Type 1 hepatorenal syndrome (HRS) is a functional

Transcription

1 GASTROENTEROLOGY 2002;122: Terlipressin in Patients With Cirrhosis and Type 1 Hepatorenal Syndrome: A Retrospective Multicenter Study RICHARD MOREAU,* FRANCOIS DURAND,* THIERRY POYNARD, CHRISTIAN DUHAMEL, JEAN PAUL CERVONI, PHILIPPE ICHAÏ, ARMAND ABERGEL, # CHANTAL HALIMI,** MATHIEU PAUWELS, JEAN PIERRE BRONOWICKI, EMILE GIOSTRA, CATHY FLEUROT, DANIELLE GURNOT, ## OLIVIER NOUEL,*** PHILIPPE RENARD, MICHEL RIVOAL, PIERRE BLANC, DIMITRI COUMAROS, SYLVIE DUCLOUX, ### STEPHANE LEVY,**** ALEXANDRE PARIENTE, JEAN MARC PERARNAU, JEAN ROCHE, MYRIAM SCRIBE OUTTAS, DOMINIQUE VALLA,* BRIGITTE BERNARD, DIDIER SAMUEL, JOËL BUTEL, ANTOINE HADENGUE, ANDRZEJ PLATEK, DIDIER LEBREC,* and JEAN FRANCOIS CADRANEL #### *INSERM U-481 et Service d Hépatologie, Hôpital Beaujon, Clichy, France; Service d Hépato-Gastroentérologie, Groupe Hospitalier Pitié- Salpétrière, Paris, France; Service de Gastroentérologie, Centre Hospitalier, Le Havre, France; Service d Hépato-Gastroentérologie, Hôpital Fontenoy, Chartres, France; Centre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, France; # Service d Hépato-Gastroentérologie, CHU Hôtel- Dieu, Clermont-Ferrand, France; **Service d Hépato-Gastroentérologie, Centre Hospitalier, Senlis, France; Service de Médecine B, Centre Hospitalier, Abbeville, France; Service d Hépato-Gastroentérologie, CHU Nancy-Brabois, Vandoeuvre-les-Nancy, France; Unité de Greffe, HUG-Hôpital Cantonal, Genève, Switzerland; Service d Anesthésie-Réanimation, CHU de Bordeaux, Centre Médico-Chirurgical de la Maison du Haut Lévêque, Groupe Hospitalier Sud, Pessac, France; ## Service de Gastroentérologie, Centre Hospitalier Jean Coulon, Gourdon, France; ***Service d Hépato-Gastroentérologie, Centre Hospitalier la Beauché, Saint-Brieuc, France; Service d Hépato-Gastroentérologie, Centre Hospitalier Victor Dupouy, Argenteuil, France; Service d Anesthésie-Réanimation-Urgences, Etablissement Public de Santé, Arpajon, France; Service d Hépato-Gastroentérologie, Hôpital Saint-Eloi, Montpellier, France; Service d Hépato-Gastroentérologie, CHU-Hôpital Civil, Strasbourg, France; ### Service d Alcoologie, CHU-Hôpital Jean Minjoz, Besançon, France; ****Service d Hépato-Gastroentérologie, Hôpital Robert Debré, CHU de Reims, Reims, France; Service d Hépato-Gastroentérologie, Centre Hospitalier, Pau, France; Service de Médecine B, CHR Notre-Dame de Bon-Secours, Metz, France; Service de Médecine Interne, Centre Hospitalier, Roanne, France; Service d Hépato-Gastroentérologie, Centre Hospitalier, Ussel, France; and #### Unité d Hépatologie, Centre Hospitalier Laënnec, Creil, France Background & Aims: Type 1 hepatorenal syndrome (HRS) is a severe complication of cirrhosis associated with a short median survival time (<2 weeks). Although the administration of terlipressin improves renal function, its effect on survival is unknown. This study investigated predictive factors of survival in patients with type 1 HRS treated with terlipressin. Methods: Ninety-nine patients with type 1 HRS treated with terlipressin in 24 centers were retrospectively studied. Terlipressininduced improved renal function was defined as a decrease in serum creatinine value to <130 mol/l or a decrease of at least 20% at the end of treatment. Results: At inclusion, the Child Pugh score was (mean SD). Terlipressin ( mg/day) was administered for days. Renal function improved in 58% of patients (serum creatinine decreased by 46% 17% from mol/l). Median survival time was 21 days. Survival rate was 40% at 1 month. Multivariate analysis showed that improved renal function and Child Pugh score <11 at inclusion were independent predictive factors of survival (P < and 0.02, respectively). Thirteen patients underwent liver transplantation (92 95 days after HRS onset), 10 of whom had received terlipressin and had had improved renal function. Conclusions: This retrospective uncontrolled study shows that in patients with type 1 HRS, terlipressin-induced improved renal function is associated with an increase in survival. Thus, a randomized trial investigating the effect of terlipressin on survival in patients with type 1 HRS should be performed. Type 1 hepatorenal syndrome (HRS) is a functional renal failure complicating end-stage cirrhosis. 1 3 The best treatment of HRS is liver transplantation. 4 However, patients with HRS who receive transplants have more complications and a higher in-hospital mortality rate than those without HRS. 4 Therefore, in patients with HRS, it might be better to treat renal function abnormalities before liver transplantation. Results of nonrandomized studies in small series of patients with HRS suggest that long-term administration of the vasopressin analog terlipressin may improve renal func- Abbreviations used in this paper: HRS, hepatorenal syndrome; SBP, spontaneous bacterial peritonitis by the American Gastroenterological Association /02/$35.00 doi: /gast

2 924 MOREAU ET AL. GASTROENTEROLOGY Vol. 122, No. 4 tion. 5 8 However, in patients with HRS, the duration of terlipressin administration and the factors involved in the renal response to this drug are unknown. In addition, although spontaneous mortality in patients with type 1 HRS is very high, 9 the mortality in patients with HRS receiving terlipressin has not yet been investigated. Therefore, the aim of this large, retrospective study was to investigate the clinical course, predictive factors of improved renal function, and survival in patients with type 1 HRS treated with terlipressin to plan a randomized, controlled trial. The adverse effects of terlipressin administration were also studied. Materials and Methods Patients This retrospective study included 99 patients with cirrhosis and type 1 HRS treated with terlipressin (Glypressine; Laboratoire Ferring SAS, Gentilly, France) admitted between April 1996 and April 2000 to 24 centers. All patients with HRS who had been treated with terlipressin in these centers were evaluated. To be included, patients were required to meet the following criteria: older than 18 years; diagnosis of HRS according to the International Ascites Club diagnosis criteria (i.e., low glomerular filtration rate, on the basis of serum creatinine 130 mol/l; absence of shock, uncontrolled bacterial infection, fluid loss, and treatment with nephrotoxic drugs; no improvement in renal function after diuretic withdrawal and plasma volume expansion; proteinuria 500 mg/l; and no ultrasonographic evidence of parenchymal renal disease or urinary tract infection) 2 ; an acute increase in serum creatinine of more than 50% above the pre-hrs value; and treated with terlipressin for at least 24 hours after diagnosis of HRS. Study Design The day of inclusion in the study (day 0) was the day of diagnosis of HRS. The following data were recorded. Within the month before day 0. In this month, the following data were recorded: the value of pre-hrs serum creatinine; previous diabetes mellitus; development of complications (bacterial infection, gastrointestinal bleeding, or acute alcoholic hepatitis); and previous treatment (therapeutic paracentesis, plasma expansion to prevent paracentesis-induced circulatory dysfunction, prevention of bacterial infection with fluoroquinolones, and prevention of portal hypertensive bleeding with -blockers). Severe bacterial infection was defined by the presence of at least 1 of the following: spontaneous bacterial peritonitis (SBP), septicemia, and pneumonia. At day 0. On this day, the following data were recorded: demographic data; etiology and duration of cirrhosis; and clinical and biological values (mean arterial pressure, heart rate, presence of encephalopathy, serum bilirubin, serum albumin, prothrombin, serum creatinine, serum sodium, and serum potassium). After day 0. After day 0, the following data were recorded: serum creatinine on the first and last day of terlipressin administration; duration of terlipressin treatment (and the dose used); concomitant treatment during terlipressin administration (therapeutic paracentesis, intravenous albumin [number of patients and daily dose], dopamine, and renal replacement therapy); date of liver transplantation; and outcome (survival or death) on the date of the last follow-up visit. Adverse effects to terlipressin administration were also noted. Terlipressin is extensively used to treat variceal bleeding in patients with cirrhosis, and the adverse effects caused by this drug are well known Terlipressin was routinely used to treat variceal bleeding in all centers of this study; thus, the participating hepatologists were aware of its adverse effects. The centers were classified into 2 categories (i.e., large and small centers) according to the number of patients enrolled ( 7 and 7, respectively). End Points The primary end points of the study were to investigate improvement of renal function during terlipressin administration and survival. Improvement of renal function was assessed between the first and last day of terlipressin administration and was defined as a decrease in serum creatinine, either to 130 mol/l or of at least 20% compared with the pretreatment value measured 6,8 at day 0. The secondary end point was to describe the adverse effects of terlipressin. Statistical Analysis Values were analyzed with SAS (version 8) for Windows (SAS Institute, Cary, NC). Unless otherwise specified, values are expressed as mean SD. Changes in serum creatinine during terlipressin therapy were compared between patients with improved renal function and those without, by use of 2-factor repeated-measures analysis of variance. Univariate analysis to identify factors predictive of improved renal function used the Student t test for quantitative variables and the Fisher exact test for qualitative variables. For certain analyses, patients were divided into 3 groups according to their Child Pugh score ( 11, 12 13, and 13). Variables reaching statistical significance (see below) in univariate analysis were included in multivariate logistic regression analysis to identify independent predictive factors of improved renal function. The probability of survival was estimated by the Kaplan Meier method. Patients who received transplants were censored on the date of transplantation. Univariate analysis to determine predictive factors of survival used the log-rank test. Variables reaching statistical significance (see below) in univariate analysis were included in a proportional hazards regression to identify independent predictive factors of survival. A 2-tailed P value of 0.05 was considered to be significant. Results Fifty-five patients were enrolled in 5 large centers, and 44 patients were enrolled in 19 small centers. In 10

3 April 2002 TERLIPRESSIN AND HEPATORENAL SYNDROME 925 centers (including the 5 large centers), information was available on all the causes of renal failure in patients with cirrhosis admitted during the study period. In these 10 centers, 71 of 423 (17%) patients with cirrhosis and renal failure met the inclusion criteria. Three hundred fifty-two of 423 patients were excluded because of acute renal failure sensitive to plasma volume expansion (135 patients); acute renal failure that was insensitive to plasma volume expansion and caused by septic or hemorrhagic shock (139 patients) or nephrotoxins (9 patients); chronic renal failure caused by glomerulopathies (26 patients); or renal failure caused by obstructive uropathy (1 patient). Patients with type 1 HRS complicating end-stage hepatocarcinoma, for whom no treatment was available (14 patients), and with type 2 HRS (28 patients) were also excluded. Characteristics of Patients Cirrhosis was caused by alcohol intake in 90% of patients (Table 1). Diabetes mellitus was a concomitant disorder in 21 patients. Certain patients had had at least 1 complication within the month preceding the onset of HRS: bacterial infections (41 episodes [13 SBP, 13 lower-urinary tract infections, 11 septicemia, 3 pneumonia, and 1 empyema] in 36 patients); acute alcoholic hepatitis (26 patients); and portal hypertensive bleeding (22 patients). None of the patients with bacterial infections or bleeding had had circulatory shock. Infections or bleeding were controlled at day 0. Within the month before the onset of HRS, 54 patients were treated by paracentesis. In 41 patients, paracentesis was followed by the administration of a plasma expander; intravenous albumin was the plasma expander used in 25 patients. During the pre-hrs period, 47 patients were receiving Table 1. Characteristics of the 99 Patients at the Time of Diagnosis of Type 1 HRS Variable Data Age (yr) Male sex 72 Duration of cirrhosis (yr) Alcoholic cirrhosis 89 Ascites 99 Serum bilirubin ( mol/l) Serum albumin (g/l) Prothrombin (%) Child Pugh score Child Pugh grade (B/C) 7/69 Serum creatinine ( mol/l) Mean arterial pressure (mm Hg) Heart rate (beats/min) Serum sodium (mmol/l) Serum potassium (mmol/l) NOTE. Data are presented as mean SD or n. a diuretic treatment furosemide in 24 patients, spironolactone in 5, and a combination of both in the remaining 18. During the same period, 38 patients were receiving a fluoroquinolone, and 21 were receiving a nonselective -blocker. The existence of a hepatocellular carcinoma was known in 7 patients at day 0; all these patients had solitary tumors. At day 0, 3 other patients had unknown advanced hepatocellular carcinoma. At day 0, diuretics were stopped (on the basis of the diagnostic criteria of HRS; see previously), and renal failure was indicated by a significant increase in serum creatinine (by 262% 147% from a pre-hrs serum creatinine value of mol/l; P ). At day 0, the Child Pugh score was obtained in 76 patients; certain results were not available in the remaining patients (serum bilirubin was lacking in 19 patients, serum albumin in 9, and prothrombin in 11). Evolution of Renal Failure During Terlipressin Administration The dose of terlipressin used was mg/day, and the duration of administration was days. Because values for serum creatinine on the first and last day of terlipressin treatment were available in all patients but 8, the analysis of renal response was performed in 91 patients. Fifty-eight of the 91 patients (64%) had improved renal function during terlipressin therapy, whereas the remaining 38 (39%) patients did not. Values for serum creatinine on the first day of terlipressin administration did not significantly differ between the 2 groups ( mol/l and mol/l; P 0.64). Patients with improved renal function had a significant decrease in serum creatinine (by 46% 17% to mol/l; P 0.001). Patients without improved renal function had a significant increase in serum creatinine (by 30% 38% to mol/l; P 0.03). Changes in serum creatinine significantly differed between the group with improved renal function and the group without (P ). The dose of terlipressin did not significantly differ between patients with improved renal function and those without (Table 2). The duration of terlipressin administration was significantly longer in patients with improved renal function than in patients without. There were no significant differences between the 2 groups in terms of concomitant treatments; in particular, the doses of albumin were similar (38 48 g/day vs g/day, respectively; P 0.85). Univariate analysis. Patients were significantly younger in the group with improved renal function than

4 926 MOREAU ET AL. GASTROENTEROLOGY Vol. 122, No. 4 Table 2. Treatments in Patients With Type 1 HRS Who Did or Did Not Have Improved Renal Function During Terlipressin Administration Treatments after day 0 a Patients with improved renal function (n 53) Patients without improved renal function (n 38) P value Terlipressin Dose (mg/day) Duration of administration (days) Concomitant treatments Intravenous albumin 42 (79%) 26 (68%) 0.35 For compensation of paracentesis 22 (42%) 11 (29%) 0.31 Dopamine 19 (36%) 13 (34%) 0.95 Renal replacement therapy 1 (2%) 2 (5%) 0.77 NOTE. Data are presented as mean SD or n (%). a Day 0 was the day of diagnosis of HRS. in the group without (Table 3). The proportion of patients with alcoholic cirrhosis was significantly higher in the former group than in the latter. The proportion of patients with a Child Pugh score 13 was significantly lower in the group with improved renal function than in that without. This difference between responders and nonresponders was not observed in patients with a score of 12 or 13 or in those with a score 11 (Table 3). Multivariate analysis. The independent predictors of improved renal function were younger age (P Table 3. Characteristics of Patients With Type 1 HRS Who Did or Did Not Have Improved Renal Function During Terlipressin Administration Variable Patients with improved renal function (n 53) Patients without improved renal function (n 38) P value Age (yr) Male sex 37 (70%) 30 (79%) 0.33 Duration of cirrhosis (yr) Alcoholic cirrhosis 51 (96%) 31 (82%) Diabetes mellitus 11 (21%) 7 (18%) 0.99 Previous complication a Bacterial infection 23 (43%) 12 (32%) 0.25 Severe bacterial infection b 16 (30%) 9 (24%) 0.49 Acute alcoholic hepatitis 13 (25%) 12 (32%) 0.46 Portal hypertensive bleeding 9 (17%) 11 (29%) 0.17 Previous treatments a Paracentesis 29 (55%) 21 (55%) 0.77 With plasma volume expansion 22 (42%) 15 (39%) 0.26 With albumin as plasma expander 15 (28%) 8 (21%) 0.36 Diuretics 24 (45%) 20 (53%) 0.63 Loop diuretics 13 (25%) 10 (26%) 0.96 Spironolactone 2 (4%) 3 (8%) 0.70 Both 9 (17%) 7 (18%) 0.92 Fluoroquinolones 22 (42%) 15 (39%) Blockers 13 (25%) 8 (21%) 0.70 Previous serum creatinine ( mol/l) a Serum creatinine at day 0 (mmol/l) c Increase in serum creatinine at day 0 (%) Child Pugh classification at day 0 d Score Grade (B/C) 4/34 3/ Patients with Score 13 1 (3%) 8 (24%) 0.02 Score (58%) 13 (39%) Score (39%) 12 (36%) NOTE. Data are presented as mean SD, n, or n (%). a Within the month before development of HRS; patients may have had more than 1 complication. b Patients with at least 1 of the following infections: spontaneous bacterial peritonitis, septicemia, or pneumonia. c Day 0 was the day of diagnosis of HRS. d Child Pugh score was available in 38 patients with improved renal function and 33 patients without improved renal function.

5 April 2002 TERLIPRESSIN AND HEPATORENAL SYNDROME ) and a Child Pugh score no greater than 13 at day 0(P 0.02). Survival The mean follow-up was days (range, days). Seventy-five patients died during the study. Death occurred days after day 0 (range, days). Multiorgan failure (mainly caused by bacterial infections) was the cause of death in 68 patients. Circulatory support and mechanical ventilation were used in 13 patients. The remaining causes of death were portal hypertensive bleeding in 3 patients, end-stage hepatocellular carcinoma in 3, and cerebral hemorrhage in 1. Thirteen patients underwent liver transplantation days (range, days) after day 0. Among the patients who received transplants, renal function had improved during terlipressin therapy in 10 (77%) patients and had not improved in 1 (8%), and 2 (15%) patients belonged to the subgroup of 8 patients in whom the renal response to terlipressin could not be evaluated (see previously). Among the 86 patients who did not receive transplants, 11 were alive at the last follow-up visit. Thus, 24 patients were censored for the estimation of the probability of survival. As shown in Figure 1, the estimated probability of survival was 61% at day 15, 40% at day 30, 28% at day 60, 22% at day 90, and 19% at day 365. The median survival time (95% confidence interval) was 21 days (13 34 days). Univariate analysis. The probability of survival was significantly higher in the group that had improved renal function during terlipressin therapy than in the group without (P ; Figure 2). The probability of survival was also significantly higher in patients with a Child Pugh score 11 than in patients with a score 11 (P ; Figure 3). In patients with a Child Figure 2. Probability of survival in patients with type 1 HRS who had improved renal function during terlipressin therapy and in those who did not. P value was determined by the log-rank test. Pugh score 11, the probability of survival was significantly higher in responders than in nonresponders (P 0.01). In patients with a Child Pugh score 11, the probability of survival was significantly higher in responders than in nonresponders (P ). In patients who received a plasma expander to compensate for paracentesis during the pre-hrs period (i.e., within the month before day 0), the probability of survival was significantly higher in those who received albumin than in those who received another plasma expander (P ; Figure 4). The probability of survival was significantly higher in patients who did not have severe bacterial infections (i.e., without SBP, septicemia, or pneumonia) during the pre-hrs period than in patients with this type of infection (P 0.042; Figure 5). The 11 patients who did not receive transplants and who were alive at the last follow-up visit were compared with the 35 patients who died within the first 15 days after day 0 (Table 4). The 2 groups were significantly Figure 1. Probability of survival in 99 patients with type 1 HRS treated with terlipressin. Vertical bars correspond to patients who underwent liver transplantation. Figure 3. Probability of survival in patients with type 1 HRS treated with terlipressin according to the value of the Child Pugh score ( 11 vs. 11) at inclusion. P value was determined by the log-rank test.

6 928 MOREAU ET AL. GASTROENTEROLOGY Vol. 122, No. 4 Figure 4. Probability of survival in patients with type 1 HRS treated with terlipressin according to the type of plasma expander (human albumin vs. other plasma expanders) administered to compensate for paracentesis during the pre-hrs period (i.e., within the month before inclusion). P value was determined by the log-rank test. different for Child Pugh score at day 0 and the proportion of patients with improved renal function during terlipressin therapy. Similar differences were found when the 11 patients who did not receive transplants were compared with the other 88 patients in the study (data not shown). The 11 patients who did not receive transplants and who were alive had a serum creatinine of mol/l at the last follow-up visit. Multivariate analysis. The independent predictive factors of survival were an improved renal function during terlipressin therapy (P ) and a Child Pugh score 11 at day 0 (P 0.018). Comparison Between Centers That Participated in the Study Large and small centers were not significantly different for number of patients with improved renal function during terlipressin therapy (32 [58%] vs. 26 [59%]; P 0.90) or number of patients who survived (10 [18%] vs. 10 [23%]; P 0.53). Adverse Effects Twenty-three patients had adverse events that may have been caused by terlipressin administration. Most adverse effects occurred with the lowest doses of terlipressin (i.e., 3 mg/day; Table 5). Discussion Although this was a retrospective study, it is the first investigation of the clinical course, predictive factors of improved renal function, and survival in a large series of patients with cirrhosis and type 1 HRS treated with terlipressin. At diagnosis of type 1 HRS before treatment, patients had severe cirrhosis, i.e., an increased Child Pugh score and arterial hypotension. In this study, the mean dose of terlipressin was 3 mg, and the mean duration of administration was 11 days. Terlipressin administration was associated with improved renal function in 60% of patients with HRS. Our results support preliminary results, which showed that terlipressin improved renal function in 60% 70% of patients with HRS. 6,8 This study also shows that renal failure continued to worsen in patients who did not respond to terlipressin. The duration of terlipressin administration was significantly shorter in patients who did not have improved renal function than in those with improved renal function. Treatment failure may explain the rapid discontinuation of terlipressin administration in the former group of patients. Univariate analyses showed that there were 3 predictive factors of improved renal function during terlipressin administration. First, patients who responded to terlipressin were younger than those who did not. Second, the proportion of patients with a Child Pugh score 13 was significantly lower in responders than in nonresponders. However, the proportion of patients with a Child Pugh score of 12 or 13 and those with a score 11 was similar in patients who responded to terlipressin and in those who did not. Moreover, the mean value for the Child Pugh score was not significantly different between responders and nonresponders. Together, these findings suggest that improved renal function during terlipressin therapy was independent of the Child Pugh score if this score was 13. Finally, this study showed that the Figure 5. Probability of survival in patients with type 1 HRS treated with terlipressin according to the presence or absence of severe infection (SBP, septicemia, or pneumonia) during the pre-hrs period (i.e., within the month before inclusion). P value was determined by the log-rank test.

7 April 2002 TERLIPRESSIN AND HEPATORENAL SYNDROME 929 Table 4. Comparison Between Patients Who Were Alive Without Need for Liver Transplantation and Patients Who Died During the First 15 Days of HRS Variable Patients who were alive without need for liver transplantation (n 11) Patients who rapidly died (n 35) P value Age (yr) Male sex 9 (82%) 23 (66%) 0.31 Child Pugh score Serum creatinine at day 0 (nmol/l) Dose of terlipressin (mg/day) No. of patients with improved renal function 10 (91%) 12 (34%) NOTE. Data are presented as mean SD or n (%). number of patients with alcoholic cirrhosis was higher in responders than in nonresponders. However, multivariate analysis showed that the only independent predictors of improved renal function were younger age and a Child Pugh score no greater than 13. The median survival time was 3 weeks in this study, and the chance of survival was 60%, 40%, and 28% at 15 days, 1 month, and 2 months, respectively. After the first 3 months, the probability of survival remained stable, at 19%, for several months. In a previous study investigating the natural history of HRS, the median survival time was 1.7 weeks, and the probability of survival was 40%, 25%, and 18% at 15 days, 1 month, and 2 months, respectively. 9 Taken together, these findings suggest that terlipressin may improve survival in patients with type 1 HRS. Table 5. Adverse Effects During Terlipressin Administration Adverse effect All patients (n 99) No. of events Terlipressin dose (mg/day) 3 (n 70) 3 4 (n 13) 4 (n 16) Cardiovascular system Cyanosis of extremities Ischemia (lower limbs) 2 2 Hypertension 2 2 Bradycardia Tachycardia Atrial fibrillation 1 1 Chest pain 1 1 Other systems Diarrhea Abdominal pain 3 3 Nausea, vomiting 2 2 Bronchospasm 2 2 Dyspnea 1 1 Skin lymphangitis 2 2 Increases of aspartate and alanine aminotransferases 1 1 NOTE. Complications expected to occur during the course of HRS were not considered adverse effects. Patients may have had more than 1 adverse effect. Univariate analyses identified 4 factors predictive of survival. First, Child Pugh score at the time of diagnosis of HRS was a factor predictive of survival. The probability of survival was significantly higher in patients with a Child Pugh score 11 than in patients with a score 11. This indicates that survival in patients with type 1 HRS was dependent on the severity of the underlying cirrhosis at the time of the development of HRS. Finally, for the first time, this study showed that the renal response to terlipressin was a predictive factor for prognosis. Indeed, the chance of survival was significantly higher in patients who had improved renal function than in those who did not. This relationship between survival and improved renal function might exist because both variables are a consequence of the severity of cirrhosis. Thus, survival may depend more on the severity of cirrhosis than on terlipressin-induced improved renal function. However, improved renal function was largely independent of changes in Child Pugh score (see previously). Moreover, the probability of survival was significantly higher in responders than in nonresponders, not only in the group of patients with a Child Pugh score 11, but also in the group of those with a score 11. Most important, multivariate analysis showed that improved renal function during terlipressin therapy and less severe cirrhosis at the onset of HRS were independent predictive factors of survival. Thus, in patients with type 1 HRS, terlipressin-induced improved renal function may have a beneficial effect on survival independently of the severity of the underlying liver disease. Univariate analyses also showed that among patients who underwent paracentesis during the pre-hrs period, the chance of survival was significantly higher in those receiving intravenous albumin than in those who received another plasma expander. Second, the occurrence of severe bacterial infection (i.e., SBP, septicemia, or pneumonia) during the pre-hrs period was a factor predictive of survival. Indeed, the probability of survival was significantly higher in patients without bacterial infection than in those with infection. However, para-

8 930 MOREAU ET AL. GASTROENTEROLOGY Vol. 122, No. 4 centesis followed by albumin administration and the lack of severe bacterial infection were not independent predictive factors of survival. In this study, a significant number of patients underwent liver transplantation. Most of these patients had had improved renal function during terlipressin administration. The mean delay for transplantation was 3 months. Thus, terlipressin may be a bridge to liver transplantation in patients with type 1 HRS. In this study, terlipressin was well tolerated in most patients, as in previous studies in patients with HRS 5 8,13 or refractory ascites. 14 The most common complications were cyanosis of the extremities and diarrhea. In conclusion, this retrospective study shows that 60% of patients with cirrhosis and type 1 HRS treated by terlipressin had improved renal function. The probability of improved renal function during terlipressin therapy significantly increased in younger patients who did not have severe underlying liver disease at the onset of HRS. The chance of survival significantly increased in patients with terlipressin-induced improvement in renal function without severe cirrhosis at the onset of HRS. These findings suggest that a randomized trial investigating the effects of terlipressin on survival in patients with type 1 HRS should be performed. References 1. Laffi G, La Villa G, Gentilini P. Pathogenesis and management of the hepatorenal syndrome. Semin Liver Dis 1994;14: Arroyo V, Ginès P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, Reynolds TB, Ring-Larsen H, Schölmerich J. Definition and diagnostic criteria of refractory ascites and hepatorenal syndrome in cirrhosis. Hepatology 1996;23: Moore K. The hepatorenal syndrome. Clin Sci 1997;92: Bataller R, Ginès P, Guevara M, Arroyo V. Hepatorenal syndrome. Semin Liver Dis 1997;17: Uriz J, Ginès P, Cardenas A, Sort P, Jimenez W, Salmeron JM, Battaler R, Mas A, Navasa M, Arroyo V, Rodès J. Terlipressin plus albumin infusion: an effective and safe therapy of hepatorenal syndrome. J Hepatol 2000;33: Bonnard P, Bernard B, Halimi C, Mathurin P, Demontis R, di Martino V, Henry-Biabaud E, Mofredj A, Poynard T, Cadranel JF. Effect of terlipressin (Glypressine ) on hepatorenal syndrome (HRS) in cirrhotic patients: results of a pilot study (abstr). Gastroenterology 1998;114:A Le Moine O, Mulkay JP, Bourgeois N, Adler M, Devière J. Longterm terlipressin administration improves renal function in cirrhotic patients with hepatorenal syndrome (HRS): a pilot study (abstr). Gastroenterology 1999;116:A1258 A Moreau R, Durand F, Colle I, Pessione F, Bernuau J, Barrière E, Lebrec D, Valla D. Type 1 hepatorenal syndrome in cirrhosis. Predictive factors for improvement of renal function and survival in patients treated with terlipressin (abstr). J Gastroenterol Hepatol 2000;15(Suppl):F Ginès A, Escorsell A, Ginès P, Salo J, Jimenez W, Inglada L, Navasa M, Claria J, Rimola A, Arroyo V, Rodes J. Incidence, predictive factors, and prognosis of the hepatorenal syndrome in cirrhosis with ascites. Gastroenterology 1993;105: Levacher S, Letoumelin P, Pateron D, Blaise M, Lapandry C, Pourriat JL. Early administration of terlipressin plus glyceryl trinitrate to control active upper gastrointestinal bleeding in cirrhotic patients. Lancet 1995;346: Feu F, Ruiz del Arbol L, Banares R, Planas R, Bosch J. Doubleblind randomized controlled trial comparing terlipressin and somatostatin for acute variceal hemorrhage. Variceal Bleeding Study Group. Gastroenterology 1996;111: Escorsell A, Ruiz del Arbol L, Planas R, Albillos A, Banares R, Calès P, Pateron D, Bernard B, Vinel JP, Bosch J. Multicenter randomized controlled trial of terlipressin versus sclerotherapy in the treatment of acute variceal bleeding: the TEST study. Hepatology 2000;32: Hadengue A, Gadano A, Moreau R, Giostra E, Durand F, Valla D, Erlinger S, Lebrec D. Beneficial effects of the two-day administration of terlipressin in patients with cirrhosis and hepatorenal syndrome. J Hepatol 1998;29: Gadano A, Moreau R, Vachiery F, Soupison T, Yang S, Cailmail S, Sogni P, Hadengue A, Durand F, Valla D, Lebrec D. Natriuretic response to the combination of atrial natriuretic peptide and terlipressin in patients with cirrhosis and refractory ascites. J Hepatol 1997;26: Received March 5, Accepted December 13, Address requests for reprints to: Richard Moreau, M.D., INSERM U-481, Hôpital Beaujon, Clichy, France. rmoreau@bichat. inserm.fr; fax: (33) The study was supported by Ferring SAS (Gentilly, France). The authors thank Patrick Blandin and Alain Truskolaski for their expert assistance.