Terlipressin and albumin combination treatment in patients with hepatorenal syndrome type 2

Abstract

Background

Hepatorenal syndrome (HRS) is associated with a poor prognosis. In HRS type 1, loss of renal function is rapidly progressive, while HRS type 2 is characterised by chronic ascites and more moderately elevated renal parameters. While treatment with terlipressin/albumin is well established in type 1, its effectiveness in chronic HRS is less clear.

Objective

The aim of this study was to evaluate the effectiveness of terlipressin/albumin treatment in patients with HRS type 2.

Methods

All patients with a first episode of HRS between April 2013 and February 2016 were included in this observational study. Relevant clinical and laboratory parameters were recorded and patients were followed.

Results

A total of 106 patients with HRS were included. With terlipressin therapy reversal of HRS types 1 and 2 was achieved in 48% and 46% of patients (p = 0.84) with relapse rates of 8% vs 50% (p = 0.001). Overall survival (OS) and survival free of liver transplantation (LTx) were similar in HRS types 1 and 2 (p = 0.69; p = 0.64). In multivariate analysis response to treatment was independently associated with better OS in HRS type 2, in addition to established risk factors such as lower Model for End-Stage Liver Disease score, absence of hepatic encephalopathy and eligibility for LTx.

Conclusion

A terlipressin treatment course seems to be justified in selected patients with HRS type 2, especially in countries and settings with long transplant waiting lists. In addition treatment response might also help to identify HRS type 2 patients with a more favourable outcome.

Key summary

Summarise the established knowledge on this subject

• Hepatorenal syndrome (HRS) is a severe complication of advanced liver cirrhosis associated with a high short-term mortality.

• While HRS type 1 typically develops after a trigger event such as gastrointestinal bleeding or infection, HRS type 2 is characterised by recurrent or refractory ascites and a slower progression of renal insufficiency.

• Several studies have shown an association between response to terlipressin treatment and survival in type 1 HRS; however, less is known about the effects of vasoconstrictor treatment in HRS type 2.

What are the significant and/or new findings of this study?

• Terlipressin and albumin treatment was associated with reversal of renal failure in approximately half of all patients with HRS type 2; however, there is a high relapse rate of 50%.

• In addition to established risk factors such as hepatic encephalopathy and Model for End-Stage Liver Disease score, response to terlipressin treatment seems to be associated with improved survival in HRS type 2.

• A course of vasoconstrictor treatment can be justified in selected HRS type 2 patients with an acute-on-chronic deterioration of renal function.

Introduction

Acute kidney injury (AKI) is a frequent complication in patients with advanced liver cirrhosis. AKI can be of prerenal, intrarenal or postrenal origin, in addition patients with liver cirrhosis can suffer from chronic kidney disease (CKD). Hepatorenal syndrome (HRS) has long been defined as a functional renal failure which develops as a consequence of splanchnic arterial vasodilatation and renin-angiotensin-aldosterone system–activation with marked renal vasoconstriction in decompensated cirrhosis.^1,2 Guidelines distinguish two types of HRS. Type 1 HRS is characterised by a rapidly progressive loss of renal function with a serum-creatinine >2.5 mg/dl, often induced by a precipitating event such as gastrointestinal (GI) bleeding or spontaneous bacterial peritonitis (SBP). HRS type 2 is characterised by recurrent or refractory ascites and a slower progression of renal insufficiency.^3,4 Recent guidelines emphasise the role of pro-inflammatory cytokines and chemokines and microvascular dysfunction in HRS development and have classified HRS type 1 as HRS-AKI, HRS type 2 as renal impairment which fulfils the criteria of HRS but not of AKI (non–AKI-HRS; NAKI) and HRS-CKD.^5,6 Median survival of HRS type 2 patients is four to six months.⁷

Vasoconstrictor therapy with terlipressin and albumin can induce remission of HRS type 1 in 40% to 60% of patients, and renal recovery can be maintained in more than 70% of patients after treatment withdrawal.^8–11 Several studies have shown an association between response to treatment and survival in type 1 HRS. Few studies have examined the efficacy of terlipressin treatment in HRS type 2, and results have been ambiguous.^9,10,12–15 While some reported a benefit of terlipressin treatment, others concluded that treatment does not affect outcome. Studies were limited by small patient cohorts, ranging from 3 to 31 type 2 patients treated with terlipressin. Meta-analyses concluded that the number of HRS type 2 patients in published studies addressing the role of terlipressin was too small for conclusive treatment recommendations.^16–18 In consequence current guidelines of the European Association for the Study of Liver Disease (EASL) as well as the American Association for the Study of Liver Disease do not recommend vasoconstrictor treatment in HRS Type 2/HRS-NAKI.^5,19

The aim of this study was to examine the effectiveness of terlipressin treatment in HRS type 2 and identify factors that are associated with response and survival.

Patients and methods

Patients

All patients presenting with a first episode of HRS between April 2013 and February 2016 were included in this observational study. Patient follow-up continued until September 2016. Diagnosis of HRS was established according to the EASL clinical practice guideline 2010, which was based on recommendations of the International Club of Ascites (ICA) updated in 2007 and 2015.^4,6,20 Diagnosis of HRS required presence of liver cirrhosis with ascites or alcoholic steatohepatitis, serum creatinine >1.5 mg/dl (133 µmol/l) and no evidence of parenchymal renal disease on urine analysis and renal ultrasound. Treatment with nephrotoxic drugs had to be discontinued for at least 48 hours, and patients received volume expansion with albumin before diagnosis of HRS was established. HRS type 1 was characterised by a doubling of serum creatinine levels from patient’s baseline within two weeks. In accordance with previously published cohorts and with the 2015 update of ICA recommendations, treatment could be initiated before serum creatinine was >2.5 mg/dl in type 1 once the diagnosis was established. Patients with HRS type 2 had to suffer from recurrent/refractory ascites with chronically elevated serum creatinine levels with gradual or intermittent increase. Classification of HRS type was independently performed by two hepatologists after review of each patient’s medical records and laboratory data.

The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki was approved by the ethics committee of the State Physician Chamber of Rhineland-Palatinate, Germany (9 July 2012). The study was registered at the German Clinical Trials Register (DRKS; DRKS00013117). All patients provided written, informed consent to participate in this study.

Treatment and outcome

After diagnosis of HRS was established, terlipressin treatment was initiated at a dose of 4 mg/day given via continuous intravenous infusion. All patients received concomitant albumin at a dose of 40 g/day. Decisions on treatment duration and dose escalation were made by a hepatology consultant.

Complete response (CR) was defined as a decrease of serum creatinine to <1.5 mg/dl, partial response (PR) was defined as a decrease of serum creatinine of at least 50% but to a level of >1.5 mg/dl, both without need for renal replacement therapy (RRT). All other patients were classified as non-responders (NRs).

Data and statistics

Patient, clinical and laboratory characteristics were retrieved and electronically recorded from hospital medical records (Table 1). During inpatient treatment patients were continuously followed with monitoring of terlipressin dose, renal parameters, need for RRT or other complications. During inpatient treatment as well as after patient discharge, patient outcomes such as need for RRT, LTx or death were followed-up on a regular basis by a review of hospital records and telephone calls to patients and/or health care providers.

Table 1.

Baseline characteristics of patients with HRS type 1 and type 2 (n = 106).

	HRS type 1 (n = 54)	HRS type 2 (n = 52)	p
Patient characteristics
Age – years	55.7 ± 10.1	60.4 ± 9.4	0.02*
Male sex – no. (%)	34 (63)	37 (71)	0.37
Child-Pugh – no. (%)
• A	0 (0)	0 (0)	0.10
• B	6 (11)	12 (23)
• C	48 (89)	40 (77)
Aetiology – no. (%)
• Alcoholic	39 (72)	42 (81)	0.30
• Non-alcoholic	15 (28)	10 (19)
Alcohol abuse – no. (%)
• Ongoing	7 (13)	15 (29)	0.04*
• None or history of	47 (87)	37 (71)
HCC – no. (%)	12 (22)	9 (17)	0.53
LTx-eligible – no. (%)	23 (43)	16 (31)	0.21
ICU admission – no. (%)	22 (41)	16 (31)	0.29
Laboratory characteristics
Serum creatinine – mg/dl	3.0 (IQR 2.5–3.5)	2.3 (IQR 1.9–3.2)	0.08
Serum bilirubin – mg/dl	6.6 (IQR 2.8–19.1)	3.0 (IQR 1.7–8.1)	0.01
INR	1.6 (IQR 1.3–1.9)	1.5 (IQR 1.3–1.8)	0.56
Serum albumin – mg/dl	22 (IQR 18–28)	23 (IQR 18–26)	0.86
MELD	29 (IQR 24–35)	24 (IQR 20–31)	0.007

HCC: hepatocellular carcinoma; HE: hepatic encephalopathy; HRS: hepatorenal syndrome; ICU: intensive care unit; INR: international normalised ratio; IQR: interquartile range; LTx: liver transplantation; MELD: Model for End-Stage Liver Disease. *p-value < 0.05.

The Student t-test was used for normally distributed data, Mann-Whitney test was used for non-parametric data, chi-squared test was used for analysis of categorical variables. For related samples with non-parametric distribution, the Friedman two-factorial variance analysis according to rank was performed. Survival analysis was performed using the Kaplan–Meier method with log-rank test for between-group comparisons. Cox regression models were used to assess the association of clinical characteristics with overall survival (OS), variables with a p ≤ 0.10 in univariate analysis were included in the regression model using the enter method. p values are two tailed, and p values < 0.05 were considered statistically significant.

Statistical analysis was performed using IBM SPSS version 23 and GraphPad Prism version 5.

Results

Patient characteristics

Overall 106 patients with liver cirrhosis presenting with a first episode of HRS were included between April 2013 and February 2016. Median follow-up was 80 (interquartile range (IQR) 21–450) days. Baseline characteristics at admission are shown in Table 1.

In 54 patients, HRS was classified as type 1, while 52 patients had HRS type 2 (49%). Alcoholic liver disease was the most common underlying aetiology of liver cirrhosis. Most patients in both groups had advanced liver cirrhosis, as indicated by Child-Pugh score, high serum bilirubin and low serum albumin levels. HRS type 1 patients had higher serum bilirubin concentration at baseline than patients with HRS type 2 (p = 0.01) and a trend toward higher serum creatinine levels (p = 0.08), resulting in higher Model for End-Stage Liver Disease (MELD) scores in this subgroup (p = 0.007). Patients with HRS type 2 were older and more often presented with ongoing alcohol abuse. In HRS type 2 serum creatinine and bilirubin levels were lower than in HRS type 1.

Terlipressin treatment and response

Median treatment duration with terlipressin was 4.5 (IQR 3.0–7.0) vs 5.0 (IQR 3.0–7.0) days in HRS type 1 and type 2 (p = 0.51). Cumulative terlipressin dosage was comparable between both groups (HRS type 1: median 20.0 mg; IQR 12.0–30.5 mg, HRS type 2: median 24.0 mg; IQR 14.3–29.5 mg; p = 0.97). Overall response to treatment (CR/PR) was observed in 50 patients (47%) regardless of HRS type. Terlipressin treatment was effective in reversing HRS type 2 in 24/52 patients (46%) and in 26/54 HRS type 1 patients (48%), thus effectiveness was comparable in both subgroups (Figure 1(a); p = 0.84). In HRS type 2 serum creatinine levels were significantly lower in patients with CR/PR three days after treatment initiation; at discharge mean serum creatinine was 1.16 ± 0.26 mg/dl compared with 2.62 ± 1.29 mg/dl in patients with NR (Figure 1(b); p < 0.001)). However, a relapse rate of 50% after initial response and subsequent terlipressin withdrawal was observed in type 2 HRS (Figure 1(c); HRS type 2: 12/24 patients; 50% vs HRS type 1: 2/26; 8%; p = 0.001). In these 14 patients with HRS relapse, median time to relapse was 19 days (median time to relapse: HRS type 1: 5 days; HRS type 2: 25 days). All 14 patients with relapse regardless of HRS type received a second treatment course, and remission was again achieved in six patients (43%).

Figure 1.

(a) Response to terlipressin and albumin treatment in hepatorenal syndrome (HRS) type 1 (n = 54) and type 2 patients (n = 52). (b) Changes in serum creatinine during inpatient treatment in HRS type 2 patients (n = 52). Error bars indicate 95% confidence interval. (c) HRS relapse after initial response in HRS type 1 (n = 26) and type 2 (n = 24).

Twenty-eight patients with HRS type 2 did not respond to terlipressin treatment, and 10/28 patients died during inpatient treatment. Out of 18 patients who survived and were discharged alive, seven patients underwent haemodialysis after terlipressin treatment failure. Three other patients with NR underwent LTx during inpatient treatment after failure of terlipressin. Therefore only eight of 28 patients survived inpatient treatment without being placed on RRT or having received an LTx.

With a response rate of 46% and a relapse rate of 50% in HRS type 2, upfront identification of patients with a higher probability of response is desirable. Therefore, we performed univariate analyses of clinical and laboratory variables at baseline to identify parameters associated with treatment response in HRS type 2. We observed a trend toward higher baseline serum creatinine in non-responders vs patients with CR or PR, but neither age, sex, hepatocellular carcinoma (HCC), ongoing alcohol abuse, overt HE, urinary output, systolic or diastolic blood pressure or any other baseline laboratory variable was associated with treatment response (Table 2).

Table 2.

Association of clinical and laboratory variables with treatment response in HRS type 2.

	Terlipressin response group (n = 24)	Terlipressin non-response group (n = 28)	p
Age – years	58.2 ± 9.2	62.3 ± 9.4	0.13
Male sex – no. (%)	17 (71)	20 (71)	0.96
Child-Pugh – score	11.1 ± 1.6	10.9 ± 1.7	0.69
Ongoing alcohol abuse – no. (%)	6 (25)	9 (32)	0.57
HCC – no. (%)	3 (13)	6 (21)	0.40
Overt HE – no. (%)	5 (21)	10 (36)	0.24
Systolic blood pressure – mmHg	105 ± 13.2	109 ± 17.5	0.34
Diastolic blood pressure – mmHg	61 ± 8.7	64 ± 13.8	0.18
Serum creatinine – mg/dl	2.3 ± 0.6	3.0 ± 1.4	0.08
Serum bilirubin – mg/dl	5.7 ± 7.8	8.8 ± 11.5	0.61
INR	1.6 ± 0.5	1.7 ± 0.6	0.75
MELD	23.7 ± 6.8	26.7 ± 7.6	0.17
Urinary sodium <20 mmol/l – no. (%)	16/20 (80)	16/23 (70)	0.43
Urinary output – ml	864 ± 470	750 ± 526	0.47

HCC: hepatocellular carcinoma; HE: hepatic encephalopathy; HRS: hepatorenal syndrome; INR: international normalised ratio; MELD: Model for End-Stage Liver Disease.

Patient outcome

With terlipressin treatment, OS and LTx-free survival did not differ significantly between HRS types 1 and 2. Median OS in HRS type 2 was 239 ± 174 compared with 89 ± 53 days in type 1 (p = 0.69; Figure 2(a)); median LTx-free survival in type 2 HRS was 72 ± 18 compared with 58 ± 16 days in type 1 (p = 0.64; Figure 2(b)). OS was poor, with a 12-month survival in type 1 HRS of 38% and 47% in HRS type 2.

Figure 2.

(a) Kaplan–Meier analysis of overall survival in hepatorenal syndrome (HRS) type 1 (n = 54) and type 2 (n = 52). (b) Kaplan–Meier analysis of liver transplantation-/renal replacement therapy-free survival in HRS type 1 (n = 54) and type 2 (n = 52).

Response to terlipressin was associated with survival in HRS type 2 (Figure 3). OS was significantly better in patients with CR or PR vs NR (median OS: 792 ± 369 vs 61 ± 63 days; p = 0.02), and LTx-free survival was longer, 147 ± 20 days compared with 25 ± 11 days (p < 0.0001). Overall 25 out of 39 LTx-eligible patients underwent LTx during follow-up. Patients were deemed LTx–non-eligible because of advanced age, cardiovascular or pulmonal comorbidities, ongoing alcohol abuse or HCC exceeding the Milano criteria. Thirty-two of 56 LTx–non-eligible patients with non-response to terlipressin therapy underwent some form of RRT; however, median OS in this subgroup did not differ from LTx–non-eligible patients who did not receive RRT (61 vs 89 days; p = 0.66).

Figure 3.

(a) Kaplan–Meier analysis of overall survival (OS) according to terlipressin treatment response (complete response (CR)/partial response (PR); n = 24) vs non-response (NR; n = 28) in hepatorenal syndrome (HRS) type 2. (b) Kaplan–Meier analysis of liver transplantation-/renal replacement therapy-free survival according to CR/PR (n = 24) vs NR (n = 28) in HRS type 2.

In multivariate analysis using Cox regression, response to terlipressin treatment and LTx eligibility were independent predictors of longer survival, while overt hepatic encephalopathy (HE) and higher MELD were associated with increased risk of death (Table 3). Of note, hazard ratio for death was 0.36 in patients with HRS type 2 with treatment response compared with NRs (p = 0.026).

Table 3.

Cox regression – overall survival.

	Death
Variables	Hazard ratio	95% confidence interval	p value
LTx-eligibility	0.038	0.005–0.291	0.002
Terlipressin response	0.363	0.149–0.883	0.026
Overt HE (≥grade 2, West-Haven)	3.073	1.266–7.456	0.004
MELD	1.095	1.029–1.165	0.004
HCC	1.213	0.388–3.795	0.740

HCC: hepatocellular carcinoma; HE: hepatic encephalopathy; LTx: liver transplantation; MELD: Model for End-Stage Liver Disease.

Discussion

Terlipressin and albumin therapy in patients with HRS type 1 has improved survival. However, only few and small HRS type 2 cohorts with inconsistent results have been published. To our knowledge this observational study including 52 HRS type 2 patients represents the largest prospective cohort reported so far.

In a retrospective review of 69 patients with HRS, with 20 patients having HRS type 2, Testro and colleagues reported a response rate of 20% in this subtype, and no patient survived without LTx. The authors concluded that terlipressin therapy in non-LTx candidates might not be justified.¹⁴ However, several other studies reported that terlipressin therapy was beneficial in HRS type 2. In one of the early pilot studies on terlipressin, renal recovery was achieved in four of five patients with HRS type 2.⁹ Alessandria et al. evaluated the effectiveness of terlipressin in 11 patients with HRS type 2 with a response rate of 73%.²¹ Ghosh and colleagues conducted a randomised trial comparing noradrenaline with terlipressin in 46 HRS type 2 patients with a response rate of 76%.¹³ In a retrospective analysis of 31 patients with type 2 HRS awaiting LTx treated with terlipressin and albumin, the response rate was 61%; however, 11/19 patients with response relapsed.¹⁵ In our study the overall response rate was 47%, and therapy was similarly effective in HRS types 2 and 1 (46% vs 48%). While these response rates were generally lower than those reported in the two randomised controlled studies by Alessandria and Gosh et al., they probably reflect the effectiveness of terlipressin treatment in a real-world setting in non-selected patients.

HRS relapse was common in HRS type 2. In line with previous studies which have reported relapse rates between 35% and >50%, half of all treated patients in our study relapsed.^13,21 While HRS type 1 commonly develops after an acute complication such as infection or GI bleeding with a potentially reversible decompensation of liver and renal function, HRS type 2 develops in a state of chronic portal hypertension in patients with refractory ascites, and the higher recurrence rate after vasoconstrictor withdrawal most likely represents this difference in pathophysiology. Retreatment in patients with relapse yielded a response in 43%, supporting the concept of terlipressin treatment as a bridge to transplant in LTx-eligible patients. In patients deemed LTx ineligible, we observed no difference in survival in the subgroup of patients who underwent RRT compared with patients who received best supportive care and therefore such therapy does not seem to be justified.

Prediction of response in HRS type 2 patients before initiating treatment would be desirable. Several studies demonstrated that response to treatment in HRS type 1 is associated with Child-Pugh score, baseline serum creatinine concentration and MELD.^22,23 In HRS type 2, Ghosh and colleagues reported an association of response to noradrenalin or terlipressin with an increase in mean arterial pressure, urinary output and urinary sodium during treatment.¹³ While we observed a trend toward higher baseline serum creatinine concentrations in NR patients, neither clinical variables such as age, sex, HCC, ongoing alcohol abuse, overt HE, urinary output, systolic or diastolic blood pressure nor laboratory baseline variables such as serum bilirubin concentration, urinary sodium or MELD were associated with treatment response. It should be noted that blood pressure was measured non-invasively in most patients. Therefore, associations of haemodynamic variables with terlipressin response might have been revealed with more extensive invasive monitoring. However, our results indicate that predicting individual treatment response is difficult in daily clinical practice.

OS as well as LTx-free survival reported in this study were longer than reported in earlier studies, suggesting that widespread use of vasoconstrictor treatment for HRS, LTx and RRT have changed the course of the disease.^24,25 Several reports demonstrated that terlipressin treatment was associated with improved survival in HRS type 1 and in mixed cohorts with HRS types 1 and 2.^22,26,27 However, no study could conclusively establish this association in an HRS type 2-only cohort. Rodriguez and colleagues found no evidence for an improvement in outcome in HRS type 2, while in our study, terlipressin response was independently associated with improved OS in addition to other generally accepted prognostic clinical characteristics such as LTx eligibility, absence of overt HE and lower MELD.¹⁵ This difference might be explained by the impressively high efficiency of the Spanish organ donation system with >40 compared with nine individual donors per million people in Germany, thus reducing the necessity to medically bridge longer LTx waiting times in patients with advanced liver disease.²⁸ Therefore, a course of terlipressin therapy might be justified especially in countries with long transplant waiting lists. In combination with other well-established mortality risk factors, a vasoconstrictor treatment course might also facilitate identification of patients with a more favourable outcome.

Limitations of our study include its observatory nature. In the absence of a control group, we cannot causally attribute improved survival to terlipressin response. In addition our study was initiated in 2013 and includes patients recruited until February 2016 and therefore the definition of HRS and subclassification into HRS types 1 and 2 were performed using the German guideline on management of patients with ascites, HRS and SBP published in 2011. In subsequent years the definition and classification of renal failure in patients with liver cirrhosis has evolved based on a better understanding of the underlying pathophysiology and recognition of the potential for overlap forms of renal failure in these patients. Revised criteria for the diagnosis of AKI and HRS now emphasise the differentiation between HRS subtypes HRS-AKI (type 1) and non–HRS-AKI/HRS-CKD (type 2), which can be challenging in some cases since it depends on glomerular filtration rate estimates within the previous three months and requires a renal biopsy at least in some cases, which is not performed in daily clinical practice in patients with decompensated liver cirrhosis and renal impairment.^5,6 In the future urinary and serum biomarkers of renal tubular damage and HRS could help to better differentiate different types of kidney injury in these patients. The latest guidelines challenge the concept of HRS as a purely functional renal failure as well as the common classification into HRS types 1 and 2. Tubular damage can occur in HRS-AKI (type 1), and patients with HRS-AKI can have underlying CKD. Therefore, future studies have to examine how evolving definitions of HRS and renal failure in patients with liver cirrhosis in studies and guidelines actually transfer to daily clinical practice.

In conclusion terlipressin treatment seems to be effective in HRS type 2 and can reverse renal failure in approximately 50% of patients. Our results suggest that HRS subclassification does not have a large influence with regard to effectiveness of terlipressin treatment. While individual response is difficult to predict and rate of HRS recurrence is higher than in HRS type 1, a course of therapy seems justified in selected type 2 patients, especially in countries with long transplant waiting lists, and response to treatment is associated with better outcome.

Declaration of conflicting interests

None declared.

Ethics approval

The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki was approved by the ethics committee of the State Physician Chamber of Rhineland-Palatinate, Germany (9 July 2012). The study was registered at the German Clinical Trials Register (DRKS; DRKS00013117).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Informed consent

All patients provided written, informed consent to participate in this study.