Effects of candesartan and propranolol combination therapy versus propranolol monotherapy in reducing portal hypertension

Background/Aims: Angiotensin receptor blockers (ARBs) inhibit activated hepatic stellate cell contraction and are thought to reduce the dynamic portion of intrahepatic resistance. This study compared the effects of combined treatment using the ARB candesartan and propranolol versus propranolol monotherapy on portal pressure in patients with cirrhosis in a prospective, randomized controlled trial. Methods: Between January 2008 and July 2009, 53 cirrhotic patients with clinically significant portal hypertension were randomized to receive either candesartan and propranolol combination therapy (26 patients) or propranolol monotherapy (27 patients). Before and 3 months after the administration of the planned medication, the hepatic venous pressure gradient (HVPG) was assessed in both groups. The dose of propranolol was subsequently increased from 20 mg bid until the target heart rate was reached, and the candesartan dose was fixed at 8 mg qd. The primary endpoint was the HVPG response rate; patients with an HVPG reduction of >20% of the baseline value or to <12 mmHg were defined as responders. Results: The mean portal pressure declined significantly in both groups, from 16 mmHg (range, 12-28 mmHg) to 13.5 mmHg (range, 6-20 mmHg) in the combination group (P<0.05), and from 17 mmHg (range, 12-27 mmHg) to 14 mmHg (range, 7-25 mmHg) in the propranolol monotherapy group (P<0.05). However, the medication-induced pressure reduction did not differ significantly between the two groups [3.5 mmHg (range, -3-11 mmHg) vs. 3 mmHg (range, -8-10 mmHg), P=0.674]. The response rate (55.6% vs. 61.5%, P=0.435) and the reductions in mean blood pressure or heart rate also did not differ significantly between the combination and monotherapy groups. Conclusions: The addition of candesartan (an ARB) to propranolol confers no benefit relative to classical propranolol monotherapy for the treatment of portal hypertension, and is thus not recommended.