BACKGROUND: Pazopanib is an oral tyrosine kinase inhibitor with demonstrated efficacy and tolerability in patients with advanced renal cell carcinoma (RCC). OBJECTIVE: To examine pazopanib persistence and compliance (adherence) and other drug utilization patterns in both treatment-naïve (first-line) patients and those previously treated with RCC therapy in the real-world setting. Key factors affecting persistence and compliance were also explored.
METHODS: This was a retrospective claims analysis using the Truven Health MarketScan Databases to cover claims activity from October 2007 through March 2012. Patients with advanced RCC aged ≥ 18 years who had received pazopanib with 180 days of follow-up were included. Bivariate comparisons of results from first-line and previously treated patients with RCC were conducted. Pazopanib persistence was measured using (a) estimated level of persistence with therapy (ELPT; i.e., the percentage of patients remaining on therapy at 30, 60, and 90 days [patients were censored at 180 days]); (b) time to discontinuation (i.e., duration of therapy); and (c) proportion of days covered (PDC; i.e., the ratio of [total days drug available minus days' supply of last prescription] to [last prescription date minus first prescription date]). Compliance was measured by medication possession ratio (MPR; i.e., the ratio of [total days' supply minus days' supply of last prescription] to [last prescription date minus first prescription date]). Other drug utilization measures included days' supply, time to initiation, time to switching, and dose-related measures. Random forest models were used to explore key factors of pazopanib persistence and compliance.
RESULTS: A total of 143 patients met all inclusion criteria; 43.3% were treated with pazopanib first line (first-line cohort), and 56.6% had ≥ 1 prior lines of therapy (previously treated cohort). The mean (± standard deviation [SD]) age of patients was 62.9 (± 10.3) years, and 71.3% of them were males. Continuous pazopanib therapy for up to 90 days was observed in greater than 50% of patients in both cohorts. In the first-line cohort, ELPT at 30, 60, and 90 days was 98.39%, 70.97%, and 56.45%, respectively; the mean (± SD) number of days to discontinuation was 112.2 (± 62.8); the mean (± SD) PDC was 84.7% (± 16.7%); and the mean (± SD) MPR was 85.2% (± 16.9%). Similar results were observed in the previously treated population: ELPT at 30, 60, and 90 days was 98.77%, 75.31%, and 58.02%, respectively; the mean (± SD) number of days to discontinuation was 118.7 (± 61.4); the mean (± SD) PDC was 87.8% (± 13.5%); and the mean (± SD) MPR was 90.1% (± 13.9%). Differences between the 2 cohorts were not statistically significant. More than 90% of patients in both cohorts had at least a 30-day supply of therapy (91.9% of first-line versus 90.2% of previously treated; P = 0.153). The mean (± SD) time from metastatic diagnosis to start of pazopanib therapy was 104.7 (± 199.3) days in the first-line cohort and 360.9 (± 187.0) days in previously treated patients (P = 0.001). Forty-six patients switched to another therapy: 17 patients in the first-line cohort and 29 patients in the previously treated cohort; the mean (± SD) time to switching therapy from each cohort was 94.7 (± 41.4) days and 87.8 (± 49.6) days (P = 0.146), respectively. Statistically significant differences were observed for the starting and ending doses between the 2 cohorts. The average daily dosage of pazopanib was greater than 700 mg in both cohorts (P = 0.055), with a maximum dose of 800 mg. Random forest models demonstrated that younger age and higher comorbidity predicted both higher persistence and compliance.
CONCLUSIONS: In this observational study, greater than 50% of patients with advanced RCC were on pazopanib for almost 4 months, with the majority of both cohorts achieving high persistence and high compliance. Additionally, younger age and higher comorbidity index were the strongest predictors of both greater persistence and compliance. Further studies with larger cohorts and longer follow-up are needed to validate these findings.
