What the number means, and what it does not
Biomarkers are useful precisely to the degree you understand their limits. Each entry states what the marker measures and, separately, what it cannot tell you.
- Measures and limitations, side by side
- No reference ranges or targets
Panel
Glucose regulation
HbA1c
Measures. Reflects average glucose exposure over roughly three months.
Limitation. Distorted by anything that changes red-cell lifespan — anaemia, haemoglobin variants, recent blood loss.
Fasting insulin
Measures. Used to infer insulin resistance, often via HOMA-IR.
Limitation. Assay methods are not standardised between laboratories, so absolute values are hard to compare across sites.
Continuous glucose monitoring
Measures. Shows glucose variability that point measurements miss.
Limitation. In people without diabetes, the clinical meaning of most variability metrics is not established.
Panel
Cardiovascular risk
ApoB
Measures. Counts atherogenic particles directly rather than inferring from cholesterol content.
Limitation. A risk marker, not a diagnosis. It informs a decision; it does not make one.
Lp(a)
Measures. Largely genetically determined and measured once in most people.
Limitation. Few interventions change it meaningfully, which limits what a repeat measurement adds.
hs-CRP
Measures. A non-specific marker of systemic inflammation.
Limitation. Rises with any recent infection, injury or intense exercise. A single elevated value usually means very little.
Panel
Body composition and ageing
DEXA lean mass
Measures. Tracks lean tissue — particularly relevant during rapid weight loss.
Limitation. Scan-to-scan variability and hydration status can produce changes that are not real tissue changes.
VO2 max
Measures. Among the strongest single predictors of all-cause mortality in observational data.
Limitation. Observational. Improving it is sensible on many grounds; the causal share of that association is debated.
Epigenetic age clocks
Measures. Research tools estimating biological age from methylation patterns.
Limitation. Substantial measurement variability, and no clock has been validated as a treatment target. Treat results as interesting, not actionable.