Biomarkers worth tracking after 35 (and the ones to ignore)

Five years ago, a comprehensive blood panel ran around 50 markers. Today, several consumer panels offer 100-200 markers for a few hundred dollars, and direct-to-consumer continuous glucose monitors, sleep trackers, and at-home hormone panels are everywhere. The marketing message is that more data equals better health decisions. The actual evidence is more limited.

Most biomarkers carry redundant or weak prognostic information. Many of them are tightly correlated with simpler markers, so adding them to a panel does not change the picture. A meaningful fraction shift around enough between draws that the reading on a given day reflects noise more than signal. And many of them have no validated treatment that responds to changes in the number, which means seeing the result does not actually inform a decision.

The useful question for any biomarker is: does this number change my plan? If the answer is no, the test is mostly entertainment. If the answer is yes, it is worth running.

For a generally healthy 35-55 year old man, a tight panel with strong evidence behind it includes the following: a comprehensive metabolic panel (CMP) for kidney function, liver enzymes, and electrolytes; a complete blood count (CBC) for anemia, infection, and platelet abnormalities; a fasting lipid panel including apoB; HbA1c and fasting glucose; high-sensitivity CRP (hsCRP) for systemic inflammation; thyroid stimulating hormone (TSH); and a one-time Lp(a) measurement.

For specific contexts, additional markers earn their slot: a vitamin D level if there is reason to suspect deficiency; ferritin in patients with possible iron issues; a uric acid level in patients with gout history or metabolic syndrome; and a coronary calcium score (CT-based) for men in their forties or fifties with elevated cardiovascular risk.

That is roughly 15 numbers. They cover cardiovascular risk, metabolic health, basic organ function, and inflammation. Adding 100 more biomarkers on top of this baseline rarely changes management. The marginal information is small and the cost of confusion (chasing false positives, treating numbers instead of patients) is real.

Three biomarkers carry an outsized share of the long-term predictive value: apoB, hsCRP, and HbA1c.

apoB measures the number of atherogenic particles in your bloodstream. It is the most accurate single number for cardiovascular risk and it predicts events better than LDL-C in nearly every comparison. A reasonable target for a healthy adult is below 90 mg/dL; lower for patients with established disease or strong family history.

hsCRP is a marker of systemic inflammation. Persistently elevated hsCRP (above 2.0 mg/L) is associated with higher cardiovascular and mortality risk independent of cholesterol numbers. It is not a perfect measurement (it is sensitive to acute illness, recent infection, and intense exercise), but a sustained high level is a real signal worth investigating.

HbA1c reflects the average blood sugar over the prior 2-3 months. It is the standard marker for diabetes diagnosis and it captures pre-diabetes and impaired glucose handling well before fasting glucose alone. An HbA1c of 5.7-6.4 is pre-diabetic; over 6.4 is diabetic. For non-diabetic adults, HbA1c trends are useful for tracking metabolic health across years.

These three together give you most of the long-term cardiometabolic picture. Adding the rest of the basic panel for completeness is reasonable. Adding 50 more obscure markers usually does not.

Hormone panels are popular and frequently misused. The most common scenario is a middle-aged man requesting a testosterone panel because he feels tired or has lost interest in workouts and assumes the cause is low T. In a meaningful percentage of these cases, the actual cause is poor sleep, stress, depression, or general deconditioning, and the testosterone number is normal or borderline.

Total testosterone in adult men typically ranges from 300 to 1000 ng/dL, with significant variation by lab assay, time of day, and individual. A single low-ish number is not by itself a diagnosis. The clinical picture matters: persistent fatigue, loss of morning erections, reduced libido, depressed mood, loss of muscle and increase in fat, all over a sustained period, alongside repeatedly low total or free testosterone, is the picture that justifies a testosterone replacement conversation.

For most men whose lifestyle is intact (good sleep, regular strength training, normal weight, low alcohol, managed stress), testosterone levels are in a reasonable range. For men whose lifestyle is broken in those domains, the testosterone number tends to drift down, and the most effective intervention is fixing the lifestyle, not skipping to TRT.

Other hormone panels (cortisol, thyroid hormones beyond TSH, growth hormone, insulin) have narrower indications. Out-of-context "full hormone panels" advertised online are mostly entertainment for healthy patients and produce more confusion than insight.

Two of the strongest single predictors of healthspan and lifespan in middle-aged adults are not blood markers. They are VO2 max (a measure of cardiovascular fitness) and grip strength (a proxy for overall muscle strength and function).

VO2 max is the maximum rate at which your body can use oxygen during exercise. It captures cardiovascular function, mitochondrial function, and overall fitness in a single number. Across multiple large cohort studies, VO2 max is one of the strongest predictors of all-cause mortality and cardiovascular mortality, often outperforming traditional risk factors. The good news is that it is highly trainable, even in middle age. Regular aerobic training, particularly with periodic higher-intensity intervals, raises VO2 max meaningfully across months.

Grip strength is a simple test (squeeze a dynamometer as hard as you can) that has been studied across hundreds of cohorts. Lower grip strength predicts higher mortality, higher cardiovascular event rates, higher fall rates, and worse cognitive trajectories in older adults. Grip strength reflects total-body muscle mass and neuromuscular function, both of which are trainable through resistance exercise.

Neither of these requires a lab visit. VO2 max can be estimated with a few minutes on a rower or treadmill in any reasonable gym, and grip strength can be measured with an inexpensive dynamometer. Tracking these once or twice a year is more useful than tracking most of the obscure blood markers people get excited about online.

Continuous glucose monitors have become a popular consumer product. The original use case was diabetes management, where they are unambiguously useful. The new pitch is for non-diabetic adults to wear them for 2-4 weeks to "see what spikes their blood sugar" and adjust eating accordingly.

There is some real signal here. CGMs do reveal individual variation in glucose responses to specific foods. They can highlight patterns that fasting glucose and HbA1c miss. Some patients find the visual feedback motivating in ways that abstract numbers are not.

There is also a lot of noise. Glucose spikes after meals are not, by themselves, harmful in non-diabetic adults; they are the body doing what it is supposed to do. Many of the recommendations that come out of CGM data (eat protein first, walk after meals, avoid bananas) are reasonable but were already supported by simpler evidence. And chasing a flat glucose curve in a healthy adult can drift into disordered eating territory if taken too literally.

The honest assessment is that a 2-4 week CGM trial is potentially useful for a curious adult who wants concrete data on how their body responds to specific foods, and not particularly useful as a long-term tool unless there is an actual metabolic concern. For most healthy adults, fasting glucose and HbA1c twice a year is enough.

For a stable, generally healthy adult, once-a-year testing of the core panel is enough. Numbers do not change quickly when nothing is changing, and frequent testing of stable values produces noise without insight.

Quarterly testing makes sense during active interventions: starting a new medication, working on weight loss, changing diet patterns, training a new fitness goal. The shorter intervals let you see whether the intervention is moving the relevant numbers and adjust if it is not.

After a significant change has stabilized, it is reasonable to drop back to once a year. Most middle-aged men do not need to be on a quarterly testing cadence indefinitely.

The biggest practical mistake in the longevity space is treating biomarker tracking as a substitute for the underlying actions. Sleep, exercise, nutrition, and not smoking are the durable interventions. The numbers are useful for confirming that the actions are working and for catching things you cannot feel. They are not useful as a replacement for doing the actions in the first place.

A useful frame for any new biomarker, panel, or wearable that catches your attention is to ask three questions before paying for it. What decision will this number help me make? What is the evidence that the number is reproducible and meaningful in someone like me? And what would I actually do differently with this information that I am not already doing? If any of those answers is unclear, the honest move is usually to skip it and put the same time and money into the boring fundamentals.