Testing for a Molecule the Body Already Makes

Somewhere in an accredited laboratory there is a rack of small tubes, each holding a few millilitres of somebody's blood or urine, and the morning's job is to decide whether any of them contains something that should not be there. For most of the prohibited list this is difficult but clean. A synthetic anabolic steroid is a stranger in human tissue: find its fingerprint and the argument is essentially over. Peptides do not behave that way, and the reason they are so hard to test for fits into a single sentence. A small peptide administered to a person is, chemically, all but indistinguishable from the peptides that person already makes — and in some cases it is not merely similar but identical, atom for atom. So the analyst cannot ask whether the substance is present. They have to ask something far slipperier: is there more of this than this particular body ought to be producing, and can that be shown to a standard that will survive being argued over by lawyers?

That shift — from presence to excess, from a yes to a judgement about what counts as normal — is the whole difference between peptide doping control and every other category on the list.

A steroid is a stranger; a peptide is a relative

Consider what a conventional doping test does. The workhorse instrument is the mass spectrometer, which sorts molecules by mass and then by the way they shatter when deliberately broken apart. The resulting pattern is about as individual as a signature. Match it against a reference standard and you have an identification defensible in front of a tribunal, because the compound has no business being in a human being at all.

Peptides ruin that logic by being family. They are short chains of amino acids — small proteins, essentially — and the body runs on them: hormones instructing the pituitary, signals governing appetite and growth and repair, fragments shed as larger proteins break down. All endogenous, which simply means made inside the organism rather than introduced from outside. A laboratory manufacturing a peptide drug is very often manufacturing a copy of something already circulating in every sample it will ever examine. Where a drug is a modified version of a natural sequence, the modification can be a single substituted amino acid — visible in principle, provided the instrument has been told to look for exactly that difference and enough of the molecule survives the journey into the tube 1.

Four problems, arriving together

Suppose the best case: the target really is foreign, with a feature an instrument can be tuned to. Four obstacles remain, and they never arrive one at a time.

• Concentration. Peptide hormones work at levels hard to grasp intuitively — often picograms per millilitre, a millionth of a millionth of a gram in less than a teaspoon — against a plasma background of thousands of far more abundant proteins.
• Clearance. The body disposes of small peptides briskly, so the molecule can be gone from circulation while the biological effect of having taken it has not faded. What can be measured and what matters are not present at the same time.
• Instability in the sample. Blood carries proteases, the enzymes whose job is dismantling peptides, and they do not stop working when the tube is capped. Without prompt cooling and careful handling, a sample can quietly digest its own evidence.
• Endogenous background. Even a perfect measurement often returns a number rather than a verdict, because the molecule was going to be there anyway. The question becomes whether the number is higher than it should be — which requires knowing what should be.

Together these push peptide analysis away from identification and towards inference. Much of the field's effort has gone not into more sensitive instruments, though it has done that too, but into establishing what normal looks like across large populations, so an individual result can be placed against a distribution rather than a threshold plucked from nowhere 1.

Growth hormone, the case everyone learned on

The clearest worked example is growth hormone, partly because the science is unusually well documented and partly because it defeated the field for years before it yielded. Human growth hormone made by recombinant technology — grown in engineered cells rather than extracted from tissue — is, in its principal form, the same molecule the pituitary secretes. Not similar. The same. There is no foreign fingerprint to find, and no amount of sensitivity conjures one into existence 3.

For a while that was treated as a dead end. Then the reasoning changed shape: if you cannot identify the molecule, look for the signature that giving it leaves on the system around it. There turned out to be two such signatures, found along independent lines 4.

Two ways round an identical molecule

The first is the isoform differential approach, and it is a lovely piece of thinking. The pituitary does not secrete one clean product. It releases a mixture: a dominant form of the hormone plus a family of variants — shorter versions, differently processed ones, forms that clump together — in proportions that hold roughly steady between people. Manufactured hormone is not a mixture. It is a single uniform species, because that is what a purified pharmaceutical product is. Give it to someone and two things happen at once: the abundance of that one form rises, and the body's own secretion is turned down by the feedback loops regulating it, thinning the natural variety. So the test does not measure how much hormone is present. It measures the ratio of one form to the others, and asks whether the mixture in the tube looks like something a gland would produce 4.

The second is the biomarker approach, which abandons the hormone altogether. Growth hormone works largely by instructing other tissues to make other things, chief among them insulin-like growth factor I, usually shortened to IGF-I. It also drives the turnover of collagen, the structural protein of bone and connective tissue, which sheds measurable fragments as it is laid down; one such procollagen marker became the second half of the pairing. Both respond to the hormone, both are easier to measure than the hormone itself, and both move more slowly than a circulating peptide does. Measure the pair, compare against reference limits built from large populations, and you have evidence of a system that has been pushed 34.

Neither produces the clean binary a steroid test produces. Both are arguments about probability, and both needed years of population work first. They are complementary rather than competing: one asks what is in the sample, the other what the body has been doing.

The compounds that add nothing foreign at all

If a copy of a natural hormone is hard, a secretagogue is harder. The word is plainer than it looks: a secretagogue makes a gland secrete. Analogues of growth hormone-releasing hormone, and compounds mimicking ghrelin at its receptor, do not deliver a hormone at all. They deliver an instruction. The hormone that follows is the person's own, made by their own pituitary, in their own characteristic mixture of forms.

That guts the isoform logic, which depends on manufactured material being uniform where natural material is varied. Here there is nothing manufactured in the bloodstream to be uniform. Downstream markers may shift, but markers rarely carry a case alone, because many innocent things move them. What remains is detecting the secretagogue itself — small, foreign, quickly cleared, near the floor of what instruments reach, in a sample whose own enzymes work against you 1. Every problem above, in one package.

A sequence is an easy thing to change

Then there is the problem the field cannot engineer its way out of. Making a short peptide is not difficult. Solid-phase synthesis has been routine for decades, and once a sequence is known, altering it means choosing different building blocks: swap a residue, flip one to its mirror image, cap an end, close the chain into a ring. Each change gives a different mass and a different fragmentation pattern — a molecule existing assays are not looking for, because an assay finds what it has been told to find 2.

The analytical response is necessarily reactive. Methods can be broadened, instruments run in modes that capture unexpected masses for later re-examination, findings shared between reference laboratories. But a method is always built after the fact of a molecule existing, and the annual reviews tracking this work read as a record of coverage extending outward towards compounds the field has heard about rather than compounds it anticipated 2.

This is why prohibited-list categories are drafted by class rather than as an inventory of names. A rule listing individual compounds is obsolete the moment somebody makes one that is not on it; a rule covering a hormone with its fragments, releasing factors, analogues and substances of comparable biological effect reaches molecules nobody has yet made. The tension is worth naming rather than glossing over. Breadth is bought at the cost of precision: whether a given molecule belongs to a prohibited class becomes a scientific judgement about structure and effect rather than a lookup, and judgements can be contested in ways a list cannot. Regulators took that trade knowingly, because a tidy list always one synthesis behind is worse.

Why any of this is published at all

An obvious question hangs over an article like this, and it deserves a direct answer rather than a nervous silence. Does describing the difficulty of detection amount to a map for getting around it? This piece declines to go near that, deliberately. It gives no figures for how long any compound remains measurable, ranks nothing by how readily it is found, and does not speculate about what would or would not be caught. Those specifics are what would make the material actionable, and the omission is the point rather than an oversight.

The general shape of the science, though, is published in full — in journals, method papers and annual reviews stating openly what laboratories can and cannot currently do 2. That openness is a requirement, not carelessness. An anti-doping result is an accusation with consequences, and it will be challenged, as it should be. A method that cannot be set out, validated, reproduced in an independent laboratory and cross-examined by an expert for the defence cannot support a sanction; secrecy would make results unfalsifiable precisely when they matter most. Methods also improve only because they are published and picked apart: the isoform and biomarker approaches were built over years of open endocrinology, argued over in print by people who disagreed about reference ranges and study design 34.

The label on the vial

One last thing, because it sits underneath the whole subject. Much of the material discussed here circulates legitimately as research chemical supply, labelled for research use only. That phrase deserves to be taken seriously rather than treated as a wink. It is a real regulatory category with real content: the substance is supplied for laboratory work, it has not been assessed for safety in humans, no approved label exists, and nobody has taken responsibility for what happens if a person administers it to themselves. Endocrinology could not function without a route to compounds that are not medicines; the category exists because the science needs it. It is also, unmistakably, the channel through which most of this material reaches people who are not researchers and have no laboratory. Both things are true at once, and writing that acknowledges only one of them is dishonest in one direction or the other.

Which returns, in the end, to the analytical problem. These compounds are hard to test for and attractive for the same reason: they sit so close to the body's own chemistry that the body barely registers them as foreign, and neither, without a great deal of ingenuity, does an instrument. That closeness is what makes the science interesting. It is also what makes a molecule with no safety assessment and no label a poor thing to put into a system whose ordinary signalling it was built to resemble.