Frances Sutcliffe

NICE guidance on sorafenib for treating advanced hepatocellular carcinoma

On Sept 6, 2017, the UK National Institute of Health and Care Excellence (NICE) published guidance that recommends sorafenib (Bayer PLC; Reading, UK), as an option for treating advanced hepatocellular carcinoma only for people with Child-Pugh grade A liver impairment, and only if the manufacturing company provides sorafenib within the agreed commercial access arrangement. The appraisal committee’s remit was to assess the clinical and cost effectiveness of sorafenib for advanced hepatocellular carcinoma compared with standard National Health Service (NHS) care. NICE appraised sorafenib, an oral multikinase inhibitor, first in 2009, and then again in 2016 when reconsidering drugs paid for by the Cancer Drugs Fund in England. When NICE originally appraised sorafenib, the committee met four times to review the evidence provided by Bayer, the manufacturer, and the critique of Bayer’s submission by the West Midlands Health Technology Assessment Collaboration, the evidence review group (ERG). The evidence centred around one trial, SHARP, designed to compare 400 mg of sorafenib twice daily with best supportive care against placebo with best supportive care in patients with Child-Pugh liver function grade A, with the primary endpoints being time to death or to symptomatic progression. The committee noted that SHARP had shown that sorafenib increased median survival by more than 2·8 months compared with placebo, yet, at the price the company had set, sorafenib was not a good use of limited NHS resources. Thus, the committee did not recommend sorafenib. Bayer appealed the decision, and an appeal panel dismissed all the points. Because NICE issued final guidance not recommending sorafenib, it was provided in England by the Cancer Drug Fund rather than by routine NHS commissioning. To permit the committee to reconsider sorafenib, Bayer submitted additional evidence to support its clinical and cost effectiveness. The committee met three times accompanied by clinical experts, representatives from Bayer, the ERG (NICE Decision Support Unit, University of Sheffield) and the public. The committee concluded that doxorubicin, local resection, ablation using radiofrequency, and chemoembolisation were not usual treatment for first-line treatment of advanced hepatocellular carcinoma, and that best supportive care was still the only relevant treatment with which to compare sorafenib. The committee recognised that between 2010 and 2016, there were no new comparators available for first-line treatment of hepatocellular carcinoma in the NHS, no new randomised controlled trial comparing sorafenib with best supportive care, and no further analyses of SHARP reflecting longer follow-up. For the committee’s first meeting, Bayer submitted: a Commercial Medicines Unit price lower than the price in the original appraisal; data from two observational studies used to validate the company’s extrapolation of overall survival beyond the end of SHARP (GIDEON, in which patient characteristics were unmatched to the characteristics of the SHARP participants, and a study by Palmer and colleagues, in which patient characteristics were also unmatched); an estimate of the duration of treatment from SHARP based on time to disease progression; costs based on the committee’s preferences from the original appraisal; and updated estimates of resource use costs. The committee discussed Palmer and colleagues’ study, recognising it as a small unpublished retrospective UK study comparing patients with hepatocellular carcinoma who were funded for (and received) sorafenib with those who did not receive funding or sorafenib, and considered that the study’s reported association between sorafenib funding and outcomes may have been confounded. The committee noted that the population in GIDEON, a Bayer-sponsored multinational post-marketing uncontrolled safety study of over 3000 people, differed from that in SHARP; for example, only 62% of participants in GIDEON had Child-Pugh grade A liver function at baseline whereas nearly all did in SHARP. The committee stated that it would have been appropriate for the company to modify the GIDEON population to reflect the characteristics of the population enrolled into SHARP. For the committee’s second meeting, Bayer’s submission included: a price for sorafenib lower than that provided at the first meeting; evidence from GIDEON, now matched to the SHARP population, using propensity scores for patient characteristics that might influence mortality; and an estimated duration of treatment using data on time-to-treatment discontinuation from SHARP. At the committee’s third meeting, Bayer’s submission included: a price for sorafenib lower than that provided at the second meeting; UK audit data for treatment for hepatocellular carcinoma; and an estimated duration of treatment using data on time-to-treatment discontinuation from GIDEON matched to the SHARP population. In discussing the appropriate patients for sorafenib, the committee noted that people with Child-Pugh grade A liver function were the only patients specified in the inclusion criteria for SHARP, comprised the majority of patients contributing Lancet Gastroenterol Hepatol 2017