Background: A microwave generator and delivery system for endoscopic use was built. Using a 650-W, 2450-MHz magnetron, 0-160 W were generated from the tip of a 180-cm flexible coaxial cable (2.1 mm diameter).
Methods: Three methods of achieving hemostasis with microwaves were identified studying standard bleeding canine ulcers: (1) interstitial method: inserting the coaxial tip into the tissue and heating slowly until bleeding stopped; (2) contact method: tip held in contact, light pressure applied; (3) noncontact method: microwave-induced sparking (dielectric breakdown) with tip held 1 mm from tissue.
Results: Studies of optimal energy levels for hemostasis showed that high power (70 W) noncontact methods required significantly less energy to stop bleeding than contact or interstitial methods. The noncontact method was more effective than the contact method, stopping 20 of 20 bleeding ulcers vs. 10 of 20 (P < 0.001) and was more rapidly effective causing less tissue damage (P < 0.05) than the interstitial method.
Conclusions: In a randomized comparison using a non-contact method, microwave coagulation was superior (P < 0.001) to a polidocanol 1% + adrenaline 1:10,000 injection and control treatment stopping 40 of 40 vs. 0 of 20 and 0 of 20 standard bleeding ulcers. Microwaves stopped bleeding from 10 of 10 severed mesenteric vessels, whereas injection was ineffective (0 of 10, P < 0.001). Microwaves look promising for hemostasis at flexible endoscopy.