What Blood Tests Tell You About Health and Fitness

A useful fitness blood panel is about twenty markers, drawn four times a year, under conditions you keep identical every time. The panel is below.

• CBC with differential
• comprehensive metabolic panel
• ferritin
• transferrin saturation
• hs-CRP
• HbA1c
• fasting insulin
• fasting glucose
• lipid panel with ApoB and Lp(a)
• TSH with free T4
• total and free testosterone with SHBG
• 25-hydroxyvitamin D
• vitamin B12
• folate
• creatine kinase

That costs $150–$400 out of pocket depending on where you order it. The 160-marker panels sold as “wellness profiles” add mostly markers with poor analytic precision, no established reference change value, or no action you or a clinician could take from a single reading. The value is in repeat measurements of the same markers, drawn the same way, so you can distinguish a real trend from lab noise. Marker count is not where the value lies.

Everything below is measurement and interpretation. Deciding what an abnormal result means for you is a clinician’s job, and some of the results here are ones you should not sit on alone.

The panel, and why each marker is on it

Iron status: ferritin, serum iron, TIBC and transferrin saturation. This is the highest-yield block for anyone training seriously, and endurance athletes in particular. Ferritin is an acute-phase reactant, so it rises with inflammation and can look normal in someone who is iron-depleted. Order transferrin saturation alongside it (serum iron / TIBC) and read the two together. If ferritin is low-normal but saturation is under 20%, that combination is more informative than either number alone. Draw iron studies fasted and in the morning, and not within 24 hours of an iron-containing supplement. Serum iron swings wide with recent intake.

CBC with differential. Hemoglobin and hematocrit for oxygen-carrying capacity, MCV to separate iron-deficient from B12/folate-deficient patterns, and the white count differential. Plasma volume expands with endurance training, which dilutes hemoglobin. A hemoglobin that drifts down over a training block is not necessarily a loss of red cell mass.

Inflammation: hs-CRP. Use the high-sensitivity assay, because the standard assay’s floor is around 5 mg/L and everything interesting sits below that. The standard CRP assay is not sensitive enough for this. After hard eccentric work hs-CRP spikes for days, so a single elevated reading in an athlete is usually training. It is not usually pathology. The signal is the median of four quarterly draws taken under matched conditions.

Glucose regulation: fasting glucose, fasting insulin, HbA1c. Compute HOMA-IR from the first two (glucose mg/dL × insulin µIU/mL / 405). HbA1c is a 90-day average weighted toward the last month, and it is distorted by anything that changes red cell lifespan: iron deficiency raises it, hemolysis and recent blood loss lower it. If you have continuous glucose data, you already have better time-resolution than HbA1c gives you, and the two disagree often enough that the disagreement itself is informative.

Lipids: standard panel plus ApoB and Lp(a). ApoB counts atherogenic particles directly, which LDL-C only estimates. Lp(a) is largely genetically set, so measure it once, confirm it once, and then stop. Ordering it quarterly wastes money.

Endocrine: TSH, free T4, total testosterone, SHBG and free testosterone (calculated or equilibrium dialysis). Testosterone has a strong diurnal rhythm, so draw between 7 and 10 a.m. or the number means little. Free testosterone by immunoassay is unreliable. Either calculate it from total plus SHBG plus albumin, or pay for equilibrium dialysis.

Micronutrients: 25-OH vitamin D, B12, folate. Three with defensible assays and established deficiency thresholds. Most of the other vitamin and mineral panels sold in wellness bundles measure serum concentrations that are homeostatically buffered and tell you little about tissue stores.

Creatine kinase. Useful only in context. CK rises into the thousands after unaccustomed eccentric loading and means nothing on its own. It is interpretable as a training-load marker when you have a personal baseline from rested draws. Work in English Premier League footballers found that subjective wellness reports and blood biomarkers including CK track together over a season, which is the sort of question a personal panel can answer for you.¹

Pre-analytic control: the part that determines whether your data is worth anything

Most variance in a repeated blood panel comes from how the sample was collected. The analyzer is not the main source. Fix these and your quarterly series becomes readable.

• Fast 10–12 hours, water only. Fasting matters for glucose, insulin, triglycerides, and iron studies. It changes lipids less than people assume, but consistency beats optimization.
• Draw between 7 and 9 a.m., same window every time. Testosterone, cortisol, and iron all have diurnal rhythms.
• No hard training for 48 hours. This is the single most common reason a panel looks alarming. Heavy eccentric work elevates CK, AST, ALT, hs-CRP, and white count for days.
• No alcohol for 72 hours (GGT, triglycerides).
• Hydrate normally. Dehydration concentrates everything and raises hemoglobin, hematocrit, albumin, and BUN together.
• Sit for five minutes before the draw and avoid prolonged tourniquet time. Postural change alone shifts protein-bound analytes by several percent.
• Hold biotin for 72 hours. Biotin at supplement doses interferes with streptavidin-biotin immunoassays and can produce dramatically false TSH, free T4, and troponin results.
• Use the same lab and the same assay platform. Cross-lab comparison of ferritin, free testosterone, or vitamin D is a recipe for chasing phantom changes.

Reading a change: reference change value and not reference range

A result inside the reference range is not necessarily normal for you, and a change between two draws is not necessarily real. The number you need is the reference change value: RCV = 2^(1/2) × Z × (CV_A^2 + CV_I^2)^(1/2), where CV_A is analytic variation (your lab publishes it, typically 2–5% for chemistry, higher for immunoassays) and CV_I is within-person biological variation.

Biological variation is large for some of these markers. Ferritin and hs-CRP have high within-person CVs, which means quarter-to-quarter swings of 30–50% can be pure noise. Sodium and albumin are tightly regulated, so a 5% move is real. Before you react to a delta, compute the RCV for that specific marker. This is the single most useful analytic habit for anyone reading their own labs, and it is the thing the 160-marker dashboards do not do for you.

Keep the results as structured data and not as PDFs. Most US labs will export HL7 or FHIR Observation resources with LOINC codes. If yours will not, transcribe into a CSV with columns date, loinc, analyte, value, unit, ref_low, ref_high, lab_id, assay_platform, fasted, hours_since_training. The last two columns are what let you explain outliers a year later. Store units explicitly, because vitamin D comes in ng/mL and nmol/L, testosterone in ng/dL and nmol/L, and silent unit mismatches are the most common bug in personal health datasets.

What the panel cannot do

Blood chemistry panels are sometimes assembled into composite “biological age” scores. Compared against DNA-methylation clocks in older US adults, the blood-chemistry and methylation approaches capture overlapping but distinct signals, and the two families of measures do not agree closely at the individual level.² Treat any single-number age score built from your CMP and CBC as a summary statistic. It is not a measurement of anything biological.

Blood markers also predict physical capacity imperfectly. In adults with heart failure with preserved ejection fraction, obesity-related biomarkers including leptin and adiponectin were associated with exercise intolerance measured by cardiopulmonary exercise testing.³ The association exists at population scale. It does not tell you your own VO2max. If you want that number, do the test.

Some results need a clinician the same week and cannot wait for your next annual visit. Examples are hemoglobin below roughly 11 g/dL in a man and a platelet count outside 100–450 K/µL. Others are potassium outside 3.5–5.0 mmol/L and a new fasting glucose above 126 mg/dL. So are liver enzymes above three times the upper reference limit that persist on a rested repeat draw. Do not manage any of those from a dashboard.

Questions people also ask

What blood tests are included in a wellness profile blood test? Typically it includes the tests below, which is a reasonable floor.

• CBC
• comprehensive metabolic panel
• lipid panel
• HbA1c
• TSH
• vitamin D

For fitness purposes add iron studies with transferrin saturation, hs-CRP, fasting insulin, ApoB, and a morning testosterone with SHBG.

What labs are considered wellness labs? There is no regulatory definition. In practice “wellness labs” means panels ordered without a specific clinical indication, usually through a direct-to-consumer ordering service. The markers are the same assays a physician would order. The difference is who interprets them and whether anyone follows up.

What not to do before a wellness check? No food for 10–12 hours, no alcohol for 72 hours, no hard or unaccustomed training for 48 hours, no biotin for 72 hours, and no iron supplements for 24 hours if iron studies are on the requisition. Do not switch labs between draws if you plan to compare results.

How often should I repeat the panel? Quarterly if you are training hard or changing something deliberately, annually otherwise. Measure Lp(a) once. Four points a year is enough to separate trend from noise for most markers, and it is cheap enough to sustain.

Footnotes

1. Marc Cleary, Fearghal McHugh, Kyle F. Paradis. The Association Between Subjective Wellness Symptoms and Blood Biomarker Data in English Premier League Footballers. Journal of Science in Sport and Exercise, 2022. https://doi.org/10.1007/s42978-022-00189-7 ↩

2. Gloria Huei-Jong Graf, Christopher L. Crowe, Meeraj Kothari, et al. Testing Black-White Disparities in Biological Aging Among Older Adults in the United States: Analysis of DNA-Methylation and Blood-Chemistry Methods. American Journal of Epidemiology, 2021. https://doi.org/10.1093/aje/kwab281 ↩

3. Mariana F. Ramirez, Emily S. Lau, Juhi K. Parekh, et al. Obesity-Related Biomarkers Are Associated With Exercise Intolerance and HFpEF. Circulation Heart Failure, 2023. https://doi.org/10.1161/circheartfailure.123.010618 ↩