Polyvinyl chloride (PVC) endotracheal tube segments were exposed to a 5.0-W CO2laser beam in the presence of different fractions of oxygen and either helium or nitrogen. Time from onset of exposure until ignition was recorded, and mean time to ignition (MTI) was calculated after 10 exposures with the same gas mixture. A second series was done with 40% oxygen in either nitrogen or helium and a laser intensity of 7.5, 10.0, or 12.5 W; a third with 40% oxygen, 60% helium, and 2% halothane and a 10.0-W laser beam; and a fourth with 40% oxygen and 60% helium and a 10.0-W laser beam directed at the radioopaque barium sulfate stripe on the tube. With 5.0-W and 20% oxygen in either nitrogen or helium, segments did not ignite. With concentrations of oxygen greater than 20% in nitrogen, segments ignited sooner than with comparable concentrations in helium: MTIHe= 55.6 s and MTIN2= 27.6 s in 40% oxygen (P< 0.05). Sixty per cent helium remained protective at laser intensities up to 10.0 W (MTIHe= 42.6 svs.MTIN2= 14.3 s) (P< 0.05).However, at 12.5 W, MTIHe= 11.5 s and MTIN2= 11.3 s. Two per cent halothane in 40% oxygen and 60% helium reduced MTIHeto 25.3 s compared with 42.3 s without halothane. With the laser directed at the barium stripe, MTIHewas 7.2 s and MTIN21.1 s. Adding helium in concentrations ≥ 60% to anesthetic gases delays laser-induced PVC endotracheal tube fires, if laser intensity is ≤10.0 W and laser bursts are ≤10 s. The barium sulfate stripe should be avoided when using this technique.