Low Agreement Between Two Core Small-Bowel Tests in Mid-Gut Bleeding
In patients with mid-gut gastrointestinal bleeding, capsule endoscopy and double-balloon enteroscopy do not line up as closely as clinicians often hope. A retrospective study from a referral hospital in Lima, Peru, found only low overall diagnostic agreement between the two procedures, even though concordance was more respectable when the analysis focused on lesion type and lesion location.
The cohort included 51 patients, 28 of them women, with a median age of 65 years. Most presented with overt bleeding, while the remainder had occult bleeding, and the mean hemoglobin level was 7.3 ± 2.24 g/dL. This was a clinically sick group, not a screening population. In a setting like this, every diagnostic discrepancy carries real consequences: delayed therapy, repeated procedures, and uncertainty about whether a lesion is truly present or has already been missed by the next test in sequence.
Video capsule endoscopy was followed by double-balloon enteroscopy in all cases, with 39 antegrade procedures and 12 retrograde procedures. The median interval between the two tests was 13 days, with an interquartile range of 5 to 29 days. That gap is long enough for active bleeding to quiet, for subtle mucosal lesions to become harder to recognize, or for the endoscopic appearance of a transient lesion to change between examinations. In a bleeding disorder driven by intermittent pathology, timing can alter what each modality captures.
Discordance Was Common After a Positive Capsule Study
The most clinically striking result came from the mismatch itself. In 10 patients, or 19.6% of the cohort, capsule endoscopy showed positive findings but double-balloon enteroscopy found no lesion. In one additional patient, capsule endoscopy was negative while double-balloon enteroscopy was positive. The overall diagnostic agreement was low, with a kappa coefficient of 0.37, 95% CI 0.10-0.64, and p = 0.002.
One sentence captures the bedside problem: a positive capsule does not guarantee confirmation on enteroscopy.
That discordance does not automatically mean either test is wrong. Capsule endoscopy surveys the small bowel noninvasively and may detect fleeting bleeding stigmata, small vascular lesions, or mucosal abnormalities that are difficult to relocate later with a device that depends on route selection and segmental reach. Double-balloon enteroscopy, by contrast, permits direct inspection and intervention, but its yield depends heavily on knowing where to go, how far to advance, and whether the target lesion is still visible by the time the procedure occurs. In a patient with substantial anemia and intermittent hemorrhage, the two tools may be measuring slightly different moments in the disease process.
The study design reinforces that possibility. This was observational, retrospective, and comparative, using standard statistical methods with IBM SPSS Statistics v.20.0 and R version 4.1.2. A retrospective series cannot fully separate a true false-positive capsule result from a lesion that disappeared, evolved, or sat beyond the reach of the second procedure. Even so, the low kappa value is a clear signal that the tests should not be treated as interchangeable confirmation methods.
Lesion Type and Location Matched Better Than Overall Diagnosis
Agreement improved when the focus narrowed from overall diagnosis to more granular features of the bleeding source. Concordance was moderate for lesion type, with kappa 0.58, 95% CI 0.42-0.74, and p < 0.001. Concordance for lesion location was also moderate, with kappa 0.50, 95% CI 0.31-0.69, and p < 0.001. Vascular lesions and tumoral lesions each showed moderate agreement, with kappa values of 0.60.
This pattern makes biological sense. Lesion classification is often easier than exact case-by-case matching, especially when the pathology is conspicuous. A vascular lesion remains a vascular lesion even if its appearance changes slightly between tests, and a tumor may still be recognizable despite differences in visualization angle or preparation quality. The harder task is confirming that the exact abnormality seen on capsule truly corresponds to the same finding reached later by deep enteroscopy, especially when bleeding has slowed and the bowel has already been traversed once.
The data also fit the practical realities of referral-center care. In mid-gut bleeding, clinicians often need a diagnostic sequence rather than a single definitive test. Capsule endoscopy can guide the route for enteroscopy, but this series suggests that the handoff between the two modalities is imperfect. A capsule result may support suspicion and help localize disease, yet the follow-up procedure can still fail to recover the same lesion. For patients with severe anemia, that can translate into repeated workups and a longer path to definitive treatment.
The lower overall agreement should not be read as a dismissal of either test. Each has a different clinical role. Capsule endoscopy is valuable for surveying the small bowel, while double-balloon enteroscopy remains critical when tissue diagnosis or therapy is needed. The tension lies in using one test to direct the other when the disease process is dynamic and the interval between tests is not trivial.
In that sense, the Peruvian series is less about a winner and more about diagnostic friction. The lesion may be real, but the trail it leaves can shift before the second scope arrives. For patients with overt or occult bleeding and hemoglobin near 7 g/dL, that gap matters at the bedside because the next decision often depends on whether the lesion can be found twice, not once.
The study leaves one practical limitation in plain view: retrospective agreement studies cannot tell clinicians how often a capsule-positive, enteroscopy-negative result represents a missed lesion, a transient abnormality, or a target outside the reach of the second procedure. A prospective design with tighter timing between tests would be better suited to answer that question.
For now, the message is straightforward. Capsule endoscopy and double-balloon enteroscopy provide overlapping but not identical information in mid-gut gastrointestinal bleeding, and the overlap is modest at the level that matters most for diagnosis. The discordance is not rare, and it should shape how teams interpret a negative follow-up enteroscopy after a positive capsule study.