ORIGINAL ARTICLE—LIVER, PANCREAS, AND BILIARY TRACT

Prophylactic impact of endoscopic treatment for esophagealvarices in liver resection: a prospective study

Shintaro Yamazaki • Tadatoshi Takayama •

Masahiko Nakamura • Tokio Higaki • Shunichi Matsuoka •

Shigeaki Mizuno • Mitsuhiko Moriyama

Received: 15 March 2013 / Accepted: 21 May 2013

� Springer Japan 2013

Abstract

Background Prophylactic treatment for esophageal vari-

ces has been performed without adequate supporting evi-

dence. We assessed the feasibility of prophylactic and

follow-up treatment for high-risk esophageal varices in

patients with hepatocellular carcinoma (HCC).

Methods Patients with HCC were screened prospectively

and followed up for esophageal varices and gastroduodenal

ulceration. High-risk esophageal varices (huge F3 varices or

intermediate F2 varices positive for red color signs) were

treated prophylactically. Follow-up endoscopy was per-

formed to assess the impact of prophylaxis and changes in

varices at 1 week, 1 month, and 6 months after operation. If

high-risk varices were found during follow-up, secondary

prophylaxis was performed according to the same criteria.

Results Among 251 patients with HCC, 81 (32.3 %) had

esophageal varices on screening endoscopy. Prophylactic

endoscopic treatment was required by 13 patients (1 with F3

varices and 12 with F2 varices positive for red color signs).

Ten varices worsened, and 4 varices progressed to high-risk

varices requiring endoscopic treatment. No F0 or F1 varices

at screening endoscopy progressed to high-risk varices, and

no bleeding event occurred during 6 months of preplanned

follow-up. A preoperative platelet count of less than

10 9 104/lL (odds ratio: 4.21, 95 % confidence interval

3.11–10.6; p \ 0.001), the presence of splenomegaly (2.87,

2.16–21.8; p = 0.011), and an indocyanine green retention

rate at 15 min of greater than 30 % (2.31, 1.88–24.6; p =

0.026) were independent predictors of worsening varices.

Conclusions Our protocol for prophylactic and follow-up

treatment of high-risk esophageal varices was feasible in

patients with HCC.

Keywords Esophageal varices � Liver resection �Prophylactic treatment

Introduction

Improved outcomes of liver resection, partially attributed

to the use of strict selection criteria, have broadened the

indications for surgery in patients with advanced liver

disease [1–3]. The presence of esophageal varices is

strongly associated with the severity of cirrhosis and is an

established risk factor for postoperative adverse events.

Esophageal varices are found in 20–30 % of patients with

chronic hepatitis and 30–50 % of those with liver cirrhosis

[4–7]. Mortality rates of patients who have bleeding varices

remain as high as 20–50 % [8–12]. Therefore, good evi-

dence had been established for prophylactic intervention of

high-risk varices in patients with liver cirrhosis.

Perioperative variceal bleeding remains an important

issue in patients who undergo liver resection. Portal

hemodynamics may be altered by surgical procedures (e.g.,

liver mobilization and hepatic pedicle clamping during

liver resection). Moreover, decreased parenchymal volume

after liver resection leads to hyperperfusion of the remnant

liver. These factors are thought to exacerbate or promote

the development of portal hypertension. Preoperative

S. Yamazaki (&) � T. Takayama � M. Nakamura � T. Higaki

Department of Digestive Surgery, Nihon University School of

Medicine, 30-1 Oyaguchikami-machi, Itabashi-ku,

Tokyo 173-8610, Japan

e-mail: takayama.tadatoshi@nihon-u.ac.jp

S. Matsuoka � S. Mizuno � M. Moriyama

Division of Gastroenterology and Hepatology, Department of

Medicine, Nihon University School of Medicine, Tokyo, Japan

123

J Gastroenterol

DOI 10.1007/s00535-013-0841-y

endoscopic screening and prophylactic treatment of varices

are thus important issues in patients who undergo liver

resection [2, 13, 14].

The size, form, and color of varices are independent

predictors of bleeding [11–15]. However, standard proto-

cols for the prophylactic treatment of high-risk varices

before liver resection are unavailable, in part owing to the

lack of prospective studies providing a firm basis for cur-

rently used criteria [2, 16]. Moreover, no study has shown

which category of patients are at risk for the growth of

varices after surgical procedures. We therefore prospec-

tively assessed the impact of prophylactically treating

esophageal varices and studied postoperative changes in

such varices in patients with HCC who underwent liver

resection.

Methods

Inclusion criteria

The eligibility criteria included liver function of Child-

Pugh class A or B and curative hepatic resection for

hepatocellular carcinoma (HCC) without severe comor-

bidities. Exclusion criteria included a history of endo-

scopic, radiologic, or surgical treatment for esophageal

varices. Patients who had large gastric varices which

require balloon-occluded retrograde transvenous oblitera-

tion and who had active bleeding gastroduodenal ulceration

were excluded. Patients who had a history of esophageal or

gastric surgery or serious medical conditions that precluded

repeat endoscopy were also excluded.

Three-phase, contrast-enhanced, dynamic computed

tomographic (CT) scanning was performed in all patients

preoperatively. The diameter of the spleen was measured

on abdominal CT scans, and a major diameter of more than

12 cm was defined as splenomegaly [17]. Written informed

consent was obtained from all patients. An English-lan-

guage summary of the protocol is available at the Clinical

Trials Registry managed by the University Hospital Med-

ical Information Network in Japan (http://www.umin.ac.jp/

ctr/index.htm: UMIN000008616).

Screening and follow up endoscopy

Upper gastrointestinal endoscopy was performed at 4 dif-

ferent times (before operation and at 1 week, 1 month, and

6 months after operation). Preoperative endoscopy was

done to assess the need for prophylactic treatment of

esophageal varices [11, 12]. At follow-up endoscopy, the

change in form, color, and presence of red-color (RC) signs

of varices were evaluated [10–12, 18]. The effectiveness of

prophylactic therapy and the need for additional treatment

were assessed at the same times as follow-up endoscopy.

Liver resection was scheduled to be performed after an

interval of at least 2 weeks in patients who underwent

esophageal variceal ligation (EVL) or endoscopic clipping

of ulcers.

Indications and treatment protocol

Prophylactic treatment of esophageal varices was per-

formed according to the General Rules for Recording

Endoscopic Findings of Esophagogastric Varices (The

Japan Society for Portal Hypertension), based on the

Beppu’s classification [11]. EVL was indicated for the

management of F3 varices (largest size) or F2 varices

(enlarged tortuous) or blue varices positive for red color

signs. EVL was performed using an endoscopic ligation

device (Pneumatic EVL Device; Sumitomo Bakelite,

Tokyo, Japan). The ligation band was placed around the

varix in a spiral fashion, starting from the proximal side

and then progressing superiorly at 1-cm intervals. Addi-

tional sessions of EVL were performed every 2 weeks

using the same method until the varices were eradicated. At

follow-up endoscopy, secondary prophylaxis was per-

formed as required according to the same criteria as pri-

mary prophylaxis in patients who had signs of worsening

varices, ulceration, or both [18].

Proton-pump inhibitors were given to patients in whom

ulcers were diagnosed. Otherwise, the study protocol

required that all patients receive a proton-pump inhibitor

for 3 days after liver resection to prevent operation-related

stress ulcers and portal hypertensive gastropathy.

Surgical procedure

The indications for surgical resection and the operative

procedures were in accordance with Makuuchi’s criteria

[19]. Anatomic resection of Couinaud’s segment was the

first-line operative procedure for hepatocellular carcinoma.

Minor hepatectomy was defined as limited resection or

resection of up to two Couinaud’s segments, and major

hepatectomy was defined as resection of more than two

segments. Hepatic parenchymal transection was guided

ultrasonographically and performed by the clamp-crushing

method with the inflow-blood-occlusion technique, and

Glisson’s pedicle was tied and divided with silk thread, as

described previously [20].

Statistical analysis

Data are expressed as medians and ranges or as absolute

values and percentages. Student’s t test, the v2 test, and

Fisher’s exact test were used for univariate analysis as

required. Multivariate analysis was performed by logistic

J Gastroenterol

123

regression. Odds ratios (OR) with 95 % confidence intervals

(CI) were derived from logistic regression analysis. The per-

iod to the onset of infection was analyzed using the Kaplan–

Meier method and compared by the log-rank test. p values of

\0.05 were considered to indicate statistical significance. All

analyses were performed using a statistical software package

(JMP v.9.0; SAS Institute, Cary, NC, USA).

Results

Patients

Between 2008 and 2010, a total of 371 consecutive patients

were studied prospectively (Fig. 1). Preoperative endos-

copy performed in all patients. The underlying diagnosis

was HCC in 269 patients, metastatic liver cancer in 87,

gallbladder cancer in 8, bile duct cancer in 5, and others in

2. We excluded the non-HCC patients from this study

because the risk of esophageal varices was obviously low

(Table 1). Also, a total of 18 patients with HCC were

excluded: 9 patients had a history of esophageal or gastric

surgery, 4 had a history of variceal treatment, 3 had large

gastric varices that required preoperative treatment, and 2

were excluded for other reasons. At screening endoscopy

of 253 HCC patients, gastroduodenal ulceration was pres-

ent in 29 patients (11.6 %), including 2 who required

endoscopic treatment. Thus, two patients were excluded

from this study. Finally, 251 HCC patients were included

and followed in this study.

Laboratory background in presence or absence

of esophageal varices

The following laboratory variables differed significantly

according to the presence (n = 81) or absence (n = 170)

of esophageal varices (Table 2): platelet count [median:

10.8 9 104/lL, (range 5.6–27.9) vs. 15.9, (4.2–54.9),

p \ 0.001], albumin [3.6 g/dL, (2.7–5.7) vs. 3.9, (2.8–6.1),

p = 0.031], total bilirubin [0.84 mg/dL, (0.28–1.75) vs.

0.61, (0.19–1.60), p \ 0.001], cholinesterase [200 IU/L,

(39–382) vs. 241, (19–481), p \ 0.001], ammonia [49 lg/

dL, (16–155) vs. 42, (15–121), p = 0.005], and indocya-

nine green retention rate at 15 min [15.4 %, (3.7–46.0) vs.

10.9 %, (2.1–37.7), p \ 0.001]. The prevalences of Child

class B disease (21.8 vs. 7.6 %; p = 0.004) and spleno-

megaly (79.5 vs. 45.7 %; p \ 0.001) were significantly

higher among patients with varices.

Screening endoscopy and prophylactic treatment

On screening endoscopy, esophageal varices were found in

81 patients (32.3 %) (Table 3). There were 49 cases of

color blue varices. One patient had a huge varix (F3), and

19 had intermediate varices (F2). Twelve of the F2 patients

had RC signs on varices, but 5 had no RC signs. Conse-

quently, a total of 13 high-risk varices (1 of F3 varix and 12

of F2 varices with RC signs) were prophylactically treated

by the present protocol. EVL was effective in all patients.

There was no bleeding event related to prophylactic

treatment by EVL. Ulceration of the esophagus was found

after EVL in only 1 patient, in whom surgery was post-

poned for 2 weeks.

Follow-up endoscopy and additional treatment

The 81 varices found on screening endoscopy were fol-

lowed up for 6 months (Fig. 2). One week after operation,

269 Hepatocellular carcinoma patients

371 Liver tumor patients102 Excluded

87 Metastatic liver cancer8 Gallbladder cancer5 Bile duct cancer2 Others

18 Excluded9 Esophageal or gastric operation4 History of esophageal varices treatment 3 Gastric varices which required treatment2 Others

2 Excluded2 Endoscopic clipping for gastric ulceration

251 Follow-up endoscopy1 week, 1 month, 6 month

Flow of patients

253 Screening endoscopyProphylactic treatment performed: F3 or F2 red color signs

Fig. 1 Flow chart of the patients. There were 371 consecutive

patients who were scheduled to undergo liver resection for cancer. A

total of 102 patients were excluded because of non-hepatocellular

carcinoma (HCC). Among 269 HCC patients, 18 were excluded

because they had a history of esophageal or gastric operation, varices

treatment, or large gastric varices that required preoperative treat-

ment. At screening endoscopy, two patients were excluded because

they had an active gastric ulcer that required endoscopic treatment.

The 251 patients underwent preoperative screening endoscopy and

prophylactic treatment was performed for high-risk varices. Then, the

treatment effect and degree of varices were followed at 1 week,

1 month, and 6 months after operation

Table 1 Findings at screening endoscopy

Hepatocellular

carcinoma (HCC)

patients

Non-HCC

patients

p value

(n = 251) (n = 102)

Esophageal varices

(%)

81 (32.3) 6 (5.9) \0.001

Portal hypertensive

gastritis (%)

32 (12.7) 2 (2.0) \0.001

Gastric or duodenum

ulceration (%)

29 (11.6) 16 (15.6) 0.381

J Gastroenterol

123

65 patients (25.9 %) had varices, and 23 (9.2 %) had

ulcerations. A total of 10 varices worsened 1 week after

operation. Two F2 varices progressed to F3 varices (patient

nos. 41 and 54), and RC signs were found on two F2

varices (patient nos. 17 and 77) 1 week after operation. An

additional session of EVL was thus performed in 4 patients.

At 1 month and 6 months, no varices progress to high-risk

varices during 6 months’ follow-up. Then, there was no

need for additional endoscopic treatment. No F0 or F1

varices progressed to high-risk varices requiring endo-

scopic treatment during the 6 months after treatment.

There was no postoperative bleeding event, and the

effect of treatment according to the study protocol con-

tinued for 6 months. There were two duodenal ulcers that

required endoscopic clipping 1 week after operation.

Risk analysis of worsening esophageal varices

Multivariate analysis revealed 3 independent predictors of

worsening varices (Table 4). A platelet count of less than

10 9 104/lL was the strongest predictor (OR: 4.21, 95 %

CI 3.11–10.6; p \ 0.001). The presence of splenomegaly

(OR: 2.87, 95 % CI 2.16–21.8; p = 0.011) and an indo-

cyanine retention rate at 15 min of greater than 30 % (OR:

2.31, 95 % CI 1.88–24.6; p = 0.026) were also significant

predictors of worsening esophageal varices.

Discussion

None of the 276 patients with HCC who were consecu-

tively enrolled in this prospective study had bleeding

events. Our results indicated that F3 varices or F2 varices

positive for RC signs should undergo prophylactic EVL.

The treatment effect continued for 6 months after liver

resection in accordance with our protocol. In contrast, F0

or F1 varices at screening endoscopy did not progress to

high-risk varices requiring additional EVL.

The feasibility of prophylactic EVL has been demon-

strated in patients who have liver cirrhosis with or without

HCC [18, 21, 22]. Therefore, prophylactic EVL has been

performed in patients scheduled to undergo liver resection

for HCC [2, 15, 18]. However, postoperative changes in the

severity of varices for patients who receive prophylactic

treatment is not well understood. To date, no study has

confirmed these changes in varices or prospectively eval-

uated the characteristics of varices that require prophylactic

treatment. Our study was designed to address these unan-

swered questions.

Our results showed that the prophylactic therapy of high-

risk varices according to our protocol was feasible and may

minimize the risk of postoperative bleeding events. Our risk

analysis demonstrated that a platelet count of less than

Table 2 Preoperative characteristics of esophageal varices

Varices p value

Presence

(n = 81)

Absence

(n = 170)

Hepatitis viral infection

(%)

62 (79.5) 142 (82.1) 0.411

Tumor diameter (mm) 24 (11–105) 28 (10–205) 0.344

Number of tumors 1 (1–3) 1 (1–4) 0.974

Phenotype (expansive

growth) (%)

17 (22.1) 40 (25.5) 0.569

Vascular invasion (%) 16 (20.5) 45 (26.0) 0.311

Platelets (104/lL) 10.8

(5.6–27.9)

15.9

(42–54.9)

\0.001

Albumin (g/dL) 3.6 (2.7–5.7) 3.9 (2.8–6.1) 0.031

Total bilirubin (mg/dL) 0.84

(0.28–1.75)

0.61

(0.19–1.60)

\0.001

Aspartate

aminotransferase (IU/L)

38 (11–96) 34 (5–296) 0.523

Cholinesterase (IU/L) 200 (39–382) 241 (19–481) \0.001

Ammonia (lg/dL) 49 (16–155) 42 (15–121) 0.005

Prothrombin time (INR) 1.00

(0.83–1.19)

0.97

(0.86–1.08)

0.508

ICG R15 (%) 15.4

(3.7–46.0)

10.9

(2.1–37.7)

\0.001

Child-Pugh B (%) 17 (21.8) 13 (7.6) 0.004

Splenomegaly (%) 62 (79.5) 79 (45.7) \0.001

ICG R15 indocyanine green retention rate at 15 min, INR international

normalized rate

Table 3 Endoscopic screening and prophylactic treatment in 251

patients with HCC

Screening Follow-up

1 week 1 month 6 months

Esophageal

varices

81

(32.3 %)

65

(25.9 %)

59

(23.5 %)

58

(23.1 %)

Form (F0/F1/F2/

F3)

12/51/17/

1

17/39/7/2 12/37/7/0 9/40/9/0

Blue color

varices

49 52 32 29

Presence of red

color signs

12 2 0 0

Endoscopic

treatment

13 4 0 0

Ulcerationa 29

(11.6 %)

23

(9.2 %)

14

(5.6 %)

6 (2.4 %)

Gastric 25 21 18 4

Endoscopic

treatment

2 0 0 0

Duodenum 6 5 3 2

Endoscopic

treatment

0 2 0 0

a Total number of patients

J Gastroenterol

123

10 9 104/lL, the presence of splenomegaly, and an indo-

cyanine retention rate at 15 min of greater than 30 % were

independent predictors of postoperative worsening of vari-

ces. However, no bleeding event occurred when patients

were treated prophylactically according to present protocol.

Moreover, no F0 or F1 varices at screening progressed to

high-risk varices requiring EVL. Thus, follow-up endos-

copy may be unnecessary in these patients.

Liver resection and related maneuvers were thought to

have a negative impact on varices. Factors such as inter-

ruption of collateral blood flow on mobilization of the

liver, intermittent clamping of the portal vein during liver

transection, decreased parenchymal volume after liver

resection, and intraoperative transfusions may destabilize

portal pressure. Our results revealed that worsening of

varices depends on the severity of liver cirrhosis and the

presence of portal hypertension. Thus, if varices are treated

prophylactically and followed up according to the present

protocol, any type of procedure can be performed safely in

patients who have adequate liver function [19].

Portal hypertension was considered a negative factor for

liver resection [17, 23]. However, recent works have shown

that the presence of portal hypertension is no longer a

contraindication [1–3, 10]. Postoperative bleeding occurs

in 3–5 % of patients with portal hypertension, and the risk

might be higher in patients who have high-risk varices [6,

9]. We therefore recommend screening endoscopy for

patients with liver cirrhosis and prophylactic treatment of

high-risk F3 varices or F2 varices positive for RC signs.

Moreover, follow-up endoscopy should be performed

1 week after operation in patients who have F2 varices

negative for RC signs at preoperative screening endoscopy.

The incidence of esophageal variceal bleeding in patients

who undergo liver resection remains unknown. However,

the significant treatment impact of EVL on high-risk varices

in patients with liver cirrhosis was shown in many studies.

Postoperative bleeding events are a serious hazard after

liver resection because they lead to secondary liver failure.

Thus, to ensure patients’ safety, we performed follow-up

endoscopy and follow-up prophylaxis 1 week after opera-

tion. We believe that our treatment protocol may contribute

to zero postoperative bleeding events in patients with HCC.

This prospective study defined the characteristics of

patients with esophageal varices associated with liver cir-

rhosis who are most likely to benefit from prophylactic

treatment before liver resection. Adherence to strict

selection criteria in liver resection and the prophylactic

treatment of varices may contribute to postoperative zero

mortality.

Fig. 2 Change in degree of

varices. Patients in whom

varices were found at screening

endoscopy were followed up for

6 months after operation. The

severities and changes of

varices are shown. A total of 10

varices worsened after

operation, and 2 F2 varices

negative for red color signs

required endoscopic treatment

(patient nos. 41 and 54). No F0

or F1 varices required

endoscopic treatment during the

6 months after operation

Table 4 Multivariate analysis of worsening varices

Odds

ratio

95 % confidence

interval

p value

Lower Upper

Platelet \10 9 104/lL 4.21 3.11 10.6 \0.001

Splenomegaly Presence 2.87 2.16 21.8 0.011

ICG R15 [30 % 2.31 1.88 24.6 0.026

Total bilirubin [1.5 mg/dL 1.76 0.97 16.9 0.053

Albumin \3.0 g/dL 1.22 0.88 32.5 0.088

ICG R15 indocyanine green retention rate at 15 min

J Gastroenterol

123

Acknowledgments This study was supported by the grant from

2013 Hagiwara fund, Nihon University School of Medicine.

Conflict of interest The authors declare that they have no conflict

of interest.

References

1. Capussotti L, Ferrero A, Vigano L, et al. Portal hypertension:

contraindication to liver surgery? World J Surg. 2006;30:992–9.

2. Ishizawa T, Hasegawa K, Aoki T, et al. Neither multiple tumors

nor portal hypertension are surgical contraindications for hepa-

tocellular carcinoma. Gastroenterology. 2008;134:1908–16.

3. Kawano Y, Sasaki A, Kai S, et al. Short- and long-term outcomes

after hepatic resection for hepatocellular carcinoma with con-

comitant esophageal varices in patients with cirrhosis. Ann Surg

Oncol. 2008;15:1670–6.

4. Brocchi E, Caletti G, Brambilla G, et al. Prediction of the first

variceal hemorrhage in patients with cirrhosis of the liver and

esophageal varices. A prospective multicenter study. North Ital-

ian Endoscopic Club for the Study and Treatment of Esophageal

Varices. N Engl J Med. 1988;319:983–9.

5. D’Amico G, Pagliaro L, Bosch J. The treatment of portal

hypertension: a meta-analytic review. Hepatology. 2003;22:

332–54.

6. de Franchis R. Evolving consensus in portal hypertension. Report

of the Baveno IV consensus workshop on methodology of diag-

nosis and therapy in portal hypertension. J Hepatol. 2005;43:

167–76.

7. Bambha K, Kim WR, Pedersen R, et al. Predictors of early re-

bleeding and mortality after acute variceal haemorrhage in

patients with cirrhosis. Gut. 2008;57:814–20.

8. Spina GP, Arcidiacono R, Bosch J, et al. Gastric endoscopic

features in portal hypertension: final report of a consensus con-

ference, Milan, Italy, September 19, 1992. J Hepatol. 1994;21:

461–7.

9. del Olmo JA, Pena A, Serra MA, et al. Predictors of morbidity

and mortality after the first episode of upper gastrointestinal

bleeding in liver cirrhosis. J Hepatol. 2000;32:19–24.

10. Nidegger D, Ragot S, Berthelemy P, et al. Cirrhosis and bleeding:

the need for very early management. J Hepatol. 2003;39:509–14.

11. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal

hemorrhage by esophageal endoscopy. Gastrointest Endosc.

1981;27:213–8.

12. Merkel C, Zoli M, Siringo S, et al. Prognostic indicators of risk

for first variceal bleeding in cirrhosis: a multicenter study in 711

patients to validate and improve the North Italian Endoscopic

Club (NIEC) index. Am J Gastroenterol. 2000;95:2915–20.

13. Tomikawa M, Shimabukuro R, Okita K, et al. Propranolol alone

may not be acceptable to prevent first esophageal variceal

bleeding in Japanese cirrhotic patients: randomized controlled

trial. J Gastroenterol Hepatol. 2004;19:576–81.

14. Carbonell N, Pauwels A, Serfaty L, et al. Improved survival after

variceal bleeding in patients with cirrhosis over the past two

decades. Hepatology. 2004;40:652–9.

15. Snady H, Feinman L. Prediction of variceal hemorrhage: a pro-

spective study. Am J Gastroenterol. 1988;83:519–25.

16. Yeo W, Sung JY, Ward SC, et al. A prospective study of upper

gastrointestinal hemorrhage in patients with hepatocellular car-

cinoma. Dig Dis Sci. 1995;40:2516–21.

17. Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carci-

noma: the BCLC staging classification. Semin Liver Dis.

1999;19:329–38.

18. Chen WC, Hou MC, Lin HC, et al. Feasibility and potential

benefit of maintenance endoscopic variceal ligation in patients

with unresectable hepatocellular carcinoma and acute esophageal

variceal hemorrhage: a controlled trial. Gastrointest Endosc.

2001;54:18–23.

19. Makuuchi M, Kosuge T, Takayama T, et al. Surgery for small

liver cancers. Semin Surg Oncol. 1993;9:298–304.

20. Takayama T, Makuuchi M, Kubota K, et al. Randomized com-

parison of ultrasonic vs clamp transection of the liver. Arch Surg.

2001;136:922–8.

21. Chalasani N, Kahi C, Francois F, et al. Improved patient survival

after acute variceal bleeding: a multicenter, cohort study. Am J

Gastroenterol. 2003;98:653–9.

22. Imperiale TF, Chalasani N. A meta-analysis of endoscopic vari-

ceal ligation for primary prophylaxis of esophageal variceal

bleeding. Hepatology. 2001;33:802–7.

23. Bruix J, Castells A, Bosch J, et al. Surgical resection of hepato-

cellular carcinoma in cirrhotic patients: prognostic value of pre-

operative portal pressure. Gastroenterology. 1996;111:1018–22.

J Gastroenterol

123