Predictors for thyroid dysfunction after discontinuation of levothyroxine in children and adolescents with Hashimoto thyroiditis
Article information
Abstract
Purpose
Few data on the clinical course after levothyroxine (L-T4) discontinuation in pediatric patients with Hashimoto thyroiditis (HT) are available. We investigated outcomes and predictors for successful withdrawal from L-T4 among children with HT.
Methods
Among 168 patients diagnosed with HT between January 2000 and March 2021 at Seoul National University Children’s Hospital and in whom L-T4 therapy was initiated during childhood, we attempted to discontinue this therapy in 47, 3 boys and 44 girls. L-T4 was restarted when patients developed overt or subclinical hypothyroidism (thyroid-stimulating hormone [TSH] levels≥10 mIU/L) after L-T4 discontinuation.
Results
Median age at discontinuation was 15.4 years (12.7–18.4 years) with a median duration of L-T4 therapy of 47 months (20.3–80.3 months). During the median 30 months of follow-up (10.6–61.0 months) after L-T4 discontinuation, 33 (70.2%) developed thyroid dysfunction. Among these patients, 17 were eventually restarted on L-T4. TSH levels over 50 mIU/L at L-T4 initiation (hazard ratio, HR 3.5, P=0.002), age under 12 years at L-T4 discontinuation (HR 11.1, P=0.0001), and TSH levels higher than the upper 50% of normal (above 2.25 mIU/L in the present study) at L-T4 discontinuation (HR 2.7, P=0.014) were significantly predictive for overt hypothyroidism or subclinical hypothyroidism after L-T4 discontinuation. In addition, age under 12 years at L-T4 discontinuation was only predictive factor for restarting L-T4 medication (HR 4.3, P=0.012).
Conclusions
L-T4 discontinuation in pediatric patients with HT resulted in thyroid dysfunction in 70.2% of cases; 36.2% of patients who attempted discontinuation required resumption of L-T4. Older age and lower TSH levels at L-T4 discontinuation were advantageous for successful withdrawal.
Highlights
· Levothyroxine (L-T4) discontinuation in pediatric patients with Hashimoto thyroiditis resulted in thyroid dysfunction in 70.2% of cases; 36.2% of patients who attempted discontinuation required resumption of L-T4. Older age and lower thyroid-stimulating hormone levels at L-T4 discontinuation were advantageous for successful withdrawal.
Introduction
Hashimoto thyroiditis (HT) is the most common cause of acquired hypothyroidism in the pediatric population. The prevalence of HT in this group is approximately 1% to 3% and varies according to ethnicity, age, sex, iodine status, and presence of coexisting disease. HT has a peak incidence during adolescence [1-3]. At diagnosis, thyroid functional status varies in pediatric HT patients: 52.1% present as euthyroid, 22.1% have overt hypothyroidism, 19.1% have subclinical hypothyroidism (SCH), and 6.5% present with overt or subclinical hyperthyroidism [4].
Overt hypothyroidism is immediately treated with levothyroxine (L-T4) to facilitate normal growth and neurodevelopment. Indications for initiation of L-T4 in SCH patients are less clear; the cutoff thyroid-stimulating hormone (TSH) level indicating a need for L-T4 therapy has not been determined because the effect of SCH on children's growth and health is unclear [5]. In clinical practice, most pediatricians prefer to start L-T4 therapy if TSH levels are above 10 mIU/L [5,6]. Since pediatric HT patients are at a higher risk for developing overt hypothyroidism than those with idiopathic hyperthyrotropinemia [5,7,8], HT patients with SCH require careful monitoring. One study reported that 15.5% of euthyroid patients and 73.8% of SCH patients with HT at diagnosis eventually required L-T4 therapy [9]. However, not all children who initiated L-T4 therapy required lifelong therapy. After discontinuation of therapy, approximately one-third of patients remained euthyroid and approximately 20%–30% of patients required no treatment [10-13].
To date, data on the clinical course after L-T4 discontinuation in the pediatric population are limited [10,11]. Therefore, we aimed to ascertain the outcomes after discontinuation of L-T4 in pediatric HT patients and to determine the predictive factors for both reinitiation of thyroid dysfunction and for successful withdrawal.
Materials and methods
1. Subjects
The medical records of 254 patients who had been prescribed L-T4 following a diagnosis of HT between January 1, 2000 and March 31, 2021 at Seoul National University Children's Hospital were reviewed retrospectively. Clinical diagnosis of HT was established by identifying the presence of antithyroid peroxidase antibodies (TPO Abs) and/or antithyroglobulin antibodies (TG Abs) and by demonstrating the abnormal appearance of the thyroid gland on sonogram [14]. After excluding 86 patients with underlying diseases such as chromosomal or genetic disorders, other autoimmune diseases, or malignancies, data from 168 patients were evaluated. Among these patients, L-T4 discontinuation was attempted in 47 (28.0%). These 47 subjects were included in this study (Fig. 1).
2. Clinical and biochemical data
Height to the first decimal place was measured with a standard stadiometer (Holtain Ltd., Crymych, UK), and body mass index (BMI) was calculated as weight divided by squared height (kg/m2). Subjects were classified as lean (<85th BMI percentile), overweight (85th–95th BMI percentile), or obese (95th ≥ BMI percentile) based on the 2017 Korean National Growth Charts [15]. History of thyroid diseases in first- and second-degree relatives was also determined. Pubertal status was assessed by Tanner staging for breasts (females) and genitals (males). Goiter was assessed according to the World Health Organization criteria: (1) grade 0=not palpable and not visible, (2) grade 1=palpable but not visible (1a=not visible with neck extension and 1b=visible with neck extension), and (3) grade 2=palpable and visible. Data on weight-based L-T4 doses were collected at the initiation and discontinuation of treatment. Thyroid sonogram results at diagnosis of HT were also collected.
Regular follow-up evaluations of serum thyroid function and thyroid autoantibody levels were performed every 3–6 months. The serum concentrations of free thyroxine (T4) and TSH were measured by immunoradiometric kits (RIAKEY; Shin Jin Medics, Seoul, Korea). The serum levels of T3, TPO Abs, and TG Abs were determined by radioimmunoassay kits (Brahms DYNOTest; Diagnostica GmbH, Berlin, Germany). The normal ranges of serum free T4, TSH, and T3 were defined as 0.70–1.8 ng/dL, 0.4–4.1 mIU/L, and 87–184 ng/dL, respectively. Titers for both TPO and TG Abs were positive when values were higher than 60 U/mL.
SCH was defined as serum free T4 levels within normal range and TSH level above normal range, and overt hypothyroidism was defined as TSH level above normal range and serum free T4 levels below normal range. SCH was further classified as mild SCH when the TSH level was less than 10 mIU/L or severe SCH when the TSH level was more than 10 mIU/L [16]. Subclinical hyperthyroidism was defined as serum free T4 levels within the normal range and TSH levels below the normal range, and overt hyperthyroidism was defined as TSH below the normal range and serum free T4 above the normal range [14].
We used "thyroid dysfunction" to describe both overt hypothyroidism and SCH. L-T4 therapy was started in pediatric patients with overt hypothyroidism or severe SCH. L-T4 therapy was started in pediatric patients with overt hypothyroidism or severe SCH. L-T4 therapy was also administered in pediatric patients who complained of symptoms (for example, fatigue) and in cases of SCH that requested treatment (with TSH levels between 4.1 and 10 mIU/L). L-T4 therapy was discontinued in patients who were stable on L-T4 therapy for at least several months while maintaining normal thyroid function without changing the L-T4 dosage (Table 1). L-T4 was also discontinued in patients who developed subclinical hyperthyroidism or overt hyperthyroidism and in patients who wanted to attempt discontinuation. After L-T4 withdrawal, L-T4 was restarted when patients developed overt hypothyroidism or severe SCH.
3. Statistical analysis
All continuous variables are described using median with interquartile range. To compare the 3 groups, the Kruskal-Wallis test was used for continuous variables and the chi-square test was performed for categorical variables. Event-free survival was constructed using the Kaplan-Meier method. A Cox proportional hazard model was used to assess predictors of event-free survival. Two events after L-T4 discontinuation were defined: (1) the development of any SCH and (2) the development of overt hypothyroidism or severe SCH requiring L-T4 medication. Several continuous variables were converted to categorical variables for analysis: these were age at L-T4 discontinuation ≤12 years (n=8), a TSH value at L-T4 initiation of ≥ 50 mIU/L (n=15), a TSH value at L-T4 discontinuation (≥50% of the upper limit of normal; 2.25mIU/L here, n= 13), and a treatment period of ≥36 months (n=29). A multivariate model was constructed with significant variables identified in the univariate model. Hazard ratios (HRs) are presented with 95% confidence intervals (CIs). P-values of <0.05 were deemed to indicate statistical significance. All continuous variable outliers were treated as missing values. All statistical analyses were performed using IBM SPSS Statistics ver. 25.0 (IBM Co., Armonk, NY, USA).
4. Ethics statement
The study protocol was approved by the Institutional Review Board of Seoul National University Hospital (2104-031-1209), and the requirement for informed consent was waived.
Results
1. Baseline characteristics
Table 1 shows the baseline and follow-up characteristics of the 47 patients, 3 boys and 44 girls. Eighteen patients (38.3%) had a family history of thyroid disease. At the time of L-T4 initiation, the median age was 10.4 years (range, 8.6–12.3 years) and 29 (61.7%) were postpubertal. Forty-two of 46 patients (91.3%) (one missing value) had a goiter. Twenty-one (44.7%) patients were diagnosed with SCH and the remaining 26 (55.3%) were diagnosed with overt hypothyroidism.
At the time of L-T4 discontinuation, 26 patients (55.3%) were euthyroid, 3 (6.4%) had SCH, and 18 (38.3%) had subclinical hyperthyroidism (n=15) or overt hyperthyroidism (n=3). Of these 18 patients, 16 were on a consistent L-T4 dose until subclinical hyperthyroidism or overt hyperthyroidism occurred; the other 2 had been on a reduced L-T4 dose for 2 months due to subclinical hyperthyroidism . None of the cases were attributed to overdose. The median age at L-T4 discontinuation was 15.4 years (range, 12.7–18.4 years), and 45 patients (95.7%) were postpubertal. Median duration of L-T4 therapy was 47 months (range, 20.3–80.3 months).
2. Outcomes after L-T4 discontinuation
During the follow-up period after L-T4 discontinuation (median, 30 months; range, 10.6–61.0 months), 14 patients (29.8%) maintained euthyroid status; however, 33 patients (70.2%) developed thyroid dysfunction. Five patients developed overt hypothyroidism and 28 patients developed SCH. Among those with SCH, 18 had mild SCH; the remaining 10 patients had severe SCH. As shown in Fig. 2, L-T4 therapy was restarted in 5 patients with overt hypothyroidism and 12 patients with SCH. There were 10 patients who were followed for more than 24 months without restarting L-T4 treatment. Three of these developed thyroid dysfunction after 2 years, and L-T4 treatment was re-initiated due to overt hypothyroidism.
Fig. 3 shows event-free survival curves. Among 33 patients who developed thyroid dysfunction after L-T4 discontinuation, the median period from discontinuation to thyroid dysfunction was 3.2 months (range, 1.2–8.2 months). When thyroid function was evaluated at 12 and 24 months after L-T4 discontinuation, 39.8% and 31.5% of patients remained euthyroid without medication, respectively.
3. Comparison of clinical and biochemical characteristics according to outcome after L-T4 discontinuation
Clinical and biochemical characteristics were compared according to the 3 outcome groups (euthyroidism, mild SCH, and overt hypothyroidism or severe SCH) after L-T4 discontinuation. At the time of L-T4 initiation and discontinuation, no significant differences in age, sex, proportion with family history of thyroid disease, being overweight or obese, being postpubertal, and goiter status were found among the 3 groups. While there were no significant differences in free T4, T3, and TSH levels at L-T4 initiation, free T4 levels at discontinuation of L-T4 was higher in the overt hypothyroidism or severe SCH group than the other 2 groups (P=0.020); there were no differences in T3 levels, however.
4. Predictors for developing thyroid dysfunction after L-T4 discontinuation
Next, predictors for thyroid dysfunction after L-T4 discontinuation were evaluated using the Cox proportional hazard model. (Table 2)The risk for developing thyroid dysfunction increased when TSH levels were more than 50 mIU/L at L-T4 initiation (P=0.046), when patient age was less than 12 years at discontinuation of L-T4 (P=0.0001), and when TSH levels were more than 2.25 mIU/L at L-T4 discontinuation (P=0.029) in a univariate analysis. Multivariate-adjusted HRs (95% CI) for developing thyroid dysfunction after L-T4 discontinuation were 3.5 (1.6–7.9) for TSH levels ≥ 50 mIU/L at L-T4 initiation (P=0.002 when compared to the TSH level < 50 mIU/L group), 11.1 (3.8–32.2) for age <12 years at discontinuation of L-T4 (P=0.0001 when compared to the ≥12 years of age group), and 2.7 (1.2–5.8) for those in the higher than 50th percentile for TSH level at L-T4 discontinuation group (P=0.014 when compared to the lower half TSH level group). Furthermore, age below 12 years at L-T4 discontinuation was the only independent predictor for developing overt hypothyroidism or severe SCH after L-T4 discontinuation (HR, 4.3; 95% CI, 1.4–13.2, P=0.012) in a univariate analysis; however, the difference was not significant when adjusted with other factors in multivariate analysis.
Discussion
After L-T4 discontinuation in pediatric HT patients, 70.2% developed thyroid dysfunction and 36.2% were restarted on L-T4 during the median of 30 months’ follow-up. Higher TSH levels at the time of L-T4 initiation or discontinuation and age under 12 years at L-T4 discontinuation increased the risk for developing thyroid dysfunction after L-T4 discontinuation, and discontinuation before 12 years of age independently increased the risk for overt hypothyroidism or severe SCH.
While 39.8% and 31.5% remained euthyroid 12 and 24 months after L-T4 discontinuation, 70.2% eventually developed thyroid dysfunction, 15% with overt hypothyroidism and 85% with SCH, during the follow-up period. This result was comparable to data from previous studies [11,12]. One meta-analysis that included both adult and children primary hypothyroidism patients suggested that up to one-third remained euthyroid after discontinuation of L-T4 [11]. In a retrospective study that included 382 adults with primary hypothyroidism, 22.5% successfully remained euthyroid without reinitiation of treatment for more than 1 year after L-T4 discontinuation. In that study, treatment was resumed for patients with overt hypothyroidism or severe SCH; reinitiation was also effected in some mild SCH cases in whom hypothyroid symptoms such as fatigue, facial edema, or constipation reappeared [12]. In a prospective pediatric study that included 148 children with HT, 30% of children required no treatment 24 months after L-T4 discontinuation. In that study, treatment was resumed only in patients with overt hypothyroidism or severe SCH [10]. Half of our patients developed thyroid dysfunction within 3.2 months after L-T4 withdrawal; these cases stressed a need for ongoing monitoring of patients within 3 months after L-T4 discontinuation. As there were some patients who restarted treatment at longer than 24 months after L-T4 withdrawal due to thyroid dysfunction, long-term thyroid function monitoring is required after LT4 discontinuation.
Higher TSH level (more than 50 mIU/L at L-T4 initiation, and the higher half at L-T4 discontinuation) was a significant predictor for developing thyroid dysfunction after L-T4 withdrawal. According to previous adult studies with primary hypothyroidism [12], TSH level at the time of L-T4 initiation was significantly lower in the group of patients who successfully discontinued L-T4 than in those with unchanged L-T4 dose. In that study, TSH level at the time of L-T4 tapering was significantly lower in the group of patients with reduced L-T4 dose than in those with unchanged L-T4 dose. However, in that study, neither TSH levels at L-T4 initiation nor TSH levels at L-T4 discontinuation were predictors for successful withdrawal of therapy. A prospective pediatric study showed that TSH level >10 IU/L at diagnosis was the only predictive factor for deterioration of thyroid function after withdrawal of L-T4 [10].
Our study showed that the risk for developing thyroid dysfunction increased when patients discontinued L-T4 before the age of 12 years. This was consistent with a previous pediatric study that showed that patients who successfully discontinued L-T4 were older than those who restarted [10]. Although pediatric patients who required resumption of L-T4 after discontinuation tended to be diagnosed at a younger age in that previous pediatric study [10], age at L-T4 initiation was not predictive for developing thyroid dysfunction after L-T4 withdrawal in our study. Shorter duration of L–T4 therapy was one of predictors for successful L–T4 discontinuation in a retrospective data study that included adult patients with a mean treatment duration of 7.3 years [12]. However, treatment duration was not significantly predictive for thyroid dysfunction in our study. The inconsistency of these results may have resulted due to patient age, sample size variation, and/or a different follow-up duration.
This study was limited by its retrospective design, small sample size, and being conducted at a single center. Although this retrospective study was also limited by inconsistent criteria for attempting to discontinue medication depending on the patient’s situation, the inclusion of homogeneous pediatric HT patients without other chronic diseases and the long duration of follow-up were strengths of this study.
In patients who had initiated L-T4 due to childhood HT, L-T4 discontinuation led to development of thyroid dysfunction (70.2%) during the median 30 months of follow-up after discontinuation, and 36.2% of patients required restart of L-T4. Attempting L-T4 discontinuation in patients diagnosed with HT during childhood may more successful in older age children and in children with lower TSH levels at initiation and discontinuation.
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Funding
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Data availability
The data that support the findings of this study can be provided by the corresponding author upon reasonable request.
Author contribution
Conceptualization: YAL, MJK, YJL, CHS; Data curation: MJK, YJL; Formal analysis: MJK, YC; Methodology: YAL, MJK, YJL, YC; Project administration: YAL, CHS; Visualization: MJK; Writing - original draft: MJK; Writing - review & editing: YAL, YJL, YC