Some conditions have no middle ground. A woman is pregnant or she is not; there is no partially pregnant, no early-stage gestation you could plausibly wave off. Diabetes is not that kind of condition, and yet the word “prediabetes” is built to sound like it is — a second diagnosis, a separate room you get moved into before the real one. Patients hear the prefix and smell a racket: yesterday you were fine, today a lab value nudges past a line someone drew, and suddenly you have a disease you didn’t have last week. The suspicion is reasonable. The biology it’s aimed at is not what people think it is.

I. A threshold, not a state

Insulin resistance does not arrive. It accumulates. Cells that once responded promptly to insulin’s signal start requiring more of it to do the same job; the pancreas compensates by secreting more; blood glucose creeps upward as the compensation strains; the strain continues, often for years, before any clinician measures it. There is no morning on which a person crosses from metabolically sound to metabolically compromised. There is only a slope, and a number on that slope that a professional body decided to call a boundary.

That number is not arbitrary in the way a conspiracy would require. An A1C between 5.7 and 6.4 percent predicts, with real statistical weight, who is likely to progress to a formal type 2 diagnosis and who is likely to benefit from intervention now rather than later. The clinical utility is genuine. What is not genuine is the grammar the label borrows — the suggestion, built into the word itself, that “prediabetes” names a distinct condition the way “diabetes” names one. It doesn’t. It names a coordinate on a curve that was already disease before anyone put a number on it, dressed in the vocabulary of a separate diagnosis because medicine needed a lever to pull for prevention. The threshold is useful. The naming convention pretends the threshold is a wall.

II. Two failures wearing one word

The deeper problem is upstream of prediabetes entirely: type 1 and type 2 diabetes are not the same disease sharing a spectrum. They are two different failures that happen to converge on the same downstream symptom, hyperglycemia, the way two different engine faults both leave a car dead in a driveway.

Type 1 is autoimmune. The body’s own immune system destroys the beta cells that manufacture insulin, and production collapses — sometimes slowly, sometimes with brutal speed in childhood. The factory burns down. Type 2 is a signaling failure. Insulin is present, often in excess for years, but the cells stop listening to it; the pancreas compensates by shouting louder, and the shouting itself becomes unsustainable. The factory is running. The floor has stopped taking orders.

These are not two intensities of one problem. They diverge at the level of mechanism — autoimmune destruction versus receptor resistance — and only converge at the level of the symptom a nineteenth-century clinician could actually observe: sugar in the urine. Prediabetes, whatever its statistical merit, is a category built entirely inside the type 2 half of that split. Calling it “pre-diabetes” as if it anticipated “diabetes” in general quietly erases the fact that type 1 was never on this curve at all.

III. A word older than the mechanism

The reason the vocabulary fights the biology this hard is that the vocabulary predates the biology by roughly two thousand years. Aretaeus of Cappadocia, the second-century Greek physician usually credited with naming the condition, called it diabetes — a siphon, from the verb “to pass through” — because the defining clinical sign he could observe was fluid moving through the body and out again in volumes that alarmed him. He had no concept of insulin, no concept of an immune system, no concept of a cell membrane refusing a hormone’s signal. He had a symptom and a metaphor for plumbing.

Everything since has been retrofit. “Mellitus,” honey-sweet, got appended once physicians started tasting the urine and noticed the sugar (a diagnostic method one does not miss). “Type 1” and “type 2” got appended once immunology and endocrinology worked out that the plumbing metaphor was hiding two unrelated engines. “Prediabetes” got appended once epidemiologists needed a cut point for prevention programs and insurance coding. Each addition solved a real problem for the institution making it. None of them touched the root noun, which still organizes the whole taxonomy around a Greek word for a symptom nobody uses to diagnose anything anymore. The classification is a settlement between eras, not a description handed down from a single coherent understanding of the disease.

IV. The scam is the taxonomy, not the medicine

None of this means patients are imagining the manipulation. The complaint — what a scam, they moved the goalposts and called it a diagnosis — is picking up something real, just misidentifying where the trick lives. The trick isn’t in the biology. Insulin resistance is a genuine, measurable, worsening process, and catching it early genuinely changes outcomes. The trick is in a naming system that keeps using disease-shaped words for what are, underneath, points on a trajectory: a threshold masquerading as a category, sitting on top of a category (diabetes generally) that was itself never a single mechanism to begin with.

A more honest vocabulary would say plainly that type 1 is a disorder of insulin production, that type 2 is a disorder of insulin response which can later impair production too, and that prediabetes is simply an earlier reading on the same type 2 curve — not a waiting room, not a milder cousin, just the same disease seen sooner. None of that requires new science. It requires medicine to admit that a word coined to describe a Greek doctor’s astonishment at how much fluid a patient could produce is still, two millennia later, doing the naming for a set of cellular mechanisms it was never built to hold. What looks like a bureaucratic sleight of hand is really a very old label straining under a very new understanding — and straining, predictably, at the seams a patient is the first to notice.

V. A number worth watching

If the categories are the problem, the fix is not to distrust the biology underneath them. It is to watch the biology directly, on its own continuous terms, instead of waiting for a clinician to announce which side of a threshold a person has landed on. The triglyceride-to-HDL ratio, a simple figure already sitting inside any routine fasting lipid panel, tracks the exact process this essay has been describing: insulin resistance accumulating, year over year, long before any diagnostic code gets attached to it. A high ratio moves with rising insulin resistance and with the shift toward the small, dense LDL particles that do more damage per unit of cholesterol; a low one moves the other way. It is not a diagnosis, and it does not replace one. It is a slope, visible early, sitting on a lab report most people file away without reading.

A closing disclosure: I am not a doctor, and nothing above is medical advice. Bring your own numbers, and your own history, to someone qualified to read them.

Further reading