Professional Documents
Culture Documents
Ugib 1
Ugib 1
Protected by
CLINICAL REVIEW
Upper gastrointestinal bleeding (UGIB) is an un- sive care, advances in diagnosis and treatment, as
copyright.
common but potentially serious and life-threaten- well as in the stabilization and management of
ing clinical condition in children. Anatomically, the critically ill patients. Mortality can be decreased
upper gastrointestinal (GI) tract includes the by early identification of UGIB, and improved
esophagus to the ligament of Treitz; therefore morbidity and mortality are most often the result
UGIB includes bleeding that originates throughout of a multidisciplinary approach to care.3 Among
this region. Common signs and symptoms at pre- children admitted to pediatric intensive care
sentation include hematemesis (73%), melena units (PICUs), those with UGIB require more
(21%), and coffee-ground emesis (6%); however, red blood cell transfusions, have longer hospital
patients may also experience epigastric pain, ab- stays, and have a longer duration of mechanical
dominal tenderness, or dizziness.1–3 ventilation.6
The worldwide mortality rate for UGIB in
children can range from 5% to 21%, which re-
flects the diverse populations that experience Literature Search
conditions associated with UGIB.4,5 In the PubMed was searched in Clinical Queries using the
United States mortality is on the lower end of the key search terms children or infants, gastrointestinal
spectrum as a result of improved pediatric inten- bleeding, and gastrointestinal hemorrhage, pathogene-
sis, diagnosis, and treatment. The search included
meta-analyses, randomized controlled trials, clini-
This article was externally peer reviewed. cal trials, and reviews. We also searched Clinical
Submitted 15 May 2014; revised 9 October 2014; accepted Evidence, the Cochrane database, Essential Evi-
14 October 2014.
From the Department of Family Medicine, Medical Col- dence Plus, the National Center for Biotechnology
lege of Georgia, Georgia Regents University, Augusta. Information at the US National Library of Medi-
Funding: none.
Conflict of interest: none declared. cine, the National Guideline Clearinghouse, and
Corresponding author: Thad Wilkins, MD, Department of DynaMed. Our initial search date was November 5,
Family Medicine, Medical College of Georgia, Georgia Re-
gents University, 1120 15th Street, HB-4032, Augusta, GA 2012, with follow-up searches preformed in De-
30912 (E-mail: jwilkins@gru.edu). cember 2012, May 2013, and February 2014.
copyright.
sion, death, or transfusion occurring within 24 Risk factors for UGIB vary. NSAID use and Heli-
hours after the bleeding episode.9 In this study cobacter pylori infection should be considered in
43% were female and their mean age was 5.3 children with severe UGIB. The observed risk of
years.9 The highest frequency of developing UGIB UGIB with NSAIDs is 7.2 per 100,000, based on a
(69.8%) occurred within the first 72 hours of ad- study of 55,785 children.13 The risk of developing
mission to the PICU.9 UGIB with NSAID use is greater in the 2-month-
old to 7-year-old age group (odds ratio [OR], 14.1)
versus the 8- to 16-year-old age group (OR, 3.4).7
Causes H. pylori has been found in up to 49% (41 of 84) of
The causes of UGIB can be categorized by age children presenting with UGIB.14 This infection
groups. In newborns the predominant causes in- represents an important risk factor, especially in
clude coagulation disorders such as vitamin K de- children with hereditary hemorrhagic disorders
ficiency, cow milk intolerance, gastritis from stress, such as hemophilia15 (Table 1). Other risk factors
sepsis, and trauma from the placement of nasogas- include peptic ulcer disease, portal hypertension or
tric tubes.10 From 1 month to 1 year of age, the varices, and bleeding disorders.3 Children who re-
most prevalent causes are caustic ingestions, dupli- quire mechanical ventilation during the course of
cation cysts, foreign body ingestion, stress esoph- their hospitalization have a higher incidence of
agitis, medication-induced bleeding (eg, nonsteroi- UGIB if they also experience a high pressure ven-
dal anti-inflammatory drug [NSAID] use), and tilator setting (relative risk, 3.73) or organ failure
peptic ulcer bleeding.10 From 1 to 5 years of age, (relative risk, 2.85).20 Other risk factors include
causes include erosive esophagitis, gastritis, caustic trauma (OR, 20.9), shock (OR, 17.4), and operative
ingestions, peptic ulcer bleeding, varices, and vom- procedures ⬎3 hours long (OR, 3.6).5
iting-induced bleeding, for example, from a Mal-
lory-Weiss tear.10 From ages 5 to 18 years, bleed- Associated Conditions
ing can arise from coagulation disorders, gastritis, Multiple disorders can contribute to the develop-
Dieulafoy lesions (an anomalous artery located in ment of an UGIB in pediatric patients. Hemato-
logic disorders, such as hemophilia A and B and the key points are an extensive history and examina-
Von Willebrand disease, predispose patients to tion, laboratory evaluations, and diagnostic proce-
UGIB secondary to the increased risk of bleeding dures.
mucosal membranes.15 Conditions such as biliary
atresia, portal vein thrombosis, primary sclerosing Differential Diagnosis
cholangitis, autoimmune hepatitis, Budd-Chiari The differential diagnosis for UGIB includes
syndrome, and cystic fibrosis predispose patients to causes for lower GI bleeding (LGIB), non-GI
UGIB through the development of portal hyper- sources, ingested maternal blood, or food sources
tension, which can lead to the formation of varices, imitating hematemesis or melena. Causes of LGIB
copyright.
a known cause of UGIB.21 include conditions such as Crohn’s disease, ulcer-
ative colitis, hemorrhoids, anal fissures, and Meckel
Screening and Prevention diverticulum. Non-GI sources of bleeding include
There are no guideline-based screening or preven- epistaxis or hemoptysis that may present with
tion recommendations. However, screening endos- symptoms similar to those of UGIB, such as he-
copy is recommended for patients with conditions matemesis, melena, and positive occult blood tests
that have a high incidence of portal hypertension in (Table 2). Maternal sources include ingestion of
an effort to identify variceal formation as early as blood during delivery or from cracked nipples dur-
possible.22 Endoscopic findings of red markings or ing breastfeeding; infants who ingest maternal
gastric varices are of concern for the presence or blood may present with hematemesis or melena.10
future development of UGIB.3 Children who have
been diagnosed with esophageal varices, red mark- History
ings, or gastric varices on initial screening endos- Children with a history of concurrent major illness
copy should undergo endoscopic sclerotherapy or that require PICU care, such as sepsis and respiratory
band ligation to prevent hemorrhage.22 In addition, failure, may present with stress gastritis or stress ul-
the use of -blockers in at-risk children is recom- cers.3 Indications for possible variceal bleeding in-
mended as prophylaxis for variceal bleeding.23 clude a history of autoimmune hepatitis, Budd-Chiari
Children in a PICU should be considered for pro- syndrome, cystic fibrosis, biliary atresia, portal vein
phylaxis with histamine 2 receptor antagonists or thrombosis, or primary sclerosing cholangitis.22 Fam-
proton pump inhibitors.24 ily history should be assessed for inheritable diseases
that may increase risk for UGIB, for example, liver
Diagnosis disease, any history of requirements for clotting factor
There are no randomized controlled trials, Cochrane replacement, hemophilia A, hemophilia B, and Von
reviews, or systematic reviews of the diagnostic ap- Willebrand disease.15
proach in children with UGIB. The diagnostic ap- Defining the amount of bleeding and any as-
proach is mostly extrapolated from studies of adults; sociated symptoms is important. If the patient is
copyright.
example, slow bleeding may result in a loss of
1–5 Years Peptic ulceration
13% of total blood volume with no change in
Stress gastritis
Medication-induced gastritis (eg, NSAIDs
hemodynamic status.25 Furthermore, a loss of
or aspirin use) palmar crease with wrist hyperextension may in-
Varices dicate a ⬎50% loss of blood volume.25
Epistaxis A comprehensive review of medications
Hemoptysis should be performed, with a special focus on
Mallory-Weiss tear those medications that may increase the risk of
Gastroesophageal reflux
UGIB. These include aspirin, NSAIDs, cortico-
Caustic ingestion
steroids, and selective serotonin reuptake inhib-
Bowel obstruction
itors (SSRIs).3 NSAIDs increase the risk of
Vasculitis
Crohn disease
UGIB by damaging gastric mucosa and promot-
Hemophilia ing tissue friability.26 The greatest risk of devel-
5–18 Years Varices
Peptic ulceration
Coagulation disorders Table 4. Normal Systolic Blood Pressure According to
Immune thrombocytopenic purpura Age8
Chemotherapy Systolic Blood Pressure (mmHg)*
Crohn disease
Age Normal Lower Limit
H. pylori gastritis
Gastroesophageal reflux 0–1 Month ⬎60 50
Mallory-Weiss tear 1–12 Months ⬎80 70
Caustic ingestion 1–10 Years 90 ⫹ (2 ⫻ age in 70 ⫹ (2 ⫻ age in
years) years)
NSAID, nonsteroidal anti-inflammatory drug. ⬎10 Years 110–130 90
Values in Values in
Age Males (%) Females (%)
Values in Values in
Age Males (g/dL) Females (g/dL)
copyright.
adult 109/L 109/L
Platelets:
Values in Values in
Age Males (U/L) Females (U/L)
Continued
Alanine aminotransferase:
Values in Values in
Age Males (U/L) Females (U/L)
Values in Values in
Males Females
Age (mg/dL) (mg/dL)
copyright.
2–4 Years 0.3–0.6 0.2–0.7
5–7 Years 0.2–0.7 0.2–0.8
8–10 Years 0.3–0.8 0.3–0.9
11–12 Years 0.3–0.9 0.3–1
13–17 Years 0.3–1.1 0.3–1.2
⬎17 Years 0.3–1.1 0.5–1.3
Creatinine:
Values in Values in
Males Females
Age (mg/dL) (mg/dL)
Coagulation Venous sample from PT: 11–15 sec Prolonged PT/INR or PTT
studies citrated tube for PT/ INR: 1 may indicate preexisting
INR, PTT PTT: 25–35 sec coagulopathy, liver
dysfunction, or acute illness
such as sepsis or
disseminated intravascular
coagulation.
BUN, blood urea nitrogen; INR, international normalized ratio; PT, prothrombin time; PTT, partial thromboplastin time; UGIB,
upper gastrointestinal bleeding.
oping gastric complications typically occurs costeroid use in neonates has been associated
within 30 days of initiating NSAID therapy.26 with increased mortality from GI hemorrhage.28
SSRIs inhibit platelet aggregation, and the con- One study found high-dose dexamethasone used
current use of NSAIDs and SSRIs further in- in the treatment of ventilator-dependent prema-
creases the risk for UGIB.27 In addition, corti- ture infants was associated with major complica-
Angiography Arterial contrast study No extravascular extravasation Has an overall good diagnostic
of dye rate of 64% but has better
diagnostic accuracy in acute
UGIB (71%) compared with
chronic or recurrent UGIB
(55%).
Apt-Downing test Stool specimen from neonate Negative Important to distinguish
between maternal and
neonatal blood.
Endoscopy Fiber-optic visualization of No bleeding sites noted; no Urgent endoscopy is indicated
esophageal, gastric, and varices for bleeding requiring
duodenal mucosa transfusion or hemodynamic
instability; otherwise
endoscopy can be performed
within the first 24 hours of
admission.
Gastric aspirate Aspirate from nasogastric No blood detected Place nasogastric tube for
tube gastric lavage to improve the
accuracy of endoscopy.
Consider testing gastric
aspirate for occult blood
using Gastrocult (Beckman
Coulter, Inc., Palo Alto, CA).
Stool for occult/frank blood Stool specimen from rectal Negative Alpha guaiaconic acid reacts
(eg, hemoccult) examination with hydrogen peroxide in
the presence of heme and
produces a blue quinone
compound. This denotes a
positive test.
copyright.
UGIB, upper gastrointestinal bleeding.
tions, such as perforated gastric ulcers, duodenal capillary refill (Tables 3 and 4). Tachycardia is the
ulcers, and upper GI hemorrhage.28 That 5-year most sensitive indicator for blood loss in children.8
review found a 66% mortality rate among chil- General presentation should be noted, including
dren who received high-dose dexamethasone and confusion, irritability, and respiratory distress. Ec-
subsequently developed a GI complication.28 chymosis may signal a poorly controlled bleeding
disorder or trauma. Pallor may indicate severe
Physical Examination blood loss but may not be present in an acute
Airway, breathing, and circulation should be as- UGIB. The abdomen should be assessed for guard-
sessed to evaluate hemodynamic stability. Vital ing, epigastric or rebound tenderness, surgical
signs should be monitored for tachycardia, tachy- scars, hepatomegaly, right upper quadrant tender-
pnea, hypotension, orthostatic hypotension, and ness, or other signs or sequelae of chronic liver
†
Strength of recommendation taxonomy (SORT) taken from ref. 47.
UGIB, upper gastrointestinal bleeding.
copyright.
Fluid Lactated Ringers or normal Hemodynamic instability Congestive heart failure Obtain 2 large-bore IV
resuscitation saline: 20 mL/kg boluses lines.
for ⬍5 min, for total of Place a Foley catheter to
80 mL/kg in the first 20 monitor urine output.
minutes; in patients with If poor response, use
cardiac insufficiency, crystalloid solutions;
dose 5–10 mL/kg bolus consider colloid solutions,
such as albumin or
plasma, and place
intraosseous access
immediately.
Proton pump Omeprazole: 1 mg/kg/24 Duodenal or gastric Drug hypersensitivity Children 1–6 years old may
inhibitors hr by mouth in 1 or 2 ulcer; stress gastritis require higher doses
divided doses or IV once Prophylaxis is an because of enhanced drug
daily; reported effective off-label indication clearance.
range: 0.2–3.5 mg/kg/24 PPIs have a longer duration
hours of action than H2
receptor antagonists.
Limited safety and efficacy
information in children.
Duration of therapy is
unknown.
H2 receptor Ranitidine: Duodenal or gastric No absolute Duration of therapy is
antagonist Oral: 2–4 mg/kg BID ulcer; stress gastritis contraindications unknown.
IV or IM: 2–4 mg/kg/day Prophylaxis is an
divided and administered off-label indication
every 6–8 hours
Maximum dose: 50 mg
every 6–8 hours
Vasoactive Octreotide: 1 g/kg IV Variceal bleeding is an No absolute No randomized controlled
copyright.
drug bolus, followed by off-label indication contraindications trials on use in children
infusion 1to 2 g/kg/
hour
-Blockers Propranolol: Portal hypertension and Asthma, atrioventricular A meta-analysis found that
Oral: 0.5–2 mg/kg/day in esophageal varices are block, bradycardia, combining endoscopic
2–4 divided doses, with an off-label indications cardiogenic shock, therapy and -blockers
the goal of reducing sick sinus syndrome reduced overall rebleeding
heart rate to 75% of more than endoscopic
baseline therapy alone or -blocker
use alone in patients with
cirrhosis and bleeding
esophageal varices.
BID, twice a day; H2, histamine 2; IM, intramuscular; IV, intravenous; PPI, proton pump inhibitor.
(eg, aspartate transaminase and alanine aminotrans- gastric aspirate because of the pH of gastric con-
ferase), and type and crossmatch8 (Table 5). In tents.11,12 Undercooked meats and raw fruits or
newborns with suspected UGIB, an Apt test can vegetables may create false-positive hemoccult re-
differentiate neonatal blood from maternal blood.29 sults; therefore a positive hemoccult result warrants
A blood urea nitrogen—to– creatinine ratio ⬎30 investigation.11,12
may be helpful to distinguish UGIB from LGIB Urgent endoscopy, which is performed ⬍12
(specificity, 98%; sensitivity, 69%; likelihood ratio, hours after admission, is indicated for bleeding
⫹34.4 and ⫺0.32, respectively).30 Placing a naso- that requires transfusion or for hemodynamic
gastric tube for gastric lavage can improve the ac- instability; otherwise, endoscopy can be per-
curacy of endoscopy.11 In addition, gastric aspirate formed within the first 24 hours of admission8
can be assessed for occult blood using a Gastroccult (Tables 6 and 7). The reported efficacy of endos-
(Beckman Coulter, Inc., Palo Alto, CA).32 Stool copy for controlling UGIB is approximately
specimens should be obtained to evaluate the pres- 90%.34 There are no randomized controlled tri-
ence of heme using hemoccult testing; however, als or systematic reviews of follow-up endoscopy
hemoccult is not accurate in testing for blood in in children with significant UGIB. Based on
Injection therapy Injection of solutions including Variceal and nonvariceal bleeding Tachycardia, cardiac
hypertonic saline with arrhythmias, hypertension
epinephrine, normal saline
with epinephrine, thrombin
in normal saline, and
ethanol
Thermocoagulation Heater probe; monopolar, Variceal and nonvariceal bleeding Heat-related mucosal injury,
bipolar, and multipolar bleeding, or perforation
coagulators May have delayed
hemorrhage from site of
therapy for up to 4 weeks
Laser photocoagulation Argon and neodymium: Variceal and nonvariceal bleeding Very expensive equipment;
yttrium-aluminum-garnet not widely used outside of
lasers specialized endoscopy
centers
Hemostatic clips Endoscopically placed clips Variceal and nonvariceal bleeding Bleeding and perforation;
that are deployed at the site clips can migrate off site
of the bleed of bleed, although rarely
Endoscopic band ligation Use of elastic bands on Variceal bleeding and Dieulafoy lesions Postprocedural pain,
bleeding lesion ulceration, secondary
hemorrhage
Retrospective study stated
that 27% of patients had
rebleeding after band
ligation and 1% had
esophageal perforation
Adhesive closure with Injection of tissue adhesive Variceal bleeding, especially for gastric Rebleeding, sepsis, arterial
N-butyl-cyanoacrylate varices embolization (rare)
Transjugular intrahepatic Tract created within the liver Biliary atresia, variceal bleeding Limited data and experience
copyright.
portosystemic shunt using radiographic guidance in children
to connect 2 veins
Surgical shunt placement Attachment of autologous or Bleeding is uncontrolled by therapeutic Loss of shunt patency,
synthetic vein to vein endoscopy and angiography repeat procedures
Balloon tamponade Balloon inflated at the site of Uncontrolled UGIB Limited experience in
bleeding children; should not be
used for more than 24
hours
adult studies, repeat endoscopy in children with promise require intravenous access for fluid resus-
life-threatening UGIB should be considered citation and transfusion, as well as cardiopulmonary
within 48 to 72 hours after the initial endos- and urine output monitoring; they may also require
copy.34 In children with severe peptic ulcer intubation and mechanical ventilation for airway
bleeding, follow-up endoscopy may be consid- protection.8 Patients requiring emergent measures
ered within 4 to 6 weeks to assess ulcer healing.8 for survival should be transferred to a PICU (Fig-
ure 1). The pediatric physiology-based score for
Treatment mortality (PRISM) can be used to assess the risk of
There are no randomized controlled trials or sys- mortality by taking clinical parameters and labora-
tematic reviews of the therapeutic approach to chil- tory values into consideration.29 A score ⬎10 (OR,
dren with UGIB. Although many pediatric patients 13.4) has been shown to be clinically significant for
present with UGIB that is not hemodynamically elevated mortality in the pediatric population.36
significant, rapid assessment, stabilization, and re- Blood transfusion is appropriate for unstable pa-
suscitation should precede diagnostic evaluation in tients and those with hemoglobin ⱕ8 g/dL.8 The
unstable children. This includes assessment of the amount of blood transfused should be determined
airway, breathing, and circulation.8 Patients with according to age and weight.8 Children with active
active bleeding that leads to hemodynamic com- bleeding and coagulopathy should be considered
copyright.
26:227–30.
cated when endoscopic therapy is unsuccessful.34
15. Dolatkhah R, Khoshbaten M, Asvadi Kermani I, et
Children with bleeding that is not controlled with al. Upper gastrointestinal bleedings in patients with
endoscopic or angiographic interventions should hereditary coagulation disorders in northwest of
be evaluated for surgery34 (Figure 1). Iran: prevalence of Helicobacter pylori infection.
Eur J Gastroenterol Hepatol 2011;23:1172–7.
16. Chey WD, Wong BC; Practice Parameters Com-
References mittee of the American College of Gastroenterology.
1. Cleveland K, Ahmad N, Bishop P, Nowicki M. Up- American College of Gastroenterology guideline on
per gastrointestinal bleeding in children: an 11-year the management of Helicobacter pylori infection.
retrospective endoscopic investigation. World J Pe- Am J Gastroenterol 2007;102:1808 –25.
diatr 2012;8:123– 8.
17. de Toledo CH. The patient with scoliosis. The de-
2. Houben CH, Chiu PW, Lau JY, et al. Duodenal fect: classification and detection. Am J Nurs 1979;
ulcers dominate acute upper gastrointestinal tract 79:1588 –91.
bleeding in childhood: a 10-year experience from 18. Di Rienzo TA, D’Angelo G, Ojetti V, et al. 13C-
Hong Kong. J Dig Dis 2008;9:199 –203. Urea breath test for the diagnosis of Helicobacter
3. Boyle JT. Gastrointestinal bleeding in infants and pylori infection. Eur Rev Med Pharmacol Sci 2013;
children. Pediatr Rev 2008;29:39 –52. 17(Suppl 2):51– 8.
4. Cox K, Ament ME. Upper gastrointestinal bleed- 19. Syam AF, Rani AA, Abdullah M, et al. Accuracy of
ing in children and adolescents. Pediatrics 1979; Helicobacter pylori stool antigen for the detection of
63:408 –13. Helicobacter pylori infection in dyspeptic patients.
5. Cochran EB, Phelps SJ, Tolley EA, Stidham GL. World J Gastroenterol 2005;11:386 – 8.
Prevalence of, and risk factors for, upper gastroin- 20. Deerojanawong J, Peongsujarit D, Vivatvakin B,
testinal tract bleeding in critically ill pediatric pa- Prapphal N. Incidence and risk factors of upper
tients. Crit Care Med 1992;20:1519 –23. gastrointestinal bleeding in mechanically ventilated
6. Gauvin F, Dugas MA, Chaibou M, Morneau S, children. Pediatr Crit Care Med 2009;10:91–5.
Lebel D, Lacroix J. The impact of clinically signifi- 21. Goncalves ME, Cardoso SR, Maksoud JG. Prophy-
cant upper gastrointestinal bleeding acquired in a lactic sclerotherapy in children with esophageal var-
pediatric intensive care unit. Pediatr Crit Care Med ices: long-term results of a controlled prospective
2001;2:294 – 8. randomized trial. J Pediatr Surg 2000;35:401–5.
copyright.
1991;48:346 –7. bleed: a randomised controlled trial. Gut 2010;59:
30. Urashima M, Toyoda S, Nakano T, et al. BUN/Cr 729 –35.
ratio as an index of gastrointestinal bleeding mass in 43. Ljubicic N, Biscanin A, Nikolic M, et al. A random-
children. J Pediatr Gastroenterol Nutr 1992;15:89 –92. ized-controlled trial of endoscopic treatment of
31. Rosenthal P, Thompson J, Singh M. Detection of acute esophageal variceal hemorrhage: N-butyl-2-
occult blood in gastric juice. J Clin Gastroenterol cyanoacrylate injection vs. variceal ligation. Hepato-
1984;6:119 –21. gastroenterology 2011;58:438 – 43.
44. Huppert PE, Goffette P, Astfalk W, et al. Tran-
32. Byers SE, Chudnofsky CR, Sorondo B, Dominici P,
sjugular intrahepatic portosystemic shunts in chil-
Parrillo SJ. Incidence of occult upper gastrointestinal
dren with biliary atresia. Cardiovasc Intervent Radiol
bleeding in patients presenting to the ED with he-
2002;25:484 –93.
matochezia. Am J Emerg Med 2007;25:340 – 4.
45. Krebs-Schmitt D, Briem-Richter A, Grabhorn E, et
33. Meyerovitz MF, Fellows KE. Angiography in gas-
al. Effectiveness of Rex shunt in children with portal
trointestinal bleeding in children. AJR Am J Roent-
hypertension following liver transplantation or with
genol 1984;143:837– 40.
primary portal hypertension. Pediatr Transplant
34. Fox VL. Gastrointestinal bleeding in infancy and 2009;13:540 – 4.
childhood. Gastroenterol Clin North Am 2000;29: 46. Edwards MJ, Kollenberg SJ, Brandt ML, et al. Sur-
37– 66, v. gery for peptic ulcer disease in children in the post-
35. Pollack MM, Patel KM, Ruttimann UE. PRISM III: histamine2-blocker era. J Pediatr Surg 2005;40:
an updated Pediatric Risk of Mortality score. Crit 850 – 4.
Care Med 1996;24:743–52. 47. Ebell MH, Siwek J, Weiss BD, et al. Strength of
36. Chaibou M, Tucci M, Dugas MA, Farrell CA, recommendation taxonomy (SORT): a patient-cen-
Proulx F, Lacroix J. Clinically significant upper gas- tered approach to grading evidence in the medical
trointestinal bleeding acquired in a pediatric inten- literature. Am Fam Physician 2004;69:548 –56.