Intended for healthcare professionals

Practice Rapid Recommendations

Thyroid hormones treatment for subclinical hypothyroidism: a clinical practice guideline

BMJ 2019; 365 doi: https://doi.org/10.1136/bmj.l2006 (Published 14 May 2019) Cite this as: BMJ 2019;365:l2006

Visual summary of recommendation

or No thyroid hormones Thyroid hormones Levothyroxine Interventions compared Recommendation Population Adults with subclinical hypothyroidism Does not apply to: May not apply to: Including: Women who are or trying to become pregnant Patients with no symptoms (diagnosed after screening) Patients with non-specific symptoms Patients with severe symptoms Patients with TSH above 20 mIU/L Young adults (such as <30 years) Elevated levels of thyroid stimulating hormone (TSH) Normal free T4 (thyroxine) levels

We recommend against thyroid hormone therapy for patients with subclinical hypothyroidism Moredetails Strong All or nearly all informed people would likely want this option. Benefits outweigh harms for almost everyone. Weak Most people would likely want this option. Benefits outweigh harms for the majority, but not for everyone. Weak Most people would likely want this option. Benefits outweigh harms for the majority, but not for everyone. Strong All or nearly all informed people would likely want this option. Benefits outweigh harms for almost everyone.

COMPARISON OF BENEFITS AND HARMS
For the elderly - about 65 years and older All the evidence, including from the largest TRUST trial conducted among an elderly population with comorbidities (see Figure 2)
No thyroid hormones Thyroid hormones Evidence quality EQ-5D score: -0.59-1 (High better) After 1 year No important difference The panel found that this difference was not important for most patients, because the intervention effects were negligible and/or very imprecise, for example confidence intervals that include both important benefit and harm

No important difference General quality of life High More 0.85 0.83

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on general quality of life High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-100 (Low better)

No important difference Thyroid-related symptoms High More 16.7 16.5

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on thyroid-related symptoms High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low

No important difference Fatigue / tiredness High More 28.6 29.0

Minimum clinically important difference is 9 points 9 CID Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on fatigue / tiredness High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-21 (Low better)

3.6 Depressive symptoms High More No important difference 3.3

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on depressive symptoms High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-infinity (High better) After 1.5 years

No important difference Cognitive function High More 27.1 28.1

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on cognitive function High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Events per 1000 people After 2 years

27 Mortality Low More 14 No important difference

Risk of Bias No serious concerns Imprecision Very serious Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone may have little or no effect on mortality Low GRADE score, because of: Only a few deaths were observed, in a single trial. We are 95% confident that the difference is between 5 fewer to 62 more deaths per 1000 patients taking levothyroxine GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low

48 Cardiovascular events Low More 54 No important difference

Risk of Bias No serious concerns Imprecision Very serious Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone may have little or no effect on cardiovascular events Low GRADE score, because of: Only a few cardiovascular events were observed, in a single trial. We are 95% confident that the difference is between 28 fewer to 62 more events per 1000 patients taking levothyroxine GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-100 (Low better)

10.9 Side effects Moderate More No important difference 10.3

Risk of Bias No serious concerns Imprecision Serious Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone probably has little or no effect on side effects Hyperthyroidism due to overdosing Moderate GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
See patient decision aids
See all outcomes
For younger people (such as 65 and younger) Lorem ipsum dolor sit amet, consectetuer adipiscing elit, sed diam nonummy nibh euismod tincidunt ut Evidence excluding the large TRUST trial, conducted among the elderly, which dominated the results of the systematic review
No thyroid hormones Thyroid hormones Evidence quality EQ-5D score: -0.59-1 (High better) After 1 year No important difference The panel found that this difference was not important for most patients, because the intervention effects were negligible and/or very imprecise, for example confidence intervals that include both important benefit and harm

No important difference General quality of life Moderate More 0.85 0.82

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency Serious Publication bias No serious concerns Thyroid hormone probably has little or no effect on general quality of life Moderate GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-100 (Low better)

No important difference Thyroid-related symptoms High More 16.7 16.4

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on thyroid-related symptoms High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low

No important difference Fatigue / tiredness Moderate More 28.6 29.0

Minimum clinically important difference is 9 points 9 CID Risk of Bias No serious concerns Imprecision No serious concerns Indirectness Serious Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone probably has little or no effect on fatigue / tiredness Moderate GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-21 (Low better)

3.6 Depressive symptoms High More No important difference 3.3

Risk of Bias No serious concerns Imprecision No serious concerns Indirectness No serious concerns Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone has little or no effect on depressive symptoms High GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-infinity (High better) After 1.5 years

No important difference Cognitive function Low More 27.1 29.7

Risk of Bias No serious concerns Imprecision Serious Indirectness Serious Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone may have little or no effect on cognitive function Low GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Events per 1000 people After 2 years

27 Mortality Very low More 14 No important difference

Risk of Bias No serious concerns Imprecision Very serious Indirectness Serious Inconsistency No serious concerns Publication bias No serious concerns We are uncertain whether thyroid hormone increases or decreases mortality Very Low GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low

48 Cardiovascular events Very low More 54 No important difference

Risk of Bias No serious concerns Imprecision Very serious Indirectness Serious Inconsistency No serious concerns Publication bias No serious concerns We are uncertain whether thyroid hormone increases or decreases cardiovascular events Very Low GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
Mean score: 0-100 (Low better)

10.9 Side effects Low More No important difference 10.3

Risk of Bias No serious concerns Imprecision Serious Indirectness Serious Inconsistency No serious concerns Publication bias No serious concerns Thyroid hormone may have little or no effect on side effects Hyperthyroidism due to overdosing Low GRADE score, because of: GRADE certainty ratings The authors have a lot of confidence that the true effect is similar to the estimated effect The authors believe that the true effect is probably close to the estimated effect High The true effect might be markedly different from the estimated effect The true effect is probably markedly different from the estimated effect Moderate Low Very low
See patient decision aids
See all outcomes
Daily oral medication, normally tablets, often long-term treatment Overdosage can lead to hyperthyroidism symptoms Should be taken 4 hours apart from any products containing calcium or iron Long-term regular visits and blood samples to monitor hormone levels Regular visits and blood samples to monitor progression or resolution The panel expects little variability in how patients weigh the lack of benefit against the possible harms Potential harms, and in particular risk of dying, may be valued differently by patients depending on their age, quality of life and comorbidities Values and preferences TSH levels may vary with stress, transient disease or with age. Elevated levels thus often revert to normal without treatment There is no clear evidence on how to reliably attribute symptoms to subclinical hypothyroidism TSH levels and symptoms Key practical issues NO THYROID HORMONES THYROID HORMONES

©BMJ Publishing Group Limited.

Disclaimer: This infographic is not a validated clinical decision aid. This information is provided without any representations, conditions or warranties that it is accurate or up to date. BMJ and its licensors assume no responsibility for any aspect of treatment administered with the aid of this information. Any reliance placed on this information is strictly at the user's own risk. For the full disclaimer wording see BMJ's terms and conditions: https://www.bmj.com/company/legal-information/

Find recommendations, evidence summaries and consultation decision aids for use in your practice
  1. G E Bekkering, guideline expert, methodologist1 2,
  2. T Agoritsas, doctor in general internal medicine, expert in shared decision making research3 4,
  3. L Lytvyn, patient partnership expert, methodologist4,
  4. A F Heen, general internist5,
  5. M Feller, general practitioner6 7,
  6. E Moutzouri, doctor in general internal medicine6 7,
  7. H Abdulazeem, patient partner8,
  8. B Aertgeerts, general practitioner1 2,
  9. D Beecher, patient partner9,
  10. J P Brito, endocrinologist10,
  11. P D Farhoumand, doctor in general internal medicine11,
  12. N Singh Ospina, endocrinologist12,
  13. N Rodondi, doctor in general internal medicine6 7,
  14. M van Driel, general practitioner13,
  15. E Wallace, general practitioner14,
  16. M Snel, endocrinologist15,
  17. P M Okwen, general practitioner16,
  18. R Siemieniuk, doctor in general internal medicine17,
  19. P O Vandvik, doctor in general internal medicine18 19 20,
  20. T Kuijpers, guideline expert, methodologist21,
  21. M Vermandere, general practitioner1
  1. 1Academic Centre for General Practice, Department of Public Health and Primary Care, KU Leuven, Belgium
  2. 2Belgian Centre for Evidence-Based Medicine, Cochrane Belgium
  3. 3Division of General Internal Medicine and Division of Clinical Epidemiology, University
  4. Hospitals of Geneva, Geneva, Switzerland
  5. 4Department of Health Research Methods, Evidence and Impact, McMaster University, Hamilton, Canada
  6. 5Department of Medicine, Innlandet Hospital Trust-division, Gjøvik, Norway
  7. 6Institute of Primary Health Care (BIHAM), University of Bern, Bern, Switzerland
  8. 7Department of General Internal Medicine, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland
  9. 8Munich, Germany
  10. 9Milan, Italy
  11. 10Knowledge and Evaluation Research Unit in Endocrinology (KER_Endo), Division of Endocrinology, Diabetes, Metabolism and Nutrition, Department of Medicine, Mayo Clinic, Rochester, MN 55905, USA.
  12. 11Division General Internal Medicine, University Hospitals of Geneva, 1205 Geneva, Switzerland
  13. 12Department of Medicine, Division of Endocrinology, University of Florida, Gainesville, Florida, USA
  14. 13Primary Care Clinical Unit, Faculty of Medicine, University of Queensland, Brisbane Qld 4029, Australia
  15. 14HRB Centre for Primary Care Research and Department of General Practice, Royal College of Surgeons in Ireland (RCSI), Dublin, Ireland
  16. 15Department of Endocrinology/General Internal Medicine, Leiden University Medical Center, Leiden, Netherlands
  17. 16Effective Basic Services (eBASE), Bamenda, Cameroon
  18. 17Department of Health Research Methods, Evidence and Impact, McMaster University, Hamilton, Ontario, Canada
  19. 18Institute of Health and Society, Faculty of Medicine, University of Oslo, Oslo, Norway
  20. 19Department of Medicine, Innlandet Hospital Trust-division, Gjøvik, Norway
  21. 20Norwegian Institute of Public Health, Oslo, Norway
  22. 21Dutch College of General Practitioners, Utrecht, Netherlands
  1. Correspondence to: G E Bekkering trudy.bekkering{at}kuleuven.be

Abstract

Clinical question What are the benefits and harms of thyroid hormones for adults with subclinical hypothyroidism (SCH)? This guideline was triggered by a recent systematic review of randomised controlled trials, which could alter practice.

Current practice Current guidelines tend to recommend thyroid hormones for adults with thyroid stimulating hormone (TSH) levels >10 mIU/L and for people with lower TSH values who are young, symptomatic, or have specific indications for prescribing.

Recommendation The guideline panel issues a strong recommendation against thyroid hormones in adults with SCH (elevated TSH levels and normal free T4 (thyroxine) levels). It does not apply to women who are trying to become pregnant or patients with TSH >20 mIU/L. It may not apply to patients with severe symptoms or young adults (such as those ≤30 years old).

How this guideline was created A guideline panel including patients, clinicians, and methodologists produced this recommendation in adherence with standards for trustworthy guidelines using the GRADE approach.

The evidence The systematic review included 21 trials with 2192 participants. For adults with SCH, thyroid hormones consistently demonstrate no clinically relevant benefits for quality of life or thyroid related symptoms, including depressive symptoms, fatigue, and body mass index (moderate to high quality evidence). Thyroid hormones may have little or no effect on cardiovascular events or mortality (low quality evidence), but harms were measured in only one trial with few events at two years’ follow-up.

Understanding the recommendation The panel concluded that almost all adults with SCH would not benefit from treatment with thyroid hormones. Other factors in the strong recommendation include the burden of lifelong management and uncertainty on potential harms. Instead, clinicians should monitor the progression or resolution of the thyroid dysfunction in these adults. Recommendations are made actionable for clinicians and their patients through visual overviews. These provide the relative and absolute benefits and harms of thyroid hormones in multilayered evidence summaries and decision aids available in MAGIC (https://app.magicapp.org/) to support shared decisions and adaptation of this guideline.

What you need to know

  • Thyroid hormones should not be routinely offered to adults with SCH (strong recommendation according to GRADE).

  • Thyroid hormones do not lead to important benefits for adults with SCH for quality of life or thyroid related symptoms including depressive symptoms and fatigue

  • Taking a pill and attending periodic testing on an ongoing or lifelong basis is burdensome

  • This recommendation builds on a recent systematic review and meta-analysis, which included the results of a new randomised controlled trial

  • If implemented, this recommendation may substantially alter prescribing trends, which show that thyroid hormones are increasingly prescribed, most probably due to SCH

Subclinical hypothyroidism (SCH) is a biochemical state. The thyroid stimulating hormone (TSH) level is elevated, but the free T4 (thyroxine) level is normal. Some people may experience symptoms linked to the abnormality. Other data have suggested links to overt hypothyroidism and adverse outcomes such as increased risk of coronary heart disease. So it is reasonable to ask whether treatment with thyroid hormones might help symptoms, prevent overt hypothyroidism, or avoid longer term heart problems. Box 1 extends and references this understanding.

Box 1

Overview of subclinical hypothyroidism (SCH)

What is SCH?

The definition of SCH varies. About 90% of all patients with SCH have TSH levels between 4 and 10 mIU/L.1 TSH levels may increase with age,2 and a slight increase of TSH may be normal for older people.

About 62% of TSH levels between 4 and 10 mIU/L normalise without intervention within five years.3 There is biological variation in TSH levels. Levels may rise in response to stress and transient disease.4 This biological variation in TSH values, means that one abnormal TSH level should be followed by a repeat blood test to confirm the diagnosis.5

According to the International Classification of Diseases (ICD), SCH does not have a separate code, but is typically labelled as “hypothyroidism, unspecified”.

How common is it?

It affects 4-20% of the adult population.6 This wide variation is due to poor consensus about the cut-off level for the diagnosis of SCH and regional variation between populations. It is more common in women, in older people, and those of white ethnicity.6

What are the symptoms?

Around 1 in 3 patients with SCH have no symptoms at all.7 The type of symptoms people link to SCH include those of overt hypothyroidism: fatigue, muscle cramps, cold sensitivity, dry skin, voice changes, and constipation.8 Other symptoms include poor memory, slowed thinking, weak muscles, puffy eyes, anxiety, and depression.7910 Many of these symptoms are not specific to hypothyroidism. Around 20-25% of people with normal TSH levels report one or two of these symptoms.7 The relation between symptoms and biochemical TSH levels remains unclear.

What is the long term outlook?

The risk of progression to overt hypothyroidism ranges between 2% and 5% a year.11 Presence of antibodies to thyroid peroxidase and, in particular, higher TSH levels increase this risk.111213

Observational data suggest that SCH is associated with an increased risk of coronary heart disease, heart failure, and cardiovascular mortality, particularly in those with TSH levels >10 mIU/L.114 Such associations were not found for most adults with TSH levels of 5-10 mIU/L.114

RETURN TO TEXT

This guideline was triggered by a systematic review,15 summarising all studies on this question. It includes a large and new trial specifically in older people with TSH.16 The results of the review might change practice. The main infographic provides an overview of the relative and absolute benefits and harms of treating SCH with thyroid hormones in standard GRADE format. Box 2 shows all of the articles and evidence linked in this Rapid Recommendation package.

Box 2

Linked sources in this BMJ Rapid Recommendations cluster

  • Bekkering GE, Agoritsas T, Lytvyn L, et al. Thyroid hormones for subclinical hypothyroidism: a clinical practice guideline. BMJ 2019;365:l2006

    • Summary of the results from the Rapid Recommendations process

  • Feller M, Snel M, Moutzouri E, et al. Association of thyroid hormone therapy with quality of life and thyroid-related symptoms in patients with subclinical hypothyroidism: a systematic review and meta-analysis. JAMA 2018;320:1349-59 doi:10.1001/jama.2018.13770

    • Review and meta-analysis of all available randomised trials that assessed thyroid hormones for subclinical hypothyroidism

  • MAGICapp (https://app.magicapp.org/public/guideline/nyqWPn)

    • Expanded version of the results with multilayered recommendations, evidence summaries, and decision aids for use on all devices

RETURN TO TEXT

Current practice

When to test for SCH

Historically, US guidelines recommended five-yearly screening of asymptomatic adults aged 35 years and older to identify thyroid dysfunction,17 but it is uncertain if such screening has any clinical benefits.18

In clinical practice, thyroid function can be checked as part of routine screening or for diagnostic purposes in those with possible hypothyroidism based on physical or mental health signs and symptoms. In the UK about 25% of adults have thyroid function tests every year.19 A recent overview showed an increase in the use of thyroid function tests over time.20

Patients and clinicians (general practitioners, internists, and endocrinologists) are commonly faced with abnormal thyroid function tests consistent with SCH. All parties collaboratively need to decide if and how to act.

When to treat SCH

Guidelines generally recommend thyroid hormones for adults with TSH levels above 10 mIU/L. For those with lower TSH levels, most guidelines recommend treatment only when people are younger, symptomatic, or have other indications for prescribing (such as cardiovascular disease or antibodies to thyroid peroxidase). Table 1 summarises current guidance from various organisations.582122

Table 1

Current guidance on thyroid hormone treatment for subclinical hypothyroidism

View this table:

In many countries, the use of levothyroxine is increasing,4 with a top ranking among the most prescribed drugs in the US in 2015.23 Increasing treatment of SCH with thyroid hormone, and of levothyroxine in particular, is the most likely explanation for this increase. Research showed that prevalence of treated SCH has doubled from 1996 to 2006 and that people with TSH <10 mIU/L were prescribed levothyroxine 1.3 times more in 2009 than in 2001 in the UK.19 This increased prevalence of treated SCH was confirmed in Norwegian population surveys, despite a stable prevalence of the condition itself.24 Other evidence includes a study by Taylor19 showing that a third of adults were offered treatment after a single TSH testing. Not all of these adults may actually have SCH as TSH levels fluctuate and may revert to normal without treatment.

Finally, some patients with symptoms may receive a trial of levothyroxine to evaluate improvement of symptoms, but in such an approach it is difficult to separate real from placebo effects. Once levothyroxine is started, most adults stay on the drug for several years.19

How was this recommendation created?

Our international panel included methodologists, general practitioners, internists, endocrinologists, and patient partners with subclinical hypothyroidism (SCH) (see appendix 1 on bmj.com for details of panel members). They decided on the scope of the recommendation and identified patient-important outcomes to inform the recommendations.

The panel met online to discuss the evidence and formulate a recommendation. No member had a financial conflict of interest; intellectual and professional conflicts were minimised and are transparently described (appendix 2 on bmj.com). The panel followed the BMJ Rapid Recommendations procedures for creating a trustworthy recommendation,27 including using the GRADE approach to critically appraise the evidence and create recommendations (appendix 3 on bmj.com).28 The panel considered the benefits, harms, and burdens and other practical issues related to thyroid hormones in the context of SCH, as well as expected variations in patient values and preferences.29 Within the GRADE approach, recommendations can be either strong or weak (also known as conditional), and for or against a specific course of action.30

The evidence

The systematic review that triggered this guideline compared the effects of thyroid hormone treatment to that of no treatment or placebo in adults with SCH.15Figure 2 presents an overview of the characteristics of the randomised controlled trials (RCTs) and participants included in the review.

Fig 2
Fig 2

Characteristics of participants and trials included in the systematic review of the effects of thyroid hormone treatment for subclinical hypothyroidism (SCH)

The systematic review includes 21 studies; the largest of which is the TRUST trial. This study examined the effects of thyroid hormone for SCH in over 700 elderly people aged 65-93 years.16 Many participants had common comorbidities: 14% had ischaemic heart disease, 12% atrial fibrillation, 51% hypertension, 16% diabetes, and 12% had osteoporosis,16 representative of a general elderly population. Because of its size, results of the TRUST trial dominate the results of the systematic review. For that reason, and to estimate the effect in younger patients, the guideline panel also looked at the results of the systematic review excluding the TRUST trial (see below).

The evidence applies to a broad range of adults with SCH as included in the systematic review. Some adults consulted physicians because of symptoms, others did not and were diagnosed after routine screening. The severity of symptoms, reported in seven studies (1263 adults) was mild to moderate. No studies included only patients with severe symptoms. For younger people (that is, those <30 years of age) and for patients with unusually high TSH levels (>20 mIU/L with normal T4 levels) the evidence remains more indirect, although this concerns only a small minority of patients.

Understanding the recommendation

Who does it apply to?

The recommendation applies to most adults, with SCH after at least two consecutive thyroid function tests, with or without mild to moderate symptoms, who are considering initiating thyroid hormone treatment. Box 3 shows situations where the guideline does not or may not apply.

Box 3

Exceptions to this BMJ Rapid Recommendation

This recommendation does not apply to

  • Women who are trying to become pregnant. Such women were excluded from the studies. A systematic review of observational studies suggests that pregnant women with SCH may be at increased risk of adverse outcomes for both mother and baby.25 Guidelines recommend levothyroxine for pregnant women depending on TSH level and presence of antibodies to thyroid peroxidase26

  • Those with very high TSH levels (>20 mIU/L) and with normal T4 (thyroxine) levels. These findings could suggest overt hypothyroidism but would affect only a few patients

This recommendation may not apply to

  • Those with severe symptoms, as few were included in the studies reviewed. However, there is no clear evidence on how to attribute symptoms to SCH reliably, even with severe symptoms

  • Very young adults (such as ≤30 years old). Few of these patients were included in the studies, probably because SCH is so uncommon at younger ages

  • Women at risk of unplanned pregnancy. Clinicians may consider offering thyroid hormones because pregnant women with SCH may be at increased risk of adverse outcomes for mother and baby25

  • Patients who already take thyroid hormones. The evidence presented here looked at the effect of starting medication and only indirectly informs stopping it

RETURN TO TEXT

Absolute benefits and harms

The panel made a strong recommendation against thyroid hormones for SCH, because there were no important benefits from treatment. In addition, we cannot rule out the possibility of harms.

For older people (≥65 years)

There was high certainty that there is little to no difference in general quality of life (QoL), thyroid related symptoms, depressive symptoms, fatigue, cognitive function, muscle strength, and body mass index (BMI). The results are consistent across these outcomes, which strengthens our confidence that there really is a lack of benefit (see main infographic).

On top of the absence of benefit, the panel were concerned about a signal of harm in those treated. There were between five fewer and 62 more deaths per year in the treatment group (this is the 95% confidence interval). This interval includes the possibility of benefit (5 fewer deaths) as well as harm (62 more deaths). Additionally, these deaths were evaluated in only one trial with a two year follow-up.16 For these reasons, the panel had low certainty in this estimate. None the less, the panel agreed that the possibility of harms contributes towards the strong recommendation.

For younger people (such as <65)

There was no important benefit shown in younger groups. However, the panel’s certainty in the estimates was slightly lower. There is moderate to high certainty that such patients experience little or no benefit from thyroid hormone therapy for SCH.

The panel re-analysed the data without TRUST (the largest trial, performed exclusively in older people). The panel examined the age distribution of each study’s participants. The results in younger people remain consistent: probably no evidence of any benefits, and possibly little or no difference in risk for harms. For some outcomes, uncertainty increased: for example, for fatigue, certainty was rated down to moderate due to indirectness (the evidence only comes from older adults). There is low certainty about the lack of effect on cognitive function, but the panel recognises that this outcome is less relevant to younger, healthier patients.

The same is true for harms. However, the panel was concerned about the burden of lifelong treatment and the limited evidence about possible long term harms of thyroid hormones (such as adverse cardiovascular effects). In addition, patients may experience a delay in diagnosis of another condition (such as mood disorder).

Values and preferences

The panel expects little variability in how patients perceive the lack of benefit. Harms may be more important as SCH is not a fatal disease, and most people are reasonably well when they are diagnosed. In addition, potential harms, and in particular risk of dying, may be valued differently by patients depending on their quality of life and comorbidities.

Practical issues

Figure 3 outlines the key practical issues about the use or non-use of thyroid hormones. The option to treat is more burdensome for patients as treatment requires daily and possibly long term medication, follow-up, and blood tests. Both treating and not treating may result in anxiety.

Fig 3
Fig 3

Practical issues about the use or non-use of thyroid hormones for subclinical hypothyroidism (SCH)

Cost and resources

Although we did not take costs and resources into account beyond direct costs to patients (such as out-of-pocket costs), thyroid hormones cannot be cost effective given the lack of important benefit, potential for harm, and associated costs.

Uncertainty

Future research could explore whether there is an unidentified subgroup of patients who do benefit from treatment. No evidence of a potential subgroup or even a trend was observed in the current body of evidence, consistently across outcomes. Such research could consider whether there is more benefit in groups of people for whom there is less direct evidence and therefore more uncertainty, such as

  • Younger people (about ≤30 years old)

  • People with more severe symptoms.

There is uncertainty about potential harms, as these were studied only in the TRUST trial, which found only a few events after a follow-up of only two years. However, this uncertainty becomes important only when there is evidence of benefit.

Updates to this article

Table 2 shows the evidence that has emerged since the publication of this article. As new evidence is published, a group will assess this new evidence and make a judgment on the extent that it is expected to alter the recommendation.

Table 2

New evidence which has emerged after initial publication

View this table:

At the time of publication, we identified one new trial in trial registries:

  • IEMO, assessing the effect of thyroid hormones versus placebo in elderly aged 80 years with SCH, 80-plus thyroid trial (NTR3851 in Netherlands Trial Register).

How patients were involved in the creation of this article

Two people with lived experience of subclinical hypothyroidism were members of the panel and participated in the whole process. They identified and rated outcomes, and helped lead the discussion on values and preferences in a videoconference and in email discussions with the full panel. They noted patients may feel anxious about deteriorating or developing overt hypothyroidism when no treatment was given. To address this, regular follow-up is very important. They also mentioned that it is difficult for patients to make a decision when feeling unwell. We thank them for their contribution.

Footnotes

  • This BMJ Rapid Recommendation article is one of a series that provides clinicians with trustworthy recommendations for potentially practice changing evidence. BMJ Rapid Recommendations represent a collaborative effort between the MAGIC group (http://magicproject.org/) and The BMJ. A summary is offered here, and the full version including decision aids is on the MAGICapp (https://app.magicapp.org/public/guideline/nyqWPn), for all devices in multilayered formats. Those reading and using these recommendations should consider individual patient circumstances, and their values and preferences and may want to use consultation decision aids in MAGICapp to facilitate shared decision making with patients. There will no active dissemination of these results to patients. We encourage adaptation and contextualization of our recommendations to local or other contexts. Those considering use or adaptation of content may go to MAGICapp to link or extract its content or contact The BMJ for permission to reuse content in this article.

  • Competing interests: All authors have completed the BMJ Rapid Recommendations interests disclosure form and a detailed, contextualised description of all disclosures is reported in appendix 2 on bmj.com. As with all BMJ Rapid Recommendations, the executive team and The BMJ judged that no panel member had any relevant financial conflict of interest. Professional and academic interests are minimised as much as possible, while maintaining necessary expertise on the panel to make fully informed decisions. M Feller, M Snel, E Moutzouri, and N Rodondi participated in writing the systematic review that formed the evidence base for this guideline. JP Brito and N Singh Ospina wrote an editorial about the overuse of levothyroxine.

  • Funding: This guideline was not funded.

  • Transparency: G E Bekkering affirms that the manuscript is an honest, accurate, and transparent account of the recommendation being reported; that no important aspects of the recommendation have been omitted; and that any discrepancies from the recommendation as planned (and, if relevant, registered) have been explained.

References

View Abstract