Capsule Endoscopy and Double-Balloon Enteroscopy Do Not Match Reliably in Mid-Gut Bleeding
In patients with mid-gut gastrointestinal bleeding, two of the most widely used diagnostic tools did not track closely with one another. At a referral hospital in Lima, Peru, video capsule endoscopy and subsequent double-balloon enteroscopy showed low overall agreement, even though both tests are central to evaluating bleeding from the small bowel.
The study reviewed 51 patients treated between 2020 and 2025. Twenty-eight were women, and the median age was 65 years, with an interquartile range of 55 to 71. Most patients, 34 of 51, presented with overt bleeding rather than occult blood loss. Their mean hemoglobin was 7.3 ± 2.24 g/dL, a level that places the clinical problem squarely in the realm of severe anemia and urgent source localization. In practical terms, this is the kind of patient population in which a missed lesion can prolong transfusion needs, delay therapy, and leave the bleeding source untreated.
The timing between the two procedures also matters. The median interval was 13 days, with a spread from 5 to 29 days. In a bleeding disorder driven by transient mucosal injury, intermittent vascular lesions, or small tumors that may bleed unevenly, a gap of nearly two weeks can easily separate what one test sees from what the next can no longer find.
A Discordance Built Into the Diagnostic Pathway
Investigators compared the two procedures using the Kappa coefficient, a standard way to measure agreement beyond chance. The overall diagnostic agreement was low, with κ = 0.37, 95% CI 0.10-0.64, and p = 0.002. In plain clinical terms, the two procedures often did not tell the same story.
The discordance ran in both directions. In 10 patients, representing 19.6% of those with positive capsule endoscopy findings, double-balloon enteroscopy did not identify a lesion. In one patient, or 2%, capsule endoscopy was negative while double-balloon enteroscopy found a lesion. Those mismatches are not trivial. They reflect the central challenge of small-bowel bleeding: lesions can be subtle, intermittent, or beyond the reach of the segment entered, and the bowel is not a static organ when the diagnostic clock is ticking.
The split between antegrade and retrograde enteroscopy also frames the problem. Thirty-nine procedures were antegrade and 12 were retrograde, suggesting that lesion location and endoscopic reach likely influenced which route was selected. In small-bowel bleeding, localization is not simply a technical detail. It determines whether the endoscopist reaches the lesion, whether bowel preparation is adequate, and whether the finding persists by the time therapeutic enteroscopy begins.
Lesion Type and Location Agreed Better Than the Global Result
The broader discordance did not mean the tests were uniformly disconnected. Agreement improved when investigators examined lesion characteristics rather than the global diagnosis. Concordance was moderate for lesion type, with κ = 0.58, 95% CI 0.42-0.74, and p < 0.001, and moderate for lesion location, with κ = 0.50, 95% CI 0.31-0.69, and p < 0.001.
Vascular lesions and tumoral lesions each showed moderate agreement, with κ = 0.60 for both. For clinicians, that pattern is informative. Vascular abnormalities may be easier to recognize across modalities when they are present and reachable, while tumors tend to produce more consistent structural signatures than faint erosions or transient bleeding points. Even so, the modest agreement on type and location does not restore confidence in the overall concordance between the two tests.
A one-sentence bedside takeaway follows directly from the data: a positive capsule study does not guarantee that double-balloon enteroscopy will find the same lesion days later.
What These Findings Mean for Small-Bowel Bleeding Workups
The findings fit the biology of small-bowel hemorrhage. Bleeding in the mid-gut often comes from lesions that are intermittently active, difficult to localize, or anatomically remote. Capsule endoscopy visualizes the mucosal surface along the entire small bowel, but it cannot biopsy, cauterize, or insufflate. Double-balloon enteroscopy can treat and sample lesions, yet it is constrained by route selection, reach, operator experience, and the exact segment accessed during the procedure. When there is a delay between tests, the visible lesion may resolve, clot over, or shift from the window seen on capsule imaging.
The study design was observational, retrospective, and comparative, so it captures real-world diagnostic performance rather than a tightly controlled experimental setting. That format is useful because it reflects how these tests are actually used in referral centers, where patients often arrive after anemia has already become advanced and the diagnostic sequence depends on availability. The tradeoff is familiar: retrospective work can reveal discordance clearly, but it cannot prove which procedure is more correct in every case or whether the interval between them systematically altered results.
Even with that limitation, the message is clinically concrete. In this Peruvian cohort, video capsule endoscopy and double-balloon enteroscopy were complementary, but they were not interchangeable. The low overall Kappa value argues against assuming that one test simply validates the other in a routine way. Moderate agreement on lesion type and location suggests that the two methods can still converge on the broad diagnostic category, especially for vascular and tumoral lesions, yet the mismatch rate shows why clinicians managing obscure bleeding often need sequential imaging, careful timing, and readiness to re-evaluate when the second test conflicts with the first.
The work leaves one practical question unresolved for day-to-day care: how much of the discordance came from true lesion disappearance, and how much came from the limits of procedural reach and timing? The study does not answer that directly, but the data make the concern impossible to ignore.