- Design
- Prospective diagnostic and prognostic cohort study
- Population
- 3,379 patients with suspected functionally relevant coronary artery disease (median age 68)
- Primary outcome
- Diagnostic accuracy for adjudicated functional CAD; all-cause and CV death over 5 years
- Effect
- Diagnostic AUC 0.598; 2-year AUC 0.815 (all-cause death), 0.834 (CV death)
In a prospective cohort of 3,379 patients (median age 68) investigated for suspected functionally relevant coronary artery disease, all had myocardial perfusion imaging and many had angiography with fractional flow reserve. Diagnosis was centrally adjudicated; 34% had functionally relevant disease.
Growth differentiation factor 15 (GDF-15), a stress-response protein, was higher in those with disease (1,325 vs 1,041 ng/L) but diagnosed it poorly (AUC 0.598; 0.659 with clinical judgement). Over five years, 437 patients (14%) died. GDF-15 predicted death strongly (time-dependent AUC 0.815 for all-cause and 0.834 for cardiovascular death at two years), at least as well as high-sensitivity troponin T and better than troponin I.
GDF-15 behaves as a marker of general frailty and illness burden rather than of coronary ischaemia. It is not routinely available and is not a test to order for chest pain.
- Do not use GDF-15 to diagnose or exclude ischaemic coronary disease.
- A raised GDF-15 signals high overall mortality risk, similar to troponin T in this cohort.
- Functional testing or angiography remains the way to diagnose significant coronary disease.
- GDF-15 is a research marker; it is not part of routine chest pain work-up.
Why it matters
It is a reminder that a marker that predicts death is not necessarily a marker of the disease you are testing for.
Don't overread it
Observational; associations, not causes.
The statistics, in plain English
An AUC of 0.5 is a coin toss and 1.0 is perfect. At 0.60, GDF-15 barely beats chance for diagnosis; at 0.82 for predicting death, it discriminates well. A time-dependent AUC measures how well the marker separates who dies by a given time from who does not.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for internal medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free