Sumitra Shantakumar1, Samantha St. Laurent2, Myrthe van Herk-Sukel3, Jeanenne J. Nelson4
Affiliations: 1GlaxoSmithKline, Worldwide Epidemiology, Research Triangle Park, NC, USA; 2GlaxoSmithKline, Worldwide Epidemiology, Waltham, MA, USA; 3PHARMO Institute for Outcomes Research, Utrecht, Netherlands; 4GlaxoSmithKline, Collegeville Road, Collegeville, Pennsylvania, 19426-0989, United States
Pazopanib is an angiogenesis inhibitor approved by the FDA in 2009 for metastatic renal cell carcinoma (RCC). This observational study describes the frequency of cardiovascular outcomes in adult RCC patients treated with pazopanib. A retrospective cohort study was conducted using data from the United States (Clinformatics™ DataMart Multiplan, IMPACT) and The Netherlands (PHARMO Database Network). The following outcomes were captured via ICD-9 diagnostic codes: myocardial infarction (MI), unstable angina (UA), Torsades de Pointes (TP), cerebrovascular accident (CVA), and transient ischemic attack (TIA). Incidence proportions were evaluated in pazopanib users and a comparator group of RCC patients treated with bevacizumab, sorafenib, or sunitinib. When multiple events occurred for the same cardiovascular outcome in the same patient, only the first event was used. Among 104 RCC patients treated with only pazopanib in the US population, 13 events occurred among 10 patients, as follows: five CVA [4.8% (95% CI: 1.6-10.9%)], three events each of MI and TIA [2.9%, (95% CI: 0.6-8.20%], and one event each of TP and UA [1.0% (95% CI: 0.02-5.2%)]. In comparison, the frequency of events among the comparator group (n=556) were: CVA (2.2%), MI (4.5%), TIA (1.4%), TP (1.6%), UA (1.6%). In the Dutch population, 21 pazopanib users and 96 patients in the comparator group met the eligibility criteria; one pazopanib patient experienced a CVA event and one sunitinib user had an UA event. The frequency of selected cardiovascular events among RCC patients was similar for users of pazopanib and similar drugs.
Clinical Practice Points
As part of pharmacovigilance activities, we evaluated ‘real world’ rates of cardiovascular events in RCC patients treated with pazopanib and other marketed anti-VEGF agents.
Based on data from a large administrative claims database in the United States and hospital records in The Netherlands, cardiovascular ischemic events occurred in 0-9.5% of RCC patients treated with pazopanib and other anti-VEGF agents over a mean of 379 days of follow-up.
Pazopanib is an angiogenesis inhibitor approved in 2009 by the United States Food and Drug Administration (FDA) for treatment of metastatic renal cell carcinoma (RCC). The safety profiles of other marketed anti-vascular endothelial growth factor (anti-VEGF) drugs at the time of initial pazopanib approval suggested an increased occurrence of cardiovascular ischemic events and Torsades de pointes (TP).1-6 As pazopanib is used to treat highly vascularized tumors, we designed a descriptive study as part of pharmacovigilance and post-marketing drug safety evaluations in the post-approval period to evaluate the occurrence of cardiac events and TP in RCC patients treated with pazopanib.
The primary objective of this study was to estimate the frequency of specific cardiovascular events and TP after drug initiation in adult (18+ years) pazopanib users with RCC and in a comparator group of RCC patients using three other marketed anti-VEGF agents at the time of pazopanib FDA approval in 2009: bevacizumab, sorafenib, and sunitinib. A comparator group allowed for descriptive evaluation of three other drugs in the anti-VEGF class in order to place the pazopanib-specific results in context.
Patients and Methods
A descriptive, retrospective cohort study was conducted using two de-identified data sources: a US healthcare claims database, the Clinformatics™ DataMart Multiplan (IMPACT), a product of OptumInsight Life Sciences Inc., and the PHARMO Database Network in The Netherlands. The IMPACT system is a comprehensive, de-identified inpatient and outpatient health insurance claims database with more than 107 million non-elderly, insurance-carrying patients participating in 46 different healthcare plans, serving members across nine census regions in the US. The PHARMO Database Network is a population-based medical record tracking system covering 65 municipal areas with 3.2 million community-dwelling inhabitants of the Netherlands. The two specific PHARMO databases linked for this study were the Dutch National Medical Register that reflects all hospital admissions, and the outpatient pharmacy database containing drug dispensing information from community pharmacies.
Inclusion criteria for the anti-VEGF cohorts were tailored to each data source due to differences in database structure. For IMPACT, RCC patients were included if prescribed or administered pazopanib or one of the other specified anti-VEGF agents on or after October 1, 2009, the FDA approval date for pazopanib in the US Drugs were identified by text string and/or National Drug Codes. A one year screening window from October 1, 2008 until September 30, 2009 was used to identify any anti-VEGF drugs used prior to the study start date. The study index date was the first administration or prescription for the only or first anti-VEGF agent after October 1, 2009. Identified using an ICD-9 diagnostic code (189.0), the RCC diagnosis had to occur October 1, 2008 or later and must have occurred no later than 30 days after the anti-VEGF index date.
Patients using only one anti-VEGF drug during the study period were examined separately from those with more than one drug, as were patients using an anti-VEGF drug in a first-line (1L) setting versus a subsequent line of treatment. A first-line therapy was defined as drug treatment received before another anti-VEGF drug. Second-line or higher therapies are those received following another anti-VEGF drug. For patients with multiple anti-VEGF use, outcomes are listed by the therapy that was taken closest prior to or at the time of the cardiovascular event.
Patients were followed from the index date to the earliest event for each cardiovascular outcome, death, or until the end of study follow-up, September 30, 2012, whichever was reached first. Patients with another primary cancer diagnosis in the five years prior to the index date were excluded and continuous IMPACT enrolment with pharmacy benefits since October 1, 2009 was required for study inclusion. Patients with a history of cardiovascular events prior to the index date were not omitted.
For the PHARMO Database Network, RCC patients were included if an anti-VEGF had been dispensed or administered on or after February 1, 2011 (date of first pazopanib prescription in The Netherlands). The first dispensing or administration within this period was defined as the index date; drugs were identified by ATC codes (pazopanib, L01XE11; sorafenib, L01XE05; sunitinib, L01XE04). Hospitalization for RCC was used as a proxy for RCC diagnosis, based on primary or secondary hospital discharge diagnoses codes occurring at any time before or after the index date. Patients were excluded if diagnosed with a secondary cancer between the RCC hospitalization and the index date or if less than one year of medical history data prior to the index date was available. RCC was not an approved indication for bevacizumab in the Netherlands at the time of pazopanib approval, and hence was not assessed with the PHARMO data. Primary hospitalization codes were used to capture the cardiovascular outcomes; secondary codes were not included as they may relate to prevalent disease. Patients were followed until the date of cardiovascular event, death, end of the study period (December 31, 2012), or end of follow-up in the database (either by moving outside the PHARMO catchment area or end of data collection in the pharmacy), whichever occurred first.
The following cardiovascular outcomes were captured via ICD-9 diagnostic codes, as follows: myocardial infarction (MI), 410.9; unstable angina (UA), 411.1; Torsades de Pointes (TP), 426.82; cerebrovascular accident (CVA), 434.91; and transient ischemic attack (TIA), 435.9. Separate incidence proportions (cumulative incidence) and corresponding 95% confidence intervals (CI) were calculated for each type of cardiovascular event in the pazopanib and comparator groups. Each CI was calculated using the exact binomial method. When multiple events for a patient occurred for the same cardiovascular outcome, only the first event was used for the incidence analyses.
The study population was 71% male and the median age was 61 years. Among pazopanib users, 40% (4/10) of patients who experienced an outcome during the study period had a history of cardiovascular events in the year prior to index date (Table 1).
No bevacizumab-only patient with cardiovascular events occurring during the study period experienced cardiovascular events in the prior year. However, 38% (5/13) of sorafenib-only and 14% (4/29) of sunitinib-only patients with cardiovascular events during the study period also had a cardiovascular event in the year prior to the index date.
In IMPACT, a total of 226 pazopanib users and 610 patients treated with the other anti-VEGF agents met the eligibility criteria. The mean and median follow-up time was 413 and 347 days, respectively, for pazopanib patients and 366 and 268 days, respectively, for patients treated with the other agents. Among 104 RCC patients treated solely with pazopanib following the index date (Table 2), 13 total events occurred among 10 patients as follows: five CVA events [4.8% (95% CI: 1.6-10.9%)], three events each of MI and TIA [2.9%, (95% CI: 0.6-8.20%], and one event each of TP and UA [1.0% (95% CI: 0.02-5.2%)]. The median time to event among pazopanib-only patients was shortest for the five CVA events (29 days) and longest for patients experiencing a TP event (580 days).
In comparison, the incidence proportions for cardiovascular events among the comparator group with users of one other anti-VEGF drug (n=556) were: CVA [2.2% (95% CI: 1.1-3.7%)], MI [4.5% (95% CI: 2.9-6.6%)], TIA [1.4% (95% CI: 0.6-2.8%)], TP [1.6% (95% CI: 0.7-3.1%)], UA [1.6% (95% CI: 0.7-3.1%)]. Among the comparator group using only one drug, the shortest median time to event was observed among patients experiencing a TIA event (96 days); the longest median time to event was found for UA events (337 days). The majority (69%) of the comparator group was comprised of sunitinib-only users and most cardiovascular events (38/63, 60%) were observed among patients treated with sunitinib only. However, the highest incidence proportion observed among non-pazopanib users with only one line of therapy was found for MI events in patients treated only with sorafenib [8/84, 9.5% (95% CI: 4.2-17.9%)].
The incidence proportions for cardiovascular events occurring in RCC patients treated with multiple anti-VEGF therapies are shown in Table 3.
Events are listed by the anti-VEGF therapy that was taken closest prior to or at the time of the cardiovascular event; the denominator is comprised of all subjects who received each therapy either prior to another therapy (first line) or following another therapy (second line or higher). Among first line pazopanib patients (n=22), no events occurred. Nine events were observed among patients treated with pazopanib in the second line or higher setting (n=100), 5 CVA events [5% (95% CI: 1.6-11.3%)] and one event each for MI, TIA, TP and UA [1% (95% CI: 0.03-5.5%)]. Results for the other anti-VEGF agents used in the first or second line setting were similar.
In the PHARMO Database Network, a total of 21 pazopanib users, 6 sorafenib users, and 90 sunitinib users met the eligibility criteria. One pazopanib patient had a CVA event and one sunitinib user experienced a UA event; no other outcomes were observed during the study period [data not shown]. These patient and event numbers were too low to perform incidence analyses.
The analyses from this descriptive study indicate that the cardiovascular events under investigation occurred in 0-9.5% of RCC patients treated with pazopanib or other anti-VEGF agents. To our knowledge, population-based estimates for these outcomes in pazopanib users do not exist. Cardiac or cerebral ischemic events were rare, but occurred in clinical trials of pazopanib among patients with advanced RCC.7-9 In the pivotal phase III trial, arterial thrombotic events consisting of MI, CVA or TIA occurred in 3% of patients receiving pazopanib compared to none in the placebo arm.10 A recent meta-analysis of arterial thrombotic events in patients from 19 randomized controlled trials of cancer patients treated with anti-VEGF drugs, including pazopanib, reported an overall incidence proportion of 1.5%; cardiac infarction/ischemia was the most common event type.11
Considering the low survival rates associated with advanced metastatic RCC, it is possible that mortality is a competing risk with non-fatal events observed only among survivors. This study is primarily focused on pazopanib users; data for the three other drug groups are included to provide a general pattern and context for the pazopanib results. This study was not designed to estimate head-to-head comparisons of event rates between drugs. Thus, no formal drug-to-drug statistical analyses were conducted.
This study was limited by the small number of subjects in several analytic groups, including the line of therapy analyses, resulting in some imprecise estimates. The estimates of effect presented are proportions; thus, assessments accounting for differing time on the prescribed agents cannot be made. Further, the IMPACT claims database lacks key cardiovascular risk factor information such as smoking, alcohol use, obesity, ethnicity, or family history. The elderly population are under-represented in the U.S. data as their healthcare is covered by Medicare. In PHARMO, out-patient pharmacy data were used to define drug exposure. As a consequence, patients starting oral anti-VEGFs during their hospital stay would not have inpatient initiation captured, thus truncating duration for the anti-VEGF drug. Further, the use of hospitalization data to identify cardiovascular outcomes may have missed events treated in an outpatient setting (such as TIA) or patients who died prior to hospitalization.
A strength of this observational study is the focus on patients from two real-world population-based settings, representing drug use in the broader general population compared to clinical trial patients, where a large number of inclusion and exclusion criteria create a more select population. Cardiovascular and cerebrovascular outcomes are well captured via ICD-9 diagnosis codes in administrative data.12,13 The IMPACT database includes a large number of patients and covers a wide variety of geographic areas.
In conclusion, the frequency of selected cardiovascular events among population-based RCC patients was similar for users of pazopanib and other anti-VEGF drugs in this study. Results from clinical trials suggest that anti-VEGF treatments may be associated with an increased risk of developing cardiovascular events among cancer patients.11 More observational epidemiologic studies of are needed to better understand this relationship in the general cancer population.
Page Abrahamson provided manuscript writing and editing support. This study was sponsored by GlaxoSmithKline; pazopanib is an asset of Novartis AG as of March 1, 2015. GlaxoSmithKline provided financial support for the conduct of the research and preparation of the article, including the collection, analysis and interpretation of data and writing of the report.
Conflict of Interest
SS, SSL, and JJN are current employees of GlaxoSmithKline; they also hold GSK shares. MvHS is an employee of the PHARMO Institute. This independent research institute performs financially supported studies for government and related healthcare authorities and several pharmaceutical companies.
SS and JJN were responsible for the study design and drafting of the manuscript. SSL and MvHS were responsible for data acquisition and statistical analysis. All authors were involved with data interpretation and critical revision of the manuscript content.
This study was sponsored by GlaxoSmithKline; pazopanib is an asset of Novartis AG as of March 1, 2015.
Role of the Funding Source
GlaxoSmithKline provided financial support for the conduct of the research and preparation of the article, including the collection, analysis and interpretation of data and writing of the report.
Ethical Approvals and Informed Consent
IMPACT data were anonym-ized and compliant with U.S. Health Insurance Portability and Accountability Act (HIPAA) guidelines; ethical board review was obtained at PHARMO.
1. Escudier B, Eisen T, Stadler WM, et al. Sorafenib in advanced clear-cell renal-cell carcinoma. N Engl J Med. 2007; 356:125-134.
2. Chen HX, Mooney M, Boron M, et al. Phase II multicenter trial of bevacizumab plus fluorouracil and leucovorin in patients with advanced refractory colorectal cancer: An NCI Treatment Referral Center trial TRC-0301. J Clin Oncol. 2006; 24:3354-3360.
3. Motzer RJ, Michaelson MD, Redman BG, et al. Activity of SU11248, a multitargeted inhibitor of vascular endothelial growth factor receptor and platelet-derived growth factor receptor, in patients with metastatic renal cell carcinoma. J Clin Oncol. 2006; 24:16-24.
4. Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer. N Engl J Med. 2006; 355:2542-2550.
5. Shah MA, Ilson D, Kelsen DP. Thromboembolic events in gastric cancer: High incidence in patients receiving irinotecan- and bevacizumab-based therapy. J Clin Oncol. 2005; 23:2574-2576.
6. Johnson DH, Fehrenbacher L, Novotny WF, et al. Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non-small-cell lung cancer. J Clin Oncol. 2004; 22:2184-2191.
7. Motzer RJ, Hutson TE, Cella D, et al. Pazopanib versus Sunitinib in Metastatic Renal-Cell Carcinoma. N Engl J Med. 2013; 369:722-731.
8. Sternberg CN, Davis ID, Mardiak J, et al. Pazopanib in locally advanced or metastatic renal cell carcinoma: results of a randomized phase III trial. J Clin Oncol. 2010:28:1061-1068.
9. Sternberg CN, Hawkins RE, Wagstaff J, et al. A randomized, double-blind phase III study of pazopanib in patients with advanced and/or metastatic renal cell carcinoma: final overall survival results and safety update. Eur J Can. 2013; 49:1287-1296.
10. McCormack PL. Pazopanib: a review of its use in the management of advanced renal cell carcinoma. Drugs. 2014; 74:1111-1125.
11. Qi WX, Shen Z, Tang LN, Yao Y. Risk of arterial thromboembolic events with vascular endothelial growth factor tyrosine kinase inhibitors: an up-to-date meta-analysis. Crit Rev Oncol Hematol. 2014; 92:71-82.
12. McCormick N, Lacaille D, Bhole V, et al. Validity of Myocardial Infarction Diagnoses in Administrative Databases: A Systematic Review. PLoS ONE. 2014; 9(3):e92286.
13. McCormick N, Bhole V, Lacaille D, et al. Validity of Diagnostic Codes for Acute Stroke in Administrative Databases: A Systematic Review. PLoS ONE. 2015; 10(8):e0135834. KCJ
Keywords: cardiovascular, ischemia, pazopanib, renal cell carcinoma, Torsades de Pointes
Corresponding Author: Sumitra Shantakumar, PhD; Worldwide Epidemiology, Glaxosmithkline R&D; 150 Beach Road, Gateway West, #26-00; Singapore 189720, SINGAPORE Tel: +065-6232-8258; E-mail: email@example.com