Changing clinical practice with measurements of portal pressure

The technique for determining the portal vein pressure by measuring the wedged hepatic vein pressure (WHVP) was developed in the 1950s.1 Although the technique was shown to be relatively easy to perform, and it gave accurate estimates of portal pressure in most situations, measurement of WHVP is considered a research tool and it is used in a limited number of clinical situations.2, 3 Three developments have lead to a reevaluation of the utility of WHVP measurement in the management of patients with portal hypertension. The first development is the widespread use of the transjugular intrahepatic portal systemic shunt (TIPS) for the treatment of complications of portal hypertension.4 The experience gained by interventional radiologists in the measurement of portal pressure during the performance of a TIPS has provided training of a large group of individuals in the measurement of portal pressure and its clinical applications. Secondly, and perhaps most importantly, has been the recent observations that the hemodynamic response to pharmacologic therapy correlates with risk of bleeding. If the hepatic venous pressure gradient (HVPG) falls to less than 12 mm Hg or by more than 20%, the risk of bleeding in the responders is significantly less than in the patients who failed to have a meaningful decline in the HVPG.5 Lastly, a number of studies have shown that the higher the level of the HVPG, especially following an acute bleed, the greater the risk of recurrent bleeding.6 Given these developments it is now reasonable to ask whether measurement of the response of the HVPG to therapy or following an episode of variceal bleeding should be used in clinical practice. This is not a trivial issue, as death from bleeding varices remains a significant problem for patients with cirrhosis.7 WHVP, wedged hepatic vein pressure; HVPG, hepatic venous pressure gradient; TIPS, transjugular intrahepatic portosystemic shunt; VBL, variceal band ligation. Primary prophylaxis is the prevention of bleeding from varices in patients who have never bled. Options for the treatment of these patients include pharmacologic and endoscopic therapies.5, 8, 9 Treatment of patients with cirrhosis and varices with pharmacologic therapy reduces the risk of bleeding by about 50% to an incidence of approximately15% over two years. Treatment of patients with large varices increases the benefits of pharmacologic therapy. These observations have led to recommendations that all patients with cirrhosis undergo a screening endoscopy and that patients with large varices be treated with beta-blockers ± nitrates, or with variceal band ligation (VBL) if intolerant of beta-blockers.10, 11 Recently, this approach has been questioned on the basis of cost-effectiveness analysis.12, 13 In the two analyses using the Markov approach, no screening paradigm was found to be cost-effective when compared with treatment of all cirrhotic patients with beta-blockers irrespective of the presence or absence of varices. Only if the risk of bleeding over two years could be reduced to about 5% was the screening approach followed by VBL found to be cost-effective.13 Pharmacologic therapy is not as effective as expected, because approximately half of the patients fail to have a hemodynamic response to therapy. If therapy could be based on hemodynamic response, then efficacy should be improved, and perhaps bleeding rates of 5% or less could be achieved. Merkel et al.14 administered a beta-blocker ± nitrates to 49 cirrhotic patients with varices. They measured the HVPG before and 1 and 3 months after initiation of therapy and followed the patients for up to 5 years. Sixty-one percent of the patients had a hemodynamic response, and the three-year risk of bleeding was 7%. In contrast, the risk of bleeding for the nonresponders was 41%. In a second study, none of 21 patients who achieved a HVPG of 20% fall in pressure also greatly reduces the risk of bleeding. Unfortunately, all but one16 of the aforementioned studies analyzed the data on pressure measurement retrospectively; i.e. the findings were not used to make therapeutic decisions. Thus, it is unknown whether or not the use of the hemodynamic response to make therapeutic decisions will have a positive effect on bleeding rates. This is not a trivial question, as in the study of Bureau et al.,16 treating patients who failed to respond to pharmacologic therapy with VBL had no impact on rebleeding. The situation with secondary prophylaxis is complicated, because the risk of rebleeding is greatest during the first six months following the index bleed.21 Thus, if the measurement of the repeat WHVP is delayed, then many of the patients will bleed before their response to therapy can be determined. For example, in the study by Villanueva et al.,22 only 49 of the 72 patients who received drug therapy had repeat pressure measurements (1–3 months following the initial procedure). The failure to have the second measurement was either because of rebleeding or failure to consent to the second procedure. Thus, if the response to pharmacologic therapy is to be used to determine efficacy, repeat measurements must be obtained as soon as the maximum dose of drug is achieved (2–4 weeks following index measurement). The alternative for this group of patients is to use both pharmacologic and endoscopic therapy and forgo pressure measurements altogether.23 Measuring the HVPG following an acute episode of variceal bleeding may also be useful, as it may predict which patients are likely to rebleed in hospital and therefore should be managed with more definitive therapy such as a TIPS.6, 24 As most patients with an acute variceal bleed are placed in an intensive-care-unit setting where the technology required for obtaining a WHVP is frequently present, this use of pressure measurements is possible. However, controlled trials in which pressure measurements are used to determine therapy need to be performed before this approach can be taken. The most likely situation in which measurement of portal pressure will find the greatest use is in the primary prevention of variceal bleeding. The appeal of determining the response in this group is that the timing of the repeat pressure measurement is not as critical. In addition, many patients will remain on the medication for a prolonged period of time, and therefore one would like to be certain that the therapy is effective, and if not, to apply another form of treatment, the so-called “A La Carte” approach.16 It seems likely that bleeding rates of 5% over two years can be achieved if pharmacologic therapy is limited to those with a hemodynamic response and the nonresponders undergo VBL.14, 25 However, randomized controlled trials are required before this approach can be taken. It is unlikely that the use of WHVP measurements will be cost-effective.26 However, the purpose of the controlled trials will be to establish whether or not this “A La Carte” approach is significantly more effective than the current approach of treating everyone with large varices with beta-blockers ± nitrates.11 Once it has been proven that the “A La Carte” approach is significantly better than the current approach, then the next step is to develop accurate, noninvasive techniques to measure portal vein pressure. Measurement of WHVP is safe and relatively simple. The information obtained may be predictive of new or recurrent bleeding and potentially can help in determining whether or not pharmacologic therapy is effective. However, controlled trials are required before measurement of portal pressure can be used to make decisions in clinical practice. Although new agents are likely to be developed that also reduce portal pressure in a predictive manner,9 it is unlikely that one drug or combination of drugs will be effective in all patients. As with systemic hypertension, we need a safe and accurate noninvasive method for the measurement of portal pressure. Once we can measure portal pressure easily, the management of patients with portal hypertension will be much more straightforward. Until that goal is achieved, the WHVP measurement remains the only way to assess responses to pharmacologic therapy and to develop a tailored approach to prevent variceal bleeding in patients with portal hypertension.