ArticlePDF Available

The prognostic value of HVPG-response to non-selective beta-blockers in patients with NASH cirrhosis and varices

Abstract and Figures

Background and Aims Non-alcoholic steatohepatitis has become a leading cause of cirrhosis. The prognostic value of (HVPG)-guided NSBB prophylaxis remains to be investigated in the setting of NASH cirrhosis. Methods Patients with NASH cirrhosis and varices undergoing HVPG-guided NSBB therapy were included. HVPG-response to NSBBs was evaluated within a median 52 (IQR:28–71) days after baseline measurement. HVPG-Response was defined as a decrease of ≥10% from baseline or below <12 mmHg. The composite endpoint was defined as variceal bleeding, decompensation, and liver-related death. Results Thirtyeight patients were included: Child-A/B:33(87%), Child-C:5(13%) median HVPG:19.7 ± 4.7 mmHg. 21(55.3%) patients achieved HVPG-response to NSBB. Presence of diabetes(aOR:0.16, p = 0.038) and arterial blood pressure (aOR:1.07, p = 0.044) were independently associated with NSBB-response. While NSBB-HVPG-responders showed fewer decompensations within 90 days (n = 1(5%) vs. n = 3(29%), p = 0.172), only Child-Pugh stage B/C (p = 0.001), MELD ≥ 15(p = 0.021) and HVPG ≥ 20 mmHg(p = 0.011) predicted the composite endpoint at 90 days. Similarly, after 2years of follow-up, only Child-Pugh stage (B:p = 0.001, C:p < 0.001), MELD ≥ 15 (p = 0.021), HVPG≥20 mmHg (p = 0.011) predicted the composite endpoint. Importantly, all bleeding events occurred in HVPG-NSBB non-responders. Conclusion HVPG-response to NSBB was achieved in 55.3% of NASH patients with varices and this seemed to protect from variceal bleeding. However, only baseline HVPG ≥ 20 mmHg, Child-Pugh stage B/C and MELD ≥ 15 were predictors of decompensation/death in patients with NASH cirrhosis and varices.
Content may be subject to copyright.
JID: YDLD [m5G; October 4, 2021;11:58 ]
Digestive and Liver Disease xxx (xxxx) xxx
Contents lists available at ScienceDirect
Digestive and Liver Disease
journal homepage:
Liver, Pancreas and Biliary Tract
The prognostic value of HVPG-response to non-selective beta-blockers
in patients with NASH cirrhosis and varices
, ✩✩
Rafael Paternostro
a , b
, Jeannette Becker
a , b
, Benedikt Silvester Hofer
a , b , c
, Vera Panagl
a , b
Helena Schiffke
a , b
, Benedikt Simbrunner
a , b , c
, Georg Semmler
a , b
, Philipp Schwabl
a , b , c
Bernhard Scheiner
a , b
, Theresa Bucsics
a , b
, David Bauer
a , b , c
, Teresa Binter
a , b
Michael Trauner
, Mattias Mandorfer
a , b
, Thomas Reiberger
a , b , c ,
Divison of Gastroenterolo gy and Hepatology, Department of Medicine III, Medical University of Vienna, Waehringer Guertel 18-20, Vienna A-1090, Austria
Vienna Hepatic Hemodynamic Lab, Medical University of Vienna, Vienna, Austri a
Christian Doppler Laboratory for Portal Hypertension and Liver Fibrosis, Medical University of Vienna, Vienna, Austria
a r t i c l e i n f o
Article history:
Received 30 May 2021
Accepted 14 September 2021
Available online xxx
Keywo rds:
Non-selective beta blockers
Portal hypertension
a b s t r a c t
Background and Aims: Non-alcoholic steatohepatitis has become a leading cause of cirrhosis. The prognos-
tic value of (HVPG)-guided NSBB prophylaxis remains to be investigated in the setting of NASH cirrhosis.
Methods: Patients with NAS H cirrhosis and varices undergoing HVPG-guided NSBB therapy were in-
cluded. HVPG-response to NSBBs was evaluated within a median 52 (IQR:28–71) days after baseline mea-
surement. HVPG-Response was defined as a decrease of 10% from baseline or below < 12 mmHg. The
composite endpoint was defined as variceal bleeding, decompensation, and liver-related death.
Results: Thirtyeight patients were included: Child-A/B:33(87%), Child-C:5(13%) median HVPG:19.7
±4.7 mmHg. 21(55.3%) patients achieved HVPG-response to NSBB. Presence of diabetes(aOR:0.16,
p = 0.038) and arterial blood pressure (aOR:1.07, p = 0.044) were independently associated with NSBB-
response. While NSBB-HVPG-responders showed fewer decompensations within 90 days ( n = 1(5%)
vs. n = 3(29%), p = 0.172), only Child-Pugh stage B/C ( p = 0.001), MELD 15( p = 0.021) and
HVPG 20 mmHg( p = 0.011) predicted the composite endpoint at 90 days. Similarly, after 2years of
follow-up, only Child-Pugh stage (B: p = 0.001, C: p < 0.001), MELD 15 ( p = 0.021), HVPG 20 mmHg
( p = 0.011) predicted the composite endpoint. Importantly, all bleeding events occurred in HVPG-NSBB
Conclusion: HVPG-response to NSBB was achieved in 55.3% of NASH patients with varices and this
seemed to protect from variceal bleeding. However, only baseline HVPG 20 mmHg, Child-Pugh stage
B/C and MELD 15 were predictors of decompensation/death in patients with NA SH cirrhosis and varices.
©2021 The Author(s). Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l.
This is an open access article under the CC BY license ( )
1. Introduction
Due to the global epidemic of obesity, non-alcoholic fatty liver
disease (NAFLD) has become one of the most common liver disease
in the Western World [1] . The prevalence of NAFLD is reported
to be as high as 15–35% [2] . Recent epidemiological studies sug-
gest that NASH will soon be the number one indication for liver
transplantation in the United States [3] . Whether NASH alone or
Funding: No financial support was rec eived for this study
✩✩ Guarantor of the article: Thomas Reiberger
Corresponding author at: Division of Gastroenterology and Hepatology, Depart-
ment of Internal Medicine III, Medical University of Vienna, Waehringer Guertel 18-
20, Vienna A-1090, Austri a.
E-mail address: (T. Reiberger).
presence of advanced fibrosis is the main driver of hepatic decom-
pensation and mortality remains controversial, however, advanced
stage of fibrosis, i.e. bridging fibrosis (F3) and cirrhosis (F4), is a
major risk factor for mortality in NASH patients [4] .
In patients suffering from other etiologies of cirrhosis, clinically
significant portal hypertension (CSPH) is the main driver of com-
plications [5 , 6] , such as development of ascites or variceal bleed-
ing. Importantly, CSPH –diagnosed by a hepatic venous pressure
gradient (HVPG) of 10 mmHg - can be present before portal-
hypertensive complications have developed. In turn, CSPH indicates
a considerably increased risk for decompensation and mortality
[7] . In general, CSPH is present in about 50–60% of patients with
compensated cirrhosis without varices [5 , 8] – while patients with
varices are usually found with an HVPG of at least 10 mmHg [5 , 6] ,
1590-8658/© 2021 The Author(s). Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l. This is an open access article under the CC BY license
( )
Please cite this article as: R. Paternostro, J. Becker, B.S. Hofer et al., The prognostic value of HVPG-response to non-selective beta-blockers
in patients with NASH cirrhosis and varices, Digestive and Liver Disease,
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
patients with an HVPG > 12 mmHg are at risk for variceal bleeding
[9] . The risk for hepatic decompensation (compensated cirrhosis)
and mortality (decompensated cirrhosis) increases at HVPG lev-
els > 16 mmHg [10] and at > 20 mmHg, failure to control-bleeding,
early rebleeding and bleeding-related death is more frequently ob-
served [5 , 11 , 12] .
Non-selective beta-blocker (NSBB) therapy is the cornerstone in
the treatment of portal hypertension [5 , 6] . In primary- and sec-
ondary prophylaxis of variceal bleeding, NSBB therapy has shown
efficacy in reducing variceal bleeding or re-bleeding [5 , 13 , 14 ] alone
or in combination with endoscopic band ligation [15] . Moreover,
NSBB therapy has been shown to impact on other outcomes in
patients with CSPH such as reducing the incidence of ascites
[16] . Nevertheless, potential harmful effects of NSBB therapy are
still openly debated and several studies have suggested impaired
outcomes in certain subpopulations of cirrhotic patients [17 , 18 ] .
Measuring HVPG-response to NSBB therapy has been shown to
predict outcomes in different etiologies of cirrhosis [19] . Ac-
cording to recent guidelines, a HVPG-reduction 10 % (primary
prophylaxis)/ 20% (secondary prophylaxis) or < 12 mmHg is clas-
sified as NSBB-response [6 , 20] .
In NASH cirrhosis per se it has been shown that an HVPG 10
predicts clinical decompensation [21] . However it seems that NASH
patients have lower HVPG and WHVP levels for each stage of fibro-
sis, compared with HCV patients [22] . Optimal HVPG cut-offs that
predict clinical decompensation in NASH cirrhosis remain to be re-
Importantly, the efficacy and outcomes of NSBB induced HVPG-
response in patients with NASH cirrhosis has yet to be investigated.
Therefore, aim of our study is to investigate (i) NSBB-response
rates in primary- and secondary prophylaxis of variceal bleeding
and (ii) the impact of NSBB-response on clinical outcomes in pa-
tients with NASH cirrhosis.
2. Methods
2.1. Study population
Patients with NASH cirrhosis undergoing HVPG-guided NSBB
therapy at the Vienna Hepatic Hemodynamic Lab between 2003
and 2019 were retrospectively included. HVPG-guided NSBB ther-
apy was part of the clinical routine, as previously published
[13 , 23] .
Only patients in primary- (presence of esophageal varices, no
previous variceal bleeding) or secondary- (previous variceal bleed-
ing) –prophylaxis of variceal bleeding were included in the study.
According to local standard-operating procedures, NSBB therapy
was initiated after baseline HVPG-measurement and dose-titrated
according to heart-rate and blood-pressure [6] . NSBB-response was
evaluated at a median 52 days (IQR 28–71d) after baseline HVPG-
measurement. Response was defined as 10% or to < 12 mmHg
according to recent guidelines [6 , 20] . NSBB non-responders were
treated by endoscopic band ligation at 4–6 weeks interval until
eradication of all varices and according to the guidelines present
at the time [6 , 20] , patients in secondary prophylaxis were treated
with endoscopic band ligation irrespective of NSBB-response sta-
tus. Choice of NSBB agent (carvedilol or propranolol) was at the
discretion of the treating physician.
The diagnosis of NASH was based one of following criteria:
(i) biopsy-proven NASH on liver histology, (ii) liver disease in pa-
tients with features of the metabolic syndrome after excluding any
other etiology of liver disease, (iii) cryptogenic cirrhosis with dia-
betes mellitus (DM), pronounced hepatic steatosis or controlled at-
tenuation parameter (CAP) > 300 dB/m. Patients with hepatocellu-
lar carcinoma, HVPG values < 10 mmHg, HVPG-response measure-
ment 200 days from baseline HVPG and post liver-transplantation
were excluded from the study. All patients were either in primary-
or secondary-prophylaxis of variceal bleeding, hence esophageal
varices were present in all patients.
2.2. Clinical follow-up
Patients were followed for decompensating events and liver-
related death until Q4/2019. Primary outcome was defined as: the
first event that occurred during follow-up: large-volume paracen-
tesis (LVP) for ascites and/or spontaneous bacterial peritonitis, hos-
pital admission for hepatic encephalopathy grade III-IV, variceal
bleeding or re-bleeding, or liver-related death. In patients that
were already decompensated at study inclusion further decompen-
sation and/or liver-related death were the clinical endpoints of the
study. Patients undergoing liver transplantation were recorded and
censored with the day of transplantation.
2.3. Clinical characteristics and laboratory parameters
Age, gender, diabetes (non-insulin-dependent, NIDDM, or
insulin-dependent = IDDM) and dates of HVPG measurements and
of liver biopsy were recorded from medical history. The diagnosis
(date) of hepatocellular carcinoma (HCC) or any other malignancy
was recorded and patients with baseline HCC excluded. Informa-
tion on medical comorbidities and relevant co-medication (statins,
metformin, diuretics, encephalopathy medication) were extracted.
Laboratory parameters from routine testing were recorded: hema-
tology including platelet count, chemistry including serum albu-
min, bilirubin, creatinine, levels of AST, ALT and GGT as well as
coagulation tests including PTT and INR. Furthermore Child-Pugh-
and MELD-scores were recorded.
2.4. Measurement of the hepatic venous pressure gradient (HVPG)
and endoscopy
Measurement of HVPG and upper-gastrointestinal endoscopy
was performed according to the standard operating procedure of
the Vienna Hepatic Hemodynamic Lab [24] . Briefly, the right in-
ternal jugular vein was accessed under ultrasound guidance and
local anesthesia with Seldinger technique using a catheter intro-
ducer set (8.5 F, Arrow International, Reading, PA, USA). Then a
balloon catheter (7F, Ferlitsch HVPG catheter, Pejcl Medizintechnik,
Austria) was used to cannulate a hepatic vein via the transjugular
access [25] . CSPH was defined as an HVPG 10 mmHg [6] . Varices
were graded according to size and presence/absence of red-spot
signs as recommended by the respective guidelines [6] .
2.5. Statistical analysis
The IBM SPSS 24.0 statistic software (SPSS Inc., Armonk, NY)
and GraphPad Prism (GraphPad software Inc.) was used for all sta-
tistical analyzes. The characteristics of patients within the differ-
ent NASH severity strata (HVPG 20 vs. < 20 mmHg) and respon-
ders vs. non-responders were reported using descriptive statis-
tics: continuous variables were reported as mean ±standard de-
viation (SD) or median (interquartile range, IQR). Categorical vari-
ables were reported as numbers (n) and proportion (%) of pa-
tients with the certain characteristic. Student’s t -test was used for
group comparisons of parametric data, while Mann-Whitney-U test
was applied for group comparisons of non-parametric data. Chi-
squared or Fisher’s exact tests were used to compare proportions
of a certain characteristic between groups. Binary regression analy-
sis was conducted to find factors associated with NSBB-response
first, univariable analysis was done for clinical relevant variables
(BMI, baseline HVPG, NSBB-type [carvedilol or propranolol], dia-
betes mellitus, MELD-score, baseline mean arterial pressure [MAP],
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
Tabl e 1
Patient characteristics.
All patients ( n = 38) HVPG 10–19 mmHg ( n = 21) HVPG 20 mmHg ( n = 17) p -value
Age, mean ±SD 60 ±12 61 ±9 58 ±16 0.605
Sex, n (%)
Female 12 (31.6%) 6 (28.6%) 6 (35.3%) 0.658
Male 26 (68.4%) 15 (71.4%) 11 (64.7%)
BMI, n (%)
< 30 17 (44.7%) 9 (42.9%) 8 (47.1%) 0.796
30 21 (55.3%) 12 (57.1%) 9 (52.9%)
n (%)
No 13 (34.2%) 6 (28.6%) 7 (41.2%) 0.415
Yes 25 (65.8%) 15 (71.4%) 10 (58.8%)
Arterial hypertension, n (%)
No 18 (47.4%) 10 (47.6%) 8 (47.1%) 0.973
Yes 20 (52.6%) 11 (52.4%) 9 (52.9%)
Statin intake, n(%) 9 (23.7%) 6 (28.6%) 3 (17.6%) 0.431
ACEi/A RB intake, n(%) 8 (21.1%) 6 (28.6%) 2 (11.8%) 0.206
Metformin intake, n(%) 14 (36.8%) 8 (38.1%) 6 (35.3%) 0.859
Child-Pugh Score median points (IQR) 6 (5–9) 6 (5–8) 9 (6–10) 0.019
Child-Pugh Stage, n(%)
-A 20 (52.6%) 15 (71.4%) 5 (29.4%) 0.008
-B 13 (34.2%) 6 (28.6%) 7 (41.2%)
- C 5 (13.2%) 0
(0%) 5 (29.4%)
MELD (points), median (IQR) 11 (9–15) 10 (8–13) 14 (9–17) 0.026
MELD, n(%)
< 15 28 (73.7%) 19 (90.5%) 9 (52.9%) 0.009
15 10 (26.3%) 2 (9.5%) 8 (47.1%)
Severe/Refractory ascites, n(%) 3 (7.9%) 0 (0%) 3 (17.6%) 0.022
Varices needing treatment (VNT), n(%) 24 (63.2%) 11
(52.4%) 13 (76.5%) 0.126
History of variceal bleeding, n(%) 7 (18.4%) 4 (19%) 3 (17.6%) 0.912
NSBB-response, N(%) 21 (55.3%) 10 (47.6%) 11 (64.7%) 0.292
Abbreviations: BMI –body mass index, ACEi – Angiotensin Converting Enzyme Inhibitor, ARB – Angiotensin Receptor Blocker, MELD – Model
for End-Stage Liver Dease, NSBB – Non-Selective Beta Blockers, SD – Standard Deviation, IQR –Interquartile Range.
baseline heart rate [HR]). Then all variables with a p -value 0.1
were included in a multivariable binary regression model to find
independent factors associated with NSBB-response.
Kaplan-Meier Curves are shown for comparison of the inci-
dences of hepatic decompensation and/or death between the re-
spective subgroups (CPS A vs. B vs. C, HVPG 20 vs. < 20 mmHg,
MELD 15 vs. < 15 points, NSBB-responders vs. non-responders).
Log-rank tests were conducted to find differences in the probabil-
ity of the composite endpoint. Two-sided p values 0.05 denote
statistical significance.
2.6. Ethics
This study was approved by the local ethic committee (EK
1493/2016 and EK 1262/2017) and performed in accordance with
the Declaration of Helsinki. The need for written informed consent
was waived due to the retrospective nature of the study and the
anonymized reporting of results.
3. Results
3.1. Clinical characteristics Table 1
Thirty-two patients with NAFLD/NASH showed an HVPG
< 10 mmHg and were therefore excluded from the study. Finally,
n = 38 patients with NASH cirrhosis and CSPH (primary pro-
phylaxis: n = 31 [82%], secondary prophylaxis: n = 7 [18%]) and
available data on NSBB response were included in our study:CPS
A: 20 (52.6%), CPS B: 13 (34.2%), and CPS: C 5 (13.2%); HVPG
20 mmHg was present in 17 (44.7%) and < 20 mmHg in 21
(55.3%) of patients. When stratified according to HVPG ( 20 mmHg
vs. < 20 mmHg), CPS, and MELD-score were significantly higher in
the HVPG 20 mmHg (CPS B/C - HVPG < 20 mmHg: 6 [28.5%] vs.
20 mmHg: 12 [70.5%], p = 0.008; MELD 15 –HVPG < 20 mmHg:
2 [9.5%] vs. 20 mmHg: 8 [47.1%], p = 0.009).
During a median follow-up of 16.9 months (4.4–49), the pri-
mary endpoint was reached in 19 (50%) patients: Variceal (re-
)bleeding n = 2 (5.3%), LVP n = 8 (21.2%), LV P with SBP n = 2
(5.3%), hospital admission for hepatic encephalopathy grade III-IV
n = 5 (13.2%) and liver-related death n = 2 (5.3%).
Non-selective beta-blocker induced HVPG-response: efficacy
Tables 1 and 2 and Supp. Table TS3
Median time between baseline HVPG and HVPG-response mea-
surement under NSBB therapy was 52 days (IQR 28–71d). Twen-
tyone (55.3%) patients responded to NSBB therapy as defined by
our methods and international guidelines (HVPG decrease 10% or
below 12 mmHg). Seven (18.4%) patients were under propranolol
therapy (median dose 80 mg/day [80–100]) and 31 (81.5%) un-
der carvedilol therapy (median dose 12.5 mg/day [12.5–25]). NSBB
doses were not different between responders and non-responders
(propranolol: p = 0.571, carvedilol: p = 0.086). Prevalence of di-
abetes mellitus was significantly higher in NSBB non-responder
(82.4% vs. 52.4% in NSBB-responders, p = 0.05). In contrast, MELD
( p = 0.275), CPS ( p = 0.086), and HVPG ( p = 0.292) were not
significantly different between responders vs. non-responders. To
find factors independently associated with NSBB-response, we con-
ducted uni- and multivariable binary logistic regression analyzes.
We found that presence of diabetes mellitus was an independent
negative predictive factor for NSBB-response (OR 0.16, 95%CI: 0.03–
0.91, p = 0.038), whereas a higher MAP was independently associ-
ated with NSBB-response (per 1 mmHg; OR 1.07, 95%CI: 1.02–1.15,
p = 0.044). Neither Statin, ACEi/ARB or Metformin intake was as-
sociated with HVPG-response to NSBBs ( Table 2 ).
Non-selective beta-blocker induced HVPG-response: Hepatic hemo-
dynamics Fig. 1 , Supplemental Table S2
Median HVPG-decrease in HVPG-responders was 5.5 ±2.7
mmHg (or 26 ±11%): mean baseline-HVPG 20.5 ±5.2 mmHg
vs. response-HVPG 15.1 ±4.3 mmHg ( p < 0.001). Furthermore,
wedged hepatic venous pressure(WHVP; baseline: 29.3 ±6.3 vs.
response-HVPG: 22.9 ±5.8 mmHg, p < 0.001), mean arterial pres-
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
Tabl e 2
Predictors of achieving HVPG-response to NSBB ( 10% decrease or HVPG < 12 mmHg).
Univariable Binary Logistic Regression Multivariable Binary Logistic Regression
OR 95%CI P -value OR 95%CI P -value
BMI [per unit] 0.93 0.81–1.07 0.304
Baseline HVPG [per mmHg] 1.10 0.95–1.28 0.220
NSBB-type [carvedilol vs. propranolol] 0.16 0.02–1.46 0.103
Statin intake [yes] 0.57 0.13–2.55 0.458
ACEi/A RB intake
[yes vs. no] 1.46 0.29–7.23 0.644
Metformin intake [yes vs. no] 0.45 0.12–1.72 0.244
Diabetes [vs. no diabetes] 0.24 0.05–1.07 0.061 0.16 0.03–0.91 0.038
MELD [per point] 1.11 0.94–1.31 0.216
Baseline MAP [per mmHg] 1.05 0.99–1.11 0.076 1.07 1.02–1.15 0.044
Baseline HR
[per point] 0.99 0.95–1.05 0.983
Abbreviations: BMI –Body Mass Index, HVPG – Hepatic Venous Pressure Gradient, NSBB – Non-Selective Beta Blocker, ACEi
Angiotensin Converting Enzyme Inhibitor, ARB – Angiotensin Receptor Blocker, MELD – Model for End-Stage Liver Disease, MAP
– Mean Arterial Pressure, HR –Heart Rate.
Fig. 1. HVPG course and systemic hemodynamics (mean arterial pressure [MAP] and heart ra te [HR]) at baseline and at HVPG-response evaluation in (A) all NASH patients,
in (B) NSBB-responders, and in (C) NSBB-non-responders.
sure (MAP; baseline: 106 ±14 vs. response-HVPG: 93 ±16 mmHg,
p < 0.001) and heart rate (baseline: 80 ±16 vs. response-HVPG:
67 ±10/min, p < 0.001) all significantly decreased in NSBB-
responders. Free hepatic venous pressure however did not signif-
icantly change (0.174).
In HVPG-non-responders mean HVPG values signifi-
cantly increased (baseline: 18. 6 ±3.8 vs. response-HVPG:
20.6 ±5.7 mmHg, p = 0.021), however WHVP (0.733) and
FHVP (0.238) did not significantly change –even though there
was a numerical trend towards increasing WHVP and decreasing
FHVP. MAP was not significantly different ( p = 0.329) –nev-
ertheless non-responders already started NSBB-therapy with a
lower MAP ( p = 0.0 6 6). Finally, HR did significantly decrease
also in non-responders (baseline: 80 ±12 vs. response-HVPG:
69 ±9/min, p = 0.001) indicating a good compliance for NSBB
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
Fig. 2. Composite endpoint bleeding/decompensation within 90 days stratified according to (A) Child-Pugh-Stage, (B) MELD, (C) HVPG-strata and (D) NSBB-response.
Short-term outcomes: bleeding/decompensation during 90 days of
follow-up Fig. 2 A–D , Supplementary Fig. FS-1
During the first 90-days of clinical follow-up, CPS-strata
(CPS A vs. B: p = 0.109, CPS A vs. C: p = 0.030, CPS B
vs. C: 0.559), MELD 15/ < 15 ( p = 0.008) and HVPG-strata
( 20/ < 20 mmHg; p = 0.037) were predictive of the occurrence of
bleeding/decompensation. NSBB-response did not attain statistical
significance –however out of four events during the first 90 days,
n = 3 (75%) occurred in NSBB-non-responders compared to n = 1
(15%) in NSBB-responders ( p = 0.172).
Long-term outcomes: bleeding/decompensation during 2 years of
follow-up Fig. 3 A–E
At 2 years of clinical follow-up, CPS strata (CPS A vs. B:
p = 0.010, CPS A vs. C: p < 0.001, CPS B vs. C: 0.143),
MELD 15/ < 15 ( p = 0.021), and HVPG-strata ( 20/ < 20 mmHg,
p = 0.011) were all good indicators for the occurrence of bleed-
ing/decompensation. However, NSBB-response was not predic-
tive ( p = 0.993) and also absolute numbers of events were
the same between responders ( n = 5) vs. non-responders
( n = 5). However when only variceal-bleeding was taken
into account, all n = 2 events occurred in non-responders
( p = 0.108).
Long-term outcomes: Transplant-free survival (TFS) during 2 years
of follow-up Fig. 4 A–D
Within 2 years of follow-up, n = 5 patients died. CPS (CPS A vs.
B: p = 0.129, A vs. C: 0.042, B vs. C: 0.455; Kaplan-Meier Survival
Estimates: CPS A: 94%, CPS B: 83%, CPS C: 53%) and MELD 15/ < 15
( p = 0.004; Kaplan-Meier Survival Estimates: MELD < 15: 92%,
MELD 15: 68%) were good indicators for predicting TFS. However,
both HVPG-strata ( 20/ < 20 mmHg) and NSBB-response status did
not predict long-term TFS (HVPG-strata: p = 0.101, NSBB-response
status: p = 0.707).
4. Discussion
This is the first study reporting on the efficacy of NSBB in low-
ering HVPG in NASH patients. Moreover, our study provides infor-
mation on the potential prognostic impact of NSBB-induced hep-
atic venous pressure gradient (HVPG)-response in patients with
NASH cirrhosis.
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
Fig. 3. Incidence of the composite endpoint bleeding/decompensation after 24 months according to (A) Child-Pugh-Stage, (B) MELD, (C) HVPG, and HVPG-response and
occurrence of (D) any decompensation or (E) va riceal (re)-bleeding only.
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
Fig. 4. Transplant-free survival stratified according to (A) Child-Pugh stage, (B) MELD, (C) HVPG-strata and (D) NSBB-response.
Firstly, we could show that a NSBB-response of 10% or
< 12 mmHg was achieved in 55.3% of patients within the specific
etiology of NASH cirrhosis. These NSBB-HVPG response rates ob-
served in our NASH patients are similar to other etiologies where
response rates of 49–76% [13 , 23 , 26–28] have been reported. Inter-
estingly, we found that presence of diabetes was independently
associated with a lower likelihood whereas higher arterial blood
pressure was associated with an increased probability of HVPG-
response to NSBB therapy. This finding may be explained by the
fact that NSBB therapy is dose-titrated not only according to heart
rate but also to arterial blood pressure (i.e. as long as systolic
blood pressure is not < 110 mmHg). Therefore, a low baseline ar-
terial blood pressure might have prevented effective NSBB dosing
in some patients potentially resulting in lower rates of HVPG re-
sponse [23] . Particularly in NASH most patients present with com-
ponents of the metabolic syndrome such as overweight/obesity, di-
abetes mellitus, arterial hypertension, and dyslipidemia [29] . In our
cohort 55.3% of subjects were obese, and 53% and 66%, respec-
tively, suffered from arterial hypertension and diabetes.
Accordingly, we have observed that NSBB non-responders
started NSBB therapy already with a lower baseline MAP which
may have prevented sufficient NSBB dose titration. Thus, one obvi-
ous explanation for the association of arterial blood pressure with
HVPG response may include higher arterial blood pressure allow-
ing for a more effective NSBB dose escalation.
We can only speculate about the pathophysiological mech-
anism underlying the association of diabetes with NSBB-non-
response. Diabetes per-se has been associated with impaired out-
comes in cirrhosis across several etiologies [30–33] . Moreover,
proper glycemic-control seems to be associated with a reduced
risk for developing complications such as spontaneous-bacterial
peritonitis [33] or hepatocellular carcinoma [34] . Diabetes per-se
might increase portal pressure and NSBB-response may be coun-
teracted by clinical deterioration. Moreover, previous studies have
shown an HVPG-increase with alcohol intake [35] and acute-on-
chronic liver failure [36] , while HVPG decreased in patients post
HCV treatment [37 , 38] – similar patterns could be present in dia-
betic patients with impaired glycemic control –however, we can-
not provide data on this matter and this association remains to
be answered in future studies. Additionally, we have assessed po-
tential effects of co-medications for dyslipidemia (statin-therapy),
diabetes (metformin) and arterial hypertension (ACEi/ARBs) that
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
may impact on NASH and thus, on portal hypertension. Interest-
ingly, we did not identify any association of these co-medications
with severity of portal hypertension ( Table 1 ), likelihood to achieve
hemodynamic HVPG response ( Table 2 ) or development of liver-
related events (decompensation, death). Finally, we did not observe
a difference in the proportion of NASH patients achieving hemody-
namic response to propranolol versus carvedilol.
Secondly, we reported short- and long-term outcome data in
NASH patients with cirrhosis stratified according to NSBB-response
status. We found that variceal bleeding was only seen in NSBB-
HVPG non-responders. This finding is in line with the previous
literature in other etiologies of advanced chronic-liver disease re-
porting a lower probability of variceal bleeding in NSBB-responders
[5 , 19 ] . In contrast, NSBB-responders did not differ in regard to the
risk of overall decompensation and transplant-free mortality, how-
ever this lack of difference might result from limited sample size
and a limited number of clinical events in the short-term follow-
up rather than the ineffectiveness of NSBB therapy. As NASH often
takes a very heterogenous clinical course [21 , 39] , the HVPG-NSBB
response at short term might not protect from long-term outcomes
that are likely more impact by natural disease course. Moreover
several studies have suggested that only a certain group of cir-
rhotic patients benefit from NSBB therapy [40 , 41] , whilst others
such as those with refractory ascites [42] or spontaneous bacterial
peritonitis [17] might not.
Finally, we still found that the presence of high-risk portal hy-
pertension, i.e. HVPG 20 mmHg was an excellent predictor of
both short term and long-term decompensation. Importantly, an
HVPG 20 mmHg was of similar prognostic value as compared to
established prognostic scores such as the Child-Pugh stage and the
MELD [39] . We are only aware of two other studies that have re-
ported on the prognostic value of HVPG in NASH before [43 , 44] .
However, both studies [43 , 44] included NASH patients with less
pronounced portal hypertension, since some patients had not even
developed varices.
While our study reports on novel data on the value of NSBB
and HVPG-reponse in NASH patients with varices, some limita-
tions have to be discussed. Firstly, HVPG-response measurements
were not always done within the same timeframe after baseline-
measurement. However, with a median duration of 52 days be-
tween baseline and response measurements, the timeframe is very
similar to previous studies [19] . Secondly indication for HVPG-
measurement might be questioned, nevertheless it has been clearly
shown that HVPG-response-guided NSBB therapy improves out-
comes in patients, with or without ascites [45] and is safe [24 , 46] .
Thirdly, one patient had a HVPG of 11 mmHg and therefore be-
low 12 mmHg, however studies have demonstrated that as soon
as there is CSPH, there is hyperdynamic circulation which ren-
ders patients more susceptible to NSBB therapy [47] , thus the
benefit of (HVPG-guided-) NSBB therapy in patients with “only”
CSPH but values < 12 mmHg is given [8 , 16 ] . It should also be
noted that considering our data, well-validated clinical scores such
as MELD or CPS are strong predictors also in NASH cirrhosis
and therefore in absence of HVPG availability these scores seem
to be sufficient to estimate prognosis, but especially in patients
with varices requiring bleeding prophylaxis, HVPG provides use-
ful prognostic information when choosing NSBB for bleeding pro-
phylaxis, as no single patient achieving HVPG response to NSBB
bled in our population of NASH cirrhotics. Fourthly, two types
of beta-blockers were used in the study, however this is mainly
due to the fact that for compensated cirrhotics we mostly chose
carvedilol while for decompensated/secondary prophylaxis, where
safety of carvedilol is not well established, we usually prefer pro-
pranolol [41 , 48] . Importantly, both patients using NSBB for pri-
mary ( n = 31) and secondary ( n = 7) prophylaxis of variceal
bleeding were included in the study. However, the low num-
ber of patients in the secondary prophylaxis group did not al-
low for a separate analysis in the two settings. Still, according
to the Baveno VI consensus conference [20] an HVPG change of
at least 10% is considered to indicate a clinically-relevant bene-
fit and thus, justifies analyzing the 10% or < 12 mmHg response
cut-off for both patient groups. Finally, we included only a small
number of patients and therefore our study might be underpow-
ered to draw firm conclusions regarding the prognostic value of
NSBB-response in NASH, due to potential type II error. However,
HVPG measurements are only available in some centers worldwide
and up-to-date no data on NSBB-HVPG response have been re-
ported so far, thus, setting this limitation into a relative academic
To conclude, a baseline HVPG 20 mmHg predicted bleed-
ing/decompensation in patients with NASH cirrhosis and varices.
While 55.3% of NASH patients achieved HVPG-response to NSBB
therapy and did not experience any variceal bleeding, the rates
of decompensation and transplant-free mortality were similar to
HVPG non-responders.
Declaration of Competing Interest
The authors report no real or potential conflict of interest re-
lated to this study. The following conflicts of interests outside of
this study exist R.P.: no conflict of interests. J.B., B.H., V.P. and
H.S.: no conflict of interests. B.SIM.: received travel support from
AbbVie and Gilead. G.S.: no conflict of interests. P.S .: received
speaking honoraria from Bristol-Myers Squibb and Boehringer-
Ingelheim, consulting fees from PharmaIN, and travel support
from Falk. B.SCH.: no conflict of interests. T.BU.: Speaker hono-
raria from Bristol-Myers Squibb. Travel support from Gilead , Ab-
bvie, and Bristol-Myers Squibb. D.B.: Has received travel support
from AbbVie and Gilead and served as speaker for AbbVie. T.BI.:
no conflict of interests. M.M.: Speaker and/or consultant and/or
advisory board member for AbbVie, Bristol-Myers Squibb, Collec-
tive Acumen, Gilead, and W. L. Gore & Associates. Travel sup-
port from AbbVie , Bristol-Myers Squibb, and Gilead. M.T.: Speaker
for BMS, Falk, Gilead, Intercept, and MSD; advisory boards for
Albireo, BiomX, Boehringer Ingelheim, Falk, Genfit, Gilead, Inter-
cept, Janssen, MSD, Novartis, Phenex, and Regulus. He further re-
ceived travel grants from Abbvie, Falk, Gilead and Intercept and
unrestricted research grants from Albireo, Cymabay, Falk, Gilead,
Intercept, MSD and Take da. He is also co-inventor of patents
on the medical use of norUDCA filed by the Medical University
of Graz. T.R.: grant support from Abbvie , Boehringer-Ingelheim,
Gilead, MSD, Philips Healthcare, Gore; speaking honoraria from
Abbvie, Gilead, Gore, Intercept, Roche, MSD; consulting/advisory
board fee from Abbvie, Bayer, Boehringer-Ingelheim, Gilead, Inter-
cept, MSD, Siemens; and travel support from Boehringer-Ingelheim,
Gilead and Roche.
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi: 10.1016/j.dld.2021.09.009 .
[1] Pimpin L , Cortez-Pinto H , Negro F , et al. Burden of liver disease in Europe: epi-
demiology and analysis of risk factors to identify prevention policies. J Hepatol
2018;69:718–35 .
[2] Younossi Z , Anstee QM , Marietti M , et al. Global burden of NAFL D and NASH:
trends, predictions, risk factors and prevention. Nat Rev Gastroenterol Hepatol
2018;15:11–20 .
[3] Goldberg D , Ditah IC , Saeian K , et al. Changes in the prevalence of hepatitis C
virus infection, nonalcoholic steatohepatitis, and alcoholic liver disease among
patients with cirrhosis or liver failure on the waitlist for
liver transplantation.
Gastroenterology 2017;152:1090–9 e1 .
R. Paternostro, J. Becker, B.S. Hofer et al. Digestive and Liver Disease xxx (xxxx) xxx
JID: YDLD [m5G; October 4, 2021;11:58 ]
[4] Dulai PS , Singh S , Patel J , et al. Increased risk of mortality by fibrosis stage in
nonalcoholic fatty liver disease: systematic review and meta-analysis. Hepatol-
ogy 2017;65:1557–65 .
[5] Garcia-Tsao G , Abraldes JG , Berzigotti A , et al. Portal hypertensive bleeding in
cirrhosis: risk stratification, diagnosis, and management: 2016 practice guid-
ance by the American association for the study of liver diseases. Hepatology
2017;65:310–35 .
[6] Reiberge r T , Puspok A , Schoder M , et al. Austrian consensus guidelines on
the management and treatment of portal hypertension (Billroth III). Wien Klin
Woch enschr 2017;129:135–58 .
[7] D’Amico G , Morabito A , D’Amico M , et al. Clinical states of cirrhosis and com-
peting risks. J Hepatol 2018;68:563–76 .
[8] Groszmann RJ , Garcia-Tsao G , Bosch J , et al. Beta-blockers to prevent gastroe-
sophageal varices in patients with cirrhosis. N Engl J Med 2005;353:2254–61 .
[9] Bosch J , Garcia-Pagan JC . Prevention of variceal rebleeding. Lancet
2003;361:952–4 .
[10] Merkel C , Bolognesi M , Sacerdoti D , et al. The hemodynamic response to
medical treatment of portal hypertension as a predictor of clinical effective-
ness in the primary prophylaxis of var iceal bleeding
in cirrhosis. Hepatology
20 0 0;32:930–4 .
[11] Abraldes JG , Villanueva C , Banares R , et al. Hepatic venous pressure gradient
and prognosis in patients with acute variceal bleeding treated with pharma-
cologic and endoscopic therapy. J Hepatol 2008;48:229–36 .
[12] Moitinho E , Escorsell A , Bandi JC , et al. Prognostic value of early mea-
surements of portal pressure in acute variceal bleeding. Gastroenterology
1999;117:626–31 .
[13] Reiberger T , Ulbrich G , Ferlitsch A , et al. Carvedilol for primary prophylaxis
of varicea l bleeding in cirrhotic patients with haemodynamic non-response to
propranolol. Gut 2013;62:1634–41 .
[14] Pfisterer N , Dexheimer C , Fuchs EM , et al. Betablockers do not increase efficacy
of band ligation in primary prophylaxis but they improve survival in secondary
prophylaxis of variceal bleeding. Aliment Pharmacol Ther 2018;47:966–79 .
[15] Sharma M , Singh S , Desai V , et al. Comparison of therapies for primary preven-
tion of esophageal variceal bleeding: a systematic review and network meta–
analysis. Hepatology 2019;69:1657–75 .
[16] Villanueva C , Albillos A , Genesca J , et al. beta blockers to prevent decompen-
sation of cirrhosis in patients with clinically significant portal hypertension
(PREDESCI): a randomised, double-blind, placebo-controlled, multicentre trial.
Lancet 2019;393:1597–608 .
[17] Mandorfer M , Bota S , Schwabl P , et al. Nonselective beta blockers increase risk
for hepatorenal syndrome and death in patients with cirrhosis and sponta-
neous bacterial peritonitis. Gastroenterology 2014;146:1680–90 e1 .
[18] Mandorfer M, Reiberger T. Beta blockers and cirrhosis, 2016. Dig Liver Dis
2017;49(1):3–10. doi: 10.1016/j.dld.2016.09.013 .
[19] Mandorfer M , Hernández-Gea V , Reiber ger T , et al. Hepatic venous pressure
gradient response in non-selective beta-blocker treatment—is it worth mea-
suring? Curr Hepatol Rep 2019;18:174–86 .
[20] de Franchis R . Expanding consensus in portal hypertension: report of the
Baveno VI Consensus Workshop: stratifying risk and individualizing care for
portal hypertension. J Hepatol 2015;63:743–52 .
[21] Sanyal AJ , Harrison SA , Ratziu V , et al. The natural history of advanced fibrosis
due to nonalcoholic steatohepatitis: data from the simtuzumab trials. Hepatol-
ogy 2019;70:1913–27 .
[22] Sourianarayanane A , Talluri J , Humar A , et al. Stage of fibrosis and portal pres-
sure correlation in nonalcoholic steatohepatitis. Eur J Gastroenterol Hepatol
2017;29:516–23 .
[23] Schwarzer R , Kivaranovic D , Paternostro R , et al. Carvedilol for reducing portal
pressure in primary prophylaxis of variceal bleeding: a dose-response study.
Aliment Pharmacol Ther 2018;47:1162–9 .
[24] Reiberger T, Schwabl P, Trauner M, et al. Measurement of the hepatic venous
pressure gradient and transjugular liver biopsy. J Vis Exp 2020;18(160). doi: 10.
3791/58819 .
[25] Ferlitsch A , Bota S , Paternostro R , et al. Evaluation of a new balloon
sion catheter specifically designed for measurement of hepatic venous pres-
sure gradient. Liver Int 2015 .
[26] Banares R , Moitinho E , Matilla A , et al. Randomized comparison of long-term
carvedilol and propranolol administration in the treatment of portal hyperten-
sion in cirrhosis. Hepatology 2002;36:1367–73 .
[27] Kim SG , Kim TY , Sohn JH , et al. A randomized, multi-center, open-label study
to evaluate the efficacy of carvedilol vs. propranolol to reduce portal pressure
in patients with liver cirrhosis. Am J Gastroenterol 2016;111:1582–90 .
[28] Banares R , Moitinho E , Piqueras B , et al. Carvedilol, a new nonselective be-
ta-blocker with intrinsic anti- alpha1-adrenergic activity, has a greater por-
tal hypotensive effect than propranolol in patients with cirrhosis. Hepatology
1999;30:79–83 .
[29] Lonardo A , Nascimbeni F , Mantovani A , et al. Hypertension, diabetes,
atherosclerosis and NASH: cause or consequence? J Hepatol 2018;68:335–52
[30] Hsu YC , Yip TC , Ho HJ , et al. Development of a scoring system to predict hep-
atocellular carcinoma in Asians on antivirals for chronic hepatitis B. J Hepatol
2018;69:278–85 .
[31] Huang YW , Yang SS , Fu SC , et al. Increased risk of cirrhosis and its decompen-
sation in chronic hepatitis C patients with new-onset diabetes: a nationwide
cohort study. Hepatology 2014;60:807–14 .
[32] Elkrief L , Chouinard P , Bendersky N , et al. Diabetes mellitus is an independent
prognostic factor for major liver-related outcomes in patients with cirrhosis
and chronic hepatitis C.
Hepatology 2014;60:823–31 .
[33] Tergast TL , Laser H , Gerbel S , et al. Association between type 2 diabetes
mellitus, HbA1c and the risk for spontaneous bacterial peritonitis in pa-
tients with decompensated liver cirrhosis and ascites. Clin Transl Gastroenterol
2018;9:189 .
[34] Cheuk-Fung Yip T , Wai-S un Wong V , Lik-Yuen Chan H , et al. Effects of diabetes
and glycemic control on risk of hepatocellular carcinoma after seroclearance
of hepatitis B surface antigen. Clin Gastroenterol Hepatol 2018;16:765–73 e2 .
[35] Luca A , García-Pagán JC , Bosch J, , et al. Effects of ethanol consumption on
hepatic hemodynamics
in patients with alcoholic cirrhosis. Gastroenterology
1997;112:1284–9 .
[36] Mehta G , Mookerjee RP , Sharma V , et al. Systemic inflammation is associ-
ated with increased intrahepatic resistance and mortality in alcohol-related
acute-on-chronic liver failure. Liver Int 2015;35:724–34 .
[37] Mandorfer M , Kozbial K , Schwabl P , et al. Sustained virologic response to in-
terferon-free therapies ameliorates HCV-induced portal hypertension. J Hepatol
2016;65:692–9 .
[38] Mandorfer M , Kozbia l K , Schwabl P , et al. Changes in hepatic venous pressure
gradient predict hepatic decompensation in patients who achieved sustained
virologic response to interferon-free therapy. Hepatology 2020;71:1023–36 .
[39] Vilar-Gomez E , Calzadilla-Bertot L , Wai-Su n Won g V , et al. Fibrosis severity
as a determinant of cause-specific mortality in patients with advanced non-
alcoholic fatty liver disease: a multi-national cohort Study. Gastroenterology
2018;155:443–57 e17 .
[40] Krag A , Wiest R , Albillos A , et al. The window hypothesis: haemodynamic and
non-haemodynamic effects of
β-blockers improve survival of patients with cir-
rhosis during a window in the disease. Gut 2012;61:967–9 .
[41] Reiberger T , Mandorfer M . Beta adrenergic blockade and decompensated cir-
rhosis. J Hepatol 2017;66:849–59 .
[42] Sersté T , Melot C , Francoz C , et al. Deleterious effects of beta-blockers
on survival in patients with cirrhosis and refractory ascites. Hepatology
2010;52:1017–22 .
[43] Sebastiani G , Alshaalan R , Wo ng P , et al. Prognostic va lue of non-invasive fibro-
sis and steatosis tools, hepatic venous pressure gradient (HVPG) and Histology
in nonalcoholic steatohepatitis. PLoS ONE 2015;10:e0128774 .
[44] Harrison SA , Abdelmalek MF , Caldwell S , et al. Simtuzumab is ineffective for
patients with bridging fibrosis or compensated cirrhosis caused by nonalco-
holic steatohepatitis. Gastroenterology 2018;155:1140–53 .
[45] Turco L , Villanueva C , La Mura V , et al. Lowering portal pressure improves out-
comes of patients with cirrhosis, with or without ascites: a meta-analysis. Clin
Gastroenterol Hepatol 2020;18:313–27 e6 .
[46] Reiberge r T , Bucsics T , Paternostro R , et al. Small esophageal var ices in patients
with cirrhosis-should we treat them? Curr Hepatol Rep 2018;17:301–15 .
[47] Villanueva C , Albillos A , Genescà J , et al. Development
of hyperdynamic circu-
lation and response to
β-blockers in compensated cirrhosis with portal hyper-
tension. Hepatology 2016;63:197–206 .
[48] Téllez L , Ibáñez-Samaniego L , Pérez Del Villar C , et al. Non-selective be-
ta-blockers impair global circulatory homeostasis and renal function in cir-
rhotic patients with refractory ascites. J Hepatol 2020;73:1404–14 .