Support Care Cancer (2014) 22:1685–1697 DOI 10.1007/s00520-014-2175-6
REVIEW ARTICLE
Efficacy and safety of palonosetron for the prophylaxis of chemotherapy-induced nausea and vomiting (CINV): a systematic review and meta-analysis of randomized controlled trials Marko Popovic & David G. Warr & Carlo DeAngelis & May Tsao & Kelvin K. W. Chan & Michael Poon & Cheryl Yip & Natalie Pulenzas & Henry Lam & Liying Zhang & Edward Chow
Received: 1 October 2013 / Accepted: 17 February 2014 / Published online: 4 March 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Palonosetron, a 5-hydroxytryptamine 3 receptor antagonist (5-HT3RA) with a strong binding affinity and long half-life, has been used in numerous trials for the prophylaxis of chemotherapy-induced nausea and vomiting (CINV). We systematically reviewed the efficacy and safety of palonosetron compared to other 5-HT 3 RAs in CINV prophylaxis. Methods A literature search of Ovid MEDLINE, EMBASE, and CENTRAL was conducted to identify randomized controlled trials (RCTs) comparing palonosetron to other 5HT3RAs in CINV prophylaxis. Primary endpoints were the percentage of patients achieving a complete response M. Popovic : M. Tsao : M. Poon : C. Yip : N. Pulenzas : H. Lam : L. Zhang : E. Chow (*) Rapid Response Radiotherapy Program, Department of Radiation Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada e-mail: [email protected] D. G. Warr Department of Medical Oncology, Princess Margaret Cancer Centre, University of Toronto, 610 University Avenue, Toronto, Ontario M5T 2M9, Canada C. DeAngelis Department of Pharmacy, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada K. K. W. Chan Department of Medical Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada
(CR), complete control (CC), no emesis, no nausea, or taking no rescue medications. Secondary endpoints were the percentage of patients suffering from 5-HT3RA-related adverse events. Results Sixteen RCTs were identified with 2,896 patients randomized to palonosetron and 3,187 patients randomized to other 5-HT3RAs. Palonosetron was consistently statistically superior in CR, CC, no emesis, or no nausea and was sometimes superior in no rescue medication. Subgroup analyses demonstrated similarity in efficacy between highly and moderately emetogenic chemotherapy cohorts. In the acute phase, statistical superiority of palonosetron was found for trials that did not allow dexamethasone; conversely, RCTs that administered dexamethasone to all patients were nonsignificant. Palonosetron was statistically significantly safer in dizziness and mean QTc interval change and similar in constipation, headache, and diarrhea. Clinical superiority of palonosetron was reached in 3 of 19 analyzed efficacy and safety endpoints. Conclusions Palonosetron is safer and more efficacious than other 5-HT 3 RAs. Future antiemetic guidelines should discuss the merits of including palonosetron as a first-line treatment. Keywords Palonosetron . 5-HT3 receptor antagonist . Chemotherapy-induced nausea and vomiting . Efficacy . Safety
Introduction Even though great progress has been made in the control of chemotherapy-induced nausea and vomiting (CINV), more
1686
Support Care Cancer (2014) 22:1685–1697
research is still needed in the delayed management and highrisk settings [1]. CINV may have the potential to impair a patient’s quality of life and disrupt the preferred treatment plan [2, 3]. As such, evaluating the safest and efficacious antiemetic regimen remains paramount to ameliorating the outcomes of future patients. Serotonin from the enterochromaffin cells in the small intestine binds to 5-hydroxytryptamine 3 (5-HT3) receptors on vagal afferents and thus results in the pathogenesis of CINV [4]. The end of the twentieth century saw the introduction of several 5-HT3 receptor antagonists (5-HT3RAs), including ondansetron, granisetron, tropisetron, and dolasetron. Trials that used these agents concluded that they were able to successfully control rates of vomiting; however, patientreported nausea, especially in the delayed phase, still remained a significant concern [3]. In an attempt to improve response rates, another 5-HT 3 RA named palonosetron (Aloxi®) was developed. Palonosetron is characterized by a markedly longer plasma elimination half-life of about 40 h (half-life of ondansetron 5.7 h) and a high selective binding affinity to the 5-HT3 receptor [4]. Several randomized controlled trials (RCTs) comparing palonosetron to other 5-HT3RAs concluded that a single dose of palonosetron was noninferior and sometimes superior in preventing CINV [4–9]. In antiemetic clinical trials, five endpoints are commonly used:
analysis is to comprehensively investigate the efficacy and safety of palonosetron in relation to other 5HT3RAs in the prophylaxis of CINV.
1. Complete response (CR; defined as no emetic episode and no use of rescue medication) 2. Complete control (CC; defined as no emetic episode, no use of rescue medication, and no more than mild nausea) 3. No emetic episodes 4. No episodes of nausea 5. No use of rescue medication
Endpoints
Despite the variance in endpoints, there have been three previous meta-analyses of RCTs that aimed to compare the efficacy and safety of palonosetron in relation to other 5HT3RAs [10–12]. Recent years have seen a more widespread use of palonosetron, which has led to a greater amount of fruitful research. As such, there is much more data available today than reported in the last meta-analysis, which completed data collection over 2 years ago [12]. As well, past analyses often had confined and incomprehensive endpoints. For efficacy, only CR and sometimes CC were previously analyzed. With respect to safety, only certain meta-analyses considered all of the most common 5-HT3RA-related adverse events— dizziness, diarrhea, constipation, and headaches—and none investigated differences in QTc prolongation. In view of these uncertainties, the purpose of this meta-
Patients and methods Data collection A literature search was conducted using the following databases: Ovid MEDLINE and OLDMEDLINE (1946–June Week 1 2013), Ovid EMBASE Classic and EMBASE (1947–2013 Week 26), and the Cochrane Central Register of Controlled Trials (CENTRAL) (May 2013). Reference lists of included RCTs and past meta-analyses were additionally searched. The following terms were employed to elicit relevant literature: “palonosetron,” “antineoplastic agents,” “neoplastic,” “chemotherapy,” “chemoradiotherapy,” “malignant,” “neoplasms,” “cancer,” “cancer chemotherapy,” “oncology,” “medical oncology,” and “drug therapy.” RCTs were included if they aimed to compare palonosetron to other 5-HT3RAs in the prophylaxis of CINV in at least one of the following five endpoints: CR, CC, no nausea, no emesis, or no rescue medication. Studies were excluded if they contained repeat data.
The primary endpoints were the percentage of patients achieving either a CR, CC, no emesis, no nausea, or taking no rescue medications within the acute, delayed, or overall phases. Secondary endpoints included the percentage of patients suffering from 5-HT3RA-related constipation, dizziness, diarrhea, or headache, as well as the mean change in the QTc interval. Each included study must have reported distinct acute, delayed, and overall phases with accompanying time frames; included trials were grouped into these phases based on what the authors reported and irrespective of time frame. Subgroup analysis For our primary endpoints, we stratified trials based on how the authors’ reported their chemotherapy emetogenicities. We sorted trials into either moderately emetogenic chemotherapy (MEC) or highly emetogenic chemotherapy (HEC) groups. In a second subgroup analysis, we sorted studies based on whether all or no patients received adjuvant dexamethasone on the day of palonosetron and chemotherapy administration.
Support Care Cancer (2014) 22:1685–1697
Sensitivity analysis A sensitivity analysis was conducted for cohorts with reasonably different clinical characteristics or for studies that were performed with reasonably different designs. Statistical analysis Statistical analysis was performed using Review Manager (RevMan 5.2) by Cochrane IMS. The Mantel-Haenszel method was applied and a fixed effect analysis model was used to generate odds ratio (OR) effect measures, absolute risk differences (RD), and accompanying 95 % confidence intervals (CI). A p value less than 0.05 was considered statistically significant in the test for overall effect, whereas a heterogeneity test p>0.05 was considered suitable. Absolute risk differences were compared to the 2010 Multinational Association of Supportive Care in Cancer/European Society of Medical Oncology (MASCC/ESMO) antiemetic guidelines, which state that “as a general rule, the panel considered changes of >10 % to be sufficient to warrant changing a guideline, given that the evidence supported this magnitude of benefit” [13]. The mean interval QTc change was computed and analyzed through a weighted analysis of variance test. PROC GLM was performed for the unbalanced data, and the total number of patients from each study was considered a weighted variable. ∑wx The weighted mean was defined as xw ¼ ∑i wi i , where wi is i i the weight for the ith study and xi is the ith variable value. This analysis was performed using Statistical Analysis of Software (SAS version 9.3 for Windows, Cary, NC). The intention-to-treat principle was utilized in all statistical analyses.
Results Upon literature search of 822 articles, 16 eligible RCTs were identified (Appendix 1) [3–9, 14–22]. Of note, an abstract [23] included in a previous meta-analysis [12] was excluded due to no specified randomization procedure. Of the 16 included RCTs, there were a total of 37 study arms, 2,896 patients randomized to receive palonosetron, and 3,187 patients randomized to any other 5-HT3RA. Characteristics of included studies Clinical and sociodemographic features of the included RCTs can be found in Table 1. Nine trials contained
1687
predominantly female patients [3–7, 9, 19, 21, 22], while three RCTs had mostly male patients [8, 14, 17]. The range of mean ages was 50.3 to 58.4 years for 15 of the 16 included RCTs; the only exception was a trial that enrolled only patients younger than 17 years of age [8]. Three RCTs included only patients receiving MEC [5, 6, 20], whereas eight studies included solely HEC patients [4, 7, 8, 14, 15, 17, 19, 21]. Corticosteroids were administered to all, some, or no patients in five [3, 4, 9, 20, 21], two [6, 7], and three [5, 14, 18] studies, respectively. In six studies, it was unknown whether corticosteroids were administered [8, 15–17, 19, 22]. Only four studies [3, 4, 18, 21] discussed the use of aprepitant as part of protocol; of these, three trials [3, 4, 18] did not allow aprepitant, while one trial [21] administered aprepitant to all patients, making any potential subgroup analysis unfeasible. Efficacy—acute phase Odds ratios demonstrated statistical superiority of palonosetron in four of five endpoints: complete response (OR=1.32, 95 % CI 1.15–1.52, Fig. 1a), complete control (OR=1.33, 95 % CI 1.12–1.58, Fig. 1b), no emesis (OR= 1.28, 95 % CI 1.10–1.50, Fig. 1c), and no nausea (OR=1.21, 95 % CI 1.01–1.45, Fig. 1d). The fifth endpoint, no rescue medication, was statistically nonsignificant (Fig. 1e). RD computations for all five endpoints in the acute phase fell short of the MASCC/ESMO threshold of >10 % (range 4– 6 %) (Table 2). Upon subgroup emetogenicity analysis, palonosetron demonstrated superiority in complete response (HEC: OR=1.26, 95 % CI 1.04–1.52; MEC: OR=1.43, 95 % CI 1.10–1.86), complete control for HEC cohorts (OR=1.23, 95 % CI 1.01– 1.50), no emesis for MEC cohorts (OR=1.36, 95 % CI 1.06– 1.74), and MEC cohorts without nausea (OR=1.56, 95 % CI 1.11–2.18). Corticosteroid subgroup analysis revealed that CR in cohorts not receiving adjuvant dexamethasone (OR=1.52, 95 % CI 1.15–2.02) and CC in cohorts receiving adjuvant dexamethasone (OR=1.30, 95 % CI 1.03–1.63) were the only significant parameters. Efficacy—delayed phase Palonosetron was statistically superior to other 5HT3RAs in four of five endpoints: complete response (OR=1.63, 95 % CI 1.42–1.87, Fig. 2a), complete control (OR=1.60, 95 % CI 1.39–1.84, Fig. 2b), no emesis (OR=1.57, 95 % CI 1.36–1.82, Fig. 2c), and no nausea (OR =1.39, 95 % CI 1.14–1.69, Fig. 2d). Again, no rescue medication was statistically nonsignificant
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg (1) Palonosetron 0.25 mg (2) Granisetron 3 mg
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Blind, parallel
Parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Blind, crossover
Nonparallel
Double-blind, noninferiority, crossover
Parallel
Crossover
Parallel
Double-blind, noninferiority, parallel
Saito et al. [4]
Gralla et al. [5]
Eisenberg et al. [6]
Aapro et al. [7]
Sepulveda-Vildosola et al. [8]
Ghosh et al. [9]
Yu et al. [14]
Chen et al. [15]
Li et al. [16]
Dong et al. [17]
Tian et al. [18]
Mattiuzzi et al. [19]
Kaushal et al. [20]
Wenzell et al. [21]
Huang et al. [22]
Corticosteroids administered to all patients
HEC
HEC MEC and HEC
(1) 223 (2) 223 (3) 221 (1) 50 (2) 50 (1) 157 (2) 616 (3) 440
(1) 117 (2) 119
(1) 20 (2) 20
(1) 30 (2) 30
(1) 48 (2) 48 (3) 47
(1) 72 (2) 72
(1) 44 (2) 45
(1) 104 (2) 104
(1) 111 (2) 112
MEC and HEC
HEC
MEC
HEC
MEC and HEC
HEC
MEC and HEC
HEC
Not specified
Corticosteroids administered to all patients
Corticosteroids administered to all patients
Not specified
Corticosteroids not allowed
Not specified
Not specified
Not specified
Corticosteroids not allowed
Corticosteroids permitted: (1) 6.3 % (2) 5.8 % (3) 4.2 % Corticosteroids permitted: (1) 67.3 % (2) 67.3 % (3) 66.5 % Not specified
MEC
(1) 189 (2) 189 (3) 191
HEC
Corticosteroids not allowed
MEC
(1) 189 (2) 189 3) 185
(1) 104 (2) 104
Corticosteroids administered to all patients
Corticosteroids administered to all patients
Role of adjuvant corticosteroids
HEC
MEC and HEC
Chemotherapy emetogenicity
(1) 555 (2) 559
(1) 234 (2) 234
Efficacy sample size
(1) 73.5 (2) 73.1
(1) 100 (2) 100
(1) n/a (2) n/a
(1) 52.1 (2) 58.3 (3) 46.8
(1) n/a (2) n/a
(1) 32.0 (2) 40.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 36.5 (2) 36.5
(1) 54.8 (2) 55.4 (3) 52.3
(1) 30.0 (2) 32.0
(1) 50.7 (2) 51.6 (3) 51.1
(1) 82.5 (2) 82.0 (3) 81.7
(1) 73.0 (2) 71.4 (3) 71.9
(1) 58.7 (2) 58.0
(1) 78.0 (2) 83.0
Percentage female
(1) 52.1 (2) 52.2
(1) 50.9 (2) 52.9
(1) n/a (2) n/a
(1) n/a (2) n/a (3) n/a
(1) n/a (2) n/a
(1) 54.0 (2) 52.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 29.0 (2) 17.0 (3) 17.0
(1) n/a (2) n/a
(1) 59.0 (2) 64.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 52.1 (2) 48.7
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 90.0 (2) 85.0 (3) 85.0
(1) n/a (2) n/a
(1) 32.0 (2) 36.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 51.0 (2) 44.2
(1) n/a (2) n/a (3) n/a (1) n/a (2) n/a (3) n/a (1) n/a (2) n/a (3) n/a (1) 52.6 (2) 50.3
(1) 12.0 (2) 16.0
(1) n/a (2) n/a
(1) 57.8 (2) 59.6 (3) 59.3 (1) 49.3 (2) 51.6 (3) 52.5 (1) 50.6 (2) 53.4 (3) 50.9 (1) 14.0 (2) 7.7
(1) 69.3 (2) 65.6 (3) 65.4 (1) 66.1 (2) 65.6 (3) 70.2
(1) 42.3 (2) 40.2 (3) 42.2
(1) 44.4 (2) 47.1 (3) 42.2
(1) 54.8 (2) 56.1 (3) 55.3 (1) 55.2 (2) 53.3 (3) 53.6
(1) 93.5 (2) 92.3
(1) 100 (2) 100
Percentage chemotherapy-naïve
(1) 42.0 (2) 42.8
(1) n/a (2) n/a
Percentage nonusers of alcohol
(1) 58.4 (2) 58.0
(1) 57.0 (2) 57.0
Mean age (years)
MEC moderately emetogenic chemotherapy, HEC highly emetogenic chemotherapy, n/a data not available or extractable, CR complete response, CC complete control
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 24 mg
(1) Palonosetron 0.25 mg administered from day 1 up to day 5 (2) Palonosetron 0.25 mg administered on days 1, 3, and 5 (3) Ondansetron 8 mg (1) Palonosetron 0.25 mg (2) Ondansetron 16 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 16 mg
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg (3) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Ondansetron 32 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Dolasetron 100 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Ondansetron 32 mg
(1) Palonosetron 0.5 mg (2) Granisetron 40 μg/kg
(1) Palonosetron 0.25 mg (2) Granisetron 1 mg
Double-blind, parallel
Roscoe et al. [3]
Intervention
Study design
Trial
Table 1 Characteristics of randomized controlled trials included in the present meta-analysis
1688 Support Care Cancer (2014) 22:1685–1697
Support Care Cancer (2014) 22:1685–1697
Fig. 1 a Efficacy of palonosetron compared with other 5-hydroxytryptamine 3 receptor antagonists in the prophylaxis of chemotherapy-induced nausea and vomiting—complete response in the acute phase. b
1689
Complete control in the acute phase. c No emesis in the acute phase. d No nausea in the acute phase. e No rescue medications in the acute phase
1690
Support Care Cancer (2014) 22:1685–1697
(Fig. 2e). The RD analyses that surpassed the >10 % guideline of clinical significance were that of complete response (RD=12 %, 95 % CI 9–15 %, Table 2) and complete control (RD = 11 %, 95 % CI 8–15 %, Table 2). Upon subgroup emetogenicity analysis, statistically significant parameters included both endpoints of complete response (HEC: OR=1.60, 95 % CI 1.33–1.93; MEC: OR= 1.90, 95 % CI 1.47–2.45), both endpoints of complete control (HEC: OR=1.63, 95 % CI 1.32–2.01; MEC: OR=1.87, 95 % CI 1.45–2.41), one endpoint of no emesis (HEC: OR=1.58, 95 % CI 1.32–1.90), and one endpoint of no nausea (HEC: OR=1.66, 95 % CI 1.31–2.11). Corticosteroid analysis revealed statistical superiority of palonosetron in all analyzed parameters: CR in cohorts not receiving dexamethasone (OR=1.57, 95 % CI 1.18–2.10), CR in cohorts receiving dexamethasone (OR=1.65, 95 % CI 1.31–2.08), CC in cohorts not receiving dexamethasone (OR=1.58, 95 % CI 1.19– 2.11), CC in cohorts receiving dexamethasone (OR=1.49, 95 % CI 1.22–1.82), no emesis in cohorts receiving dexamethasone (OR=1.46, 95 % CI 1.19–1.80), and no nausea in cohorts receiving dexamethasone (OR=1.34, 95 % CI 1.09– 1.63).
Efficacy—overall phase Statistical significance favoring palonosetron was reached for all five endpoints: complete response (OR=1.54, 95 % CI 1.34–1.77, Fig. 3a), complete control (OR=1.54, 95 % CI 1.31–1.81, Fig. 3b), no emesis (OR=1.54, 95 % CI 1.32–1.80, Fig. 3c), no nausea (OR=1.51, 95 % CI 1.20–1.88, Fig. 3d), and no rescue medications (OR=1.53, 95 % CI 1.11–20.13, Fig. 3e). Only complete control in the overall phase exceeded the >10 % threshold for RD analysis (RD=11 %, 95 % CI 7– 14 %, Table 2). Emetogenicity analysis showed that palonosetron was superior in the following parameters: both endpoints of complete response (HEC: OR=1.56, 95 % CI 1.30–1.87; MEC: OR= 1.72, 95 % CI 1.34–2.22), both endpoints of complete control (HEC: OR=1.46, 95 % CI 1.16–1.84; MEC: OR=1.86, 95 % CI 1.44–2.40), both endpoints of no emesis (HEC: OR=1.56, 95 % CI 1.30–1.88; MEC: OR=1.43, 95 % CI 1.01–2.03), and one endpoint of no nausea (HEC: OR=1.49, 95 % CI 1.18–1.87). Corticosteroid subgroup analysis revealed that palonosetron was superior in all analyzed parameters: CR in cohorts not receiving dexamethasone (OR=1.50, 95 % CI 1.13–2.00), CR in cohorts receiving dexamethasone (OR=
Table 2 Absolute risk differences between palonosetron and other 5-hydroxytryptamine 3 receptor antagonist intervention arms for all included chemotherapy-induced nausea and vomiting endpoints Endpoint
Absolute risk difference (%)
95 % confidence interval (%)
Test for overall effect
Heterogeneity test
Satisfies MASCC/ESMO antiemetic guidelines requirement
CR, acute phase CR, delayed phase CR, overall phase CC, acute phase CC, delayed phase
6 12 10 6 11
3–8 9–15 7–14 2–9 8–15
p=0.0001 p
REVIEW ARTICLE
Efficacy and safety of palonosetron for the prophylaxis of chemotherapy-induced nausea and vomiting (CINV): a systematic review and meta-analysis of randomized controlled trials Marko Popovic & David G. Warr & Carlo DeAngelis & May Tsao & Kelvin K. W. Chan & Michael Poon & Cheryl Yip & Natalie Pulenzas & Henry Lam & Liying Zhang & Edward Chow
Received: 1 October 2013 / Accepted: 17 February 2014 / Published online: 4 March 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Palonosetron, a 5-hydroxytryptamine 3 receptor antagonist (5-HT3RA) with a strong binding affinity and long half-life, has been used in numerous trials for the prophylaxis of chemotherapy-induced nausea and vomiting (CINV). We systematically reviewed the efficacy and safety of palonosetron compared to other 5-HT 3 RAs in CINV prophylaxis. Methods A literature search of Ovid MEDLINE, EMBASE, and CENTRAL was conducted to identify randomized controlled trials (RCTs) comparing palonosetron to other 5HT3RAs in CINV prophylaxis. Primary endpoints were the percentage of patients achieving a complete response M. Popovic : M. Tsao : M. Poon : C. Yip : N. Pulenzas : H. Lam : L. Zhang : E. Chow (*) Rapid Response Radiotherapy Program, Department of Radiation Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada e-mail: [email protected] D. G. Warr Department of Medical Oncology, Princess Margaret Cancer Centre, University of Toronto, 610 University Avenue, Toronto, Ontario M5T 2M9, Canada C. DeAngelis Department of Pharmacy, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada K. K. W. Chan Department of Medical Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada
(CR), complete control (CC), no emesis, no nausea, or taking no rescue medications. Secondary endpoints were the percentage of patients suffering from 5-HT3RA-related adverse events. Results Sixteen RCTs were identified with 2,896 patients randomized to palonosetron and 3,187 patients randomized to other 5-HT3RAs. Palonosetron was consistently statistically superior in CR, CC, no emesis, or no nausea and was sometimes superior in no rescue medication. Subgroup analyses demonstrated similarity in efficacy between highly and moderately emetogenic chemotherapy cohorts. In the acute phase, statistical superiority of palonosetron was found for trials that did not allow dexamethasone; conversely, RCTs that administered dexamethasone to all patients were nonsignificant. Palonosetron was statistically significantly safer in dizziness and mean QTc interval change and similar in constipation, headache, and diarrhea. Clinical superiority of palonosetron was reached in 3 of 19 analyzed efficacy and safety endpoints. Conclusions Palonosetron is safer and more efficacious than other 5-HT 3 RAs. Future antiemetic guidelines should discuss the merits of including palonosetron as a first-line treatment. Keywords Palonosetron . 5-HT3 receptor antagonist . Chemotherapy-induced nausea and vomiting . Efficacy . Safety
Introduction Even though great progress has been made in the control of chemotherapy-induced nausea and vomiting (CINV), more
1686
Support Care Cancer (2014) 22:1685–1697
research is still needed in the delayed management and highrisk settings [1]. CINV may have the potential to impair a patient’s quality of life and disrupt the preferred treatment plan [2, 3]. As such, evaluating the safest and efficacious antiemetic regimen remains paramount to ameliorating the outcomes of future patients. Serotonin from the enterochromaffin cells in the small intestine binds to 5-hydroxytryptamine 3 (5-HT3) receptors on vagal afferents and thus results in the pathogenesis of CINV [4]. The end of the twentieth century saw the introduction of several 5-HT3 receptor antagonists (5-HT3RAs), including ondansetron, granisetron, tropisetron, and dolasetron. Trials that used these agents concluded that they were able to successfully control rates of vomiting; however, patientreported nausea, especially in the delayed phase, still remained a significant concern [3]. In an attempt to improve response rates, another 5-HT 3 RA named palonosetron (Aloxi®) was developed. Palonosetron is characterized by a markedly longer plasma elimination half-life of about 40 h (half-life of ondansetron 5.7 h) and a high selective binding affinity to the 5-HT3 receptor [4]. Several randomized controlled trials (RCTs) comparing palonosetron to other 5-HT3RAs concluded that a single dose of palonosetron was noninferior and sometimes superior in preventing CINV [4–9]. In antiemetic clinical trials, five endpoints are commonly used:
analysis is to comprehensively investigate the efficacy and safety of palonosetron in relation to other 5HT3RAs in the prophylaxis of CINV.
1. Complete response (CR; defined as no emetic episode and no use of rescue medication) 2. Complete control (CC; defined as no emetic episode, no use of rescue medication, and no more than mild nausea) 3. No emetic episodes 4. No episodes of nausea 5. No use of rescue medication
Endpoints
Despite the variance in endpoints, there have been three previous meta-analyses of RCTs that aimed to compare the efficacy and safety of palonosetron in relation to other 5HT3RAs [10–12]. Recent years have seen a more widespread use of palonosetron, which has led to a greater amount of fruitful research. As such, there is much more data available today than reported in the last meta-analysis, which completed data collection over 2 years ago [12]. As well, past analyses often had confined and incomprehensive endpoints. For efficacy, only CR and sometimes CC were previously analyzed. With respect to safety, only certain meta-analyses considered all of the most common 5-HT3RA-related adverse events— dizziness, diarrhea, constipation, and headaches—and none investigated differences in QTc prolongation. In view of these uncertainties, the purpose of this meta-
Patients and methods Data collection A literature search was conducted using the following databases: Ovid MEDLINE and OLDMEDLINE (1946–June Week 1 2013), Ovid EMBASE Classic and EMBASE (1947–2013 Week 26), and the Cochrane Central Register of Controlled Trials (CENTRAL) (May 2013). Reference lists of included RCTs and past meta-analyses were additionally searched. The following terms were employed to elicit relevant literature: “palonosetron,” “antineoplastic agents,” “neoplastic,” “chemotherapy,” “chemoradiotherapy,” “malignant,” “neoplasms,” “cancer,” “cancer chemotherapy,” “oncology,” “medical oncology,” and “drug therapy.” RCTs were included if they aimed to compare palonosetron to other 5-HT3RAs in the prophylaxis of CINV in at least one of the following five endpoints: CR, CC, no nausea, no emesis, or no rescue medication. Studies were excluded if they contained repeat data.
The primary endpoints were the percentage of patients achieving either a CR, CC, no emesis, no nausea, or taking no rescue medications within the acute, delayed, or overall phases. Secondary endpoints included the percentage of patients suffering from 5-HT3RA-related constipation, dizziness, diarrhea, or headache, as well as the mean change in the QTc interval. Each included study must have reported distinct acute, delayed, and overall phases with accompanying time frames; included trials were grouped into these phases based on what the authors reported and irrespective of time frame. Subgroup analysis For our primary endpoints, we stratified trials based on how the authors’ reported their chemotherapy emetogenicities. We sorted trials into either moderately emetogenic chemotherapy (MEC) or highly emetogenic chemotherapy (HEC) groups. In a second subgroup analysis, we sorted studies based on whether all or no patients received adjuvant dexamethasone on the day of palonosetron and chemotherapy administration.
Support Care Cancer (2014) 22:1685–1697
Sensitivity analysis A sensitivity analysis was conducted for cohorts with reasonably different clinical characteristics or for studies that were performed with reasonably different designs. Statistical analysis Statistical analysis was performed using Review Manager (RevMan 5.2) by Cochrane IMS. The Mantel-Haenszel method was applied and a fixed effect analysis model was used to generate odds ratio (OR) effect measures, absolute risk differences (RD), and accompanying 95 % confidence intervals (CI). A p value less than 0.05 was considered statistically significant in the test for overall effect, whereas a heterogeneity test p>0.05 was considered suitable. Absolute risk differences were compared to the 2010 Multinational Association of Supportive Care in Cancer/European Society of Medical Oncology (MASCC/ESMO) antiemetic guidelines, which state that “as a general rule, the panel considered changes of >10 % to be sufficient to warrant changing a guideline, given that the evidence supported this magnitude of benefit” [13]. The mean interval QTc change was computed and analyzed through a weighted analysis of variance test. PROC GLM was performed for the unbalanced data, and the total number of patients from each study was considered a weighted variable. ∑wx The weighted mean was defined as xw ¼ ∑i wi i , where wi is i i the weight for the ith study and xi is the ith variable value. This analysis was performed using Statistical Analysis of Software (SAS version 9.3 for Windows, Cary, NC). The intention-to-treat principle was utilized in all statistical analyses.
Results Upon literature search of 822 articles, 16 eligible RCTs were identified (Appendix 1) [3–9, 14–22]. Of note, an abstract [23] included in a previous meta-analysis [12] was excluded due to no specified randomization procedure. Of the 16 included RCTs, there were a total of 37 study arms, 2,896 patients randomized to receive palonosetron, and 3,187 patients randomized to any other 5-HT3RA. Characteristics of included studies Clinical and sociodemographic features of the included RCTs can be found in Table 1. Nine trials contained
1687
predominantly female patients [3–7, 9, 19, 21, 22], while three RCTs had mostly male patients [8, 14, 17]. The range of mean ages was 50.3 to 58.4 years for 15 of the 16 included RCTs; the only exception was a trial that enrolled only patients younger than 17 years of age [8]. Three RCTs included only patients receiving MEC [5, 6, 20], whereas eight studies included solely HEC patients [4, 7, 8, 14, 15, 17, 19, 21]. Corticosteroids were administered to all, some, or no patients in five [3, 4, 9, 20, 21], two [6, 7], and three [5, 14, 18] studies, respectively. In six studies, it was unknown whether corticosteroids were administered [8, 15–17, 19, 22]. Only four studies [3, 4, 18, 21] discussed the use of aprepitant as part of protocol; of these, three trials [3, 4, 18] did not allow aprepitant, while one trial [21] administered aprepitant to all patients, making any potential subgroup analysis unfeasible. Efficacy—acute phase Odds ratios demonstrated statistical superiority of palonosetron in four of five endpoints: complete response (OR=1.32, 95 % CI 1.15–1.52, Fig. 1a), complete control (OR=1.33, 95 % CI 1.12–1.58, Fig. 1b), no emesis (OR= 1.28, 95 % CI 1.10–1.50, Fig. 1c), and no nausea (OR=1.21, 95 % CI 1.01–1.45, Fig. 1d). The fifth endpoint, no rescue medication, was statistically nonsignificant (Fig. 1e). RD computations for all five endpoints in the acute phase fell short of the MASCC/ESMO threshold of >10 % (range 4– 6 %) (Table 2). Upon subgroup emetogenicity analysis, palonosetron demonstrated superiority in complete response (HEC: OR=1.26, 95 % CI 1.04–1.52; MEC: OR=1.43, 95 % CI 1.10–1.86), complete control for HEC cohorts (OR=1.23, 95 % CI 1.01– 1.50), no emesis for MEC cohorts (OR=1.36, 95 % CI 1.06– 1.74), and MEC cohorts without nausea (OR=1.56, 95 % CI 1.11–2.18). Corticosteroid subgroup analysis revealed that CR in cohorts not receiving adjuvant dexamethasone (OR=1.52, 95 % CI 1.15–2.02) and CC in cohorts receiving adjuvant dexamethasone (OR=1.30, 95 % CI 1.03–1.63) were the only significant parameters. Efficacy—delayed phase Palonosetron was statistically superior to other 5HT3RAs in four of five endpoints: complete response (OR=1.63, 95 % CI 1.42–1.87, Fig. 2a), complete control (OR=1.60, 95 % CI 1.39–1.84, Fig. 2b), no emesis (OR=1.57, 95 % CI 1.36–1.82, Fig. 2c), and no nausea (OR =1.39, 95 % CI 1.14–1.69, Fig. 2d). Again, no rescue medication was statistically nonsignificant
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg (1) Palonosetron 0.25 mg (2) Granisetron 3 mg
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Blind, parallel
Parallel
Double-blind, noninferiority, parallel
Double-blind, noninferiority, parallel
Blind, crossover
Nonparallel
Double-blind, noninferiority, crossover
Parallel
Crossover
Parallel
Double-blind, noninferiority, parallel
Saito et al. [4]
Gralla et al. [5]
Eisenberg et al. [6]
Aapro et al. [7]
Sepulveda-Vildosola et al. [8]
Ghosh et al. [9]
Yu et al. [14]
Chen et al. [15]
Li et al. [16]
Dong et al. [17]
Tian et al. [18]
Mattiuzzi et al. [19]
Kaushal et al. [20]
Wenzell et al. [21]
Huang et al. [22]
Corticosteroids administered to all patients
HEC
HEC MEC and HEC
(1) 223 (2) 223 (3) 221 (1) 50 (2) 50 (1) 157 (2) 616 (3) 440
(1) 117 (2) 119
(1) 20 (2) 20
(1) 30 (2) 30
(1) 48 (2) 48 (3) 47
(1) 72 (2) 72
(1) 44 (2) 45
(1) 104 (2) 104
(1) 111 (2) 112
MEC and HEC
HEC
MEC
HEC
MEC and HEC
HEC
MEC and HEC
HEC
Not specified
Corticosteroids administered to all patients
Corticosteroids administered to all patients
Not specified
Corticosteroids not allowed
Not specified
Not specified
Not specified
Corticosteroids not allowed
Corticosteroids permitted: (1) 6.3 % (2) 5.8 % (3) 4.2 % Corticosteroids permitted: (1) 67.3 % (2) 67.3 % (3) 66.5 % Not specified
MEC
(1) 189 (2) 189 (3) 191
HEC
Corticosteroids not allowed
MEC
(1) 189 (2) 189 3) 185
(1) 104 (2) 104
Corticosteroids administered to all patients
Corticosteroids administered to all patients
Role of adjuvant corticosteroids
HEC
MEC and HEC
Chemotherapy emetogenicity
(1) 555 (2) 559
(1) 234 (2) 234
Efficacy sample size
(1) 73.5 (2) 73.1
(1) 100 (2) 100
(1) n/a (2) n/a
(1) 52.1 (2) 58.3 (3) 46.8
(1) n/a (2) n/a
(1) 32.0 (2) 40.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 36.5 (2) 36.5
(1) 54.8 (2) 55.4 (3) 52.3
(1) 30.0 (2) 32.0
(1) 50.7 (2) 51.6 (3) 51.1
(1) 82.5 (2) 82.0 (3) 81.7
(1) 73.0 (2) 71.4 (3) 71.9
(1) 58.7 (2) 58.0
(1) 78.0 (2) 83.0
Percentage female
(1) 52.1 (2) 52.2
(1) 50.9 (2) 52.9
(1) n/a (2) n/a
(1) n/a (2) n/a (3) n/a
(1) n/a (2) n/a
(1) 54.0 (2) 52.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 29.0 (2) 17.0 (3) 17.0
(1) n/a (2) n/a
(1) 59.0 (2) 64.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 52.1 (2) 48.7
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 90.0 (2) 85.0 (3) 85.0
(1) n/a (2) n/a
(1) 32.0 (2) 36.0
(1) n/a (2) n/a
(1) n/a (2) n/a
(1) 51.0 (2) 44.2
(1) n/a (2) n/a (3) n/a (1) n/a (2) n/a (3) n/a (1) n/a (2) n/a (3) n/a (1) 52.6 (2) 50.3
(1) 12.0 (2) 16.0
(1) n/a (2) n/a
(1) 57.8 (2) 59.6 (3) 59.3 (1) 49.3 (2) 51.6 (3) 52.5 (1) 50.6 (2) 53.4 (3) 50.9 (1) 14.0 (2) 7.7
(1) 69.3 (2) 65.6 (3) 65.4 (1) 66.1 (2) 65.6 (3) 70.2
(1) 42.3 (2) 40.2 (3) 42.2
(1) 44.4 (2) 47.1 (3) 42.2
(1) 54.8 (2) 56.1 (3) 55.3 (1) 55.2 (2) 53.3 (3) 53.6
(1) 93.5 (2) 92.3
(1) 100 (2) 100
Percentage chemotherapy-naïve
(1) 42.0 (2) 42.8
(1) n/a (2) n/a
Percentage nonusers of alcohol
(1) 58.4 (2) 58.0
(1) 57.0 (2) 57.0
Mean age (years)
MEC moderately emetogenic chemotherapy, HEC highly emetogenic chemotherapy, n/a data not available or extractable, CR complete response, CC complete control
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 24 mg
(1) Palonosetron 0.25 mg administered from day 1 up to day 5 (2) Palonosetron 0.25 mg administered on days 1, 3, and 5 (3) Ondansetron 8 mg (1) Palonosetron 0.25 mg (2) Ondansetron 16 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 16 mg
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg (3) Granisetron 3 mg
(1) Palonosetron 0.25 mg (2) Ondansetron 8 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Ondansetron 32 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Dolasetron 100 mg
(1) Palonosetron 0.75 mg (2) Palonosetron 0.25 mg (3) Ondansetron 32 mg
(1) Palonosetron 0.5 mg (2) Granisetron 40 μg/kg
(1) Palonosetron 0.25 mg (2) Granisetron 1 mg
Double-blind, parallel
Roscoe et al. [3]
Intervention
Study design
Trial
Table 1 Characteristics of randomized controlled trials included in the present meta-analysis
1688 Support Care Cancer (2014) 22:1685–1697
Support Care Cancer (2014) 22:1685–1697
Fig. 1 a Efficacy of palonosetron compared with other 5-hydroxytryptamine 3 receptor antagonists in the prophylaxis of chemotherapy-induced nausea and vomiting—complete response in the acute phase. b
1689
Complete control in the acute phase. c No emesis in the acute phase. d No nausea in the acute phase. e No rescue medications in the acute phase
1690
Support Care Cancer (2014) 22:1685–1697
(Fig. 2e). The RD analyses that surpassed the >10 % guideline of clinical significance were that of complete response (RD=12 %, 95 % CI 9–15 %, Table 2) and complete control (RD = 11 %, 95 % CI 8–15 %, Table 2). Upon subgroup emetogenicity analysis, statistically significant parameters included both endpoints of complete response (HEC: OR=1.60, 95 % CI 1.33–1.93; MEC: OR= 1.90, 95 % CI 1.47–2.45), both endpoints of complete control (HEC: OR=1.63, 95 % CI 1.32–2.01; MEC: OR=1.87, 95 % CI 1.45–2.41), one endpoint of no emesis (HEC: OR=1.58, 95 % CI 1.32–1.90), and one endpoint of no nausea (HEC: OR=1.66, 95 % CI 1.31–2.11). Corticosteroid analysis revealed statistical superiority of palonosetron in all analyzed parameters: CR in cohorts not receiving dexamethasone (OR=1.57, 95 % CI 1.18–2.10), CR in cohorts receiving dexamethasone (OR=1.65, 95 % CI 1.31–2.08), CC in cohorts not receiving dexamethasone (OR=1.58, 95 % CI 1.19– 2.11), CC in cohorts receiving dexamethasone (OR=1.49, 95 % CI 1.22–1.82), no emesis in cohorts receiving dexamethasone (OR=1.46, 95 % CI 1.19–1.80), and no nausea in cohorts receiving dexamethasone (OR=1.34, 95 % CI 1.09– 1.63).
Efficacy—overall phase Statistical significance favoring palonosetron was reached for all five endpoints: complete response (OR=1.54, 95 % CI 1.34–1.77, Fig. 3a), complete control (OR=1.54, 95 % CI 1.31–1.81, Fig. 3b), no emesis (OR=1.54, 95 % CI 1.32–1.80, Fig. 3c), no nausea (OR=1.51, 95 % CI 1.20–1.88, Fig. 3d), and no rescue medications (OR=1.53, 95 % CI 1.11–20.13, Fig. 3e). Only complete control in the overall phase exceeded the >10 % threshold for RD analysis (RD=11 %, 95 % CI 7– 14 %, Table 2). Emetogenicity analysis showed that palonosetron was superior in the following parameters: both endpoints of complete response (HEC: OR=1.56, 95 % CI 1.30–1.87; MEC: OR= 1.72, 95 % CI 1.34–2.22), both endpoints of complete control (HEC: OR=1.46, 95 % CI 1.16–1.84; MEC: OR=1.86, 95 % CI 1.44–2.40), both endpoints of no emesis (HEC: OR=1.56, 95 % CI 1.30–1.88; MEC: OR=1.43, 95 % CI 1.01–2.03), and one endpoint of no nausea (HEC: OR=1.49, 95 % CI 1.18–1.87). Corticosteroid subgroup analysis revealed that palonosetron was superior in all analyzed parameters: CR in cohorts not receiving dexamethasone (OR=1.50, 95 % CI 1.13–2.00), CR in cohorts receiving dexamethasone (OR=
Table 2 Absolute risk differences between palonosetron and other 5-hydroxytryptamine 3 receptor antagonist intervention arms for all included chemotherapy-induced nausea and vomiting endpoints Endpoint
Absolute risk difference (%)
95 % confidence interval (%)
Test for overall effect
Heterogeneity test
Satisfies MASCC/ESMO antiemetic guidelines requirement
CR, acute phase CR, delayed phase CR, overall phase CC, acute phase CC, delayed phase
6 12 10 6 11
3–8 9–15 7–14 2–9 8–15
p=0.0001 p
