Clinical survey of current practice regarding treatment of children with borderline thyroid abnormalities

Abstract

Aim

Recently, there has been debate about reducing newborn screening (NBS) thyroid‐stimulating hormone (TSH) cut‐offs to identify children with mild, but potentially clinically significant, thyroid deficiency. Once identified by NBS, these children will be referred to paediatric endocrinologists for further testing and possible treatment; however, variation in current clinical practice is not known. The aim of this study is to survey Paediatric Endocrinologists in Australia and New Zealand to gain insight into clinical practice for the treatment of mild thyroid deficiency.

Methods

A piloted questionnaire was sent to members of the Australasian Paediatric Endocrinologist Group. The survey asked the Australasian Paediatric Endocrinologist Group members about the investigations performed, treatment and follow‐up for infants with different confirmatory serum TSH levels.

Results

There were 42 completed surveys, a response rate of 34%. When presented with four case studies, 7% of clinicians would treat a child with confirmatory serum TSH of 8.7 mU/L with thyroxine, 69% would treat a child with confirmatory serum TSH 21.4 mU/L, 76% would treat a child with confirmatory serum TSH 24.3 mU/L and 95% would treat a child with confirmatory serum TSH 44.7 mU/L.

Conclusion

This contemporary survey of clinicians regarding the treatment of mild thyroid deficiency in children has shown that clinical practice varies extensively. International and national guidelines on the treatment of congenital hypothyroidism should be updated to incorporate new evidence and ensure consistency across clinical practice.

What is already known on this topic

1. Over the past 10 years there has been increasing research examining the outcomes of children with mild thyroid abnormalities at birth.

2. Some newborn screening programs have lowered the thyroid stimulating hormone (TSH) threshold for congenital hypothyroidism (CH) screening to identify infants with milder thyroid abnormalities.

3. These children are referred to endocrinologists for further investigations for CH, however the contemporary clinical practice for the treatment of children with mild thyroid abnormalities is not known.

What this paper adds

1. There is a wide range of practice between clinicians regarding treatment of children with borderline TSH elevation in Australia ad New Zealand.

2. Local guidelines for diagnosis and treatment of CH should be reviewed and updated to ensure consistent practice across clinicians.

Newborn screening (NBS) of neonatal thyroid function, predominantly thyroid‐stimulating hormone (TSH), in the first few days of life to identify and treat congenital hypothyroidism (CH), has virtually eliminated significant intellectual disability associated with CH in developed countries. ¹ However, recently, there has been international debate about lowering the TSH threshold for CH screening to identify infants with milder, but potentially still clinically significant, thyroid abnormalities. ² Recent research examining the cognitive and developmental outcomes of children with mild thyroid deficiency has had conflicting results. ³ , ⁴ , ⁵ , ⁶ A large study of over 500 000 infants by our research group found children with a neonatal TSH just below NBS thresholds had poorer education and developmental outcomes compared to those with normal TSH levels ⁷ ; however, the benefit of treatment for these children is unknown.

The lowering of NBS thresholds for CH will result in the identification of additional infants that will be referred to paediatric endocrinologists for further testing, including potentially healthy infants that will not require treatment. Within Australia and New Zealand, NBS programmes use a range of TSH thresholds to identify infants, from 8 to 15 mIU/L. Once referred from NBS to an endocrinologist or general paediatrician for further investigations, the treatment of children with mildly elevated thyroid levels depends on the results of further tests and the clinical judgement of the clinician.

There is little published information regarding the diagnosis and treatment of children with mildly elevated TSH levels once they have been referred to a clinician. One UK study asked endocrinologists if they would treat patients with thyroxine when presented with four case studies detailing results from serum samples ⁸ ; however, this study was small (n = 16) and performed over 12 years ago. Given the recent research examining the outcomes of children with mild thyroid deficiency, and the potential for more children to be referred to endocrinologists for further investigations for CH, it is important to understand contemporary clinical practice. The aim of this study is to survey clinicians who currently manage children with CH in Australia and New Zealand to gain insight into clinical practice for the treatment of mild thyroid deficiency.

Methods

This was a cross‐sectional survey of current members of the Australasian Paediatric Endocrine Group (APEG). APEG is the premier professional body that supports clinical care, advocacy, education and research in paediatric endocrinology in Australia and New Zealand. At the time of this study, APEG membership included 96 paediatric endocrinologists and 28 general paediatricians.

Survey instrument and enrolment

The questionnaire included questions regarding investigations, treatment and follow‐up of infants with different mildly elevated NBS blood spot results and confirmatory serum TSH levels. APEG members were given four clinical scenarios, based on cases from a UK study, ⁸ with initial TSH levels modified due to timing of NBS collection of blood spots on day 5 in the UK compared to days 2–3 in Australia and New Zealand. The survey was piloted on members of the APEG Thyroid Working Group.

Current members of the APEG were emailed an invitation to participate in the study with a link to complete an anonymous online survey hosted on the RedCap online platform. The invitation was included in the fortnightly newsletter and sent to members three times over 6 weeks. Ethics approval was gained from the University Human Research Ethics Committee (Project number 2020/156).

Statistical analysis plan

Descriptive statistics of investigations, treatment and follow‐up in response to each clinical scenario were calculated and differences between factors were examined using χ² test for categorical data. This was used to compare responses in treatment of CH between sub‐groups of: (i) Fellows trained in Australia versus other countries; (ii) Paediatric endocrinologists versus other clinicians; and (iii) clinicians by length of time since they began practicing in paediatric endocrinology (less than or greater than 15 years).

Results

There were 42 respondents that completed the survey, including 30 paediatric endocrinologists, 8 paediatricians and 4 endocrinology fellows/trainees, giving a response rate of 34%. The majority of respondents were female (67%), paediatric endocrinologists (71%), trained in Australia (60%), and had a range of clinical experience (Table 1).

Table 1.

Characteristics of respondents

	n (%)
Gender
Male	13 (31)
Female	28 (67)
Different identity	1 (2)
Professional role
Paediatric endocrinologist	30 (71)
General paediatrician	8 (19)
Fellows/Trainees	4 (10)
State/Country of practice
NSW	8 (19)
Queensland	8 (19)
Victoria	8 (19)
WA	5 (12)
New Zealand	13 (31)
State/Country of training
NSW	6 (14)
Queensland	8 (19)
Victoria	7 (17)
WA	4 (10)
New Zealand	12 (28)
Other	5 (12)
Length of practice in paediatric endocrinology, years
<5	10 (24)
5–15	13 (31)
15–30	13 (31)
>30	6 (14)

NSW, New South Wales; WA, Western Australia.

Investigations for CH

The details and responses to the four case studies presented are provided in Table 2. When presented with an infant with mild TSH elevation (case 1: 8.7 mIU/L at confirmatory serum) and asked which investigations would be performed (more than one could be chosen); 81% of respondents would repeat thyroid function tests (TFTs) at a later date, 17% would perform a thyroid ultrasonography and 17% would perform a thyroid 99mT or 123I scan (Table 2). For children with elevated TSH levels (case 4: 44.7 mIU/L at confirmatory serum), over three quarters (77%) of clinicians would perform a thyroid 99mT or 123I scan, and 38% would perform a thyroid ultrasound.

Table 2.

Details of case studies presented to respondents and investigations and treatment reported

Case study details				Number (%) of investigative tests reported by respondents (respondents could chose more than 1)							APEG survey results, n (%)	UK survey results, 7 n (%)
	First NBS bloodspot TSH, mU/L WB†	Repeat NBS bloodspot TSH, mU/L WB†	Serum results: TSH and FT4	Repeat TFTs	Thyroid US	Knee X‐ray‡	Thyroid 99mT or 123I scan	Thyroid antibodies	Urinary iodine	Maternal thyroid antibody	Would treat with thyroxine	Would treat with thyroxine
Case 1: Baby born at 36 weeks GA, 2.1 kgs Serum obtained at 17 days old	8.3	8.1 mIU/L	TSH: 8.7 mU/L FT4: 18.0 pmol/l	34 (81)	7 (17)	1 (2)	7 (17)	7 (17)	6 (14)	13 (31)	3 (7)	0/17 (0)
Case 2: Baby born at 42 weeks GA, 3.3 kgs Serum obtained at 16 days old	8.1	8.4 mU/L	TSH: 21.4 mU/L FT4: 18.4 pmol/l	31 (74)	15 (36)	2 (5)	23 (55)	12 (29)	7 (17)	13 (31)	29 (69)	10/17 (59)
Case 3: Baby born at 38 weeks GA, 3.1 kgs Serum obtained at 35 days old	8.2	9.4 mU/L	TSH: 24.3 mU/L FT4: 17.2 pmol/l	31 (74)	17 (40)	1 (2)	25 (60)	14 (33)	9 (21)	12 (29)	32 (76)	16/17 (94)
Case 4:Baby born at 40 weeks GA, 3.5 kgs Serum obtained at 29 days old	8.7	20.8 mU/L	TSH: 44.7 mU/L FT4: 18.0 pmol/l	25 (60)	16 (38)	2 (5)	32 (77)	11 (26)	6 (14)	13 (31)	40 (95)	17/17 (100)

^†Conversion factor: Neonatal TSH in milliunits per litre whole blood × 2.2 = milliunits per litre serum.

^‡Knee X‐rays are used to identify bone maturation, with the absence of epiphyses may reflect the degree of intra‐uterine hypothyroidism.

FT4, free thyroxine; GA, gestational age; mU/L, milliunits per litre; NBS, newborn screening; pmol/L, picomol per litre; TFT, thyroid function test; TSH, thyroid‐stimulating hormone; US, ultrasound; WB, whole blood.

Treatment of children with borderline TSH levels at confirmatory serum

Seven percent of respondents would treat case study 1 with thyroxine, 69% would treat case 2, 76% would treat case 3 and 95% would treat case 4 (Table 2). The three clinicians that would treat case 1 were all paediatric endocrinologists. If clinicians stated they would not treat with thyroxine, they would perform repeat TFTs for all case studies, except for two (5%) clinicians that would not treat or repeat TFTs for case 1. A separate question asked about clinical management of a term infant with confirmatory serum TSH of 8.9 mU/L and FT4 14 pmol/L and half (51%) of respondents stated they would ‘repeat TFTs in 2–4 weeks and treat if TSH >10 mIU/L’. While 22% stated they would ‘treat if they had symptoms consistent with hypothyroidism, congenital abnormalities or relevant family history’, 10% would ‘treat if TSH >20 mIU/L’, 10% would ‘treat only if further investigations suggest a thyroid abnormality’ and 10% would ‘treat if serum FT4 was below the normal range for age regardless of TSH level’.

Follow‐up

Clinical reviews were commonly performed every 3 months for children aged 3 months to 3 years, and every 6 months for children aged 3 years to beyond puberty. For children aged less than 3 years, half of the clinicians (46%) would adjust the dose of thyroxine to aim for TSH in the lower half of normal age‐appropriate range, while 30% of respondents aim for FT4 > 15 mmol/L and 16% aim for TSH levels in the normal age‐appropriate range (Fig. 1). There was a range in responses as to how soon physicians recheck TFTs after initiating treatment with thyroxine; 26% would check 5–7 days after, 12% after 8–10 days, 40% after 11–14 days and 14% more than 14 days after. For children with a normally sited thyroid gland who had not required increasing doses of thyroxine with growth, 35% of respondents would trial thyroxine cessation at 2 years of age, and 55% would trial stopping treatment at 3 years of age.

Fig. 1.

Aim for adjusting dose of thyroxine by age of child with congenital hypothyroidism. Respondents could only choose one response per age group. (), Aim for TSH <10 mU/L; (), aim for TSH in the normal age range; (), aim for TSH in the lower half of the normal range; (), aim for FT4 >15 pmol/L; (), other.

Difference by clinician type, country of training and length of practice

The only significant difference between responses by clinician type was regarding trialling thyroxine cessation for children with normally sited thyroid gland. The majority of paediatric endocrinologists would trial thyroxine cessation at 3 years of age (61%), while the majority of other clinicians (54%) would trial thyroxine cessation at 2 years of age (P = 0.04). There was no difference in responses by country/state of training or length of practice in paediatric endocrinology.

Discussion

This survey of paediatric endocrinologists and general paediatricians in Australia and New Zealand highlights the wide range of practice between clinicians regarding treatment of children with borderline TSH elevation. Identifying children via NBS is only the first stage of ensuring at‐risk children are treated and this survey has shown that the frequency of follow‐up and the levels of thyroid hormone at which treatment is commenced may differ by clinician.

APEG developed guidelines in 2007 for the management and treatment of CH; however, the guidelines do not specify a level of elevated serum TSH at which treatment should begin. ⁹ The European Society for Paediatric Endocrinology (ESPE) has recently updated their guidelines and recommends that if serum TSH concentration is >20 mU/l at confirmatory testing then treatment should be started, ¹⁰ while the American Academy of Pediatrics (AAP) guidelines suggest if ‘persistent basal TSH’ concentration is higher than 10 mU/L after the first 2 weeks of age then the infant should be treated. ¹¹ According to both guidelines, cases 2, 3 and 4 in our study, and in the UK study on which the details of the cases were based, ⁸ should have been treated with thyroxine. In the UK, 59, 94 and 100% of respondents reported they would treat cases 2, 3 and 4, respectively (Table 1). This compares to 69, 76 and 95% of our Australia and New Zealand respondents who would treat cases 2, 3 and 4 with thyroxine, respectively. Although our study included both paediatric endocrinologists, paediatricians and fellows, compared to just paediatric endocrinologists surveyed in the UK study, there was no significant difference in those who would treat the cases with thyroxine by clinician type.

For infants with mildly elevated TSH levels, the published guidelines also differ. ESPE guidelines suggest treatment if serum TSH is between 6 and 20 mU/L in a well‐baby for 3–4 weeks ¹⁰ while the AAP suggest for persistent elevated TSH (6–10 mU/L at 1 month), treat if TSH >10 mU/L at 6 weeks. ¹¹ For case study 1 with mild TSH elevation (serum TSH of 8.7 mU/L), in the UK, no clinicians would have treated this infant, but three paediatric endocrinologists (7%) trained in Australia answered that they would treat this infant with thyroxine (Table 1). Our survey and the UK survey of clinicians were carried out over 12 years apart and there have been a number of studies examining cognitive outcomes of untreated children with mildly elevated NBS TSH levels (8–15 mU/LWB) in the interim, however, studies show conflicting results. ⁴ , ⁵ , ⁶ , ⁷

For children with thyroid abnormalities identified with a newborn TSH level above NBS cut‐offs, prompt initiation of treatment that normalises thyroid levels has been shown to result in mostly normal neurocognitive outcomes. ¹² The recently updated European guidelines recommend monitoring treatment according to age‐specific TSH and fT4 reference intervals; however, if TSH is in age‐specific interval then ‘fT4 concentrations above the upper limit of the reference interval can be accepted’. ¹⁰ Depending upon the age of patients, clinicians surveyed adjusted the dose of thyroxine according to different treatment goals, including aiming for TSH <10 mU/L or fT4 >15 pmol/L. As thyroid hormone levels change markedly across childhood, age‐specific ranges should be applied, and clinicians need to be aware of the impact of different laboratory analytical methods on reference ranges. ¹³

This survey provides insight into contemporary practices of paediatric endocrinologists and general paediatricians in Australia and New Zealand, in the diagnosis and treatment of children with CH. A strength of the study is the broad range of geographical areas, level of clinical experience of the respondents and including both paediatric endocrinologists and general paediatricians. Although the survey response rate of 34% was lower than the response rates for similar surveys in the USA and UK, 59 and 86%, respectively, ¹⁴ , ¹⁵ it is similar to an online survey of paediatricians across all states of Australia performed in 2017 (36%). ¹⁶ The sample size is also considerably larger than the 16 respondents of the UK study that surveyed paediatric endocrinologists. ⁸

Conlusion

The responses to the survey show that for a child with mildly elevated TSH levels, their choice of clinician could greatly impact the subsequent management they receive. The recent ‘Consensus Guidelines Update’ by ESPE ¹⁰ incorporates recent literature for the diagnosis and treatment of CH however other national guidelines for CH, such as AAP and APEG guidelines, should also be reviewed and expanded to include new evidence, particularly on the cognitive outcomes of children with mildly elevated TSH levels. This will ensure up‐to‐date evidence‐based guidelines are available to ensure consistent practice across clinicians.