Approach to the Patient Considering Thyroid Hormone Deprescribing

Abstract

There is a widespread and increasing thyroid hormone overuse. Levothyroxine, a synthetic form of thyroid hormone, is one of the most prescribed medications in the United States. Many patients are started on levothyroxine for the treatment of mild subclinical hypothyroidism or nonevidence based indications (e.g., euthyroid individuals with weight gain or depression). High-certainty evidence does not show a benefit of thyroid hormone initiation in important patient outcomes (e.g., quality of life or hypothyroid symptoms) for patients with subclinical hypothyroidism. Moreover, thyroid hormone treatment is associated with potential patient burden, harms, and costs. Yet, once levothyroxine is initiated, it is continued lifelong for most patients. Therefore, active interventions are needed to reduce unnecessary thyroid hormone treatment when it is clinically appropriate. Deprescribing is a structured and evidence-based approach to safely taper or discontinue unnecessary or potentially inappropriate medications, aiming to reduce polypharmacy, improve outcomes, and decrease healthcare costs and patient harm. Recent limited data support the feasibility and safety of thyroid hormone deprescribing in selected patients and describe clinician-, patient-, and system-level barriers and facilitators to implementing this intervention. We present three illustrative clinical cases of patients taking levothyroxine to exemplify the rationale and key considerations in thyroid hormone deprescribing, emphasizing that a structured, patient-centered, and closely supervised plan is essential for safe and effective thyroid hormone de-escalation in clinically appropriate situations.

INTRODUCTION

Thyroid hormone use has increased over the past several years, with levothyroxine (a synthetic form of thyroxine) being one of the most frequently prescribed medications in the United States (U.S.) (1–6). In 2023, an estimated 18 million Americans had an active prescription for levothyroxine, with more than 80 million prescriptions dispensed in the same year (7). However, the steady rise in levothyroxine prescriptions is disproportional to the stable low rates of overt hypothyroidism (8), the main indication for thyroid hormone replacement, suggesting that some patients may be receiving levothyroxine for other, potentially inappropriate, indications (4,9,10). Studies have shown that once levothyroxine is initiated, it is continued lifelong for the majority of patients (6).

In this context, it has been suggested that deprescribing can be considered in select patients (4,11,12). The term “deprescribing” refers to the systematic, supervised and evidence-based process of reducing the dose or discontinuing potentially inappropriate or unnecessary medications in a safe and effective manner, aimed to decrease polypharmacy, decrease healthcare costs, improve patient outcomes and reduce patient harm (13–17). Despite a surge of research focusing on deprescribing practices in other settings and contexts, published data on guiding the patient selection and process for deprescribing levothyroxine are limited, posing a major challenge in clinical practice. In this article, we present illustrative clinical cases of patients taking levothyroxine to exemplify the rationale and highlight key considerations in our approach to thyroid hormone deprescribing.

CASE PRESENTATIONS

Case 1

A 75-year-old man is establishing care with a new primary care physician. He feels well except for some minor arthritic pain. Medical history is significant for hypertension treated with lisinopril 40 mg daily and hypothyroidism on a stable dose of levothyroxine 50 mcg daily. On physical examination, his weight is 88 kg with body mass index (BMI) 23 kg/m² and blood pressure is 137/73 mm Hg. The thyroid has normal size and texture on palpation. Current laboratory studies show a serum thyroid stimulating hormone (TSH) 3.8 mIU/L (reference range: 0.34–4.2 mIU/L). Review of his medical records shows that the patient was started on levothyroxine 50 mcg daily 6 years ago based on one set of labs showing TSH 6.0 mIU/L, free thyroxine (FT4) 8.6 pmol/L (reference range: 6.3–14.0 pmol/L), and thyroid peroxidase antibodies (TPOAb) < 2.0 IU/mL (reference range: <16).

Case 2

A 58-year-old post-menopausal woman presents with tiredness and weight gain. Although she follows a daily 1400-kcal diet and exercises for 30 minutes three times a week, she has gained 2 kg in the last month. She has hypercholesterolemia requiring statin therapy. There is no family history of thyroid disease. Her BMI is 38.2 kg/m² and her physical exam is unremarkable. There is no goiter. Three months ago, her TSH was 6.9 mIU/L. At that time, observation was recommended with repeat labs in 3 months. Current thyroid function testing shows TSH 6.6 mIU/L, FT4 10.2 pmol/L, and TPOAb 6 IU/mL. In view of these labs, a therapeutic trial of levothyroxine 50 mcg daily was initiated for treatment of subclinical hypothyroidism. At her follow-up visit 8 weeks later, the patient reports no improvement in her energy level and she has gained an additional 3 kg, while her TSH and FT4 levels are now within the normal reference range. Six months after levothyroxine initiation, TSH and FT4 remain normal, but the patient still complains of fatigue and an additional 2 kg weight gain.

Case 3

A 72-year-old-man with a 10-year history of hypothyroidism was recently diagnosed with atrial fibrillation with rapid ventricular rhythm. He weighs 77 kg and has been on a stable dose of levothyroxine 100 mcg daily for several years. Thyroid function tests prior to levothyroxine initiation are unavailable. Recent laboratory results show TSH 0.51 mIU/L, FT4 12 pmol/L, and TPOAb 52 IU/mL.

Thyroid hormone overuse and overtreatment

Several studies have shown that the increase in thyroid hormone use in recent years is mainly driven by initiation of thyroid hormone for potentially inappropriate indications, including for subclinical hypothyroidism, and sometimes even in euthyroid individuals (4,9,10,18–21). A retrospective, cross-sectional analysis of deidentified administrative claims data linked with laboratory results from OptumLabs Data Warehouse (N=110,842) between 2008 and 2018, showed that 60% of patients had levothyroxine initiated for treatment of mild subclinical hypothyroidism and 30% for normal thyroid function without significant change over time, suggesting substantial overuse of levothyroxine (4). In a recent retrospective multicenter study of 977 U.S. adults who were prescribed levothyroxine for the first time between 2017 and 2020, nearly half of the levothyroxine prescriptions were considered nonevidence based with the most common reason for prescription being an index TSH value of less than 10 mIU/L without previous TSH or thyroxine values (9).

Existing guidelines do not currently support initiation of levothyroxine in patients with mild subclinical hypothyroidism, or other controversial indications, such as thyroid hormone suppression for thyroid nodules, weight gain or depression in euthyroid individuals (22). Because of variability in thyroid function tests, if subclinical hypothyroidism is suspected, labs should be repeated to confirm the diagnosis as the majority often spontaneously revert to euthyroidism (18,22–26). In an observational study including 346,549 patients not taking thyroid specific medication, 3% had mildly elevated TSH levels higher than 5.5 mIU/L but equal to or lower than 10 mIU/L (27). Serum TSH normalized in 62% of these patients without intervention within the following five years (27). In addition, current evidence does not support a consistent benefit from treatment of mild subclinical hypothyroidism. A systematic review and meta-analysis of 21 randomized clinical trials including 2,192 nonpregnant adults with subclinical hypothyroidism showed that thyroid hormone therapy did not improve thyroid-related symptoms (e.g., fatigue, depressive symptoms), cognitive function or general quality of life (28).

Furthermore, levothyroxine treatment requires modification of daily habits, e.g., dosing 30–60 minutes before a meal, monitoring of effects, frequent dose titration, clinic and laboratory visits, and financial costs to patients and society (22). Healthcare expenditures associated with levothyroxine treatment in the U.S. increased from $1.1 billion in 1997 to $3.2 billion in 2016, with an average cost of $1600/year for each patient (including the costs of medical visits and laboratory tests), and up to $4,100/year for patients requiring 3 or more levothyroxine dose adjustments (5,29). Moreover, two-thirds of pregnant women with hypothyroidism report levothyroxine -related treatment burdens, such as financial hardship and difficulty in taking a pill (30).

In addition to inappropriate thyroid hormone initiation, overtreatment with thyroid hormone occurs often, justifying the need for dose de-escalation. Thyroid hormone over-replacement is particularly prominent in older adults with rates up to 48% in some studies (19,31,32). Furthermore, thyroid hormone overtreatment, which may lead to iatrogenic thyrotoxicosis, has been shown to be associated with increased risk of atrial fibrillation, stroke, cardiovascular and all-cause mortality (33–37).

Owing to the treatment burden, potential patient harm and lack of benefit, a multidisciplinary panel employing a novel metric to identify low-value prescribing practices assigned the highest low-value prescription score to the initiation of levothyroxine in patients without overt hypothyroidism (38).

Available data to support thyroid hormone deprescribing in select patients

Discontinuation of medications has been recognized as one of the four most important innovations in healthcare by experts of the Innovation Network—a collaboration of The Commonwealth Fund and the Institute of Health Care Improvement (39). Recently, data have emerged to support thyroid hormone deprescribing in select patients. A systematic review and meta-analysis assessed clinical outcomes following discontinuation of thyroid hormone replacement and identified predictors of successful discontinuation (12). Seventeen observational studies (N=1103 patients) were included. Overall, 37.2% of patients remained euthyroid at follow-up following thyroid hormone discontinuation; a higher proportion of patients with an initial diagnosis of subclinical hypothyroidism were more likely to remain euthyroid compared to those with overt hypothyroidism (35.6% versus 11.8%) (12). Findings on patient self-reported symptoms following levothyroxine discontinuation were mixed, even though none of the included studies systematically assessed symptomatology, and no major adverse events were reported. In a prospective observational study of 802 patients who were on levothyroxine treatment without a documented diagnosis of hypothyroidism, all patients had levothyroxine discontinued and were followed for up to 60 months (20). Over a median follow-up of 18 months, only 23% of patients became hypothyroid necessitating re-initiation of levothyroxine. The authors also concluded that if hypothyroidism did not occur during the first 4 months of levothyroxine discontinuation, the probability of progression to hypothyroidism was negligible (3%).

Finally, a single institution, pilot, double-blind, placebo-controlled randomized controlled trial with a 6-month follow-up that randomized 45 adults with subclinical hypothyroidism on levothyroxine ≤75 mcg/day to continue levothyroxine (N=21) or switch to placebo (N=24) demonstrated feasibility of levothyroxine discontinuation (40). Albeit underpowered, there was no statistically significant difference between the two groups in overall quality of life using validated measures, tiredness, weight, BMI or lipid levels and no significant adverse events were reported. After 6 months of follow-up, 65% of patients who had discontinued levothyroxine (placebo) had normal thyroid tests, 35% had subclinical hypothyroidism, while none developed overt hypothyroidism. Two participants in the placebo group were restarted on levothyroxine, one due to TSH elevation above 10 mIU/L (patient asymptomatic) and another due to persistent fatigue (patient preference), which was subsequently attributed to metastatic rectal adenocarcinoma.

Taken together, the above data suggest that levothyroxine deprescribing is feasible in select patients without any adverse effects. Close clinical follow-up is needed to assess for development of symptoms and need for levothyroxine re-initiation for biochemically demonstrated hypothyroidism.

Barriers and facilitators to thyroid hormone deprescribing

Despite evidence of widespread inappropriate thyroid hormone use and overtreatment, and data supporting the effectiveness of deprescribing in this and other contexts emerging from both observational studies and randomized controlled trials (40–47), several barriers exist to implementing thyroid hormone deprescribing. In a qualitative study where semi-structured interviews were conducted with 19 clinicians practicing in the U.S., including endocrinologists, geriatricians, and primary care providers, the most commonly reported barriers were related to patient-level factors, followed by clinician- and system-level factors (48). Specifically, patient-level barriers included the patients’ perceived need for thyroid hormone use, patient anxiety about potential unfavorable effects related to levothyroxine dose reduction or discontinuation, patient lack of knowledge and misinformation about deprescribing. Physician inertia was also highlighted as a key barrier to thyroid hormone deprescribing, perhaps related to difficultly reconciling benefit with risk, fear of negative effects resulting from deprescribing, and clinic visit time limitations. Moreover, lack of deprescribing guidelines and absence of standardized protocols, insufficient follow up as well as limited availability of older electronic medical records were also reported to hinder deprescribing efforts (48).

Facilitators to thyroid hormone deprescribing have also been reported by clinicians (48). Effective physician-patient communication including direct discussions about potential harms from thyroid hormone overtreatment, long-term plans and setting realistic expectations, as well as establishing a trusting physician-patient relationship to promote familiarity with patient preferences and goals of care were reported to enhance deprescribing conversations. At the patient level, motivation to reduce medication burden and polypharmacy, presence of comorbid medical conditions, such as atrial fibrillation and osteoporosis, and recent initiation of low-dose thyroid hormone replacement were described as facilitators towards thyroid hormone deprescribing (48).

Developing a thyroid hormone deprescribing framework

A patient-centered, stepwise and supervised deprescribing plan that follows a structured framework is essential to successfully deprescribe thyroid hormone in select patients (Figure 1). The deprescribing process is complex and has to follow a systematic approach, starting with clearly outlining the rationale behind the decision to deprescribe and identifying patients for whom thyroid hormone deprescribing would be appropriate. Patients who have an unclear or inappropriate indication for thyroid hormone use (such as transient hypothyroidism due to an episode of subacute thyroiditis, mild elevations of TSH related to obesity, initiation of thyroid hormone in euthyroid patients with non-specific symptoms, etc.), those who are on low or moderate doses of levothyroxine (≤75 mcg/day) and older, frail patients who are asymptomatic, may be candidates for deprescription. Shared decision-making is integral to a successful deprescribing process because it integrates the patients’ values and preferences with the clinicians’ expertise, a collaboration that fosters trust and builds therapeutic alliance. Conversations about thyroid hormone deprescribing should focus on relaying the risks and benefits of deprescribing and take into account patient context (e.g., higher likelihood of progression of subclinical to overt hypothyroidism in younger patients, coexisting hyperlipidemia, thyroid antibody positivity, etc.), discussing alternative options, addressing patients’ fears and concerns and goals of care, and providing support through the deprescribing process. The goal of reducing polypharmacy, patient harm and costs should be clearly indicated.

Figure 1. Proposed Thyroid Hormone Deprescribing Framework.

This figure outlines a stepwise framework to guide clinicians in the deprescribing of thyroid hormone therapy, including rationale, appropriate patient selection and process of deprescribing.

Once the decision is made to deprescribe thyroid hormone, a clear and detailed plan for tapering or discontinuing the medication, frequency and aspects of monitoring, steps to take in the event of adverse effects, and criteria for increasing the dose or resuming the medication if discontinued should be formed. Specifically, it is reasonable to initially monitor thyroid function tests and patient symptoms every 6–8 weeks to make any necessary dose adjustments, with less frequent follow-up once the patient is clinically and biochemically euthyroid.

Overall, successful implementation of thyroid hormone deprescribing requires shared decision-making between patients and clinicians, robust communication across healthcare teams, patient and provider education, high-certainty evidence regarding the effects of discontinuing levothyroxine, structured deprescribing frameworks, and effective integration into routine clinical care. Figure 2 demonstrates key factors relating to implementation of thyroid hormone deprescribing.

Figure 2. Key factors relating to implementation of thyroid hormone deprescribing.

This figure illustrates the main factors at the clinician, patient and system levels affecting the safe and effective reduction or discontinuation of thyroid hormone therapy in clinical practice.

BACK TO THE CASES

Case 1

This patient was initiated on levothyroxine based on one set of labs that showed a mildly elevated serum TSH, normal FT4 and negative TPOAb (suggesting that an underlying autoimmune thyroid etiology is less likely). Repeat thyroid function tests were not performed prior to levothyroxine prescription, thus subclinical hypothyroidism was not confirmed prior to treatment. Considering that mild TSH elevations may be a normal manifestation of aging and not necessarily represent thyroid dysfunction, and that spontaneous resolution of mild TSH elevations commonly occurs on repeat testing, deprescribing levothyroxine in this patient by discontinuation of the medication is reasonable to pursue. In addition, the patient is older than 65 years-old and studies have not shown that treating subclinical hypothyroidism in this population confers benefit. Following levothyroxine discontinuation, the patient remained clinically and biochemically euthyroid a year later.

Case 2

This 58-year-old patient has confirmed subclinical hypothyroidism with at least two consecutive mildly elevated TSH levels and normal FT4, but negative TPOAb and normal thyroid exam suggesting that progression to overt hypothyroidism is less likely. In addition, obesity may be associated with mild TSH elevations irrespective of underlying autoimmune thyroid disease. A therapeutic levothyroxine trial did not lead to symptomatic improvement in this patient, despite resulting in biochemical euthyroidism. Levothyroxine was subsequently discontinued and the patient was evaluated for obstructive sleep apnea and started on a glucagon-like peptide-1 receptor agonist (GLP1RA). Six months following management of her sleep apnea and initiation of weight loss medication, the patient reported a weight loss of 13 kg, improvement in her energy level and thyroid function tests had remained in the normal reference range.

Case 3

The indication for levothyroxine initiation in this 72-year-old man is unclear due to lack of available medical records. His current evaluation confirms the presence of Hashimoto thyroiditis due to positive TPOAb, a low normal serum TSH, and a high normal serum FT4. In the context of new-onset cardiac arrhythmia and his current labs, it is reasonable to consider thyroid hormone deprescription. Given his long-standing levothyroxine treatment at a dose that is almost 80% of his estimated weight-based dose, the deprescribing plan involved gradual tapering of the levothyroxine dose with frequent laboratory and symptoms assessments during follow-up (Table 1). The patient remained asymptomatic during the deprescribing process; however, due to a rise in his TSH above 10 mIU/L when reducing the levothyroxine dose from 50 mcg to 25 mcg daily, it was elected that the patient would continue 50 mcg levothyroxine daily.

Table 1.

Laboratory and symptoms monitoring during levothyroxine deprescribing (Case 3)

Timing	Levothyroxine dose	TSH (mIU/L)	FT4 (pmol/L)	Symptoms of Hypothyroidism
Initial visit	100 mcg	0.51	12	None reported
8 weeks later	75 mcg	3.3	10	Asymptomatic
16 weeks later	50 mcg	6.1	8.7	Asymptomatic
30 weeks later	25 mcg	11.0	7.8	Asymptomatic

Reference ranges: TSH, 0.34–4.2 mIU/L; FT4, 6.3–14.0 pmol/L.

CONCLUSION

Inappropriate use of thyroid hormone and thyroid hormone overtreatment are increasingly encountered in clinical practice. Despite the lack of proven benefit, associated patient risks and burden, and clinical practice guidelines advising against routine use of levothyroxine for patients without overt hypothyroidism, a uniform approach to thyroid hormone deprescribing and its effective implementation in clinical settings has not yet been established. Limited data support the feasibility and safety of thyroid hormone deprescribing in select patients. A structured, patient-centered, and closely supervised deprescribing plan is essential for the safe and effective de-escalation of thyroid hormone therapy in patients for whom it is clinically appropriate (e.g., older adults, patients with mild subclinical hypothyroidism or unclear indication). Further studies are needed to help overcome barriers to adoption of thyroid hormone deprescribing in clinical practice and inform clinical guidelines.

DATA AVAILABILITY

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.