PCR-restriction endonuclease analysis (PRA) was used for direct identification of Mycobacterium haemophilum in clinical specimens from immunocompromised patients. PRA correctly identified M. haemophilum in four smear-positive specimens. Direct identification by PRA takes 2 to 3 working days compared to the 3 to 5 weeks required for culture isolation and identification by conventional methods.In recent years, Mycobacterium haemophilum has emerged as an important human pathogen (6, 10, 12), causing mainly opportunistic infections in severely immunocompromised patients with AIDS and those receiving immunosuppressive therapy after transplantation (1,4,7,8,12,16). M. haemophilum has also been isolated from localized lesions in immunocompetent pediatric patients with cervical lymphadenopathy (3, 9). Superficial lesions such as cutaneous lesions and multiple skin nodules are not uncommon clinical presentations of M. haemophilum infection (6, 10, 12). Though necessary, culture isolation and identification of mycobacteria by conventional methods, especially for M. haemophilum, are time-consuming and laborious, usually taking 3 to 5 weeks (5). PCR-restriction endonuclease analysis (PRA) of an amplified 439-bp segment of the hsp65 gene encoding the 65-kDa heat shock protein has been successfully used for the rapid identification of mycobacterial isolates to the species level and has gained wide acceptance (11,13,14). This study is aimed at applying PRA as a means of rapid identification of M. haemophilum directly from acid-fast bacillus (AFB) smear-positive skin lesion specimens from immunocompromised patients.Organisms. We used four superficial lesion specimens from three immunocompromised patients with suspected nontuberculous mycobacterial infections (two skin biopsy specimens from erythematous nodules, one skin abscess specimen from a patient's left foot, and one pus drainage specimen from a patient with skin bursitis). Reference strains M. haemophilum ATCC 29548, M. intracellulare ATCC 13950, M. tuberculosis ATCC 27294, and M. kansasii ATCC 35775 were utilized as controls for PRA.Specimen processing. Specimens (1 g of skin biopsy specimen and 0.3 to 0.6 ml of pus and abscess drainage specimens) were digested and decontaminated using MycoPrep (Becton Dickinson). The derived sediment was used for smears and cultures. An aliquot of the original specimen was kept at Ϫ70°C for molecular analysis. AFB microscopy. Auramine O fluorescent stain was used, and positive smears were counterstained with Ziehl-Neelsen stain to confirm the presence of AFB (5).Mycobacterial cultures. Each specimen was inoculated into two sets of BACTEC 460 12B vials (Becton Dickinson) and Lö-wenstein-Jensen (LJ) medium slants (BBL, Becton Dickinson), and each set was incubated at either 30 or 37°C. In addition, one blood agar plate was inoculated and incubated at 30°C. Broth and solid medium cultures were held for 6 and 8 weeks, respectively, and examined periodically for growth. The culture isolates were identified using DNA probes (AccuProbe; Gen-Pr...