The complaint arrives in a consistent shape. Cold when nobody else in the room is. Tired in a way that a good night does not fix. Weight creeping up on the same food and the same training. And a thyroid result, often more than one, that came back normal.

The useful place to start is with what that result measured. In most cases it is a single number: TSH. It is worth being precise about where that number comes from, because it is not made by the thyroid. TSH is released by the pituitary, and it is the instruction the pituitary sends when it judges circulating thyroid hormone to be low. It climbs as the gland struggles to keep up. That makes it an early and sensitive detector of a failing gland, which is exactly the job it was selected for.

What it does not describe is the amount of active hormone in circulation, the conversion step that produces it, or the immune activity around the gland. A normal TSH is strong evidence that the gland is not failing. It is thin evidence about anything downstream of that.

Symptoms on their own do not close the gap either, and it would be dishonest to imply otherwise. In a cross-sectional study of 25,862 adults at a Colorado health fair, hypothyroid symptoms were reported more often by people with biochemical hypothyroidism than by euthyroid people, but the sensitivity of any individual symptom was low.¹ Cold, tired and heavier is a pattern with a long list of possible causes. It points at a question rather than at an answer.

A reference range is a description of a population. It is built from the spread of results across a sample of people assumed to be healthy, and it flags the outer tails. Nothing in that construction is designed to say whether a given result is normal for the person holding it.

How much that matters was quantified in a study that is small but unusually clean. Sixteen healthy men were sampled monthly for twelve months, measuring T4, T3, free T4 index and TSH. Each man varied around his own stable mean, and the width of each individual’s 95% interval was approximately half that of the group. The index of individuality was correspondingly low: 0.58 for T4, 0.54 for T3, 0.49 for TSH.²

The authors put the conclusion plainly: a test result within laboratory reference limits is not necessarily normal for an individual.²

In practice this means someone whose set point sits in the lower part of the range can drift a long way from their own baseline and still land comfortably mid-range on paper. Nothing on the report flags it, because the report has no memory of where that person started. This is the same problem we cover from the other direction in optimal versus reference range, and it is the reason a first panel is a starting position rather than a verdict.

The thyroid gland mostly secretes T4, which does very little in the form it leaves in. T3 is the active hormone, and a substantial share of the body’s T3 is produced outside the thyroid altogether, by deiodinase enzymes that strip an iodine atom from T4 in peripheral tissue. Type 2 deiodinase in particular contributes significantly to plasma T3 production in humans.³

So there is a step between what the gland releases and what your tissues run on, and that step happens somewhere else. A gland can be performing perfectly well while the conversion behind it underperforms. TSH watches the gland. It does not watch the step.

That a normal TSH and an adequate T3 position can come apart has been shown directly, though in a specific circumstance worth naming. Analysing 469 participants taking levothyroxine within the US national survey data, all with normal TSH, researchers found higher free T4 and lower free T3 than in matched controls, amounting to roughly 15 to 20% lower T3 to T4 ratios. Those participants differed from their matched controls on 12 of 52 measures, including a higher body mass index despite reporting fewer calories per kilogram per day.⁴

The relevance here is not the treatment. It is that a normal TSH sat alongside a measurably different hormone position and a measurably different clinical picture, in a large sample, and the TSH gave no indication of either.

Deiodination runs in more than one direction. Type 3 deiodinase converts T4 into reverse T3, a mirror-image molecule that fills the space without doing the work. It is best understood as a disposal route, and the body uses it deliberately.³

Under illness, fasting and physiological stress the whole pattern shifts in a recognisable way: low or low-normal T4, low T3, raised reverse T3, and no compensating rise in TSH. This is non-thyroidal illness syndrome, described most completely in intensive care.⁵

The honest reading of that literature is more interesting than the popular version of it. The reviewer’s own conclusion is that the acute changes are partly brought about by concurrent macronutrient restriction, that this early response appears adaptive and beneficial, and that tolerating it is the wise course. The picture in prolonged critical illness is a different one, involving suppressed hypothalamic signalling.⁵ The shift is not automatically a fault to be corrected.

What makes it relevant well outside an intensive care unit is direction. Months of short sleep, hard training on inadequate fuel, sustained restriction and prolonged psychological load push the same levers, at a fraction of the intensity. Measuring reverse T3 alongside free T3 tells you whether that redirection is happening. Reading it in isolation tells you very little, and I will come back to that.

In iodine-sufficient countries, most hypothyroidism is autoimmune in origin, and the antibodies are measurable long before the gland’s output changes enough to move a TSH.

They are also common. In the third US National Health and Nutrition Examination Survey, which measured thyroid antibodies in 17,353 people, thyroid peroxidase antibodies were positive in 11.3% and thyroglobulin antibodies in 10.4%. Both were more prevalent in women and both increased with age. Thyroid peroxidase antibodies were significantly associated with thyroid dysfunction; thyroglobulin antibodies on their own were not.⁶

Their forward value is considerable. The twenty-year follow-up of the Whickham cohort in England calculated the odds of developing hypothyroidism from risk factors recorded two decades earlier. Positive thyroid antibodies alone carried odds ratios of 8 in women and 25 in men. Raised TSH together with positive antibodies carried odds ratios of 38 in women and 173 in men, and increasing TSH values above 2 mU/L raised the probability further.⁷

The study that speaks most directly to how someone feels, though, looked at women who were euthyroid at the time. In 426 women undergoing surgery for benign thyroid disease, those with anti-thyroid peroxidase levels above a calculated cut-off reported a mean of 6.7 symptoms against 4.1 in the rest of the group, with no significant difference in TSH between them. Chronic fatigue, dry hair, chronic irritability, chronic nervousness and lower quality-of-life scores all tracked with antibody level. The authors concluded that hypothyroidism is only a contributing factor to the symptom burden.⁸

Antibody-positive with a normal TSH is a real position. It carries a measurable symptom load now and a measurable risk later. It is also not yet hypothyroidism, and most people in it will want monitoring rather than treatment.

The deiodinase enzymes that produce T3 are selenoproteins, which is to say selenium is built into their structure rather than helping from the sidelines.³ Iron matters a step earlier. Iron deficiency impairs thyroid hormone synthesis by reducing the activity of heme-dependent thyroid peroxidase, the enzyme that assembles the hormone in the first place, and iron-deficiency anaemia blunts the response to iodine while iron repletion improves it.⁹

This is worth measuring because both are cheap to check and both run low in exactly the group this article describes: women with heavy periods, anyone eating in a restricted way, anyone training hard on modest fuel. A structurally normal gland can underperform for want of materials, and that shortfall will not show on a TSH.

The correction has limits that deserve stating. Supplementation trials in chronic autoimmune thyroiditis have generally lowered antibody titres without producing clear improvement in the clinical course of the disease, and selenium shows adverse effects at both extremes of intake, with a narrow optimum in between.¹⁰ This is a question of correcting a measured deficiency, not of taking more.

Reverse T3 is where this subject is most often oversold, so it is the place to be most careful. The ratio calculations that circulate widely are not validated diagnostic tests, and presenting a threshold as though it decides something would be overstating what is known. Reverse T3 is useful as context beside free T3 and the clinical picture. On its own it settles nothing.

The field’s own position cuts in both directions and is worth reading properly. The American Thyroid Association task force on thyroid hormone replacement concluded that levothyroxine should remain the standard of care, and found no consistently strong evidence that alternative preparations improve health outcomes. In the same document, the task force named the development of superior biomarkers of euthyroidism to supplement thyrotropin measurement as a research need.¹¹ TSH is acknowledged as incomplete by the people who rely on it most. Nothing has yet replaced it.

Association also has to stay association. In a population study of 4,082 adults with no overt thyroid dysfunction, TSH was positively associated with body mass index and free T4 negatively associated, with no association found for free T3. The difference between the highest and lowest TSH groups was 1.9 kg/m², which corresponded to 5.5 kg of body weight among women.¹² That tells you thyroid function within the normal range travels with body weight across a population. It does not establish which way the arrow points for any individual, and body fat influences thyroid signalling as well as the reverse.

The last limit is the broadest one. Tired, cold and gaining weight is not a thyroid-specific pattern. Iron deficiency, accumulated sleep debt, insulin resistance, the perimenopausal transition and low mood all produce a version of it, and more than one is usually in play at once. We look at two of those elsewhere, in insulin resistance in women over 40 and in why you are exhausted with normal blood tests. A thyroid panel narrows the question considerably. It does not close it.