We are thankful to Procopet et al. for their comments and for sharing their data with us. We are also happy to see that our results were confirmed by an independent series and found a similar proportion of patients responding to treatment as we did.
In addition, they were able to show a long-term effect of sorafenib on portal hypertension in some patients. We were unable to do so because our patients underwent trans-arterial chemoembolisation after 2 weeks of sorafenib treatment, which precluded us from obtaining unbiased hepatic venous pressure gradient (HVPG)-measurements and survival data thereafter.
We are not surprised to see a lack of correlation between HVPG-response and survival, considering the overall patient number was 18, 11 patients underwent repeat HVPG-measurement only, and 4 of those showed a drop in HVPG. In view of the potential confounders in cirrhotic patients with hepatocellular carcinoma (HCC) with major impact on survival (cirrhosis stage, performance status, tumour size, vascular invasion, extrahepatic spread, among others), it would be really astonishing to find a significant survival difference within such a small patient sample.
However, the findings extend nicely the principle observation from animal experiments to the clinical setting, with about one-third of treated individuals showing an improvement in portal hypertension with sorafenib, an effect of about the same magnitude as can be expected from the treatment of portal hypertensive patients with nonselective β-blockers like propranolol or nadolol. We agree with Procopet et al. that for studying the antihypertensive effects of sorafenib, patients with HCC are not an ideal patient population, but for lack of other options to test the drug so far, both of our groups did resort to this group of patients to perform a proof-of-principle study on the haemodynamic effects of sorafenib in humans.
In our opinion, two scenarios for the use of sorafenib for treatment of portal hypertension could be envisioned: as variceal bleeding is a major cause of death in patients with advanced stage HCC, haemodynamic responders to sorafenib might not need the addition of a nonselective β-blocker for prevention of variceal bleeding, which could help to improve their quality of life significantly. On the other hand, in cirrhotic patients without HCC and nonresponse to nonselective β-blockers, lower dose sorafenib (for safety and tolerability reasons) could be investigated as an alternative medical treatment option (albeit currently not feasible for economic reasons).
We think that in view of the human data presented as well as our previous preclinical data from a model of prehepatic portal hypertension, a pilot trial on the effects of lower dose sorafenib on portal hypertension in patients with cirrhosis would be reasonable, in particular, in patients not responding to nonselective β-blockers. Prospective randomised trials do not seem to be warranted to date as long as we do not have data on the response in this selected and difficult to treat patient group.