MALDI-TOF Mass Spectrometry Evaluations of performance and options for workflow improvements
Barbara M. Willey Mount Sinai Hospital/UHN, Toronto
VITEK USER GROUP Toronto 3 June 2014
What’s up for MALDI today……. • Reproducibility studies – VITEK MS PLUS, Bruker BioTyper
• Prospective evaluations – VITEK MS PLUS, Bruker BioTyper, Conventional ID/16S
• Retrospective evaluations – Candida spp. – Streptococcus pneumoniae from closely related spp.
• Simple workflow solutions – Latex agglutination for Escherichia coli vs. Shigella – Chromogenic agars for Escherichia coli from urines
• Blood cultures – Algorithm options for MS testing
• Mixed bacteria/yeast challenges
+
– MS from short incubation poly-microbial blood cultures
• Still to be done…. Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Reproducibility studies (1) STUDY DESIGN • Bruker BioTyper vs. VITEK MS PLUS • Strains: 6 ATCC, 4 clinical wild-type – GNB: E. coli, P. aeruginosa – GPC: S. aureus, E. faecium/faecalis • Media: 11 common agars (Oxoid) – 5%BA, CNA, CHOC, MAC, ESBL, MHA, Brucella – Chromogenic Staph, Denim Blue, Brilliance VRE, Brilliance CRE • Incubation conditions as appropriate: 37oC in O2, CO2, AnO2 • MS testing at: 4h, 18h, 24h, 36h T=250 tests – exact same colony picks tested in parallel Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Reproducibility studies (2) RESULTS • VITEK MS: 100% correct ID with >98% confidence – 247 (98.9%) with 99.9% conf – 2 with 99.7% conf; 1 with 98.7% conf
• Bruker: 87.6% correct ID scores >1.9 (100% GNB, 89% Entero, 74.4% SA) – Correct ID but low scores (1.7-1.9): 8.8% – Unreliable scores (1.7 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Retrospective - Candida (2) RESULTS While both MS equally able to ID common Candida as 100% ID were correct, the MS systems varied in: • Numbers of repeat tests required to obtain ID: – SAB 16h: VITEK MS 2.5% vs. Bruker 1.2% – BA 22h: VITEK MS 2.3% vs. Bruker 20.1% (P1.9 (with scores >1.7: 88%) • 96.7% vs. 40.2% P98% (range 98.9-99.9%; 94.2% strains=99.9%) • S. pseudopneumoniae: 60 (100%) - all NT/rpoB-, with 59/60 GP-pos! (FP) • S. mitis/oralis: 26 (100%) – all BS/opt=V; NT/rpoB/GP-; 12 pneumolysin+) – 16 conf >99%, 6 conf 93-89.9%, and 4 with conf 68.9-78-8%
• Uncertain spp. ID (NT/rpoB-): n=3; VITEK MS = SPN with low conf (64-79%) – below 98% threshold > would be rejected – >> not ID errors by MSH/UHN criteria
CONCLUSIONS • Highly accurate ID - 100% (95%CI: 96.7-100%) S. pneumoniae • Overall ability to ID isolates: 98.7% (95%CI 96-99.7%) • Correctly discriminated S. pseudopneumoniae – all GP False+ • Effectively excluded 100% non-SPN, even those of uncertain ID Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Simple solutions for reducing workload: No MALDI-TOF for Urine E. coli (1) STUDY DESIGN Urine Shigella extremely rare - 0 in 12y review • Burgundy pink colonies on chromogenic agar reputedly 100%-specific for E. coli – Evaluated 3 brands x 5 lots each • Colorex Orientation (Alere) • CPS4 ChromID (bioMerieux) • Brilliance UTI Clarity (Oxoid)
• 2500 prospective urines • 300 retrospective isolates x 5 = ~1500 tests per agar Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Simple solutions for reducing workload:
No MALDI-TOF for Urine E. coli (1) RESULTS Burgundy pink colonies =100% specific for E. coli both prospective and retrospective studies
Agars verified: CPS4 ChromID (bioMerieux), Brilliance UTI Clarity (Oxoid), Colorex Orientation (Alere)
CONCLUSIONS No further testing for burgundy pink colonies identified from urine cultures – report as E. coli No MS needed – improves workflow - reduces workload –reduces cost Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Simple solution for MALDI-TOF limitation of Shigella = E. coli (1) STUDY DESIGN • Evaluated the Wellcolex Colour Shigella kit (Oxoid) to prevent misidentification of Shigella as E. coli without incurring any reporting delays – Simple 2 min latex agglutination test – Provides species-level ID of 4 Shigella spp. • Only need to test NLF with MS ID of E. coli – No agglutination with E. coli • Prospective blood/sterile site NLF isolates – including from short-incubation subcultures
• Retrospective blood and enteric isolates of confirmed Shigella
T=350 tests
Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Simple solution for MALDI-TOF limitation of Shigella = E. coli (2) RESULTS • Agglutinated all 88 Shigella, accurately assigning species-level ID – Overall sensitivity: 100% (95% CI: 95-100%) • 260/262 (99.2%) NLF E. coli – no agglutination – 2/262 (0.8%) stringy agglutination S. flexneri latex particles
– Overall specificity: 99.3% (95% CI: 97.1-99.97%) • specificity for latex other than S. flexneri = 100%
CONCLUSIONS The Wellcolex Colour Shigella latex kit rapidly and reliably ruled out Shigella in NLF isolates ID by MALDI-TOF MS as E. coli Well-designed, easily integrated into blood bench workflow Replaces more laborious and costly Shigella anti-sera Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Blood Culture MS Algorithms (1) Three possible MS testing options available: 1) Direct from each positive blood bottle as it becomes positive 2) From short-incubation agar cultures (~3-6h) 3) From visibly pure colonies after overnight incubation
• Option 1: Testing each bottle as it becomes positive – Sample prep requires processing +BC broth – Not yet automated -pipetting, centrifugation steps etc, finally pellets to MS (~15 min hands-on/spec) – Presently ~70-85% success at best – Labour-intensive, erratic workflow, costly – Most labs land up batch-processing then running MS every 3-4h – Quickest if successful (moderate), high rpt rate (esp. for GP) *Perez, K. Journal of Infection 2013, update 2014 (Bruker GNB only; batch-testing) Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Blood Culture MS Algorithms (2) Three MS testing options available: 1) Direct from each positive blood bottle as it becomes positive 2) From short-incubation agar cultures (~3-6h) 3) From visibly pure colonies after overnight incubation
• Option 2: MS from “short incubation cultures” (typically 3-6h) – Gram strain, incubate plates as usual – Plates examined every 3-6h • MS at first sign of visible growth
Smooth work-flow as high success rate/few repeats and can proceed directly to AST including rapid tests - PBP2a, BLACTA, CARBA, vanA/B etc Same day ID, no extra labour, no added costs, best fit for large labs – BUT full benefit to patients only realized if MS results are available 24h/7d • Must also be acted on right away by physicians/pharmacists to maximize pt benefit
– 24h/7d VERY hard on techs - but to be effective for patients, lab needs move to night shifts • Initial compromise – day hours “short-incubation”, then overnight for rest Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Blood Culture MS Algorithms (3) Three MS testing options available: 1) Direct from each positive blood bottle as it becomes positive 2) From short-incubation agar cultures (~3-6h) 3) From visibly pure colonies after overnight incubation
• Option 3: MS from colonies after overnight incubation – Loss of benefit to patients – well-documented that timely ID/change to appropriate therapy save lives (lots of lives!) – But MS at 24h is still an improvement on traditional non-MALDI-TOF identification algorithms
However, Option 2 leads to questions regarding testing “blind” as ~15% cases of positive bloods are poly-microbial So, does MALDIMALDI-TOF MS from films of early growth lead to ID errors when cultures are mixed? Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Highly accurate MALDI-TOF ID using VITEK MS PLUS when challenged with intentional mixtures prepared from bacteria and yeast colonies
B.M. Willey, E. Ramoutar , P. Lo, S. M. Poutanen Mount Sinai Hospital/University Health Network Toronto, Ontario, Canada
O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Background & Objectives • Identification (ID) from shortshort-incubation cultures (3(3-6h) is a strategy to provide rapid TAT for organism ID from positive blood cultures (BC) • Since positive BC may be polymicrobial, the accuracy of MALDIMALDI-TOF mass spectrometry ID in this setting has been questioned • By intentionally subjecting mixed organisms to MALDIMALDI-TOF MS, the objective of this study was to challenge the bioMé bioMérieux VITEK MS PLUS (VMS) to produce accurate ID regardless of culture purity O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Methods • 31 ATCC strains selected for genus/species diversity challenged VITEK MS PLUS to produce correct ID 376 (87%) mixed bacteria tests • 243 = 2 species mixes • 133 = 3 species mixes – No extraction
56 (13%) mixed yeast tests • All = 2 species mixes – OnOn-slide extraction
• T=432 mixtures O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
ATCC Bacterial species n=24 A. aphrophilus C. jeikeium
K. pneumoniae L. mesenteroides
S. aureus S. epidermidis
C. spp not jeikeium E. faecalis
M. catarrhalis N. gonorrhoeae
S. lugdunensis S. saprophyticus
E. faecium E. gallinarum
N. meningitidis P. mirabilis
S. agalactiae S. equi ssp. equi
E. coli H. influenzae
P. vulgaris P. aeruginosa
S. pneumoniae S. pyogenes
O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
ATCC Candida species n=7 C. albicans C. glabrata C. guillerimondii C. krusei C. lusitaniae C. parapsilosis C. tropicalis O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Test Procedure
Step1: Pick well isolated colonies from 2 or 3 fresh ATCC cultures
Step 2: Blend colonies on moist surface of BA plate
Step 3: Apply the smooth mixture to VITEK MS PLUS slide spot
Step 4: Place in VITEK MS PLUS for MALDI-TOF
•Mixes prepared in predetermined approximated proportions • 2-species mixtures: 1:1, 2:1, or 1:2 • 3-species mixtures: 1:1:1, 1:2:3, 3:2:1, etc O188 ECCMID Barcelona 12 May 2014 Barbara Willey, Mount Sinai Hospital/University Health Network
VITEK USER GROUP MEETING Toronto 2 June 2014
Results - Bacterial mixes Of 376 bacterial mixtures, VITEK MS PLUS produced: – 31 (8%) = “Bad Spectrum During Analysis” Analysis” with no ID
Of 345 bacterial mixtures with ID: • 340 (98.6%; 95% CI: 96.696.6-99.5%) = High confidence ID (>98%) – 293 (84.9%; 95% CI: 80.680.6-88.3%) = 11-species correct ID • 186 from 22-species mixes • 107 from 33-species mixes
– 47 (13.6%; 95% CI: 10.410.4-17.7%) = 22-species correct ID • 38 from 22-species mixes • 9 from 33-species mixes
– 5 (1.5%; 95% CI: 0.50.5-3.5%) = Low confidence ID (98%) • 39 = 11-species correct ID • 10 = 22-species correct ID – all from 22-species mixes – 1 (2%; 95%CI:
