1. Brigham and Women’s Hospital
High-sensitivity Cardiac Troponin at Any Detectable Concentration Identifies Higher Risk of Major Cardiovascular Events in Patients with Stable Ischemic Heart Disease
Nicholas Marston, MD
TIMI Study Group
Brigham and Women’s Hospital
Nicholas A. Marston, Marc P. Bonaca, Petr Jarolim, Deepak L. Bhatt, Philippe Gabriel Steg, Marc Cohen, Robert F. Storey, Per Johanson, Erica Goodrich, Steve D. Wiviott, Eugene Braunwald, Marc S. Sabatine, David A. Morrow

2. Disclosures Presenter Disclosure: None
PEGASUS-TIMI 54 supported by a grant to BWH from AstraZeneca
Biomarker testing received reagent support from Abbott Laboratories

3. hsTn is an emerging tool to better risk-stratify these patients
Background
Stable patients with a history of MI (SIHD) have a heterogeneous risk of recurrent major CV events
hsTn is an emerging tool to better risk-stratify these patients
PROVE IT-TIMI 22
Subset of 4,162 patients
30 days out from ACS
hsTnI
CV Death/MI/Stroke
TRA 2P - TIMI 50
15,833 patients
Prior MI, stroke, or PAD
Stable patients with a history of MI (SIHD) have a heterogeneous risk of recurrent major CV events
…and high-sensitivity troponin is an emerging application to help risk stratify these patients.
-Our group has previously shown that in a subset of over 4,000 patients 30 days out from MI, an elevated hsTn was associated with increasing risk of CV events
-Similarly, in nearly 16,000 pts with either prior MI, stroke, or PAD, there was a gradient of increased CV risk with higher hsTn
Objective: To further explore this relationship in the stable ischemic heart disease population
Bonaca MP, et al. J Am Coll Cardiol 2016;68:
Eisen A, et al. Clin Chem 2017;63:

4. Methods
Nested prospective cohort study (n= 8,635) from PEGASUS- TIMI 54
Randomized, double-blinded, placebo controlled trial of ticagrelor vs. placebo
≥55 years old
SIHD with a MI 1-3 years prior to enrollment
≥1 high-risk feature: ≥65 yo, DM, ≥2 MIs, multivessel CAD, CKD
Median F/U 33 months
Primary endpoint (PEP): CV death, MI, or stroke
All elements of PEP centrally adjudicated
We aimed to study this relationship further in a dedicated SIHD population
-To do this, we performed a Nested prospective cohort study (n= 8,635) from PEGASUS-TIMI 54
-This was a Randomized, double- blinded, placebo controlled trial of ticagrelor vs. placebo
-Patients were ≥ 55 years old, SIHD with a MI 1-3 years prior to enrollment
+ ≥1 high risk feature: ≥65 yo, DM, ≥2 MIs, multivessel disease, or CKD
-Median F/U 33 months
-Primary endpoint was a composite of CV death, MI, or stroke
-All elements of PEP centrally adjudicated

5. Methods (2) Serum hsTnI (Abbott ARCHITECT) drawn at enrollment
Limit-of-detection 2 ng/L
MI cut-point 26 ng/L (99th percentile)
hsTnI was analyzed as categorical variable using a priori categories within the normal reference range, defined as:
Undetectable (<2 ng/L)
Low (2-6 ng/L)
High (>6 ng/L)
All patients had baseline serum hsTnI drawn at enrollment using Abbott ARCHITECT assay
-Limit-of-detection 2 ng/L, MI cut-point 26 ng/L
-hsTnI was analyzed as categorical variable using a priori categories that were within the normal range,
Including: Undetectable (<2 ng/L), Low (2-6 ng/L), High (>6 ng/L)
-The cut-point of 2 is based on the limit of detection and the cut-point of 6 is based on numerous prior studies demonstrating a range of 5-7 as a risk threshold.
Population studies with risk thresholds between 5-7 ng/L
WOSCOPS (Ford, et al. JACC. 2016) BiomarCare (Blankenberg, et al. EHJ. 2016)
JUPITER (Everett, et al. Circ.2015) HUNT (Omland, et al Clin Chem. 2015)
CAPS (Patterson, et al. IJC. 2015) FINNRISK (Neumann, et al. PLOS One. 2014)
ADVANCE (Iribarren, et al. Heart. 2016) AGES-Rej (Thorstein, et al. Clin Chem. 2016)

6. hsTn Distribution 2 6 Median hsTnI (ng/L) Overall 4.2 [2.7,7.2]
Total N
8,635
Undetectable
920 (10.7%)
Detectable 7,715 (89.3%)
2-6 ng/L
5,028 (58.2%)
>6 ng/L
2,687 (31.1%)
<2 ng/L
Median hsTnI (ng/L)
Overall 4.2 [2.7,7.2]
Female 3.6 [2.3,6.5]
Male 4.3 [2.8,7.3]
Here is the distribution of hsTn among the study cohort and divided by male (in blue) and female (in red)
-The median in the total cohort was 4.2 ng/L but females had significantly lower hsTn, with median of 3.6.
Using our cut-points of 2 and 6, we divide the cohort into:
920 (10.7%) in the undetectable group
5,028 (58.2%) in the 2-6 group
and 2,687 (31.1%) in the >6 group
Sex

7. Baseline Characteristics
by hsTnI Level
<2
(N=920) %
2-6 (N=5028)
>6 (N=2687)
Demographics
Age (yrs) (median, IQR)
61 (55,66)
65 (58,70) 
67 (60,73)
Male 64 77 79
History
Hypertension 70 78 82
Diabetes Mellitus 29 28 34
History of CABG 1 5 9
History of 2nd prior MI 14 21
Peripheral Artery Disease 3 6
Congestive Heart Failure 7 12 25
eGFR <60 20 33
-There were significant differences in a number of the baseline characteristics across the 3 troponin subgroups
-Patients with higher troponin tended to be:
older, more likely male, more CV risk factors, more advanced CAD, CHF, and CKD
-We included these covariates in our adjusted analysis
*All p-values <0.001

8. Results: Unadjusted Analysis
13.5%*
6.0%^
2.8%
120
240
360
480
600
720
840
960
1080 0% 5%
10%
15%
20%
<2 (n=920)
>6 (n=2,687)
2-6 (n=5,028)
Days
CV Death, MI, Stroke (%)
CV Death, MI, Stroke by hsTnI Level (ng/L)
hsTnI Level
* p< vs. undetectable
^p= vs. undetectable
-Looking at the primary, unadjusted results by hsTn group, there are a 3 main findings
the event rates diverged immediately and continue to do so over 3 years
Those with hsTnI >6 had a 13.5% rate of CV Death, MI, Stroke at 3 years
Those in the 2-6 range had a 6% CV event rate
(both of which were significantly increased compared to 2.8% in undetectable group)

9. Results: Unadjusted Analysis
18.6%#
12.5%*
6.0%^
2.8% 5%
≥26 (n=433, 5.0%)
Days
20%
120
240
360
480
600
720
840
960
1080
15%
10% 0%
CV Death, MI, Stroke by hsTnI Level (ng/L)
# p< vs. undetectable
* p< vs. undetectable
^ p= vs. undetectable
<2 (n=920)
>6-<26 (n=2,254)
2-6 (n=5,028)
hsTnI Level
CV Death, MI, Stroke (%)
-When we separate the patients whose troponin is >26, we can appreciate that the risk in the >6 group is stable and therefore not driven by the risk in patients with troponin >26.
-In other words, the high risk seen in the 6-26 group are all in patients with “normal troponin”

10. Components of PEP P<0.0001* P<0.0001* P<0.0001^ P<0.0001^
-When looking at the components of the primary endpoints, we see striking consistency
1) Higher troponin was associated with a gradient of increased risk across CV death, coronary heart death, MI, and stroke
2) Importantly, in the 920 patients with undetectable troponin, there were no coronary heart deaths or CV deaths over 3 year follow up
P<0.0001*
P<0.0001*
P<0.0001^
P<0.0001^
*>6 vs 2-6
^P-Trend

11. PEP, Mortality, MI, and Stroke
Adjusted Hazard Ratios by hsTnI Level
Adjusted for congestive heart failure, hypertension, age >=75, diabetes mellitus, history of stroke, history of CABG, PAD, eGFR <60, current smoker
Hazard Ratio
<0.0001
0.0007
P-Trend
-In adjusted analysis, we see that the significant gradient of risk across troponin subgroups is maintained
1) such that trop >6 confers a:
4.3-fold risk in PEP, 3.9-fold increase in mortality, 3.6-fold increase in MI, and a nearly 5-fold increase in stroke
2) Likewise, those with troponin in the 2-6 range were also at a:
~2-fold increased risk for PEP, mortality and Myocardial infarction

12. hsTnI and TRS 2ᵒP Score *P-Trend <0.0001 n=1094 n=769 n=806
-Another way of assessing whether troponin adds to clinical risk factors is to add it to TRS 2P clinical risk score.
-Here we have a 3x3 chart with TRS 2P on the X axis and hsTn on the z-axis.
-We see 3 important things:
No matter what the clinical risk, troponin provides a significant improvement in risk stratification
If hsTn <2, the patient is low risk no matter what their clinical score, all with CV risk of <1%/year
High clinical risk patients (who you may not think would benefit from biomarker testing) appear to get the greatest risk stratification from hsTn
n=1660
n=481
n=135
n=292
*P-Trend <0.0001

13. Ticagrelor by hsTnI
0.84
P=0.001
0.91
0.94
0.91
Lastly, I will address the effect of Ticagrelor by hsTn
-As this audience will recall, PEGASUS showed an overall significant reduction in major CV events w/ long-term ticagrelor, HR 0.84. 
-In this biomarker cohort, there was a consistent effect of ticagrelor but w/ predictably larger CIs in this smaller sample. 
-Similarly, there was no statistical heterogeneity in the relative effect of ticagrelor across troponin subgroups.
-However, we noted a possible trend toward lack of benefit in those very low risk pts with undetectable troponin.
1.20
3.1

14. Conclusion
hsTn is a strong independent predictor of recurrent major CV events in SIHD
Gradient of risk is present at any detectable level
As low as 2-6 ng/L
hsTnI >6 ng/L present in nearly 1/3 of patients 
>10% 3-year risk of MACE
Extrapolated 10-year risk of MACE = 45%
In combination with TRS 2ºP, hsTnI provides complementary information
In conclusion, hsTn is a strong independent predictor of recurrent major CV events in SIHD
The gradient of risk is present at any detectable level
Even as low as 2-6 ng/L
A hsTnI >6 ng/L is present in nearly 1/3 of SIHD patients and confers a >10% 3-year risk of MACE
Which extrapolated over 10 years becomes a 45% risk of MACE
Finally, in combination with TRS 2ºP, hsTnI provides complementary information that we can use to better risk-stratify our patients with SIHD.

15. THANK YOU
Thank you for your attention

16. 99th Percentile by Sex Women Men
18.4%
12.0%
5.6%
3.3%
120
240
360
480
600
720
840
960
1080
Days 0% 5%
10%
15%
20%
25%
<2
2-6
>6-<34
≥34
Women
Men
CV Death, MI, Stroke (%)
21.5%
14.5%
7.1%
1.9%
>6-<16
≥16
Sex-specific 99th %ile URL: 34 ng/L in men and 16 ng/L in women
-When we look at men and women separately, we see that hsTn provides good risk stratification in both, however it seems to be better in women (1.9->21.5 vs 3.3->18.4)

17. PEP, Mortality, MI, and Stroke
Adjusted Hazard Ratios by hsTnI Level
Adjusted for congestive heart failure, hypertension, age >=75, diabetes mellitus, history of stroke, history of CABG, PAD, eGFR <60, current smoker
-Another way of looking at the same data is to use 2-6 as the reference, after all this is the where the majority of SIHD patients fall.
-hsTn <2 identifies patients with 50% lower risk of events while >6 identifies those who are 2-fold higher risk.