longevity researchLongevity biomarkers panel: what to test, what to skip
A practical longevity biomarkers panel — the markers with mortality data behind them, and the popular tests that fail to add information.
The Mandsager 2018 analysis tracked 122,007 patients undergoing graded exercise testing at the Cleveland Clinic for an average of 8.4 years. The mortality difference between the lowest and highest cardiorespiratory fitness quintiles was 5-fold. That gap is larger than the mortality difference between smokers and non-smokers. No blood marker, no scan, no genetic test currently competes with VO2 max as a single predictor of all-cause death.
That single finding should reshape how a longevity panel is built. Most direct-to-consumer "longevity testing" services lead with 30+ biomarkers, of which 5 carry the bulk of the prognostic signal. The other 25 add cost, anxiety, and noise without changing what a competent clinician would recommend. This guide separates the markers with mortality data from the markers that sound sophisticated but fail to predict anything actionable.
The five tests that carry the signal
These five tests cover the largest fraction of modifiable mortality variance for the lowest combined cost. Order them first, optimize them, then consider whether anything else is worth adding.
1. VO2 max via graded exercise test. Roughly $150-300 in most clinical exercise labs. The Mandsager data shows hazard ratios for all-cause mortality of 5.04 for low fitness vs. elite fitness, with continuous improvement across the range. Even moving from below-average to above-average fitness produces a hazard ratio improvement of 1.41. Sub-maximal estimates from wearables carry 10-15 percent error and are useful for tracking trends but not for absolute classification. See the VO2 max biomarker page and the VO2 max longevity link for the mechanism detail.
2. ApoB. Apolipoprotein B counts the total number of atherogenic lipoprotein particles in circulation. Allan Sniderman's 2019 JAMA Cardiology analysis demonstrated that ApoB outperforms LDL-C in roughly 20 percent of patients due to particle-cholesterol discordance, particularly in people with metabolic syndrome or high triglycerides. The test costs $25-40 and is widely available.
3. HbA1c and fasting insulin. HbA1c reflects average glucose over the prior 90 days and has a continuous association with mortality even within the "normal" range. The Selvin 2010 NEJM analysis showed that HbA1c values between 5.7 and 6.4 percent carried a 1.7-fold higher cardiovascular mortality risk than values under 5.5 percent. Fasting insulin adds resolution to early insulin resistance that HbA1c can miss for years. See the fasting glucose optimization guide for the optimization detail.
4. High-sensitivity CRP (hs-CRP). Paul Ridker's two decades of work, culminating in the CANTOS trial, established that residual inflammatory risk persists after LDL is lowered, and that hs-CRP captures it. Levels above 2 mg/L mark elevated cardiovascular and all-cause mortality risk. Levels below 1 mg/L mark the favorable range. The test costs $15-30 and is one of the most informative single inflammation markers available.
5. Lp(a), once. Lipoprotein(a) is roughly 90 percent genetically determined and stable across adult life. Levels above 75 nmol/L (or 50 mg/dL by mass) carry independent cardiovascular risk that no lifestyle intervention reliably modifies. Knowing the number changes prevention strategy, even though no targeted therapy is yet approved. Test it once. Do not retest annually.
Three secondary tests worth considering
Homocysteine. Elevated levels (>10 µmol/L) are associated with cardiovascular and dementia risk. The mechanism may involve endothelial damage and methylation. B-vitamin supplementation lowers levels but the mortality benefit in trials has been mixed. Test if family history of premature cardiovascular disease or cognitive decline is present.
Vitamin D (25-OH). Deficiency below 20 ng/mL is associated with elevated mortality across multiple cohorts, though the causal relationship remains debated. The Mendelian randomization data is weaker than the observational data, which suggests reverse causation explains part of the association. Still cheap to test and cheap to correct.
eGFR and cystatin C. Kidney function decline predicts cardiovascular and all-cause mortality with strength comparable to traditional risk factors. Cystatin C is a more sensitive early marker than creatinine-based eGFR alone, particularly in muscular individuals or older adults.
A useful longevity panel does not measure everything. It measures the markers with mortality data behind them, and ignores the markers that sound impressive but predict nothing.
What to deprioritize or skip
A surprising amount of popular longevity testing fails to add prognostic information once the core panel is run.
Methylation-based biological age tests. GrimAge, PhenoAge, and similar tools have legitimate research validity. The problem is measurement noise. Morgan Levine's own 2018 paper acknowledges that single-reading variance can be 2-5 years, which makes interpreting a single result difficult and tracking small interventions nearly impossible. They become useful only with repeated annual readings showing a directional trend. For most people, optimizing ApoB, HbA1c, hs-CRP, and VO2 max produces more actionable change than chasing an epigenetic age number.
Telomere length. Strong association in cross-sectional studies, weak prospective predictive value once age is controlled for. Commercial telomere tests have not been shown to inform clinical decisions.
Heavy metal panels. Useful in occupational exposure cases. Not useful as a routine longevity screen in adults with no exposure history.
Comprehensive hormone panels in asymptomatic men. Free testosterone, estradiol, DHEA-S, and IGF-1 are worth testing when symptoms are present. As a longevity ranking exercise in asymptomatic adults, the mortality data is weaker than for the core panel. Optimize them for function and symptoms, not for a longer expected lifespan.
Most genetic risk scores. Polygenic risk scores for cardiovascular disease and dementia are real and improving, but their incremental value over a well-measured ApoB, hs-CRP, and family history is currently small in clinical settings. Worth doing once if cost is not a barrier. Not worth obsessing over.
The longevity extension protocol integrates these markers into a structured optimization sequence. See also the biological age testing breakdown for the epigenetic measurement detail.
Frequency
Annual testing is sufficient for the core panel in adults without active disease. ApoB, HbA1c, hs-CRP, fasting insulin, and lipid subfractions all show meaningful change over 3-12 months when interventions are working. VO2 max can be retested every 6-12 months. Lp(a) and one-time genetic panels do not need repeating.
Resist the urge to test every 4-8 weeks. Biomarker noise — particularly hs-CRP, which can spike with infection or training — produces apparent changes that are not real. Annual testing under standardized conditions (fasted, similar time of day, no recent illness, no acute training within 48 hours) produces more interpretable data than monthly testing under variable conditions.
Protocol
- Order the core five. VO2 max test, ApoB, HbA1c + fasting insulin, hs-CRP, and Lp(a). Total cost in the US: roughly $250-500 out of pocket if uninsured.
- Establish baseline under standardized conditions. Fasted, morning, no acute illness or hard training in the prior 48 hours.
- Add secondary tests only with a reason. Homocysteine if family history. Vitamin D if symptoms. eGFR routinely.
- Skip the noise. Methylation age, telomere length, heavy metals, and broad hormone panels unless symptoms or specific indications drive the order.
- Retest annually. Quarterly testing wastes money on biomarker noise.
- Track interventions one at a time. Changing diet, exercise, and three supplements simultaneously makes it impossible to attribute changes to any one input.
- Use absolute thresholds, not just trends. ApoB under 80 mg/dL, hs-CRP under 1 mg/L, HbA1c under 5.5 percent, VO2 max in the top two age-adjusted quintiles. These are defensible targets backed by mortality data.
A longevity panel that fits on one page and costs under $500 will outperform a 60-marker dashboard for almost everyone. The signal-to-noise ratio is what matters, and the signal is concentrated in a small number of tests.