Sorafenib improved OS among patients with hepatocellular carcinoma who were hepatitis C positive but hepatitis B negative, suggesting that the beneficial effects of sorafenib may be based on hepatitis status, according to results of a meta-analysis.
“Irrespective of the mechanism, our data suggest that in future trials in HCC, particularly where sorafenib is the control arm, there should be stratification according to etiology,” Richard Jackson, BSc, MSc, from the Liverpool Cancer Trials Unit of University of Liverpool in United Kingdom, and colleagues wrote. “Etiological differences have already been considered as factors in the interpretation of clinical trials.”
The analysis included 3,256 patients with HCC treated in three randomized phase 3 clinical trials that compared sorafenib (Nexavar, Bayer) with brivanib (BMS-540215, Bristol-Myers Squibb; n = 1,155), sunitinib (Sutent, Pfizer; n = 1,070) or linifanib (ABT-869, Abbott; n = 1,031).
In total, 1,643 patients (50%) received sorafenib as part of a control arm, 577 (18%) received brivanib, 526 (16%) received sunitinib and 510 (16%) received linifanib.
Researchers sought to determine the effect of etiology — based on hepatitis B virus (HBV) and HCV status — on OS.
A smaller proportion of patients in the linifanib trial had HBV (49%) compared with patients in the brivanib trial (58.6%) and sunitinib trial (55.7%). The number of patients with HCV was similar among all three clinical trials (26.9% in brivanib; 22.6% in sunitinib; and 25.1% in linifanib).
A Bayesian hierarchical approach for individual patient data meta-analyses was used with a piecewise exponential model. Researchers evaluated data from 2,863 patients who had defined HBV and HCV status and OS data available.
The median follow-up was 20 months overall, 20.8 months in the brivanib trial, 22.2. months in the sunitinib trial and 15.9 months in the linifanib trial.
Researchers evaluated outcomes in subgroups of four levels of mixed HBV and HCV status:
- HBV and HCV negative (m = 689; median OS, 10.4 months, 95% CI, 9.38-11.46);
- HBV positive and HCV negative (n = 1,474; median OS, 11.41 months; 95% CI, 10.19-12.63);
- HBV negative and HCV positive (n = 628; median OS, 7.76 months; 95% CI, 7.1-8.36); and
- both HBV and HCV positive (n = 72; median OS, 8.5 months; 95% CI, 6.22-10.78).
Patients who were HBV negative and HCV positive were the only group who showed improved OS after sorafenib use (log HR = 0.27; 95% CI, 0.46 to 0.06). The researchers note that the results remained consistent across all trials with heterogeneity assessed using Cochran’s Q statistic (Q = 2.032).
Patients HBV and HCV negative showed some evidence of improved OS after sorafenib, but this was not considered significant (log HR = 0.11; 95% CI, 0.28 to 0.09). Patients HBV and HCV positive had similar magnitude of OS effect as the negative group, but with a much larger CI due to difference in number of patients (log HR = 0.11; 95% CI, 0.45 to 0.25).
Pooled HR data favoring the comparator showed that HBV–positive and HCV–negative patients had no evidence of improved OS after sorafenib (log HR = 0.05; 95% CI, 0.1 to 0.21).
The median unadjusted OS was 12.6 months (95% CI, 11.15-13.8) for sorafenib and 10.2 months (95% CI, 8.88-12.2) for all other therapies. Individual median adjusted OS estimates were 9.67 (95% CI, 8.78-10.65) for brivanib, 8.45 (95% CI, 7.76-9.24) for sunitinib and 9.1 (95% CI, 8.2-9.92) for linifanib.
When comparing sorafenib with all alternative therapies, hierarchical modeling did not produce enough evidence to support an OS benefit associated with sorafenib (log HR = 0.13; 95% CI, 0.38 to 0.07).
“Here, by a meta-analysis, we arrive at an answer to a question that was not considered when the trials were conceived and could not have been answered by any of the trials individually,” the researchers wrote. “It also illustrated that the clinical benefit arising when enlightened pharmaceutical companies are prepared to share their data with an academic unit, even when the primary question does not relate to their products.” – by Melinda Stevens
Disclosures: Jackson reports no relevant financial disclosures. One researcher reports honoraria from Wako Life Sciences, a consultant/advisory role with Merck Sharp & Dohme, and speakers bureau roles with Gilead Sciences, Sirtex-Europe and Wako Life Sciences.