Predictors of response to therapy with terlipressin and albumin in patients with cirrhosis and type 1 hepatorenal syndrome

Authors

  • André Nazar,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Gustavo Henrique Pereira,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Mónica Guevara,

    Corresponding author
    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    • Liver Unit, Hospital Clínic, Villarroel 170, 08036 Barcelona, Spain
    Search for more papers by this author
    • fax: +34 93 451 5522.

  • Marta Martín-Llahi,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Marie-Noëlle Pepin,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    Search for more papers by this author
  • Marcella Marinelli,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Elsa Solá,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • María Eugenia Baccaro,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Carlos Terra,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Vicente Arroyo,

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author
  • Pere Ginès

    1. Liver Unit, Hospital Clínic, University of Barcelona, Barcelona, Catalunya, Spain
    2. Institut d'Investigacions Biomèdiques August-Pi-Sunyer (IOIBAPS)
    3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD)
    Search for more papers by this author

  • Potential conflict of interest: Dr. Ginès received research funding from Otsuka. Dr. Ginès is a consultant for Sanofi-Aventis and Otsuka.

Abstract

Terlipressin plus albumin is an effective treatment for type 1 hepatorenal syndrome (HRS), but approximately only half of the patients respond to this therapy. The aim of this study was to assess predictive factors of response to treatment with terlipressin and albumin in patients with type 1 HRS. Thirty-nine patients with cirrhosis and type 1 HRS were treated prospectively with terlipressin and albumin. Demographic, clinical, and laboratory variables obtained before the initiation of treatment as well as changes in arterial pressure during treatment were analyzed for their predictive value. Response to therapy (reduction in serum creatinine <1.5 mg/dL at the end of treatment) was observed in 18 patients (46%) and was associated with an improvement in circulatory function. Independent predictive factors of response to therapy were baseline serum bilirubin and an increase in mean arterial pressure of ≥5 mm Hg at day 3 of treatment. The cutoff level of serum bilirubin that best predicted response to treatment was 10 mg/dL (area under the receiver operating characteristic curve, 0.77; P < 0.0001; sensitivity, 89%; specificity, 61%). Response rates in patients with serum bilirubin <10 mg/dL or ≥10 mg/dL were 67% and 13%, respectively (P = 0.001). Corresponding values in patients with an increase in mean arterial pressure ≥5 mm Hg or <5 mm Hg at day 3 were 73% and 36%, respectively (P = 0.037). Conclusion: Serum bilirubin and an early increase in arterial pressure predict response to treatment with terlipressin and albumin in type 1 HRS. Alternative treatment strategies to terlipressin and albumin should be investigated for patients with type 1 HRS and low likelihood of response to vasoconstrictor therapy. (HEPATOLOGY 2009.)

Hepatorenal syndrome (HRS) is a severe complication of patients with advanced cirrhosis characterized by marked renal failure due to vasoconstriction of the renal circulation in the absence of significant morphological abnormalities in the kidneys.1–5 In the overall population of patients with cirrhosis, HRS is a strong predictor of mortality. Moreover, in patients listed for transplantation, the development of HRS either is a frequent cause of not receiving a transplant or, in patients who receive a transplant, is associated with increased morbidity and mortality after transplantation.6–10 HRS is pathogenically related to a marked arterial vasodilation of the splanchnic circulation, due to portal hypertension, that causes impairment of the effective arterial blood volume with activation of systemic vasoconstrictor factors that lead to marked reduction of renal blood flow and glomerular filtration rate.1–5 In recent years, several studies have shown that the administration of terlipressin, a vasopressin analogue that is not available in all countries (including the United States), together with intravenous albumin improves renal function in patients with HRS. Other vasoconstrictor drugs, including midodrine and norepinephrine, have also been investigated, yet the information available is limited.1–4, 11, 12 The results of the studies show that not all patients with HRS respond to terlipressin and albumin. A recent meta-analysis indicates that 52% of patients with HRS respond to treatment with terlipressin.13 In the remaining patients, terlipressin therapy is not associated with improvement of renal function. So far, no specific studies have been published investigating predictive factors of response to therapy in patients with HRS treated with terlipressin and albumin. The identification of patients with low likelihood of response to treatment is of important clinical interest, particularly in patients awaiting transplantation, and also in the design of new therapies for HRS. In the current study, we assessed predictive factors of response to terlipressin and albumin in patients with cirrhosis and type 1 HRS treated with the same therapeutic protocol in a single institution.

Abbreviations

CVP, central venous pressure; HRS, hepatorenal syndrome; MAP, mean arterial pressure; MELD, Model for End-Stage Liver Disease.

Patients and Methods

Study Population.

Since 1998, all patients with cirrhosis and renal failure admitted to the Liver Unit of the Hospital Clínic of Barcelona (Catalunya, Spain) were evaluated using the same diagnostic algorithm,14 which includes diuretic withdrawal, assessment of possible causes of hypovolemia, a trial of plasma expansion with intravenous albumin to rule out the existence of renal failure due to volume depletion, and evaluation of possible drug nephrotoxicity, infection, or renal parenchymal diseases. Patients meeting the criteria of type 1 HRS, as proposed by the International Ascites Club,15 were treated with terlipressin and albumin. The new criteria for definition of HRS were not used because patients included in this study were treated before the new criteria were published.3 Reasons for not receiving treatment include terminal condition (death expected in less than 48 hours), presence of severe cardiovascular diseases, advanced hepatocellular carcinoma, and active infection. Presence of infection was evaluated by means of blood, ascitic fluid, and urine cultures, measurement of leukocyte concentration in the ascitic fluid, and chest radiography. Patients with infection were excluded from the current study and were treated with terlipressin and albumin only if renal failure persisted after the resolution of the infection. The current study includes 39 consecutive patients with cirrhosis and type 1 HRS treated between January 1998 and November 2007. In 22 of the 39 patients (56%), HRS was triggered by a bacterial infection, including seven cases of spontaneous bacterial peritonitis. Some patients have been included in previous prospective studies of management of HRS.16–18 A small number of patients (n = 7) who were treated with terlipressin without albumin during the same period are not included in the current study.17

Treatment Protocol.

Terlipressin (Glypressin, Ferring S.A., Saint-Prex, Switzerland) was administered initially at a dose of 0.5 to 1 mg/4 hours as an intravenous bolus for 3 days. If after the first 3 days serum creatinine had decreased at least 25% of the pretreatment values, the dose was not modified. In patients in whom serum creatinine had not decreased at least 25% of the pretreatment values within the first 3 days, the dose was increased up to a maximum of 2 mg/4 hours. Terlipressin was given until serum creatinine had decreased below 133 μmol/L (1.5 mg/dL) or for a maximum of 15 days. Terlipressin administration was withheld if patients developed signs or symptoms compatible with ischemic complications. All patients received albumin (Albúmina 20%, Grífols International, Barcelona, Spain) at a dose of 1 g per kilogram of body weight during the first 24 hours, followed by 40 g per day, targeted to obtain a central venous pressure (CVP) between 10 and 15 cm of water. CVP was measured at least once a day throughout the treatment period. When CVP increased over 15 cm of water, the albumin dose was reduced to 20 g/day and was withheld when CVP increased above 18 cm of water or there were clinical or radiological signs of pulmonary edema. In addition of stopping albumin administration, these latter patients received intravenous boluses of furosemide.

Physical examination, chest radiography, and routine laboratory tests were performed in all patients before the initiation of therapy and at regular intervals during treatment. Arterial pressure was measured several times per day in all patients. In addition, in 19 of the 39 patients, blood samples were obtained before the initiation of therapy to measure plasma renin activity, and the plasma concentrations of aldosterone, norepinephrine, and atrial natriuretic factor.

Complications of cirrhosis developing during treatment of HRS were treated according to standardized therapeutic measures.19 Only patients with a past history of spontaneous bacterial peritonitis were treated with prophylactic antibiotics (norfloxacin 400 mg/day).

Statistical Analysis.

The primary endpoint was response to treatment, defined either as a reduction in serum creatinine below 133 μmol/L (1.5 mg/dL) during treatment or a reduction in serum creatinine of greater than 50% of the pretreatment value but with an end-of-treatment value equal to or greater than 133 μmol/L (1.5 mg/dL). The probability of response was calculated using the Kaplan–Meier method. Patients treated with liver transplantation (n = 4) were included in the calculation of overall response and were considered censored at the time of transplantation. Comparisons of variables between patients were made using the Student t test for continuous data and the χ2 test for categorical data. Comparisons of variables obtained at different time points were performed using a paired Student t test and Wilcoxon test. A multivariate analysis including variables with predictive value in the univariate analysis (P < 0.10) was performed using stepwise logistic regression. The best cutoff values for parameters with independent predictive value were calculated using receiver operating characteristic curves (AUC). Statistical analysis was performed using SPSS version 14 for Windows (SPSS Inc., Chicago, IL). Results are expressed as the mean ± standard deviation. P < 0.05 was considered statistically significant.

Results

Patient Characteristics.

The baseline characteristics of patients with cirrhosis and type 1 HRS before the initiation of therapy with terlipressin and albumin are shown in Table 1. As expected, most patients had severe liver failure, as indicated by high serum bilirubin and prothrombin time and high Child-Pugh and Model for End-Stage Liver Disease (MELD) scores, and severe renal failure with high serum creatinine levels and low glomerular filtration rate, estimated using Modification of Diet in Renal Disease equation.20 Finally, patients showed marked circulatory dysfunction with low mean arterial pressure (MAP) and marked activation of vasoactive systems.

Table 1. Baseline Characteristics of the 39 Patients with Type 1 HRS Included in the Study
Variables 
  • *

    Estimated with the Modification of Diet in Renal Disease formula.20

  • Urine volume was lower than 500 mL/day in 14 patients (36%).

  • Obtained in 19 patients. Normal values in healthy subjects are: plasma renin activity, 1.2 ± 0.1 ng/mL/h; plasma aldosterone, 24 ± 2 ng/dL; plasma norepinephrine, 233 ± 17 pg/mL; plasma atrial natriuretic factor, 6 ± 0.5 fmol/mL.

Age (years)56 ± 9
Sex (M/F)29/10
Etiology of cirrhosis: alcohol/virus/alcohol plus virus/others (n)16/14/5/4
Serum bilirubin (mg/dL)16 ± 17
Serum albumin (g/L)29 ± 6
Prothrombin time (%)42 ± 16
Aspartate aminotransferase (U/L)143 ± 239
Alanine aminotransferase (U/L)91 ± 160
Child-Pugh score11 ± 1.4
MELD score30 ± 7
Serum creatinine (mg/dL)3.7 ± 1.4
Glomerular filtration rate* (mL/min/1.73m2)19 ± 7
Serum sodium (mEq/L)125 ± 6
Serum potassium (mEq/L)4.5 ± 1.0
Urine volume (mL/day)639 ± 359
Leukocyte count (/mm3)8,944 ± 4,418
Hematocrit (%)28 ± 5
MAP (mm Hg)77 ± 11
Heart rate (bpm)78 ± 10
Plasma renin activity (ng/mL/h)12 ± 23
Plasma aldosterone (ng/dL)256 ± 207
Plasma norepinephrine (pg/mL)1,116 ± 541
Plasma atrial natriuretic factor (fmol/mL)118 ± 101

Response to Treatment and Predictive Factors.

Eighteen out of 39 patients (46%) had response to treatment. In 16 of the 18 patients, serum creatinine decreased below 1.5 mg/dL at the end of therapy, whereas in the other two patients serum creatinine decreased by more than 50% compared with pretreatment values but did not end up below 1.5 mg/dL (from 4.7 and 3.5 mg/dL to 1.7 and 1.6 mg/dL, respectively). The remaining 21 patients did not meet the criteria of response to treatment. Values of serum creatinine throughout treatment in responders and nonresponders are shown in Fig. 1. The probability of response during treatment in the entire series of patients is shown in Fig. 2. Median time to response was 14 days. In most patients, response to treatment was persistent. HRS recurred in five of the 18 patients who responded to treatment (mean time to recurrence, 14 days [range, 2–46 days]).

Figure 1.

Serum creatinine concentration (mean and standard deviation) during treatment with terlipressin and albumin in responders and nonresponders.

Figure 2.

Probability of response to therapy in patients with type 1 HRS treated with terlipressin and albumin. Dotted lines represent 95% confidence intervals. The numbers under the curve are patients at risk at each time point.

Response to treatment was associated with an improvement in circulatory function, as indicated by an increase in arterial pressure at the end of treatment period and marked suppression in the activity of the renin-angiotensin-aldosterone system and sympathetic nervous system (Table 2). There was also an increase in plasma atrial natriuretic peptide levels, but it did not reach statistical significance. By contrast, no increase in mean arterial pressure (MAP) was observed in patients not responding to treatment (79 ± 9 versus 73 ± 14 mm Hg; P value not significant).

Table 2. Effects of Treatment on Arterial Pressure and Activity of Vasoactive Systems in Responders to Terlipressin and Albumin (n = 18)
VariablesBaselineEnd of TreatmentP
  • *

    Values in 10 patients. Normal values in healthy subjects are: plasma renin activity, 1.2 ± 0.1 ng/mL/h; plasma aldosterone, 24 ± 2 ng/dL; plasma norepinephrine, 233 ± 17 pg/mL; plasma atrial natriuretic factor, 6 ± 0.5 fmol/mL.

MAP (mm Hg)75 ± 1384 ± 130.017
Plasma renin activity (ng/mL/h)*16 ± 323.4 ± 5.50.007
Plasma aldosterone (ng/dL)*272 ± 21748 ± 540.005
Plasma norepinephrine (pg/mL)*1,108 ± 543409 ± 6460.005
Plasma atrial natriuretic factor (fmol/mL)*114 ± 87185 ± 900.051

Several variables obtained at baseline were analyzed for predictive value of response to treatment (Table 3). Variables associated with response to treatment (P < 0.10) were serum aspartate aminotrasferase, serum alanine aminotranferase, MELD score, urine volume, leukocyte count, and serum bilirubin. Of note, neither serum creatinine nor arterial pressure levels at baseline were associated with response to therapy. In the multivariate analysis, only serum bilirubin (odds ratio, 0.913; 95% confidence interval, 0.853-0.978; P = 0.01) was associated with an independent predictive value of response to treatment. The cutoff level of serum bilirubin that best predicted response to treatment, as assessed by receiver operating characteristic curves, was 10 mg/dL (area under the curve, 0.77; P < 0.0001; sensitivity, 89%; specificity, 61%). Response rates in patients divided according to baseline serum bilirubin ≥10 mg/dL or <10 mg/dL were 13% (2/15) and 67% (16/24), respectively (P = 0.001). There was a trend for an association between baseline leukocyte levels and response to treatment, but the difference did not reach statistical significance in the multivariate analysis (odds ratio, 0.825; 95% confidence interval, 0.674-1.009; P = 0.061).

Table 3. Univariate Analysis of Baseline Variables According to Response to Treatment with Terlipressin and Albumin
VariablesResponders (n = 18)Nonresponders (n = 21)P
  • *

    Estimated with the Modification of Diet in Renal Disease formula.20

  • Values in 10 responders and nine nonresponders.

  • Values in 10 responders and eight nonresponders.

  • §

    Values in 10 responders and six nonresponders.

Age (years)58 ± 955 ± 100.430
Sex (M/F)12/617/40.308
Alcoholic cirrhosis (%)8 (44)13 (62)0.276
Serum bilirubin (mg/dL)6 ± 724 ± 200.000
Serum albumin (g/L)30 ± 628 ± 70.367
Prothrombin time (%)44 ± 1841 ± 150.621
Aspartate aminotransferase (U/L)63 ± 36212 ± 3100.040
Alanine aminotransferase (U/L)40 ± 27134 ± 2090.056
Child-Pugh score11 ± 211 ± 10.808
MELD score28 ± 633 ± 70.029
Serum creatinine (mg/dL)3.5 ± 1.43.9 ± 1.40.378
Glomerular filtration rate*20 ± 719 ± 80.630
Serum sodium (mEq/L)125 ± 7125 ± 60.736
Serum potassium (mEq/L)4.4 ± 1.14.5 ± 0.80.674
Urine volume (mL/day)752 ± 341541 ± 3530.066
Leukocyte count (mm3)7,237 ± 3,04410,407 ± 4,9340.020
Hematocrit (%)27 ± 429 ± 60.361
Mean arterial pressure (mm Hg)76 ± 1379 ± 90.335
Heart rate (bpm)80 ± 1076 ± 110.247
Plasma renin activity (ng/mL/h)16 ± 328 ± 50.461
Plasma aldosterone (ng/dL)272 ± 217237 ± 2070.732
Plasma norepinephrine (pg/mL)1,108 ± 5431,124 ± 5770.952
Plasma atrial natriuretic factor (fmol/mL)§114 ± 86126 ± 1290.826

To investigate whether changes in arterial pressure during the early period of treatment with terlipressin and albumin could be useful as a predictive factor of response, we analyzed response rates in patients divided according to changes in arterial pressure measured at day 3 of treatment. The value of MAP used was the average value of all measurements of arterial pressure obtained at day 3; an increase in MAP of 5 mm Hg was considered relevant and used as a cutoff value. Patients with an increase in MAP equal to or greater than 5 mm Hg at day 3 of treatment had a response rate at the end of therapy of 73% (8/11) compared with 36% (10/28) in patients with an increase that did not reach 5 mm Hg or a decrease in MAP (P = 0.037). When the increase in arterial pressure of 5 mm Hg at day 3 was included in the multivariate analysis together with the same baseline variables mentioned above, the independent predictive factors of response to therapy were baseline serum bilirubin levels and an increase in MAP ≥5 mm Hg at day 3 (Table 4). Response rates in the subgroups of patients divided according to the selected cutoff values for these two parameters are shown in Table 5.

Table 4. Variables with Independent Predictive Value of Response to Treatment with Terlipressin and Albumin in Patients with Type 1 HRS
VariablesOdds Ratio95% Confidence IntervalP
Baseline serum bilirubin0.9010.834–0.9740.009
Δ MAP at day 3 ≥5 mm Hg9.4821.007–89.3160.049
Table 5. Response Rate According to Variables with Independent Predictive Value of Response to Treatment with Terlipressin and Albumin in Patients with Type 1 HRS
VariableSerum bilirubin <10 mg/dLSerum bilirubin ≥10 mg/dL
  1. Values are expressed as the number of patients per total number of patients, with percentages in parentheses.

Δ MAP at day 3 ≥5 mm Hg7/7 (100)1/4 (25)
Δ MAP at day 3 <5 mm Hg9/17 (53)1/11 (9)
Total16/24 (67)2/15 (13)

Finally, to assess whether an early reduction in serum creatinine during treatment was predictor of response to therapy, we analyzed the relationship between changes in serum creatinine at day 3 compared with baseline with response at the end of treatment. Response to treatment was observed in 13 of the 17 patients (76%) in whom serum creatinine decreased at least 0.5 mg/dL at day 3, compared with only five of the 22 patients (23%) in whom serum creatinine did not decrease 0.5 mg/dL or increased at day 3 compared with baseline (P = 0.001). Similar figures were observed when the cutoff value of change in serum creatinine used was 1 mg/dL instead of 0.5 mg/dL (80% and 36%, respectively; P = 0.016). The value of the reduction in serum creatinine at day 3 as predictor of response to therapy was confirmed in a multivariate analysis (odds ratio, 6.5; P = 0.047).

Complications of Cirrhosis and Side Effects of Treatment.

Thirty-five patients (90%) developed a total of 43 major complications of cirrhosis during treatment: 31 cases of hepatic encephalopathy, nine cases of bacterial infection, and three cases of gastrointestinal bleeding. There was no significant difference between the frequency of bacterial infections between nonresponders and responders (33% equation image versus 11% equation image, respectively; P = 0.10). The development of bacterial infections during treatment with terlipressin and albumin was slightly more frequent (yet not significantly different) in patients with a baseline leukocyte count above the median value of 7,900/mm3 than in those with a leukocyte count below the median value (32% equation image versus 15% equation image, respectively; P = 0.27). Patients with a baseline leukocyte count above the median value developed pneumonia (n = 3), sepsis (n = 2), and spontaneous bacterial peritonitis (n = 1) after a mean of 5 days (range, 2–13 days), whereas patients with a baseline leukocyte count below the median value developed urinary tract infections (n = 2) and pneumonia (n = 1) after a mean of 5 days (range, 2–8 days). It should be noted that none of these patients had proven bacterial infection at the time of initiation of therapy with terlipressin and albumin.

Eleven patients (28%) developed side effects likely related to treatment with terlipressin and albumin. Five patients developed signs of circulatory overload, which improved after temporary suppression of albumin together with administration of furosemide and did not require discontinuation of terlipressin. Two patients developed abdominal signs suggestive of intestinal ischemia (one of which was associated with circulatory overload), which subsided after discontinuation of treatment. Two patients developed transient arrhythmia (bradycardia and ventricular extrasystolia, respectively) that did not require permanent treatment discontinuation. One patient developed myocardial infarction and signs suggestive of intestinal ischemia (associated with circulatory overload). One patient developed signs of tongue ischemia. In the two latter patients, symptoms disappeared after treatment withdrawal.

Survival.

Three months after the start of therapy, 28 (72%) patients had died, four (10%) had undergone transplantation, five (13%) were alive, and two (5%) were lost to follow-up. The probability of survival at 3 months was significantly greater in responders compared with nonresponders (44% and 14%, respectively; median survival, 70 days and 7 days, respectively; P = 0.001). The causes of death were multiorgan failure (n = 14), septic shock (n = 6), liver failure (n = 5), acute respiratory distress syndrome (n = 1), and unknown (n = 2).

Discussion

This study confirms that terlipressin and albumin is an effective therapy for the management of type 1 HRS in patients with cirrhosis.1–5, 11, 12, 21 Forty-six percent of patients included responded to treatment with a marked improvement of renal function. This efficacy rate is similar to that reported in a recent meta-analyses.13 Moreover, the results of the current study confirm the previous observations of studies including lower numbers of patients, indicating that response to treatment is associated with an improvement of circulatory function that is markedly impaired in patients with HRS.16, 17, 21–24 In fact, patients who responded to therapy showed a significant increase in arterial pressure and a suppression of the markedly increased activity of the renin-angiotensin-aldosterone system and sympathetic nervous system at the end of treatment; these findings are consistent with an improvement of the low effective arterial blood volume characteristic of HRS.1–5, 11, 12 By contrast, no increase in arterial pressure was observed in patients who did not show an improvement in renal function. These findings strongly suggest that the beneficial effect of terlipressin in the management of HRS is related to its capacity of improving systemic hemodynamics. Reasons for the lack of improvement of systemic hemodynamics in some patients with type 1 HRS treated with terlipressin are unknown but may include, among others, increased levels of vasodilator cytokines, increased bacterial products or latent infections, and presence of concomitant adrenal insufficiency. These possible causes deserve investigation in order to improve the efficacy of treatment.

The current study was intended to assess predictive factors of response to terlipressin and albumin in a consecutive series of patients with type 1 HRS treated with the same standardized protocol. Independent predictive factors of response to treatment were baseline serum bilirubin levels and an increase in MAP of 5 mm Hg at day 3 of treatment. Seven of the 7 patients (100%) with baseline serum bilirubin <10 mg/dL who showed an increase in MAP ≥5 mm Hg at day 3 responded to treatment with terlipressin and albumin. By contrast, only one of the 11 (9%) patients with baseline serum bilirubin ≥10 mg/dL and a change in MAP <5 mm Hg had response to treatment. Predictive factors of response reported in previous studies in patients with HRS included baseline Child-Pugh and MELD scores, serum creatinine, and arterial pressure.21, 23, 24–26 However, in these studies either the number of patients treated was low or the analyses included not only patients treated with terlipressin but also control patients not treated with vasoconstrictors or patients treated with vasoconstrictors other than terlipressin (such as norepinephrine).

The relationship between the presence of an early increase in MAP and the renal response to terlipressin stresses the importance of the improvement of systemic hemodynamics in achieving a reversal of type 1 HRS. These data are in keeping with those of a recent study reported in abstract form analyzing the effects of terlipressin versus placebo on arterial pressure and renal function in patients with cirrhosis and type 1 HRS.27 Nevertheless, it is important to emphasize that not all patients showing an early increase in arterial pressure ended up with a renal response. Conversely, approximately one-third of patients without the early hemodynamic response showed an improvement of renal function at the end of therapy. Therefore, our data indicate that treatment with terlipressin should not be stopped after day 3 if there is no improvement in arterial pressure.

In the current study, baseline serum bilirubin levels were also an independent predictive factor of response to therapy. The mechanisms by which high serum bilirubin levels are associated with a poor response to therapy is unknown and seems to be independent of the hemodynamic response to terlipressin. This relationship between high serum bilirubin levels and lack of response to terlipressin is intriguing and deserves investigation.

We also analyzed the relationship between an early reduction in serum creatinine during treatment with terlipressin and the response at the end of treatment. As it could be anticipated, patients with an early (at day 3) reduction in serum creatinine of at least 0.5 mg/dL compared with baseline had a higher probability of response at the end of treatment compared with patients who did not meet this criterion. Nevertheless, it is important to note that a significant proportion of patients (up to one-third) without an early reduction in serum creatinine show a response at the end of treatment. The cause of this may be either a renal response delayed with respect to the hemodynamic improvement or related to the fact that the dose of terlipressin was increased in our protocol in patients not having an early reduction in serum creatinine. In any case, terlipressin treatment should be maintained after 3 days even if there is no reduction in serum creatinine.

The results of the current study confirm data from previous reports indicating that patients with type 1 HRS who respond to treatment with terlipressin and albumin have longer survival compared with that of nonresponders.17, 18, 21, 23–25 In fact, in the current series, 3-month probability of survival in responders was 44% compared with only 14% in nonresponders. These data, together with the observation that patients with HRS in whom renal function improves with terlipressin and albumin have an excellent posttransplantation outcome similar to that of patients without HRS,28 suggest that terlipressin plus albumin is an effective therapeutic option for patients with HRS awaiting liver transplantation.

An important feature of the current study that deserves a comment is that according to our treatment protocol for type 1 HRS, patients with renal failure with associated bacterial infections were not treated with terlipressin and albumin until the infection resolved. Patients with renal failure and active bacterial infections (without septic shock) are currently considered as having HRS according to the new diagnostic criteria reported in 2007.3 However, these patients were not included in our treatment protocol because patients were treated before these criteria were published, and we used the previous diagnostic criteria of HRS, which deliberately excluded patients with ongoing bacterial infections.15 Therefore, the results of this study cannot be extrapolated to patients with HRS and associated bacterial infections. Moreover, to our knowledge, there are no reports published on the management of HRS in this patient population. Therefore, it would be important to perform studies in this subset of patients before treatment with terlipressin and albumin can be recommended for this particular clinical situation.

The current study has some limitations. First, the assessment of predictive factors of response to therapy should ideally be performed in a large patient population. Nonetheless, because type 1 HRS is not a common condition and terlipressin is not available in many countries, the recruitment of a large series of patients for such a study is difficult. In fact, this is one of the largest series of patients reported to date on the management of type 1 HRS. Second, the accuracy of the variables reported in the current study in predicting response to therapy would require prospective validation in other series of patients either from the same institution or, ideally, from other institutions. We are prospectively validating these predictive factors in patients with type 1 HRS treated with terlipressin and albumin in our institution, but it will take several years to accumulate a sufficient number of patients for analysis. As terlipressin becomes available in more countries for the treatment of type 1 HRS, the evaluation of predictors of response in external series of patients would be easier to perform.

In conclusion, the results of the current study indicate that baseline serum bilirubin and an increase in MAP at day 3 of treatment are predictive factors of response to therapy with terlipressin and albumin in patients with type 1 HRS. Future research on management of type 1 HRS should be focused on the analysis of mechanisms of impaired response to pharmacological therapy and on the implementation of new therapies for nonresponders.

Acknowledgements

We thank Raquel Cela, R.N., and the nursing staff of the Liver Unit and Intensive Care Unit for their participation in the study and Marco Pavesi for statistical advice.

Ancillary