Mycobacterium xenopi typically accounts for less than 0.3% of all clinical mycobacterial isolates. Over a 37-mo period, 21 (35%) of 60 mycobacterial isolates from a Michigan hospital were identified as M. xenopi. Hospital, laboratory, and bronchoscopy records were reviewed to determine case characteristics, develop a case series, and calculate procedure-specific M. xenopi isolation rates. A case-control study was conducted to elucidate aspects of the bronchoscopy procedure associated with M. xenopi isolation. Bronchoscope cleaning procedures were reviewed, and hospital water systems were cultured. Four isolates were from three patients with disease attributable to M. xenopi. Of the other isolates, specimens obtained by bronchoscopy were more likely to yield M. xenopi than were specimens obtained by other routes (relative risk, 9.7; 95% confidence intervals, 3.2, 29.6). Bronchoscopes were disinfected in a 0.13% glutaraldehyde-phenate and tap-water bath and then were rinsed in tap water. Water from the hot water tank supplying this area yielded M. xenopi. Mycobacteria were cultured from bronchoscopes after disinfection. M. xenopi in the tap water appears to have contaminated the bronchoscopes during cleaning. Adequate disinfection of contaminated bronchoscopes and careful collection of specimens to avoid contamination with contaminated water are essential, both for limiting diagnostic confusion caused by mycobacterial pseudoinfections and for reducing risks of disease transmission.