Article Text
Abstract
Background Matrix-assisted laser desorption ionisation–time of flight mass spectrometry (MALDI-TOF MS) has been described as a rapid, accurate method for bacterial identification.
Aims To investigate the ability of the technique, using the unamended database supplied with the system, to identify bacteria commonly isolated in cystic fibrosis (CF) patients.
Methods Organisms commonly isolated from CF patients identified by MALDI-TOF MS were compared to conventional phenotypic and genotypic analyses. For MALDI-TOF MS, the direct colony technique was used routinely with one extraction procedure performed on a mucoid Pseudomonas aeruginosa. For 24 unique CF specimens, workload comparison and time to identification were assessed.
Results Of 464 tested isolates, conventional (phenotypic and genotypic) identification compared to MALDI-TOF MS showed complete genus, species agreement in 92%, with genus agreement in 98%. This included 29 isolates within the Burkholderia cepacia complex. All 29 were correctly identified to the genus level and 24 of these were speciated. Time to identification with 47 bacterial isolates from 24 CF patients showed identification of 85% of isolates by MALDI-TOF MS at 48 h of incubation, compared to only 34% with conventional methods.
Conclusions Using the unamended database supplied with the system, MALDI-TOF MS provides rapid and reliable identification of bacteria isolated from CF specimens. Time to identification studies showed that the use of same day, same method for organism identification will decrease time to result and optimise microbiology workflow.
- MALDI-TOF MS
- cystic fibrosis
- bacterial identification
- microbiology
- microbial genetics
- laboratory tests
- clinical infectious diseases
- antimicrobial resistance
- Helicobacter pylori
- MRSA
Statistics from Altmetric.com
- MALDI-TOF MS
- cystic fibrosis
- bacterial identification
- microbiology
- microbial genetics
- laboratory tests
- clinical infectious diseases
- antimicrobial resistance
- Helicobacter pylori
- MRSA
Introduction
Progress in rapid diagnostics has seen an unparalleled growth with the development of innovative technologies. Molecular diagnostic testing has evolved to the point where PCR amplification testing for certain organisms can be performed in a routine clinical laboratory 24 hours a day, 7 days a week.1 Additionally, the development of 16S rRNA gene sequencing has led to a new ‘gold standard’ for identifying bacteria.2 However, it is still more expensive than most traditional identification methods and is limited to use in large research and reference laboratories.
One of the latest methods for bacterial identification is matrix-assisted laser desorption ionisation–time of flight mass spectrometry (MALDI-TOF MS).3 This method involves the use of mass spectrometry to generate a unique peptidic spectrum for each isolate tested. This spectrum is compared to an internal database for bacterial identification. This technique has allowed for the rapid identification of bacteria and mycobacteria, as well as yeast.4–9 MALDI-TOF MS has proven to be cost-effective and accurate for identification of microbial organisms, obviating the need for routine biochemical methods.10
The purpose of this study was to examine the use of the MALDI-TOF MS in the clinical microbiology laboratory to identify the group of organisms commonly isolated from patients with cystic fibrosis (CF). These patients are prone to develop infections with difficult to identify non-fermenting Gram-negative rods including Burkholderia cepacia complex (BCC). We sought to identify species in the BCC, compared to gene sequencing. To the best of our knowledge, this is the first publication looking at BCC isolates with the unabridged database.
Materials and methods
Microbial strains
Organisms were obtained from routine clinical specimens cultivated on a variety of different media (Remel, Lenexa, Kansas, USA) including: trypticase soy blood agar with 5% sheep blood (TSA); chocolate agar (CHOC); MacConkey agar (MAC); mannitol salt agar (MS); and B cepacia selective agar (BCSA). Plates were incubated in ambient air or 5% CO2 at 35–37°C and examined daily until all isolates were identified. Phenotypic identification of organisms conformed to routine microbiological procedures (Clinical Microbiology Procedures Handbook, ASM Press, 2010). The subset of non-fermenting Gram-negative rods was identified according to manufacturer's instructions for MicroScan (Siemens, Newark, Delaware, USA) and bioMérieux API 20 NE (Durham, North Carolina, USA).
Reference method for Burkholderia spp.
Definitive identification of Burkholderia spp. was obtained from genetic analyses performed at the B cepacia Research Laboratory and Repository (University of Michigan). The methods used a combination of species-specific 16S ribosomal DNA PCR assays, recA-RFLP analysis and recA sequence analysis as previously described.11–18
MALDI-TOF MS
MALDI-TOF MS measurements were performed on a Microflex LT (Bruker Daltonics, Billerica, Massachusetts, USA) with a 60-Hz nitrogen laser and spectra analysed over a mass range of 2000–20 000 Da. Each series of measurements was preceded by weekly calibration with a bacterial test standard (BTS 255343; Bruker Daltonics) and daily QC organisms. All specimens were processed as per manufacturer's guidelines.
Routine testing
The tops of individual colonies were transferred to 1–3 wells on a steel template carrier (96-well target). After air drying, 1 μl of matrix solution was added to each well. The matrix was a saturated solution of α-cyano-4-hydroxycinnamic acid dissolved in a basic organic solvent composed of 50% acetonitrile and 2.5% trifluoroacetic acid. Spots were measured with the MBT FC.par flexControl method and the MBT-autoX.exe autoXecute method. All measurements resulted from six series of 40 laser shots at multiple positions on the spotted target. Spectra were analysed with the MALDI BioTyper database version 3.0.2, containing spectra for 3995 organisms, including various bacterial and fungal species. Spectra were scored with a logarithmic value resulting from the alignment of the unknown raw spectrum and the best matching database spectrum. A high log score of >2 was required for species level identification and an intermediate log score lying between ≤2 and ≥1.7 was required for genus level identification. A low score of <1.7 was considered unreliable for identification.
Extraction procedure
An extraction procedure was performed for the single organism which demonstrated no peaks on initial testing. Briefly, approximately three colonies were scraped from the agar and added to 300 μl distilled water, followed by the addition of 900 μl of 100% ethanol. The suspension was centrifuged twice at 13 000 rpm for 2 min (Eppendorf Centrifuge model 5425) and the supernatant was removed. Caps were left open for 10 min to allow the pellet to dry. The pellet was suspended in 50 μl of formic acid (70%) and incubated for 15 min at room temperature followed by the addition of 50 μl of 100% acetonitrile. The mixture was centrifuged for 2 min at 13 000 rpm and 1 μl of the supernatant was placed on each of two wells on the target. After air drying, each well was overlaid with matrix and subjected to analysis as above.
Workflow analysis
We executed hand-motion time studies documenting steps for individual phenotypic tests performed on each isolate until full identification was achieved. Twenty-four respiratory specimens submitted from CF patients were evaluated for turn-around time (TAT) and compared to MALDI-TOF MS TAT.
Results
A total of 464 isolates representing 17 bacterial species were evaluated. Conventional identification compared to MALDI-TOF MS showed complete genus, species agreement in 92% of tested isolates, with genus agreement in 98%. The remaining 2% were either organisms with known limitations of the MALDI-TOF MS database or discrepant results (table 1).
In summary, 100% of 124 isolates of Pseudomonas aeruginosa, 189 Staphylococcus aureus, and 25 Haemophilus influenzae matched to the species level. Of two isolates identified by conventional methods as Ralstonia pickettii, one showed complete agreement. The other was identified by MALDI-TOF MS and confirmed with 16S rRNA sequencing as Herbaspirillum huttiense. For 58 Stenotrophomonas maltophilia isolates, 84% showed complete genus, species agreement. The remaining 16% were identified as Pseudomonas hibiscola (n=6), Pseudomonas geniculata (n=2) and Pseudomonas betelei (n=1). Only one organism (a mucoid Pseudomonas aeruginosa) required the additional extraction procedure.
MALDI-TOF MS provided accurate identification for all tested isolates of the BCC to the genus level, with 24 of 29 (83%) being correctly speciated. Species included Burkholderia cenocepacia, Burkholderia cepacia, Burkholderia multivorans and Burkholderia vietnamiensis. In addition, one isolate of Burkholderia gladioli (not a member of the BCC) was accurately identified.
Among the 15 tested isolates of Acinetobacter spp., 53% were accurately identified to the species level as Acinetobacter baumannii, with the remaining 47% identified to the genus level.
Workflow analysis was performed on CF respiratory specimens to compare the hands on time, complexity, and time to result between the routine culture and biochemical methods and the MALDI-TOF MS method. A total of 47 bacterial isolates obtained from 24 patients were examined. Considering the additional time needed for sub-culturing, special stains, spot tests, quality control, and constant motion for various subculture to additional plates, MALDI-TOF MS was shown to require less technical time, was of lower complexity, and resulted in a faster time to organism identification (figure 1).
The simple, single MALDI-TOF procedure replaced 48 subcultures, 20 cetrimide plates, 25 oxidase spot tests, 19 staphylococcus agglutination tests, five MicroScan identifications, and one streptococcus agglutination test. As shown in figure 1, among total isolates, 85% of isolates were identified by MALDI-TOF MS at 48 h of incubation, and all were identified at 5 days. Using the conventional method for organism identification, only 34% were identified at 48 h, with the remainder taking up to 7 days.
From the 24 specimens, 21 isolates of P aeruginosa, 19 isolates of S aureus, 3 isolates of S maltophilia, and single isolates of Achromobacter xylosoxidans, B cepacia, group A streptococcus and Klebsiella oxytoca were identified. Seventy-one per cent of P aeruginosa isolates were identified by MALDI-TOF MS by day 2 of incubation, compared to none by conventional methods. Only 62% had been identified by conventional methods by day 4 of incubation. Eighty-nine per cent of S aureus isolates were identified by day 1 of incubation by MALDI-TOF MS, 100% were identified by day 2. Comparatively, only 68% were identified by conventional methods by day 2. MALDI-TOF MS was able to identify all three isolates of S maltophilia by day 2 of incubation, whereas identification was delayed until day 4 by conventional methods.
Discussion
Using the unamended database for MALDI-TOF MS compared to routine conventional methodology for commonly identified pathogens from CF patients, we were able to show accurate results, with faster time to identification.
We demonstrated complete genus, species agreement with MALDI-TOF MS compared to conventional identification in 92% of isolates, with genus agreement to 98%. One hundred per cent of our tested isolates of P aeruginosa, S aureus and H influenzae matched to the species level. Among CF workflow isolates, 85% of organisms were identified by MALDI-TOF MS at 48 h of incubation, compared to only 34% by conventional methods. This is consistent with other publications showing the improvement in cost and time to detection in using MALDI-TOF MS for routine bacterial identification.10
Others have reported the use of the MALDI-TOF MS in identification of various organisms, including yeast and mycobacteria.4–9 Degand et al used MALDI-TOF MS for identification of organisms commonly isolated from CF patients, using a modified database to include additional bacterial isolates commonly isolated from CF patients.19 They reported the improved identification of CF organisms when the database was supplemented with CF-specific additional reference strains. Our study is the first that we know of to identify CF organisms using the standard database with no modification.
Regarding Burkholderia spp., definitive identification was achieved for B cenocepacia, B cepacia, B multivorans, B vietnamiensis and B gladioli. This was compared to the definitive identification obtained from genetic analyses performed at the B cepacia Research Laboratory and Repository (University of Michigan) as described in Methods. These findings are significant in that current phenotypic ‘kit methods’ when accurate, only identify this class of organisms as BCC. Accurate distinction between BCC and other Gram-negative non-fermenting rods is critical for appropriate management of CF patients. These species have been correlated with increased morbidity and mortality, may represent a contraindication to lung transplant, and require attention to infection control to prevent spread within both the hospital and the community.20 ,21
Several noteworthy findings emerged from testing these organisms. A xylosoxidans by phenotypic kit assay was commonly identified as Achromobacter ruhlandii by MALDI-TOF MS. In addition, the Acinetobacter complex of organisms identifying as A baumannii or Acinetobacter lwoffii by MicroScan or API NE were often identified as Acinetobacter genomospecies or A ursingii, respectively by MALDI-TOF MS. As the taxonomy of both Achromobacter and Acinetobacter spp. continues to be updated, modifications in the databases of both MALDI-TOF MS and phenotypic systems will be warranted.
S maltophilia identified by biochemical parameters was commonly called P hibiscola, P geniculata or P betelei by MALDI-TOF MS. These Pseudomonas species show high levels of homology with S maltophilia and are considered by some to be heterotypic synonyms.22 ,23
A limitation of our study is the lack of sequencing confirmation for all glucose non-fermenters, apart from the Burkholderia species and one H huttiense. MALDI-TOF MS accurately identified H huttiense, while both API NE and MicroScan identified the organism as R pickettii. This is significant in that this organism is uncommonly isolated from CF patients, often initially identified as a member of the BCC, and is under scrutiny as a potentially significant pathogen.24
Conclusion
MALDI-TOF MS is a fast, reliable tool that can be used in a paediatric clinical microbiology laboratory. The database supplied by the company was used with no modification to assess its ability to identify organisms commonly isolated in the CF patient population. This is in contrast to prior studies, which used a supplemented database for identification of hard to identify bacteria. MALDI-TOF MS allows for faster identification of organisms that historically have been thought of as hard-to-identify, time-consuming and costly. This study demonstrates that the use of MALDI-TOF MS can replace routine methods of identification for organisms cultured from CF patients. Adoption of this technique will allow for the earlier report of culture results to the clinical care team, and ultimately lead to improved patient outcomes.
Take-home messages
Matrix-assisted laser desorption ionisation–time of flight mass spectrometry can replace routine methods of identification for difficult to identify organisms cultured from cystic fibrosis patients, with faster time to identification.
The unamended database supplied with the system was used with no modification to assess its ability to identify organisms commonly isolated in the cystic fibrosis patient population.
References
Footnotes
A preliminary report of this work was presented at the 110th General Meeting of the American Society for Microbiology in May 2010 in San Diego, California.
Funding This work was supported by a grant from Bruker Daltonics, Billerica, Massachusetts, USA. The company supplied the MALDI-TOF MS instrument, reagent support and funds for technical support. They were not involved in experimental design, data analysis or manuscript preparation.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.