Clinical Gastroenterology and Hepatology 2014;12:571–582

SYSTEMATIC REVIEWS AND META-ANALYSESFasiha Kanwal, Section Editor

Medical and Endoscopic Therapies for Angiodysplasia and GastricAntral Vascular Ectasia: A Systematic Review

Eric Swanson,* Amar Mahgoub,*,‡ Roderick MacDonald,§ and Aasma Shaukat*,‡

*Department of Medicine and ‡Section of Gastroenterology, Veterans Affairs Medical Center and University of Minnesota,Minneapolis; and §Minnesota Evidence based Practice Center, Center for Chronic Disease and Outcome Research,Minneapolis, Minnesota

BACKGROUND & AIMS:

Few studies have compared the efficacy and complications of endoscopic or medical therapiesfor bleeding angiodysplasias or gastric antral vascular ectasias (GAVE). We conducted a sys-tematic review to evaluate therapies.

METHODS:

We performed a PubMed search for studies (written in English from January 1, 1980, throughJanuary 1, 2013) of medical or endoscopic treatment of bleeding angiodysplasias and GAVE.Measured outcomes included levels of hemoglobin, transfusion requirements, rebleeding rates,complications, treatment failures, and overall mortality.

RESULTS:

We analyzed data from 63 studies that met inclusion criteria; 50 evaluated endoscopic treat-ment (1790 patients), 13 evaluated medical treatment (392 patients), and 12 were comparativestudies. In patients with angiodysplasias, the combination of estrogen and progesterone did notsignificantly reduce bleeding episodes, compared with placebo (0.7/y vs 0.9/y, respectively),and increased mortality, compared with conservative therapy (33% vs 21%). A higher per-centage of patients receiving octreotide were free of rebleeding at 1 and 2 years vs placebo(77% vs 55% and 68% vs 36%, respectively; P [ .03). Thalidomide reduced the number ofbleeding episodes (–8.96/y), compared with iron therapy (–1.38/y, P < .01), but neither treat-ment reduced mortality. More patients with GAVE treated by endoscopic band ligation were freefrom rebleeding (92%) than those treated with argon plasma coagulation (32%, P [ .01).

CONCLUSIONS:

In a systematic review, we found a low quality of evidence to support treatment of angiodys-plasias with thalidomide or the combination of estrogen and progesterone and insufficientevidence to support treatment with octreotide. There is also insufficient evidence for endo-scopic therapy of angiodysplasia or GAVE. Well-designed randomized controlled trials areneeded to study the efficacy and complications of medical and endoscopic treatments for pa-tients with angiodysplasias or GAVE.

Keywords: Systematic Review; Angiodysplasia; Gastric Antral Vascular Ectasia; Comparison Analysis.

Abbreviations used in this paper: APC, argon plasma coagulation; EBL,endoscopic band ligation; GAVE, gastric antral vascular ectasia; GI,gastrointestinal; HHT, hereditary hemorrhagic telangiectasia; Nd:YAG,neodymium-doped yttrium aluminum garnet; RCT, randomized controlledtrial; VWD, von Willebrand disease.

© 2014 by the AGA Institute1542-3565/$36.00

http://dx.doi.org/10.1016/j.cgh.2013.08.038

Angiodysplasias and gastric antral vascular ectasia(GAVE) are 2 common causes of iron deficiency

anemia. Angiodysplasias are thin-walled tortuous vesselsfound throughout the gastrointestinal (GI) tract, whereasGAVE is a collection of dilated tortuous vessels thatappear to streak out from the pylorus. Patients with he-reditary hemorrhagic telangiectasias (HHTs) and vonWillebrand disease (VWD) have a high rate of GI malfor-mations, which in turn increases the frequency of clini-cal ly significant bleeding episodes.1 Idiopathicangiodysplasias are more common, typically occur in pa-tients older than 60 years, and are found incidentally in1%–2% of all asymptomatic patients undergoing colo-noscopy.2,3 Most angiodysplasias will never bleed, butthose that do make up 5% of overt upper GI bleeds,4

7% of overt lower GI bleeds,5 and between 65% and

80% of lesions found in obscure GI bleeds.6–9 Many pa-tients with GAVE are also asymptomatic, so the preva-lence in the general population is unknown, but ascreening study in cirrhotic patients reported a preva-lence of 12%.10 GAVE accounts for 4% of all non-variceal bleeds in the general population11 and 6% ofupper GI bleeds in cirrhotic patients.10 These etiologies

572 Swanson et al Clinical Gastroenterology and Hepatology Vol. 12, No. 4

of GI bleeding have significant morbidity and financialimpact, because patients are often hospitalized with eachepisode of bleeding for investigation of anemia and un-dergo endoscopic procedures to find and treat the sourceof bleeding. Surgery is occasionally used for patients whohave failed medical and endoscopic therapies but is asso-ciated with high morbidity and mortality. Only 35% ofpatients were free of complications after surgery for an-giodysplasias,12 with an associated 33% mortality.13 Sur-gery for GAVE has a 50% thirty-day mortality14 andperioperative mortality of 7.4%.15 Transjugular intrahe-patic portosystemic shunt has also been attempted forGAVE with limited success because 7 of 8 people rebledafter transjugular intrahepatic portosystemic shunt inone study,14 and all 14 patients in another studycontinued to require transfusions.16 Even though GAVEis often seen in patients with portal hypertension, suc-cessful reduction in the portal hypertension does notresult in reduced risk of bleeding.

The treatment goals for bleeding angiodysplasias orGAVE include resolution of anemia, reduced transfusionrequirements, and reduction in the incidence ofrebleeding. The current treatment of choice appears tobe endoscopic intervention with argon plasma coagula-tion (APC), whereas systemic medical therapies aretraditionally reserved for patients who have either failedendoscopic treatment or challenging lesions in terms oflocation and number to be managed endoscopically.Endoscopic techniques destroy lesions locally, whereasmedical therapies work systemically. A substantialamount of literature exists on the use of endoscopic andmedical therapies, but there are no recent comprehen-sive systematic reviews evaluating efficacy and harm ofendoscopic or medical therapies. Our goal was toconduct a systematic review on the efficacy and harms ofmedical and endoscopic therapies for angiodysplasia andGAVE.

Methods

Data Collection and Selection

A PubMed search was performed on the Medlinedatabase for all studies in the English language publishedbetween January 1, 1980, and January 1, 2013, includingthe medical subject heading (MeSH) angiodysplasia/therapy (GAVE is categorized within the MeSH termangiodysplasia) and angiodysplasia or gastric antralvascular ectasia cross-referenced with thalidomide,octreotide, estrogen, progesterone, lenalidomide, ator-vastatin, tranexamic acid, argon plasma coagulation,neodymium: yttrium aluminum garnet laser, cryo-therapy, band ligation, bipolar, coagulation, heater probe,and radiofrequency ablation. We searched by hand thereferences of studies included from the original searchand included additional studies that fit inclusion criteria.We also searched the Cochrane database.

Inclusion Criteria

For treatment efficacy, we included studies with acomparator, such as randomized controlled trials (RCTs),and cohort studies with comparators that reported atleast one of our predefined primary outcomes: change inhemoglobin, transfusion requirements, and/or bleedingdata for the therapeutic modality being studied. Forcomplication outcomes we included case series withmore than 5 patients. Complications included adverseevents, perforations, failed therapy, and overallmortality.

Exclusion Criteria

We excluded reviews, editorials, animal studies,nonclinical studies or those with less than 5 patients,studies that were not GAVE or angiodysplasia of the GIsystem, and studies where patients were managed pri-marily with surgery or interventional radiology. Inaddition, we excluded studies where patients wereasymptomatic (without evidence of active bleeding) andhad incidental finding of GAVE and/or angiodysplasia. Aflow chart of this search strategy is shown in Figure 1.

Definitions

We synthesized the results by using the exact defi-nitions the authors used for assessing efficacy andcomplications.

� Change in hemoglobin: Studies reported hemoglobinas the groups’ average change in hemoglobin aftertherapy.

� Transfusion requirements: Studies reported trans-fusion requirements in 3 different ways: the numberof patients requiring transfusions after therapy, thegroups’ average number of units of blood transfusedduring a set period of time after therapy, or thechange in the groups’ average blood transfusion re-quirements after therapy.

� Bleeding: Studies reported bleeding as the totalnumber of bleeding episodes or change in fre-quency of bleeding episodes after therapy during aset time period; the ratio of patients who rebled inthe treatment group; the number of patients withcessation of bleeding or the percent of patients freeof rebleeding during a certain time period. Whenpossible, we converted to a percent of patients freeof rebleeding by dividing the number of patientswho rebled by the total number of patients in thestudy group minus those lost to follow-up (ifavailable).

� Failure rates: A failure of therapy was defined as anypatient who stopped therapy or changed therapy to adifferent endoscopic, surgical, or medical therapy.

Figure 1. Search strategy and results.

April 2014 Therapies for Management of AVMs and GAVE 573

Types of Therapies

Therapies included in this review are estrogen withprogesterone, octreotide, thalidomide, APC, neodymium-doped yttrium aluminum garnet (Nd:YAG), heater probe,monopolar coagulation, bipolar coagulation, sclerother-apy, endoscopic band ligation (EBL), and cryotherapy.

Types of Comparators

We used comparators as reported by the individualstudies. These included placebo, conservative or sup-portive treatment, or another active agent such as heaterprobe or surgical resection.

Data Synthesis and Analysis

This was completed by using recently publishedguidelines.17 We rated study quality by the followingcriteria: adequate allocation concealment based on theapproach by Schulz and Grimes,18 blinding methods(participant, investigator, or outcome assessor), analysisby intention to treat, and lost to follow-up. The quality ofthe overall evidence was rated by the investigators ac-cording to the Grading of Recommendations, Assessment,Development, and Evaluation group.19 A consensus onthe evidence was reached by all investigators on thebasis of the 4 domains (risk of bias, consistency, direct-ness, and precision), and subsequently, the level of evi-dence on each modality of therapy for angiodysplasiaand GAVE was ranked as high, moderate, low, orinsufficient.

Results

A total of 429 citations resulted from the originalsearch, of which only 49 studies fit inclusion criteria,with the largest percentage (35%) being excludedbecause of a sample size less than 5 patients. Fourteenadditional studies that fit inclusion criteria were foundby manually reviewing citations of other included arti-cles. Of the 63 reviewed studies on GAVE and angio-dysplasia, 50 were endoscopic (n ¼ 1790) and 13 weremedical (n ¼ 392), with only 12 studies that used com-parators included in the systematic review (Figure 1).

Angiodysplasia

Medical therapies. These studies are summarized inTable 1.

Estrogen with progesterone. One double-blind RCT(n ¼ 68) showed that compared with placebo, there wasno difference in the number of units of blood transfusedper year (0.9 vs 0.7), change in bleeding episodes peryear (0.7 vs 0.9), or percent of patients free ofrebleeding at 1 and 2 years, respectively (69% vs 55%and 50% vs 36%).20 A second, retrospective study(n ¼ 64) found no difference in units of blood trans-fused per month compared with conservative manage-ment (1.5 vs 1.6).21

Octreotide. A prospective study (n ¼ 65) found thatoctreotide failed to show a difference in units of bloodtransfused per year (1.1 vs 0.7) compared with placebo;however, it did show fewer chronic bleeding episodesper year (0.03 vs 0.2, P ¼ .04) and a higher percent of

Table 1. Studies With a Comparator

Authors(year)

Studytype

Type oftherapy Comparator

No. ofpatientsrequiring

transfusions

Duration ofdrug therapyor mean no.of sessions

(range)

Meanfollow-up,mo (range)

Outcomes

% Withside

effects% Failedtherapy

%Overallmortality

Change inhemoglobin

(g/dL)Transfusionrequirements Bleeding

% Free ofrebleeding

Junqueraet al(2001)

DB-RCT Estrogen withprogresterone(Ethinylestradiol0.01 mg andnorethisterone2 mg daily)

Placebo 68 (33 vs 35) 13.5 mo 13.5 (12–36) NR Units/y: 0.9vs 0.7

Episodes/y:0.7 vs 0.9

1 y: 69 vs 55 45 vs 14a NR 0 vs 32 y: 50 vs 36

Lewiset al(1992)

Retrocohort

Estrogen withprogresterone(Premarin0.625 mg/dayfor 6 patientsand Enovid10 mg/day for24 patients)

Conservative 64 (30 vs 34) 15.6 mo 15.6 (2–31) vs13.4 (1–23)

NR Units/mo: 1.5vs 1.6

NR NR 57 vs NR Stopped: 40vs NRsurgery:17 vs 24

33 vs 21

Junqueraet al(2007)

Prosp cohort Octreotide Placebo 65 (30 vs 35) 13 mo 13 (12–36) NR Units/y: 1.1vs 0.7

Episodes:0.03 vs0.2a

1 y: 77 vs 55a 53 vs 14a Stopped: 3vs 0

0 vs 32 y: 68 vs 36a

Ge et al(2011)

Open-labelRCT

Thalidomide Iron therapy 52 (26 vs 26) 4 mo 39 (8–52) 2.88 vs –0.05a No. of patients:3 vs 13a

D episodes/y: –8.96vs –1.38a

Cessationat 39 mo:42.3 vs 0

73 vs 35 Stopped: 8vs 0

0b

Saperaset al(2009)

Retro cohort APC Conservative 57 (34 vs 23) NR 33 NR NR NR 1 y: 87 vs 73 9 vs NR NR 12b

2 y: 74 vs 52

Askinet al(1996)

Retro cohort Heater probe Conservative 83 (55 vs 28) 1.94 (1–7) 30 vs 26 NR D units/mo: 2.1vs 0.96

NR NR NR Changed:11 vs 25

NR

Guptaet al(1995)

Retro cohort Heater probe,surgery

Transfusion,observation

32 (16, 9 vs 4, 3) 1.25 (1–2) 13.5 (3–42) NR NR NR 13.5 mo: 66,100 vs66, NR

0, NR vsNR,NR

Surgery: 6,0vs 0,0

6, 33 vs0, 0

Richteret al(1989)

Retro cohort Monopolarcoagulation,surgery

Conservative,asymptomatic

101 (19, 31vs 36, 15)

NR 22 (1–120) NR NR NR 1 y: 66, 84 vs74, 100

5, 0 vs0, 0

Surgery: 11,0 vs 0, 0

22b

3 y: 47, 76 vs54, 100

574Swanson

etal

Clinical

Gastroenterology

andHepatology

Vol.

12,No.

4

Trud

el etal

(198

8)

Retro

coho

rtMon

opolar

coag

ulation,

surgery

Con

servative

56(28,

17vs

11)

1.2(1–2)

17.4

(1–79

.5)

NR

NR

NR

17.4

mo:

50,

76vs

550,

0vs

0Cha

nged

:25

,18

vs6

23b

Satoet

al(201

2)c

Retro

coho

rtEBL

APC

34(12vs

22)

3.0vs

2.3

14.6

vs16

.6NR

NR

NR

EBL:

14.6

mo,

928vs

00vs

016

vs32

APC:16

.6mo,

32a

Wells

etal

(200

8)c

Retro

coho

rtEBL

APC

22(9

vs13

)1.4vs

4.7a

10.1

(2–24

)vs

15.3

(1–35

)

2.81

vs0.9a

D–12

.7vs

–5.2a

Ces

satio

n:5/9vs

3/13

a

NR

11vs

8Surge

ry:

0vs

033

vs38

Van

cutsem

etal

(199

3)d

RCT cr

osso

verEstroge

nwith

progres

terone

(Ethinyles

trad

iol

0.01

mgan

dno

rethisterone

2mgdaily)

Place

bo

14(8

HHT,

1HHT

andVWD,5

angiod

ysplasia)

6mo

20(12–

29)

NR

No.

ofpatients:

3/13

(trea

tmen

t)vs

12/13

(place

bo)

NR

20mo:

2538

vs23

7vs

07vs

14

DB,dou

ble-blind;NR,no

treported;Prosp

,prosp

ectiv

e;Retro,retros

pec

tive;

D,Cha

nge.

a Significa

ntly

differen

t.bStudyco

mbined

grou

psformortality.

cStudyon

GAVE.

dStudyon

HHT.

April 2014 Therapies for Management of AVMs and GAVE 575

patients free of rebleeding at 1 year and 2 years (77% vs55% and 68% vs 36%, P ¼ .03).22

Thalidomide. An open-label RCT (n ¼ 52) comparedlow-dose thalidomide with a positive control of ironsupplements for 4 months.23 Compared with iron,thalidomide had a significant rise in hemoglobin (2.88 vs–0.05 g/dL, P < .01), fewer patients needing trans-fusions (3 vs 13, P < .01), and a larger change in thenumber of bleeding episodes per year (–8.96 vs –1.38,P < .01).23

The quality of the studies was low, with only 2 RCTsfound. Studies were small in size, and reporting of out-comes was variable. We found low evidence for efficacyof thalidomide and estrogen with progesterone, withinsufficient evidence of efficacy of octreotide for angio-dysplasia (Supplementary Table 1).

Complications. The RCT found estrogen with proges-terone had more side effects than placebo (45% vs 14%,P < .01) and similar mortality (0% vs 3%) (Table 1).20

The prospective study found side effects in 57% oftreated patients, with similar rates to placebo requiringsurgery (17% vs 24%) and a higher overall mortalityrate (33% vs 21%).21 Octreotide had more side effectsthan placebo (53% vs 14%, P ¼ .01), diarrhea being mostfrequent, with similar percentages stopping therapy (3%vs 0%) and similar mortality (0% vs 3%).22 Thalidomidehad more side effects than iron therapy (73% vs 35%,P < .01); fatigue, constipation, and dizziness being mostcommon, with similar stopping rates (8% vs 0%) and thesame mortality rate (0% vs 0%).23

Case series. Three case series (n ¼ 66) evaluatedestrogen with progesterone and found side effects in32% (0%–44%) of patients, a 20% (0%–44%) stoppingrate, and a 23% (0%–28%) overall mortality(Table 2).8,24,25 Octreotide had 3 case series (n ¼ 45)that found a complication rate of 2.2% (0%–8%), failurerate of 13% (0%–46%), and no mortality (0%).26–28 Asingle case series on thalidomide (n ¼ 12) found 17% ofpatients had side effects, 25% stopped therapy, and nomortality (0%).29

Endoscopic therapies. These studies are summarizedin Table 1.

Argon plasma coagulation. A retrospective study(n ¼ 57) found no difference in the percent of patientsfree of rebleeding compared with conservative therapyat 1 or 2 years, respectively (87% vs 73% and 74% vs52%, P ¼ .06).30

Heater probe. Two retrospective studies fit inclusioncriteria.13,31 The larger study (n ¼ 83) evaluated thechange in transfusion requirements before and aftertherapy in patients treated with heater probe (n ¼ 55)versus those managed conservatively (n ¼ 28) and foundno difference (2.1 vs 0.96, P ¼ .19).31 The second study(n ¼ 32) had 4 arms of therapy: heater probe (n ¼ 16),surgery (n ¼ 9), transfusions only (n ¼ 4), and obser-vation (n ¼ 3). There was no difference in percent free ofrebleeding between heater probe and the 3 comparators,respectively (66%, 100%, 66%, 100%).13

Table 2. Complications, Perforations, Failures, and Mortality in Case Series

Lesion Type of therapyNo. ofstudies

Sum ofpatients

Weighted mean ofaverage follow-up, moa

% Withcomplicationsa

% Withperforationsa Failure ratea (%)

% Overallmortalitya

% of overall mortalityfrom bleedingor therapy

complicationa

AngiodysplasiaEstrogen with progesterone 3 66 16 (12–18) 32 (0–44) NR 20 (0–44) 23 (0–28) 0Octreotide 3 45 29 (14–39) 2.2 (0–8) NR 13 (0–46) 0 0Thalidomide 1 12 24–36 17 NR 25 0 0EBL 2 29 16 (12–18) 14 (11–18) 0.0 0.0 14 (11–18) 0APC 7 493 19 (6–55) 2.4 (0–6) 0.6 (0–2) 1.6 (0–14) 9 (5–34) 19 (0–100)Nd:YAG laser 6 359 15 (12–24) 13 (4–32) 2.6 (0–5) 8 (1–17) 8 (3–27) 6 (0–9)Heater probe 2 34 7 4 0 0 9 0Monopolar coagulation 2 50 17 4 (0–9) 2 6 NR NRBipolar coagulation 1 11 11 0 0 18 0 0Sclerotherapy 1 8 29 25 0 0 13 100

GAVEEstrogen with progesterone 1 6 8 50.0 NR 0 NR NRAPC 10 192 17 (9–30) 21 (0–80) 0.0 1.8 (1–14) 25 (0–57) 12 (0–50)Nd:YAG laser 8 185 30 (9–55) 20 (0–51) 1.6 (0–14) 6 (0–13) 28 (9–63) 10 (0–25)Heater probe 1 12 21 33 0 0 0 0Monopolar coagulation 1 6 NR 33 0 0 NR NRCryotherapy 2 26 3 20 (7–33) 0 0 NR NR

aWhen reported.

576Swanson

etal

Clinical

Gastroenterology

andHepatology

Vol.

12,No.

4

April 2014 Therapies for Management of AVMs and GAVE 577

Monopolar coagulation. Two retrospective studies fitinclusion criteria.32,33 One study (n ¼ 101) had 4 arms oftherapy: monopolar coagulation (n ¼ 19), right hemi-colectomy with ileal resection (n ¼ 31), conservativetherapy (n ¼ 36), and observation of asymptomatic pa-tients (n ¼ 15).32 The percent of patients free ofrebleeding was no different between monopolar coagu-lation, surgery, and medical management at 1 year (66%,84%, 74%) or 3 years (47%, 76%, 54%), whereas 100%of patients with asymptomatic lesions were free ofbleeding.32 The second study (n ¼ 56) had 3 arms oftherapy: monopolar coagulation (n ¼ 28), surgicalresection (n ¼ 17), and transfusions only (n ¼ 11).33

This study found no difference in the percent free ofrebleeding between monopolar coagulation and trans-fusions (50% vs 55%, P ¼ .79) or coagulation or trans-fusion versus surgery (76%; P ¼ .08 and P ¼ .21respectively).33

The quality of the studies was low with no RCTs.Studies were all retrospective and small in size, andreporting of outcomes was variable. We found insuffi-cient evidence of efficacy for any single endoscopictherapy for angiodysplasias (Supplementary Table 1).

Complications. The study on APC only reported APC-induced hemorrhage and found a higher rate with ther-apy (9% vs 0%) (Table 1).30 One study found fewerpatients change therapy with heater probe than withconservative management (11% vs 25%).31 The secondstudy reported no side effects with heater probe (0%), aneed for surgery in 6% of patients, and a lower mortalitywith heater probe than surgery but higher than trans-fusion and observation (6%, 33% vs 0%, 0%).13

Monopolar coagulation had higher adverse events thanall 3 comparators in one study (5%, 0% vs 0%, 0%),32

whereas it had equal percentages to surgery and con-servative management (0%, 0% vs 0%) in the secondstudy.33 Surgery to abolish rebleeding was required at ahigher rate in monopolar coagulation than in the 3comparators (10%, 0% vs 0%, 0%).32 Monopolar coag-ulation had more patients change therapy (25%, 18% vs6%),33 and both studies reported overall mortality of allpatients in the study at 22% and 23%.32,33

Case series. There were 2 case series on EBL (n ¼ 29),which found complications in 14% (11%–18%) of pa-tients, no perforations or failed therapy (0%), and anoverall mortality of 14% (11%–18%) (Table 2).34,35

There were 7 case series on APC (n ¼ 493) that foundcomplications in 2.4% (0%–6%) of patients, perforationsin 0.6% (0%–2%), 1.6% (0%–14%) failed therapy,and 9% (5%–34%) overall mortality, whereas 19% (0%–100%) of the mortality was bleeding related.36–42 Therewere 6 case series on Nd:YAG (n ¼ 359) that found13% of patients (0%–32%) had complications,2.6% (0%–5%) had perforations, 8% (1%–17%) failedtherapy, and 8% (3%–27%) overall mortality, with 6%(0%–9%) of mortality therapy or bleeding related.43–48

There were 2 case series on heater probe (n ¼ 34),with a 4% complication rate, no perforations or therapy

failures (0%), and 9% overall mortality.49,50 There were2 case series on monopolar coagulation (n ¼ 50); 4% ofpatients (0%–9%) had complications, 2% had perfora-tions, 6% failed therapy, and neither study reportedmortality data.51,52 Bipolar coagulation had a single caseseries (n ¼ 11), with no complications or perforations(0%), whereas 18% of patients required surgery, andthere was no mortality (0%).53 A single case series onsclerotherapy with ethanolamine (n ¼ 8) found 25% ofpatients had complications, and there were no perfora-tions or therapy failures (0%) and 13% overall mortality(1 death), with 100% of mortality being a therapy-related complication.54

Gastric Antral Vascular Ectasia

Medical therapies. No studies with comparatorsevaluated medical therapy for GAVE.

Case series. A single case series (n ¼ 6) on estrogenwith progesterone found 50% of patients had side ef-fects, none failed therapy, and no mortality data werereported (Table 2).55

Endoscopic therapies. These studies are summarizedin Table 1.

Endoscopic band ligation. Two retrospective studiescompared EBL with APC.56,57 Of the patients in the EBLtreatment arms, 50% and 44% had previously failed APCtherapy before enrolling in the study.56,57 The first studyincluded 34 cirrhotic patients with class A (n ¼ 6), classB (n ¼ 21), and class C (n ¼ 7) and found a higherpercent of patients free of rebleeding with EBL than withAPC (92% vs 32%, P ¼ .01).56 The second study (n ¼ 22)included 9 patients treated with EBL and 13 treated withendoscopic thermal therapy: APC only (n ¼ 10), bipolarcoagulation only (n ¼ 1), or both APC and bipolarcoagulation (n ¼ 2). Ten of 22 patients (45%) hadcirrhosis. This study found EBL required fewer treatmentsessions (1.9 vs 4.7, P ¼ .05), had a greater increase inhemoglobin (2.8 vs 0.9 g/dL, P ¼ .05), a greater decreasein the number of transfused units of blood (–12.7 vs –5.2,P ¼ .02), and a higher percent of patients who hadcessation of bleeding (56% vs 23%, P ¼ .04).57 Outcomesby underlying portal hypertension were not reported.

The quality of the studies was low with no RCTs.Studies were both retrospective and small in size, andreporting of outcomes was variable. We found insuffi-cient evidence of efficacy for either endoscopic therapyfor GAVE (Supplementary Table 1).

Complications. Higher rates of complications werereported with EBL than with APC in both studies (8% vs0% and 11% vs 8%), with no failures in EBL or APC anda lower mortality with EBL than with APC (16% vs 32%and 33% vs 38%) (Table 1).56,57

Case series. There were 10 case series on APC(n ¼ 192) that found 21% of patients (0%–80%) hadadverse events, no perforations (0%), 2% (1%–14%)failed therapy with 25% (0%–57%) overall mortality,and 12% (0%–50%) of mortality was therapy or

Table 3. Evidence for Therapies

Type of therapy Angiodysplasia GAVE

Estrogen with progesterone – IOctreotide þ XThalidomide þ XEBL I IAPC I –

Nd:YAG laser I –

Heater probe I IMonopolar coagulation I IBipolar coagulation I XSclerotherapy I XCryotherapy X I

þ, Evidence for use; –, evidence against use; I, insufficient data; X, no data.

578 Swanson et al Clinical Gastroenterology and Hepatology Vol. 12, No. 4

bleeding related (Table 2).58–67 There were 10 case se-ries on Nd:YAG therapy (n ¼ 185) that found 20% ofpatients (0%–51%) had complications, 1.6% (0%–14%)had perforations, 6% (0%–13%) failed therapy, andthere was 28% (9%–63%) overall mortality, and10% (0%–25%) of mortality was therapy or bleedingrelated.68–75 A single case series on heater probe(n ¼ 12) found 33% of patients had complications, butthere were no perforations, therapy failures, or mortality(0%).76 A single case series on monopolar coagulation(n ¼ 6) found 33% of patients had complications, noperforations or therapy failures (0%), and mortality datawere not reported.10 Two case series on cryotherapy(n ¼ 26) found 20% of patients (7%–33%) had com-plications, no perforations or therapy failures (0%), andmortality data were not reported.77,78

We also pooled all studies on APC and Nd:YAG(angiodysplasia and GAVE) that reported perforationoutcomes. There were fewer perforations with APCcompared with Nd:YAG (3 of 662 vs 11 of 484).

Hereditary Hemorrhagic Telangiectasiaand Von Willebrand Disease

There were 2 randomized controlled crossover trialsthat included patients with HHT; however, the firststudy was not included because it was replicated in thesecond study plus 3 additional patients (Table 1).79,80

The crossover trial (n ¼ 14) had 8 patients with HHT,1 with HHT and VWD, and 5 with angiodysplasia.Therapy consisted of 6 months of estrogen with pro-gesterone, followed by 6 months of placebo to a groupof 8 patients, and vice versa to a group of 6. This studyfound fewer patients needed transfusions when activelybeing treated with estrogen and progesteronecompared with placebo (3 of 13 vs 12 of 13, P < .01),but all patients in follow-up (8 of 8) required trans-fusions after therapy was completed, even though thestudy reported 25% of patients were free ofrebleeding.79

Complications. The single study reported more com-plications in patients who received estrogen with pro-gesterone than placebo (38% vs 23%), a higher rate ofpatients requiring surgery (7% vs 0%), and similarmortality (7% vs 14%) (Table 1).

Case series. There were 6 larger case series that hadsubsets of patients with HHT. Because most patients inthose studies had either angiodysplasia or GAVE and thecomplications were reported for all patients in the study,these results were discussed in the GAVE or angiodys-plasia sections.40,43–46,48

Discussion

Overall, we found insufficient or low evidenceregarding the efficacy for currently used modalitiesand agents. Most studies were small, retrospective,

heterogeneous in population and setting, did not reportthe number or location where lesions were found, andhad considerable variation in reporting outcomes acrossand within treatment modalities. Pooling results was notpermissible, and determining generalized estimates ofclinically relevant treatment effects was not feasible. Ourfindings are summarized in Table 3.

The current standard of care, APC therapy, failed toshow a difference in rebleeding versus conservativemanagement in angiodysplasia.30 The other 2 endoscopictherapies studied (monopolar coagulation and heaterprobe) failed to show benefit versus conservative ther-apy in angiodysplasia. For studies on GAVE, APC wasinferior to EBL at reducing rebleeding in GAVE,56,57 andno study has compared the efficacy of EBL with con-servative therapy. Of note, these studies included a largepercentage (100% and 45%) of patients with cirrhosisand therefore may not be applicable to all patients withGAVE.

We found low evidence for efficacy of estrogen withprogesterone for angiodysplasia because 2 studiesdemonstrated no improvement in clinical outcomes andshowed high complication rates and higher mortalitythan conservative therapy.20,21 Octreotide showeddecreased risk of bleeding with equal rates of mortalityversus conservative therapy but was of insufficient evi-dence because of the study design, small number of pa-tients, and unconcealed allocation of patients.22 Althoughthalidomide showed a decrease in the risk of rebleedingand equal rates of mortality versus conservative therapy,we rated it as low evidence because the RCT was openlabeled, with a small number of patients and large con-fidence intervals.23

Because of the small number of patients, retrospec-tive nature, and uncontrolled patient populations, wefound insufficient evidence to determine the efficacy ofAPC, heater probe, and monopolar coagulation.13,30–33

Although EBL appears superior to APC with regard tobleeding and mortality for the treatment of GAVE, thedata are from 2 very small, nonrandomized patientpopulations and therefore of insufficient evidence.

April 2014 Therapies for Management of AVMs and GAVE 579

For patients with GI bleeding from HHT, we foundinsufficient evidence to determine the efficacy of any onetherapy.

Endoscopic techniques appear effective for initialhemostasis by destruction of the lesion, but none havebeen shown to prevent long-term bleeding better thanconservative therapy. Although octreotide and thalido-mide work by different mechanisms than endoscopictherapies, they similarly have been shown to resolve le-sions.23,27,28,81,82 A possible mechanism for the long-termrebleeding superiority of octreotide and thalidomideover endoscopic therapies is that endoscopic therapiestreat locally, and 17% of patients who rebleed afterendoscopic therapy bleed from new lesions that were notseen during the original endoscopy37; thus, systemictreatment theoretically offers better control of new ormissed lesions on the initial endoscopy. Another possiblemechanism is the prevention of new lesion formationbecause both octreotide and thalidomide inhibit vascularformation.83,84 However, we found no studies thatassessed the rate of formation of new lesions while onthese therapies.

Complications

There was great variability between studies on whatwas reported as a complication. Medical therapies ten-ded to report all side effects of therapy even if veryminor (eg, bitter taste and dry eyes)23; therefore, it wasnot surprising to see more side effects than for thosereceiving placebo. Studies on endoscopic therapyseemed to only report major complications. It is worthnoting that the complications from thalidomide andoctreotide seem to be less serious because no patientsrequired antibiotics for bacteremia, surgery for perfo-rations, or antral stenosis. In fact, the most severecomplication reported in thalidomide was a reversiblepolyneuropathy,29 whereas 1 patient on octreotidesuffered a stroke, which was the same frequency as forplacebo.22

Perforations

Similar to other reports,85,86 we found that APC hasfewer perforations than Nd:YAG, and notably no perfo-rations were observed in EBL studies.

Failure Rates

There was substantial variability in what was re-ported as a failure of therapy. Medical therapies tendedto report the number of patients stopping because ofintolerance to medications. Estrogen with progesteronehad the highest rate of patients stopping because ofintolerance, followed by thalidomide and octreotide.Endoscopic therapies reported the percentage of patientswho went to surgery or changed therapy. Because of the

variability of how a failure was defined, it is difficult tocompare among studies.

Mortality

Estrogen with progesterone was associated with ahigher rate of mortality, whereas octreotide and thalid-omide had similar rates to conservative therapy. Only1 endoscopic study on angiodysplasia compared mor-tality with transfusion only therapy, whereas the rest didnot report mortality data. Two small studies on GAVE(n ¼ 56) that compared EBL with APC found lowermortality with EBL. Additional well-designed prospectivestudies are needed to determine whether there is atherapy-related mortality benefit.

Limitations of our review include only searchingstudies in the English language, not contacting authors,and limitations of the data because of the lack of uniformreporting among articles, most importantly includinglocation and number of angiodysplasias and use of con-founding drugs such as steroids, nonsteroidal anti-inflammatory drugs, antiplatelet, and anticoagulantsthat can affect initial bleeding, rebleeding, and trans-fusion requirements.

Future rigorous double-blind RCTs are needed toevaluate treatment efficacy and should clearly delineatethe location of angiodysplasias and the number of lesionstreated and use standardized measures for reportingoutcomes. These studies should compare medical andendoscopic agents against one another, placebo, andconservative management. Last, there needs to berigorous long-term safety data for these interventionswith standardized complications reporting, which pres-ently is lacking in most studies.

Supplementary Material

Note: To access the supplementary material accom-panying this article, visit the online version of ClinicalGastroenterology and Hepatology at www.cghjournal.org,and at http://dx.doi.org/10.1016/j.cgh.2013.08.038.

References

1. Kjeldsen AD, Kjeldsen J. Gastrointestinal bleeding in patients

with hereditary hemorrhagic telangiectasia. Am J Gastroenterol2000;95:415–418.

2. Foutch PG, Rex DK, Lieberman DA. Prevalence and naturalhistory of colonic angiodysplasia among healthy asymptomaticpeople. Am J Gastroenterol 1995;90:564–567.

3. Diggs NG, Holub JL, Lieberman DA, et al. Factors thatcontribute to blood loss in patients with colonic angiodysplasiafrom a population-based study. Clin Gastroenterol Hepatol2011;9:415–420, quiz e49.

4. Fleischer D. Etiology and prevalence of severe persistentupper gastrointestinal bleeding. Gastroenterology 1983;84:538–543.

5. Tedesco FJ, Waye JD, Raskin JB, et al. Colonoscopic evaluationof rectal bleeding: a study of 304 patients. Ann Intern Med 1978;89:907–909.

580 Swanson et al Clinical Gastroenterology and Hepatology Vol. 12, No. 4

6. Lewis BS, Waye JD. Chronic gastrointestinal bleeding ofobscure origin: role of small bowel enteroscopy. Gastroenter-ology 1988;94:1117–1120.

7. Lewis BS, Kornbluth A, Waye JD. Small bowel tumours: yield ofenteroscopy. Gut 1991;32:763–765.

8. Schmit A, Gay F, Adler M, et al. Diagnostic efficacy of push-enteroscopy and long-term follow-up of patients with smallbowel angiodysplasias. Dig Dis Sci 1996;41:2348–2352.

9. Ell C, Remke S, May A, et al. The first prospective controlled trialcomparing wireless capsule endoscopy with push enteroscopyin chronic gastrointestinal bleeding. Endoscopy 2002;34:685–689.

10. Yamada M. Gastroduodenal vascular ectasia in patients withliver cirrhosis. Dig Endosc 1999;11:241–245.

11. Dulai GS, Jensen DM, Kovacs TO, et al. Endoscopic treatmentoutcomes in watermelon stomach patients with and withoutportal hypertension. Endoscopy 2004;36:68–72.

12. Smith GF, Ellyson JH, Parks SN, et al. Angiodysplasia of thecolon: a review of 17 cases. Arch Surg 1984;119:532–536.

13. Gupta N, Longo WE, Vernava AM 3rd. Angiodysplasia of thelower gastrointestinal tract: an entity readily diagnosed by co-lonoscopy and primarily managed nonoperatively. Dis ColonRectum 1995;38:979–982.

14. Spahr L, Villeneuve JP, Dufresne MP, et al. Gastric antralvascular ectasia in cirrhotic patients: absence of relation withportal hypertension. Gut 1999;44:739–742.

15. Borsch G. Diffuse gastric antral vascular ectasia: the “water-melon stomach” revisited. Am J Gastroenterol 1987;82:1333–1334.

16. Kamath PS, Lacerda M, Ahlquist DA, et al. Gastric mucosalresponses to intrahepatic portosystemic shunting in patientswith cirrhosis. Gastroenterology 2000;118:905–911.

17. Kanwal F, White D. Systematic reviews and meta-analyses inClinical Gastroenterology and Hepatology. Clin GastroenterolHepatol 2012;10:1184–1186.

18. Schulz KF, Grimes DA. Allocation concealment in randomisedtrials: defending against deciphering. Lancet 2002;359:614–618.

19. Atkins D, Best D, Briss PA, et al. Grading quality of evidence andstrength of recommendations. BMJ 2004;328:1490.

20. Junquera F, Feu F, Papo M, et al. A multicenter, randomized,clinical trial of hormonal therapy in the prevention of rebleedingfrom gastrointestinal angiodysplasia. Gastroenterology 2001;121:1073–1079.

21. Lewis BS, Salomon P, Rivera-MacMurray S, et al. Does hor-monal therapy have any benefit for bleeding angiodysplasia?J Clin Gastroenterol 1992;15:99–103.

22. Junquera F, Saperas E, Videla S, et al. Long-term efficacy ofoctreotide in the prevention of recurrent bleeding from gastro-intestinal angiodysplasia. Am J Gastroenterol 2007;102:254–260.

23. Ge ZZ, Chen HM, Gao YJ, et al. Efficacy of thalidomide for re-fractory gastrointestinal bleeding from vascular malformation.Gastroenterology 2011;141:1629–1637.

24. Barkin JS, Ross BS. Medical therapy for chronic gastrointestinalbleeding of obscure origin. Am J Gastroenterol 1998;93:1250–1254.

25. Bronner MH, Pate MB, Cunningham JT, et al. Estrogen-progesterone therapy for bleeding gastrointestinal telangiecta-sias in chronic renal failure: an uncontrolled trial. Ann Intern Med1986;105:371–374.

26. Bon C, Aparicio T, Vincent M, et al. Long-acting somatostatinanalogues decrease blood transfusion requirements in pa-tients with refractory gastrointestinal bleeding associatedwith angiodysplasia. Aliment Pharmacol Ther 2012;36:587–593.

27. Nardone G, Rocco A, Balzano T, et al. The efficacy of octreotidetherapy in chronic bleeding due to vascular abnormalities of thegastrointestinal tract. Aliment Pharmacol Ther 1999;13:1429–1436.

28. Scaglione G, Pietrini L, Russo F, et al. Long-acting octreotide asrescue therapy in chronic bleeding from gastrointestinal angio-dysplasia. Aliment Pharmacol Ther 2007;26:935–942.

29. Garrido A, Sayago M, Lopez J, et al. Thalidomide in refractorybleeding due to gastrointestinal angiodysplasias. Rev EspEnferm Dig 2012;104:69–71.

30. Saperas E, Videla S, Dot J, et al. Risk factors for recurrence ofacute gastrointestinal bleeding from angiodysplasia. Eur JGastroenterol Hepatol 2009;21:1333–1339.

31. Askin MP, Lewis BS. Push enteroscopic cauterization: long-term follow-up of 83 patients with bleeding small intestinalangiodysplasia. Gastrointest Endosc 1996;43:580–583.

32. Richter JM, Christensen MR, Colditz GA, et al. Angiodysplasia:natural history and efficacy of therapeutic interventions. Dig DisSci 1989;34:1542–1546.

33. Trudel JL, Fazio VW, Sivak MV. Colonoscopic diagnosis andtreatment of arteriovenous malformations in chronic lowergastrointestinal bleeding: clinical accuracy and efficacy. DisColon Rectum 1988;31:107–110.

34. Junquera F, Brullet E, Campo R, et al. Usefulness of endoscopicband ligation for bleeding small bowel vascular lesions. Gas-trointest Endosc 2003;58:274–279.

35. Ljubicic N. Endoscopic detachable mini-loop ligation for treat-ment of gastroduodenal angiodysplasia: case study of 11 pa-tients with long-term follow-up. Gastrointest Endosc 2004;59:420–423.

36. Cello JP, Grendell JH. Endoscopic laser treatment for gastro-intestinal vascular ectasias. Ann Intern Med 1986;104:352–354.

37. Eickhoff A, Enderle MD, Hartmann D, et al. Effectiveness andsafety of PRECISE APC for the treatment of bleeding gastroin-testinal angiodysplasia: a retrospective evaluation.Z Gastroenterol 2011;49:195–200.

38. Chiu YC, Lu LS, Wu KL, et al. Comparison of argon plasmacoagulation in management of upper gastrointestinal angio-dysplasia and gastric antral vascular ectasia hemorrhage. BMCGastroenterol 2012;12:67.

39. Kwan V, Bourke MJ, Williams SJ, et al. Argon plasma coagu-lation in the management of symptomatic gastrointestinalvascular lesions: experience in 100 consecutive patients withlong-term follow-up. Am J Gastroenterol 2006;101:58–63.

40. May A, Friesing-Sosnik T, Manner H, et al. Long-term outcomeafter argon plasma coagulation of small-bowel lesions usingdouble-balloon enteroscopy in patients with mid-gastrointestinalbleeding. Endoscopy 2011;43:759–765.

41. Olmos JA, Marcolongo M, Pogorelsky V, et al. Long-termoutcome of argon plasma ablation therapy for bleeding in 100consecutive patients with colonic angiodysplasia. Dis ColonRectum 2006;49:1507–1516.

42. Olmos JA, Marcolongo M, Pogorelsky V, et al. Argon plasmacoagulation for prevention of recurrent bleeding from GI angio-dysplasias. Gastrointest Endosc 2004;60:881–886.

April 2014 Therapies for Management of AVMs and GAVE 581

43. Bown SG, Swain CP, Storey DW, et al. Endoscopic lasertreatment of vascular anomalies of the upper gastrointestinaltract. Gut 1985;26:1338–1348.

44. Gostout CJ, Bowyer BA, Ahlquist DA, et al. Mucosal vascularmalformations of the gastrointestinal tract: clinical observationsand results of endoscopic neodymium: yttrium-aluminum-garnet laser therapy. Mayo Clin Proc 1988;63:993–1003.

45. Mathus-Vliegen EM. Laser treatment of intestinal vascularabnormalities. Int J Colorectal Dis 1989;4:20–25.

46. Naveau S, Aubert A, Poynard T, et al. Long-term results oftreatment of vascular malformations of the gastrointestinal tractby neodymium YAG laser photocoagulation. Dig Dis Sci 1990;35:821–826.

47. Rutgeerts P, Van Gompel F, Geboes K, et al. Long term resultsof treatment of vascular malformations of the gastrointestinaltract by neodymium Yag laser photocoagulation. Gut 1985;26:586–593.

48. Sargeant IR, Loizou LA, Rampton D, et al. Laser ablation ofupper gastrointestinal vascular ectasias: long term results. Gut1993;34:470–475.

49. Hayat M, Axon AT, O’Mahony S. Diagnostic yield and effect onclinical outcomes of push enteroscopy in suspected small-bowel bleeding. Endoscopy 2000;32:369–372.

50. Shackel NA, Bowen DG, Selby WS. Video push enteroscopy inthe investigation of small bowel disease: defining clinical in-dications and outcomes. Aust N Z J Med 1998;28:198–203.

51. Howard OM, Buchanan JD, Hunt RH. Angiodysplasia of thecolon: experience of 26 cases. Lancet 1982;2:16–19.

52. Rogers BH. Endoscopic diagnosis and therapy of mucosalvascular abnormalities of the gastrointestinal tract occurringin elderly patients and associated with cardiac, vascular,and pulmonary disease. Gastrointest Endosc 1980;26:134–138.

53. Foutch PG, Sawyer R, Sanowski RA. Push-enteroscopy fordiagnosis of patients with gastrointestinal bleeding of obscureorigin. Gastrointest Endosc 1990;36:337–341.

54. Bemvenuti GA, Julich MM. Ethanolamine injection for sclero-therapy of angiodysplasia of the colon. Endoscopy 1998;30:564–569.

55. Tran A, Villeneuve JP, Bilodeau M, et al. Treatment of chronicbleeding from gastric antral vascular ectasia (GAVE) withestrogen-progesterone in cirrhotic patients: an open pilot study.Am J Gastroenterol 1999;94:2909–2911.

56. Sato T, Yamazaki K, Akaike J. Endoscopic band ligationversus argon plasma coagulation for gastric antral vascularectasia associated with liver diseases. Dig Endosc 2012;24:237–242.

57. Wells CD, Harrison ME, Gurudu SR, et al. Treatment of gastricantral vascular ectasia (watermelon stomach) with endoscopicband ligation. Gastrointest Endosc 2008;68:231–236.

58. Bhatti MA, Khan AA, Alam A, et al. Efficacy of argon plasmacoagulation in gastric vascular ectasia in patients with livercirrhosis. J Coll Physicians Surg Pak 2009;19:219–222.

59. Chaves DM, Sakai P, Oliveira CV, et al. Watermelon stomach:clinical aspects and treatment with argon plasma coagulation.Arq Gastroenterol 2006;43:191–195.

60. Fuccio L, Zagari RM, Serrani M, et al. Endoscopic argon plasmacoagulation for the treatment of gastric antral vascular ectasia-related bleeding in patients with liver cirrhosis. Digestion 2009;79:143–150.

61. Herrera S, Bordas JM, Llach J, et al. The beneficial effectsof argon plasma coagulation in the management of differenttypes of gastric vascular ectasia lesions in patientsadmitted for GI hemorrhage. Gastrointest Endosc 2008;68:440–446.

62. Lecleire S, Ben-Soussan E, Antonietti M, et al. Bleeding gastricvascular ectasia treated by argon plasma coagulation: a com-parison between patients with and without cirrhosis. Gastro-intest Endosc 2008;67:219–225.

63. Naga M, Esmat S, Naguib M, et al. Long-term effect of argonplasma coagulation (APC) in the treatment of gastric antralvascular ectasia (GAVE). Arab J Gastroenterol 2011;12:40–43.

64. Probst A, Scheubel R, Wienbeck M. Treatment of watermelonstomach (GAVE syndrome) by means of endoscopic argonplasma coagulation (APC): long-term outcome. Z Gastroenterol2001;39:447–452.

65. Roman S, Saurin JC, Dumortier J, et al. Tolerance and efficacyof argon plasma coagulation for controlling bleeding in patientswith typical and atypical manifestations of watermelon stomach.Endoscopy 2003;35:1024–1028.

66. Sebastian S, McLoughlin R, Qasim A, et al. Endoscopic argonplasma coagulation for the treatment of gastric antral vascularectasia (watermelon stomach): long-term results. Dig Liver Dis2004;36:212–217.

67. Yusoff I, Brennan F, Ormonde D, et al. Argon plasma coagula-tion for treatment of watermelon stomach. Endoscopy 2002;34:407–410.

68. Bourke MJ, Hope RL, Boyd P, et al. Endoscopic laser therapyfor watermelon stomach. J Gastroenterol Hepatol 1996;11:832–834.

69. Calamia KT, Scolapio JS, Viggiano TR. Endoscopic YAG lasertreatment of watermelon stomach (gastric antral vascular ecta-sia) in patients with systemic sclerosis. Clin Exp Rheumatol2000;18:605–608.

70. Geller A, Gostout CJ, Balm RK. Development of hyperplasticpolyps following laser therapy for watermelon stomach. Gas-trointest Endosc 1996;43:54–56.

71. Gostout CJ, Viggiano TR, Ahlquist DA, et al. The clinical andendoscopic spectrum of the watermelon stomach. J Clin Gas-troenterol 1992;15:256–263.

72. Liberski SM, McGarrity TJ, Hartle RJ, et al. The watermelonstomach: long-term outcome in patients treated with Nd:YAGlaser therapy. Gastrointest Endosc 1994;40:584–587.

73. Mathou NG, Lovat LB, Thorpe SM, et al. Nd:YAG laser induceslong-term remission in transfusion-dependent patients withwatermelon stomach. Lasers Med Sci 2004;18:213–218.

74. Potamiano S, Carter CR, Anderson JR. Endoscopic lasertreatment of diffuse gastric antral vascular ectasia. Gut 1994;35:461–463.

75. Selinger RR, McDonald GB, Hockenbery DM, et al. Efficacy ofneodymium:YAG laser therapy for gastric antral vascular ectasia(GAVE) following hematopoietic cell transplant. Bone MarrowTransplant 2006;37:191–197.

76. Petrini JL Jr, Johnston JH. Heat probe treatment for antralvascular ectasia. Gastrointest Endosc 1989;35:324–328.

77. Kantsevoy SV, Cruz-Correa MR, Vaughn CA, et al. Endoscopiccryotherapy for the treatment of bleeding mucosal vascular le-sions of the GI tract: a pilot study. Gastrointest Endosc 2003;57:403–406.

582 Swanson et al Clinical Gastroenterology and Hepatology Vol. 12, No. 4

78. Cho S, Zanati S, Yong E, et al. Endoscopic cryotherapy for themanagement of gastric antral vascular ectasia. GastrointestEndosc 2008;68:895–902.

79. Van Cutsem E. Long-term effect of hormonal therapy forbleeding gastrointestinal vascular malformations. Eur J Gastro-enterol Hepatol 1993;5:439–443.

80. Van Cutsem E, Rutgeerts P, Vantrappen G. Treatment ofbleeding gastrointestinal vascular malformations with oes-trogen-progesterone. Lancet 1990;335:953–955.

81. Bauditz J, Lochs H, Voderholzer W. Macroscopic appearance ofintestinal angiodysplasias under antiangiogenic treatment withthalidomide. Endoscopy 2006;38:1036–1039.

82. Kamalaporn P, Saravanan R, Cirocco M, et al. Thalidomide forthe treatment of chronic gastrointestinal bleeding from angio-dysplasias: a case series. Eur J Gastroenterol Hepatol 2009;21:1347–1350.

83. Barrie R, Woltering EA, Hajarizadeh H, et al. Inhibition ofangiogenesis by somatostatin and somatostatin-like com-pounds is structurally dependent. J Surg Res 1993;55:446–450.

84. D’Amato RJ, Loughnan MS, Flynn E, et al. Thalidomide is aninhibitor of angiogenesis. Proc Natl Acad Sci U S A 1994;91:4082–4085.

85. Ripoll C, Garcia-Tsao G. Management of gastropathy andgastric vascular ectasia in portal hypertension. Clin Liver Dis2010;14:281–295.

86. Ripoll C, Garcia-Tsao G. The management of portal hyperten-sive gastropathy and gastric antral vascular ectasia. Dig LiverDis 2011;43:345–351.

Reprint requestsAddress requests for reprints to: Aasma Shaukat, MD, MPH, GastroenterologySection 111-D, 1 Veterans Drive, Minneapolis, Minnesota 55417. e-mail:shaukat@umn.edu; fax: (612) 725-2248.

Conflicts of interestThe authors disclose no conflicts.

FundingSupported by VA Career Development Award (CDA-2) (A.S.).