Interpreting a single nutrient result: reference ranges, assay variation, and why one value rarely justifies an intervention

A reference range is a population interval, not a target. This note sets out what a single out-of-range result does and does not support, and the three sources of variation that precede any clinical interpretation.

A nutrient result is frequently presented to the patient as a verdict and received as one. It is more accurately a single observation carrying three distinct sources of variation, only the last of which is the thing anyone wants to measure.

This note sets out those sources and what follows from them for interpretation. It concerns methodology rather than any specific nutrient, and it is not clinical guidance for an individual.

The reference interval is a population descriptor

Reference intervals are typically constructed to contain the central 95% of values from a defined reference population.

That construction has a consequence that is arithmetically unavoidable: approximately one in twenty individuals in that population falls outside the interval while being entirely healthy. Being outside a reference range is therefore not, on its own, evidence of pathology — it is evidence of being in the tails of a distribution, which one person in twenty is.

A second and separate point: the reference population may not resemble the individual. Intervals derived from one demographic applied to another can shift the apparent prevalence of “abnormality” substantially without anything about the individuals changing.

Analytical variation sits on top of that

Two laboratories measuring the same sample may return systematically different values, for reasons that are well characterised: different analytical platforms, different calibrators, different reference materials.

Each platform also carries its own imprecision — repeated measurement of an identical sample does not return an identical number.

The practical consequence for serial monitoring is specific and frequently overlooked. A change between two results from different laboratories may reflect the laboratories. Where a trend is the clinical question, method and laboratory should be held constant, and where they cannot be, the comparison should be interpreted with that limitation stated.

Within-subject biological variation is the largest term for several nutrients

An individual’s value fluctuates around their own homeostatic set point over time, independent of any change in status. Time of day, recent intake, hydration, acute illness and inflammation all contribute.

For several nutrients this within-subject variation is substantial relative to the width of the reference interval. Where that holds, a single measurement is a poor estimator of the individual’s usual value, and two or three measurements under standardised conditions are required before an individual estimate becomes reliable.

This is the strongest methodological argument for repeat testing before intervention, and it is independent of any argument about cost or clinical caution.

What this implies for practice

A single out-of-range result supports further assessment. It does not support intervention on its own.

Repeat under standardised conditions — same laboratory, same method, comparable time of day and fasting state — before treating a difference as real.

Interpret alongside intake and presentation rather than in isolation. A biochemical value and a plausible dietary explanation are considerably more informative together than either alone.

A note on intake data as corroboration

Intake information improves interpretation and carries its own measurement error, which should be stated rather than assumed away.

Self-reported recall under-reports systematically, and the magnitude is not small. Consumer food-logging tools vary widely: independent benchmarking of such applications has reported calorie errors ranging from approximately 1% to over 12% depending on the tool, with the better-characterised figures coming from groups with no commercial relationship to the products tested.

Intake data is therefore corroborative context for a biochemical result, not a second measurement of status. Treating it as the latter is a common interpretive error and it runs in both directions — apparently adequate intake does not exclude deficiency, and apparently low intake does not establish it.

Scope

This note addresses measurement and interpretation methodology. It does not provide diagnostic criteria, reference values or treatment thresholds for any nutrient, and it is not a substitute for assessment by a qualified clinician.