Diagnostic and Therapeutic Utility of Double-Balloon Endoscopy in Small-Bowel Bleeding

Diagnostic Yield of DBE in Patients with OGIB 

Published studies report relatively wide-ranging diagnostic yields, varying between 38% and 91% (Table 1). In the first reported series, Yamamoto and coworkers6 investigated 66 patients with OGIB, in whom the bleeding source was identified in 50 (76%). In a study of 52 patients, May and coworkers14 identified pathological lesions in 38 (73%): the bleeding source was identified with oral DBE in 32 of the 52 patients (61.5%), whereas in 6 (11.5%), including 2 colon cancers, diagnosis was made with anal DBE alone. Similarly, an abnormality was seen by DBE in 45 (75%) of the 60 patients reported by Kaffes and coworkers.15 In a large cohort of 152 patients, Sun and coworkers16 identified a potential bleeding site by DBE in 115 (76%). In another large study by Heine and coworkers,10 including 168 patients with OGIB, an explanation was found by DBE in 123 (73%). Similar yields were also reported in 4 small series; in a group of 31 patients, Manabe and coworkers17 were able to identify bleeding points by DBE in 23 (74%); the bleeding source was successfully identified in 15 (79%) of 19 patients investigated by Suzuki and coworkers18; small-bowel lesions potentially responsible for bleeding were identified in 15 (75%) of 20 patients reported by Hsu and coworkers19 and by Safatle-Ribeiro and coworkers.20

Table 1. Summary of DBE Studies for Obscure GI Bleeding

	Author (Ref.)	No. of Patients	Diagnostic Yield (%)	Endoscopic Therapeutic Yield
	Yamamoto et al. (6)	66	50(76)	20%
	Di Caro et al. (9)	33	29(88)	55%
	Heine et al. (10)	168	123(73)	36%
	Monkemüller et al. (11)	29	11(38)	28%
	May et al. (14)	52	38(73)	50%
	Kaffes et al. (15)	60	45(75)	57%
	Sun et al. (16)	152	115(76)	12%
	Manabe et al. (17)	31	23(74)	10%
	Suzuki et al. (18)	19	15(79)	16%
	Hsu et al. (19)	20	15(75)	55%
	Safatle-Ribeiro et al. (20)	20	15(75)	15%
	Zhi et al. (21)	57	52(91)	NR⁎
	Hadithi et al. (22)	35	21(60)	60%
	Mehdizadeh et al. (23)	130	66(51)	27%
	Nakamura et al. (24)	28	12(43)	11%

⁎ NR, not reported.

Two studies report remarkably high diagnostic yields. Di Caro and coworkers9 described small-intestine abnormalities in 29 (88%) of 33 patients submitted to DBE, whereas Zhi and coworkers21 identified a bleeding source in 52 (91%) of 57 patients. Lower yields were reported in studies performed elsewhere: of 35 patients described by Hadithi and coworkers,22 DBE identified pathological findings in 21 (60%), whereas the procedure was diagnostic in 66 (51%) of 130 patients reported by Mehdizadeh and coworkers.23 Nakamura and coworkers24 reported on 28 patients, of whom DBE revealed the bleeding source in only 12 (43%). Lastly, a potential bleeding site was discovered in 11 (38%) of 29 patients described by Mönkemüller and coworkers.11

The significant difference in diagnostic yield depends on a combination of factors, including nonuniform inclusion criteria, the quality and completeness of pre-DBE endoscopic investigations, the timing of DBE, as well as a lack of clarity over what constitutes a clinically significant finding. In addition, there are still few large studies. Interestingly, small-bowel ulcers (Fig. 1) and tumors (Fig. 2) are the most common diagnostic findings in series originating from Far Eastern countries, whereas in European and North/South American studies, the most common diagnostic finding was angioectasia (Fig. 3A). To date, few studies have assessed factors that may be predictive of the diagnostic yield of DBE. In a series of 162 OGIB patients,25 the clinical relevance of findings at prior CE and the oral access route were found to be associated with a significantly higher yield of DBE.

View Large Image Download to PowerPoint

Figure 1. Jejunal ulcerations and swollen villi in a patient with Crohn’s disease. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 2. Mass in the ileum. Histology revealed a gastrointestinal stromal tumor. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 3. (A) Large jejunal angioectasia. (B) The same lesion after injection of diluted epinephrine and cauterization with APC. (Color version of figure is available at www.techgiendoscopy.com.)

Comparison of DBE with Other Diagnostic Techniques 

DBE has been compared with PE and CE in patients with OGIB. In a preliminary retrospective study,26 DBE was reported to be superior to PE in terms of endoscope insertion depth. A subsequent controlled prospective trial14 on patients with suspected small-bowel bleeding confirmed that oral DBE is significantly superior to PE with regard to the length of small-bowel visualized (230 cm versus 80 cm), as well as the detection of pathological lesions (63% versus 44%). Moreover, in the same study, DBE identified additional lesions in deeper parts of the small bowel in 78% of patients who had positive findings at PE. This latter finding was also reported in another smaller study, in which patients who did not benefit from PE had a significant diagnostic and therapeutic outcome from DBE.27

Few studies have compared DBE with CE in patients with OGIB (Table 2). In a preliminary prospective blinded study, findings at DBE were concordant with those at CE in 12 of 13 patients.28 In a prospective trial, 52 of 90 patients who had chronic GI bleeding had previously undergone CE; the additional diagnostic yield of DBE compared with CE was reported to be approximately 20%, although a formal comparison between CE and DBE was not made.7 In a prospective study of 35 patients with OGIB, the diagnostic rate of small-bowel abnormalities by CE (80%) was significantly higher than by DBE (60%).22 Only 1 patient had an abnormality detected by DBE that had not been seen by CE, whereas DBE could not confirm findings detected by CE in 8 patients. These results are hypothesized to depend on an overestimation of CE findings. Another prospective study estimated the detection rates of CE and DBE in a series of 32 patients with suspected small-bowel bleeding: the diagnostic rate by CE was higher than that by DBE (59.4% versus 42.9%), although not significantly different.24 A multicenter retrospective study compared DBE with CE in a group of 115 patients with OGIB.23 CE found a potential bleeding source in 63 patients (55%; group 1) and was negative in 45% of patients (group 2). DBE was positive in 41 (65%) of the 63 patients in group 1, and in 16 (30%) of the 52 in group 2. Of note, DBE detected 4 large adenocarcinomas that were missed by CE. There was a significant agreement between DBE and CE in diagnosing ulcers, angioectasias, and large masses. Agreement between the 2 studies for mucosal and submucosal polyps was poor, potentially because of false readings of mucosal bulges on CE.

Table 2. Studies Comparing Diagnostic Yields of CE and DBE in Patients with Obscure GI Bleeding

	Author (Ref.)	No. of Patients	Results
	May et al. (7)	52	DBE>CE
	Hadithi et al. (22)	35	CE>DBE
	Mehdizadeh et al. (23)	115	CE≈DBE
	Nakamura et al. (24)	32	CE≈DBE
	Matsumoto et al. (28)	13	CE≈DBE

The Place of CE and DBE in the Management Algorithm of OGIB 

Overall, the diagnostic values of CE and DBE have been reported to be similar in this clinical setting, and this has been confirmed by a recent meta-analysis.29 A pragmatic approach would thus be to recommend that the diagnostic work-up of patients with OGIB be individualized on the basis of clinical presentation.30 Currently available data suggest that, in patients with obscure-occult bleeding, provided that there is no suspected obstruction, CE should be the first diagnostic test, to identify or exclude a bleeding intestinal lesion, and thus direct subsequent management.31, 32 DBE should be the confirmatory/therapeutic technique of choice. Since DBE can detect lesions that are missed by CE,15, 23, 33 it should be closely considered in patients with a strong suspicion of a small-bowel lesion despite negative CE.

Although CE has a higher performance when it is done during an active bleeding episode,34 the more evident the bleeding (“obscure–overt”), the more logical approach would seem to perform a “therapeutic DBE” without prior “purely diagnostic CE.” DBE should also be applied first if a small-bowel stricture is suspected (clinically or by other imaging techniques) and in patients with surgically modified anatomy (especially those with an afferent intestinal loop).

It must, however, be said that the most efficient and cost-effective ways to use these two procedures in patients with OGIB must be further investigated in randomized, prospective, controlled clinical studies.

Determination of the Primary Insertion Route 

The choice of either oral or anal route depends on the suspected location of the disease within the small bowel, in turn depending on clinical manifestations and the results of other diagnostic examinations. CE may indicate the preferential endoscope insertion route for DBE. Gay and coworkers35 reported the positive predictive value of CE in indicating the DBE route to be 95%, with a negative predictive value as high as 98%. Pennazio and coworkers36 used a similar capsule-directed DBE approach in a group of 44 patients, in whom a one-sided procedure (oral or anal) was sufficient to reach the lesion of interest in almost 90% of the DBE examinations. Obviously, in certain countries, the high cost of CE may preclude its use as a first-line diagnostic modality for small-bowel disease, or as a screening test before DBE.37 Stool color has also been used to select the DBE route: the oral route in the case of tarry stool and the anal route in the case of bright or dark red stool. However, this method appears somewhat empirical and remains to be validated. When stool color is indeterminate and CE is unavailable, one alternative strategy is to examine most of the intestine through the anus and the remainder through the mouth, because the anal approach usually causes less discomfort.38 In the same situation, given the technical difficulties of the anal approach, others39 recommend commencing with the oral approach. It is also easier to identify bleeding points with the oral approach in cases of active ongoing bleeding, because the blood flows toward the anus from bleeding points in the small intestine.38

In general, total enteroscopy by DBE is not required in the majority of the patients, as the potential bleeding source can be identified without it. Only about one-third of patients will require two separate DBEs to make diagnosis. Unfortunately, it may not always be possible to perform total enteroscopy: reported rates of total small-bowel intubation with DBE range from 0% to 86%.6, 27 Reasons for failing to achieve total enteroscopy include marked intestinal adhesion caused by previous abdominal or pelvic surgery,40 type of enteroscope used (the floppy nature of the P5 endoscope makes the retrograde approach more difficult, and passing the ileocecal valve can be troublesome), and the endoscopist’s level of experience. As discussed above, the strategy of performing CE before DBE may avoid unnecessary double DBE procedures.

Therapeutic Possibilities of DBE in Patients with OGIB 

Endoscopic therapy is an important aspect of DBE. Cauterization with an argon plasma coagulator (APC) or a heat probe, metal clips, and injection are all useful and effective options for endoscopic hemostasis. Devices for APC and injection can easily be introduced into the working channel of either the EN-450P5 or the EN-450T5. Metal clip devices can be introduced into the working channel of the EN-450T5. Angioectasias are the main bleeding source in the mid-GI-tract and are best treated by APC with or without prior injection of saline or diluted epinephrine (Fig. 3B). For large polyps with a broad base or a thick stalk, piecemeal resection is in general recommended to reduce the risk of complications. Before polypectomy, submucosal injection of diluted epinephrine–saline solution may be useful.41 Injection of sclerosing agents by DBE has been used to treat small-bowel varices.42

As far as angioectasias are concerned, doubt remains over whether they should be treated during endoscope advancement or withdrawal. There is a theoretical risk of perforation caused by radial pressure exerted by the balloons on cauterized areas if cauterization is performed during scope insertion. On the other hand, it can be difficult to locate previously discovered lesions for therapy on scope withdrawal. Mucosal trauma from enteroscope passage may also induce erythematous changes that mimic angioectasias. Therefore, it is advisable to treat all lesions during forward scope passage.43 Furthermore, because angioectasias may be found throughout the entire small bowel, it has been suggested to use the diagnostic device (P-type) because deep insertion into the small bowel is easier and faster.41

Safety of DBE in Patients with OGIB 

Multiple studies have demonstrated that the complication rate of diagnostic DBE appears to be acceptably low (<1%).6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 However, as more centers have reported their experience and the results of an increasing number of DBE procedures are known, it has become evident that DBE may carry a risk especially during or after therapeutic interventions. APC appears to be a safe treatment, associated with a low complication rate of approximately 1%; complications include perforation, enteritis, paralytic intestinal ileus.41, 44, 45 However, caution is important when coagulation procedures are performed at high-energy settings, due to the very thin wall of the small intestine. A case of intestinal necrosis as a complication of epinephrine injection therapy has also been reported.46 Complications such as bleeding or perforation may occur after polypectomy (Fig. 4A, B), especially in polyps larger than 3 cm. In this situation, the complication rate may reach 10%.41 In cases where endoscopic hemostasis is difficult, a tattoo placed on the luminal side allows identification of the small-bowel bleeding site from the serosal side, thereby limiting the extent of intestinal resection if it is later required.

View Large Image Download to PowerPoint

Figure 4. (A) Pedunculated jejunal polyp in a patient with Peutz–Jeghers syndrome. (B) Polypectomy. (Color version of figure is available at www.techgiendoscopy.com.)

Although the majority of endoscopic interventions using the DBE device can be performed safely, therapeutic endoscopists should be aware of the complication rate of endoscopic small-bowel therapy, which is higher than that of interventional colonoscopy.45

Impact on Patient Management and Outcomes 

The therapeutic impact of DBE may consist in the decision to start new treatment, change existing treatment, carry out surgical intervention, or therapeutic endoscopy.7, 47 As far as the latter point is concerned, reports show that endoscopic interventions by DBE are possible in approximately 35% (range 10-60%) of OGIB patients (Table 1). Importantly, some factors, such as a history of blood transfusions,15 the oral access route, patients with overt-OGIB, and those with highly relevant findings at prior CE,25 have been found to be significantly associated with a therapeutic benefit from DBE.

Few studies have examined the impact of DBE on the clinical outcome of OGIB patients. Hadithi and coworkers22 described the outcome of 35 patients followed during a short follow-up period (median 5 months, range 2-12). Twenty-six (74%) patients remained clinically stable at the end of the study, 18 (51%) of whom had received APC therapy. Manabe and coworkers17 reported that therapeutic measures induced by DBE led to complete resolution of bleeding in 17 (89%) of 19 patients with OGIB, after a mean follow-up period of 8.5 months. Kaffes and coworkers15 used a targeted approach to a group of 60 OGIB patients and a positive finding on CE. Over a mean follow-up period of 10 ± 5.2 months, the treatment performed as a result of the DBE led to the resolution of the bleeding problem in 47 (80%) patients; the reduction in transfusion rate and iron requirements was also significant. In a group of 20 patients with OGIB, Hsu and coworkers19 assessed outcome changes after DBE in terms of further bleeding episodes. Over a mean follow-up period of 12.6 months (range 3-28), rebleeding occurred significantly more frequently in patients in whom no definite lesion had been detected (80%) compared with those who had a lesion identified that was further treated (20%). Similar conclusions were reached in another larger study: in this retrospective report, Zhong and coworkers37 investigated 191 OGIB patients, of whom DBE revealed positive findings in 154 (81%). Patients were followed-up during a 6-month period, and a special scoring system for assessing the severity of OGIB was developed (variables included hemoglobin values, duration of disease, units of blood transfused, frequency of defecation, and features of stool). The mean score in the 154 patients with positive findings before DBE was 6.8 ± 2.2, and this dropped significantly to 1.5 ± 0.5 in patients who underwent specific (medical, surgical, or endoscopic) treatment. Although these preliminary results are promising, larger prospective randomized studies with adequate follow-up are warranted. Since methods such as DBE potentially offer greater guarantees of endoscopic therapeutic radicality than those used to date in patients who have OGIB, studies designed to specifically evaluate outcomes in subgroups of patients (ie, those with angioectasias) would also be of great interest.

Conclusions 

DBE is recognized as a milestone in gastrointestinal endoscopy. With the timely and widespread use of DBE, small-bowel bleedings are today more precisely diagnosed and, especially, more efficaciously treated than was possible previously. Whether this will ultimately result in enduring positive patient outcomes and cost savings remains to be determined through carefully designed studies.