The prognostic value of hepatic venous pressure gradient in patients with cirrhosis is highly dependent on the accuracy of the technique

Hepatic venous pressure gradient (HVPG), the difference between wedged (WHVP) and free hepatic vein pressure (FHVP), predicts survival in patients with cirrhosis. It has been suggested for the use of inferior vena cava (IVC) value instead of FHVP to calculate HVPG when the difference between proximal FHVP (obtained at 2 cm from the hepatic vein outlet) and IVC (measured at the level of the hepatic ostium) is >2 mm Hg. However, there are no data supporting this recommendation. The main aim of the study was to establish which gradient, WHVP-FHVP (HVPG-Free) or WHVP-IVC (HVPG-IVC), better correlates with orthotopic liver transplantation (OLT)-free survival. This work was a retrospective evaluation of hepatic hemodynamic studies of 380 consecutive patients with cirrhosis performed from January 2006 to December 2012 with follow-up until December 2013. Patients had a mean age of 56 +/- 10 years and 64.7% were men. Mean Child-Pugh was 7 +/- 2. HVPG-Free (16 +/- 5 mm Hg) was significantly lower than HVPG-IVC (17 +/- 5.5 mm Hg; P < 0.001). During a mean follow-up of 43 months, 40 patients were transplanted and 111 died. A total of 285 (75%) patients had an FHVP-IVC difference within +/- 2 mm Hg (no discrepancy) and 95 (25%) patients <-2 mm Hg or >2 mm Hg (discrepancy). In patients without discrepancy, 16 mm Hg was the best cut-off value predicting survival, independently of being calculated as HVPG-Free or HVPG-IVC. However, in those patients with discrepancy, 16 mm Hg was still the best cut-off value for HVPG-Free, but not for HVPG-IVC, among which 25 patients (26%) were misclassified regarding their risk of OLT/death. Conclusions: Given that WHVP-FHVP was more accurate in assessing prognosis than WHVP-IVC, HVPG should be calculated as the gradient between WHVP and FHVP, but not with IVC, in order to optimize its prognostic value and in identifying different risk population. (Hepatology 2015;62:1584-1592)