Kovac, Anthony L. MD*; Eberhart, Leopold MD†; Kotarski, Jan MD‡; Clerici, Giuseppina MD§; Apfel, Christian MD, PhD∥; The Palonosetron 04-07 Study Group
Palonosetron is a novel 5-HT3 receptor antagonist with a greater binding affinity and longer half-life than older 5-HT3 antagonists. Recent receptor binding studies suggest that palonosetron is further differentiated from other 5-HT3 antagonists by interacting with 5-HT3 receptors in an allosteric, positively cooperative manner at sites different from those that bind with ondansetron and granisetron.1 In addition, this unique form of receptor interaction may be associated with long-lasting effects on receptor ligand binding and functional responses to serotonin. These pharmacologic differences, combined with clinical evidence that palonosetron provides better protection against chemotherapy-induced nausea and vomiting than the older 5-HT3 drugs (including ondansetron) throughout the 5-day postchemotherapy period prompted the study of palonosetron as an antinausea and antiemetic drug with potentially long-lasting efficacy against postoperative nausea and vomiting (PONV) in at-risk patients.2,3
This study was conducted to evaluate the safety and efficacy of IV palonosetron in the prevention of PONV for up to 3 days in women who were scheduled to be hospitalized for at least 72 h after gynecologic or breast surgery. The primary efficacy end-point was complete response (CR) defined as no emesis and no use of rescue medications during the 0–24 h and 24–72 h time intervals. An earlier dose-ranging study provided the rationale for the selection of the doses of palonosetron used in this clinical trial.4,5
After IRB approval and written informed consent, we enrolled patients at 28 centers in Europe: 13 in Germany, 10 in Poland, and 5 in the Czech Republic. Eligible patients were women ASA physical status 1–3 and at least 18 yr of age who were scheduled to undergo elective inpatient gynecological surgery (e.g., abdominal hysterectomy or vaginal hysterectomy) or breast surgery (except aesthetic/cosmetic plastic surgery). Surgery was expected to last a minimum of 1 h and patients were scheduled to be hospitalized for at least 72 h after surgery. In addition to being female, patients had at least one of the two following risk factors for PONV: 1) history of PONV and/or currently prone to motion sickness, and 2) nonsmoking status (never smoked or quit ≥12 mo ago). Excluded from the study were patients who had received cancer chemotherapy within 4 wk or emetogenic radiotherapy within 8 wk before study entry.
Patients were randomized to one of the three active treatment groups (0.025, 0.05, or 0.075 mg) or placebo. Since rescue medication for the relief of PONV was permitted, withholding active treatment was not considered detrimental to patients randomized to placebo. To achieve a homogeneous risk of PONV in each of the four study groups, treatment randomization was stratified by type of surgery, history of PONV, and smoking status. A dynamic adaptive stratification type of randomization method was used to balance the four treatment groups across the entire study and not within each individual site.
At each study site, palonosetron and placebo were prepared for administration to patients by an unblinded research pharmacist or designee who was otherwise not involved in the conduct of the study. A single IV dose of palonosetron (0.025 mg, 0.050 mg, or 0.075 mg), with an adequate volume of saline solution added to bring the total injectable volume to 2 mL, was administered as a 10-S IV bolus immediately (no more than 5 minutes) before induction of anesthesia. Subjects in the placebo group received a single IV dose of normal saline as a 2-mL bolus in a similar manner.
Premedication with midazolam or fentanyl was permitted. Induction was performed with propofol or barbiturates. Anesthetic maintenance consisted of any inhaled anesthetic combined with nitrous oxide (concentration 50%–70%). Neuromuscular blocking drugs were used at the discretion of the anesthesiologist with reversal by neostigmine (≤2.5 mg) and glycopyrrolate. Medication for the prevention of nausea and vomiting or any other medication with antiemetic properties (except propofol) was prohibited in the 24 h before the induction of anesthesia and throughout the entire study period. However, rescue medication for the treatment of nausea and vomiting after surgery, with the exception of palonosetron and droperidol, was permitted at the discretion of the investigator.
Baseline patient assessment for study eligibility took place within 1 wk before surgery and follow-up assessments began postoperatively when the patient woke up and was able to respond to verbal commands. The final study visit occurred between day 6 and 10 after surgery.
Emetic episodes and intensity of nausea (based on a four-point categorical scale of “none” to “severe”) and use of rescue therapy were measured at the time points of 2, 6, 24, 48, and 72 h in reference to symptom occurrence during the previous observation period. These outcomes were captured on patient-reported diary cards and were used to determine CR as defined by no emesis and no use of rescue medication.
CR was determined for the 0–24 h and 24–72 h postoperative time intervals as an efficacy end-point evaluated at two separate primary time points. Because inpatients become increasingly ambulatory after surgery, additional time intervals of 0–6 and 6–72 h were also evaluated to study the extent of emesis and nausea control during a time when patients had minimal ambulation (including time in the postanesthesia unit: 0–6 h) and also a time interval when they were more mobile (after discharge to the hospital floor: 6–72 h).
The primary efficacy hypotheses were that at least one dose of palonosetron was superior to placebo, for the CR rate, in the 0–24 h and 24–72 h time periods. To account for multiple comparisons of treatments (placebo versus 0.025 mg, 0.050 mg, and 0.075 mg), the multiple type-I error level was guaranteed by the Holm-Bonferroni method. To claim a significant difference, the smallest of the 3 2-sided P values could not exceed 0.017 (0.05/3). If this was achieved, the second smallest P value could not exceed 0.025 (0.05/2) to be significant, and if this was also achieved, the significance threshold for the third P value was 0.05. This sequential procedure was to be stopped if the respective threshold was exceeded. The procedure guarantees the multiple type I error level of 0.05. All differences in secondary time intervals compared with placebo were considered significant if the P value was <0.05.
The sample size was calculated based on a responder rate of 55% in the palonosetron group and 35% in the placebo group. For a two-sided test of difference, using α = 0.017 (obtained as error type I/number of comparisons = 0.05/3) and β = 0.2 for each comparison, the sample size was estimated at 131 evaluable patients per group, rounded up to 136 patients per group.
Multiple logistic regression adjusted for stratification criteria and risk factors was performed for the primary and secondary end-points of CR and incidence of emesis, where each dose of palonosetron was compared with placebo. The incidence of nausea was analyzed using the χ2 test for overall and pairwise comparisons. The distribution of nausea severity was analyzed using Cochran-Mantel-Haenszel test with adjustment for stratification criteria. Time to emesis and treatment failure was analyzed using the log-rank test. All secondary end-points were tested at a significance level of 0.05. The incidence rates for adverse events were summarized; formal statistical comparisons were not performed.
After the first 130 patients were enrolled in this study, a potential unblinding event with the clinical supplies label was discovered by a representative of the sponsor. If the labels were closely examined, there was a remote possibility that unblinding information of palonosetron versus placebo could have been revealed to the pharmacist. However, neither patient nor investigator was likely unblinded and the dose of palonosetron could not have been revealed. The decision was made to exclude these 130 patients and enroll an additional 130 patients to replace them. For this reason, the first 130 patients were included in the safety analysis (n = 673) but excluded from the efficacy analysis (n = 544).
Patient Enrollment and Disposition
The efficacy analysis population consisted of patients receiving placebo (n = 136), palonosetron 0.025 mg (n = 136), palonosetron 0.050 mg (n = 137), and palonosetron 0.075 mg (n = 135). Age, height, weight, and PONV risk factors had no statistically significant differences for patients across all treatment groups. Additional baseline characteristics for all treatment arms are shown in Table 1.
Approximately 45% of the patients in this study received postoperative opioids, and there were no significant differences among the various treatment groups (placebo = 43%, palonosetron 0.025 mg, palonosetron 0.050 mg = 46%, and palonosetron 0.075 mg = 46%).
Compared with placebo, a trend in the proportion of patients with a CR was observed with increasing doses of palonosetron across the 72-h study interval. In the 0–24 h postoperative time interval, CR rates were placebo, 36%; palonosetron 0.025 mg, 46% (P = 0.073); palonosetron 0.050 mg, 47% (P = 0.069); and palonosetron 0.075 mg, 56% (P = 0.001) (Fig. 1). For the 24–72 h interval after surgery, CR rates were placebo, 52%; palonosetron 0.025 mg, 56% (P = 0.511); palonosetron 0.050 mg, 61% (P = 0.151); and palonosetron 0.075 mg, 70% (P = 0.002). Of the three doses of palonosetron used in this study, only the 0.075-mg dose was consistently statistically superior to placebo during the secondary time intervals evaluated. During the 0–6 h, 6–72 h, and overall 0–72 h time intervals, patients who received palonosetron 0.075 mg had significantly higher CR rates compared with placebo (Fig. 1).
Time to Treatment Failure and Use of Rescue Therapy
Time to treatment failure was defined as the time to the first emetic episode or time to first administration of rescue medication, whichever occurred first. A significantly longer time to treatment failure was observed in patients who received palonosetron 0.075 mg versus placebo (P = 0.004) (Fig. 2). There was no significant difference in the time to treatment failure between placebo and the palonosetron 0.025 mg (P = 0.112) or palonosetron 0.050 mg (P = 0.060) groups. Over the entire 0–72 h time course of observation, 62 of the 136 patients (46%) who received placebo required rescue therapy, compared with 36 of the 135 patients (27%) (P < 0.001) who received palonosetron 0.075 mg.
By the end of the 0–24 h interval, 82 of 136 patients (60%) in the placebo group had experienced emesis, compared with only 54 of 135 patients (40%) in the palonosetron 0.075-mg group (P < 0.001) (Fig. 3). This corresponds to a relative risk reduction of 34% for patients treated with palonosetron 0.075 mg.6 Although there were significantly fewer patients in the palonosetron 0.075 mg group who experienced emesis (versus placebo) during the 0–6 h (37% vs 54%) (P = 0.006) and 0–72 h (44% vs 60%) (P = 0.006) intervals, there were no significant differences between placebo and palonosetron 0.075 mg during the 24–72 h (10% vs 4%) (P = 0.061) or 6–72 h (23% vs 29%) (P = 0.200) intervals. Palonosetron 0.075 mg and 0.050 mg were associated with a significantly longer median time to first emesis (>72 h) compared with placebo (3.9 h) (P = 0.002 and P = 0.014). Palonosetron 0.025 mg was not associated with a significantly longer median time to first emesis compared with placebo.
On the day of surgery (0–24 h), significantly fewer patients treated with palonosetron 0.075 mg (50%) experienced nausea compared with patients who received placebo (71%) (P < 0.001) (Fig. 3). This corresponds to a relative risk reduction of 31% for patients treated with palonosetron 0.075 mg. There was also a significantly less intense distribution of nausea over a four-point categorical scale (“none” to “severe”) in the group of patients who received palonosetron 0.075 mg compared with patients who received placebo during the 0–24 h interval (P < 0.001) and also during the 0–6 h (P < 0.001), 6–72 h (P = 0.011), and 0–72 h (P < 0.001) intervals (Fig. 4). Significant differences in the distribution of nausea intensity were reported for both the 0.025 mg and 0.050 mg palonosetron doses during the 0–24 h interval (P = 0.036 and P = 0.004). During the 24–72 h interval, a significant difference was reported between palonosetron 0.050 mg (P = 0.049) and placebo but no significant difference was reported between palonosetron 0.075 mg or 0.025 mg and placebo.
In the safety population (n = 673), all doses of palonosetron were well tolerated in this study, most adverse events were judged to be mild or moderate in intensity, and no patients were withdrawn from the study due to adverse events. In the placebo group, 5% of patients experienced severe adverse events, compared with 4% in the palonosetron 0.075-mg group and 7% in both the palonosetron 0.025-mg and 0.050-mg groups. The proportion of patients with treatment-related adverse events was similar among all treatment groups, including placebo.
Among the treatment-related adverse events were headache and constipation, consistent with this class of drugs. The palonosetron study group rates were not statistically significant from placebo. Treatment-related headache rates were placebo, 2%; palonosetron 0.025 mg, 2%; palonosetron 0.050 mg, 2%; and palonosetron 0.075 mg, 1%. Treatment-related constipation rates were placebo, 2%; palonosetron 0.025 mg, 2%; palonosetron 0.050 mg, 4%; and palonosetron 0.075 mg, 2%. There was no prolongation of the QTc interval observed in any palonosetron study group compared to placebo.
Despite the frequent use of the wide variety of prophylactic antiemetics currently available for PONV, including older 5-HT3 receptor antagonists, many patients may still experience nausea and vomiting either in the PACU or after discharge from the PACU, hospital floor or home.7–11 Most published, randomized, clinical trials have reported results for the 0–2 h (PACU) and 0–24 h discharge intervals. The inclusion of longer postoperative time intervals in this study reflects a more recent trend to further evaluate the time course of PONV. The time interval of 0–6 h was chosen to reflect the period during which the patient was less mobile (i.e., including time spent in the PACU), versus 6–72 h when patients were increasingly mobile (i.e., after discharge to hospital floor).
This randomized, double-blind, multicenter, stratified phase III study evaluated the dose-response efficacy of three different single IV doses of palonosetron compared with placebo for the prevention of PONV in an inpatient setting. An earlier dose-ranging study that evaluated effects of IV palonosetron out to 24 h provided the rationale for the selection of the doses used in this study.4,5
A previous weight-based (μg/kg) study of palonosetron (RS-25259) for preventing PONV was conducted on 218 hysterectomy patients by Tang et al.4 Evaluating palonosetron doses of 0.1, 0.3, 1.0, 3.0, and 30 μg/kg, these researchers determined that the 30 μg/kg dose was most effective. However, the Tang et al.4 study was a subset of a study reported by White and Scuderi5 who determined that palonosetron 1.0 μg/kg was the lowest effective dose and displayed a risk:benefit ratio superior to that of the two higher doses. The 1 mcg/kg dose was comparable to the highest dose studied in the current trial. The 0.075 mg was chosen as the highest dose, with the lower 0.025 mg and 0.050 mg doses included to test for the presence of a dose response. The absolute difference between the palonosetron 0.075 mg dose and placebo for this trial (20% difference) was similar to the absolute difference observed between palonosetron 1.0 μg/kg IV and placebo (25% difference) in an earlier phase II dose-ranging trial.5
In this study, the highest palonosetron dose (0.075 mg) was found to have the best treatment effect compared with the lower doses evaluated. The palonosetron 0.075 mg dose was statistically significant compared to placebo for most end-points, including CR rate and reduction in nausea severity through the 72-h study evaluation period. The palonosetron 0.075 mg dose was statistically superior to placebo for all end-points during the first 24 h, including CR, emesis, and nausea rates, and a reduction in nausea severity. In addition, palonosetron 0.075 mg was associated with a significantly longer median time to first emesis and a significantly longer time to treatment failure. Similar results were seen for the 0–72 h interval, although there was no significant difference between palonosetron 0.075 mg and placebo in the emesis rates or nausea severity during the 24–72 h time interval (i.e., the efficacy of palonosetron appears to be mainly in the first 24 h).
In addition to emesis, this trial was designed to evaluate palonosetron's effect on nausea. Instead of measuring only “none” or “no more than mild” nausea, all four levels of a categorical scale of severity (none, mild, moderate, severe) were assessed versus placebo. Various time intervals were included to evaluate PONV up to 72 h postoperatively. Compared with placebo, palonosetron 0.075 mg was associated with an overall downward shift toward less severe nausea (e.g., toward “mild” to “none”) in the 0–6, 6–72, and 0–72 h time intervals (Fig. 4).
The use of opioids in the surgical setting is associated with the development of late postoperative emesis.7,8 An interesting finding from this study was that only 45% of the patients in this study received postoperative opioids. The use of postoperative opioids is highly variable and is dependent on cultural background. In Apfel's risk score for predicting PONV, 81% of patients in the Finnish population received postoperative opioids compared to 10% of German patients.8 The type of surgery performed also greatly affects the rate of opioid use. In the Netherlands, for example, approximately 40% of inpatients versus only 6%–8% of outpatients received postoperative opioids.12 The high percentage of patients in the placebo group who did not achieve a CR during the 6–72 h interval (57%), suggests that opioid use is not the only contributing factor to the presence of postoperative symptoms of nausea and vomiting, contrary to the perception that opioids may be the predominant cause of emetic symptoms during this later time period.
In contrast to the older 5-HT3 receptor antagonists which may have an effect and prolongation of the QTc interval,13–15 this was not observed with any of the palonosetron doses.
A possible limitation of this study was the modest population size for a dose-response trial, with perhaps too few patients in each treatment arm for a robust evaluation of the treatment effect on all end-points and all time intervals for the purposes of statistical analyses. Although this study was placebo controlled, the lack of an active comparator limits the ability to directly compare the results for palonosetron with older 5-HT3 receptor antagonists. In addition, end-points were measured during preset time intervals (e.g., 0–6 h and 6–72 h), rather than evaluated by occurrence in a particular setting such as in the PACU (0–2 h) or post-PACU (2–24 h) time intervals. Patient satisfaction was not included as an end-point in this trial.
In the inpatient surgical setting, a single 0.075-mg IV dose of palonosetron significantly reduced emesis, intensity of nausea and the use of rescue antiemetics in addition to delaying the time to emesis and treatment failure, particularly during the first 24 h after surgery. The lower 0.025 mg and 0.050 mg doses of palonosetron tended not to be significantly different from placebo. Palonosetron 0.075 mg is now the Food and Drug Administration-approved dose for prevention of PONV up to 24 h after surgery.
The authors thank Rebecca Gerber, Susan Weidner, Jennifer Vanden Burgt, Don Tessier, and Tim DeGroot for their help in preparation and review of the manuscript.
1. Rojas C, Stathis M, Thomas A, Massuda E, Alt J, Zhang J, Rubenstein E, Sebastiani S, Cantoreggi S, Snyder SH, Slusher B. Palonosetron exhibits unique molecular interactions with the 5-HT3 receptor. Anesth Analg 2008;107:469–78
2. Gralla R, Lichinitser M, Van der Vegt S, Sleeboom H, Mezger J, Peschel C, Tonini G, Labianca R, Macciocchi A, Aapro M. Palonosetron improves prevention of chemotherapy-induced nausea and vomiting following moderately emetogenic chemotherapy: results of a double-blind randomized phase III trial comparing single doses of palonosetron with ondansetron. Ann Oncol 2003;14:1570–7
3. Eisenberg P, Figueroa-Vadillo J, Zamora R, Charu V, Hajdenberg J, Cartmell A, Macciocchi A, Grunberg S; 99–04 Palonosetron Study Group. Improved prevention of moderately emetogenic chemotherapy-induced nausea and vomiting with palonosetron, a pharmacologically novel 5-HT3 receptor antagonist. Cancer 2003;98:2473–82
4. Tang J, D'Angelo R, White PF, Scuderi PE. The efficacy of RS-25259, a long-acting selective 5-HT3 receptor antagonist, for preventing postoperative nausea and vomiting after hysterectomy procedures. Anesth Analg 1998;87:462–7
5. White PF, Scuderi PE. Prevention of postoperative nausea and vomiting (PONV): a dose-ranging study involving palonosetron, a potent 5-HT3 receptor antagonist. Anesthesiology 2005;103:A703
6. Barrat A, Wyer PC, Hatala R, McGinn T, Dans AL, Keitz S, Moyer V, For GG. Evidence-Based Med Teaching Tips Working Group. Tips for learners of evidence-base medicine: 1. Relative risk reduction, absolute risk reduction and number needed to treat. Can Med Assoc J 2004;171:353–8
7. Apfel CC, Korttila K, Abdalia M. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004;350:2441–51
8. Apfel CC, Läärä E, Koivuranta M, Greim C-A, Roewer N. A simplified risk score for predicting postoperative nausea and vomiting. Anesthesiology 1999;91:693–700
9. Hirsch J. Impact of postoperative nausea and vomiting in the surgical setting. Anaesthesia 1994;49:30–3
10. Gupta A, Wu CL, Elkassabany N, Krug CE, Parker SD, Fleisher LA. Does the routine prophylactic use of antiemetics affect the incidence of postdischarge nausea and vomiting following ambulatory surgery? Anesthesiology 2003;99:488–95
11. Parlow JL, Meikle AT, van Vlymen J, Avery N. Post discharge nausea and vomiting after ambulatory laparoscopy is not reduced by promethazine prophylaxis. Can J Anaesth 1999;46:719–24
12. Visser K, Hassink EA, Bonsel GJ, Moen J, Kalkman CJ. Randomized controlled trial of total intravenous anesthesia with propofol versus inhalation anesthesia with isoflurane-nitrous oxide. Anesthesiology 2001;95:616–26
13. Zofran® (ondansetron hydrochloride injection) prescribing information. Research Triangle Park, NC: GlaxoSmithKline, 2006
14. Anzemet® (dolasetron mesylate injection) prescribing information. Bridgewater, NJ: Aventis Pharmaceuticals, 2003
15. Kytril® (granisetron hydrochloride injection) prescribing information. Nutley, NJ: Roche Laboratories, 2002