Acute Lower Gastrointestinal

Bleeding
Jonathan P. Terdiman, M.D.
University of California, San
Francisco

Lower GI Bleeding
Epidemiology, Etiology and Outcomes
Presentation and Diagnosis
Therapy
Management strategy by clinical scenario
Putting it all together
 Epidemiology and outcome
Annual incidence is 20/100, 000
1/10 to 1/3 of all acute bleeds requiring hospital stay
Disease of the elderly
200 fold increase from the 3rd to 9th decades of life
Comorbid medical conditions are common
NSAID use is common
> 50%

Strate L; B&W Strate L; B&W Schmulewitz N;

1996-99 2001-2003 Duke, 1993-2000

Retrospective Prospective Retrospective

Number 275 252 415
Mean age 70 66 67
Cause
Diverticulosis 41% 30% 41%
Rectal ulcers: stercoral, 8% 9% 6%
solitary ulcer
Postpolypectomy 6% 7% 2%
AVM 1% 3% 3%
Hemorrhoids 11% 11% 13%
Ischemic colitis 11% 10% 8%
Other colitis (IBD, 12% 15% 7%
infectious, radiation)
Neoplasm 3% 6% 7%
No source found 7% 9% 11%
 Epidemiology and outcome
Outcome depends on etiology and
comorbidities
> 80% of bleeding will stop spontaneously and
not recur
5-10% will have persistent or severe bleed
Mortality is < 5%

Outcomes
Author/ N Continued or Died Surg pRBCs LOS
year rebleeding Tx (SD) (Days)

Strate 275 4.0% 2.6% 2.5

2005 (4.5)

Strate 252 7% 2.4% 3.6% 2.0 4.3

2003 (3.0)

Schmule- 565 11% 3% 5% 3.1 6.7

witz 2003 (3.9)

Das 332 19% 5% 2.2 4.4

2003
 Diverticular Bleeding
30

25

20
Rebleed
15
Death
10

0
1 2 3 4
Years after discharge

Intractable bleed in hospital = 7%

Longstreth Am J Gastro
In hospital mortality = 2% 1997;92:419

Presentation
Vital signs
20-30% with shock or orthostasis
Form of bleeding
Hematochezia versus melena
Abdominal pain
Present versus absent, location
Directed history and Exam
Comorbid conditions
Labs
 Risks for Ongoing Bleeding

Strate et al. Arch Int Med, 2003

HR > 100 OR, 3.67
Sys BP < 115 OR, 3.45
Syncope OR, 2.82
Painless OR, 2.43
Overt bleed (4 hr) OR, 2.32
ASA use OR, 2.07
2 active comorbid OR, 1.93

Risks for Severe Bleed

Severe bleeding (ongoing bleed and/or > 2 units
transfusion) occurs in:
79-84% with > 3 risk factors 17% of total
43% with 1-3 risk factors 78% of total
6-9% with 0 risk factors 5% of total

Strate et al. Am J Gastro, 2005

 Outcomes Based on Risk
Low Risk = 0 factors
Surgery = 0%
Death = 0%
LOS = 2.8 days
Moderate Risk = 1-3 risk factors
Surgery = 1.5%
Death = 2.9%
LOS = 3.1 days
High Risk = > 3 risk factors
Surgery = 7.7%
Death = 9.6%
LOS = 4.6 days

Risk of Death
Strate, Clin Gastro Hepatol, 2008
Nationwide audit in US
Mortality = 3.9%
Risk factors
Age > 70 OR = 4.9
Int ischemia OR = 3.5
>/= 2 comorbid OR = 3.0
Nosocomial bleed OR = 2.4
Coagulopathy OR = 2.3
Hypovolemia OR = 2.2
Transfusion OR = 1.6
Men OR = 1.5
 Hospitalization
Abnormal vital signs
Ongoing rectal bleeding
Active/multiple comorbid conditions
Suspicion of upper tract bleed
Previous aortic surgery
Severe anemia (HgB < 8)
Fever, leukocytosis
Abdominal pain/tenderness

Triage/LOS
Clinical criteria
High, Moderate, Low
HIGH Risk: shock or > 3-4 units blood/day
Endoscopic or Angiographic Criteria?
High Risk
Active arterial bleed , vessel, (clot ?) from TIC,
ulcer
Cancer
Lower Risk
Polyp/polypectomy, ectasia, colitis, anorectal
 Triage and Optimal Length of
Stay
Data are scarce compared with upper GI
bleed
Expert opinion
High risk
ICU for 24 hours, hospital for 72 hours
Moderate Risk
Hospital for 24-48 hours, early refeeding
Low Risk
Feed and early discharge

Critical Initial Diagnostic Steps

Upper versus lower tract bleed
Color of bleed
NG aspirate
History
Labs
EGD
Anorectal versus other lower source of
bleed
History
Bedside anoscopy
 Nasogastric Aspirate
> 90% of those with red, pink or black
aspirates have upper GI source
> 60% of those with negative (bilious)
aspirate have lower source, < 1% with
upper source
Equivocal aspirate?
10% or more of upper tract bleeds (DU)

Observe: no bleed,
colonoscopy
w/in 1-2 days

Nuclear bleeding scan;

Rapid purge If neg, colonoscopy
colonoscopy If positive, angio

?
Sigmoidoscopy Angiography
 One Division:
Parallel Practices
Observe; prep Scintigraphy: Angiography
non-emergent Angiography vs.
colonoscopy Elective colonoscopy

None Recurrent/ Continuous Severe/

Intermittent Rapid

Rapid purge;
Rapid-purge; Angiography
non-emergent
Urgent colonoscopy
colonoscopy

Nuclear scintigraphy

Two purposes:
Screening prior to angiography
Increase likelihood of positive angio
Localization for surgery
Assessing accuracy in clinical studies
Variable techniques
Variable thresholds for performing study
Variable times to angiography or surgery
Variable criteria for determining accurate
localization
 99mTc RBC for LGIB: Recent Studies
Author Total Positive scans Correct Positive
Year Scans localization angiograms
Olds 127 39% 48% 42%
2005
Levy 40* 70% 45% 0%
2003
Ng 160 54% - 43%
1997
Suzman 224 51% 78% 44%
1996
Rantis 80 48% 73% -
1995
Voeller 59 32% 69% -
1991
Hunter 203 26% 41% 44%
1990
*

RBC Scintigraphy
Details matter
O.1 ml/min = 1 unit rbc/2-4 hours
Summary of 14 studies: 78% accurate versus
22% inaccurate
Active bleed at time of scan
Technetium Tc 99m-labeled in vitro
Early positive (2 hours) versus late positive
Upper tract source excluded
 Angiography
Diagnosis
Femoral access
5 Fr catheters with steerable wires
Selective access of SMA, IMA catheterization
(sometimes celiac)
Endoscopic identification/marking of bleeding
lesion with clips facilitates
Endovascular therapy
Vasopressin infusion no longer used
Sub-3 Fr catheter placed to most peripheral arteries
Microcoils (1-2 mm) for colon
Polyvinyl microspheres (350-500 um) for small
intestine

Angiography: UCSF Experience

17 patients with angiographically detected
lower tract bleeding
Subselective embolization possible in 14
Tracker 2.5 Fr coaxial microcatheter
metallic coils for embolization
Durable hemostasis in 13/14
Bowel infarction or other major procedure
related morbidity in 0
Am J Surg 1997;174:24-28
 Meta-analysis of Angiography
for LGIB
Khanna A et al: J Gastrointest Surg 2005;9:343

Included:
7 cases series; all with > 10 pts with major
LGIB txed with attempted embolization
Results:
Median 30 d rebleeding rate: 14% (0-75)
Rebleed w/ Non-diverticular source: 45%
(OR 3.4 vs diverticular bleeding)
75 % rebleed w/in 3.5 days

Urgent Colonoscopy
Colonoscopy w/in 6-24 hours of admission
Rapid purge: Get serious!
Polyethylene glycol-based preps
1 Liter q 30-45 minutes
Median 6 L (range: 4-14L)
Time required: 3-4 hrs
NG tube: required in one-third
Consider: metoclopramide 10 mgIV
Goal: clear effluent (if not, give more)
Colonoscopy w/in 1 hr of clearance
If ongoing bleeding, colonoscopy when effluent
is pink with no clots
 Urgent (W/in 24 h)
Colonoscopy in LGIB
Study (year) N Specific Dx Endoscopic Complications
Tx
Green, 2005 50 48 17 2%
Angtuaco, 2001 39 29 4 -
Kok, 1998 190 148 10 0%
Chaudhry, 1998 85 82 17 1%

TOTAL 364 307 (84%) 48 (13%)

Urgent Colonoscopy: UCLA

Experience
Urgent colonoscopy after rapid purge
diagnostic yield
80%; endoscopic
treatment in 40%
complications in 0%
Retrospective Results
angio rate from 50 to < 5%
BE rate from 25 to 0%
surgery rate from 20 to < 5%
LOS from 10 to 5 days and ICU stay from 3 to
1 day
Cost reduced $10, 000 per patient
 Bleeding diverticula (n=3) Rxd with Gold Probe (10-
15W, 1 sec pulses X 6-18 pulses)

Gold probe applied

Flattened VV

VV at edge of tic

Savides et al. GIE 1994;40:70-72

Colonoscopy and Severe Diverticular

Bleed: UCLA Experience
Study 1 - 73 patients (medical/surgical)
Study 2 - 48 patients (medical/colonoscopy)
Definite TIC bleed: 17/73 versus 10/48
Study 2: severe hematochezia = 150
Outcomes
Study 1 - 9/17 with ongoing bleed, 6/17 to OR
Study 2 - 0/10 with ongoing bleed
NEJM 2000;342:78-82
 Urgent Colonoscopy?
Green, Rockey et al., Am J Gastro, 2005

RCT of urgent colonoscopy versus standard

care with angio for ongoing bleed
Urgent colonoscopy in 50
Endo Rx in 17
Standard care in 50
Angio Rx in 10

Results
Urgent Colonoscopy Standard Care Statistics
Definite bleed source 42 % 21 % OR 2.6
(CI, 1.1-6.2)
Presumptive bleed 26 % 20 % OR 1.6
source (CI, 1.1-6.2)
No diagnosis 4% 24% P < 0.05
Hospital stay (days) 5.8 6.6 NS
PRBCs 4.2 (0.4) 5.0 (0.5) NS
Surgery 14% 12% NS
Early rebleed 22 % 30 % NS
Late rebleed 16% 14% NS
 Early Colonoscopy
Strate et al. (Am J Gastroenterol, 2003; GIE, 2005)
252 patients admitted with LGIB
No benefit with respect to need for surgery,
death
Colonoscopy within 24 hours associated with
less transfusion and shorter LOS (hazards ratio,
2.02; 1.5-2.6)
< 24 hours = 2.1 days
24-48 hours = 2.7 days
> 48 hours = 4.4 days

Urgent Surgery
Segmental resection after localization of
bleed
Complication rate < 10%
Rebleeding at 1 year < 15%
Blind segmental resection
Rebleeding > 40% during hospital stay
Emergency total colectomy
Rebleeding > 25%
Mortality > 25%
 Case #1
85 year old woman
multiple medical problems
hematochezia and tachycardia
vital signs normalize with IV fluid and NG
lavage is bilious
initial Hct is 28%
no further hematochezia is passed in the ED.

Question #1: What test to order?

1) EGD and Flex Sig
2) Colonoscopy
3) RBC scan
4) Angiography
 Case #1
While being prepared for colonoscopy the
patient passes more BRBPR and her BP
drops. Her vital signs normalize with an
increase in her transfusion rate.

Question #2
RBC scan is (+) for activity at the splenic
flexure of the colon. Now what
intervention?
1) Colonoscopy
2) Angiogram and embolization of site if active
bleeding seen
3) Angiogram as prelude to surgery if bleed
localized
4) Surgery now
 Case #2
27 year old man
hematochezia, normal BP and tachycardia
HR remains elevated despite IV fluids
no further hematochezia in the ED
NG lavage is clear and initial Hct is 31%.

Question #3
What test (s) should be undertaken first?
1) EGD + Flex Sig
2) Colonoscopy
2) RBC scan
4) Angiography
 Case #2
You perform and urgent EGD and flex
sigmoidoscopy.
results are negative
no further bleeding over the next 12 hours
Colonoscopy is negative
after the colonoscopy more BRBPR with
tachycardia and drop in hematocrit
NG lavage is bilious again

Question #4
What should you do now?
1) Repeat colonoscopy
2) RBC scan
3) Angiography
4) Enteroscopy
5) Capsule endoscopy
 Case #2
RBC scan is positive in ileum
Angiogram is performed and active
bleeding seen and vessel embolized
After angiogram, CT enterography
demonstrates mass in ileum
Elective operative resection reveals GIST

Case #3
64 year old man
hematochezia, tachycardia, bilious NG
lavage and Hct of 28%
two further episodes of hematochezia in
the ED
 Case #4
While being prepared for colonoscopy an
rbc scan is obtained and is negative
Colonoscopy
diverticula throughout the colon, L >> R
No stigmata of ongoing or recent
bleeding are seen, no therapy given

Case #4
While being observed in the hospital
several more discrete episodes of
hematochezia with change in VS
Rbc scans obtained again, positive in
LLQ
Angiogram, negative for active bleed
Patient has received a total of 8 units of
blood
 Question #5
What intervention now?
1) Left hemicolectomy?
2) Total abdominal colectomy?
3) Repeat Angiography?
4) Repeat colonoscopy?

Hematochezia

Any of following?
Pulse > 100/min
BP < 100 mmHg
No Yes
BRBPR w/in 4hrs

Admit to Floor
Observe
If no further bleeding
Elective colonoscopy;
If recurrent BRBPR,
Initiate rapid purge
NG lavage
Consider EGD in high risk groups
Admit to TCU/ICU
Initiate rapid purge
Colonoscopy w/in 1 hr of prep
If massive bleeding/unable
to clear,
Angiography (no RBC scan)
Surgery consult