Over a 15-month period, 74 trauma patients who were expected to require extended intubation were studied prospectively to evaluate the appropriateness of tracheostomy. Patients were randomized to receive either early (34) or late (40) tracheostomies. The patients also were grouped to determine the difference of early versus late tracheostomy on the development of laryngotracheal pathology and respiratory infections; length of intubation and type of patient injury were studied as possible differential factors. Fifteen per cent (11/74) of the patients developed major laryngotracheal pathology as identified by endoscopy, and respiratory infections developed in 54% (40/74), but there was no significant difference in the complication incidence between the early and late tracheostomy groups. Significantly more complications occurred in rigid-posture, head-injured patients than in any other trauma grouping, but there was no significant difference in the complication incidence between the two tracheostomy groups within that classification. We conclude that patients can undergo translaryngeal intubation for up to 2 weeks without significantly increasing complications relative to transtracheal intubation.