- Email: [email protected]
Prognostic significance of hepatic encephalopathy in patients with cirrhosis
Journalof Hepatology ISSN1116X-827X
Prognostic significance of hepatic encephalopathy in patients with cirrhosis Javier
Bustamante,
Antoni
Rimola,
Pere-Joan
Ventura, and Juan
Liver Unit, und Epidemiology
nnd Biostutistics
N
UMEROUSstudies have been published concerning the natural history and prognostic factors in different populations of cirrhotic patients, particularly those with ascites or gastrointestinal hemorrhage (l5). The results obtained in these investigations have been useful in the selection of candidates for liver transplantation (69). In contrast, much less attention has been paid to the prognostic significance of hepatic encephalopathy (HE) as a single sign of liver disease, despite the high frequency of this complication in cirrhosis (10) and the inclusion of HE in the Child-Pugh score and classification, a common method to estimate the degree of the severity of liver disease in cirrhotic 13 July: revised 2 Drcrmher;
accepted
7 Drcembrr
Correspondence: Antoni Rimola, Liver Unit. Clinic, Villarroel 170, 08036 Barcelona, Spain. Tel: 34 93 227 5499. Fax: 34 93 451 5522. E-mail: [email protected]
890
Navasa,
Isabel
Cirera,
Unit, Institut d’lnvestigacions BiomGdiqurs August Pi i Sun)vr, Barcelona, Barcelona, Spain
Background: There are numerous studies concerning the natural history and prognostic factors in cirrhosis, the results of which are useful in selecting liver transplant candidates. However, little attention has been paid to the prognostic significance of hepatic encephalopathy despite the high frequency of this complication. Methods: We reviewed the charts of Ill cirrhotic patients who developed a first episode of acute hepatic encephalopathy to determine their survival probability and to identify prognostic factors. Results: During follow-up (12st17 months), 82 (74%) patients died. The survival probability was 42% at 1 year of follow-up and 23% at 3 years. With univariate analyses followed by a multivariate analysis, 7 out of 30 clinical and standard laboratory variables were significantly associated with poor prognosis: male sex, increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, and decreased serum
Recrived
Miquel
1998
Hospital
Virginia
Reggiardo
Rod& Hospital
Clinic, Uniwrsii~
of
albumin and prothrombin activity. Patients were classified into two groups according to a prognostic index calculated from these 7 variables. Survival probability at 1 and 3 years was 73O/oand 38%, respectively, in patients with a low prognostic index, and 10% and 3% in patients with a high prognostic index. Can&&n: Hepatic encephalopathy is associated with short survival in cirrhotic patients. Although these patients can be classified into several groups with a different prognosis, the survival probability in every group is lower than that currently expected after liver transplantation. Therefore, cirrhotic patients developing a first episode of acute hepatic encephalopathy should be considered as potential candidates for this therapeutic procedure.
Key words: Cirrhosis; Hepatic encephalopathy; Prognosis.
patients (11). In most studies investigating the natural history and prognostic factors in cirrhosis, it is only mentioned that patients with HE did worse than those without HE (1,4,12-20). To our knowledge, survival of cirrhotic patients with HE has been specifically investigated in only two studies (21,22). These two studies showed that the cumulative survival in these patients was very short: approximately 20~ 40% at 1 year and 15% at 3 years of follow-up. In one of these studies (22), several prognostic factors were identified and a prognostic index was calculated in order to group patients according to their estimated survival probability. Unfortunately, these prognostic calculations were obtained from a series of patients which included a mixture of cirrhotic patients with gastrointestinal bleeding, coma or both conditions, thus making the interpretation of the results difficult for their applicability to patients with HE alone. Furthermore, patients included in the two afore-
Prognosis of encephalopathy
mentioned studies (21,22) corresponded to relatively old series (hospital admission from 1959 to 1976, and from 1976 to 1983, respectively). Therefore, the current investigation was aimed to determine the survival expectancy of cirrhotic patients with HE admitted to hospital in the 90’s and to identify possible prognostic factors in this specific population, with the final goal being to obtain data useful for selecting candidates for liver transplantation.
Patients and Methods The medical records from all cirrhotic patients consecutively admitted to our Liver Unit between January 1990 and December 1993 were retrospectively reviewed. One hundred and eleven of these patients had their first episode of acute HE either at the time of hospital admission or during hospitalization, and were included in the study. The diagnosis and grade of HE were established on the basis of classical neuropsychiatric signs (10). No patient with proven or highly suspected hepatocellular carcinoma was included in the study. Diagnosis or high suspicion of hepatocellular carcinoma was established by standard imaging techniques (mainly ultrasonography) and/or histological or cytological examination, either during hospital admission, early after discharge or during previous admissions and outpatient controls, or at necropsy in most patients who died early after admission. In all patients the treatment of HE consisted in both the correction of possible precipitating factors and the administration of oral lactulose and cleansing enemas. In nine patients oral neomycin was also administered. Twenty-seven patients died during hospitalization (15 before and 12 after the resolution of the episode of HE, respectively). Sixty-eight of the 84 patients discharged from hospital were closely followed throughout their illness by staff members of our Liver Unit. In nine other patients we learned whether they were alive or dead (and the date of death in the latter case) by contacting the patients, their relatives or their physicians by mail or telephone. The remaining seven patients were lost to follow-up immediately after hospital discharge (three cases) or after several control outpatient visits in our Unit (four cases; time from discharge ranging between 4 and 34 months). Overall, 18 of the 111 patients underwent liver transplantation during the follow-up period. Statistical analysis Data of continuous variables are presented as meankstandard deviation. Survival probability curves were calculated using the KaplanMeier method. The curves were compared using the Mantel-Cox test. The prognostic significance of 30 variables, including demographic data, etiology of cirrhosis, maximal grade of HE, precipitating factors of HE, and clinical and standard laboratory data at the time of HE, was analyzed using a Cox regression model. Variables reaching statistical significance (psO.05) in this univariate analysis were subsequently introduced as covariates in a multivariate analysis (stepwise Cox’s multiple regression procedure; p-value to enter: ~0.10, p-value to remove: ~0.15) to identify those variables independently associated with survival. With the variables reaching statistical significance in the multivariate analysis, a prognostic index (PI) was calculated for each patient following the equation: PI=exp
(8, ~x1+fiZ~x2+...+fin~xn)
population studied (Group I). After calculating a de now PI in each patient from Group I and after classifying Group I patients into two subgroups according to whether their PI was lower/equal to or greater than the median value of the de now PI (subgroup Ia and subgroup Ib), survival probability curves were constructed for each of the two subgroups. The equation used in Group I patients was also used for calculating a de now PI in the remaining 50% of the population (Group II). Group II patients were also classified into two groups (subgroups Ha and IIb) according to the same cutpoint PI value used in the Group I patients, and survival probability curves were also constructed for each of these two subgroups. Finally, each survival curve for the subgroups of Group II was compared with the survival curve of the corresponding subgroups of Group I. The statistical analysis was carried out by using the BMDP statistical package.
Results Characteristics
of patients
There were 71 male and 40 female patients with ages of 60t12 years. The diagnosis of cirrhosis was made by histological data in 53 cases and by clinical and laboratory data in the remaining 58 cases. The etiology of cirrhosis was alcoholic in 41 patients, HCV infection in 18, HBV infection in six (associated with HDV infection in one case), primary biliary cirrhosis in two, autoimmune in one, combined factors in 23 (HCV infection and alcohol in 12, HBV infection and alcohol in six, HCV and HBV infection in two, and HCV and HBV infection and alcohol in three), and cryptogenic in 17 (in eight of these 17 cases the serological markers for HCV infection were not determined). In three patients the etiology of cirrhosis was not completely investigated. In nine (8%) patients a surgical portal-systemic shunt had been performed. Potential precipitating factors for HE were present in 100 (90%) patients. Alone or in different combinations, these factors were the following: gastrointestinal hemorrhage in 40 patients, severe infection in 45,
1.0 0.8 n
0.0
1 0
where /3i to j& are the Cox’s regression coefftcients of the variables, and xi to x, are the values of the variables in a particular patient. Higher values of PI mean a worse prognosis and lower PI values mean a better prognosis. The predictive power of the model was tested using a split-sample technique. For this purpose, regression coefficients of the independent prognostic variables were obtained in a random sample of 50% of the
in cirrhosis
I
12
24
38
48
MONTHS
Fig. I. Survival probability of Ill cirrhotic patients presenting their first episode of acute hepatic encephalopathy. Time zero of the curve corresponds to the time of diagnosis of hepatic encephalopathy.
891
J. Bustamante TABLE
et
al.
1
Hazard ratio for the different variables investigated by univariate with their first episode of hepatic encephalopthy (HE)
and multivariate
analyses
as possible prognostic
Variable
Univariate
Age
0.999 (0.980 to 1.018) 0.394 (0.231 to 0.673)”
Sex (female vs. male) Etiology of cirrhosis:b HBV infection HCV infection Alcoholic Maximal grade of HE (grade III-IV vs. grade Portal-systemic shuntb Precipitating factorsb Type of precipitating factor:b Gastrointestinal hemorrhage Infection Renai/electrolyte disturbances Diuretic therapy Constipation Ascites at the time of HEb Serum values:” Bilirubin Aspartate aminotransferase Alanine aminotransferase Alkaline phosphatase Gammaglutamyl transpeptidase Albumin Blood urea nitrogen Creatinine Sodium Potassium Other laboratory data:’ Prothrombin ratio Hemoglobin White blood cell count (/lOOO) Platelet count (/lOOO) Child-Pugh class (C vs. A+B) Year of admission (1990-91 vs. 1992-93)
I-11)
hazard
ratio”
2.001 0.653 1.289 2.794 0.948 1.415
(1.149 (0.417 (0.808 (1.706 (0.435 (0.649
to to to to to to
3.486)d 1.023) 2.054) 4.576)d 2.065) 3.086)
0.783 1.098 1.596 2.227 1.097 2.649
(0.493 (0.704 (1.023 (1.366 (0.564 (1.629
to to to to to to
1.246) 1.714) 2.489)“ 3.631)d 2.132) 4.307)d
1.124 1.001 0.999 1.002 1.000 0.933 1.022 1.216 0.969 1.316
(1.084 (0.993 (0.994 (1.001 (0.997 (0.891 (1.010 (0.969 (0.947 (1.000
to to to to to to to to to to
1.165)d 1.008) 1.009) 1.003)d 1.003) 0.977)d 1.034)d 1.526) 0.992)” 1.748)“
0.969 0.976 1.025 1.001 3.494 0.662
(0.956 (0.879 (0.985 (0.998 (1.857 (0.413
to to to to to to
0.982)d 1.083) 1.066) 1.005) 6.573)d 1.060)
factors
Multivariate
in 111 cirrhotic
hazard
patients
ratio”
0.416 (0.231 to 0.747)”
1.156 (1.088 to 1.228)d
1.002 (1.001 to 1.003)d 0.928 (0.878 to 0.981)” 1.030 (1.015 to 1.046)“
1.410 (1.084 to 1.834)d 0.981 (0.966 to 0.996)”
“In brackets: 95% confidence interval. bPresence vs. absence. CHazard ratio per unit increase. dps0.05.
renal function or hydroelectrolytic disturbances in 6 1, diuretic therapy in 73, constipation in 12, and miscellaneous factors in eight. The maximal grade of HE was I in 25 (23%) patients, II in 57 (51%), III in 23 (21%) and IV in six (5%). HE was resolved in 96 patients (424 days after the diagnosis of this complication), whereas the remaining 15 patients died before the resolution of HE. At the time of inclusion in the study, 65 (59%) patients had ascites. At inclusion, serum bilirubin was 5.5k5.7 mg/dl, serum albumin 29+5 g/l and prothrombin activity 52218%. Two (2%) patients had ChildPugh class A cirrhosis, 24 (22%) had class B and 77 (69%) had class C. In the remaining eight patients the Child-Pugh score could not be calculated. Blood urea nitrogen was 26? 16 mg/dl, serum creatinine 1.2kO.8 mg/dl, serum sodium 13128 mmol/l, and serum potassium 4.1 kO.9 mmol/l. 892
Mortality and prognostic factors The follow-up period of the study (time zero: diagnosis of HE; end-points: January 1997, death, liver transplantation or loss to follow-up) was 122 17 months. During this period, 82 (74%) out of the 111 patients died. The main causes of death were liver failure in 50 patients (in five cases liver failure was associated with hepatocellular carcinoma development), severe infectious complications in 12, gastrointestinal hemorrhage in six, miscellaneous in five (traffic accident, hemoperitoneum, renal failure, cardiac arrest during a status epilepticus, and pulmonary thromboembolism) and unknown in nine. The probability of survival in the whole series of patients was 42”/0 at 1 year of followup, 27% at 2 years, and 23% at 3 years (Fig. 1). Among the 30 variables analyzed, 14 were significantly (p~O.05) associated with a poor prognosis in univariate analyses: male sex, HBV infection, presence
Prognosis of encephalopathy in cirrhosis
of ascites at the time of HE, renal/electrolyte disturbances and diuretic therapy as possible precipitating factors of HE, grade III or IV HE (maximal grade during the episode of HE), increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, decreased serum albumin, prothrombin activity and serum sodium at the time of HE, and Child-Pugh class C at the time of HE (Table 1). All these variables were introduced into the multivariate analysis, with the exception of renal and/or electrolyte disturbances and diuretic therapy as possible precipitating factors of HE because they provide information already implied in other variables considered, namely ascites, blood urea nitrogen and serum sodium and potassium concentration. The multivariate analysis (performed in the 93 patients in whom the values of the remaining 12 variables could be collected) identified seven variables as significant ($10.05) independent prognostic factors: sex, serum bilirubin, alkaline phosphatase, albumin and potassium, blood urea nitrogen and prothrombin activity (Table 1). By using the seven variables with independent prognostic value and their regression coefficients, a PI was individually calculated in each patient according to the following equation (described in the Patients and Methods section): PI=exp{(-0.8782Xsex)+(O.l449Xbilirubin) +(0.0021 XAP)-(0.0746Xalbumin) +(0.3437Xpotassium)+(O.O299xBUN) -(O.O193Xprothrombin)) where bilirubin, alkaline phosphatase (AP), albumin, potassium, blood urea nitrogen (BUN) and prothrom-
5cq&‘;l;:“““p”’ la
GROUP
&.__--------_______
I
0.2-
------I 0.0 1 0
GROUP lb
8
1--------n 12
GROUP Ilb 24
36
3 46
MONTHS Fig. 3. Survival probability in four subsets of cirrhotic patients randomly grouped following a split-sample technique (Groups I and II) and classified according to their prognostic index (low in Groups Ia and IIa, and high in Groups Ib and IIb). No significant differences were observed between curves from Groups Ia versus IIa andfrom Groups Ib versus IIb.
bin were introduced as continuous variables, and sex was introduced as a categorical variable (1 =male, and 2=female). The 96 patients in whom their PI could be calculated were subsequently classified into two groups according to whether they had low or high PI (~0.4175 or >0.4175, respectively; 0.4175 was the median value for the whole series of patients). Fig. 2 shows the survival probability in these two groups of patients. The survival curves calculated in the four subgroups obtained by the split-sample technique (described in the Patients and Methods section) are depicted in Fig. 3. No significant differences were observed between the survival curves of each pair of subgroups, thus showing the internal validity of the model described.
Discussion
HIGH PI 0.0
1 0
I_________
_____( I
12
24
36
46
MONTHS Fig. 2. Survival probability of a series of cirrhotic patients with afirst episode of acute hepatic encephalopathy, classified into two groups according to whether their prognostic index (PI) was low (PI~O.417.5; n=48) or high (PDO.4175; n=48). PI of 0.4175 was the median value of the whole series. p=O.OOOO.
The results of the present study indicate that the development of HE in cirrhotic patients is a sign associated with short life expectancy. In our investigation the cumulative survival at 1 and 3 years after the presentation of the first episode of acute HE was only 42% and 23%, respectively (Fig. 1). These data are in agreement with the results previously reported by other authors (21,22) and suggest that the prognosis in cirrhotic patients developing HE has not substantially changed during the last decades. Furthermore, we identified several prognostic factors in these patients. Among the 30 variables analyzed, male sex and six standard liver and renal function tests, such as increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, and decreased serum albumin and prothrombin activity,
893
J. Bustamante
et al.
were independently associated with a poor prognosis (Table 1). The relationship between prognosis and serum bilirubin, serum albumin and prothrombin activity is reasonable since these parameters are commonly considered as reliable markers of the degree of liver failure in cirrhosis. Such a relationship agrees with the results obtained in most studies (1,4,12-20,2232). Similarly, the prognostic significance of blood urea nitrogen and serum potassium found in the present investigation is not surprising since a strong relationship between renal function parameters and prognosis has been reported in other studies in which renal function tests have been included in the list of variables investigated as possible prognostic factors in cirrhotic patients, particularly in those with ascites or in critical condition (1,20,22,33,34). The correlation between increased serum alkaline phosphatase and poor survival found in the present investigation is less comprehensible, although it has also been observed in other studies (24,25,32). In our study, the survival of male patients was shorter than for female patients. The explanation of this different survival is uncertain, although this finding was also observed in two previously published studies (23,25). The Child-Pugh classification reached statistical significance as a prognostic factor in the univariate analysis but not in the multivariate analysis, regardless of the fact that three of its components did (serum bilirubin, albumin and prothrombin activity). This finding was also observed by Merkel et al. (16). Furthermore, a lower accuracy of the Child-Pugh classification in predicting survival as compared with the variables comprising the score expressed in their original units was also reported by Infante-Rivard et al. (35). One of the most remarkable findings in the present study was the fact that, after classifying our patients into two groups according to their different prognostic index, the survival probability in each group (Fig. 2) was lower than that currently reported for patients undergoing liver transplantation, which is approximately 80% at 1 year and 70% at 3 years (36,37). Even in the group with a better outcome in our investigation (group with low prognostic index, Fig. 2) the actuarial survival rate was 73% at 1 year of follow-up, but only 38% at 3 years. From comparison of these figures the conclusion may be drawn that cirrhotic patients developing the first episode of acute HE should be considered as potential candidates for liver transplantation. Nevertheless, although the mathematical model described in the study was internally validated by the split-sample technique, further studies are needed to support this conclusion.
894
The very low survival observed in the current study in cirrhotic patients with acute HE was somewhat unexpected to our group. Although we knew that HE was associated with a poor prognosis in series of cirrhotic patients studied in the past decades (21,22), we thought that the survival expectancy of these patients in the 1990’s was much better because of the improvements in both the general management of severely ill cirrhotic patients and the management of complications which frequently act as precipitating factors for HE and which can strongly influence the final outcome of cirrhotic patients with HE (i.e. gastrointestinal hemorrhage or serious infections). This could explain why only 23 out of the 111 patients included in the study were evaluated as possible liver transplant candidates (data not shown in the results) in spite of our hospital having a well-established liver transplant program at the time of the study. This proportion can be considered as clearly small, even after excluding those patients with potential contraindications for liver transplantation (i.e. old age or serious extrahepatic diseases). One factor which could have theoretically influenced the short survival observed in our series is the low proportion of patients included with surgical portal-systemic shunting, who are commonly thought to have a relatively good prognosis (38). The results of our study, however, do not support this idea since the cumulative survival of patients with and without portal-systemic shunts after their first episode of hepatic encephalopathy was similar: 77% versus 62% at 3 months, 33% versus 42% at 1 year, and 22% versus 23% at 3 years of follow-up, respectively @-value=0.89; data not reported in the Results). Nevertheless, any conclusion on the prognosis of patients with surgical portal-systemic shunts after the first episode of hepatic encephalopathy should be made cautiously because the number of these patients in our series was small (nine patients). In conclusion, in this study we found that the development of the first episode of acute HE in cirrhotic patients is associated with short survival and that, although these patients can be classified into several subsets with different prognoses according to a mathematically calculated Pl, the survival probability in each subset is lower than that presently expected for liver transplant recipients. Therefore, cirrhotic patients developing HE should be considered as potential candidates for liver transplantation.
References 1. Llach J, Gin&s P, Arroyo V, Rimola A. Tit6 L, Badalamenti S, et al. Prognostic value of arterial pressure, endogenous vasoactive systems, and renal function in cirrhotic patients admitted to the
Prognosis of encephalopathy in cirrhosis hospital for the treatment of ascites. Gastroenterology 1988; 94: 482-7. 2. Tit6 L, Rimola A, Gin&s P, Llach J, Arroyo V, Rod& J. Recurrence of spontaneous bacterial peritonitis in cirrhosis: frequency and predictive factors. Hepatology 1988; 8: 27-31. 3. Gin& P, Gin& A, Arroyo V Prognosis of patients with cirrhosis and ascites. In: Rod&s J, Arroyo V, editors. Therapy in Liver Disease. Barcelona: Ediciones Doyma S.A.; 1992. p. 16672. 4. Salerno F, Borroni G, Moser E Badalamenti S, Cassara L, Maggi A, et al. Survival and prognostic factors of cirrhotic patients with ascites: a study of 134 outpatients. Am J Gastroenterol 1993; 88: 5149. 5. Pagliaro L, D’Amico G, Pasta L, Politi F, Vizzini G, Traina M, et al. Portal hypertension in cirrhosis: natural history. In: Bosch J, Groszmann RJ, editors. Portal Hypertension: Pathophysiology and Treatment. Oxford: Blackwell Scientific Publications; 1994. p. 72-92. 6. Lake JR. Changing indications for liver transplantation. Med Clin N Am 1993; 22: 213-29. 7. Consensus statement on indications for liver transplantation: Paris, June 22-23, 1993. Hepatology 1994; 2O(suppl.): 63S-68s. 8. Rimola A, Navasa M, Rod& J, Elias E, Neuberger J, Lucey MR, et al. Specific indications. Chronic parenchymal liver disease. In: Neuberger J, Lucey MR, editors. Liver Transplantation: Practice and Management. London: BMJ Publishing Group; 1994. p. 34 104. 9. Wiesner RH. Current indications, contraindications, and timing for liver transplantation. In: Busuttil RW, Klintmahn GB, editors. Transplantation of the Liver. Philadelphia: W.B. Saunders Co.; 1996. p. 71-84. 10. Conn HO. Hepatic encephalopathy. In: Schiff L, Schiff ER, editors. Diseases of the Liver, Vol 2, 7th ed. Philadelphia: J.B. Lippincott Company; 1993. p. 103660. 11. Pugh RNH, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: 6469. 12. Ekindjian OG, Devanlay M, Duchassaing D, Leluan G, Kammerer J, Fouet P, et al. Multivariate analysis of clinical and biological data in cirrhotic patients: application to prognosis. Eur J Clin Invest 1981; 11: 213-20. 13. Sauerbruch T, Weinzierl M, Ktipcke W, Paumgartner G. Longterm sclerotherapy of bleeding esophageal varices in patients with liver cirrhosis. An evaluation of mortality and rebleeding risk factors. Stand J Gastroenterol 1985; 20: 51-8. 14. Milani A, Marra L, Sicilian0 M, Rossi L. Prognostic significance of clinical and laboratory parameters in liver cirrhosis: a multivariate statistical approach. Hepato-Gastroenterol 1985; 32: 270-2. 15. D’Amico G, Morabito A, Pagliaro L, Marubini E, and the Study Group of “V Cervello” Hospital. Survival and prognostic indicators in compensated and decompensated cirrhosis. Dig Dis Sci 1986; 31: 468-75. 16. Merkel C, Bolognesi M, Angeli E Noventa F, Caregaro L, Sacerdoti D, et al. Prognostic indicators of survival in patients with cirrhosis and esophageal varices, without previous bleeding. Am J Gastroenterol 1989; 84: 717-22. 17. Adler M, Van Laethem J, Glibert A, Gelin M, Bourgeois N, Vereerstraeten E et al. Factors influencing survival at one year in patients with nonbiliary hepatic parenchymal cirrhosis. Dig Dis Sci 1990; 35: 1-5. 18. De Jongh FE, Janssen HL, De Man RA, Hop WC, Schalm SW, Van Blankenstein M. Survival and prognostic indicators in hepatitis B surface antigen-positive cirrhosis of the liver. Gastroenterology 1992; 103: 1692-4. 19. Arrigoni A, Gindro T, Aimo G, Cape110 N, Meloni A, Bendetti E et al. Monoethylglicinexylidide test: a prognostic indicator of survival in cirrhosis. Hepatology 1994; 20: 383-7.
20. Magliocchetti N, Torchio P, Corrao G, Aricb S, Favilli S. Prognostic factors for long-term survival in cirrhotic patients after the first episode of liver decompensation. Ital J Gastroenterol Hepatol 1997; 29: 3845. 21. Saunders JB, Walters JRF, Davies P, Paton A. A 20-year prospective study of cirrhosis. Br Med J 1981: 282: 263-6. 22. Christensen E, Krintel JJ, Hansen SM, Johansen JK, Juhl E. Prognosis after the first episode of gastrointestinal bleeding or coma in cirrhosis. Survival and prognostic factors. Stand J Gastroenterol 1989; 24: 999-1006. 23. Schlichting P, Christensen E, Andersen PK, Fauerholdt L, Juhl E, Poulsen H, et al. Prognostic factors in cirrhosis identified by Cox’s regression model. Hepatology 1983; 3: 889-95. 24. Christensen E, Schlichting P, Andersen PK, Fauerholdt L, Schou G, Pedersen BV, et al. Updating prognosis and therapeutic effect evaluation in cirrhosis with Cox’s multiple regression model for time-dependent variables. Stand J Gastroenterol 1986; 21: 16374. 25. Gin& P, Quintero E, Arroyo V, Teres J, Bruguera M, Rimola A, et al. Compensated cirrhosis: natural history and prognostic factors. Hepatology 1987; 7: 122-8. 26. Tage-Jensen U, Henriksen JH, Christensen E, Widding A, RingLarsen H, Christensen NJ. Plasma catecholamine level and portal venous pressure as guides to prognosis in patients with cirrhosis. J Hepatol 1988; 6: 35&8. 27. Merkel C, Bolognesi M, Finucci GF, Angeli P, Caregaro L, Rondana M, et al. Indocyanine green intrinsic hepatic clearance as a prognostic index of survival in patients with cirrhosis. J Hepatol 1989; 9: 1622. 28. Oellerich M, Burdelski M, Lautz H-U, Rodeck B, Duewel J, Schultz M, et al. Assessment of pretransplant prognosis in patients with cirrhosis. Transplantation 1991; 51: 801-6. 29. Ferro D, Saliola M, Quintarelli C, Alessandri C, Basili S, Cordova A. l-year survey of patients with advanced liver cirrhosis. Prognostic value of clinical and laboratory indexes identified by the Cox regression model. Stand J Gastroenterol 1992; 27: 8526. 30. Navasa M, Garcia-Pagan JC, Bosch J, Rod& J. Valor pronostico de1 aclaramiento hepatico de1 Verde de indocianina en pacientes con cirrosis hepatica y hemorragia digestiva por varices eosfagicas. Gastroenterol Hepatol 1992; 98: 290-4. 31. Realdi G, Fattovich G, Hadziyannis S, Schalm SW, Almasio E Sanchez-Tapias J, et al. Survival and prognostic factors in 336 patients with compensated cirrhosis type B: a multicenter study. J Hepatol 1994; 21: 65666. 32. Adler M, Verset D, Bouhdid H, Bourgeois N, Gulbis B, Le Moine 0, et al. Prognostic evaluation of patients with parenchymal cirrhosis. Proposal of a new simple score. J Hepatol 1997; 26: 6429. 33. Salerno F, Borroni G, Moser P, Sangiovanni A, Almasio P, Budillon G, et al. Prognostic value of the galactose test in predicting survival of patients with cirrhosis evaluated for liver transplantation. J Hepatol 1996; 25: 47480. 34. Castera L, Pauwels L, Levy V-G. Indicateurs pronostiques chez les malades atteints de cirrhose admis en service de reanimation. Gastroenterol Clin Biol 1996; 20: 263-8. 35. Infante-Rivard C, Esnaola S, Villeneuve J-P Clinical and statistical validity of conventional prognostic factors in predicting shortterm survival among cirrhotics. Hepatology 1987; 7: 66&l. 36. Kilpe VE, Krakauer H, Wren RE. An analysis of liver transplant experience from 37 transplant centers as reported to Medicare. Transplantation 1993; 56: 554-61. 37. European Liver Transplant Registry. Data analysis from 0511968 to 0611997. Hopital Paul Brousse, Villejuif, France. 38. Sherlock S, Dooley J. Diseases of the Liver and Biliary System, 10th ed. Oxford: Blackwell Science Ltd.; 1997. p. 87-102.
895
Prognostic significance of hepatic encephalopathy in patients with cirrhosis Javier
Bustamante,
Antoni
Rimola,
Pere-Joan
Ventura, and Juan
Liver Unit, und Epidemiology
nnd Biostutistics
N
UMEROUSstudies have been published concerning the natural history and prognostic factors in different populations of cirrhotic patients, particularly those with ascites or gastrointestinal hemorrhage (l5). The results obtained in these investigations have been useful in the selection of candidates for liver transplantation (69). In contrast, much less attention has been paid to the prognostic significance of hepatic encephalopathy (HE) as a single sign of liver disease, despite the high frequency of this complication in cirrhosis (10) and the inclusion of HE in the Child-Pugh score and classification, a common method to estimate the degree of the severity of liver disease in cirrhotic 13 July: revised 2 Drcrmher;
accepted
7 Drcembrr
Correspondence: Antoni Rimola, Liver Unit. Clinic, Villarroel 170, 08036 Barcelona, Spain. Tel: 34 93 227 5499. Fax: 34 93 451 5522. E-mail: [email protected]
890
Navasa,
Isabel
Cirera,
Unit, Institut d’lnvestigacions BiomGdiqurs August Pi i Sun)vr, Barcelona, Barcelona, Spain
Background: There are numerous studies concerning the natural history and prognostic factors in cirrhosis, the results of which are useful in selecting liver transplant candidates. However, little attention has been paid to the prognostic significance of hepatic encephalopathy despite the high frequency of this complication. Methods: We reviewed the charts of Ill cirrhotic patients who developed a first episode of acute hepatic encephalopathy to determine their survival probability and to identify prognostic factors. Results: During follow-up (12st17 months), 82 (74%) patients died. The survival probability was 42% at 1 year of follow-up and 23% at 3 years. With univariate analyses followed by a multivariate analysis, 7 out of 30 clinical and standard laboratory variables were significantly associated with poor prognosis: male sex, increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, and decreased serum
Recrived
Miquel
1998
Hospital
Virginia
Reggiardo
Rod& Hospital
Clinic, Uniwrsii~
of
albumin and prothrombin activity. Patients were classified into two groups according to a prognostic index calculated from these 7 variables. Survival probability at 1 and 3 years was 73O/oand 38%, respectively, in patients with a low prognostic index, and 10% and 3% in patients with a high prognostic index. Can&&n: Hepatic encephalopathy is associated with short survival in cirrhotic patients. Although these patients can be classified into several groups with a different prognosis, the survival probability in every group is lower than that currently expected after liver transplantation. Therefore, cirrhotic patients developing a first episode of acute hepatic encephalopathy should be considered as potential candidates for this therapeutic procedure.
Key words: Cirrhosis; Hepatic encephalopathy; Prognosis.
patients (11). In most studies investigating the natural history and prognostic factors in cirrhosis, it is only mentioned that patients with HE did worse than those without HE (1,4,12-20). To our knowledge, survival of cirrhotic patients with HE has been specifically investigated in only two studies (21,22). These two studies showed that the cumulative survival in these patients was very short: approximately 20~ 40% at 1 year and 15% at 3 years of follow-up. In one of these studies (22), several prognostic factors were identified and a prognostic index was calculated in order to group patients according to their estimated survival probability. Unfortunately, these prognostic calculations were obtained from a series of patients which included a mixture of cirrhotic patients with gastrointestinal bleeding, coma or both conditions, thus making the interpretation of the results difficult for their applicability to patients with HE alone. Furthermore, patients included in the two afore-
Prognosis of encephalopathy
mentioned studies (21,22) corresponded to relatively old series (hospital admission from 1959 to 1976, and from 1976 to 1983, respectively). Therefore, the current investigation was aimed to determine the survival expectancy of cirrhotic patients with HE admitted to hospital in the 90’s and to identify possible prognostic factors in this specific population, with the final goal being to obtain data useful for selecting candidates for liver transplantation.
Patients and Methods The medical records from all cirrhotic patients consecutively admitted to our Liver Unit between January 1990 and December 1993 were retrospectively reviewed. One hundred and eleven of these patients had their first episode of acute HE either at the time of hospital admission or during hospitalization, and were included in the study. The diagnosis and grade of HE were established on the basis of classical neuropsychiatric signs (10). No patient with proven or highly suspected hepatocellular carcinoma was included in the study. Diagnosis or high suspicion of hepatocellular carcinoma was established by standard imaging techniques (mainly ultrasonography) and/or histological or cytological examination, either during hospital admission, early after discharge or during previous admissions and outpatient controls, or at necropsy in most patients who died early after admission. In all patients the treatment of HE consisted in both the correction of possible precipitating factors and the administration of oral lactulose and cleansing enemas. In nine patients oral neomycin was also administered. Twenty-seven patients died during hospitalization (15 before and 12 after the resolution of the episode of HE, respectively). Sixty-eight of the 84 patients discharged from hospital were closely followed throughout their illness by staff members of our Liver Unit. In nine other patients we learned whether they were alive or dead (and the date of death in the latter case) by contacting the patients, their relatives or their physicians by mail or telephone. The remaining seven patients were lost to follow-up immediately after hospital discharge (three cases) or after several control outpatient visits in our Unit (four cases; time from discharge ranging between 4 and 34 months). Overall, 18 of the 111 patients underwent liver transplantation during the follow-up period. Statistical analysis Data of continuous variables are presented as meankstandard deviation. Survival probability curves were calculated using the KaplanMeier method. The curves were compared using the Mantel-Cox test. The prognostic significance of 30 variables, including demographic data, etiology of cirrhosis, maximal grade of HE, precipitating factors of HE, and clinical and standard laboratory data at the time of HE, was analyzed using a Cox regression model. Variables reaching statistical significance (psO.05) in this univariate analysis were subsequently introduced as covariates in a multivariate analysis (stepwise Cox’s multiple regression procedure; p-value to enter: ~0.10, p-value to remove: ~0.15) to identify those variables independently associated with survival. With the variables reaching statistical significance in the multivariate analysis, a prognostic index (PI) was calculated for each patient following the equation: PI=exp
(8, ~x1+fiZ~x2+...+fin~xn)
population studied (Group I). After calculating a de now PI in each patient from Group I and after classifying Group I patients into two subgroups according to whether their PI was lower/equal to or greater than the median value of the de now PI (subgroup Ia and subgroup Ib), survival probability curves were constructed for each of the two subgroups. The equation used in Group I patients was also used for calculating a de now PI in the remaining 50% of the population (Group II). Group II patients were also classified into two groups (subgroups Ha and IIb) according to the same cutpoint PI value used in the Group I patients, and survival probability curves were also constructed for each of these two subgroups. Finally, each survival curve for the subgroups of Group II was compared with the survival curve of the corresponding subgroups of Group I. The statistical analysis was carried out by using the BMDP statistical package.
Results Characteristics
of patients
There were 71 male and 40 female patients with ages of 60t12 years. The diagnosis of cirrhosis was made by histological data in 53 cases and by clinical and laboratory data in the remaining 58 cases. The etiology of cirrhosis was alcoholic in 41 patients, HCV infection in 18, HBV infection in six (associated with HDV infection in one case), primary biliary cirrhosis in two, autoimmune in one, combined factors in 23 (HCV infection and alcohol in 12, HBV infection and alcohol in six, HCV and HBV infection in two, and HCV and HBV infection and alcohol in three), and cryptogenic in 17 (in eight of these 17 cases the serological markers for HCV infection were not determined). In three patients the etiology of cirrhosis was not completely investigated. In nine (8%) patients a surgical portal-systemic shunt had been performed. Potential precipitating factors for HE were present in 100 (90%) patients. Alone or in different combinations, these factors were the following: gastrointestinal hemorrhage in 40 patients, severe infection in 45,
1.0 0.8 n
0.0
1 0
where /3i to j& are the Cox’s regression coefftcients of the variables, and xi to x, are the values of the variables in a particular patient. Higher values of PI mean a worse prognosis and lower PI values mean a better prognosis. The predictive power of the model was tested using a split-sample technique. For this purpose, regression coefficients of the independent prognostic variables were obtained in a random sample of 50% of the
in cirrhosis
I
12
24
38
48
MONTHS
Fig. I. Survival probability of Ill cirrhotic patients presenting their first episode of acute hepatic encephalopathy. Time zero of the curve corresponds to the time of diagnosis of hepatic encephalopathy.
891
J. Bustamante TABLE
et
al.
1
Hazard ratio for the different variables investigated by univariate with their first episode of hepatic encephalopthy (HE)
and multivariate
analyses
as possible prognostic
Variable
Univariate
Age
0.999 (0.980 to 1.018) 0.394 (0.231 to 0.673)”
Sex (female vs. male) Etiology of cirrhosis:b HBV infection HCV infection Alcoholic Maximal grade of HE (grade III-IV vs. grade Portal-systemic shuntb Precipitating factorsb Type of precipitating factor:b Gastrointestinal hemorrhage Infection Renai/electrolyte disturbances Diuretic therapy Constipation Ascites at the time of HEb Serum values:” Bilirubin Aspartate aminotransferase Alanine aminotransferase Alkaline phosphatase Gammaglutamyl transpeptidase Albumin Blood urea nitrogen Creatinine Sodium Potassium Other laboratory data:’ Prothrombin ratio Hemoglobin White blood cell count (/lOOO) Platelet count (/lOOO) Child-Pugh class (C vs. A+B) Year of admission (1990-91 vs. 1992-93)
I-11)
hazard
ratio”
2.001 0.653 1.289 2.794 0.948 1.415
(1.149 (0.417 (0.808 (1.706 (0.435 (0.649
to to to to to to
3.486)d 1.023) 2.054) 4.576)d 2.065) 3.086)
0.783 1.098 1.596 2.227 1.097 2.649
(0.493 (0.704 (1.023 (1.366 (0.564 (1.629
to to to to to to
1.246) 1.714) 2.489)“ 3.631)d 2.132) 4.307)d
1.124 1.001 0.999 1.002 1.000 0.933 1.022 1.216 0.969 1.316
(1.084 (0.993 (0.994 (1.001 (0.997 (0.891 (1.010 (0.969 (0.947 (1.000
to to to to to to to to to to
1.165)d 1.008) 1.009) 1.003)d 1.003) 0.977)d 1.034)d 1.526) 0.992)” 1.748)“
0.969 0.976 1.025 1.001 3.494 0.662
(0.956 (0.879 (0.985 (0.998 (1.857 (0.413
to to to to to to
0.982)d 1.083) 1.066) 1.005) 6.573)d 1.060)
factors
Multivariate
in 111 cirrhotic
hazard
patients
ratio”
0.416 (0.231 to 0.747)”
1.156 (1.088 to 1.228)d
1.002 (1.001 to 1.003)d 0.928 (0.878 to 0.981)” 1.030 (1.015 to 1.046)“
1.410 (1.084 to 1.834)d 0.981 (0.966 to 0.996)”
“In brackets: 95% confidence interval. bPresence vs. absence. CHazard ratio per unit increase. dps0.05.
renal function or hydroelectrolytic disturbances in 6 1, diuretic therapy in 73, constipation in 12, and miscellaneous factors in eight. The maximal grade of HE was I in 25 (23%) patients, II in 57 (51%), III in 23 (21%) and IV in six (5%). HE was resolved in 96 patients (424 days after the diagnosis of this complication), whereas the remaining 15 patients died before the resolution of HE. At the time of inclusion in the study, 65 (59%) patients had ascites. At inclusion, serum bilirubin was 5.5k5.7 mg/dl, serum albumin 29+5 g/l and prothrombin activity 52218%. Two (2%) patients had ChildPugh class A cirrhosis, 24 (22%) had class B and 77 (69%) had class C. In the remaining eight patients the Child-Pugh score could not be calculated. Blood urea nitrogen was 26? 16 mg/dl, serum creatinine 1.2kO.8 mg/dl, serum sodium 13128 mmol/l, and serum potassium 4.1 kO.9 mmol/l. 892
Mortality and prognostic factors The follow-up period of the study (time zero: diagnosis of HE; end-points: January 1997, death, liver transplantation or loss to follow-up) was 122 17 months. During this period, 82 (74%) out of the 111 patients died. The main causes of death were liver failure in 50 patients (in five cases liver failure was associated with hepatocellular carcinoma development), severe infectious complications in 12, gastrointestinal hemorrhage in six, miscellaneous in five (traffic accident, hemoperitoneum, renal failure, cardiac arrest during a status epilepticus, and pulmonary thromboembolism) and unknown in nine. The probability of survival in the whole series of patients was 42”/0 at 1 year of followup, 27% at 2 years, and 23% at 3 years (Fig. 1). Among the 30 variables analyzed, 14 were significantly (p~O.05) associated with a poor prognosis in univariate analyses: male sex, HBV infection, presence
Prognosis of encephalopathy in cirrhosis
of ascites at the time of HE, renal/electrolyte disturbances and diuretic therapy as possible precipitating factors of HE, grade III or IV HE (maximal grade during the episode of HE), increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, decreased serum albumin, prothrombin activity and serum sodium at the time of HE, and Child-Pugh class C at the time of HE (Table 1). All these variables were introduced into the multivariate analysis, with the exception of renal and/or electrolyte disturbances and diuretic therapy as possible precipitating factors of HE because they provide information already implied in other variables considered, namely ascites, blood urea nitrogen and serum sodium and potassium concentration. The multivariate analysis (performed in the 93 patients in whom the values of the remaining 12 variables could be collected) identified seven variables as significant ($10.05) independent prognostic factors: sex, serum bilirubin, alkaline phosphatase, albumin and potassium, blood urea nitrogen and prothrombin activity (Table 1). By using the seven variables with independent prognostic value and their regression coefficients, a PI was individually calculated in each patient according to the following equation (described in the Patients and Methods section): PI=exp{(-0.8782Xsex)+(O.l449Xbilirubin) +(0.0021 XAP)-(0.0746Xalbumin) +(0.3437Xpotassium)+(O.O299xBUN) -(O.O193Xprothrombin)) where bilirubin, alkaline phosphatase (AP), albumin, potassium, blood urea nitrogen (BUN) and prothrom-
5cq&‘;l;:“““p”’ la
GROUP
&.__--------_______
I
0.2-
------I 0.0 1 0
GROUP lb
8
1--------n 12
GROUP Ilb 24
36
3 46
MONTHS Fig. 3. Survival probability in four subsets of cirrhotic patients randomly grouped following a split-sample technique (Groups I and II) and classified according to their prognostic index (low in Groups Ia and IIa, and high in Groups Ib and IIb). No significant differences were observed between curves from Groups Ia versus IIa andfrom Groups Ib versus IIb.
bin were introduced as continuous variables, and sex was introduced as a categorical variable (1 =male, and 2=female). The 96 patients in whom their PI could be calculated were subsequently classified into two groups according to whether they had low or high PI (~0.4175 or >0.4175, respectively; 0.4175 was the median value for the whole series of patients). Fig. 2 shows the survival probability in these two groups of patients. The survival curves calculated in the four subgroups obtained by the split-sample technique (described in the Patients and Methods section) are depicted in Fig. 3. No significant differences were observed between the survival curves of each pair of subgroups, thus showing the internal validity of the model described.
Discussion
HIGH PI 0.0
1 0
I_________
_____( I
12
24
36
46
MONTHS Fig. 2. Survival probability of a series of cirrhotic patients with afirst episode of acute hepatic encephalopathy, classified into two groups according to whether their prognostic index (PI) was low (PI~O.417.5; n=48) or high (PDO.4175; n=48). PI of 0.4175 was the median value of the whole series. p=O.OOOO.
The results of the present study indicate that the development of HE in cirrhotic patients is a sign associated with short life expectancy. In our investigation the cumulative survival at 1 and 3 years after the presentation of the first episode of acute HE was only 42% and 23%, respectively (Fig. 1). These data are in agreement with the results previously reported by other authors (21,22) and suggest that the prognosis in cirrhotic patients developing HE has not substantially changed during the last decades. Furthermore, we identified several prognostic factors in these patients. Among the 30 variables analyzed, male sex and six standard liver and renal function tests, such as increased serum bilirubin, alkaline phosphatase, potassium and blood urea nitrogen, and decreased serum albumin and prothrombin activity,
893
J. Bustamante
et al.
were independently associated with a poor prognosis (Table 1). The relationship between prognosis and serum bilirubin, serum albumin and prothrombin activity is reasonable since these parameters are commonly considered as reliable markers of the degree of liver failure in cirrhosis. Such a relationship agrees with the results obtained in most studies (1,4,12-20,2232). Similarly, the prognostic significance of blood urea nitrogen and serum potassium found in the present investigation is not surprising since a strong relationship between renal function parameters and prognosis has been reported in other studies in which renal function tests have been included in the list of variables investigated as possible prognostic factors in cirrhotic patients, particularly in those with ascites or in critical condition (1,20,22,33,34). The correlation between increased serum alkaline phosphatase and poor survival found in the present investigation is less comprehensible, although it has also been observed in other studies (24,25,32). In our study, the survival of male patients was shorter than for female patients. The explanation of this different survival is uncertain, although this finding was also observed in two previously published studies (23,25). The Child-Pugh classification reached statistical significance as a prognostic factor in the univariate analysis but not in the multivariate analysis, regardless of the fact that three of its components did (serum bilirubin, albumin and prothrombin activity). This finding was also observed by Merkel et al. (16). Furthermore, a lower accuracy of the Child-Pugh classification in predicting survival as compared with the variables comprising the score expressed in their original units was also reported by Infante-Rivard et al. (35). One of the most remarkable findings in the present study was the fact that, after classifying our patients into two groups according to their different prognostic index, the survival probability in each group (Fig. 2) was lower than that currently reported for patients undergoing liver transplantation, which is approximately 80% at 1 year and 70% at 3 years (36,37). Even in the group with a better outcome in our investigation (group with low prognostic index, Fig. 2) the actuarial survival rate was 73% at 1 year of follow-up, but only 38% at 3 years. From comparison of these figures the conclusion may be drawn that cirrhotic patients developing the first episode of acute HE should be considered as potential candidates for liver transplantation. Nevertheless, although the mathematical model described in the study was internally validated by the split-sample technique, further studies are needed to support this conclusion.
894
The very low survival observed in the current study in cirrhotic patients with acute HE was somewhat unexpected to our group. Although we knew that HE was associated with a poor prognosis in series of cirrhotic patients studied in the past decades (21,22), we thought that the survival expectancy of these patients in the 1990’s was much better because of the improvements in both the general management of severely ill cirrhotic patients and the management of complications which frequently act as precipitating factors for HE and which can strongly influence the final outcome of cirrhotic patients with HE (i.e. gastrointestinal hemorrhage or serious infections). This could explain why only 23 out of the 111 patients included in the study were evaluated as possible liver transplant candidates (data not shown in the results) in spite of our hospital having a well-established liver transplant program at the time of the study. This proportion can be considered as clearly small, even after excluding those patients with potential contraindications for liver transplantation (i.e. old age or serious extrahepatic diseases). One factor which could have theoretically influenced the short survival observed in our series is the low proportion of patients included with surgical portal-systemic shunting, who are commonly thought to have a relatively good prognosis (38). The results of our study, however, do not support this idea since the cumulative survival of patients with and without portal-systemic shunts after their first episode of hepatic encephalopathy was similar: 77% versus 62% at 3 months, 33% versus 42% at 1 year, and 22% versus 23% at 3 years of follow-up, respectively @-value=0.89; data not reported in the Results). Nevertheless, any conclusion on the prognosis of patients with surgical portal-systemic shunts after the first episode of hepatic encephalopathy should be made cautiously because the number of these patients in our series was small (nine patients). In conclusion, in this study we found that the development of the first episode of acute HE in cirrhotic patients is associated with short survival and that, although these patients can be classified into several subsets with different prognoses according to a mathematically calculated Pl, the survival probability in each subset is lower than that presently expected for liver transplant recipients. Therefore, cirrhotic patients developing HE should be considered as potential candidates for liver transplantation.
References 1. Llach J, Gin&s P, Arroyo V, Rimola A. Tit6 L, Badalamenti S, et al. Prognostic value of arterial pressure, endogenous vasoactive systems, and renal function in cirrhotic patients admitted to the
Prognosis of encephalopathy in cirrhosis hospital for the treatment of ascites. Gastroenterology 1988; 94: 482-7. 2. Tit6 L, Rimola A, Gin&s P, Llach J, Arroyo V, Rod& J. Recurrence of spontaneous bacterial peritonitis in cirrhosis: frequency and predictive factors. Hepatology 1988; 8: 27-31. 3. Gin& P, Gin& A, Arroyo V Prognosis of patients with cirrhosis and ascites. In: Rod&s J, Arroyo V, editors. Therapy in Liver Disease. Barcelona: Ediciones Doyma S.A.; 1992. p. 16672. 4. Salerno F, Borroni G, Moser E Badalamenti S, Cassara L, Maggi A, et al. Survival and prognostic factors of cirrhotic patients with ascites: a study of 134 outpatients. Am J Gastroenterol 1993; 88: 5149. 5. Pagliaro L, D’Amico G, Pasta L, Politi F, Vizzini G, Traina M, et al. Portal hypertension in cirrhosis: natural history. In: Bosch J, Groszmann RJ, editors. Portal Hypertension: Pathophysiology and Treatment. Oxford: Blackwell Scientific Publications; 1994. p. 72-92. 6. Lake JR. Changing indications for liver transplantation. Med Clin N Am 1993; 22: 213-29. 7. Consensus statement on indications for liver transplantation: Paris, June 22-23, 1993. Hepatology 1994; 2O(suppl.): 63S-68s. 8. Rimola A, Navasa M, Rod& J, Elias E, Neuberger J, Lucey MR, et al. Specific indications. Chronic parenchymal liver disease. In: Neuberger J, Lucey MR, editors. Liver Transplantation: Practice and Management. London: BMJ Publishing Group; 1994. p. 34 104. 9. Wiesner RH. Current indications, contraindications, and timing for liver transplantation. In: Busuttil RW, Klintmahn GB, editors. Transplantation of the Liver. Philadelphia: W.B. Saunders Co.; 1996. p. 71-84. 10. Conn HO. Hepatic encephalopathy. In: Schiff L, Schiff ER, editors. Diseases of the Liver, Vol 2, 7th ed. Philadelphia: J.B. Lippincott Company; 1993. p. 103660. 11. Pugh RNH, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: 6469. 12. Ekindjian OG, Devanlay M, Duchassaing D, Leluan G, Kammerer J, Fouet P, et al. Multivariate analysis of clinical and biological data in cirrhotic patients: application to prognosis. Eur J Clin Invest 1981; 11: 213-20. 13. Sauerbruch T, Weinzierl M, Ktipcke W, Paumgartner G. Longterm sclerotherapy of bleeding esophageal varices in patients with liver cirrhosis. An evaluation of mortality and rebleeding risk factors. Stand J Gastroenterol 1985; 20: 51-8. 14. Milani A, Marra L, Sicilian0 M, Rossi L. Prognostic significance of clinical and laboratory parameters in liver cirrhosis: a multivariate statistical approach. Hepato-Gastroenterol 1985; 32: 270-2. 15. D’Amico G, Morabito A, Pagliaro L, Marubini E, and the Study Group of “V Cervello” Hospital. Survival and prognostic indicators in compensated and decompensated cirrhosis. Dig Dis Sci 1986; 31: 468-75. 16. Merkel C, Bolognesi M, Angeli E Noventa F, Caregaro L, Sacerdoti D, et al. Prognostic indicators of survival in patients with cirrhosis and esophageal varices, without previous bleeding. Am J Gastroenterol 1989; 84: 717-22. 17. Adler M, Van Laethem J, Glibert A, Gelin M, Bourgeois N, Vereerstraeten E et al. Factors influencing survival at one year in patients with nonbiliary hepatic parenchymal cirrhosis. Dig Dis Sci 1990; 35: 1-5. 18. De Jongh FE, Janssen HL, De Man RA, Hop WC, Schalm SW, Van Blankenstein M. Survival and prognostic indicators in hepatitis B surface antigen-positive cirrhosis of the liver. Gastroenterology 1992; 103: 1692-4. 19. Arrigoni A, Gindro T, Aimo G, Cape110 N, Meloni A, Bendetti E et al. Monoethylglicinexylidide test: a prognostic indicator of survival in cirrhosis. Hepatology 1994; 20: 383-7.
20. Magliocchetti N, Torchio P, Corrao G, Aricb S, Favilli S. Prognostic factors for long-term survival in cirrhotic patients after the first episode of liver decompensation. Ital J Gastroenterol Hepatol 1997; 29: 3845. 21. Saunders JB, Walters JRF, Davies P, Paton A. A 20-year prospective study of cirrhosis. Br Med J 1981: 282: 263-6. 22. Christensen E, Krintel JJ, Hansen SM, Johansen JK, Juhl E. Prognosis after the first episode of gastrointestinal bleeding or coma in cirrhosis. Survival and prognostic factors. Stand J Gastroenterol 1989; 24: 999-1006. 23. Schlichting P, Christensen E, Andersen PK, Fauerholdt L, Juhl E, Poulsen H, et al. Prognostic factors in cirrhosis identified by Cox’s regression model. Hepatology 1983; 3: 889-95. 24. Christensen E, Schlichting P, Andersen PK, Fauerholdt L, Schou G, Pedersen BV, et al. Updating prognosis and therapeutic effect evaluation in cirrhosis with Cox’s multiple regression model for time-dependent variables. Stand J Gastroenterol 1986; 21: 16374. 25. Gin& P, Quintero E, Arroyo V, Teres J, Bruguera M, Rimola A, et al. Compensated cirrhosis: natural history and prognostic factors. Hepatology 1987; 7: 122-8. 26. Tage-Jensen U, Henriksen JH, Christensen E, Widding A, RingLarsen H, Christensen NJ. Plasma catecholamine level and portal venous pressure as guides to prognosis in patients with cirrhosis. J Hepatol 1988; 6: 35&8. 27. Merkel C, Bolognesi M, Finucci GF, Angeli P, Caregaro L, Rondana M, et al. Indocyanine green intrinsic hepatic clearance as a prognostic index of survival in patients with cirrhosis. J Hepatol 1989; 9: 1622. 28. Oellerich M, Burdelski M, Lautz H-U, Rodeck B, Duewel J, Schultz M, et al. Assessment of pretransplant prognosis in patients with cirrhosis. Transplantation 1991; 51: 801-6. 29. Ferro D, Saliola M, Quintarelli C, Alessandri C, Basili S, Cordova A. l-year survey of patients with advanced liver cirrhosis. Prognostic value of clinical and laboratory indexes identified by the Cox regression model. Stand J Gastroenterol 1992; 27: 8526. 30. Navasa M, Garcia-Pagan JC, Bosch J, Rod& J. Valor pronostico de1 aclaramiento hepatico de1 Verde de indocianina en pacientes con cirrosis hepatica y hemorragia digestiva por varices eosfagicas. Gastroenterol Hepatol 1992; 98: 290-4. 31. Realdi G, Fattovich G, Hadziyannis S, Schalm SW, Almasio E Sanchez-Tapias J, et al. Survival and prognostic factors in 336 patients with compensated cirrhosis type B: a multicenter study. J Hepatol 1994; 21: 65666. 32. Adler M, Verset D, Bouhdid H, Bourgeois N, Gulbis B, Le Moine 0, et al. Prognostic evaluation of patients with parenchymal cirrhosis. Proposal of a new simple score. J Hepatol 1997; 26: 6429. 33. Salerno F, Borroni G, Moser P, Sangiovanni A, Almasio P, Budillon G, et al. Prognostic value of the galactose test in predicting survival of patients with cirrhosis evaluated for liver transplantation. J Hepatol 1996; 25: 47480. 34. Castera L, Pauwels L, Levy V-G. Indicateurs pronostiques chez les malades atteints de cirrhose admis en service de reanimation. Gastroenterol Clin Biol 1996; 20: 263-8. 35. Infante-Rivard C, Esnaola S, Villeneuve J-P Clinical and statistical validity of conventional prognostic factors in predicting shortterm survival among cirrhotics. Hepatology 1987; 7: 66&l. 36. Kilpe VE, Krakauer H, Wren RE. An analysis of liver transplant experience from 37 transplant centers as reported to Medicare. Transplantation 1993; 56: 554-61. 37. European Liver Transplant Registry. Data analysis from 0511968 to 0611997. Hopital Paul Brousse, Villejuif, France. 38. Sherlock S, Dooley J. Diseases of the Liver and Biliary System, 10th ed. Oxford: Blackwell Science Ltd.; 1997. p. 87-102.
895