Page 36 - Gastrointestinal Bleeding (Xuất huyết tiêu hóa)

310 PART III Symptoms, Signs, and Biopsychosocial Issues

should be palpated (“running the bowel”) to detect mass lesions. (42% vs. 6%). 363,364 A small series has found capsule endoscopy
In general, a standard exploratory laparotomy or laparoscopy is to be superior to CT enteroclysis for the diagnosis of obscure GI
performed first to lyse any adhesions and look for obvious tumors, bleeding because of its ability to identify angioectasias. 365
a Meckel diverticulum, or large vascular lesions. The small bowel An evaluation of published studies that have compared push
is usually extracted through the abdominal incision to allow the enteroscopy with capsule endoscopy in patients with obscure
surgeon to assist with advancement of an endoscope within the bleeding (79% overt, 21% occult) found that the average rate
lumen of the GI tract, which allows mucosal visualization as well of positive findings was 23% for push enteroscopy and 63%
as transillumination. Various endoscopes can be used (standard for capsule endoscopy. 309 A similar result was found in a meta-
upper endoscope and colonoscope, pediatric colonoscope, or analysis of published trials and abstracts; the diagnostic yield for
push enteroscope), depending on the route of access. The endo- push enteroscopy was 28% and 63% for capsule endoscopy. 364
scope can be passed transorally for a natural orifice luminal exam- A randomized trial that compared push enteroscopy with cap-
ination or via an enterotomy with use of a sterile endoscope. 359 sule endoscopy as a first-line approach to obscure GI bleeding
Because air insufflation will distend the entire small intestine and reported identification of a bleeding source in 24% of the push
thereby make laparoscopic or open visualization difficult, the sur- enteroscopy examinations and 50% of the capsule studies (P =
geon should pinch the intestine, manually or with an atraumatic .02). 366 In this study, capsule endoscopy missed lesions in 8% of
clamp, distal to the tip of the endoscope, to trap enough air to patients, and all the missed lesions were within reach of a stan-
permit visualization. Additionally, insufflation of the bowel with dard upper endoscope.
carbon dioxide, rather than room air, allows faster diffusion of A study of patients with acute, overt, unexplained GI bleeding
gas out of the bowel. The surgeon helps advance the endoscope (melena or hematochezia with nondiagnostic EGD and colonos-
by pleating the small bowel over the endoscope. Any lesion iden- copy) who were randomized to capsule endoscopy or angiography
tified can be addressed surgically or endoscopically, depending reported a significantly higher diagnostic rate for capsule endos-
on the nature of the lesion. Most series report complete enteros- copy than for angiography (53% vs. 20%) but no difference in the
copy of the entire small bowel in 50% to 75% of cases. 360,361 The long-term outcomes, including transfusions, hospitalizations, and
diagnostic yield of intraoperative enteroscopy ranges from 58% mortality. 367 Capsule endoscopy was compared with intraopera-
to 88%, but rebleeding after intraoperative enteroscopy has also tive endoscopy in one study of 47 patients who underwent both
been reported in 13% to 60% of patients. 309 The moderate per- procedures, primarily for obscure overt GI bleeding. 368 Using
formance characteristics, as well as risks of surgical exploration, intraoperative endoscopy as the gold standard, capsule endoscopy
limit this procedure as a diagnostic tool, but in selected patients, had a sensitivity of 95%, specificity of 75%, positive predictive
combined endoscopic and surgical evaluation can be useful and value of 95%, and negative predictive value of 85%. Most of the
definitive. bleeding lesions were angioectasias.
The role of intraoperative endoscopy in the management of Several studies have found that the diagnostic yield of cap-
severe obscure GI bleeding before versus after the introduction sule endoscopy increases in the setting of ongoing or recent (<2
of capsule endoscopy and deep enteroscopy has been reported. 361 weeks) overt GI bleeding or severe chronic GI bleeding (hemo-
Before an operation in the precapsule endoscopy era, a presump- globin <10 g/dL, iron deficiency anemia, or more than one overt
tive diagnosis or localization of bleeding site was achieved in 36% bleeding episode). 368-370 In a study from Greece of 34 patients
of patients compared with 63% in the postcapsule endoscopy who had active mild to moderate overt GI bleeding and nega-
era. In the precapsule endoscopy era, a definitive diagnosis was tive EGD and colonoscopy results, and who underwent an urgent
made intraoperatively in 100% of patients compared with 76% capsule endoscopy study while still in the hospital, the diagnostic
in the postcapsule endoscopy era. For lesions that were surgically yield was 92%, as defined by the identification of a bleeding lesion
resectable—small bowel tumors, Meckel diverticula, aortoenteric (18 angioectasias, 3 ulcers, 2 tumors) or the segment of intestine
29
fistula, and focal ischemic ulcers—no patient experienced postop- with bleeding (11 patients). By contrast, the same group from
erative bleeding during long-term follow up; however, rebleed- Greece found that the diagnostic yield of capsule endoscopy in
ing rates were high in other patients—67% with vascular lesions, patients with obscure occult bleeding and iron deficiency anemia
44% with small bowel ulcers, 50% with Crohn disease, and 63% was 57% (angioectasias in 24%, multiple jejunal or ileal ulcers in
with no definitive diagnosis. 361 Preoperative diagnosis with cap- 12%, multiple erosions in 8%, solitary ulcers in 6%, polyps in
sule endoscopy or deep enteroscopy has become more important, 4%, and other tumors in 4%). 371
particularly in older patients who may have significant compli-
cations or death from surgery. Careful selection of patients is Deep Enteroscopy of the Jejunum and Ileum
required (see Fig. 20.5). 363 Specially designed ultraflexible, 200-cm-long enteroscopes are
used in conjunction with an overtube to advance the endoscope by
Capsule Endoscopy pleating the small intestine over it. The available systems include
With capsule endoscopy, the patient ingests a pill camera that a double-balloon endoscope (with a balloon on the tip of the
transmits images of the small intestine for 8 hours or more. In endoscope and another balloon on the overtube), a single-balloon
patients with severe recurrent GI bleeding, this technique can system (a balloon on the overtube only), and a spiral overtube
identify a transition point at which fresh blood appears in the (no balloon used). All enteroscopes work by pleating the small
small bowel, and thereby localize the bleeding site and sometimes intestine over the endoscope. These enteroscopes can be inserted
identify a specific source lesion. 363 Capsule endoscopy does not orally (antegrade) and advanced into the proximal to midileum or
permit the application of therapy and can only localize a lesion inserted rectally (retrograde) and advanced to the distal to midi-
in the small bowel on the basis of the time of passage down the leum. Rarely, a complete enteroscopy of the small intestine to the
small intestine, as determined by sensors on the abdomen and cecum can be performed via the antegrade approach using the
telemetry. The information can be useful, however, in directing double-balloon enteroscope, whereas total enteroscopy (complete
subsequent therapeutic procedures such as deep enteroscopy, visualization of entire small bowel) using a combined antegrade
angiography, or surgery. Although capsule endoscopy may occa- and retrograde occurred in 44% of cases in a large review of pub-
sionally detect gastric, duodenal, or colonic lesions, it is not a lished studies. 372 Deep enteroscopy allows not only visualization
substitute for EGD and colonoscopy. but also interventions such as biopsy, hemostasis, and tattooing of
Compared with small bowel barium studies, capsule endoscopy lesions. The endoscopes used for deep enteroscopy have standard
has significantly improved detection rates for small bowel lesions working channels that allow passage of accessories such as biopsy
(67% vs. 8%) and findings that influence clinical management forceps, MPEC probes, hemoclips, and injection needles that fit