Lower Gastrointestinal Bleeding, Surgical Treatment
Introduction
Acute lower gastrointestinal (GI) hemorrhage accounts for
approximately 20% of all cases of GI hemorrhage. The annual
incidence is about 20-27 cases per 100,000 population in
westernized countries. Lower GI hemorrhage continues to be
a frequent cause of hospital admission and is a factor in
hospital morbidity and mortality. Mortality rates are
reportedly 10-20% and are dependent on age (>60 y),
multiorgan system disease, transfusion requirements in
excess of 5 units, need for operation, and recent stress (eg,
surgery, trauma, sepsis).
Localization of hemorrhage relative to the Treitz ligamentum
directs the initial evaluation and resuscitation. The passage of
maroon stools or bright red blood from the rectum is usually
indicative of massive lower GI hemorrhage. Lower GI
hemorrhage can be due to numerous conditions,
including diverticulosis,
anorectal
diseases,
carcinomas, inflammatory
bowel
disease (IBD),
and angiodysplasias. For types of GI bleeding and an
algorithm for massive bleeding, see the images below.
History of the Procedure
Understanding of the pathogenesis, diagnosis, and treatment
of lower GI bleeding has drastically changed during the last
50 years. In the first half of the 20th century, large intestinal
neoplasms were believed to be the most common cause of
lower GI bleeding. In the 1950s, lower GI hemorrhage was
commonly attributed to diverticulosis. In this period, surgical
treatment consisted of blind segmental bowel resections,
with disappointing results. Patients who underwent blind
segmental bowel resection suffered from a prohibitively high
rebleeding rate (up to 75%), morbidity (up to 83%), and
mortality (up to 60%).
In the last 4 decades, diagnostic methods for locating the
precise bleeding point greatly improved. In 1965, Baum et al
described selective mesenteric angiography, which permitted
the identification of vascular abnormalities and the precise
bleeding point.1 Experience with mesenteric angiography in
the late 1960s and 1970s suggested that angiodysplasias and
diverticulosis were the most common reasons for lower GI
bleeding. Since its discovery, mesenteric angiography
remains the criterion standard in precise localization of the
bleeding.
Rösch et al described superselective visceral arteriography
for infusion of vasoconstrictors in 1971 and superselective
embolization of the mesenteric vessels as an alternative
technique to treat massive lower GI bleeding in 1972.2 The
most feared complication of embolization of the mesenteric
vessels is ischemic colitis, which has limited its use for GI
bleeding.
The initial experience with vasopressin infusion was reported
in 1973-1974. Vasopressin causes vasoconstriction and
arrests the bleeding in 36-100% of patients. The recurrence
rate following completion of vasopressin infusion can be as
high as 71%; therefore, vasopressin is used to temporize the
acute event and to stabilize patients before surgery.
The flexible endoscope was developed in 1954. The fulllength colonoscope was developed in 1965 in Japan. The first
anal colonoscopy was performed in 1969. Endoscopic control
of bleeding with thermal modalities or sclerosing agents has
been in use since the 1980s. One of the advantages of upper
(or lower) endoscopic evaluation is that it provides a means
to administer therapy in patients with GI bleeding. Nuclear
scintigraphy has been used since the early 1980s as a very
sensitive diagnostic tool to evaluate bleeding from GI tract.
Nuclear scintigraphy can detect hemorrhage at rates as low
as 0.1 mL/min.
The average age of patients with lower GI bleeding is 60
years in most series. Etiology varies according to the age of
the patient. Segmental bowel resection following precise
localization of the bleeding point is a well-accepted surgical
practice today. Despite improvement in diagnostic imaging
and procedures, 10-20% of patients with lower GI bleeding
have no demonstrable bleeding source. Subtotal colectomy is
the procedure of choice in patients who are actively bleeding
from an unknown source.
Problem
Lower GI hemorrhage is defined as an abnormal intraluminal
blood loss from a source distal to the Treitz ligamentum.
Lower GI bleeding is classified under 3 groups according to
the amount of bleeding. These groups are shown in the
image below. Massive hemorrhage is a life-threatening
condition and requires transfusion of at least 5 units of
blood.
Patients with massive hemorrhage present with a systolic
blood pressure of less than 90 mm Hg and a hemoglobin level
of 6 g/dL or less. These patients are usually aged 65 years and
older, have multiple medical problems, and are at risk of
death from acute hemorrhage or its complications.
Therefore, the overall mortality rate for massive lower GI
hemorrhage ranges from 0-21%. Occult bleeding manifests as
microcytic hypochromic anemia and intermittent guaiac
reaction.
Definition of massive lower GI bleeding
1. Passage of a large volume of red or maroon blood
through the rectum
2. Hemodynamic instability and shock
3. Initial decrease in hematocrit (Hct) level of 6 g/dL or less
4. Transfusion of at least 2 units of packed RBCs
5. Bleeding that continues for 3 days
6. Significant rebleeding in 1 week
Frequency
The total incidence of lower GI bleeding in the United States
is not known; although lower GI bleeding is common, most
patients do not require hospitalization.
Vernava et al reviewed Department of Veterans Affairs' (VA)
databases for a 4-year period to study the incidence and
etiology of lower GI bleeding. They found that less than 1% of
5.1 million hospital admissions were for lower GI
hemorrhage. Another study, in which the Kaiser Permanente
database was reviewed, estimated an annual incidence rate
of 20.5 patients per 100,000 (24.2 in males vs 17.2 in
females). The rate of lower GI bleeding increased more than
200-fold from the third to the ninth decades of life.
Etiology
Bleeding from diverticular disease has been reported as the
most common reason for massive lower GI bleeding in most
of the single-institution publications. However, the reported
frequency of various other etiologies of lower GI bleeding is
not consistent in these manuscripts because of the small
number of cases and the highly selective referral pattern and
patient populations. Comprehensive knowledge of the
etiology of lower GI bleeding is essential for patient
management and, ultimately, for patient outcome.
In a retrospective review of medical records from
approximately 1100 patients with acute lower GI bleeding, all
of whom were admitted to the surgical service of a single
urban emergency hospital, Gayer et al determined that the
most common etiologies for bleeding in these patients were
diverticulosis (33.5%), hemorrhoids (22.5%), and carcinoma
(12.7%).3 The investigators also found that most patients in
the study (55.5%) presented with hematochezia, with the
next most frequent presentations being maroon stools
(16.7%) and melena (11%).
Vernava and colleagues' review found that patients with
lower GI bleeding made up only 0.7% of all hospital
admissions (17,941 patients). The average age of these
patients was 64 years. Only 24% of these patients (4410) had
a diagnostic workup, including colonoscopy, barium enema,
and/or mesenteric angiography. Among the patients who
underwent a diagnostic workup, the most common causes of
bleeding were diverticular disease (60%), IBD (13%), and
anorectal diseases (11%) (see Table 1), the figures differing
somewhat from the above-mentioned study by Gayer et al.
Although some publications have reported arteriovenous
malformations as a common cause of lower GI bleeding, the
true incidence of arteriovenous malformations is insignificant
(3%), as stated by Vernava et al.
Table 1: Common Causes of Lower GI Bleeding in Adults*
LOWER GI HEMORRHAGE IN ADULTS
% OF PATIENTS
Diverticular disease
60%
-Diverticulosis/diverticulitis of small
intestine
-Diverticulosis/diverticulitis of colon
IBD
13%
-Crohn's disease of small bowel, colon, or
both
-Ulcerative colitis
-Noninfectious gastroenteritis and colitis
Benign anorectal diseases
11%
-Hemorrhoids
-Anal fissure
-Fistula-in-ano
Neoplasia
9%
-Malignant neoplasia of small intestine
-Malignant neoplasia of colon, rectum,
and anus
Coagulopathy
4%
Arteriovenous malformations (AVM)
3%
TOTAL
100%
*From Vernava and colleagues' survey of 4410 patients
Longstreth reviewed the discharge summary and
colonoscopy data from a large health maintenance
organization with members in the San Diego, Calif, area. In
all, 235 hospital admissions for 219 patients were reviewed.
The estimated hospital admission rate for lower GI bleeding
was found to be 20 patients per 100,000 admissions.
Bleeding from diverticular disease was the most common
reason for lower GI bleeding (42%), followed by colorectal
malignancies (9%) and ischemic colitis (8.7%). The incidence
of lower GI bleeding due to colonic angiodysplasias was 6%.
These findings were consistent with those of the VA database
study, although that study was limited to males.
The common causes of lower GI bleeding in infants, children,
and adolescents differ from those found in adults. Meckel
diverticulum, intussusception, polyposis syndromes, and IBD
are the common causes of GI bleeding in children and
adolescents (see Table 2).4
Table 2: Common Causes of Lower GI Hemorrhage in Children
and Adolescents
LOWER GI HEMORRHAGE IN CHILDREN & ADOLESCENTS
Intussusception
Polyps and polyposis syndromes
-Juvenile polyps and polyposis
-Peutz-Jeghers syndrome
-Familial adenomatous polyposis (FAP)
IBD
-Crohn's disease
-Ulcerative colitis
-Indeterminate colitis
Meckel diverticulum
Many other causes of lower GI bleeding have been
documented, including hemorrhage from small bowel
diverticulosis, Dieulafoy lesions of the colon or small bowel,
portal colopathy with colonic and rectal varices,
endometriosis, solitary rectal ulcer syndrome, and
vasculitides with small bowel or colonic ulcerations (see Rare
causes of lower GI bleeding).
Pathophysiology
Diverticulosis is a common acquired condition in Western
societies. Approximately 50% of adults older than 60 years
have radiologic evidence of diverticulosis. Diverticula are
most commonly located in the sigmoid and descending colon.
Diverticular bleeding originates from vasa rectae located in
submucosa, which can rupture at the dome or the neck of
the diverticulum. Up to 20% of patients with diverticular
disease experience bleeding. In 5% of patients, bleeding from
diverticular disease can be massive. Hemorrhage from
diverticular disease stops spontaneously in 80% of patients.
Although diverticulosis is a left colonic condition,
approximately 50% of diverticular bleeding originates from a
diverticulum located proximal to the splenic flexure.
Diverticula located on the right side may expose the larger
portions of vasa rectae to injury because they have wider
necks and larger domes compared to the typical left-sided
colonic diverticulum.
Colonic angiodysplasias are arteriovenous malformations
located in the cecum and ascending colon. Colonic
angiodysplasias are an acquired lesion affecting elderly
persons older than 60 years. These lesions are composed of
clusters of dilated vessels, mostly veins, in the colonic
mucosa and submucosa. Colonic angiodysplasias are believed
to occur as a result of chronic, intermittent, low-grade
obstruction of submucosal veins as they penetrate the
muscular layer of the colon. The characteristic angiographic
findings are clusters of small arteries during the arterial
phase of the study, accumulation of contrast media in
vascular tufts, early opacification, and persistent
opacification due to the late emptying of the draining veins. If
mesenteric angiography is performed at the time of active
bleeding, extravasation of contrast media is visualized.
Unlike diverticular bleeding, angiodysplasia tends to cause
slow but repeated episodes of bleeding. Therefore, patients
with angiodysplasia present with anemia and syncopal
episodes. Infrequently, angiodysplasias can cause an abrupt
loss of large quantities of blood. Angiodysplasias can be easily
recognized by colonoscopy as 1.5- to 2-mm red patches in
the mucosa. Actively bleeding lesions can be treated with
colonoscopic electrocoagulation. Incidentally discovered
lesions should be left alone.
Massive hemorrhage due to IBD is rare. Ulcerative colitis
causes bloody diarrhea in most cases. In up to 50% of
patients with ulcerative colitis, mild-to-moderate lower GI
bleeding occurs, and approximately 4% of patients with
ulcerative colitis have massive hemorrhage.
Lower GI bleeding in patients with Crohn’s disease is not as
common as in patients with ulcerative colitis; 1-2% of
patients with Crohn’s disease may experience massive
bleeding. The frequency of bleeding in patients with Crohn’s
disease is significantly more common with colonic
involvement than with small bowel involvement alone.
Ischemic colitis, the most common form of ischemic injury to
the digestive system, frequently involves the watershed
areas, including the splenic flexure and the rectosigmoid
junction. In most cases, the precipitating event cannot be
identified. Colonic ischemia is a disease of the elderly
population and is commonly observed after patients' sixth
decade of life. Ischemia causes mucosal and partial-thickness
colonic wall sloughing, edema, and bleeding. Ischemic colitis
is not associated with significant blood loss or hematochezia,
although abdominal pain and bloody diarrhea are the main
clinical manifestations.
Colorectal adenocarcinoma is the third most common cancer
in the United States. Colorectal carcinoma causes occult
bleeding, and patients usually present with anemia and
syncopal episode. The incidence of massive bleeding due to
colorectal carcinoma varies from 5-20% in different series.
Postpolypectomy hemorrhage is reported to occur up to 1
month following colonoscopic resection. The reported
incidence is between 0.2-3%. Postpolypectomy hemorrhage
can be managed by electrocoagulation of the polypectomy
site/bleeding with either snare or hot biopsy forceps or by
epinephrine injection.
Benign anorectal disease (eg, hemorrhoids, anal fissures,
anorectal fistulas) can cause intermittent rectal bleeding.
Massive rectal bleeding due to benign anorectal disease has
also been reported. The VA database review revealed that
11% of patients with lower GI bleeding had hemorrhage from
anorectal disease. Patients who have rectal varices with
portal hypertension may develop painless massive lower GI
bleeding; therefore, examining the anorectum early in the
workup is important. If active bleeding is identified, treat it
aggressively. Note that the discovery of benign anorectal
disease does not exclude the possibility of more proximal
bleeding from the lower GI tract.
Presentation
Massive lower GI bleeding is a life-threatening condition.
Although massive lower GI bleeding manifests as maroon
stools or bright red blood from the rectum, patients with
massive upper GI bleeding may also present with similar
findings. Regardless of the level of the bleeding, one of the
most important elements of the management of patients
with massive upper or lower GI bleeding is the initial
resuscitation. These patients should receive 2 large-bore
intravenous catheters and isotonic crystalloid infusions.
Meanwhile, rapid assessment of vital signs, including heart
rate, systolic blood pressure, pulse pressure, and urine
output, should be performed. Orthostatic hypotension (ie, a
blood pressure fall of >10 mm Hg) is usually indicative of
blood loss of more than 1000 mL.
History and physical examination are essential parts of an
initial evaluation. Document prior episodes of GI bleeding as
well as significant medical history and prior medications,
including peptic ulcer disease, liver disease, cirrhosis,
coagulopathy, IBDs, and use of nonsteroidal antiinflammatory drugs (NSAIDs) and/or warfarin. Symptoms are
also important in identifying the source of bleeding. The
symptoms of young patients with abdominal pain, rectal
bleeding, diarrhea, and mucous discharge may be associated
with IBD. On the other hand, symptoms of elderly patients
with abdominal pain, rectal bleeding, and diarrhea can be
associated with ischemic colitis. Stools streaked with blood,
perianal pain, and blood drops on the toilet paper or in the
toilet bowl may be associated with perianal pathology, such
as anal fissure or hemorrhoidal bleeding.
The physical examination must include careful inspection and
examination of the oropharynx, nasopharynx, abdomen,
perineum, and anal canal. Nasogastric aspirates usually
correlate well with upper gastric hemorrhage proximal to the
Treitz ligamentum; therefore, insert a nasogastric tube to
confirm the presence or absence of blood in the stomach. If
necessary, perform gastric lavage with warm isotonic fluids
to obtain bilious discharge from the nasogastric tube to
exclude any upper GI bleeding beyond the pylorus.
Nasogastric tube aspirates can provide false-negative results
in approximately 50% of cases if the aspirate contains no bile
or if the bleeding is intermittent. These patients eventually
need esophagogastroduodenoscopy (EGD) to obtain a more
specific evaluation of the upper GI tract. Place a Foley
catheter to monitor urine output. Careful digital rectal
examination, anoscopy, and rigid proctosigmoidoscopy
should exclude an anorectal source of bleeding.
Indications
Surgical treatment is indicated if the patient continues to
bleed and if nonoperative management is unsuccessful or
unavailable. Segmental colectomy is indicated if the bleeding
point is localized by preoperative diagnostic studies. Subtotal
colectomy is the procedure of choice if the bleeding point
cannot be localized with preoperative or intraoperative
diagnostic studies. Subtotal colectomy is associated with
negligibly higher perioperative morbidity and mortality
compared to segmental colonic resection. In addition,
postoperative diarrhea can be a significant problem in elderly
patients who undergo subtotal colectomy and ileorectal
anastomosis.
Relevant Anatomy
The average length of the large intestine is 135-150 cm.
Ascending and descending segments of the colon are fixed to
the retroperitoneum. On the other hand, the transverse and
sigmoid colon are supported by a mesentery in the abdomen.
A comprehensive understanding of small bowel and colonic
vascular anatomy is essential for any surgeon performing
primary lower GI surgery for hemorrhage or other diseases.
The ileocolic, right colic, and middle colic branches of the
superior mesenteric artery supply blood to the cecum,
ascending, and proximal transverse colon, respectively. The
superior mesenteric vein drains the right side of the colon,
joining the splenic vein to form the portal vein. The inferior
mesenteric artery supplies blood to the distal transverse,
descending, and sigmoid colon. The inferior mesenteric vein
carries blood from the left side of the colon to the splenic
vein. A rich network of vessels from the superior, middle, and
inferior hemorrhoidal vessels supplies the rectosigmoid
junction and rectum.
Contraindications
No contraindications exist with regard to surgery in
hemodynamically unstable patients with active bleeding.
Surgery is warranted even in the absence of accurate
preoperative localization for patients who require transfusion
of 5 units or more blood in the first 24 hours. Surgery is also
necessary in patients with recurrent bleeding during the
same hospitalization.
Workup
Laboratory Studies
Appropriate blood tests include CBC; serum electrolytes
(sequential multiple analysis 7 [SMA7]); and coagulation
profile, including activated partial thromboplastin time
(aPTT), prothrombin time (PT), manual platelet count, and
bleeding time.
Imaging Studies
1. The role of nuclear scintigraphic imaging in the diagnosis
and treatment of patients who present with lower GI
bleeding remains controversial. Nuclear scintigraphy is a
sensitive diagnostic tool (86%) and can detect
hemorrhage at rates as low as 0.1 mL/min. Nuclear
scintigraphy is reportedly 10 times more sensitive than
mesenteric angiography in detecting ongoing bleeding.
The scintigraphic imaging suffers from a low specificity
(50%) due to its limited resolution; this has led many
investigators to recommend that scintigraphic imaging
be used primarily as a screening examination to select
patients for mesenteric angiography. See the algorithm
for GI bleeding in the image below.
2. No preparation is required for99m sulfur colloid. This
agent has a very short half-life (2.5-3.5 min) because it is
rapidly cleared by the reticuloendothelial system.
Because it enhances the liver and spleen, bleeding from
both the hepatic flexures and the splenic flexures may be
obscured.99m Tc-labeled RBC scintigraphy is the preferred
technique because its half-life is longer. Images delayed
up to 24 hours can be taken with labeled RBC scanning.
3. The sensitivity of the99m Tc-labeled RBC scintigraphy is
reportedly 20-95%. The bleeding site can be identified
accurately when intraluminal accumulation of 99m Tclabeled RBCs is observed during the dynamic phase of
scanning. Although nuclear scintigraphy is sensitive
enough to diagnose ongoing bleeding at a rate as low as
0.1 mL/min, it is not highly accurate in locating the
bleeding point. The bleeding point is accurately localized
in 52-90% of positive cases, with an average of 86% and
incorrect localization of 14%, as reported in 24
publications. Because of the high false localization rate
(10-60%) for the bleeding site, performing segmental
resections based solely on scintigraphy results is not
recommended.
4. Ng and colleagues reviewed 86 patients with
positive99m Tc-labeled RBC scintigraphy findings.5 Patients
with an immediate blush (within 2 min of the study)
revealed a positive predictive value of 75% for
angiography. Patients with a delayed blush (after 2 min
of the study) had a negative predictive value of 93% for
angiography. Thus, patients with delayed blush should
proceed with colonoscopic evaluation instead of
mesenteric
angiography.
Use99m Tc-labeled
RBC
scintigraphy as a prescreening test for selective
mesenteric angiography.
5.



6.


In 1992, Ryan et al published their experience with 99m Tclabeled RBC scintigraphy. In this study, 29 patients with
lower GI bleeding were identified. Scintigraphy identified
the site of bleeding accurately in 9 patients with massive
lower GI bleeding. In 6 of 9 patients, the scintigraphy
finding was positive in the first 5 minutes of the study. In
3 patients, the scintigraphy finding was positive at 14-45
minutes.
Another study was performed to evaluate the efficacy of
RBC scintigraphy in confirming the location of the lower
GI bleeding. Twenty-one patients with positive
scintigraphy results were included in the study. Of these,
the bleeding site was confirmed in 16 patients by various
methods. RBC scintigraphy findings were positive within
the continuous phase of the study in 10 of the confirmed
studies and in none of the incorrectly localized studies.
Therefore, in carefully selected cases, patients can
undergo segmental resections only if scintigraphy
findings are strongly positive in the very initial part of the
test.
Cinematic99m Tc-labeled RBC scintigraphy (real-time
scanning) has been described as a noninvasive
alternative to mesenteric angiography. Continuous
dynamic imaging using sequential computer acquisition
provides more accurate localization of the bleeding point
because it enables cinematic playback. More studies are
necessary to identify the success of real-time
scintigraphic evaluations.
Recurrent lower GI bleeding occurs after negative99m Tclabeled RBC scintigraphy. Hammond et al conducted
retrospective evaluations of 84 patients with
negative99m Tc-labeled RBC scintigraphy. The overall
rebleeding rate was found to be 27% (n=23 patients).
Hammond et al concluded that age, gender, bleeding
source, use of anticoagulant/antiplatelet agents, length
of hospital stay, admission Hct, Hct nadir, and
transfusion requirements are not predictive of the
patients who will rebleed.
The use of111 indium–labeled RBC scintigraphy to detect
intermittent bleeding has been described in the medical
literature in a handful of publications.
Ferrant and colleagues initially used111 indium–labeled
RBC scintigraphy in patients with lower GI bleeding in
1980; however, it remains underutilized because of a
prolonged half-life of 67 hours. This scintigraphy is more
expensive and also is a more labor-intensive technology
than99m Tc labeling. The image quality and localization of
bleeding can be less than desirable because of the
prolonged half-life and intestinal motility. However, the
longer half-life of111 indium–labeled RBC scintigraphy can
be useful in locating intermittent bleeding points,
particularly when conventional methods have failed.
Schmidt et al published a report on 6 patients in
whom99m Tc
scanning
was
initially

7.
8.
9.

unrewarding.6Subsequent scintigraphy with111 indium–
labeled RBCs located the site of bleeding in all patients.
Mole et al detected synchronous, small and large
intestinal adenocarcinomas with111 indium–labeled RBC
scintigraphy in a 70-year-old patient with intermittent GI
bleeding and profound blood loss anemia.7
In 1965, Baum et al described selective mesenteric
angiography in the diagnosis of GI bleeding. 1 Since then,
the value of mesenteric angiography in the diagnosis and
management of lower GI bleeding has been well
established. The extravasation of contrast material
indicates a positive study finding. Selective mesenteric
angiography can detect bleeding at a rate of more than
0.5 mL/min. In a patient with active GI bleeding, the
radiologist concentrates on the major mesenteric vessel
most likely to be responsible (eg, the inferior mesenteric
artery in bright red rectal bleeding). If no bleeding is
identified, the other major mesenteric vessels, including
the superior mesenteric artery and celiac axis, are
studied. In some cases, aberrant vascular anatomy can
contribute to colonic or small bowel circulation; in other
cases, patients with upper GI bleeding may present in an
uncommon clinical fashion.
Helical CT scan of the abdomen and pelvis can also be
used when routine workup fails to determine the cause
of active GI bleeding. Multiple criteria, including vascular
extravasation of the contrast medium, contrast
enhancement of the bowel wall, thickening of the bowel
wall, spontaneous hyperdensity of the peribowel fat, and
vascular dilatations, are used to establish the bleeding
site with helical CT. The presence of diverticula alone
was not enough to define the bleeding site. Three-phase
helical CT should be performed using intravenous
contrast. Water can be used as an oral contrast in the
workup of patients who are actively bleeding. Therefore,
helical CT could be a good diagnostic tool in acute lower
GI bleeding to help the physician identify the bleeding
site.
A pilot study was done in Sydney, Australia, to evaluate
CT as a diagnostic tool for acute lower GI bleeding.
Helical CT was compared to selective mesenteric
angiography and colonoscopy in the diagnosis and
detection of a bleeding site. Seven patients with acute
lower GI bleeding were included. All patients underwent
mesenteric angiography following CT. Colonoscopies
were also performed on 5 patients investigated with
both CT and mesenteric angiography. Both modalities
had concordant findings of 2 active bleeding sites, 1
nonbleeding rectal tumor, and 1 negative result. In 3
patients, the source of bleeding was found on CT,
whereas the mesenteric angiography finding was
negative. Colonoscopies performed in these 3 patients
confirmed blood in the colon/ileum.

10.
11.
12.
13.
Sabharwal et al concluded that helical CT is a safe,
convenient, and accurate diagnostic tool for acute lower
GI hemorrhage.8 The authors proposed a new
management algorithm for acute lower GI hemorrhage
using CT as the preselective mesenteric angiography
screening tool.
Frattaroli et al examined the sensitivity of multi-detector
row CT (MDCT) scanning in identifying the site and
etiology of acute upper (11 patients) and lower (18
patients) GI bleeding.9 Comparing this modality with
endoscopy, the investigators reported that, in terms of
identifying the anatomic location and etiology of upper
GI bleeding, MDCT had a sensitivity of 100% and 90.9%,
respectively, while endoscopy had a sensitivity of 72.7%
and 54.5%. For lower GI bleeding, MDCT had a sensitivity
for site and etiology identification of 100% and 88.2%,
respectively, while endoscopy had a sensitivity of 52.9%
for both identifications.
Once the bleeding point is identified, angiography offers
potential treatment options, such as selective
vasopressin drip and embolization. Thirteen publications
reported experiences with selective mesenteric
angiography. When 657 patients underwent mesenteric
angiography, the percentage of positive study findings
fluctuated between 27-86%, with an average of 45%.
Because of the intermittent nature of lower GI bleeding,
the number of positive study findings is significantly less
with this invasive diagnostic modality.
Emergency angiography as an initial study is indicated in
a highly selected group of patients with massive ongoing
lower GI bleeding. Browder et al used 2 criteria to triage
patients for emergency angiography.10 The criteria were
at least 4 units of blood transfusion in the first 2 hours
following hospital admission and systolic blood pressure
of less than 100 mm Hg with aggressive resuscitation.
Fifty patients underwent emergency angiography, and
bleeding was localized in 72% of patients. Vasopressin
infusion was successful in 91%; however, half
experienced bleeding following cessation of the
vasopressin infusion. Thus, patients with ongoing
hemorrhage, emergency angiography, and vasopressin
infusion have improved operative morbidity, mortality,
and outcome.
Five to 10% of patients may present with recurrent
episodes of massive lower GI bleeding without any
diagnosis of the bleeding site. These patients experience
multiple hospital admissions; they also undergo
recurrent blood transfusions and several invasive studies
repeatedly. Ryan et al performed 17 elective provocative
bleeding studies for occult lower GI bleeding in 16
patients.11 Although an abnormality was identified in
50% of patients, bleeding was provoked in 6 (37.5%)
patients. Most of the positively provoked patients (ie, 5
patients) had a previously positive tagged red cell
scintigraphy.11 Of the 6 patients with provoked bleeding,
3 were treated with superselective embolization at the
time of provoked bleeding, 2 were treated with estrogen
therapy, and 1 was treated with palliative therapy. 11 Ten
patients did not bleed during the provoked study.11
14. Widlus and Salis reported 9 patients who underwent
provocative angiography with Reteplase, a new
fibrinolytic agent.12 An initial diagnostic visceral
arteriogram was performed and failed to identify the
source of bleeding in each patient.12 Reteplase was
administered, and provocative arteriography was
repeated. Bleeding was identified in 8 (89%) patients,
and these patients were treated with microembolization,
segmental resection, or conservatively.12 It was
concluded that the use of Reteplase is safe and effective
as a provocative agent, stimulating bleeding to allow
localization, in patients with occult, recurrent, massive
lower GI bleeding.12
Other Tests
1. Double-contrast barium enema examinations can be
justified only for elective evaluation of unexplained
lower GI bleeding. Do not use barium enema
examination in the acute hemorrhage phase because it
makes subsequent diagnostic evaluations, including
angiography and colonoscopy, impossible.
2. Elective contrast radiography of the small bowel and/or
enteroclysis is often valuable in investigation of longterm, unexplained lower GI bleeding.
Diagnostic Procedures
Colonoscopy has an important role in the diagnosis and
treatment of lower GI bleeding. Rapid colonic lavage with
GoLYTELY clears the intraluminal blood, clot, and stool,
providing an adequate environment for visualization of the
lower GI mucosa and lesions. GoLYTELY can be administered
orally or by nasogastric tube. The best candidates for
colonoscopic evaluation are patients who are bleeding slowly
or who have already stopped bleeding.
Histologic Findings
Most colonic diverticula are false pulsion diverticula and are
composed only of mucosa and submucosa herniated through
the colonic wall musculature. Hemorrhage associated with
diverticula comes from perforated vasa rectae located at the
neck or the apex of the diverticula.
Colonic angiodysplasias are vascular ectasias commonly
located on the right side of the colon. Microscopically,
vascular ectasia consists of dilated thin-walled venules and
capillaries localized in the submucosa of the colonic wall.
Treatment
Medical Therapy
Vasoconstrictive agents
Initially, vasoconstrictive agents, such as vasopressin
(Pitressin), can be used. An experimental study of treatment
of lower GI bleeding by selective arterial infusion of
vasoconstrictors, such as epinephrine with propranolol and
vasopressin, was reported. Although epinephrine and
propranolol drastically reduced mesenteric blood flow, they
also caused a rebound increase in blood flow and recurrent
bleeding. Vasopressin is a pituitary hormone that causes
severe vasoconstriction in the splanchnic bed.
Vasoconstriction reduces the blood flow and facilitates
hemostatic plug formation in the bleeding vessel. The results
are less than satisfactory in patients with severe
atherosclerosis and coagulopathy.
Following positive angiogram findings, the angiographic
catheter is left in place and vasopressin infusion is started at
a rate of 0.2 unit/min. A repeat angiogram is obtained every
half hour, and the rate of infusion is increased up to 0.4
unit/min if bleeding continues. Vasopressin doses above 0.4
unit/min are not recommended because of the high rate of
potential complications. If hemorrhage remains controlled,
the dose of vasopressin is reduced to half every 6-12 hours.
The angiographic catheter, following an additional 6-12 hours
of saline infusion, is removed. If vasopressin infusion fails to
control the hemorrhage, patients should undergo a
segmental resection.
The initial experience with vasopressin infusion was reported
in 1973-1974. Twenty-four patients were included in this
study. In 22 of the patients, bleeding was controlled. Of
these, 12 received no further therapy and were discharged.
Three patients (25%) developed recurrent bleeding within 212 months of discharge. Selective vasopressin infusion was
used as the sole treatment and arrested the bleeding in 36100% of the cases. Because the rebleeding rates fluctuated
between 27-71%, vasopressin infusion was used in the acute
event to stabilize patients prior to surgery.
During vasopressin infusion, monitor patients for recurrent
hemorrhage,
myocardial
ischemia,
arrhythmias,
hypertension, and volume overload with hyponatremia.
Nitroglycerine paste or drip can be used to overcome cardiac
complications. Selective mesenteric infusion induces bowel
wall contraction and spasms, which should not be confused
with bowel wall ischemia. Do not administer vasopressin into
systemic circulation intravenously because this causes
coronary vasoconstriction, diminished cardiac output, and
tachyphylaxis.
Superselective embolization
Superselective embolization of the mesenteric vessels is an
alternative technique for treating massive lower GI bleeding.
Rösch and colleagues first described this technique in 1972.
Autologous clot, Gelfoam, polyvinyl alcohol, microcoils,
ethanolamine, and oxidized cellulose can be used as embolic
agents. Embolization involves superselective catheterization
of the bleeding vessel to minimize necrosis, the most feared
complication of ischemic colitis.
Rosenkrantz et al reported 3 cases of colonic infarction. 14 One
patient died following segmental colectomy, and the other
patients revealed full-thickness bowel wall injury in the
resected specimen. Intestinal ischemia and infarction have
also been reported. To prevent this complication, perform
embolization beyond the marginal artery as close as possible
to the bleeding point in the terminal mural arteries. A total of
139 cases have been collected from the medical literature
since 1972.
Overall bleeding was controlled in 115 patients (83%), with a
rebleeding rate of 11% (15 patients). Complications were
observed in 20%, and bowel injury and perforation were
observed in 12% (16 patients). The overall mortality rate was
11% (15 patients); thus, careful patient selection is necessary
for this procedure. Use embolization in high-risk patients
whose conditions are refractory to conservative
management. If terminal mural branches of the bleeding
vessel cannot be catheterized, abort the procedure and
immediately perform surgery.
Kuo et al evaluated the safety and effectiveness of
superselective microcoil embolization for the treatment of
lower GI bleeding in 2003.15 Twenty-two patients with
angiographic evidence of lower GI bleeding underwent
superselective microcoil embolization during a 10-year
period. Complete clinical success was achieved in 86% of
patients with a rebleeding rate of 14%. Minor and major
ischemic complication rates were reported as 4.5% and 0%,
respectively.
The authors also reviewed the data from 122 cases of lower
GI superselective microcoil embolization in the literature. The
meta-analysis was performed in 144 patients. This combined
analysis revealed a minor ischemic complication rate of 9%
and a major ischemic complication rate of 0%. It was
concluded that superselective microcoil embolization is a
safe and effective treatment of acute lower GI hemorrhage.
Colonoscopy
Colonoscopy has become the first choice of diagnostic
modality following rapid purge with volume cathartics, such
as GoLYTELY. Jensen and Machicado have evaluated the role
of urgent colonoscopy after purge prospectively in 80
consecutive patients with severe hematochezia. 16 Urgent
colonoscopies were performed in the intensive care unit.
Seventy-four percent of patients had colonic lesions, 11% had
upper GI lesions, and 9% had presumed small bowel lesions;
in 6%, no bleeding site was identified. Although Jensen and
Machicado recommended that EGD be performed prior to
colonoscopy, upper and lower endoscopies can be performed
simultaneously.
In another study, colonoscopy yielded a diagnosis in 90% of
the patients, which provided opportunity for therapy at the
same time. The patients who underwent colonoscopic
evaluation had a significantly shorter hospital stay. Perform
the urgent colonoscopy in the operating room or endoscopy
suite on hemodynamically stable patients. If patients become
unstable or colonoscopy reveals an active fulminant
inflammation, abort the procedure.
Endoscopic coagulation
The treatment options for angiodysplasias are numerous,
including segmental bowel resection and selective
mesenteric embolization. Endoscopic coagulation of
angiodysplasias is becoming a treatment of choice using
either heated probe or lasers, such as Nd:YAG and argon.
Argon laser treatment is recommended for mucosal or
superficial lesions because the energy penetrates only 1 mm.
Nd:YAG lasers are more useful for deeper lesions because
they penetrate 3-4 mm.
Hunter et al evaluated 222 GI endoscopic laser procedures in
122 patients. Hemorrhage was arrested in 84% of the
patients with GI bleeding. No perforations were reported in
this series. One death occurred and was attributed to laser
therapy in a patient with duodenal ulcer and gastroduodenal
artery bleeding.
Forty patients with GI arteriovenous malformations
underwent 72 photocoagulation sessions with mostly argon
laser. Of those 40 patients, 15 had significant hemorrhage
from colonic arteriovenous malformations. No deaths
occurred in ablation of GI arteriovenous malformations in 15
patients with colonic lesion.
One of the advantages of upper or lower endoscopic
evaluation is that it provides access to therapy in patients
with GI bleeding. Endoscopic control of bleeding can be
achieved using thermal modalities or sclerosing agents.
Absolute alcohol, morrhuate sodium, and sodium tetradecyl
sulfate can be used for sclerotherapy of upper and lower GI
lesions.
Endoscopic thermal modalities (eg, laser photocoagulation,
electrocoagulation, heater probe) can also be used to arrest
hemorrhage. Endoscopic control of hemorrhage is suitable
for GI polyps and cancers, arteriovenous malformations,
mucosal
lesions,
postpolypectomy
hemorrhage,
endometriosis, and colonic and rectal varices.
The medical literature has also been reviewed for endoscopic
treatment of significant lower GI bleeding. A total of 286
patients were identified in 8 publications. Hemorrhage was
successfully arrested in 70% of patients, with a rebleeding
rate of 15%. Endoscopic therapy for lower GI bleeding is a
minimally invasive and viable option in carefully selected
patients.
Surgical Therapy
An emergency operation is required in approximately 10% of
patients with lower GI bleeding. When the bleeding point is
localized, perform a limited segmental resection of the small
or large bowel. The crude outcome analysis was applied to
483 cumulative cases of limited segmental resection derived
from 23 publications since 1974. The rebleeding rate was 7%
(0-21%), and the mortality rate was 10% (0-15%). A morbidity
rate of 0-33% was reported in only a very few publications;
thus, limited segmental resection is preferred because it can
be performed with low morbidity, mortality, and rebleeding
rates.
If the patient is hemodynamically unstable because ongoing
hemorrhage, perform an emergency operation before any
diagnostic study.
In these cases, make every attempt to diagnose the bleeding
point intraoperatively. Intraoperative EGD, surgeon-guided
enteroscopy, and colonoscopy may be helpful in diagnosing
undiagnosed massive GI bleeding. Depending on the
availability of local resources and the patient's condition, it
may sometimes be better to perform subtotal colectomy
with distal ileal inspection than to try to achieve these other
tests, particularly if the surgeon is not privileged or
comfortable with endoscopy.
If the bleeding point cannot be diagnosed following a
thorough intraoperative endoscopy and examination and if
evidence points to colonic bleeding, perform a subtotal
colectomy with ileorectal anastomosis. Subtotal colectomy is
a rational option because it is associated with a very low
rebleeding rate (3%) and with acceptable average morbidity
(32%) and mortality (19%) rates.
Practitioners must understand that blind segmental resection
should not be performed because of a prohibitively high
rebleeding rate of up to 75%, a morbidity rate up to 83%, and
a mortality rate up to 60%. Once the bleeding point is
identified, a limited segmental resection should be
performed.
Patients who have experienced multiple episodes of lower GI
bleeding without a known source or diagnosis should
undergo elective mesenteric angiography, upper and lower
endoscopy, Meckel scan, upper GI with small bowel series,
and enteroclysis. Elective evaluation of the entire GI tract
may identify uncommon lesions and undiagnosed
arteriovenous malformations.
Rare causes of lower GI bleeding
1. Chronic radiation enteritis/proctitis
2. Ischemic colitis/mesenteric vascular insufficiency
3. Small bowel diverticulosis
4. Meckel diverticulum
5. Colonic/rectal varices
6. Portal colopathy
7. Solitary rectal ulcer syndrome
8. Diversion colitis
9. Dieulafoy lesion of colon
10. Dieulafoy lesion of small bowel
11. Vasculitides
12. Small bowel ulceration
13. Intussusception
14. Endometriosis
15. GI bleeding in runners
If the bleeding point is diagnosed by mesenteric angiography,
vasopressin infusion can be temporarily used to control the
hemorrhage to stabilize the patient in anticipation of
semiurgent segmental bowel resection. Use selective
mesenteric embolization in high-risk patients for whom the
operative management is associated with prohibitive risk of
morbidity and mortality. If mesenteric embolization is used,
these patients must be carefully monitored for bowel
ischemia and perforation. Any evidence of ongoing bowel
ischemia and/or unexplained sepsis following mesenteric
embolization requires exploratory laparotomy to resect the
affected bowel segment. Perform subtotal colectomy with
ileoproctostomy in patients with multiple episodes of
nonlocalized lower GI bleeding or bilateral sources of colonic
hemorrhage.
Preoperative Details
Acute lower GI hemorrhage is a common clinical entity and is
associated with significant morbidity and mortality. Mortality
rates associated with lower GI hemorrhage are reported to
be 10-20% and are dependent on age (>60 y), multiorgan
system disease, transfusion requirements (>5 units), need for
operation, and recent stress (eg, surgery, trauma, sepsis).
Three major aspects are involved in managing lower GI
hemorrhage. The initial priority is to treat the shock. Second,
localization of the source of bleeding is required to perform
the third task—formulating an interventional plan. Insert a
nasogastric tube in all patients. A clear bile-stained aspirate
generally excludes bleeding proximal to the Treitz
ligamentum. After initial resuscitation, undertake a search for
the cause of the bleeding to precisely locate the bleeding
point.
Following accurate localization by angiogram, bleeding can
be temporarily controlled with either angiographic
embolization or vasopressin infusion. Segmental bowel
resection is performed in the next 24-48 hours following
correction of the patient's physiologic parameters, which
include hypotension, hypothermia, acute hemorrhagic
anemia, and deficient coagulation factors.
Intraoperative Details
Surgical intervention is required in only a small percentage of
patients with lower GI hemorrhage. The surgical option
depends on whether the bleeding source has been accurately
identified preoperatively; if so, it is then possible to perform
segmental intestinal resection.
If the bleeding source is unknown, an upper GI endoscopy
should be performed prior to any surgical exploration. At
celiotomy, identifying the bleeding point is often impossible,
as blood refluxes into the proximal and distal bowel. The
abdominal cavity is explored through a midline vertical
incision. The assistance of a gastroenterologist is required for
intraoperative endoscopic evaluation. The colonoscope is
introduced, and the surgeon assists its passage. On-table
colonic lavage and colonoscopy may identify the colonic
source of bleeding. Surgeon-guided intraoperative small
bowel enteroscopy is also performed when no colonic source
of bleeding is identified. Again, the colonoscope can be used
for this procedure.
Unlike colonoscopy, enteroscopy is performed during the
advancement of the scope. Colonoscopic manipulation of the
small bowel may cause iatrogenic mucosal tears and
hematomas, which may be mistakenly identified as a source
of bleeding. Another intraoperative strategy is to clamp
segments of the bowel with noncrushing intestinal clamps to
identify the segment that fills with blood. If the bleeding
point cannot be diagnosed through intraoperative panintestinal endoscopy and examination and if evidence points
to a colonic bleeding, perform a subtotal colectomy with end
ileostomy.
Postoperative Details
Hypotension and shock are the eventual consequences of
blood loss, but this depends on the rate of bleeding and the
patient's response. Clinical development of shock may
precipitate myocardial infarction, cerebrovascular accident,
and renal or hepatic failure. Azotemia occurs in patients with
GI blood loss.
Follow-up
Postoperative office visits every 2 weeks are essential to
ensure proper wound healing. Upon discharge, a general diet
abundant in fruits and vegetables is recommended. Patients
are instructed to drink 6-8 glasses of fluid per day. Psyllium
seed preparations should also be started.
Complications
Patients who have had surgery of the lower GI tract are
prone to the development of complications. The most
common early postoperative complications are intraabdominal or anastomotic bleeding, ileus, mechanical small
bowel obstruction (SBO), intra-abdominal sepsis, localized or
generalized
peritonitis,
wound
infection
and/or
dehiscence, Clostridium difficilecolitis, pneumonia, urinary
retention, urinary tract infection (UTI), deep venous
thrombosis (DVT), and pulmonary embolus (PE).
Intra-abdominal sepsis following colorectal surgery is a lifethreatening complication
and requires aggressive
resuscitation. Systemic conditions (eg, severe blood loss and
shock, poor bowel preparation, irradiation, diabetes,
malnutrition, hypoalbuminemia) may adversely affect
anastomotic healing. Changes in anatomy and physiology of
the large bowel, high bacterial content, improper operative
technique, tension, and ischemia can cause anastomotic leak
associated with abscess and intra-abdominal sepsis. This
condition requires either laparotomy (if the sepsis is
generalized) or percutaneous drainage (if the sepsis is
localized).
Delayed complications usually occur more than a week after
surgery. The most common delayed complications are
anastomotic stricture, incisional hernia, and incontinence.
Outcome and Prognosis
Identification of the bleeding point is the most important
initial step in treatment. Once the bleeding point is localized,
the treatment options are straightforward and curative.
Although diagnostic methods for precisely locating the
bleeding point have greatly improved over the last 3 decades,
10-20% of patients with lower GI bleeding have no
demonstrable bleeding source. Therefore, this complex
problem requires systematic and orderly evaluation to
reduce the percentage of undiagnosed and untreated cases
of lower GI bleeding.
Future and Controversies
The evolution of more sophisticated diagnostic imaging (eg,
angiography, bleeding scan, flexible fiberoptic colonoscope)
offers the promise of precise localization of the bleeding site.
These advances also provide nonoperative and less invasive
control of bleeding using angiographic techniques or
colonoscope. Pharmacologic discoveries are also improving
patient care and outcome. Therefore, the therapeutic
armamentaria have expanded greatly in the last 50
years.Precise localization of the bleeding point is essential for
treatment of lower GI bleeding. Despite the improvement in
diagnostic imaging and procedures, up to 10-20% of the
patients with lower GI bleeding have no demonstrable
bleeding source; therefore, noninvasive diagnostic images
and techniques should be developed to improve patient
outcome.