Efficacy of Oral Ondansetron in the Prevention of Emesis in Outpatients Receiving Cyclophosphamide-based Chemotherapy

Methods

Study Design

This study was a multicenter, randomized, double-blind, dose-comparison, stratified, placebo-controlled trial designed to evaluate the efficacy and safety of oral ondansetron (Zofran; Glaxo, Inc., Research Triangle Park, North Carolina) for the control of nausea and vomiting in patients receiving cyclophosphamide-based chemotherapy. Patients were randomly assigned to one of four treatment groups: 1 mg, 4 mg, and 8 mg of ondansetron, or placebo. The study was stratified to compare doxorubicin- with nondoxorubicin-containing regimens. All study medications were matched tablets prepackaged by the pharmacy (Glaxo Inc.) before shipment to study centers.

Criteria for Entry

Patients with histologically confirmed cancer were eligible for this clinical trial if they were chemotherapy naive, 18 years or older, had a Karnofsky performance status 60% or more, and were scheduled to receive cyclophosphamide-based chemotherapy. Women of childbearing age had a negative serum or urine β-human chorionic gonadotropin-pregnancy test before entry. Patients were excluded if they had a serum alanine aminotransferase (ALT) level greater than twice the institution's upper limit of normal, had any retching or vomiting during the 24 hours before study entry, received any investigational agent in the previous 30-day period, or were scheduled to receive pelvic or abdominal radiation therapy during the 24 hours before or during the 72 hours after study entry. Concomitant administration of cisplatin, carboplatin, dacarbazine, nitrogen mustard, procarbazine, carmustine, or ifosfamide, or both made the patient ineligible. In addition, use of any antiemetic agents including phenothiazines, butyrophenones, lorazepam, cannabinoids, metoclopramide, corticosteroids, and trimethobenzamide during the previous 24 hours or during the 3-day study period excluded the patient. All patients gave written informed consent, and the protocol was reviewed and approved by the institutional review board of each participating center.

Pretreatment and Follow-up Examinations

Pretreatment evaluations were done within 7 days of study initiation. They included a complete medical history and physical examination, a complete blood count and serum chemical analysis, including blood urea nitrogen (BUN), creatinine, total bilirubin, lactic dehydrogenase, alkaline phosphatase, aspartate aminotransferase (AST), and ALT. Complete blood counts and all serum chemical analyses were repeated within 1 week of chemotherapy completion. Patients received cyclophosphamide (≥ 450 mg/m²) as a short (≤ 2 hours) infusion plus either doxorubicin (≥ 35 mg/m²) or methotrexate (≥ 30 mg/m²). Other low to moderately emetogenic chemotherapy was allowed, including 5-fluorouracil (≥ 400 mg/m²), vincristine, and etoposide (≥ 35 mg/m²).

Antiemetic Schedule and Rescue

Patients were randomized to one of four study groups (placebo or 1 mg, 4 mg, or 8 mg of ondansetron). The first dose of study medication was taken 30 minutes before the start of cyclophosphamide administration. Two additional doses of study drug were taken 4 and 8 hours after the start of the cyclophosphamide infusion. On study days 2 and 3, study drug was taken at approximately 7:00 am, 3:00 pm, and 10:00 pm. Patients repeated the dose of study medication if an episode of vomiting occurred within 30 minutes of ingesting any scheduled dose of medication. Patient compliance with the antiemetic schedule was assessed by a daily telephone call for 3 consecutive days beginning on study day 2, review of the patient's diary, and pill counts.

The study sponsor provided prochlorperazine tablets to investigators for use as “rescue” medication if patients had persistent nausea or vomiting. The approved Food and Drug Administration (FDA) package insert dose (10 mg, three times per day) was recommended. However, the rescue medication and regimen used was determined by the investigator. A patient could be withdrawn at any time; no minimum number of emetic episodes nor minimum degree of nausea needed to occur before a patient could be withdrawn. Study drug was discontinued if patients were rescued.

Evaluation of Response

Each patient was monitored by research personnel for the 3 days after chemotherapy. Patients were either seen in person or contacted by telephone. Patients were provided a diary for each study day and were instructed to record the time each study drug was taken, to record the time and total number of emetic episode(s), to record the assessment of nausea by a visual analog scale, and to record any adverse events. Emetic episodes and nausea were assessed independently.

The primary efficacy parameter measured was the number of emetic episodes. An emetic episode was defined as a single vomit (productive of stomach contents) or retch (not productive of stomach contents) or any number of continuous vomits or retches or both separated by less than 1 minute. If more than 1 minute elapsed between vomits or retches, this was counted as two emetic episodes. A complete response was defined as no emetic episodes and no use of rescue medication during the 3-day study period. A major response was 1 to 2 emetic episodes in the 3-day period, and a minor response was 3 to 5 emetic episodes. Patients who had more than five emetic episodes or who were withdrawn due to the severity of nausea or vomiting or adverse events were deemed treatment failures. The secondary efficacy parameter analyzed was the severity of nausea. Before treatment and in the evening of each study day, patients assessed their own nausea. A 100-mm visual analog scale was used. Zero mm was no nausea and 100 mm was nausea as bad as it could be.

Safety Assessments

An adverse event was defined as any untoward medical occurrence in a patient (except nausea or vomiting) or an exacerbation of a preexisting medical condition. The investigator and patient were instructed to record all adverse events regardless of severity or assessment of the relation to the study drug. Blood samples to monitor hematology and blood chemistry variables were obtained within 7 days before chemotherapy and within 1 week after chemotherapy.

Statistical Assessment

All statistical tests were pairwise comparisons of the respective ondansetron dose groups to placebo. No adjustments for multiple comparisons between ondansetron doses were planned or done. As per protocol, patients with severe emesis were given rescue medication before the end of the study. Therefore, the number of emetic episodes occurring during the entire study period was not available for these patients; hence, a statistical analysis using means was not appropriate. The number of emetic episodes was analyzed in two ways, using nonparametric procedures.

1. Complete Response. The Cochran-Mantel-Haenszel Test (stratified by study center) was used to compare each ondansetron dose group to placebo with respect to the proportion of patients having a complete response (no emetic episodes, no use of rescue medication).

2. Number of Emetic Episodes. The van Elteren test [15, 16] (Wilcoxon rank-sum test stratified by study center) was used to compare each dose group to the placebo group.

We compared the ondansetron dose groups to placebo with respect to nausea severity using the van Elteren test (Wilcoxon rank-sum test stratified by study center).

Results

A total of 349 cancer patients from 27 centers receiving their first cycle of chemotherapy were enrolled in this trial. Intravenous cyclophosphamide was given to 346 patients at a dose ≥ 450 mg/m². All patients received other chemotherapeutic agents (Table 1). Three patients did not receive intravenous cyclophosphamide. Ninety-two patients received placebo, 86 patients received 1 mg of ondansetron, 89 received 4 mg, and 82 received 8 mg. Seventy-one percent of patients had breast cancer, 14% had lung cancer, and 10% had leukemia or lymphoma. The patients ranged in age from 26 to 87 years (see Table 1). No significant differences [P > 0.05] existed between treatment groups with regard to age, sex, smoking status, or alcohol consumption. The mean weight of the 8-mg treatment groups was less than the placebo group (P = 0.018).

All 349 patients who received study medications were included in demographic and safety evaluations. Efficacy analyses were done on both the intent-to-treat group (n = 349) and the efficacy subgroup (for example, the 318 patients who had no protocol violations). All inclusion or exclusion decisions were made without knowledge of treatment group assignments. Thirty-one (8.9%) patients were excluded from efficacy analyses due to protocol violations. The most common reason for exclusion was improper dose of chemotherapeutic agents (10 patients). Patients who received more than three doses of study drug or fewer than two doses of study drug on any study day were deemed noncompliant and excluded (nine patients). Other reasons for exclusion included concomitant antiemetic therapy (n = 6), previous chemotherapy (n = 4), uncontrolled vomiting at study entry due to an organic cause (n = 1), and early withdrawal due to administrative error (n = 1). The exclusions were evenly distributed throughout the treatment groups and strata. The efficacy data are presented here for the 318 assessable patients. The intent-to-treat analysis was done on all 349 patients enrolled in the study. Both analyses resulted in a similar outcome; all results significantly different (P < 0.05) from placebo in the efficacy analysis were also significant in the intent-to-treat analysis, and the resulting P values were similar. Therefore, the elimination of patients with protocol violations had no effect on the overall response rates for each treatment group and no effect on the conclusions drawn from this study.

During the 72-hour period after chemotherapy, oral ondansetron was superior to placebo in preventing nausea and vomiting. In the efficacy subgroup, the complete response rate (no emetic episodes during the 3-day study period) was greater (P < 0.001) with each ondansetron dose compared with placebo (Figure 1). A complete response was observed in 15 (19%), 44 (57%), 53 (65%), and 52 (66%) of patients in the placebo, 1-mg, 4-mg, and 8-mg ondansetron groups, respectively. The number of patients considered treatment failures was also greater (P ≤ 0.001) for the placebo group compared with each of the ondansetron treatment groups. Treatment failures were reported for 51 (63%), 20 (26%), 19 (23%), and 13 (16%) patients in the placebo, 1-mg, 4-mg, and 8-mg groups, respectively. In addition, the number of emetic episodes reported by patients receiving placebo was greater (P ≤ 0.001) compared with ondansetron-treated groups (Table 2). The largest percentage of patients assigned to placebo who were withdrawn due to nausea and vomiting were withdrawn on study day 1 (Table 3). For the ondansetron treatment groups, the largest number withdrawn due to nausea or vomiting occurred on day 2. Five study participants withdrawn because of nausea or vomiting were hospitalized and received intravenous fluids as well as additional antiemetics (two patients from the 1-mg ondansetron treatment group and three from the placebo group); all five patients recovered without sequelae.

In the efficacy subgroup, 174 patients received ≥ 35 mg/m² of doxorubicin in addition to cyclophosphamide, and 108 patients received ≥ 600 mg/m² of cyclophosphamide (Figures 2 and 3). In all three treatment groups (Figures 1, 2, and 3), oral ondansetron (1, 4, or 8 mg) was more efficacious than placebo in preventing emesis. No statistical comparison was done among the ondansetron groups.

Nausea scores were recorded on a 100-mm visual analog scale (see Methods). A score of zero mm was no nausea, and a score of 100 mm was nausea as bad as it could be. Nausea scores across all study days for each ondansetron treatment group were lower than placebo (Figure 4). Adverse effects due to ondansetron were mild and infrequent. Headache and constipation were the most common side effects reported (Table 4); no clear dose-response relationship was evident for either effect. No extrapyramidal reactions were reported. Laboratory variables including hemoglobin, hematocrit, leukocyte count, neutrophils, bands, monocytes, platelets, BUN, creatinine, total bilirubin, alkaline phosphatase, lactate dehydrogenase, AST, and ALT for each ondansetron group were compared with the placebo group. The only difference noted was an increase in the proportion of patients in the 1-mg ondansetron group with a clinically significant increase in granulocytes.

Discussion

Our results show that oral ondansetron (1, 4, or 8 mg taken three times per day for 3 consecutive days) is effective and safe in preventing cyclophosphamide-induced nausea and vomiting. Greater protection from nausea and vomiting was observed in all three ondansetron treatment groups compared with the placebo group (see Figures 1, 2, 3). More than 70% of the patients treated with cyclophosphamide plus methotrexate and randomized to ondansetron had no vomiting during the 3-day study period. Additionally, 53% of patients receiving cyclophosphamide plus doxorubicin had no emetic episodes. Only 26% and 13% of the nondoxorubicin and doxorubicin patients, respectively, in the placebo-treated group had no vomiting during the study period.

To date, no oral compound has been approved by the FDA as efficacious in the prevention of emesis associated with cyclophosphamide or doxorubicin. Additionally, few studies evaluating the effectiveness of oral phenothiazines, steroids, and cannabinoid compounds have used homogenous chemotherapy populations, a double-blind design, or conventional efficacy end points, thus limiting the value of the results [17-19]. Some antiemetic studies have shown a significant placebo response, with 22% to 47% of placebo-treated patients reporting no emetic episodes [3, 20, 21]. Based on this lack of an FDA-approved oral agent and lack of valid data from clinical trials, a placebo arm was thought to be required in the current trial to validate the study model and to evaluate fully the antiemetic efficacy of ondansetron in this group of patients. The study sponsor, fully appreciating the ethical concerns of a placebo arm in such a trial, provided approved package-insert doses of oral prochlorperazine to patients for use as needed. Additionally, the protocol did not restrict the withdrawal of any patient from the study, thereby allowing investigators to withdraw patients at any time.

Most patients participating in our trial were treated as outpatients. The safety of any pharmacologic entity in the outpatient setting and in patients with a systemic illness is of obvious importance. Few side effects attributable to ondansetron were reported in this study. Extrapyramidal reactions were not observed; no sedation, sleepiness, or diarrhea developed. Mild headache and constipation were the most frequently reported toxicities. Patient compliance with the three-times-per-day schedule was very good; only nine (2.6%) patients deviated from the planned course of antiemetic therapy.

Nausea and vomiting secondary to the administration of chemotherapy has been classified as acute, delayed, and anticipatory [22-24]. Control of nausea and vomiting is particularly important during the initial cycle of chemotherapy as the development of anticipatory vomiting occurs more frequently in patients with poor control during previous chemotherapy treatments. It is probably more effective to prevent anticipatory emesis than to try to treat this conditioned response. Our study shows that oral ondansetron is able to prevent vomiting in most patients having their first cycle of cyclophosphamide-based chemotherapy.

Current evidence [25, 26] suggests that serotonin is involved in the mechanism by which chemotherapy induces nausea and vomiting. Antagonism of 5-hydroxytryptamine subtype-3 receptors has been shown to reduce the incidence of chemotherapy- and of radiation therapy-induced nausea and vomiting [27, 28]. In the ferret model, cisplatin-induced emesis is completely controlled if serotonin stores are depleted by inhibition of serotonin synthesis [29]. Ondansetron is a highly selective 5-hydroxytryptamine subtype-3 receptor-antagonist that is effective in preventing cisplatin-induced nausea and vomiting when administered intravenously [27, 30-32]. Kaasa and colleagues [33] have shown intravenous ondansetron to be effective in the prevention of emesis associated with cyclophosphamide-based chemotherapies. Ondansetron is rapidly and completely absorbed when administered as a tablet. Absolute bioavailability averages 56% in young, healthy volunteers [34].

The efficacy of oral ondansetron had been suggested in an open-label, single-institution study [35]. This earlier trial evaluated the antiemetic efficacy of 1 mg, 4 mg, and 8 mg of oral ondansetron three times daily for 2 days in patients receiving cyclophosphamide and doxorubicin chemotherapy. These data showed ondansetron prevented emesis in 57%, 65%, and 66% of the patients who received 1 mg, 4 mg, and 8 mg of ondansetron, respectively. The results of this preliminary trial [35] and the current study show that oral ondansetron, when used as a single agent, is both safe and effective.

Appendix

The members of the Ondansetron Study Group included Neil Abramason, MD, Baptist Hospital, Jacksonville, Florida; Jane Alavi, MD, University of Pennsylvania, Philadelphia, Pennsylvania; Thomas M. Beck, MD, Mountain States Tumor Institute, Boise, Idaho; Alan Cartmell, MD, San Joaquin Medical Group; Alex Chang, MD, University of Rochester Cancer Center-The Genesee Hospital, Rochester, New York; Eric M. Chevlen, MD, Good Samaritan Medical Group, San Jose, California; Bruce Dana, MD, Oregon Health Sciences University, Portland, Oregon; Joseph DiBenedetto, MD, Oncology/Hematology Associates, Providence, Rhode Island; John Feldmann, MD, Mobile Infirmary Center, Mobile, Alabama; William Grace, MD, St. Vincent's Hospital, New York, New York; Walter H. Harvey, DO, University of Texas Medical Branch, Galveston, Texas; Stephen E. Jones, MD, Baylor University Medical Center, Dallas, Texas; Basil Kasimis, MD, Veterans Affairs Medical Center, East Orange, New Jersey; John F. Kessler, MD, David Grant United States Air Force Medical Center, Travis Air Force Base, California; J. Philip Kuebler, MD, Oklahoma University, Health Sciences Center, Oklahoma City, Oklahoma; Stefan Madajewicz, MD, Medical Oncology Division, Stony Brook, New York; Frank Meyskens, MD, Irvine Medical Center, Orange, California; Douglas Reding, MD, Marshfield Clinic, Marshfield, Wisconsin; Bahu Shaikh, MD, Toledo Clinic Oncology Department, Toledo, Ohio; Daniel Shevrin, MD, University of Illinois, Chicago, Illinois; Mary Simmonds, MD, Milton S. Hershey Medical Center, Hershey, Pennsylvania; Robert Sponzo, MD, Albany Regional Cancer Center, Albany, New York; James Stewart, MD, Vermont Regional Cancer Center, Burlington, Vermont; Ronald Stoller, MD, Montefiore Hospital, Pittsburgh, Pennsylvania; Stephen Strum, MD, Culver City, California; N. Simon Tchekmedyian, MD, Pacific Coast Hematology/Oncology Medical Group, Long Beach, California; and Stephen J. Tremont, MD, Rex Cancer Center, Raleigh, North Carolina.