Hepatic hydrothorax as a manifestation of decompensated cirrhosis: An update on current management and future directions

Abstract

Hepatic hydrothorax (HH) is an uncommon yet severe manifestation of portal hypertension which develops in 5%-10% of patients with liver cirrhosis. It typically presents as a unilateral, right-sided pleural effusion and in the context of end-stage liver disease and concomitant ascites. The most widely accepted explanatory model for HH accumulation is the formation of small diaphragmatic defects (pleuroperitoneal connections) facilitating migration of ascitic fluid from the peritoneal cavity directly to the pleural cavity. Medical management involves sodium restriction and diuretic therapy, with thoracentesis also offering symptomatic relief. In cases of refractory HH, a transjugular intrahepatic portosystemic shunt is considered either as definitive treatment or as a bridge to liver transplantation, which remains the only curative treatment option. HH refractory to medical therapy presents a challenging clinical dilemma, particularly in those who are ineligible for liver transplantation. In this mini-review, we aim to highlight the pathophysiology, clinical presentation, diagnosis and management of HH. Additionally, we discuss and appraise novel therapeutic options and offer future directions.

Key Words: Hepatic hydrothorax; Cirrhosis; Ascites; Pleural effusion; Decompensation

Core Tip: Hepatic hydrothorax (HH) commonly presents as a right-sided pleural effusion in patients with cirrhosis after excluding cardiopulmonary and renal causes. The most widely accepted mechanism is migration of ascitic fluid via small diaphragmatic defects directly into the pleural space. Medical management is centred around sodium restriction and diuretics, with thoracentesis also offering symptomatic relief. A transjugular intrahepatic portosystemic shunt (TIPS) is generally first line when HH becomes refractory to medical therapy. Beyond TIPS and liver transplantation, there are few alternatives for those who are ineligible. However, promising modalities such as indwelling pleural catheters, albumin infusions, and continuous terlipressin require further validation.

INTRODUCTION

Hepatic hydrothorax (HH) is an uncommon cause for a transudative pleural effusion in the setting of cirrhosis and portal hypertension, after exclusion of cardiopulmonary or renal causes. It requires a high index of clinical suspicion and timely intervention in order to improve long-term survival and prognosis. Medical management of HH is limited, and includes sodium restriction and diuretic therapy[1-3]. Thoracentesis can be performed but effusions can re-accumulate within days to weeks. Transjugular intrahepatic portosystemic shunt (TIPS) placement is first-line for refractory HH and can also serve as a bridge to liver transplantation[2,4,5]. As these therapies have variable rates of response, liver transplantation is ultimately the only curative option. Given the management complexity and heterogeneity of treatment response, HH poses a considerable challenge to clinicians. This mini-review aims to comprehensively outline HH pathophysiology and to explore treatment options for patients who are ineligible for liver transplantation. Considering the limitations this inherently presents, we also provide a thorough discussion on future advances in the management of HH.

EPIDEMIOLOGY AND RISK FACTORS

HH accounts for only 2%-3% of all pleural effusions[6]. In the liver cirrhosis population, it has an incidence of 5%-10% and carries a high mortality rate with a median survival of only nine months[1,4,7]. The effusion itself characteristically accumulates on the right side and in the context of end-stage liver disease and ascites[1,4]. HH typically presents in the fifth decade of life and has a slight male predominance reported at 63.6%[1].

A recent retrospective study of 947 patients with cirrhosis found that a requirement for recurrent paracentesis, high serum bilirubin, diabetes mellitus, and non-use of a non-selective β-blocker (NSBB) all predicted the development of HH[8]. Of note, recurrent paracentesis requirement reflects massive and diuretic-refractory ascites which conceivably increases the risk of HH by raising intraabdominal pressure. In addition, Bai et al[9] report that a higher model for end-stage liver disease (MELD) score, a dilated portal vein, and biochemical correlates such as D-dimer, immunoglobulin G and prothrombin activity may confer an increased HH risk. Patients with portal vein thrombosis also appear to develop bilateral as opposed to unilateral pleural effusions.

PATHOPHYSIOLOGY

Pleural effusions result from fluid accumulating in the pleural cavity at a rate that exceeds the pleural membrane’s capacity to reabsorb it. HH occurs on a background of chronic liver disease complicated by portal hypertension, and is often accompanied by ascites[4,10]. The accepted pathophysiological basis for HH is direct movement of ascitic fluid from the peritoneal cavity to the pleural cavity via small (typically < 1 cm) diaphragmatic defects[10]. Flow is generally unidirectional into the pleural space, owing to the negative intrathoracic pressures during inspiration, and raised intraabdominal pressure secondary to ascites. This is supported by peritoneal scintigraphy studies using ⁹⁹Tc-sulphor-colloid demonstrating migration of this radioisotope from the peritoneal to the pleural cavity[11,12]. The morphological appearance of diaphragmatic defects is also useful for classification: (1) Type I (no obvious defect); (2) Type II (blebs); (3) Type III (fenestrations); and (4) Type IV (multiple gaps)[13]. Other mechanisms including hypoalbuminaemia, azygous vein hypertension (due to portosystemic shunting), and trans-diaphragmatic drainage of ascitic fluid via lymphatics have also been implicated and likely contribute to varying degrees (Figure 1)[10,14].

Figure 1 Proposed mechanisms of hepatic hydrothorax. Among the proposed explanatory models for hepatic hydrothorax (HH) accumulation, migration of ascitic fluid through diaphragmatic defects is the most widely accepted. Progressive fibrosis and cirrhosis lead to portal hypertension and ascites. Long-standing and uncontrolled ascites forms blebs and fenestrations which permit direct communication between the peritoneal and pleural spaces. In addition, portal hypertension promotes the formation of portosystemic collaterals which in turn worsens azygous venous congestion. Impaired hepatic synthetic function results in hypoalbuminaemia and decreased colloid osmotic pressure. This exacerbates fluid accumulation and third spacing in potential spaces such as the pleural cavity. Finally, redistribution of interstitial fluid via lymphatic networks into the pleural space also contributes to HH accumulation.

There is also a well-described predilection for HH to develop on the right side. It is estimated that 73% of people with HH develop a right-sided pleural effusion, followed by left-sided (17%) and bilateral (10%) effusions[1]. This is thought to be a consequence of embryological development. The left hemidiaphragm, being thicker and having a higher collagen content, makes it intrinsically more resistant to bleb formation and subsequent rupture[14].

CLINICAL PRESENTATION AND DIAGNOSIS

In a patient with known cirrhosis and portal hypertension presenting with a unilateral pleural effusion, a diagnosis of HH should be strongly considered. Patients with HH typically present with dyspnoea at rest (34%), non-productive cough (22%), and pleuritic chest pain (8%) but may also complain of non-specific symptoms such as nausea (11%) and fatigue (7%)[1]. HH is rarely seen outside of decompensated cirrhosis, with just 3.9% of all patients being Child-Pugh class A at the time of presentation. Of note, profound dyspnoea, hypoxia, and even frank respiratory failure can develop following only a modest accumulation of fluid in the pleural space[1,4]. By contrast, some patients can accumulate several litres of ascites and remain relatively symptom-free with only modest abdominal distension. While HH generally co-presents with ascites, up to 9% of patients do not have clinically detectable ascites[1]. In the latter case, HH can develop without ascites owing to the higher resorptive capacity of the peritoneum compared to the pleura[10,15]. In rare cases, HH has been reported as the index presentation for decompensated cirrhosis[16].

After a pleural effusion is confirmed on chest radiography, aspiration of pleural fluid follows for differentiation between a transudative and exudative effusion. This is routinely performed according to Light’s criteria[17]. It is well-documented that some transudates are misclassified as exudates according to Light’s criteria, particularly among patients on long-term diuretic therapy (Table 1)[18,19]. Bielsa et al[18] therefore recommend an adjusted cut-off of a pleural fluid-serum albumin ratio < 0.6 g/dL (6.0 g/L) for transudative effusions when the clinical picture is concerning for HH.

Table 1 Biochemical profiles of transudative vs exudative pleural aspirates.

	Transudate	Exudate
Pleural: Serum protein	< 0.5 g/dL	> 0.5 g/dL
Pleural: Serum LDH	< 0.6 g/dL	> 0.6 g/dL
Pleural fluid LDH	< 2/3 ULN serum LDH	> 2/3 ULN serum LDH
Pleural: Serum albumin1	< 0.6 g/dL	> 0.6 g/dL

¹Adjunct criteria proposed by Bielsa et al[18].

Light’s criteria, as adapted from the British Thoracic Guideline for Thoracic Disease[19]. Pleural fluid is considered an exudate if at least one of the above criteria is met. LDH: Lactate dehydrogenase; ULN: Upper limit of normal.

An important complication of HH is spontaneous bacterial empyema (SBEM), a superimposed infection of the migrated ascitic fluid which is reported in up to 16% of patients[20]. It is a leading cause of death in HH patients with mortality during the acute treatment period up to 20%[21]. A fever with features of spontaneous bacterial peritonitis (SBP) such as abdominal pain and distension should raise clinical concern for SBEM. However, absence of these features should not be used to exclude SBEM. As Xiol et al[21] caution, 43% of SBEM cases are not associated with underlying SBP. A pleural tap can readily distinguish between HH and SBEM. Pleural fluid that is turbid and grossly purulent suggests a superimposed bacterial infection. Microbiological testing yielding a polymorphonuclear leukocyte (PMN) > 250 cells/mm³ with a positive fluid culture, or PMN > 500 cells/mm³ alone favours a diagnosis of SBEM[21,22]. While 81% have a negative pleural fluid culture, the most commonly detected microorganisms are Escherichia coli, Streptococcus species, Enterococcus species, Klebsiella pneumoniae, and Pseudomonas[21]. Pleural fluid biochemical testing is also useful whereby any of low pH (< 7.2), low glucose (< 3 mmol/L), and raised lactate dehydrogenase (> 1000 U/L) are suggestive of SBEM[19,23,24]. Determining which patients have an increased risk of SBEM is challenging as there is a paucity of high-quality studies on risk factors[25]. One study reported that low serum albumin level, prolonged prothrombin time, low pleural fluid protein level, and higher Child-Pugh scores confer higher SBEM risk[26].

MANAGEMENT

Medical management

The management of HH is multidisciplinary, involving collaboration among hepatologists, respiratory physicians, interventional radiologists, and thoracic surgeons (Table 2)[1,7,20,27-39]. Medical management is offered regardless of clinical stage and involves adherence to a salt-restricted diet and diuresis (Figure 2). A moderate restriction of sodium intake is generally recommended (80-120 mmol/day which is equivalent to 4.6-6.9 g of salt)[2,40]. In practice, this can be achieved by no added salt and avoiding pre-prepared meals. The use of diuretics in HH parallels that for ascites. Mineralocorticoid antagonists like spironolactone are the cornerstone of diuretic therapy, counteracting an overactive renin-angiotensin-aldosterone system due to portal hypertension[2,40,41]. Starting doses of 100 mg/day are generally used with step-wise titration up to 400 mg/day, although this is rarely tolerated[40]. Furosemide can be added if there is inadequate control or hyperkalaemia, starting at 40 mg/day up to a maximum of 160 mg/day. Fluid balance should be carefully monitored to ensure a loss of no more than 1 kg/day in patients with peripheral oedema, and 0.5 kg/day in patients without. This is to avoid plasma volume contraction, renal failure, and hyponatraemia. Of note, long-term resolution is achieved in only 12% of patients with medical therapy alone[27]. In addition, addressing underlying factors that contribute to ongoing liver decompensation-such as treating viral hepatitis, promoting alcohol cessation, and encouraging weight loss-should be a key management priority. Treatment of portal pressures with NSBBs such as propranolol or carvedilol have been shown to delay progression to decompensation mainly be decreasing the incidence of ascites, however data is lacking as a therapeutic modality for HH specifically[42].

Figure 2 Management algorithm for hepatic hydrothorax. Following hepatic hydrothorax (HH) diagnosis, medical management is initiated first. This includes dietary sodium restriction, diuresis, suppression of aetiological factors, and thoracentesis. If there is good response to medical therapy, patients should be followed-up and monitored closely. If patients remain symptomatic and/or demonstrate recurrent pleural effusions (e.g. radiographic or via pleural fluid aspirate) despite optimal medical therapy, this defines refractory HH. At this stage, transjugular intrahepatic portosystemic shunt (TIPS) eligibility should be considered in all patients. Successful TIPS placement may even serve as a bridge to liver transplantation. Eligibility for transplantation should be assessed particularly if the patient is not a TIPS candidate. If the patient is ineligible for TIPS or transplantation, best supportive or palliative care should be initiated. Treatment options for the former include indwelling pleural catheters and surgical approaches such as video-assisted thoracoscopic surgery or pleurodesis. Experimental treatments including a low-flow ascites pump device, albumin infusions, and continuous terlipressin are novel options which require validated trials in pure HH populations to support their use. HH: Hepatic hydrothorax; IPCs: Indwelling pleural catheters; TIPS: Transjugular intrahepatic portosystemic shunt; VATS: Video-assisted thoracoscopic surgery.

Table 2 Summary of key studies reporting on treatment outcomes relevant to hepatic hydrothorax.

Ref.	Sample size (n)	Age (years)	Sex (male/female)	Hepatic function	Summary of key outcomes
Badillo and Rockey[1], 2014	77	52 ± 9	49/28	MELD: 16 ± 9. Child-Pugh: 2/37/35	83% managed with diuretics/thoracentesis, 10% TIPS, 7% underwent liver transplantation. TIPS prolonged survival from 368 days to 845 days
Liu et al[7], 2010	52	62.2 ± 12.6	23/29	Child-Pugh: 4/20/28	50% of patients undergoing surgical repair and/or pleurodesis achieved resolution of HH. Median survival of patients with successful intervention was 22.5 months
Chen et al[20], 2016	24	59.8 ± 8.5	NA	MELD: 19.4 ± NA	No patient required further pleural drainage procedures post-IPC placement. Pleural fluid infection in 16.7%
Romero et al[27], 2022	84	58.3 ± 11.5	46/38	MELD: 26.8 ± 7.1	Overall and transplant-free survival at 12 months were 68% and 41%, respectively. Current smoking and acute kidney injury increased mortality
Young et al[28], 2019	147 (32 with HH)	56.1 ± 9.7	97/50	MELD: 15.2 ± 4.4. Child-Pugh: 9.4 ± 1.2	No difference in overall response or survival rates post-TIPS in refractory HH vs refractory ascites
Orman and Lok[29], 2009	17	55 ± 9.4	7/10	MELD: 16 ± 7.1	16/17 patients had a complication due to chest tube placement and 3-month mortality was 35%
Liu et al[30], 2004	59 (24 with HH)	53.8 ± 11.2	31/28	Child-Pugh: 3/31/25	47/59 patients had ≥ 1 complication due to chest tube placement and mortality was 25.4%
Kniese et al[31], 2019	62	60.7 ± 10.8	34/28	MELD-Na: 24 ± 6.5	36% had a complication due to IPCs and 10/62 were transplanted successfully post-IPC
Shojaee et al[32], 2019	79	60 ± 10.7	43/36	MELD: 18.1 ± 5.1	10% pleural space infection risk and 2.5% mortality post-IPC. Older age is a predictor of post-IPC mortality
Huang et al[33], 2016	63	60.4 ± 15	32/31	MELD: 14.6 ± 5.1. Child-Pugh: 12/36/15	4/63 patients experienced recurrence after VATS and 1-month mortality was 9.5%. Renal dysfunction and higher MELD predicted 3-month mortality
Luh and Chen[34], 2009	12	NA	NA	NA	4/12 had HH recurrence and massive ascites despite VATS. 6/12 died of end-stage liver-disease (including 4 patients with favourable early post-operative course) by 23 months
Lee et al[35], 2011	11	63 (38-84)	6/5	MELD: 16 (9-21). Child-Pugh: 0/2/9	Successful chemical pleurodesis in 72.7% and remained HH-free for a median of 16 weeks. 10/11 had AEs and 5/11 had procedure-related mortality
Bellot et al[36], 2013	40 (0 with HH)	59 (34-80)	28/12	MELD: 12.6 ± 4. Child-Pugh: 0/30/10	ALFApump® reduced median LVP per month. Main AEs: Liver-related (75%), infections (100%), catheter dislodgement (12.5%)
Bureau et al[37], 2020	58 (0 with HH)	61.9	46/12	MELD: 11.7 ± NA	ALFApump® group did not require LVP in first 6 months. Serious AEs were more common with ALFApump® vs LVP (85.2% vs 45.2%)
Hannah et al[38], 2023	24 (0 with HH)	59.5 (53.3-63.3)	17/7	MELD: 14.5 (11.3-18). Child-Pugh: 1/12/11	Albumin infusions associated with reduced hospital admission and paracentesis requirement
Gow et al[39], 2022	23 (2 with HH)	59.6 (42.5-68.6)	23/0	MELD: 22.5 (14-41). Child-Pugh: 10 (8-13)	Significant reduction in MELD score and median LVP/thoracentesis requirement with continuous terlipressin infusion. No serious AEs reported

Reported as mean ± SD, median (interquartile range), or median (range). Child-Pugh reported as classes A/B/C or as numerical scores. AEs: Adverse events; ALFApump^®: Automated low-flow ascites pump; HH: Hepatic hydrothorax; IPC: Indwelling pleural catheter; LVP: Large-volume paracentesis; MELD: Model for end-stage liver disease; NA: Not available; TIPS: Transjugular intrahepatic portosystemic shunt; VATS: Video-assisted thoracoscopic surgery.

Therapeutic thoracentesis is routinely performed in HH, is generally well-tolerated, and can often still be performed despite thrombocytopenia or coagulopathy[43]. Thoracentesis tends to only provide transient relief as fluid often re-accumulates. Repeated thoracentesis is encouraged in this setting but is impractical as a long-term option. The pleural space is able to be completely drained safely as the risk of re-expansion pulmonary oedema following thoracentesis is low[44,45]. The main risk of thoracentesis is pneumothorax but this can reduce significantly if performed under ultrasound guidance[46]. When thoracentesis has to be repeated frequently, particularly in patients on maximally tolerated diuresis and sodium restriction, alternative treatments must be sought.

TIPS and other interventions

Despite optimising medical management, some patients have recurrent pleural effusions. This is defined as refractory HH. In these cases, TIPS is often considered standard of care[2]. TIPS relieves portal hypertension and reduces ascites formation by creating a communication between the portal and hepatic veins[5]. In a systematic review and meta-analysis of 198 patients with medically-refractory HH, 73.4% of patients demonstrate symptomatic improvement and/or a decreased thoracentesis requirement post-TIPS placement[47]. Of these, 55.8% achieved complete resolution over a mean duration of 10 months. TIPS-related hepatic encephalopathy occurred at a rate of 11.7% which is comparable to what is observed following TIPS insertion for other common indications such as acute variceal haemorrhage and refractory ascites[48]. Another study reported the post-TIPS response rate (complete or partial) as 54.8%, 66.7%, and 64.3% at one month, three months, and six months, respectively[28]. Clinical response in refractory HH did not differ significantly compared to a propensity matched cohort with refractory ascites, nor did overall survival. In a retrospective study of 73 post-TIPS patients for refractory HH, survival at one year was 42% and fell to 15% at five years[49]. Pre-TIPS MELD scores < 15 and clinical response were all independent predictors of long-term survival, with higher pre-TIPS creatinine levels also associated with an increased 30-day mortality.

Of note, there is no established role for chest tube insertion in HH as it is associated with significant mortality and morbidity, due to massive fluid shifts, electrolyte disturbance and risk of infection. One series of 17 patients with HH reported a three-month mortality rate of 35% for patients in whom a chest tube was placed[29]. In another study, chest tube insertion was similarly associated with a mortality of 27%, and a complication rate of 80%[30].

Indwelling pleural catheters (IPCs) appear to be more appropriate for simple effusions and may offer a less invasive alternative to chest tubes. A large body of evidence already supports their use in malignant pleural effusions, particularly in patients without a non-expandable lung[19,45]. There is also increasing evidence for IPCs in HH management. They can be an effective palliative option in non-transplantable patients with refractory HH. IPCs have also been proposed as a potential bridge to transplantation in patients where TIPS is contraindicated[2,20]. A good baseline performance status is generally recommended for consideration of an IPC[19]. Retrospective and prospective studies note spontaneous pleurodesis is achieved in 15%-33% of patients post-IPC placement[20,31,32]. These rates are likely overestimated due to some patients in these studies being transplanted during the period between IPC insertion and spontaneous pleurodesis. Beyond this observation, Kniese et al[31] report a complication rate of 34.5% associated with IPC insertion. Infection is the most frequent complication with empyema developing at a rate of 16.1%. Inclusion of quality-of-life measures such as dyspnoea scores should also be a priority of future research[44]. Ultimately, HH as an indication for IPCs remains underexplored and warrants investigation in large, preferably multicentre, prospective trials.

Surgical interventions

Non-transplant surgical options also exist although these are only considered for refractory HH where TIPS is contraindicated or has failed[14]. Despite reports of excellent resolution, complication rates are significant. A recognition of macroscopic diaphragmatic defects has led to efforts to perform surgical repair via video-assisted thoracoscopic surgery (VATS) (Table 3)[33,34,50-52]. A 10-year surgical case series of the outcomes of 63 patients who underwent a thoracoscopic mesh repair for HH using either a combination of mesh covering and suturing or mesh covering alone[33]. Huang et al[33] reported resolution in 93.7% across a median follow-up period of 20.5 months. Despite a remarkable resolution rate, diaphragmatic defect repair was associated with a complication rate of 31.7% and three-month mortality of 26.4%. This was mostly due to septic shock (9.5%), acute kidney injury (6.3%) and gastrointestinal haemorrhage (6.3%). Appropriate selection of patients for thoracoscopic surgery may improve survival as baseline poor renal function and higher MELD scores predicted higher three-month mortality rates. In a case series of twelve patients undergoing VATS for refractory HH, 8/12 (67%) had an uncomplicated early post-operative course and no re-accumulation within the first three months on medical therapy alone[34]. Three patients who initially responded favourably required further interventions either with a re-do VATS or tube thoracostomy. Ultimately, mortality was 50% by 23 months post-VATS. Other case reports[50,51] and a case series of two patients[52] who received thoracoscopic diaphragmatic defect repair reported complete resolution and no further accumulations during follow-up.

Table 3 Summary of outcomes from published studies on video-assisted thoracoscopic surgery.

Ref.	Study population (n, intervention)	Resolution rate	Mortality post-VATS
Huang et al[33], 2016	63, VATS only	At median 20.5 months: 93.7%	At 3 months: 25.4%
Luh and Chen[34], 2009	12, VATS ± pleurodesis	At 36 months: 33.3%	At 23 months: 50%
Ikeda et al[50], 2021	1, VATS only	100%1	NA
Nakamura et al[51], 2012	1, VATS only	100%1	NA
Ibi et al[52], 2008	2, VATS ± pleurodesis	100%2	Case 1 NA; case 2 0% at 1 year

¹Follow-up period not reported by studies.

²Follow-up period not reported for case 1 whereas the follow-up period for case 2 was one year.

NA: Not available; VATS: Video-assisted thoracoscopic surgery.

Chemical pleurodesis is rarely performed in the setting of refractory HH and is usually only considered in patients who are ineligible for more definitive treatment options. This technique uses an irritant such as talc slurry or tetracycline to obliterate the pleural space[53]. Evidence supporting pleurodesis in HH limited. Reported success rates of chemical pleurodesis vary and this is likely explained by how rapidly ascitic fluid re-accumulates in the pleural cavity[54]. This prevents the parietal and visceral surfaces of the pleura from adhering long enough to achieve pleurodesis. In at least one small series of 11 cases, chemical pleurodesis with or without VATS demonstrated a resolution rate of 72.7% (defined as absence of symptoms and no pleural effusion at one month)[35]. Long term follow-up confirmed that overall survival was better in responders with 62.5% still in remission after a median of 16 weeks. Literature on the utility of pleurodesis in HH is relatively low-level and in the form of case reports and case series. One systematic review and meta-analysis[55] found the pooled response rate was 72% (95%CI: 65%-79%). The chemical agent utilised did show varied responses with talc alone demonstrating a complete response rate of 71% across seven studies (115 patients) but up to 93% when using compound OK-432 (Picibanil) alone or in combination with minocycline. However, the latter was only reported in two studies comprising only 19 patients. Complications were relatively common with a pooled incidence of 82% (95%CI: 66%-94%). Common complications included: (1) Fever (47%); (2) Renal failure (17%); (3) Pneumothorax (16%); and (4) Hepatic encephalopathy (11%). Pneumonia and empyema were less common occurring in 9% and 6%, respectively.

Liver transplantation

Ultimately, liver transplantation is the only curative option for cirrhosis, and by extension, HH[2]. It remains the cornerstone of long-term treatment of end-stage liver disease. There are however several post-transplantation challenges such as immunosuppression, steroid-induced metabolic dysfunction and malignancy risk. Additionally, in patients with alcohol-related liver disease, prolonged abstinence (traditionally six months) is required outside of very specific criteria for alcoholic hepatitis[56]. There is however regional variation in acknowledgement of reasonable outcomes in appropriately selected patients. Practice should ultimately be dictated by local policies and guidelines. Despite this, all patients with cirrhosis should be evaluated to determine suitability and success of a transplant. In general, indications for liver transplantation include any of: (1) Diuretic refractory ascites; (2) MELD > 15; (3) A history of recurrent hepatic encephalopathy; (4) Gastrointestinal haemorrhage or SBP; and (5) Others[57]. In a recent retrospective series of 84 patients with newly diagnosed HH, 33% of patients underwent liver transplantation within 12 months of their index admission[27]. HH was also associated with a 12-month transplant-free survival of only 41%, with no deaths in the transplanted group recorded during follow-up. These findings reinforce the need to early and prompt referral to a transplant centre due to improved survival.

A key limitation of current therapeutic approaches for HH is their reliance on expert opinion rather than evidence from large randomised controlled trials[10]. This is likely due to the rarity of HH and the consequent lack of adequately powered studies to establish an optimal treatment strategy. As a result, management is often guided by individual clinician experience rather than robust clinical evidence. Beyond liver transplantation, evidence supporting definitive treatments remains limited, posing significant challenges for the long-term management of patients who are not transplant candidates. Treatment options are further restricted in those for whom TIPS is contraindicated.

EXPERIMENTAL THERAPIES AND FUTURE DIRECTIONS

In HH, TIPS placement and liver transplantation are the only validated definitive treatment options. A novel automated low-flow ascites pump (ALFApump^®) has been trialled in refractory ascites with the aim of eliminating the need for recurrent large-volume paracentesis (LVP) and improving functional status[36,37]. It is a battery-operated and subcutaneously implanted pump. The device is connected to a catheter which directly moves ascitic fluid from the peritoneal cavity to the urinary bladder for elimination in the urine. In a multicentre, nonrandomised landmark trial of forty patients with refractory ascites, 40% of patients no longer needed LVP, and 70% required less than one paracentesis per month post-device implantation[36]. This was followed by a randomised control trial (NCT01528410) comparing the ALFApump^® and LVP standard of care in 60 patients with refractory ascites[37]. This trial demonstrated that implanted patients had a substantially reduced requirement for LVP with only 10/27 patients requiring LVP at six months. Complications of the ALFApump^® can include peritoneal and/or bladder catheter migration, blockage, and kinking. Despite some evidence in refractory ascites, there are currently no published trials evaluating ALFApump^® in HH. While there is a potential for treatment benefit to be extrapolated to HH, the ALFApump^® lacks validated trial data for this indication. There is however a growing body of experience which has been published in case series indicating a possible avenue for further research.

Regular outpatient human albumin infusions show promise in the management of decompensated cirrhosis complicated by diuretic-refractory ascites, however there is minimal data on its use in HH[58]. Emerging evidence supports a therapeutic and perhaps even disease-modifying role for long-term outpatient human albumin infusions in decompensated cirrhosis complicated by ascites. The landmark multicentre randomised trial (ANSWER) demonstrated a significant mortality benefit for regular albumin infusions and standard medical therapy compared to medical therapy alone (77% vs 66% 18-month survival, respectively)[58]. A recent study by Hannah et al[38] also found that long-term outpatient albumin infusions were associated with reduced portal hypertensive-related admissions as well as improved biochemical and clinical markers among patients with cirrhosis and diuretic-refractory ascites. However, variation in albumin 20% preparations from different international suppliers is well-documented and is likely to influence findings reported between studies[59]. Sodium content in particular has been implicated, with comparatively lower sodium concentrations found in preparations used in Australia compared to those in Europe and the United Kingdom. Higher sodium content albumin, especially if given as large infusions, is associated with poorer outcomes in cirrhosis, likely by exacerbating fluid retention and worsening pre-existing renal dysfunction. Ultimately, there is only a theoretical benefit of albumin infusions from improved ascites control as there are no published trials reporting specifically on HH.

Terilipressin is a synthetic vasopressin analogue and a potent splanchnic vasoconstrictor leading to reduced portal inflow[60]. However, because of its short half-life (50-80 minutes), bolus administration offers limited clinical benefit. A recent open label phase IIa trial found that continuous terlipressin infusion increased the time interval between LVP by more than 50% in 4/6 patients with refractory ascites, and reduced the volume of paracentesis by over 30% in all patients[61]. Three (50%) patients experienced serious adverse events (bacteraemia, recurrence of hepatic encephalopathy, umbilical hernia leak) during the treatment period but none were deemed by the investigators to be treatment-related. In a small series, five patients with refractory ascites established on a 3.4 mg/24-hour continuous terlipressin infusion demonstrated a significant reduction in LVP requirement over 28 days compared to baseline[62]. No serious adverse events were reported. A prospective observational study observed a 57% reduction in median LVP frequency in their subgroup of six patients with refractory ascites or HH over three months[39]. Efforts to manufacture more stable terlipressin preparations or alternative long-acting vasopressin analogues are needed[62]. The role for continuous terlipressin infusions in refractory ascites and HH is currently not established but merits further exploration in confirmatory trials.

There is a clear need to develop and validate appropriate treatments for bridging the gap to liver transplantation for candidates. Although it is expected that a reduction in ascites will minimise HH burden with these novel approaches, they still require validation in trials to support use in long-term management.

CONCLUSION

HH is an uncommon complication of decompensated cirrhosis that is associated with high mortality. It tends to present as a right-sided transudative pleural effusion and its management is predominantly symptomatic, except for liver transplantation which is curative. For patients who are not candidates for liver transplantation there are currently no truly disease-modifying treatments apart from TIPS, which greatly limits survival. Consequently, re-accumulation and readmission to hospital is common. The role for an automated low-flow ascites pump device, regular outpatient albumin infusions, and continuous terlipressin in the management of HH, however promising, requires validation. High quality evidence from well-designed, multicentre clinical trials must be seen as a priority in future research. This will support clinicians in managing a complex condition resulting in high morbidity and mortality. Better characterisation of patient-reported outcomes such as treatment tolerability and patient perception may similarly optimise treatment selection strategies.