- Email: [email protected]
Patterns of amiodarone-induced thyroid dysfunction in infants and children
Patterns of amiodarone-induced thyroid dysfunction in infants and children Ana Creo, MD,* Heather Anderson, MD,† Bryan Cannon, MD, FHRS,† Aida Lteif, MD,* Seema Kumar, MD,* Peter Tebben, MD,*‡ Anoop Mohamed Iqbal, MBBS,* Akhila Ramakrishna, MBBS,* Siobhan Pittock, MBBCh* From the *Division of Pediatric Endocrinology and Metabolism, Mayo Clinic, Rochester, Minnesota, † Division of Pediatric Cardiology, Mayo Clinic, Rochester, Minnesota, and ‡Division of Endocrinology, Diabetes and Metabolism, Mayo Clinic, Rochester, Minnesota. BACKGROUND Heart Rhythm Society guidelines recommend obtaining thyroid function tests (TFTs) at amiodarone initiation and every 6 months thereafter in adults, with no specific pediatric recommendations. Untreated hypothyroidism in young children negatively affects brain development and somatic growth, yet the optimal screening frequency for pediatric patients remains unclear, and limited data exist on pediatric amiodarone-induced thyroid dysfunction. OBJECTIVE The purpose of this study was to describe the patterns of amiodarone-induced thyroid dysfunction in pediatric patients. METHODS We established a retrospective cohort of 527 pediatric patients who received amiodarone between 1997 and 2017. We defined amiodarone therapy lasting 3–30 days as “short term” and .30 days as “long term.” RESULTS The final cohort (n 5 150) consisted of 27 neonates (18%), 25 infants (16%), 27 young children (18%), and 71 children (47%). Of the children in whom TFTs were checked, half (50.8%) developed a thyroid-stimulating hormone (TSH) value above the
Introduction Amiodarone is a potent drug used to manage both atrial and ventricular arrhythmias in children and adults. Amiodarone is a fat-soluble molecule with a half-life as long as 100 days, such that systemic effects may remain long after the drug is discontinued.1–4 Amiodarone also has a very high iodine content, which may interfere with normal thyroid hormone metabolism.5 Amiodarone typically disrupts thyroid function in 3 ways: (1) it inhibits monodeiodination of thyroxine (T4), decreasing triiodothyronine (T3) levels6; (2) it blocks the ability of T3 to bind to its target nuclear receptors7; and (3) it is known to have a direct thyrotoxic effect resulting in chronic thyroiditis.8–10 The Heart Rhythm Society The authors have no financial relationships or conflicts of interest relevant to this article to disclose. Address reprint requests and correspondence: Dr Siobhan Pittock, Division of Pediatric Endocrinology and Metabolism, Mayo Clinic, 200 First St SW, Rochester, MN 55905. E-mail address: [email protected].
1547-5271/$-see front matter © 2019 Heart Rhythm Society. All rights reserved.
reference for age. Neonates had the highest median peak TSH values in both short- and long-term groups: 23.5 mIU/L (interquartile range 11.4–63.1) and 28.8 mIU/L (interquartile range 11.4– 34.4), respectively. Although concurrent use of inotropic support was significantly associated with lower initial TSH values, no variable related to cardiac illness or type of heart disease was associated with peak TSH values. CONCLUSION Neonates and infants receiving amiodarone had more thyroid dysfunction with greater degrees of TSH elevation than older children. TSH elevations occurred early, even with short-term exposure. Given the concern for brain development and growth in hypothyroid children, our results suggest the need for more rigorous pediatric-specific thyroid monitoring guidelines. KEYWORDS Amiodarone; Congenital heart disease; Pediatrics; Thyroid dysfunction; Thyroid-stimulating hormone (Heart Rhythm 2019;16:1436–1442) © 2019 Heart Rhythm Society. All rights reserved.
published Adult Thyroid Screening Guidelines that recommend obtaining thyroid function tests (TFTs) before amiodarone initiation and every 6 months thereafter.11 No thyroid monitoring guidelines exist specifically for pediatric patients, and the optimal thyroid screening frequency in children remains unclear. Unlike adults, children may have irreversible consequences of untreated hypothyroidism.12–15 Children, and especially infants, may develop impaired brain development and stunted growth due to hypothyroidism, no matter the etiology. Multiple case reports have described pediatric amiodarone-induced hypothyroidism.16–21 Although older studies estimate the overall prevalence of hypothyroidism in children receiving amiodarone between 2% and 20%, TFTs were inconsistently checked and often were checked only after a year or more of therapy.22–24 To our knowledge, no contemporary cohort studies detail amiodarone-induced thyroid dysfunction. Moffett et al25 explored amiodarone monitoring practices and reported that https://doi.org/10.1016/j.hrthm.2019.03.015
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
only 12.2% of children receiving chronic amiodarone therapy ever had a complete thyroid function panel checked. The lack of consistent thyroid function testing in children receiving amiodarone limits our knowledge regarding the effects of amiodarone on the pediatric thyroid gland. Understanding the subtle patterns of pediatric amiodarone induced-thyroid dysfunction may guide clinical practice and development of more rigorous pediatric-specific monitoring guidelines. In this study we aimed to describe the patterns of amiodarone-induced thyroid dysfunction in our pediatric patients.
Methods Study design and setting We established a retrospective cohort by identifying pediatric patients (17 years old) who received amiodarone therapy between 1997 and 2017. Patients were seen at the Mayo Clinic Children’s Center (Rochester, Minnesota), an academic tertiary care medical center in outpatient and inpatient settings. Data were abstracted from the medical records. The Mayo Clinic Institutional Review Board approved this study.
Subjects To be included in the cohort, all subjects needed to receive 3 days of amiodarone therapy. Children were excluded if they had an abnormal newborn screen, previously abnormal thyroid-stimulating hormone (TSH) or free thyroxine (FT4) values, a history of prematurity (,36 weeks’ gestation), a history of positive thyroid peroxidase antibodies, or known maternal thyroid disease. In total, 150 children were included in the final cohort.
Definitions We defined amiodarone exposure as including either intravenous or oral therapy. Typical amiodarone doses used at our institution are 5 mg/kg as an initial dose, followed by an amiodarone infusion as a continuous drip (typically between 10 and 20 mg/kg/d) or oral load consisting of 5–10 mg/kg twice daily for 1 week. Patients are then typically converted over to an oral maintenance dose between 5 and 10 mg/kg once daily thereafter. We defined amiodarone therapy lasting 3–30 days as “short term” and .30 days as “long term.” The child’s age group was based on the age when the first dose of amiodarone was administered. By convention, we categorized “neonates” as infants ,30 days old, “infants” as 1–11 months old, “young children” as 12–35 months, and “children” as 3–17 years old. Potential confounders and outcomes were also defined and identified, including age at the time of amiodarone administration, sex of the patient, type of cardiac lesion, and age at the time thyroid tests were obtained. We defined “iodine contrast exposure” as any documented use of contrast containing iodine within 2 weeks before amiodarone initiation or any time while on amiodarone therapy. Inotropic support was defined as the need for any intravenous ionotropic support (.1 day) concurrently at the time of amiodarone
1437 initiation. Extracorporeal membrane oxygenation (ECMO) use was defined as any need for ECMO 2 weeks before amiodarone or any time while on amiodarone therapy. The first “on-therapy TFTs” were defined as the first laboratory values of TSH and/or FT4 after amiodarone administration was initiated. The “peak” TSH was the highest value encountered after amiodarone therapy in patients who had .1 TSH value. FT4 at the time of TSH peak was the value obtained on the same day as the TSH peak. The FT4 and TSH assay was performed using a Beckman Coultier Dx1600 (Maryfort, O’Callaghans Mills, Co. Clare, Ireland) and subsequently by the Roche Cobas immunoassay (Roche Diagnostics, Indianapolis, IN). Cardiac diagnoses were categorized into the major lesion groups classified by the American Heart Association.26
Statistical analysis Statistical analysis was performed using Microsoft Excel (Microsoft Corporation, Redmond, WA) and JMP (Cary, NC). Descriptive statistics included demographic and clinical characteristics. Test specific analysis was performed using Excel formulas for odds ratio, confidence interval, and significance from typical contingency tables. P values were based on 2-tailed testing or the Fisher exact test. P ,.05 was considered significant.
Results
Our final cohort (Table 1) consisted of 150 patients: 109 patients received short-term amiodarone therapy (3–30 days; median 6 days; interquartile range [IQR] 4–9 days) and 41 received long-term amiodarone therapy (.30 days; median 95 days; IQR 60–380 days). There was a wide distribution of patient ages, with 27 neonates (18% of the total cohort), 25 infants (17%), 27 young children (18%), and 71 older children (47%). A greater percentage of neonates (78%) and infants (68%) received short-term amiodarone therapy compared to the distribution of older children receiving short-term amiodarone therapy, but this was not statistically significant. The 4 most common cardiac lesions were Ebstein anomaly, cardiomyopathy, single ventricle defect, and transposition of the great vessels. There was a high prevalence of mortality, with nearly 10% of patients dying during the study period from causes not directly related to amiodarone therapy and primarily related to their structural heart disease. There was no statistically significant difference in mortality between long- and short-term amiodarone groups. Over half of all patients in the cohort (58%) had TSH values measured while on amiodarone therapy (Table 2). Most patients (78%) receiving long-term therapy had TSH and FT4 levels measured, in compliance with current adult guidelines, suggesting at least a TSH be obtained after 6 months of amiodarone therapy.11 In addition, 28% of patients receiving short-term therapy had a TSH level measured. Overall median time to initial TSH after amiodarone exposure was 7 days (IQR 2.5–17 days): 6 days in neonates (IQR 2–10 days), 9 days in infants (IQR 4–13 days), 6 days in young
1438 Table 1
Heart Rhythm, Vol 16, No 9, September 2019 Characteristics of children receiving amiodarone
Clinical feature Age at amiodarone initiation Neonates (0–30 d) Infants (1–11 mo) Young children (1–2 y) Older children (3–17 y) Female gender Cardiac diagnosis Ebstein anomaly Cardiomyopathy Single ventricle defect Transposition Complete atrioventricular canal defect Primary recurrent atrial tachycardia Aortic root abnormality Tetralogy of Fallot Truncus arteriosus Total anomalous pulmonary venous connection Septal defect Channelopathy Miscellaneous Death
Entire cohort (n5150)
Short-term therapy (3–30 days) (n 5 109)
Long-term therapy (.30 days) (n 5 41)
27 (18) 25 (17) 27 (18) 71 (47) 70 (47)
21 (78)* 17 (68)* 10 (37)* 61 (86)* 56 (51)
6 (22)* 8 (32)* 17 (63)* 10 (14)* 18 (44)
47 (31) 16 (11) 15 (10) 14 (9) 13 (9) 10 (7) 8 (5) 5 (3) 4 (3) 2 (1) 2 (1) 2 (1) 12 (8) 14 (9)
34 (31) 12 (11) 13 (12) 13 (12) 10 (9) 4 (4) 7 (6) 3 (3) 4 (4) 1 (1) 0 (0) 2 (2) 6 (6) 11 (10)
13 (32) 4 (10) 2 (5) 1 (2) 3 (7) 6 (15) 1 (2) 2 (5) 0 1 (2) 2 (5) 0 6 (15) 5 (12)
Values are given as n (%). *Percent of age group.
children (IQR 3–33 days), and 6 days in older children (IQR 2–12 days). The median time to developing an abnormal TSH value after amiodarone exposure was within the first 2 weeks across all ages. In addition, the upper end of the IQR for time (in days) to first abnormal TSH value on amiodarone was within the first 3 weeks in all ages except young children (neonate IQR 6–20 days, infant IQR 3–17 days, young children IQR 5–57 days, older child IQR 2–17 days). Many patients who had at least 1 TSH value checked had elevated TSH values (50.8% overall; 66.7% of neonates, Table 2
36.4% of infants, 38.9% of young children, 56.5% of older children) defined by values elevated above the normal upper limit for age in which the assay was performed. The median peak TSH level was elevated in all children receiving amiodarone for all groups (11.4 mIU/L; IQR 6.4–23.3 mIU/L) (Table 2 and Figure 1). Neonates had the highest initial median TSH values after amiodarone exposure among all age groups and the highest median peak TSH values in both short- and long-term amiodarone therapy groups: 23.5 mIU/L (IQR 11.4–63.1) and 28.8 mIU/L (IQR 11.4–34.4),
Thyroid function tests in children receiving amiodarone Patients with TFTs checked
All patients (n 5 150) 63 (58) Neonates (n 5 27) 12 (44) Infants (n 5 25) 11 (46) Young children (n 5 27) 18 (67) Older children (n 5 71) 23 (32) Short-term therapy (3–30 days) All ages (n 5 109) 31 (28) Neonates (n 5 21) 7 (33) Infants (n 5 17) 5 (29) Young children (n 5 10) 3 (30) Older children (n 5 61) 16 (26) Long-term therapy (.30 days) All ages (n 5 41) 32 (78) Neonates (n 5 6) 5 (83) Infants (n 5 8) 6 (75) Young children (n 5 17) 15 (88) Older children (n 5 10) 7 (70)
Patients developing an elevated TSH*
Time to first abnormal TSH (d)
Initial TSH (mIU/L)
Peak TSH (mIU/L)
50.8 66.7 36.4 38.9 56.5
7 (3–19) 12 (6–20) 4 (3–17) 7 (5–57) 10 (2–17)
4.8 (2.8–8.8) 10.3 (6.8–26.4) 4.3 (2.9–5.1) 3.0 (2.3–4.9) 5.9 (2.2–9.1)
11.4 (6.4–23.3) 28.8 (11.4–40) 14.3 (9.4–211.7) 6.0 (5.0–8.0) 9.6 8.3–18.3)
1.6 (1.2–1.8) 1.5 (1.1–1.6) 1.6 (1.2–1.9) 1.8 (1.6–2.4) 1.2(1.2–1.4)
54.8 57.1 40.0 33.3 62.5
5 (2–14) 6 (4–17) N/A N/A 5 (2–11)
5.9 (2.7–9.8) 9.2 (6.0–22.9) 4.2 (1.1–4.4) 2.8 (2.5–5.8) 6.15 (2.4–9.5)
15 (8.7–26.8) 23.5 (11.4–63.1) N/A N/A 10.3 (9.2–15.6)
1.35 (1.0–1.6) 1.25 (0.9–1.5) N/A N/A 1.2 (1.0–1.5)
46.9 80.0 33.3 40.0 42.9
15 (6–20) 17 (7–20) 4 (3–17) 7 (5–57) 19 (18–30)
4.5 (2.9–6.9) 11.4 (7.5–25.6) 4.6 (3.3–5.3) 3.1 (2.3–4.8) 4.4 (2.2–7.7)
8.9 (5.2–18.8) 28.8 (11.4–34.4) 14.3 (9.4–211.7) 6.0 (5.0–8.0) 8.9 (5.2–16.1)
1.6 (1.4–2.3) 1.85 (1.6–2.1) 1.6 (1.2–1.9) 1.8(1.6–2.4) 1.3 (1.3–1.4)
Values are given as n (%), %, or median (interquartile range). FT4 5 free thyroxine; TFT 5 thyroid function test; TSH 5 thyroid-stimulating hormone. *In patients who had thyroid testing performed.
FT4 at peak TSH (ng/dL)
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
Figure 1
1439
Thyroid function tests by age. FT4 5 free thyroxine; TSH 5 thyroid-stimulating hormone.
respectively. The peak median TSH value was significantly elevated for age across all age groups, with neonate median TSH 28.8 mIU/L (IQR 11.4–40 mIU/L; P 5 .008), infant median TSH 14.3 mIU/L (IQR 9.4–211.7 mIU/L; P 5 .02), young children median TSH 6.0 mIU/L (IQR 5.0–8.0 mIU/L; P 5 .04), and older children TSH 9.6 mIU/L (IQR 8.3–18.3 mIU/L; P 5 .008). Despite the elevated peak TSH, FT4 values (obtained in 70% who had TSH levels checked) remained in the normal range (median 1.6 ng/dL; IQR 1.2–1.8 ng/dL). Approximately 21.6% of patients had 1 TSH level obtained before amiodarone initiation. Fifty-eight percent of patients had a TSH level checked at least once after amiodarone initiation, and approximately 10.6% of all patients had .1 TSH level checked at our institution after amiodarone administration. A total of 10% of patients had a TSH level checked both before and after amiodarone administration, and among those 15 patients, 6 (40%) developed an elevated TSH for age after amiodarone exposure. Most patients did not have amiodarone discontinued while at our institution (as many transitioned to outpatient care closer to home on discharge) such that no patients had a TSH value performed at our institution after amiodarone discontinuation. Based on our current endocrine practice, we estimate that as many as 50.8% of all patients (66.7% of neonates and 36.4% of infants) who had a TSH checked may warrant further thyroid function testing and monitoring. No patient in our cohort developed suppressed TSH levels in conjunction with elevated FT4 levels. T3 levels were not routinely obtained.
Lastly, we explored variables that served both as potential markers of critical illness as well as possible confounders to amiodarone-induced thyroid dysfunction (Table 3). Iodine contrast exposure, history of heart surgery, and heart surgery within 14 days of amiodarone exposure were all significantly associated with increased duration of amiodarone use (P 5 .004, .004, and .001, respectively). Inotropic support at the time of amiodarone initiation was significantly associated with a lower initial TSH value (P 5 .008). We also explored the relationship with specific types of heart disease and did not find any single type of heart condition that was associated with any measure of amiodarone-induced thyroid dysfunction, although our power was very small and may have not been sufficient to detect meaningful associations.
Discussion Amiodarone is an antiarrhythmic agent that may disrupt thyroid function. In our cohort of pediatric patients with significant cardiac disease, morbidity, and mortality, amiodarone therapy was associated with elevated TSH values. In this critically ill population, other factors such as iodine exposure (independent of amiodarone), inotropic medication use, and ECMO may also impact thyroid function. Furthermore, not all patients had TFTs checked, given this was a retrospective review, so patients more likely to have abnormally high TSH values may have had TFTs checked more often. Despite these potential limitations and confounding factors, clear patterns of thyroid dysfunction emerged. First, neonates had the highest TSH values of any age group and were most likely to
1440 Table 3
Heart Rhythm, Vol 16, No 9, September 2019 Association of exposures related to critical illness and thyroid dysfunction
Iodine contrast exposure (n 5 27) Inotropic support at time of amiodarone initiation (n 5 42) ECMO use (n 5 25) Any history of heart surgery (n 5 126) Heart surgery within 14 days before amiodarone initiation (n 5 115)
Duration of amiodarone use
Initial TSH
Initial FT4
Peak TSH
FT4 at time of TSH peak
.004* .08 .4 .005* .001*
.58 .008† .46 .34 .75
.89 .49 .53 .49 .91
.84 .33 .14 .21 .74
.4 .85 .21 .69 .5
All data are given as P value. ECMO 5 extracorporeal membrane oxygenation; FT4 5 free thyroxine. *Exposure associated with significantly increased duration of amiodarone use. † Associated with significantly lower initial thyroid-stimulating hormone (TSH) values.
develop elevated TSH values. Peak median TSH values in neonates receiving both short- and long-term therapy were substantially elevated, approximately 2.5 times the upper limit of normal for age. Second, in patients who had .1 TFT performed, the peak median TSH value was above age-appropriate limits in all age groups (neonates, infants, young children, and older children). Finally, although current adult screening guidelines recommend the first TFT be performed after 6 months of amiodarone therapy, we found that most patients who developed an elevated TSH did so within the first 2 weeks of amiodarone exposure. Pediatric amiodarone-induced thyroid dysfunction has been previously described in various case reports and historical cohorts. More than 20 years ago, Figa et al1 reported one of the largest case series to date, describing 3 of 30 patients (10%) who developed amiodarone-induced hypothyroidism. Costigan et al16 estimated that 3 of 15 pediatric patients (20%) developed amiodarone-induced hypothyroidism. Although defining what truly constitutes the definition of hypothyroidism in this population of sick children remains challenging, our cohort suggests a higher rate of TSH elevations and possible hypothyroidism compared to these previous studies. At least 21% of all children in the cohort developed elevated TSH levels; however, this may underrepresent the true prevalence, as only 58% of patients had TFTs performed. Other older studies reported lower rates of hypothyroidism compared to our study; however, in many of these studies thyroid tests were checked only after 1 year of amiodarone therapy, so earlier hypothyroidism may have been missed.22–24 Finally, few cohorts included many young children and infants; and our higher rate of thyroid dysfunction also may be due to the inclusion of many more young children and infants. Several case reports suggest that infants receiving amiodarone may be particularly susceptible to developing elevated TSH values.18,27 A neonate with hypoplastic left heart developed TSH of 208 mIU/L 3 weeks into amiodarone therapy.18 In both reports, it seems that initial TSH testing was performed 3–4 weeks after initiation of amiodarone such that the time to abnormal TSH could have been much sooner, as seen in our patients. Even fetal amiodarone exposure may result in neonatal hypothyroidism.17,19–21,28,29 Challenges remain in fully defining amiodarone-induced hypothyroidism in neonates because
they generally have higher physiological TSH values; however, the TSH levels in our patients were elevated to a median value approximately 2.5 times the upper limit of normal, physiological age-based values. All patients with elevated TSH values, corrected for age, had a normal FT4 value. Amiodarone typically decreases T4 to T3 metabolism, so a normal FT4 is not necessarily reassuring and low T3 levels likely drive the TSH elevation. Regardless of T3 levels, by definition, these patients receiving amiodarone at the least have subclinical hypothyroidism with elevated TSH values and normal FT4 values. In older children, untreated subclinical hypothyroidism does not impair brain development or growth but may be an independent risk factor for developing cardiovascular disease and dyslipidemia.30,31 Typical recommendations, even in the presence of normal T4 levels, are to treat subclinical hypothyroidism in children .3 years when TSH .10 mIU/L and even .5 mIU/L in some cases.32,33 In our cohort, 26.1% of children age 3–17 years had TSH values .10 mIU/L, and 56.5% of children age 3–17 years had TSH values .5 mIU/L. In neonates and young children, however, even in the presence of a normal FT4, more subtle TSH abnormalities warrant further evaluation and close follow-up because of the potentially negative impact of hypothyroidism on brain development. The Heart Rhythm Society Adult Thyroid Screening Guidelines suggest TFTs be performed at amiodarone initiation and every 6 months thereafter. Although 6 months is the recommended interval, acute TSH changes have been described in adults with as few as 5 days of amiodarone exposure.3 Even with the potential for acute thyroid changes after amiodarone therapy, transient hypothyroidism in adults is often tolerable and not aggressively treated. Transient hypothyroidism is not tolerable in neonates, infants, and young children, who may accrue potentially irreversible changes in brain development.12,14 We recognize that amiodarone is often the optimal antiarrhythmic choice in acutely ill patients, particular postoperatively because of to its pharmacologic properties and ability to treat various atrial and ventricular arrhythmias. Alternative antiarrhythmic therapy should be considered when possible to avoid the side effects noted with amiodarone, including thyroid dysfunction. However, in the many cases in which amiodarone use is unavoidable, we propose the need for
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
Table 4 Suggested timetable for thyroid function screening in children receiving amiodarone When to obtain thyroid-stimulating hormone and free thyroxine levels Age ,3 y Baseline* Week 2 Week 4 Week 8 Every 2 months thereafter 3 months after amiodarone discontinuation Age 3 y Baseline* Week 4 Every 3 months for the first year on therapy Every 6 months thereafter 3 months after amiodarone discontinuation *Obtained before any amiodarone exposure when possible.
more rigorous pediatric-specific thyroid monitoring guidelines in children receiving amiodarone and screening for thyroid dysfunction early in the course of amiodarone treatment given the concern for poor brain development due to hypothyroidism and the short lag time between amiodarone initiation and development of elevated TSH values in our patients. Given the risk of cerebral effects of hypothyroidism and the frequency of significant TSH elevations in our study, we propose more frequent thyroid monitoring in children.12–15 Other investigators also endorse more frequent thyroid monitoring in children than in adults.16–18 Costigan et al16 suggested measuring T4, T3, and TSH “often” during the first year of treatment, and even more frequently in children ,1 year. Trudel et al18 proposed performing TSH monitoring in children ,4 years of age at baseline, 1 week, every 2 weeks of the next month, every 4 weeks for the next 3 months, and every 3 months thereafter. We have modified Trudel’s protocol based on our data and propose a similar timetable for thyroid monitoring in children (Table 4).18 Our data demonstrate that of the children who developed an elevated TSH level, most did so within 2–4 weeks (IQR 3–19 days for all children); however, there was a wide distribution, with some children developing elevated TSH levels months after initial exposure.
Study limitations Our results should be interpreted in the context of several limitations. First, the children in our cohort were sick and likely had been exposed to other confounding agents. We performed analysis of variables, which could potentially be associated with greater degrees of amiodarone dysfunction and did not find any significant associations. However, we still had a relatively small statistical sample size. Also, at our institution many children undergo heart surgery and then follow-up closer to home on discharge, so our sample mostly captured hospitalized patients and may not be generalizable to less ill children receiving amiodarone in the outpatient setting. This study was also limited due to the retrospective
1441 nature of the review. Although the majority of patients receiving amiodarone had TFTs performed, some did not. Also, most patients did not have TFTs performed before amiodarone initiation, so the true percentage of children who would have had abnormal thyroid function even before amiodarone initiation remains unknown. However, we excluded patients with known risk factors for abnormal TFTs and a low likelihood of having an elevated TSH value at baseline. Furthermore, sick children tend to have lower, not higher, TSH values due to sick euthyroid syndrome, such that our cohort would be at higher risk for developing a low TSH value. It is possible that TFTs were checked or were checked more frequently in patients displaying possible symptoms of hypothyroidism, potentially artificially increasing the prevalence and degree of TSH elevation among patients. Lastly, although TFTs were obtained, they were not drawn frequently enough to confidently capture the true timing of the peak TSH value or time to initial abnormal TSH value after amiodarone exposure.
Conclusion Amiodarone has the potential to disrupt normal thyroid function in pediatric patients. In this cohort of patients receiving amiodarone who had TFTs checked, the majority developed elevated TSH values, in contrast to older reports. Neonates and infants were most prone to developing elevated TSH values. In patients who had .1 TFT performed, peak median TSH values were above age-appropriate limits in the neonate, infant, young children, and older child groups. Most patients who developed elevated TSH values did so within the first 2 weeks of amiodarone exposure. Given the concern for poor brain development and growth in hypothyroid children, our results suggest the need for more rigorous pediatric-specific thyroid monitoring guidelines and screening early during therapy to monitor for thyroid dysfunction.
References 1. Figa FH, Gow RM, Hamilton RM, Freedom RM. Clinical efficacy and safety of intravenous Amiodarone in infants and children. Am J Cardiol 1994;74:573–577. 2. Ha HR, Stieger B, Grassi G, Altorfer HR, Follath F. Structure-effect relationships of amiodarone analogues on the inhibition of thyroxine deiodination. Eur J Clin Pharmacol 2000;55:807–814. 3. Iervasi G, Clerico A, Bonini R, et al. Acute effects of amiodarone administration on thyroid function in patients with cardiac arrhythmia. J Clin Endocrinol Metab 1997;82:275–280. 4. Latini R, Tognoni G, Kates RE. Clinical pharmacokinetics of amiodarone. Clin Pharmacokinet 1984;9:136–156. 5. Institute of Medicine (US) Panel on Micronutrients. Dietary Reference Intakes for Vitamin D, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Washington, DC: National Academy Press; 2001. 6. Rao RH, McCready VR, Spathis GS. Iodine kinetic studies during amiodarone treatment. J Clin Endocrinol Metab 1986;62:563–568. 7. Franklyn JA, Davis JR, Gammage MD, Littler WA, Ramsden DB, Sheppard MC. Amiodarone and thyroid hormone action. Clin Endocrinol 1985;22:257–264. 8. Roti E, Minelli R, Gardini E, Bianconi L, Braverman LE. Thyrotoxicosis followed by hypothyroidism in patients treated with amiodarone. A possible consequence of a destructive process in the thyroid. Arch Intern Med 1993; 153:886–892. 9. Czarnywojtek A, Plazinska MT, Zgorzalewicz-Stachowiak M, et al. Dysfunction of the thyroid gland during amiodarone therapy: a study of 297 cases. Ther Clin Risk Manag 2016;12:505–513.
1442 10.
11. 12. 13. 14.
15. 16. 17. 18. 19.
20.
21.
22.
Trip MD, Wiersinga W, Plomp TA. Incidence, predictability, and pathogenesis of amiodarone-induced thyrotoxicosis and hypothyroidism. Am J Med 1991; 91:507–511. Goldschlager N, Epstein AE, Naccarelli GV, et al. A practical guide for clinicians who treat patients with amiodarone: 2007. Heart Rhythm 2007;4:1250–1259. Koibuchi N, Chin WW. Thyroid hormone action and brain development. Trends Endocrinol Metab 2000;11:123–128. Rovet JF. The role of thyroid hormones for brain development and cognitive function. Endocr Dev 2014;26:26–43. Zoeller RT, Rovet J. Timing of thyroid hormone action in the developing brain: clinical observations and experimental findings. J Neuroendocrinol 2004; 16:809–818. Williams GR. Neurodevelopmental and neurophysiological actions of thyroid hormone. J Neuroendocrinol 2008;20:784–794. Costigan DC, Holland FJ, Daneman D, Hesslein PS, Vogel M, Ellis G. Amiodarone therapy effects on childhood thyroid function. Pediatrics 1986;77:703–708. Niinikoski H, Hamalainen AM, Ekblad H, Soukka H. Neonatal hypothyroidism after amiodarone therapy. Acta Paediatr 2007;96:773–774. Trudel K, Sanatani S, Panagiotopoulos C. Severe amiodarone-induced hypothyroidism in an infant. Pediatr Crit Care Med 2011;12:e43–e45. Lomenick JP, Jackson WA, Backeljauw PF. Amiodarone-induced neonatal hypothyroidism: a unique form of transient early-onset hypothyroidism. J Perinatol 2004;24:397–399. Pavan-Senn CC, Nesi-Franca S, Pelaez J, et al. [Transient neonatal hypothyroidism due to amiodarone administration during pregnancy—two cases report and review of literature]. Arq Bras Endocrinol Metabl 2008;52:126–130. Vanbesien J, Casteels A, Bougatef A, et al. Transient fetal hypothyroidism due to direct fetal administration of amiodarone for drug resistant fetal tachycardia. Am J Perinatol 2001;18:113–116. Coumel P, Fidelle J. Amiodarone in the treatment of cardiac arrhythmias in children: one hundred thirty-five cases. Am Heart J 1980;100:1063–1069.
Heart Rhythm, Vol 16, No 9, September 2019 23.
24.
25. 26.
27. 28.
29.
30.
31. 32.
33.
Guccione P, Paul T, Garson A Jr. Long-term follow-up of amiodarone therapy in the young: continued efficacy, unimpaired growth, moderate side effects. J Am Coll Cardiol 1990;15:1118–1124. Vignati G, Danzi GB, Austoni P, Fancini P, Mascarello M, Figini A. [Amiodarone therapy in childhood: efficacy and side effects]. G Ital Cardiol 1985; 15:786–794. Moffett BS, Valdes SO, Kim JJ. Amiodarone monitoring practices in pediatric hospitals in the United States. Pediatr Cardiol 2013;34:1762–1766. American Heart Association. Congenital heart defects. 2018. Available at: http:// www.heart.org/HEARTORG/Conditions/CongenitalHeartDefects/CongenitalHeart-Defects_UCM_001090_SubHomePage.jsp. Gandy J, Wonko N, Kantoch MJ. Risks of intravenous amiodarone in neonates. Can J Cardiol 1998;14:855–858. Bartalena L, Bogazzi F, Braverman LE, Martino E. Effects of amiodarone administration during pregnancy on neonatal thyroid function and subsequent neurodevelopment. J Endocrinol Invest 2001;24:116–130. Matsumura LK, Born D, Kunii IS, Franco DB, Maciel RM. Outcome of thyroid function in newborns from mothers treated with amiodarone. Thyroid 1992; 2:279–281. Dahl AR, Iqbal AM, Lteif AN, Pittock ST, Tebben PJ, Kumar S. Mild subclinical hypothyroidism is associated with pediatric dyslipidaemia. Clin Endocrinol (Oxf) 2018;89:330–335. Marwaha RK, Tandon N, Garg MK, et al. Dyslipidemia in subclinical hypothyroidism in an Indian population. Clin Biochem 2011;44:1214–1217. Salerno M, Capalbo D, Cerbone M, De Luca F. Subclinical hypothyroidism in childhood - current knowledge and open issues. Nat Rev Endocrinol 2016; 12:734–746. Jonklaas J, Bianco AC, Bauer AJ, et al. American Thyroid Association Task Force on Thyroid Hormone Replacement. Guidelines for the treatment of hypothyroidism: prepared by the American Thyroid Association Task Force on Thyroid Hormone Replacement. Thyroid 2014;24:1670–1751.
Introduction Amiodarone is a potent drug used to manage both atrial and ventricular arrhythmias in children and adults. Amiodarone is a fat-soluble molecule with a half-life as long as 100 days, such that systemic effects may remain long after the drug is discontinued.1–4 Amiodarone also has a very high iodine content, which may interfere with normal thyroid hormone metabolism.5 Amiodarone typically disrupts thyroid function in 3 ways: (1) it inhibits monodeiodination of thyroxine (T4), decreasing triiodothyronine (T3) levels6; (2) it blocks the ability of T3 to bind to its target nuclear receptors7; and (3) it is known to have a direct thyrotoxic effect resulting in chronic thyroiditis.8–10 The Heart Rhythm Society The authors have no financial relationships or conflicts of interest relevant to this article to disclose. Address reprint requests and correspondence: Dr Siobhan Pittock, Division of Pediatric Endocrinology and Metabolism, Mayo Clinic, 200 First St SW, Rochester, MN 55905. E-mail address: [email protected].
1547-5271/$-see front matter © 2019 Heart Rhythm Society. All rights reserved.
reference for age. Neonates had the highest median peak TSH values in both short- and long-term groups: 23.5 mIU/L (interquartile range 11.4–63.1) and 28.8 mIU/L (interquartile range 11.4– 34.4), respectively. Although concurrent use of inotropic support was significantly associated with lower initial TSH values, no variable related to cardiac illness or type of heart disease was associated with peak TSH values. CONCLUSION Neonates and infants receiving amiodarone had more thyroid dysfunction with greater degrees of TSH elevation than older children. TSH elevations occurred early, even with short-term exposure. Given the concern for brain development and growth in hypothyroid children, our results suggest the need for more rigorous pediatric-specific thyroid monitoring guidelines. KEYWORDS Amiodarone; Congenital heart disease; Pediatrics; Thyroid dysfunction; Thyroid-stimulating hormone (Heart Rhythm 2019;16:1436–1442) © 2019 Heart Rhythm Society. All rights reserved.
published Adult Thyroid Screening Guidelines that recommend obtaining thyroid function tests (TFTs) before amiodarone initiation and every 6 months thereafter.11 No thyroid monitoring guidelines exist specifically for pediatric patients, and the optimal thyroid screening frequency in children remains unclear. Unlike adults, children may have irreversible consequences of untreated hypothyroidism.12–15 Children, and especially infants, may develop impaired brain development and stunted growth due to hypothyroidism, no matter the etiology. Multiple case reports have described pediatric amiodarone-induced hypothyroidism.16–21 Although older studies estimate the overall prevalence of hypothyroidism in children receiving amiodarone between 2% and 20%, TFTs were inconsistently checked and often were checked only after a year or more of therapy.22–24 To our knowledge, no contemporary cohort studies detail amiodarone-induced thyroid dysfunction. Moffett et al25 explored amiodarone monitoring practices and reported that https://doi.org/10.1016/j.hrthm.2019.03.015
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
only 12.2% of children receiving chronic amiodarone therapy ever had a complete thyroid function panel checked. The lack of consistent thyroid function testing in children receiving amiodarone limits our knowledge regarding the effects of amiodarone on the pediatric thyroid gland. Understanding the subtle patterns of pediatric amiodarone induced-thyroid dysfunction may guide clinical practice and development of more rigorous pediatric-specific monitoring guidelines. In this study we aimed to describe the patterns of amiodarone-induced thyroid dysfunction in our pediatric patients.
Methods Study design and setting We established a retrospective cohort by identifying pediatric patients (17 years old) who received amiodarone therapy between 1997 and 2017. Patients were seen at the Mayo Clinic Children’s Center (Rochester, Minnesota), an academic tertiary care medical center in outpatient and inpatient settings. Data were abstracted from the medical records. The Mayo Clinic Institutional Review Board approved this study.
Subjects To be included in the cohort, all subjects needed to receive 3 days of amiodarone therapy. Children were excluded if they had an abnormal newborn screen, previously abnormal thyroid-stimulating hormone (TSH) or free thyroxine (FT4) values, a history of prematurity (,36 weeks’ gestation), a history of positive thyroid peroxidase antibodies, or known maternal thyroid disease. In total, 150 children were included in the final cohort.
Definitions We defined amiodarone exposure as including either intravenous or oral therapy. Typical amiodarone doses used at our institution are 5 mg/kg as an initial dose, followed by an amiodarone infusion as a continuous drip (typically between 10 and 20 mg/kg/d) or oral load consisting of 5–10 mg/kg twice daily for 1 week. Patients are then typically converted over to an oral maintenance dose between 5 and 10 mg/kg once daily thereafter. We defined amiodarone therapy lasting 3–30 days as “short term” and .30 days as “long term.” The child’s age group was based on the age when the first dose of amiodarone was administered. By convention, we categorized “neonates” as infants ,30 days old, “infants” as 1–11 months old, “young children” as 12–35 months, and “children” as 3–17 years old. Potential confounders and outcomes were also defined and identified, including age at the time of amiodarone administration, sex of the patient, type of cardiac lesion, and age at the time thyroid tests were obtained. We defined “iodine contrast exposure” as any documented use of contrast containing iodine within 2 weeks before amiodarone initiation or any time while on amiodarone therapy. Inotropic support was defined as the need for any intravenous ionotropic support (.1 day) concurrently at the time of amiodarone
1437 initiation. Extracorporeal membrane oxygenation (ECMO) use was defined as any need for ECMO 2 weeks before amiodarone or any time while on amiodarone therapy. The first “on-therapy TFTs” were defined as the first laboratory values of TSH and/or FT4 after amiodarone administration was initiated. The “peak” TSH was the highest value encountered after amiodarone therapy in patients who had .1 TSH value. FT4 at the time of TSH peak was the value obtained on the same day as the TSH peak. The FT4 and TSH assay was performed using a Beckman Coultier Dx1600 (Maryfort, O’Callaghans Mills, Co. Clare, Ireland) and subsequently by the Roche Cobas immunoassay (Roche Diagnostics, Indianapolis, IN). Cardiac diagnoses were categorized into the major lesion groups classified by the American Heart Association.26
Statistical analysis Statistical analysis was performed using Microsoft Excel (Microsoft Corporation, Redmond, WA) and JMP (Cary, NC). Descriptive statistics included demographic and clinical characteristics. Test specific analysis was performed using Excel formulas for odds ratio, confidence interval, and significance from typical contingency tables. P values were based on 2-tailed testing or the Fisher exact test. P ,.05 was considered significant.
Results
Our final cohort (Table 1) consisted of 150 patients: 109 patients received short-term amiodarone therapy (3–30 days; median 6 days; interquartile range [IQR] 4–9 days) and 41 received long-term amiodarone therapy (.30 days; median 95 days; IQR 60–380 days). There was a wide distribution of patient ages, with 27 neonates (18% of the total cohort), 25 infants (17%), 27 young children (18%), and 71 older children (47%). A greater percentage of neonates (78%) and infants (68%) received short-term amiodarone therapy compared to the distribution of older children receiving short-term amiodarone therapy, but this was not statistically significant. The 4 most common cardiac lesions were Ebstein anomaly, cardiomyopathy, single ventricle defect, and transposition of the great vessels. There was a high prevalence of mortality, with nearly 10% of patients dying during the study period from causes not directly related to amiodarone therapy and primarily related to their structural heart disease. There was no statistically significant difference in mortality between long- and short-term amiodarone groups. Over half of all patients in the cohort (58%) had TSH values measured while on amiodarone therapy (Table 2). Most patients (78%) receiving long-term therapy had TSH and FT4 levels measured, in compliance with current adult guidelines, suggesting at least a TSH be obtained after 6 months of amiodarone therapy.11 In addition, 28% of patients receiving short-term therapy had a TSH level measured. Overall median time to initial TSH after amiodarone exposure was 7 days (IQR 2.5–17 days): 6 days in neonates (IQR 2–10 days), 9 days in infants (IQR 4–13 days), 6 days in young
1438 Table 1
Heart Rhythm, Vol 16, No 9, September 2019 Characteristics of children receiving amiodarone
Clinical feature Age at amiodarone initiation Neonates (0–30 d) Infants (1–11 mo) Young children (1–2 y) Older children (3–17 y) Female gender Cardiac diagnosis Ebstein anomaly Cardiomyopathy Single ventricle defect Transposition Complete atrioventricular canal defect Primary recurrent atrial tachycardia Aortic root abnormality Tetralogy of Fallot Truncus arteriosus Total anomalous pulmonary venous connection Septal defect Channelopathy Miscellaneous Death
Entire cohort (n5150)
Short-term therapy (3–30 days) (n 5 109)
Long-term therapy (.30 days) (n 5 41)
27 (18) 25 (17) 27 (18) 71 (47) 70 (47)
21 (78)* 17 (68)* 10 (37)* 61 (86)* 56 (51)
6 (22)* 8 (32)* 17 (63)* 10 (14)* 18 (44)
47 (31) 16 (11) 15 (10) 14 (9) 13 (9) 10 (7) 8 (5) 5 (3) 4 (3) 2 (1) 2 (1) 2 (1) 12 (8) 14 (9)
34 (31) 12 (11) 13 (12) 13 (12) 10 (9) 4 (4) 7 (6) 3 (3) 4 (4) 1 (1) 0 (0) 2 (2) 6 (6) 11 (10)
13 (32) 4 (10) 2 (5) 1 (2) 3 (7) 6 (15) 1 (2) 2 (5) 0 1 (2) 2 (5) 0 6 (15) 5 (12)
Values are given as n (%). *Percent of age group.
children (IQR 3–33 days), and 6 days in older children (IQR 2–12 days). The median time to developing an abnormal TSH value after amiodarone exposure was within the first 2 weeks across all ages. In addition, the upper end of the IQR for time (in days) to first abnormal TSH value on amiodarone was within the first 3 weeks in all ages except young children (neonate IQR 6–20 days, infant IQR 3–17 days, young children IQR 5–57 days, older child IQR 2–17 days). Many patients who had at least 1 TSH value checked had elevated TSH values (50.8% overall; 66.7% of neonates, Table 2
36.4% of infants, 38.9% of young children, 56.5% of older children) defined by values elevated above the normal upper limit for age in which the assay was performed. The median peak TSH level was elevated in all children receiving amiodarone for all groups (11.4 mIU/L; IQR 6.4–23.3 mIU/L) (Table 2 and Figure 1). Neonates had the highest initial median TSH values after amiodarone exposure among all age groups and the highest median peak TSH values in both short- and long-term amiodarone therapy groups: 23.5 mIU/L (IQR 11.4–63.1) and 28.8 mIU/L (IQR 11.4–34.4),
Thyroid function tests in children receiving amiodarone Patients with TFTs checked
All patients (n 5 150) 63 (58) Neonates (n 5 27) 12 (44) Infants (n 5 25) 11 (46) Young children (n 5 27) 18 (67) Older children (n 5 71) 23 (32) Short-term therapy (3–30 days) All ages (n 5 109) 31 (28) Neonates (n 5 21) 7 (33) Infants (n 5 17) 5 (29) Young children (n 5 10) 3 (30) Older children (n 5 61) 16 (26) Long-term therapy (.30 days) All ages (n 5 41) 32 (78) Neonates (n 5 6) 5 (83) Infants (n 5 8) 6 (75) Young children (n 5 17) 15 (88) Older children (n 5 10) 7 (70)
Patients developing an elevated TSH*
Time to first abnormal TSH (d)
Initial TSH (mIU/L)
Peak TSH (mIU/L)
50.8 66.7 36.4 38.9 56.5
7 (3–19) 12 (6–20) 4 (3–17) 7 (5–57) 10 (2–17)
4.8 (2.8–8.8) 10.3 (6.8–26.4) 4.3 (2.9–5.1) 3.0 (2.3–4.9) 5.9 (2.2–9.1)
11.4 (6.4–23.3) 28.8 (11.4–40) 14.3 (9.4–211.7) 6.0 (5.0–8.0) 9.6 8.3–18.3)
1.6 (1.2–1.8) 1.5 (1.1–1.6) 1.6 (1.2–1.9) 1.8 (1.6–2.4) 1.2(1.2–1.4)
54.8 57.1 40.0 33.3 62.5
5 (2–14) 6 (4–17) N/A N/A 5 (2–11)
5.9 (2.7–9.8) 9.2 (6.0–22.9) 4.2 (1.1–4.4) 2.8 (2.5–5.8) 6.15 (2.4–9.5)
15 (8.7–26.8) 23.5 (11.4–63.1) N/A N/A 10.3 (9.2–15.6)
1.35 (1.0–1.6) 1.25 (0.9–1.5) N/A N/A 1.2 (1.0–1.5)
46.9 80.0 33.3 40.0 42.9
15 (6–20) 17 (7–20) 4 (3–17) 7 (5–57) 19 (18–30)
4.5 (2.9–6.9) 11.4 (7.5–25.6) 4.6 (3.3–5.3) 3.1 (2.3–4.8) 4.4 (2.2–7.7)
8.9 (5.2–18.8) 28.8 (11.4–34.4) 14.3 (9.4–211.7) 6.0 (5.0–8.0) 8.9 (5.2–16.1)
1.6 (1.4–2.3) 1.85 (1.6–2.1) 1.6 (1.2–1.9) 1.8(1.6–2.4) 1.3 (1.3–1.4)
Values are given as n (%), %, or median (interquartile range). FT4 5 free thyroxine; TFT 5 thyroid function test; TSH 5 thyroid-stimulating hormone. *In patients who had thyroid testing performed.
FT4 at peak TSH (ng/dL)
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
Figure 1
1439
Thyroid function tests by age. FT4 5 free thyroxine; TSH 5 thyroid-stimulating hormone.
respectively. The peak median TSH value was significantly elevated for age across all age groups, with neonate median TSH 28.8 mIU/L (IQR 11.4–40 mIU/L; P 5 .008), infant median TSH 14.3 mIU/L (IQR 9.4–211.7 mIU/L; P 5 .02), young children median TSH 6.0 mIU/L (IQR 5.0–8.0 mIU/L; P 5 .04), and older children TSH 9.6 mIU/L (IQR 8.3–18.3 mIU/L; P 5 .008). Despite the elevated peak TSH, FT4 values (obtained in 70% who had TSH levels checked) remained in the normal range (median 1.6 ng/dL; IQR 1.2–1.8 ng/dL). Approximately 21.6% of patients had 1 TSH level obtained before amiodarone initiation. Fifty-eight percent of patients had a TSH level checked at least once after amiodarone initiation, and approximately 10.6% of all patients had .1 TSH level checked at our institution after amiodarone administration. A total of 10% of patients had a TSH level checked both before and after amiodarone administration, and among those 15 patients, 6 (40%) developed an elevated TSH for age after amiodarone exposure. Most patients did not have amiodarone discontinued while at our institution (as many transitioned to outpatient care closer to home on discharge) such that no patients had a TSH value performed at our institution after amiodarone discontinuation. Based on our current endocrine practice, we estimate that as many as 50.8% of all patients (66.7% of neonates and 36.4% of infants) who had a TSH checked may warrant further thyroid function testing and monitoring. No patient in our cohort developed suppressed TSH levels in conjunction with elevated FT4 levels. T3 levels were not routinely obtained.
Lastly, we explored variables that served both as potential markers of critical illness as well as possible confounders to amiodarone-induced thyroid dysfunction (Table 3). Iodine contrast exposure, history of heart surgery, and heart surgery within 14 days of amiodarone exposure were all significantly associated with increased duration of amiodarone use (P 5 .004, .004, and .001, respectively). Inotropic support at the time of amiodarone initiation was significantly associated with a lower initial TSH value (P 5 .008). We also explored the relationship with specific types of heart disease and did not find any single type of heart condition that was associated with any measure of amiodarone-induced thyroid dysfunction, although our power was very small and may have not been sufficient to detect meaningful associations.
Discussion Amiodarone is an antiarrhythmic agent that may disrupt thyroid function. In our cohort of pediatric patients with significant cardiac disease, morbidity, and mortality, amiodarone therapy was associated with elevated TSH values. In this critically ill population, other factors such as iodine exposure (independent of amiodarone), inotropic medication use, and ECMO may also impact thyroid function. Furthermore, not all patients had TFTs checked, given this was a retrospective review, so patients more likely to have abnormally high TSH values may have had TFTs checked more often. Despite these potential limitations and confounding factors, clear patterns of thyroid dysfunction emerged. First, neonates had the highest TSH values of any age group and were most likely to
1440 Table 3
Heart Rhythm, Vol 16, No 9, September 2019 Association of exposures related to critical illness and thyroid dysfunction
Iodine contrast exposure (n 5 27) Inotropic support at time of amiodarone initiation (n 5 42) ECMO use (n 5 25) Any history of heart surgery (n 5 126) Heart surgery within 14 days before amiodarone initiation (n 5 115)
Duration of amiodarone use
Initial TSH
Initial FT4
Peak TSH
FT4 at time of TSH peak
.004* .08 .4 .005* .001*
.58 .008† .46 .34 .75
.89 .49 .53 .49 .91
.84 .33 .14 .21 .74
.4 .85 .21 .69 .5
All data are given as P value. ECMO 5 extracorporeal membrane oxygenation; FT4 5 free thyroxine. *Exposure associated with significantly increased duration of amiodarone use. † Associated with significantly lower initial thyroid-stimulating hormone (TSH) values.
develop elevated TSH values. Peak median TSH values in neonates receiving both short- and long-term therapy were substantially elevated, approximately 2.5 times the upper limit of normal for age. Second, in patients who had .1 TFT performed, the peak median TSH value was above age-appropriate limits in all age groups (neonates, infants, young children, and older children). Finally, although current adult screening guidelines recommend the first TFT be performed after 6 months of amiodarone therapy, we found that most patients who developed an elevated TSH did so within the first 2 weeks of amiodarone exposure. Pediatric amiodarone-induced thyroid dysfunction has been previously described in various case reports and historical cohorts. More than 20 years ago, Figa et al1 reported one of the largest case series to date, describing 3 of 30 patients (10%) who developed amiodarone-induced hypothyroidism. Costigan et al16 estimated that 3 of 15 pediatric patients (20%) developed amiodarone-induced hypothyroidism. Although defining what truly constitutes the definition of hypothyroidism in this population of sick children remains challenging, our cohort suggests a higher rate of TSH elevations and possible hypothyroidism compared to these previous studies. At least 21% of all children in the cohort developed elevated TSH levels; however, this may underrepresent the true prevalence, as only 58% of patients had TFTs performed. Other older studies reported lower rates of hypothyroidism compared to our study; however, in many of these studies thyroid tests were checked only after 1 year of amiodarone therapy, so earlier hypothyroidism may have been missed.22–24 Finally, few cohorts included many young children and infants; and our higher rate of thyroid dysfunction also may be due to the inclusion of many more young children and infants. Several case reports suggest that infants receiving amiodarone may be particularly susceptible to developing elevated TSH values.18,27 A neonate with hypoplastic left heart developed TSH of 208 mIU/L 3 weeks into amiodarone therapy.18 In both reports, it seems that initial TSH testing was performed 3–4 weeks after initiation of amiodarone such that the time to abnormal TSH could have been much sooner, as seen in our patients. Even fetal amiodarone exposure may result in neonatal hypothyroidism.17,19–21,28,29 Challenges remain in fully defining amiodarone-induced hypothyroidism in neonates because
they generally have higher physiological TSH values; however, the TSH levels in our patients were elevated to a median value approximately 2.5 times the upper limit of normal, physiological age-based values. All patients with elevated TSH values, corrected for age, had a normal FT4 value. Amiodarone typically decreases T4 to T3 metabolism, so a normal FT4 is not necessarily reassuring and low T3 levels likely drive the TSH elevation. Regardless of T3 levels, by definition, these patients receiving amiodarone at the least have subclinical hypothyroidism with elevated TSH values and normal FT4 values. In older children, untreated subclinical hypothyroidism does not impair brain development or growth but may be an independent risk factor for developing cardiovascular disease and dyslipidemia.30,31 Typical recommendations, even in the presence of normal T4 levels, are to treat subclinical hypothyroidism in children .3 years when TSH .10 mIU/L and even .5 mIU/L in some cases.32,33 In our cohort, 26.1% of children age 3–17 years had TSH values .10 mIU/L, and 56.5% of children age 3–17 years had TSH values .5 mIU/L. In neonates and young children, however, even in the presence of a normal FT4, more subtle TSH abnormalities warrant further evaluation and close follow-up because of the potentially negative impact of hypothyroidism on brain development. The Heart Rhythm Society Adult Thyroid Screening Guidelines suggest TFTs be performed at amiodarone initiation and every 6 months thereafter. Although 6 months is the recommended interval, acute TSH changes have been described in adults with as few as 5 days of amiodarone exposure.3 Even with the potential for acute thyroid changes after amiodarone therapy, transient hypothyroidism in adults is often tolerable and not aggressively treated. Transient hypothyroidism is not tolerable in neonates, infants, and young children, who may accrue potentially irreversible changes in brain development.12,14 We recognize that amiodarone is often the optimal antiarrhythmic choice in acutely ill patients, particular postoperatively because of to its pharmacologic properties and ability to treat various atrial and ventricular arrhythmias. Alternative antiarrhythmic therapy should be considered when possible to avoid the side effects noted with amiodarone, including thyroid dysfunction. However, in the many cases in which amiodarone use is unavoidable, we propose the need for
Creo et al
Amiodarone-Induced Thyroid Dysfunction in Children
Table 4 Suggested timetable for thyroid function screening in children receiving amiodarone When to obtain thyroid-stimulating hormone and free thyroxine levels Age ,3 y Baseline* Week 2 Week 4 Week 8 Every 2 months thereafter 3 months after amiodarone discontinuation Age 3 y Baseline* Week 4 Every 3 months for the first year on therapy Every 6 months thereafter 3 months after amiodarone discontinuation *Obtained before any amiodarone exposure when possible.
more rigorous pediatric-specific thyroid monitoring guidelines in children receiving amiodarone and screening for thyroid dysfunction early in the course of amiodarone treatment given the concern for poor brain development due to hypothyroidism and the short lag time between amiodarone initiation and development of elevated TSH values in our patients. Given the risk of cerebral effects of hypothyroidism and the frequency of significant TSH elevations in our study, we propose more frequent thyroid monitoring in children.12–15 Other investigators also endorse more frequent thyroid monitoring in children than in adults.16–18 Costigan et al16 suggested measuring T4, T3, and TSH “often” during the first year of treatment, and even more frequently in children ,1 year. Trudel et al18 proposed performing TSH monitoring in children ,4 years of age at baseline, 1 week, every 2 weeks of the next month, every 4 weeks for the next 3 months, and every 3 months thereafter. We have modified Trudel’s protocol based on our data and propose a similar timetable for thyroid monitoring in children (Table 4).18 Our data demonstrate that of the children who developed an elevated TSH level, most did so within 2–4 weeks (IQR 3–19 days for all children); however, there was a wide distribution, with some children developing elevated TSH levels months after initial exposure.
Study limitations Our results should be interpreted in the context of several limitations. First, the children in our cohort were sick and likely had been exposed to other confounding agents. We performed analysis of variables, which could potentially be associated with greater degrees of amiodarone dysfunction and did not find any significant associations. However, we still had a relatively small statistical sample size. Also, at our institution many children undergo heart surgery and then follow-up closer to home on discharge, so our sample mostly captured hospitalized patients and may not be generalizable to less ill children receiving amiodarone in the outpatient setting. This study was also limited due to the retrospective
1441 nature of the review. Although the majority of patients receiving amiodarone had TFTs performed, some did not. Also, most patients did not have TFTs performed before amiodarone initiation, so the true percentage of children who would have had abnormal thyroid function even before amiodarone initiation remains unknown. However, we excluded patients with known risk factors for abnormal TFTs and a low likelihood of having an elevated TSH value at baseline. Furthermore, sick children tend to have lower, not higher, TSH values due to sick euthyroid syndrome, such that our cohort would be at higher risk for developing a low TSH value. It is possible that TFTs were checked or were checked more frequently in patients displaying possible symptoms of hypothyroidism, potentially artificially increasing the prevalence and degree of TSH elevation among patients. Lastly, although TFTs were obtained, they were not drawn frequently enough to confidently capture the true timing of the peak TSH value or time to initial abnormal TSH value after amiodarone exposure.
Conclusion Amiodarone has the potential to disrupt normal thyroid function in pediatric patients. In this cohort of patients receiving amiodarone who had TFTs checked, the majority developed elevated TSH values, in contrast to older reports. Neonates and infants were most prone to developing elevated TSH values. In patients who had .1 TFT performed, peak median TSH values were above age-appropriate limits in the neonate, infant, young children, and older child groups. Most patients who developed elevated TSH values did so within the first 2 weeks of amiodarone exposure. Given the concern for poor brain development and growth in hypothyroid children, our results suggest the need for more rigorous pediatric-specific thyroid monitoring guidelines and screening early during therapy to monitor for thyroid dysfunction.
References 1. Figa FH, Gow RM, Hamilton RM, Freedom RM. Clinical efficacy and safety of intravenous Amiodarone in infants and children. Am J Cardiol 1994;74:573–577. 2. Ha HR, Stieger B, Grassi G, Altorfer HR, Follath F. Structure-effect relationships of amiodarone analogues on the inhibition of thyroxine deiodination. Eur J Clin Pharmacol 2000;55:807–814. 3. Iervasi G, Clerico A, Bonini R, et al. Acute effects of amiodarone administration on thyroid function in patients with cardiac arrhythmia. J Clin Endocrinol Metab 1997;82:275–280. 4. Latini R, Tognoni G, Kates RE. Clinical pharmacokinetics of amiodarone. Clin Pharmacokinet 1984;9:136–156. 5. Institute of Medicine (US) Panel on Micronutrients. Dietary Reference Intakes for Vitamin D, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Washington, DC: National Academy Press; 2001. 6. Rao RH, McCready VR, Spathis GS. Iodine kinetic studies during amiodarone treatment. J Clin Endocrinol Metab 1986;62:563–568. 7. Franklyn JA, Davis JR, Gammage MD, Littler WA, Ramsden DB, Sheppard MC. Amiodarone and thyroid hormone action. Clin Endocrinol 1985;22:257–264. 8. Roti E, Minelli R, Gardini E, Bianconi L, Braverman LE. Thyrotoxicosis followed by hypothyroidism in patients treated with amiodarone. A possible consequence of a destructive process in the thyroid. Arch Intern Med 1993; 153:886–892. 9. Czarnywojtek A, Plazinska MT, Zgorzalewicz-Stachowiak M, et al. Dysfunction of the thyroid gland during amiodarone therapy: a study of 297 cases. Ther Clin Risk Manag 2016;12:505–513.
1442 10.
11. 12. 13. 14.
15. 16. 17. 18. 19.
20.
21.
22.
Trip MD, Wiersinga W, Plomp TA. Incidence, predictability, and pathogenesis of amiodarone-induced thyrotoxicosis and hypothyroidism. Am J Med 1991; 91:507–511. Goldschlager N, Epstein AE, Naccarelli GV, et al. A practical guide for clinicians who treat patients with amiodarone: 2007. Heart Rhythm 2007;4:1250–1259. Koibuchi N, Chin WW. Thyroid hormone action and brain development. Trends Endocrinol Metab 2000;11:123–128. Rovet JF. The role of thyroid hormones for brain development and cognitive function. Endocr Dev 2014;26:26–43. Zoeller RT, Rovet J. Timing of thyroid hormone action in the developing brain: clinical observations and experimental findings. J Neuroendocrinol 2004; 16:809–818. Williams GR. Neurodevelopmental and neurophysiological actions of thyroid hormone. J Neuroendocrinol 2008;20:784–794. Costigan DC, Holland FJ, Daneman D, Hesslein PS, Vogel M, Ellis G. Amiodarone therapy effects on childhood thyroid function. Pediatrics 1986;77:703–708. Niinikoski H, Hamalainen AM, Ekblad H, Soukka H. Neonatal hypothyroidism after amiodarone therapy. Acta Paediatr 2007;96:773–774. Trudel K, Sanatani S, Panagiotopoulos C. Severe amiodarone-induced hypothyroidism in an infant. Pediatr Crit Care Med 2011;12:e43–e45. Lomenick JP, Jackson WA, Backeljauw PF. Amiodarone-induced neonatal hypothyroidism: a unique form of transient early-onset hypothyroidism. J Perinatol 2004;24:397–399. Pavan-Senn CC, Nesi-Franca S, Pelaez J, et al. [Transient neonatal hypothyroidism due to amiodarone administration during pregnancy—two cases report and review of literature]. Arq Bras Endocrinol Metabl 2008;52:126–130. Vanbesien J, Casteels A, Bougatef A, et al. Transient fetal hypothyroidism due to direct fetal administration of amiodarone for drug resistant fetal tachycardia. Am J Perinatol 2001;18:113–116. Coumel P, Fidelle J. Amiodarone in the treatment of cardiac arrhythmias in children: one hundred thirty-five cases. Am Heart J 1980;100:1063–1069.
Heart Rhythm, Vol 16, No 9, September 2019 23.
24.
25. 26.
27. 28.
29.
30.
31. 32.
33.
Guccione P, Paul T, Garson A Jr. Long-term follow-up of amiodarone therapy in the young: continued efficacy, unimpaired growth, moderate side effects. J Am Coll Cardiol 1990;15:1118–1124. Vignati G, Danzi GB, Austoni P, Fancini P, Mascarello M, Figini A. [Amiodarone therapy in childhood: efficacy and side effects]. G Ital Cardiol 1985; 15:786–794. Moffett BS, Valdes SO, Kim JJ. Amiodarone monitoring practices in pediatric hospitals in the United States. Pediatr Cardiol 2013;34:1762–1766. American Heart Association. Congenital heart defects. 2018. Available at: http:// www.heart.org/HEARTORG/Conditions/CongenitalHeartDefects/CongenitalHeart-Defects_UCM_001090_SubHomePage.jsp. Gandy J, Wonko N, Kantoch MJ. Risks of intravenous amiodarone in neonates. Can J Cardiol 1998;14:855–858. Bartalena L, Bogazzi F, Braverman LE, Martino E. Effects of amiodarone administration during pregnancy on neonatal thyroid function and subsequent neurodevelopment. J Endocrinol Invest 2001;24:116–130. Matsumura LK, Born D, Kunii IS, Franco DB, Maciel RM. Outcome of thyroid function in newborns from mothers treated with amiodarone. Thyroid 1992; 2:279–281. Dahl AR, Iqbal AM, Lteif AN, Pittock ST, Tebben PJ, Kumar S. Mild subclinical hypothyroidism is associated with pediatric dyslipidaemia. Clin Endocrinol (Oxf) 2018;89:330–335. Marwaha RK, Tandon N, Garg MK, et al. Dyslipidemia in subclinical hypothyroidism in an Indian population. Clin Biochem 2011;44:1214–1217. Salerno M, Capalbo D, Cerbone M, De Luca F. Subclinical hypothyroidism in childhood - current knowledge and open issues. Nat Rev Endocrinol 2016; 12:734–746. Jonklaas J, Bianco AC, Bauer AJ, et al. American Thyroid Association Task Force on Thyroid Hormone Replacement. Guidelines for the treatment of hypothyroidism: prepared by the American Thyroid Association Task Force on Thyroid Hormone Replacement. Thyroid 2014;24:1670–1751.