Design: Population cohort study with biochemical verification and external replication
Population: 19,499 adults with paired high-sensitivity troponin I and T (median age 49, 58% women)
Primary outcome: Future MI, ischaemic stroke or cardiovascular death
Effect: Myocardial injury: aHR 2.90 (cTnI), 1.99 (cTnT); suspected macrotroponin: aHR 1.59 (0.94–2.70) and 1.27 (1.00–1.61)

In a general-population cohort of 19,499 adults (median age 49), both high-sensitivity troponin I and troponin T were measured. Troponin I was raised in 1% and troponin T in 6%. About half of these raised results were discordant, with a more than threefold difference between the two assays.

Biochemical testing (immunoglobulin depletion and ultracentrifugation) in 203 samples attributed 92.5% of discordant troponin I and 15.4% of discordant troponin T to macrotroponin: troponin bound to antibodies, which clears slowly and reads high without any extra heart injury.

People with concordant raised troponin, taken as true myocardial injury, had higher future risk of MI, ischaemic stroke or cardiovascular death (adjusted HR 2.90 for troponin I, 1.99 for troponin T). Those with suspected macrotroponin did not have clearly raised risk (HR 1.59, 95% CI 0.94 to 2.70 for I; 1.27, 1.00 to 1.61 for T). A second cohort replicated the findings.

This matters as troponin moves into screening and outpatient risk assessment. A stable, isolated raised troponin I in a well person may be an assay artefact.

A persistently raised troponin I with no symptoms and no dynamic change may be macrotroponin rather than injury.
Measuring the other troponin (T if I is raised) can reveal discordance; a more than threefold gap suggests interference.
Ask the laboratory about polyethylene glycol precipitation or other interference testing before labelling a patient.
This applies to stable outpatient results, not to suspected acute coronary syndrome, where serial change still decides.

Why it matters

It changes what a raised troponin I means in screening: often an antibody complex, not a damaged heart.

The statistics, in plain English

The hazard ratio for suspected macrotroponin I was 1.59, but its 95% confidence interval (0.94 to 2.70) crosses 1.0, so it is not clearly different from normal risk. For true injury the whole interval sits well above 1.0. The biochemical confirmation was done on only 203 samples.