- Email: [email protected]
Prognostic usefulness of ultrasonographic signs of portal hypertension in patients with Child–Pugh stage A liver cirrhosis
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 1999 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 94, No. 12, 1999 ISSN 0002-9270/99/$20.00 PII S0002-9270(99)00607-3
Prognostic Usefulness of Ultrasonographic Signs of Portal Hypertension in Patients With Child–Pugh Stage A Liver Cirrhosis Manuel A. Macı´as-Rodrı´guez, M.D., Paloma Rendo´n-Unceta, M.D., M. Carmen Martı´nez-Sierra, M.D., Inocencı´o Teyssiere-Blas, M.D., Fernando Dı´az-Garcı´a, M.D., and Leopoldo Martı´n-Herrera, M.D. Servicio de Aparato Digestivo, Hospital Universitario Puerta del Mar, Ca´diz, Spain
OBJECTIVE: The aim of this study was to identify factors related with mortality in patients with cirrhosis in the absence of habitual biochemical markers of liver dysfunction. METHODS: Seventy-five cirrhotic patients in Child–Pugh stage A, without hepatocellular carcinoma, were followed until death or the end of the study period. We analyzed the association between cumulative survival and 15 variables determined at the moment of inclusion: age, sex, time from diagnosis of cirrhosis, alcohol abuse, history of variceal bleeding, hepatitis B and C virus infection, Child–Pugh score, plasma albumin and bilirubin levels, prothrombin activity, and four sonographic parameters (size of liver, portal vein diameter, size of spleen, and presence of collateral circulation). RESULTS: Mean follow-up was 38.7 ⫾ 10 months. Eighteen patients died. Four-year cumulative survival was 77.4 ⫾ 5%. Only five variables had a significant influence on survival according to log-rank test: sex, previous variceal bleeding, hepatitis B virus infection, portal vein diameter, and size of the spleen. Multivariate Cox’s model showed male sex (relative risk 4.6; 95% confidence interval 1.2– 16.8) and diameter of the portal vein ⬎13 mm, splenomegaly ⬎145 mm, or both together (relative risk 6.0; 95% confidence interval 1.3–27.2) as independent predictors of the risk of death. CONCLUSIONS: Child–Pugh stage A cirrhotic patients have substantial variability in mid-term survival. Ultrasonography is a useful aid in establishing their prognosis. Men with dilation of the portal vein, splenomegaly, or both, form a group with a significantly higher risk of death. (Am J Gastroenterol 1999;94:3595–3600. © 1999 by Am. Coll. of Gastroenterology)
INTRODUCTION Currently, the development of decompensated liver disease and the classification of Child as modified by Pugh are the most widely accepted parameters for evaluating the prognosis in patients with cirrhosis (1, 2). Once the first episode of decompensation occurs, the 5-yr survival rate is less than
40% (3– 6). In contrast, this figure ranges from 40% to 90% for patients with compensated disease (3– 8). Although the Child–Pugh classification makes it possible to evaluate the degree of progression of cirrhosis, the rate of progression is not uniform, and patients with a similar score can have very different prognoses. Recent studies have investigated the usefulness of different quantitative methods for assessing liver function. These methods are relatively complex and involve the infusion of substances into the body. The results of some studies that compared their usefulness to that of traditional parameters for evaluating the prognosis in patients with cirrhosis have been contradictory (9 –11). Together with hepatocellular insufficiency and possible appearance of liver cancer, portal hypertension also affects the prognosis in patients with cirrhosis. Portal vein pressure is directly related to the degree of progression of the disease as reflected by the Child–Pugh score (12). Although it is not normally used in clinical practice and involves an invasive procedure, measurement of portal vein hypertension by catheterization of the suprahepatic veins has been found useful in predicting the prognosis in patients with cirrhosis, with regard to both risk of death and the likelihood of hemorrhage as a result of rupture of esophageal varices (13). Ultrasonography is a widely used technique in the study and follow-up of patients with cirrhosis, because of its ability to detect the appearance of hepatocellular carcinoma. It also makes it possible to measure the size of the liver, which several studies have related with functional capacity of the organ and prognosis (7, 14, 15). Moreover, this technique can reveal signs of portal hypertension, such as dilation of the portal vein, the presence of collateral circulation, or splenomegaly, all of them of considerable value in staging patients with chronic liver disease (16 –18). In this study we analyzed the usefulness of measuring the size of the liver and determining sonographic signs indicative of the presence of portal hypertension for evaluating the prognosis in patients with cirrhosis in the absence of conventional biochemical signs of hepatocellular insufficiency.
3596
Macı´as-Rodrı´guez et al.
MATERIALS AND METHODS Patients We studied 75 consecutive patients with liver cirrhosis in stage A according to the Child–Pugh classification (2) (48 patients had a score of 5 points, and the remaining 27 patients had a score of 6 points). None of the patients had evidence of liver carcinoma, and all underwent echographic examination between February 1992 and September 1993. These subjects were part of a group of 110 patients included in a prospective, controlled, cross-sectional study designed to establish the usefulness of conventional abdominal echography in diagnosing portal hypertension in patients with cirrhosis (18). Cirrhosis was diagnosed on the basis of anatomopathological findings in 66 patients (laparoscopy, laparotomy, or biopsy), and on the basis of clinical and biological evidence in the remaining 9 patients, against a background of chronic liver disease with clinical signs of portal hypertension. Forty-three patients were women. Age at the start of the study ranged from 27 to 76 yr (mean, 58.7 ⫾ 10 yr). In 38 patients hepatitis C virus antibodies (anti-HCV) were demonstrated, 25 had history of alcohol abuse (8 of them were active, although moderate, drinkers during follow-up), and 8 were carriers of hepatitis B virus surface antigen (HBsAg). Eight patients had both previous alcoholism and viral infection. Cirrhosis was classified as cryptogenic in 8 patients, primary biliary cirrhosis was demonstrated in 3, and 1 patient had Wilson disease. Time from diagnosis of cirrhosis ranged from 1 to 146 months, with a mean of 42 ⫾ 37 months. Seventeen patients had history of bleeding as a result of rupture of esophageal varices; all of them had been treated by endoscopic sclerotherapy, with as many sessions as were needed to eradicate the varices. At the time of inclusion, mean plasma albumin was 3.7 ⫾ 0.4 g/dl, bilirubin 1.1 ⫾ 0.6 mg/dl, and prothrombin activity 74.3 ⫾ 9.6% (normal values: ⬎3.5 g/dl, ⬍1 mg/dl, and ⬎70%, respectively). Protocol At the moment of inclusion in the study each patient underwent a complete clinical study, echographical examination, and routine tests of hepatic function to establish the Child–Pugh score. Patients were followed until death or the end of the study period (December 1996). Patients who died of causes unrelated with their liver disease were considered censored at the time of death. Most patients underwent physical assessment and ultrasonographical study every 6 months. Ultrasonographical Study We used two real-time apparatus with a 3.5-MHz sectorial transducer (Sonoline SL-2, Siemens, Germany, and Sigma 1 AC, Kontron, Italy). All studies were done with the patient in supine decubitus after a night fast. Size of the liver was measured in longitudinal section at the level of the midclavicular line. Caliber of the portal vein was measured at the level of the hepatic artery. Maximum diameter of the spleen
AJG – Vol. 94, No. 12, 1999
was measured at the level of the hilum, and collateral vessels were searched for carefully around the stomach and distal end of the esophagus, the splenic hilum, the umbilicus, along the splenic vein, and around the gallbladder. Statistical Analyses Results were expressed as mean ⫾ SD. Cumulative survival (⫾SE) was calculated with the method of Kaplan and Meier. We used the log-rank test to analyze the possible association between likelihood of survival and 15 variables: age, sex, time from diagnosis of cirrhosis, antecedents of alcoholism, hepatitis B and C virus infection, ruptured esophageal varices, Child–Pugh score, plasma albumin and bilirubin concentration and prothrombin activity, and the echographic parameters size of the liver, portal vein diameter, longitudinal diameter of spleen, and collateral circulation. Continuous variables were transformed into two-level categories using as a cut-off value the median or the value with the largest capacity to discriminate between patients who survived and patients who had died by the end of the study. The variables that showed a significant association with chances of survival in the univariate analysis (p ⬍ 0.05) were included in the multivariate analysis, which was done with the Cox’s proportional hazard model using a forward stepwise procedure. Proportionality of the risks was demonstrated by drawing log-minus-log curves, and interaction between the variables included in the model was ruled out. The order of entry or exit of each variable was determined by the maximum log likelihood. We used Wald’s test to compare the hypotheses. A difference of p ⬍ 0.05 was considered significant.
RESULTS Mean follow-up time was 38.7 ⫾ 10.5 months, with a range of 4 –56 months. At the end of the study period 14 patients had died of causes related with their liver disease (hepatocellular failure in 8 patients, hepatocellular carcinoma in 4, and variceal bleeding in 2). Another 4 patients had died of other causes (meningitis, respiratory insufficiency, gallbladder carcinoma, and gastric carcinoma). After 4 yr, cumulative survival was 72.6 ⫾ 5.8%; when only patients who died of liver disease were considered, the 4-yr cumulative survival rate was 77.4 ⫾ 5.6% (Fig. 1). Table 1 summarizes the results of the univariate analysis. Only five variables showed a statistically significant association with survival: sex, HBsAg, previous variceal bleeding, portal vein caliber, and longitudinal diameter of the spleen. In the Cox’s regression model constructed with the nonsonographical variables that were significantly associated with the probability of survival according to the log-rank test, only sex showed an independent predictive value for risk of death. Inclusion of the variables “portal vein caliber” and “longitudinal diameter of the spleen” in the multivariate analysis led to an increase in prognostic usefulness to the
AJG – December, 1999
US and Prognosis of Liver Cirrhosis
3597
Figure 1. Cumulative survival in 75 patients with Child–Pugh stage A liver cirrhosis.
limit of significance (Table 2). In contrast, when we grouped the patients into two categories depending on whether both echographical parameters yielded favorable values (38 patients), or whether at least one of these parameters surpassed the limit associated with increased mortality in the univariate analysis (37 patients), sonographical signs of portal hypertension were found to be the main independent factor that predicted risk of death, with a relative risk of 6.02 (Table 3).
DISCUSSION Mid-term prognosis in cirrhotic population, such as that included in our study, is difficult to assess. Our patients had good liver function at enrollment (Child–Pugh stage A), but 17 of 75 had been previously decompensated, as they suffered bleeding from esophageal varices. Such patients generally have a good mid-term prognosis, with chances of survival better than 90% after 2 yr, and 70% after 4 yr. Our results show that values of portal vein dilation or splenomegaly above the considered cut-off points are parameters with independent predictive value for risk of death in patients who do not have the clinical and biological indicators usually used to detect hepatocellular dysfunction. In addition, these sonographical signs of portal hypertension were, together with male sex, the only risk factors in the multivariate analysis. Portal hypertension influences the course of cirrhosis, not only because of the risk of variceal bleeding, but also because this factor is at least partly responsible for other complications such as infection, encephalopathy, ascites, or renal insufficiency (19). It is also involved in the progression of hepatic insufficiency, as the portal–systemic shunt in the splanchnic circulation and the thickening and loss of
permeability of the sinusoidal barrier diminish the supply of nutrients to the hepatocyte (20). This may explain the apparently paradoxical finding that sonographical signs of portal hypertension were the best predictors of the risk of death, although death was caused by liver failure in more than half of the patients who died of causes related with liver disease, whereas only 2 patients died of hemorrhage secondary to portal hypertension. Although many studies have analyzed the usefulness of conventional sonography for demonstrating portal hypertension, few have considered its usefulness in determining the prognosis in patients with cirrhosis. In two earlier studies, portal vein dilation and splenomegaly were among the variables related with the size of esophageal varices (21) and the risk of rebleeding after the first episode of rupture (22). In contrast, Zoli et al. (7), found no association between portal vein dilation and survival in patients with compensated cirrhosis, and this factor did not correlate well with the portal pressure gradient (23, 24). Although splenomegaly found during physical examination has no prognostic value (4, 14, 25), when size of the spleen has been determined with imaging techniques, patients with a larger one are at greater risk of death (8, 26) and progression of the disease (27). The variable significance of portal vein dilation and splenomegaly in different studies may reflect the fact that these parameters do not increase steadily as liver disease progresses. Because the formation of collateral vessels decreases blood flow to the liver, patients with poor hepatocellular function can show smaller alterations in these signs of portal hypertension (28). Another finding that deserves attention is the lack of association between collateral circulation and the probability of survival in our patients. Lafortune et al. (23) found that the caliber of the left gastric vein and the number of
3598
Macı´as-Rodrı´guez et al.
AJG – Vol. 94, No. 12, 1999
Table 1. Characteristics of Patients and Results of Univariate Analysis of Risk Factors by the Log Rank Test Characteristic Age (y) ⱕ61 ⬎61 Sex Female Male Time from diagnosis (mo) ⱕ28 ⬎28 Previous variceal bleeding No Yes History of alcohol abuse No Yes HbsAg (⫺) (⫹) Anti-HCV (⫺) (⫹) Child–Pugh score 5 6 Albumin (g/dl) ⬎3.7 ⱕ3.7 Prothrombin activity (%) ⬎75 ⱕ75 Bilirrubin (mg/dl) ⱕ1 ⬎1 Right hepatic lobe size (mm) ⬎128 ⱕ128 Collateral circulation Absent Present Portal vein caliber (mm) ⱕ13 ⬎13 Longitudinal diameter of spleen (mm) ⱕ145 ⬎145
No. of Patients
4-yr Survival
37 38
0.72 ⫾ 0.09 0.81 ⫾ 0.07
43 32
0.92 ⫾ 0.04 0.54 ⫾ 0.11
37 38
0.77 ⫾ 0.06 0.76 ⫾ 0.07
58 17
0.84 ⫾ 0.05 0.56 ⫾ 0.13
50 25
0.84 ⫾ 0.05 0.61 ⫾ 0.13
67 8
0.81 ⫾ 0.04 0.50 ⫾ 0.17
37 38
0.81 ⫾ 0.03 0.73 ⫾ 0.08
48 27
0.77 ⫾ 0.07 0.77 ⫾ 0.08
39 36
0.79 ⫾ 0.07 0.76 ⫾ 0.07
38 37
0.83 ⫾ 0.07 0.73 ⫾ 0.07
38 37
0.84 ⫾ 0.07 0.70 ⫾ 0.07
39 36
0.76 ⫾ 0.08 0.79 ⫾ 0.07
67 8
0.76 ⫾ 0.06 0.87 ⫾ 0.07
44 31
0.90 ⫾ 0.04 0.58 ⫾ 0.08
p Value NS 0.0016 NS 0.02 NS 0.033 NS NS NS NS 0.067 NS NS 0.0088 0.015
50 25
0.83 ⫾ 0.06 0.58 ⫾ 0.09
NS ⫽ not significant.
portal–systemic collateral vessels (determined with umbilicoportography) were directly related with the degree of portal hypertension determined as the portal– hepatic gradient. Vilgrain et al. (24) also found that the number of Table 2. Significance of Adding Sonographic Parameters to the Variable “Sex” in Cox’s Model Variables
2
p Value
Sex vs sex ⫹ portal vein diameter Sex vs sex ⫹ longitudinal diameter of spleen
3.7 2.6
0.052 0.1014
Table 3. Variables Predicting Risk of Death in the Final Cox’s Model Variable Sex Female Male Sonographic parameters PV ⱕ 13 mm and LDS ⱕ 145 mm PV ⬎ 13 mm and/or LDS ⬎ 145 mm
Relative Risk
95% Confidence Interval
1 4.65
1.28–16.88
1 6.02
1.33–27.22
PV ⫽ portal vein diameter; LDS ⫽ longitudinal diameter of spleen.
collateral vessels visualized with ultrasonography was significantly associated with the hepatic vein pressure gradient. These investigators demonstrated collateral circulation in 80% of their patients, and considered this finding to be the most sensitive for establishing a sonographical diagnosis of portal hypertension. We found portal–systemic collateral circulation in only 10% of our patients with cirrhosis, none of them showing more than one route or large caliber vessels (data not shown). The difference in data may again reflect the inclusion of patients with different degrees of hepatocellular dysfunction. Size of the liver in our patients was not associated with the likelihood of survival; however, other studies have shown a worse prognosis in patients with a smaller liver (7, 15). This discrepancy may reflect differences in the methods used to measure the size of the organ. We measured the longitudinal diameter of the right lobe, which is the part of the liver most clearly affected by the scarring that characterizes cirrhosis. The variability of this measure may have been smaller in our study because of the greater homogeneity of our sample of patients in comparison with other studies that involved patients with different degrees of hepatic insufficiency. A noteworthy finding in our results is the risk of death associated with male sex. This association was also suggested in other studies (3, 4, 10, 14, 29), although no explanation has been proposed thus far. Some researchers have not concluded that male sex is associated with a greater risk of death (5, 6, 8, 25, 30). In relation with age, studies that used longer follow-up periods reported that risk increased with age (3, 4, 8, 14). The lack of association between survival and time from diagnosis of cirrhosis is probably little more than a reflection of the weak relationship between that moment and actual duration of cirrhosis, in view of the absence of symptoms during the initial phases. In contrast with other investigators (3, 4), we have found no prognostic value for previous decompensations of liver disease. In patients who have no other clinical or biological signs of hepatocellular dysfunction, rupture of esophageal varices can be the initial manifestation of cirrhosis, especially in alcoholic patients in whom abstinence may be accompanied by a decrease in the portal pressure
AJG – December, 1999
gradient (31) and increased survival (32). We believe that these findings account for the association between this complication and the increased mortality we found in the univariate analysis, and for the loss of statistical significance when it was included in the multivariate model together with the variable sex. There are no conclusive data regarding the influence of the etiology of cirrhosis on prognosis. Discrepant findings have been frequent, especially with regard to the influence of alcoholism, probably because of differences in the populations studied, the criteria used to define alcoholism, and the proportion of patients who were abstaining at the time of the study (10, 15). Few patients in our study were active, although moderate, drinkers during follow-up; therefore, we could not assess differences between them and former drinkers. No differences in survival have been found between patients infected with hepatitis virus B or C (6). We conclude that in patients with Child–Pugh stage A liver cirrhosis, conventional ultrasonography provides valid information on the risk of death. A finding of portal vein dilation or splenomegaly distinguishes patients with increased mid-term mortality, who should be closely followed. However, our results cannot be considered applicable to patients with more severe degrees of hepatocellular dysfunction.
ACKNOWLEDGMENTS We thank Karen Shashok for translating the original manuscript into English. Reprint requests and correspondence: Leopoldo Martı´n-Herrera, M.D., Servicio de Aparato Digestivo, Hospital Universitario Puerta del Mar, Avenida Ana de Viya 21, E-11009 Ca´diz, Spain. Received Feb. 5, 1999; accepted June 28, 1999.
REFERENCES 1. Child C, Turcotte J. Surgery and portal hypertension. In: Child C, ed. The liver and portal hypertension. Philadelphia: WB Saunders, 1964:50–1. 2. Pugh R, Murray-Lyon I, Dawson J, et al. Transection of the esophagus for bleeding esophageal varices. Br J Surg 1973; 60:646 –9. 3. D’Amico G, Morabito A, Pagliaro L, et al. Survival and prognostic indicators in compensated and decompensated cirrhosis. Dig Dis Sci 1986;31:468 –75. 4. Gine´s P, Quintero E, Arroyo V, et al. Compensated cirrhosis: Natural history and prognostic factors. Hepatology 1987;7: 122– 8. 5. Propst A, Propst T, Zangerl G, et al. Prognosis and life expectancy in chronic liver disease. Dig Dis Sci 1995;40: 1805–15. 6. Gentilini P, Laffi G, La Villa G, et al. Long course and prognostic factors of virus-induced cirrhosis of the liver. Am J Gastroenterol 1997;92:66 –72. 7. Zoli M, Cordiani M, Marchesini G, et al. Prognostic indicators in compensated cirrhosis. Am J Gastroenterol 1991;86:1508 – 13. 8. Realdi G, Fattovich G, Hadziyannis S, et al. Survival and
US and Prognosis of Liver Cirrhosis
9. 10. 11. 12.
13. 14. 15.
16. 17.
18. 19. 20. 21. 22. 23. 24. 25. 26.
27. 28. 29.
3599
prognostic factors in 366 patients with compensated cirrhosis type B: A multicenter study. J Hepatol 1994;21:656 – 66. Oellerich M, Burdelski M, Lautz H, et al. Predictors of oneyear pretransplant survival in patients with cirrhosis. Hepatology 1991;14:1029 –34. Merckel C, Bolognesi M, Bellon S, et al. Aminopyrine breath test in the prognostic evaluation of patients with cirrhosis. Gut 1992;33:836 – 42. Reichen J. MEGX test in hepatology: The long-sought ultimate quantitative liver function test? J Hepatol 1993;19:4 –7. Braillon A, Cales P, Valla D, et al. Influence of the degree of liver failure on systemic and splanchnic haemodynamics and on response to propranolol in patients with cirrhosis. Gut 1986;27:1204 –9. Armonis A, Patch D, Burroughs A. Hepatic venous pressure measurement: An old test as a new prognostic marker in cirrhosis? Hepatology 1997;25:245– 8. Schlichting P, Christensen E, Andersen P, et al. Prognostic factors in cirrhosis identified by Cox’s regression model. Hepatology 1983;3:889 –95. Llach J, Gine`s P, Arroyo V, et al. Prognostic value of arterial pressure, endogenous vasoactive systems, and renal function in cirrhotic patients admitted to the hospital for the treatment of ascites. Gastroenterology 1988;94:482–7. Zoli M, Dondi C, Marchesini G, et al. Splanchnic vein measurements in patients with liver cirrhosis. A case-control study. J Ultrasound Med 1985;4:641– 6. Goyal A, Pokharma D, Sharma S. Ultrasonic measurements of portal vasculature in diagnosis of portal hypertension. A controversial subject reviewed. J Ultrasound Med 1990;9: 45– 8. Macı´as M, Rendo´n P, Soria M, et al. Utilidad de la ecografı´a abdominal en el diagno`stico de hipertensio`n portal en pacientes cirro`ticos. Rev Esp Enf Dig 1994;86:655– 60. Groszmann R, Atterbury C. The pathophysiology of portal hypertension: A basis for its classification. Sem Liver Dis 1982;2:177– 86. Schaffner F, Popper H. Capillarization of hepatic sinusoids in man. Gastroenterology 1963;41:239 – 43. Medhat A, Iber F, Dunne M, et al. Ultrasonographic findings with bleeding and nonbleeding esophageal varices. Am J Gastroenterol 1988;83:58 – 63. Schmassmann A, Zuber M, Livers M, et al. Recurrent bleeding after variceal hemorrhage: Predictive value of portal venous duplex sonography. Am J Roentgenol 1993;160:41–7. Lafortune M, Marleau D, Breton G, et al. Portal venous system measurements in portal hypertension. Radiology 1984;151: 27–30. Vilgrain V, Lebrec D, Menu Y, et al. Comparison between ultrasonographic signs and the degree of portal hypertension in patients with cirrhosis. Gastrointest Radiol 1990;15:218 –22. Serfaty L, Aumaıˆtre H, Chazouille`res O, et al. Determinants of outcome of compensated hepatitis C virus-related cirrhosis. Hepatology 1998;27:1435– 40. De Vries P, Hooge P, Hoekstra J, et al. Blunted postprandial reaction of portal venous flow in chronic liver disease, assessed with duplex Doppler: Significance for prognosis. J Hepatol 1994;21:966 –73. Ito K, Mitchell D, Hann H, et al. Compensated cirrhosis due to viral hepatitis: Using MR imaging to predict clinical progression. Am J Roentgenol 1997;169:801–5. Zoli M, Cordiani M, Marchesini G, et al. Ultrasonographic follow-up of liver cirrhosis. J Clin Ultrasound 1990;18: 91– 6. Merli M, Riggio O, Dally L, and PINC. Does malnutrition affect survival in cirrhosis? Hepatology 1996;23:1041– 8.
3600
Macı´as-Rodrı´guez et al.
30. Zoli M, Iervese T, Merckel C, et al. Prognostic significance of portal hemodynamics in patients with compensated cirrhosis. J Hepatol 1993;17:56 – 61. 31. Poynard T, Degott C, Mun˜oz C, et al. Relationship between degree of portal hypertension and liver histologic lesions in
AJG – Vol. 94, No. 12, 1999
patients with alcoholic cirrhosis. Effect of acute alcoholic hepatitis on portal hypertension. Dig Dis Sci 1987;32:337– 43. 32. Caballerı´a L, Pare`s A, Montull S, et al. Prognostic value of the aminopyrine breath test and abstinence in sever alcoholic liver disease. J Hepatol 1994;S81:
Vol. 94, No. 12, 1999 ISSN 0002-9270/99/$20.00 PII S0002-9270(99)00607-3
Prognostic Usefulness of Ultrasonographic Signs of Portal Hypertension in Patients With Child–Pugh Stage A Liver Cirrhosis Manuel A. Macı´as-Rodrı´guez, M.D., Paloma Rendo´n-Unceta, M.D., M. Carmen Martı´nez-Sierra, M.D., Inocencı´o Teyssiere-Blas, M.D., Fernando Dı´az-Garcı´a, M.D., and Leopoldo Martı´n-Herrera, M.D. Servicio de Aparato Digestivo, Hospital Universitario Puerta del Mar, Ca´diz, Spain
OBJECTIVE: The aim of this study was to identify factors related with mortality in patients with cirrhosis in the absence of habitual biochemical markers of liver dysfunction. METHODS: Seventy-five cirrhotic patients in Child–Pugh stage A, without hepatocellular carcinoma, were followed until death or the end of the study period. We analyzed the association between cumulative survival and 15 variables determined at the moment of inclusion: age, sex, time from diagnosis of cirrhosis, alcohol abuse, history of variceal bleeding, hepatitis B and C virus infection, Child–Pugh score, plasma albumin and bilirubin levels, prothrombin activity, and four sonographic parameters (size of liver, portal vein diameter, size of spleen, and presence of collateral circulation). RESULTS: Mean follow-up was 38.7 ⫾ 10 months. Eighteen patients died. Four-year cumulative survival was 77.4 ⫾ 5%. Only five variables had a significant influence on survival according to log-rank test: sex, previous variceal bleeding, hepatitis B virus infection, portal vein diameter, and size of the spleen. Multivariate Cox’s model showed male sex (relative risk 4.6; 95% confidence interval 1.2– 16.8) and diameter of the portal vein ⬎13 mm, splenomegaly ⬎145 mm, or both together (relative risk 6.0; 95% confidence interval 1.3–27.2) as independent predictors of the risk of death. CONCLUSIONS: Child–Pugh stage A cirrhotic patients have substantial variability in mid-term survival. Ultrasonography is a useful aid in establishing their prognosis. Men with dilation of the portal vein, splenomegaly, or both, form a group with a significantly higher risk of death. (Am J Gastroenterol 1999;94:3595–3600. © 1999 by Am. Coll. of Gastroenterology)
INTRODUCTION Currently, the development of decompensated liver disease and the classification of Child as modified by Pugh are the most widely accepted parameters for evaluating the prognosis in patients with cirrhosis (1, 2). Once the first episode of decompensation occurs, the 5-yr survival rate is less than
40% (3– 6). In contrast, this figure ranges from 40% to 90% for patients with compensated disease (3– 8). Although the Child–Pugh classification makes it possible to evaluate the degree of progression of cirrhosis, the rate of progression is not uniform, and patients with a similar score can have very different prognoses. Recent studies have investigated the usefulness of different quantitative methods for assessing liver function. These methods are relatively complex and involve the infusion of substances into the body. The results of some studies that compared their usefulness to that of traditional parameters for evaluating the prognosis in patients with cirrhosis have been contradictory (9 –11). Together with hepatocellular insufficiency and possible appearance of liver cancer, portal hypertension also affects the prognosis in patients with cirrhosis. Portal vein pressure is directly related to the degree of progression of the disease as reflected by the Child–Pugh score (12). Although it is not normally used in clinical practice and involves an invasive procedure, measurement of portal vein hypertension by catheterization of the suprahepatic veins has been found useful in predicting the prognosis in patients with cirrhosis, with regard to both risk of death and the likelihood of hemorrhage as a result of rupture of esophageal varices (13). Ultrasonography is a widely used technique in the study and follow-up of patients with cirrhosis, because of its ability to detect the appearance of hepatocellular carcinoma. It also makes it possible to measure the size of the liver, which several studies have related with functional capacity of the organ and prognosis (7, 14, 15). Moreover, this technique can reveal signs of portal hypertension, such as dilation of the portal vein, the presence of collateral circulation, or splenomegaly, all of them of considerable value in staging patients with chronic liver disease (16 –18). In this study we analyzed the usefulness of measuring the size of the liver and determining sonographic signs indicative of the presence of portal hypertension for evaluating the prognosis in patients with cirrhosis in the absence of conventional biochemical signs of hepatocellular insufficiency.
3596
Macı´as-Rodrı´guez et al.
MATERIALS AND METHODS Patients We studied 75 consecutive patients with liver cirrhosis in stage A according to the Child–Pugh classification (2) (48 patients had a score of 5 points, and the remaining 27 patients had a score of 6 points). None of the patients had evidence of liver carcinoma, and all underwent echographic examination between February 1992 and September 1993. These subjects were part of a group of 110 patients included in a prospective, controlled, cross-sectional study designed to establish the usefulness of conventional abdominal echography in diagnosing portal hypertension in patients with cirrhosis (18). Cirrhosis was diagnosed on the basis of anatomopathological findings in 66 patients (laparoscopy, laparotomy, or biopsy), and on the basis of clinical and biological evidence in the remaining 9 patients, against a background of chronic liver disease with clinical signs of portal hypertension. Forty-three patients were women. Age at the start of the study ranged from 27 to 76 yr (mean, 58.7 ⫾ 10 yr). In 38 patients hepatitis C virus antibodies (anti-HCV) were demonstrated, 25 had history of alcohol abuse (8 of them were active, although moderate, drinkers during follow-up), and 8 were carriers of hepatitis B virus surface antigen (HBsAg). Eight patients had both previous alcoholism and viral infection. Cirrhosis was classified as cryptogenic in 8 patients, primary biliary cirrhosis was demonstrated in 3, and 1 patient had Wilson disease. Time from diagnosis of cirrhosis ranged from 1 to 146 months, with a mean of 42 ⫾ 37 months. Seventeen patients had history of bleeding as a result of rupture of esophageal varices; all of them had been treated by endoscopic sclerotherapy, with as many sessions as were needed to eradicate the varices. At the time of inclusion, mean plasma albumin was 3.7 ⫾ 0.4 g/dl, bilirubin 1.1 ⫾ 0.6 mg/dl, and prothrombin activity 74.3 ⫾ 9.6% (normal values: ⬎3.5 g/dl, ⬍1 mg/dl, and ⬎70%, respectively). Protocol At the moment of inclusion in the study each patient underwent a complete clinical study, echographical examination, and routine tests of hepatic function to establish the Child–Pugh score. Patients were followed until death or the end of the study period (December 1996). Patients who died of causes unrelated with their liver disease were considered censored at the time of death. Most patients underwent physical assessment and ultrasonographical study every 6 months. Ultrasonographical Study We used two real-time apparatus with a 3.5-MHz sectorial transducer (Sonoline SL-2, Siemens, Germany, and Sigma 1 AC, Kontron, Italy). All studies were done with the patient in supine decubitus after a night fast. Size of the liver was measured in longitudinal section at the level of the midclavicular line. Caliber of the portal vein was measured at the level of the hepatic artery. Maximum diameter of the spleen
AJG – Vol. 94, No. 12, 1999
was measured at the level of the hilum, and collateral vessels were searched for carefully around the stomach and distal end of the esophagus, the splenic hilum, the umbilicus, along the splenic vein, and around the gallbladder. Statistical Analyses Results were expressed as mean ⫾ SD. Cumulative survival (⫾SE) was calculated with the method of Kaplan and Meier. We used the log-rank test to analyze the possible association between likelihood of survival and 15 variables: age, sex, time from diagnosis of cirrhosis, antecedents of alcoholism, hepatitis B and C virus infection, ruptured esophageal varices, Child–Pugh score, plasma albumin and bilirubin concentration and prothrombin activity, and the echographic parameters size of the liver, portal vein diameter, longitudinal diameter of spleen, and collateral circulation. Continuous variables were transformed into two-level categories using as a cut-off value the median or the value with the largest capacity to discriminate between patients who survived and patients who had died by the end of the study. The variables that showed a significant association with chances of survival in the univariate analysis (p ⬍ 0.05) were included in the multivariate analysis, which was done with the Cox’s proportional hazard model using a forward stepwise procedure. Proportionality of the risks was demonstrated by drawing log-minus-log curves, and interaction between the variables included in the model was ruled out. The order of entry or exit of each variable was determined by the maximum log likelihood. We used Wald’s test to compare the hypotheses. A difference of p ⬍ 0.05 was considered significant.
RESULTS Mean follow-up time was 38.7 ⫾ 10.5 months, with a range of 4 –56 months. At the end of the study period 14 patients had died of causes related with their liver disease (hepatocellular failure in 8 patients, hepatocellular carcinoma in 4, and variceal bleeding in 2). Another 4 patients had died of other causes (meningitis, respiratory insufficiency, gallbladder carcinoma, and gastric carcinoma). After 4 yr, cumulative survival was 72.6 ⫾ 5.8%; when only patients who died of liver disease were considered, the 4-yr cumulative survival rate was 77.4 ⫾ 5.6% (Fig. 1). Table 1 summarizes the results of the univariate analysis. Only five variables showed a statistically significant association with survival: sex, HBsAg, previous variceal bleeding, portal vein caliber, and longitudinal diameter of the spleen. In the Cox’s regression model constructed with the nonsonographical variables that were significantly associated with the probability of survival according to the log-rank test, only sex showed an independent predictive value for risk of death. Inclusion of the variables “portal vein caliber” and “longitudinal diameter of the spleen” in the multivariate analysis led to an increase in prognostic usefulness to the
AJG – December, 1999
US and Prognosis of Liver Cirrhosis
3597
Figure 1. Cumulative survival in 75 patients with Child–Pugh stage A liver cirrhosis.
limit of significance (Table 2). In contrast, when we grouped the patients into two categories depending on whether both echographical parameters yielded favorable values (38 patients), or whether at least one of these parameters surpassed the limit associated with increased mortality in the univariate analysis (37 patients), sonographical signs of portal hypertension were found to be the main independent factor that predicted risk of death, with a relative risk of 6.02 (Table 3).
DISCUSSION Mid-term prognosis in cirrhotic population, such as that included in our study, is difficult to assess. Our patients had good liver function at enrollment (Child–Pugh stage A), but 17 of 75 had been previously decompensated, as they suffered bleeding from esophageal varices. Such patients generally have a good mid-term prognosis, with chances of survival better than 90% after 2 yr, and 70% after 4 yr. Our results show that values of portal vein dilation or splenomegaly above the considered cut-off points are parameters with independent predictive value for risk of death in patients who do not have the clinical and biological indicators usually used to detect hepatocellular dysfunction. In addition, these sonographical signs of portal hypertension were, together with male sex, the only risk factors in the multivariate analysis. Portal hypertension influences the course of cirrhosis, not only because of the risk of variceal bleeding, but also because this factor is at least partly responsible for other complications such as infection, encephalopathy, ascites, or renal insufficiency (19). It is also involved in the progression of hepatic insufficiency, as the portal–systemic shunt in the splanchnic circulation and the thickening and loss of
permeability of the sinusoidal barrier diminish the supply of nutrients to the hepatocyte (20). This may explain the apparently paradoxical finding that sonographical signs of portal hypertension were the best predictors of the risk of death, although death was caused by liver failure in more than half of the patients who died of causes related with liver disease, whereas only 2 patients died of hemorrhage secondary to portal hypertension. Although many studies have analyzed the usefulness of conventional sonography for demonstrating portal hypertension, few have considered its usefulness in determining the prognosis in patients with cirrhosis. In two earlier studies, portal vein dilation and splenomegaly were among the variables related with the size of esophageal varices (21) and the risk of rebleeding after the first episode of rupture (22). In contrast, Zoli et al. (7), found no association between portal vein dilation and survival in patients with compensated cirrhosis, and this factor did not correlate well with the portal pressure gradient (23, 24). Although splenomegaly found during physical examination has no prognostic value (4, 14, 25), when size of the spleen has been determined with imaging techniques, patients with a larger one are at greater risk of death (8, 26) and progression of the disease (27). The variable significance of portal vein dilation and splenomegaly in different studies may reflect the fact that these parameters do not increase steadily as liver disease progresses. Because the formation of collateral vessels decreases blood flow to the liver, patients with poor hepatocellular function can show smaller alterations in these signs of portal hypertension (28). Another finding that deserves attention is the lack of association between collateral circulation and the probability of survival in our patients. Lafortune et al. (23) found that the caliber of the left gastric vein and the number of
3598
Macı´as-Rodrı´guez et al.
AJG – Vol. 94, No. 12, 1999
Table 1. Characteristics of Patients and Results of Univariate Analysis of Risk Factors by the Log Rank Test Characteristic Age (y) ⱕ61 ⬎61 Sex Female Male Time from diagnosis (mo) ⱕ28 ⬎28 Previous variceal bleeding No Yes History of alcohol abuse No Yes HbsAg (⫺) (⫹) Anti-HCV (⫺) (⫹) Child–Pugh score 5 6 Albumin (g/dl) ⬎3.7 ⱕ3.7 Prothrombin activity (%) ⬎75 ⱕ75 Bilirrubin (mg/dl) ⱕ1 ⬎1 Right hepatic lobe size (mm) ⬎128 ⱕ128 Collateral circulation Absent Present Portal vein caliber (mm) ⱕ13 ⬎13 Longitudinal diameter of spleen (mm) ⱕ145 ⬎145
No. of Patients
4-yr Survival
37 38
0.72 ⫾ 0.09 0.81 ⫾ 0.07
43 32
0.92 ⫾ 0.04 0.54 ⫾ 0.11
37 38
0.77 ⫾ 0.06 0.76 ⫾ 0.07
58 17
0.84 ⫾ 0.05 0.56 ⫾ 0.13
50 25
0.84 ⫾ 0.05 0.61 ⫾ 0.13
67 8
0.81 ⫾ 0.04 0.50 ⫾ 0.17
37 38
0.81 ⫾ 0.03 0.73 ⫾ 0.08
48 27
0.77 ⫾ 0.07 0.77 ⫾ 0.08
39 36
0.79 ⫾ 0.07 0.76 ⫾ 0.07
38 37
0.83 ⫾ 0.07 0.73 ⫾ 0.07
38 37
0.84 ⫾ 0.07 0.70 ⫾ 0.07
39 36
0.76 ⫾ 0.08 0.79 ⫾ 0.07
67 8
0.76 ⫾ 0.06 0.87 ⫾ 0.07
44 31
0.90 ⫾ 0.04 0.58 ⫾ 0.08
p Value NS 0.0016 NS 0.02 NS 0.033 NS NS NS NS 0.067 NS NS 0.0088 0.015
50 25
0.83 ⫾ 0.06 0.58 ⫾ 0.09
NS ⫽ not significant.
portal–systemic collateral vessels (determined with umbilicoportography) were directly related with the degree of portal hypertension determined as the portal– hepatic gradient. Vilgrain et al. (24) also found that the number of Table 2. Significance of Adding Sonographic Parameters to the Variable “Sex” in Cox’s Model Variables
2
p Value
Sex vs sex ⫹ portal vein diameter Sex vs sex ⫹ longitudinal diameter of spleen
3.7 2.6
0.052 0.1014
Table 3. Variables Predicting Risk of Death in the Final Cox’s Model Variable Sex Female Male Sonographic parameters PV ⱕ 13 mm and LDS ⱕ 145 mm PV ⬎ 13 mm and/or LDS ⬎ 145 mm
Relative Risk
95% Confidence Interval
1 4.65
1.28–16.88
1 6.02
1.33–27.22
PV ⫽ portal vein diameter; LDS ⫽ longitudinal diameter of spleen.
collateral vessels visualized with ultrasonography was significantly associated with the hepatic vein pressure gradient. These investigators demonstrated collateral circulation in 80% of their patients, and considered this finding to be the most sensitive for establishing a sonographical diagnosis of portal hypertension. We found portal–systemic collateral circulation in only 10% of our patients with cirrhosis, none of them showing more than one route or large caliber vessels (data not shown). The difference in data may again reflect the inclusion of patients with different degrees of hepatocellular dysfunction. Size of the liver in our patients was not associated with the likelihood of survival; however, other studies have shown a worse prognosis in patients with a smaller liver (7, 15). This discrepancy may reflect differences in the methods used to measure the size of the organ. We measured the longitudinal diameter of the right lobe, which is the part of the liver most clearly affected by the scarring that characterizes cirrhosis. The variability of this measure may have been smaller in our study because of the greater homogeneity of our sample of patients in comparison with other studies that involved patients with different degrees of hepatic insufficiency. A noteworthy finding in our results is the risk of death associated with male sex. This association was also suggested in other studies (3, 4, 10, 14, 29), although no explanation has been proposed thus far. Some researchers have not concluded that male sex is associated with a greater risk of death (5, 6, 8, 25, 30). In relation with age, studies that used longer follow-up periods reported that risk increased with age (3, 4, 8, 14). The lack of association between survival and time from diagnosis of cirrhosis is probably little more than a reflection of the weak relationship between that moment and actual duration of cirrhosis, in view of the absence of symptoms during the initial phases. In contrast with other investigators (3, 4), we have found no prognostic value for previous decompensations of liver disease. In patients who have no other clinical or biological signs of hepatocellular dysfunction, rupture of esophageal varices can be the initial manifestation of cirrhosis, especially in alcoholic patients in whom abstinence may be accompanied by a decrease in the portal pressure
AJG – December, 1999
gradient (31) and increased survival (32). We believe that these findings account for the association between this complication and the increased mortality we found in the univariate analysis, and for the loss of statistical significance when it was included in the multivariate model together with the variable sex. There are no conclusive data regarding the influence of the etiology of cirrhosis on prognosis. Discrepant findings have been frequent, especially with regard to the influence of alcoholism, probably because of differences in the populations studied, the criteria used to define alcoholism, and the proportion of patients who were abstaining at the time of the study (10, 15). Few patients in our study were active, although moderate, drinkers during follow-up; therefore, we could not assess differences between them and former drinkers. No differences in survival have been found between patients infected with hepatitis virus B or C (6). We conclude that in patients with Child–Pugh stage A liver cirrhosis, conventional ultrasonography provides valid information on the risk of death. A finding of portal vein dilation or splenomegaly distinguishes patients with increased mid-term mortality, who should be closely followed. However, our results cannot be considered applicable to patients with more severe degrees of hepatocellular dysfunction.
ACKNOWLEDGMENTS We thank Karen Shashok for translating the original manuscript into English. Reprint requests and correspondence: Leopoldo Martı´n-Herrera, M.D., Servicio de Aparato Digestivo, Hospital Universitario Puerta del Mar, Avenida Ana de Viya 21, E-11009 Ca´diz, Spain. Received Feb. 5, 1999; accepted June 28, 1999.
REFERENCES 1. Child C, Turcotte J. Surgery and portal hypertension. In: Child C, ed. The liver and portal hypertension. Philadelphia: WB Saunders, 1964:50–1. 2. Pugh R, Murray-Lyon I, Dawson J, et al. Transection of the esophagus for bleeding esophageal varices. Br J Surg 1973; 60:646 –9. 3. D’Amico G, Morabito A, Pagliaro L, et al. Survival and prognostic indicators in compensated and decompensated cirrhosis. Dig Dis Sci 1986;31:468 –75. 4. Gine´s P, Quintero E, Arroyo V, et al. Compensated cirrhosis: Natural history and prognostic factors. Hepatology 1987;7: 122– 8. 5. Propst A, Propst T, Zangerl G, et al. Prognosis and life expectancy in chronic liver disease. Dig Dis Sci 1995;40: 1805–15. 6. Gentilini P, Laffi G, La Villa G, et al. Long course and prognostic factors of virus-induced cirrhosis of the liver. Am J Gastroenterol 1997;92:66 –72. 7. Zoli M, Cordiani M, Marchesini G, et al. Prognostic indicators in compensated cirrhosis. Am J Gastroenterol 1991;86:1508 – 13. 8. Realdi G, Fattovich G, Hadziyannis S, et al. Survival and
US and Prognosis of Liver Cirrhosis
9. 10. 11. 12.
13. 14. 15.
16. 17.
18. 19. 20. 21. 22. 23. 24. 25. 26.
27. 28. 29.
3599
prognostic factors in 366 patients with compensated cirrhosis type B: A multicenter study. J Hepatol 1994;21:656 – 66. Oellerich M, Burdelski M, Lautz H, et al. Predictors of oneyear pretransplant survival in patients with cirrhosis. Hepatology 1991;14:1029 –34. Merckel C, Bolognesi M, Bellon S, et al. Aminopyrine breath test in the prognostic evaluation of patients with cirrhosis. Gut 1992;33:836 – 42. Reichen J. MEGX test in hepatology: The long-sought ultimate quantitative liver function test? J Hepatol 1993;19:4 –7. Braillon A, Cales P, Valla D, et al. Influence of the degree of liver failure on systemic and splanchnic haemodynamics and on response to propranolol in patients with cirrhosis. Gut 1986;27:1204 –9. Armonis A, Patch D, Burroughs A. Hepatic venous pressure measurement: An old test as a new prognostic marker in cirrhosis? Hepatology 1997;25:245– 8. Schlichting P, Christensen E, Andersen P, et al. Prognostic factors in cirrhosis identified by Cox’s regression model. Hepatology 1983;3:889 –95. Llach J, Gine`s P, Arroyo V, et al. Prognostic value of arterial pressure, endogenous vasoactive systems, and renal function in cirrhotic patients admitted to the hospital for the treatment of ascites. Gastroenterology 1988;94:482–7. Zoli M, Dondi C, Marchesini G, et al. Splanchnic vein measurements in patients with liver cirrhosis. A case-control study. J Ultrasound Med 1985;4:641– 6. Goyal A, Pokharma D, Sharma S. Ultrasonic measurements of portal vasculature in diagnosis of portal hypertension. A controversial subject reviewed. J Ultrasound Med 1990;9: 45– 8. Macı´as M, Rendo´n P, Soria M, et al. Utilidad de la ecografı´a abdominal en el diagno`stico de hipertensio`n portal en pacientes cirro`ticos. Rev Esp Enf Dig 1994;86:655– 60. Groszmann R, Atterbury C. The pathophysiology of portal hypertension: A basis for its classification. Sem Liver Dis 1982;2:177– 86. Schaffner F, Popper H. Capillarization of hepatic sinusoids in man. Gastroenterology 1963;41:239 – 43. Medhat A, Iber F, Dunne M, et al. Ultrasonographic findings with bleeding and nonbleeding esophageal varices. Am J Gastroenterol 1988;83:58 – 63. Schmassmann A, Zuber M, Livers M, et al. Recurrent bleeding after variceal hemorrhage: Predictive value of portal venous duplex sonography. Am J Roentgenol 1993;160:41–7. Lafortune M, Marleau D, Breton G, et al. Portal venous system measurements in portal hypertension. Radiology 1984;151: 27–30. Vilgrain V, Lebrec D, Menu Y, et al. Comparison between ultrasonographic signs and the degree of portal hypertension in patients with cirrhosis. Gastrointest Radiol 1990;15:218 –22. Serfaty L, Aumaıˆtre H, Chazouille`res O, et al. Determinants of outcome of compensated hepatitis C virus-related cirrhosis. Hepatology 1998;27:1435– 40. De Vries P, Hooge P, Hoekstra J, et al. Blunted postprandial reaction of portal venous flow in chronic liver disease, assessed with duplex Doppler: Significance for prognosis. J Hepatol 1994;21:966 –73. Ito K, Mitchell D, Hann H, et al. Compensated cirrhosis due to viral hepatitis: Using MR imaging to predict clinical progression. Am J Roentgenol 1997;169:801–5. Zoli M, Cordiani M, Marchesini G, et al. Ultrasonographic follow-up of liver cirrhosis. J Clin Ultrasound 1990;18: 91– 6. Merli M, Riggio O, Dally L, and PINC. Does malnutrition affect survival in cirrhosis? Hepatology 1996;23:1041– 8.
3600
Macı´as-Rodrı´guez et al.
30. Zoli M, Iervese T, Merckel C, et al. Prognostic significance of portal hemodynamics in patients with compensated cirrhosis. J Hepatol 1993;17:56 – 61. 31. Poynard T, Degott C, Mun˜oz C, et al. Relationship between degree of portal hypertension and liver histologic lesions in
AJG – Vol. 94, No. 12, 1999
patients with alcoholic cirrhosis. Effect of acute alcoholic hepatitis on portal hypertension. Dig Dis Sci 1987;32:337– 43. 32. Caballerı´a L, Pare`s A, Montull S, et al. Prognostic value of the aminopyrine breath test and abstinence in sever alcoholic liver disease. J Hepatol 1994;S81: