Complications of portal hypertension: clinical studies

Abstract

INTRODUCTION: With over 1.32 million global liver cirrhosis related deaths annually, the burden of liver disease is still rising, predominantly due to alcohol and the metabolic syndrome. The development of portal hypertension occurs in almost 90% of patients with cirrhosis. It marks the transitin from compensated to decompensated disease, it is the pathophysiological hallmark of cirrhosis related complica4ons, and is the prequel to mortality. Oesophageal varices are the most clinically significant of these complica4ons due to their tendency to rupture and haemorrhage, contribu4ng to 25- 50% of overall deaths. Varices oYen develop before other portal hypertension related complications such as ascites, hepatorenal syndrome and hepa4c encephalopathy. Despite significant developments in understanding the pathophysiological mechanisms of portal hypertension over the last few decades, treatment options for this condition are dispropor4onately limited. Non-selec4ve beta blockers (NSBBs) have been the cornerstone of therapy for 40-years; they reduce portal pressure by reducing cardiac output and increasing splanchnic vascular resistance. NSBBs are most oYen initiated upon the finding of oesophageal varices on screening endoscopy to prevent a first bleeding episode, or aYer stabilisa4on of a patient admi;ed to hospital with oesophageal variceal bleeding (OVB) to prevent rebleeding. Despite clinicians’ familiarity with oesophageal varices and OVB, no clear strategy for primary or secondary prophylaxis of OVB exists, with endoscopic band ligation (EBL) of varices oYen used in place of, or as an adjunct to NSBBs. Furthermore, reduction in portal pressure with more invasive procedures such as transjugular intra-hepatic portosystemic shunt (TIPSS) has been used as rescue therapy in emergency OVB situations as well as for pharmacotherapy refractory ascites. However, TIPSS placement within 72 hours of successfully treated OVB has also been adopted for use “pre-emptively” to prevent rebleeding in high-risk patients. Carvedilol, a NSBB that also exhibits intrinsic anti-α1 adrenergic effects, offers interest to clinicians due to its superior effect on portal pressure reduction compared to other NSBBs, improvement in intrahepatic vascular resistance, and potential influence on pleiotropic mechanisms that contribute towards liver decompensation out with portal hypertension. Regardless, data supporting novel approaches in the management of portal hypertension are lacking, with studies oYen limited by low patient numbers, observational design or absence of long-term outcomes.

AIMS: The aims of the studies presented in this thesis were: (i) to determine if early / preemptive TIPSS (pTIPSS) offers survival benefit over modern standard of care (SOC) following OVB; (ii) to determine if the timing of pTIPSS insertion following OVB alters patient outcome; (iii) to determine the long-term outcomes for patients receiving either carvedilol or EBL following OVB; and (iv) to determine the long-term outcomes for patients receiving pTIPSS.

METHODS: 1. In a two-centre open-label parallel-group randomised controlled trial (RCT), pa4ents with cirrhosis and OVB were recruited following haemostasis with vasoac4ve drugs and EBL. Participants were randomised to SOC (EBL + carvedilol) or pTIPSS (formerly known as “early-TIPSS”). The primary outcome was 1-year survival, secondary outcomes included early and late re-bleeding, and other complica4ons of portal hypertension.

2. A multicentre cohort study of con4nuous, unselected pa4ents referred to four UK tertiary centres for pTIPSS between 01/01/2010 – 31/12/2018. Time from OVB to pTIPSS was recorded and pre-defined clinically relevant outcomes were observed relative to two groups: early pTIPSS (<72hrs of achieving endoscopic haemostasis) and late pTIPSS (72 hours – 28 days). Primary outcome was 1-year transplant free survival. Patients’ clinical progress was observed from pTIPSS inser4on to the close of the study, 31/12/2020.

3. Long-term follow up of a multi-centre, randomised controlled trial in which cirrhotic patients with OVB presenting between 2006-2011 were randomised to receive either carvedilol or EBL as secondary prophylaxis of OVB. Follow-up was undertaken to April 2020 by review of electronic patient records. The primary outcome was long-term survival. Other outcomes including variceal rebleeding and liver decompensation events were compared.

4. Long-term follow-up of a RCT (above) in which cirrhotic patients presenting with OVB were randomised to receive either SOC or pTIPSS. Follow up was extended to three years from recruitment, with the study closing on 04/01/2021. Primary outcome was 3-year transplant free survival on intention to treat and per-protocol analysis. Secondary outcomes included cause of death, rates of variceal rebleeding, rates of ascites and rates of encephalopathy.

RESULTS: 1. pTIPSS RCT: 58 patients (58±11.12 years; 32.7% female) were randomised. AYer one year, seven patients died in the standard of care group and six in the early-TIPSS group, a 1-year survival of 75.9% versus 79.3% respec4vely (p=0.79). Variceal rebleeding occurred in eight patients in the standard of care group compared with three patients in the early-TIPSS group (p=0.09). Not all participants randomised to early-TIPSS received the intervention within the required 72 hours (13 within 72hrs, ten between 3-5 days, and no TIPSS placed in six. For those receiving TIPSS perprotocol, variceal rebleeding rates were reduced (0.0% versus 27.6%, p=0.04) but this had no effect on survival (76.9% versus 75.9%, p=0.91). Serious adverse events were similar in both treatment groups, except rates of hepatic encephalopathy which were higher in patients receiving TIPSS (46.1% vs. 20.7%, p<0.05).

2. UK pTIPSS cohort study: 171 pa4ents were observed, 83 received early pTIPSS and 88 received late pTIPSS. Baseline characteris4cs were similar with no requirement for propensity score matched analysis. There was no difference between early and late pTIPSS groups for the predefined outcomes; 1-year transplant free survival rate (69.9% vs. 71.6%, p=0.73, HR 0.91, 95% CI 0.52-1.58), long-term survival (p=0.52, HR 1.132, 95% CI 0.77 – 1.65), variceal rebleeding (4.8% vs. 11.4%, p=0.09, HR 0.411, 95% CI 0.14-1.17), hepatic encephalopathy (43.9% vs. 34.6%, p=0.26), and new or worsening ascites (16.6% vs. 13.5%, p=0.79). Death due to liver failure was significantly more prevalent in those undergoing early pTIPSS compared to late pTIPSS (44.0% vs. 16.0%, p=0.046, HR 2.79, 95%CI 1.02-8.32).

3. Carvedilol vs. EBL secondary prophylaxis RCT – long-term follow-up: Of the 64 patients recruited and randomised, 26 out of 33 received carvedilol in the follow-up period and 28 out of 31 attended for regular EBL sessions. The median number of follow-up days for all patients recruited was 1459 (SE = 281.74). On intention to treat analysis, there was a trend towards improved survival in the carvedilol group (p=0.09). On per-protocol analysis, carvedilol use was associated with improved longterm survival (p=0.005, HR 3.083, 95%CI 1.397-6.809), fewer liver related deaths (0% vs. 22.8%, p=0.013, OR ∞, 95%CI 1.565 - ∞), and fewer unscheduled hospital admissions with decompensated liver disease (12.0% vs. 64.3% (p=0.0002, OR 13.2, 95%CI 3.026 – 47.23) compared to the EBL group. There was no statistically significant difference in variceal rebleeding rates.

4. pTIPSS RCT – long-term follow-up: On intention to treat analysis, 3-year transplant free survival rate in the SOC group was significantly higher than that of the pTIPSS group (55.2% vs. 20.1%, p=0.006, HR 2.5, 95%CI 1.3-4.87). On per-protocol analysis, 3-year transplant free survival rate in the SOC group was, again, significantly higher than that of the pTIPSS group (55.2% vs. 15.4%, p=0.03, HR 2.93, 95%CI 1.27-7.94). There were significantly higher rates of sepsis related death or sepsis induced liver decompensation related death in the pTIPSS group compared to the SOC group (48.2% vs. 3.6%, p<0.001, reciprocal of RR 13.0, 95% CI 2.46 – 75.45). There were no differences in other outcomes associated with portal hypertension on intention to treat analysis.

CONCLUSION: In a randomised trial, early / pTIPSS had no effect on 1-year transplant free survival in high-risk patients presenting with OVB when compared to SOC using EBL and carvedilol. An important finding was that pTIPSS within the recommend 72 hours timeframe may not be feasible in many centres due to its semi-elective nature set amongst real world clinical practice and emergency care. The 72-hour timeframe is somewhat arbitrary and based on historical variceal rebleeding data. In a novel, dedicated cohort study to investigate this, placement of pTIPSS within 72 hours offered similar short and long-term survival benefit compared to pTIPSS placed between 72 hours – 28 days. This observation may help improve access to pTIPSS in centres wishing to pursue this service for their patients. However, early pTIPSS may be associated with increased risk of liver failure related mortality which raises concerns over the current patient selection recommendations. Given the reasonable 1-year survival rates reported for cirrhotic patients following OVB, it is sensible to explore the impact of the interventions used over a long-term period, particularly as corresponding data are lacking. Carvedilol use was associated with long-term survival benefit, fewer liver failure related deaths and fewer hospital admissions with decompensated disease when compared to EBL, in the setng of a RCT. Furthermore, in a separate long-term follow-up study, pTIPSS was associated with significantly reduced rates of transplant free survival at 3-years compared to SOC using EBL and carvedilol together. This may be due to higher rates of sepsis related mortality observed for those receiving pTIPSS. However, it is unclear whether pTIPSS is an independent risk factor for sepsis, or whether carvedilol protects against sepsis and subsequent acute decompensation in an otherwise at-risk population. Further large, randomised studies are required to validate these findings.