Terlipressin for Hepatorenal Syndrome in Patients With Early‐Stage Acute‐on‐Chronic Liver Failure

ABSTRACT

Background & Aims

Hepatorenal syndrome‐acute kidney injury (HRS‐AKI) is a life‐threatening complication of decompensated cirrhosis. The US Food and Drug Administration approved terlipressin use for HRS‐AKI based on the CONFIRM study, which demonstrated a significant improvement in HRS reversal with terlipressin versus placebo. The label notes elevated risk of respiratory failure in patients with volume overload or acute‐on‐chronic liver failure (ACLF) grade 3 and limited benefit when serum creatinine (SCr) exceeds 5 mg/dL.

Methods

We performed a post hoc analysis of CONFIRM excluding patients with ACLF grade 3 or SCr ≥ 5 mg/dL. This allowed us to assess the efficacy and safety of terlipressin in a population where the benefit‐to‐risk profile is more favourable. Efficacy outcomes included HRS reversal, renal replacement therapy (RRT), liver transplantation (LT), RRT‐free survival, LT‐free survival and overall survival. Changes in SCr, Model for End Stage Liver Disease (MELD) and sodium were also assessed.

Results

HRS reversal occurred in 43% (60/141) of patients with terlipressin versus 17% (13/75) with placebo (p < 0.001). Terlipressin was associated with significantly larger reductions (vs. placebo) in SCr (p < 0.001) and increases in serum sodium (p < 0.001). Importantly, LT rates were similar even though MELD scores decreased. 90‐day survival was similar between treatment arms. Notably, selecting patients with a favourable benefit‐to‐risk profile led to a similar incidence of respiratory failure between treatment arms (11% with terlipressin vs. 7% with placebo; p = 0.360).

Conclusions

In patients with HRS‐AKI without baseline ACLF grade 3 or SCr ≥ 5 mg/dL, terlipressin improved clinical outcomes and was not associated with an increased risk of respiratory failure.

Trial Registration

CONFIRM, ClinicalTrials.gov identifier: NCT02770716

Abbreviations

AASLD

American Association for the Study of Liver Diseases

ACLF

acute‐on‐chronic liver failure

AE

adverse event

ANOVA

analysis of variance

CI

confidence interval

EASL

European Association for the Study of the Liver

EOT

end of treatment

FDA

US Food and Drug Administration

HRS

hepatorenal syndrome

HRS‐1

hepatorenal syndrome type 1

HRS‐AKI

hepatorenal syndrome‐acute kidney injury

ICA

International Club of Ascites

INR

international normalised ratio

LT

liver transplantation

MAP

mean arterial pressure

MELD

Model for End‐Stage Liver Disease

MODS

multiple organ dysfunction syndrome

Na

sodium

OS

overall survival

RRT

renal replacement therapy

SAE

serious adverse event

SAS

statistical analysis software

SCr

serum creatinine

SD

standard deviation

SIRS

systemic inflammatory response syndrome

SLKT

simultaneous liver‐kidney transplant

SpO₂/FiO₂

pulse oximetry/fraction of inspired oxygen

TFS

transplant‐free survival

V₁

vasopressin receptor type 1

V₂

vasopressin receptor type 2

Summary.

• Hepatorenal syndrome‐acute kidney injury (HRS‐AKI) is a serious condition in patients with advanced liver disease in which kidney function is markedly impaired.

• Terlipressin—the only US Food and Drug Administration‐approved drug to treat patients with HRS‐AKI to improve kidney function—can cause respiratory (breathing) problems in certain patients.

• This post hoc analysis of the CONFIRM trial demonstrates that terlipressin improves kidney function and does not appear to cause increased respiratory problems when the condition is treated at an earlier or less severe stage.

• The results emphasise the importance of carefully selecting those patients with HRS‐AKI who will most benefit from terlipressin treatment.

1. Background

Hepatorenal syndrome (HRS) is a life‐threatening deterioration in kidney function that occurs in patients with advanced cirrhosis, portal hypertension and ascites [1], and is characterised by splanchnic and systemic arterial vasodilation and hyperdynamic circulation [2, 3]. Left untreated, patients with HRS have a very poor prognosis, with median survival estimated from days to weeks [3].

Historically, a potentially reversible, rapidly progressive form of HRS was classified as HRS type 1 (HRS‐1) and was diagnosed based on a fixed serum creatinine (SCr) threshold (i.e., a doubling of SCr to a level > 2.5 mg/dL in < 2 weeks) [4]. To facilitate an earlier diagnosis and treatment of patients with HRS, in 2015, the International Club of Ascites (ICA) revised the criteria to define HRS‐acute kidney injury (HRS‐AKI) [5] based on the degree of change in SCr (i.e., an absolute increase in SCr of ≥ 0.3 mg/dL within 48 h or an increase in SCr of ≥ 1.5 times from baseline) and a lack of response 48 h after plasma volume expansion with 1 g/kg albumin and withdrawal of diuretics [6]. The criteria for HRS‐AKI have replaced the criteria for HRS‐1 [3, 5].

Based on data indicating that terlipressin improves renal function in patients with HRS‐AKI, the American Association for the Study of Liver Diseases (AASLD) and the European Association for the Study of the Liver (EASL) recommend a combination of terlipressin and albumin as first‐line treatment for patients with HRS‐AKI [6, 7]. By improving kidney function, terlipressin may prolong short‐term survival in patients with HRS‐AKI who respond to treatment [8], thereby providing a bridge to liver transplantation (LT)—the only definitive treatment for critically decompensated cirrhosis [9]. HRS reversal improves LT outcomes [10] and reduces the need for peritransplant renal replacement therapy (RRT), which is the most important predictive factor of posttransplant renal recovery and survival [11, 12].

CONFIRM, the largest randomised, placebo‐controlled study of terlipressin in patients with HRS‐AKI (i.e., HRS‐1 at the time of enrolment; N = 300), was designed prior to the 2015 ICA diagnostic criteria and demonstrated a significant improvement in renal function in patients given terlipressin compared with placebo [13]. Based on the results from CONFIRM, terlipressin is the first and only vasoconstrictor approved by the US Food and Drug Administration (FDA) to improve kidney function for adults with HRS and a rapid reduction in kidney function [14]. In addition to kidney failure, patients with HRS are at risk of other organ failures, cumulatively referred to as acute‐on‐chronic liver failure (ACLF) [15]. ACLF is characterised by the functional failure of one or more of the six major organ systems (i.e., liver, kidney, brain, coagulation, circulation and respiration), and depending on the number and severity of affected organs, is graded from 1 to 3 [15, 16]. Based on the eligibility criteria in CONFIRM (i.e., patients with HRS‐AKI and a doubling of SCr to a level ≥ 2.25 mg/dL in less than 2 weeks before randomisation), all patients in the study also met criteria for ACLF grade 1 or higher based on the threshold for kidney failure of an SCr level of ≥ 2 mg/dL [16]. Patients with more advanced liver disease at enrolment (i.e., bilirubin > 12 mg/dL, international normalised ratio [INR] ≥ 2.5, hepatic encephalopathy stages 3–4, pulse oximetry/fraction of inspired oxygen [SpO₂/FiO₂] ≤ 214) were categorised as ACLF grade 2 or 3 [17]. Several studies have demonstrated that the severity of ACLF is inversely associated with treatment success with terlipressin (i.e., achieving HRS reversal) [17, 18, 19]; additionally, the cumulative 90‐day mortality of HRS increases in a stepwise fashion according to ACLF grade (30%, 50% and 79% for ACLF grades 1, 2 and 3, respectively) [18].

Although the mechanisms are not entirely understood, terlipressin therapy has previously been associated with serious adverse events (SAEs) of respiratory failure in certain patients with HRS [13]. For example, a post hoc analysis of CONFIRM data revealed that patients with ACLF grade 3 at study entry were at the highest risk of developing respiratory failure [17, 20]. Additionally, HRS reversal is more likely to occur in patients with a lower SCr at the start of therapy [8, 20, 21]. Taken together, these findings prompted the FDA label to recommend against terlipressin therapy in patients with ACLF grade 3 or SCr ≥ 5 mg/dL, which inspired this post hoc analysis in the corresponding subpopulation of patients from CONFIRM. Herein, we report the effect of terlipressin (vs. placebo) on the incidence of HRS reversal, the need for RRT, changes in electrolyte concentrations, Model for End‐Stage Liver Disease (MELD) score, the rate of LT, survival and adverse events (AEs) in patients with HRS‐AKI without baseline ACLF grade 3 or SCr ≥ 5 mg/dL.

2. Methods

2.1. Patients and Treatment

This is a post hoc analysis of the subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL (excluded patients that met either or both of these exclusion criteria) from the CONFIRM study [13] (NCT02770716). In brief, adult patients (≥ 18 years of age) with cirrhosis, ascites and HRS‐AKI were eligible for the study. Patients were diagnosed according to slightly modified historic HRS‐1 criteria, which, to facilitate an earlier diagnosis, included a lower threshold of SCr ≥ 2.25 mg/dL (compared with SCr > 2.5 mg/dL [4]) and a doubling of SCr within 14 days prior to randomisation (or a slope of increase in SCr levels consistent with expected doubling within 2 weeks), with no sustained improvement in renal function (defined as < 20% decrease in SCr, and SCr ≥ 2.25 mg/dL), 48 h after diuretic withdrawal and plasma volume expansion with albumin [13]. Patients were randomly assigned (2:1) to treatment with either terlipressin 1 mg intravenously or matched placebo every 6 h; per the study protocol, on Day 4, the dose could be increased to up to 2 mg every 6 h if SCr decreased, but by < 30% from baseline after a minimum of 10 doses of the study drug [13]. Concomitant albumin administration (1 g/kg to ≤ 100 g on Day 1, followed by 20–40 g/day) was strongly recommended for all patients [13].

2.2. Efficacy

Efficacy outcomes were retrospectively assessed by treatment arm (terlipressin vs. placebo). HRS reversal was defined as the percentage of patients with an SCr ≤ 1.5 mg/dL while on treatment (i.e., up to 24 h after the final dose of study drug up to Day 14 or discharge); any SCr values obtained posttransplant or after RRT were excluded. The incidence of RRT, RRT‐free survival, incidence of LT, transplant‐free survival (TFS), and overall survival (OS) were assessed in the subpopulation of all randomised patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. Changes in SCr, bilirubin, INR and sodium (Na) were evaluated from baseline to the end of treatment (EOT) in the safety analysis population, defined as all randomised patients without ACLF grade 3 or SCr ≥ 5 mg/dL who received at least 1 dose of study drug (terlipressin or placebo). Baseline was defined as Day 0 of the study period; however, a prestudy period value could be used if the Day 0 value was missing.

2.3. Safety

Safety data were collected prospectively during the original clinical study (for non‐serious AEs up to 7 days posttreatment, and SAEs up to 30 days after the end of the treatment period) and retrospectively assessed in the safety analysis population. Respiratory failure was reported by the study investigators using the preferred terms of ‘respiratory failure’ or ‘acute respiratory failure’ and was established based on the study investigators' clinical assessment. A subsequent individual patient review showed no meaningful difference between the characteristics of the AEs that were reported using these 2 terms [17], therefore, the terms were combined and reported collectively as ‘respiratory failure’.

2.4. Statistical Analyses

Comparisons of the nominal outcomes were performed by a Chi‐square test (if the number of events per cell was ≥ 5) or a Fisher's exact test (if the number of events per cell was < 5). Numerical baseline characteristics, use of albumin, changes in MELD score, and compound concentrations from baseline to the EOT were compared between treatment arms by Analysis of Variance (ANOVA) or Kruskal–Wallis tests following testing for normality.

Survival estimates were assessed using the Kaplan–Meier method; p values comparing survival estimates were derived from a 2‐sample log‐rank test stratified by qualifying SCr (< 3.4 mg/dL vs. ≥ 3.4 mg/dL) and prior large‐volume paracentesis within 14 days of randomisation (at least 1 single event of ≥ 4 L vs. < 4 L). All statistical tests were 2‐sided with a final significance level of 0.05. Statistical analyses were performed using Statistical Analysis Software (SAS) version 9.4.

All authors had access to the study data and reviewed and approved the final manuscript.

3. Results

3.1. Patients

Out of 300 patients enrolled in CONFIRM, 141 patients in the terlipressin arm and 75 patients in the placebo arm of the overall CONFIRM population met the analysis population criteria (i.e., without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline; excluded patients met either or both of these exclusion criteria). Baseline patient characteristics in both treatment arms were similar, with mean MELD scores of 31 in both arms and mean SCr levels of 3.2 mg/dL and 3.3 mg/dL in the terlipressin and placebo arms, respectively (Table 1).

TABLE 1.

Baseline demographics and clinical characteristics.

Parameter at baseline	Terlipressin (n = 141)	Placebo (n = 75)	p
Age (years), mean ± SD	56 ± 10.6	54 ± 12.0	0.347
Sex, male, n (%)	84 (60)	44 (59)	0.897
Alcohol‐associated hepatitis, n (%)	48 (34)	24 (32)	0.762
SIRS subgroup, n (%)	54 (38)	32 (43)	0.532
MAP (mmHg), mean ± SD	78 ± 12.2	77 ± 9.3	0.488
SCr (mg/dL), mean ± SD	3.2 ± 0.7	3.3 ± 0.8	0.413
Bilirubin (mg/dL)	n = 131	n = 73	0.075
Mean ± SD	9.1 ± 10.8	12.6 ± 13.2
Child–Pugh score	n = 138	n = 71	0.397
Mean ± SD	10 ± 2	10 ± 2
MELD score	n = 120	n = 62	0.406
Mean ± SD	31 ± 6	31 ± 6
ACLF grade, n (%)
1	89 (63)	39 (52)	0.113
2	52 (37)	36 (48)

Note: CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. p values were calculated from a Chi‐square or Fisher's exact test for categorical data, and from ANOVA or a Kruskal–Wallis test for numerical data.

Abbreviations: ACLF, acute‐on‐chronic liver failure; ANOVA, analysis of variance; MAP, mean arterial pressure; MELD, Model for End‐Stage Liver Disease; SCr, serum creatinine; SD, standard deviation; SIRS, systemic inflammatory response syndrome.

Albumin was administered before randomisation to 99% of patients in the analysis population, while 87% and 93% of patients in the terlipressin and placebo arms, respectively, received albumin during active treatment (Table 2). The total amount of albumin (mean ± standard deviation [SD]) received prior to randomisation was 326 ± 167 g in the terlipressin arm and 349 ± 293 g in the placebo arm. During treatment, on average, patients received 204 ± 151 g and 252 ± 200 g of albumin in each treatment arm, respectively.

TABLE 2.

Albumin received before and during treatment, and use of midodrine and octreotide prior to randomisation.

Parameter	Terlipressin (n = 141)	Placebo (n = 75)	p
Total albumin prior to randomisation
n (%)	140 (99)	74 (99)	0.934
Mean ± SD, (g)	326 ± 167	349 ± 293
Total albumin during treatment after randomisation
n (%)	122 (87)	70 (93)	0.139
Mean ± SD, (g)	204 ± 151	252 ± 200
Midodrine and/or octreotide use prior to randomisation, a n (%)	113 (80)	58 (77)	0.628

Note: CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. p values were calculated from a Chi‐square or Fisher's Exact test.

Abbreviations: ACLF, acute‐on‐chronic liver failure; SCr, serum creatinine; SD, standard deviation.

^a Patients receiving a vasopressor other than midodrine within 24 h of a qualifying SCr level were excluded from the trial; midodrine and octreotide were allowed but were to be stopped prior to randomisation.

Patients with ACLF grade 3 or SCr ≥ 5 mg/dL at baseline were excluded from this post hoc analysis. Among the excluded patients, the total amount of albumin (mean ± SD) received prior to randomisation was 357 ± 174 g in the terlipressin arm and 431 ± 210 g in the placebo arm. Total albumin received during treatment after randomisation was 186 ± 136 g in the terlipressin arm and 201 ± 115 g in the placebo arm (Table S1).

If discontinued before randomisation, prior use of midodrine and/or octreotide was allowed (after completion of 24‐h washout period) per the CONFIRM study protocol; 80% and 77% of patients in the analysis population received midodrine and/or octreotide before study treatment in the terlipressin and placebo arms, respectively (Table 2).

3.2. HRS Reversal and RRT

Significantly more patients achieved HRS reversal in the terlipressin arm versus placebo (terlipressin: 43% [60/141] vs. placebo: 17% [13/75], p < 0.001, Figure 1). The incidence of RRT was 23% versus 31% by Day 30; 27% versus 32% by Day 60; and 28% versus 33% by Day 90 in the terlipressin and placebo arms, respectively; the incidences were not significantly different (Figure 1). By the end of the study, 4 patients (3 terlipressin; 1 placebo) had received a simultaneous liver‐kidney transplant (SLKT). Excluding patients who had an SLKT, 41% (56/138) versus 35% (26/74) of patients in the respective treatment arms were alive and RRT‐free at Day 90. Median RRT‐free survival (95% confidence interval [CI]) was 44 (24–72) days in the terlipressin arm and 18 (10–37) days in the placebo arm (p = 0.102, Figure 2). Post‐LT (among patients who were listed for LT at baseline and excluding those with an SLKT), the need for RRT was 13% versus 20% by Day 30; 17% versus 30% by Day 60; and 21% versus 40% by Day 90 in the terlipressin and placebo arms, respectively; however, the difference was not statistically significant (Table S2).

FIGURE 1.

HRS reversal and RRT. CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. The p value to compare HRS reversal incidence was calculated from a Chi‐square test, and for comparisons of RRT incidence from a Cochran–Mantel–Haenszel test stratified by qualifying SCr (< 3.4 mg/dL vs. ≥ 3.4 mg/dL) and prior large volume paracentesis within 14 days of randomisation. ACLF, acute‐on‐chronic liver failure; CI, confidence interval; HRS, hepatorenal syndrome; RRT, renal replacement therapy; SCr, serum creatinine.

FIGURE 2.

RRT‐free survival. CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline, excluding those who received an SLKT. p values were calculated from a 2‐sample log‐rank test comparing terlipressin vs. placebo. RRT and death were considered ‘events’. ACLF, acute‐on‐chronic liver failure; RRT, renal replacement therapy; SCr, serum creatinine; SLKT, simultaneous liver‐kidney transplant.

3.3. The Effect of Terlipressin on Hyponatraemia

Since 1 patient who had been assigned to receive placebo inadvertently received 1 dose of terlipressin [13], the safety analysis population included 142 patients in the terlipressin arm and 74 patients in the placebo arm. In the safety analysis population, serum Na concentration at baseline was 133 ± 6 mmol/L (mean [±SD]) in the terlipressin arm and 133 ± 5 mmol/L in the placebo arm. From baseline to the EOT, serum Na increased significantly more in terlipressin‐treated patients compared with placebo: the mean (±SD) change in Na was 5 ± 5.0 mmol/L versus 2 ± 5.3 mmol/L, respectively (p < 0.001) (Table S3).

3.4. The Effect of Treatment on MELD Score and Liver Transplantation Outcomes

In the safety analysis population, the MELD score (mean ± SD) at baseline was 31 ± 6 in both the terlipressin and placebo arms (Figure 3A, Table S4). Information on MELD score at the EOT was available for 175 patients (terlipressin, n = 118; placebo, n = 57). From baseline to the EOT, MELD score decreased significantly more among terlipressin‐treated patients: the mean (±SD) change was −4 ± 5 versus −1 ± 4 in the terlipressin and placebo arms, respectively; p < 0.001 (Figure 3B).

FIGURE 3.

(A) MELD scores and (B) changes in MELD scores from baseline to the EOT. CONFIRM safety subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. The p value to compare changes in MELD score was calculated from an ANOVA and a Kruskal–Wallis test following testing for normality. ACLF, acute‐on‐chronic liver failure; ANOVA, analysis of variance; EOT, end of treatment; MELD, Model for End‐Stage Liver Disease; SCr, serum creatinine; SD, standard deviation.

Changes in MELD score were primarily due to a decrease in SCr levels: the mean (±SD) change in SCr among terlipressin‐treated patients was −0.9 ± 1.1 mg/dL versus 0 ± 1.3 mg/dL among placebo‐treated patients (p < 0.001); changes in the bilirubin concentration or INR values from baseline and the EOT were not significantly different between treatment arms (Table S3).

Despite a greater decrease in MELD score from baseline to the EOT among terlipressin‐treated patients (vs. placebo), the incidence of LT was similar in both treatment arms (19% vs. 21% by Day 30; 26% vs. 24% by Day 60; and 28% vs. 25% by Day 90, respectively; all p > 0.70, Figure 4).

FIGURE 4.

Liver transplantation. CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. p values were calculated from a Chi‐square test. ACLF, acute‐on‐chronic liver failure; LT, liver transplantation; SCr, serum creatinine.

3.5. Transplant‐Free and Overall Survival

TFS and OS were similar in both treatment arms. The median TFS (95% CI) was 31 (22–40) days in the terlipressin arm and 24 (15–44) days in the placebo arm (p value not significant); 29% versus 33% of patients were transplant‐free and alive by Day 90 (Figure 5A). There were no significant differences in OS estimates, with 57% of patients in both treatment arms alive by Day 90; median OS was not estimable as the OS function did not reach 50% by Day 90 (Figure 5B).

FIGURE 5.

(A) Transplant‐free survival^a and (B) overall survival up to 90 days by treatment arm. CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline. The p values were calculated from a 2‐sample log‐rank test comparing terlipressin versus placebo. ^aDeath and transplant were counted as ‘events’. ACLF, acute‐on‐chronic liver failure; SCr, serum creatinine.

3.6. Safety

The majority of patients in both treatment arms experienced an AE (terlipressin: 94%, placebo: 97%); drug discontinuation due to AEs was reported in 12% of terlipressin‐treated and 4% of placebo‐treated patients (p = 0.081). The incidence of respiratory failure—an SAE of particular concern—was comparable between treatment arms (terlipressin: 11% [15/142]; placebo: 7% [5/74], p = 0.360) (Table 3). A separate post hoc analysis was performed among the patients who experienced respiratory failure and it was observed that the baseline demographics and clinical characteristics were generally similar in the treatment arms (Table S5).

TABLE 3.

Adverse events.

Event, n (%)	Terlipressin (n = 142)	Placebo (n = 74)	p
AEs of any grade	134 (94)	72 (97)	0.500
AEs leading to study drug discontinuation	17 (12)	3 (4)	0.081
SAEs of > 3% frequency in either treatment arm	95 (67)	44 (60)	0.278
SAEs of > 3% by preferred term
Abdominal pain	6 (4)	1 (1)	0.427
MODS	5 (4)	3 (4)	1.00
Chronic hepatic failure	11 (8)	7 (10)	0.666
Hepatic cirrhosis	3 (2)	3 (4)	0.414
Hepatic failure	7 (5)	4 (5)	1.00
Hepatorenal syndrome	2 (1)	3 (4)	0.341
Sepsis	6 (4)	0	0.097
Septic shock	6 (4)	0	0.097
Hepatic encephalopathy	7 (5)	3 (4)	1.00
Acute kidney injury	5 (4)	1 (1)	0.667
Respiratory failure a	15 (11)	5 (7)	0.360

Note: CONFIRM subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL at baseline.

Abbreviations: ACLF, acute‐on‐chronic liver failure; AE, adverse event; MODS, multiple organ dysfunction syndrome; SAE, serious adverse event; SCr, serum creatinine.

^a Acute respiratory failure or respiratory failure.

4. Discussion

In this analysis of the subpopulation of patients without ACLF grade 3 or SCr ≥ 5 mg/dL from the CONFIRM study, the incidence of HRS reversal was 43% in the terlipressin arm compared with 17% in the placebo arm (p < 0.001). In the overall CONFIRM population, in which 19% of patients had ACLF grade 3 and 10% of patients had SCr ≥ 5 mg/dL, HRS reversal was achieved in 39% of patients in the terlipressin arm and 18% of patients in the placebo arm (p < 0.001) [13, 17].

The safety of terlipressin in the current study population was similar to that reported in CONFIRM and reflects the seriousness of HRS: over 94% of patients experienced AEs of any grade, and over 60% experienced SAEs [13]. In the current analysis, the incidence of respiratory failure AEs in terlipressin‐treated patients was 11%, which was not statistically significantly different from the 7% incidence among placebo‐treated patients. As previously reported, in the overall CONFIRM population, 13.5% of patients experienced respiratory failure [17]. Baseline demographics and clinical characteristics between this subpopulation (i.e., patients without ACLF grade 3 or SCr ≥ 5 mg/dL) and the patients who experienced respiratory failure were similar. Judicious patient selection (as per FDA package insert recommendations) appears to lead to clinically meaningful mitigation of the risk of respiratory failure in patients receiving terlipressin.

The higher numerical incidence of HRS reversal and TFS in the current subpopulation, as compared with the terlipressin arm in the entire CONFIRM population, stems from the exclusion of patients with ACLF grade 3 or SCr ≥ 5 mg/dL, who—based on the current understanding of the clinical factors associated with renal recovery in patients with decompensated cirrhosis—were less likely to derive clinical benefit and more likely to experience SAEs [3, 5]. Similarly, several previous studies demonstrated a reduction in clinical benefit from terlipressin treatment with an increase in baseline ACLF grade and SCr. For example, in a retrospective analysis of 298 consecutive patients with HRS‐AKI, ACLF grade and baseline SCr were the only independent predictors of response to treatment, with the probability of HRS reversal decreasing and the likelihood of 90‐day mortality increasing in a stepwise fashion from ACLF grade 1 to grades 2 and 3 [18]. Further, numerous individual trials [8, 20, 21], as well as a recent meta‐analysis of seven studies [22], confirmed a significant association between a higher baseline SCr level and nonresponse to terlipressin treatment. Our results are in line with these findings, thereby highlighting the importance of an earlier diagnosis and therapy initiation while SCr is relatively low.

Among patients with HRS‐AKI, the need for RRT is associated with an increased AE risk and confers a poor prognosis, with a 3‐month mortality rate of 80% [23, 24]. However, in some patients, RRT may improve short‐term survival and be clinically justified, especially for LT‐listed patients [6, 24]. Unfortunately, while life‐sustaining, each additional day of dialysis is associated with a 6% increased probability of renal nonrecovery posttransplant [11]. Response to terlipressin was previously shown to reduce the need for RRT before and after LT [25, 26], and to decrease the risk of chronic kidney disease within a year after LT [26]. In the current analysis of the CONFIRM subpopulation without ACLF grade 3 or SCr ≥ 5 mg/dL, the overall incidence of RRT was not significantly different between treatment arms, which is not surprising because this analysis was not pre‐specified in the CONFIRM study and was conducted post hoc in a smaller subpopulation that was not sufficiently powered for statistical comparisons. Notably, the RRT‐free‐survival curve (Figure 2) indicates the possibility of a beneficial effect of terlipressin on the need for RRT that can be observed within approximately the first 30–60 days, which is consistent with the understanding that terlipressin is not a definitive treatment for the underlying liver disease that causes HRS‐AKI [9]. Given these data, it is notable that analysis of the larger data set (N = 608) from the pooled Phase III placebo‐controlled clinical studies of terlipressin (i.e., CONFIRM, REVERSE and OT‐0401) demonstrated a significant reduction in RRT incidence by Days 30 and 60 overall, and by Days 60 and 90 post‐LT (after the last dose of study drug) in the entire population of patients treated with terlipressin (vs. placebo) [27].

In this CONFIRM study subpopulation, MELD score decreased to a greater extent in the terlipressin arm than in the placebo arm. Notably, changes in MELD score were predominantly due to improved renal function, manifested by the significant reduction in SCr; as expected, changes in other MELD score components (i.e., INR and bilirubin) were not significantly different between treatment arms. Additionally, many patients in the study had hyponatraemia at baseline (with mean Na levels of 133 mmol/L in both treatment arms). This electrolyte abnormality is commonly observed in patients with advanced cirrhosis who have severe portal hypertension leading to activation of the renin‐angiotensin system, and is associated with a poor prognosis [6, 28, 29]. Terlipressin—a vasopressin analogue—has an affinity for both vasopressin receptor type 1 (V₁), which mediates vasoconstriction, and vasopressin receptor type 2 (V₂), which mediates the antidiuretic effect of vasopressin, with an approximate 6‐fold higher affinity for V₁ receptors versus V₂ [30]. It was previously noted that among patients with acute portal‐hypertensive bleeding, terlipressin may worsen hyponatraemia via an antidiuretic effect by activating the V₂ receptors; however, worsening of hyponatraemia has rarely been observed in patients treated with terlipressin for HRS‐AKI [31]. This difference could be explained by a high endogenous concentration of vasopressin, which is characteristic of patients with very advanced liver disease and results in saturation of the V₂ receptors and, consequently, the absence of a significant antidiuretic effect [31]. Consistent with this explanation, our analysis demonstrates an increase in serum Na concentration among patients with HRS‐AKI following treatment with terlipressin: while patients in both treatment arms had improved serum Na levels by the EOT, terlipressin treatment resulted in a significantly larger median increase in serum Na concentration compared with placebo.

SCr and Na are components of both MELD‐Na and MELD 3.0 scores, the latter of which is currently used for LT allocation [32]. HRS reversal may lead to a reduction in MELD scores—driven by changes in SCr and serum Na. Notably, in our study, terlipressin treatment did not impact LT rates; rates were similar in both treatment arms. Additionally, specific strategies, such as using the baseline MELD score for patients with HRS‐AKI who respond to pharmacological treatment, have been proposed as mitigation strategies [26, 33].

We recognise the limitations of this study. First, caution is urged when comparing safety and efficacy outcomes of this analysis with the previously published CONFIRM results, since our analysis was conducted retrospectively in a subset of patients from the overall CONFIRM population. Additionally, a relatively low sample size in the subgroups has negatively affected statistical power, and, due to the post hoc nature of the statistical analysis, the interpretation of p values is limited to signal screening. A follow‐up duration of only 90 days is a further limitation, especially for the post‐LT analysis. The patients in this study had longer treatments with substantial amounts of albumin and midodrine than international guideline recommendations limiting generalisability of the findings. However, the amount of albumin received was generally similar between the terlipressin‐treated patients included in this study and those excluded from this study (patients with ACLF grade 3 or SCr ≥ 5 mg/dL), which suggests that the improved terlipressin safety profile in patients without ACLF grade 3 or SCr ≥ 5 mg/dL compared with the population in CONFIRM are unlikely to be driven by differences in the amount of albumin received.

In summary, this analysis reinforces the utility of terlipressin as an effective treatment for patients with HRS‐AKI, particularly in the subgroup of patients without ACLF grade 3 or SCr ≥ 5 mg/dL, where terlipressin is expected to improve renal function and reverse HRS in over 40% of these patients. Results also support that patients with earlier‐stage disease (i.e., without ACLF grade 3 and without SCr ≥ 5 mg/dL) do not have an increased risk of respiratory failure when treated with terlipressin compared with placebo. Importantly, the data suggest that early diagnosis and initiation of therapy, coupled with careful patient selection (as suggested by the FDA package insert) and monitoring for changes in respiratory status, are likely to lead to optimal treatment outcomes.

Author Contributions

All authors participated in the study design and data interpretation of this post hoc study. The study sponsor performed the data analysis. All authors discussed the results, contributed to review and revisions, and approved the manuscript for submission.

Ethics Statement

The original CONFIRM study was conducted in accordance with the ethical principles of the Declaration of Helsinki, received prior institutional review board and/or independent ethics committee approval at each study site, and written informed patient consent prior to study initiation.

Conflicts of Interest

Don C. Rockey received grant/research support from AstraZeneca, Boehringer Ingelheim, Galectin Therapeutics, Gilead Sciences, Intercept Pharmaceuticals, Mallinckrodt Pharmaceuticals, Novo Nordisk, and Pfizer; received consulting fees from AstraZeneca, Resolution Therapeutics, and Takeda. Fredric Gordon serves on a Speakers Bureau for Mallinckrodt Pharmaceuticals, Ipsen, Madrigal, and Gilead. Paul J. Thuluvath received consulting/speaker fees from AbbVie, Gilead, Mallinckrodt Pharmaceuticals; grant support from AbbVie, Alnylam, AstraZeneca, Bayer, BioVie Pharma, Bristol Myers Squibb, Cirius Therapeutics, Conatus, Corona, Cosmic, CStone Pharmaceuticals, Cymabay, Eisai, Enanta Pharmaceuticals, Enyo Pharma, Exact Sciences, Exelixis, Gilead, Hanmi Pharmaceutical, Intercept, Inventiva, Laboratory for Advanced Medicine Inc. (LAM), NGM Bio, Novo Nordisk, Roche, Salix, Takeda, Target, and Viking. David Victor serves on a Speakers Bureau for Gilead and Madrigal; serves on an Advisory Board for Intercept Pharmaceuticals. Nyingi Kemmer has no conflicts of interest to declare. Sanaz Cardoza is an employee of Mallinckrodt Pharmaceuticals. Khurram Jamil was an employee of Mallinckrodt Pharmaceuticals at the time of the study. R. Todd Frederick received grant/research support from AstraZeneca, Mallinckrodt Pharmaceuticals, and Salix; received consulting fees from Mallinckrodt Pharmaceuticals and TenNor Therapeutics.

Supporting information

Table S1: Albumin received before and during treatment, in patients with ACLF grade 3 or SCr ≥ 5 mg/dL, who were excluded from the study.

Table S2: Summary of posttransplant RRT in patients who were listed for a liver transplant at baseline and received a transplant through Day 90.

Table S3: Changes in serum sodium and MELD score components from baseline to the end of treatment.

Table S4: MELD scores and the change from baseline to the end of treatment.

Table S5: Baseline demographics and clinical characteristics and albumin use in patients who experienced respiratory failure.

Rockey D. C., Gordon F., Thuluvath P. J., et al., “Terlipressin for Hepatorenal Syndrome in Patients With Early‐Stage Acute‐on‐Chronic Liver Failure,” Liver International 45, no. 12 (2025): e70399, 10.1111/liv.70399.

Data Availability Statement

Summary aggregate (basic) results (including adverse event information) and the study protocol will be available on clinicaltrials.gov (CONFIRM, NCT02770716) when required by regulation. Individual de‐identified patient data will not be disclosed. Requests for additional information should be directed to the company at medinfo@mnk.com.