The timing of tracheotomy. An evolving consensus.

pulmonary complications of the acquired immunodeficiency syndrome. lavage in the diagnosis of diffuse pulmonary infiltrates in the immunosuppressed host. of transbronchial and open lung biopsies in chronic infiltrative lung diseases. phantom for quantitative CT analysis of pulmonary nodules. DJ, Hudson LD. Acute lobar atelaectasis: a prospective comparison of fiberoptic bronchoscopy and respiratory therapy. I '' 1807, Napoleon Bonaparte's great-nephew died of diphtheria, prompting the Emperor to offer a prize for research to combat that illness, thus stimulating work on airway management. ' Tracheotomy was identified as useful therapy and became the standard of care. When Bouchut suggested translaryngeal intu-bation as a better alternative, he was condemned by an investigative committee of the Paris Academy. However, an ocean away, O'Dwyer in New York routinely maintained tracheal intubation for diphtheria for up to two weeks. Thus, not only have both techniques been used for more than a century and a half, but clinicians have argued their relative merits for a century. Because of the complexity and severity of disease in critically ill patients, definitive studies of airway management have been hard to obtain and controversies therefore difficult to resolve. Instead, the development of a consensus has been an evolutionary process with conclusions based on the preponderance of evidence from a series of imperfect clinical and animal studies. The article by Colice and colleagues in this issue (see page 877) adds further weight to the evolving consensus that tracheotomy is not routinely medically indicated in the patient requiring a temporary artificial airway. Although convictions have been strong, evidence has often been lacking as to the need for converting a translaryngeal intubation to a tracheotomy. A survey of critical care practice revealed that while most clinicians felt that a tracheotomy should be performed after a fixed period of translaryngeal intubation, the " ideal " time varied between one and four weeks with marked differences depending on primary specialty.2 Colice et al confirm that translaryngeal intubation routinely causes laryngeal injury. Pressures at the interface of the tube with the mucosa covering the posterior laryngeal cartilages, the arytenoids and the cricoid, far exceed mucosal perfusion pressures.5 It is not surprising, therefore, that the authors found a 94 percent incidence oftypical mucosal ulcerations along the posterior cords. The article by Colice and coauthors reaches several important conclusions. The lack ofcorrelation between severity of injury as seen at the time of extubation and long-term complications argues against the …