Three Longevity Experts Debate Which Biomarkers Actually Matter
Three longevity specialists with very different backgrounds compare notes on blood work: what it can tell you, what it cannot, and where each of them disagrees. The recurring theme is context — no single number tells the story, and chasing one marker in isolation can quietly damage another system. They converge on a surprisingly cheap core panel plus repeat measurement over time.
Overview
Siim Land brings together three people who look at blood work from opposite ends of the telescope: Matt Kaeberlein, a biologist of ageing turned health-technology CEO; Brad Stanfield, a practising GP in Auckland working inside public-health budgets and clinical guidelines; and Michael Lustgarten, a former research scientist who has run dozens of his own blood panels alongside tracked diet data.
Brad's framing anchors the conversation — measure things you can actually change, where good evidence links the change to a better outcome. Michael pushes further, arguing that trends and rate of change over many tests reveal more than any single reading inside a reference range. Matt adds a systems view: blood work is one signal among several, alongside body composition, strength, mobility and cognitive assessment.
The panel disagrees openly about hormone screening, continuous glucose monitors in non-diabetics, and how much value biological age clocks really carry. All three warn that optimising one number can harm another — insulin driven too hard causes hypos, very low-carb diets can flatten triglycerides while pushing apoB up. On lipids they largely converge on apoB, with Michael arguing that lipoprotein(a) and VLDL sharpen the picture further. They close on cost, and agree the essentials are far cheaper than the headline panels suggest.
Key quotes
5This idea that if you live a healthy lifestyle, everything's going to be hunky-dory and nothing's ever going to go wrong — that's just not the way it works in the real world.
If you try and target these biomarkers too much, you can actually do unintended harm.
I'm a big fan of the philosophy: measure, intervene, measure again.
You can be in the reference range and your data getting worse over time, and you're still in the reference range.
If your biomarkers are great but you feel terrible — what are you doing?
Key ideas
9Measure what you would actually act on
Brad's clinical filter is simple: test something only if a change in that number reliably changes an outcome. Blood pressure and apoB pass that test because lowering them is linked to fewer heart attacks and strokes.
Reference range is a low bar
Michael points out that a value can drift steadily worse for years while never leaving the normal range. The direction and speed of travel carry information that a single in-range result hides completely.
Blood work is one signal, not the whole picture
Matt treats blood work as one input in a systems view of health, alongside strength, mobility, body composition by DEXA and cognitive assessment. Any single number is probabilistic, never definitive.
Optimising one marker can damage another system
Pushing HbA1c down hard with insulin caused hypoglycaemic falls in Brad's patients. Very low-carb diets can produce beautiful triglycerides while apoB climbs. The panel's rule is to watch the net systemic effect.
The hormone disagreement
Matt suspects population-level androgen decline is under-measured and worth investigating; Brad counters that guidelines only support testosterone testing when there is clinical suspicion, and that trial results for replacement were underwhelming.
Not all atherogenic particles are equal
Michael cites Mendelian randomisation work suggesting VLDL is roughly four to five times more atherogenic than LDL, and lipoprotein(a) six to seven times. Two people with identical apoB can carry very different risk.
HDL follows a J-curve
Both very low and very high HDL associate with elevated risk, and HDL tends to drift down with age. Brad notes that drugs raising HDL failed in trials, so the number is most useful inside other calculations.
Insulin alone can mislead
Insulin traces an inverted U across the lifespan, peaking in the early thirties and falling again in very old age as beta-cell function fails. A low reading in isolation is not automatically good news.
GGT as a steadier liver signal
ALT and AST follow J-shaped curves where very low values are hard to interpret. Michael prefers GGT for a general read on liver status because it tracks more linearly with age.
Practical takeaways
6- 1
Start with the cheap essentials 32:00
A standard chemistry panel plus complete blood count already covers kidney, liver, immune and metabolic signals. Adding hs-CRP keeps it modest, and a blood pressure cuff costs less than most single tests.
- 2
If you can only add one lipid test, add apoB 35:00
All three agree apoB is the single most informative lipid marker, and Brad names it as the first extra test he would suggest a patient pay for.
- 3
Build a loop, not a snapshot 26:40
Measure, change something, measure again — and log the other variables in your life at the same time, or you will not know which change moved the number.
- 4
Read glucose curves, not spikes 1:08:00
A rise that falls back quickly reads very differently from one that stays elevated for hours. The panel warns against turning a CGM into a reason to cut out whole fruit and oats.
- 5
Set testing frequency to your goal 1:11:00
Brad works on six months to a year for otherwise healthy people; Matt prefers quarterly with a broader panel; Michael tests far more often because he is hunting for the fastest signal. All three call it individual.
- 6
Know yourself before you quantify 1:21:00
Matt's closing advice is that if measuring energises you, lean in — and if it generates anxiety, step back. Data that makes you miserable is not a health gain.
Topics & chapters
16Three perspectives at one table
Matt, Brad and Michael introduce their very different routes into blood work — basic science, general practice and long-run self-quantification.
Measure what you can change
Brad sets out the clinical filter he uses, with blood pressure and apoB as the model cases.
Biological age clocks enter the room
The panel calls them useful research tools while warning against chasing a lower score with supplements or single foods.
One piece of the puzzle
Michael places diet, movement, sleep and air quality ahead of blood work, then asks what is happening under the hood.
When optimising backfires
Insulin dosing, hypoglycaemia and the low-carb triglyceride-versus-apoB trade-off illustrate the cost of tunnel vision.
Measure, intervene, measure again
Matt's iterative framework, and Michael's addition: track the surrounding variables or the correlations are meaningless.
Pick only three
Brad chooses weight, blood pressure and how the person feels, then creatinine, a lipid panel and a full blood count.
The testosterone debate
The panel's sharpest disagreement: population screening for hormones versus guideline-led, symptom-triggered testing.
Lipids and apoB
Why apoB beats LDL as a single number, when the two diverge, and what lipoprotein(a) adds.
HDL, ageing and reverse causation
The J-curve, the failed HDL-raising trials, and why epidemiological associations mislead when comorbidities go unmodelled.
Blood sugar: insulin and HbA1c
Metabolic dysfunction as a driver of chronic disease, and why insulin needs HbA1c beside it to be readable.
Continuous glucose monitors
Strong in diabetes and pre-diabetes, contested in healthy people — educational tool or a route to needless food fear.
How often should you test?
Three different cadences, three different goals, and agreement that the answer is individual.
Liver markers
ALT, AST, albumin and bilirubin, and Michael's case for GGT — plus his own trial-and-error with green tea and fruit intake.
Obsession, anxiety and knowing yourself
Matt on backing off when data creates stress; Michael on why he happily lives inside his own numbers.
Cancer detection and the cost question
Multi-cancer early detection tests, inflammatory patterns around cancer, and the panel's shared conclusion that the essentials are cheap.