Did it work? Reading a marker before and after a change

One result before a change and one after can't tell you much on their own. Every marker moves from draw to draw, because the body varies and so does the measurement, so a difference only means something once it is bigger than that normal movement. You read it best with more than two results, drawn at the same lab under the same conditions, with a recheck timed to how quickly the marker can respond.

I build Sera, a lab-review workspace for practitioners, so read the last section with that in mind. Everything before it applies however you keep your records.

Why the same marker moves when nothing has changed

Two kinds of variation sit under every result.

Biological variation. A person's level of a marker isn't fixed. It moves around their own set point from day to day and through the day. Some markers have a strong daily pattern. The TSH range note says TSH runs highest overnight and early morning, and serum iron varies through the day. Others react to short-term events. Ferritin is an acute-phase protein that inflammation raises regardless of iron stores, and a single raised hsCRP can reflect any recent infection or injury.

Analytical variation. The same sample run twice on the same analyzer won't give exactly the same number. Every method has some imprecision, and different methods at different labs can also sit at slightly different levels.

Laboratory medicine has a name for putting the two together: the reference change value. In his review of reference change values in Clinical Chemistry and Laboratory Medicine, Callum Fraser describes it as a way to judge whether the difference between a person's serial results is significant, calculated from the lab's analytical imprecision and the marker's within-person biological variation. The detail is for labs, but the idea is the useful part: a change smaller than the usual noise for that marker isn't evidence that anything happened.

Keep the conditions the same

You can't remove biological variation, but you can avoid adding to it. The more the two draws differ in how they were taken, the harder the comparison.

• Same lab, where you can. Reference intervals and methods differ between labs. The TSH note gives lab limits from about 0.3 to 0.45 at the bottom and 4.2 to 5.3 at the top, and TPO antibody numbers aren't comparable between labs at all. A move from one lab to another can look like a change in the client.
• Same time of day. This matters most for markers with a daily rhythm. Morning cortisol ranges are for a sample drawn at about 8 am, and the testosterone note says to test men in the morning and repeat a low result before acting on it.
• Same fasting state. A non-fasting sample runs higher for triglycerides. If the first draw was fasting and the second wasn't, a lipid "improvement" or "worsening" may be neither.
• Same point in the cycle, for female hormones. Estradiol, LH and FSH swing across the cycle, so comparing a day-3 result with a mid-cycle one compares two different things.
• Well, not recovering. If the client had a cold the week before one draw, inflammatory markers and ferritin from that draw are hard to compare with anything. The CRP note says CRP rises within hours of infection or injury and falls over days, so repeat it once the person is well.

Time the recheck to the marker

A retest too soon can show no change simply because the marker hasn't had time to move.

HbA1c is the clearest example. MedlinePlus explains that it shows average blood glucose over the past two to three months, because red blood cells live about three months. An HbA1c a few weeks after a diet change mostly reflects the time before it. Fructosamine covers a two-to-three-week window instead, and Sera's range note on it names that as the better measure for judging a change made a few weeks ago. It falls with low albumin for reasons unconnected to glucose, so check protein status before reading a low value as improvement.

For most other markers there's no single published waiting time, and I won't invent one. What you can do is write down, before the change, when you plan to recheck and why that interval makes sense for that marker. It stops the retest date from being set by whenever the client happens to book.

How many results make a trend

Two results give you a difference. Whether it's a trend depends on how big that difference is against the marker's normal movement, and two points can't show you that movement.

Three or more results, taken under matched conditions, start to show a direction. A baseline helps most: if you have two or three results from before the change, you know roughly where the client usually sits, and a later result can be read against that rather than against one earlier number. The researchers behind the reference change value make a related point from the statistics side. In a 2019 paper in Annals of Clinical Biochemistry, Lund, Hyltoft Petersen and Fraser show that applying the standard two-result test over and over to a longer series raises the rate of false alarms, and they set out a method that uses several results to establish the steady state first.

In practice:

• Treat a single surprising result as a reason to repeat, not a conclusion. The hsCRP and testosterone notes both say this directly.
• Look for the same direction across several draws before calling something a response.
• If you change several things at once, the trend can't tell you which one did it.

What to record with each draw

The result alone won't tell you what changed. Write these next to every draw, at the time, while the client still remembers:

The last two rows are the ones that most often explain a result that doesn't fit.

When to refer

If a marker moves sharply, or keeps moving in a direction that worries you, it isn't a question about whether an intervention worked any more. Refer the client to their doctor. Several Sera range notes set this line: a morning cortisol below 5 μg/dL, for example, needs medical follow-up.

Where Sera fits

Sera is built for this kind of reading. Each client's results sit on one timeline across years and labs, against optimal and sufficient bands whose sources and evidence grades are published. Diet, supplement and medication changes are marked on the trend charts, so you see the result and the change on the same screen, and you can put up to three markers on one chart. Every result stays in the unit the lab reported. Sera suggests a retest date for any result out of range, and a yearly one for the markers on a typical annual check-up list, and you can set your own date for any marker, such as the recheck you planned before the change. It doesn't diagnose or suggest treatment.

It's free for three clients. There's a 3-minute tour if you'd rather watch first.

Alina Radu is the founder of Sera and a software engineer. The idea started with her own Hashimoto's tests, scattered across years without the big picture.