BACKGROUND Although expandable esophageal endoprostheses may be easier to insert and are associated with fewer procedure-related perforations, data comparing clinical results with these stents to those obtained with conventional prostheses are sparse. METHODS We reviewed the records of all patients undergoing esophageal stent placement at our institution between October 1983 and July 1995 to define relative risks, clinical results, and need for reintervention prior to death, contrasting conventional to a variety of expandable esophageal endoprostheses. RESULTS Over the period of review, 47 patients had conventional prostheses (CP) and 38 had expandable prostheses (EP) placed. Fifteen of 44 patients with CP and 14 of 38 with EP for malignancy also had esophago-airway fistulas. Insertion complications, prestent and poststent dysphagia scores, and complete fistula occlusion rates were comparable. Subacute complications were higher in the patients receiving EP (80%) than in those receiving CP (60%), possibly related to the prospective accumulation of data in patients receiving EP. Survival data were virtually identical and approximated 3 months for either group. CONCLUSIONS Although both CP and EP improve dysphagia and occlude tracheoesophageal fistulae, patients ill enough to require a prosthesis do poorly regardless of prosthesis design. Moreover, although EP may be easier to insert than CP, stent- and patient-related problems persist and may require additional intervention.