Capsule Endoscopy or Angiography in Patients With Acute Overt Obscure Gastrointestinal Bleeding

A Prospective Randomized Study With Long-Term Follow-Up

Wai K Leung MD; FACG; Simon S M Ho MBBS; Bing-Yee Suen BN; Larry H Lai MBChB; Simon Yu MD; Enders K W Ng MD; Simon S M Ng MD; Philip W Y Chiu MD; Joseph J Y Sung MD; PhD; Francis K L Chan MD; James Y W Lau MD


Am J Gastroenterol. 2012;107(9):1370-1376. 

In This Article


This was the first clinical study to prospectively compare CE with angiography in patients with acute overt OGIB. We determined both the diagnostic yield and the long-term outcomes of patients randomized to two different modalities. The diagnostic yield of CE was found to be significantly higher than mesenteric angiography (53.3% vs. 20%). Although the cumulative rebleeding rate was higher in the angiography group, the difference did not reach statistical significance (Figure 4). There was also no difference on other long-term outcomes between the two groups. Our data support the idea that CE may be preferred over angiography, even if it does not improve clinical efficacy, because it is less invasive, avoids radiation exposure, and has comparable long-term outcome.

There are very few prospective randomized studies that directly compare CE with other modalities in patients with OGIB. There are only two prospective randomized studies comparing CE with push enteroscopy or small-bowel series.[13,14] Both studies, however, also recruited patients with occult bleeding or unexplained iron deficiency anemia, which may have different implications on diagnostic yield and long-term rebleeding risk. Although previous nonrandomized studies demonstrated the high diagnostic yield of CE,[5,7,8,15] it remains to be determined whether it could be reproduced in randomized trial and translated into clinical benefits with emphasis on reduction of subsequent bleeding.[16] In this study, we found no significant difference on long-term rebleeding, further hospitalization, and transfusion rates between the two groups. In a recent prospective randomized study, CE was shown to be superior to dedicated small-bowel radiography in detecting small-bowel lesions.[13] There was also no significant difference on rebleeding rates, further blood transfusions, and hospitalization requirement between the two groups at 12 months. Our findings are therefore consistent with their observation that the long-term outcomes of patients receiving initial CE or angiography were comparable.

The higher sensitivity of CE in patients with overt OGIB in this study could be explained by several factors. All CEs were performed immediately after negative upper and lower endoscopy, typically within the first 24 h of hospitalization. It has been shown that CEs performed during active or recent bleeding have higher diagnostic yield.[5] Apart from immediate CE that would enhance the diagnostic yield, CE also enables direct visual examination of the small-bowel mucosa. Unlike angiography that only detects gross vascular or structural abnormality, CE could detect both bleeding and nonbleeding mucosal lesions. Notably, CE also identifies non-small-bowel bleeding lesions like bleeding lesions in stomach or proximal colon that were overlooked by conventional endoscopy. In this study, three patients were actually found to have bleeding from stomach during CE. In keeping with this observation, a previous enteroscopy study also found that up to 24% of lesions identified during enteroscopy examination were within reach of upper endoscopy.[17] Hence, the latest AGA Institute Technical Review actually recommends repeat standard endoscopy for lesions overlooked by prior upper endoscopy as part of diagnostic work up.[1]

A recent meta-analysis reported the overall sensitivity of CE to be 60% in patients with OGIB,[8] which was slightly higher than our findings. It may be accounted by the stringent criteria we used for classifying a positive CE. We only included CE findings that were shown to be associated with high risk of bleeding to be positive findings.[12] Although it remains arguable whether fresh blood in small bowel is a positive finding of CE, our experiences showed that finding of fresh blood in small bowel other than proximal duodenum usually indicate active small bleeding. In this study, four patients were found to have fresh blood in small bowel during CE examinations. The small-bowel bleeding sources were subsequently identified in three of the four patients. The remaining patient who failed to have bleeding sites determined despite extensive investigations developed rebleeding at 8 months. All these findings support the validity of including fresh blood in small bowel as positive findings for CE.

On the other hand, there are very few data on the diagnostic yield of mesenteric angiography in patients with OGIB. The reported sensitivity of angiography in previous case series ranged from 40 to 80%.[18,19] Although angiography was performed immediately after negative endoscopy in this study, bleeding might have been stopped in some patients at the time of angiography. We were unable to randomize 2 patients (out of 91) who failed to respond to resuscitation and required immediate surgical intervention. As these patients have a 50% chance of randomizing into CE group, they were excluded in the original protocol. The exclusion of these patients may possibly lower the sensitivity of angiography in this study.

This study had other limitations. Although angiography is usually recommended in patients with massive bleeding, the recruitment of patients with overt but nonmassive bleeding to either CE or angiography may not be the ideal patient groups for comparison. As the current guidelines[1,2,10] are based on very limited data from prospective clinical trials, the exact positioning of CE and angiography in patients with different degree of OGIB are largely based on expert opinion. Although this study helps to answer some of the uncertainties, it is arguable that some of the patients recruited in this study would not undergo angiography in clinical practices because of the moderate degree of bleeding. This would limit the utility and relevance of our findings. Second, we based our sample size estimation on the diagnostic yield of two investigations rather than clinical outcomes of the two groups. Although our findings are in favor of CE, this study is underpowered to detect any significant difference on clinical outcome that would limit the impact of our findings. Based on the current rebleeding rate (16.7% vs. 33.7%) and a power of 0.8, a total sample size of 166 is needed to demonstrate a potential difference on long-term rebleeding between the two groups. As overt OGIB is not that common, this sample size is not easily accomplished. On the other hand, the use of safety end point would be even more difficult because of the rarity of complications related to these two investigations (usually <5%). The small sample size also makes subgroup analysis difficult, particularly on the risk of rebleeding and the detection rate of different small-bowel pathology. Last, another limitation of this study is the lack of gold standard for bleeding source. There is so far no gold standard for OGIB. Even with the use of double-balloon enteroscopy and CT angiography, not all causes of bleeding could be identified. In a recent meta-analysis, the diagnostic yield of double-balloon enteroscopy may be even lower than CE.[20] As shown in this study, the bleeding sources remained unidentified in three rebleeding patients despite further extensive investigations ( Table 3). Although the use of multiple investigations as a gold standard may help to better determine the diagnostic yield, it would make subsequent assessment of long-term clinical outcomes in patients assigned to a particular investigation not possible. In this study, all patients with positive findings or on-going bleeding would receive further investigations for determination of bleeding source and control of bleeding. However, further investigations were not performed in patients with negative findings on initial assigned investigation unless they have evidence of clinical rebleeding. This is to simulate the real-life clinical situations and is also consistent with most recommendations when patients with no further rebleeding are put on expectant management.

With the relatively low diagnostic yield of angiography and increasing availability of deep enteroscopy and other competing technology like CT enterography, we believe that future clinical trials should be directed to compare the performance of CE with these new modalities. Deep enteroscopy is increasingly used as a first-line investigation for OGIB.[21] Although recent data suggested that the diagnostic yield of enteroscopy is comparable to CE, randomized clinical data are lacking.[20] Moreover, the use of CT enterography has been shown to identify more small intestinal masses than CE.[22] Further studies are necessary to characterize the role of CT enterography in patients with OGIB.

In conclusion, we have shown that the diagnostic yield of immediate CE is superior to angiography in patients with active overt OGIB. There was also no significant difference on the long-term outcomes between the two groups. With the less invasive nature of CE, it may be preferred over angiography in patients with active overt OGIB.