GDF-15 Unit Converter
GDF-15 Unit Converter
Convert growth differentiation factor 15 between pg/mL, ng/L and ng/mL — the first two are the same number. What the converted figure is worth is the harder question: GDF-15 predicts outcome about as well as any single biomarker in cardiology and diagnoses nothing at all, because almost everything raises it.
GDF-15 converter
Mass units onlyGDF-15 950 pg/mL
The ladder, and the unit the trials use
1 ng/mL = 1,000 pg/mL, so pg/mL ÷ 1,000 = ng/mL
no pmol/L is offered
- pg/mL = ng/L
- a picogram in a millilitre is a nanogram in a litre. Prefix and volume both change a thousandfold and cancel. The cardiology literature almost always writes ng/L; assay inserts and most laboratory reports write pg/mL; they are the same number and neither is more correct
- ng/mL
- the pg/mL figure divided by a thousand. A typical result becomes something like 0.95, which is easy to misread as a low result on a scale it does not belong to — if a GDF-15 arrives in ng/mL, convert it before comparing it with anything published
- no molar unit
- GDF-15 circulates as a disulphide-linked dimer of the TGF-β superfamily and is measured against a protein calibrator. Its grams per mole is not the constant a molar conversion needs, so no pmol/L appears on this page
- what the unit cannot fix
- GDF-15 assays are not harmonised. One review states plainly that "the values shown in these studies are method-dependent and cannot be indistinctively used", so a figure from one platform is not interchangeable with a threshold derived on another, however carefully the units are converted
Worked example
GDF-15 950 pg/mL
950 pg/mL is 950 ng/L — the same number, and ng/L is how the anticoagulation trials print it
950 ÷ 1,000 = 0.950 ng/mL
Inside the trials' lowest stratum, below the 1,200 pg/mL that was chosen as the 90th percentile in healthy individuals
That places the patient in the lowest of three risk groups for major bleeding and for death — and says nothing whatever about what is wrong with them
For a marker with a diagnostic threshold rather than a prognostic gradient, see the NT-proBNP converter and read it on the NT-proBNP heart failure interpreter
The three strata the anticoagulation trials used, and what they predicted
| GDF-15 stratum | Share of patients | Major bleeding | All-cause mortality | Stroke |
|---|---|---|---|---|
| Under 1,200 ng/L (= pg/mL) | 31.2% | Reference group | Reference group | Reference group |
| 1,200 to 1,800 ng/L | 31.9% | Intermediate | Intermediate | No gradient |
| Over 1,800 ng/L | 36.9% | "hazard ratio [95% CI] group 3 vs group 1 1.76 [1.28-2.42], P < .0005" | "hazard ratio 1.72 [1.30-2.29], P < .0005" | None |
The ABC bleeding score, which is where a GDF-15 is usually ordered
| Component | What it contributes | Where the figures come from |
|---|---|---|
| Age | The A | Derived in ARISTOTLE (14,537 patients), externally validated in RE-LY (8,468) |
| GDF-15 | One of the three biomarkers | Roche Elecsys GDF-15 in the published validations; a real-world cohort reports medians of 783 ng/L in its low-risk group and 2,075 ng/L in its high-risk group |
| High-sensitivity cardiac troponin T | One of the three biomarkers | See the troponin T converter |
| Haemoglobin | One of the three biomarkers | — |
| Prior bleeding | The clinical history | — |
| Discrimination, derivation cohort | ABC-bleeding 0.68 (0.66–0.70) | against HAS-BLED 0.61 (0.59–0.63) and ORBIT 0.65 (0.62–0.67) |
| Discrimination, validation cohort | ABC-bleeding 0.71 (0.68–0.73) | against HAS-BLED 0.62 (0.59–0.64) and ORBIT 0.68 (0.65–0.70) |
Everything else that raises GDF-15, which is the reason it diagnoses nothing
| Cause | What the source says |
|---|---|
| Age | "concentrations of GDF-15 in healthy individuals increase slowly with aging"; a cohort study reports a moderate positive correlation with age (ρ = 0.457) |
| Renal impairment | Among the "Major determinants of GDF-15 concentrations", listed alongside age, diabetes, NT-proBNP, hs-TnT and NYHA class III/IV |
| Heart failure severity | Same list — NYHA class III/IV is a determinant, which is exactly why it predicts outcome in heart failure |
| Cancer | GDF-15 rises "in several physiological and pathological conditions, including aging, cancer, cardiovascular disease (CVD), metabolic disorders, and pregnancy" |
| Any inflammatory or stress state | "rapidly produced by various cell types … in response to other cytokines, cellular stress or tissue injury, hypoxia and oncogene activation" |
| Smoking | In a cohort of people with HIV, current smokers 682.7 against 410.2 pg/mL in never-smokers (validation cohort 544.3 against 397.9); "tobacco exposure strongly induces GDF-15 expression" |
| Multimorbidity as such | In the same cohort, 771.5 against 390.0 pg/mL with and without multimorbidity; GDF-15 "appears to reflect a broader state of multisystem physiological stress" |
| Pregnancy | Named in the list of physiological causes |
| Metformin | "GDF-15 levels increase following administration of metformin, the most prescribed antidiabetic drug worldwide, and serve as an important mediator of its effects" |
A number that forecasts well and explains nothing
The unit question is short. A picogram per millilitre and a nanogram per litre are the same concentration: the prefix changes by a thousand, the volume changes by a thousand, and the two cancel. The cardiology literature writes ng/L and assay inserts usually write pg/mL, and no conversion stands between them. Nanograms per millilitre is the pg/mL figure divided by a thousand, which turns a typical result into something like 0.95 — a number that invites misreading, and the one case on this ladder where converting before comparing actually matters. There is no molar unit, because GDF-15 is a disulphide-linked dimer measured against a protein calibrator rather than against a molar mass.
The interesting part is what the number is for. GDF-15 is weakly expressed in healthy tissue and is induced rapidly by cellular stress, injury, hypoxia, inflammatory cytokines and oncogene activation. That makes it an excellent forecaster and a hopeless diagnostician. It rises with age, with falling renal function, with diabetes, with heart failure severity, with cancer, with smoking, in pregnancy and on metformin; one review calls it an integrated biomarker of multiple comorbidities rather than a specific reflection of cardiovascular health, and records that it has shown only limited diagnostic usefulness in chest pain, breathlessness, heart failure or myocardial infarction while remaining a strong predictor of cardiovascular events. Those two sentences are the whole clinical profile. If a raised GDF-15 comes back on a breathless patient, it has not told you whether the cause is cardiac, renal, malignant or inflammatory. It has told you that the patient is more likely to do badly whichever it is.
That property turns out to be useful in one specific place. In atrial fibrillation the clinician is balancing stroke risk against bleeding risk, and a marker that tracks one and not the other is worth having. The RE-LY analysis found exactly that: patients above 1,800 ng/L had hazard ratios of 1.76 for major bleeding and 1.72 for death against the group below 1,200 ng/L, and no gradient at all for stroke. GDF-15 is therefore one of the three biomarkers in the ABC bleeding score, alongside high-sensitivity troponin T and haemoglobin, with age and prior bleeding — a score derived in ARISTOTLE and validated in RE-LY, where it discriminated better than HAS-BLED and ORBIT by roughly six points of c-index. That is a real improvement in a decision where the downside runs in both directions, and it is a modest one, and both halves of that sentence belong in any account of why the assay exists.
What the page deliberately does not give you is a reference interval. One review states that no specific cut-points have been proposed and that published values are method-dependent and cannot be used interchangeably; another gives a typical healthy range of roughly 100 to 1,200 pg/mL without proposing it as a decision threshold. The single comparison figure offered above is the lowest of the three strata the trials pre-specified, whose 1,200 ng/L bound was chosen as the ninetieth percentile in healthy people — a population statistic used to cut a trial cohort into thirds, not a limit above which anything is treated. Reading it as a normal range would be reading a research design as a laboratory interval, and the honest position is that GDF-15 is a continuous prognostic variable that happens to have been stratified for analysis.
Practically: if you want a number that tells you whether this patient has heart failure, that is the NT-proBNP converter read on the NT-proBNP heart failure interpreter, not this one. If you want to know whether myocardium has been injured, that is the troponin T converter. GDF-15 answers a different question — how much cumulative physiological burden is this patient carrying — and it answers it well enough to be worth converting correctly and worth interpreting cautiously.
Frequently asked questions
Is pg/mL the same as ng/L for GDF-15?
Yes, exactly. A picogram per millilitre and a nanogram per litre are the same concentration, because the prefix and the volume both change by a factor of a thousand and cancel. The trial literature almost always uses ng/L and assay inserts usually use pg/mL, so a threshold quoted in one can be applied directly to a result in the other. Nanograms per millilitre is the same figure divided by a thousand.
What is a normal GDF-15 level?
No agreed reference interval exists. One review states that "no specific cutpoints have been proposed" and that published values are method-dependent and cannot be used interchangeably; another gives a typical healthy range of roughly 100 to 1,200 pg/mL. The 1,200 ng/L figure most often quoted is the 90th percentile in healthy individuals, adopted by the anticoagulation trials as the boundary of their lowest stratum — a research cut, not a laboratory limit.
What raises GDF-15?
Almost everything that makes a patient unwell. Age, renal impairment, diabetes, heart failure severity, cancer, any inflammatory or hypoxic state, smoking, pregnancy and metformin therapy all raise it. That breadth is the point rather than a limitation: GDF-15 behaves as an integrated marker of accumulated physiological burden, which is why it forecasts outcome so well across so many diseases and why it cannot identify which disease is present.
Is GDF-15 used to diagnose heart failure?
No. One review records that it has "shown only limited diagnostic usefulness in patients with chest pain, dyspnea, HF or MI" while remaining "a strong predictor of cardiovascular events". For diagnosis and rule-out in breathlessness the natriuretic peptides carry the validated thresholds. GDF-15 adds prognostic information on top of a diagnosis already made by other means.
Why is GDF-15 in the ABC bleeding score?
Because it predicts bleeding and death without predicting stroke, which is precisely the discrimination an anticoagulation decision needs. The score is age, three biomarkers (GDF-15, high-sensitivity troponin T and haemoglobin) and prior bleeding; it was derived in ARISTOTLE and validated in RE-LY, where its c-index for major bleeding was 0.71 against 0.62 for HAS-BLED and 0.68 for ORBIT. The score uses GDF-15 as a continuous variable rather than applying a threshold.
Can I compare a GDF-15 result between laboratories?
Only with care. The units convert exactly, but the assays are not harmonised: a published review states that the values "are method-dependent and cannot be indistinctively used". The trial thresholds were generated on one manufacturer’s electrochemiluminescence assay, so applying 1,200 or 1,800 ng/L to a result from a different platform imports an assumption the literature does not support. Within one laboratory, a trend is more informative than a single value.
Related calculators
References
- Hijazi Z, Oldgren J, Lindbäck J, et al. The novel biomarker-based ABC (age, biomarkers, clinical history)-bleeding risk score for patients with atrial fibrillation: a derivation and validation study. Lancet. 2016;387:2302–11. Components "age, biomarkers [GDF-15, cTnT-hs, and hemoglobin], and clinical history [prior bleeding]"; derived in ARISTOTLE (n=14,537), validated in RE-LY (n=8,468); c-index 0.71 (0.68–0.73) against HAS-BLED 0.62 and ORBIT 0.68.
- Growth-differentiation factor 15 and risk of major bleeding in atrial fibrillation: insights from the RE-LY trial. Am Heart J. 2017. Strata <1,200 ng/L (31.2%), 1,200–1,800 ng/L (31.9%), >1,800 ng/L (36.9%); 1,200 ng/L "represented the 90th percentile in healthy individuals"; major bleeding HR 1.76 (1.28–2.42), death HR 1.72 (1.30–2.29), no association with stroke.
- Growth differentiation factor-15, a novel systemic biomarker of oxidative stress, inflammation, and cellular aging. Am J Prev Cardiol. 2021. "an integrated biomarker of multiple comorbidities rather than a specific reflection of cardiovascular health"; "only limited diagnostic usefulness in patients with chest pain, dyspnea, HF or MI"; "no specific cutpoints have been proposed"; "the values shown in these studies are method-dependent and cannot be indistinctively used".
- Growth differentiation factor 15 (GDF-15) as an emerging biomarker for cardiovascular and kidney diseases. Front Pharmacol. 2026. "Circulating GDF-15 levels typically range from 100 to 1,200 pg/mL"; rises in "aging, cancer, cardiovascular disease (CVD), metabolic disorders, and pregnancy"; "GDF-15 levels increase following administration of metformin".
- Elevated plasma GDF-15 levels are associated with aging, multimorbidity, and tobacco exposure in people with HIV. Infect Dis Ther. 2026. Current smokers 682.7 vs never-smokers 410.2 pg/mL (validation 544.3 vs 397.9); multimorbidity 771.5 vs 390.0 pg/mL; correlation with age ρ = 0.457.
- Performance of the ABC-bleeding risk score in Chinese patients with atrial fibrillation on oral anticoagulation: a real-world study. Front Cardiovasc Med. 2022. Roche Elecsys GDF-15; median 783 ng/L in the low-risk group, 2,075 ng/L in the high-risk group.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.