Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has proved to be a useful diagnostic method for identifying conventional bacteria. In the case of mycobacteria, a good protein extraction protocol is essential in order to obtain reliable identification results. To date, no such protocol has been definitively established. The aim of this study was to compare the manufacturer's recommended protein extraction protocol (protocol A) with two novel protocols (protocols B and C), which apply different freezing temperatures and mechanical disruption times using an automatic tissue homogenizer. A total of 302 clinical isolates, comprising 41 nontuberculous mycobacteria (NTM) species, were grown in parallel on solid and liquid media and analyzed: 174 isolates were slow-growing mycobacteria (SGM) and 128 isolates were rapid-growing mycobacteria (RGM). Overall, MALDI-TOF MS identified a higher number of NTM isolates from solid than from liquid media, especially with protocol C (83.4 and 68.2%, respectively; < 0.05). From solid media, this protein extraction method identified 57.9 and 3.9% more isolates than protocols A ( < 0.001) and B ( < 0.05), respectively. In the case of liquid media, protocol C identified 49.7 and 6.3% more isolates than protocols A and B, respectively ( < 0.001). With regard to the growth rate, MALDI-TOF MS identified more RGM isolates than SGM isolates in all of the protocols studied. In conclusion, the application of freezing and automatic tissue homogenizer improved protein extraction of NTM and boosted identification rates. Consequently, MALDI-TOF MS, which is a cheap and simple method, could be a helpful tool for identifying NTM species in clinical laboratories.

Conventional identification of mycobacteria species is slow, laborious and has low discriminatory power. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has proved highly effective for identifying conventional bacteria, and it may also be useful for identifying mycobacteria. The aim of this study was to evaluate and compare MALDI-TOF MS with currently recommended molecular methods for the identification of nontuberculous mycobacteria (NTM), applying Mycobacteria Libraries v3.0 (ML3.0) and v2.0 (ML2.0). A total of 240 clinical isolates of 41 NTM species grown on solid media were analysed: 132 isolates of slow-growing mycobacteria and 108 of rapid-growing mycobacteria. MALDI-TOF MS, using ML3.0, identified 192 (80%) NTM isolates with a score ≥1.7, encompassing 35 (85.4%) different species, that is, 17 (7.1%; p = 0.0863) isolates and 15 (36.6%; p = 0.0339) species more than currently recommended molecular techniques (polymerase chain reaction reverse hybridization). All these isolates were correctly identified according to molecular identification methods. The application of ML3.0 also identified 15 (6.2%) NTM isolates more than ML2.0 (p < 0.01). The scores obtained with MALDI-TOF MS using ML3.0 (mean score: 1.960) were higher in 147 (61.2%) isolates than when using ML2.0 (mean score: 1.797; p < 0.01). Three of the species analysed were not included in either database, so they were not recognized by this system. In conclusion, MALDI-TOF MS identified more isolates and species than the recommended polymerase chain reaction reverse hybridization assays. Although the new ML3.0 is not the definitive database, it yielded better results than ML2.0. This shows that the updating of the MALDI-TOF MS database plays an essential role in mycobacterial identification. Copyright © 2017 John Wiley & Sons, Ltd.