After 15 patients had positive cultures for Mycobacterium abscessus without evidence of infection (i.e., pseudoinfection) following endoscopy, retrospective cohort studies of patients undergoing endoscopy and microbiologic sampling of the environment were done to examine potential nosocomial transmission and to identify the source and risk factors for M. abscessus pseudoinfection. In the epidemic period, M. abscessus-positive cultures were significantly more likely to be obtained during bronchoscopy than gastroendoscopy (16/149 vs. 1/860, P < .001) and during procedures using bronchoscopes disinfected in an automated washer rather than by other methods (16/54 vs. 0/95, P < .001). M. abscessus was recovered from the automated washer, the inlet water feeding the washer, and a flexible bronchoscope. Environmental and case-patient isolates had identical large restriction fragment (LRF) patterns of genomic DNA separated by pulsed-field gel electrophoresis. Molecular typing using LRF analysis supported the epidemiologic results and demonstrate the utility of combined epidemiologic and laboratory investigations in nosocomial outbreaks of nontuberculous mycobacteria.