Kevin M. Lee
Clinically Appraised Topic
July 6, 2001
55 y/o F with dyspnea presents to UNC ED. EKG demonstrates no ischemic changes; However, an initial set of cardiac enzymes is obtained and Troponin I is 15 ng/ml.
What is the predictive value of Troponin I for diagnosing acute MI/cardiac death?
(any prognostic value for risk stratification?)
Ability of Troponin I to predict cardiac events in
patients admitted from the Emergency Department.
Kontos, M et al. Journal of the
American College of Cardiology. November 15, 2000. Vol 36 (6), pp
1818-1823.
A positive Troponin I (TnI) has a high sensitivity and is roughly 14x more likely to be found in patients with MI/cardiac death, than not. This sensitivity was unchanged even when ischemic EKG’s were excluded from the analysis. In patients without MI who subsequently experienced significant complications (i.e. underwent revascularization w/n one month or had cardiac/respiratory arrest, intubation, pressors, IABP, PA cath, transvenous pacer, VF/VT during admission), only 21% had a positive TnI. Likelihood ratios revealed no predictive ability of TnI to identify those with significant CAD/significant complications. This study did not assess the effect of gradations in TnI elevation for predicting different cardiac endpoints.
Methods
· 2194 consecutive patients enrolled as part of chest pain protocol for suspected MI/acute coronary syndrome at MCV in Richmond, VA
· 125 pts excluded when dx of MI high (persistent ST elev/H&P suggestive)
· 124 pts excl - incomplete documentation of cardiac enzymes/16 pts left AMA
· Remaining 1929 pts admitted for ongoing observation/mgmt
· CK/CK-MB at 0, 3, 6, and 8 hours (+ if CK/MB > 8.0 ng/ml)
· Troponin I at 0 and 8 hours (+ if TnI > 1.0 ng/ml)
· Serial enzymes beyond 8 hrs - if any MI dx, Sx continued, or enzyme set +
· Ischemic EKG (transient ST elev, ST depr >1mm, or TWI >2mm) determined by cardiologist blinded to clinical variables, TnI results, and outcomes
Critical
Review
Are the results valid?
Was there an independent, blind comparison w/ a
reference standard?
Yes, includes outcomes and blinded EKG readings by cardiologist. However, only 534/1754 of pts w/o MI had cath…definitive absence/presence of sig. CAD unknown
Is test clearly described? Can the methods described
permit successful replication?
Somewhat. Cut-off value for positive TnI determined by ROC but time from onset of Sx to test not always stated nor adjusted for.
Did pt sample include appropriate spectrum of pts to
whom dx test is applied?
Yes (see Table 1) – though does not make distinction re: pts who presented w/ arrhythmia, CHF, others in diff dx (myo/pericarditis, PE, recent trauma, etc)
Did results of test being evaluated influence decision
to perform reference standard?
Yes, though other factors such as vital signs (tachy/bradycardias, hypotension, etc) not mentioned in analysis. Other facets of CP protocol may have led to interventions such as revascularization/cardiac cath independent of troponin values.
What are the results? (see Table 2)
Cardiac events in 513/1929 pts (22%)
· MI w/n 7 days – 175 pts (9%)
- 96% pts had positive TnI
- 33% of pts who had MI/death had 1st TnI positive
· Cardiac death w/n 5 weeks – 34 pts (2%)
- 26% pts had negative TnI
· Significant CAD (>50% sten L main, >70% other maj cor.)
in pts without MI who underwent cath – 323 pts (17%)
- 45/323 pts had a positive TnI
· Significant complications – 248 pts (13%)
- 20% pts had positive TnI
- revasc via CABG/PTCA in pts w/o MI – 193 pts
- others – arrested, needed pressors, PA cath, etc
Is test performance summarized by Sensitivities/Specificities/LR?
|
All Patients |
Non-ischemic EKG |
||
|
LR+ |
LR- |
LR+ |
LR- |
|
13.7 |
0.04 |
16 |
0.03 |
|
13.8 |
0.1 |
15 |
0.1 |
|
8.75 |
0.6 |
8 |
0.3 |
|
5 |
0.8 |
3.8 |
0.8 |
|
2.8 |
0.9 |
2.4 |
0.9 |
Troponin
I > 1
1. MI
2. MI/Death
3. MI/D/CAD