HH is a serious and difficult to manage complication of cirrhosis and portal hypertension that can progress to end-stage liver disease. However, the specific prognostic survival impact of HH patients, and the interrelationship between its long-term impact on mortality and on cirrhosis-related decompensated events are currently unknown. Here, we aim to examine the clinical characteristics, natural history and factors associated with long-term prognostic survival. More complete clinical data from hospital records of patients with decompensated cirrhosis and ascites admitted to the First Hospital of Jilin University from January 2013 to June 2021 were extracted for retrospective analysis, and risk factors affecting prognostic survival were further analysed based on follow-up survival time. The diagnosis of hepatic pleural fluid was based on currently accepted clinical features of the disease, including a known diagnosis of cirrhosis, the presence of portal hypertension, the analysis of pleural effusion, and the absence of primary cardiopulmonary disease. It is intended to further guide clinicians to clarify the clinical features of HH and the natural history of long-term survival and to seek more appropriate diagnosis and treatment.
In this study, we selected a large number of patients with decompensated cirrhosis in combination with one or both pleural effusions, for a total of 131 patients with HH according to the inclusion criteria. The current study compares with the previous literature on HH and our data represent one of the largest study cohorts[6,13−15] with a comprehensive analysis of demographic, clinical presentation, laboratory test and examination findings data in this unique patient population.
In this study, the ratio of men to women with HH was 1.67:1, the youngest age was 25 years, the oldest was 81 years, the mean age was (52.76 ± 11.88) years and the median age was 58 years. Hepatitis B virus was the most common cause of HH, which is more consistent with most studies[6,13−15], and HH was more common in men than in women. The most common symptoms of HH are dyspnoea and abdominal distention, which may be associated with massive pleural effusion compressing the lung tissue and massive accumulation of abdominal fluid resulting in. Splenomegaly and jaundice are the most common signs in patients with HH and are also common signs of decompensated cirrhosis and are not clearly specific. In the thoracoabdominal imaging of HH patients a large amount of ascites combined with a large amount of pleural fluid can be observed commonly, and pleural effusion is commonly found on the right side in about 87 cases (66.4%), with only a few located on the left side and bilaterally. The pathophysiology of HH is not yet fully understood, and the most prominent view is that pleural effusion is aided by the negative intrathoracic pressure generated during inspiration, with ascites entering the thoracic cavity directly through diaphragmatic defects of various sizes and is produced[16, 17], as the liver is anatomically similar to the diaphragm and the right diaphragm is less muscular than the left. Most patients with HH have compressive atelectasis on CT of the lungs (73.3%), which is closely related to compression of lung tissue by a large pleural effusion.
HH as a complication in patients with decompensated cirrhosis can often be combined with other complications, with peritoneal effusion being the most common, followed by electrolyte disturbances, but with hepatic encephalopathy and liver failure being the most severe and with a poor prognosis, in general agreement with the data reported in the literature. Based on our data, the prognosis of patients with HH may be comparable to that predicted based on the MELD model, with a mean MELD score of 10.20 for the entire cohort at first diagnosis and a MELD score > 15 in 19% of the population, indicating a relatively short survival time and a significantly higher risk of death. Our MELD score data were somewhat lower than in a previous study in the literature, where the mean MELD score was 16, possibly due to the larger cohort of our data and the early diagnosis of patients. When using the MELD-Na score for prediction, the mean score was 10.53 and the number of people with > 16 was 22%, which is generally consistent with the MELD score. We also performed Child-Pugh grading and ALBI grading, where Child-Pugh grade C and ALBI grade 3 were 65% and 70.5%, respectively, both higher than 54% and 60% in a recent retrospective study in 2021, also indicating that most patients had poor liver function. Although our follow-up data suggest that most patients with HH eventually die from severe liver failure, some patients instead die from complications related to their lung disease, as we were unable to determine the specific cause of death for all patients who died. Therefore, based on our data, caution is needed in identifying the most likely cause of death in patients with HH.
We performed a statistical analysis of treatment modalities at the time of first diagnosis in 131 patients with HH and found that HH management was similar to portal hypertensive ascites, with restriction of sodium intake and the use of diuretics, drainage of puncture placement, and TIPS and liver transplantation as options for later treatment[3,18−21]. Pleural catheter drainage has been previously reported and no significant efficacy has been observed, but it is notable that all our patients did not receive this treatment. Successful treatment of patients with refractory hepatic pleural fluid who did not meet the criteria for TIPS by repairing the diaphragmatic defect with televised thoracoscopic surgery (VATS) has also been reported in the literature, but none of our patients attempted this treatment. All our patients received diuretics at first diagnosis and their combination ratio of furosemide and spironolactone was individualized, commonly furosemide 60 mg combined with spironolactone 100 mg in a single dose, with the addition of tolvaptan, a selective vasopressin V2 receptor antagonist with an affinity for vasopressin V2 receptors that is 1.8 times higher than that of natural arginine vasopressin (AVP), when diuresis was not effective. 1.8 times the affinity of natural arginine vasopressin (AVP). Of these, 20 (15%) were treated with drugs alone and 111 (75%) were treated with drugs combined with thoracic intubation, which also suggests that most pleural effusions were massive and required puncture and drainage to relieve the compression of lung tissue and thereby relieve the symptoms of dyspnoea. There was no statistically significant difference in MELD scores or survival time between patients first diagnosed with drugs alone and those with drugs combined with thoracic intubation. Only one patient in our study was found to have undergone TIPS at a later stage of treatment, three then underwent liver transplantation and the three patients who underwent liver transplantation are still alive today.
By following the survival time of patients, we found that the overall prognosis of HH patients was poor, with approximately more than half dying within one year of the onset of symptoms and only 34 cases (33.7%) surviving after a year. However, patients survived significantly longer after receiving liver transplantation, considering that liver transplantation may be a better treatment option than drugs and chest tube drainage, but caution is needed in interpreting these data as only a very small number of patients, 2.2% of the entire study population, received liver transplantation, and it is also possible that patients with a better prognosis were more likely to opt for liver transplantation treatment. In addition, the selection of liver transplant recipients is extremely complex and although patients with advanced liver failure are selected based on screening criteria, transplant candidates are usually rarely comorbid with other diseases. Therefore, retrospective conclusions based on a very small amount of data may be inaccurate.
The long-term natural history of this complication of HH in patients with decompensated cirrhosis described in this study excludes the radical treatment of portal hypertension with liver transplantation, and there are not many studies in this direction. There are few studies on long-term mortality in patients with HH. In this study 101 patients with known clinical endpoint events had mortality rates of 70.3%, 94.1%, 98.1% and 100% at 2, 3, 4 and 5 years respectively, with close to more than two thirds of patients dying within 2 years and almost all facing death within 3 years, with one patient surviving for up to 5 years. Three patients underwent liver transplantation late in their treatment and are currently alive, with survival times of 33, 35, and 38 months after their first diagnosis of HH, respectively, with more follow-up still needed to clarify the ultimate survival time. Therefore, based on the data available to us, it is not possible to determine exactly to what extent liver transplantation has an impact on the survival time of patients with HH and there is a lack of reference in the previous literature.
Gender, liver encephalopathy, hyponatremia, and MELD score had a significant effect on mortality when assessing the prognosis of patients with HH. After multifactorial analysis, the findings indicated that liver encephalopathy, hyponatraemia, and MELD score were independent risk factors for prognostic survival in HH, and the MELD score appeared to be a good predictor of mortality in patients with HH. On the contrary, the Child-Pugh score and the ALBI score were not as good as the MELD score in predicting survival outcomes in this subgroup, possibly because the MELD takes more factors into account. Additionally, the presence of a MELD score above the threshold of 15, liver encephalopathy or hyponatraemia in patients with HH was more likely to result in a poorer prognosis, and the survival gap increased further with increasing MELD scores. In this context, early identification of patients with HH in patients with decompensated cirrhosis and timely diagnosis and consequently timely management of patients with a combination of other decompensated events, HH can be a marker presentation of poor prognosis. In patients with HH, early identification of patients with a poor prognosis based on factors affecting the prognosis, together with targeted treatment, is particularly important.
In conclusion, to our knowledge, we have the largest volume of data on patients with HH of any published study to date, and typically the majority of patients usually have predominantly liver and lung symptoms. We believe that an early and accurate diagnosis is the key to guiding personalized treatment. Patients with HH generally have a poor prognosis, and those who receive liver transplants generally have a relatively good prognosis. Future studies may require multiple centre, prospective studies that critically evaluate the different diagnostic criteria and treatment modalities typically used in these patients.
Limitations in this study: (1) This study is a single-centre retrospective analysis and does not exclude the influence of subjective factors by patients and medical record keepers; Additionally, not all patients were followed up in our department and therefore detailed information on the cause of death or other relevant clinical events is lacking. (2) Data collection throughout follow-up taking into account changes in compliance with sodium restriction and diuretic dose was relatively difficult, therefore throughout the natural history of HH we did not specify the impact of various treatment modalities on their prognosis, which can certainly be an important prognostic modifier, nor the impact of clinical presentations related to HH on their prognosis, but this may help better understand the impact of decompensated events on the prognosis of patients with HH. (3) The small number of patients undergoing TIPS and liver transplantation did not allow inclusion in the Cox proportional risk model survival analysis.