Currently, fiberoptic bronchoscopy (FB) is recommended for correct positioning of double-lumen endobronchial tubes (DLTs) because of the high incidence of malpositions not appreciated by clinical signs. The aims of this study were to assess whether clinical signs allow accurate confirmation of adequate positioning with left red rubber (RR) or polyvinyl-chloride (PVC) double-lumen tubes and to compare the incidence of malpositions between the two tubes. Another goal was to assess whether these malpositions, not appreciated by clinical assessment, adversely affected outcome. Twenty-one adult patients scheduled for elective thoracic surgery were randomly assigned to the RR (11 patients) or PVC group (10 patients). After endobronchial intubation, the position of the tubes was adjusted until clinically satisfactory lung separation had been achieved. A single investigator performed all the FB assessments were performed in the supine (SUP) and lateral positions. The anesthesiologists responsible for the clinical evaluation were "blinded" to the bronchoscopic findings. While in the SUP position, the tube was "too deep" to permit visualization of the carina during tracheal bronchoscopy in 5 patients (2 RR, 3 PVC). In 17 of 21 (10 RR, 7 PVC), the bronchial cuff could not be visualized, although in 1 patient (RR group), the cuff was overinflated and bulged out to partially obstruct the right main bronchus orifice. Bronchial bronchoscopy showed 4 of 11 patients in the RR group in whom the left upper lobe orifice was occluded compared with 1 only in the PVC group.(ABSTRACT TRUNCATED AT 250 WORDS)