We compared the effectiveness of transtracheal oxygen delivery with that of nasal cannula delivery in 10 subjects with severe hypoxemia who had previously used a portable liquid oxygen system for 20 +/- 12 months. Arterial blood gas measurements, spirometry, and an exercise test were performed prior to catheter insertion and at 2 and 6 months after insertion. With transtracheal oxygen delivery, subjects significantly (p less than 0.0001) reduced their oxygen requirement while maintaining previous exercise capacity. No change occurred in pH, PaCO2, or in spirometry. Of the 10 subjects, 4 have continued catheter use with excellent acceptance. Three experienced complications after insertion, which necessitated catheter removal. Two elected to discontinue the catheter despite subjective improvement in dyspnea. One subject died after an exacerbation of COPD. Our experience confirms findings of others that transtracheal oxygen delivery significantly reduces oxygen requirement. However, in this study, we encountered complications requiring removal of the catheter and variability in patient acceptance. Results indicate the need for additional controlled studies to identify which patients will benefit most from those potentially promising system for delivery of chronic oxygen therapy.