Bone Marrow Transplantation, (1999) 24, 561–566  1999 Stockton Press All rights reserved 0268–3369/99 $15.00 http://www.stockton-press.co.uk/bmt Safety and efficacy of a continuous infusion, patient controlled anti- emetic pump to facilitate outpatient administration of high-dose

SP Dix, MK Cord, SJ Howard, JL Coon, RJ Belt and RB Geller

Blood and Marrow Transplant Program, Oncology and Hematology Associates and Saint Luke’s Hospital of Kansas City, Kansas City, MO, USA

Summary: and colleagues1 described an outpatient BMT care model utilizing intensive clinic support following inpatient admin- We evaluated the combination of , lor- istration of HDC. This approach facilitated early patient azepam, and delivered as a continuous discharge and significantly decreased the total number of i.v. infusion via an ambulatory infusion pump with days of hospitalization associated with BMT. More patient-activated intermittent dosing (BAD pump) for recently, equipped BMT centers have extended the out- prevention of acute and delayed / in patient care approach to include administration of HDC in patients receiving high-dose chemotherapy (HDC) for the clinic setting. Success of the total outpatient care peripheral blood progenitor cell (PBPC) mobilization approach is dependent upon the availability of experienced (MOB) or prior to autologous PBPC rescue. The BAD staff and necessary resources as well as implementation of pump was titrated to patient response and tolerance, supportive care strategies designed to minimize morbidity and continued until the patient could tolerate oral anti- in the outpatient setting. emetics. Forty-four patients utilized the BAD pump Despite improvements in supportive care strategies, during 66 chemotherapy courses, 34 (52%) for MOB chemotherapy-induced nausea and vomiting continues to be and 32 (48%) for HDC with autologous PBPC rescue. a side-effect causing significant distress to BMT patients The median number of days on the BAD pump during and one that could limit the use of outpatient care during MOB and HDC was 3 (1–6) and 9 (2–19) days, respect- HDC administration. In a survey evaluating symptom dis- ively. Complete overall or complete emesis control tress in BMT patients receiving high-dose chemotherapy in occurred on 94% of MOB and 89% of HDC treatment the inpatient setting, frequency and severity of nausea were days during chemotherapy administration and 72% and among the symptoms associated with the highest degree of 43%, respectively, following chemotherapy adminis- stress upon discharge from hospital.2 To date, limited data tration. Eighty-three percent of MOB and 55% of HDC are available on anti-emetic regimens that can be safely and treatment days were associated with no nausea. While effectively administered to patients receiving HDC in the on the BAD pump, no patient experienced severe tox- outpatient setting. Thus, the development of such regimens icity or required hospitalization for management of is essential to help facilitate outpatient administration of nausea/vomiting. The BAD pump was safe and effective HDC during BMT. The objective of this study was to in minimizing nausea and vomiting associated with evaluate the safety and efficacy of a combination of three HDC, and thus, eliminated the need for hospitalization anti-emetics with different mechanisms of anti-emetic for management of chemotherapy-induced nausea and action, diphenhydramine (Benadryl, Parke-Davis, Morris vomiting. Plains, NJ, USA), (Ativan, Wyeth-Ayerst, Phila- Keywords: high-dose chemotherapy; autologous periph- delphia, PA, USA), and dexamethasone (Decadron, Merck, eral blood progenitor cell transplantation; outpatient BMT; West Point, PA, USA), also called the BAD pump, deliv- anti-emetic therapy ered as a continuous i.v. infusion with patient-controlled i.v. bolus doses, in patients undergoing chemotherapy mob- ilization (MOB) or HDC followed by autologous peripheral blood progenitor cell (PBPC) transplantation in the out- Introduction patient setting. Historically, patients undergoing blood or marrow trans- plantation (BMT) have required prolonged hospitalization Methods for administration of high-dose chemotherapy (HDC) and recovery from regimen-related toxicities. In 1994, Peters Patients were enrolled and evaluated over a 1 year period from October 1996 to October 1997. All patients consented to institutional review board approved treatment protocols Correspondence: SP Dix, Blood and Marrow Transplant Program, Oncology and Hematology Associates of Kansas City, 4320 Wornall including assessments of toxicity and supportive care stra- Road, Suite 220, Kansas City, Missouri 64111, USA tegies. In order to participate, patients were required to have Received 22 May 1998; accepted 28 January 1999 met routine clinical eligibility criteria to undergo autolog- Safety and efficacy of anti-emetic infusion pump SP Dix et al 562 1 mg or BAD pump with basal 32 mg i.v. and = 0.2–0.3 ml/h with 2–3 ml Dexamethasone 10 mg i.v. q 24 h bolus q 15 min lockout on days of i.v. chemotherapy

N/V = no N/V = yes

no change appropriate use inappropriate use

re-evaluate q 24 h and re-educate patient continue BAD pump and caregiver until at least 24 h after PBPC infusion

constant N/V intermittent N/V

increase basal increase bolus by 25–50% by 25–50% (max = 0.6 ml/h) (max = 6 ml/dose)

N/V = no N/V = yes

no change individualize or add rescue anti-emetics

re-evaluate q 24 h and continue BAD pump until at least 24 h after PBPC infusion and patient able to tolerate oral anti-emetics

Additional guidelines:

1 If assessment in outpatient clinic, BMT nurse to make adjustment. If assessment is after hours or on weekend, BMT home health care nurse to make adjustment in conjunction with BMT on-call MD.

2 Options for rescue anti-emetics include: (1) continue to increase basal and/or bolus and titrate to effect; increments must be done in clinic setting; (2) add 0.5–1.0 mg p.o. q 6 h; or (3) add 10 mg p.o.q 6 h

Figure 1 Outpatient anti-emetic algorithm: BAD = diphenhydramine 400 mg i.v., lorazepam 16 mg i.v., and dexamethasone 40 mg i.v. in 100 ml D5W non-PVC at 0.2–0.3 ml/h with 2–3 ml bolus q15 min lockout. Doses for each drug as follows: diphenhydramine 0.8–1.2 mg/h with 8–12 mg bolus; lorazepam 0.032–0.048 mg/h with 0.32–0.48 mg bolus; dexamethasone 0.08–0.12 mg/h with 0.8–1.2 mg bolus.

ous BMT based on disease status and organ function along Regimens for PBPC mobilization or HDC preparative with the following criteria to undergo outpatient care during regimen were selected based on institutional disease-spe- BMT: (1) the patient must have undergone a psychosocial cific protocols. All chemotherapy doses were initiated in assessment by the BMT social worker with or without the morning in the BMT outpatient clinic. Patients received additional assessment by the BMT-designated clinical psy- standard premedications consisting of a antagon- chologist or psychiatrist; (2) the patient must have an ist, ondansetron 32 mg i.v. or granisetron 1 mg i.v., plus approved caregiver or caregivers available 24 h a day start- dexamethasone 10 mg i.v. each day beginning 30 min ing at initiation of MOB or HDC; (3) the patient must have before the first dose of chemotherapy. The BAD pump was lodging, either an approved home or hotel, within a 30-min initiated in the afternoon on the first day of high-dose driving radius of the BMT center along with transportation; chemotherapy and administered as outlined in Figure 1. The (4) the patient and caregiver must have attended the BMT BAD pump was prepared by mixing diphenhydramine 400 Caregiver Class and met class objectives prior to initiation mg i.v., lorazepam 16 mg i.v., and dexamethasone 40 mg of MOB or HDC; (5) the patient must have insurance bene- i.v. in a total volume of 100 ml D5W in a non-poly- fits for outpatient care during BMT; and (6) the patient must vinylchloride bag to minimize adsorption. All patients were have signed an informed consent to receive MOB or HDC initiated at a dose of 0.2–0.3 ml/h continuous basal rate followed by autologous PBPCT in the outpatient setting. with a 2–3 ml patient-activated bolus option programmed Safety and efficacy of anti-emetic infusion pump SP Dix et al 563 with a 15-min lockout between boluses. Dosing was based Table 1 Demographics and results on patient weight with patients weighing more than 80 kg initiated at the higher end of the dosing range. The pump Total No. of BAD pump courses 66 was refilled every 48 h or sooner if needed. Dexamethasone No. of mobilization courses 34 (52%) was deleted after a maximum duration of 96 h. The pump (Regimen = cyclophosphamide 4000 mg/m2 i.v. ␮ was continued until at least 24 h after chemotherapy or on day 1, then G-CSF 10 g/kg/d s.c. day 2 or GM-CSF 500 ␮g/d s.c. day 2 to pheresis) PBPC infusion and the patient was able to tolerate oral anti- No. of HDC with autologous PBPCT courses 32 (48%) emetics. The BAD combination was infused via the Abbott Breast CA Provider 6000 ambulatory infusion pump (Abbott Labora- Cyclophosphamide 2000 mg/m2/d days −5, 19 tories, Abbott Park, IL, USA). Patients and caregivers were −4, −3 instructed on battery alarms and changes, and a back-up set Thiotepa 167 mg/m2/d days −5, −4, −3 2 − − − of 9 V batteries were dispensed with each pump. Flumaz- Carboplatin 267 mg/m /d days 5, 4, 3 enil (Romazicon; Roche Laboratories, Nutley, NJ, USA), Multiple myeloma: Melphalan 100 mg/m2/d 6 days −3, −2 the benzodiazepine reversing agent, with necessary supplies NHL: for administration, was also dispensed at the time of Carmustine 300 mg/m2/d day −73 initiation of the pump and kept in the patient’s home or Cytarabine 100 mg/m2 bid days −6, −5, −4, hotel room. −3 2 − − − Patients were assessed daily in the outpatient BMT clinic Etoposide 100 mg/m bid days 6, 5, 4, −3 while on the BAD pump and after discontinuation. After Cyclophosphamide 35 mg/kg/d days −6, clinic hours, patients were instructed to call the on-call −5, −4, −3 BMT physician or home health care nurse for any problems Other regimena 4 with the pump or refractory nausea and/or vomiting. The Initiation dose = 0.2 ml/h with 2 ml bolus Ȱ 49 (74%) number of emetic episodes (EE) per day was recorded by q15 min lockout = Ȱ the nurses while in the clinic. When outside the clinic, care- Initiation dose 0.3 ml/h with 3 ml bolus 17 (26%) q15 min lockout givers recorded the number of EE on a diary sheet. On the Median duration of BAD pump during MOB 3 days (1–6) following day, the nurse recorded the total number of EE (range) for each 24-h period in the patient chart. An EE was defined Median duration of BAD pump during HDC 9 days (2–19) as a single vomit or retch or any number of continuous (range) episodes of vomiting or occurring within 1 min of No. of BAD courses requiring dose adjustment 17 (26%) each other. Response per 24 h was categorized as follows: No. of BAD courses requiring rescue anti- 5 (7.6%) complete overall control = no EE, no nausea, and no rescue emetics medications; complete emesis control = no EE, no rescue No. of hospitalization for nausea/vomiting 0 medications; major emesis control = 1–2 EE per day. Nau- during BAD courses sea and other toxicities were also assessed daily and graded based on National Cancer Institute Common Toxicity Cri- aHigh-dose combination chemotherapy – cyclophosphamide based. teria. Other data endpoints collected included the duration of the BAD pump, the number of BAD pump dose adjust- ments required, the name and duration of rescue anti- treatment courses and 0.3 ml/h with a 3 ml bolus every 15 emetics prescribed, and the number of hospitalizations for min lockout for the remaining 17 (26%) of the treatment nausea and/or vomiting. courses (Table 1). Four patients (9%) did not receive dexamethasone because of a history of diabetes (n = 3) or an allergy (n = 1). During MOB regimens, the median num- Results ber of days on the BAD pump was 3 days (range 1–6 days). During HDC regimens, the median number of days on the Over the 1 year period, the BAD pump was used with a pump was 9 days (range 2–19 days). Of the 66 treatment total of 66 high-dose chemotherapy treatment regimens in courses, 57 (86%) had data evaluable for efficacy in pre- 44 patients. Demographic information and chemotherapy venting nausea and vomiting, 33 MOB regimens and 24 regimens are outlined in Table 1. In terms of emetogenic HDC regimens. Nine courses were not evaluable for effi- potential, all chemotherapy regimens were classified as cacy due to incomplete or missing data. highly emetogenic based on the Hesketh criteria with the During MOB, complete overall or complete emesis con- exception of high-dose melphalan which is not included in trol occurred on all days of chemotherapy administration the Hesketh proposal.3 Of the 66 treatment regimens, 34 in 88% of BAD treatment courses and on all days following (52%) were for PBPC mobilization and 32 (48%) treatment chemotherapy administration while on the BAD pump in regimens were for HDC followed by autologous PBPC 64% of courses. During HDC, complete overall or complete transplant. All patients received hydration with 1–2 liters emesis control occurred on all days of chemotherapy overnight on the days of chemotherapy. On cyclophos- administration in 75% of BAD treatment courses and on phamide days, hydration included mesna given as a con- all days following chemotherapy administration while on tinuous 24-h infusion at 100% of the cyclophosphamide the BAD pump in 35% of courses. These data are subdiv- dose. ided for complete overall and complete emesis control and The initiation dose of the BAD pump was 0.2 ml/h with depicted graphically in Figure 2. When evaluating treat- a 2 ml bolus every 15 min lockout for 49 (74%) of the ment days on the BAD pump, 59% and 81% of treatment Safety and efficacy of anti-emetic infusion pump SP Dix et al 564 100 the outpatient clinic daily; however, no patient required 90 CEC hospitalization for management of refractory nausea and/or 80 COC vomiting while on the BAD pump. 70 Seventeen of the 66 (25%) treatment courses required a 60 50 median of one dose adjustment in the BAD pump with 82% 40 of dose adjustments occurring during the HDC treatment 30 courses. The maximum dose delivered was 0.4 ml/h with 20 4 ml bolus every 15 min lockout and was given to only 10 % of treatment courses one patient. No patient required rescue anti-emetics while 0 on the BAD pump during MOB. During HDC, five (16%) Chemo Post Total Chemo Chemo Total MOB MOB MOB HDC HDC HDC BAD treatment courses required the addition of rescue anti- emetics (Table 1). COC 26/33 (79%) 11/33 (33%) 12/33 (36%) 7/24 (29%) 3/20 (15%) 5/24 (21%) In terms of toxicity, patients reported mild (grade I) CEC 9/33 (29%) 10/33 (31%) 10/33 (31%) 11/24 (46%) 4/20 (20%) 5/24 (21%) headache during 12.5% of treatment courses. On at least 1 day on the BAD pump, all patients experienced grade I–II Figure 2 Response rate by treatment course: percentage of patients per depressed levels of consciousness consisting of somnolence phase of treatment course who experienced complete overall control (COC) or complete emesis control (CEC) on all days of treatment course. or sedation with or without interference in function, but not interfering with activities of daily living. Dexamethasone was discontinued early during four treatment courses sec- 100 ondary to indigestion (n = 2), hiccoughs (n = 1), and rest- 90 lessness (n = 1). No severe toxicities related to the BAD 80 70 pump occurred. No pump malfunctions occurred which 60 MEC resulted in inaccurate dosing or administration of the BAD 50 CEC infusion. One patient had the pump replaced three times 40 COC because of an inability to infuse; however, this was later 30 determined to be a central line problem requiring line 20 replacement. % of treatment days 10 0 Chemo Post Total Chemo Post Total MOB MOB MOB HDC HDC HDC Discussion

COC 29/35 (83%) 21/58 (36%) 50/93 (54%) 40/72 (56%) 25/145 (17%) 65/217 (30%) CEC 4/35 (11%) 21/58 (36%) 25/93 (27%) 24/72 (33%) 38/145 (26%) 62/217 (29%) Given the complexity of BMT and differences in resources MEC 2/35 (6%) 12/58 (21%) 14/93 (15%) 8/72 (11%) 66/145 (46%) 74/217 (34%) available among centers, various models have been developed and implemented to facilitate outpatient care Figure 3 Response rate by treatment days: percentage of treatment days per phase where patient experienced complete overall control, complete during BMT. The first model to be widely publicized and emesis control, or major emesis control (MEC). noted to decrease hospitalization was the model developed at Duke University where the HDC regimen is delivered in the inpatient setting and then patients are discharged to the days had complete overall or complete emesis control dur- outpatient setting for supportive care management during ing HDC and mobilization courses, respectively. Control neutropenia and the early recovery phase.1 More recently, rates achieved based on treatment days are subdivided for certain BMT centers have attempted to expand the out- complete overall control, complete emesis control, and patient care model to include delivery of high-dose chemo- major emesis control and depicted graphically in Figure 3. therapy in the outpatient setting.4 In order to implement this During mobilization, 67% (22/33) of courses were associa- total outpatient care approach, supportive care protocols ted with no EE while on the BAD pump. While on the must be implemented and include management of acute BAD pump during HDC regimens, 38% (9/24) of courses chemotherapy-related toxicity, particularly gastrointestinal were also associated with no EE on any day, 21% (5/24) complications. While various studies have described suc- with a maximum of one EE per day, and 29% (7/24) with cessful outpatient management of other BMT-related sup- a maximum of two EE per day. portive care issues such as infectious complications, limited During and following MOB, patients experienced no data are available describing successful strategies to prevent nausea on 83% (29/35) and 36% (21/58) of treatment days, and manage acute nausea and vomiting associated with respectively. During and following HDC, patients experi- delivery of HDC in the outpatient setting.5 enced no nausea on 55% (40/72) and 17% (25/145) of treat- In the Duke University outpatient model, the anti-emetic ment days, respectively. In terms of severity of nausea dur- regimen initially implemented consisted of a continuous ing mobilization, the maximum nausea grade experienced infusion of prochlorperazine in combination with lora- on any given day while on the BAD pump was grade 0 for zepam or other individualized rescue anti-emetics.1 Follow- 44% of courses, grade I for 35%, grade 2 for 18%, and ing completion of HDC in the inpatient setting, patients grade 3 for 3%. During HDC regimens, the maximum nau- were often discharged on continuous or twice daily doses sea grade experienced on any given day while on the BAD of i.v. ondansetron delivered via an ambulatory infusion pump was grade 0 for 20%, grade 1 for 20%, grade 2 for pump for management of delayed nausea and vomiting. 40%, and grade 3 for 20%. All patients were evaluated in Miyahara et al6 evaluated the efficacy of an oral anti-emetic Safety and efficacy of anti-emetic infusion pump SP Dix et al 565 strategy of granisetron 1 mg twice daily combined with least major emesis control in the majority of patients. lorazepam or other agents as first-line therapy for manage- Although limited, other data evaluating anti-emetic prophy- ment of acute and delayed nausea and vomiting in a pro- laxis consisting of a serotonin antagonist alone during stem spective study conducted in the inpatient setting. The objec- cell transplantation report complete or major response rates tive of this pilot study was to evaluate the feasibility of of 51–58% during the HDC regimen.10,11 Our response outpatient care during and after delivery of the STAMP V rates were at least comparable to these reports, but more preparative regimen followed by autologous bone marrow importantly facilitated outpatient admininstration of HDC, rescue. Despite aggressive efforts to manage patients with the primary goal of this study. oral medications, Miyahara et al6 reported that over 40% In terms of safety, this regimen was not associated with of patients required a median of 7 days of i.v. anti-emetics, any serious toxicity, despite the continuous use of medi- primarily lorazepam, for management of chemotherapy- cations with sedative properties. No patient experienced induced nausea and vomiting. These data suggest a need to any serious adverse effect due to overuse or pump malfunc- develop a strategy incorporating i.v. anti-emetics in order tion. Patients did experience sedation or somnolence at to successfully deliver HDC in the outpatient setting. some time on the pump, but this was easily managed by Meisenberg et al4 recently published data describing dose titration and the continued presence of a caregiver for their experience on a subtotal outpatient BMT care model assistance. The ambulatory infusion pumps were associated where HDC was administered in the hospital and a total with only minor maintenance needs, primarily battery outpatient BMT care model where HDC was administered changes, once the patient was discharged from the clinic. in the clinic. Meisenberg et al4 report that most patients Routine nursing procedures to minimize potential pump received a combination of continuous infusions of lora- malfunctions such as requiring two nurses to check the set- zepam, prochlorperazine, and diphenhydramine for anti- tings prior to discharge and daily heparin flushes of the emetic prophylaxis. These authors do not provide specific BAD infusion lumen of the central catheter were most information on dosing or administration and do not report likely responsible for the surprisingly low rate of ambulat- data on specific response rates to this regimen in terms of ory infusion pump-related problems. management of nausea and vomiting. Of note, no patient While on the BAD pump, four patients did experience in the total outpatient program was admitted for refractory adverse effects presumed to be related to the dexame- nausea and vomiting. These data are, thus, encouraging and thasone including indigestion, hiccoughs, and restlessness. suggest that HDC can be safely administered in the out- These adverse effects resolved upon discontinuation of the patient setting, yet fail to provide descriptive details on dexamethasone, and patients were then able to be effec- outpatient management of HDC-associated nausea and tively controlled on just the combination of diphenhydram- vomiting. ine and lorazepam in the pump. No other clinically signifi- In our study, the objective was to evaluate the safety cant, potentially steroid-related adverse effects were and efficacy of a combination i.v. anti-emetics given as a observed. The low rate of steroid-induced adverse effects continuous infusion with patient-activated bolus dosing in is most likely related to the short duration of the therapy, patients receiving HDC in the outpatient setting. This but may also have been due to careful patient screening. evaluation also provides practical and descriptive infor- To minimize adverse effects, patients with diabetes (n = 3) mation to the clinician using such treatment regimens. The or a previously documented intolerance to steroids (n = 1) combination of diphenhydramine, lorazepam, and dexame- did not receive dexamethasone in the anti-emetic pump. thasone, called the BAD pump, was delivered as a very low Other initial concerns with the use of the BAD pump basal rate with patient-activated, bolus doses, much like a were the lack of compatibility information available on this patient-controlled analgesia pump administered for control combination and the potential for the BAD pump to con- of acute pain. The drugs used in this combination were stantly occupy i.v. access. Although chemical compatibility selected for their different mechanisms of action against the data are lacking, this combination was physically compat- emetic center and potential to provide a broad scope of anti- ible for at least 48 h based on visual observation. To mini- emetic action. Potential advantages of this regimen include mize adsorption of lorazepam to the plastic containers, non- the psychological benefit of patient control by allowing for polyvinyl-chloride containers were used.12 In addition, dex- self-dosing and the relative ease of delivery by using the trose was used as a diluent instead of saline to minimize ambulatory pump.7 Possible disadvantages include the decomposition of lorazepam. With these concessions to potential for excess sedation with the use of diphenhydram- enhance potency of the lorazepam and based on the premise ine and lorazepam, short-term side-effects related to that the combination was titrated to effect, we felt a chemi- dexamethasone, overuse by the patient, and pump mal- cal compatibility test was not warranted in order to proceed function. with the study objective. In terms of efficacy, we found the BAD pump in combi- In addition to being a safe and effective anti-emetic regi- nation with standard pre-medications of a serotonin anta- men, our regimen is also less costly than some other combi- gonist plus dexamethasone to provide at least comparable nations implemented during stem cell transplantation. For efficacy to first-line anti-emetic regimens used with non- example, the regimen published by Barbounis et al11 utiliz- transplant, highly emetogenic regimens such as high-dose ing a total of 28 i.v. doses of ondansetron 8 mg over an 8- .8,9 The BAD pump was most effective on days of day period is associated with a total drug cost of approxi- chemotherapy administration. On days following chemo- mately $1400 based on 1998 average wholesale prices.12 therapy administration, the BAD pump was less effective Our regimen was associated with a comparable response in providing complete control; however, it did provide at rate and incorporated more drugs, yet costs less than the Safety and efficacy of anti-emetic infusion pump SP Dix et al 566 regimen of Barbounis et al. Considering drug costs only, 2 Lawrence CC, Gilbert CJ, Peters WP. Evaluation of symptom our regimen, including granisetron 1 mg i.v. plus dexame- distress in a bone marrow transplant outpatient environment. thasone 10 mg i.v. daily during chemotherapy for 3 days Ann Pharmacother 1996; 30: 941–945. plus the BAD pump for a median of 9 days, costs approxi- 3 Hesketh PJ, Kris MG, Grunberg SM et al. Proposal for classi- mately $700 for the entire regimen. In addition, our regi- fying the acute emetogenicity of cancer chemotherapy. J Clin Oncol 1997; 15: 103–109. men facilitates outpatient administration of HDC and pre- 4 Meisenberg BR, Miller WE, McMillan R et al. Outpatient vents the need for hospitalization, which has the potential high-dose chemotherapy with autologous stem-cell rescue for to further substantially decrease the overall cost of care hematologic and nonhematologic malignancies. J Clin Oncol during BMT. 1997; 15: 11–17. In summary, we found this combination of diphenydram- 5 Gilbert C, Meisenbert B, Vrendenburgh J et al. Sequential ine, lorazepam, and dexamethasone, the BAD pump, along prophylactic oral and empiric once-daily parenteral antibiotics with a serotonin antagonist plus dexamethasone on days of for neutropenia and fever after high-dose chemotherapy and chemotherapy to be effective in preventing or minimizing autologous bone marrow support. J Clin Oncol 1994; 12: nausea and vomiting during chemotherapy-based MOB and 1005–1011. HDC with autologous PBPC transplantation. In addition, 6 Miyahara TT, Dix SP, Devine SM et al. Evaluation of sup- the safety and effectiveness of this regimen facilitated out- portive care guidelines for monitoring breast cancer bone mar- row transplant patients in the outpatient setting. Blood 1995; patient care and eliminated the need for hospitalization to 86 (Suppl. 1): 213a (Abstr.). administer HDC, as has been the standard of care in most 7 Thomas V, Heath M, Rose D et al. Psychological character- BMT centers. By continuing the BAD pump until the istics and the effectiveness of patient-controlled analgesia. Br patient could be maintained on oral medication, we were J Anaesthesia 1995; 74: 271–276. also able to control and eliminate hospitalization for man- 8 Hesketh PJ, Harvey WH, Harket WG et al. A randomized, agement of delayed nausea and vomiting. Further studies double-blind comparison of intravenous ondansetron alone are warranted to evaluate the role of the BAD pump to and in combination with intravenous dexamethasone in the facilitate outpatient care in other patient populations receiv- prevention of emesis of high-dose cisplatin-induced emesis. J ing highly emetogenic chemotherapy regimens. Clin Oncol 1994; 12: 596–600. 9 Kris MG, Pendergrass KB, Navari R et al. Prevention of acute emesis in cancer patients following high-dose cisplatin with Acknowledgements the combination of oral and dexamethasone. J Clin Oncol 1997; 15: 2135–2138. We are grateful to Keely Hotchkiss RN, and Elaine Stenstrup RN, 10 Okamoto S, Takahaski S, Tanosaki R et al. Granisetron in the for their assistance in development and implementation of this prevention of vomiting induced by conditioning for stem cell research protocol. transplantation: a prospective randomized study. Bone Mar- row Transplant 1996; 17: 679–683. 11 Barbounis V, Koumakis G, Vassilomanolakis M et al. A phase References II study of ondansetron as prophylaxis in patients receiving high-dose polychemotherapy and stem cell trans- 1 Peters WP, Ross M, Vredenburgh JJ et al. The use of intensive plantation. Support Care Cancer 1995; 3: 301–306. clinic support to permit outpatient autologous bone marrow 12 Trissel LA. Handbook on Injectable Drugs. American Society transplantation for breast cancer. Semin Oncol 1994; 21 of Health-System Pharmacists: Bethesda, MD, 1996, pp (Suppl. 7): 25–31. 651–654.