- Email: [email protected]
Morbidity and mortality in compensated cirrhosis type C: A retrospective follow-up study of 384 patients
GASTROENTEROLOGY 1997;112:463–472
Morbidity and Mortality in Compensated Cirrhosis Type C: A Retrospective Follow-up Study of 384 Patients GIOVANNA FATTOVICH,* GIULIANO GIUSTINA,‡ FRANCOISE DEGOS,§ FEDERICO TREMOLADA,‡ GIULIO DIODATI,‡ PIERO ALMASIO,x FREDERIK NEVENS,Ø ANTONIO SOLINAS,# DANIELA MURA,** JOHANNES T. BROUWER,‡‡ HOWARD THOMAS,§§ CORINNE NJAPOUM,§ CARLA CASARIN,‡ PAOLA BONETTI,‡ PATRIZIA FUSCHI,x JOVAN BASHO,Ø ANDREINA TOCCO,# ABHA BHALLA,‡‡ RACHEL GALASSINI,§§ FRANCO NOVENTA,‡ SOLKO W. SCHALM,‡‡ and GIUSEPPE REALDI** *Istituto di Semeiotica e Nefrologia Medica, University of Verona, Verona, Italy; ‡Istituto di Medicina Clinica, Clinica Medica 2a, Universita` di Padova, Padova, Italy; §Fe`de`ration d’He`patogastroente`rologie et INSERM Unite´ 24, Hopital Beaujon, Paris, France; xCattedra di Clinica Medica R, University of Palermo, Palermo, Italy; ØInternal Medicine, University Hospital Gasthuisberg, Leuven, Belgium; #Istituto di Patologia Speciale Medica and **Istituto di Clinica Medica, University of Sassari, Sassari, Italy; ‡‡Internal Medicine II, University Hospital Rotterdam, Rotterdam, The Netherlands; and §§Department of Medicine, St. Mary’s Hospital Medical School, London, England
See editorial on page 651. Background & Aims: Few data are available concerning the long-term prognosis of chronic liver disease associated with hepatitis C virus infection. This study examined the morbidity and survival of patients with compensated cirrhosis type C. Methods: A cohort of 384 European cirrhotic patients was enrolled at seven tertiary referral hospitals and followed up for a mean period of 5 years. Inclusion criteria were biopsy-proven cirrhosis, abnormal serum aminotransferase levels, absence of complications of cirrhosis, and exclusion of hepatitis A and B viruses and of metabolic, toxic, or autoimmune liver diseases. Results: Antibodies against hepatitis C virus were positive in 98% of 361 patients tested. The 5-year risk of hepatocellular carcinoma was 7% and that of decompensation was 18%. Death occurred in 51 patients (13%), with 70% dying of liver disease. Survival probability was 91% and 79% at 5 and 10 years, respectively. Two hundred five patients (53%) were treated with interferon alfa. After adjustment for clinical and serological differences at baseline between patients treated or not treated with interferon, the 5-year estimated survival probability was 96% and 95% for treated and untreated patients, respectively. Conclusions: In this cohort of patients, life expectancy is relatively long, in agreement with the morbidity data showing a slowly progressive disease.
about 8%.6 However, few data are available concerning the clinical course of compensated cirrhosis associated with hepatitis C, and at present, the question is how frequently serious complications of cirrhosis occur. The rate of development of decompensation or hepatocellular carcinoma (HCC) and the degree to which they contribute to the mortality rate are not well known. Moreover, there is uncertainty about the long-term mortality rate of chronic hepatitis caused by hepatitis C virus (HCV). On the basis of these considerations, the primary objective of this study was to evaluate the morbidity and survival of histologically documented cirrhosis associated with HCV in patients who had never had clinical manifestations of decompensation, such as jaundice, ascites, variceal bleeding, or encephalopathy, but who had biochemical signs of disease activity. Morbidity was studied in relation to the incidence of HCC and the rate of appearance of hepatic decompensation. The knowledge of the clinical course and prognosis of this clinical condition is of particular importance in clinical practice and can influence medical decision making, such as rational application of therapeutic measures.8 The outcome of the disease was assessed retrospectively in a cohort of Western European patients who were referred to tertiary-care centers and followed up longitudinally. We also analyzed the serological and clinical data of these patients as prognostic factors for survival.
Patients and Methods
T
he prognosis of non-A, non-B (NANB)/C chronic hepatitis is a matter of controversy. It is well documented that NANB/C chronic hepatitis may progress to cirrhosis,1 – 7 and the risk of developing cirrhosis has been reported to be 40% at 5 years and 60% at 8 years, as calculated by actuarial analysis, with an annual rate of / 5e18$$0016
01-13-97 17:45:57
gasa
The study was planned within the European Concerted Action on Viral Hepatitis (EUROHEP) and involved seven Abbreviation used in this paper: IFN-a, interferon alfa. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
464 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
university hospitals in Europe that serve a tertiary referral function. The data were to be collected according to a protocol that followed basic epidemiological principles required for conducting prognostic studies,8,9 with clear definitions of the study population, time of entry, referral pattern, and follow up.
Study Population Several studies have shown that most patients with NANB chronic liver disease have HCV infection.10 – 12 Therefore, it was decided to include patients with NANB hepatitis in the study and to retrospectively test stored serum for HCV. The study included all consecutive patients admitted to the participating centers between January 1982 and December 1992 and who fulfilled the following criteria: (1) biopsy-proven cirrhosis according to international criteria13; (2) persistent elevations of alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) concentrations (at least 1.5 times the upper limit of normal) for at least 6 months before entry into the study; (3) absence of hepatitis B surface antigen and all other potential causes of chronic liver disease; (4) no history or clinical evidence at enrollment of complications of cirrhosis, i.e., ascites, variceal bleeding, encephalopathy, or jaundice; (5) no evidence of HCC at entry into the study on the basis of ultrasonography, a-fetoprotein levels (õ400 mg/L), and/or histology; and (6) follow-up at the enrolling center for a minimum period of 6 months. Patients with other forms of liver disease were excluded by history, serological parameters, and histology. Patients with a history of alcohol abuse, defined as alcohol intake of ¢80 g/day for more than 5 years, were also excluded from the study. Information on serum anti-HCV antibodies was obtained, but anti-HCV testing was not considered an inclusion criterion for eligibility into the study.
Time of Entry Into the Study and Referral Pattern Entry into the study was defined as the time of diagnosis of NANB/C compensated cirrhosis. The four referral patterns to the enrolling center were defined as follows: incidental finding of elevated aminotransferase levels; presence of nonspecific symptoms; diagnosis of cirrhosis made at the enrolling center during follow-up for NANB/C chronic hepatitis (in this case, the time of entry into the study was established as the time of histological diagnosis of cirrhosis); and referral to the enrolling center after the histological diagnosis of cirrhosis was made elsewhere previously (in this case, the start of follow-up was established as the time of referral to the enrolling center).
Follow-up The time of observation was calculated from the date of entry until death or the end of the observation period (June 30, 1993). All patients underwent clinical assessment and standard liver biochemical tests at least once a year or more frequently if indicated. Death was classified as caused by liver failure if progressive impairment of liver function occurred, or if it occurred within 40 days of variceal bleeding, regardless
/ 5e18$$0016
01-13-97 17:45:57
gasa
of the severity,14 or if it was a result of HCC. Patients who died of conditions not related to liver disease and patients who were lost to follow-up were censored at the time of death or at the time of dropout, respectively, in the statistics.15 HCC was diagnosed on the basis of histological findings and/or elevated a-fetoprotein levels (ú400 mg/L) and ultrasonographic findings consistent with HCC. Decompensation was defined as the appearance of at least one episode of ascites, jaundice, hepatic encephalopathy, or gastrointestinal bleeding of variceal origin. Biochemical remission was defined as the normalization of serum aminotransferase values on at least two consecutive determinations obtained at least 3 weeks apart and persisting at the time of last observation. A response to interferon alfa (IFN-a) therapy was considered as the normalization of aminotransferase levels by the end of treatment and persisting at the time of last observation. The actual total dose of IFN-a was defined as low when õ200 million units was administered and as medium/high when values greater than this limit were administered.
Laboratory Tests Results of anti-HCV tests were obtained from medical records of patients, or anti-HCV testing was performed on stored serum samples by a second-generation enzyme-linked immunosorbent assay according to the manufacturer’s instructions.
Prognostic Factors The following clinical and serological variables at entry were evaluated as potential prognostic factors for survival: sex, age, referral pattern, presumed source of infection, age at the time of transfusion (õ50 years or ¢50 years), alcohol consumption, symptoms, hepatic stigmata, liver firmness, splenomegaly, esophageal varices, AST, ALT, AST/ALT ratio, bilirubin, albumin, prothrombin time, g-globulin, platelets, and a-fetoprotein. Moreover, the prognostic value for survival of IFN-a therapy was analyzed. Liver firmness was defined as an increased consistency of the liver from firm to rock hard. Splenomegaly was documented by physical examination and/or ultrasonography (ú13 cm in longitudinal diameter).
Statistical Analysis The x2 test and Student’s t test were used to compare frequencies and means. The cumulative probability of survival and of developing HCC or decompensation was calculated using the method of Kaplan and Meier.15 Univariate analysis of survival was performed by computing survival curves according to the Kaplan–Meier method. In the univariate analysis of survival between patients treated with IFN-a and those not treated with IFN-a therapy, the time zero for IFN-a–treated patients was the time of starting treatment. Curves were compared statistically using the log rank test16 and Breslow test. Multivariate analysis of prognostic factors for survival was also studied by stepwise forward Cox regression model.17 In this analysis, IFN-a therapy during follow-up was evaluated as a
WBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 465
time-dependent variable, taking into account the time of starting treatment. In this model, the order of insertion of any variable was determined by using the maximum log-likehood value, and the statistical significance was assessed by the likelihood ratio test.18 Because the analysis requires that all patients are represented by a complete set of variables and to not reduce the number of patients, missing data were replaced by neutral estimates.19 The estimated curves of the probability of survival in the two subgroups of patients treated with IFN-a and those not treated with antiviral therapy were calculated according to a statistical method that adjusted for the influence of significant variables as found by Cox model.20 The statistical software used was BMDP.21
Results A total of 384 patients fulfilling the inclusion criteria was enrolled at seven European centers from 1982 to 1992 and followed up for a mean period of 61 months (range, 6–153 months). Sixteen patients (4%) were lost to follow-up. Patient Characteristics The clinical and serological features of the patients at entry are shown in Table 1. The patients ranged in age from 19 to 78 years (mean age, 54 years). There were 223 men and 161 women; 370 of the patients were white, 11 were Asian, and 3 were black. At entry, 138 patients had a history of occasional or daily alcohol intake of õ80 g, whereas the remaining 246 patients denied any alcohol consumption. At presentation, nonspecific symptoms such as dyspepsia, asthenia, and/or upper abdominal discomfort were present in 166 patients (44%). However, only 60 patients were referred to the enrolling center because of these nonspecific symptoms. On physical examination, liver firmness was detected in 79% of 299 patients whose liver was palpable. Hepatic stigmata and splenomegaly were detected in 31% and 34% of the patients, respectively. Histological activity on liver biopsy was present in the large majority of the patients (95%). Of the 75 patients in whom the histological diagnosis of cirrhosis was made elsewhere before being referred to the enrolling center, 38 had undergone the liver biopsy within 12 months (range, 1–12 months) of enrollment, whereas 37 had a histological diagnosis of cirrhosis from 2 to 17 years before entry into the study. The baseline serological features showed that most patients had normal values of bilirubin (61%), albumin (90%), g-globulin (69%), and prothrombin time (67%). None of the 128 patients tested was anti–human immunodeficiency virus positive. Three hundred fifty-five of the 361 patients tested (98%) were positive for serum antibody to HCV. Of the 384 patients enrolled in the study, 159 (41%) / 5e18$$0016
01-13-97 17:45:57
gasa
Table 1. Baseline Clinical and Serological Features in Patients With NANB/C Compensated Cirrhosis Agea (yr ) (n Å 384) Men (%) (n Å 384) Referral pattern (%) (n Å 384) Diagnosis of cirrhosis during follow-up for chronic hepatitis Incidental finding of abnormal transaminase levels Diagnosis of cirrhosis made elsewhere previously Symptoms Histology (%) (n Å 384) Active cirrhosis Inactive cirrhosis AST (%) (n Å 380) °1.5 unv ú1.5–3 unv ú3 unv ALT (%) (n Å 371) °1.5 unv ú1.5–3 unv ú3 unv Bilirubin (%) (n Å 365) °17 mmol/L ú17 õ51 mmol/L Albumin (%) (n Å 344) ¢35 g/L õ35 g/L g-Globulin (%) (n Å 315) °20 g/L ú20 g/L Platelets (%) (n Å 372) ú130 1 109/L °130 1 109/L Prothrombin time (%) (n Å 343) ¢70% õ70% a-Fetoprotein (%) (n Å 200) °10 mg/L ú10 mg/L
54 { 4.93 223 (58)
130 119 75 60
(34) (31) (19) (16)
363 (95) 21 (5) 33 (9) 177 (46) 170 (45) 22 (6) 126 (34) 223 (60) 224 (61) 141 (39) 309 (90) 35 (10) 217 (69) 98 (31) 181 (49) 191 (51) 230 (67) 113 (33) 114 (57) 86 (43)
NOTE. n values are the total numbers of patients with analyzable data. unv, upper limit of normal value. a Value is expressed as mean { SD.
had a history of blood transfusion, 9 (2%) had a history of intravenous drug abuse, 11 (3%) had occupational exposure to blood, 63 (16%) had a history of major surgery, 29 (8%) had a family member with chronic liver disease, and 113 patients (30%) had no risk factors for infection. Of the 159 patients with a history of blood transfusion, 14 had received multiple transfusions. In 125 of the remaining 145 patients, the year of the blood transfusion was documented according to the patient’s recollection, and the mean age at the time of the transfusion was 43 years (range, 8–66 years). A total of 81 patients received the transfusion before the age of 50 years (mean age, 35 years) and 44 at the age of 50 years or older (mean age, 56 years). The mean interval between the time of the blood transfusion and the diagnosis of cirrhosis was 12 WBS-Gastro
466 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
Figure 1. Cumulative probability of HCC appearance in patients with NANB/C compensated cirrhosis. The 5-year appearance rate was 7%, and the 10-year appearance rate was 14%.
years (range, 1–43 years) among all patients, and 5.7 and 15.8 years, respectively, for patients who were 50 years or older at the time of transfusion or who were younger than 50 years. Therapy During follow-up, 205 patients (53.5%) were treated with IFN-a, 2 (0.5%) with ribavirin, and 19 (5%) with steroids, whereas the remaining 158 patients (41%) remained untreated. The majority of patients treated with IFN-a received a medium/high dose of IFN-a (181 of 205 patients; 88%). A response to IFN-a therapy with biochemical remission lasting for 5–70 months (mean, 43 months) after stopping treatment was observed in 18 of 205 patients (9%). Persistent biochemical remission for 5–126 months (mean, 36 months) occurred in an additional 12 patients, namely, 7 of 158 untreated patients (5%) and 5 patients at variable intervals after stopping IFN-a treatment. None of the patients treated with steroids showed sustained aminotransferase level normalization.
(72%) were men. At diagnosis, patients were between 44 and 72 years of age (mean age, 63 years). Twentyfour of the 29 patients were tested for antibody to HCV and were antibody to HCV positive. Seventeen of the 29 patients died 1–19 months after the diagnosis of HCC, whereas 12 patients were still alive at the end of the observation. Development of decompensation. Of the 355 patients who remained tumor free, 65 (18%) developed at least one episode of ascites, jaundice, hepatic encephalopathy, or variceal bleeding. The mean interval between the time of entry into the study and the appearance of the first episode of decompensation was 37 months (range, 3–137 months). In the remaining 290 patients, the cirrhosis remained compensated during follow-up. The first episode of decompensation was caused by ascites, variceal bleeding, encephalopathy, or jaundice in 31, 14, 5, and 4 patients, respectively, and by more than one complication in the remaining 11 patients. The cumulative probability of developing decompensated cirrhosis after diagnosis was 12% at 3 years, 18% at 5 years, and 29% at 10 years with a yearly incidence of 3.9% during the first 5 years (Figure 2). Survival. During follow-up, 51 patients (13%) died, including 3 of the 6 patients who underwent liver transplantation. Complications related to HCC accounted for 33% of the deaths (17 cases), liver failure for 31% (16 cases), and bleeding for 6% (3 cases). Fifteen patients (30%) died of causes unrelated to cirrhosis. The rate of liver-related mortality during the course of the study was 9% (36 of 384 patients). The mean length of time from entry into the study to death from liver disease was 50 months (range, 9–124 months). The probability of survival after diagnosis of compensated cirrhosis was
Morbidity Development of HCC. During observation, HCC developed in 29 of the 384 patients (8%). Liver cancer was diagnosed in 20 patients on histology and in 9 patients on the basis of elevated a-fetoprotein values (ú400 mg/L) and compatible ultrasonography. The mean interval between the time of entry into the study and the diagnosis of HCC was 48 months (range, 7–134 months). The cumulative probability of HCC appearance among the patients was 4% at 3 years after diagnosis of cirrhosis, 7% at 5 years when 157 patients were still in followup, and 14% at 10 years with a yearly incidence of 1.4% during the first 5 years (Figure 1). Twenty-one patients
/ 5e18$$0016
01-13-97 17:45:57
gasa
Figure 2. Cumulative probability of developing decompensation in 355 patients with NANB/C compensated cirrhosis who remained tumor free. The 5-year probability rate was 18%, and the 10-year probability rate was 29%.
WBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 467
Figure 3. Curve A represents the probability of survival after diagnosis in all patients studied with NANB/C compensated cirrhosis. The 5year probability rate was 91%. Curve B represents the probability of survival after the appearance of the first major complication of the disease in 65 patients who developed decompensated cirrhosis. The 5-year probability rate was 50%.
5, top). However, multivariate analysis by Cox’s model failed to show any independent prognostic significance of IFN-a therapy for survival rates (Table 4). The difference in adjusted estimated survival curves, based on the Cox’s model for median values of bilirubin, age, platelets, and hepatic stigmata, between the two groups was not significant (Figure 5, bottom). The 5year adjusted estimated survival probability was 96% for IFN-a–treated patients and 95% for patients not receiving IFN-a. As shown in Table 5, IFN-a–treated patients were significantly younger and their illness was significantly less severe with fewer findings of splenomegaly or esophageal varices; with less elevated bilirubin and g-globulin levels; with higher albumin and platelet levels; and with greater prothrombin time compared with patients not receiving IFN-a.
Discussion 96%, 91%, and 79% at 3, 5, and 10 years, respectively (Figure 3). The yearly incidence of mortality was 1.9 during the first 5 years. The probability of survival after the onset of the first major complication of the disease was 50% at 5 years (Figure 3). Factors correlating with survival. Among the clinical and biochemical features studied, 16 variables significantly predicted survival at univariate analysis: age, referral pattern, source of infection, alcohol consumption, symptoms, hepatic stigmata on physical examination, splenomegaly, esophageal varices, AST, AST/ALT ratio, bilirubin, albumin, g-globulin, platelets, prothrombin time, and interferon-a therapy (Tables 2 and 3). The probability of survival was significantly different according to the presumed source of infection, with a better survival observed in patients with a history of intravenous drug abuse or blood transfusion (Table 2 and Figure 4). Among posttransfusion cirrhotic patients, at univariate analysis, the probability of survival did not differ between patients who had been administered the transfusion before the age of 50 years and those who had the transfusion at age 50 years or older (P Å 0.67). After multivariate analysis with Cox’s model, bilirubin, hepatic stigmata on physical examination, age, and platelets remained as factors independently related to survival (Table 4). Survival in Relation to Antiviral Therapy At univariate analysis, the probability of survival was significantly different (P Å 0.016) between the two subgroups of patients, those treated with IFN-a and those not receiving this drug, with a better survival rate observed in IFN-a–treated patients (Table 2 and Figure / 5e18$$0016
01-13-97 17:45:57
gasa
Chronic HCV infection is a well-known cause of chronic liver disease both in patients with histories of exposure to parenterally transmitted viral diseases10,11 and in patients who have no recognizable source of infection.12 In this study, 98% of patients with posttransfusion or cryptogenic cirrhosis had detectable antibody to HCV, thus confirming the high prevalence of antibody to HCV in these patients and indicating that chronic HCV infection is a very important cause of liver cirrhosis in Europe. The long-term outcome of this chronic condition has not yet been fully investigated. Previous longitudinal studies have only included a small number of cirrhotic patients5 – 7 and/or have not distinguished between compensated and decompensated cirrhosis.22 In this study, we evaluated retrospectively the clinical evolution of a large cohort of predominantly white Western European patients with compensated cirrhosis caused by HCV who were followed up for a mean period of 5 years with a low dropout rate. The main findings regard frequency and timing of HCC, decompensation, and mortality. Chronic HCV infection is considered one of the major causes of liver cancer in Western countries. Previous cross-sectional studies have shown a markedly high prevalence of antibody to HCV in patients with HCC in Europe.23,24 Follow-up studies of posttransfusion hepatitis have documented the sequential development of cirrhosis and HCC in some patients.3,4,6 However, the magnitude of the risk for HCC has not yet been quantified in chronic HCV infection. In the current study, HCC developed in 8% of 384 Western European patients with compensated cirrhosis type C with a yearly incidence of 1.4%. Other factors promoting liver cancer, such as alcoWBS-Gastro
468 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
Table 2. Clinical Variables Showing Prognostic Significance for Survival at Univariate Analysis Probability of survival Variable
P value
No. of patients
5 yr (SE)
10 yr (SE)
166 218
0.95 (0.02) 0.88 (0.03)
0.83 (0.06) 0.75 (0.06)
0.028a
130 119 75 60
0.94 0.94 0.87 0.85
(0.02) (0.02) (0.04) (0.05)
0.81 0.83 0.82 0.73
(0.07) (0.06) (0.06) (0.09)
0.019b
9 159 103 113
1.00 0.94 0.90 0.89
(0.00) (0.02) (0.03) (0.03)
— 0.94 (0.02) 0.76 (0.08) 0.66 (0.09)
0.019c
138 246
0.88 (0.03) 0.96 (0.06)
0.77 (0.06) 0.79 (0.06)
0.0439b
166 207
0.87 (0.03) 0.95 (0.02)
0.67 (0.08) 0.90 (0.03)
0.012a
113 252
0.81 (0.04) 0.96 (0.01)
0.71 (0.06) 0.83 (0.06)
0.001a
124 235
0.85 (0.04) 0.95 (0.02)
0.70 (0.08) 0.85 (0.05)
0.0062a
79 115
0.77 (0.06) 0.95 (0.02)
0.53 (0.12) 0.85 (0.07)
0.00001a
205 179
0.94 (0.02) 0.87 (0.03)
— 0.71 (0.06)
0.016a
Age (yr ) (n Å 384) °54 ú54 Referral pattern (n Å 384) Diagnosis of cirrhosis during follow-up for chronic hepatitis Incidental finding of abnormal transaminase levels Diagnosis of cirrhosis made elsewhere previously Symptoms Source of infection (n Å 384) Intravenous drug abuse Blood transfusion Otherd Unknown Alcohol intake (õ80 g/day) (n Å 384) Yes No Symptoms (n Å 373) Present Absent Hepatic stigmata (n Å 365) Present Absent Splenomegaly (n Å 359) Present Absent Esophageal varices (n Å 194) Present Absent IFN-a therapy (n Å 384) Yes No
NOTE. Values in parentheses represent the total number of patients on which survival calculation is based because initial data were not available in all patients. a P value derived from log rank test and Breslow test. b P value significant by Breslow test. c P value significant by log rank test. d Occupational exposure to blood, major surgery, or family member with chronic liver disease.
hol abuse or hepatitis B virus coinfection, were excluded in all cases. In a study conducted in Italy analyzing the prevalence of HCC among 417 patients with compensated cirrhosis of viral and nonviral origin, liver cancer developed in 4% of 189 HCV-related cases.25 This lower prevalence of HCC is most likely caused by the shorter follow-up of the Italian patients, averaging 33 months, compared with a mean follow-up of 61 months in our series. The rate of liver cancer may be greater for a longer period. Indeed, in our cirrhotic patients, the risk of HCC increased progressively. At 3 and 5 years, the probability of liver cancer appearance was 4% and 7%, respectively. A recent study conducted in Japan has reported a greater 5-year probability of liver cancer occurrence of approximately 25% in untreated Asian patients with compensated cirrhosis type C.26 Liver cancer associated with HCV infection appears to be increasing in various / 5e18$$0016
01-13-97 17:45:57
gasa
countries, particularly in Japan,27,28 suggesting geographic variation in the incidence of HCC that may relate to viral strain, genetic factors, or yet unknown environmental factors. Interestingly, most patients did not seem to have evidence of rapidly progressive liver disease, and our data indicate that a high proportion of patients with compensated cirrhosis do well for at least 10 years. This finding together with the absence of subjective complaints in most patients suggest that the disease has a relatively long subclinical course until features of overt decompensation or cancer develop. In our series, the liver-related mortality rate was 9% during a mean follow-up of 5 years, and HCC and liver failure were the main causes of death. In other reports, the overall liver-related mortality rate varies from 2.5% to 14%.5 – 7,22,29 This variability may be related to differWBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 469
Table 3. Serological Variables Showing Prognostic Significance for Survival at Univariate Analysis Probability of survival Variable AST (n Å 380) °1.5 unva ú1.5–3 unva ú3 unva AST/ALT ratio (n Å 367) °1 ú1 Bilirubin (n Å 365) °17 mmol/L ú17 õ51 mmol/L Albumin (n Å 344) ¢35 g/L õ35 g/L g-Globulin (n Å 315) °20 g/L ú20 g/L Platelets (n Å 372) ¢130 1 109/L õ130 1 109/L Prothrombin time (n Å 343) ¢70% õ70%
P value
No. of patients
5 yr (SE)
10 yr (SE)
33 177 170
1.00 (0.00) 0.93 (0.02) 0.87 (0.03)
1.00 (0.00) 0.71 (0.10) 0.77 (0.05)
0.0285
261 106
0.94 (0.02) 0.85 (0.04)
0.87 (0.04) 0.61 (0.09)
0.0016
224 141
0.96 (0.01) 0.81 (0.04)
0.86 (0.05) 0.67 (0.07)
0.00001
309 35
0.93 (0.02) 0.80 (0.07)
0.80 (0.05) 0.64 (0.12)
0.0053
217 98
0.93 (0.02) 0.87 (0.04)
0.81 (0.06) 0.68 (0.08)
0.047
186 186
0.96 (0.02) 0.86 (0.03)
0.87 (0.04) 0.71 (0.07)
0.0009
230 113
0.93 (0.02) 0.87 (0.04)
0.92 (0.02) 0.53 (0.11)
0.0033
NOTE. Values in parentheses represent the total number of patients on which survival calculation is based because initial data were not available in all patients. P value derived from log rank test and Breslow test. a Multiplied by the upper limit of normal value (unv).
ences among studies in the duration of follow-up as well as to the inclusion of HCV-infected patients at different stages of liver disease, with the lowest mortality rates observed in studies tracing patients from acute hepatitis7,28 and the highest mortality rates reported in studies including patients with features of advanced liver diseases.22 It is of note that most of our patients are middle aged
at the time of diagnosis of compensated cirrhosis and have continued to do well with regard to their livers, thus explaining the finding of the relatively high proportion of deaths unrelated to liver disease (30%). In this cohort of patients, life expectancy is relatively long (91% probability of survival at 5 years). However, once clinical decompensation has occurred, the prognosis is rather poor (50% probability of survival at 5 years). Of interest is the observation that, at univariate analysis using the Kaplan–Meier method that corrects for differences in follow-up, the probability of survival was greater in the cirrhotic patients with a history of intravenous drug abuse or blood transfusion than in patients with unknown source of HCV infection. The reasons for the possible different prognosis according to the presumed
Table 4. Significant Prognostic Variables for Survival and Their Regression Coefficient
Figure 4. Cumulative probability of survival according to the presumed source of infection. A, Intravenous drug abuse; B, blood transfusion; C, occupational exposure to blood, major surgery, or family member with chronic liver disease; and D, unknown.
/ 5e18$$0016
01-13-97 17:45:57
gasa
Variable
b coefficient regression
SE
P valuea
Bilirubin Hepatic stigmata Age Platelets
0.0901 0.9072 0.0422 00.0072
0.0163 0.3446 0.0203 0.0038
0.00001 0.0083 0.0306 0.043
a
P value derived from the likelihood ratio test.
WBS-Gastro
470 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
presumed acquisition of HCV infection, did not influence survival rates, and this finding may be caused by the low number of events observed during the study period (6 liver-related deaths of 159 posttransfusion patients). It should be noted that, in our study population, some of the patients were cirrhotic when they first presented, and in these cases, there is no information on how long the cirrhosis had been present. However, in a proportion of patients, the diagnosis of cirrhosis was made during follow-up for chronic hepatitis, and in these cases, the time of development of cirrhosis may be more accurately known. Accordingly, at univariate analysis, the probability of survival was better in the latter group of patients most likely followed up since an earlier stage of cirrhosis compared with the former category of patients. Multivariate analysis by Cox’s model found that the only factors contributing significantly and independently to the estimation of prognosis were bilirubin, presence of hepatic stigmata, age, and platelets, all parameters indicating an advanced stage of cirrhosis. The fact that age is a prognostic variable is not surprising, probably reflecting a longer duration of liver disease. IFN-a therapy has been shown to induce sustained biochemical remission in 15%–20% of treated patients with chronic hepatitis associated with HCV and amelioration in the degree of hepatic inflammatory activity,30,31 but the presence of cirrhosis markedly reduces the rate of response.32 In this retrospective study, about half of the cirrhotic patients were treated with IFN-a with rates of biochemical remission similar to that reported in other
Figure 5. Cumulative probability of survival in (top) patients treated with IFN-a and (bottom) patients not treated with IFN-a. The difference in unadjusted survival curves between the two groups is significant (P Å 0.016) (top ). The difference in adjusted estimated survival curves, based on the Cox’s model for median values of bilirubin, age, platelets, and absent hepatic stigmata between the two groups is not significant (bottom ).
mode of acquisition of infection are unknown and need further investigation. In the subgroup of posttransfusion cirrhotics, the age at the time of the blood transfusion, i.e., the age of the
Table 5. Initial Demographic Data and Clinical Status of Patients With NANB/C Compensated Cirrhosis in Relation to IFN-a Therapy Group A (n Å 205)
Group B (n Å 179)
P value
119/86 53 { 0.6 79/201 (39) 54/198 (27) 126/161 (78) 51/191 (27) 27/92 (29)
104/75 56 { 0.7 87/172 (50) 59/167 (35) 111/138 (80) 73/168 (43) 52/102 (51)
NS 0.002 0.029 NS NS 0.0009 0.0022
{ { { { { { { {
NS 0.0003 0.0002 0.0029 0.017 0.0004 0.048 0.012
Patient characteristics Sex (M/F) Age (yr)a Symptoms (%)b Hepatic stigmata (%)b Liver firmness (%)b Splenomegaly (%)b Esophageal varices on endoscopy (%)b Serum parametersa AST (1 unv)c ALT (1 unv)c Bilirubin (mmol/L) Albumin (g/L) g-Globulin (g/L) Platelets (109/L) Prothrombin time (%) Follow-up (mo)a
3.3 4.7 15.8 42 17.7 146 77 57
{ { { { { { { {
0.14 0.21 0.5 0.3 0.4 4 1 2
3.3 3.6 19 40 19.3 125 74 65
Group A, IFN-a–treated patients; group B, patients not treated with IFN-a therapy. a Values are expressed as means { SE. b Values are expressed as numbers positive/total numbers of patients with analyzable data. c Multiplied by the upper limit of normal value (unv).
/ 5e18$$0016
01-13-97 17:45:57
gasa
WBS-Gastro
0.16 0.18 0.7 0.4 0.5 4 1 3
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 471
studies.32 At univariate analysis, patients treated with antiviral therapy had a more favorable prognosis compared with patients not treated with IFN-a. However, the prognosis in cirrhosis varies widely, being dependent on patient characteristics. The multivariate analysis was designed to adjust for differences between the two groups, those treated or not treated with IFN-a. The significant difference in survival curves for the two groups disappeared after adjustment for the influence of bilirubin, age, hepatic stigmata, and platelets, all factors independently related to survival. The main reasons for this effect are that IFN-a–treated patients were significantly younger and showed clinical and serological features at presentation reflecting a less advanced stage of cirrhosis; these differences were adjusted for by Cox’s model. To date, no studies of IFN-a therapy in chronic hepatitis caused by HCV have included the end point of improved survival rates, and we provide some information on this question in a subgroup of patients with wellcompensated cirrhosis. However, it should be pointed out that, although a powerful tool, statistical adjustment for the influence of prognostic variables cannot replace randomization and that a long-term study of randomized IFN-a–treated and control patients with similar clinical and serological features at entry is required to establish clinically significant efficacy of this treatment in white patients. In conclusion, our results contribute to defining the timing of appearance of HCC and decompensation in cirrhosis associated with HCV, showing a slowly progressive increase with the duration of the liver disease. This study indicates that the disease often has an indolent course, thus explaining the relatively long life expectancy observed in most patients.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17. 18. 19. 20. 21.
References 1. Realdi G, Alberti A, Rugge M, Rigoli AM, Tremolada F, Schivazappa L, Ruol A. Long-term follow-up of acute and chronic nonA, non-B post-transfusion hepatitis: evidence of progression to liver cirrhosis. Gut 1982;23:270–275. 2. Alter HJ, Hoofnagle JH. Non-A, non-B: observations on the first decade. In: Vyas GN, Dienstag JH, Hoofnagle JH, eds. Viral hepatitis and liver disease. Philadelphia: Grune & Stratton, 1984: 345–354. 3. Wejstal R, Lindberg J, Lundin P, Norkrans G. Chronic non-A, nonB hepatitis: a long-term follow-up study in 49 patients. Scand J Gastroenterol 1987;22:1115–1122. 4. Kiyosawa K, Sodeyama T, Tanaka E, Gibo Y, Yoshizawa K, Nakano Y, Furuta S, Akahane Y, Nishioka K, Purcell RH. Interrelationship of blood transfusion, non-A, non-B hepatitis and hepatocellular carcinoma: analysis by detection of antibody to hepatitis C virus. Hepatology 1990;12:671–675. 5. Di Bisceglie AM, Goodman ZD, Ishak KG, Hoofnagle JH, Melpolder JJ, Alter HJ. Long-term clinical and histopathological followup of chronic posttransfusion hepatitis. Hepatology 1991;14: 969–974. 6. Tremolada F, Casarin C, Alberti A, Drago C, Tagger A, Ribero
/ 5e18$$0016
7.
01-13-97 17:45:57
gasa
22.
23.
24.
25.
26.
27.
ML, Realdi G. Long-term follow-up of non-A, non-B (type C) posttransfusion hepatitis. J Hepatol 1992;16:273–281. Koretz RL, Abbey H, Coleman E, Gitnick G. Non-A, non-B posttransfusion hepatitis: looking back in the second decade. Ann Intern Med 1993;119:110–115. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology. A basic science for clinical medicine. Boston: Little, Brown, 1991. Infante-Rivard C, Villeneuve JP, Esnaola S. A framework for evaluating and conducting prognostic studies: an application to cirrhosis of the liver. J Clin Epidemiol 1989;42:791–805. Alter HJ, Purcell RH, Shih JW, Melpolder JC, Houghton M, Choo QL, Kuo G. Detection of antibody to hepatitis C virus in prospectively followed transfusion recipients with acute and chronic nonA, non-B hepatitis. N Engl J Med 1989;321:1494–1500. Tremolada F, Casarin C, Tagger A, Ribero ML, Realdi G, Alberti A, Ruol A. Antibody to hepatitis C virus in post-transfusion hepatitis. Ann Intern Med 1991;114:277–281. Jeffers LJ, Hasan F, De Medina M, Reddy R, Parker T, Silvas M, Mendez L, Schiff ER, Manns M, Houghton M, Choo Q-l, Kuo G. Prevalence of antibodies to hepatitis C virus among patients with cryptogenic chronic hepatitis and cirrhosis. Hepatology 1992; 15:187–190. Bianchi L, De Groote J, Desmet VJ, Gedik P, Korb G, Popper H, Poulsen H, Scheuer PJ, Thaler H, Wepler W. Acute and chronic hepatitis revisited. Lancet 1977;2:914–919. Sorensen TIA. Definition of death in relation to variceal bleeding. In: Burroughs AK, ed. Methodology and reviews of clinical trials in portal hypertension. Amsterdam: Excerpta Medica, 1987:31– 35. Kaplan EL, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc 1958;53:457–481. Peto R, Pike MC. Conservatism of the approximation (0-E)/E in the logrank test for survival data on tumor incidence data. Biometrics 1973;29:579–584. Cox DR. Regression models and life tables (with discussion). J R Stat Soc B 1972;34:187–220. Rao CR. Linear statistical inference and its application. New York: Wiley, 1973. Beale EML, Little RJA. Missing values in multivariate analysis. J R Stat Soc B 1975;37:129–145. Christensen E. Multivariate survival analysis using Cox’s regression model. Hepatology 1987;7:1346–1358. Dixon WH, ed. BMDP statistical software. Los Angeles: University of California, 1983. Tong MJ, El-Farra NS, Reikes AR, Co RL. Clinical outcomes after transfusion-associated hepatitis C. N Engl J Med 1995;332: 1463–1466. Colombo M, Kuo G, Choo QL, Donato MF, Del Ninno E, Tommasini MA, Dioguardi N, Houghton M. Prevalence of antibodies to hepatitis C virus in Italian patients with hepatocellular carcinoma. Lancet 1989;2:1006–1008. Bruix J, Barrera JM, Calvet X, Ercilla G, Costa J, Sanchez-Tapias JM, Vall M, Bru C, Rodes J. Prevalence of antibodies to hepatitis C virus in Spanish patients with hepatocellular carcinoma and hepatic cirrhosis. Lancet 1989;2:1004–1006. Colombo M, De Franchis R, Del Ninno E, Sangiovanni A, De Fazio C, Tommasini M, Donato MF, Piva A, Di Carlo V, Dioguardi N. Hepatocellular carcinoma in Italian patients with cirrhosis. N Engl J Med 1991;325:675–680. Nishiguchi S, Kuroki T, Nakatani S, Morimoto H, Takeda T, Nakajima S, Shiomi S, Seki S, Kobayashi K, Otani S. Randomized trial of effects of interferon alfa on incidence of hepatocellular carcinoma in chronic active hepatitis C with cirrhosis. Lancet 1995;346:1051–1055. Okuda K, Fujimoto I, Hanai A, Urano Y. Changing incidence of
WBS-Gastro
472 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
hepatocellular carcinoma in Japan. Cancer Res 1987;47:4967– 4972. 28. Nishioka K, Watanabe J, Furuta S, Tanaka E, Iino S, Suzuki H, Tsuji T, Yano M, Kuo G, Choo QL, Houghton M, Oda T. A high prevalence of antibody to the hepatitis C virus in patients with hepatocellular carcinoma in Japan. Cancer 1991;67:429–433. 29. Seef LB, Buskell-Bales Z, Wright EC, Durako SJ, Alter HJ, Iber FL, Hollinger FB, Gitnik G, Knodell RG, Perillo RP, Stevens CE, Hollingsworth CG, the National Heart, Lung, and Blood Institute Study Group. Long-term mortality after transfusion-associated non-A, non-B hepatitis. N Engl J Med 1992;327:1906–1911. 30. Saracco G, Rosina F, Abate ML, Chiandussi L, Gallo V, Cerutti E, Di Napoli A, Solinas A, Deplano A, Tocco A, Cossu P, Chien D, Kuo G, Polito A, Weiner AJ, Houghton M, Verme G, Bonino F, Rizzetto M. Long-term follow-up of patients with chronic hepatitis C treated with different doses of interferon-a2b. Hepatology 1993;18:1300–1305.
/ 5e18$$0016
01-13-97 17:45:57
gasa
31. Poynard T, Bedossa P, Chevallier M, Mathurin P, Lemonnier C, Trepo C, Couzigou P, Payen JL, Sajus M, Costa JM, Vidaud M, Chaput JC, the Multicenter Study Group. A comparison of three interferon alfa-2b regimens for the long-term treatment of chronic non-A, non-B hepatitis. N Engl J Med 1995;332:1457–1462. 32. Jouet P, Roudot-Thoraval F, Dhumeaux D, Me`treau JM, Le Group Francais Pour L’Etude Du Traitment Des He`patitis Chroniques NANB/C. Comparative efficacy of interferon alfa in cirrhotic and noncirrhotic patients with non-A, non-B, C hepatitis. Gastroenterology 1994;106:686–690.
Received December 26, 1995. Accepted October 8, 1996. Address requests for reprints to: Giuseppe Realdi, M.D., Istituto di Clinica Medica, University of Sassari, Viale San Pietro 8, 07100 Sassari, Italy.
WBS-Gastro
Morbidity and Mortality in Compensated Cirrhosis Type C: A Retrospective Follow-up Study of 384 Patients GIOVANNA FATTOVICH,* GIULIANO GIUSTINA,‡ FRANCOISE DEGOS,§ FEDERICO TREMOLADA,‡ GIULIO DIODATI,‡ PIERO ALMASIO,x FREDERIK NEVENS,Ø ANTONIO SOLINAS,# DANIELA MURA,** JOHANNES T. BROUWER,‡‡ HOWARD THOMAS,§§ CORINNE NJAPOUM,§ CARLA CASARIN,‡ PAOLA BONETTI,‡ PATRIZIA FUSCHI,x JOVAN BASHO,Ø ANDREINA TOCCO,# ABHA BHALLA,‡‡ RACHEL GALASSINI,§§ FRANCO NOVENTA,‡ SOLKO W. SCHALM,‡‡ and GIUSEPPE REALDI** *Istituto di Semeiotica e Nefrologia Medica, University of Verona, Verona, Italy; ‡Istituto di Medicina Clinica, Clinica Medica 2a, Universita` di Padova, Padova, Italy; §Fe`de`ration d’He`patogastroente`rologie et INSERM Unite´ 24, Hopital Beaujon, Paris, France; xCattedra di Clinica Medica R, University of Palermo, Palermo, Italy; ØInternal Medicine, University Hospital Gasthuisberg, Leuven, Belgium; #Istituto di Patologia Speciale Medica and **Istituto di Clinica Medica, University of Sassari, Sassari, Italy; ‡‡Internal Medicine II, University Hospital Rotterdam, Rotterdam, The Netherlands; and §§Department of Medicine, St. Mary’s Hospital Medical School, London, England
See editorial on page 651. Background & Aims: Few data are available concerning the long-term prognosis of chronic liver disease associated with hepatitis C virus infection. This study examined the morbidity and survival of patients with compensated cirrhosis type C. Methods: A cohort of 384 European cirrhotic patients was enrolled at seven tertiary referral hospitals and followed up for a mean period of 5 years. Inclusion criteria were biopsy-proven cirrhosis, abnormal serum aminotransferase levels, absence of complications of cirrhosis, and exclusion of hepatitis A and B viruses and of metabolic, toxic, or autoimmune liver diseases. Results: Antibodies against hepatitis C virus were positive in 98% of 361 patients tested. The 5-year risk of hepatocellular carcinoma was 7% and that of decompensation was 18%. Death occurred in 51 patients (13%), with 70% dying of liver disease. Survival probability was 91% and 79% at 5 and 10 years, respectively. Two hundred five patients (53%) were treated with interferon alfa. After adjustment for clinical and serological differences at baseline between patients treated or not treated with interferon, the 5-year estimated survival probability was 96% and 95% for treated and untreated patients, respectively. Conclusions: In this cohort of patients, life expectancy is relatively long, in agreement with the morbidity data showing a slowly progressive disease.
about 8%.6 However, few data are available concerning the clinical course of compensated cirrhosis associated with hepatitis C, and at present, the question is how frequently serious complications of cirrhosis occur. The rate of development of decompensation or hepatocellular carcinoma (HCC) and the degree to which they contribute to the mortality rate are not well known. Moreover, there is uncertainty about the long-term mortality rate of chronic hepatitis caused by hepatitis C virus (HCV). On the basis of these considerations, the primary objective of this study was to evaluate the morbidity and survival of histologically documented cirrhosis associated with HCV in patients who had never had clinical manifestations of decompensation, such as jaundice, ascites, variceal bleeding, or encephalopathy, but who had biochemical signs of disease activity. Morbidity was studied in relation to the incidence of HCC and the rate of appearance of hepatic decompensation. The knowledge of the clinical course and prognosis of this clinical condition is of particular importance in clinical practice and can influence medical decision making, such as rational application of therapeutic measures.8 The outcome of the disease was assessed retrospectively in a cohort of Western European patients who were referred to tertiary-care centers and followed up longitudinally. We also analyzed the serological and clinical data of these patients as prognostic factors for survival.
Patients and Methods
T
he prognosis of non-A, non-B (NANB)/C chronic hepatitis is a matter of controversy. It is well documented that NANB/C chronic hepatitis may progress to cirrhosis,1 – 7 and the risk of developing cirrhosis has been reported to be 40% at 5 years and 60% at 8 years, as calculated by actuarial analysis, with an annual rate of / 5e18$$0016
01-13-97 17:45:57
gasa
The study was planned within the European Concerted Action on Viral Hepatitis (EUROHEP) and involved seven Abbreviation used in this paper: IFN-a, interferon alfa. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
464 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
university hospitals in Europe that serve a tertiary referral function. The data were to be collected according to a protocol that followed basic epidemiological principles required for conducting prognostic studies,8,9 with clear definitions of the study population, time of entry, referral pattern, and follow up.
Study Population Several studies have shown that most patients with NANB chronic liver disease have HCV infection.10 – 12 Therefore, it was decided to include patients with NANB hepatitis in the study and to retrospectively test stored serum for HCV. The study included all consecutive patients admitted to the participating centers between January 1982 and December 1992 and who fulfilled the following criteria: (1) biopsy-proven cirrhosis according to international criteria13; (2) persistent elevations of alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) concentrations (at least 1.5 times the upper limit of normal) for at least 6 months before entry into the study; (3) absence of hepatitis B surface antigen and all other potential causes of chronic liver disease; (4) no history or clinical evidence at enrollment of complications of cirrhosis, i.e., ascites, variceal bleeding, encephalopathy, or jaundice; (5) no evidence of HCC at entry into the study on the basis of ultrasonography, a-fetoprotein levels (õ400 mg/L), and/or histology; and (6) follow-up at the enrolling center for a minimum period of 6 months. Patients with other forms of liver disease were excluded by history, serological parameters, and histology. Patients with a history of alcohol abuse, defined as alcohol intake of ¢80 g/day for more than 5 years, were also excluded from the study. Information on serum anti-HCV antibodies was obtained, but anti-HCV testing was not considered an inclusion criterion for eligibility into the study.
Time of Entry Into the Study and Referral Pattern Entry into the study was defined as the time of diagnosis of NANB/C compensated cirrhosis. The four referral patterns to the enrolling center were defined as follows: incidental finding of elevated aminotransferase levels; presence of nonspecific symptoms; diagnosis of cirrhosis made at the enrolling center during follow-up for NANB/C chronic hepatitis (in this case, the time of entry into the study was established as the time of histological diagnosis of cirrhosis); and referral to the enrolling center after the histological diagnosis of cirrhosis was made elsewhere previously (in this case, the start of follow-up was established as the time of referral to the enrolling center).
Follow-up The time of observation was calculated from the date of entry until death or the end of the observation period (June 30, 1993). All patients underwent clinical assessment and standard liver biochemical tests at least once a year or more frequently if indicated. Death was classified as caused by liver failure if progressive impairment of liver function occurred, or if it occurred within 40 days of variceal bleeding, regardless
/ 5e18$$0016
01-13-97 17:45:57
gasa
of the severity,14 or if it was a result of HCC. Patients who died of conditions not related to liver disease and patients who were lost to follow-up were censored at the time of death or at the time of dropout, respectively, in the statistics.15 HCC was diagnosed on the basis of histological findings and/or elevated a-fetoprotein levels (ú400 mg/L) and ultrasonographic findings consistent with HCC. Decompensation was defined as the appearance of at least one episode of ascites, jaundice, hepatic encephalopathy, or gastrointestinal bleeding of variceal origin. Biochemical remission was defined as the normalization of serum aminotransferase values on at least two consecutive determinations obtained at least 3 weeks apart and persisting at the time of last observation. A response to interferon alfa (IFN-a) therapy was considered as the normalization of aminotransferase levels by the end of treatment and persisting at the time of last observation. The actual total dose of IFN-a was defined as low when õ200 million units was administered and as medium/high when values greater than this limit were administered.
Laboratory Tests Results of anti-HCV tests were obtained from medical records of patients, or anti-HCV testing was performed on stored serum samples by a second-generation enzyme-linked immunosorbent assay according to the manufacturer’s instructions.
Prognostic Factors The following clinical and serological variables at entry were evaluated as potential prognostic factors for survival: sex, age, referral pattern, presumed source of infection, age at the time of transfusion (õ50 years or ¢50 years), alcohol consumption, symptoms, hepatic stigmata, liver firmness, splenomegaly, esophageal varices, AST, ALT, AST/ALT ratio, bilirubin, albumin, prothrombin time, g-globulin, platelets, and a-fetoprotein. Moreover, the prognostic value for survival of IFN-a therapy was analyzed. Liver firmness was defined as an increased consistency of the liver from firm to rock hard. Splenomegaly was documented by physical examination and/or ultrasonography (ú13 cm in longitudinal diameter).
Statistical Analysis The x2 test and Student’s t test were used to compare frequencies and means. The cumulative probability of survival and of developing HCC or decompensation was calculated using the method of Kaplan and Meier.15 Univariate analysis of survival was performed by computing survival curves according to the Kaplan–Meier method. In the univariate analysis of survival between patients treated with IFN-a and those not treated with IFN-a therapy, the time zero for IFN-a–treated patients was the time of starting treatment. Curves were compared statistically using the log rank test16 and Breslow test. Multivariate analysis of prognostic factors for survival was also studied by stepwise forward Cox regression model.17 In this analysis, IFN-a therapy during follow-up was evaluated as a
WBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 465
time-dependent variable, taking into account the time of starting treatment. In this model, the order of insertion of any variable was determined by using the maximum log-likehood value, and the statistical significance was assessed by the likelihood ratio test.18 Because the analysis requires that all patients are represented by a complete set of variables and to not reduce the number of patients, missing data were replaced by neutral estimates.19 The estimated curves of the probability of survival in the two subgroups of patients treated with IFN-a and those not treated with antiviral therapy were calculated according to a statistical method that adjusted for the influence of significant variables as found by Cox model.20 The statistical software used was BMDP.21
Results A total of 384 patients fulfilling the inclusion criteria was enrolled at seven European centers from 1982 to 1992 and followed up for a mean period of 61 months (range, 6–153 months). Sixteen patients (4%) were lost to follow-up. Patient Characteristics The clinical and serological features of the patients at entry are shown in Table 1. The patients ranged in age from 19 to 78 years (mean age, 54 years). There were 223 men and 161 women; 370 of the patients were white, 11 were Asian, and 3 were black. At entry, 138 patients had a history of occasional or daily alcohol intake of õ80 g, whereas the remaining 246 patients denied any alcohol consumption. At presentation, nonspecific symptoms such as dyspepsia, asthenia, and/or upper abdominal discomfort were present in 166 patients (44%). However, only 60 patients were referred to the enrolling center because of these nonspecific symptoms. On physical examination, liver firmness was detected in 79% of 299 patients whose liver was palpable. Hepatic stigmata and splenomegaly were detected in 31% and 34% of the patients, respectively. Histological activity on liver biopsy was present in the large majority of the patients (95%). Of the 75 patients in whom the histological diagnosis of cirrhosis was made elsewhere before being referred to the enrolling center, 38 had undergone the liver biopsy within 12 months (range, 1–12 months) of enrollment, whereas 37 had a histological diagnosis of cirrhosis from 2 to 17 years before entry into the study. The baseline serological features showed that most patients had normal values of bilirubin (61%), albumin (90%), g-globulin (69%), and prothrombin time (67%). None of the 128 patients tested was anti–human immunodeficiency virus positive. Three hundred fifty-five of the 361 patients tested (98%) were positive for serum antibody to HCV. Of the 384 patients enrolled in the study, 159 (41%) / 5e18$$0016
01-13-97 17:45:57
gasa
Table 1. Baseline Clinical and Serological Features in Patients With NANB/C Compensated Cirrhosis Agea (yr ) (n Å 384) Men (%) (n Å 384) Referral pattern (%) (n Å 384) Diagnosis of cirrhosis during follow-up for chronic hepatitis Incidental finding of abnormal transaminase levels Diagnosis of cirrhosis made elsewhere previously Symptoms Histology (%) (n Å 384) Active cirrhosis Inactive cirrhosis AST (%) (n Å 380) °1.5 unv ú1.5–3 unv ú3 unv ALT (%) (n Å 371) °1.5 unv ú1.5–3 unv ú3 unv Bilirubin (%) (n Å 365) °17 mmol/L ú17 õ51 mmol/L Albumin (%) (n Å 344) ¢35 g/L õ35 g/L g-Globulin (%) (n Å 315) °20 g/L ú20 g/L Platelets (%) (n Å 372) ú130 1 109/L °130 1 109/L Prothrombin time (%) (n Å 343) ¢70% õ70% a-Fetoprotein (%) (n Å 200) °10 mg/L ú10 mg/L
54 { 4.93 223 (58)
130 119 75 60
(34) (31) (19) (16)
363 (95) 21 (5) 33 (9) 177 (46) 170 (45) 22 (6) 126 (34) 223 (60) 224 (61) 141 (39) 309 (90) 35 (10) 217 (69) 98 (31) 181 (49) 191 (51) 230 (67) 113 (33) 114 (57) 86 (43)
NOTE. n values are the total numbers of patients with analyzable data. unv, upper limit of normal value. a Value is expressed as mean { SD.
had a history of blood transfusion, 9 (2%) had a history of intravenous drug abuse, 11 (3%) had occupational exposure to blood, 63 (16%) had a history of major surgery, 29 (8%) had a family member with chronic liver disease, and 113 patients (30%) had no risk factors for infection. Of the 159 patients with a history of blood transfusion, 14 had received multiple transfusions. In 125 of the remaining 145 patients, the year of the blood transfusion was documented according to the patient’s recollection, and the mean age at the time of the transfusion was 43 years (range, 8–66 years). A total of 81 patients received the transfusion before the age of 50 years (mean age, 35 years) and 44 at the age of 50 years or older (mean age, 56 years). The mean interval between the time of the blood transfusion and the diagnosis of cirrhosis was 12 WBS-Gastro
466 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
Figure 1. Cumulative probability of HCC appearance in patients with NANB/C compensated cirrhosis. The 5-year appearance rate was 7%, and the 10-year appearance rate was 14%.
years (range, 1–43 years) among all patients, and 5.7 and 15.8 years, respectively, for patients who were 50 years or older at the time of transfusion or who were younger than 50 years. Therapy During follow-up, 205 patients (53.5%) were treated with IFN-a, 2 (0.5%) with ribavirin, and 19 (5%) with steroids, whereas the remaining 158 patients (41%) remained untreated. The majority of patients treated with IFN-a received a medium/high dose of IFN-a (181 of 205 patients; 88%). A response to IFN-a therapy with biochemical remission lasting for 5–70 months (mean, 43 months) after stopping treatment was observed in 18 of 205 patients (9%). Persistent biochemical remission for 5–126 months (mean, 36 months) occurred in an additional 12 patients, namely, 7 of 158 untreated patients (5%) and 5 patients at variable intervals after stopping IFN-a treatment. None of the patients treated with steroids showed sustained aminotransferase level normalization.
(72%) were men. At diagnosis, patients were between 44 and 72 years of age (mean age, 63 years). Twentyfour of the 29 patients were tested for antibody to HCV and were antibody to HCV positive. Seventeen of the 29 patients died 1–19 months after the diagnosis of HCC, whereas 12 patients were still alive at the end of the observation. Development of decompensation. Of the 355 patients who remained tumor free, 65 (18%) developed at least one episode of ascites, jaundice, hepatic encephalopathy, or variceal bleeding. The mean interval between the time of entry into the study and the appearance of the first episode of decompensation was 37 months (range, 3–137 months). In the remaining 290 patients, the cirrhosis remained compensated during follow-up. The first episode of decompensation was caused by ascites, variceal bleeding, encephalopathy, or jaundice in 31, 14, 5, and 4 patients, respectively, and by more than one complication in the remaining 11 patients. The cumulative probability of developing decompensated cirrhosis after diagnosis was 12% at 3 years, 18% at 5 years, and 29% at 10 years with a yearly incidence of 3.9% during the first 5 years (Figure 2). Survival. During follow-up, 51 patients (13%) died, including 3 of the 6 patients who underwent liver transplantation. Complications related to HCC accounted for 33% of the deaths (17 cases), liver failure for 31% (16 cases), and bleeding for 6% (3 cases). Fifteen patients (30%) died of causes unrelated to cirrhosis. The rate of liver-related mortality during the course of the study was 9% (36 of 384 patients). The mean length of time from entry into the study to death from liver disease was 50 months (range, 9–124 months). The probability of survival after diagnosis of compensated cirrhosis was
Morbidity Development of HCC. During observation, HCC developed in 29 of the 384 patients (8%). Liver cancer was diagnosed in 20 patients on histology and in 9 patients on the basis of elevated a-fetoprotein values (ú400 mg/L) and compatible ultrasonography. The mean interval between the time of entry into the study and the diagnosis of HCC was 48 months (range, 7–134 months). The cumulative probability of HCC appearance among the patients was 4% at 3 years after diagnosis of cirrhosis, 7% at 5 years when 157 patients were still in followup, and 14% at 10 years with a yearly incidence of 1.4% during the first 5 years (Figure 1). Twenty-one patients
/ 5e18$$0016
01-13-97 17:45:57
gasa
Figure 2. Cumulative probability of developing decompensation in 355 patients with NANB/C compensated cirrhosis who remained tumor free. The 5-year probability rate was 18%, and the 10-year probability rate was 29%.
WBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 467
Figure 3. Curve A represents the probability of survival after diagnosis in all patients studied with NANB/C compensated cirrhosis. The 5year probability rate was 91%. Curve B represents the probability of survival after the appearance of the first major complication of the disease in 65 patients who developed decompensated cirrhosis. The 5-year probability rate was 50%.
5, top). However, multivariate analysis by Cox’s model failed to show any independent prognostic significance of IFN-a therapy for survival rates (Table 4). The difference in adjusted estimated survival curves, based on the Cox’s model for median values of bilirubin, age, platelets, and hepatic stigmata, between the two groups was not significant (Figure 5, bottom). The 5year adjusted estimated survival probability was 96% for IFN-a–treated patients and 95% for patients not receiving IFN-a. As shown in Table 5, IFN-a–treated patients were significantly younger and their illness was significantly less severe with fewer findings of splenomegaly or esophageal varices; with less elevated bilirubin and g-globulin levels; with higher albumin and platelet levels; and with greater prothrombin time compared with patients not receiving IFN-a.
Discussion 96%, 91%, and 79% at 3, 5, and 10 years, respectively (Figure 3). The yearly incidence of mortality was 1.9 during the first 5 years. The probability of survival after the onset of the first major complication of the disease was 50% at 5 years (Figure 3). Factors correlating with survival. Among the clinical and biochemical features studied, 16 variables significantly predicted survival at univariate analysis: age, referral pattern, source of infection, alcohol consumption, symptoms, hepatic stigmata on physical examination, splenomegaly, esophageal varices, AST, AST/ALT ratio, bilirubin, albumin, g-globulin, platelets, prothrombin time, and interferon-a therapy (Tables 2 and 3). The probability of survival was significantly different according to the presumed source of infection, with a better survival observed in patients with a history of intravenous drug abuse or blood transfusion (Table 2 and Figure 4). Among posttransfusion cirrhotic patients, at univariate analysis, the probability of survival did not differ between patients who had been administered the transfusion before the age of 50 years and those who had the transfusion at age 50 years or older (P Å 0.67). After multivariate analysis with Cox’s model, bilirubin, hepatic stigmata on physical examination, age, and platelets remained as factors independently related to survival (Table 4). Survival in Relation to Antiviral Therapy At univariate analysis, the probability of survival was significantly different (P Å 0.016) between the two subgroups of patients, those treated with IFN-a and those not receiving this drug, with a better survival rate observed in IFN-a–treated patients (Table 2 and Figure / 5e18$$0016
01-13-97 17:45:57
gasa
Chronic HCV infection is a well-known cause of chronic liver disease both in patients with histories of exposure to parenterally transmitted viral diseases10,11 and in patients who have no recognizable source of infection.12 In this study, 98% of patients with posttransfusion or cryptogenic cirrhosis had detectable antibody to HCV, thus confirming the high prevalence of antibody to HCV in these patients and indicating that chronic HCV infection is a very important cause of liver cirrhosis in Europe. The long-term outcome of this chronic condition has not yet been fully investigated. Previous longitudinal studies have only included a small number of cirrhotic patients5 – 7 and/or have not distinguished between compensated and decompensated cirrhosis.22 In this study, we evaluated retrospectively the clinical evolution of a large cohort of predominantly white Western European patients with compensated cirrhosis caused by HCV who were followed up for a mean period of 5 years with a low dropout rate. The main findings regard frequency and timing of HCC, decompensation, and mortality. Chronic HCV infection is considered one of the major causes of liver cancer in Western countries. Previous cross-sectional studies have shown a markedly high prevalence of antibody to HCV in patients with HCC in Europe.23,24 Follow-up studies of posttransfusion hepatitis have documented the sequential development of cirrhosis and HCC in some patients.3,4,6 However, the magnitude of the risk for HCC has not yet been quantified in chronic HCV infection. In the current study, HCC developed in 8% of 384 Western European patients with compensated cirrhosis type C with a yearly incidence of 1.4%. Other factors promoting liver cancer, such as alcoWBS-Gastro
468 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
Table 2. Clinical Variables Showing Prognostic Significance for Survival at Univariate Analysis Probability of survival Variable
P value
No. of patients
5 yr (SE)
10 yr (SE)
166 218
0.95 (0.02) 0.88 (0.03)
0.83 (0.06) 0.75 (0.06)
0.028a
130 119 75 60
0.94 0.94 0.87 0.85
(0.02) (0.02) (0.04) (0.05)
0.81 0.83 0.82 0.73
(0.07) (0.06) (0.06) (0.09)
0.019b
9 159 103 113
1.00 0.94 0.90 0.89
(0.00) (0.02) (0.03) (0.03)
— 0.94 (0.02) 0.76 (0.08) 0.66 (0.09)
0.019c
138 246
0.88 (0.03) 0.96 (0.06)
0.77 (0.06) 0.79 (0.06)
0.0439b
166 207
0.87 (0.03) 0.95 (0.02)
0.67 (0.08) 0.90 (0.03)
0.012a
113 252
0.81 (0.04) 0.96 (0.01)
0.71 (0.06) 0.83 (0.06)
0.001a
124 235
0.85 (0.04) 0.95 (0.02)
0.70 (0.08) 0.85 (0.05)
0.0062a
79 115
0.77 (0.06) 0.95 (0.02)
0.53 (0.12) 0.85 (0.07)
0.00001a
205 179
0.94 (0.02) 0.87 (0.03)
— 0.71 (0.06)
0.016a
Age (yr ) (n Å 384) °54 ú54 Referral pattern (n Å 384) Diagnosis of cirrhosis during follow-up for chronic hepatitis Incidental finding of abnormal transaminase levels Diagnosis of cirrhosis made elsewhere previously Symptoms Source of infection (n Å 384) Intravenous drug abuse Blood transfusion Otherd Unknown Alcohol intake (õ80 g/day) (n Å 384) Yes No Symptoms (n Å 373) Present Absent Hepatic stigmata (n Å 365) Present Absent Splenomegaly (n Å 359) Present Absent Esophageal varices (n Å 194) Present Absent IFN-a therapy (n Å 384) Yes No
NOTE. Values in parentheses represent the total number of patients on which survival calculation is based because initial data were not available in all patients. a P value derived from log rank test and Breslow test. b P value significant by Breslow test. c P value significant by log rank test. d Occupational exposure to blood, major surgery, or family member with chronic liver disease.
hol abuse or hepatitis B virus coinfection, were excluded in all cases. In a study conducted in Italy analyzing the prevalence of HCC among 417 patients with compensated cirrhosis of viral and nonviral origin, liver cancer developed in 4% of 189 HCV-related cases.25 This lower prevalence of HCC is most likely caused by the shorter follow-up of the Italian patients, averaging 33 months, compared with a mean follow-up of 61 months in our series. The rate of liver cancer may be greater for a longer period. Indeed, in our cirrhotic patients, the risk of HCC increased progressively. At 3 and 5 years, the probability of liver cancer appearance was 4% and 7%, respectively. A recent study conducted in Japan has reported a greater 5-year probability of liver cancer occurrence of approximately 25% in untreated Asian patients with compensated cirrhosis type C.26 Liver cancer associated with HCV infection appears to be increasing in various / 5e18$$0016
01-13-97 17:45:57
gasa
countries, particularly in Japan,27,28 suggesting geographic variation in the incidence of HCC that may relate to viral strain, genetic factors, or yet unknown environmental factors. Interestingly, most patients did not seem to have evidence of rapidly progressive liver disease, and our data indicate that a high proportion of patients with compensated cirrhosis do well for at least 10 years. This finding together with the absence of subjective complaints in most patients suggest that the disease has a relatively long subclinical course until features of overt decompensation or cancer develop. In our series, the liver-related mortality rate was 9% during a mean follow-up of 5 years, and HCC and liver failure were the main causes of death. In other reports, the overall liver-related mortality rate varies from 2.5% to 14%.5 – 7,22,29 This variability may be related to differWBS-Gastro
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 469
Table 3. Serological Variables Showing Prognostic Significance for Survival at Univariate Analysis Probability of survival Variable AST (n Å 380) °1.5 unva ú1.5–3 unva ú3 unva AST/ALT ratio (n Å 367) °1 ú1 Bilirubin (n Å 365) °17 mmol/L ú17 õ51 mmol/L Albumin (n Å 344) ¢35 g/L õ35 g/L g-Globulin (n Å 315) °20 g/L ú20 g/L Platelets (n Å 372) ¢130 1 109/L õ130 1 109/L Prothrombin time (n Å 343) ¢70% õ70%
P value
No. of patients
5 yr (SE)
10 yr (SE)
33 177 170
1.00 (0.00) 0.93 (0.02) 0.87 (0.03)
1.00 (0.00) 0.71 (0.10) 0.77 (0.05)
0.0285
261 106
0.94 (0.02) 0.85 (0.04)
0.87 (0.04) 0.61 (0.09)
0.0016
224 141
0.96 (0.01) 0.81 (0.04)
0.86 (0.05) 0.67 (0.07)
0.00001
309 35
0.93 (0.02) 0.80 (0.07)
0.80 (0.05) 0.64 (0.12)
0.0053
217 98
0.93 (0.02) 0.87 (0.04)
0.81 (0.06) 0.68 (0.08)
0.047
186 186
0.96 (0.02) 0.86 (0.03)
0.87 (0.04) 0.71 (0.07)
0.0009
230 113
0.93 (0.02) 0.87 (0.04)
0.92 (0.02) 0.53 (0.11)
0.0033
NOTE. Values in parentheses represent the total number of patients on which survival calculation is based because initial data were not available in all patients. P value derived from log rank test and Breslow test. a Multiplied by the upper limit of normal value (unv).
ences among studies in the duration of follow-up as well as to the inclusion of HCV-infected patients at different stages of liver disease, with the lowest mortality rates observed in studies tracing patients from acute hepatitis7,28 and the highest mortality rates reported in studies including patients with features of advanced liver diseases.22 It is of note that most of our patients are middle aged
at the time of diagnosis of compensated cirrhosis and have continued to do well with regard to their livers, thus explaining the finding of the relatively high proportion of deaths unrelated to liver disease (30%). In this cohort of patients, life expectancy is relatively long (91% probability of survival at 5 years). However, once clinical decompensation has occurred, the prognosis is rather poor (50% probability of survival at 5 years). Of interest is the observation that, at univariate analysis using the Kaplan–Meier method that corrects for differences in follow-up, the probability of survival was greater in the cirrhotic patients with a history of intravenous drug abuse or blood transfusion than in patients with unknown source of HCV infection. The reasons for the possible different prognosis according to the presumed
Table 4. Significant Prognostic Variables for Survival and Their Regression Coefficient
Figure 4. Cumulative probability of survival according to the presumed source of infection. A, Intravenous drug abuse; B, blood transfusion; C, occupational exposure to blood, major surgery, or family member with chronic liver disease; and D, unknown.
/ 5e18$$0016
01-13-97 17:45:57
gasa
Variable
b coefficient regression
SE
P valuea
Bilirubin Hepatic stigmata Age Platelets
0.0901 0.9072 0.0422 00.0072
0.0163 0.3446 0.0203 0.0038
0.00001 0.0083 0.0306 0.043
a
P value derived from the likelihood ratio test.
WBS-Gastro
470 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
presumed acquisition of HCV infection, did not influence survival rates, and this finding may be caused by the low number of events observed during the study period (6 liver-related deaths of 159 posttransfusion patients). It should be noted that, in our study population, some of the patients were cirrhotic when they first presented, and in these cases, there is no information on how long the cirrhosis had been present. However, in a proportion of patients, the diagnosis of cirrhosis was made during follow-up for chronic hepatitis, and in these cases, the time of development of cirrhosis may be more accurately known. Accordingly, at univariate analysis, the probability of survival was better in the latter group of patients most likely followed up since an earlier stage of cirrhosis compared with the former category of patients. Multivariate analysis by Cox’s model found that the only factors contributing significantly and independently to the estimation of prognosis were bilirubin, presence of hepatic stigmata, age, and platelets, all parameters indicating an advanced stage of cirrhosis. The fact that age is a prognostic variable is not surprising, probably reflecting a longer duration of liver disease. IFN-a therapy has been shown to induce sustained biochemical remission in 15%–20% of treated patients with chronic hepatitis associated with HCV and amelioration in the degree of hepatic inflammatory activity,30,31 but the presence of cirrhosis markedly reduces the rate of response.32 In this retrospective study, about half of the cirrhotic patients were treated with IFN-a with rates of biochemical remission similar to that reported in other
Figure 5. Cumulative probability of survival in (top) patients treated with IFN-a and (bottom) patients not treated with IFN-a. The difference in unadjusted survival curves between the two groups is significant (P Å 0.016) (top ). The difference in adjusted estimated survival curves, based on the Cox’s model for median values of bilirubin, age, platelets, and absent hepatic stigmata between the two groups is not significant (bottom ).
mode of acquisition of infection are unknown and need further investigation. In the subgroup of posttransfusion cirrhotics, the age at the time of the blood transfusion, i.e., the age of the
Table 5. Initial Demographic Data and Clinical Status of Patients With NANB/C Compensated Cirrhosis in Relation to IFN-a Therapy Group A (n Å 205)
Group B (n Å 179)
P value
119/86 53 { 0.6 79/201 (39) 54/198 (27) 126/161 (78) 51/191 (27) 27/92 (29)
104/75 56 { 0.7 87/172 (50) 59/167 (35) 111/138 (80) 73/168 (43) 52/102 (51)
NS 0.002 0.029 NS NS 0.0009 0.0022
{ { { { { { { {
NS 0.0003 0.0002 0.0029 0.017 0.0004 0.048 0.012
Patient characteristics Sex (M/F) Age (yr)a Symptoms (%)b Hepatic stigmata (%)b Liver firmness (%)b Splenomegaly (%)b Esophageal varices on endoscopy (%)b Serum parametersa AST (1 unv)c ALT (1 unv)c Bilirubin (mmol/L) Albumin (g/L) g-Globulin (g/L) Platelets (109/L) Prothrombin time (%) Follow-up (mo)a
3.3 4.7 15.8 42 17.7 146 77 57
{ { { { { { { {
0.14 0.21 0.5 0.3 0.4 4 1 2
3.3 3.6 19 40 19.3 125 74 65
Group A, IFN-a–treated patients; group B, patients not treated with IFN-a therapy. a Values are expressed as means { SE. b Values are expressed as numbers positive/total numbers of patients with analyzable data. c Multiplied by the upper limit of normal value (unv).
/ 5e18$$0016
01-13-97 17:45:57
gasa
WBS-Gastro
0.16 0.18 0.7 0.4 0.5 4 1 3
February 1997
MORBIDITY AND MORTALITY IN CIRRHOSIS C 471
studies.32 At univariate analysis, patients treated with antiviral therapy had a more favorable prognosis compared with patients not treated with IFN-a. However, the prognosis in cirrhosis varies widely, being dependent on patient characteristics. The multivariate analysis was designed to adjust for differences between the two groups, those treated or not treated with IFN-a. The significant difference in survival curves for the two groups disappeared after adjustment for the influence of bilirubin, age, hepatic stigmata, and platelets, all factors independently related to survival. The main reasons for this effect are that IFN-a–treated patients were significantly younger and showed clinical and serological features at presentation reflecting a less advanced stage of cirrhosis; these differences were adjusted for by Cox’s model. To date, no studies of IFN-a therapy in chronic hepatitis caused by HCV have included the end point of improved survival rates, and we provide some information on this question in a subgroup of patients with wellcompensated cirrhosis. However, it should be pointed out that, although a powerful tool, statistical adjustment for the influence of prognostic variables cannot replace randomization and that a long-term study of randomized IFN-a–treated and control patients with similar clinical and serological features at entry is required to establish clinically significant efficacy of this treatment in white patients. In conclusion, our results contribute to defining the timing of appearance of HCC and decompensation in cirrhosis associated with HCV, showing a slowly progressive increase with the duration of the liver disease. This study indicates that the disease often has an indolent course, thus explaining the relatively long life expectancy observed in most patients.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17. 18. 19. 20. 21.
References 1. Realdi G, Alberti A, Rugge M, Rigoli AM, Tremolada F, Schivazappa L, Ruol A. Long-term follow-up of acute and chronic nonA, non-B post-transfusion hepatitis: evidence of progression to liver cirrhosis. Gut 1982;23:270–275. 2. Alter HJ, Hoofnagle JH. Non-A, non-B: observations on the first decade. In: Vyas GN, Dienstag JH, Hoofnagle JH, eds. Viral hepatitis and liver disease. Philadelphia: Grune & Stratton, 1984: 345–354. 3. Wejstal R, Lindberg J, Lundin P, Norkrans G. Chronic non-A, nonB hepatitis: a long-term follow-up study in 49 patients. Scand J Gastroenterol 1987;22:1115–1122. 4. Kiyosawa K, Sodeyama T, Tanaka E, Gibo Y, Yoshizawa K, Nakano Y, Furuta S, Akahane Y, Nishioka K, Purcell RH. Interrelationship of blood transfusion, non-A, non-B hepatitis and hepatocellular carcinoma: analysis by detection of antibody to hepatitis C virus. Hepatology 1990;12:671–675. 5. Di Bisceglie AM, Goodman ZD, Ishak KG, Hoofnagle JH, Melpolder JJ, Alter HJ. Long-term clinical and histopathological followup of chronic posttransfusion hepatitis. Hepatology 1991;14: 969–974. 6. Tremolada F, Casarin C, Alberti A, Drago C, Tagger A, Ribero
/ 5e18$$0016
7.
01-13-97 17:45:57
gasa
22.
23.
24.
25.
26.
27.
ML, Realdi G. Long-term follow-up of non-A, non-B (type C) posttransfusion hepatitis. J Hepatol 1992;16:273–281. Koretz RL, Abbey H, Coleman E, Gitnick G. Non-A, non-B posttransfusion hepatitis: looking back in the second decade. Ann Intern Med 1993;119:110–115. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology. A basic science for clinical medicine. Boston: Little, Brown, 1991. Infante-Rivard C, Villeneuve JP, Esnaola S. A framework for evaluating and conducting prognostic studies: an application to cirrhosis of the liver. J Clin Epidemiol 1989;42:791–805. Alter HJ, Purcell RH, Shih JW, Melpolder JC, Houghton M, Choo QL, Kuo G. Detection of antibody to hepatitis C virus in prospectively followed transfusion recipients with acute and chronic nonA, non-B hepatitis. N Engl J Med 1989;321:1494–1500. Tremolada F, Casarin C, Tagger A, Ribero ML, Realdi G, Alberti A, Ruol A. Antibody to hepatitis C virus in post-transfusion hepatitis. Ann Intern Med 1991;114:277–281. Jeffers LJ, Hasan F, De Medina M, Reddy R, Parker T, Silvas M, Mendez L, Schiff ER, Manns M, Houghton M, Choo Q-l, Kuo G. Prevalence of antibodies to hepatitis C virus among patients with cryptogenic chronic hepatitis and cirrhosis. Hepatology 1992; 15:187–190. Bianchi L, De Groote J, Desmet VJ, Gedik P, Korb G, Popper H, Poulsen H, Scheuer PJ, Thaler H, Wepler W. Acute and chronic hepatitis revisited. Lancet 1977;2:914–919. Sorensen TIA. Definition of death in relation to variceal bleeding. In: Burroughs AK, ed. Methodology and reviews of clinical trials in portal hypertension. Amsterdam: Excerpta Medica, 1987:31– 35. Kaplan EL, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc 1958;53:457–481. Peto R, Pike MC. Conservatism of the approximation (0-E)/E in the logrank test for survival data on tumor incidence data. Biometrics 1973;29:579–584. Cox DR. Regression models and life tables (with discussion). J R Stat Soc B 1972;34:187–220. Rao CR. Linear statistical inference and its application. New York: Wiley, 1973. Beale EML, Little RJA. Missing values in multivariate analysis. J R Stat Soc B 1975;37:129–145. Christensen E. Multivariate survival analysis using Cox’s regression model. Hepatology 1987;7:1346–1358. Dixon WH, ed. BMDP statistical software. Los Angeles: University of California, 1983. Tong MJ, El-Farra NS, Reikes AR, Co RL. Clinical outcomes after transfusion-associated hepatitis C. N Engl J Med 1995;332: 1463–1466. Colombo M, Kuo G, Choo QL, Donato MF, Del Ninno E, Tommasini MA, Dioguardi N, Houghton M. Prevalence of antibodies to hepatitis C virus in Italian patients with hepatocellular carcinoma. Lancet 1989;2:1006–1008. Bruix J, Barrera JM, Calvet X, Ercilla G, Costa J, Sanchez-Tapias JM, Vall M, Bru C, Rodes J. Prevalence of antibodies to hepatitis C virus in Spanish patients with hepatocellular carcinoma and hepatic cirrhosis. Lancet 1989;2:1004–1006. Colombo M, De Franchis R, Del Ninno E, Sangiovanni A, De Fazio C, Tommasini M, Donato MF, Piva A, Di Carlo V, Dioguardi N. Hepatocellular carcinoma in Italian patients with cirrhosis. N Engl J Med 1991;325:675–680. Nishiguchi S, Kuroki T, Nakatani S, Morimoto H, Takeda T, Nakajima S, Shiomi S, Seki S, Kobayashi K, Otani S. Randomized trial of effects of interferon alfa on incidence of hepatocellular carcinoma in chronic active hepatitis C with cirrhosis. Lancet 1995;346:1051–1055. Okuda K, Fujimoto I, Hanai A, Urano Y. Changing incidence of
WBS-Gastro
472 FATTOVICH ET AL.
GASTROENTEROLOGY Vol. 112, No. 2
hepatocellular carcinoma in Japan. Cancer Res 1987;47:4967– 4972. 28. Nishioka K, Watanabe J, Furuta S, Tanaka E, Iino S, Suzuki H, Tsuji T, Yano M, Kuo G, Choo QL, Houghton M, Oda T. A high prevalence of antibody to the hepatitis C virus in patients with hepatocellular carcinoma in Japan. Cancer 1991;67:429–433. 29. Seef LB, Buskell-Bales Z, Wright EC, Durako SJ, Alter HJ, Iber FL, Hollinger FB, Gitnik G, Knodell RG, Perillo RP, Stevens CE, Hollingsworth CG, the National Heart, Lung, and Blood Institute Study Group. Long-term mortality after transfusion-associated non-A, non-B hepatitis. N Engl J Med 1992;327:1906–1911. 30. Saracco G, Rosina F, Abate ML, Chiandussi L, Gallo V, Cerutti E, Di Napoli A, Solinas A, Deplano A, Tocco A, Cossu P, Chien D, Kuo G, Polito A, Weiner AJ, Houghton M, Verme G, Bonino F, Rizzetto M. Long-term follow-up of patients with chronic hepatitis C treated with different doses of interferon-a2b. Hepatology 1993;18:1300–1305.
/ 5e18$$0016
01-13-97 17:45:57
gasa
31. Poynard T, Bedossa P, Chevallier M, Mathurin P, Lemonnier C, Trepo C, Couzigou P, Payen JL, Sajus M, Costa JM, Vidaud M, Chaput JC, the Multicenter Study Group. A comparison of three interferon alfa-2b regimens for the long-term treatment of chronic non-A, non-B hepatitis. N Engl J Med 1995;332:1457–1462. 32. Jouet P, Roudot-Thoraval F, Dhumeaux D, Me`treau JM, Le Group Francais Pour L’Etude Du Traitment Des He`patitis Chroniques NANB/C. Comparative efficacy of interferon alfa in cirrhotic and noncirrhotic patients with non-A, non-B, C hepatitis. Gastroenterology 1994;106:686–690.
Received December 26, 1995. Accepted October 8, 1996. Address requests for reprints to: Giuseppe Realdi, M.D., Istituto di Clinica Medica, University of Sassari, Viale San Pietro 8, 07100 Sassari, Italy.
WBS-Gastro