Hepatorenal syndrome (HRS) is defined as a potentially reversible kidney failure in patients with liver cirrhosis, acute liver failure, or alcoholic hepatitis [
Type I HRS is a subtype with poor prognosis [
A number of interventional studies indicate that a combined intravenous therapy with vasopressor drugs such as terlipressin and human albumin improves kidney function and enhances survival in type I HRS patients [
Combined terlipressin/albumin treatment has emerged as standard medical treatment for patients with HRS type I in the last years. Guidelines released by the German Society of Gastroenterology (Deutsche Gesellschaft für Verdauungs- und Stoffwechselerkrankungen (DGVS)) in 2011 recommend combined terlipressin/albumin i.v. treatment in patients with HRS type I [
In this retrospective, single-center analysis of a tertiary care center (University Hospital of Münster, Departments of Gastroenterology and Transplantation Medicine), we aimed to delineate treatment patterns, doses, and clinical outcomes of patients with HRS intravenously treated with the vasopressin analogue terlipressin. Furthermore, we conducted regression analysis in order to identify predictors of survival in patients with HRS.
The study was conducted as a retrospective, single-center analysis using the International Classification of Diseases (ICD) endorsed by the World Health Organization (WHO). From January 2005 to March 2014, inpatients at the Departments of Gastroenterology and Transplantation Medicine, University of Münster, were coded as hepatorenal syndrome (ICD code K76.7). A total of 119 complete files of patients treated from HRS were retrieved and were appropriate for retrospective analysis. HRS patients with response to treatment were compared to the patient cohort without improvement. Baseline characteristics (e.g., age, gender, and underlying disease) were retrieved. The present study has been performed in accordance with the ethical standards laid down in the 2000 Declaration of Helsinki.
For this study, complete patients files coded for ICD code K76.7 (hepatorenal syndrome) were screened for inclusion in this study. Reasons for exclusion of analysis were incomplete patient files, patients not having received terlipressin treatment, multiple files per patient, or miscoding/misclassification with regard to hepatorenal syndrome as classified by the criteria established by Salerno et al. in 2007 [
Retrospectively, patient files were assessed for the presence of hepatorenal syndrome types I or II using the modified Salerno criteria; all patients had to present with serum creatinine of >1.5 mg/dL, and patients with other reasons for acute renal dysfunction (e.g., SIRS or sepsis) were excluded.
The treatment response was defined as any serum creatinine of 1.5 mg/dL or below after therapy with terlipressin was commenced. In all patients, the retention parameters serum creatinine and urea were measured the first day after treatment initiation. Both retention parameters were then monitored every 48–72 hrs.
Descriptive analysis was used to document the demographic and clinical data of the patients. Data were analyzed using SPSS 17.0 (Chicago, IL, USA). Results are expressed as means ± standard deviation or medians [interquartile range]. Comparisons between groups were performed by using the Mann-Whitney
Within the study period of nine years, 65 HRS patients (55%) with response to terlipressin therapy were compared to a cohort of 54 patients (45%) without adequate response. The median age and gender distribution were similar in both groups. No statistically significant difference could be observed in terms of the underlying disease. When considering the distribution of the Child stage, significantly more patients with Child A cirrhosis can be found in the group of responders (
Baseline characteristics of the patient cohort.
Variable | Response | No response |
|
---|---|---|---|
Number of patients | 65 (54%) | 54 (45%) | |
Age, median (IQR) | 56 |
59 |
0.325 |
Age range | 25–78 | 25–74 | |
Gender (m/f) | 44/21 | 37/17 | 0.580 |
Etiology, |
|||
Alcohol abuse | 46 | 31 | 0.130 |
Hepatitis B | 0 | 3 | 0.055 |
Hepatitis C | 3 | 6 | 0.184 |
AIH | 1 | 0 | 0.362 |
Hemochromatosis | 3 | 1 | 0.407 |
PBC/PSC | 2 | 5 | 0.155 |
Cryptogenic | 11 | 9 | 0.970 |
Child stage, |
0.693 | ||
A | 20 | 8 |
|
B | 20 | 30 |
|
C | 25 | 16 | 0.404 |
MELD score, median (IQR) | 26 |
26 |
0.505 |
#: more than one etiology possible.
In univariate analysis, only initial serum protein was statistically different between the two groups (responders:
Univariate analysis of the patient cohort.
Variable | Response | No response |
|
---|---|---|---|
HRS type 1/2, |
47/18 | 45/9 | 0.165 |
Ascites, grade 1/2/3 | 9/38/18 | 11/25/16 | 0.466 |
Treatment, median days (IRQ) | 9 |
10.5 |
0.350 |
HE, grade 1/2/3 | 46/15/3 | 39/14/1 | 0.699 |
Serum protein, g/L | 5.84 ± 1.27 | 5.30 ± 1.47 |
|
INR | 1.60 ± 0.52 | 1.64 ± 0.44 | 0.698 |
Serum sodium (mmol/L) | 133.60 ± 6.21 | 133.12 ± 5.73 | 0.714 |
Serum potassium (mmol/L) | 4.34 ± 0.80 | 4.22 ± 0.71 | 0.381 |
Serum creatinine (mg/dL) | 2.79 ± 1.24 | 3.07 ± 1.47 | 0.207 |
Urea (mg/dL) | 55.30 ± 24.68 | 67.47 ± 51.59 | 0.189 |
Bilirubin (mg/dL) | 8.12 ± 10.23 | 10.59 ± 12.29 | 0.901 |
Terlipressin dose (mg) | 26.43 ± 30.86 | 32.11 ± 31.57 | 0.450 |
Albumin dose (g) | 266.26 ± 236.31 | 298.14 ± 252.02 | 0.612 |
HE: hepatic encephalopathy; INR: international normalized ratio.
Kaplan-Meier calculation indicates that response to terlipressin therapy in HRS patients is a significant predictor of survival. The mean short-time survival (30 days) for the responder group is significantly longer compared to the group of nonresponders (28.4 days [95% CI 27.3–29.4] versus 25.6 days [95% CI 23.3–27.8],
Kaplan-Meier survival analysis of the patient cohort: mean short-term survival was significantly longer in the responder group compared with nonresponders to terlipressin therapy: 28.4 days [95% CI 27.3–29.4] versus 25.6 days [95% CI 23.3–27.8] (responder versus nonresponder group, log-rank test, and
Kaplan-Meier survival analysis of the patient cohort: median overall survival was significantly longer in the responder group compared with nonresponders to terlipressin therapy: 29 months [95% CI 20.7–38.4] versus 8 months [95% CI 0.0–16.3] (treatment responders versus nonresponder group, log-rank test,
Patients with alcohol abuse had a significant lower short-time survival compared to those without alcohol problems (mean survival, 26.7 days [95% CI 25.2–28.1] versus 27.9 days [95% CI 26.1–29.8],
Kaplan-Meier survival analysis of the patient cohort: mean short-term survival was significantly longer in the group without alcohol abuse compared with abusers: 27.9 days [95% CI 26.1–29.8] versus 26.7 days [95% CI 25.2–28.1] (alcohol abuse versus nonalcohol abuse, log-rank test,
Kaplan-Meier survival analysis of the patient cohort: mean short-term survival was significantly longer in group with treatment duration >18 days compared with those less than 18 days treated: 29.2 days [95% CI 28.4–30.0] versus 23.4 days [95% CI 20.9–25.9] (treatment days ≥ 18 versus treatment days < 18 days, log-rank test,
Kaplan-Meier survival analysis of the patient cohort: mean short-term survival was significantly longer in patients with MELD score less than 27 compared with those having a MELD score >27: 28.5 days [95% CI 27.3–29.7] versus 25.4 days [95% CI 23.4–27.5] (MELD score < 27 versus MELD score ≥ 27, log-rank test,
In the Cox proportional hazard model, only age (HR 1.05 [95% CI 1.005–1.093]), alcohol abuse (HR 3.05, [95% CI 1.111–8.384]), duration of treatment (HR 0.92 [95% CI 0.875–0.964]) and MELD score (HR 1.08 [95% CI 1.019–1.141]) proved to be independent prognostic survival factors (Table
Cox regression analysis and predictors of one-month mortality.
Variable | HR | 95% CI |
|
---|---|---|---|
Age | 1.05 | 1.005–1.093 |
|
alcohol abuse | 3.05 | 1.111–8.384 |
|
duration of therapy | 0.92 | 0.875–0.964 |
|
response to therapy | 0.48 | 0.199–1.146 | 0.098 |
MELD score | 1.08 | 1.019–1.141 |
|
Cirrhotic patients with portal hypertension are at high risk to develop a multitude of renal dysfunction patterns, including paracentesis-induced circulatory dysfunction (PICD) and fully established hepatorenal syndrome (HRS). Leithead et al. have extensively reviewed the recent progress in the pathophysiology and treatment of portal hypertension-related renal dysfunction, which occurs as a multifactorial pathophysiological sequence on the background of profound circulatory and neurohumoral alterations in cirrhotic patients [
Therefore, the prevention of any renal further impairment in cirrhotic patients with portal hypertension-related renal dysfunction is of utmost importance for prognosis and survival. These include avoidance of nephrotoxic drugs (e.g., NSAID and ACE-Inhibitors), prevention and treatment of infection (including spontaneous bacterial peritonitis), prevention and treatment of gastrointestinal bleeding, avoidance of large volume paracentesis without albumin replenishment [
Therapeutic agents tested for the prevention of hepatorenal syndrome include vasoactive compounds, human albumin infusion, antibacterial substances such as rifaximin [
Despite a small randomized trial which indicates that prophylaxis with oral pentoxifylline, an oral phosphodiesterase inhibitor, was able to prevent HRS in some patients with alcoholic steatohepatitis [
Based on recommendations endorsed by the leading societies of gastroenterology and hepatology [
The criteria of when to initiate terlipressin therapy and how to judge sufficient treatment response are poorly defined. In many cirrhotic patients presenting with acute kidney failure, preclinical creatinine levels are rarely available, which leads to speculation about how quickly deterioration of renal function occurred. Furthermore, it has been criticized that the diagnosis of HRS is based on a rigid cutoff value of serum creatinine (1.5 mg/dL), because creatinine synthesis in patients is known to vary widely with regard to cachexia/muscle mass, ethnicity, gender, and age. Furthermore, it was shown that, in cirrhotic patients, serum creatinine levels are falsely low due to reduced creatinine production in liver and wasting muscles, as well as increased renal tubular secretion despite the fact that the actual glomerular filtration rate is low [
Overall treatment response in our patient series, as judged by dropping creatinine values to or below 1.5 mg/dL was approximately 55% overall, which is in well accordance with previous studies which have indicated terlipressin treatment response in 40–60% of HRS patients [
While short-term survival in acute HRS (i.e., type I) patients widely depends on acute clinical measures such as terlipressin/albumin treatment, calculated hydration, and differential diuretic therapy, long-term survival in patients with HRS is depending on the restitution of liver function. Therefore, liver transplantation is considered to be the first line treatment for both types of HRS [
In our patient series, short-term survival (30 days) was significantly higher in terlipressin responders versus nonresponders (28.4 versus 25.6 days,
Interestingly, patients with Child stage B cirrhosis had a lower likelihood to respond to therapy, leading to the assumption that other renoparenchymal disorders not responsive to terlipressin treatment (e.g., diabetic nephropathy) might have mimicked hepatorenal syndrome in some patients with concomitant liver cirrhosis (Table
Finally, Cox regression analysis identified the factors age, alcohol abuse, shorter duration of therapy, and MELD score as independent variables worsening the probability of survival in patients with HRS (Table
Finally, corroborated by the results from earlier studies, a higher MELD score was an independent predictor of lower one-month survival (HR 1.08, 95% CI 1.019–1.141,
In our patient analysis, many patients with hepatorenal syndrome were treated very early with combined albumin and terlipressin. Treatment response was approximately 55%, as expected from older studies. Our data suggest that older age, higher Child stage, alcohol abuse, and higher initial MELD score are clinical parameters associated with less favorable patient outcomes, whereas the differentiation between HRS types I and II did not influence treatment response rates in our patient series. Moreover, patients showing no sufficient treatment response show significantly higher one-month mortality than patients with terlipressin response. Finally, patients responding to terlipressin therapy have a significantly higher estimated median survival as compared to nonresponders. Age, duration of treatment, MELD score, and alcohol abuse are independent predictors of short-term survival. Given the notion that hepatorenal syndrome is today widely regarded as a potentially preventable condition, prophylaxis of any causes that precipitate of renal dysfunction should be the focus of patient care for cirrhotic patients.
We acknowledge that our patient cohort was rather small and possible patient selection bias from a single center and misclassification or information bias as a result of the retrospective study design might impact the veracity of the findings of our study. Nevertheless, our study adds valuable new findings and endorses established knowledge in the identification of predictors of successful response to HRS therapy and survival in cirrhotic patients for better future patient care. Clearly, further studies with prospective, multicenter design will be needed in order to more closely define the predictors of treatment response and survival in patients with HRS.
All the authors declare that there is no conflict of interests.