Most people have had blood work done, been told everything looks normal, and still felt tired, foggy, or off. The test was not wrong. The reference range simply answered a different question than the one being asked. A lab reference range describes where the middle 95% of a reference population falls. It is a statistical boundary, not a health target.
This hub covers the markers that change decisions: what to test, what the optimal target is rather than the lab-normal one, how markers group into systems, and how often each is worth repeating. Every guide here states a number you can compare your own result against.
The Verdict
Normal versus optimal — the framework underneath every guide here
A reference range is built by testing a reference population and reporting the central 95%. If that population is broadly unhealthy, the range is broadly unhealthy. Nothing in the method asks whether the result is good — only whether it is common. Normal vs optimal covers the full framework. The table below shows where the two diverge enough to change a decision.
| Marker | Typical lab-normal range | Optimal target | Why the gap matters |
|---|---|---|---|
| Fasting insulin | Usually not ordered; flagged only above ~25 µIU/mL | 2–5 µIU/mL | Rises 5–10 years before glucose does. The earliest measurable signal of insulin resistance. |
| Fasting glucose | 70–99 mg/dL | 75–85 mg/dL | A 96 mg/dL result is "normal" and also within a few points of prediabetes. |
| HbA1c | Below 5.7% | 5.0–5.4% | Reflects roughly 90 days of average glucose. Skewed high by short red-cell lifespan or anemia. |
| ApoB | Rarely ordered; LDL-C reported instead | Below 80 mg/dL (below 60 if high risk) | Counts atherogenic particles directly. Normal LDL-C with high ApoB is a common miss. |
| hsCRP | Below 3.0 mg/L | Below 1.0 mg/L | A single value above 10 usually means an acute infection, not chronic risk. Retest before acting. |
| Homocysteine | Below 15 µmol/L | Below 8 µmol/L | Elevation often reflects B12, B6, or folate status rather than a primary problem. |
| Vitamin D (25-OH) | 30–100 ng/mL | 50–70 ng/mL | The bottom of the reference range is set to prevent rickets, not to support hormone function. |
| Ferritin | 30–400 ng/mL | 50–150 ng/mL | Both ends carry risk. Ferritin also rises with inflammation, so pair it with hsCRP. |
| TSH | 0.45–4.50 mIU/L | 1.0–2.0 mIU/L | A TSH of 4.2 is inside the range and is where many people report fatigue and cold intolerance. |
| SHBG | 10–57 nmol/L | 20–45 nmol/L | High SHBG makes total testosterone look adequate while free testosterone runs low. |
| Omega-3 index | Not part of any standard panel | Above 8% | Measures red-cell EPA and DHA. Intake is a poor proxy — absorption varies two- to threefold. |
| Lp(a) | Not part of any standard panel | Below 75 nmol/L (below 30 mg/dL) | Genetically set. One lifetime test answers it; diet and exercise barely move it. |
Two cautions on reading this table. Optimal targets are population-level associations, not personal prescriptions — an ApoB of 85 means something different at 35 than at 65 with a family history. And reference ranges themselves vary between labs, because each lab derives them from its own instruments and its own population. Comparing a result from one lab against a target from another introduces error before interpretation even starts.
What standard panels leave out, and why
A typical primary-care panel runs 10–20 analytes: a complete blood count, a metabolic panel, a standard lipid panel, and often TSH. That set is designed to detect or rule out disease efficiently. It is not designed to characterize function.
The systematic omissions follow a pattern. Fasting insulin is absent because it carries no diagnostic threshold for diabetes, while glucose and HbA1c do — even though insulin rises first, often 5–10 years earlier. ApoB is skipped because LDL-C has longer guideline history and a cheaper assay. Lp(a) is skipped because it is not modifiable by lifestyle, which makes it feel actionless despite being a strong independent risk factor. Free testosterone, SHBG, homocysteine, and the omega-3 index are skipped because no reimbursement rule requires them.
None of that means the markers are uninformative. It means the standard panel optimizes for diagnosis under reimbursement constraints, and optimization asks a question reimbursement does not fund.
Group markers by system, not by test tube
A results portal lists markers alphabetically or by the panel they were bundled into. That ordering hides the thing that matters. Grouping by physiological system is what turns a list of values into a picture.
| System | Core markers | What a cluster signals | Retest cadence |
|---|---|---|---|
| Metabolic | Fasting insulin, fasting glucose, HbA1c, HOMA-IR, triglycerides, HDL | Insulin resistance developing years before a diabetes diagnosis | Every 3–6 months while intervening; annually once stable |
| Cardiovascular | ApoB, Lp(a), hsCRP, LDL particle number, blood pressure | Atherogenic particle burden that LDL-C alone underestimates | ApoB every 3 months on therapy; Lp(a) once per lifetime |
| Hormonal | Total and free testosterone, SHBG, estradiol, LH, FSH, TSH, free T3, free T4 | Binding-protein and pituitary problems masquerading as low output | Every 3–4 months while titrating; every 6–12 months once stable |
| Inflammatory | hsCRP, homocysteine, ferritin, white cell count, uric acid | Chronic low-grade inflammation driving vascular and cognitive aging | Every 6 months; sooner if a value was drawn during an illness |
| Nutrient | Vitamin D, B12, folate, RBC magnesium, omega-3 index, iron panel | Correctable deficiencies that explain fatigue and mood symptoms | 8–12 weeks after starting a supplement, then every 6–12 months |
Why a pattern across three markers beats one flagged value
Markers within a system move together because they share upstream drivers. That correlation is what makes patterns readable.
- Metabolic: high fasting insulin plus high triglycerides plus low HDL plus elevated waist circumference is one coherent picture of insulin resistance. Any one of those alone is ambiguous.
- Hormonal: normal total testosterone plus high SHBG plus low free testosterone explains why a "normal" result can accompany real symptoms. Reading total testosterone alone misses it entirely.
- Inflammatory: high ferritin plus high hsCRP usually means inflammation raised the ferritin, not that iron stores are genuinely high. Treating that as iron overload is the wrong intervention.
- Thyroid: elevated TSH plus normal free T4 plus low free T3 points at conversion rather than glandular output, which changes what to do about it.
The practical consequence: a panel is worth more than the sum of its markers only if something reads it as a set. That is the difference between a results dashboard and an interpretation, and it is the main axis on which testing platforms actually differ.
How often to retest
Retest cadence should follow the biology of each marker, not a subscription's billing cycle. HbA1c reflects roughly 90 days of average glucose, so retesting it at six weeks measures mostly the previous quarter and wastes the draw. Fasting insulin, hsCRP, and free testosterone respond within weeks and are worth checking at 8–12 weeks after a change. Vitamin D takes about 8–12 weeks to reach a new steady state on a fixed dose. Lp(a) needs one test, ever.
Draw conditions matter as much as timing. Fasting insulin and glucose require a genuine 10–12 hour fast. Testosterone should be drawn between 7 and 10 a.m., when levels peak — an afternoon draw can read 20–30% lower and produce a diagnosis that a morning draw would not support. Comparing results across different draw times or different labs is a common source of imaginary trends.
Biomarker guides
The advanced markers below are the ones worth adding to a standard panel. Guides to the markers already on the panels most people receive — the red cell indices and the liver and metabolic components of a CMP — follow underneath.
- ApoB — counts atherogenic particles; a stronger cardiovascular risk marker than LDL-C alone
- ApoB test cost — cash price, what insurance and Medicare cover, and how to order it without a physician
- Fasting insulin — the earliest signal of insulin resistance, rarely on standard panels
- Insulin resistance vs prediabetes: two conditions defined on different numbers, and the years between them
- Triglyceride to HDL ratio: the insulin-resistance surrogate you can calculate free from any lipid panel
- HOMA-IR — calculated from fasting insulin and glucose; quantifies resistance before glucose rises
- Homocysteine — associated with cardiovascular and cognitive aging; often a B-vitamin status marker
- hsCRP — chronic inflammation load, and a major driver of cardiovascular aging
- Lp(a) — inherited risk factor; test once, since it rarely changes with lifestyle
- Normal vs optimal ranges — the framework underneath every guide here
- Omega-3 index — red-cell EPA and DHA; more accurate than tracking fish oil intake
- SHBG — binds testosterone and controls bioavailability; explains normal-total, low-free results
- Heavy metals test — which sample is valid for which metal, and the two methods to avoid
Markers already on your standard panel
- CMP blood test — what a comprehensive metabolic panel measures, the six common misreadings, and the gaps it leaves
- CO2 blood test: the metabolic-panel line that measures bicarbonate rather than the gas you breathe out
- Creatine kinase: the muscle enzyme a hard workout raises for days
- Vitamin D blood test: which of the two assays to order, and where the 20 and 30 ng/mL thresholds came from
- ALT — liver-cell injury, and the clearest case of a reference range more generous than the evidence supports
- Alkaline phosphatase — grouped with the liver tests but produced equally by bone; GGT identifies the source
- MCV — average red cell size; sorts anaemia towards iron on one side, B12 and folate on the other
- RDW — red cell size variability; moves earlier than MCV, and tracks mortality risk inside the normal range
- MCHC — haemoglobin concentration per cell; insensitive to early iron deficiency, and a high result is usually a sample artifact
- MCH — average haemoglobin per red cell; confirms the story MCV tells, and outlasts it when a sample is delayed
- RBC — the red cell count every index above is derived from; a high count is usually contracted plasma, not extra cells
- MPV — average platelet size; only interpretable read against the platelet count, and it drifts in the tube
- AST — cell injury, but not liver-specific; muscle, heart and a haemolysed sample all raise it
- GGT — the marker that splits a raised alkaline phosphatase into liver or bone, and the most alcohol-sensitive enzyme on a panel
- BUN — filed as a kidney test, but hydration and dietary protein move it as often as the kidneys do
- eGFR — calculated from creatinine rather than measured, which is why muscle mass and creatine shift it
- TSH — a pituitary hormone, not a thyroid one; the clearest case where lab-normal and optimal diverge
- TIBC — with UIBC and transferrin saturation, the iron panel that stays readable when ferritin is distorted by inflammation
- Ferritin — iron stores, and the marker that falls first; inflammation pushes it back up and hides the deficiency
- Uric acid — the gout and kidney-stone marker; usually a story about kidney excretion and insulin resistance rather than diet
Which platforms test these
Coverage varies widely, and more markers does not automatically mean more useful. Every platform reviewed on this site is listed, alphabetically: Fountain Life (200+, clinic-based), Function Health (100+, self-directed), Hone Health (hormone panel plus therapy safety markers), InsideTracker (~48, with a recommendation engine), Lifeforce (~40 plus hormones), Marek Health (a la carte — you select the panel), Maximus (hormone panel plus therapy safety markers), Opt Health (55+, physician-led), and Superpower (100+, app-first). Ordering carries no editorial meaning. The full side-by-side is on the platform hub, ordered alphabetically.
First-mention official sites, alphabetically: Fountain Life, Function Health, Hone Health, InsideTracker, Lifeforce, Marek Health, Maximus, Opt Health, Superpower.
Related
- Biological age — what the tests measure and what they do not
- Doctor-led options — programs where a clinician reads the panel with you
- Longevity platforms compared — every option on the same rubric
- Supplement guides — what to take once you know what is low
Frequently Asked Questions
What biomarkers should I test for longevity?
A defensible core panel spans five systems: metabolic (fasting insulin, glucose, HbA1c, HOMA-IR), cardiovascular (ApoB, Lp(a), hsCRP), hormonal (total and free testosterone, SHBG, a full thyroid panel), inflammatory (hsCRP, homocysteine, ferritin), and nutrient status (vitamin D, B12, RBC magnesium, omega-3 index). That is roughly 20–25 analytes. Panels advertising 100+ markers add breadth, not necessarily decision-relevant information.
What is the difference between a normal and an optimal range?
A normal reference range is a statistical description of a reference population, usually the middle 95% of people tested at that lab. It answers "is this result unusual?" An optimal range is a target associated with the lowest measured risk or the best function. The two diverge most for fasting insulin, TSH, vitamin D, and ferritin — markers where a result can sit comfortably inside normal while still being far from ideal.
Why do standard doctor panels leave out fasting insulin and ApoB?
Standard panels are built around what changes clinical management under insurance reimbursement and existing guidelines. Fasting glucose and HbA1c already carry diagnostic criteria for diabetes; fasting insulin does not, so it is rarely ordered. LDL-C has decades of guideline history and a cheaper assay, so ApoB is often skipped even where it is the better risk marker. The omission is procedural, not a judgment that the marker is uninformative.
How often should I retest my biomarkers?
Match the cadence to the biology. Markers that respond to intervention within weeks — fasting insulin, hsCRP, vitamin D, free testosterone — are worth retesting at 8–12 weeks after a change. HbA1c needs at least 90 days to reflect anything, so retesting it at 6 weeks wastes a draw. Lp(a) is genetically determined and needs one lifetime test. Once markers are stable and in range, annual testing is usually enough.
Is one abnormal biomarker a problem?
Usually not on its own. Single-marker variability is large: fasting insulin can move 20–30% between draws, hsCRP spikes with any infection, and ferritin rises as an acute-phase reactant. A pattern across three or more related markers is far more informative than one flagged value. High fasting insulin plus high triglycerides plus low HDL is a coherent metabolic picture; an isolated borderline value is often measurement noise.
Do I need a physician to interpret biomarker results?
For a small, clean panel with everything in range, a careful reader can handle it. Interpretation gets harder when markers conflict — high ferritin with high hsCRP, normal total testosterone with high SHBG, elevated TSH with normal free T4. Those patterns require weighing several results against each other, and they are where testing-only platforms leave the work to you. Anyone on prescription medication or with a flagged result should have a clinician review the panel.