Why one abnormal result usually gets repeated

Being asked to repeat a blood test is unsettling. It reads as though something was found, or something went wrong. Usually neither is true.

Repeating is a deliberate strategy for dealing with two facts about laboratory results: they vary for reasons unrelated to health, and a single value cannot distinguish variation from change.

Two kinds of variation

Analytical variation is the imprecision of the measurement itself. Run the same sample twice and the analyser will not return exactly the same number. For most markers this is small, in the region of a few per cent, and laboratories track it continuously.

Biological variation is larger and more interesting. The amount of most substances in blood fluctuates naturally within a person, hour to hour and day to day. Some markers are extremely stable — sodium barely moves. Others swing widely, and serum iron can differ substantially between morning and afternoon in the same person.

Put together, these mean that two results taken a week apart can differ noticeably with nothing at all having changed.

Why this matters most at the edge of a range

A result deep inside a reference range and a result far outside it are both fairly unambiguous. A result sitting just past a limit is the difficult case, and it is exactly where variation is most likely to be the explanation.

Because reference ranges are built to cover the middle 95 per cent of a population, marginal results are also common in people with nothing wrong. Acting on every one of them would generate a great deal of investigation for very little benefit.

Repeating separates the two possibilities cheaply. If the second result sits comfortably inside the range, variation is the likely explanation. If it confirms the first, the finding is real and worth pursuing.

Sample problems that look like results

Some repeats are requested because the laboratory suspects the sample rather than the person. Haemolysis raises potassium and several enzymes. A tight tourniquet concentrates protein-bound markers. A delay in transport lowers glucose and raises potassium.

Laboratories detect many of these and either flag or reject the sample, but not all of them are visible. Where a result is out of keeping with the rest of a panel, a fresh sample is often the quickest diagnostic step available.

How long to wait

The interval depends on what is being measured and why. A suspected sample artefact is repeated immediately. A borderline thyroid result is typically repeated after some weeks, because the system it measures moves slowly. A liver enzyme raised after a viral illness may be rechecked after a month or two to allow recovery.

A repeat that is scheduled rather than urgent is usually a sign that nobody is alarmed. Where a clinician wants a test repeated within days, that is worth clarifying rather than assuming.

What to do differently the second time

Where timing, fasting or activity could have influenced the first result, matching conditions makes the comparison cleaner. Same time of day, same fasting state, same laboratory if practical.

Using the same laboratory removes method differences from the comparison, which matters when the question is whether a value has genuinely changed rather than whether it is high.

Mentioning anything unusual about the first draw — a difficult vein, a long tourniquet, a delay before the sample was collected — gives useful context that will not otherwise reach the person interpreting the result.