Background and aims: The position of capsule endoscopy (CE) relative to push enteroscopy (PE) in the diagnostic algorithm of obscure gastrointestinal bleeding is unclear, as previous studies involved the use of both techniques in all patients. We therefore conducted a trial in which patients were randomized to undergo one or other exploration.
Methods: All consecutive patients referred for obscure gastrointestinal bleeding were randomized between CE and PE as the first-line exploration. The alternative method was only used if the first-line method revealed no definite bleeding source, or if required for clinical reasons during follow-up.
Results: CE and PE, used as the first-line exploration, identified a bleeding source in 20 of 40 patients and 9 of 38 patients, respectively (50% vs 24%; P = .02). CE missed lesions in 8% of patients, and all these lesions were located in sites accessible to standard endoscopy. PE missed lesions in 26% of patients. At the end of the 12-month follow-up period, the strategy based on CE as first-line exploration followed by PE if necessary only was similar to PE followed by CE in terms of diagnostic yield, clinical outcome, and therapeutic impact, but reduced the percentage of patients needing the alternative exploration (25% vs 79%; P < .001).
Conclusions: CE has a higher diagnostic yield than PE in obscure gastrointestinal bleeding, and a strategy based on CE as first-line exploration avoids unnecessary explorations.