sign in sign up
<<
Home , Capsule endoscopy, Fecal occult blood

Evaluation of Occult Gastrointestinal Bleeding KATHY BULL-HENRY, MD, and FIRAS H. AL-KAWAS, MD, Georgetown University Hospital, Washington, District of Columbia

Occult gastrointestinal bleeding is defined as gastrointestinal bleeding that is not visible to the patient or physician, resulting in either a positive fecal occult test, or with or without a positive test. A stepwise evaluation will identify the cause of bleeding in the majority of patients. Esophagogastroduode- noscopy (EGD) and will find the bleeding source in 48 to 71 percent of patients. In patients with recur- rent bleeding, repeat EGD and colonoscopy may find missed lesions in 35 percent of those who had negative initial findings. If a cause is not found after EGD and colonoscopy have been performed, capsule has a diagnostic yield of 61 to 74 percent. Deep reaches into the mid and distal small bowel to further investigate and treat lesions found during or computed tomographic enterography. Evaluation of a patient who has a positive fecal occult blood test without iron deficiency anemia should begin with colonoscopy; asymptomatic patients whose colonoscopic findings are negative do not require further study unless anemia develops. All men and postmenopausal women with iron deficiency anemia, and premenopausal women who have iron deficiency anemia that cannot be explained by heavy menses, should be evaluated for occult gastrointestinal bleeding. Physicians should not attribute a positive fecal occult blood test to low-dose aspirin or anticoagulant medications without further evalu- ation. (Am Fam Physician. 2013;87(6):430-436. Copyright © 2013 American Academy of Family Physicians.)

astrointestinal (GI) bleeding may In a review of five prospective stud- be classified as overt, obscure, or ies of upper endoscopy and colonoscopy occult. Overt GI bleeding is vis- in patients with occult GI bleeding, 20 to ible, such as (bloody 30 percent of patients had a colorectal source, G or coffee-ground emesis), (the whereas 29 to 56 percent had an upper GI presence of blood and blood clots in the ), tract source. No source was found in 29 to or (black tarry stools). Obscure GI 52 percent of patients. Synchronous lesions, bleeding refers to recurrent bleeding in which or simultaneous sources of occult bleeding in a source is not identified on upper endos- the upper GI tract and the colon, were found copy, colonoscopy, or small bowel radiogra- in 1 to 17 percent of patients.4 Colonic lesions phy. Obscure bleeding may be either overt or include colon , colon polyps, vascular occult, with the source of bleeding often found ectasias, and colitis. Causes of occult bleed- in the small bowel. Occult bleeding is not vis- ing in the upper GI tract include esophagi- ible to the patient or physician. This review tis, Cameron ulcers (a linear erosion in a focuses on the causes and diagnostic investi- hiatal ), gastric and duodenal ulcers, gation of occult GI bleeding, manifested as a vascular ectasias, gastric cancer, and gastric positive fecal occult blood test (FOBT) or iron antral vascular ectasia (Tables 14 and 25). deficiency anemia with or without a positive A small bowel source is likely in a high per- FOBT. centage of patients with recurrent bleeding and negative findings on esophagogastrodu- Etiology odenoscopy (EGD) and colonoscopy. Occult GI bleeding may occur anywhere in In patients younger than 40 years, small the GI tract, from the oral cavity to the ano- bowel tumors are the most common cause .1,2 A review of multiple studies has of occult GI bleeding; other causes include shown that lesions in the upper GI tract and celiac disease6 and Crohn disease.5 In small bowel are often the cause of iron defi- patients older than 40 years, vascular ecta- ciency anemia.3 sias and nonsteroidal anti-inflammatory

430Downloaded American from Family the American Physician Family Physician Web site at www.aafp.org/afp.www.aafp.org/afp Copyright © 2013 American AcademyVolume of Family87, Number Physicians. 6 ◆ For March the private, 15, 2013 non- commercial use of one individual user of the Web site. All other rights reserved. Contact [email protected] for copyright questions and/or permission requests. Occult GI Bleeding SORT: KEY RECOMMENDATIONS FOR PRACTICE

Evidence Clinical recommendation rating References

Patients who have a positive FOBT without iron deficiency anemia should be evaluated with C 4, 8, 10, 24 colonoscopy. If colonoscopy is negative, asymptomatic patients do not require further studies unless anemia develops. A positive FOBT should not be attributed to low-dose aspirin or anticoagulation without further C 4, 8, 10, 24 GI evaluation. Premenopausal women who have iron deficiency anemia that cannot be explained by heavy C 4, 8, 10 menses, or those who have GI symptoms, should be evaluated for a GI cause. Men and postmenopausal women with iron deficiency anemia should undergo progressive C 4, 8, 10 evaluation for GI blood loss with colonoscopy, esophagogastroduodenoscopy, and capsule endoscopy, as clinically indicated.

FOBT = fecal occult blood test; GI = gastrointestinal. A = consistent, good-quality patient-oriented evidence; B = inconsistent or limited-quality patient-oriented evidence; C = consensus, disease- oriented evidence, usual practice, expert opinion, or case series. For information about the SORT evidence rating system, go to http://www.aafp. org/afpsort.xml.

Table 1. Prospective Studies of EGD Table 2. Potential Causes of Occult and Colonoscopy in Patients with Gastrointestinal Bleeding Occult GI Bleeding Mass lesions Vascular Findings on EGD and Carcinoma (any Gastric antral vascular colonoscopy Frequency (%) site)* ectasia Large Hemangioma Colorectal source 20 to 30 (> 1.5 cm) Portal hypertensive 1 to 9 Inflammation gastropathy or Colitis 1 to 2 Cameron ulcers colopathy Colon cancer 5 to 11 (linear erosions Vascular ectasias Polyps () 5 to 14 within a hiatal (any site)* Upper source 29 to 56 hernia) Infection Celiac disease Angiodysplasia 1 to 8 Amebiasis Colitis (nonspecific) Celiac disease 0 to 6 Ascariasis Crohn disease Hookworm Duodenal ulcer 1 to 11 Erosive Strongyloidiasis Esophagitis 6 to 18 esophagitis* Tuberculous Gastric cancer 1 to 4 Erosive gastritis enterocolitis Gastric ulcer 4 to 6 Idiopathic cecal Whipworm Gastritis 3 to 16 ulcer Other Synchronous lesions (lesions 1 to 17 Ulcer (any site)* Munchausen syndrome found in both upper GI Ulcerative colitis Long-distance running tract and colon)

No source found 29 to 52 *—Most common. Adapted with permission from Rockey DC. Occult and EGD = esophagogastroduodenoscopy; GI = gastro- obscure gastrointestinal bleeding: causes and clini- intestinal. cal management. Macmillan Publishers Ltd.: Nat Rev Information from reference 4. Gastroenterol Hepatol. 2010;7(5):270.

drug–induced ulcers are the most com- (e.g., hookworm) and long-distance run- mon causes (Figures 1 through 5). However, ning. Researchers hypothesize that long- celiac disease remains a possible cause in distance running induces GI blood loss symptomatic7 and asymptomatic8 adults because of transient intestinal ischemia older than 50 years. Less common causes from decreased splanchnic perfusion during of occult GI bleeding include infections exercise.9

March 15, 2013 ◆ Volume 87, Number 6 www.aafp.org/afp American Family Physician 431 Figure 1. Mid ileal mass (arrow) seen during single-balloon enteroscopy.

Figure 4. Small bowel vascular ectasia (arrow) seen during capsule endoscopy.

Figure 2. Example of colon cancer (arrow) seen during colonoscopy.

Figure 5. Example of gastric antral vascular ectasia (arrows) seen during eso­pha­go­gas­tro­duode­no­scopy.

anti-inflammatory drug use suggests ulcer- ative mucosal injury. Anticoagulants or antiplatelet medications may precipitate bleeding in an undiagnosed lesion. A family history of GI bleeding may suggest heredi- tary hemorrhagic telangiectasia (associated Figure 3. Small bowel ulcers (arrows) seen with vascular lesions on the lips, tongue, or during capsule endoscopy. palms) or blue rubber bleb nevus syndrome (a syndrome with venous malformations History and in the GI tract, soft tissues, and skin). A A targeted history and physical examina- history of may sug- tion should be performed. A history of GI gest impaired iron absorption.5 A history bleeding, surgery, or pathology may reveal of disease or stigmata of liver disease important diagnostic clues. Unintentional suggests portal hypertensive gastropathy or weight loss suggests a . Abdomi- colopathy. Other helpful physical exami- nal pain with aspirin or other nonsteroidal nation findings that could indicate the

432 American Family Physician www.aafp.org/afp Volume 87, Number 6 ◆ March 15, 2013 Occult GI Bleeding

presence of an underlying condition include GI bleeding.11 In a meta-analysis of 14 stud- dermatitis herpetiformis (celiac disease); ies, the diagnostic yield of capsule endos- erythema nodosum (painful erythema- copy was superior to push enteroscopy tous nodules seen in Crohn disease); an (63 versus 28 percent) and barium stud- atrophic tongue and brittle, spoon-shaped ies (42 versus 6 percent).12 The nails (Plummer-Vinson syndrome); hyper- sensitivity, specificity, positive In patients younger than extensible joints and ocular and dental predictive value, and nega- 40 years, small bowel abnormalities (Ehlers-Danlos syndrome); tive predictive value of capsule tumors are the most and freckles on the lips and in the endoscopy in obscure GI bleed- (Peutz-Jeghers syndrome).4 ing were 95, 75, 95, and 86 per- common cause of occult cent, respectively, without the gastrointestinal bleeding. Diagnostic Studies high rates of morbidity and Multiple diagnostic procedures are avail- mortality associated with intraoperative able to investigate the GI tract in patients enteroscopy.13 Complications related to cap- with occult bleeding. The choice and sule endoscopy are rare and include capsule sequence of procedures will depend on retention (occurring in less than 1 percent of clinical suspicion and any associated symp- patients who do not have Crohn disease).14 toms. Upper GI bleeding (identified as the Two wireless capsule endoscopy systems source of bleeding proximal to the ampulla are available in the United States: the Endo of Vater) can be detected by EGD. Proxi- Capsule (Olympus America, Inc., Center mal small bowel bleeding can be detected Valley, Pa.) and the PillCam SB II (Given with push enteroscopy, which reaches the Imaging Ltd., Duluth, Ga.). Both capsules proximal jejunum. Bleeding of the mid measure 11 mm × 26 mm and capture two and distal small bowel can be detected with images per second. The diagnostic yields of capsule endoscopy, deep enteroscopy, and each system are similar, ranging from 55 to computed tomographic (CT) enterogra- 74 percent.15 Limitations of capsule endos- phy. Lower GI bleeding (colonic bleeding) copy include lack of therapeutic capability can be detected with colonoscopy.10 Intra- and poor visualization in the periampullary operative enteroscopy remains an option region. Videos depicting capsule endoscopy for those rare patients who have recurrent may be viewed at: bleeding from a source not yet identified • http://www.youtube.com/watch?v= with the previously mentioned methods. YqE98avDCt8 Small bowel barium studies have a very low • http://www.youtube.com/watch?v= yield and have been largely replaced by cap- DDza5tt7lX8 sule endoscopy. • http://www.youtube.com/watch?v= EGD and colonoscopy will find the bleed- Y5aHGfDFZh0 ing source in 48 to 71 percent of patients.4 In patients with recurrent bleeding, repeat EGD DEEP ENTEROSCOPY and colonoscopy may find missed lesions in Push enteroscopy using longer endoscopes 35 percent of those who had negative initial called enteroscopes can evaluate the GI findings.4 If a cause is not found after EGD tract to the proximal jejunum. However, and colonoscopy have been performed, cap- with the availability of deep enteroscopy, sule endoscopy has a diagnostic yield of 63 to which typically can reach the mid and dis- 74 percent.10 tal small bowel, the use of push enteroscopy has diminished significantly. Currently, CAPSULE ENDOSCOPY there are three deep enteroscopy systems Wireless capsule endoscopy is a noninva- approved in the United States: the Double sive method to evaluate the entire length Balloon Endoscopy System (Fujifilm Medi- of the small bowel. Capsule endoscopy can cal Systems, Stamford, Conn.), the Single identify vascular ectasias, ulcers, and masses Balloon Enteroscope System (Olympus in the small bowel in patients with occult America, Inc., Center Valley, Pa.), and the

March 15, 2013 ◆ Volume 87, Number 6 www.aafp.org/afp American Family Physician 433 Evaluation of the Patient with a Positive FOBT

Positive FOBT without iron deficiency anemia No findings were reported in 0 to 57 percent of patients. The diagnostic yield of double- balloon enteroscopy ranged from 41 to Upper GI tract Upper GI tract 80 percent, with therapeutic success vary- symptoms not present symptoms present ing from 43 to 76 percent.10 Perforation rates range from 0.3 to 3.6 percent.10,16-18 Stud- ies comparing double-balloon enteroscopy Colonoscopy EGD and colonoscopy with single-balloon enteroscopy 19 and spiral enteroscopy 20 are lacking.

OTHER STUDIES Source No source No source Source found found found found Capsule endoscopy and deep enteroscopy have higher overall diagnostic yield com-

Treat as No further evaluation Treat as pared with small bowel barium studies and indicated unless anemia develops indicated CT scans, most likely because vascular ecta- sias are a common cause. However, advances Figure 6. Evaluation of the patient with a in CT techniques using multiphasic CT scans positive FOBT. (EGD = esophagogastroduo- with neutral oral contrast media (CT enterog- denoscopy; FOBT = fecal occult blood test; raphy) are more specific for small bowel GI = gastrointestinal.) pathology and may be useful in evaluating the Information from reference 10. small bowel for tumors and bleeding.21,22 Use of this technology may be limited to refer- Endo-Ease Discovery SB small bowel entero- ral to an imaging center based on the cen- scope or spiral enteroscope (Gyrus ACMI, ter’s expertise and availability. Nuclear scans Inc., Southborough, Mass.). Double-balloon and angiography are best suited for overt GI enteroscopy uses two balloons (one on the bleeding. The high diagnostic yield, therapeu- overtube and one at the end of the entero- tic capability, and safety of deep enteroscopy scope). Single-balloon enteroscopy uses one have decreased the need for intraoperative balloon at the end of the enteroscope. Spi- enteroscopy and barium studies. ral enteroscopy uses an overtube with raised helices over an enteroscope, which is rotated Evaluation clockwise into the small bowel. All three sys- POSITIVE FOBT WITHOUT tems permit therapeutic interventions such IRON DEFICIENCY ANEMIA as coagulation, polypectomy, or presurgical The American Gastroenterological Asso- tattooing of lesions. ciation recommends a stepwise approach to Deep enteroscopy may be performed via the evaluation of patients who have a posi- the oral (antegrade) or anal (retrograde) tive FOBT without iron deficiency anemia route. Lesions noted in the proximal two- (Figure 6).10 The evaluation begins with thirds of the small bowel transit time on colon­oscopy. Colonoscopy is preferred capsule endoscopy generally may be reached because of its high sensitivity for mass lesions orally. Lesions noted in the distal one-third and mucosal lesions, and its intervention of the small bowel transit time on capsule capabilities with biopsy, polypectomy, and endoscopy are approached anally. treatment of bleeding lesions. Barium stud- In eight studies of double-balloon enter- ies have lower sensitivity for mucosal lesions oscopy in patients with obscure GI bleed- and are generally not recommended. The ing, the most common findings in the cost-effectiveness of routinely evaluating small bowel were vascular ectasias (6 to the upper GI tract in the setting of a nega- 55 percent), ulcerations (3 to 35 percent), tive colonoscopy is unclear. If colonoscopy and (3 to 26 percent).10 Other is negative, further studies are not required conditions such as small bowel diverticula in the asymptomatic patient unless anemia were present in 2 to 22 percent of patients. develops.4,8,10,23,24 For patients with upper

434 American Family Physician www.aafp.org/afp Volume 87, Number 6 ◆ March 15, 2013 Occult GI Bleeding

GI tract symptoms, evaluation with EGD should be performed at the time of colon­ Diagnosis of Iron Deficiency Anemia with or oscopy. Health care professionals should not Without a Positive FOBT attribute a positive FOBT to low-dose aspi- rin or anticoagulation without further eval- Iron deficiency anemia with or without a positive FOBT uation.4,8,10,24 In one prospective study, 15 of 16 patients receiving anticoagulants with a positive FOBT had significant lesions, and EGD and colonoscopy 20 percent of the lesions were malignant.24

IRON DEFICIENCY ANEMIA WITH OR WITHOUT A POSITIVE FOBT Positive Negative Figure 7 illustrates a recommended approach Positive to the evaluation of patients who have iron Management Consider repeat deficiency anemia with or without a posi- as indicated endoscopy tive FOBT.25 Until proven otherwise, men and postmenopausal women with iron defi- Negative ciency anemia are assumed to have GI blood loss. In premenopausal women, menses may be the cause of anemia. However, colonic Capsule endoscopy and upper GI tract lesions, including cancer, have been reported in this group.26 Thus, pre- menopausal women who have iron deficiency anemia that cannot be explained by heavy Negative Positive menses, or those who have GI symptoms, should be evaluated for a GI cause.4,8,10 Is further Push enteroscopy, deep All patients should be evaluated for extrain- evaluation enteroscopy, or surgery needed? as indicated testinal causes such as epistaxis, hematuria, and heavy gynecologic bleeding. Endoscopic evaluation should include EGD and colo- No Yes noscopy.27 Biopsies should be performed Observation, medical management in the to evaluate for celiac dis- ease. Patients with obscure occult GI bleed- ing should undergo repeat upper endoscopy Recurrence? and colonoscopy. If these repeat are negative, capsule endoscopy should be No Yes performed to evaluate the small bowel. If a small bowel lesion is found, the patient may Follow-up Repeat routine endoscopy, capsule be evaluated with push enteroscopy, deep as indicated endoscopy; consider CT enterography enteroscopy, or surgery, as clinically indi- cated. If capsule endoscopy does not reveal Positive a lesion, second-look capsule endoscopy and CT enterography should be considered. In Push enteroscopy, deep enteroscopy, summary, men and postmenopausal women or surgery as indicated with iron deficiency anemia should undergo progressive evaluation for GI blood loss with colonoscopy, EGD, and capsule endoscopy, Figure 7. Proposed algorithm for diagnosis of iron deficiency ane- as clinically indicated.4,8,10 mia with or without a positive FOBT. (CT = computed tomographic; EGD = esophagogastroduodenoscopy; FOBT = fecal occult blood test.) Data Sources: A PubMed search was completed in Clini- Adapted with permission from American Gastroenterological Association medical position cal Queries using the key terms occult gastrointestinal statement: evaluation and management of occult and obscure gastrointestinal bleeding. bleeding and obscure gastrointestinal bleeding. The search . 2000;118(1):199.

March 15, 2013 ◆ Volume 87, Number 6 www.aafp.org/afp American Family Physician 435 Occult GI Bleeding

included meta-analyses, clinical trials, and reviews. Other 12. Triester SL, Leighton JA, Leontiadis GI, et al. A databases searched included the American Gastroentero- meta-analysis of the yield of capsule endoscopy com- logical Association, American Society for Gastrointestinal pared to other diagnostic modalities in patients with Endoscopy, Clinical Evidence, Essential Evidence Plus, obscure gastrointestinal bleeding. Am J Gastroenterol. Evidence-Based Medicine, OvidSP Medline, and UpToDate. 2005;100(11):2407-2418. Search dates: August 2010 and October 2011. 13. Hartmann D, Schmidt H, Bolz G, et al. A prospective two-center study comparing wireless capsule endoscopy with intraoperative enteroscopy in patients with obscure The Authors GI bleeding. Gastrointest Endosc. 2005;61(7):826-832. 14. Sears DM, Avots-Avotins A, Culp K, Gavin MW. Fre- KATHY BULL-HENRY, MD, is an associate professor in the quency and clinical outcome of capsule retention Department of Medicine at Georgetown University Hospi- during capsule endoscopy for GI bleeding of obscure tal, Washington, DC. origin. Gastrointest Endosc. 2004;60(5):822-827. FIRAS H. AL-KAWAS, MD, is a professor and chief of 15. Cave DR, Fleischer DE, Leighton JA, et al. A multicenter endoscopy in the Department of Medicine at Georgetown randomized comparison of the Endocapsule and the University Hospital. Pillcam SB. Gastrointest Endosc. 2008;68(3):487-494. 16. May A, Nachbar L, Ell C. Double-balloon enteroscopy Address correspondence to Kathy Bull-Henry, MD, (push-and-pull enteroscopy) of the small bowel: feasi- Georgetown University Hospital, Department of Medi- bility and diagnostic and therapeutic yield in patients cine, 3800 Reservoir Road NW, Washington, DC 20007 with suspected small bowel disease. Gastrointest (e-mail: [email protected]). Reprints are Endosc. 2005;62(1):62-70. not available from the authors. 17. Nakamura M, Niwa Y, Ohmiya N, et al. Preliminary Author disclosure: No relevant financial affiliations. comparison of capsule endoscopy and double-balloon enteroscopy in patients with suspected small-bowel bleeding. Endoscopy. 2006;38(1):59-66. REFERENCES 18. Manabe N, Tanaka S, Fukumoto A, Nakao M, Kamino D, Chayama K. Double-balloon enteroscopy in patients 1. Mitchell SH, Schaefer DC, Dubagunta S. A new view of with GI bleeding of obscure origin. Gastrointest Endosc. occult and obscure gastrointestinal bleeding. Am Fam 2006;64(1):135-140. Physician. 2004;69(4):875-881. 19. Upchurch BR, Sanaka MR, Lopez AR, Vargo JJ. The clini- 2. Rockey DC. Occult gastrointestinal bleeding. N Engl J cal utility of single-balloon enteroscopy: a single-center Med. 1999;341(1):38-46. experience of 172 procedures. Gastrointest Endosc. 3. Rockey DC. Occult gastrointestinal bleeding. Gastroen- 2010;71(7):1218-1223. terol Clin North Am. 2005;34(4):699-718. 20. Akerman PA, Agrawal D, Cantero D, Pangtay J. Spiral 4. Zuckerman GR, Prakash C, Askin MP, Lewis BS. AGA enteroscopy with the new DSB overtube: a novel tech- technical review on the evaluation and management of nique for deep peroral small-bowel intubation. Endos- occult and obscure gastrointestinal bleeding. Gastroen- copy. 2008;40(12):974-978. terology. 2000;118(1):201-221. 21. Huprich JE, Fletcher JG, Alexander JA, Fidler JL, Burton 5. Rockey DC. Occult and obscure gastrointestinal bleed- SS, McCullough CH. Obscure gastrointestinal bleeding: ing: causes and clinical management. Nat Rev Gastro- evaluation with 64-section multiphase CT enterog- enterol Hepatol. 2010;7(5):265-279. raphy—initial experience. Radiology. 2008;246(2): 6. Fine KD. The prevalence of occult gastrointesti- 562-571. nal bleeding in celiac sprue. N Engl J Med. 1996; 22. Hakim FA, Alexander JA, Huprich JE, Grover M, 334 (18 ):1163-1167. Enders FT. CT enterography may identify small bowel 7. Rostom A, Murray JA, Kagnoff MF. American Gastroen- tumors not detected by capsule endoscopy: eight terological Association (AGA) Institute technical review years experience at Mayo Clinic Rochester. Dig Dis Sci. on the diagnosis and management of celiac disease. 2011;56(10):2914-2919. Gastroenterology. 2006;131(6):1981-2002. 23. Rasmussen M, Kronborg O. Upper gastrointestinal can- 8. Godfrey JD, Brantner TL, Brinjikji W, et al. Mor- cer in a population-based program with fecal bidity and mortality among older individuals with occult blood test for . Scand J Gastro- undiagnosed celiac disease. Gastroenterology. enterol. 2002;37(1):95-98. 2010;139(3):763-769. 24. Jaffin BW, Bliss CM, LaMont JT. Significance of occult 9. Stewart JG, Ahlquist DA, McGill DB, Ilstrup DM, Schwartz gastrointestinal bleeding during anticoagulation ther- S, Owen RA. Gastrointestinal blood loss and anemia in apy. Am J Med. 1987;83(2):269-272. runners. Ann Intern Med. 1984;100(6):843-845. 25. American Gastroenterological Association medical 10. Raju GS, Gerson L, Das A, Lewis B; American Gastro- position statement: evaluation and management of enterological Association. American Gastroentero- occult and obscure gastrointestinal bleeding. Gastroen- logical Association (AGA) Institute technical review on terology. 2000;118(1):197-201. obscure gastrointestinal bleeding. Gastroenterology. 26. Green BT, Rockey DC. Gastrointestinal endoscopic eval- 2007;133(5):1697-1717. uation of premenopausal women with iron deficiency 11. Pennazio M, Santucci R, Rondonotti E, et al. Outcome anemia. J Clin Gastroenterol. 2004;38(2):104-109. of patients with obscure gastrointestinal bleeding after 27. Ioannou GN, Rockey DC, Bryson CL, Weiss NS. Iron defi- capsule endoscopy: report of 100 consecutive cases. ciency and gastrointestinal malignancy: a population- Gastroenterology. 2004;126(3):643-653. based cohort study. Am J Med. 2002;113(4):276-280.

436 American Family Physician www.aafp.org/afp Volume 87, Number 6 ◆ March 15, 2013