Portal pressure reductions induced by nonselective beta‐blockers improve outcomes and decrease mortality in patients with cirrhosis with and without ascites

Short abstract

Content available: Audio Recording

Listen to an audio presentation of this article.

INTRODUCTION

Portal hypertension is the most common complication of cirrhosis and is the main determinant of the transition from an asymptomatic compensated stage to a symptomatic decompensated stage. ¹ Decompensation is characterized by the development of ascites, variceal bleeding, and/or hepatic encephalopathy. These two stages of cirrhosis (compensated and decompensated) have entirely different mortalities and should therefore be considered distinct entities. ² , ³

Nonselective beta‐blockers (NSBB) are the first‐line therapy to reduce portal pressure and prevent first variceal hemorrhage (primary prophylaxis) and recurrent variceal hemorrhage (secondary prophylaxis). By decreasing portal pressure, NSBB are also able to prevent other decompensating events (mainly ascites) ⁴ and improve survival.

In a meta‐analysis by D’Amico et al., ⁵ patients with cirrhosis who achieved a response to NSBB (determined by a predefined reduction of the hepatic venous pressure gradient [HVPG]) had a significantly lower risk for variceal bleeding and death. However, all the studies analyzed in this meta‐analysis combined results from patients with both compensated and decompensated cirrhosis. Moreover, concerns have been recently raised on the possible deleterious effect of NSBB in patients with refractory ascites. ⁶

SUMMARY OF FINDINGS

In our meta‐analysis, ⁷ we included studies that assessed the difference in all clinically relevant outcomes between HVPG “responders” and “non‐responders” to NSBB therapy, stratifying patients by the presence or absence of ascites. The presence of ascites was chosen as indicative of decompensation because it is the most frequent decompensating event in cirrhosis and the one that carries the highest mortality risk.

We identified 15 studies of primary or secondary prophylaxis of variceal bleeding that had reported on variceal bleeding and death in HVPG responders versus non‐responders (HVPG response was defined as a decrease >10%–20% from baseline or to <12 mm hg) published until November 2015 with original data still available. Principal investigators of eligible studies provided data regarding the presence or absence of ascites and clinical outcomes specific for each of these subgroups and for each of the responder groups (responder versus non‐responder), separately for patients enrolled in primary or secondary prophylaxis studies.

Primary outcomes for patients without ascites were the development variceal bleeding (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies), ascites, or encephalopathy. Primary outcomes for patients with ascites were the development of variceal bleeding (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies) or refractory ascites, spontaneous bacterial peritonitis, hepatorenal syndrome, or encephalopathy.

Secondary outcome was death for both subgroups.

Data from 1113 unique patients were analyzed. The majority were male (49%–79%) with a mean age ranging from 47 to 63 years (Table 1). The main etiology of cirrhosis was alcohol (37%–78% of patients). Patients with ascites (n = 452; 40.6%) had higher HVPG levels (both at baseline and at follow‐up) compared with patients without ascites (n = 661, 59.4%).

TABLE 1.

Evidence profile

No. of trials: 15 No. of randomized clinical trials: 5 No. of case‐series studies: 10 Studies years: 1990–2013 No. of patients included in the 15 studies: 1341 No. of patients included in the meta‐analysis: 1113 Population: patients with cirrhosis and varices needing treatment Men (range): 49%–79% Mean age (range): 47–63 years Race/ethnicity: not reported Main etiology of cirrhosis: alcohol (37%–78% of patients) Intervention: treatment with NSBB (87% of the patients) to lower portal pressure Comparison: patient HVPG responders versus HVPG non‐responders to NSBB therapy a Time between baseline and follow‐up HVPG: 0.6–5 months Primary outcomes: In patients without ascites: development of ascites, variceal hemorrhage (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies), or encephalopathy In patients with ascites: development of variceal bleeding (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies) or refractory ascites, spontaneous bacterial peritonitis, hepatorenal syndrome, or encephalopathy Secondary outcomes: In patients without ascites: all‐cause mortality In patients with ascites: all‐cause mortality Follow‐up time: 15–70 months Prespecified subgroup analyses: Meta‐analyses were performed separately for patients without ascites and for those with ascites, and also performed separately for patients enrolled in primary prophylaxis and those enrolled in secondary prophylaxis for variceal bleeding studies.

^aIn 13 studies, HVPG response was defined as a decrease to <12 mm Hg or a reduction of >20% from baseline; in 2 studies (with all patients taking NSBB for primary prophylaxis), HVPG response was defined as a decrease to <12 mm Hg or a reduction of >10% from baseline.

In patients without ascites, responders (50%) had a significantly lower rate of clinical events than non‐responders (odds ratio [OR], 0.35; 95% confidence interval [CI]: 0.22–0.56), both in patients included in primary (n = 332) (OR, 0.28; 95% CI: 0.13–0.58) or in secondary (n = 329) prophylaxis studies (OR, 0.41; 95% CI: 0.22–0.78) without significant heterogeneity (p = 0.10). Death/transplant rates were also significantly lower among responders (OR, 0.50; 95% CI: 0.32–0.78) in both patients in primary (OR, 0.44; 95% CI: 0.20–0.98) or secondary prophylaxis (OR, 0.55; 95% CI: 0.32–0.95) studies without significant heterogeneity (p = 0.28) (Table 2, Figure 1).

TABLE 2.

Summary of findings

	Benefit of HVPG Responders Versus Non‐responders
	Any Clinical Event a					Death/Transplant
	Overall Effect	p	Primary Versus Secondary Prophylaxis	p	Heterogeneity	Overall Effect	p	Primary Versus Secondary Prophylaxis	p	Heterogeneity
Patients without ascites (n = 661)	OR, 0.35 (0.22–0.56)	<0.0001	Primary (n = 332) OR, 0.28 (0.13–0.58)	0.0007	0.10	OR, 0.50 (0.32–0.78)	0.002	Primary OR, 0.44 (0.20–0.98)	0.04	0.28
			Secondary (n = 329) OR, 0.41 (0.22–0.78)	0.006				Secondary OR, 0.55 (0.32–0.95)	0.03
Patients with ascites (n = 452)	OR, 0.27 (0.16–0.43)	<0.00001	Primary (n = 172) OR, 0.38 (0.16–0.89)	0.03	0.40	OR, 0.47 (0.29–0.75)	0.002	Primary OR, 0.74 (0.34–1.63)	0.46	0.69
			Secondary (n = 280) OR, 0.24 (0.12–0.48)	<0.0001				Secondary OR, 0.36 (0.20–0.75)	0.0007

Numbers in parentheses represent 95% CI. In patients with cirrhosis and without ascites, NSBB‐induced reduction in portal pressure improves outcomes (variceal bleeding, ascites, or encephalopathy) and decreases mortality. In patients with cirrhosis and with ascites, NSBB‐induced reduction in portal pressure improves outcomes (variceal bleeding, refractory ascites, spontaneous bacterial peritonitis, or hepatorenal syndrome) and decreases mortality.

^a In patients without ascites: variceal bleeding (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies), ascites, or encephalopathy; in patients with ascites: variceal bleeding (first in primary prophylaxis studies, recurrent in secondary prophylaxis studies) or refractory ascites, spontaneous bacterial peritonitis, hepatorenal syndrome, or encephalopathy.

FIGURE 1.

In patients with compensated cirrhosis, portal pressure reductions by NSBB lower the risk of decompensation and improve survival. In patients with decompensated cirrhosis, portal pressure reductions by NSBB lower the risk of "further" decompensation and improve survival. *Ascites, variceal hemorrhage, and encephalopathy. **Refractory ascites, variceal hemorrhage, encephalopathy, spontaneous bacterial peritonitis, and hepatorenal syndrome. ^§Compared with patients of the same group without a significant decrease in portal pressure.

In patients with ascites, responders (42%) had a significantly lower rate of clinical events than non‐responders (OR, 0.27; 95% CI: 0.16–0.43), both in patients included in primary (n = 172) (OR, 0.38; 95% CI: 0.16–0.89) or in secondary (n = 280) prophylaxis studies (OR, 0.24; 95% CI: 0.12–0.48) without significant heterogeneity (p = 0.40). Death/transplant rates were lower among responders (OR, 0.47; 95% CI: 0.29–0.75) both in patients in primary (OR, 0.74; 95% CI: 0.34–1.63) or secondary prophylaxis (OR, 0.36; 95% CI: 0.20–0.65) studies without significant heterogeneity (p = 0.69) (Table 2, Figure 1).

Patients without ascites had a greater reduction in HVPG (a decrease of 19%) than patients with ascites (a decrease of 14%). Patients without ascites and without previous variceal bleeding had the highest HVPG response rate (50% of patients were HVPG responders), whereas patients with ascites and prior variceal bleeding had the lowest HVPG response rate (only 36% of them were responders).

DISCUSSION

This meta‐analysis ⁷ shows that decreases in portal pressure (a reduction in HVPG to levels <12 mm hg or >10%–20% from baseline), determined mainly by NSBB treatment, are associated with better outcomes both in patients with and without ascites.

D’Amico et al. ⁵ had already shown an association between an HVPG‐identified reduction of portal hypertension and a reduced risk for variceal bleeding and death, analyzing studies that included both patients with and without ascites. However, it has become clear that these patients are different prognostically ² , ³ and should therefore be analyzed separately.

Results of this study ⁷ show that, in the subgroup of patients without ascites (all of them with varices), decompensation (defined as development of ascites, variceal bleeding, or encephalopathy) is significantly reduced in HVPG responders to NSBB. Even in the absence of ascites, a large percentage of these patients have been shown to have a hyperdynamic circulatory state ⁸ , ⁹ with a more pronounced effect of NSBB on HVPG, leading to a significantly reduced risk for decompensation and a reduced risk for death and liver transplantation.

In the subgroup of patients with ascites, portal pressure reduction, achieved mainly by NSBB, was not only associated with a lower rate of “further” decompensation (defined as development of variceal bleeding, refractory ascites, spontaneous bacterial peritonitis or hepatorenal syndrome) but also to a decrease in death/transplant rate. This decrease was particularly significant in patients with ascites and a previous variceal bleeding event, which have the lowest life expectancy. ³ Results in this subgroup of patients are of particular interest in light of concerns regarding a potential deleterious effect of NSBB in patients with refractory ascites. ⁶ Based on careful analysis of available literature, the recent Baveno VII consensus conference on portal hypertension ¹⁰ concluded that NSBB should not be avoided in patients with refractory ascites. Attention should, however, be paid in those patients with persistently low blood pressure (systolic blood pressure < 90 mm Hg or mean arterial pressure < 65 mm Hg) and/or hepatorenal syndrome‐acute kidney injury (HRS‐AKI) in whom NSBB should be dose‐reduced or discontinued until blood pressure returns to baseline and/or HRS‐AKI resolves.

In conclusion, findings in this meta‐analysis are proof of concept that reduction in portal pressure (through the use of NSBB) is associated with better clinical outcomes in patients with both compensated and decompensated cirrhosis. However, they do not imply that HVPG should be measured before and after initiating NSBB in an individual patient. In fact, recent data show that, although repeat measurements of HVPG are reliable at a study level, they are not reliable at an individual level and should not be used to tailor response to NSBB in the individual patient. ¹¹ , ¹²

LIMITATIONS AND FUTURE AREAS OF STUDY

Due to the retrospective nature of every meta‐analysis and the absence of uniformly collected data in the analyzed studies, it was not possible to explore the impact on outcomes of possible confounders, such as comorbidities, hepatocellular carcinoma, Child‐Pugh score, and Model for End‐Stage Liver Disease, as well as the potential confounder of the use of alcohol. However, only 9 of 15 studies reported on alcohol use/abstinence during follow‐up, and of these only 2 studies, comprising just 110 patients, found a higher percentage of alcohol abstinence among HVPG responders compared with non‐responders.

In the studies included in the meta‐analysis, decreases in HVPG were dichotomized into “responders” and “non‐responders,” but more granularity would be obtained if response would be described as absolute changes in HVPG (in mm Hg), and this should be the goal of future studies.

The second most common etiology of cirrhosis was viral hepatitis, with all the studies analyzed carried out before direct antiviral agent (DAA) therapy for hepatitis C virus–related cirrhosis. The impact of NSBB‐related HVPG reduction on outcomes in patients treated with DAA should be analyzed. There is nowadays a tremendous increase in the number of patients with nonalcoholic steatohepatitis (NASH)‐related cirrhosis. It has been recently demonstrated that, in decompensated patients with NASH‐related cirrhosis, HVPG may underestimate portal pressure. ¹³ , ¹⁴ The real impact of different thresholds of HVPG reduction or the absolute changes in millimeters of mercury (mm hg) on outcomes in these patients is still unknown.

Finally, studies to determine whether changes/degree of change in portal pressure (determined invasively or noninvasively) can determine the risk (or lack thereof) of clinical outcomes in an individual patient will be essential before portal pressure measurements can be recommended in a clinical setting.

CONFLICT OF INTEREST

Nothing to report.

Supporting information

Supplementary Material

Turco L, García‐Tsao G. Portal pressure reductions induced by nonselective beta‐blockers improve outcomes and decrease mortality in patients with cirrhosis with and without ascites. Clin Liver Dis. 2022;20:1–4. 10.1002/cld.1210