Introduction Capsule endoscopy (CE) is the first line modality to investigate the small bowel (SB) in patients with obscure gastrointestinal bleeding (OGB). We investigated the largest cohort of patients to date with OGB who underwent CE.

Methods All patients referred for CE for OGB were included retrospectively. Data was collected for demographics, clinical indication (overt bleeding (OB) or iron deficiency anaemia (IDA)) the presence of co-morbidity, medications including anticoagulation, diagnostic yield (DY) and subsequent follow up.

Results A total of 1324 patients were identified. The mean age was 60 years with females consisting of 55%. The indication for CE was IDA in 73%(n = 971) and OB in 27%. Significant co-morbidities were present in 38% of patients. Whilst 8% of patients were on warfarin, 6%(n = 84) were transfusion dependent. The cumulative number of lesions identified was 1086.The overall DY was 36%(n = 473) for lesions deemed to be the source of OGB. On logistic regression, increasing age (p < 0.001) and transfusion dependence (p = 0.02) were associated with a DY with CE. The commonest diagnosis was angioectasia (AE) in 23%(n = 302).Other SB findings included ulcers and erosions in 25%(n = 326), fresh blood without a definite source in 6%(n = 79) whilst tumours, polyps and crohns disease represented 2% each respectively. Miscellaneous findings in the SB included diverticulae (14), SB strictures (15), SB varices (9), dielaufoy (1) and others (28). There was no significant difference in the DY between those with IDA and OB (p = 0.35) and between the sexes (p = 0.44). Management was altered in 78%(n = 367) of those with a positive DY, in the form of further procedures (42%, n = 153). Of these, 35%(n = 53) had double balloon enteroscopy, 24%(n = 36) had push enteroscopy, 23%(n = 35) had repeat standard endoscopy, 17 patients were referred for surgery whilst 2 patients had intra-operative enteroscopy. 34%(n = 52) of patients received argon photocoagulation therapy for their AE. On logistic regression, factors associated with a subsequent change in management included the indication of OB (p = 0.03), the presence of comorbidity (p = 0.03), transfusion dependence (p = 0.02) and the presence of AE (p < 0.001) on the CE. In 14% of patients, significant lesions were found outside the SB (upper GI tract 12% and colon 2%).

Conclusion This is the largest series on the use of CE in OGB in the published literature. A significant number of patients with a positive DY had a change in management. Factors that were more likely to predict an alteration of outcome were the presence of co-morbidity, transfusion dependence and the presence of AE. Although, the DY did not differ significantly between those with IDA and OB, management was more likely to be altered in those with OB.

Disclosure of Interest None Declared.