TSH Normal Range: What the Numbers Mean and What It Means for Athletes
Table of Contents
Key Takeaways: TSH Normal Range
- The TSH normal range is a statistical reference interval, not a diagnosis or treatment threshold. Results need to be interpreted in clinical context.
- TSH reference limits can change with age. In NHANES III, the 97.5th percentile increased from 3.56 mIU/L at ages 20–29 to 7.49 mIU/L at age 80 and older.
- Time of day, biotin and pregnancy can all affect how a TSH result should be interpreted, which is why borderline values are more useful when viewed in context and, when relevant, over time.
- A mildly abnormal TSH does not automatically mean treatment is needed. In practice, I look at the trend, free T4 and the broader clinical picture, and may also consider TPO antibodies or free T3 in selected situations.
- In athletes, lower thyroid hormone values do not prove that endurance training itself suppresses thyroid function. Low energy availability is one plausible contributor, but the available studies do not establish a single causal explanation.
- TSH is not a diagnostic test for overtraining syndrome. Its practical value in that setting is mainly as part of the differential diagnosis, helping identify thyroid disease that could otherwise be mistaken for a training-related problem.
Introduction: What Is the TSH Normal Range?
TSH is one of the blood tests I use most often in everyday clinical practice. On a typical day, I may order several TSH measurements as part of the initial work-up for patients presenting with symptoms such as fatigue, feeling unusually cold, hair loss, sleep problems, low mood, gastrointestinal symptoms, or concerns about overtraining or burnout. These symptoms are nonspecific and can have many causes, but thyroid function is often one of the areas worth considering. In practice, I usually interpret TSH together with free T4, and in selected situations free T3 can also add context. Patients frequently ask a simple but important question when they see the result: what is actually normal, and when does a TSH value become abnormal?
A TSH normal range is a reference interval. It describes where results fall in a defined reference population, rather than providing a stand-alone diagnosis or treatment threshold. One major US example is NHANES III, which measured TSH, total T4 and thyroid antibodies in 17,353 people aged 12 and older. Its reference population of 13,344 was drawn from the disease-free group after additionally excluding people who were pregnant, taking androgens or estrogens, positive for thyroid antibodies, or biochemically hypothyroid or hyperthyroid.[1] In a later age-based analysis of the same dataset, measurements across all age groups had originally been used to define the upper limit of the reference range.[2]
That helps explain why a single “normal” TSH number can be misleading. A result only becomes clinically useful when it is interpreted in context: alongside the laboratory range, other thyroid measurements when relevant, age, symptoms and the broader clinical picture. This is also why patients’ questions about whether their TSH is “normal” are often more nuanced than they first appear. In this article, I will first explain how the TSH normal range is defined and how it changes with age and other factors, and then look more specifically at what TSH can mean in athletes. For a deeper discussion of low-normal values in trained people, see my article on TSH in athletes.
How Laboratories Set the TSH Normal Range
A TSH normal range is usually a statistically defined reference interval, commonly covering the middle 95% of values in a selected healthy reference population. In practical terms, the lower limit is typically set at the 2.5th percentile and the upper limit at the 97.5th percentile. This is an important distinction, because a reference interval describes how laboratory values are distributed in a reference population; it is not automatically the same thing as a threshold for disease or treatment.
A reference interval and a clinical decision threshold are therefore not interchangeable. For example, the laboratory in a 2024 study of elite athletes used a TSH reference interval of 0.27 to 4.2 uUI/mL.[4] A 2004 scientific review and guideline, by contrast, discussed clinical management in TSH bands of 0.1–0.45 mIU/L and 4.5–10.0 mIU/L.[3] The first represents a laboratory reference range, whereas the second represents clinically defined categories used when discussing subclinical thyroid disease. In practice, I start by looking at the reference interval reported by the laboratory and then interpret the result in the wider clinical context rather than treating the cut-off itself as a diagnosis.
TSH Normal Range by Age
The upper limit is not fixed across life. In NHANES III, the 97.5th centile was 3.56 mIU/L in 20- to 29-year-olds and 7.49 mIU/L in the 80+ group. Among people without thyroid disease, 10.6% of 20- to 29-year-olds had a TSH above 2.5 mIU/L, rising to 40% in the 80+ group, and 14.5% of that oldest group exceeded 4.5 mIU/L.[2] The authors reported that 70% of older patients with a TSH above 4.5 mIU/L were within their age-specific reference range. They concluded that subclinical hypothyroidism may be significantly overestimated unless an age-specific TSH range is used.[2]
A 2025 analysis in Annals of Internal Medicine of 8,308 NHANES participants also found that the 97.5th percentile of TSH increased with age. With current reference intervals, the prevalence of subclinical hypothyroidism rose from 2.4% at ages 20 to 29 to 5.9% at ages 70 and older.[5]
A 2026 UK study built intervals in 1,364 anti-thyroid peroxidase antibody-negative adults with no self-reported conditions or medications and applied them to 21,286 primary care tests. With an all-adult interval, 8.2% of over-60s had an elevated TSH and 7.0% were classed as subclinical hypothyroid. With age-related intervals these fell to 4.4% with an elevated TSH, 3.7% subclinical hypothyroid and 0.4% overt hypothyroid. Minimal changes were seen in the 18–60 years group.[6]
In practice, this is often less complicated for patients and clinicians than the age-related data might suggest. In Finland, laboratory reports commonly provide the reference interval that applies to the patient rather than requiring the clinician to calculate an age adjustment manually. From my perspective, this is useful because the clinician can focus on interpreting the reported value in its clinical context instead of trying to memorize different numerical cut-offs for every age group. The important point is still to recognize that the reference interval itself may change with age, even if that adjustment is already reflected in the laboratory report.
What Shifts a TSH Result Besides Thyroid Disease
Time of day. In a large outpatient sampling study, mean TSH was 2.76 mIU/L in samples taken at 7–9 am (n = 5,448) and 2.22 mIU/L in samples taken at 1–3 pm (n = 3,572). These are group means from samples collected at different times, not serial within-person measurements, and the authors state that mean TSH varied nearly 28% depending on the time of day. They judged the diurnal effect on TSH clinically significant and found no clinically significant effect of resting time before sampling.[7] A 2024 paper likewise describes a marked circadian rhythm, with TSH lower in the afternoon and evening than in the early morning, and notes that intra-day variation largely exceeds inter-vendor variation and assay variability.[8]
In Finland, laboratory instructions commonly recommend that TSH sampling is done before 4 pm. In my own practice, however, I usually advise patients to have the test taken in the morning when that is practical. TSH has a clear circadian rhythm, so keeping the sampling time reasonably consistent can make serial results easier to compare. I would not usually treat the exact clock time as the most important part of the test, but if a patient is being followed over time, taking the sample at a similar time of day can reduce one avoidable source of variation.
Biotin. Over-the-counter biotin products containing 2–20 mg are readily available. In a prospective study, 13 adults took biotin 10 mg/day for eight days. TSH, fT4 and TT3 measurements changed significantly on the Roche Cobas 6000 platform, and at the time points tested, interference was greatest two hours after intake.[9] The authors stressed that physicians must be aware of erroneous results in people taking biotin supplements.
Pregnancy. The largest fall in serum TSH occurs in the first trimester, because hCG directly stimulates the TSH receptor. TSH and its reference range then rise gradually but stay lower than in non-pregnant women, and the downward shift is greater in twin pregnancies.[10]
When TSH Falls Outside the Normal Range
In NHANES III, subclinical hypothyroidism was defined as a normal serum T4 with a TSH above 4.5 mIU/L.[2] The 2004 review concluded that the consequences of subclinical thyroid disease, at TSH 0.1–0.45 or 4.5–10.0 mIU/L, are minimal, and it recommended against routine treatment in those bands. It also noted that evidence linking subclinical disease to symptoms or outcomes is limited, and that pregnancy is a special case in which aggressive case finding and treatment can be justified.[3]
The TRUST trial tested treatment directly. It randomized 737 adults aged 65 or older with persisting subclinical hypothyroidism (TSH 4.60–19.99 mIU/L, free T4 within the reference range) to levothyroxine or placebo. Mean baseline TSH was 6.40 mIU/L. At one year it was 3.63 mIU/L on levothyroxine and 5.48 mIU/L on placebo. The Hypothyroid Symptoms and Tiredness scores did not differ between groups, and no benefit appeared on secondary outcomes.[11] These findings come from older adults and should not be transferred to younger or athletic people.
At the other end, the 2004 review rated the data linking a TSH below 0.1 mIU/L to atrial fibrillation and progression to overt hyperthyroidism as good, but found no data supporting treatment to prevent those outcomes.[3]
Subclinical hypothyroidism is a finding I encounter fairly often in clinical practice. When the pattern persists, I tend to follow thyroid function a little more closely rather than drawing conclusions from a single mildly abnormal result. Depending on the situation, this may mean repeating thyroid tests after a few months to see whether the abnormality resolves, remains stable or progresses. I also have a relatively low threshold for checking thyroid peroxidase (TPO) antibodies when the clinical picture makes autoimmune thyroid disease relevant. For me, the key point is that a mildly elevated TSH is rarely interpreted in isolation; the trend over time, free T4 and the broader clinical context usually matter more than one borderline value.
TSH Normal Range in Athletes
Endurance training and lower values. In a 2024 study of 1,453 elite athletes tested between 2012 and 2022 in one laboratory, 26 were on thyroid hormones and 85 had a TSH outside the laboratory range, leaving 1,342 with normal TSH for analysis. Endurance athletes had the lowest TSH (p < 0.0001), fT3 (p = 0.007) and fT4 (p < 0.0001) compared with the other disciplines. Samples were taken early in the morning after at least 10 hours of fasting. The authors linked chronic aerobic training to lower thyroid hormone levels and suggested this may be an adaptive mechanism.[4] Because this was a selected elite-athlete cohort rather than a prevalence study, it does not establish the population prevalence of thyroid disease among athletes.
Energy availability. In a controlled experiment in exercising women (an experimental protocol, not a cohort of athletes), four days at low energy availability (8 versus 30 kcal per kg body weight per day) lowered T3 by 15% and free T3 by 18%, raised T4 by 7% and reverse T3 by 24%, and left fT4 unchanged. Exercise quantity (0 versus 1,300 kcal/day) and intensity (40 versus 70% of aerobic capacity) did not affect any thyroid hormone, and raising dietary energy intake prevented the low-T3 state.[12] A review of endocrine effects in relative energy deficiency in sport reports that low energy availability consistently lowers T3, while effects on TSH and T4 are less consistent.[13]
A 2026 meta-analysis of 15 fasting studies in adults (fasts of at least 24 hours; the inclusion criteria did not target athletes) found a pooled standardized mean change of −1.979 for FT3 (7 studies; 95% CI −3.054 to −0.905) and −0.581 for TSH (10 studies; 95% CI −0.882 to −0.280). FT4 changed by 0.073 (7 studies; 95% CI −0.237 to 0.382), which was not significant.[14]
Taken together, these findings make low energy availability a plausible contributor to lower thyroid hormone values in some athletes, but they do not prove that it explains the endurance-athlete pattern seen in the 2024 cohort. This is where I separate the research finding from the clinical interpretation. In an under-fuelled athlete, a low fT3 with a low-normal TSH can be compatible with adaptation to low energy availability, so energy status and training load are factors I consider alongside thyroid disease and other clinical explanations. I therefore avoid assuming that endurance exercise itself has suppressed thyroid function simply because the athlete has a lower-normal thyroid profile. That is also why, when the broader clinical picture raises concern about under-fuelling, I look at fT3 and fT4 together with TSH; see free T3 in athletes, free T4 in athletes and RED-S blood work.
Overtraining. A systematic review of 38 studies found that basal hormone levels were mostly normal in athletes with overtraining syndrome, functional or non-functional overreaching compared with healthy athletes. It concluded that basal hormone levels are not good predictors, whereas blunted ACTH and GH responses to stimulation tests may be.[15] A single basal TSH should not be expected to identify overtraining syndrome reliably. For the wider picture, see thyroid function in athletes and fatigue in athletes.
In practice, overtraining syndrome remains a clinical diagnosis rather than something that can be confirmed with a single blood test. A normal TSH does not rule it in or out, and an isolated TSH value should not be expected to identify overtraining syndrome reliably. However, thyroid testing can still be clinically useful as part of the differential diagnosis, because overt hypo- or hyperthyroidism can produce symptoms that overlap with fatigue, reduced performance and impaired recovery. When I assess an athlete with suspected overtraining, I therefore see TSH primarily as part of the work-up for alternative explanations rather than as a biomarker of overtraining itself. For the wider picture, see thyroid function in athletes and fatigue in athletes.
Conclusion: Reading the TSH Normal Range Wisely
TSH is useful precisely because it is simple to measure, but interpreting it is rarely as simple as deciding whether a number falls inside or outside a printed range. The reference interval is a statistical construct, not a diagnosis, and factors such as age, time of day, pregnancy, supplements and the wider clinical situation can all affect how a result should be understood.
In clinical practice, I find the trend and the context more useful than an isolated borderline value. A mildly abnormal TSH may warrant follow-up, interpretation alongside free T4 and, in selected situations, consideration of TPO antibodies or free T3 rather than an immediate conclusion that thyroid disease is present. This becomes particularly important in athletes, where lower thyroid hormone values can reflect several overlapping influences. Low energy availability is one plausible contributor, but the available studies do not justify assuming that endurance training itself causes thyroid suppression.
The same principle applies when overtraining is suspected. TSH is not a biomarker that confirms overtraining syndrome, but it can still be useful in the differential diagnosis so that genuine thyroid disease is not mistaken for a training-related problem. For me, that is the most practical way to read the TSH normal range: start with the laboratory reference interval, but interpret the number together with the patient, the other thyroid results and the clinical question that led to testing in the first place.
References
- https://doi.org/10.1210/jcem.87.2.8182
- https://doi.org/10.1210/jc.2007-1499
- https://pubmed.ncbi.nlm.nih.gov/14722150/
- https://doi.org/10.3390/biomedicines12071530
- https://doi.org/10.7326/ANNALS-24-01559
- https://doi.org/10.1177/00045632261426852
- https://doi.org/10.1371/journal.pone.0140475
- https://doi.org/10.1515/hmbci-2024-0043
- https://doi.org/10.1089/thy.2020.0866
- https://doi.org/10.1089/thy.2016.0457
- https://doi.org/10.1056/nejmoa1603825
- https://pubmed.ncbi.nlm.nih.gov/8498602/
- https://www.sciencedirect.com/science/article/abs/pii/S2451965019300420
- https://doi.org/10.3389/fendo.2026.1876045
- https://doi.org/10.1186/s13102-017-0079-8
