GDF15 as a Biomarker of Biological Aging

GDF15 is a circulating cellular-stress signal that rises with age and predicts mortality across large cohorts, but it is useful only when read as a slow risk index rather than as a target to lower.

Also known as: Growth Differentiation Factor 15, MIC-1, macrophage inhibitory cytokine 1, NAG-1, PLAB

GDF15 appears on longevity-clinic panels and direct-to-consumer testing menus for the same reason it appears in geroscience papers: a single immunoassay produces a number that tracks accumulated cellular stress better than any one upstream marker can. The clinical and biological context is still settling. The popular framing has run ahead of it.

Context

GDF15 is a TGF-β superfamily protein discovered in the late 1990s. It has been renamed several times (NAG-1, PLAB, MIC-1) because different groups found it doing different things in different tissues. The current consensus is that GDF15 is a stress-response protein. Most cell types express it at low levels under ordinary conditions. Mitochondrial dysfunction, oxidative stress, hypoxia, tissue injury, low-grade inflammation, integrated stress response activation, and severe acute illness can all drive expression upward and push protein into circulation.

Serum levels are very low in healthy young adults, often near 200–400 pg/mL, and rise approximately exponentially with chronological age. By the eighth and ninth decades, levels above 1,500 pg/mL are common in otherwise healthy adults. Levels above 3,000 pg/mL appear in frailty, advanced cancer, severe cardiovascular disease, and end-stage renal disease. The biology under that age curve is cumulative tissue-stress integration, not one tissue’s senescence signature.

The biomarker shows up in three reader paths. A clinician orders it as part of an extended cardiovascular or oncology workup. A direct-to-consumer testing service includes it on an aging-focused panel. A longevity clinic adds it to an annual deep screen beside ApoB, hsCRP, IL-6, Lp(a), and methylation-age estimates. The number often arrives without much context, and the reader is left to decide what it means. That isn’t a small ask.

Problem

GDF15 produces two confusing reactions, and the published framing rarely separates them.

The first is over-treatment of the number. A high GDF15 result reads as alarm. Wellness writers translate “predicts mortality” into “should be lowered,” and supplement vendors propose mitochondrial cofactors and senolytic stacks as if a lower GDF15 were the goal. No human trial has shown that lowering GDF15 improves a longevity endpoint. The protein is downstream of upstream causes; lowering the readout without changing the causes is unlikely to change the outcome. In some contexts, the body may be using GDF15 signaling for adaptive purposes.

The second is dismissal. A skeptical reader hears that GDF15 has no proven intervention and concludes the number is not useful. That is too strong. The protein has the largest single age effect of any well-validated circulating biomarker. Several large prospective cohorts have shown that it adds independent prognostic information after adjustment for age, sex, traditional risk factors, and other inflammation markers. The signal is real even if the action is not yet defined.

Both errors live in the gap between “this measures something” and “this is a treatment target.” The pattern is to use GDF15 as an integrated risk index without crossing that gap.

Forces

• Large prospective cohorts show GDF15 independently predicts all-cause mortality, cardiovascular death, and cancer incidence, but no intervention RCT has tested whether lowering GDF15 changes outcomes.
• The protein is mechanistically interesting (mitochondrial-stress integration, GFRAL receptor signaling, anorexia and weight-loss pathways), but the popular framing is often just “elevated is bad.”
• Metformin substantially raises GDF15 by drug effect; any reader already on metformin has elevated GDF15 that does not reflect their aging trajectory.
• Severe acute illness, cancer, advanced heart failure, and chronic kidney disease all elevate GDF15; without that context, a single number from a person in any of those states can be misread.
• The test is widely offered on longevity panels but is not on standard primary-care panels; cost, reference ranges, and assay technique vary by lab.

Solution

The pattern is clinician-interpreted GDF15 measurement read as a slow-moving cellular-stress index, not as a target to titrate. Useful GDF15 measurement is bounded by four discipline points.

First, the result is interpreted alongside the readings GDF15 is not. hsCRP and IL-6 capture acute-and-chronic inflammatory load. ApoB and Lp(a) capture atherogenic-particle risk. Ferritin, hemoglobin A1c, fasting insulin, eGFR, and a basic metabolic panel capture iron, glucose, and renal-function context that can move GDF15 independently of aging. Body composition, blood pressure, fitness, and sleep set the baseline. GDF15 supplies a different axis than any of these. A high GDF15 in a person whose other markers are clean means something different from a high GDF15 in a person whose hsCRP is also high and whose renal function is declining.

Second, the metformin confound is named explicitly. Metformin raises GDF15 in a dose-dependent manner, often doubling or tripling baseline values in chronic users, through a direct effect on the mitochondrial integrated stress response. A metformin user’s GDF15 can’t be read as a clean aging signal. Metformin-aging research has had to ask whether GDF15-mediated appetite suppression is part of a hypothesized geroprotective mechanism or a side channel to control for. For the reader, the practical point is simpler: stop the inference at the drug effect.

Third, severe acute illness is excluded from the read. A measurement during or shortly after a hospitalization, an active infection, a recent surgical recovery, or an oncology treatment course describes the acute state, not the aging baseline. Cancer in particular can drive GDF15 upward, sometimes dramatically. A high GDF15 in an undiagnosed person whose other markers are unusual deserves a workup, not a supplement intervention.

Fourth, the retest interval is conservative. GDF15 is a slow-moving signal compared with hsCRP or IL-6. A meaningful change at the individual level requires intervals of months to a year and stable life conditions in between. Quarterly testing in a healthy adult is too frequent and will surface assay noise and ordinary variation as if they were change.

What GDF15 does not support is a titration plan. No published intervention RCT shows that any drug, supplement, exercise protocol, or diet lowers GDF15 in healthy adults and produces a longevity benefit. Exercise studies are mixed: some report acute elevations after intense training, while chronic effects are less clear. Caloric restriction in animal models raises GDF15 in some contexts because the protein is part of the body’s stress signal to caloric scarcity. Senolytics, NAD+ precursors, and mitochondrial-cofactor stacks have not been shown to durably move GDF15 in humans in any way that maps to outcomes.

Evidence

Evidence tier: Observational (human, large) for the age and mortality associations; Mechanistic / animal model for the cellular-stress-integrator framing; no intervention RCT has tested whether lowering GDF15 in healthy adults improves longevity outcomes.

The age signal is the strongest finding in the literature. Tanaka and colleagues, in a 2021 analysis pooling several large cohorts, reported that GDF15 had the largest age effect of any plasma protein they examined. Serum levels rose roughly threefold to fourfold across the adult lifespan, even after adjustment for sex, BMI, and traditional risk factors. Conte and colleagues’ 2022 review in Ageing Research Reviews synthesized that age signal alongside the cellular-stress-integrator biology, drawing on the Italian centenarian and offspring cohorts. It remains the most useful single starting reference on GDF15 in aging.

The mortality signal has been validated in several large prospective cohorts. Wiklund and colleagues in 2010 reported that elevated GDF15 was associated with all-cause mortality in older women independent of traditional risk factors. Subsequent studies in coronary disease cohorts, heart failure cohorts, and renal disease cohorts found the same independent prognostic value. A 2025 Hypertension analysis of the Sardinian SardiNIA cohort (n=4,736) reported that baseline GDF15 was independently associated with all-cause mortality over 10 years. The association survived adjustment for age, sex, hypertension status, eGFR, and several cardiovascular and renal-stress biomarkers.

The mechanistic story has been updated by recent work on cell-free mitochondrial DNA. A 2025 bioRxiv preprint reported that circulating cell-free mitochondrial DNA correlates with GDF15 across human aging and that both rise together during periods of mitochondrial stress. That supports the framing that GDF15 indexes accumulated mitochondrial damage rather than acting as a free-standing inflammatory marker. The work is preprint, not yet peer reviewed, and the precise causal sequencing remains open.

The relationship between GDF15 and epigenetic age clocks has been examined cross-sectionally. A 2024 Biogerontology analysis reported modest correlations between GDF15 and several second-generation methylation clocks (GrimAge, PhenoAge) in adult cohorts. The correlations weakened after adjustment for chronological age and chronic disease burden. The result is consistent with GDF15 and the methylation clocks capturing overlapping but distinguishable facets of biological aging.

The metformin elevation is well documented and dose-related. Studies in type 2 diabetes patients have shown chronic metformin use roughly doubles serum GDF15 versus matched non-users; the elevation tracks dose and persists for as long as the drug is taken. Mechanistic work has linked the drug-induced elevation to GDF15’s GFRAL hindbrain receptor, which mediates appetite suppression and may account for part of metformin’s modest weight effect. For aging research, this is a confound: the TAME hypothesis (testing metformin for healthy aging) has had to consider whether GDF15-mediated anorexia is part of the proposed mechanism or a side effect to control for.

What the evidence does not support is the inverse claim. No published RCT has shown that lowering GDF15 in healthy adults extends lifespan, slows disease incidence, or improves any patient-centered longevity endpoint. The protein has been a strong observational risk marker for fifteen years without becoming a longevity intervention target. The clinical-trial pipeline that does target GDF15 is in oncology: anti-GDF15 antibodies for cancer cachexia, where high GDF15 drives appetite loss. None of that work has produced a longevity-relevant intervention RCT.

How It Plays Out

A 58-year-old reader orders an extended longevity panel and finds GDF15 reported at 1,420 pg/mL, with a lab reference range listed as “less than 1,200 pg/mL.” The reader’s hsCRP is 1.1 mg/L, ApoB is at target, Lp(a) is unremarkable, eGFR is 88, hemoglobin A1c is 5.5%, and resting blood pressure is 118/76. The age-adjusted percentile for that value is roughly the 60th-70th percentile, slightly above the cohort median for the age band but not alarming. The disciplined response is to record the value, repeat in 6-12 months under stable conditions, and not change the existing plan based on a single reading.

A 64-year-old reader on metformin 1,000 mg twice daily for prediabetes finds GDF15 reported at 2,800 pg/mL. The first interpretation is the drug effect, not the aging signal. The clinical conversation is whether metformin is still doing what the prescriber wanted: the A1c trend, the weight trend, and the side-effect picture. The question is not whether GDF15 should be “lowered.” The number adds nothing to the metformin decision in this reader and can’t serve as an aging readout while the drug is on board.

A 51-year-old reader recovering from a viral pneumonia three weeks ago finds GDF15 at 2,100 pg/mL on a longevity panel ordered during recovery. The acute illness has driven the value upward, and a measurement during recovery cannot be read as a baseline. The right response is to retest in 8–12 weeks once the inflammatory recovery is complete, not to attribute the elevation to accelerated aging.

A clinician using GDF15 in a structured longitudinal way might track it once a year alongside the rest of an annual panel. The change that matters is not the absolute number but the trajectory. Stable GDF15 across several years, in an adult whose other markers are also stable, is different from a year-over-year increase of 20-30% in a person whose hsCRP and renal function are drifting in the same direction. The trajectory carries the information; a single isolated reading does not.

A longevity clinic that includes GDF15 alongside ApoB, Lp(a), hsCRP, IL-6, methylation age, and a coronary calcium score is doing something the reader should evaluate by the governance, not by the count of markers. Who explains how the markers relate to each other? What changes if GDF15 is high but other markers are clean? Who decides which markers warrant follow-up and which do not? If the answers are vague, the panel becomes a premium dashboard rather than a clinical instrument.

Consequences

Benefits. GDF15 supplies a different axis than any other widely available biomarker. Traditional inflammation markers (CRP, IL-6) track acute and chronic immune activity; ApoB and Lp(a) track atherogenic-particle risk; methylation clocks track an aggregated chronological-age or mortality signal in DNA. GDF15 sits closer to a slow-moving cellular-stress integrator. It captures accumulated mitochondrial, oxidative, and tissue-injury load across the months and years a single measurement reflects. For a reader building a stable risk picture, the protein adds information that the rest of the panel does not.

Used carefully, GDF15 can also discipline the conversation around metformin and aging. The drug raises the marker by mechanism, not by aging effect, and naming that relationship in advance prevents a metformin user from misreading the elevated number as evidence of accelerated decline. The same discipline applies in reverse. A low GDF15 in someone not on metformin is not proof that aging has been slowed, and a falling GDF15 over time should be checked against other markers before being attributed to a recent intervention.

Liabilities. The most common harm is treating GDF15 as a target. The protein is downstream of multiple upstream stress signals; titrating supplements or interventions against the GDF15 readout is Single-Biomarker Tunnel Vision with a fashionable marker. No intervention has been shown to durably lower GDF15 in healthy adults in a way that maps to outcomes. Treating the readout without addressing upstream causes, such as sleep, training load, body composition, glucose regulation, cardiovascular risk, chronic inflammation, or occult disease, is unlikely to change the underlying biology.

The second harm is testing pressure. GDF15 is a slow-moving signal, and quarterly testing surfaces noise as signal. Repeated measurement at short intervals will produce values that drift up and down because of assay variability, recent acute stressors, transient inflammation, recent training load, and ordinary biological variation. A reader who chases those movements with stack changes is on the Biomarker Treadmill regardless of whether the marker itself is useful.

The third harm is false alarm. A high GDF15 in someone with active but undiagnosed cancer, advanced heart failure, or progressing renal disease is a signal worth following clinically. The marker is doing its observational job, and the right response is a clinical workup, not a supplement intervention. A reader who reads “high GDF15” as “I’m aging fast” misses what the number is actually saying about the body’s current state.

The fourth harm is false reassurance. A GDF15 within the age-adjusted band doesn’t erase high ApoB, high Lp(a), low fitness, high visceral adiposity, hypertension, smoking exposure, sleep apnea, or family history of premature disease. The protein adds one axis; it doesn’t replace any of the others.

The useful posture is bounded. GDF15 is a slow-moving cellular-stress index. Its value comes from being one stable axis in a broader risk picture, read with the metformin and acute-illness confounds named, and not used to titrate interventions in healthy adults until intervention trials support that use.

Related Articles

Sources

• Conte, Maria, Cristina Giuliani, Antonio Chiariello, Vincenzo Iannuzzi, Claudio Franceschi, and Stefano Salvioli. “GDF15, an Emerging Key Player in Human Aging.” Ageing Research Reviews 75 (2022): 101569. https://doi.org/10.1016/j.arr.2022.101569
• Tanaka, Toshiko, Ann Z. Moaddel, Marta Zukley, Eleanor M. Simonsick, Stephanie Studenski, and Luigi Ferrucci. “GDF15 as a Biomarker of Ageing.” Experimental Gerontology 146 (2021): 111228. https://doi.org/10.1016/j.exger.2021.111228
• Wiklund, Fredrik E., Antony M. Bennet, Per K. Magnusson, Ulrika K. Eriksson, Filip Lindmark, Lars Wu, Niranjana Yaghoutyfam, et al. “Macrophage Inhibitory Cytokine 1 (MIC-1/GDF15): A New Marker of All-Cause Mortality.” Aging Cell 9, no. 6 (2010): 1057-1064. https://doi.org/10.1111/j.1474-9726.2010.00629.x
• Casati, Martina, Beatrice Arosio, Margherita Concas, Marco Masala, Michele Marongiu, Luca Carrara, Maristella Steri, et al. “Growth Differentiation Factor-15 and All-Cause Mortality in a Sardinian Population.” Hypertension 82 (2025): 1023-1031. https://doi.org/10.1161/HYPERTENSIONAHA.124.24235
• Coleman, Mitchell B., Hayden L. Hyatt, Caitlin Frye, Andrew R. Sims, Lauren M. Berg, Brett A. Dolezal, Yuliya R. Lokshina, et al. “Cell-Free Mitochondrial DNA and GDF15 in Human Aging.” bioRxiv preprint, January 28, 2025. https://doi.org/10.1101/2025.01.28.635306
• Galkin, Fedor, Aleksandr Aliper, Evgeny Putin, Igor Kuznetsov, Vadim N. Gladyshev, and Alex Zhavoronkov. “Aging Clocks, Entropy, and Biological Age: A Cross-Sectional Analysis with GDF-15 and Epigenetic Clocks.” Biogerontology 25 (2024): 1015-1031. https://doi.org/10.1007/s10522-024-10165-z
• Coll, Anthony P., Michael Chen, Pranali Taskar, Debra Rimmington, Satish Patel, John A. Tadross, Irene Cimino, et al. “GDF15 Mediates the Effects of Metformin on Body Weight and Energy Balance.” Nature 578, no. 7795 (2020): 444-448. https://doi.org/10.1038/s41586-019-1911-y

Medical and Legal Boundary

This entry is a reference, not medical advice. It describes published evidence, regulatory status, and common clinical practice patterns. It does not diagnose, prescribe, or replace a clinician’s judgment for a specific person.

GDF15 is a laboratory measurement used in aging research, cardiovascular and oncology prognostication, and some longevity-clinic panels. It is not an FDA-approved diagnostic for aging or any longevity indication. Values should be interpreted by a qualified clinician in the context of age, sex, medication list (especially metformin), recent acute illness, kidney function, known cardiovascular or oncologic disease, and the rest of the lipid, glycemic, and inflammatory panel. A markedly elevated value in an unexplained context warrants a clinical workup for upstream contributors, not a self-directed supplement intervention. This entry does not recommend starting, stopping, or changing any specific therapy, supplement, or diagnostic schedule.