Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: The accuracy of troponin T and I tests in acute MI depends on timing after onset of symptoms


ACP J Club. 2001 Jan-Feb;134:28. doi:10.7326/ACPJC-2001-134-1-028

Source Citation

Ebell MH, Flewelling D, Flynn CA. A systematic review of troponin T and I for diagnosing acute myocardial infarction. J Fam Pract. 2000 Jun;49:550-6. [PubMed ID: 10923557] (All 2001 articles were reviewed for relevancy, and abstracts were last revised in 2006.)



Are troponin T and troponin I tests accurate for detecting acute myocardial infarction (MI) at varying times from onset of symptoms or presentation to the emergency department (ED)?

Data sources

Studies were identified by searching MEDLINE to December 1999 with terms and phrases related to troponin, diagnosis, and diagnostic use.

Study selection

English-, German-, French and Spanish-language studies were selected if troponin tests were used in the assessment of adults with acute chest pain, the study design included prospective data collection, assessment of acute MI was made by clinicians blinded to the results of the troponin tests, MI was diagnosed using the World Health Organization or similar criteria, and sensitivity and specificity data were reported for ≥ 1 time point from the onset of pain or presentation to the ED.

Data extraction

Data were extracted on settings, consecutive or nonconsecutive patient enrollment, patient characteristics, test manufacturer, cut point values for abnormal levels, and timing of the test from onset of pain or presentation to the ED. Study quality was classified as level I if patients were enrolled consecutively and the rest were classified as level II. Data on sensitivity, specificity, and likelihood ratios were extracted or calculated and combined in meta-analyses.

Main results

19 studies met the inclusion criteria. Of the 11 level-I studies (3736 patients), 7 evaluated troponin T, and 4 evaluated troponin I. Of the 8 level-II studies (1998 patients), 6 evaluated troponin T, 1 evaluated troponin I, and 1 evaluated both. The sensitivity of both tests increased from the time of onset of symptoms, whereas specificity stayed the same or decreased slightly. Sensitivity was highest at 8 hours and beyond for troponin T and at 6 hours for troponin I. Troponin T ≤ 0.2 ng/mL at 8 hours best ruled out MI. Summary data are presented in the Table.


The sensitivity of troponin T and troponin I tests for detecting acute myocardial infarction is dependent on the timing of the test; the tests are insensitive within 6 hours of onset of symptoms. A negative test result 8 or more hours after onset best rules out myocardial infarction.

Source of funding: Michigan Consortium for Family Practice Research.

For correspondence: Dr. M.H. Ebell, Department of Family Practice, Michigan State University, B101 Clinical Center, East Lansing, MI 48824-1315, USA. FAX 706-549-8121.

Table. Test characteristics for troponin T and I for detecting acute myocardial infarction*

Test and cut point for abnormal levels Time in hours from onset of symptoms Sensitivity Specificity +LR -LR
Troponin T > 0.1 ng/mL 1 47% 87% 3.7 0.6
6 74% 83% 4.4 0.3
8 84% 81% 4.5 0.2
Troponin T > 0.2 ng/mL 1 14% 87% 1.1 1.0
6 86% 83% 5.1 0.2
8 96% 81% 5.2 0.05
Troponin I > 0.1 ng/mL (based on only 1 study) 1 13% 95% 2.7 0.9
4 67% 95% 13 0.3
6 90% 95% 18 0.1

*LRs defined in Glossary.


Serum troponin T and I tests are often used to help diagnose acute MI. The diagnostic performance of the tests varies greatly, however, depending on the cut point used to define abnormal levels and the timing of the test from onset of symptoms. The review by Ebell and colleagues highlights the difficulties in using troponin tests for the diagnosis of MI and provides summary sensitivity and specificity data for specific cut point values and times from symptom onset. These values are derived by using data from studies of varying quality and from a wide spectrum of patients, including patients presenting to the ED with symptoms suggestive of MI and those admitted to coronary care units because of high suspicion of MI. The reported test sensitivities may be inflated by the inclusion of patients with a high likelihood of MI. The summary diagnostic performance values are “best-fit” estimates from data from several studies and are not derived by statistically rigorous methods.

The only conclusions that should be drawn from this review are that a positive troponin T test result can be useful for ruling in MI when blood is drawn ≥ 6 hours after symptom onset and a negative troponin test result can effectively rule out MI when blood is drawn ≥ 8 hours after symptom onset.

The review fails to mention the difficulty in using these tests to diagnose patients with unstable angina; a negative test result that rules out MI should not necessarily be used to send patients home from the ED. Furthermore, the diagnostic performance of the tests is greatly affected by changing the cut points for abnormal values: A lower cutoff (0.1 ng/mL) increases sensitivity but decreases specificity.

Catherine Milch, MD
Tufts University School of Medicine
New England Medical CenterBoston, Massachusetts, USA