Abstract
Background and aims
Hemodynamic response to pharmacotherapy improves survival in patients with cirrhosis post variceal bleeding, but long-term outcomes remain unexplored especially in this part of the world. We aimed to study the long-term impact of portal pressure reduction on liver-related outcomes after index variceal bleed.
Methods
Patients with hepatic venous pressure gradient (HVPG) more than 12 mm Hg after index variceal bleed were given non-selective beta-blockers in combination with variceal band ligation. HVPG response was assessed after 4 weeks. Patients were followed up for rebleed events, survival, additional decompensation events and safety outcomes. Rebleed and other decompensations were compared using competing risks analysis, taking death as competing event, and survival was compared using Kaplan–Meier analysis.
Results
Forty-eight patients (29 responders and 19 non-responders) were followed up for a median duration of 45 (24–56) months. Rebleeding rates at 1, 3 and 5 years were 10.3%, 20.7% and 20.7% in responders and 15.8%, 44.7% and 51.1% in non-responders, respectively (Gray's test, P = 0.044). Survival rates at 1, 3 and 5 years were 89.7%, 72.1% and 51.9% in responders and 89.5%, 44% and 37.7% in non-responders, respectively (log-rank test, P = 0.1). Both severity of liver disease (MELD score, multivariate sub-distributional hazards ratio: 1.166 [1.014–1.341], P = 0.030) and HVPG non-response (multivariate sub-distributional hazards ratio: 2.476 [1.87–7.030], P = 0.045) predicted rebleeding risk while survival was dependent only on severity of liver disease (MELD > 12, multivariate hazards ratio: 2.36 [1.04–5.38], P = 0.041).
Conclusion
Baseline severity of liver disease predicted survival and rebleed in these patients. Hemodynamic response, although associated with lower rebleeding rate, had limited impact on survival.
Keywords: non-selective beta-blockers, acute variceal bleed, hepatic venous pressure gradient, hemodynamic response and carvedilol
Abbreviations: AFP, alpha-fetoprotein; AVB, acute variceal bleed; ACLF, acute on chronic liver failure; CT, computed tomography; CTP, Child–Turcotte–Pugh; EASL-CLIF, European Association of Study of Liver Disease – Chronic Liver Failure Consortium; EBL, endoscopic band ligation; EGD, esophagogastroduodenoscopy; HE, hepatic encephalopathy; HRS, hepatorenal syndrome; HVPG, hepatic venous pressure gradient; MELD, model for end-stage liver disease; NSBB, non-selective beta-blockers; SBP, spontaneous bacterial peritonitis
Pharmacotherapy with non-selective beta-blockers (NSBB) occupies an important place in the management of portal hypertension related to chronic liver disease, and these drugs have been shown to be effective across all stages of clinically significant portal hypertension (CSPH).1,2 Apart from being useful for primary and secondary prophylaxis of variceal bleed, NSBB also delays the onset of further decompensation events in asymptomatic patients with CSPH.1,2 Reduction in portal pressure (as assessed by hepatic venous pressure gradient [HVPG]) is the most robust predictor in this context of response to pharmacotherapy.3 Hemodynamic response to drugs including NSBB has been shown to prevent future clinical decompensations in patients with compensated cirrhosis.4, 5, 6, 7, 8 In patients with decompensated cirrhosis, however, the prognostic value of HVPG response in improving survival and preventing further decompensations is less established,9, 10, 11, 12, 13, 14 and long-term data on its efficacy and adverse events are limited.15 Although many trials and a recently published meta-analysis have shown HVPG response to improve outcomes in these patients, the narrow therapeutic window with NSBB makes their judicious use important to avoid adverse events in patients with advanced disease, particularly those with uncontrolled ascites.16,17
The present study is a long-term follow-up analysis of a cohort of decompensated cirrhosis after index variceal bleed given pharmacotherapy with NSBB, in whom HVPG was measured after bleed and after 4 weeks. The clinical outcomes and overall survival in patients who achieved HVPG response with pharmacotherapy were compared with those who did not achieve HVPG response, and predictors of rebleed and survival were identified in these patients.
Patients and Methods
Study Design
The present study is the post-hoc analysis of a randomised controlled trial comparing carvedilol and propranolol (CTRI/2013/10/004119) to assess the efficacy of these agents to reduce HVPG after 4 weeks of pharmacotherapy in patients with cirrhosis after index acute variceal bleed (AVB).18 For patients presenting with index variceal bleed, who were willing to participate in the study, measurement of HVPG was done on day 3–5 of AVB episode. Patients with HVPG more than 12 mmHg were included in the original study. Patients with 1) HVPG was less than 12 mm Hg, 2) invalid measurement of HVPG, 3) acute on chronic liver failure (ACLF) at presentation, 4) hepatocellular carcinoma, 5) renal dysfunction, 6) Budd–Chiari syndrome and with aetiology of portal hypertension other than cirrhosis and 7) patients with isolated gastric variceal bleeding were excluded from study. These patients were randomised to receive either carvedilol or propranolol after first HVPG assessment, and the dose was titrated to achieve a target heart rate of 55–60 beats per minute, as described later. The first HVPG measurement was done during the episode of AVB, 12 h after discontinuation of vasoactive agents. A repeat HVPG measurement was done after 4 weeks of NSBB therapy. Depending on hemodynamic response (described later), patients were stratified as HVPG responders and non-responders. Maximal tolerated doses of NSBB were continued in both groups, and no other agent was added in patients without hemodynamic response.
The present study is the long-term follow-up of these patients conducted over a period of 7 years from January 2013 to October 2019 and aimed to establish the outcomes (stratified by HVPG response) while maintained at optimum drug dose. For the purpose of this study, these patients were clinically followed up till either the end of the study period (October 2019) or earlier if they experienced a liver-related event. A liver-related event was defined as mortality related to liver disease or requirement of liver transplant. For patients who required discontinuation of drugs, the follow-up period was restricted till the duration when the patient was on the drug. Informed consent was taken from all patients, and all procedures performed in studies were in accordance with the ethical standards of the institutional ethics committee and are in accordance with the 1975 Helsinki declaration and its later amendments or comparable ethical standards. All authors approved the final version of the manuscript.
Methodology
Management Strategy
Patients were followed up after dose titration of NSBB and documentation of HVPG response on out-patient basis at intervals of 1–3 months and as inpatient when required. Missing data from patients lost to follow-up was completed through telephonic interviews as far as possible, wherein pre-defined outcomes were assessed. The dose in propranolol group was increased in increments of 20–40 mg every third day (maximum dose 160 mg) until they achieved the target heart rate or had intolerance, requiring dose reduction. Similarly, doses in carvedilol group was increased in increments of 3.125 mg (maximum dose 25 mg). All patients received EBL prophylaxis for variceal eradication along with NSBB. In addition, standard management for treatment of aetiology (wherever feasible) and for other complications of liver disease was given. At each visit clinical, laboratory evaluations were done and prognostic scores including Child–Turcotte–Pugh (CTP) and model for end-stage liver disease (MELD) were calculated. Screening for hepatocellular carcinoma was done using regular serum alpha-fetoprotein (AFP) level assessment, supplemented by imaging including ultrasound examinations and multiphase computed tomography (CT) as needed. Information on prespecified liver-related complications like rebleeding, new onset/worsening ascites, hepatic encephalopathy (HE), spontaneous bacterial peritonitis (SBP), hepatorenal syndrome (HRS) and ACLF that may have occurred since the previous visit was collected. For patients who had rebleeding, esophagogastroduodenoscopy (EGD) was done to ascertain the cause of bleed. The bleed was categorised as variceal and non-variceal. Adverse events attributable to beta-blocker use were assessed at each visit and were categorized by type, and for those having recalcitrant symptoms, dose reduction of drug was done. In case of no response to dose reduction, beta-blockers were discontinued.
Measurement of HVPG and Secondary Prophylaxis With Endoscopic Band Ligation (EBL)
Both the procedures are described in detail in supplementary information.
Definitions
HVPG response in each group was defined as in the original study, i.e. more than 20% reduction in HVPG or an absolute reduction to less than 12 mm Hg after 4 weeks of treatment with NSBB when compared with baseline (post index bleed).19 Rebleeding from failure of secondary prophylaxis was defined as a single episode of clinically significant rebleeding from portal hypertensive sources after day 5 following an episode of AVB. Variceal bleed was defined as per Baveno VI consensus.1 All other aetiologies of bleeding were clubbed as non-variceal bleed. Grades of ascites was defined as per EASL definition.20 Varices at baseline were graded according to Westaby classification. For the purpose of this study, new onset ascites was defined as appearance of ascites in a patient who did not have it previously. Worsening ascites was defined as increasing grades of ascites on follow-up. HE was defined according to West-Haven criteria. HRS21 and SBP were assessed as per standard definitions. ACLF was defined as per European Association of Study of Liver Disease – Chronic Liver Failure Consortium (EASL-CLIF) definition.22
Study Outcomes
The primary end point of this study was comparison of rebleeding and survival during the course of follow-up between HVPG responders and non-responders. The secondary end points were development of additional liver-related decompensation like new onset/worsening ascites, HE SBP, HRS and ACLF between two groups. Safety outcomes like incidence of adverse events (irrespective of type of NSBB used) and compliance to these drugs was assessed in both groups.
Statistical Analysis
Baseline data of the patients available for follow-up from the original study was recorded as number (%) or mean ± SD/median (inter quartile range) as appropriate, based on normalcy of distribution. Parameters recorded at the time of first bleed were compared between carvedilol and propranolol group using chi-square test/Fischer exact test for categorical variables and Student's t-test for continuous variables with normal distribution, whereas continuous variables with non-normal distribution were compared using independent samples Kruskal-Wallis test. For all statistical tests, a P value < 0.05 was considered statistically significant.
Potential predictors of rebleeding were assessed using univariate and multivariable adjusted Fine and Gray's sub-distributional hazards model. Variables that were significantly different on univariate analysis between those who did and did not rebleeding were considered for inclusion in the multivariate model. Variables measuring similar characteristics were excluded to avoid multicollinearity, and those with limited clinical significance were not included to maintain event per variable rate less than 5, thus preventing unstable estimates. Results were expressed as sub-distributional hazards ratio (sHR) with 95% confidence intervals and P-values for both univariate and multivariate analyses. Univariate and multivariate analysis was also done for identifying predictors of survival using Cox-proportional hazards model, with variable selection similar to that for predictors of rebleeding. Results were expressed as hazards ratio with 95% confidence intervals and were presented using Forest plot.
Cumulative incidence of rebleeding, survival and other decompensating events was calculated via competing risk analysis. Further details of competing risk analysis are described in supplementary text.
All data was entered using Microsoft Excel 2011 and was analysed using RStudio. In addition to the base packages in R, ggplot2, survival, survminer, cmprsk, tidyverse and readxl packages were used.
Results
A total of 129 patients of cirrhosis presenting with variceal bleeding were evaluated for the purpose of inclusion during the study period from June 2013 to November 2013 (recruitment period of original RCT). Sixty-four patients underwent HVPG measurement within 3–5 days of presentation. Fifty-nine patients with HVPG more than 12 mm Hg were included in the original study and started on pharmacotherapy. With 2 additional patients not followed up, hemodynamic response was seen in 36 out of 57 patients started on pharmacotherapy. For the present study, 48 patients with follow-up longer than 1 year following index variceal bleed were available for analysis. Of these 48 patients, 29 (60.4%) were HVPG responders, whereas remaining 19 (39.6%) were non-responders (Figure 1).
Figure 1.
Overview of the recruitment of patients for the current study. Out of 129 patients, 59 patients with hepatic venous pressure gradient (HVPG) >12 mm Hg were randomised to achieve hemodynamic response. For the present study 48 of these 59 patients were available. Out of 48 patients, 29 were HVPG responders and 19 were HVPG non-responders.
Baseline Characteristics of Hepatic Venous Pressure Gradient Responders and Non-responders
Baseline clinical and demographic characteristics, aetiology and severity of liver disease in both the groups were similar (Table 1). Patients were predominantly male (89.6%), and alcohol (58.3% overall) was the most common aetiology in both groups. Among both groups, number of patients actively consuming alcohol were comparable (10% responders vs 15% non-responders, P = 0.462). Patients who had abstained from alcohol were compliant with no reported consumption of alcohol over the follow-up period. Among other aetiologies, eight patients with hepatitis C virus (HCV) had detectable HCV RNA and were administered antivirals (newly diagnosed case of HCV). Seven out of eight patients achieved sustained virological response after treatment. The other three patients with HCV were previously diagnosed patients and had received treatment in past. Their HCV RNA levels were undetectable and hence antivirals were not given. Severity of liver disease as estimated using Child–Pugh score (responders 7 [6–8] vs non-responders 9 [8–10], P = 0.406) and MELD score (responders 11.4 ± 3.8 vs non-responders 12 ± 2.9, P = 0.545) were comparable between both the groups. Overall, ascites was present at baseline in 63.3% of patients (62.1% responders vs 63.3% non-responders, P = 0.109). Of these patients, 30% of them had grade 2/3 ascites. Apart from ascites and bleed, no other overt clinical decompensations were present at the time of recruitment in both the groups. The type of NSBB (carvedilol or propranolol) and the dose administered were statistically similar in both groups, with median dose of carvedilol being 6.25 mg/day (Inter-quartile range [IQR]: 3.125–12.5 mg) and that of propranolol being 40 mg/day (IQR: 40–80 mg). Target heart rate achieved was also similar in both groups (responders 62 ± 3 bpm and non-responders 63 ± 3 bpm, P = 0.462). Baseline EGD findings were similar including grades of varices and proportion of patients with red colour signs on varices. Patients were followed up for a median duration of 51 (30–72) months and 45 (27–60) months in responder and non-responder group, respectively.
Table 1.
Comparison of Demographic Profile and Baseline Parameters in Patients With HVPG Response (n = 29) and Without HVPG Response (n = 19) Group.
| Variables | HVPG responders (n = 29) | HVPG non-responders (n = 19) | P value |
|---|---|---|---|
| Demographic details | |||
| Age (years) (Mean ± SD) | 42.2 ± 12.1 | 46.7 ± 10.5 | 0.192 |
| Male: female (%) | 27 (93.1):2 (6.9) | 16 (84.2):3 (15.8) | 0.324 |
| Aetiology | 0.661 | ||
| HBV (%) | 3 (10.3) | 0 (0) | |
| HCV (%) | 7 (24.1) | 4 (21.1) | |
| Alcohol (%) | 16 (55.2) | 12 (63.2) | |
| NAFLD (%) | 1 (3.4) | 1 (5.3) | |
| Cryptogenic (%) | 2 (6.9) | 2 (10.5) | |
| Baseline parameters | |||
| Child–Pugh score | 7 (6–8) | 9 (8–10) | 0.406 |
| Child A | 9 (31%) | 3 (15.8%) | 0.479 |
| Child B | 17 (58.6%) | 14 (73.7%) | 0.191 |
| Child C | 3 (10.3%) | 2 (10.5%) | 0.18 |
| MELD | 11.4 ± 3.8 | 12 ± 2.9 | 0.545 |
| Ascites (%) | 18 (62.1) | 12 (63.2) | 0.939 |
| Grade 1 | 6 (33.3) | 3 (25) | |
| Grade 2 | 9 (50) | 6 (50) | |
| Grade 3 | 3 (16.7) | 3 (25) | |
| Oesophageal varices | 0.456 | ||
| Grade 1 | 1 | 1 | |
| Grade 2 | 12 | 8 | |
| Grade 3 | 16 | 10 | |
| Red colour signs | 24 | 15 | |
| Haemoglobin (g/dl)) (Mean ± SD) | 9.3 ± 1.9 | 9.8 ± 2 | 0.469 |
| Platelet count (mm3) | 85,000 (55,000–132,000) | 75,000 (55,000–123,000) | 0.41 |
| Bilirubin (mg/dl) | 2 (0.9–3) | 1.7 (0.9–2.8) | 0.519 |
| AST (IU/L) | 58 (45–98) | 56 (36–97) | 0.435 |
| ALT (IU/L) | 52 (35–80) | 48 (36–66) | 0.448 |
| ALP (IU/L) | 287 (150–400) | 272 (160–360) | 0.67 |
| Creatinine (mg/dl) (Mean ± SD) | 0.8 ± 0.1 | 0.8 ± 0.1 | 0.959 |
| Albumin (g/dl) (Mean ± SD) | 3.4 ± 0.5 | 3.37 ± 0.6 | 0.742 |
| INR | 1.2 ± 0.19 | 1.3 ± 0.2 | 0.153 |
| LSM at baseline (kPa) | 28 (18–66) | 34 (25–56) | 0.914 |
| Treatment parameters | |||
| Type of NSBB | 0.186 | ||
| Carvedilol (%) | 18 (62.1) | 7 (36.8) | |
| Dose (mg/day) | 6.25 (3.125–12.5) | 6.25 (3.125–12.5) | |
| Propranolol (%) | 11 (37.9%) | 12 (63.2%) | |
| Dose (mg/day) | 40 (40–80) | 40 (40–80) | |
| HVPG (baseline) (mmHg) | 17 ± 2.9 | 17.7 ± 2.7 | 0.51 |
| HVPG (4 weeks) (mmHg) | 11.5 ± 3.2 | 15.4 ± 2.4 | 0.001 |
| Heart rate achieved (beats/min) (Mean ± SD) | 62 ± 3 | 63 ± 3 | 0.462 |
Data are presented as median (IQR) for quantitative variables and as n(%) for qualitative variables unless otherwise specified. List of abbreviations: AST, aspartate transaminase; ALT, alanine transaminase; ALP, alkaline phosphatase; CTP, Child–Turcotte–Pugh; HBV, hepatitis B virus; HCV, hepatitis C virus; HVPG, hepatic venous pressure gradient; INR, international normalised ratio; MELD, model for end-stage liver disease; NAFLD, non-alcoholic fatty liver disease.
Rebleeding (Table 2)
Table 2.
Comparison of Follow-up, Decompensation Events, Liver-Related Outcomes and Adverse Effects Between Patients With HVPG Response (n = 29) and Without HVPG Response (n = 19) Group.
| Outcomes | HVPG responders (n = 29) | HVPG non-responders (n = 19) | P value |
|---|---|---|---|
| Rebleed (%) | 6 (20.7) | 9 (47.4) | 0.051 |
| Site of rebleed | |||
| Esophageal varix (%) | 3 (50) | 6 (66.3) | 0.051 |
| Fundal varix | 0 | 0 | |
| Post EBL ulcer | 3 (50) | 3 (33.7) | 0.132 |
| Ascites (new onset/worsening) (%) | 13 (44.8) | 13 (68.4) | 0.109 |
| HE (%) | 6 (20.7) | 8 (42.1) | 0.11 |
| HRS (%) | 7 (24.1) | 4 (21.1) | 0.804 |
| SBP (%) | 9 (31) | 1 (5.3) | 0.032 |
| HCC (%) | 1 (3.4) | 2 (10.5) | 0.322 |
| ACLF (%) | 4 (13.8) | 2 (10.5) | 0.738 |
| Deaths (%) | 17 (58.6) | 7 (36.8) | 0.14 |
| Variceal eradication (%) | 26 (89.6) | 15 (78.9) | 0.132 |
| Side effects | |||
| Fatigue (%) | 10 (34.5) | 6 (31.6) | 0.835 |
| Hypotension (%) | 6 (20.6) | 5 (26.3) | 0.253 |
| Depression (%) | 2 (6.9) | 1 (5.3) | 0.451 |
| Dose reduction needed | 7 (24.1%) | 4 (22.2%) | 0.88 |
| Drug discontinuation needed | 4 (13.8%) | 6 (33.3%) | 0.112 |
| Follow-up duration (months) | 51 (30–72) | 45 (27–60) | 0.456 |
Data are presented as n (%) for qualitative variables and as mean ± SD for quantitative variables. List of abbreviations: ACLF, acute on chronic liver failure; EBL, endoscopic band ligation; HCC, hepatocellular carcinoma; HE, hepatic encephalopathy; HRS, hepatorenal syndrome; HVPG, hepatic venous pressure gradient; SBP, spontaneous bacterial peritonitis.
Overall, six (20.1%) patients among HVPG responders and nine (44.7%) among HVPG non-responders had rebleeding (P = 0.044) over follow-up. Bleed from oesophageal varices was seen in nine patients (three responders and six non-responders), and six patients had non-variceal bleed (all patients had post band ligation ulcers: three responders and three non-responders). The 1-year, 3-year and 5-year cumulative rebleeding rates calculated taking death as a competing event were 10.3%, 20.7% and 20.7% for responders and 15.8%, 44.7% and 51.1% for non-responders, respectively (Figure 2) (Gray's test; P = 0.044). On univariate analysis, CTP (P = 0.024) and MELD (P = 0.033) score at first bleed, with lower serum albumin levels (P = 0.029), the presence of ascites (P = 0.004) and HVPG non-response (P = 0.036) were predictive of rebleeding. On multivariable adjusted competing risks analysis, only higher MELD score (sub-distributional hazards ratio [95% CI]: 1.166 [1.014–1.341]; P = 0.031) and HVPG non-response (2.476 [1.87–7.030], P = 0.045) could predict rebleeding, whereas other variables were statistically insignificant (Table 3). Rebleeding risk among hemodynamic responders stratified by type of beta-blocker is outlined in supplementary information. In addition, rebleeding risks stratified by post-bleed HVPG ≥15 mm Hg (the value with best discriminatory power for rebleeding in our analysis) and HVPG response are outlined in supplementary information.
Figure 2.
Competing-risk plots showing cumulative rates of rebleeding (green) and death (blue) across patients with hepatic venous pressure gradient (HVPG) response and HVPG non-response. The red colour represents absence of either of the event. The 1-year, 3-year and 5-year cumulative rebleeding rates calculated taking death as a competing event were 10.3%, 20.7% and 20.7% for responders and 15.8%, 44.7% and 51.1% for non-responders, respectively (Gray's test; P = 0.044).
Table 3.
Multivariate Analysis of Predictors of Rebleeding and Mortality. Prediction of Rebleeding was Done using Fine and Gray's Sub-distributional Hazards model, While Prediction of Mortality was Done Using Cox-proportional Hazards Model With Univariate and Multivariate Hazards Ratios Along With Significance Shown.
| Predictors of rebleeding | ||||
|---|---|---|---|---|
| Univariate sub-distributional hazards ratio (95% CI) | P value | Multivariate sub-distributional hazards ratio (95% CI) | P value | |
| MELD | 1.16 (1.03–1.31) | 0.015 | 1.166 (1.014–1.341) | 0.030 |
| No HVPG response | 2.77 (1.04–7.43) | 0.042 | 2.476 (1.87–7.030) | 0.045 |
| Predictors of mortality | ||||
| Univariate hazards ratio (95% CI) | P value | Multivariate hazards ratio (95% CI) | P value | |
| MELD > 12 | 2.84 (1.27–6.37) | 0.011 | 2.36 (1.04–5.38) | 0.041 |
| No HVPG response | 1.94 (0.87–4.33) | 0.107 | 1.86 (0.83–4.19) | 0.136 |
| Grade-2/3 Ascites at presentation | 3.09 (1.15–8.29) | 0.025 | 2.73 (1.01–7.45) | 0.049 |
HVPG, hepatic venous pressure gradient; MELD, model for end-stage liver disease.
Survival
The 1-year, 3-year and 5-year survival was 89.7%, 72.1% and 51.9%, in responders and 89.5%, 44% and 37.7% in non-responders, respectively (log-rank test; P = 0.1) (Figure 3). Half of the patients followed up (24 of 48) had liver-related death, with no patients undergoing liver transplant. Progressive decompensation of underlying chronic liver disease (12/24 deaths, 50%) was the most common cause of death, followed by repeat gastrointestinal bleeding (7/24, 27.3%), ACLF (4/24, 16.6%) and HCC (1/24, 4.1%). Among deaths associated with gastrointestinal bleeding, four out of seven patients (57.1%) had post band ligation-related ulcers. No difference was seen in the cause of deaths when stratified by HVPG response. On univariate analysis, baseline prognostic scores like CTP (P = 0.024) and MELD (P = 0.002) scores as well as their individual components including serum albumin (P = 0.017) and INR (P = 0.023) were associated with mortality. On multivariable adjusted Cox-proportional hazards model, while MELD score more than 12 (hazard ratio [95% CI]: 2.4 [1.04–5.4]; P = 0.04) and grade 2/3 ascites at baseline (HR: 2.7 [1.01–7.4]; P = 0.049) were associated with mortality, HVPG non-response (HR: 1.9 [0.83–4.2]; P = 0.134) could not independently predict decreased survival (Supplementary Figure 1). None of the patient underwent liver transplant as our centre is a low-volume transplant centre with predominantly deceased donor transplant programme. Survival in hemodynamic responders stratified by type of beta-blocker is outlined in supplementary information.
Figure 3.
Kaplan–Meier plots demonstrating overall survival rates in patients with and without hepatic venous pressure gradient response. The 1-year, 3-year and 5-year survival was 89.7%, 72.1% and 51.9%, in responders and 89.5%, 44% and 37.7% in non-responders, respectively (log-rank test; P = 0.1).
Other Decompensations
Multiple additional decompensation events were noted in both groups over the course of follow-up (Table 2). While the incidence of new onset/worsening ascites, HE and HRS was comparable in both groups, SBP developed more frequently among HVPG responders using actuarial probability rates and using competing risk analysis (Gray's test, P = 0.032) (Supplementary Figure 2). Clinical decompensations stratified by type of beta-blocker are outlined in supplementary information.
Safety Outcomes
Adverse events were common on both drugs and occurred at near-similar frequency. The most common adverse events were fatigue (34.5% for responders and 31.6% for non-responders, P = 0.835), and hypotension (20.6% responders and 26.3% non-responders, P = 0.253). Overall, 37.3% of patients in responder group and 31.4% of patients required reduction/discontinuation of pharmacotherapy due to adverse events.
Discussion
In the present study of long-term follow-up of patients with decompensated cirrhosis, hemodynamic response in portal pressure with pharmacotherapy although associated with reduced rebleeding, did not significantly affect the risk of future decompensation events and the impact on overall survival was limited. Baseline severity of liver disease was the most important predictor of survival and risks of future decompensations and rebleeding.
The complex interplay of higher portal pressure, hyperdynamic circulation and lower cardiac output makes pharmacotherapy a challenge in patients with decompensated cirrhosis.23,24 Its use in these patients is further complicated due to narrow therapeutic window16 with frequent adverse effects. Additionally, logistics of performing a repeat invasive procedure in these patients to determine response restricts the widespread use of assessment of hemodynamic response in patients with decompensated cirrhosis. We in our study attempted to study the long-term impact of HVPG reduction in patients with decompensated cirrhosis. In our study, majority of included patients were of Child B class with a median MELD score of 12. In addition to variceal bleed, 60% of patients had ascites at baseline at the time of inclusion with 30% of these patients having grade 2/3 (uncontrolled) ascites, indicating fairly advanced liver disease. We had used propranolol and carvedilol in tolerated doses as a part of the original trial, and the same doses were continued for the purpose of this study.
Rebleeding rate over follow-up was lower in HVPG responders, as anticipated. However, survival and incidence of other decompensations were comparable with both groups except SBP, which was significantly more common in HVPG responders. On multivariate analysis, only higher MELD score (>12) and the presence of uncontrolled ascites could predict survival whereas HVPG non-response was not associated with increased non-bleed-related decompensations and improved survival, although there was a trend in that direction. Our findings suggest that in decompensated cirrhosis, baseline severity of liver disease plays a major role in deciding the prognosis rather than reduction in portal pressure. These findings are different from the previously published literature on prognostic influence of portal pressure reduction in decompensated cirrhosis.7,9, 10, 11,13, 14, 15 These studies and a recently published meta-analysis showed improvement in survival with hemodynamic response in patients with decompensated cirrhosis with/without ascites. We believe that a number of reasons were responsible for these differences. First, follow-up duration of the previous studies except those of Abraldes et al10 and Augustin et al13 was shorter than our study. The present study has a median follow-up of nearly 4 years. Second, we had used propranolol and carvedilol for attaining HVPG response while the previous study had used propranolol/nadolol with or without isosorbide mononitrate. Although a greater number of patients in carvedilol had HVPG response, this could still not translate into improvement in end points barring rebleeding risk in responders. Third, higher proportion of patients with ascites were included in the present study, with only one previous study by Villanueva et al14 having ascites in more than 50% of included patients. Moreover, half of the patients included in our study had grade 2/3 ascites at presentation, whereas severity of ascites was not addressed in any previous study. Fourth, the present study was conducted in the era when directly acting antivirals (DAAs) were available. The treatment of HCV can itself lower the HVPG and prevent clinical decompensation.25,26 Due to a small sample size, we could not demonstrate an impact of DAA in determining the outcome in both groups as seven out of eight patients achieved sustained virological remission. Fifth, discontinuation of alcohol in these patients is known to cause reduction in portal pressures.13 As majority of the patients (90% responders and 85% non-responders) with ALD abstained from alcohol with good compliance, the reduction in HVPG brought upon by alcohol abstinence could have resulted in comparable survival in both groups. However, active alcohol consumption was not a significant factor on univariate analysis for determining the outcome, likely due to a small sample size.
There is a growing evidence that NSBB may be harmful in certain subsets of patients with decompensated cirrhosis like those with refractory ascites, SBP and HRS.17,27,28 Our results support these findings. The incidence of SBP was significantly more in patients who were HVPG responders. Further compromise of already affected hemodynamic responses with NSBB can be harmful in these patients. Attempts to achieve hemodynamic response to pharmacotherapy in these patients can result in intolerance as seen in our study where 30% of our patients in both groups required either dose reduction/discontinuation in view of adverse events.
The strengths of the present study are 1) longer duration of follow-up than most of the previously published literature and 2) use of competing risk analysis to analyse the individual decompensation and overall transplant free survival. The findings of this study also emphasise the need to refer the decompensated patients for evaluation of transplant particularly when they have ascites as nearly 50% of patient had liver-related mortality over the period of follow-up. The study had some obvious limitations. First, the sample size of the present study was small, and some of the patients from the original study were lost to follow-up, and it is possible that end points in these patients were missed and overall outcomes therefore did not reach statistical significance. While detection of decline in rebleeding rates is extensively known with HVPG response, it would have been interesting to see the benefit of HVPG reduction in other clinical end points in this group of patients with a larger sample size. However, the study was a post-hoc analysis of a RCT powered for detecting changes in HVPG response to different beta-blockers. Hence, clinical end points other than rebleeding analysed over the follow-up period may have been underpowered to detect differences. The incidence of SBP was higher among HVPG responders. However, beta-blockers are known to reduce the incidence of this complication. Due to a small sample size, there is a possibility of Type 2 error and this finding needs clarification in prospective studies with larger sample size. Second, the durability of attained hemodynamic response. It is possible that some of the responders may have eventually lost the response over the follow-up as repeat HVPG measurements was not done in the follow-up period. Third, the study was carried out in a time where therapies for aetiology like HCV were available and is known to cause reduction in HVPG.26 Similarly, alcohol abstinence is known to be associated with maintenance of HVPG response.13 As majority of these patients discontinued alcohol, the overall benefit of HVPG reduction on pharmacotherapy might have been confounded by these factors.
In conclusion, baseline severity of liver disease is the most important predictor of bleed-related and non-bleed-related outcomes in patients with decompensated cirrhosis. Benefit of attaining hemodynamic response to pharmacotherapy may be restricted to prevention of rebleed with limited impact on survival and other decompensations particularly in patients with increased ascites.
Ethical approval
Obtained (IESC/T-120/01.03.2013, RTx-09/03.05.2013 and RT-36/15.06.2013.)
CRediT authorship contribution statement
Sanchit Sharma: Conceptualization, data collection, writing of original draft and revision. Samagra Agarwal: Conceptualization, writing of original draft and revision. Deepak Gunjan: Critical revision of draft. Kanav Kaushal: Writing of draft, revision of draft. Abhinav Anand: Writing of draft, revision of draft. Srikant Gopi: Writing of draft, revision of draft. Srikant Mohta: Writing of draft, revision of draft. Anoop Saraya: Study design, Conceptualization, Visualization, study overview, critical revision of draft.
Conflicts of interest
The authors have none to declare.
Funding
None.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jceh.2020.11.001.
Appendix A. Supplementary data
The following are the supplementary data to this article:
figs1.
figs2.
References
- 1.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;63:743–752. doi: 10.1016/j.jhep.2015.05.022. [DOI] [PubMed] [Google Scholar]
- 2.Villanueva C., Albillos A., Genescà J. β blockers to prevent decompensation of cirrhosis in patients with clinically significant portal hypertension (PREDESCI): a randomised, double-blind, placebo-controlled, multicentre trial. Lancet. 2019;393:1597–1608. doi: 10.1016/S0140-6736(18)31875-0. [DOI] [PubMed] [Google Scholar]
- 3.Albillos A., Bañares R., González M. Value of the hepatic venous pressure gradient to monitor drug therapy for portal hypertension: a meta-analysis. Am J Gastroenterol. 2007;102:1116–1126. doi: 10.1111/j.1572-0241.2007.01191.x. [DOI] [PubMed] [Google Scholar]
- 4.Groszmann R.J., Bosch J., Grace N.D. Hemodynamic events in a prospective randomized trial of propranolol versus placebo in the prevention of a first variceal hemorrhage. Gastroenterology. 1990;99:1401–1407. doi: 10.1016/0016-5085(90)91168-6. [DOI] [PubMed] [Google Scholar]
- 5.Merkel C., Bolognesi M., Sacerdoti D. The hemodynamic response to medical treatment of portal hypertension as a predictor of clinical effectiveness in the primary prophylaxis of variceal bleeding in cirrhosis. Hepatology. 2000;32:930–934. doi: 10.1053/jhep.2000.19322. [DOI] [PubMed] [Google Scholar]
- 6.Turnes J., Garcia-Pagan J.C., Abraldes J.G., Hernandez-Guerra M., Dell'Era A., Bosch J. Pharmacological reduction of portal pressure and long-term risk of first variceal bleeding in patients with cirrhosis. Am J Gastroenterol. 2006;101:506–512. doi: 10.1111/j.1572-0241.2006.00453.x. [DOI] [PubMed] [Google Scholar]
- 7.Villanueva C., Aracil C., Colomo A. Acute hemodynamic response to beta-blockers and prediction of long-term outcome in primary prophylaxis of variceal bleeding. Gastroenterology. 2009;137:119–128. doi: 10.1053/j.gastro.2009.03.048. [DOI] [PubMed] [Google Scholar]
- 8.Sharma P., Kumar A., Sharma B.C., Sarin S.K. Early identification of haemodynamic response to pharmacotherapy is essential for primary prophylaxis of variceal bleeding in patients with “high-risk” varices. Aliment Pharmacol Ther. 2009;30:48–60. doi: 10.1111/j.1365-2036.2009.04015.x. [DOI] [PubMed] [Google Scholar]
- 9.Villanueva C., Miñana J., Ortiz J. Endoscopic ligation compared with combined treatment with nadolol and isosorbide mononitrate to prevent recurrent variceal bleeding. N Engl J Med. 2001;345:647–655. doi: 10.1056/NEJMoa003223. [DOI] [PubMed] [Google Scholar]
- 10.Abraldes J.G., Tarantino I., Turnes J., Garcia-Pagan J.C., Rodés J., Bosch J. Hemodynamic response to pharmacological treatment of portal hypertension and long-term prognosis of cirrhosis. Hepatology. 2003;37:902–908. doi: 10.1053/jhep.2003.50133. [DOI] [PubMed] [Google Scholar]
- 11.García-Pagán J.C., Villanueva C., Albillos A. Nadolol plus isosorbide mononitrate alone or associated with band ligation in the prevention of recurrent bleeding: a multicentre randomised controlled trial. Gut. 2009;58:1144–1150. doi: 10.1136/gut.2008.171207. [DOI] [PubMed] [Google Scholar]
- 12.Villanueva C., Aracil C., Colomo A. Clinical trial: a randomized controlled study on prevention of variceal rebleeding comparing nadolol + ligation vs. hepatic venous pressure gradient-guided pharmacological therapy. Aliment Pharmacol Ther. 2009;29:397–408. doi: 10.1111/j.1365-2036.2008.03880.x. [DOI] [PubMed] [Google Scholar]
- 13.Augustin S., González A., Badia L. Long-term follow-up of hemodynamic responders to pharmacological therapy after variceal bleeding. Hepatology. 2012;56:706–714. doi: 10.1002/hep.25686. [DOI] [PubMed] [Google Scholar]
- 14.Villanueva C., Graupera I., Aracil C. A randomized trial to assess whether portal pressure guided therapy to prevent variceal rebleeding improves survival in cirrhosis. Hepatology. 2017;65:1693–1707. doi: 10.1002/hep.29056. [DOI] [PubMed] [Google Scholar]
- 15.Turco L., Villanueva C., Mura V.L. Lowering portal pressure improves outcomes of patients with cirrhosis, with or without ascites: a meta-analysis. Clin Gastroenterol Hepatol. 2020;18:313–327.e6. doi: 10.1016/j.cgh.2019.05.050. [DOI] [PubMed] [Google Scholar]
- 16.Ge P.S., Runyon B.A. The changing role of beta-blocker therapy in patients with cirrhosis. J Hepatol. 2014;60:643–653. doi: 10.1016/j.jhep.2013.09.016. [DOI] [PubMed] [Google Scholar]
- 17.Sersté T., Melot C., Francoz C. Deleterious effects of beta-blockers on survival in patients with cirrhosis and refractory ascites. Hepatology. 2010;52:1017–1022. doi: 10.1002/hep.23775. [DOI] [PubMed] [Google Scholar]
- 18.Gupta V., Rawat R., null Shalimar, Saraya A. Carvedilol versus propranolol effect on hepatic venous pressure gradient at 1 month in patients with index variceal bleed: RCT. Hepatol Int. 2017;11:181–187. doi: 10.1007/s12072-016-9765-y. [DOI] [PubMed] [Google Scholar]
- 19.D'Amico G., Garcia-Pagan J.C., Luca A., Bosch J. Hepatic vein pressure gradient reduction and prevention of variceal bleeding in cirrhosis: a systematic review. Gastroenterology. 2006;131:1611–1624. doi: 10.1053/j.gastro.2006.09.013. [DOI] [PubMed] [Google Scholar]
- 20.Angeli P., Bernardi M., Villanueva C. EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis. J Hepatol. 2018;69:406–460. doi: 10.1016/j.jhep.2018.03.024. [DOI] [PubMed] [Google Scholar]
- 21.Salerno F., Gerbes A., Ginès P., Wong F., Arroyo V. Diagnosis, prevention and treatment of hepatorenal syndrome in cirrhosis. Gut. 2007;56:1310–1318. doi: 10.1136/gut.2006.107789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Moreau R., Jalan R., Gines P. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology. 2013;144 doi: 10.1053/j.gastro.2013.02.042. 1426–1437, 1437.e1–9. [DOI] [PubMed] [Google Scholar]
- 23.Krag A., Bendtsen F., Henriksen J.H., Møller S. Low cardiac output predicts development of hepatorenal syndrome and survival in patients with cirrhosis and ascites. Gut. 2010;59:105–110. doi: 10.1136/gut.2009.180570. [DOI] [PubMed] [Google Scholar]
- 24.Møller S., Bendtsen F. The pathophysiology of arterial vasodilatation and hyperdynamic circulation in cirrhosis. Liver Int. 2018;38:570–580. doi: 10.1111/liv.13589. [DOI] [PubMed] [Google Scholar]
- 25.Mandorfer M., Kozbial K., Schwabl P. Changes in hepatic venous pressure gradient predict hepatic decompensation in patients who achieved sustained virologic response to interferon-free therapy. Hepatology. 2020;71:1023–1036. doi: 10.1002/hep.30885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lens S., Alvarado-Tapias E., Mariño Z. Effects of all-oral anti-viral therapy on HVPG and systemic hemodynamics in patients with hepatitis C virus-associated cirrhosis. Gastroenterology. 2017;153:1273–1283.e1. doi: 10.1053/j.gastro.2017.07.016. [DOI] [PubMed] [Google Scholar]
- 27.Kim S.G., Larson J.J., Lee J.S., Therneau T.M., Kim W.R. Beneficial and harmful effects of nonselective beta blockade on acute kidney injury in liver transplant candidates. Liver Transplant. 2017;23:733–740. doi: 10.1002/lt.24744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mandorfer M., Bota S., Schwabl P. Nonselective β blockers increase risk for hepatorenal syndrome and death in patients with cirrhosis and spontaneous bacterial peritonitis. Gastroenterology. 2014;146:1680–1690.e1. doi: 10.1053/j.gastro.2014.03.005. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.