Evaluating the link between thyroid function test (TFT) results and levothyroxine dose in the management of hypothyroidism: can we improve dosing regimes?

Adrian Heald; Lakdasa Premawardhana; Peter Taylor; Suhani Bahl; Buchi Okosieme; Anthony Fryer; Antonio Bianco; Colin Dayan; Mike Stedman

doi:10.1530/endoabs.117.P231

P231

Prev Next

Section Contents Cite

Endocrine Abstracts (2026) 117 P231 | DOI: 10.1530/endoabs.117.P231

SFEBES2026 Poster Presentations Thyroid (34 abstracts)

Evaluating the link between thyroid function test (TFT) results and levothyroxine dose in the management of hypothyroidism: can we improve dosing regimes?

Adrian Heald ^1,2 , Lakdasa Premawardhana ³ , Peter Taylor ³ , Suhani Bahl ³ , Buchi Okosieme ³ , Anthony Fryer ⁴ , Antonio Bianco ⁵ , Colin Dayan ³ & Mike Stedman ⁶

Author affiliations

¹The School of Medicine and Manchester Academic Health Sciences Centre, Manchester, United Kingdom; ²Salford Royal Hospital, Salford, United Kingdom; ³Thyroid Research Group, Systems Immunity Research Institute, Cardiff University School of Medicine, Cardiff, United Kingdom; ⁴School of Medicine, Keele University, Keele, United Kingdom; ⁵Medicine- Endocrinology, University of Chicago, Chicago, USA; ⁶Res Consortium, Andover, United Kingdom

Introduction: Over 10 million thyroid function tests (TFTs) are conducted annually in England, the majority requested from primary care. Previous research found that only 25% of patients treated with Levothyroxine had TSH and FT4 levels within the range encompassing 75% of untreated individuals. This study aimed to further investigate differences in thyroid hormone levels, accounting for diagnostic codes and prescribed doses.

Methods: Using a comprehensive city-wide population dataset, we analysed simultaneous TSH and FT4 results from 47,869 diagnosed hypothyroid patients on Levothyroxine and 393,101 untreated (euthyroid) individuals collected over a 14-year period.

Results: The FT4 distribution in treated individuals closely resembled that of untreated individuals but was consistently shifted toward higher FT4, even at the lowest Levothyroxine dose. This separation increased progressively with dose (F value rising from 1.5 to 4.2). In contrast, TSH distribution differed markedly: while untreated individuals showed a near-normal pattern, those on Levothyroxine displayed a pronounced “hockey stick” distribution, heavily skewed toward low or undetectable TSH, with increasing skewness as dose increased. Among untreated individuals, 90.3% had TSH within the reference range, with only 0.8% exhibiting low FT4. Conversely, just 43.8% of treated patients had TSH within range. Median Levothyroxine doses were higher in men than women across all age groups, peaking at 107 mg/day (men) and 93 mg/day (women) at ages 50–59. With treatment, FT4 rose and TSH fell progressively with age, both effects more pronounced in women. TSH levels in treated and untreated populations aligned only around FT4 = 20 pmol/l—below this, treated individuals had higher TSH, and above it, lower TSH for equivalent FT4.

Conclusion: FT4 and TSH distributions differ substantially between treated and untreated individuals. Only 43.8% of those on Levothyroxine maintained TSH within the reference range, with increasingly unphysiological TSH suppression at higher doses.