Abstract 5025: Diagnostic Value of Serial Changes in Low-Positive Extra-Sensitive Troponin I Levels for Acute Myocardial Infarction

Abstract

Background: The use of high sensitivity troponin I (TnI-ES) assays results in higher rates of low-positive (LP-TnI-ES) results. Diagnostic value of serial changes in LP-TnI-ES levels remains unclear. Methods: Results of serial TnI-ES tests (VITROS ECi, Ortho-Clinical Diagnostics) were obtained from July 2008 to March 2009. Demographic and clinical data were recorded. Positive (>0.034 ng/ml [the 99 th percentile]) results were divided into low (0.035– 0.100 ng/ml), intermediate (0.101– 0.300 ng/ml), and high (>0.300 ng/ml) categories. Serial LP-TnI-ES results were tracked in hospitalized patients for progression to a higher category, regression to normal, or no significant change. Time interval for such changes was calculated. ANOVA and Pearson’s chi-Square tests were used to analyze association with outcome variables. Results: TnI-ES was performed in a total of 4090 unique patients, of whom 1134 (27.7%) had at least one positive TnI-ES value. At least one LP-TnI-ES value was noted in 818 (73.5%; 59% male, 37% African American; age 68±11 years) of these 1134 pts. Patients diagnosed with myocardial infarction (MI) comprised 7.3% of the LP-TnI-ES group. Predictors of MI by univariate analysis included higher initial LP-TnI-ES level ( P =0.002), progression from lower or regression from higher values ( P <0.001), higher GFR ( P =0.05), lower creatinine ( P =0.05), and prior history of MI ( P <0.001); while age, gender, diabetes, and race were not significant. Results from a multivariate regression model are shown in Table 1 . Progression of LP-TnI-ES to high in less than 12 hours was most predictive of MI (mean LP-TnI-ES change: 10.646±27.054; mean time to change: 6:24±3:01; P <0.001). Conclusions: In patients with a low-positive TnI-ES result, serial changes and higher initial LP-TnI-ES levels, prior history of MI, and higher GFR were predictive of MI. These variables may be utilized to construct a clinical algorithm to aid decision-making in patients with LP-TnI-ES in the ER. Table 1- Results of Multivariate Logistic Regression Model