Several reasons may explain that other studies using

Several reasons may explain that other studies using PDE-5-Is yielded negative or contradictory results [30], [31], [32]. Lee [30] applied 50mg sildenafil to seven patients with liver cirrhosis. NO and cGMP in the hepatic CDK4 inhibitor significantly increased, hepatic and pulmonary resistance decreased, and hepatic blood flow increased. These data show the predicted effect of sildenafil on hepatic haemodynamics, except for HVPG. Importantly, MAP remained unchanged. Clemmesen [31] tested 50mg sildenafil in ten patients with liver cirrhosis. While HVPG did not significantly change (from 18 to 16mmHg), in seven patients there was a marked decrease of HVPG. The patients’ details suggest that Clemmesen had included patients with advanced and decompensated cirrhosis. In advanced cirrhosis, the effect of sildenafil on the level of sinusoids may be diminished, while systemic haemodynamic side effects may worsen due to a diminished first pass effect and by collateral blood flow. Data on hepatic blood flow were not given, but mean arterial pressure decreased by 14%. Furthermore, the starting level of 77mmHg MAP was very low, also suggesting that patients with progressed liver cirrhosis were included here. Tandon [32] investigated the acute effect of 25mg sildenafil in 12 patients. Sildenafil did not significantly lower HVPG. However, five of the patients had a portal pressure of <10mmHg, i.e. only mild portal hypertension. 25mg sildenafil is equivalent to about 50mg udenafil and may not be an appropriate dose. Only 75mg and 100mg udenafil effectively decreased HVPG. Mean arterial pressure dropped significantly in the Tandon study from 92mmHg by 7%. Our study differed in some aspects from the studies mentioned above, which might explain our positive results. (1) Only patients with no or mild decompensation of liver cirrhosis were included. (2) Minimum HVPG was 12mmHg. (3) Doppler sonography demonstrated an antegrade portal flow in all patients. (4) We used udenafil and tried to find the effective doses. Therefore, the present study is the first to show unequivocally that a PDE-5-I lowers portal pressure to an extent that may have a beneficial long-term effect. Studies with NSBB compared baseline HVPG to HVPG after a >2 weeks application time. It was derived that a decrease of HVPG by ≥20% from baseline or to ≤12mmHg is the best predictor for a clinically relevant response [19], [33]. Even when applying community succession criterion, 4/10 patient in the 100mg udenafil group (udenafil given for only seven days) fulfilled it. Meanwhile, several studies have shown that a decrease of HVPG by >10% in the acute i.v. treatment with propranolol (0.15mg/kg) followed by a continuous infusion (0.2mg/h) is an adequate indicator for a clinical effect, at least in primary prophylaxis of variceal bleeding [34], [35], [36]. Noteworthy, all studies dealing with PDE-5-I and portal pressure focused on the acute setting [27], [30], [31], [32]. This evaluation eliminates drawbacks of comparing HVPG at the beginning of a study to values after a longer time. HVPG determination is influenced by position of the catheter in the hepatic vein, choice of hepatic vein, sedation or hydration state of the patient, and varying systemic haemodynamics. Therefore, we decided to focus on the acute effects of udenafil. In the combined dosage group, on day 0 (by −20%; p=0.0006) as well as day 6 (by −12%; p=0.042), the acute relative change of HVPG was statistically significant. The treatment interval of 7 days is too short to draw a reliable conclusion on a medium- or long-term effect of udenafil. Previously, we had described that a PDE-5-I increases portal venous flow by approximately 20% [26], [27]. Measurements of liver perfusion, e.g. by indocyanine green clearance, could not be integrated in the study protocol.