Testing your own lactate: what a meter tells you that power can't

A portable lactate analyser costs less than a mid-range power meter and comes with a box of strips, a lancing device and a strong sense that you are now doing this properly. It also comes with a large number of ways to generate confident, precise, wrong answers.

This post is about the practice rather than the theory. If you want to know what LT1 and LT2 are and how the domains between them behave, lactate threshold explained covers that ground. What follows is the part nobody explains: how to design a step test that is worth the fingers, where the error actually comes from, and whether a rider who already owns a power meter has any business buying a meter at all.

What the analyser is measuring

A portable lactate meter is an enzymatic electrochemical device. A drop of whole capillary blood — typically under a microlitre — goes onto a strip containing lactate oxidase. The reaction produces a current proportional to lactate concentration, and the meter reports it in millimoles per litre, usually in under a minute.

Two things follow from that. First, it is a real measurement of a real variable, not an estimate: there is no model between you and the number in the way that FTP or Critical Power sit behind a model. Second, it is a measurement of concentration in blood at a fingertip, which is downstream of the muscle producing and consuming the lactate in question. That gap is where most of the interpretive trouble lives.

The meters themselves vary. When three widely used portable analysers were evaluated against a laboratory reference, they differed in both systematic bias and reproducibility, and they differed from each other. The practical rule is unglamorous: your meter measures your trend. Do not compare its numbers with a training partner's different model, and do not compare them with the figures in a paper.

Stage length decides your answer

This is the part that surprises people, and it is the reason a lot of home testing produces numbers that drift for no physiological reason.

Blood lactate lags what is happening in the muscle. At a new, higher workload, muscle lactate rises quickly, but the appearance and clearance in blood take several minutes to find a steady value. Sample after three minutes and the blood has not caught up: every point reads low, and the curve therefore appears to inflect at a higher power. Sample after eight minutes and the values are closer to the steady state, and the inflection sits lower.

Nothing about that shift is physiological. The rider did not change. Only the protocol did. This is also why a threshold quoted without its protocol is close to meaningless, and why the single most valuable habit in home testing is writing the protocol down and never touching it again.

A protocol you can repeat

The following is a conventional incremental step test, biased towards reproducibility rather than laboratory elegance. Warm up for fifteen minutes at an easy intensity before stage one, and take a resting sample before you start.

Reading the two lines

LT1 is the first departure from baseline — the point where lactate begins to climb out of the flat resting region. Common working definitions are the first stage where the value rises 0.3–0.5 mmol/L above the lowest reading, or the break point in a log–log plot. For an endurance cyclist this is the more useful of the two lines, because it marks the top of the intensity you can accumulate hours at, and because power and heart rate proxy it poorly.

LT2 is the upper boundary, intended to approximate the maximal lactate steady state. Here the honest position is that there is no consensus method. Reviews of threshold concepts have catalogued more than twenty published definitions, and the fixed 4 mmol/L convention — visible as the dashed line in the diagram — is a group average that fits a substantial proportion of individuals badly. Curve-fitting approaches such as modified Dmax are more defensible, but they are more sensitive to how many stages you ran and where they were placed.

If you are going to use one number to set zones, use the same definition every time and treat the absolute value with suspicion. The change between tests is the signal. The number itself is a convention.

Where the error comes from

The glycogen one deserves emphasis because it points the wrong way. Turn up depleted and your curve shifts right, your apparent threshold rises, and you conclude the block worked. This has been known since the early 1980s, when work on glycogen depletion showed the lactate response to incremental exercise moving without any change in fitness. If you test after three days of low carbohydrate availability, you have measured your diet.

Re-testing so the comparison means something

Everything that made the first test what it was has to be copied. Same time of day, same breakfast, same caffeine, same warm-up, same trainer, same fan, same cadence, same stage length, same increment, same sampling site, same meter, and ideally the same box of strips. Keep a card with all of it written down and read it before every test.

Then re-test infrequently. Six to eight weeks is a sensible interval: long enough for a real change to exceed the noise, short enough to catch a block that is not working. Testing monthly mostly measures how well you controlled the conditions.

Does a rider with a power meter need one?

For most riders, no. The blunt case against is this: you will use the result to set zone boundaries, and your zone boundaries need to be right to perhaps ±5%, which a well-executed field test already achieves. Spending £150 and a finger every eight weeks to refine a number to a precision you cannot act on is not a good trade. How to test and raise your FTP and how many training zones you actually need between them cover what most people are trying to get out of a meter.

There are genuine exceptions. The first is LT1. Power-based methods estimate the upper threshold reasonably and the lower one poorly, and if your training distribution depends on knowing where the aerobic threshold sits — as it does for anyone taking polarised or pyramidal training literally — then a lactate curve tells you something a 20-minute test genuinely cannot. The second is diagnosis: a rider whose performance and power numbers disagree, or who suspects they are riding their easy days too hard, gets a clear answer in one afternoon. The third is coaching several riders, where the cost per athlete falls and the protocol control improves.

If you do go without, the thing to replace it with is not a single number but a curve from your own riding. Moveee's power profile fits your best efforts across durations, which gives you the upper boundary from data you have already collected — and the diagnostic that costs nothing is aerobic decoupling, which tells you whether a long ride was genuinely below your aerobic threshold without a single drop of blood.

The meter is not a status object and it is not a shortcut. It is a good instrument attached to a demanding protocol, and the protocol is where almost all the value and almost all the error live. If you are not going to run it the same way every time, the honest answer is to keep the money and ride.