Prognostic Value of Portal Pressure in PBC

In This Article

Abstract and Introduction

Abstract

Background and Aims: In other forms of chronic liver disease, measurement of portal pressure is of prognostic value, but this has not yet been established in primary biliary cholangitis (PBC). The aim of the study is to determine the prognostic value of hepatic venous pressure gradient (HVPG) in relation to liver-related survival outcomes, as well as to the development of hepatic decompensation, oesophageal varices and variceal bleeding.

Methods: Baseline HVPG and liver biopsies were obtained in 86 patients followed for 10 years in a controlled trial of colchicine treatment, and subsequently in a long-term observational cohort study for a further 30 years.

Results: There were 49 Hepatic deaths in addition to 10 Liver Transplants (Hepatic death/transplant; n = 59). Some of these were associated with a significant variceal bleed within 3 months of death or transplant (Portal hypertension-associated death or transplant; n = 19). There were 63 deaths from all causes. During follow-up, oesophageal varices developed in 26 patients, whilst 17 bled from varices and 32 developed hepatic decompensation over a median follow-up of 18.1 years (1.9–28.5). Baseline HVPG was highly predictive of all 6 clinical outcomes and contributed significant predictive information additional to that provided by Mayo score and Ludwig stage.

Conclusion: Measurement of baseline portal pressure is of significant prognostic value in primary biliary cholangitis.

Introduction

Around 40 per cent of patients with PBC, particularly those with elevated serum bilirubin levels and stage 3 or 4 disease histologically, fail to respond to ursodeoxycholic acid (UDCA)^[1] and are at risk of progression to cholestatic liver failure associated with complications of portal hypertension including bleeding oesophageal varices, encephalopathy, hepato-renal syndrome, hepatocellular carcinoma and ultimately, cholestatic liver failure with progressively rising serum bilirubin levels. For such patients, second-line treatments such as obeticholic acid^[2] and fibrates^[3] need to be considered and prognostic models developed and validated in patients treated with UDCA, such as Paris 1^[4] and the UK—PBC risk score^[5] and Globe score,^[6] are not applicable.

Recent studies have clarified the histological correlates and prevalence of portal hypertension in PBC. This commences in the early stages of the disease, long preceding both rises in serum bilirubin and the development of cirrhosis and around 34% of patients with pre-cirrhotic PBC have "high-risk" portal hypertension, defined as an HVPG greater than 12 mmHg.^[7] Portal hypertension is related to cholestasis, interface hepatitis and portal tract and sinusoidal fibrosis and Ludwig stage shows a significant correlation with clinical outcomes such as transplant-free survival and hepatic decompensation.^[8]

Whilst there have been a number of studies documenting the prevalence and development of oesophageal varices and variceal bleeding in PBC, no measurements of portal pressure were reported in these studies.^[9–11] In other forms of chronic liver disease, evidence supporting a correlation between portal pressure, the presence and size of varices and the subsequent occurrence of variceal bleeding is conflicting.^[12,13]

Measurement of the HVPG is a safe and reproducible technique to assess portal pressure, and is the method of choice in PBC.^[7,14] In other forms of chronic liver disease, it is recognised to be of major value in assessment of prognosis, monitoring efficacy of medical treatment and evaluation of progression of portal hypertension^[15–17] but such data for PBC is largely lacking.

The present study aims to evaluate the prognostic value of portal hypertension in a large cohort of PBC patients with long-term follow-up.

Table 1. Characteristics of patients and biopsies at the time of baseline HVPG measurement
Patients	86
Biopsies	85
Age	59.2 (52.2–64.9) [86]
Male	6/86 (7%)
Classification
Presymptomatic	14/86 (16%)
Symptomatic	56/86 (65%)
PBC/AIH overlap	5/86 (6%)
Portal hypertensive PBC	5/86 (6%)
Atypical PBC	6/86 (7%)
Childs class
A	55/83 (66%)
B	23/83 (28%)
C	5/83 (6%)
Mayo Score	5.6 (4.7–6.7) [80]
Ludwig Stage
1	14/85 (16%)
2	15/85 (18%)
3	35/85 (41%)
4	21/85 (25%)
Albumin (g/l)	36.0 (33.0–40.0) [86]
Globulin (g/l)	40.0 (37.0–46.0) [86]
Bilirubin (µmol/l)	19.5 (11.2–38.8) [86]
c ALP (% ULN)	323 (168–458) [85]
ALT (g/l)	67 (42–95) [83]
AST (g/l)	77 (43–116) [59]
AMA titre	300 (300–1000) [80]
IgG (g/l)	17.1 (13.7–22.7) [84]
IgM (g/l)	5.0 (2.9–7.4) [84]
Hb (g/dl)	12.4 (11.3–13.6) [85]
WBC (x10⁹/l)	5.3 (4.0–7.2) [85]
Platelets (x10⁹/l)	231 (137–280) [84]
Prothrombin Index	1.0 (1.0–1.1) [86]

Note: Data represent median (IQR) [N] or number (%).
Abbreviations: cALP (% ULN), Serum Alkaline phosphatase as % upper limit of normal; AMA titre, reciprocal of antimitochondrial antibody titre; PBC, Primary Biliary Cholangitis; PBC/AIH overlap, Primary Biliary Cholangitis/autoimmune hepatitis overlap syndrome; Portal Hypertensive PBC, presentation with bleeding oesophageal varices in the absence of jaundice; Atypical PBC, AMA negative PBC or PBC with an initially normal serum alkaline phosphatase.

Table 2. Numbers of events and patients at risk for each outcome along with the median (inter-quartile range) length of followup
	Pre-baseline events	Never assessed	At risk	Post-baseline events	Followup years Median (IQR)
All deaths	0	0	86	63	13.1 (1.7–28.5)
Hepatic death/transplant	0	0	86	59 (49 deaths, 10 Transplants)	13.1 (1.7–28.5)
PHT-associated death/transplant	0	0	86	19 (18 deaths, 1 Transplant)	13.1 (1.7–28.5)
Decompensation	25	0	61	32	18.1 (1.9–28.5)
Varices	19	6	61	26	15.3 (1.6–28.5)
Variceal bleeds	8	3	75	17	16.4 (1.6–28.5)

Note: PHT associated death—liver-related death associated with bleeding varices within 3 months of death. P.B.C. associated death—Death from cholestatic liver failure or liver-related death associated with bleeding varices within 3 months of death. Excludes liver transplants but includes 3 HCC's and 6 co-morbidities (see text).

Table 3. HVPG as a prognostic factor
Outcome	Unadjusted		Adjusted for Mayo		Adjusted for Mayo and Ludwig
Outcome	HR	P	HR	P	HR	P
All deaths	1.08 (1.03 to 1.14)	0.003	1.02 (0.96 to 1.08)	0.48	1.03 (0.96 to 1.10)	0.40
Hepatic death/transplant	1.11 (1.06 to 1.17)	<0.001	1.07 (1.02 to 1.14)	0.012	1.08 (1.02 to 1.15)	0.011
PHT death/transplant	1.13 (1.02 to 1.25)	0.015	1.12 (0.98 to 1.28)	0.072	1.12 (0.99 to 1.28)	0.058
Decompensation	1.18 (1.10 to 1.28)	<0.001	1.16 (1.06 to 1.27)	<0.001	1.21 (1.07 to 1.36)	0.001
Varices	1.12 (1.06 to 1.19)	<0.001	1.08 (1.02 to 1.14)	0.014	1.08 (1.01 to 1.15)	0.026
Variceal bleed	1.12 (1.02 to 1.23)	0.022	1.07 (0.96 to 1.19)	0.20	1.08 (0.96 to 1.22)	0.19

Note: Cox regression analyses with and without adjustment for Mayo score (continuous variable) and Ludwig stage (categorical). Hazard ratios (HR) and associated 95%CI per mmHg HVPG and P-values. PHT associated death—liver-related death associated with bleeding varices within 3 months of death. P.B.C. associated death—Death from cholestatic liver failure or liver-related death associated with bleeding varices within 3 months of death. Excludes liver transplants but includes 3 HCC's and 6 co-morbidities (see text).