OBJECTIVES: Due to its superior ability to examine the entire small bowel mucosa, capsule endoscopy (CE) has broadened the diagnostic evaluation of patients with obscure gastrointestinal bleeding (OGIB). Published studies have revealed a numerically superior performance of CE in determining a source of OGIB compared with other modalities, but due to small sample sizes, the overall magnitude of benefit is unknown. Additionally, the types of lesions more likely to be found by CE versus alternate modalities are also unknown. The aim of this study was to evaluate the yield of small bowel findings with CE in patients with OGIB compared to other modalities using meta-analysis. METHODS: We performed a recursive literature search of prospective studies comparing the yield of CE to other modalities in patients with OGIB. Data on yield and types of lesions identified among various modalities were extracted, pooled, and analyzed. Incremental yield (IY) (yield of CE-yield of comparative modality) and 95% confidence intervals (95% CI) of CE over comparative modalities were calculated. RESULTS: A total of 14 studies (n = 396) compared the yield of CE with push enteroscopy for OGIB. The yield for CE and push enteroscopy was 63% and 28%, respectively (IY = 35%, p < 0.00001, 95% CI = 26-43%) and for clinically significant findings (n = 376) was 56% and 26%, respectively (IY = 30%, p < 0.00001, 95% CI = 21-38%). Three studies (n = 88) compared the yield of CE to small bowel barium radiography. The yield for CE and small bowel barium radiography for any finding was 67% and 8%, respectively (IY = 59%, p < 0.00001, 95% CI = 48-70%) and for clinically significant findings was 42% and 6%, respectively (IY = 36%, p < 0.00001, 95% CI = 25-48%). Number needed to test (NNT) to yield one additional clinically significant finding with CE over either modality was 3 (95% CI = 2-4). One study each compared the yield of significant findings on CE to intraoperative enteroscopy (n = 42, IY = 0%, p= 1.0, 95% CI =-16% to 16%), computed tomography enteroclysis (n = 8, IY = 38%, p= 0.08, 95% CI =-4% to 79%), mesenteric angiogram (n = 17, IY =-6%, p= 0.73, 95% CI =-39% to 28%), and small bowel magnetic resonance imaging (n = 14, IY = 36%, p= 0.007, 95% CI = 10-62%). Ten of the 14 trials comparing CE with push enteroscopy classified the types of lesions found on examination. CE had a 36% yield for vascular lesions versus 20% for push enteroscopy, with an IY of 16% (p < 0.00001, 95% CI = 9-23%). Inflammatory lesions were also found more often in CE (11%) than in push enteroscopy (2%), with an IY of 9% (p= 0.0001, 95% CI = 5-13%). There was no significant difference in the yield of tumors or "other" findings between CE and push enteroscopy. CONCLUSIONS: CE is superior to push enteroscopy and small bowel barium radiography for diagnosing clinically significant small bowel pathology in patients with OGIB. In study populations, the IY of CE over push enteroscopy and small bowel barium radiography for clinically significant findings is >or=30% with an NNT of 3, primarily due to visualization of additional vascular and inflammatory lesions by CE.
OBJECTIVES: Due to its superior ability to examine the entire small bowel mucosa, capsule endoscopy (CE) has broadened the diagnostic evaluation of patients with obscure gastrointestinal bleeding (OGIB). Published studies have revealed a numerically superior performance of CE in determining a source of OGIB compared with other modalities, but due to small sample sizes, the overall magnitude of benefit is unknown. Additionally, the types of lesions more likely to be found by CE versus alternate modalities are also unknown. The aim of this study was to evaluate the yield of small bowel findings with CE in patients with OGIB compared to other modalities using meta-analysis. METHODS: We performed a recursive literature search of prospective studies comparing the yield of CE to other modalities in patients with OGIB. Data on yield and types of lesions identified among various modalities were extracted, pooled, and analyzed. Incremental yield (IY) (yield of CE-yield of comparative modality) and 95% confidence intervals (95% CI) of CE over comparative modalities were calculated. RESULTS: A total of 14 studies (n = 396) compared the yield of CE with push enteroscopy for OGIB. The yield for CE and push enteroscopy was 63% and 28%, respectively (IY = 35%, p < 0.00001, 95% CI = 26-43%) and for clinically significant findings (n = 376) was 56% and 26%, respectively (IY = 30%, p < 0.00001, 95% CI = 21-38%). Three studies (n = 88) compared the yield of CE to small bowel barium radiography. The yield for CE and small bowel barium radiography for any finding was 67% and 8%, respectively (IY = 59%, p < 0.00001, 95% CI = 48-70%) and for clinically significant findings was 42% and 6%, respectively (IY = 36%, p < 0.00001, 95% CI = 25-48%). Number needed to test (NNT) to yield one additional clinically significant finding with CE over either modality was 3 (95% CI = 2-4). One study each compared the yield of significant findings on CE to intraoperative enteroscopy (n = 42, IY = 0%, p= 1.0, 95% CI =-16% to 16%), computed tomography enteroclysis (n = 8, IY = 38%, p= 0.08, 95% CI =-4% to 79%), mesenteric angiogram (n = 17, IY =-6%, p= 0.73, 95% CI =-39% to 28%), and small bowel magnetic resonance imaging (n = 14, IY = 36%, p= 0.007, 95% CI = 10-62%). Ten of the 14 trials comparing CE with push enteroscopy classified the types of lesions found on examination. CE had a 36% yield for vascular lesions versus 20% for push enteroscopy, with an IY of 16% (p < 0.00001, 95% CI = 9-23%). Inflammatory lesions were also found more often in CE (11%) than in push enteroscopy (2%), with an IY of 9% (p= 0.0001, 95% CI = 5-13%). There was no significant difference in the yield of tumors or "other" findings between CE and push enteroscopy. CONCLUSIONS: CE is superior to push enteroscopy and small bowel barium radiography for diagnosing clinically significant small bowel pathology in patients with OGIB. In study populations, the IY of CE over push enteroscopy and small bowel barium radiography for clinically significant findings is >or=30% with an NNT of 3, primarily due to visualization of additional vascular and inflammatory lesions by CE.