capsule endoskopi edit.docx

Dig Dis Sci (2014) 59:1036–1041DOI 10.1007/s10620-014-3036-3

ORIGINAL ARTICLE

Portal Hypertensive Enteropathy Diagnosed by CapsuleEndoscopy in Cirrhotic Patients: A Nationwide Multicenter Study

Seong Ran Jeon · Jin-Oh Kim · Ji-Beom Kim · Byong Duk Ye · Dong Kyung ChangKi-Nam Shim · Dae Young Cheung · Jin Soo Kim · Myung Gyu Choi ·Hyun Joo Song · Yun Jeong Lim · Soo Jung Park · Ji Hyun Kim ·

Jeong Seop Moon · Yoon Tae Jeen · The Korean Gut Image Study Group

Received: 27 November 2013 / Accepted: 8 January 2014 / Published online: 4 February 2014Ó Springer Science+Business Media New York 2014

AbstractBackground and Aim Due to the limited data on portalhypertensive enteropathy (PHE), the prevalence of andclinical factors related to PHE remain unclear. This studydetermined the prevalence of PHE using capsule endos-copy (CE) and PHE-related clinical factors.Methods This was a retrospective multicenter study usingthe Capsule Endoscopy Nationwide Database Registry.From 2,879 cases that underwent CE, 45 cirrhosis patientswith portal hypertension (PH) were enrolled and dividedinto PHE (n = 18) and non-PHE (n = 27) groups. Fromcomputed tomography (CT) images, six secondary changesdue to PH were scored to give a total CT score of 0–6. The

S. R. Jeon Á J.-O. Kim (&)Department of Internal Medicine, Soonchunhyang UniversityCollege of Medicine, 22 Daesagwan-Gil, Yongsan-Gu,Seoul 140-743, Koreae-mail: jokim@schmc.ac.kr

J.-B. Kim Á B. D. YeDepartment of Internal Medicine, University of Ulsan College ofMedicine, Seoul, Korea

D. K. ChangDepartment of Internal Medicine, Sungkyunkwan UniversitySchool of Medicine, Seoul, Korea

K.-N. ShimDepartment of Internal Medicine, Ewha Womans UniversitySchool of Medicine, Seoul, Korea

D. Y. Cheung Á J. S. Kim Á M. G. ChoiDepartment of Internal Medicine, The Catholic UniversityCollege of Medicine, Seoul, Korea

H. J. SongDepartment of Internal Medicine, Jeju National UniversitySchool of Medicine, Cheju, Korea

123

·

main outcome variable was the prevalence of PHE andPHE-related clinical factors.Results The prevalence of PHE was 40 %. Comparingthe PHE and non-PHE groups, the most common findingswere angiodysplasias in 55.7 % (vs. 7.4 %, p = 0.001) andvarices in 38.9 % (vs. 0 %, p = 0.001). Active bleedingwas observed in 16.6 and 3.7 %, respectively, but thisdifference was not significant. In the univariate analysis,Child–Turcotte–Pugh class C (p = 0.002) and a high CTscore (C3 vs. \3, p = 0.004) were significantly associatedwith PHE. However, only a high CT score was significantin the multivariate analysis (odds ratio 11.19; 95 % con-fidence interval, 1.59-infinity; p = 0.040).

Y. J. LimDepartment of Internal Medicine, Dongguk University Collegeof Medicine, Goyang, Korea

S. J. ParkDepartment of Internal Medicine, Yonsei University College ofMedicine, Seoul, Korea

J. H. KimDepartment of Internal Medicine, Inje University College ofMedicine, Busan, Korea

J. S. MoonDepartment of Internal Medicine, Inje University College ofMedicine, Seoul, Korea

Y. T. JeenDepartment of Internal Medicine, Korea University College ofMedicine, Seoul, Korea

Dig Dis Sci (2014) 59:1036–1041

Conclusions The prevalence of PHE was 40 %, and itmight be more prevalent in cirrhosis patients with PH whohave a high CT score. CE is a useful diagnostic tool forevaluating PHE in cirrhosis patients with PH.

Keywords Capsule endoscopy Á Enteropathy Á Portalhypertension Á Liver cirrhosis

Introduction

Portal hypertensive enteropathy (PHE) is a mucosalabnormality of the small bowel that is observed in cirrhosispatients with portal hypertension (PH) [1]. Althoughprognosis-related major bleeding of PH generally origi-nates from esophageal (EV) or gastric varices (GV), PHEcan lead to gastrointestinal (GI) bleeding or anemia incirrhosis patients [2]. As the small bowel is difficult toapproach from either the mouth or anus, it is difficult toevaluate the entire small bowel using upper and lower GIendoscopy. Consequently, endoscopic abnormalities in thesmall bowels of patients with PH have not been wellcharacterized [3]. The introduction of capsule endoscopy(CE) and double-balloon enteroscopy (DBE) has made itpossible to detect these abnormalities.

The mucosal abnormalities of PHE are defined as mucosal

inflammatory-like abnormalities and vascular lesions [1].Summarizing previous studies, the most common indicationfor CE is obscure GI bleeding (OGIB), which is defined asbleeding from an unknown origin that persists or recurs afternegative upper and lower endoscopy [4]. The reportedprevalence of PHE is 65.7–92.8 % in cirrhosis patients,which is higher than that in control groups [1, 5–7]; however,another study reported a frequency of only 18.2 % [8]. Inaddition, the clinical factors related to PHE are heteroge-neous, including grade 2 or greater EV, portal hypertensivegastropathy (PHG), portal hypertensive colopathy (PHC),Child–Turcotte–Pugh (CTP) class C cirrhosis [1], and ahistory of acute GI bleeding [5].

Due to the limited data available on PHE, the prevalence

of and clinical factors related to PHE remain unclear. Toour knowledge, no multicenter study has examined PHE in

cirrhosis patients with PH. Therefore, this study determinedthe prevalence of PHE using CE, and evaluated the clinicalfactors related to PHE in cirrhosis patients with PH.

Methods

Study Design and Patients

We analyzed the records in the Capsule EndoscopyNationwide Database Registry of cirrhosis patients with PH

1037

who underwent CE for various reasons from October 2002to May 2012. We confirmed the presence of liver cirrhosis(LC) based on a physical examination, laboratory findings,and histological or radiological features. Cirrhosis severitywas graded according to the CTP score. PH was diagnosedusing endoscopic or radiographic evidence of esophageal,gastric, or intra-abdominal varices, collateral circulation,splenomegaly, secondary gall bladder wall thickening, orascites [3]. Exclusion criteria included the following:underlying small bowel disease; concomitant severe car-diopulmonary disease or end-stage renal disease; user ofanticoagulant or nonsteroidal anti-inflammatory drugs;evidence of acute infection or sepsis; and lack of data.From 2,879 CE cases, 45 that met the inclusion criteriawere enrolled in this study. These patients were dividedinto those with (n = 18) and without (n = 27) PHE, andthen compared retrospectively. The data collection andanalysis were approved by each Institutional Review Board(IRB_2012-045).

Capsule Endoscopy Procedure

Patients were examined using the PillCamTM (SB1/SB2)(Given Imaging; Yokneam, Israel) or MiroCamTM (Intro-Medic, Seoul, Republic of Korea). Written informed con-sent was obtained from all patients before the procedure.Either no bowel preparation was used or the bowel wasprepared according to the patient’s circumstances. The CEfindings were classified as highly relevant (P2) or lessrelevant (P1 or P0) lesions according to the standardpractice guidelines [9]. A result of P2 was reported aspositive. The degree of cleansing of the small bowel duringCE was scored as excellent, good, fair, or poor [10]. Then,the quality of images was divided into acceptable (excel-lent, good, and fair) and non-acceptable (poor).

Classification and Definition of EndoscopicAbnormalities

EVs were graded using the criteria proposed by Chunget al. [11]. GVs were classified as absent or present. PHGwas classified as mild or severe according to McCormacket al. [12]. PHC was subclassified into grades I (mild), II(moderate), and III (severe). PHE was classified as grade I(mucosal inflammatory-like abnormalities; edema, ery-thema, granularity, friability) or II (vascular lesions; cherryred spots, telangiectasia, angiodysplasia-like lesions, andvarices) [1]. Small bowel varices were defined as circum-ferential raised venous lesions in the small bowel thatresemble classic images of EV or GV [13]. The mainoutcome variable was the prevalence of PHE based on theCE findings.

123

1038

Computed Tomography Findings and New ScoringSystem

Dig Dis Sci (2014) 59:1036–1041

Table 1 Clinical characteristics of the portal hypertensive enterop-athy and non-portal hypertensive enteropathy groups

We identified six secondary changes due to PH onVariable PHE Non-PHE

abdominal computed tomography (CT) because the CT Patients (%) 18 (40.0) 27 (60.0)

findings were too diverse to be consistent. A new scoringsystem was developed to evaluate whether the CT findingswere related to PHE diagnosed by CE. The six findingsincluded: (1) EV or GV; (2) other collateral circulations(i.e., periumbilical varices); (3) PHG or PHC; (4) portalhypertensive cholecystopathy; (5) splenomegaly; and (6)ascites. Each of these six findings was worth one point. Thepoints were totaled to give a score from 0 to 6 points. We

Age (mean ± SD; year)

Male (%)

Etiology of cirrhoisis

HBV

HCV

Alcohol

Others

Comorbidities

57.3 ± 10.1

12 (66.7)

7

3

7

1

56.4 ± 14.1

21 (77.8)

10

4

7

6

0.811

0.499

0.713

0.767

analyzed whether the CT score and various clinical factorswere related to PHE between the PHE and non-PHE

Diabetes mellitus

Hypertension

groups. Pulmonary tuberculosis

Hepatocellular carcinoma

1

3

1

6

Statistical Analysis

The statistical analyses were performed using SPSS vs.14.0 (SPSS, Chicago, IL) and SAS vs. 9.2 (SAS, Cary,NC). All continuous variables were compared using a two-

Previous bleeding history

(%)

Esophageal variceal bleeding

Small bowel bleeding

Other

History of endoscopic

8 (44.4)

5 (62.5)

1 (12.5)

2 (25.0)

5 (62.5)

13 (48.1)

4 (30.7)

2 (15.3)

7 (54.0)

3 (23.0)

0.807

0.592

tailed Student’s t test. All categorical variables were intervention (%)a

compared using Fisher’s exact test. Multivariate analysiswas conducted by logistic regression. A p value of \0.05was used as the cutoff for statistical significance.

Results

Clinical Characteristics

The mean age of the 45 patients with PH was57.1 ± 13.1 years, and the dominant gender was male (33/45, 73.3 %). Among these patients, the mean hemoglobinwas 7.5 ± 1.9 g/dL, and 13.3 % (6/45) of the patients wereCTP class C. CTP class C was significantly more frequentin the PHE group (p = 0.002). The mean CT score wassignificantly higher in the PHE group (1.8 ± 1.4 vs.1.0 ± 0.8; p = 0.027). The clinical characteristics aresummarized in Table 1.

Conventional Endoscopy and Capsule Endoscopy-Related Data

EV, GV, PHG, and PHC except rectal varices were more

Lowest Hb (mean ± SD,

g/dL)

CTP class (%)

Class A/B

Class C

Duration of symptom(mean ± SD; days)

Esophageal varices

Grade 1

Grade 2

Grade 3

Grade 4

Gastric varices

Portal hypertensivegastropathy

Mild

Severe

Rectal varices

Portal hypertensive

colopathy

Grade 1

Grade 2

Grade 3

CT score (mean ± SD)

7.5 ± 1.8

12 (66.7)

6 (33.3)

22.1 ± 35.2

13 (72.2)

9 (69.2)

3 (23.1)

1 (7.7)

0 (0.0)

3 (16.7)

9 (50.0)

5 (55.6)

4 (44.4)

0 (0.0)

5 (27.8)

3 (60.0)

1 (20.0)

1 (20.0)

1.8 ± 1.4

7.6 ± 1.9

27 (100.0)

0 (0.0)

10.6 ± 9.9

13 (48.1)

87 (61.5)

4 (30.8)

1 (7.7)

0 (0.0)

3 (11.1)

7 (25.9)

4 (57.1)

3 (42.9)

3 (11.1)

4 (14.8)

0 (0.0)

2 (50.0)

2 (50.0)

1.0 ± 0.8

0.807

0.002b

0.132

0.904

0.671

0.098

0.264

0.449

0.027b

frequent in the PHE group than the non-PHE group, but the

difference was not significant (Table 1). For the 45patients, the completion rate of CE was 80 %

p

4 5

3 5

a

b

(36/45) andOGIB was the most common indication for CE (44/45,97.7 %). In the non-PHE group, only one patient

123

PHE portal hypertensive enteropathy, SD standard deviation, HBVhepatitis B virus, HCV hepatitis C virus, Hb hemoglobin, CTP Child–Turcott–Pugh, CT computed tomography

Endoscopic variceal ligation and sclerotherapy

Denote significant difference (p \ 0.05)

p

Dig Dis Sci (2014) 59:1036–1041

Table 2 Capsule endoscopic findings of the portal hypertensiveenteropathy and non-portal hypertensive enteropathy groups

Outcome Variables Data

1039

Variable PHE(n = 18) (%)

Non-PHE(n = 27) (%)

The prevalence of PHE was 40 % (18/45), and vascularlesions, except inflammatory-like lesions, were detected in

Indication of CE

Overt OGIB

Occult OGIB

Other

Type of CE

PillCamTM (SB1/2)

MiroCamTM

Bowel preparation

NPO

PEG 2L or 4L

NaP

Acceptable qualityof images

Arrival at cecum

Results of CE

Positive findings (P2)

Negative findings(P1 or P0)

Diagnostic yield

17 (94.4)

1 (5.6)

0

13 (72.2)

5 (27.8)

5 (27.8)

12 (66.7)

1 (5.5)

16 (88.8)

15 (88.2)

18 (100.0)

0 (0.0)

71.1 % (32/45)

19 (70.4)

7 (25.9)

1 (3.7)

21 (80.8)

5 (19.2)

7 (25.9)

19 (70.4)

1 (3.7)

25 (92.3)

21 (77.8)

14 (51.9)

13 (48.1)

0.137

0.716

0.246

0.669

0.455

\0.001

94.4 % (17/18). In the PHE group, the most commonfindings included angiodysplasia, which was detected sig-nificantly more frequently than in the non-PHE group (55.7vs. 7.4 %, p = 0.001). Varices were significantly morecommon in the PHE group (38.9 vs. 0 %, p = 0.001).Active bleeding was more frequent in the PHE group thanthe non-PHE group, but the difference was not significant(16.6 vs. 3.7 %, p = 0.139). The CE findings in the smallbowel are presented in Table 3 and Fig. 1. In the univariateanalysis, CTP class C and a CT score [3 points weresignificantly associated with PHE. However, only a highCT score was significantly associated with PHE in themultivariate analysis [odds ratio (OR) 11.19, 95 % confi-dence interval (CI), 1.59-infinity, p = 0.040] (Table 4).

Treatment and Clinical Outcomes

In the PHE group, the frequency of interventional therapyshowed a marginally significant difference compared with thenon-PHE group (38.9 vs. 14.8 %, p = 0.066). In the 40

PHE portal hypertensive enteropathy, CE capsule endoscopy, OGIBobscure gastrointestinal bleeding, NPO nil per os, PEG polyethyleneglycol, NaP sodium phosphate

Table 3 Capsule endoscopic findings in the small bowel

patients who were followed, the rebleeding rate was not sig-nificantly different between the groups. For the 17 patientswith rebleeding, small bowel lesions were more common inthe PHE group than in the non-PHE group, although the dif-ference was not significant. During follow-up, no death due tosmall bowel bleeding occurred in either group (Table 5).

Group Finding N (%)

PHE (n = 18)

Non-PHE (n = 27)

Inflammatory-like lesion

Red spots

Angiodysplasias

Red spots ? angiodysplasias

Varices

Varices ? angiodysplasias

Angiodysplasias ? active

bleeding

Varices ? active bleeding

Vascular lesion

Mucosal lesion

Mucosal lesion ? active bleeding

Tumor

Other

No specific finding

1 (5.6)

3 (16.6)

5 (27.8)

1 (5.6)

2 (11.1)

3 (16.7)

1 (5.6)

2 (11.1)

2 (7.4)

6 (22.2)

1 (3.7)

1 (3.7)

5 (18.5)

12 (44.4)

Discussion

PH-related disease can be a cause of bleeding in cirrhosispatients [2]. Unlike PHG and PHC, data on PHE are scarcedue to the relative inaccessibility of the small bowel [14].Although our knowledge of PHE has expanded with theintroduction of CE and DBE into clinical practice, the trueprevalence of PHE and related clinical factors remainunclear. To date, there has been no multicenter study basedon the Capsule Endoscopy Nationwide Database Registry.Therefore, we evaluated the prevalence and clinical sig-nificance of PHE using CE in cirrhosis patients with PH.

According to data derived from a few small studies, theprevalence of PHE in cirrhosis patients exceeds 60 % [1,5–7]. In our study, the prevalence of PHE was 40 %, which

PHE portal hypertensive enteropathy

underwent CE to evaluate protein-losing enteropathy. Theoverall diagnostic yield was 71.1 % (32/45). The CEresults are summarized in Table 2.

was lower than reported previously. However, whenapplying the narrow definition of PHE defined as edema-tous and hyperemic lesions that are reminiscent ofinflammatory lesions [8], the prevalence of PHE was 18 %,which is lower than our result. Overestimation and differ-ent interpretations might explain this.

123

1040

Fig. 1 Capsule endoscopy findings in patients with portal hyperten-sive enteropathy. a Hyperemic mucosal change. b Herring roeappearance (edematous, swollen, and rounded villi). c Mucosal

Table 4 Univariate and multivariate analysis of the portal hyper-

Dig Dis Sci (2014) 59:1036–1041

granularity with edema. d Red spots (arrow). e Angiodysplasia(arrow). f Varix with bluish discoloration (arrow). g Saccular varixwith a bluish color change. h Active variceal bleeding in the jejunum

Table 5 Treatment and clinical outcomes

tensive enteropathy and non-portal hypertensive enteropathy groups

Variable PHE Non-PHE

Variable Univariate Multivariate (n = 18) (%) (n = 27) (%)

Bleeding history

No

Yes

CTP score

p

0.745

OR 95 % CI p Treatment

Interventional

Endoscopic

Radiologic

Surgical

7 (38.9)

2 (11.1)

4 (22.2)

1 (5.6)

4 (14.8)

3 (11.1)

0 (0.0)

1 (3.7)

0.066

Child A/B 0.002 Ref Conservative 11 (61.1) 23 (85.2)

Child C

EV severity

\Grade 2

CGrade 2

PHG

No

Yes

PHC

No

Yes

CT score

0.819

0.175

0.126

7.08 0.81–infinity 0.136 Follow up duration(mean ± SD; days)

Rebleeding after

treatmenta

Site

Small bowel

Extra-small bowel

Cause of death

Small bowel bleeding

Extra-small bowel

bleeding

Liver related

Otherb

18.6 ± 25.9

7 (46.7)

5 (71.4)

2 (28.6)

3 (16.7)

0 (0.0)

0 (0.0)

3 (100.0)

0 (0.0)

24.6 ± 25.3

10 (40.0)

5 (50.0)

5 (50.0)

4 (14.8)

0 (0.0)

1 (25.0)

2 (50.0)

1 (25.0)

0.453

0.749

0.622

0.836

\3 0.004a Ref PHE portal hypertensive enteropathy, SD standard deviation

C3 11.19 1.59–infinity 0.040 Rebleeding rate among 40 patients who were followed

PHE portal hypertensive enteropathy, OR odds ratio, CI

confidenceinterval, CTP Child–Turcott–Pugh, EV esophageal varix, PHG portalhypertensive gastropathy, PHC portal hypertensive colopathy, CTcomputed tomography

Denote significant difference (p \ 0.05)

123

p

a

a a

b

a

Aspiration pneumonia, respiratory arrest

In the PHE group, the most common finding was mul-tiple angiodysplasias, which is similar to the results of mostprevious studies. In our study, varices and active bleeding

Dig Dis Sci (2014) 59:1036–1041

were observed during CE in 38.9 and 16.6 % of thepatients with PHE, respectively, and were more frequentthan the reported rates of varices (8.1–21.4 %) and activebleeding (5.5–11 %) [1, 5, 7, 13]. This suggests that PHE isthe cause of meaningful GI bleeding or anemia in cirrhosispatients.

De Palma et al. [1] reported that grade 2 or more EV,

1041

and deciding treatment methods in cirrhosis patients with

Acknowledgments We gratefully and sincerely thank Ji Sung Lee,PhD, Biostatistical Consulting Unit, Soonchunhyang UniversityMedical Center, Seoul, Korea, who helped with the statisticalanalysis.

PHG, PHC, and CTP class C cirrhosis were all significantlyassociated with PHE. In our study, only a high CT score

Conflict of interest None.

was significantly related to PHE in cirrhosis patients withPH. It is true that CE is useful for detecting PHE. Unfor-tunately, CE cannot be used routinely for cirrhosis patientswith PH in Korea because of the cost. Therefore, we pro-posed a new CT scoring system, which could better selectpatients who should undergo CE because of the suspectedpresence of PHE, especially portal hypertensive patientswith OGIB.

Although argon plasma coagulation, thalidomide, ortransjugular intrahepatic portosystemic shunt placementhave been reported as treatments [15], no recommendationsfor the management of bleeding due to PHE exist. In ourstudy, more than 60 % of the patients with PHE weretreated conservatively. Regardless of the type of therapy,however, CE could differentiate conservative treatmentfrom the need for further treatment or intensive care. CE isimportant in cirrhosis patients with PH because it enablesthe identification of treatable small bowel lesions, and canhelp to optimize treatment on a case-by-case basis.

We recognize the potential limitations of this study.First, it included a small number of patients and had thedisadvantage that no control group was included. Second,this was a retrospective analysis and PH-related diseaseswere not well defined, so differences in interpretationbetween institutions could exist. Third, patient selectionbias might have existed because the data were selectedbased on a database that included mainly tertiary-carehospitals. In Korea, however, most cirrhosis patients withPH underwent CE and were treated at tertiary-care hospi-

tals because of the severity of disease and availability ofCE. Therefore, we believe that the domestic situation couldminimize the unpredictable selection bias. Fourth, eachfinding in our CT scoring system was not validated.Finally, we did not acquire sufficient follow-up databecause of the limitations of the registry. Nevertheless, webelieve that this study has value because the prevalence ofPHE and PHE-related bleeding, and associated clinicalfactors, were identified based on a large database.

In conclusion, the prevalence of PHE was 40 %. PHE is

more prevalent in cirrhosis patients with PH who have ahigh CT score. Although additional larger, prospectivestudies are needed, CE is a useful tool for evaluating PHE

PH.

References

1. De Palma GD, Rega M, Masone S, et al. Mucosal abnormalitiesof the small bowel in patients with cirrhosis and portal hyper-tension: a capsule endoscopy study. Gastrointest Endosc.2005;62:529–534.

2. Viggiano TR, Gostout CJ. Portal hypertensive intestinal vascu-lopathy: a review of the clinical, endoscopic, and histopathologicfeatures. Am J Gastroenterol. 1992;87:944–954.

3. Kodama M, Uto H, Numata M, et al. Endoscopic characterizationof the small bowel in patients with portal hypertension evaluatedby double balloon endoscopy. J Gastroenterol. 2008;43:589–596.

4. Zuckerman GR, Prakash C, Askin MP, Lewis BS. AGA technicalreview on the evaluation and management of occult and obscuregastrointestinal bleeding. Gastroenterology. 2000;118:201–221.

5. Figueiredo P, Almeida N, Lerias C, et al. Effect of portalhypertension in the small bowel: an endoscopic approach. DigDis Sci. 2008;53:2144–2150.

6. Goulas S, Triantafyllidou K, Karagiannis S, et al. Capsuleendoscopy in the investigation of patients with portal hyperten-sion and anemia. Can J Gastroenterol. 2008;22:469–474.

7. Akyuz F, Pinarbasi B, Ermis F, et al. Is portal hypertensive enter-opathy an important additional cause of blood loss in portal hyper-tensive patients? Scand J Gastroenterol. 2010;45:1497–1502.

8. Kovacs M, Pak P, Pak G, Feher J, Racz I. Small bowel alterationsin portal hypertension: a capsule endoscopic study. Hepatogas-troenterology. 2009;56:1069–1073.

9. Saurin JC, Delvaux M, Gaudin JL, et al. Diagnostic value ofendoscopic capsule in patients with obscure digestive bleeding:blinded comparison with video push-enteroscopy. Endoscopy.2003;35:576–584.

10. Pons Beltran V, Gonzalez Suarez B, Gonzalez Asanza C, et al.Evaluation of different bowel preparations for small bowel cap-sule endoscopy: a prospective, randomized, controlled study. DigDis Sci. 2011;56:2900–2905.

11. Chung JM, Seol SY, Jeong SJ, Choi SR, Choe HC. A study onendoscopic classification of esophageal varices. Korean J Gas-troenterol. 1985;17:51–58.

12. McCormack TT, Sims J, Eyre-Brook I, et al. Gastric lesions inportal hypertension: inflammatory gastritis or congestive gas-tropathy? Gut. 1985;26:1226–1232.

13. Canlas KR, Dobozi BM, Lin S, et al. Using capsule endoscopy toidentify GI tract lesions in cirrhotic patients with portal hyperten-sion and chronic anemia. J Clin Gastroenterol. 2008;42:844–848.

14. Rondonotti E, Villa F, Signorelli C, de Franchis R. Portalhypertensive enteropathy. Gastrointest Endosc Clin N Am.2006;16:277–286.

15. Jimenez-Saenz M, Romero-Vazquez J, Caunedo-Alvarez A,Maldonado-Perez B, Gutierrez JM. Beneficial effects and reversionof vascular lesions by thalidomide in a patient with bleeding portalhypertensive enteropathy. Dig Liver Dis. 2010;42:232–233.

123

Reproduced with permission of the copyright owner. Further reproduction prohibited without

permission.