The sense of more sensitive troponin assays Prof.dr.M.P. van - - PowerPoint PPT Presentation

The sense of more sensitive troponin assays

Prof.dr.M.P. van Dieijen-Visser

2

History of cardiac markers

SGOT/ASAT (Karmen) LDH better performance CK (Rosalki) CK-MB activity (Sobel and Roberts) CK-MB mass Troponin T (Katus) Troponin I (Ladenson) Highly sensitive troponins 1954 1960 1967 1972 1986 1987 1992 Now:

Not specific Highly specific Highly sensitive Troponin 2.0

3

Spectrum of Acute Coronary Syndrome

Stable angina Unstable Angina Non-Q-wave MI Q-wave MI Non-ST elevation ACS STE- MI atherosclerosis Plaque rupture partial occlusion complete occlusion Troponin as prognostic biomarker Troponin as diagnostic biomarker

Troponin: The lower the better

4

Diagnosis of acute myocardial infarction

Clinical guidelines: Thygesen et al. Eur Heart J / JACC / Circulation 2007

Typical rise and/or fall of biomarkers, preferably troponin T or I, with at least 1 value above the 99th percentile reference limit With at least one of the following:

• Ischemic symptoms
• ECG changes (ST-T changes, pathological Q waves)
• Imaging evidence of new loss of viable myocardium
• r new regional wall motion abnormality

Analytical guidelines: Morrow et al. Clin Chem 2007

C ut-off

99%

www.bpac.org.nz

How well do we follow the guidelines?

CARMAGUE surveys 2006 and 2010 from the EFCC working group:

• More than 90% of the

laboratories use troponin as biomarker of first choice

• 70 % of the laboratories still

combine troponin with other markers

Year Number hospitals 2006 220 2010 303

CK (%) 87 (%) 59 AST 53 34 LDH 55 30 HBD 9.0 2.4 CK-MBactivity 38 8.3 CK-MBmass 31 21 Myoglobin 21 11

5

Do we use/know the right cut-off value or delta?

6 ESC guidelines Eur Heart J 2011

D?

Cut-off

7

High-sensitivity (hs) troponin assays

1990: cut-off 0.1 μg/L 1995: 0.05 μg/L 2004: 0.03 μg/L (10% CV) or <0.01 μg/L (99th percentile) 2009: 0.014 μg/L = 14 ng/L (99th percentile)

Requirements

8

Assay requirements Assay CV at 99th percentile <10% Assay CV <20% causes no misclassification

Ratio troponin at 10% CV/ troponin at 99th percentile <1

Reference population requirements Subjects included > 300 Matched for age and gender

“cardio-healthy”: negative exercise-stress test and normal cardiac function measured by “non- invasive imaging”

Apple et al Clin Chem 2009

Cardiac troponin assay scorecard

9

Acceptance designation

Total precision at 99th percentile

Guideline acceptable <10% Clinically usable >10 to £20% Not acceptable >20% Assay designation Measurable values below the 99th percentile Level 1 <50% Level 2 50 to < 75% Level 3 75 to < 95% Level 4 ≥ 95%

Website IFCC

10 Apple et al Clin Chem 2009

How consistent are these data?

11

Roche TnT : Assay performance

Assay level (% measurable data)

• Pre-commercial assay:

– > 1ng/L 97% level 4

• Commercial Assay:

– > 3 ng/L 62% level 2

12 Mingels et al. Clin Chem 2009 Jacobs et al. ACB 2009

Guideline acceptable:

TnT at 10%CV / TnT at 99th percentile: ratio 0.56

Percentage measurable concentrations

13

Samples were frozen three times !

Level 1 <50% Level 2 50 to < 75% Level 3 75 to < 95% Level 4 ≥ 95%

Level1

Apple et al. Clin Chem 58:1574-81 (2012)

Roche hs-TnT “Slight” down-shift of 6 lot numbers

14

15

Estimation of 99th percentile (URL)

Reference population requirements Subjects included > 300 Matched for age and gender

“cardio-healthy”: negative exercise-stress test and normal cardiac function measured by “non- invasive imaging”

Limitations Large cohort, difficult/expensive to fully screen, less influence outliers Small Cohort, more influence of outliers, screening easier

CARMAGUE survey 2010: Where do laboratories get their information on decision limits,

• Most laboratories use 10%

CV or 99th percentile

• >50% of labs use Data sheet,

package insert How correct are the datasheet values?

2010

10% CV 99th percentile ROC curve Other

2010

Data sheet IFCC//NACB ESC / ACC 2000 Universal definition Peer-reviewed literature Reference limits Locally derived

Clin Chem Lab Med 2009;47:227-34. Clin Chem. 2012;58:305-6.

17

Review of 44 papers, 6 methods on estimation of troponin 99th percentiles.

• 40% of the studies included less than the required 300

subjects

• 50% of the studies insufficiently reported exclusion

criteria

• No study met all criteria of “cardio-healthy” population

Cardinaels et al. CCLM 2012;50:791-806

Assays studied

18

Assay Ratio IFCC table

Roche hs-TnT

0.93 A level 2

ADVIA centaur TnI-Ultra (Siemens)

0.75 A level 1

Accu TnI (Beckman Coulter)

Now improved

1.50 U level 2

1.00 Stratus CS Acute Care cTnI (Siemens)

0.86 A level 1

Vitros Eci cTnI ES (Ortho)

1.00 A level 1

Architect cTnI (Abbott)

Now hs-TnI improved

1.14 U level 1

0.18

19

Roche hs-TnT: 99th percentile from 12 different studies

No information Minor exclusion Some exclusion Careful exclusion cardiac disease

* * * *

* Significant difference male/ female Guideline acceptable in 11 studies (ratio<1) Variation between studies 11%

Cardinaels et al. CCLM 2012;50:791-806

• ------ Datasheet value

20

Roche Hs-TnT: Significant difference male-female

Study Number subjects Male Female Ratio male/ female Mingels 2009 479 18 8 2.3 Giannitsi s 2010 616 14.5 10 1.5 Saenger 533 15.5 8.9 1.7 Koerbin 2010 104 12.9 11 1.2

Cardinaels et al. CCLM 2012;50:791-806

21

Hs-TnI ADVIA Centaur : 99th percentile from 11 different studies

No information Minor exclusion More exclusion Careful exclusion cardiac disease

* * *

Cardinaels et al. CCLM 2012;50:791-806

* = Significant difference male/ female Guideline acceptable in 7 studies.

• ------ Datasheet value

Variation between studies 44%

22

AccuTnI Access Beckman : 99th percentile 16 different studies

No selection/examination Minor exclusion More exclusion Careful exclusion cardiac disease

* *

Cardinaels et al. CCLM 2012;50:791-806

* = Significant difference male/ female Guideline acceptable in 4 studies. *

• ------ Datasheet value

Variation between studies 48%

Variation in 99th percentiles found in different studies

Assay Mean 99th % SD CV % IFCC 99th %

Roche hs-TnT

14.5 1.61 11.1 14

ADVIA centaur TnI-Ultra (Siemens)

53.6 23.4 43.6 40

Accu TnI (Beckman Coulter) *

39.7 19.0 47.9 40

Stratus CS Acute Care cTnI (Siemens)

60 20 33.3 70

Vitros Eci cTnI ES (Ortho)

25.0 10.5 42 34

Architect cTnI (Abbott)*

22.6 7.5 33.2 28

* improved methods recently introduced Large variation in values found, especially for TnI

23

Guideline acceptability of assays

24

Assay Mean Ratio 10%CV / 99th% SD Ratio IFCC table

Roche hs-TnT

0.86 0.19 0.92 A level 2

ADVIA centaur TnI-Ultra (Siemens)

1.03 0.62 0.75 A level 1

AccU TnI (Beckman Coulter) Now hs-TnI (ratio 1.0)

1.77 1.55 1.50 U level 2

Stratus CS Acute Care cTnI (Siemens)

1.24 0.63 0.86 A level 1

Vitros Eci cTnI ES (Ortho)

2.60 2.31 1.00 A level 1

Architect cTnI (Abbott) Architect hs-TnI pre- commercial

1.88 0.18 1.22 1.11 U level 1 A level 4

Large difference between mean obtained from different studies versus IFCC info

Continuous improvements: Abbott ARCHITECT pre-commercial

25

Datasheet: pre-commercial assay M 34.2 ng/L F 15.6 ng/L Overall 26.2 ng/L Ratio 10% CV/ 99th % 0.18 Koerbin et al Level 4 (98.6% measurable), guideline acceptable Ratio 10% CV/ 99th % 0.29 Significant difference male/female Overall cut-off 13.6

Koerbin et al, Clin Chem Lab Med, 2012

male female

26

D?

ESC guidelines Eur Heart J 2011

Next Challenge….

Guidelines prescribe 20% Δ within 4-6 hours as based

• n 3x SD of analytical variances

Thygesen et al. JACC / Eur Heart J / Circulation 2007

• Reference change values
• Absolute delta values obtained from ROC-analysis

Serial troponin testing and use of Δ

Reference change values (%)

Assay RCV % RCV+ rise (%) RCV- fall (%)

cTnI Beckman Coulter

45 64

• 39

Apple, 2011

cTnI Beckman Coulter

45

• 16

Vasile, 2011

hs-TnI Abbott Architect

49 54

• 35

Nordenskjold, 2013

hs-TnT Roche Modular E170

47 64

• 39

Frankenstein

hs-TnT Roche Modular E170

85

Vasile, 2011

hs-TnT Elecsys

23 26

• 21

Nordenskjold, 2013 28

29

A rise or fall ? Change values obtained from NSTEMI patients presenting at ED

for hs-cTnT D2 hours +7 ng/L Reichlin et al. Circulation 2011 D6 hours +9 ng/L Mueller et al. Clin Chem 2012

Reichlin et al. Circulation 2011

30

Conclusions

• 99th % strongly dependent on reference population
• Large cohort (>1000): lower values
• Strong exclusion criteria for cardiac disease: lower values.
• Larger variations for Troponin I.
• Difference between male and female

– TnT ~1.7 times higher in males – TnI ~ 1.2 times higher in males

• Higher values in elder population
• Uncertainty about cut-off values complicates diagnostics
• Package insert data very important
• Estimation of D cTn for each assay

Acknowledgements

• Eline Cardinaels
• Alma Mingels
• Steven Meex
• Lieke Klinkenberg
• Vincent Kleijnen
• EFCC working group on Cardiac Markers

Laitinen,P, Pulkki,K.; Suvisaari,J.; Collinson,P.; Ravkilde,J.; Stavljenic-Rukavina,A.; Hammerer- Lercher,A.; Baum,H.;

31

32

Voorbeeld van een afbeelding