more appropriate for them and one patient progressed with AML during treatment was admitted on day+8 for salvage chemotherapy. Twenty-four outpatients were admitted on day+16 (range day+1 to day+21) for complication: 12 for sepsis, six for neutropenic fever, three for pneumonia [one aspergillosis, two not otherwise specified (NOS)], one for cerebellar toxicity, one for pain investigation, and one for 1248 Halim et al. Volume 18 No. 7 July 2007

4 Annals of Oncology original article Table 4. Septicemia incident by discharge status results, substratified by induction and consolidation chemotherapy and by and HIDAC chemotherapy Cycle Septicemia Ipt Septicemia Opt Septicemia Edc Total IND 50/294 1/12 4/22 55/328 (17%) CON 44/132 20/145 7/18 71/295 (24%) Total 94/426 (22%) 21/157 (13%) 11/40 (28%) 126/623 (20%) a 42/222 8/85 7/25 57/332 (17%) HIDAC b 52/204 13/72 4/15 69/291 (24%) Total 94/426 (22%) 21/157 (13%) 11/40 (28%) 126/623 (20%) a 7 + 3, 5 + 3, 96-02, chemotherapy regimen. b HIDAC +/2 daunorubicin chemotherapy regimen. HIDAC, high-dose cytarabine-containing regimen; IpT, inpatient; OpT, outpatient; EdC, early discharge patient; IND, induction chemotherapy; CON, consolidation chemotherapy. perianal abscess. Only one ambulatory patient required intensive care unit (ICU) admission. This patient was admitted for Serratia marcescens sepsis on the first day of chemotherapy. Less than 24 h after admission the patient was transferred to ICU for fluid resuscitation and vasopressor. The patient recovered and was discharged. The patient received complete chemotherapy as outpatient 2 weeks later. Other complications not requiring admission during OptC included one central line infection, one urinary infection, one cellulites infection, one pneumonia NOS, and three fungal pneumonia. No EdcC patient was readmitted before day+30 after their chemotherapy. Beside septicemia, in the EdcC patients were three episodes of central line infections and five pneumonia (fungal: 1, Streptococci viridans: 2, NOS: 2) after discharge. No severe bleeding and no neutropenic colitis were experienced in the ambulatory cohort. blood culture/resistance patterns A total of 152 clinically significant isolates were found, grouped as 55% Gram-positive bacteria, 39% Gram-negative bacteria, 4% fungal organisms, and 2% NOS (Table 5). Coagulasenegative Staphylococci and S. viridans were the most frequently isolated pathogens recovered in 22% and 18% of cases, respectively. This was followed by Escherichia coli in 11% and Klebsiella spp. in 8%. A significant decrease in incidence of Gram-negative septicemia occurred in ambulatory care patients (with ciprofloxacin prophylaxis) compared with inpatients (without ciprofloxacin prophylaxis) from 46% to 36% of bloodstream infections (P 0.05). Resistance to ciprofloxacin was seen in 25% of E. coli. We encountered 14 episodes of E. coli septicemia during InptC and two episodes of E. coli septicemia during OptC. Fourteen percent (2/14) of the IptC isolates displayed ciprofloxacin resistance, compared with 2/2 of the OptC E. coli isolates. In addition, two methicillin resistant Staphylococcus aureus were isolated, and no vancomycin resistance was encountered in any isolates. uni/multivariate analysis Univariate analysis identified the chemotherapy regimen (group HIDAC), as well as the chemotherapy cycle (consolidation) as a significant factor (P 0.05) for higher risk of septicemia incidence. Multivariate analysis identified chemotherapy regimen (group HIDAC) as a significant determining factor in the occurrence of septicemia. outcome One hundred and fifty-three patients are alive with a mean follow-up of 774 days [range days, 95% confidence interval (CI) days] while 141 patients have died since diagnosis. Overall, 89 patients relapsed at some point after diagnosis and 73 patients went on to receive HSCT. Survival at 12 months was 64%, (95% CI 58 69%). The cause of death was relapse or refractory AML in 78% (110/142), treatmentrelated mortality (TRM) post-hsct in 13% (19/142), and TRM after chemotherapy in 7% (10/142). Of the 10 patients who died of postchemotherapy, only one died as a direct result of septicemia (Table 6). No treatment-related deaths occurred as a result of infection in the ambulatory population. discussion Ambulatory management of neutropenic AML patients with curative intent chemotherapy is gradually increasing in acceptance. Investigations currently published differ in patient selection, treatment design, and methodology; however, the main objectives of these investigations are similar in nature and focus on the safety and feasibility of ambulatory care [4 8]. This retrospective study was designed to evaluate the incidence and microbial etiology of septicemia following the implementation of an ambulatory policy for AML patients. It also allowed us to elucidate some of the risk factors for septicemia and confirmed feasibility and safety. To observe any trends in infection due to the implementation of the new ambulatory policy, the study period included 2.5 years before and following the change in policy from primarily inpatient management to selective ambulatory management with prophylactic antibiotic therapy. Our results appear to indicate that with our current discharge guidelines and antimicrobial prophylaxis a decrease in septicemia is attainable, which translates into a feasible ambulatory policy. We recognize that patient bias was inevitable by choosing for ambulatory care only patients with favorable performance status. Multivariate statistical analysis of septicemia incidence implicated the HIDAC chemotherapy regimen as a significant Volume 18 No. 7 July 2007 doi: /annonc/mdm

5 Annals of Oncology Table 5. Breakdown of isolated microorganisms Table 6. Treatment-related mortality Isolates Total no. of cases Before 1 September 2001 Gram-positive Staphylococcus (coagulase negative) Viridans Streptococcus Enterococcus spp Stapylococcus aureus Micrococcus Corynebacterium spp Stomatococcus mucilaginosus Streptococcus pneumoniae Group B Streptococcus Gram-positive cocci NOS Gram-negative Escherichia coli Klebsiella spp Serratia spp Enterobacter spp Citrobacter spp Pseudomonas aeruginosa Acinetobacter spp Stenotrophomonas maltophilia Fusobacterium Aeromonas veronii Capnocytophaga Moraxella sp Fungi Candida albicans Candida tropicalis Candida krusei Fusarium Nonspecified bacteria Total NOS, not otherwise specified. After 1 September 2001 risk factor. Stratification of septicemia incidence by chemotherapy regimen reflects this finding, indicating an increase in septicemia from 17% for group to 24% for group HIDAC chemotherapy. Destruction of oral and gastrointestinal mucosa during intensive chemotherapy is closely associated with infection; therefore, these results were anticipated and are consistent with an increase in mucosal toxicity and prolonged neutropenia due to high-dose cytarabine [14 18]. Also of interest is the stratification of septicemia incidence by chemotherapy cycle, which shows an increase from 17% for induction to 24% for consolidation cycles. This uncharacteristic trend defied our expectations that we would anticipate higher incidence of septicemia during Patient Chemotherapy regimen Patient status Cause of death inpatient MOF 2 HIDAC inpatient MOF inpatient MOF inpatient MOF 5 HIDAC + DN inpatient Liver failure inpatient Renal failure outpatient Cardiomyopathy 8 HIDAC + DN inpatient Pneumonia 9 HIDAC + DN inpatient Aspergillosis 10 HIDAC inpatient Bacterial sepsis MOF, multiple organ failure; HIDAC, high-dose cytarabine-containing regimen; DN, daunorubicin. induction cycles due to more frequent poor performance status and comorbidities in patients at the time of diagnosis. Careful review of our data indicated that septicemia during inpatient consolidation (without antibiotic prophylaxis) is most responsible for this unexpected finding. The use of antibiotic prophylaxis in neutropenic patients has been controversial for many years. Antibiotic prophylaxis has frequently been cited as a driving force in the emergence of resistant strains. Of major concern is the reported increase in fluoroquinolone resistant strains of E. coli [19, 20]. Until recently, clinicians were faced with the dilemma of choosing between Gram-negative prophylaxis in high-risk neutropenic AML patients, which has been shown to reduce infectious complications but without a major impact on infection-related mortality or diminish the emergence of antibiotic resistance by withholding prophylactic antibiotherapy. The Infectious Diseases Society of America guidelines published in 2002 permit fluoroquinolone prophylaxis in high-risk, profoundly neutropenic patients [12]. Last year, the benefits of prophylactic antibiotics have been emphasized by two major double-blind, placebo controlled trials with levofloxacin which confirmed very significant reductions in all infection-related events (febrile episode, positive culture, bacteremias infections with Gram-negative bacilli). Bucaneve et al. [21] target 760 high-risk adults with therapy for leukemia or an autologous HSCT while Cullen et al. [22] randomized patients with solid tumor or lymphoma. Mortality was lower in the levofloxacine groups but both studies were not powered to prove a difference in mortality. A meta-analysis comparing prophylactic antibiotic therapy with placebo, confirmed a survival advantage (infection-related death and all cause mortality) and this is greatest with the use of fluoroquinolone [23]. It was recently updated to include Bucavene and Cullen studies. Among acute leukemia and HSCT patients the relative risk of death with quinolone prophylaxis was ( ) [24]. Nevertheless, vigilance is required to detect epidemiological shifts in pathogens and their susceptibility patterns to current prophylactic and empiric therapy. Taking into account these new evidences and the severe immunocompromised state of AML chemotherapy patients, our institution will continue to utilize fluoroquinolone prophylaxis in the ambulatory 1250 Halim et al. Volume 18 No. 7 July 2007

6 Annals of Oncology subgroup to prevent life-threatening Gram-negative sepsis and is considering universal bacterial prophylaxis for all AML. Over the past decade, the spectrum of pathogens that cause infections in neutropenic patients has shifted from Gramnegative bacilli to Gram-positive microorganisms possibly attributable to the increased use of prophylaxis that targets Gram-negative bacteria, intensified chemotherapy protocol, and the use of indwelling venous catheters [25, 26]. Our experience corroborates the shift in the bacterial spectrum following the implementation of selective ambulatory care with antibiotic prophylaxis targeting Gram-negative organisms. Although our dataset on ciprofloxacin resistance in E. coli is insufficient for statistical analysis over the study period, we do report a nonsignificant increasing trend in ciprofloxacin resistance (1/9 before September 1st 2001, 3/7 after September 1st 2001). This coincides with a similar hospital wide trend of decreasing ciprofloxacin susceptibility in E. coli from 92% in 1999 to 69% in 2004 (Vancouver General Hospital internal data). Notably, in the wake of reported emergence of vancomycin resistance [27], no resistance was observed in Enterococci spp. or other bacteria despite its use as front line empiric antibiotherapy. conclusion The obtained data has led us to make several important observations. First of all, we observe a decrease in septicemia in the consolidation subgroup of curative intent chemotherapy recipients postimplementation of selective discharge protocol; this shift can be correlated with a significant decrease in septicemia in the ambulatory population, leading us to conclude that the selective ambulatory management of AML chemotherapy recipients in our patient population is feasible with respect to bloodstream infections. Secondly, a notable decrease in the Gram-negative pathogens causing septicemia is experienced in the Outpatient population. We speculate that this shift is mainly due to the Gram-negative targeting ciprofloxacin antibiotic prophylaxis administered to the ambulatory population. While this practice is not advantageous in all situations due to the possible selection for antibiotic resistance, given the severe immunocompromised state of this patient group, we feel that ciprofloxacin prophylaxis is warranted. A comprehensive ambulatory management program specifically dedicated to high-risk AML proved to be feasible and safe. acknowledgements The authors wish to acknowledge the contribution of the medical and nursing staff of Leukemia/BMT Ward T15A, BMT Daycare at the Vancouver General Hospital and 6 West Ward at the British Columbia Cancer Agency. We would also like to thank Diane Roscoe for reviewing the manuscript, Janet Nitta and Alan Le for data assistance and Emily Gushe for help with manuscript preparation. Grants or external funding did not support this study. This study was presented at the American Society of Hematology Meeting, December 2004 [Blood 2004; 104: 252a (Abstr 884)]. references original article 1. Ferrara F, Palmieri S, Viola S et al. Outpatient-based peripheral blood stem cell transplantation for patients with multiple myeloma. Hematol J 2004; 5: Ruiz-Arguelles GJ, Gomez-Almaguer D, Ruiz-Argüelles A et al. Results of an outpatient-based stem cell allotransplant program using nonmyeloablative conditioning regimens. Am J Hematol 2001; 66: van Tiel FH, Harbers MM, Kessels AG, Schouten HC. Home care versus hospital care of patients with hematological malignancies and chemotherapy-induced cytopenia. Ann Oncol 2005; 16: Allan DS, Buckstein R, Imrie KR. Outpatient supportive care following chemotherapy for acute myeloblastic leukemia. Leuk Lymphoma 2001; 42: Girmenia C, Alimena G, Latagliata R et al. Out-patient management of acute myeloid leukemia after consolidation chemotherapy. Role of hematologic emergency unit. Haematologica 1999; 84: Girmenia C, Latagliata R, Tosti S et al. Outpatient management of acute promyelocytic leukemia after consolidation chemotherapy. Leukemia 1999; 13: Gillis S, Dann EJ, Rund D. Selective discharge of patients with acute myeloid leukemia during chemotherapy-induced neutropenia. Am J Hematol 1996; 51: Ruiz-Arguelles GJ, Apreza-Molina M, Aleman-Hoey DD et al. Outpatient supportive therapy after induction remission therapy in adult acute myeloid leukemia (AML) is feasible: a multicentre study. Eur J Haematol 1995; 54: Klastersky J, Paesmans M, Rubenstein EB et al. The multinational association for supportive care in cancer risk index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 2000; 18: Paesmans M. Risk factors assessment in febrile neutropenia. Int J Antimicrob Agents 2000; 16: Savoie ML, Nevill TJ, Song KW et al. Shifting to outpatient management of acute myeloid leukemia: a prospective experience. Ann Oncol 2006; 17: Hughes WT, Armstrong D, Bodey GP et al guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 2002; 34: Clinical and Laboratory Standard Institute (formerly NCCLS). Performance Standards for Antimicrobial Susceptibility Testing; Fifteenth Informational Supplement, M100-S11. Wayne, PA: Clinical and Laboratory Standard Institute Rubenstein EB, Peterson DE, Schubert M et al. Clinical practice guidelines for the prevention and treatment of cancer therapy-induced oral and gastrointestinal mucositis. Cancer 2004; 100: Tsuji E, Hiki N, Nomura S et al. Simultaneous onset of acute inflammatory response, sepsis-like symptoms and intestinal mucosal injury after cancer chemotherapy. Int J Cancer 2003; 107: Pico J, Avila-Garavito A, Naccache P. Mucositis: its occurrence, consequences, and treatment in the oncology setting. Oncologist 1998; 3: Bishop JF, Matthews JP, Young GA et al. A randomized study of highdose cytarabine in induction in acute myeloid leukemia. Blood 1996; 87: Mayer RJ, Davis RB, Schiffer CA et al. Intensive postremission chemotherapy in adults with acute myeloid leukemia. N Engl J Med 1994; 331: Lautenbach E, Strom BL, Nachamkin I et al. Longitudinal trends in fluoroquinolone resistance among enterobacteriaceae isolates from inpatients and outpatients. Clin Infect Dis 2004; 38: Baum HV, Franz U, Geiss HK. Prevalence of ciprofloxacin-resistant Escherichia coli in hematologic-oncologic patients. Infection 2000; 28: Bucaneve G, Micozzi A, Menichetti F et al. Levofloxacin to prevent bacterial infection in patients with cancer and neutropenia. N Engl J Med 2005; 353: Volume 18 No. 7 July 2007 doi: /annonc/mdm

7 22. Cullen M, Steven N, Billingham L et al. Antibacterial prophylaxis after chemotherapy for solid tumors and lymphomas. N Engl J Med 2005; 353: Gafter-Gvili A, Fraser A, Paul M, Leibovici L. Meta-analysis: antibiotic prophylaxis reduces mortality in neutropenic patients. Ann Intern Med 2005; 142: Leibovici L, Gafter-Gvili A, Paul M et al. Implications of the evidence for antibiotic prophylaxis for cancer patients with neutropenia: updated meta-analysis. Eur Soc Clin Microbiol Infect Dis 2006 Abstr 256. Annals of Oncology 25. Zinner SH. Changing epidemiology of infections in patients with neutropenia and cancer: emphasis on gram-positive and resistant bacteria. Clin Infect Dis 1999; 29: Soo RA, Gosbell IB, Gallo JH et al. Hickman catheter complications in a haematology unit, Intern Med J 2002; 32: Tenover FC, McDonald LC. Vancomycin-resistant Staphylococci and Enterococci: epidemiology and control. Curr Opin Infect Dis 2005; 18: Halim et al. Volume 18 No. 7 July 2007