Reference range is not the same as optimal
A laboratory reference range is the middle 95% of the people that laboratory tested — a population that includes the sedentary, the pre-diabetic and the unwell. Being "in range" means you are not unusual, which is a low bar in a population where most adults are overweight. An optimal range is the band the prospective cohorts and Mendelian-randomisation studies associate with the lowest risk of the diseases that end healthspan. The two are often far apart, and the gap is where most of the useful work happens.
One caution runs through this article: optimal ranges are population associations, not personal targets. Some are U-shaped, some shift with age, and chasing a number past the point of evidence is its own mistake. Read the table with your physician, not against them.
The blood panel
| Marker | Typical reference | Range associated with lower risk | Why it is on the list |
|---|---|---|---|
| ApoB | < 1.3 g/L | < 0.8 g/L; < 0.6 if high risk | Counts every atherogenic particle. Causal for heart disease in Mendelian randomisation; the lipid signal in Interheart was ApoB/ApoA1, not total cholesterol. |
| hsCRP | < 3 mg/L (some labs < 5) | < 1 mg/L | Chronic low-grade inflammation. AHA bands: < 1 low, 1–3 average, > 3 high. Rose with event risk in JUPITER and in the Women's Health Study. |
| HbA1c | < 42 mmol/mol (6.0%) | < 36 mmol/mol (5.4%) | Three-month average glucose. Mortality rises steadily from the mid-5s in EPIC-Norfolk and NHANES, well below the diabetes cut-off. |
| Fasting insulin | < 25 µIU/mL | < 8 µIU/mL | Rises a decade before glucose does. Reference ranges are famously permissive. Combine with glucose as HOMA-IR; under 1.0 is the band to aim for. |
| Cystatin C | eGFR > 60 | eGFR > 90 | Kidney function independent of muscle mass. Creatinine over-estimates function in people who have lost muscle and under-estimates it in people who have gained it. The only kidney marker we accept. |
| Ferritin | 30–400 µg/L (men) | 50–150 µg/L | Iron stores. Low explains a surprising share of fatigue; high is a common finding in men over 40 and is associated with insulin resistance and liver fat. Also rises with inflammation, so read with hsCRP. |
| Vitamin D (25-OH) | > 50 nmol/L | 75–125 nmol/L | Deficiency is associated with weaker bones, muscle and immunity. Supplement trials in replete people show little (VITAL, 25,871 adults), so the goal is correcting a deficiency, not mega-dosing. |
| Testosterone, total and free | 8.6–29 nmol/L (men) | Mid-range for age; symptoms decide, not the number | Falls about 1–2% a year after 40. The EMAS cohort tied symptoms to total < 11 nmol/L. SHBG decides how much is free and rises with age, so free T matters more than total. |
| Estradiol | Men: 40–160 pmol/L | Men: 70–120 pmol/L; women: by phase and stage | Men need some — bone, libido, mood. Crushing it with an aromatase inhibitor is a classic mistake. In women it maps the perimenopause transition with FSH. |
| IGF-1 | Age-adjusted, wide | Mid-range for age | Growth-hormone output. U-shaped with mortality in a 2011 meta-analysis of 12 cohorts — both low and high worse. Essential before and during any GH-axis protocol. |
| Thyroid (TSH, free T4) | TSH 0.4–4.0 mIU/L | TSH 0.5–2.5 in younger adults; higher is normal after 70 | Fatigue and weight are thyroid questions before they are anything else. Over-treating mild elevations in the elderly showed no benefit in the TRUST trial. |
| Full blood count | Haematocrit < 0.50 | Haematocrit < 0.50 | The safety brake on any testosterone protocol: above 0.54 the plan stops. |
Lipoprotein(a) deserves one line of its own: genetic, measured once in a lifetime, and if it is high the ApoB target drops. It costs little and most people have never had it checked.
The physical panel
Blood tells you about risk. These three tell you about function, and function is what healthspan actually is.
| Measure | How | Benchmark | Evidence |
|---|---|---|---|
| VO2max | Lab test with a gas mask, or a validated field test; watch estimates are within about 10%. | Top 25% for age and sex; "elite" is above the 97th percentile. | Mandsager 2018 (122,007 adults): bottom quartile vs elite carried a five-fold mortality difference. See the VO2max article. |
| Grip strength | Hand dynamometer, best of three, both hands. | Men > 40 kg, women > 25 kg are rough floors; aim for the top third for age. | PURE (139,691 adults): each 5 kg less grip, 16% higher all-cause mortality — a better predictor than systolic blood pressure. |
| DEXA | Whole-body scan: lean mass, fat mass, visceral fat, bone density. | Appendicular lean mass index above the sarcopenia cut-offs (7.0 kg/m² men, 5.5 women); visceral fat under about 1 kg; bone T-score above −1. | Low lean mass and high visceral fat are each independently associated with mortality and with insulin resistance. Also the only honest way to see what a diet or a compound did. |
What to measure
Everything above, once, is the baseline. After that the calendar depends on what you are changing:
- Every 12 months, always: the full blood panel, VO2max, grip, DEXA, home blood pressure.
- Every 3–6 months while something is changing: ApoB and HbA1c if you have changed diet, training or a lipid-lowering plan; fasting insulin and HbA1c during fat loss; ferritin if you were low or high; hormones after any hormonal change.
- Mid-cycle during any research protocol: the markers the protocol names — typically IGF-1 and HbA1c on a GH-axis plan, haematocrit and estradiol on anything hormonal, cystatin C on everything. The baseline-labs guide lists the house panel and the week-6 repeat.
- Once in a lifetime: lipoprotein(a), and a coronary calcium score from about 45 if ApoB or family history argue for it.
Draw fasting, in the morning, at the same laboratory each time, having not trained hard the day before — a hard session raises CRP, creatine kinase, liver enzymes and even cystatin C for 24–48 hours and produces a scare that is not real.
Where peptides and therapies fit
A biomarker panel is how a research protocol is both justified and policed. Several compounds on this site have a marker that decides whether they are sensible: the GH-axis pairings such as CJC-1295 / ipamorelin are gated on IGF-1 and HbA1c; tesamorelin is studied specifically for visceral fat, which only DEXA can see; the longevity goal page describes how the mid-cycle repeat decides whether a dose holds, steps or stops. Clinic therapies follow the same logic — an NAD infusion or a sauna block is a reasonable addition to a plan with a known baseline and a pointless one without.
Common mistakes
- Reading creatinine as kidney function. Muscular people and creatine users get flagged, sarcopenic people get missed. Cystatin C, always.
- Testing total cholesterol and LDL-C instead of ApoB. They disagree in about one adult in five, usually the ones with insulin resistance, and ApoB is the one that tracks risk.
- Accepting "normal" fasting glucose with an unmeasured insulin. Glucose is the last marker to move.
- Testing hormones at 4 p.m. after a bad night. Testosterone is a morning, rested, fasting measurement, repeated once before any conclusion is drawn.
- Running two hundred markers. More data is not more information. A panel you cannot act on is a bill.
- Changing three things and testing once. You will never know which one worked.
Frequently asked questions
Which biomarkers of ageing matter most?
ApoB, hsCRP, HbA1c, fasting insulin and cystatin C in blood; VO2max, grip strength and a DEXA scan for function and body composition. Ferritin, vitamin D, thyroid, testosterone or estradiol and IGF-1 complete the baseline. That dozen changes decisions; most of the rest does not.
What is the difference between a reference range and an optimal range?
The reference range is the middle 95% of the people a laboratory tested, which includes the unwell. The optimal range is the band that large cohorts associate with the lowest risk. For ApoB, HbA1c and fasting insulin the two can be far apart.
How often should I test?
The full panel once a year. Markers you are actively changing — ApoB, HbA1c, insulin, ferritin, hormones — every three to six months. During a research protocol, whatever the protocol names at mid-cycle. Lipoprotein(a) once in a lifetime.
Why cystatin C rather than creatinine?
Creatinine is a by-product of muscle, so anyone with a lot of muscle, or on creatine, reads as having worse kidneys than they do, and anyone who has lost muscle reads as better. Cystatin C is independent of muscle mass. It is the only kidney marker we accept before a protocol.
Are epigenetic age tests worth it?
They are interesting and improving, but results differ by years between vendors and between weeks, and no decision in a protocol changes because of one. Treat them as a trend over years, if at all, and spend the money on a DEXA and a VO2max test first.
Want a plan built on your own numbers?
Upload your labs with the intake and the protocol is written against them — every threshold read by a reviewer, every result walked through with a coach on WhatsApp. €399, follow-up included.
Build my protocol Book a 15-minute call