Monitoring of HVPG during pharmacological therapy: Evidence in favor of the prognostic value of a 20% reduction

Abstract

An objective of our commissioned paper was to stimulate debate on monitoring hepatic venous pressure gradient (HVPG) following variceal bleeding. We stress that our critique pertains to the clinical applicability of HVPG measurements based on current evidence, and not on the pathophysiological importance of a reduced portal pressure. Whilst we do not disagree with the rebleeding rates in responders with a ≥20% reduction in HVPG but without a decrease to ≤12 mmHg, we have emphasized,1 both in response to their review2 and in our peer-reviewed article,3 the need to separate primary and secondary prevention due to very different bleeding risks.4 The secondary prevention studies5-9 are a homogeneous cohort (risk of [re]bleeding is highest, these HVPG measurements should be most useful). However, many patients were excluded, not having baseline or repeat HVPG measurements; in some studies5 rebleeding was worse in responders than in nonresponders, and some patients rebled early, before remeasurement,up to 22%.9 These factors make it difficult to evaluate the utility of monitoring the reduction of HVPG in all the population at risk. The heterogeneity extends to the proportion of Child-Turcotte-Pugh C patients (6%-47%)5, 9 defining in part the risk of early rebleeding. In some studies,5, 10 rebleeding is only 7% or 8% within 3 months. The clinical applicability of HVPG measurements is dependent on the measurement's timing, because if performed too late, many patients will have rebled, whereas if performed early on, it would capture most patients at risk. Our colleagues accept this view, proposing a second HVPG measurement at 2 weeks,1 although there is no study that confirms this interval; they themselves have data based only on a median of 3 months.5, 10 We agree our study8 is “anomalous,” having the longest interval to remeasurement (mean, 5.3 months) and having many alcoholics(70%); some patients abstained, which this may account for the many responders due to improvement of liver function. Seven patients8 had HVPG measurements after having rebled, comparable to 5 patients in the study of Feu et al.,5 and we noted this fact.3 We agree there is little variability with HVPG with correct measurements of HVPG; but at lower HVPG baseline values, measurement errors are greatest, in percentage terms. We found that baseline HVPG and rebleeding were correlated9; this fact was not commented on by others and would benefit from prospective evaluation. The ability to obtain a reduction of HVPG by 20% or more may reflect a lower baseline HVPG. Unfortunately, this effect cannot be deduced from published studies, but it is a plausible scenario. In simple terms, if HVPG is reduced by 20%, from 28 to 22.3 mmHg, does this confer the same protection against rebleeding as does a 20% reduction, from 19 to 15.2 mmHg? A suggestion that baseline HVPG is important is shown in the second Villanueva study7 in which drugs were beneficial solely in Child-Turcotte-Pugh A patients, whilst in the first study,6 in which mean baseline HVPG was lower (17.7 ± 3.4 vs. 19.9 + 3.5 mm Hg), rebleeding was reduced in all Child-Turcotte-Pugh classes. The relationship between baseline and repeated HVPG, and severity of liver disease (potentially improving or worsening with time) should also be prospectively studied. The improvement in liver function or its absence was not commented on by Abraldes et al.10 in relation to reduction of HVPG. Extending HVPG measurement to assess progression, or regression with therapy, of chronic hepatitis C is one of our proposals,11 but it is separate from the clinical applicability of HVPG measurements for preventing variceal bleeding. We hold the view that new prospective HVPG monitoring studies are needed for prevention of variceal bleeding. New studies could justify the extra resources and training required and prove the cost-effectiveness of implementing routine HVPG monitoring compared to empirical use of propranolol—a very cheap and simple management strategy. Andrew K Burroughs*, Ulrich Thalheimer*, Maria Mela*, David Patch*, * Hepatobiliary Medicine & Liver Transplant, Royal Free Hospital, London, UK.