1 2 The Role of Provocative 3 4 Testing and Localization of 5 6 the Video Capsule Endoscope in 7 the Management of Small 8 9 Intestinal Bleeding Q4 10 11 a, b 12 Daniel L. Raines, MD *, Douglas G. Adler, MD Q5 Q6 Q7 13 Q1 Q2 14 15 KEYWORDS 16  Small intestinal bleeding  Provocative angiography  Provocative 17  Video capsule endoscopy  Device-assisted 18 19 20 KEY POINTS 21  Cases of small intestinal bleeding (SIB) are commonly encountered in clinical practice 22 despite advancements in medical technology that allow consistent visualization of the 23 entire . 24  The use of antithrombotic agents to provoke bleeding has been found to be revealing in 25 cases of SIB when used in combination with angiography as well as endoscopy. 26  The performance of endoscopic procedures, including video capsule endoscopy (VCE) 27 and device-assisted enteroscopy (DAE) on active antiplatelet and/or anticoagulant ther- 28 apy is considered low risk and may improve diagnostic yield. 29  The use of VCE for gastrointestinal bleeding in the inpatient setting has been found to 30 improve diagnostic yield and decrease hospital stay. 31  Because DAE is required for therapeutic intervention in patients with small bowel 32 bleeding, access to this technology is essential for effective management in many cases. 33 34 35 Video content accompanies this article at http://www.giendo.theclinics.com/. 36 37 Q8 38 OBSCURE GASTROINTESTINAL BLEEDING Definitions 39 40 Small intestinal bleeding (SIB) has been described as bleeding that is undefined in 41 cause despite evaluation by standard upper endoscopy, , and small Q9 42 bowel follow-through (SBFT). The low diagnostic yield of SBFT combined with the Q10 43 advent of higher-yield small bowel studies has resulted in the removal of SBFT from 44 45 a b 46 1542 Tulane Avenue, Suite 423, New Orleans, LA, USA; University of Utah School of Med- 47 icine, 30 North 1900 East 4R118, Salt Lake City, UT 84132, USA * Corresponding author. 200 West Esplanade, Suite 200, Kenner, LA 70112. 48 E-mail address: [email protected]

Gastrointest Endoscopy Clin N Am - (2021) -–- https://doi.org/10.1016/j.giec.2021.01.001 giendo.theclinics.com 1052-5157/21/ª 2021 Elsevier Inc. All rights reserved.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 2 Raines & Adler

49 algorithms for evaluation of gastrointestinal (GI) bleeding.1,2 Video capsule endoscopy 50 (VCE) is associated with a 38% to 83% diagnostic yield in patients with suspected 51 small bowel bleeding and is therefore recommended as the next step in management 52 after negative esophagogastroduodenoscopy (EGD)/colonoscopy.3 Computed to- 53 mography enterography (CTE) may be indicated after negative VCE, or as an alterna- 54 tive if VCE is contraindicated, and has a reported diagnostic yield of 40% in bleeding of 55 suspected small bowel origin.4,5 Device-assisted enteroscopy (DAE) by double 56 balloon, single balloon, or spiral technique is a useful study in patients with suspected 57 small bowel bleeding but is not available in all centers. Rates of total enteroscopy differ 58 by technique but the diagnostic yield for all techniques is similar, around 60% to 80%.6 59 The widespread adoption of these advanced technologies to evaluate for small 60 bowel bleeding sources resulted in a revision of definitions pertaining to obscure 61 bleeding in 2015. Examples of bleeding with negative evaluation by EGD and colonos- 62 copy are now termed suspected SIB. This group accounts for 5% of bleeding presen- 63 tations, most of which are small bowel in origin.7 Sources of blood loss in these cases 64 can be grouped into lesions within the reach of standard endoscopy that are missed, 65 accounting for 25% of sources, and lesions in the , accounting for 75%. 66 The term obscure GI bleeding is now typically reserved for cases in which a bleeding 67 source cannot be identified despite current standard-of-care evaluation, including 68 EGD, colonoscopy, VCE, and/or DAE and radiographic studies.2 69 70 Distribution of Bleeding Sources 71 Missed lesions should be considered as lesions that are difficult to detect (Table 1), 72 which includes lesions that are identifiable on closer inspection as well as vascular ab- 73 normalities, such as Dieulafoy lesions, which are difficult or impossible to detect un- 74 less actively bleeding (Fig. 1A, B).8 Bleeding sources that are commonly missed on 75 upper endoscopy include Cameron erosions, associated with a large hiatal hernia; 76 gastric antral vascular ectasia, which may appear as benign mucosal erythema; small 77 or concealed angioectasias (Figs. 2 and 3); and peptic ulcers that are positioned in a 78 location difficult to visualize. Missed lesions on colonoscopy include flat neoplasms as 79 well as angioectasias in the right colon. Missed lesions in the small bowel vary accord- 80 ing to patient age. In patients older than 50 years, angioectasias are the most common 81 cause of small bowel bleeding in the Western population and are the most common 82 missed small bowel lesion by VCE.2,9 In adults younger than 50 years, tumors of the 83 84 85 86 Table 1 Culprit sources in suspected small bowel bleeding 87 88 Missed Lesions on Upper Endoscopy Small Bowel Bleeding Sources 89 Cameron erosions Vascular lesions Ulcerative disorders 90 Peptic ulcer disease  Angioectasias  Crohn disease   91 GAVE Dieulafoy lesion NSAID enteropathy Angioectasias  Aortoenteric fistula  Ischemic enteropathy 92 Dieulafoy lesion Neoplasms  Radiation enteropathy 93 Missed lesions on  Carcinoid  Meckel diverticulum 94 colonoscopy  GIST 95 Flat neoplasms  Adenocarcinoma 96 Angioectasias  Lymphoma 97 Dieulafoy lesion 98 Abbreviations: GAVE, gastric antral vascular ectasia; GIST, GI stromal tumor; NSAID, nonsteroidal 99 antiinflammatory drug.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 150 149 148 147 146 145 144 143 142 141 140 139 138 137 136 135 134 133 132 131 130 129 128 127 126 125 124 123 122 121 120 119 118 117 116 115 114 113 112 111 110 109 108 107 106 105 104 103 102 101 100

print & web 4C=FPO print & web 4C=FPO print & web 4C=FPO i.1. Fig. i.3. Fig. 2. Fig. ( cieybedn nietsab ie capsule. video by angioectasia bleeding Actively angioectasia. gastric Dense A , B ml oe iuao einosre yVEte usqetDAE. subsequent then VCE by observed lesion Dieulafoy bowel Small ) GEC1206_proof ■ eray2021 February 5 aaeeto ml netnlBleeding Intestinal Small of Management ■ :7pm 3:57 Q3 3 4 Raines & Adler

151 152 153 154 155 156 157 158 FPO

159 = 160 161 162 163

164 print & web 4C 165 Fig. 4. Endophytic GIST marked with India ink (observed by ). 166 167 small bowel are the most common bleeding source.10 The miss rate of VCE for small 168 bowel tumors is estimated to be 19%, especially for tumors such as GI stromal tumor 169 (GIST) that can be endophytic (Fig. 4).11 Therefore CTE is recommended after VCE in 170 patients less than 50 years old or in patients more than 50 years old with recurrent 171 bleeding or progressive anemia, although some tumors may remain undetected after 172 both VCE and CTE (Fig. 5).2,12 CTE may be combined with PET to show hypermeta- 173 bolic lesions such as adenocarcinoma (Fig. 6A, B). A Meckel scan is recommended as 174 an additional test, particularly in younger patients (Fig. 7). 175 176 Outcomes in Obscure Gastrointestinal Bleeding 177 Outcomes in patients with bleeding of unknown origin after negative EGD, colonos- 178 copy, and VCE have been evaluated in several studies. Two early studies of SIB 179 patient cohorts reported rebleeding rates of only 6% to 11% over an average 18- 180 month follow-up period.13,14 These studies also showed that negative capsule 181 endoscopy (CE) was predictive of lower rebleeding rates compared with patients Q11 182 with positive CE. However, subsequent studies of this patient population conflict 183 with these results. In 1 study, 26 patients with mostly overt obscure GI bleeding 184 (OGIB) were followed over an extended median follow-up period of 32 months and 185 rebleeding was observed in 35% of patients. In addition, this study found that 186 187 188 189 190 191 192 193 194 195 FPO 196 = 197 198 199 200 print & web 4C 201 Fig. 5. Ileal carcinoid tumor missed by VCE and CTE.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Management of Small Intestinal Bleeding 5

202 203 204 205 206 207 208 209 210

211 FPO = 212 213 214 215 216 print & web 4C 217 Fig. 6. (A, B) Metastatic adenocarcinoma of the by PET–computed tomography (A) 218 and DAE (B). Q20 219 220 221 rebleeding rates between patients with negative CE and positive CE were almost iden- 222 tical and only specific therapy directed to the bleeding source resulted in a change in 223 the frequency of rebleeding.15 Although this study was retrospective and complicated 224 by a high rate of incomplete capsule examinations (67%), similar results were shown in 225 another study of patients with SIB and negative CE.16 This study found that 47% of 226 patients with a nondiagnostic first CE study experienced recurrent overt bleeding or 227 progressive anemia during a mean follow-up period of 25 months. Management of in- 228 dividual patients should therefore be tailored to the clinical course. In older patients 229 with persistent resolution of bleeding, expectant management after negative VCE 230 may be considered unless additional symptoms or clinical findings warrant further 231 testing.1,2 In younger patients and patients with recurrent overt bleeding or progres- 232 sive anemia, additional testing should be pursued. 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 Fig. 7. Positive Meckel scan with tracer uptake in the gastric body (red arrow) and Meckel Q21 252 site (yellow arrow).

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 6 Raines & Adler

253 Repeat Testing in Obscure Gastrointestinal Bleeding 254 Additional testing in patients who experience progressive anemia attributed to GI 255 blood loss or recurrent, overt bleeding after negative initial evaluation by EGD, colo- 256 noscopy, VCE, and/or DAE and CTE should be individualized. Supplemental history 257 should be obtained, including previous history of epistaxis, blood donation, menstrual 258 bleeding, frequent minor rectal bleeding, diarrhea, abdominal pain, history of radiation 259 exposure, history of abdominal aortic aneurysm repair predisposing to aortoenteric 260 fistula (Figs. 8 and 9), nonsteroidal antiinflammatory drug use, olmesartan use, and Q12 261 family history of inflammatory bowel disease or celiac disease. After review of this 262 additional history and previous studies, subsequent testing should be divided into 2 263 groups based on bleeding pattern as progressive anemia attributed to GI blood loss 264 (occult bleeding) versus discrete episodes of hematochezia or melena (overt bleeding) 265 In patients with occult bleeding, repeat upper endoscopy with push enteroscopy is 266 worthwhile to reevaluate for missed lesions in the upper GI tract as well as for angioec- 267 tasias in the proximal small bowel and to obtain duodenal biopsies, if not previously 268 performed.17 Examination of the with a side-viewing duodenoscope may 269 also be helpful in identifying bleeding sources in the duodenal wall or associated 270 with the ampulla of Vater. Repeat colonoscopy with more extensive intubation of 271 the terminal ileum can be considered to reevaluate for a flat neoplasm or angioectasias 272 in the right colon as well as to further investigate for evidence of Crohn disease 273 involving the distal ileum. In cases in which the initial CE study was incomplete or in 274 which anemia is progressive, repeat CE should be considered. One study of 275 second-look CE in 20 patients with occult bleeding reported a 35% diagnostic yield 276 of a second CE.18 Another study, which included 48 patients with occult, obscure 277 bleeding, showed that repeat CE was successful in identifying a likely bleeding source 278 in 65% of patients with progressive anemia, as defined by a 4-g/dL decrease in hemo- 279 globin level. In patients who converted from occult to overt bleeding, the diagnostic 280 yield of repeat CE was 100%.16 If 1 or more of these studies is unrevealing, evaluation 281 of the remaining small bowel through DAE should be pursued depending on availabil- 282 ity. Intraoperative enteroscopy is the test of last resort in patients with severe anemia 283 who have been evaluated by all other means. 284 Recurrent, overt SIB is often frustrating for both patients and clinicians but can usu- 285 ally be managed successfully with persistence. Repeated standard endoscopy, CE, 286 angiography, and/or DAE should be performed as soon as possible after a 287 288 289 290 291 292 293 294 295 296 297 FPO 298 = 299 300 301 302 print & web 4C 303 Fig. 8. Endoscopic view of a jejunal aortoenteric fistula (red arrow).

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Management of Small Intestinal Bleeding 7

304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 Fig. 9. Vascular tuft with early draining vein (arrow) consistent with AVM. 323 324 presentation with overt bleeding until bleeding stops or the bleeding site is localized. 325 Multiple attempts may be required to successfully identify the bleeding source. 326 Tagged red blood cell (RBC) scan is a useful adjunct to endoscopy because of its 327 availability and lack of risk, although it is often unrevealing or nonspecific. Computed 328 tomography angiography (CTA) may provide images that are easier to obtain and more 329 localizing compared with tagged RBC scan. Patients with recurrent overt bleeding are 330 considered to be the primary candidates for provocative testing. 331 332 333 PROVOCATIVE TESTING IN THE EVALUATION OF OBSCURE GASTROINTESTINAL 334 BLEEDING 335 Provocation is defined as the introduction of a stimulus in attempt to elicit a specific 336 response. Provocative testing is well established in the management of certain med- 337 ical conditions, such as the stimulation of arrhythmias in electrophysiology.19–21 Pro- 338 vocative angiography for the management of difficult cases of GI bleeding is included 339 in the American Society for Gastrointestinal Endoscopy (ASGE) algorithm for evalua- 340 tion of obscure GI bleeding but is practiced in a limited number of centers.2 The use of 341 provocation combined with endoscopy may be practiced in the community in low vol- 342 umes but the current body of literature pertaining to this practice is limited to 4 case 343 reports and 1 large retrospective study.22–26 Although these techniques seem 344 extreme, reported complication rates associated with these interventions are low, 345 with reasonable diagnostic yields.2 346 347 Provocative Angiography 348 The body of literature pertaining to provocative angiography includes 9 series and 7 349 case reports totaling 126 cases (Table 2).27–35 The first series published by Rosch 350 and colleagues27 in 1982 described 3 cases of recurrent, overt bleeding in which 351 the administration of systemic heparin, intra-arterial heparin, and/or an intra-arterial 352 vasodilator was used as an adjunct to conventional angiography with successful treat- 353 ment of the bleeding source in all 3 cases. A subsequent study of provocation by Mal- 354 den and colleagues28 involved the administration of systemic heparin and systemic

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 8 Raines & Adler

355 Table 2 356 Provocative angiography case series 357 358 Success 359 Systemic Intra-arterial Rate Series Patients Agents Agents (%) 360 26 361 Rosch et al, 2017 3 Heparin Tolazoline 100 27 362 Malden et al, 1982 9 Heparin —20 363 Urokinase 28 364 Koval et al, 1998 10 Heparin Tolazoline 80 Streptokinase 365 29 366 Bloomfeld et al, 1987 7 — Heparin 28 Tolazoline 367 Urokinase 368 Ryan et al30 16 Heparin Tolazoline tPA 37 Q22 369 Mernagh et al31 12 Heparin — 50 370 32 371 Wildus et al 9 — Reteplase 89 33 372 Kim et al 34 Heparin Nitroglycerin tPA 31 373 Zurkiya et al34 19 — Heparin 26 374 Nitroglycerin tPA 375 Abbreviation: tPA, tissue plasminogen activator. 376 377 urokinase after initial negative mesenteric angiography followed by tagged RBC 378 scan.27 Positive tagged RBC scans were documented in 4 of 10 cases, but angiog- 379 raphy was diagnostic in only 2 of these, resulting in an overall success rate of 20%. 380 Systemic heparin alone was used in another study by Mernagh and colleagues32 after 381 initial negative angiography, which was associated with positive repeat angiography in 382 6 of 12 initial negative cases (yield of 50%). All of the remaining protocols have 383 involved the use of systemic heparin accompanied by intra-arterial infusion of a vaso- 384 dilator and/or thrombolytic, resulting in a wide range of diagnostic yields (18%– 385 80%).27–35 The 2 most recent case series of provocation using a combination of hep- 386 arin, intra-arterial nitroglycerin, and intra-arterial tissue plasminogen activator (tPA) re- 387 ported success rates of 26% to 31%.34,35 388 Despite concerns regarding the risk of bleeding associated with angiography in the 389 setting of anticoagulation, the rate of complications in these reported series was low. 390 Major complications were noted in only 3 of 126 cases. Rosch and colleagues27 re- 391 ported 1 episode of hematemesis during angiography after infusion of heparin, tolazo- 392 line, and streptokinase that stabilized with protamine infusion and transfusion of 2 393 units of blood. Koval and colleagues29 reported 1 case of arterial wall puncture site he- 394 matoma after infusion of streptokinase. Kim and colleagues34 reported 1 case of intes- 395 tinal ischemia resulting in perforation requiring surgery after embolization of a lesion in 396 the ileum, which was subsequently diagnosed as a carcinoid tumor during surgery. 397 The availability of mesenteric angiography varies by center according to the exper- 398 tise and interest of the interventional radiology and interventional cardiology staff. The 399 use of provocation combined with angiography is likely to be limited to interventional- 400 ists with a specific interest in and comfort level with mesenteric angiography. Mesen- 401 teric angiography is associated with risks of vascular injury, including hematoma and 402 other bleeding complications, and these risks may be increased when combined with 403 anticoagulant, antiplatelet, or thrombolytic agents. If an intervention such as emboli- 404 zation is planned, there is a risk of subsequent intestinal ischemia requiring surgery, 405 which must be also be considered.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Management of Small Intestinal Bleeding 9

406 Provocative Endoscopy 407 The body of literature pertaining to provocative testing combined with endoscopy 408 consisted of only 4 case reports before 2017. The first, by Berkelhammer and col- 409 leagues22 in 2000, involved a 62-year-old man with an who experienced 410 recurrent overt bleeding. Upper endoscopy, ileoscopy, and provocative angiography 411 were negative. Provocation with a 10,000-unit intravenous (IV) heparin bolus followed 412 by a heparin 1000 unit/h infusion resulted in active bleeding 90 minutes later. Emer- 413 gent ileoscopy subsequently revealed a Dieulafoy lesion 5 cm proximal to the stoma, 414 which was treated definitively by distal ileal resection. In 2004, Wright and col- 415 leagues23 described a 77-year-old man who presented with melena and hypotension 416 resulting in intestinal ischemia and myocardial infarction. Repeat endoscopic testing 417 and imaging by tagged RBC scan were negative for a bleeding source despite 418 bleeding requiring transfusion of a total of 80 units of packed RBCs. Provocation 419 with 5000 units of IV heparin resulted in active bleeding 2 hours later, indicated by 420 bloody nasogastric tube output. Subsequent EGD revealed a Dieulafoy lesion in the 421 that was treated endoscopically. Reider and colleagues24 published a 422 case report in 2006 describing a 59-year-old man with a history of overt OGIB. 423 EGD, colonoscopy, enteroclysis, push enteroscopy, and VCE did not identify a 424 bleeding source. Provocative CE was performed with heparin IV infusion initiated 425 12 hours before the procedure and targeted to a partial thromboplastin time between 426 72 and 108 seconds (normal range, 26–36 seconds). Blood was observed in the prox- 427 imal jejunum by VCE, and a subsequent resulted in removal of a GIST. In 428 addition, in 2007, Kumar and colleagues25 published a case report of an unsuccessful 429 provocative endoscopy attempt in a 65-year-old man presenting with hematochezia. 430 Upper endoscopy and colonoscopy did not reveal a bleeding source. Tagged RBC 431 scintigraphy localized the bleeding to the distal small bowel, whereas angiography 432 localized the bleeding to the hepatic flexure; however, bleeding ceased and a lesion 433 was not identified on subsequent colonoscopy despite provocation with 10,000 units 434 of IV heparin and endoscopic observation for 75 minutes. The patient was taken to 435 surgery, where a distal ileal carcinoid was found as the source of bleeding. 436 The authors published a retrospective review in 2017 of a databases of 824 DAE 437 procedures. We identified 38 instances in which provocation with antiplatelet and/or 438 anticoagulant agents was used as part of an OGIB evaluation in 27 patients.26 Proced- 439 ures were divided into 3 groups based on the method of provocation: cases in which 440 an antiplatelet or anticoagulant agent was reintroduced because of a prior history of 441 bleeding exacerbation related to this agent (provocation experienced); cases in which 442 an antiplatelet or anticoagulant agent was administered without any such history 443 (provocation naive); and extreme cases of recurrent, overt OGIB in which a combina- 444 tion of clopidogrel and IV heparin was administered for provocation (LSU protocol). Q13 445 The diagnostic yield of provocative testing per procedure was 53% in the 446 provocation-experienced group, 27% in the provocation-naive group, and 71% in 447 the full-protocol group. Provocative testing was revealing in 15 out of 27 patients 448 with angioectasias, and Dieulafoy lesions were found to be the most common lesions 449 (Video 1). Provocative testing was not beneficial in 4 patients, who were eventually 450 diagnosed with bleeding caused by intestinal angioectasias (3) and 1 patient with an 451 aortoenteric fistula. The bleeding source was never found in 8 patients out of 27. There 452 were no complications associated with any of the 38 procedures performed. 453 The ideal agent for provocative testing would have no adverse effects and would 454 provoke bleeding consistently but not increase the risk of endoscopy. Antiplatelet 455 agents such as clopidogrel are attractive because of evidence of platelet dysfunction 456

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 10 Raines & Adler

457 being the culprit for precipitation of bleeding from angioectasias in several settings, 458 such as bleeding from acquired von Willebrand factor deficiency in Heyde syndrome 459 and in patients with left ventricular assist devices.36,37 There is emerging evidence that 460 endoscopy such as colonoscopy with polypectomy and DAE with argon plasma coag- 461 ulation (APC) for treatment of angioectasias may be performed safely in the setting of 462 active clopidogrel therapy.38,39 If clopidogrel is selected as the stimulus, patients may 463 be given a loading dose of 150 mg or 300 mg 8 hours before endoscopy in order to 464 achieve 35% to 45% or 40% to 50% platelet inhibition, respectively.40 In extreme 465 cases of overt bleeding, a combination of clopidogrel and IV heparin may be consid- 466 ered to allow for both antiplatelet and anticoagulant effects with the ability to withdraw 467 heparin (or reverse with protamine sulfate) resulting in endoscopy in the setting of clo- 468 pidogrel alone. The use of systemic thrombolytic agents should be avoided because 469 of the risk of major complication, such as stroke, associated with systemic administra- 470 tion of these agents and low yield observed in previous study of systemic 471 thrombolytics.28 472 Many antiplatelet, anticoagulant, and antithrombotic agents have been introduced 473 over the past 20 years. Each agent could be considered for inclusion in a provocation 474 protocol depending on its effect, pharmacokinetics, and risk. One anticoagulant that 475 might be considered for future protocols is dabigatran (Pradaxa), because a standard 476 dose of 150 mg peaks in 2 hours and can be immediately reversed with IV infusion of 477 idarucizumab (Praxbind).41 Regarding antiplatelet agents, the glycoprotein IIb/IIIa in- 478 hibitor abciximab (ReoPro) can also be considered reversible because its short plasma 479 half-life allows its hemostatic inhibition to be reversed with platelet administration.42 480 The ability of the endoscopist to reach the bleeding source is also a critical factor 481 when considering provocative endoscopy, particularly when the bleeding segment 482 has not been localized. The authors recommend that centers that might consider pro- 483 vocative endoscopy have the capability to perform total enteroscopy by DAE, prefer- 484 ably with double balloon enteroscopy. The ancillary resources of the center should 485 also be assessed, including access to intensive care, angiography, and surgery 486 services. 487 488 489 BLEEDING LOCALIZATION WITH VIDEO CAPSULE ENDOSCOPY 490 Accurate localization of the bleeding source is essential in CE interpretation. Experi- 491 enced capsule readers can identify pitfalls in which the landmarks can be mislabeled, 492 and small bowel transit times may be misleading. A fair estimate of the location of the 493 bleeding segment can typically be made if the capsule transits through the entire small 494 bowel. However, retention of the capsule endoscope in the duodenum for an 495 extended period before transit through the deep small bowel is a source of confound- 496 ing. The video capsule may reflux back and forth into the stomach after the first 497 duodenal image. Residue or blood may interfere with proper delineation of landmarks 498 as well as capsule transit. Early CE technology included a sensory array designed to 499 show the capsule location in the abdomen in addition to obtaining images. However, 500 this technology was abandoned with adoption of the sensor belt. CE technology has 501 recently advanced to estimate progress more effectively through the small bowel by 502 computer analysis of redundant images.43 This software may still be fooled by inaccu- 503 rate landmarks from the reader or by retention in the duodenum but shows promise for 504 accurate assessment of location in most studies. 505 An experienced capsule reader may rely on the original study to determine an 506 impression of where the lesion lies in the small bowel before intervening by DAE. A 507 general rule for planning DAE after VCE is selection of an oral route for lesions

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Management of Small Intestinal Bleeding 11

508 observed in the first 60% of small bowel transit versus anal route for lesions observed 509 in the last 40% of small bowel transit.44,45 In some cases, DAE may be deferred if the 510 abnormality observed on VCE appears to be within reach of standard upper endos- 511 copy, push enteroscopy, or colonoscopy. 512 513 SUMMARY 514 515 Advances in techniques to evaluate bleeding sources have advanced to the point w 516 where obscure GI bleeding is now restricted to an estimated 1% to 2% of all 517 bleeding cases. However, in this small subset of patients, the bleeding source may 518 remain undefined despite exhaustive testing. Where do clinicians go from here? 519 Attempts to intentionally stimulate bleeding for the purpose of identifying the 520 bleeding source by angiography or endoscopy should be considered for select pa- 521 tients with recurrent, severe anemia or recurrent overt bleeding in whom exhaustive 522 repeat testing is negative. A detailed discussion of the risks and benefits of provoca- 523 tive testing with the patients and their families is needed before attempting stimulation 524 of bleeding. This discussion should include a review of the risks of angiography and/or 525 endoscopy, which may be increased in the setting of active antiplatelet and/or antico- 526 agulant therapy; the risk of stimulating bleeding that might not be successfully con- 527 tained; and the increased risk of surgery in the event of uncontrollable bleeding or a 528 procedure-related complication. However, provocative testing may be a justifiable 529 intervention in highly selected cases in which the morbidity associated with recurrent 530 bleeding justifies the additional risk. 531 VCE studies should be performed while maintaining routine antiplatelet and/or anti- 532 coagulant therapy. The performance of DAE as well as other endoscopic procedures 533 should also be pursued on antithrombotics because these are considered low risk if 46 534 diagnostic or include only APC and/or hemostatic clip placement. Our study of 39 535 DAE on active thienopyridine supports the safety and efficacy of this practice. The 536 performance of inpatient VCE may be limited in certain hospitals because of concerns 537 about loss of capital. However, studies have shown that timely inpatient VCE signifi- 538 cantly decreases hospital stay to offset the cost of the video capsule and improves 47 539 diagnostic yield. 540 In the future, advancements in artificial intelligence may allow computer localization 541 of intestinal angioectasias, which may be present on only 1 of thousands of im- 48,49 542 ages. These technologies may increase capture rate for this disorder and reduce 543 read time. Progressive development of a motorized spiral enteroscope is an additional 544 technology on the horizon that has been anxiously awaited by small bowel endoscop- 50 545 ists because of its potential for rapid small bowel examination. 546 547 DISCLOSURES

548 D.L. Raines: speaker and consultant for Medtronic. D.G. Adler: consultant/Advisory Q14 549 Board for BSC, Merit, Olympus; Speaker’s Bureau for AbbVie. Q15 550 551 SUPPLEMENTARY DATA 552 553 Supplementary video related to this article can be found at https://doi.org/10.1016/j. 554 giec.2021.01.001. Q16 555 556 REFERENCES 557 1. ASGE Standards of Practice Committee. The role of endoscopy in the manage- 558 ment of obscure GI bleeding. Gastrointest Endosc 2010;72(3):471–9.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 12 Raines & Adler

559 2. Gerson LB, Fidler JL, Cave DR, et al. ACG clinical guideline: diagnosis and man- 560 agement of small bowel bleeding. Am J Gastroenterol 2015;110(9):1265–87. 561 3. Rondonotti E, Villa F, Mulder CJ, et al. Small bowel capsule endoscopy in 2007: 562 indications, risks and limitations. World J Gastroenterol 2007;13:6140–9. 563 4. Huprich JE, Fletcher JG, Fidler JL, et al. Prospective blinded comparison of wire- 564 less capsule endoscopy and multiphase CT enterography in obscure gastrointes- 565 tinal bleeding. Radiology 2011;260(3):744–51. 566 5. Huprich JE, Fletcher JG, Alexander JA, et al. Obscure gastrointestinal bleeding: 567 evaluation with 64-section multiphase CT enterography—initial experience. Radi- 568 ology 2008;246(2):562–71. 569 6. Leighton JA. The role of endoscopic imaging of the small bowel in clinical prac- 570 tice. Am J Gastroenterol 2011;106:27–36. 571 7. Raju GS, Gerson L, Das A, et al. American gastroenterological association (AGA) 572 Institute medical position statement on obscure gastrointestinal bleeding. Gastro- 573 enterology 2007;133:1694–6. 574 8. Tee HP, Kaffes AJ. Non-small-bowel lesions encountered during double-balloon 575 enteroscopy performed for obscure gastrointestinal bleeding. World J Gastroen- 576 terol 2010;16(15):1885–9. 577 Clin Gas- 578 9. Regula J. Vascular lesions of the gastrointestinal tract. Best Pract Res troenterol 579 2008;22(2):313–28. 580 10. Lewis BS, Kornbluth A, Wayne JD. Small bowel tumors: yield of enteroscopy. Gut 581 1991;32:763–5. 582 11. Lewis BS, Eisen GM, Friedman S, et al. A pooled analysis to assess results of 583 capsule endoscopy trials. Endoscopy 2005;37:960–5. 584 12. Boriskin HS, Devito BS, Hines JJ, et al. CT enterography vs capsule endoscopy. 585 Abdom Imaging 2009;34:149–55. 586 13. Macdonald J, Porter V, McNamara D. Negative capsule endoscopy in patients 587 with obscure GI bleeding predicts low rebleeding rates. Gastrointest Endosc 588 2008;68:1122–7. 589 14. Lai LH, Wong GL, Chow DK, et al. Long-term follow-up of patients with obscure 590 gastrointestinal bleeding after negative capsule endoscopy. Am J Gastroenterol 591 2006;101:1224–8. 592 15. Park JJ, Cheon JH, Kim HM, et al. Negative capsule endoscopy without subse- 593 quent enteroscopy does not predict lower long-term rebleeding rates in patients 594 with obscure GI bleeding. Gastrointest Endosc 2010;71(6):990–7. 595 16. Viazis N, Papaxoinis K, Vlachogiannakos J. Is there a role for second-look 596 capsule endoscopy in patients with obscure GI bleeding after a nondiagnostic 597 first test? Gastrointest Endosc 2009;69:850–6. 598 17. Lara LF, Bloomfield, Pineau BC. The rate of lesions found within reach of esoph- 599 agogastroduodenoscopy during push enteroscopy depends on the type of 600 obscure gastrointestinal bleeding. Endoscopy 2004;36:745–50. 601 18. Bar-Meir S, Eliakim R, Nadler M. Second capsule endoscopy for patients with se- 602 vere iron deficiency anemia. Gastrointest Endosc 2004;60:711–3. 603 19. Spurrell RAJ, Krikler DM, Sowton E. Concealed bypasses of the atrioventricular 604 node in patients with paroxysmal supraventricular tachycardia revealed by intra- 605 cardiac stimulation and verapamil. Am J Cardiol 1974;33:590–5. 606 20. Denes P, Wu D, Dhingra RC, et al. Demonstration of dual AV nodal pathways in 607 patients with supraventricular tachycardia. Circulation 1973;48:549–55. 608 21. Scheinman MM, Huang S. The NASPE prospective catheter ablation registry. 609 Pacing Clin Electrophysiol 1998;23:1020–8.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Management of Small Intestinal Bleeding 13

610 22. Berkelhammer C, Radvany A, Lin A, et al. Heparin provocation for endoscopic 611 localization of recurrent obscure GI hemorrhage. Gastrointest Endosc 2000;52: 612 555–6. 613 23. Wright CA, Petersen BT, Bridges CM, et al. Heparin provocation for identification 614 and treatment of a gastric Dieulafoy’s lesion. Gastrointest Endosc 2004;59: 615 728–30. 616 24. Rieder F, Schneidewind A, Bolder U, et al. Use of anticoagulation during wireless 617 capsule endoscopy for the investigation of recurrent obscure gastrointestinal 618 bleeding. Endoscopy 2006;38:526–8. 619 25. Kumar A, Gandolfo F, Halwan B. Provocation of bleeding during endoscopy in 620 patients with recurrent acute lower gastrointestinal bleeding. Gastroenterol Hep- 621 atol 2007;3:570. 622 26. Raines DL, Jex K, Nicaud M, et al. Pharmacologic provocation combined with 623 endoscopy in refractory cases of GI bleeding. Gastrointest Endosc 2017;85(1): 624 112–20. 625 27. Rosch J, Keller FS, et al. Pharmaco-angiography in the diagnosis of massive 626 lower gastrointestinal bleeding. Radiology 1982;145:615–9. 627 28. Malden ES, Hicks ME, Royal HD, et al. Recurrent gastrointestinal bleeding: use of 628 thrombolysis in diagnosis. Radiology 1998;207:147–51. 629 630 29. Koval G, Benner KG, Ro¨sch J, et al. Aggressive angiographic diagnosis in acute 631 lower gastrointestinal hemorrhage. Dig Dis Sci 1987;32:248–53. 632 30. Bloomfeld RS, Smith TP, Schneider AM, et al. Provocative angiography in patients 633 with gastrointestinal hemorrhage of obscure origin. Am J Gastroenterol 2000; 634 95(10):2807–12. 635 31. Ryan JM, Key SM, Dumbleton SA, et al. Nonlocalized lower gastrointestinal 636 bleeding. J Vasc Interv Radiol 2001;12:1273–7. 637 32. Mernagh JR, O’Donovan N, Somers S, et al. Use of heparin in the investigation of 638 obscure gastrointestinal bleeding. Can Assoc Radiol J 2001;52(4):232–5. 639 33. Wildus DM. Reteplase provocative visceral angiography. J Clin Gastroenterol 640 2007;41:830–3. 641 34. Kim CY, Suhocki PV, Miller Jr MJ, et al. Provocative mesenteric angiography for 642 lower gastrointestinal hemorrhage. J Vasc Interv Radiol 2010;21:477–83. 643 35. Zurkiya O, et al. Provocative mesenteric angiography for gastrointestinal hemor- 644 rhage [abstract]. J Vasc Interv Radiol 25(3):S55. Q17 645 36. Pate GE, Chandavimol M, Naiman SC, et al. Heyde’s syndrome: a review. J Heart 646 Valve Dis 2004;13:701–12. 647 37. Harvey L, Holley CT, John R. Gastrointestinal bleed after left ventricular assist de- 648 vice implantation: incidence, management, and prevention. Ann Cardiothorac 649 Surg 2014;3:475–9. 650 38. Feagins LA, Uddin FS, Davila RE, et al. The rate of post-polypectomy bleeding for 651 patients on uninterrupted clopidogrel therapy during elective colonoscopy is 652 acceptably low. Dig Dis Sci 2011;56(9):2631–8. 653 39. Bollinger E, Spera M, Raines DL. The safety and efficacy of device-assisted en- 654 teroscopy in the setting of thienopyridine therapy. J Clin Gastroenterol (Accepted 655 for publication). Q18 656 40. Fukushima K, Yoshio K, Hideki K, et al. Effect of 150-mg vs 300-mg loading doses 657 of clopidogrel on platelet function in japanese patients undergoing coronary stent 658 placement. Circ J 2008;72:1282–4. 659 41. Pollack CV, Reilly PA, Eikelboom J, et al. Idarucizumab for dabigatran reversal. 660 N Engl J Med 2015;373:511–20.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm 14 Raines & Adler

661 42. Tcheng JE. Clinical challenges of platelet glycoprotein IIb/IIIa receptor inhibitor 662 therapy: bleeding, reversal, thrombocytopenia, and retreatment. Am Heart J 663 2000;139:S38–45. 664 43. Gomes S, Valerio MT, Salgo M, et al. Unsupervised neural network for homogra- 665 666 phy estimation of capsule endoscopy frames. Procedia Computer Science 2019; 667 164:602–9. Q19 668 44. Nakamura M, Ohimya N, Shirai O, et al. Route selection for double-balloon en- 669 teroscopy, based on capsule transit time, in obscure gastrointestinal bleeding. 670 J Gastroenterol 2010;45(6):592–9. 671 45. Li X, Chen H, Dai J, et al. Predictive role of capsule endoscopy on the insertion 672 route of double-balloon enteroscopy. Endoscopy 2009;41(9):762–6. 673 46. Acosta RD, Neena SA, Vinay C, et al. The management of antithrombotic agents 674 for patients undergoing GI endoscopy. Gastrointest Endosc 2016;83(1):3–16. 675 47. Singh A, Marshall C, Chaudhuri B, et al. Timing of video capsule endoscopy rela- 676 677 tive to overt obscure GI bleeding: implications from a retrospective study. Gastro- 678 intest Endosc 2013;77:761–6. 679 48. Leenhardt R, Pauline V, Cynthia L, et al. A neural network algorithm for detection 680 of GI angioectasia during small-bowel capsule endoscopy. Gastrointest Endosc 681 2018;89(1):189–94. 682 49. Alagappan M, Brown JRG, Mori Y, et al. Artificial intelligence in gastrointestinal 683 endoscopy: the future is almost here. World J Gastrointest Endosc 2018;10: 684 239–49. 685 50. Mans L, Arvanitakis M, Neuhaus H, et al. Motorized spiral enteroscopy for occult 686 bleeding. Dig Dis 2018;36:325–7.

GEC1206_proof ■ 5 February 2021 ■ 3:57 pm Our reference: GEC 1206 P-authorquery-v9

AUTHOR QUERY FORM

Journal: GEC

Article Number: 1206

Dear Author,

Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen annotation in the PDF file) or compile them in a separate list. It is crucial that you NOT make direct edits to the PDF using the editing tools as doing so could lead us to overlook your desired changes. Note: if you opt to annotate the file with software other than Adobe Reader then please also highlight the appropriate place in the PDF file. To ensure fast publication of your paper please return your corrections within 48 hours.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in the proof.

Location Query / Remark: Click on the Q link to find the query’s location in text in article Please insert your reply or correction at the corresponding line in the proof Q1 Please verify your preferred correspondence address to be published and provide any missing information. Elsevier recommends not using your personal home address. Q2 For your co-authors, please verify their affiliations and provide a complete address for the affiliations listed. The address will appear on the footnote of the first page of your article and will be published. Once again, Elsevier recommends not using personal home addresses. Also, please note that we will send each contributing author a copy of this issue to their mentioned address. Q3 Please approve the short title to be used in the running head at the top of each right-hand page. Q4 Please add a list of Clinics Care Points to the text. Clinics Care Points are a set of evidence-based pearls and pitfalls relevant to the point of care. Please note, these are different from Key Points. Q5 Are author names and order of authors OK as set? Q6 This is how your name will appear on the contributor's list. Please add your academic title, if it is missing, as well as any other necessary titles and professional affiliations. DANIEL L. RAINES, MD, New Orleans, Louisiana, USA DOUGLAS G. ADLER, MD, University of Utah School of Medicine, Salt Lake City, Utah, USA Q7 The following synopsis is the one that you supplied, but edited down to less than 100 words. Please confirm OK, or submit a replacement (also less than 100 words). Please note that the synopsis will appear in PubMed: The cause of small intestinal bleeding (SIB) may be elusive despite exhaustive testing. This article describes the current understanding of SIB regarding evaluation, with emphasis on the use of video capsule endoscopy (VCE) as a diagnostic procedure. This article addresses the utility of provocative testing in challenging cases and the performance of endoscopic procedures on active antithrombotic therapy. Specific recommendations accompany this article, including use of antithrombotic agents to stimulate

(continued on next page) bleeding when clearly indicated; performance of endoscopic procedures on active antithrombotic therapy; and progressive adoption of VCE and device-assisted enteroscopy in the inpatient setting. Q8 As per style numbered sections are not allowed. Please verify the structure of the headings and subheadings throughout. Q9 Please verify that the language editing preserves the intended meaning throughout. Q10 If there are any drug dosages in your article, please verify them and indicate that you have done so by initialing this query. Q11 Please verify the expansion of the abbreviation “CE”. Q12 Fig. 9 was not cited in the text, hence it has been combined with the citation of Fig. 8. Please verify. Q13 Please provide the expansion of the abbreviation “LSU”. Q14 Please verify the disclosure statement. Q15 Correctly acknowledging the primary funders and grant IDs of your research is important to ensure compliance with funder policies. We could not find any acknowledgement of funding sources in your text. Is this correct? Q16 Please verify the caption for Video 1. Q17 Please note that as per the journal style, if there are more than 3 authors, the first 3 author names are listed followed by ‘et al’ if the author group consists of 3 authors or fewer, all author names should be listed. Therefore, in Ref. 35; please list all names for up to 3 authors. For more than 3 authors, use ‘et al’ after the first 3 authors. Also provide year of publication. Q18 Please update complete reference details for Ref. 39. Q19 The journal title in Ref. [43] does not match that in PubMed. Please check that the reference details are correct. Q20 Please verify the part labels A and B in Fig. 6. Q21 The arrows are not present in Figs. 7 and 9. Please verify. Q22 The author name in the reference list is different from Table 2 of many references. Please clarify.

Please check this box or indicate your approval if you have no , corrections to make to the PDF file

Thank you for your assistance.