In adults with a mildly underactive thyroid gland (known as subclinical hypothyroidism), thyroid hormones should not be routinely offered, say a panel of international experts in The BMJ today.
Their advice is part of The BMJ's 'Rapid Recommendations' initiative - to produce rapid and trustworthy guidance based on new evidence to help doctors make better decisions with their patients.
Subclinical hypothyroidism (SCH) occurs when the thyroid stimulating hormone (TSH) levels in the bloodstream are slightly raised, while the level of thyroid hormone remains normal.
Subclinical hypothyroidism can only be detected by a blood test. In the UK, about 25% of adults have thyroid function tests every year and guidelines tend to recommend hormones for those with TSH levels above 10 mIU/L (a normal TSH level is 0.4 to 4.0 mIU/L).
Research shows that hormone treatment for SCH has doubled from 1996 to 2006.
So an international panel of clinicians and patients with SCH used a recent detailed analysis of the latest evidence (a systematic review of 21 trials and about 2,200 participants comparing the effects of thyroid hormone treatment with no treatment or placebo in adults with SCH) to develop recommendations for clinical practice.
Using the GRADE approach (a system used to assess the quality of evidence), the panel makes a strong recommendation against thyroid hormones for almost all adults with SCH, because there were no benefits from treatment on fatigue, low mood, weight gain or any other outcomes that were tested.
What's more, they say taking a pill and attending lifelong check-ups is burdensome, and they cannot rule out the possibility of harms.
And while they did not take costs and resources into account beyond direct costs to patients, they say thyroid hormones cannot be cost effective.
The recommendation does not apply to women who are trying to become pregnant or patients with particularly high TSH levels (above 20 mIU/L). It may also not apply to patients with severe symptoms or some aged under 30.
Future research could explore whether there is an unidentified group of patients who might benefit from treatment, say the authors.
In the meantime, if implemented, this recommendation "may substantially alter prescribing trends" they conclude.