PTU-034 Double Blind Randomised Controlled Trial of Magnetically Steerable Gastric Capsule Endoscopy (MSGCE) Vs. Conventional Gastroscopy for Detection of Beads in a Porcine Stomach

    loading  Checking for direct PDF access through Ovid



Gastroscopy is uncomfortable for patients and incurs the risks of intubation and sedation. Capsule endoscopy is well tolerated and recently a handheld magnet has been developed to enable steering of the capsule to visualise all areas of the capacious stomach. Our preliminary data suggests that a novice can identify all beads sewn into a porcine stomach within 4 min after 40 consecutive examinations.1 We performed a double blind randomised controlled trial comparing MSGCE with conventional gastroscopy in the detection of beads in the same model.


Ex-vivo porcine stomach models were used in a standard housing unit. MSGCE was performed according to a standard protocol using 1000mls of water to distend each stomach and a combination of positional change (head down, 30° left lateral, 30° right lateral) and magnetic control to steer the capsule. Each model was examined in a standard fashion by gastroscopy and subsequently MSGCE using MiroCam Navi (Intromedic Ltd). Two blinded investigators (MFH and IR) competent to perform both procedures were allocated randomly to perform either gastroscopy or MSGCE on each model.


This was performed as a non-inferiority study with an expected sensitivity of 90% for both (0 estimated difference), a specificity of 100% and a difference of interest 10 percentage points (i.e., 80% is significantly worse). A sample size of 85 beads was needed to achieve this statistical power. Twelve porcine stomachs were prepared with beads as follows: 2 × 0 beads, 2 × 1 bead, 2 × 2 beads, 2 × 3 beads, 2 × 4 beads, 2 × 5 beads, giving a total of 30 beads. The study was conducted in three rounds, giving a total of 90 beads to be identified. Number of beads identified and procedure duration was recorded.


Gastroscopy correctly identified 88% (79/90) beads, MSGCE correctly identified 89% (80/90) beads and thus is non-inferior to gastroscopy in this setting (95% CI 82.54–95.46%). Mean examination times for gastroscopy and MSGCE were 3.34 min and 9.90 min respectively. MSGCE overestimated the number of beads present on a single occasion.


MSGCE is equivalent to conventional gastroscopy in the detection of beads placed in a porcine stomach model. Procedure duration was longer for MSGCE compared to gastroscopy. Further studies in humans are necessary to define the scope and utility of this exciting new technique.

Disclosure of Interest

None Declared.

Related Topics

    loading  Loading Related Articles