Objective This study aims to determine whether the prognostic significance of troponins in acute coronary syndrome in predicting mortality varies by age, and if so, to what extent when other prognostic indicators are considered.

Methods We analysed Myocardial Ischemia National Audit Project registry data collected between January 2006 and December 2010 and followed up this cohort for all-cause mortality until August 2011. Relationships between peak troponin levels (types I and T) and time to death in different age groups, and between age and time to death at different troponin levels were investigated using multiple variable adjusted Cox regression models.

Results Of the 322 617 patients with acute coronary syndromes included, a third (n=106 365, 33%) died during 695 334 person-years of follow-up. Within each troponin category, older age was associated with a higher mortality even in those with a troponin <0.01 ng/mL for both troponin types (HR ∼10–12 in ≥85 years cf. HR of 1.0 in <65 years). The relative potency of an elevated troponin to predict an adverse outcome compared with a low troponin attenuated with increased age (for troponin I ≥15.0 compared with troponin I <0.01 in age <65, adjusted HR 2.41 (95% CI 1.80 to 3.24); age ≥85 HR 2.01 (1.62 to 2.52)). Similar but less consistent results were observed with troponin T elevation at the higher levels.

Conclusions Clinicians should interpret the prognostic value of troponin taking into account the patient's age.