STANDARDIZATION OF CARDIAC TROPONIN I: THE ONGOING INTERNATIONAL EFFORTS MILANO, ITALY NOVEMBER 30, 2011
Secondary reference materials for ctni: The way forward Jill Tate Pathology Queensland Brisbane, AUSTRALIA Joint Committee for Traceability in Laboratory Medicine
Talk Outline Talk Outline Reference measurement system for ctni Reference materials for ctni Characterisation of secondary RM Harmonisation ability Composition Commutability Stability Value assignment Value transfer to manufacturer s calibrator.
Reference System for ctni Materials SI unit: ctni (mass) nmol/l Procedures Traceability Purified reference material: NIST SRM 2921 Serumbased (commutable) reference material: panel of 3 ctnipositive pools Manufacturer s working calibrator (master lot) Manufacturer s product calibrator Primary reference method: RPLC and amino acid analysis Secondary reference method: noncommercial immunoassay with comparable antibody specificity to commercial assays Manufacturer s selected measurement immunoassay procedure (serumbased reference material as calibrator): assay manufacturer s working calibrator (master lot) and assign ctni concentration (and uncertainty) Manufacturer s standing measurement immunoassay procedure (working calibrator as calibrator): assay manufacturer s product calibrator and assign ctni concentration (and uncertainty) Routine immunoassay Patient ctni result Tate et al. Pathology 2010; 42: 4028
Reference materials for ctni Reference materials for ctni NIST SRM 2921 Purified ICT complex Wellcharacterised Value assignment by RPLC & amino acid analysis 31.2 (1.4) mg/l and traceable to S.I. Not commutable in ~50% of commercial assays Serumbased commutable material Complex matrix Sourced from AMI / ACS patients Either serum pools or a panel of individual patient sera Requires characterisation Value assignment by candidate secondary RMP.
Secondary reference material for ctni Secondary reference material for ctni Secondary calibrators can be used by IVD medical device manufacturers to calibrate their selected or standing measurement procedures for assigning quantity values to inhouse master or commercial product calibrators The Goals: 1. Three ctni concentration levels: 0.1 µg/l, 1.0 µg/l, 10 µg/l 2. 5 liters of pooled serum at each ctni concentration level to yield at least a 5year supply of the reference material 3. Value assignment and traceability to the S.I.
AACC ctni Standardisation Study AACC ctni Standardisation Study NIST SRM 2921 ternary ITC complex 1000.00 100.00 o Manufacturer s calibration SRM 2921 Not commutable in all ctni assays ctni (µg/l) 10.00 1.00 0.10 0.01 SP 1 SP 2 SP 3 SP 4 SP 5 SP 6 Serum Pools Christenson RH et al. Clin Chem 2006; 52: 168592
ctni Harmonisation by Alignment ctni Harmonisation by Alignment 1. Calibrate methods Before After 2. Measure 6 serum pools with each of the 15 ctni systems 3. Designate common comparison system 4. Regression analysis to yield parameters 5. Align each serum pool with regression parameters 6. Determine intermethod variability of aligned results for each pool Interassay CV, % 100 90 80 70 60 50 40 30 20 10 0 1 2 3 4 5 6 Sample # Christenson RH et al. Clin Chem 2006; 52: 168592
ctni assays 2006 ctni assays 2006 ctni (ug/l) 20 18 16 14 12 10 8 6 4 2 0 Centaur Dim RxL AxSYM Architect Access Vitros ECi Linear (Centaur) Linear (Dim RxL) Linear (AxSYM) Linear (Architect) Linear (Access) Linear (Vitros ECi) y1 = 1.50x1 0.14 R 2 = 0.9804 y2 = 1.03x2 + 0.01 R 2 = 0.9929 y3 = 1.06x3 + 0.01 R 2 = 0.9947 y4 = 0.75x4 + 0.15 R 2 = 0.9145 y5 = 0.81x5 0.06 R 2 = 0.9525 y6 = 0.85x6 + 0.04 R 2 = 0.9862 0 2 4 6 8 10 12 14 16 18 20 Mean ctni (ug/l) Tate J. Clin Chem Lab Med 2008; 46: 1489500
Post realignment against ctnipositive serum Normalisation factor applied All method median for MI serum ctni (ug/l) 16 14 12 10 8 6 4 2 0 Centaur Dim RxL AxSYM Architect Access Vitros ECi Linear (Centaur) Linear (Dim RxL) Linear (AxSYM) Linear (Architect) Linear (Access) Linear (Vitros ECi) y1 = 1.09x 0.10 R 2 = 0.9757 y2 = 1.01x 0.00 R 2 = 0.991 y3 = 0.97x + 0.00 R 2 = 0.9947 y4 = 0.90x + 0.15 R 2 = 0.9209 y5 = 1.02x 0.08 R 2 = 0.9537 y6 = 1.01x + 0.03 R 2 = 0.9876 0 2 4 6 8 10 12 14 16 Corrected Mean ctni (ug/l) Tate J. Clin Chem Lab Med 2008; 46: 1489500
Post realignment against all method mean (patient panel) Post mathematical recalibration All method sample means Method ctni (ug/l) 18 16 14 12 10 8 6 4 2 0 Centaur Dim RxL AxSYM Architect Access IStat Vitros y = 1.0682x 0.0659 R 2 = 0.988 y = 1.0108x 0.0065 R 2 = 0.9947 y = 1.009x + 0.0126 R 2 = 0.9941 y = 0.8274x + 0.1135 R 2 = 0.8829 y = 1.0126x 0.0057 R 2 = 0.9868 y = 0.9641x + 0.0549 R 2 = 0.9799 y = 0.9163x + 0.1656 R 2 = 0.9046 0 2 4 6 8 10 12 14 16 18 All method sample mean ctni (ug/l) Calculations by Dietmar Stöckl
Secondary reference material Issues to address Definitions measurand antibody specificity Potential biases from noncommutability ctni isoforms and nonequimolar reactivity interferences e.g. ctni autoantibodies, HetAbs instability Harmonisation capability what about traceability at low ctni concentrations?
Pi Pi Troponin I Troponin C Sites of proteolytic degradation Pi Pi Heparin Phosphorylation Other chemical modifications phosphorylation, oxidation, reduction, Nacetylation Figure kindly supplied by Dr Alexey Katrukha
Definitions ctni measurand unique, invariant part of molecule common to all components of the mixture present in serum i.e. a specific amino acid sequence Antibody specificity antibodies should preferably recognise epitopes that are located in the stable part of the ctni molecule and not be affected by complex formation (such as ICT) and other in vivo modifications Panteghini M et al. Clin Chem Lab Med 2001 Caveat all plasma ctni forms are measured by current assays either equally, or the difference in reactivity is not clinically relevant.
Potential biases between methods Potential biases between methods 80 60 % Bias from mean ctni Architect 125 100 % Bias from mean ctni istat 40 20 0 20 40 60 TE +24% 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 TE 24% 75 50 25 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 25 50 75 100 TE +24% TE 24% Analytical performance goals for ctn I measurements using routine methods based on biological variability data for ctni (CVi 9.7 %; CVg 56.8 %) obtained by Wu A et al. Total Error = bias goal + (1.96 x imprecision goal).
120 100 % Bias from mean ctni Dim RxL 50 40 % Bias from mean ctni AxSYM 80 30 TE +24% 60 40 TE +24% 20 10 20 0 0.0 20 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 0 0.0 10 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 20 40 TE 24% 30 TE 24% 60 40 40 30 20 30 40 50 % Bias from mean ctni Centaur Ultra TE +24% 20 10 0 0.0 10 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 TE 24% 50 40 30 20 30 40 50 60 % Bias from mean ctni Vitros ECi TE +24% 20 10 0 100.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 TE 24%
ctni plasma forms and assay recognition ctni plasma forms and assay recognition ctni Assay ctnic vs. free ctni vs. ctnict Phosph. ctni vs. native I in ctnict Oxidised & reduced forms Protease degraded or heparin added Literature Reference Beckman Coulter Access ctni Equimolar Equimolar UettwillgerGeiger et al Clin Chem 2002 Abbott Equimolar Equimolar Equimolar Apple et al ISTAT ctni (± 95%) (± 95%) (± 95%) CCA 2005 BioSite Triage ctni 93125% recovery for IC, IT, ITC 93125% recovery for oxid I & red I. Package insert Radiometer AQT90 FLEX ctni ctnic 1.62 fold higher vs. free ctni 1.22fold higher Degraded ctni 9.46% vs. intact ctni. Heparin (10 IU/mL) 1.24fold higher Eriksson et al Clin Chem 2005 crmp ELISA 1+1 ctni Equimolar ctnic vs. free ctni Insensitive Degraded ctni similar response to intact. Heparin insensitive Noble et al CCLM 2010
ctni plasma forms and assay recognition ctni plasma forms and assay recognition Ratio Accu:Utlra ctn 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Changes in ctni ratio Accu:Ultra over time post chest pain onset ctnic + T ctnic + ctnitc + T ctnic + ctnitc + T 0 5 10 15 20 25 30 35 40 45 50 55 Time (h) AMI Patient A AMI Patient B AMI Patient M ctnic + T Bates KJ et al. Clin Chem 2010; 56: 9528
Pools as surrogate reference material Pools as surrogate reference material 69 patients: AccuTnI Accu: Accu: Accu: Accu: Accu: 0.0520 (µg/l) Ultra Archit AxSYM Vitros Dim Mean ratio 0.66 0.99 0.95 0.93 0.96 Ratio range 0.41.2 0.52.6 0.53.0 0.71.7 0.53.0 AACC Pools AccuTnI (µg/l) Accu: Ultra Accu: Archit Accu: AxSYM Accu: Vitros Accu: Dim Accu: Stratus SP1 SP2 SP3 SP4 SP5 SP6 SP7 SP8 0.09 0.17 0.22 0.37 0.74 2.05 3.03 6.05 0.9 0.9 0.8 0.7 0.6 0.7 0.7 0.6 1.3 1.0 0.7 0.9 1.0 0.8 1.0 0.7 1.3 1.0 0.9 0.8 0.8 0.8 0.8 0.8 1.7 1.3 1.1 1.1 0.9 0.8 1.0 0.8 1.5 1.2 1.2 1.1 1.1 0.9 0.8 0.8 1.2 0.9 0.8 0.8 0.8 0.8 0.8 0.7 Mean ratio 0.74 0.93 0.89 1.09 1.07 0.84 Ratio range 0.60.9 0.71.3 0.81.3 0.81.7 0.81.5 0.70.9 AACC information from R Christenson on behalf of AACC Standardisation Study; Australian study
ctni Reference Material Development ctni Reference Material Development Aims of IFCC WGTNI 2010/2011 Pilot Study: 1. Evaluate performance of ELISAbased crmp using pooled and individual patient sera 30 each at low, medium, and high ctni concentrations vs. commercial ctni assays 2. Evaluate pooling protocols for individual patient sera using commercial ctni assays minimise need for ctnipositive sera samples by pooling ctnipositive with normal serum evaluate commutability of pools evaluate stability of pools.
ctni Pilot Study Samples ctni Pilot Study Samples Samples from 90 AMI / ACS patients ctni concentrations in range 0.0520 µg/l Collected from patients up to 72 h post presentation 30 samples per low, medium and high level 20 ml serum ( 50 ml blood) collected per patient Aliquoted within 4 h of collection & stored at 70 C Testing by NIST, NPL, Diagnostic Industry: Method comparison (crmp vs. 23 commercial ctni assays) Current status of ctni method agreement Commutability (patient samples vs. serum pools) Stability.
Preparation of Serum Pools Preparation of Serum Pools Patient pools prepared in three ways by: addition of individual ctnipositive native patient samples dilution of high ctni concentration pool with low & medium concentration pools dilution of high & medium pools with a normal pool final concentration range 0.210 µg/l Normal Pool 500 ml pool from ~510 female donors (<30 y, BMI <25, & no reported history of heart disease) prescreen for ctnaas, HAMA, HetAbs all participating labs to also screen an aliquot.
Interference Testing of Normal Pool Interference Testing of Normal Pool Serum dilution: AMI serum : Diluent 2:1, 1:1, 1:2 Hytest internal assay using mabs to central region (sensitive to ctnaas) Conclusion: Normal pooled serum suitable for IFCC study. Data kindly supplied by Dr Alexey Katrukha
Pilot Study interference testing Pilot Study interference testing Testing for ctni AutoAntibodies Add Normal pool to ctnipositive sample in stated proportions (1:3, 1:1, 3:1) and test for recovery Testing for ctni Heterophile Antibodies If Normal pool has a ctni value above assay 99 th percentile concentration proceed to test for HetAbs using heterophile blocking tube by dilution (1:1, 1:2) in assay diluent.
ctni Stability Study ctni Stability Study To test effect on ctni stability of adding a cocktail of protease/ phosphatase inhibitors crmp and commercial ctni assays To assay 2 serum pools (low and high levels) spiked with inhibitor cocktail along with unspiked pools & determine recovery Isochronous Stability Study to be done by NIST High pools stored at +5 C, 20 C, & 80 C for t ime intervals ranging from a one day to one month to be assayed on a single platform.
ctni Stability Monitoring SRM 2921 ctni Stability Monitoring SRM 2921 ctni SRM (ng/well) 4 2 1 0.5 0.1 0.0250.01 0.25 0.05 3C7 (2540) Intact ctni 560 (8393) Intact ctni Cterminal degradation products? Nterminal degradation products? Other degradation product 267 (169178) Intact ctni Nterminal degradation products? He HJ, Lowenthal MS, Cole KD, Bunk D, Wang L. Clin Chim Acta 2011;412:10711
ctni Quantification in Positive Serum ctni Quantification in Positive Serum ctni spike in (ng) Normal serum (100 µl) Positive serum (µl) + 50 100 200 400 + 0.1 + 0.25 + 0.5 + 1 + 2 + Intact ctni Positive serum SRM 2921 1. The ctni concentration in the positive serum pool is ~2.0 µg/l (1020 % error) 2. ctni in positive serum has degraded more than that in SRM 3. Immunoprecipitation coupled with fluorescent Western blot analysis can serve as the validation tool for high ctni concentration pools in the stability study. He HJ, Lowenthal MS, Cole KD, Bunk D, Wang L. Clin Chim Acta 2011;412:10711
Commutability of serum pools Commutability of serum pools A material is commutable when it gives the same result as a patient sample with the same ctni concentration Study aims to compare serum pools and patient samples by both candidate secondary RMP and routine ctni assays Determine if ctni concentrations of 7 pools fall within the prediction interval measured by two measurement procedures (one of these is crmp) for 90 native patient samples.
Commutable if same as patients Commutable if same as patients Test Method 10 8 6 4 2 Clinical Specimen RM Commutable 95% prediction interval 0 0 2 4 6 8 10 Comparison Method (1 RMP if available) Figure kindly supplied by Dr Greg Miller
Not commutable if different to patients Not commutable if different to patients Test Method 10 8 6 4 2 0 Clinical Specimen RM Commutable RM NotCommutable 0 2 4 6 8 10 Comparison Method (1 RMP if available) Figure kindly supplied by Dr Greg Miller 95% prediction interval
Commutability testing requires optimal assay performance Imprecision internal QC throughout run(s) duplicate sample measurements quadruplicate Pool (AG) measurements Linearity curvature in curves will make it difficult to assign commutability assess with Residuals plot Exclusion of outlier samples e.g. suboptimal assay performance possible isoform differences Normal pool gives an interference in an assay further testing of specific samples shows the presence of interference.
Commutability testing of ctni serum pools 1. Assay performance: precision, linearity, sample biases 2. Method comparability before and after exclusion of outlier samples 3. Determine allowable prediction interval (95% C.I., 90% C.I., clinical limits) for acceptable commutability 4. Deming regression analysis (weighted) and/or Relative Residual Analysis or Correspondence Analysis.
Value assignment of secondary RM: different models 1. A candidate secondary immunoassaybased higherorder measurement ctni procedure (ELISA) Calibrated with NIST SRM 2921 Traceable to higherorder method and to S.I. Independent method traceable to SRM 2921 Standardisation approach 2. Mathematical recalibration All method mean (or trimmed) is surrogate RMP Harmonisation approach 3. Consensus method process Use multiple commercial methods Requires traceability to SRM 2921 and commutable assays.
Value transfer to manufacturer s calibrators and to low ctni levels Compare with value assignment of CRM470 (ERM DA470/IFCC) BlirupJensen S et al. CCLM 2008 Consensus method process uses a standardised value transfer RMP consisting of dilutions of pure material (SRM gravimetric dilution) and candidate reference material Slope and yintercept within acceptable limits Imprecision within & betweenrun within acceptable limits. Value transfer to low ctni concentration using a dilution series and specified dilution protocol e.g. hs CRP CRM Linear (R 2 >0.99) & residuals plot has no bias.
Value transfer to low ctni concentration linear dilution Measured AccuTnI (ug/l) 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 y = 0.0062x + 0.0195 R 2 = 0.9988 y = 0.0171x + 0.0187 R 2 = 0.9997 0 10 20 30 40 50 60 70 80 90 100 Proportion of 100% sample Residual AccuTnI (ug/l) 0.02 0.015 0.01 0.005 0 0.005 0.01 0.015 0 10 20 30 40 50 60 70 80 90 100 Proportion of 100% Sample Patient sample & NIST SRM 2921 diluted in normal serum pool Diluted patient sample measured: 0.020 0.631 µg/l (LoD 0.01 µg/l) NIST SRM 2921 expected range: 0.010 1.00 µg/l R 2 >0.99 Residuals plot shows no discernible trend suggesting scatter of points about the line is random
Name J Tate (Chair) (AU) J Barth (UK) D Bunk (US) R Christenson (US) A Katrukha (FI) M Panteghini (IT) R Porter (UK) J Noble (UK) H Schimmel (BE) L Wang (US) I Young IFCC WG Standardisation of Troponin I WGTNI Membership Affiliation IFCC ACB NIST AACC HyTest Ltd. CIRME NPL IRMM NIST IFCC SD Liaison Labs Participating in ctni Pilot Study ABBOTT DIAGNOSTICS BECKMAN COULTER BIORAD LABORATORIES BIOMERIEUX MITSUBISHI CHEMICAL MED CO ORTHOCLINICAL DIAGNOSTICS RANDOX LABORATORIES LTD ROCHE DIAGNOSTICS GmbH SIEMENS DIAGNOSTICS NIST NPL