VOLUME
32
䡠
NUMBER
14
䡠
MAY
10
2014
JOURNAL OF CLINICAL ONCOLOGY
O R I G I N A L
R E P O R T
Randomized, Controlled, Double-Blind, Cross-Over Trial Assessing Treatment Preference for Pazopanib Versus Sunitinib in Patients With Metastatic Renal Cell Carcinoma: PISCES Study Bernard Escudier, Institut Gustave Roussy, Villejuif; Emmanuel Sevin, Centre Franc¸ois Baclesse, Caen; Sylvie Ne´grier, Leon Berard Cancer Center, Lyon, France; Camillo Porta, Fondazione Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Policlinico S. Matteo, Pavia; Cora N. Sternberg, San Camillo Forlanini Hospital, Rome; Ugo De Giorgi, IRCCS Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori, Meldola, Italy; Petri Bono, Helsinki University Central Hospital, Helsinki, Finland; Thomas Powles, Barts Experimental Cancer Medicine Centre, Barts Cancer Institute, Queen Mary University of London, London; Tim Eisen, Cambridge University Health Partners, Cambridge; Omi Parikh, Royal Preston Hospital, Lancashire; Robert Hawkins, Christie Cancer Research UK, Manchester; Sadya Khan, Jose Diaz, and Faisal Mehmud, GlaxoSmithKline, Uxbridge, United Kingdom; Ju¨rgen E. Gschwend, Klinikum Rechts der Isar der Technischen Universita¨t Mu¨nchen, Munich, Germany; Suman Redhu, GlaxoSmithKline, Collegeville, PA; and David Cella, Northwestern University Feinberg School of Medicine, Chicago, IL. Published online ahead of print at www.jco.org on March 31, 2014. Supported by GlaxoSmithKline Pharmaceuticals, Philadelphia, PA. Presented in part at the 48th Annual Meeting of the American Society of Clinical Oncology, June 1-5, 2012, Chicago, IL. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article. Clinical trial information: NCT01064310. Corresponding author: Bernard Escudier, MD, Institut Gustave Roussy, 114 rue Edouard Vaillant, 94805 Villejuif, France; e-mail:
[email protected]. © 2014 by American Society of Clinical Oncology 0732-183X/14/3214w-1412w/$20.00 DOI: 10.1200/JCO.2013.50.8267
1412
Bernard Escudier, Camillo Porta, Petri Bono, Thomas Powles, Tim Eisen, Cora N. Sternberg, Ju¨rgen E. Gschwend, Ugo De Giorgi, Omi Parikh, Robert Hawkins, Emmanuel Sevin, Sylvie Ne´grier, Sadya Khan, Jose Diaz, Suman Redhu, Faisal Mehmud, and David Cella See accompanying editorial on page 1392 A
B
S
T
R
A
C
T
Purpose Patient-reported outcomes may help inform treatment choice in advanced/metastatic renal cell carcinoma (RCC), particularly between approved targeted therapies with similar efficacy. This double-blind cross-over study evaluated patient preference for pazopanib or sunitinib and the influence of health-related quality of life (HRQoL) and safety factors on their stated preference. Patients and Methods Patients with metastatic RCC were randomly assigned to pazopanib 800 mg per day for 10 weeks, a 2-week washout, and then sunitinib 50 mg per day (4 weeks on, 2 weeks off, 4 weeks on) for 10 weeks, or the reverse sequence. The primary end point, patient preference for a specific treatment, was assessed by questionnaire at the end of the two treatment periods. Other end points and analyses included reasons for preference, physician preference, safety, and HRQoL. Results Of 169 randomly assigned patients, 114 met the following prespecified modified intent-to-treat criteria for the primary analysis: exposure to both treatments, no disease progression before cross over, and completion of the preference questionnaire. Significantly more patients preferred pazopanib (70%) over sunitinib (22%); 8% expressed no preference (P ⬍ .001). All preplanned sensitivity analyses, including the intent-to-treat population, statistically favored pazopanib. Less fatigue and better overall quality of life were the main reasons for preferring pazopanib, with less diarrhea being the most cited reason for preferring sunitinib. Physicians also preferred pazopanib (61%) over sunitinib (22%); 17% expressed no preference. Adverse events were consistent with each drug’s known profile. Pazopanib was superior to sunitinib in HRQoL measures evaluating fatigue, hand/foot soreness, and mouth/throat soreness. Conclusion This innovative cross-over trial demonstrated a significant patient preference for pazopanib over sunitinib, with HRQoL and safety as key influencing factors. J Clin Oncol 32:1412-1418. © 2014 by American Society of Clinical Oncology
INTRODUCTION
Since 2005, several targeted agents have been approved for the treatment of metastatic renal cell carcinoma (RCC).1 Data from head-to-head studies comparing these agents are limited,2,3 and thus, current guidelines include multiple treatment options.4,5 Among the tyrosine kinase inhibitors recommended, pazopanib and sunitinib are commonly used first-line treatments, and a phase III study demonstrated similar efficacy of pazopanib versus sunitinib.2 This study also demonstrated that pazopanib was associated with a
lower incidence of certain toxicities, including hand-foot syndrome and fatigue, as well as a higher incidence of liver function test abnormalities. Moreover, 11 of 14 health-related quality-oflife (HRQoL) domains favored pazopanib over sunitinib, an important factor to consider in this palliative setting.2 Patient-reported outcomes and patient preference may reveal differences in tolerability that are not captured from standard adverse event (AE) reporting measures, especially in the context of low-grade chronic toxicities that may lead to significant morbidity.6
© 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Pazopanib v Sunitinib Patient Preference in Metastatic RCC
This randomized cross-over study was designed to evaluate patient preference for continued treatment with pazopanib or sunitinib and identify factors that influenced a patient’s choice. The overall objective of the study was to provide health care professionals with insights on the patient experience with these agents.
PATIENTS AND METHODS Patients Adult patients (ⱖ 18 years) who had no prior systemic therapy for their locally advanced or metastatic RCC and had an Eastern Cooperative Oncology Group performance status of 0 or 1 and adequate hematologic, hepatic, renal, and blood coagulation function were eligible. Patients with non– clear cell histology and nonmeasurable disease by RECIST were also eligible. Major exclusion criteria included history or evidence of brain metastases, GI abnormalities, presence of uncontrolled infection, or cardiovascular abnormality. The study protocol was reviewed and approved by local institutional review boards; the study was conducted in accordance with good clinical practice and the principles of the Declaration of Helsinki. All patients provided written informed consent before enrollment. Study Design and End Points This double-blind, multicenter, phase IIIb cross-over study randomly assigned patients (1:1) to either pazopanib during period 1 followed by sunitinib during period 2 (pazopanib-sunitinib) or the reverse (sunitinibpazopanib). Pazopanib and sunitinib were provided by GlaxoSmithKline (Philadelphia, PA; trial sponsor), and blinding was done by overencapsulation. Patients were stratified by Eastern Cooperative Oncology Group status (0 v 1) and number of metastatic sites (zero or one v ⱖ two metastatic sites). A 10-week treatment period was selected to provide the two treatments to patients under similar health conditions (eg, tumor burden and cancer-related comorbidities). Each treatment period was separated by a 2-week washout, equating to five or more half-lives for either treatment, to allow the effective systemic elimination of the drug before initiation of subsequent treatment. Pazopanib dosing was 800 mg once daily for 10 weeks. Sunitinib dosing was 50 mg once daily for 4 weeks, matching placebo for 2 weeks, and then sunitinib 50 mg once daily for the remaining 4 weeks in accordance with the approved regimen. Dose reductions for AEs were permitted. However, to avoid the potential for prolonged time off therapy as a result of the blinded placebo interval during sunitinib dosing, patients with AEs that necessitated dose interruptions had the dose interruption but were then required to either cross over early from period 1 to period 2 or come off study and complete the preference questionnaire if they were in period 2. All patients were informed of their disease assessment result after period 1, as per standard clinical practice. In addition, for ethical reasons, patients with a clinically significant objective response (defined by the study steering committee as a ⬎ 50% reduction in tumor size/complete response if nonmeasurable disease) were required to reconsent before crossing over. The primary end point was patient preference. Secondary end points and analyses included evaluation of specific reasons for preference, HRQoL, physician preference, frequency and time to dose modifications, and safety. Assessments Patient preference was assessed by questionnaire at study end, before unblinding and before patients were informed of the final disease assessment. Patients were asked whether they preferred to continue treatment with the drug administered during period 1 or period 2 or whether they had no preference, assuming both treatments were equally effective. In addition, patients were queried on prespecified factors influencing their preference and the factors’ relative importance (Data Supplement). Physicians were also asked to state their preference for a patient’s further treatment before unblinding, taking into account the patient’s preference, tumor assessments, toxicities, and laboratory tests. www.jco.org
AEs were graded according to Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. HRQoL was assessed using the Functional Assessment of Cancer Therapy–Fatigue (FACT-F) questionnaire, EuroQoL EQ-5D health questionnaire, and the Supplementary Quality of Life Questionnaire (SQLQ).7 The FACT-F and SQLQ were administered every 2 weeks. Statistical Methods and Analysis The study was designed to control the ␣ error rate at .10 with a power of ⱖ 80%. Required sample size was estimated as 160 patients based on the following assumptions: 50% of patients would express preference for one treatment, 30% would express preference for the other treatment, and 20% would have no preference (the minimum difference of 20% between the two agents was deemed adequate by the study steering committee as clinically relevant given the limited empirical data on which to base a priori statistical assumptions for the novel trial end point of patient preference); and ⱕ 20% of patients would withdraw from the study before completing the questionnaire or because of progressive disease (PD) after period 1. The intent-to-treat (ITT) population comprised patients who received ⱖ one dose of study drug in each treatment period. The modified ITT population was used to assess the primary end point and was predefined as all patients who received ⱖ one dose of study drug in each period, did not have PD at the end of period 1 (assessed by RECIST 1.1), and completed the preference questionnaire. Patients with PD at the end of period 1 were excluded from the primary analysis because their preference would potentially be confounded by their knowledge of their disease progression on the period 1 treatment. However, patients with an objective response, who were also aware of their assessment at the end of period 1, were included in the primary analysis because the preference questionnaire was administered before patients had knowledge of their final tumor assessment after period 2 and therefore could still choose either treatment based on tolerability and safety differences. Prescott’s test was used to analyze the primary end point to test the significance of difference between the two treatments in the presence of period effects. Reasons for patient preference and physician preference were summarized descriptively. Sensitivity analyses included analyses using data only from patients who completed the full 10 weeks of treatment in each period, using data from all completed questionnaires (including patients who had PD at the end of period 1) and imputation of sunitinib preference for patients who did not complete the questionnaire. Both 95% and 90% CIs for patient preference are reported. Cross-over analyses for HRQoL scores averaged the between-treatment difference for each patient within each sequence and then across both sequences, providing an estimate of treatment effect.8 The estimated treatment difference, 95% CI (90% CI in Appendix tables and Appendix Fig A1, online only), and P value were adjusted for period and sequence effects in the analysis of variance model. No adjustment was made for multiple comparisons; P values are two-sided.
RESULTS
Patient Characteristics and Disposition Of 169 randomly assigned patients, one patient was randomly assigned in error and excluded from all analyses. Baseline disease characteristics were balanced between treatment arms (Table 1). Two patients in the sunitinib-pazopanib arm were excluded from the safety population because of a dispensing error. The predefined modified ITT population comprised 114 patients (54 patients were excluded: 32 patients did not receive ⱖ one dose of study drug in each treatment period [ie, did not cross over to period 2], 12 had PD after period 1, and 10 withdrew before completing the preference questionnaire; Fig 1). Patient Preference The proportion of patients who preferred pazopanib was 70% (95% CI, 60.9% to 78.4%), whereas 22% (95% CI, 14.7% to 30.6%) © 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
1413
Escudier et al
Table 1. Patient Demographics and Baseline Disease Characteristics Demographic or Clinical Characteristic
S-P (n ⫽ 82) No. of Patients
Median age, years Male sex Ethnicity Hispanic/Latino Not Hispanic/Latino Missing Raceⴱ African American/African heritage Asian Central/South Asian heritage White ECOG performance status 0 1 Primary tumor type Renal cell Missing Time since diagnosis, months Median Range Histology Papillary adenocarcinoma Clear cell Chromophobe Other Unknown Measurable disease‡ Nontarget lesions‡ No. of metastatic sites 0 or 1 ⱖ2 Prior anticancer therapy Any therapy Nephrectomy Radiotherapy
P-S (n ⫽ 86) %
No. of Patients
62.0
Total (N ⫽ 168) %
No. of Patients
64.0
%
63.0
52
63
61
71
113
67
10 66 6
12 80 7
6 77 3
7 90 3
16 143 9
10 85 5
1 1 1 74
1 1 1 97
0 0 0 83
0 0 0 100
1 1 1 157
⬍1 ⬍1 ⬍1 99
61 21
74 26
60 26
70 30
121 47
72 28
78 4
95 5
81 5
94 6
159 9
95 5
6.7 1-222† 3 76 1 0 2 75 64
9.1 0-241 4 93 1
7.7 0-241
2 91 78
3 75 1 4 3 80 68
3 87 1 5 3 93 79
6 151 2 4 5 155 132
4 90 1 2 3 92 79
23 58
28 71
20 65
23 76
43 123
26 73
71 70 3
87 85 4
80 79 11
93 92 13
151 149 14
90 89 8
Abbreviations: ECOG, Eastern Cooperative Oncology Group; P-S, pazopanib followed by sunitinib; S-P, sunitinib followed by pazopanib. ⴱ Race data were available for 159 patients (S-P, n ⫽ 76; P-S, n ⫽ 83). †Data were not available for one patient in the S-P sequence. ‡As defined by the investigator.
preferred sunitinib and 8% expressed no preference. The proportion of patients who completed the patient preference questionnaire was similar between the sunitinib-pazopanib and pazopanib-sunitinib arms (64 patients [78%] v 62 patients [72%], respectively). There was a period effect, in that a greater proportion of patients preferred the treatment in period 1 (54%) versus period 2 (38%) versus no preference (8%). Despite this period effect, patients in the sunitinib-pazopanib arm preferred pazopanib approximately twice as often as sunitinib (62% v 32%, respectively). In the pazopanib-sunitinib arm, 80% of patients preferred pazopanib, whereas 11% preferred sunitinib. After adjusting for this period effect, which accounts for differences in preference percentages in the two sequences, the difference in patient preference for pazopanib versus sunitinib was 49.3% (95% CI, 34.7% to 63.8%) in favor of pazopanib and was statistically significant (P ⬍ .001). See Appendix Table A1 (online only) for additional primary analysis data based on 90% CIs for which the study was originally 1414
© 2014 by American Society of Clinical Oncology
powered. All sensitivity analyses were statistically significant at the predetermined 10% ␣ level and favored pazopanib (Appendix Fig A1 and Table A2, online only), including the ITT population and the worst-case sensitivity analysis, which imputed sunitinib as the preference for all patients who did not complete the preference questionnaire. Similar results were obtained in analyses at the 5% ␣ level (Fig 2), with the exception of the worst-case sensitivity analysis, which was not significant (54% preferred pazopanib v 40% preferred sunitinib; P ⫽ .065). The most common (ⱖ 10%) factors influencing patient preference included diarrhea, HRQoL, fatigue, loss of appetite, taste changes, nausea/vomiting, hand/foot soreness, stomach pain, and mouth/throat soreness (Fig 3). Eighty patients (70%) preferred pazopanib; the most common reasons included better overall quality of life (QoL) and less fatigue. Twenty-five patients (22%) preferred sunitinib; the most common reasons included less diarrhea and better overall QoL. JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Pazopanib v Sunitinib Patient Preference in Metastatic RCC
Randomly assigned 1:1 (N = 169*)
Period 1: Sunitinib 50 mg QD (n = 82) Withdrawals (n = 14) Adverse event (n = 7 ) Entered open-label phase without (n = 1) completing preference questionnaire Lack of efficacy (n = 3 ) Investigator discretion (n = 1) Patient withdrew consent (n = 2)
Period 1: Pazopanib 800 mg QD Withdrawals Adverse event Death Lack of efficacy Investigator discretion Patient withdrew consent
(n = 86) (n = 18) (n = 9) (n = 2) (n = 4) (n = 1) (n = 2)
Period 2: Pazopanib 800 mg QD (n = 68) Withdrawals (n = 4) Adverse event (n = 1) Death (n = 1 ) Entered open-label phase without (n = 1) completing preference questionnaire Lack of efficacy (n = 1 )
Period 2: Sunitinib 50 mg QD Withdrawals Adverse event Death Lack of efficacy
(n = 68) (n = 6) (n = 1) (n = 3) (n = 2)
End of randomized phase (n = 64) Excluded from primary analysis due to (n = 4) progressive disease during period 1
End of randomized phase (n = 62) Excluded from primary analysis due to (n = 8) progressive disease during period 1
Analyzed for preference (n = 60) Received ≥ 1 dose of drug during each period Did not have progressive disease after period 1 Completed preference questionnaire
Analyzed for preference (n = 54) Received ≥ 1 dose of drug during each period Did not have progressive disease after period 1 Completed preference questionnaire
When queried regarding the single most important factor for preference, patients preferring pazopanib selected “no single reason” and less fatigue, and patients preferring sunitinib selected “no single reason” and less diarrhea. Overall, approximately 25% of patients selected “no single reason.” Physician Preference Physician preferences were consistent with patient preferences. More physicians preferred to continue their patients on pazopanib (61%) than on sunitinib (22%), with 17% stating no preference.
Safety Dose modifications and treatment discontinuation. Overall, 11% (n ⫽ 18) of patients withdrew from the study as a result of an AE. The most common reason for early discontinuation was an AE, with no
Better quality of life Less fatigue Less change in food tastes Less soreness in mouth/throat
Sensitivity Analysis 1 (n = 80)
P < .001
Sensitivity Analysis 2 (n = 126)
P < .001
Sensitivity Analysis 3 (n = 166)
P < .001
Sensitivity Analysis 4 (n = 166)
P = .065
Less stomach pain
Sensitivity Analysis 5 (n = 166)
P = .035
Less diarrhea
Less nausea/vomiting Less soreness in hands/feet Less loss of appetite
0
10
20
30
40
50
60
70
80
90 100
Mean (95% CI) percentage of patients expressing preference for pazopanib Mean (95% CI) percentage of patients expressing preference for sunitinib
Increasing Rate of Preference Fig 2. Sensitivity analyses of patient preferences for pazopanib versus sunitinib. Bars represent 95% CIs. Analysis descriptions are provided in Appendix Table A3 (online only). www.jco.org
Fig 1. CONSORT diagram. QD, once daily. (*) One patient was randomly assigned in error, and no data were entered into the study for this patient; data were available for 168 patients.
Pazopanib preferred (n = 80) Sunitinib preferred (n = 25)
Other Less hair color change 0
10
20
30
40
50
60
70
No. of Patients Fig 3. Common factors influencing patient preference. Patients were able to select more than one item on the questionnaire.
© 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
1415
Escudier et al
HRQoL Cross-over analyses compared patients’ average scores on each treatment, adjusting for sequence. In a cross-over analysis of FACT-F scores, the difference in fatigue score between pazopanib and sunitinib treatment was statistically significant, with an overall treatment difference of 2.49 points (Table 3; Appendix Table A3, online only). In a cross-over analysis of SQLQ scores, the treatment differences in mouth/throat, hand, and foot soreness scores between pazopanib and sunitinib treatment significantly favored pazopanib. Similarly, a cross-over analysis of SQLQ limitation scores showed that the treatment differences in limitations as a result of mouth/throat soreness and limitations as a result of foot soreness scores between pazopanib and sunitinib treatment significantly favored pazopanib (Table 3; Appendix Table A3).
single AE dominating as a cause. The rates of premature discontinuations in period 1 as a result of AEs were 18% (15 of 82 patients) and 14% (12 of 86 patients) for sunitinib and pazopanib, respectively; and in period 2, these rates 15% (10 of 68 patients) and 31% (21 of 68 patients), respectively. These discontinuations represent patients who required a dose interruption but, because of the study design, either crossed over early or came off study. The median time to dose reduction was similar between arms (sunitinib, 3.7 weeks; pazopanib, 4.0 weeks). Overall, dose reductions for patients on sunitinib and pazopanib were 20% and 13%, respectively, and dose interruptions were 12% and 6%, respectively. Seven patients (9%) withdrew from period 1 in the sunitinib-pazopanib arm as a result of an AE, and nine patients (10%) withdrew in the pazopanib-sunitinib arm as a result of an AE. AEs. The dose-intensities of sunitinib and pazopanib were 96% and 98%, respectively. Diarrhea, fatigue, and nausea were the most common AEs reported with sunitinib or pazopanib (Table 2). Two fatal serious AEs (SAEs) were reported on sunitinib (dyspnea and worsening of general condition), and two fatal SAEs were reported on pazopanib (respiratory failure and peritonitis); none of these fatal SAEs were considered treatment related. In addition, three patient deaths during the study occurred as a result of PD (two patients on sunitinib and one patient receiving pazopanib).
DISCUSSION
This study revealed a more than three-fold greater patient preference for pazopanib over sunitinib, with preplanned sensitivity analyses confirming these results. The difference in preference exceeded the 20% threshold that was considered meaningful during trial planning, suggesting that the difference in symptomatic toxicities between the
Table 2. Adverse Events (ⱖ 10% of patients on either sunitinib or pazopanib) by Maximum Grade (safety population) Sunitinib (n ⫽ 148)ⴱ All Grades
Grade 1
Grade 2
Pazopanib (n ⫽ 153)ⴱ Grade 3
Grade 4
All Grades
Grade 1
Grade 2
Grade 3
Grade 4
Adverse Event
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
Any event Diarrhea Fatigue Nausea Dysgeusia Hypertension Palmar-plantar erythrodysesthesia Asthenia Mucosal inflammation Decreased appetite Dyspepsia Stomatitis Vomiting Constipation Hair color changes Abdominal pain upper Abdominal pain Epistaxis Headache Rash
147 47 44 44 40 38 38 35 32 28 23 23 24 22 20 19 16 16 17 17
⬎ 99† 32 30 30 27 26 26 24 22 19 16 16 16 15 14 13 11 11 11 11
20 32 21 37 34 8 21 16 18 18 18 9 16 13 20 11 14 10 13 16
14 22 14 25 23 5 14 11 12 12 12 6 11 9 14 7 9 7 9 11
57 14 16 7 6 17 11 15 12 9 5 11 7 9 0 8 1 5 4 1
39 9 11 5 4 11 7 10 8 6 3 7 5 6
58 1 7 0 0 13 6 4 2 1 0 3 1 0 0 0 1 1 0 0
39 ⬍1 5
11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
7‡
148 64 44 50 25 35 25 25 24 31 16 7 22 11 26 7 20 8 22 13
97† 42 29 33 16 23 16 16 16 20 10 5 14 7 17 5 13 5 14 8
25 41 18 37 22 8 14 16 21 25 15 3 17 9 25 2 9 8 13 12
16 27 12 24 14 5 9 10 14 16 10 2 11 6 16 1 6 5 8 8
63 22 20 12 3 15 9 8 3 6 1 3 4 1 1 5 11 0 8 0
41 14 13 8 2 10 6 5 2 4 ⬍1 2 3 ⬍1 ⬍1 3 7
51 1 6 1 0 12 2 1 0 0 0 1 1 1 0 0 0 0 1 1
33 ⬍1 4 ⬍1
8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
5§
5 ⬍1 3 3 ⬍1
9 4 3 1 ⬍1 2 ⬍1
⬍1 ⬍1
5
8 1 ⬍1
⬍1 ⬍1 ⬍1
⬍1 ⬍1
NOTE. Patient numbers for pazopanib (n ⫽ 148) and sunitinib (n ⫽ 153) represent the patients exposed to that treatment, regardless of treatment period. Adverse events are sorted in descending order based on the overall incidence for sunitinib treatment. Physical examination was performed at baseline, at weeks 2 and 6 of each treatment period, during the washout period, and at study end. Laboratory tests were assessed at baseline; at weeks 2, 6, and 8 of each treatment period; during the washout period; and at study end. ⴱ An on-therapy grade 5 adverse event was reported for one patient (⬍ 1%) on sunitinib. Grade 5 adverse events were not reported for patients on pazopanib. †Total number of adverse events of any grade reported during the randomized phase of the study. ‡Included (one each) general physical health deterioration, intracranial hemorrhage, dyspnea, pulmonary embolism, hyponatremia, blood potassium decreased, infectious peritonitis, neutropenic infection, thrombocytopenia, disorientation, and acute myocardial infarction. §Included (one each) transient ischemic attack, spinal cord compression, back pain, respiratory failure, hyponatremia, ALT elevation, AST elevation, infectious peritonitis, and acute myocardial infarction.
1416
© 2014 by American Society of Clinical Oncology
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Pazopanib v Sunitinib Patient Preference in Metastatic RCC
Table 3. Patient-Reported Quality-of-Life Measures (safety population) Pazopanib Quality-of-Life Measure
Mean Score
FACT-F‡ SQLQ Worst mouth and throat soreness Worst hand soreness Worst foot soreness SQLQ limitation scores Mouth and throat soreness Foot soreness
38.1
ⴱ
Sunitinib
No. of Patients
Mean Scoreⴱ
No. of Patients
131
35.6
131
2.49
Treatment Difference†
0.40 0.21 0.36
131 131 129
0.78 0.29 0.52
131 131 129
⫺0.38 ⫺0.08 ⫺0.16
14.32 13.82
126 129
13.72 13.24
126 129
0.60 0.58
95% CI
P
0.92 to 4.07
.002
⫺0.49 to ⫺0.27 ⫺0.15 to ⫺0.01 ⫺0.27 to ⫺0.05
⬍ .001 .026 .005
0.29 to 0.90 0.21 to 0.96
⬍ .001 .003
NOTE. Only patients who completed at least one postbaseline assessment in both treatment periods were included in the analysis. Fatigue and the impact of fatigue on life functions were assessed by the FACT-F score; mouth/throat soreness, hand/foot soreness, and work activity were assessed by SQLQ at day 1 and every 2 weeks thereafter. Abbreviations: FACT-F, Functional Assessment of Cancer Therapy–Fatigue; SQLQ, Supplementary Quality of Life Questionnaire. ⴱ Mean score was calculated for all patients as their average postbaseline score within each period averaged over periods; higher scores indicated better health. †The estimated treatment difference CI and P value were adjusted for period and sequence effects in the analysis of variance model. ‡Patients were required to complete assessments in both treatment periods to be included in the analysis.
two drugs was meaningful to patients. Patients preferred pazopanib for a variety of reasons, with better overall QoL a commonly selected reason, suggesting that the perceptions of HRQoL are complex and unlikely to be completely captured by discrete assessments of specific health dimensions in isolation. Less impact of fatigue was the second most commonly selected reason for preferring pazopanib, consistent with the FACT-F crossover analysis that showed a statistically significant difference favoring pazopanib over sunitinib by 2.5 points. The difference of 2.5 points is approximately 1 SE of measurement, which is regarded as likely to be a meaningful difference. Furthermore, a 2.5-point difference in terms of fatigue has been reported as the difference between patients with a hemoglobin of 8 g/dL versus 10 g/dL.9,10 Given the baseline starting point of fatigue in these patients, the magnitude of this difference equates to a patient response of “rarely” versus “sometimes” to the following question: “How often did you feel tired even when you hadn’t done anything?”11 Whereas the rates of asthenia were higher with sunitinib, interestingly, AE rates for fatigue were similar in both arms. There is a lack of consensus within the medical community on the precise difference between asthenia and fatigue, which may partially explain this discordance. This apparent inconsistency and the inconsistencies seen between physicians and patients in the rates of other AEs underscore the difficulty in assessing the severity of AEs in isolation, as well as the discordance between CTCAE grading and patient perception of tolerability, which is likely to be a multifaceted experience. Physician preference is an important component of this study, because the physician will have objective knowledge of the efficacy assessments and asymptomatic toxicities that are clinically relevant. These results also favored pazopanib, but slightly fewer physicians preferred pazopanib (61% of physicians; 70% of patients) and more selected no preference (17% of physicians; 8% of patients). This difference may reflect underdetection by the investigator, especially in terms of severity, of a patient’s toxicity burden, giving further importance to the primary end point of this study. To control for bias, physicians were required to state their preference before unblinding. However, awareness of patient preference may have influenced physician preference, and a significant clinical reason would be required for the physician to state an alternate preference to the patient. www.jco.org
Although efficacy remains the primary goal of treatment, assessment of efficacy was limited by the inclusion of patients with nonmeasurable and non– clear cell disease and the relatively short treatment duration. In addition, patients were exposed to both drugs during the study, which makes interpretation of the response data problematic. Details regarding period 1 responses (Appendix Table A4, online only) show a nonstatistically significant imbalance in the rate of PD between these drugs (20% for pazopanib v 11% for sunitinib; P ⫽ .139). However, data from a large randomized study comparing pazopanib versus sunitinib demonstrated comparable efficacy between these agents.2 Strengths of this study include blinding of study treatments and a cross-over design, where patients were exposed to both treatments in similar health states. This allowed for detection of differences in tolerability not confounded by differences in health states and for each patient to act as their own control. In addition, the 2-week washout period and random assignment minimized possible effects of the order of treatment and carryover. Finally, the study used an innovative patient preference questionnaire that was developed with input from leading practitioners in the RCC arena and experts in measuring HRQoL and with input from patients with metastatic RCC. Limitations of this study included the fact that the primary analysis population was 33% smaller than the enrolled population. It was anticipated that a significant proportion of patients would not be eligible for the primary analysis population, which was the only population on which the primary analysis could be appropriately conducted, and the sample size was adjusted accordingly to attempt to account for this issue. However, because of this, it must be acknowledged that the results may not be generalizable to all treatmentnaive patients. Furthermore, in this study, patients were asked to state a preference at one time point only (end of period 2). Because of the intermittent nature of sunitinib administration, this may have resulted in bias favoring pazopanib, because it coincides with the end of the 4-week treatment regimen of the 6-week sunitinib cycle and therefore may not fully capture the variations in patient perceptions of sunitinib toxicity. However, random assignment and the results from the sunitinib-pazopanib arm may address this issue, because when patient preference was not assessed at the peak of © 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
1417
Escudier et al
sunitinib toxicities, twice as many patients expressed a preference for pazopanib (62%) versus sunitinib (32%), suggesting that this was not a cause of significant overall bias. Alternative study design, where preference is assessed at multiple time points, is possible but may introduce a bias of its own. Other limitations of the study include the limited treatment exposure and the requirement to cross over to the second treatment period or come off study if an AE-related dose interruption was required. Thus, the patient experience in this study may not reflect the patient experience in clinical practice, because toxicities may decline over time as a result of patient tolerance and dose modifications.12 This innovative trial suggests that, from a patient’s perspective, and in the context of comparable efficacy, pazopanib is preferred over sunitinib as first-line treatment for metastatic RCC based on the overall impact of treatment-related toxicities and HRQoL. This study also suggests that further research is needed to evaluate the relationship between AEs as reported by physicians using the CTCAE system, HRQoL as reported using patient-reported outcome instruments, and the actual patient experience. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Although all authors completed the disclosure declaration, the following author(s) and/or an author’s immediate family member(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: Sadya Khan, GlaxoSmithKline (C); Jose Diaz, GlaxoSmithKline (C); Suman Redhu, GlaxoSmithKline (C); Faisal Mehmud, GlaxoSmithKline (C) Consultant or Advisory Role: Bernard Escudier, Bayer AG (C), Pfizer (C), Novartis (C); Camillo Porta, GlaxoSmithKline (C), Pfizer (C), Bayer Schering Pharma (C), Astellas Pharma (C), AVEO Pharmaceuticals (C), Novartis (C), Boehringer Ingelheim (C); Petri Bono, GlaxoSmithKline (C), Pfizer (C); Thomas Powles, GlaxoSmithKline (C), Pfizer (C); Tim Eisen, GlaxoSmithKline (C), Pfizer (C); Ju¨rgen E. Gschwend, GlaxoSmithKline
REFERENCES 1. Posadas EM, Figlin RA: Systemic therapy in renal cell carcinoma: Advancing paradigms. Oncology (Williston Park) 26:290-301, 2012 2. Motzer RJ, Hutson TE, Cella D, et al: Pazopanib versus sunitinib in metastatic renal-cell carcinoma. N Engl J Med 369:722-731, 2013 3. Rini BI, Escudier B, Tomczak P, et al: Comparative effectiveness of axitinib versus sorafenib in advanced renal cell carcinoma (AXIS): A randomised phase 3 trial. Lancet 378:1931-1939, 2011 4. Escudier B, Eisen T, Porta C, et al: Renal cell carcinoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 23: vii65-71, 2012 (suppl 7)
(C), Janssen Pharmaceuticals (C), Astellas Pharma (C), Bayer AG (C), Amgen (C), Novartis (C), Roche (C), Pfizer (C); Ugo De Giorgi, GlaxoSmithKline (C), Pfizer (C), Bayer AG (C), Novartis (C); Robert Hawkins, GlaxoSmithKline (C), Pfizer (C), Novartis (C); David Cella, GlaxoSmithKline (C), Pfizer (C), Bayer AG (C), AVEO Pharmaceuticals (C) Stock Ownership: Tim Eisen, AstraZeneca; Sadya Khan, GlaxoSmithKline; Jose Diaz, GlaxoSmithKline; Suman Redhu, GlaxoSmithKline; Faisal Mehmud, GlaxoSmithKline Honoraria: Bernard Escudier, Bayer AG, Roche, Pfizer, Genentech, Novartis, AVEO Pharmaceuticals, GlaxoSmithKline; Camillo Porta, GlaxoSmithKline, Pfizer, Bayer Schering Pharma, Astellas Pharma, Novartis; Petri Bono, GlaxoSmithKline, Pfizer; Thomas Powles, GlaxoSmithKline, Pfizer; Tim Eisen, Pfizer, GlaxoSmithKline; Cora N. Sternberg, GlaxoSmithKline, Pfizer, Novartis, AVEO Pharmaceuticals; Ju¨rgen E. Gschwend, GlaxoSmithKline, Janssen Pharmaceuticals, Astellas Pharma, Bayer AG, Amgen, Novartis, Roche, Pfizer; Omi Parikh, GlaxoSmithKline; Robert Hawkins, GlaxoSmithKline, Pfizer, Novartis; Sylvie Ne´grier, Pfizer, Novartis, Astellas Pharma Research Funding: Camillo Porta, Bayer Schering Pharma, Novartis, Pfizer; Petri Bono, GlaxoSmithKline, Pfizer; Thomas Powles, GlaxoSmithKline, Pfizer; Tim Eisen, Bayer AG, Pfizer, GlaxoSmithKline; Robert Hawkins, Pfizer, GlaxoSmithKline; David Cella, GlaxoSmithKline, Pfizer, AVEO Pharmaceuticals Expert Testimony: Robert Hawkins, GlaxoSmithKline (C) Patents, Royalties, and Licenses: None Other Remuneration: None
AUTHOR CONTRIBUTIONS Conception and design: Bernard Escudier, Petri Bono, Thomas Powles, Tim Eisen, Cora N. Sternberg, Ju¨rgen E. Gschwend, Robert Hawkins, Sadya Khan, Faisal Mehmud, David Cella Financial support: Faisal Mehmud Administrative support: Faisal Mehmud, Sadya Khan Provision of study materials or patients: Bernard Escudier, Camillo Porta, Petri Bono, Cora N. Sternberg, Robert Hawkins, Emmanuel Sevin, Faisal Mehmud Collection and assembly of data: Bernard Escudier, Camillo Porta, Petri Bono, Cora N. Sternberg, Ugo De Giorgi, Omi Parikh, Robert Hawkins, Emmanuel Sevin, Sylvie Ne´grier, Sadya Khan Data analysis and interpretation: Bernard Escudier, Camillo Porta, Petri Bono, Tim Eisen, Robert Hawkins, Jose Diaz, Sadya Khan, Suman Redhu, Faisal Mehmud, David Cella Manuscript writing: All authors Final approval of manuscript: All authors
5. Motzer RJ, Agarwal N, Beard C, et al: Kidney cancer. J Natl Compr Canc Netw 9:960-977, 2011 6. Trotti A, Colevas AD, Setser A, et al: Patientreported outcomes and the evolution of adverse event reporting in oncology. J Clin Oncol 25:51215127, 2007 7. Beaumont JL, Diaz J, Khan S, et al: Validation of the supplementary quality of life questionnaire for mouth/throat and hand/foot soreness. J Clin Oncol 31, 2013 (suppl; abstr 9574) 8. Senn S, Stevens L, Chaturvedi N: Repeated measures in clinical trials: Simple strategies for analysis using summary measures. Stat Med 19: 861-877, 2000 9. Cella D, Eton DT, Lai JS, et al: Combining anchor and distribution-based methods to derive minimal clinically important differences on the Func-
tional Assessment of Cancer Therapy (FACT) anemia and fatigue scales. J Pain Symptom Manage 24:547-561, 2002 10. Yellen SB, Cella DF, Webster K, et al: Measuring fatigue and other anemia-related symptoms with the Functional Assessment of Cancer Therapy (FACT) measurement system. J Pain Symptom Manage 13:63-74, 1997 11. Smith E, Lai JS, Cella D: Building a measure of fatigue: The Functional Assessment of Chronic Illness Therapy Fatigue Scale. PM R 2:359-363, 2010 12. Porta C, Szczylik C, Bracarda S, et al: Shortand long-term safety with sunitinib in an expanded access trial in metastatic renal cell carcinoma (mRCC). J Clin Oncol 26, 2008 (suppl 15s; abstr 5114)
■ ■ ■
1418
© 2014 by American Society of Clinical Oncology
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Pazopanib v Sunitinib Patient Preference in Metastatic RCC
Acknowledgment We thank Jerome F. Sah, PhD, at ProEd Communications for his medical editorial assistance with this article. Appendix
Table A1. Primary Analysis of Patient Preference Treatment Sequence S-P (n ⫽ 60) Analysis Preference Sunitinib Pazopanib No preference Statistical parameter 90% CI for pazopanib preference, % 90% CI for sunitinib preference, % Difference (pazopanib v sunitinib), % 90% CI for difference, % Pⴱ
P-S (n ⫽ 54)
Total (N ⫽ 114)
No. of Patients
%
No. of Patients
%
No. of Patients
%
19 37 4
32 62 7
6 43 5
11 80 9
25 80 9
22 70 8
62.3 to 77.2 15.7 to 29.3 49.26 37.0 to 61.5 ⬍ .001
Abbreviations: P-S, pazopanib followed by sunitinib; S-P, sunitinib followed by pazopanib. ⴱ The two-sided P value is calculated from Prescott’s test.
www.jco.org
© 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Escudier et al
Table A2. Sensitivity Analyses for Patient Preference Preference Analysis
Population
Sensitivity analysis 1, only patients who completed full 10 weeks Pazopanib Sunitinib No preference Difference† Sensitivity analysis 2, includes patients with period 1 PD Pazopanib Sunitinib No preference Difference† Sensitivity analysis 3, individualized imputed preference based on withdrawal reason Pazopanib Sunitinib No preference Difference† Sensitivity analysis 4, worst case: sunitinib preference on all missing questionnaires Pazopanib Sunitinib No preference Difference† Sensitivity analysis 5, worst case: no preference % fixed as in nonmissing; sunitinib preference on remaining missing questionnaires Pazopanib Sunitinib No preference Difference†
ⴱ
Modified ITT
ITT
No. of Patients
No. of Patients
%
90% CI (%)
58 17 5 —
73 21 6
63 to 81 14 to 30 2 to 13 39 to 68
90 26 10 —
71 21 8
64 to 78 15 to 27 4 to 13 39 to 63
94 35 37 —
57 21 22
50 to 63 16 to 27 17 to 28 25 to 46
90 66 10 —
54 40 6
48 to 61 33 to 46 3 to 10 2 to 26
90 63 13 —
54 38 8
48 to 61 32 to 45 5 to 12 4 to 28
⬍ .001
80
⬍ .001
126
Safety
⬍ .001
166
Safety
166
Safety
P
.065
166
.035
Abbreviations: ITT, intent-to-treat; PD, progressive disease. ⴱ Patients who were dosed with both treatments in the same period were excluded from these analysis populations. No imputations were made for their preference. †Approximate 90% CI for the difference in pazopanib v sunitinib preference rates.
Table A3. Patient-Reported Quality-of-Life Measures (safety population) Pazopanib
Sunitinib
Quality-of-Life Measure
Mean Scoreⴱ
No. of Patients
Mean Scoreⴱ
No. of Patients
FACT-F‡ SQLQ Worst mouth and throat soreness Worst hand soreness Worst foot soreness SQLQ limitation scores Mouth and throat soreness Foot soreness
38.1
131
35.6
131
2.49
Treatment Difference†
0.40 0.21 0.36
131 131 129
0.78 0.29 0.52
131 131 129
⫺0.38 ⫺0.08 ⫺0.16
14.32 13.82
126 129
13.72 13.24
126 129
0.60 0.58
90% CI 1.17 to 3.82
P .002
⫺0.47 to ⫺0.29 ⫺0.14 to ⫺0.02 ⫺0.25 to ⫺0.07
⬍ .001 .026 .005
0.34 to 0.85 0.27 to 0.89
⬍ .001 .003
NOTE. Only patients who completed at least one postbaseline assessment in both treatment periods were included in the analysis. Fatigue and the impact of fatigue on life functions were assessed by the FACT-F score; mouth/throat soreness, hand/foot soreness, and work activity were assessed by SQLQ at day 1 and every 2 weeks thereafter. Abbreviations: FACT-F, Functional Assessment of Cancer Therapy–Fatigue; SQLQ, Supplementary Quality of Life Questionnaire. ⴱ Mean score was calculated for all patients as their average postbaseline score within each period averaged over periods; higher scores indicated better health. †The estimated treatment difference CI and P value were adjusted for period and sequence effects in the analysis of variance model. ‡Patients were required to complete assessments in both treatment periods to be included in the analysis.
© 2014 by American Society of Clinical Oncology
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228
Pazopanib v Sunitinib Patient Preference in Metastatic RCC
Table A4. Investigator-Assessed Best Response for End of Period 1 (safety population) Sunitinib (n ⫽ 80)ⴱ Response Best response CR PR Stable disease PD Non-CR/non-PD Not evaluable† Response rate (CR ⫹ PR) 90% CI, %‡ Difference in response rate 90% CI for difference P
Pazopanib (n ⫽ 85)ⴱ
No. of Patients
%
No. of Patients
%
1 16 38 9 4 12 17
1 20 48 11 5 15 21
1 15 32 17 3 17 16
1 18 38 20 4 20 19
14.0 to 30.2
12.2 to 27.2 ⫺2.4 ⫺12.7 to 7.8 .703
Abbreviations: CR, complete response; PD, progressive disease; PR, partial response. ⴱ Patients who had a washout scan after the first dose of treatment in period 2 were excluded. †Not evaluable includes patients who had a missing scan and those with a scan showing stable disease or non-CR/non-PD from less than 49 days after the first dose date. ‡An exact binomial confidence limit method was used for both treatments.
Primary end point (n = 114)
P < .001
Patients who completed full study treatment (10 weeks in both periods, n = 80)
P < .001
All questionnaires (includes patients with Period 1 PD, n = 126)
P < .001
Worst case: imputation of sunitinib preference for study withdrawals (n = 166)
P = .065
0
10
20
30
40
50
60
70
80
90 100
Mean (90% CI) percentage of patients expressing preference for pazopanib Mean (90% CI) percentage of patients expressing preference for sunitinib
Increasing Rate of Preference Fig A1. Sensitivity analyses. Bars represent 90% CIs. PD, progressive disease.
www.jco.org
© 2014 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at NEW YORK UNIVERSITY MED CTR on October 4, 2014 Copyright © 2014 Americanfrom Society of Clinical Oncology. All rights reserved. 128.122.253.228