Two types of sheaths were designed, namely the regular and supported types (Fig. 1). The regular type of sheath was made of polytetrafluoroethylene (PTFE, Sang-A Flontec Co. Ltd., Incheon, Korea) tube, 90 cm in length with a 4 mm outer diameter and a 3 mm inner diameter. The length of the sheath was decided after consideration of the optimum working length and the distance between the incisor and the esophagogastric junction. The diameter and thickness of the sheath were decided after consideration of the physical properties of the PTFE tube. The distal end of the regular sheath was bent using a heat-gun, creating a bend that was 4 cm long with an angle of curvature of 100 degrees. A radiopaque ring was attached to the outer surface of the distal end of the sheath for the purpose of providing easy detection during fluoroscopic guidance. In order to fabricate the supported sheath, a 5 Fr catheter catheter was attached to a regular sheath to act as the side lumen for the passage of a 0.035-inch wire designed to support the sheath on the other side of the bent tip, parallel to the long axis. The supporting wire in the side lumen can be pushed against the greater curvature of the gastric wall, while supporting the sheath and keeping it oriented toward the pylorus.

To draw the outline of the stomach, the distance from the greater curvature of the stomach to the lesser curvature was measured on each section of an abdominal CT (Somatom Plus IV, Siemens, Erlangen, Germany) scan taken after the ingestion of one liter of contrast media (slice thickness = 1 cm) in a patient in whom the gastric outlet was not obstructed. The distances from the location of the greater curvature on one section of the CT scan to the same location on the lower and upper sections were also measured. Both the distances from the greater to the lesser curvature and the location of the greater curvature on one section relative to that on the next section were plotted on a 2-dimensional plane, and the contour of the greater and lesser curvature of the stomach was drawn by connecting these points on a 2-dimensional plane (Fig. 2A). The esophageal outline was drawn from the distal portion of the esophagus to the cardia, with a diameter of 2 cm and a length of 5 cm. Clay was shaped so as to have a thickness of 2 cm, in order to produce the gastric contour with a tubular shape at the esophageal portion, and it was then put into a molding flask made of acryl plate (Atoglas, Elf Atochem Korea Ltd., Seoul, Korea) (Fig. 2B). The molding flask was then filled with silicone, and the hardening process was performed. The clay in the gastric and esophageal portion was removed from the hardened silicone, and the exposed side of the silicone was covered with a lucent acryl plate. The cavity between the acryl plate and the hardened silicone having the shape of a stomach was used as the gastric phantom. PTFE tube, with an outer diameter of 3 mm and a length of 10 cm, having a blind distal end, was attached to the gastric phantom, so as to act as the duodenal and pyloric portions of the stomach.

The ability of the newly designed sheaths to change the direction of the guide wire or coil catheter from parallel to transverse to the esophageal portion, and to guide it toward the pyloric portion was evaluated, while the wire or coil catheter was advanced to the duodenal portion beyond the pyloric portion of the phantom. The ability of the sheath to prevent the guide wire and coil catheter from being pushed toward the greater curvature was also evaluated. Once the sheath was inserted into the gastric portion via the esophageal portion, the location of the sheath's tip, without any force being applied to the sheath or guide wire, was chosen as the central point. The inner circle, which was surrounded by concentric circles drawn at intervals of 0.5 cm in order to act as a scale, corresponded to the tip of the sheath. A 0.038-inch guide wire (Radiofocus, Terumo Corporation, Tokyo, Japan) was advanced to the pyloric and duodenal portions through the sheath. When the guide wire was continuously pushed toward the blind end of the duodenal portion, it reached a point where it did not advance any further, whereupon it sloped down toward the greater curvature. Subsequently, the tip of the sheath was displaced away from the central point and toward the curved guide wire. Pushing was stopped when the curved guide wire contacted the greater curvature of the phantom, in order to prevent bias due to differences in the pushing force. The distance between the central point and tip of the sheath was measured. This experiment was repeated 30 times using both the regular and supported types of sheath.

With the supported sheath, an additional experiment was performed, and this involved advancing a supporting wire through the side lumen. A supporting wire with a diameter of 0.035-inches was inserted into the stomach portion through the side lumen and the tip was located at the blind end of the duodenal portion. When the wire was pushed, it advanced against the greater curvature side of the gastric phantom, and the distal portion of the sheath was propelled toward the pyloric portion. The distance of movement of the tip of the supported sheath, from the central point to its final location, was measured, while pushing both the guide and supporting wires (Figs. 3A, B).

The statistical significance of the differences between the distance of movement of the tip of the sheath in the regular and supported sheaths, with and without the supporting wire, was evaluated by means of the ANOVA postHoc test. The same experiment was repeatedly performed, while pushing both the guide wire and coil catheter, which was custom-made at this institution for the purpose of gastroduodenal intervention. The coil catheter was made of a stainless steel coil and a heat-compressive tube (1.3 mm in internal diamenter, 2.5 mm in external diameter and 90 cm in length) (13) (Figs. 4A, B).

All experimental procedures were performed in accordance with the National Institutes of Health guidelines for the humane handing of animals and approved by the committee for animal research at our institution.

Twelve experimental sessions were performed in six mongrel dogs, in order to evaluate the torque and support-ability of the sheath in the stomach, while pushing a guide wire or a coil catheter under fluoroscopic guidance (Fig. 5). For general anesthesia, the dogs were premedicated intramuscularly with ketamine hydrochloride (20 mg/kg, Yuhan Pharmacy, Seoul, Korea) and atropine sulfate (0.5 mg, Daewon Pharmacy, Seoul, Korea). General anesthesia was maintained by an intravenous infusion of ketamine hydrochloride (1-2 mg/kg/min). The dogs were restrained on the fluoroscopic unit (MCA-901 High-End Surgical C-arm, Medison, Seoul, Korea), and a gag was put in their mouth. After the advancement of the 0.038-inch-diameter guide wire into the stomach, a regular sheath coated with lubricating jelly was pushed into the stomach over the guide wire. The sheath was then rotated, so that the tip moved in a direction which was away from the greater curvature of the body and toward the pylorus in the stomach, in order to evaluate the torque and, finally, a wire or coil catheter was pushed toward the pylorus to evaluate the supportability of the sheath. The same experiment was performed again for the supported sheath, with and without a supporting wire in the side lumen.

The dogs were sacrificed after the second session of experiments for the purpose of evaluating possible injuries caused by the sheaths, including hemorrhage or rupture of the esophagus or stomach.

The newly designed sheaths easily changed the direction of the guide wire or coil catheter from the longitudinal direction of the esophageal portion to the transverse direction. The sheaths successfully guided the tip toward the pyloric portion, with the simultaneous advancement of the wire or coil catheter toward the duodenal portion past the pyloric portion of the phantom, for both the regular and supported types. In the experiments designed to evaluate the supportability of the sheaths when the guide wire was advanced toward the greater curvature, the distance of movement of the tip from the central point was the shortest in the case of the supported sheath combined with a supporting wire situated in the side lumen (4.47±0.32 cm). The distance of movement of the supported sheaths without a supporting wire (6.23±0.41 cm) was shorter than that of the regular sheaths (8.40±0.51 cm). The differences among the different types of sheath were statistically significant (p < 0.001). According to the data obtained by pushing a coil catheter showed that the distance of movement was the shortest in the supported sheath combined with a supporting wire in the side lumen (3.91±0.59 cm), followed by the supported sheath without a supporting wire (5.61±0.31 cm) and the regular sheath (7.22±0.70 cm). The differences among the different types of sheath were statistically significant (p < 0.001) (Table 1).

In the experiment designed to evaluate the torque, the sheaths were smoothly rotated within the stomach, regardless of the type of sheath or the insertion or not of a supporting wire, in all 12 sessions for the six dogs. When the guide wire and coil catheter were pushed through the sheath, their direction was changed from the long axis of the esophagus to the transverse axis oriented toward the pylorus in the stomach, and the shape of the sheath was well maintained without straightening. The guide wire and coil catheter were easily advanced to the duodenum beyond the pylorus, for all types of sheath with and without a supporting wire. No kinking or buckling of the guide wire or coil catheter occurred in the stomach of any of the dogs.

The experiments were well tolerated by the dogs, and no procedure related complications occurred, such as massive bleeding or perforation of the esophagus or stomach. There was no definitive evidence of rupture or gross hemorrhage in the esophagus or stomach on the pathologic specimens.