Oral (15W133)

Identification of predictive factors for positive capsule endoscopy findings in obscure gastrointestinal bleeding and anaemia


Grainne Holleran, Mary Hussey, Deirdre McNamara


Department of Gastroenterology, Tallaght Hospital and Trinity Academic Gastroenterology Group


Capsule endoscopy(CE) is the recommended first second-line investigation in obscure gastrointestinal bleeding(OGIB) and iron deficiency anaemia(IDA). Currently, there are no known predictive factors to identify high risk patients or guidelines on prioritising patients.


To evaluate the diagnostic yield(DY) of CE in IDA and OGIB, and to identify predictive factors which may be useful in prioritising requests for CE.


A retrospective review of the CE database at Tallaght Hospital was performed identifying patients who underwent a CE for IDA or OGIB;either occult or overt. Data was collected by reviewing clinic and referral letters, discharge summaries and laboratory results. Information including:demographics, CE indication, lowest haemoglobin(Hb)level, interval from referral, comorbidities and medication usage, was recorded. Results were compared between groups based on CE positivity and clinical factors using a student’s t test or relative risk(RR) ratio.


From our database of 1379CEs, we identified 455patients whose indication was IDA or OGIB. Of these, 51%(n=232) were female and the mean age was 64years(13-99). Indications were IDA:62%(n=281), occult bleeding:26%(n=117) and overt bleeding:12%(n=57). The overall DY was 57%(n=258), with a significantly lower DY-49%(n=137)(p=0.01) in IDA, compared to occult(63(n=74) and overt(77%(n=44) OGIB. Between the negative and positive CE groups, there was a significant difference in age 61vs67years(p<0.001,95%CI -8.91- -2.93) and lowest Hb10.3g/dL vs9.3g/dL(p<0.003,95%CI 0.33-1.55), but there was no difference in gender. The presence of overt bleeding increased the DY from 53%to77%, with a RR of 1.46(p<0.0001). Clinical data was available in 39%(n=178) of patients, of which 56%(n=99) had positive findings. In the 56(31%) patients who underwent CE within 4days of presentation, the DY was significantly higher than beyond 4days with DYs of 68%vs49%(p=0.02, 95%CI 0.03-0.34) respectively. The use of warfarin and antiplatelet agents had RRs of 1.36(p=0.03) and 1.31(p=0.04) respectively, however no medical conditions or other medications were found to be predictive. When Hb levels were assessed, a cut-off<10g/dL was associated with a positive CE, with a RR of 1.43(p<0.03), and a DY of 61%vs44% with a Hb>10g/dL.


The DY of CE in OGIB is increased in patients with overt bleeding and when performed within 4days of presentation. Older age, Hb<10g/dL, anticoagulation and antiplatelets all increase the risk of positive findings and should be assessed when prioritising patients for CE.