Thyroid Hormone Resistance and Defects of Thyroid Hormone Action and Metabolism

Resistance to thyroid hormone β (THRB) and α (THRA), selenoprotein deficiency (SECISBP2, TRU-TCA1-1) and type 1 deiodinase deficiency (DIO1), within the classification of impaired sensitivity to thyroid hormone; MCT8 deficiency and primary congenital hypothyroidism enter the differential diagnosis.

Literature and regulatory information update: 27 October 2026.

0. Reference Summary Table

Table 0.1. Identification of the entities covered in this article. The abbreviations are those used in the text: thyroid hormones (TH), free thyroxine (FT4), free triiodothyronine (FT3), reverse T3 (rT3), thyrotropin (TSH), sex hormone-binding globulin (SHBG), triiodothyroacetic acid (TRIAC, tiratricol) and European Reference Network on Rare Endocrine Conditions (Endo-ERN).

1. Introduction

1.1. Definition and the concept of sensitivity to thyroid hormone

Thyroid hormone action depends on a chain of steps that begins in the gland and ends in the nucleus of the target cell: secreted thyroxine (T4) circulates bound to plasma proteins, enters cells through membrane transporters, is converted to the active form, triiodothyronine (T3), by the selenocysteine-containing deiodinases, and T3 binds to nuclear receptors that regulate the transcription of target genes. A defect at any of these steps reduces the biological effect of a chemically intact hormone, secreted in normal or excessive amounts, and this is the meaning of the term "reduced sensitivity to thyroid hormone" (RSTH), introduced in 2005 [1]. After the 12th International Workshop on Resistance to Thyroid Hormone, the concept came to cover three levels: defects of membrane transport, defects of intracellular metabolism (generation of T3 from T4) and defects of action mediated by the nuclear receptors [2].

Resistance to thyroid hormone β (RTHβ) is defined by the combination of genuinely raised serum total and free T4 and T3 with non-suppressed TSH, caused by a pathogenic variant in the THRB gene encoding the β receptor [2]. Resistance to thyroid hormone α (RTHα) results from variants in the THRA gene encoding the α receptor and has a clinically inverse profile: TRα-dependent tissues behave as hypothyroid with near-normal thyroid function tests, marked by a low T4/T3 ratio [1, 2]. Selenoprotein deficiency results from biallelic variants in SECISBP2 or in TRU-TCA1-1, which impair the synthesis of all selenoproteins, including the deiodinases, and produces a multisystem disease with high FT4 and rT3 and low or normal FT3 [2]. Type 1 deiodinase (D1) deficiency, caused by heterozygous variants in DIO1, is to date a biochemical phenotype with raised rT3 and rT3/T3 ratio [2, 6]. The RTH phenotype without a THRB variant is called "non-TR-RTH" (RTH without a receptor variant) [1].

1.2. History of the clinical and molecular description

In 1967, Samuel Refetoff, Loren DeWind and Leslie DeGroot described a family with deaf-mutism, stippled epiphyses, goitre and abnormally high protein-bound iodine (PBI), and proposed end-organ refractoriness to thyroid hormone [44]. The 2024 European guideline identifies this family as carrying a homozygous THRB deletion, with dysmorphic facies and auditory and visual abnormalities [2]. The thyroid hormone receptor genes were cloned in 1986 and the first THRB variants in patients were identified in 1989 [1]. Linkage between RTH and the THRB locus (then called ERBA2 or c-erbA-β) was demonstrated in 1988, without recombination, in a family with dominant inheritance, and the first variant identified, in 1989, was a glycine-to-arginine substitution in the hormone-binding domain, in a father and son [8, 28]. In 1991, Chatterjee and colleagues showed that mutant β receptors inhibit the function of the normal receptor (dominant-negative effect), which explained the dominant inheritance [45]. In the family described in 1967, Takeda and colleagues demonstrated in the affected members homozygous deletion of the entire THRB coding region and clinical and biochemical normality of the heterozygotes, which is the molecular basis of the recessive form; the conclusion that a single normal allele suffices for receptor function reinforced the dominant-negative interpretation of the heterozygous forms [20, 21]. In the same period, Usala and colleagues described another type of biallelic form, homozygous deletion of a single codon (threonine 337) in a patient with severe generalised resistance, whose mutant receptor barely bound T3 [19, 46].

The selective pituitary form has a parallel history. In 1975, Gershengorn and Weintraub studied an 18-year-old woman with clinical and biochemical hyperthyroidism and persistently raised immunoreactive TSH, and called the condition "inappropriate secretion of TSH"; almost two decades later, genetic testing showed in this patient the THRB variant now designated p.(Arg338Trp) [29]. Between 1982 and 1991 family and paediatric reports accumulated: six women from three generations of one family, with hyperthyroidism due to chronic thyrotroph stimulation, in sustained remission on continuous liothyronine, interpreted by the authors as partial insensitivity of the thyrotrophs to the hormone, possibly due to deficient intrapituitary conversion of T4 to T3; three sporadic cases, the youngest a 4-year-old girl with goitre and thyrotoxic symptoms; a 15-month-old boy treated with D-thyroxine; and a 12-year-old girl with an affected mother and sister, in a report of control with TRIAC combined with propranolol [29]. These treatments document the era; the current European guideline bases control of the thyrotoxic symptoms of RTHβ on beta-blockade alone or, rarely, on TRIAC [2].

RTHα was described in 2012, when exome sequencing of a child with growth retardation, developmental delay, skeletal dysplasia and severe constipation, with only borderline thyroid hormones, identified a heterozygous de novo nonsense variant in THRA; the mutant protein inhibited the normal receptor in a dominant-negative manner [47]. In 2014 the first patients with a variant affecting both the TRα1 and TRα2 isoforms (p.Ala263Val) were published, an index case and two children who improved with thyroxine despite normal circulating hormones [48].

The metabolic defects emerged from the study of fibroblasts. In 2005, Dumitrescu and colleagues studied a family in which 3 of 7 siblings had abnormal thyroid hormone metabolism, with reduced type 2 deiodinase (D2) activity not linked to the DIO2 locus, and identified by linkage analysis a homozygous missense variant in SECISBP2; an unrelated child with a similar phenotype was a compound heterozygote [49]. In 2010, the description of compound heterozygous patients with azoospermia, axial muscular dystrophy, photosensitivity and immune abnormalities defined the disease as multisystem selenoprotein deficiency [5]. The first variant in the human selenocysteine tRNA (tRNA-Sec) gene was published in 2016 and the second family with the same variant in 2021 [26, 50]. The first inherited D1 deficiency was described in 2021, in two unrelated families studied by exome sequencing [6].

1.3. Nomenclature and classification

The 1st International Workshop on RTH (Cambridge, United Kingdom, 1993) standardised the nomenclature of THRB variants; the 5th (Lyon, France, 2005) introduced the term RSTH; and the 10th (Quebec, Canada, 2012) proposed the designation "impaired sensitivity to thyroid hormone", because syndromes of increased sensitivity may exist [1]. The proposal published in 2014, simultaneously in three journals, kept "RTH" for the phenotype of congenitally raised FT4 with non-suppressed TSH, recommended "RTHβ" when there is a THRB variant and "RTHα" for THRA, and advised against the abbreviation "THR" as a synonym, because it is not the hormone that is resistant and the abbreviation has more than 20,000 Medline references with other meanings [1]. The 2024 European Thyroid Association (ETA) guideline updated the classification with the metabolic defects identified after 2014 (TRU-TCA1-1 and DIO1) and treats selective pituitary RTH as part of the RTHβ spectrum rather than as a distinct disease [2, 7]. Orphanet keeps the entry ORPHA:97927, "peripheral resistance to thyroid hormones", as obsolete [51].

The 1993 standardisation, published by Beck-Peccoz and colleagues in 1994, resolved two sources of confusion that persist in the older literature. The first THRB cDNA sequence predicted a 450-amino-acid protein, but the corrected sequence revealed an upstream initiation codon and a 461-amino-acid protein, so variants published before the revision have numbering 5 units lower: G340R corresponds to p.(Gly345Arg), P448H to p.(Pro453His) and A312T to p.(Ala317Thr); and the exons, numbered 00 to 8 or 1 to 10 depending on the authors, came to be numbered 1 to 10, so that the former exons 7 and 8 correspond to the current 9 and 10 [8]. OMIM keeps three phenotypic entries for THRB, the dominant generalised form (188570), the recessive generalised form (274300) and selective pituitary RTH (145650), and catalogues as selective pituitary variants that in other families cause generalised RTH, such as p.(Arg338Trp), associated with the pituitary form in 4 of 5 families in one series, which is consistent with the European guideline's reading of a single spectrum [2, 8, 29]. RTHα appears in OMIM as "congenital nongoitrous hypothyroidism 6" (614450), and the three metabolic defects as "abnormal thyroid hormone metabolism" 1, 2 and 3 (609698, 619855 and 620198) [23–25, 27].

Table 1.1. Classification of the inherited defects of sensitivity to thyroid hormone (Refetoff 2014, updated by the 2024 ETA guideline).

Table sources: Table 1 of Refetoff 2014 and the sections of the 2024 ETA guideline on the metabolic defects [1, 2, 6, 50].

1.4. Scope of the article

This article covers the class of defects of thyroid hormone action (RTHβ, non-TR-RTH, RTHα) and metabolism (selenoprotein deficiency due to SECISBP2 and TRU-TCA1-1, D1 deficiency), with a common structure and dedicated subsections for each entity where the clinical picture, diagnosis or treatment differ. MCT8 and OATP1C1 deficiencies appear here only in the differential diagnosis, as do primary congenital hypothyroidism (dysgenesis and dyshormonogenesis) and central hypothyroidism. Variants of the serum transport proteins (albumin, thyroxine-binding globulin and transthyretin), which alter thyroid function tests without altering hormone action, and TSH-secreting pituitary adenoma enter the differential diagnosis, because they are the most frequent causes of a profile similar to that of RTHβ.

2. Molecular Genetics and Pathophysiology

2.1. Receptors, isoforms and tissue distribution

Thyroid hormone receptors are nuclear transcription factors that bind T3 (and, with lower affinity, T4) and act as ligand-dependent transcriptional repressors or activators [7]. In the absence of hormone, the DNA-bound receptor recruits corepressors (NCOR1, SMRT); T3 binding releases them and recruits coactivators (such as SRC1), activating positively regulated target genes [53]. The THRB gene produces the TRβ1 and TRβ2 isoforms; TRβ2 is the pituitary isoform that mediates negative feedback on TSH, and TRβ also mediates hormone action in the liver, kidney and cochlea [7, 54]. The THRA gene produces TRα1, which binds T3 and acts in bone, gut, heart, skeletal muscle and brain, and, through alternative splicing, TRα2, which does not bind hormone and weakly antagonises TRα1 [10]. This distribution explains the clinical picture of each entity: in RTHβ, the hypothalamic-pituitary axis and the liver are resistant while the heart, which depends on TRα, receives excess hormone; in RTHα, the axis and the liver respond normally, TSH is unchanged and TRα-dependent tissues become hypothyroid [2, 7, 10].

2.2. RTHβ: the THRB gene and the dominant-negative mechanism

Most THRB variants are heterozygous and act through a dominant-negative effect: the mutant receptor, which binds hormone with lower affinity, releases the corepressor poorly or recruits the coactivator poorly, occupies the response elements on DNA and inhibits the normal receptor encoded by the other allele [2, 45]. Because pituitary feedback depends on TRβ, TSH is not suppressed by the raised hormones; continued stimulation of the thyroid maintains the hyperthyroxinaemia and the goitre, and tissues that express mainly TRα, in particular the myocardium, receive an excess of hormone to which they are sensitive [2, 7]. Tissue resistance, assessed in the prospective National Institutes of Health (NIH) study, decreases in the order pituitary, brain, bone, liver and heart [55].

Pathogenic variants cluster in three clusters of the hormone-binding domain, with two exceptions [2]. In the Chicago laboratory, 59 of 65 families had variants in exons 7 to 10, the mutagenic region of the gene [56]. The compilations of the 1990s located the variants in two hot spots of the T3-binding domain, codons 310 to 347 and 438 to 459, separated by a cold region of about 90 amino acids involved in dimerisation, where artificial mutations did not alter, or barely altered, T3 affinity and would therefore not be expected to produce detectable RTH; about 45% of the variants involved a CpG dinucleotide, and the variants recurring in unrelated families arose independently, as shown by de novo occurrences and distinct haplotypes [8]. The exon 7 variants, in the hinge domain, p.(Arg243Gln) and p.(Arg243Trp), do not significantly reduce T3 affinity in vitro, and the proposed mechanism is increased homodimer formation or a loss of affinity that manifests only after DNA binding [8]. More than 85% of families with RTH have a heterozygous THRB variant, and 23% of variants arise de novo [18]. Frameshift variants with a premature termination codon cluster at residues 435 to 453 and are associated with severe phenotypes; in the case described by van Geest and colleagues, the p.R438Lfs445X mutant (the article's nomenclature) did not respond to T3, did not release NCOR1 or SMRT and did not recruit SRC1 [53].

Biallelic forms are rare and of two types. THRB deletions behave as true recessives (OMIM 274300): heterozygous carriers of a deletion are clinically and biochemically normal, because the absence of protein exerts no dominant-negative effect [7, 19–21]. Homozygosity for missense variants, in contrast, produces an invariably severe picture, with impaired intellectual development, tachycardia, goitre, growth retardation and hearing loss, more severe than deletion RTH, because the mutant TRβ, in double dose, also interferes with TRα1 function [19]. The same ranking appears in the thyrotroph resistance index to T4 (product of the free T4 index and TSH; normal 136±73): 920 in those homozygous for the deletion, 4920 in the patient with a single mutant allele and between 10,456 and 239,120 in those homozygous for point variants, whereas the index of the heterozygotes in the deletion family is normal [19]. By 2017 five families with homozygous RTHβ had been described [57].

2.3. Mosaicism and RTH without a THRB variant

The first mosaicism in RTH was described in a father with the p.(Arg338Trp) variant present in leucocytes, buccal mucosa and prostate, with lower amplification of the mutant allele, and absent in fibroblasts, transmitted to both sons; the mutation probably arose de novo at an early embryonic stage that included the germline, and tissue sensitivity was heterogeneous [18]. In a 52-year-old woman with asthenia, palpitations and diarrhoea, Sanger sequencing in blood was negative and next-generation sequencing (NGS) identified p.(Ala317Thr) with an allele frequency of 12% in blood, 5.7% in urine, 17.9% in oral mucosa, 9.9% in nasal mucosa, 6.4% in skin and 0% in conjunctiva [58]. Before 2025, mosaicism causing mild RTHβ had been described only three times [53].

In about 10% of individuals with a clinical and biochemical RTH phenotype no THRB defect is identified [2]. In the Chicago series, 21 of 141 families (15%) had no variant detectable in genomic DNA or cDNA [18]. In five families studied in detail, all coding and non-coding exons of TRβ1 and TRβ2 were normal, linkage analysis excluded THRB in two and THRA in three, and the phenotype was indistinguishable from RTHβ, with thyrotroph hyposensitivity among the most severe; the authors proposed a cofactor defect [56]. In that analysis, non-TR-RTH accounted for about 10% of families with the classical phenotype, involvement of THRB and THRA was excluded by linkage in 5%, and transmission was dominant in 2 of the 6 families then known and compatible with a new defect or recessive inheritance in the remaining 4 [28]. In one of these families, the form not linked to THRA or THRB arose de novo in the proband and was transmitted to two children by different fathers, with pituitary and peripheral resistance requiring 8 to 10 times the usual doses of levothyroxine and liothyronine [28]. Some of the cases without a variant correspond to somatic mosaicism not detected in blood [2].

2.4. Genotype and phenotype in RTHβ

Clinical expression is highly variable, from asymptomatic to thyrotoxic, and depends on genotype, age, sex and whether or not the mother has RTH [2, 55]. For a subgroup of variants (termed "type 1"), whose loss of function is proportional to the hormone-binding defect, the rise in FT4, resting energy expenditure and low-density lipoprotein (LDL) cholesterol correlates with genotype [2]. Homozygotes, whether for deletions or point variants, have dysmorphic facies, audiovisual abnormalities, intellectual disability, tachyarrhythmias and thyrotoxic cardiomyopathy [2]. In functional assays with the TRβ2 isoform, all mutants studied in a multicentre cohort had reduced T3-dependent activity, absent in the p.(Gly432del) mutant; the potency of TRIAC was similar to that of T3 for most mutants, but five times greater for p.(Arg316His) [54]. Genotype alone does not determine the picture. The same p.(Arg316His) variant was identified in an adolescent girl with severe selective pituitary RTH, clinical hyperthyroidism, advanced bone age and raised SHBG, but also in her clinically normal father and half-sister; the mutant receptor had a T3-binding defect without detectable dominant-negative activity, and the authors considered a second cause [8, 29]. The designation of selective pituitary form rests on the predominance of thyrotoxic manifestations, such as palpitations, anxiety, tremor, heat intolerance, insomnia, weight loss and increased stool frequency, which suggests peripheral tissues less refractory than the pituitary, and not on a distinct genotype [29]. In a series of 29 families published in 1994 (20 with the generalised form and 9 with the pituitary form, sporadic or with dominant transmission), all affected individuals were heterozygous for a single-nucleotide substitution in THRB, with a single exception, a 7-nucleotide insertion; dominant cosegregation was documented in 4 families with the pituitary form, sporadic variants had a similar frequency in both forms, the same variant gave rise to one or the other phenotype in different families, the two phenotypes coexisted in members of the same family and thyrotoxic expression varied even within a single individual [29, 59]. The 20 variants (13 new and 7 already known) clustered, with one exception, in the two clusters of the hormone-binding domain; 15 families shared 6 variants, which arose independently according to haplotype analysis, and 14 of the 19 recurrent variants, against 1 of the 10 unique ones, were transitions at CpG dinucleotides [59]. The degree of reduction in T3 binding by the mutant receptor did not correlate with the magnitude of thyroid dysfunction, and the authors concluded that the generalised and pituitary forms are phenotypic variants of the same disease, modulated by factors unrelated to the variant [59]. There was no clear association between variant and pituitary form, although p.(Arg338Trp) presented with this phenotype in 4 of the 5 carrier families in the series and in 2 further published cases [29]. The substitutions at codon 453 illustrate the variability between families: p.(Pro453His) was associated with hyperactivity, learning difficulties and short stature, and p.(Pro453Thr), in another family, only with diffuse goitre, without short stature, bradycardia, hypercholesterolaemia or learning difficulties; the differences between families with the same variant suggest genetic modifiers [8].

2.5. THRB allelic disease: macular dystrophy

THRB controls the diversity and survival of retinal cones [60]. In the mouse, deletion of Thrb eliminates the medium-wavelength ("green") cones and broadens the distribution of S cones; isolated deletion of the TRβ2 isoform reproduces the central TH resistance of the complete deletion but, unlike it, does not cause deafness, which assigns the main role in hearing to TRβ1 [8]. The variants causing RTHβ lie in the 3' exons encoding the ligand-binding domain and are accompanied, when sought, by a moderate and variable colour vision deficit; non-syndromic macular dystrophies are associated with variants in a 5' exon encoding the N-terminal domain of one of the isoforms [2, 60]. In one family, a 67-year-old woman with visual impairment since childhood (occult macular dystrophy) and her 39-year-old son (vitelliform macular dystrophy) were heterozygous for c.283+1G>A [61]. In a preprint not yet peer reviewed (medRxiv, April 2026), the c.283+1G>A and c.283G>C variants cause skipping of the TRβ1-specific exon 4, with loss of 87 amino acids in the N-terminal domain and a receptor twice as active; carriers had normal FT4 and TSH, without goitre, tachycardia, hypercholesterolaemia or steatosis, but with altered colour vision and reduced bone density, and the authors propose a "TRβ1 gain-of-function syndrome" [62].

2.6. RTHα: the THRA gene

THRA variants are heterozygous, de novo or inherited, and loss-of-function; they affect TRα1 alone or both TRα1 and TRα2 [10]. Mutant TRα1 does not activate target genes and inhibits the normal receptor through a dominant-negative effect, whereas mutant TRα2 behaves like the normal one [22, 48]. Severity follows residual function: variants that eliminate the C-terminal helix of TRα1 (truncations such as p.(Glu403Ter) and p.(Glu395Ter)) give the severe forms [47, 63]. The p.(Ala263Val) variant has a partial loss of function that high T3 concentrations reverse in vitro, and the p.(Leu274Pro) mutant is practically inert, responding only to 10 µM T3; this difference correlates with the patients' phenotype and response to levothyroxine [22]. By 2019, 21 THRA variants were known [63]. OMIM lists as reference variants p.(Glu403Ter), de novo, in the first case; a 1-bp insertion inherited in a father and daughter; and, in a Polish series of 5 children, p.(Cys392Ter) and p.(Glu403Ter), de novo, p.(Glu403Lys), inherited from an affected father, and p.(Pro398Arg), de novo, with no variant in the fifth patient [11, 23]. A different substitution at the p.(Ala263Val) codon, p.(Ala263Ser), was described in a family with 7 patients with mild symptoms of hypothyroidism [38]. In the Moran 2017 series, 9 of the 14 THRA variants then recorded had arisen de novo [22]. A single case with a variant affecting TRα1 and TRα2 had severe bone malformations, hypercalcaemia with hyperparathyroidism and diarrhoea instead of constipation, and it is not known whether the whole phenotype results from the variant [1].

2.7. The selenocysteine pathway: SECISBP2 and TRU-TCA1-1

Selenium exerts its function mainly as selenocysteine (Sec), incorporated into the 25 human selenoproteins, among them the glutathione peroxidases, which remove reactive oxygen species (ROS), and the three deiodinases [2, 64]. Selenocysteine is not encoded by a codon of its own: it is incorporated during translation at a UGA codon, which would otherwise be a stop codon, thanks to a SECIS element (SElenoCysteine Insertion Sequence) in the 3' untranslated region of the mRNA [64]. SBP2, encoded by SECISBP2, binds the SECIS element when the ribosome reaches the UGA and contacts the 40S subunit, which allows the elongation factor EEFSEC to deliver the selenocysteinyl-tRNA-Sec [12]. tRNA-Sec, the only natural suppressor tRNA of the UGA codon, is encoded by the single gene TRU-TCA1-1, charged with serine and converted to selenocysteine; the last step is catalysed by SepSecS (SEPSECS) [14, 64].

Biallelic SECISBP2 variants reduce the synthesis of most selenoproteins [5]. The thyroid phenotype results from reduced deiodinase activity, demonstrated for D2 in fibroblasts, which decreases the conversion of T4 to T3; the systemic phenotype results from the lack of tissue-specific selenoproteins (for example, in the testis) and from oxidative damage through loss of the antioxidant selenoenzymes [2, 5, 49]. Severity follows the residual function of SBP2, and incomplete loss explains the mild forms [12, 49]. Human SBP2 has 854 amino acids and is expressed in all tissues, much more in the testis; besides the full-length protein, there are shorter isoforms translated from downstream initiation codons (Met139, Met233 and Met300), and the minimal functional domain, which combines SECIS binding, Sec incorporation and ribosome binding, lies between codons 399 and 774 [13, 36]. All patients described up to 2020 retained partial activity: a premature stop is partly compensated by the short isoforms, as in homozygosity for p.(Arg128Ter), with a mild phenotype, whereas variants that truncate or inactivate the C-terminal domain give more severe pictures, as in compound heterozygosity for p.(Arg120Ter) and p.(Arg770Ter) [36]. In the table of Fu and colleagues, using the numbering of transcript ENST00000375807.7, 10 families had been published, 3 with homozygosity and 7 with compound heterozygosity, including intronic variants such as c.881-155T>A and c.1283+29G>A, which create aberrant splicing sites [13, 36]. The SECISBP2 gene is complex, with several isoforms, and there are variants in coding and non-coding regions, including intronic ones causing aberrant splicing [2].

The two known patients with a TRU-TCA1-1 variant are homozygous for the same acceptor-stem variant, n.65C>G ("C65G"), in non-coding transcript nomenclature, corresponding in ClinVar to NC_000019.10:g.45478610G>C (VCV000217868, pathogenic; gene on the minus strand) [2, 14, 65]. The variant reduces tRNA-Sec expression and the 2'-O-methylribosylation of uridine 34 of the anticodon, which selectively decreases the stress-related selenoproteins (glutathione peroxidases 1 and 3) and spares the housekeeping selenoproteins [50]. The functional gene has 87 nucleotides, without introns, and there is a pseudogene on chromosome 22 [15]. In the heterozygous father of the first patient, tRNA-Sec was reduced by 40% (74% normal and 26% mutant), and in the patient by 70%, with a fivefold fall in the mcm5Um isoform, on which the synthesis of the stress-related selenoproteins depends [15]. In CRISPR-edited cells carrying C65G, UGA recoding was reduced and selenoprotein production no longer responded to selenium [66].

SEPSECS variants cause progressive cerebello-cerebral atrophy with microcephaly, with normal circulating T4 and selenium, and defects of individual selenoproteins (SELENON/SEPN1, GPX4, TXNRD2) cause myopathy, lethal skeletal dysplasia or endocrine abnormalities, without the thyroid profile of selenoprotein deficiency [2, 64].

2.8. Deiodinases: DIO1, DIO2 and DIO3

D1 is a membrane selenoenzyme that deiodinates the outer ring of T4, generating T3, and clears rT3; to a lesser degree it deiodinates the inner ring, inactivating T4 and T3; it contributes substantially to circulating T3, and its activity decreases adaptively during fasting and acute illness [2, 16]. The gene has 4 exons, and the enzyme, which is propylthiouracil-sensitive, predominates in liver and kidney [17]. The two variants described by França and colleagues, p.(Asn94Lys) and p.(Met201Ile), increase the enzyme's Km two- to threefold (lower affinity) and reduce the reaction rate; heterozygous null Dio1 mice reproduce the profile, which indicates haploinsufficiency [6]. In a family of Latin origin with congenital hypothyroidism due to compound heterozygous TPO variants, the two sisters with an additional DIO1 variant, published as NM_001039715.3:c.395G>A p.(Arg132His) and corresponding in the MANE transcript to NM_000792.7:c.539G>A p.(Arg180His), had more severe developmental delay than their brother without the DIO1 variant and disproportionately high FT4 relative to T3 on levothyroxine, compatible with reduced generation of T3 from administered T4; here DIO1 acts as a modifier [67]. OMIM places this association among those awaiting confirmation: the variant, which it designates R132H, was inherited from the father, heterozygous for TPO, and was also present in a sister without hypothyroidism, and all the siblings had raised rT3, but only the four carriers, the father included, had a raised rT3/T3 ratio [17, 27, 67]. The MANE notation was obtained in VariantValidator from the transcript given in the article (NC_000001.11:g.53906152G>A); the identifier rs1450006861 that the article and OMIM associate with the variant corresponds, in dbSNP, to a position on chromosome 2, so it does not identify this variant [9, 67]. The common DIO1 polymorphism rs2235544 is associated with the FT3/FT4 ratio in the general population, without an effect on TSH, and does not constitute disease [68]; similarly, the 3' untranslated region polymorphisms D1a-C/T and D1b-A/G are associated, in opposite directions, with rT3 and the T3/rT3 ratio in the general population [17].

The search for human deiodinase defects intensified after the knockout mice for each of the three enzymes; until 2021, the only human genetic defects affecting the deiodinases were those of SECISBP2 and TRU-TCA1-1 [2]. No human Mendelian disease has been described for DIO2 or DIO3 [2, 69, 70]. D3 inactivates T4 to rT3 and T3 to T2, is highly expressed in the pregnant uterus, the placenta and fetal tissues, and its increased activity in infantile haemangiomas causes severe consumptive hypothyroidism, an acquired condition [70].

2.9. Normalised variants

THRB variants in the literature use the numbering of the TRβ1 isoform, which coincides with that of the MANE transcript NM_001354712.2 for the variants in Table 2.1; TRβ2 numbering is different and must not be mixed [9]. In albumin, the historical numbering counts the mature protein, whereas HGVS nomenclature includes the 24-amino-acid pre-pro-peptide: the historical variant "R218H" of familial dysalbuminaemic hyperthyroxinaemia is NM_000477.7:c.725G>A p.(Arg242His) [9, 71].

Table 2.1. Variants cited in the article, normalised in VariantValidator (GRCh38, MANE Select transcripts) or in ClinVar.

The p.(Gly432del) variant was published as c.1294_1296del and VariantValidator corrects it to c.1295_1297del by the 3' alignment rule; the two descriptions designate the same deletion [9].

3. Clinical Spectrum and Manifestations by System

3.1. RTHβ: sources of natural history data

The natural history of RTHβ rests on series from reference centres and on record-linkage cohorts. The NIH prospective controlled study evaluated 104 patients from 42 families and 114 unaffected relatives; inheritance was dominant in 22 families, sporadic in 14 and unknown in 6, and variants were identified in 25 families (64 patients), 16 in exon 9 and 9 in exon 10, the only exons analysed [55]. In the Welsh SAIL Databank cohort (United Kingdom), 61 genetically confirmed patients, followed between 1997 and 2019, were compared with 2750 age- and sex-matched controls [39]. The Italian cohort assembled 284 patients diagnosed between 1984 and 2023, with longitudinal data on 249 heterozygotes [40]. Recent paediatric series come from Turkey (13 and 25 children), and a multicentre series from Spain assembled 28 patients [35, 37, 71]. These are specialist clinic populations, biased towards symptomatic forms, and do not estimate the frequency of manifestations in the population with undiagnosed RTHβ [39].

Table 3.1. Frequency of RTHβ manifestations in the main series.

The percentages in the series are not comparable with one another: they differ in age (adults and children at the NIH and in Italy, children only in Turkey), in ascertainment (prospective assessment at the NIH, retrospective records in the others) and in the definition of each manifestation.

3.2. RTHβ: thyroid

Goitre is frequent and tends to progress to nodular disease, possibly because of the greater bioactivity of circulating TSH [2]. Rare carcinomas have been described, mostly papillary microcarcinomas and one metastatic case, with a favourable course despite incomplete TSH suppression after ablation [2]. The risk of thyroid autoimmunity is increased [2]. In a family with RTHβ and coexisting autoimmune thyroiditis, a 45-year-old woman was treated for years as hyperthyroid before diagnosis, with positive anti-thyroid peroxidase (anti-TPO) and anti-thyroglobulin antibodies and negative TSH receptor antibodies (TRAb) [81].

The 2024 French National Diagnosis and Care Protocol (PNDS), written by the reference centre for rare diseases of the thyroid and hormone receptors (CRMR-TRH, Angers) and published by the Haute Autorité de Santé (HAS), places the frequency of goitre between 66% and 95% of patients and notes that recurrence after incomplete thyroidectomy is not rare [41]. In children, goitre, sometimes large, appears more frequent when the mother also has RTHβ than when the mother is unaffected (86% versus 35%), and a large goitre in a child, especially with nodules, should prompt measurement of TSH and TH [41]. The volume of a simple or multinodular goitre can cause compression, with dyspnoea, swallowing disorder, dysphonia and superior vena cava syndrome, exceptionally in the newborn and child, and these large goitres sometimes lead to thyroidectomy before the diagnosis is made [41]. The carcinomas of follicular origin described are mostly microcarcinomas, discovered before or after the diagnosis of RTHβ; RTHβ does not appear to worsen the prognosis, it has not been shown to increase cancer risk, and the role of chronic TSH stimulation in carcinogenesis remains unanswered [41]. Thyroid autoimmunity is up to 2 to 4 times more frequent than in the general population, especially in men; the proposed mechanisms are chronic TSH stimulation of the lymphocytes infiltrating the gland and activation of thymic immunity, dendritic cells and the cytotoxic and pro-inflammatory response [41]. The association of RTHβ with congenital hypothyroidism is very rare: five genetically confirmed cases had been described, mostly due to sublingual thyroid ectopy, with raised neonatal TSH and high or low FT4, and the diagnosis of the second disease was generally late, suggested by difficulty in normalising TSH despite normal FT4 [41].

3.3. RTHβ: cardiovascular system and mortality

Tachycardia and tachyarrhythmias, atrial fibrillation and heart failure are mediated by myocardial TRα, which is sensitive to the raised hormones [2]. The Welsh cohort was the first to show reduced survival: all-cause mortality with a hazard ratio (HR) of 2.84 (95% confidence interval [95% CI] 1.59–5.08), atrial fibrillation HR 10.56 (4.72–23.63), heart failure HR 6.35 (2.26–17.86) and major adverse cardiovascular events (MACE: cardiovascular death, acute myocardial infarction, heart failure or stroke) HR 3.49 (2.04–5.99) [39]. The first adverse event occurred at a median age of 56 years (95% CI 44–65), against 67 years (65–70) in controls, 11 years earlier; FT4 at diagnosis was positively associated with mortality and MACE, with increased risk from 30 pmol/L [39]. Antithyroid drugs, surgery, radioiodine or thyroxine did not control the hormone excess compared with no intervention [39].

The Italian cohort confirmed these findings in another country: MACE at a median age of 59.4 years (interquartile range 50.4–66.4) and premature mortality equivalent to a mean of 11 years of life lost; an FT4 57% above the upper limit of normal was associated with premature cardiovascular manifestations; on multivariate analysis, only age at diagnosis, raised FT4 and male sex were associated with MACE, and age at diagnosis and FT4 with mortality; previous thyroidectomy or radioiodine did not alter MACE or mortality [40]. It is not known whether the events result from atherosclerosis, from direct thyrotoxic cardiomyopathy or from both [2].

The French PNDS describes resting or exertional tachycardia and palpitations in one third to three quarters of patients, when sought in the history, sometimes since childhood and sometimes as the only complaint, and notes raised blood pressure [41]. Atrial fibrillation was historically described in about 6% of patients, and the French reference centre database found a prevalence of 20% [41]. Echocardiographic studies, compared with healthy individuals and with hyperthyroid patients, are contradictory regarding systolic and diastolic function and left ventricular mass; peripheral vascular resistance is altered (increased in the text of the protocol, decreased in the attached clinical table), without an increase in carotid intima-media thickness, and the abnormalities as a whole are less frequent and less severe than in primary hyperthyroidism, with no correlation between cardiovascular phenotype and genotype [41]. Despite the atherogenic metabolic profile, no increase in the risk of thrombotic events has been demonstrated; the protocol regards atrial flutter and fibrillation as the determinants of severity, allows for primary heart failure or failure secondary to the arrhythmia, and asks for particular cardiac surveillance in elderly patients, in those with heart disease of another cause and in those with FT4 above 30 pmol/L at diagnosis [41].

3.4. RTHβ: metabolism, bone and growth

Hepatic resistance, mediated by TRβ, explains the normal SHBG, the mixed dyslipidaemia (raised cholesterol and triglycerides) and the increase in liver fat, with a risk of metabolic dysfunction-associated steatotic liver disease (MASLD) [2]. In a cross-sectional study of 105 adults with confirmed RTHβ (mean age 40 years), age was associated with total cholesterol and LDL, BMI with the total cholesterol/HDL ratio and the HOMA-IR insulin resistance index with triglycerides; the only thyroid measure associated was rT3, inversely associated with triglycerides, and the dyslipidaemia appeared to depend on the same factors as in the general population and not on thyroid status [82]. In the first 77 patients of this sample, compared with controls matched for age, sex and BMI, LDL was only 0.3 mmol/L higher, with overlapping absolute ranges (1.2–5.4 versus 1.5–5.0 mmol/L); 75.5% had HOMA-IR above 1 and almost a quarter above 2.5, non-esterified fatty acids were not associated with triglycerides, and the models explained less than a quarter of the variability of each lipid parameter [82]. The authors propose managing insulin resistance mainly with lifestyle measures and deciding lipid-lowering therapy on conventional individual risk, and note that TRIAC, an agonist with hepatic effect, could lower LDL and triglycerides at the cost of a lower HDL [82]. In children, poor weight gain and low BMI are described; in adults, osteopenia or osteoporosis; in biallelic cases, stippled epiphyses [2]. Unlike conventional thyrotoxicosis, bone turnover markers are usually normal [2]. Saudubray includes familial resistance to thyroid hormone among the causes of stippled epiphyseal calcifications (chondrodysplasia punctata) in inborn errors [83].

The bone data in the PNDS come mainly from a cross-sectional study from Ribeirão Preto (Brazil), with 14 patients with RTHβ (8 adults and 6 children) and 24 controls: total calcium was higher in affected adults and children, with normal calciuria; only the children had lower inorganic phosphate and higher fibroblast growth factor 23 (FGF23); the adults had lower bone mineral density in the whole body and lumbar spine; and parathyroid hormone, 25-hydroxyvitamin D, alkaline phosphatase, osteocalcin and C-terminal telopeptide did not differ from controls [41, 84]. The authors attribute the pattern to the predominance of TRα1 in the skeleton, with a possible increase in calcium flux from bone to the circulation, and fracture risk remains to be studied [84]. In children, both growth delay with delayed bone maturation and advanced bone maturation are described, as are low weight with low BMI and thinness; adults with RTHβ are not significantly shorter than controls [41]. Peripheral markers of hormone action (cholesterol, creatine kinase, SHBG, angiotensin-converting enzyme, ferritin, osteocalcin, pro-BNP) are often normal, unlike in hyperthyroidism; a recent series described, besides higher LDL and triglycerides and lower HDL, increased HOMA-IR and accumulation of intrahepatic and intramuscular lipids on MRI [41].

3.5. RTHβ: neurocognition, hearing and vision

The neurocognitive manifestations include anxiety, sleep disturbance, ADHD, mild intellectual impairment (lower non-verbal IQ), language difficulties and poorer school performance; severe intellectual disability is a feature of homozygotes [2]. In a study of 18 families (49 affected and 55 unaffected) with structured psychiatric interviews by interviewers blind to clinical status, 50% of affected adults (11 of 22) had had ADHD in childhood, against 7% (2 of 30) of unaffected adults, and 70% of affected children (19 of 27) had ADHD, against 20% (5 of 25) (P <0.001 in both comparisons); the likelihood of ADHD was 3.2 times greater in affected males than in affected females [85]. Hearing loss results from a conductive deficit, linked to frequent otitis in childhood, and from cochlear dysfunction [2]. Colour perception is altered in heterozygotes when sought, and colour vision is frankly abnormal only in rare homozygotes [2, 60].

The French PNDS places ADHD in 40% to 70% of children, draws attention to isolated moderate hyperactivity and considers intellectual disability with IQ <70 very rare, although IQ <85 has been described; the proposed mechanisms are abnormalities of axons, of neuronal proliferation and migration and of catecholaminergic neurotransmission, and manifestations such as tremor, nervousness, anxiety, sleep disorders, mood disorders and autistic traits are difficult to attribute to the disease [41]. Hearing disorders affect 21% to 24% of adults and 19% of children, ear, nose and throat infections up to 60% of children, and conductive and sensorineural hearing loss coexist; in homozygous children a hearing aid was needed before age 1, whereas severe hearing impairment is rare when diagnosis is made in adulthood [41]. The warning signs are a poor reaction to noise and delayed language in the infant and, later, difficulty in noisy environments and comprehension lapses noticed by those close to the patient; hearing loss can worsen behavioural disorders and should be screened for systematically in children [41]. In adults without alteration of corrected visual acuity, a comparative study found lower chromatic sensitivity and reduced responses of rods and L and M cones, with normal S cones, independently of sex and hormones; in children, a reduction of L and M cones with an increased S-cone response was described, attributed to the role of the receptors in photoreceptor differentiation; these changes are subclinical, and no major visual involvement has been described even in biallelic patients [41].

3.6. RTHβ: severe forms (homozygosity and extreme mosaicism)

In homozygosity for missense variants, the features of the heterozygous form (goitre, non-suppressed TSH, upper respiratory infections, hyperactivity, low BMI) are joined by dysmorphic facies, winged scapulae and much higher hormone levels; in a 9-year-old boy homozygous for p.(Arg243Gln), the elder brother, with a similar picture and probable homozygous RTHβ, died of heart failure at age 13 [57]. Thyrotoxic cardiomyopathy is the main threat in these forms [2, 57]. A boy with a mosaic frameshift variant had, from 2 months of age, agitation, excessive crying and reflux, epileptic seizures from 1.5 years and, at 8 years, severe developmental delay, refractory epilepsy, tachycardia, chronic diarrhoea and low weight, with FT4 of 132.4 pmol/L (Section 10) [53].

Patients without a normal THRB allele show the effect of mutant receptor dose (Table 3.2) [19]. In family G, the original 1967 family, the three siblings homozygous for the deletion had stippled epiphyses, bird-like facies, pectus carinatum, winged scapulae, deaf-mutism and dyschromatopsia, but reached adult height above the parental mean and had normal IQ compared with deaf children; in family S, the patient homozygous for the threonine 337 deletion had at birth respiratory distress, hyperbilirubinaemia, hypofibrinogenaemia, polycythaemia and very high T4 and TSH, then goitre, cardiac malformation, non-communicating hydrocephalus, epilepsy and profound intellectual disability, with no response to beta-blockers, propylthiouracil or D-thyroxine, and died at age 8 of cardiogenic shock during staphylococcal pneumonia [19]. In the two families described in 2012, a girl homozygous for p.(Gly347Glu), the child of a consanguineous union, with very high TSH on newborn screening, mixed hearing loss from the first month, poor weight gain and IQ between 61 and 72, and two Turkish siblings of consanguineous parents, homozygous for p.(Arg316Cys), with bird-like facies, hearing loss requiring hearing aids, markedly delayed bone age and full-scale IQ of 50 and 46, had mean FT4 of 488% and 299% of the upper limit of normal and TSH of 490 and 35 mU/L, against 173% and 113% and TSH of 4.8 and 2.6 mU/L in heterozygous relatives [19]. Effects on growth, bone, cognitive development and hearing, uncommon in heterozygous forms, were present in all patients of the four families with mutant receptor in double or single dose [19].

Table 3.2. Patients without a normal THRB allele in the Ferrara 2012 review.

Source: Tables 1 and 2 of Ferrara 2012 [19].

3.7. RTHβ: pregnancy and the fetus

Women with RTHβ have more first-trimester miscarriages, and unaffected newborns of affected mothers are small for gestational age (SGA), with low birth weight and suppressed TSH, due to exposure to excess maternal hormones to which the normal fetus is sensitive [2, 72]. The figures and the implications for counselling are in Section 8. The French PNDS stresses that the excess of miscarriages has been described inconsistently, that data on fetal growth diverge, that there are no data on the risk of prematurity and that the genetic status of mother and fetus may modulate these outcomes; the rise in maternal TH would be greatest between 8 and 16 weeks of gestation and greater when the fetus is also affected [41].

3.8. RTHα: manifestations by system and by age

RTHα is not detected by TSH-based newborn screening, because TSH is not raised, and it manifests in early childhood [86]. The phenotype is highly variable and the spectrum is expanding [2]. The features in Table 3.3 summarise the table of manifestations in the European guideline.

Table 3.3. RTHα: manifestations and investigation by system (2024 ETA guideline).

Source: Table 3 of the 2024 ETA guideline [2].

In the child, the picture combines growth delay with short limbs and delayed bone age, skeletal dysplasia, broad or round face, severe constipation, motor dyspraxia, slow speech and developmental delay or variable intellectual disability [10, 86]. HR and BP are low for age and BMI is raised, with reduced basal metabolic rate [86]. In the untreated adult, fatigue, constipation, low mood and abdominal adiposity predominate: in four never-treated first-degree adult relatives with the p.(Ala263Val) variant, there had been in childhood short stature with short limbs, metacarpals and phalanges, delayed bone age and motor and language delay, and in adulthood a round face, increased head circumference, low haemoglobin and a high LDL/HDL ratio, with a more marked phenotype than in patients treated since childhood; diagnosis was delayed because TSH and FT4 were within the reference range [38]. In three Brazilian children from unrelated families, all with low FT4, mild constipation and short stature, one had hypoactivity and mildly raised TSH, another mild developmental delay interpreted as central hypothyroidism and the third a severe picture with hypotonia, hypercholesterolaemia, severe hypertriglyceridaemia, high amylase and mild pericardial effusion [87]. In other descriptions gathered in OMIM, a girl with a 1-bp insertion in THRA, inherited from her father with a similar picture, had in her first years macroglossia, omphalocele, congenital hip dislocation, absent hip ossification centres, delayed suture closure and macrocephaly, with high cholesterol and IQ of 90, and growth hormone had little effect; in a Polish series of 5 children, the picture included post-term delivery with relatively high birth weight and length, hypotonia, short limbs, hands and feet with a long thorax, deep voice, coarse facies with hypertelorism, tortuous arteries on the dorsum of the hands and feet, clubfoot, anaemia, mildly raised CK and cholesterol and, in the 2 most affected patients, mild cardiomyopathy, with no improvement on levothyroxine [23]. In the Norwegian family, the four affected members also had numerous naevi, and the mother recognised at birth in her youngest daughter the facial and bodily features of the father and siblings; the variant was identified when the index case was 18, on a panel of 849 intellectual disability genes, after 5 years of negative genetic investigation (microarray, NSD1, SHOX, PTEN) [38]. SHBG is often raised in the guideline series and was described as markedly raised in the Oxford Desk Reference, whereas in the two adolescents of Moran 2017 it was within the reference range before treatment [2, 22, 86].

3.9. Selenoprotein deficiency (SECISBP2)

The typical presentation is in childhood, with delayed growth and bone age, although one boy had normal growth and bone age [12]. Developmental delay accompanies short stature in most cases, by a mechanism not fully understood [2]. The major phenotype is weakness and hypotonia due to progressive degeneration of specific muscles (sartorius, adductors, axial paravertebral muscles), similar to SELENON-deficiency dystrophy, which can cause hypoventilation [2]. The other manifestations include azoospermia due to failure of the late stages of spermatogenesis, cutaneous photosensitivity with increased ROS and susceptibility to ultraviolet oxidative damage, Raynaud's phenomenon, progressive sensorineural hearing loss, deficient T-lymphocyte proliferation with abnormal cytokine secretion and telomere shortening, increased fat mass and, paradoxically, greater systemic and cellular insulin sensitivity [2, 5]. Selenoprotein deficiency predisposes to aneurysmal dilatation of the thoracic aorta [2, 88].

Table 3.4. Selenoprotein deficiency due to SECISBP2: manifestations and investigation (2024 ETA guideline).

Source: Table 5 of the 2024 ETA guideline [2].

The cases published in 2026 illustrate the variability: a 3-year-old Korean girl, compound heterozygous for p.(Arg120Ter) and p.(Arg672Ter), with short stature, global delay, bilateral partial ptosis and congenital sensorineural hearing loss [77]; and a girl aged 13 years and 6 months, homozygous for p.(Arg120Ter), assessed for constipation and abnormal thyroid function tests, with speech and language delay and learning difficulties, FT4 25.5 ng/L, FT3 2.3 ng/L (low), TSH 1.2 mIU/L and serum selenium of 14.37 µg/L [76]. In two families with four new compound heterozygous variants, both probands had growth and developmental delay: an Argentinian girl with normal newborn screening, poor weight gain at one year of age (weight −3.0 standard deviations), walking at 18 months and, at 5.4 years, height of −2.7 standard deviations with a bone age of 2 years; and a Turkish girl walking and talking at 2.5 years with height between the 3rd and 10th centiles at 11 years, without photosensitivity, hypoglycaemia, recurrent infections or hearing loss [36]. Serum glutathione peroxidase activity was 44 and 15 mU/mL, against means of 124 and 126 mU/mL in unaffected relatives, and the T3/T4 ratio was 5.9 and 3.5, against 19.2 and 15.5, which makes this ratio informative even with T3 at the lower limit of normal [36]. In the original families, a Saudi Bedouin boy with short stature and delayed bone age at 12 years normalised his height with the pubertal growth spurt, without treatment of the raised neonatal TSH, because T4 was not low; a Turkish boy homozygous for c.800_801insA, assessed at 10 years for obesity, had right ptosis since infancy, proximal lower limb weakness with a myopathic pattern on electromyography and fatty muscle infiltration on MRI, attention deficit and an insufficient growth hormone response to stimulation, and normalisation of selenium by supplementation did not correct the thyroid function tests [13, 24]. Most known patients are children or young adults, and it is not known whether chronic oxidative damage predisposes to neurodegeneration, premature ageing or neoplasia [2].

3.10. tRNA-Sec defect (TRU-TCA1-1)

The first patient, an 8-year-old boy, the son of consanguineous parents and homozygous for C65G, had abdominal pain, fatigue, muscle weakness and low plasma selenium, with raised FT4 and rT3 and normal FT3 and TSH (euthyroid hyperthyroxinaemia), undetectable plasma and erythrocyte glutathione peroxidase, mild signal changes on muscle MRI and increased endogenous hydrogen peroxide in fibroblasts; the heterozygous parents and a brother had normal thyroid function tests, selenium and selenoproteins [25, 26, 50]. The second, a 13-year-old boy followed for Hashimoto's thyroiditis, initially had high FT3 (4.6 pg/mL; reference 2–4.2) with normal FT4 and TSH; over 6 years of untreated follow-up FT3 fell, FT4 rose and TSH remained normal, with increased rT3 at 19 years and 4 months and low total selenium (91 µg/L; reference 95–125), without complaints [26]. The heterozygous father had normal tests, except for moderately increased rT3, and increased total selenium (143 µg/L) [26]. Polyethylene glycol precipitation excluded anti-T3 antibodies as the cause of the initial high FT3, and the thyroid was hypertrophic and homogeneous on ultrasound [26]. The patient was homozygous through paternal uniparental isodisomy of the 19q13.32 region, demonstrated with polymorphic markers, without consanguinity and with the mother not carrying the variant [15, 26]. Six months after the low measurement, total selenium was at the upper limit (128 µg/L) and serum selenium was normal on two measurements; glutathione peroxidase 3 was at the lower limit of normal in the patient and the father (168 and 180 IU/L; reference 150–558) and superoxide dismutase was low in both (1.2 and 1.8 U/mL), with normal immunoglobulins and lymphocyte immunophenotyping [26]. The authors attribute the fall in FT3 and rise in FT4 partly to the normal evolution with age, keeping the high FT4 as an expression of the defect, and conclude that the variant does not, by itself, disrupt all selenoprotein synthesis [26]. With two patients, the clinical spectrum is not defined [2].

3.11. D1 deficiency (DIO1)

D1 deficiency was identified through abnormal thyroid function tests and is not accompanied by described clinical manifestations: raised serum rT3 and rT3/T3 ratio, normal or mildly high FT4, normal T3 and, in one family, raised total cholesterol [2, 6]. In the first family, of Hispanic origin, the 3-year-old proband had Down syndrome with mildly raised TSH and anti-peroxidase antibodies, and the variant and high rT3 segregated in five relatives on the maternal side; in the second, Jewish, the proband presented at 15 years with TSH resistance, and the affected mother had very high cholesterol [27]. Loss of D1 can modify the thyroid function tests associated with other thyroid defects, as in TPO congenital hypothyroidism in which the DIO1 variant worsened the phenotype on levothyroxine [2, 67].

4. Clinical Recognition and Red Flags

4.1. When to suspect RTHβ

The situation that most often leads to diagnosis is a discordant laboratory result, raised FT4 (or total T4) and FT3 with normal or high TSH, instead of the suppressed TSH of primary thyrotoxicosis [2]. In Spain, 46.4% of the 28 patients were diagnosed by family study, 35.7% by tests requested for unrelated symptoms and only 17.8% on clinical grounds (goitre, hypertension, tachycardia) [35]. In a Turkish paediatric series, the reasons were abnormal thyroid function tests (53.8%), family screening (23.1%), raised TSH on newborn screening (15.4%) and short stature (7.7%), and one child was being treated with levothyroxine for a misdiagnosis of hypothyroidism [37].

The European guideline adds two contexts. In treated autoimmune or congenital hypothyroidism, failure of levothyroxine, even at a supraphysiological dose, to normalise TSH suggests RTHβ, but only when accompanied by raised FT4 [2]. In coexisting hyperthyroidism (Graves' disease, thyroiditis, amiodarone), RTHβ may be masked, and suspicion arises when antithyroid drugs or other treatment cause an exaggerated rise in TSH with normal hormones [2]. The finding of a pituitary macroadenoma in a patient with central hyperthyroidism does not exclude RTHβ: in a 35-year-old man, genetic testing identified p.(Val458Ala) and avoided pituitary surgery for a coincidental incidentaloma [75].

The French PNDS adds two circumstances. The first is the screening strategy: the recommendations of the French Society of Endocrinology and of the HAS advocate measuring TSH alone when thyroid dysfunction is suspected, which detects neither RTHβ nor central hypothyroidism; only the combined request of TSH with FT4 and/or FT3 reveals the profile [41]. The second is the difficulty, in a patient thyroidectomised for any reason, in normalising TSH without high-dose levothyroxine and without raised FT4 [41]. In children, the symptoms that lead to testing are frank tachycardia, insomnia, diarrhoea and heat intolerance, goitre and neuropsychological disorders (ADHD, isolated hyperactivity, learning, language and mood disorders), and family study reveals many pauci- or asymptomatic children [41].

4.2. When to suspect RTHα, selenoprotein deficiency and D1 deficiency

RTHα should be considered in a child with growth delay, macrocephaly, delayed tooth eruption, constipation and developmental delay, especially if thyroid function tests show low-normal or low FT4 with high-normal or high FT3 and normal TSH [2, 86]. The profile is often interpreted as central hypothyroidism, as in the second adolescent of Moran 2017 and in one of the Brazilian children of Jorge 2025 [22, 87]. Selenoprotein deficiency should be considered in a child with delayed growth and bone age, high FT4, low or normal FT3 and high rT3, especially with axial weakness, hearing loss, photosensitivity or male infertility [2]. D1 deficiency is a hypothesis when rT3 and the rT3/T3 ratio are raised outside non-thyroidal illness or fasting [2].

4.3. Red flags by age

Table 4.1. Red flags by age and context.

4.4. Emergency signs

The situations requiring immediate assessment are thyrotoxic cardiomyopathy or heart failure in the homozygous or extreme mosaic forms, atrial fibrillation, which may not respond to cardioversion or ablation, and fetal tachycardia or growth restriction in the pregnancy of a woman with RTHβ [2, 57]. A frequent error with irreversible consequences is treating RTHβ as primary hyperthyroidism with antithyroid drugs, radioiodine or thyroidectomy: ablation did not reduce MACE or mortality in the Welsh and Italian cohorts and necessitates supraphysiological doses of levothyroxine to avoid thyrotroph hyperplasia [39, 40, 89]. Familial dysalbuminaemic hyperthyroxinaemia, more frequent than RTHβ in paediatric series, is the other main cause of inappropriate treatment of the same profile [71, 80]. The French PNDS adds compressive goitre, with dyspnoea, dysphagia, dysphonia or superior vena cava syndrome, and asks that antithyroid drugs, amiodarone and any radical thyroid therapy (thyroidectomy or radioiodine) not be started without the opinion of the reference centre or a competence centre; the most frequent error that the protocol seeks to avoid is thyroidectomy for presumed hyperthyroidism [41].

4.5. Newborn screening

Neonatal detection of RTHβ depends on programmes that screen for T4: in Oregon, total T4 screening identified 2 cases compatible with RTH in 80,884 newborns, and in the Basque Country, where TSH and T4 are measured simultaneously, 4 newborns were diagnosed by screening [34, 35]. Programmes based on TSH alone detect neither RTHβ nor RTHα, and in Korea TSH-based screening played a limited role in the diagnosis of RTH and MCT8 deficiency [34, 86, 90]. A raised TSH on screening can, however, lead to the diagnosis of RTHβ, as in 2 of 13 Turkish children [37]. The European guideline lists the benefit of neonatal diagnosis and early intervention as an area of uncertainty [2].

In France, newborn screening for congenital hypothyroidism measures only TSH in whole blood collected between 48 and 96 hours, so RTHβ is rarely diagnosed in this period [41]. RTHβ can, however, cause an excessive TSH peak, with neonatal values of 13 to 405 mU/L described, generally in symptomatic newborns with severe forms, and more rarely a modest and transient rise; congenital hypothyroidism may then be diagnosed in error and levothyroxine started, and lack of improvement or the appearance of thyrotoxicosis with very high FT4 and FT3 redirects the diagnosis [41]. Between screening and the confirmatory sample TSH may have normalised, and TH are physiologically higher in the first weeks, which makes the neonatal profile difficult to interpret; symptomatic forms in the newborn manifest as thyrotoxicosis (tachycardia, excessive sweating, even respiratory distress), goitre, prolonged jaundice, early poor weight gain or deafness, and may correspond to biallelic genotypes [41].

5. Diagnosis: Criteria, Testing, and Algorithms

5.1. The European guideline algorithm for raised TH with non-suppressed TSH

The 2024 ETA guideline proposes a three-step approach to raised thyroid hormones (T4 and/or T3) with non-suppressed TSH [2]. In the first step, confounding factors are excluded, such as intercurrent physical or psychiatric illness and drugs (for example, amiodarone or thyroxine), and the clinical thyroid status is determined (strong recommendation; low quality of evidence) [2]. In the second, analytical interference is investigated with the laboratory before screening for rare genetic or acquired disorders (strong recommendation; moderate quality) [2]. In the third, once genuine hyperthyroxinaemia is confirmed, FT3 with the T4/T3 ratio, rT3, autoantibodies and SHBG are used to guide the diagnosis (Table 5.2), and RTHβ is distinguished from TSH-secreting pituitary adenoma (thyrotropinoma or TSHoma) by biochemistry, genetics, dynamic tests (thyrotropin-releasing hormone [TRH] stimulation, if available; liothyronine suppression), pituitary imaging (MRI, positron emission tomography) and a therapeutic trial of a long-acting somatostatin receptor ligand (strong recommendation; low quality) [2]. The guideline uses the modified GRADE classification: strong (S) or weak (W) strength and quality of evidence in four levels, from high (ØØØØ) to very low (ØOOO) [2].

5.2. Analytical interference

Antibodies that interfere in immunoassays may be analyte-specific (macro-TSH, anti-iodothyronine antibodies) or directed against the reagents (heterophile antibodies and anti-animal immunoglobulins) [2]. Comparison between platforms is strongly suggestive, and confirmation is by serial dilutions, polyethylene glycol (PEG) precipitation or gel filtration chromatography for TSH, and by equilibrium dialysis or ultrafiltration for T4 and T3; the clinical picture guides the investigation, and in a patient with signs of hyperthyroidism the first suspicion falls on the TSH measurement [2]. The 2026 ETA guideline on interference in thyroid function immunoassays classifies interference as acquired (antibodies), genetic (variant binding proteins, as in familial dysalbuminaemic hyperthyroxinaemia and in transthyretin and thyroxine-binding globulin variants) and pharmacological (binding displacers and biotin), and requires that it be excluded before diagnosing RTH or TSHoma, with liaison between clinician and laboratory and referral to a regional centre with specialised tests in a subgroup of patients [91]. The levothyroxine label recommends stopping biotin at least 2 days before thyroid testing [92]. During TRIAC treatment, T3 measured by immunoassay is falsely raised by cross-reactivity, and only liquid chromatography with tandem mass spectrometry (LC-MS/MS) measures it accurately [2, 93].

The 2026 guideline starts from a review in which, in more than 150 published cases of interference, at least half of the patients suffered negative clinical consequences and only 8% none [91]. For FT4, one-step competitive assays are more susceptible than the reference methods (equilibrium dialysis, ultrafiltration), and inter-method variability is greater when FT4 is high; anti-iodothyronine antibodies increase FT4 in one-step assays, through binding to the tracer, and affect two-step assays less [91]. Non-genetic causes include heparin, which activates lipoprotein lipase and generates ex vivo free fatty acids that displace T4, very high endogenous fatty acid concentrations, high doses of phenytoin and furosemide, chloral hydrate and biotin, which in competitive streptavidin-biotin assays falsely raises FT4 and FT3 [91]. For TSH, heterophile antibodies and human anti-animal antibodies can give falsely high or low values; macro-TSH, a complex of TSH with autoantibodies, is biologically inactive because it is confined to the intravascular compartment; and a single-nucleotide variant in TSHB produces a bioactive TSH that some assays do not recognise, with falsely low values [91]. The guideline estimates albumin variants at up to 1.8% of the population, transthyretinaemic hyperthyroxinaemia at about 1:6300 and TBG variants at about 1:5000; transthyretin interference depends on the assay and may vary in the same individual with intercurrent illness, and TBG deficiency can occasionally falsely raise FT4 [91]. Albumin variants at codons 218 and 222 (mature protein numbering) also falsely raise FT3 and total T3 in many immunoassays, including a two-step one, and the therapeutic analogues TRIAC and diiodothyropropionic acid (DITPA) have variable cross-reactivity in T3 assays [91]. The International Federation of Clinical Chemistry (IFCC) reference methods for FT3 (equilibrium dialysis with isotope dilution mass spectrometry) and for total T3 (isotope dilution LC-MS/MS) are not available in routine practice [91].

Table 5.1. Recommendations of the 2026 ETA guideline on interference most relevant to the differential diagnosis of these disorders.

Source: Table 1 of the 2026 ETA guideline, which uses the same modified GRADE classification as the 2024 guideline [91].

5.3. Biochemical signatures

Table 5.2. Biochemical signature of the genetic disorders of TH transport, metabolism and action.

Source: Table 1 of the 2024 ETA guideline, transcribed as published [2]. The "high" TSH in D1 deficiency corresponds to the mildly increased TSH described in the families; the marker of D1 deficiency is rT3 with a raised rT3/T3 ratio [2, 16]. In selenoprotein deficiency, Table 1 of Refetoff 2014 describes TSH as normal or mildly raised [1].

5.4. RTHβ: biochemical confirmation and distinction from TSHoma

The guideline's diagnostic recommendations for RTHβ are in Table 5.3. SHBG is normal in RTHβ, because the liver is resistant, and tests in first-degree relatives are the simplest discriminating test: abnormal tests in a relative increase the likelihood of RTHβ, whereas TSHoma is sporadic [2, 7]. Distinguishing RTHβ from TSHoma requires a multimodal approach; the reference guideline for TSHoma remains the 2013 ETA guideline [2, 94]. A coincidental macroadenoma can coexist with RTHβ and does not prove a TSHoma [75].

Table 5.3. Recommendations of the 2024 ETA guideline for the diagnosis of RTHβ.

Source: Box 2 of the 2024 ETA guideline [2].

5.5. RTHβ: genetic testing and its limits

Sanger sequencing of THRB identifies the heterozygous variant in most suspected cases [2]. The proportion of families with an RTH phenotype in which a THRB variant is found is 85% to 90% (diagnostic yield of sequencing in the reference series), which is not equivalent to the clinical sensitivity of an individual test [2, 18, 58]. A negative result does not exclude RTH: low-fraction somatic mosaicism escapes Sanger sequencing in blood, as in the patient with a 12% allele frequency, and NGS with greater read depth, in blood and other tissues, may allow the diagnosis [2, 58]. The historical series analysed only exons 9 and 10 or 7 to 10, the region where variants cluster, and variants outside these hotspots, non-coding THRB variants and the unknown genes of non-TR-RTH are areas of uncertainty listed by the guideline [2, 55, 56]. In resource-limited settings, Sanger sequencing may not be available [2]. For a novel THRB variant, the main evidence is cosegregation with the abnormal tests in the family and, when this is impossible, the in vivo T3 suppression test and in vitro functional studies (Table 5.3) [2].

5.6. RTHα: criteria for sequencing THRA

The RTHα spectrum is expanding, and the criteria for sequencing THRA are proposals that may evolve; a definitive diagnosis requires a pathogenic variant in THRA, whose sequencing is not available in all countries, and cases with an RTHα phenotype without an identified variant have been described [2]. Clinical assessment includes head circumference, height, sitting height and subischial leg length, compared with age references (strong; low), and investigations include full blood count, T4, T3, TSH, rT3, CK and skeletal and dental radiographs (weak; low) [2].

Table 5.4. Criteria of the 2024 ETA guideline for sequencing THRA.

Source: Box 2 of the 2024 ETA guideline [2].

For a VUS in THRA, the guideline suggests, in liaison with clinical genetics and international laboratories, assessing cosegregation, the existence of a homologous THRB variant causing RTHβ, structural modelling and functional receptor assays (weak; low); it recommends not classifying a variant on predictive algorithms alone such as PolyPhen, SIFT, REVEL or CADD, and not making a definitive diagnosis of RTHα without a pathogenic variant (strong; low) [2].

5.7. Selenoprotein deficiency

The guideline recommends documenting raised FT4, normal or low FT3, raised rT3 and low plasma selenium (strong; low) [2]. Reduced plasma glutathione peroxidase 3 activity and serum selenoprotein P distinguish a selenoprotein synthesis defect from RTH and MCT8 deficiency, but do not identify the gene [2, 12]. For genetic diagnosis, exome or genome sequencing is recommended, which detects the intronic SECISBP2 variants described in several cases; in proven deficiency with a single apparent pathogenic allele, analysis of SECISBP2 cDNA is recommended to identify aberrant splicing due to a cryptic intronic variant on the other allele, and demonstration of reduced selenoprotein expression or function in the patient's cells is informative (strong; low) [2]. D2 activity in cultured fibroblasts is a research test [12, 49].

5.8. D1 deficiency

The guideline suggests documenting raised rT3 with a raised rT3/T3 ratio in the absence of non-thyroidal illness (weak; very low) and recommends studying relatives, especially the parents, as the inheritance is dominant (strong; low) [2]. Demonstration of a DIO1 variant is essential, because selenoprotein deficiency produces the same abnormalities (strong; low), and novel variants require functional demonstration (weak; very low) [2]. There is no routine clinical assay of deiodinase activity [16].

5.9. Tests in order and report content

Table 5.5. Practical sequence of tests.

The genetic report should state the gene, the transcript with its version (for example, NM_001354712.2 for THRB), the variant in HGVS nomenclature and the correspondence with the historical nomenclature, the zygosity and, in mosaics, the allele fraction by tissue, and the class according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) criteria [9, 58]. In RTHβ, THRB nomenclature follows TRβ1 numbering, which coincides with the MANE transcript [9].

5.10. Prenatal and preimplantation diagnosis

In 18 pregnancies of 13 women with THRB variants, prenatal diagnosis by amniocentesis made it possible to identify unaffected fetuses, whose mothers could receive antithyroid drugs to keep FT4 below a threshold [95]. The European guideline suggests considering antithyroid drugs when maternal FT4 exceeds 150% of the upper limit of normal, possibly preceded by prenatal diagnosis to identify unaffected fetuses, ideally within a clinical trial (weak; very low), and lists prenatal diagnosis and the use of antithyroid drugs in pregnancy as an area of uncertainty [2]. The sources consulted do not describe preimplantation genetic testing in these disorders; prior identification of the familial variant is a prerequisite for any prenatal or preimplantation test [2].

5.11. Diagnostic pathway of the French PNDS

The French PNDS defines RTHβ by the hormone profile: inappropriate TSH, normal in the great majority of cases or mildly raised, with raised T4 and/or T3, which may be moderately or markedly above the upper limit (1.5 to more than 3 times) [41]. No other assay is discriminating, and persistence of the profile without worsening over time and its confirmation with two different assay techniques favour RTHβ [41]. Antibody interference (anti-T4, anti-T3 or heterophile) is more frequent than RTHβ, especially when the clinical picture is discreet; the assay is repeated with another reagent, and in France testing for these antibodies is done in the hormonology laboratory of the Hospices Civils de Lyon [41]. The presence of anti-peroxidase or anti-thyroglobulin antibodies does not rule out RTHβ, in which they are more frequent than in the general population [41].

Confirmation is by next-generation sequencing of a panel of genes of TH transport, action and metabolism, defined by the Agence de la biomédecine in November 2022 (THRB, THRA, SECISBP2, SLC16A2, TTR, ALB and SERPINA7), in laboratories linked to the reference centre or competence centres; clinical, hormonal and molecular correlation is sometimes needed to attribute the variant [41]. The variants lie in the hormone-binding domain and in the hinge region between it and the DNA-binding domain, grouped in three clusters in exons 10 (cluster 1), 9 (cluster 2) and 7–8 (cluster 3), and are mainly point substitutions, more rarely deletions, insertions or duplications [41]. With a negative panel and strong suspicion, the France Genomic Medicine Plan 2025 allows genome analysis under the pre-indication "dysfunction of the thyrotroph axis", after validation at a national multidisciplinary meeting [41]. Table 5.6 summarises the protocol's decision tree.

Table 5.6. Decision tree of the French PNDS for clinical suspicion or incidental finding of RTHβ.

Source: appendix 6 of the French PNDS, reorganised into steps [41].

In family studies, the protocol requests TSH and FT4 and a search for suggestive manifestations; outside the neonatal period, normal tests rule out RTHβ and need no repetition, and a compatible profile leads to targeted testing for the familial variant [41]. In the newborn, as the neonatal profile does not always allow the diagnosis to be excluded, it recommends assessing thyroid function at 3 days and 3 months of life, by consensus of the authors in the absence of literature data [41].

5.12. Initial assessment by age

The PNDS proposes an initial assessment, preferably as a day case, adapted to age at diagnosis and to the manifestations known in other relatives (Table 5.7) [41].

Table 5.7. Initial assessment of RTHβ by age at diagnosis, according to the French PNDS.

Source: section 3.9 and appendix 7 of the French PNDS [41].

6. Differential Diagnosis

6.1. Framework

The differential is organised by biochemical profile. With raised TH and non-suppressed TSH, the hypotheses are analytical interference and drugs, serum transport protein variants, RTHβ, TSH-secreting pituitary adenoma (thyrotropinoma or TSHoma) and, more rarely, selenoprotein deficiency [2, 71]. With low or low-normal FT4, high or high-normal FT3 and normal TSH, the hypotheses are RTHα and MCT8 deficiency, and central hypothyroidism is the most frequent misinterpretation [2, 33]. Isolated raised rT3 points to D1 deficiency or selenoprotein deficiency, once non-thyroidal illness and fasting have been excluded [2].

Table 6.1. Differential diagnosis of the defects of sensitivity to thyroid hormone.

6.2. RTHβ and thyrotropinoma

TSHoma accounts for less than 1% of pituitary adenomas [96]. The distinction is made by biochemistry, genetics, dynamic tests, pituitary imaging and a therapeutic trial of a long-acting somatostatin ligand [2]. The presence of relatives with the same profile and a normal SHBG point to RTHβ [2]. The 2013 ETA guideline on thyrotropin-secreting pituitary tumours remains the European reference for this investigation [94]. A macroadenoma may be coincidental: in a 35-year-old man with central hyperthyroidism, identification of p.(Val458Ala) in THRB made pituitary surgery unnecessary [75]. In pregnancy, a systematic review of 9 cases of TSHoma published between 1993 and 2024 described hyperthyroid manifestations in 55%, galactorrhoea in 33% and anxiety in 44%, with somatostatin ligands stopped during gestation and surgery deferred to the postpartum period unless there was mass effect [96].

The French PNDS estimates TSHoma at less than 1 per 100,000 people and summarises the dynamic tests (Table 6.2): an absent TSH response to TRH excludes RTHβ, whereas a response makes an adenoma unlikely; the long-acting somatostatin ligand lowers free TH in both disorders, but only transiently in RTHβ (escape); and liothyronine suppression reduces TSH in RTHβ and not in the adenoma [41]. A raised glycoprotein hormone α-subunit, with an increased α-subunit/TSH molar ratio under basal conditions or on the TRH test, points to an adenoma, and the protocol requires these tests to be validated at a multidisciplinary meeting, given their uncertain utility and the cardiovascular risk [41]. Two acquired differentials complete the picture: biotin falsely raises FT4, but with suppressed TSH, so it does not exactly reproduce RTHβ, and amiodarone, in a euthyroid patient, can raise FT4 with normal TSH, through inhibition of deiodination [41].

Table 6.2. Hormone profiles and dynamic tests in RTHβ and its differentials, according to the French PNDS.

Source: appendix 5 of the French PNDS; dashes correspond to empty cells in the original [41].

6.3. Serum transport protein variants

Familial dysalbuminaemic hyperthyroxinaemia results mainly from the c.725G>A p.(Arg242His) variant of ALB (R218H in mature albumin numbering), which creates a high-affinity site for thyroxine; the person is euthyroid and the true free hormone is normal [80, 97]. The French PNDS designates the most frequent variants by codons 218 and 222 of mature albumin, 242 and 246 in the numbering that includes the pre-pro-peptide (VariantValidator confirms an arginine at codon 246 of transcript NM_000477.7), stresses that FT3 is less affected than FT4, that the familial character makes confusion with RTHβ easier and that equilibrium dialysis, the reference method, is no longer available in France, so that dialogue with the laboratory scientist and molecular study of ALB and the other transport proteins resolve the ambiguity [9, 41]. In 37 individuals with the confirmed variant, FT4 was falsely raised in all immunoassays tested except the Ortho VITROS, with decreasing susceptibility Beckman ACCESS, Roche ELECSYS, Fujirebio Lumipulse, Siemens CENTAUR, Abbott ARCHITECT and Perkin-Elmer DELFIA; five assays also gave raised FT3, in up to 30% of cases on the Siemens CENTAUR, and some affected relatives had normal FT4 depending on the method [78]. The proportion of FT4 values above the upper limit was 100% on the ACCESS, 92% on the ELECSYS and Lumipulse, 84% on the CENTAUR, 46% on the DELFIA, 43% on the ARCHITECT and 0% on the VITROS, and that of FT3 30% on the CENTAUR, 19% on the ELECSYS, 14% on the DELFIA and ARCHITECT, 11% on the Lumipulse and 0% on the VITROS [78]. One- or two-step architecture did not explain susceptibility, because the most affected assay was two-step and the least affected one-step, which depends mainly on buffer composition, notably chloride content; a low FT4 on the VITROS with high FT4 on other methods suggests this entity, and a low FT4 on that method may lead to a misdiagnosis of central hypothyroidism [78]. Because FT4 can be normal in affected relatives, the authors recommend studying relatives with definitive methods, such as ALB sequencing, and not with FT4 [78]. In the p.(Arg242His) variant, total T4 is raised two- to threefold, owing to a tenfold greater affinity of the mutant albumin for T4 and a fivefold greater affinity for T3 [78]. In a Turkish paediatric cohort with hyperthyroxinaemia and non-suppressed TSH, sequencing identified this entity in 46.2% of children tested for ALB and RTHβ in 33.3% of those tested for THRB; FT4 and TSH were similar in the two groups, FT3 tended to be lower in the dysalbuminaemic form and none of these children had tachycardia, so the authors propose analysing the ALB hotspot first-line in asymptomatic patients [71]. The frequency of the T4 form is about 1/10,000 in Caucasian populations, where it is the most common cause of euthyroid hyperthyroxinaemia [79, 80, 97]; it has been described with unusually high frequency in Hispanics of Puerto Rican origin, and the introduction of Khoo and colleagues cites a frequency of about 1 in 100 in Hispanics [78, 79]. OMIM also lists the p.(Arg242Pro) (R218P) variant, described in a Japanese family with extremely high total T4 and relatively high total T3 and rT3; homozygosity for p.(Arg242His), identified in a consanguineous Turkish family, doubles T4 binding per gram of albumin compared with heterozygotes; and the L66P variant (mature protein numbering) causes, in a Thai family, dysalbuminaemic hypertriiodothyroninaemia, with total T3 high by radioimmunoassay and very low by other methods, and normal free hormones [79]. Coexisting acquired TBG excess (for example, on oral contraceptives) or thyroid dysfunction can mask the diagnosis, and thyroxine replacement in these people requires attention to the binding of the drug to the abnormal albumin [79]. Management is to recognise the pattern and not to treat [78, 97].

SERPINA7 variants alter only the total hormones. In Oregon's T4-based newborn screening, TBG excess was the most frequent cause of persistent hyperthyroxinaemia (10 cases, 1:8088), followed by increased T4 binding without TBG excess (5 cases, 1:16,177) [34]. TBG deficiency, all forms combined, was detected in 1/4315 newborns (1/2400 boys) among 1,367,724 screened in the US Northwest Regional Screening Program from 1975 to 1991 [101]. In other screening series, TBG deficiency affected 1 in 3600 newborns in Rome and at least 1 in 5000 (1 in 2800 boys) in Minnesota, and complete deficiency ranges from 1:1200–1:1900 newborns in Japan to 1:5000–1:15,000 in Caucasian populations [100]. The defects are fully expressed in hemizygous males and, rarely, in homozygous females or those with Turner syndrome; heterozygous females usually have intermediate values, but skewed X inactivation can give them a phenotype identical to that of affected males, and inherited excess results from gene duplication or triplication [100]. Before attributing the abnormality to a variant, acquired causes are excluded: oestrogens, pregnancy and oral contraceptives increase TBG, whereas androgens, glucocorticoids, severe liver disease, severe non-thyroidal illness and renal or intestinal protein losses decrease it [98]. TBG abnormalities are not an indication for levothyroxine or antithyroid drugs [98, 99]. Under the TTR gene, OMIM lists the benign biochemical trait of dystransthyretinaemic hyperthyroxinaemia; patients with hereditary transthyretin amyloidosis have no clinical deficit of thyroxine transport [102, 103].

6.4. RTHα, MCT8 deficiency and central hypothyroidism

RTHα and MCT8 deficiency share low or low-normal FT4 with high FT3 and normal TSH. In RTHα, intellectual disability is mild to moderate, brain MRI is described as normal in the table of the GeneReviews chapter on Allan-Herndon-Dudley syndrome and some features improve with levothyroxine; in MCT8 deficiency there is severe hypotonia, hypomyelination and peripheral thyrotoxicosis with high SHBG [2, 33]. The Oxford Desk Reference recommends reviewing the newborn screening card results and the first venous sample before starting levothyroxine in congenital hypothyroidism, because MCT8 and THRA defects are the exceptions to the usual pattern of raised TSH with low FT4 and FT3, and measuring TSH, FT4 and FT3 in all first-degree relatives of the proband [86]. The other differentials of congenital hypothyroidism include TSH resistance due to TSHR, Albright hereditary osteodystrophy, IGSF1 deficiency and combined pituitary hormone deficiencies [86].

6.5. Selenocysteine pathway defects and D1 deficiency

The thyroid profile of selenoprotein deficiency, with high FT4 and rT3 and low or normal FT3, accompanied by low plasma selenium, selenoprotein P and glutathione peroxidase 3, distinguishes it from RTH and MCT8 deficiency, but does not identify the gene: SECISBP2 and TRU-TCA1-1 defects give overlapping biochemical pictures and are separated by genetics [2, 12]. D1 deficiency produces the same rT3 abnormalities as selenoprotein deficiency, so the European guideline considers it essential to demonstrate the DIO1 variant [2]. SEPSECS variants, in the final step of tRNA-Sec synthesis, cause progressive cerebello-cerebral atrophy with normal T4 and selenium, and defects of individual selenoproteins (SELENON, GPX4, TXNRD2) do not have the thyroid profile of this class [2, 64]. The common DIO1 polymorphism rs2235544 alters the FT3/FT4 ratio in the general population and should not be confused with the pathogenic variants of the gene [68].

7. Treatment and Risk Management Strategies

7.1. General principles

There is no therapy that corrects the receptor or selenocysteine pathway defect. Most patients with RTHβ do not need specific treatment, and the aim of treatment, when indicated, is to control thyrotoxic symptoms and signs and reduce exposure of sensitive tissues without causing a rise in TSH that enlarges the goitre [2, 37]. In RTHα and selenoprotein deficiency the aim is the opposite: to supply enough hormone to overcome the resistance of TRα-dependent tissues or to compensate for deficient conversion of T4 to T3 [2]. No randomised trial has been conducted in any of these disorders, and all recommendations rest on cases, series and retrospective cohorts [2, 104].

The European guideline is explicit about what to avoid in RTHβ: it recommends avoiding antithyroid drugs and thyroid ablation, surgical or with radioiodine, in the absence of significant comorbidity (strong recommendation; moderate quality), and limiting ablation to severe, uncontrolled or life-threatening hyperthyroidism, for example in heterozygous RTHβ with thyroid autonomy due to Graves' disease or a toxic nodule, in homozygous RTHβ with thyrotoxic cardiomyopathy, in large goitre with compression and in carcinoma; after ablation, levothyroxine is needed to restore thyroid function tests to the patient's own values (strong; low) [2]. Chronic under-replacement with levothyroxine after ablation risks pituitary thyrotroph hyperplasia [2, 89]. The two large cohorts showed that ablation did not reduce MACE or mortality [39, 40].

7.2. TRIAC (tiratricol) in RTHβ

TRIAC, triiodothyroacetic acid or tiratricol, is a thyroid hormone analogue with central thyromimetic action that inhibits TSH secretion and lowers circulating hormones [2]. The European guideline recommends beta-blockade alone or, rarely, TRIAC to control thyrotoxic symptoms (anxiety, tremor, palpitations) and signs (tachycardia) (strong; moderate), and requires that the decision to use TRIAC be taken after discussion with experienced centres (strong; moderate) [2]. The recommended dose is 1.4 to 2.8 mg per day, in two or three doses, titrated to control symptoms and signs and lower FT4, the divided dosing being justified by the short half-life (strong; moderate) [2]. Monitoring is by FT4, because TRIAC cross-reacts in T3 immunoassays, but not on LC-MS/MS [2].

In an international retrospective cohort of 28 patients from 23 families followed in 6 centres, with a median age of 26.0 years (0.4–73.1), 27 heterozygotes with 18 distinct variants and 1 homozygote for p.(Arg316His), the median dose was 1400 µg/day (range 136–4200; n=27), equivalent to 24 µg/kg/day, for a median of 50 months (5–118); TSH fell from 3.23±1.78 to 1.88±0.76 mU/L (P <0.0001), FT4 from 40.0±13.5 to 25.1±10.6 pmol/L (P <0.0001), HR from 86±22 to 76±13 bpm (n=22; P=0.014) and LDL from 2.92±0.96 to 2.35±0.78 mmol/L (n=17; P=0.017), with no significant change in BP, SHBG, ferritin, total cholesterol or bone turnover markers [54]. In the single-centre Cambridge cohort, 8 adults (mean age 41 years) treated with TRIAC monotherapy reduced their hyperthyroid symptom scale score from 17.5 to 6 (P=0.0014), FT4 from 30 to 17 pmol/L (P=0.007) and total T3 from 2.5 to 1.4 nmol/L (P=0.001), with normalisation in 7 of 8, without a rise in TSH (2.1 to 1.7 mU/L; P=0.65) or worsening of cholesterol or triglycerides, and without recorded adverse effects [89]. The mean dose was 1.8 mg/day (1.3–3 mg), divided into two doses in 4 patients and three in 4, for a median of 40 months (13–143), and the dose needed to bring FT4 into the reference range was a median of 1.4 mg/day (range 0.7–3.5) [89].

In severe forms, the guideline and published cases support combining an antithyroid drug with TRIAC. In a 9-year-old boy with homozygous RTHβ and signs of early dilated cardiomyopathy, carbimazole combined with TRIAC, with careful titration to limit the rise in TSH and the increase in thyroid volume, produced a sustained reduction in hormones with normal TSH, lower basal metabolic rate, gain in lean mass and improvement in growth and cardiac function [57]. In the case of extreme mosaic RTHβ, the final regimen was methimazole 15 mg/day with TRIAC 3150 µg/day (Section 10) [53]. In the two siblings homozygous for p.(Arg316Cys), TRIAC was used to lower TSH and contain the goitre, at 0.26 mg/day in three doses, doubled one year later in the sister and increased to 0.525 mg/day at 8 years in the brother; the sister, who had been receiving propranolol 20 mg/day since age 3 for sinus tachycardia, underwent thyroidectomy at 8.75 years for goitre enlargement with macronodules (diffuse adenomatous hyperplasia) and was switched to levothyroxine 75 µg/day [19]. These are doses from individual cases, lower than the guideline's adult doses, and do not constitute a validated paediatric dosage [2, 19]. For ADHD that interferes with daily life, the guideline suggests formal diagnosis by neuropsychiatry and standard intervention or TRIAC, recalling that the latter is not authorised for RTHβ in all countries (weak; low) [2]. In one publication, liothyronine, but not methylphenidate, improved ADHD, and alternate-day T3 reduced a large goitre in one patient [2].

7.3. Beta-blockers and treatment of arrhythmias in RTHβ

The guideline recommends cardioselective beta-blockade for tachycardia, atrial fibrillation and supraventricular tachyarrhythmias in heterozygous RTHβ, and for cardiac hyperthyroidism in the homozygous form (strong; low) [2]. In atrial fibrillation, it recommends specialist cardiology advice and warns that chemical (amiodarone, flecainide) or electrical cardioversion and ablation may fail (strong; low) [2]. The guideline gives no doses, so those of the official labels are used, as a declared extrapolation: atenolol has on the Food and Drug Administration (FDA) label 50 mg once daily in hypertension, up to 100 mg/day, and 50 to 100 mg in angina, up to 200 mg in a single dose, with no established paediatric use; propranolol has 40 mg twice daily in hypertension, with maintenance of 120 to 240 mg/day, 80 to 320 mg/day in 2 to 4 doses in angina, and should be withdrawn gradually over weeks [105, 106]. Propranolol has been used in paediatric cases of RTH, for example in a school-age boy with neck swelling, nocturnal enuresis and tachycardia, with symptomatic improvement [107].

7.4. Levothyroxine in RTHα and in RTHβ with coexisting hypothyroidism

In RTHα, the guideline recommends a therapeutic trial of levothyroxine in all patients (strong; low) and suggests continuing it long term unless there are adverse effects or intolerance (weak; very low), but considers the data insufficient to recommend a dose, which should be above the usual replacement dose, or clinical and biochemical targets (weak; very low) [2]. The doses published in the series reach 3.8 µg/kg/day; TSH remains normal in some treated patients and is suppressed with raised hormones in others, and TSH-suppressive doses may be needed to overcome the resistance of TRα-dependent tissues, with a risk of thyrotoxicosis in TRβ-dependent tissues [2]. In the long follow-up of 4 children and 3 adults from two families, with a median treatment duration of 6.7 years, the dose at the last visit was 1.4 µg/kg/day (1.2–2.2) in adults and 1.9 µg/kg/day (1.2–2.4) in children [108]. According to the HGMD compilation consulted by the Norwegian authors in September 2022, thyroxine had been used in 21 patients with 14 THRA variants, with counts by age at onset published as 6 before age 3, 15 in childhood or adolescence and 3 in adulthood; the benefit was limited or nil for variants in the C-terminal segment of the binding domain and for nonsense or frameshift variants, and greater when treatment started early and for missense variants in the segment common to TRα1 and TRα2 [38]. In an open pilot study, 4 adults with p.(Ala263Val) were treated with 1.75 µg/kg/day for 3 years [109]. For comparison, the Synthroid® label indicates for adult hypothyroidism replacement 1.6 µg/kg/day, with titration of 12.5 to 25 µg every 4 to 6 weeks, doses above 200 µg/day seldom needed and inadequate response to more than 300 µg/day rare, suggesting poor adherence, malabsorption or interaction, with no maximum dose defined on the label, and in children 10 to 15 µg/kg/day from 0 to 3 months, 8 to 10 from 3 to 6 months, 6 to 8 from 6 to 12 months, 5 to 6 from 1 to 5 years, 4 to 5 from 6 to 12 years, 2 to 3 above 12 years with incomplete growth and 1.6 once growth is complete; in children at risk of hyperactivity, one quarter of the dose is used initially, increased by one quarter per week [92]. The label indicates a single daily dose on an empty stomach, 30 to 60 minutes before breakfast, with an interval of at least 4 hours from drugs that interfere with absorption [92]. RTHα is not an approved indication in any jurisdiction, so use is off-label, and the published doses exceed the label's replacement doses [2, 10].

In RTHβ with coexisting hypothyroidism, autoimmune or congenital, a raised TSH with normal free hormones means under-replacement; the guideline recommends titrating levothyroxine to an FT4 comparable to that of relatives who have RTHβ alone, or keeping TSH high-normal with raised free hormones, monitoring cardiac function to avoid overtreatment (strong; low) [2].

7.5. Liothyronine and selenium in selenoprotein deficiency

In selenoprotein deficiency, the guideline recommends liothyronine when FT3 is subnormal (strong; low), recommends not supplementing selenium (strong; low) and suggests considering an antioxidant, for example alpha-tocopherol (weak; low) [2]. Liothyronine corrects subnormal FT3 and improves linear growth, alone or combined with growth hormone, although untreated patients may reach normal height [2]. In the 2026 Korean case, liothyronine normalised thyroid function tests, but developmental and speech delays persisted [77]. The guideline gives no dose; for comparison, the Cytomel® label indicates for adult hypothyroidism 25 µg/day initially, with increases of 25 µg every 1 to 2 weeks and maintenance of 25 to 75 µg/day, 5 µg initially in the elderly and those with heart disease, and in paediatrics 5 µg/day initially with increases of 5 µg every 3 to 4 days, infants of a few months possibly needing 20 µg/day, a 1-year-old child 50 µg/day and children over 3 years the full adult dose, with no maximum dose defined on the label [110]. Use in selenoprotein deficiency is off-label and the dose is extrapolated from the hypothyroidism label [2, 110].

The recommendation not to supplement selenium rests on a supplementation trial in which 3 affected and 2 unaffected siblings received 100, 200 or 400 µg/day of selenomethionine-rich yeast and 400 µg/day of sodium selenite, each form for one month: selenomethionine increased serum selenium in all, regardless of genotype, and selenite increased selenoprotein P in the normal siblings and less in the affected ones, but neither form increased glutathione peroxidase activity or corrected the thyroid abnormalities, without adverse effects [42]. With adequate intake, selenium is not the limiting factor: the defect lies in the incorporation machinery [12, 42]. In the tRNA-Sec defect there is no established therapy; selenium has been proposed as rational, because selenium status shifts the balance between tRNA isoforms, but it has not been tested in patients, and in cells edited with C65G selenoprotein expression no longer responded to selenium, which argues against benefit [14, 66]. The first patient received only symptom relief and the second remained asymptomatic without treatment over 6 years [14, 26]. For D1 deficiency it has not been determined whether carriers need treatment; coexistence of D1 haploinsufficiency with other congenital thyroid defects may require adjustment of levothyroxine to ensure tissue T3, especially during development [2, 67].

7.6. Growth hormone and other non-pharmacological interventions

In children with RTHα and short stature, recombinant human growth hormone has been combined with levothyroxine, off-label: in three Brazilian children, two also received growth hormone, with improvement in growth velocity and height standard deviation score, and one gained 2.5 standard deviations and reached normal adult height [87]. The guideline suggests reassessing growth hormone secretion after levothyroxine is started in children with reduced secretion or low IGF-1 (weak; low) [2]. In RTHα, it recommends lifelong follow-up by an endocrinologist (strong; very low) and referral to neurology, neuropsychology, gastroenterology, haematology, dentistry, physiotherapy, speech therapy and occupational therapy (strong; very low) [2]. In RTHβ, it recommends prompt treatment of ear, nose and throat infections and audiometry at diagnosis (weak; low), neuropsychological and psychometric assessment of children for ADHD and cognitive deficits and educational support (weak; low), and assessment of growth and development by a paediatric endocrinologist in childhood (weak; very low) [2]. In selenoprotein deficiency, management includes surveillance of the muscular dystrophy with MRI and, in selected cases, muscle biopsy, with vital capacity and polysomnography for nocturnal hypoventilation (strong; low) [2].

7.7. Drugs used in these disorders

Table 7.1. Drugs, doses and sources. Except for tiratricol in France, which has a national authorisation for RTHβ, none has an approved indication for the disorders in this article and the other uses are off-label; the label doses cited serve for comparison, under the rule that the reference dose is the guideline's when one exists.

7.8. Regulatory status of tiratricol

Table 7.2. Tiratricol: regulatory status and practical consequence.

7.9. Evidence from treatment cohorts and trial registries

Table 7.3. Treatment studies in the disorders covered in this article.

None of these studies is randomised or controlled, and the results are not comparable with one another: populations, durations and outcomes differ. The TRIAC cohorts included symptomatic patients from reference centres and did not include those who stopped the drug for lack of efficacy, which favours the observed effect; in the international cohort, the homozygous patient was excluded from the analysis because values exceeded the limit of quantification, variants with a severe premature stop codon were not included, and the coordinating centre receives royalties from the manufacturer [54]. In the Cambridge cohort, the reductions in symptoms and hormones are within-group changes, without a comparator [89]. The negative results are informative: selenium supplementation did not correct selenoprotein deficiency, and thyroid ablation did not reduce cardiovascular events or mortality in the Welsh and Italian cohorts [39, 40, 42].

In trial registries, a ClinicalTrials.gov search by condition and gene identified no interventional trial specifically for RTHβ, RTHα, selenoprotein deficiency or D1 deficiency; there are two observational records, DEEPTYPE (NCT06566066, Charité Berlin, recruiting since 2021, up to 200 participants, expected until 31 July 2029, which includes SLC16A2 and THRA and uses an abnormal FT3/FT4 ratio as a screening parameter) and ADAM-THAD (NCT06307990, Istituto Auxologico Italiano, recruiting since 1 January 2023, up to 150 participants, in RTHβ and RTHα) [104]. The tiratricol trials all concern MCT8 deficiency. This is the result of a registry search and does not prove the absence of unregistered studies [104]. Selective TRβ agonists in development, such as resmetirom, are intended for steatotic liver disease and not for RTHβ, in which TRβ itself is dysfunctional [104].

7.10. Developments in the last 12 months

Between October 2025 and October 2026 the following were published: two TRIAC treatment cohorts in RTHβ, an international one with 28 patients and a single-centre one with 8, both published results of observational studies [54, 89]; a case of extreme RTHβ due to a mosaic variant treated with methimazole and TRIAC [53]; a cross-sectional study of 105 adults on the dyslipidaemia of RTHβ [82]; two Turkish paediatric cohorts and a Norwegian pilot of levothyroxine in RTHα [37, 71, 109]; two cases of selenoprotein deficiency due to SECISBP2 [76, 77]; the new 2026 ETA guideline on interference in thyroid function immunoassays, a recommendations document [91]; a review of selenoprotein genetics [123]; and, on the regulatory front, FDA approval of tiratricol on 28–29 September 2026, for MCT8 deficiency [111]. On THRB allelic disease, there is a peer-reviewed familial case of vitelliform macular dystrophy and a non-peer-reviewed preprint proposing a TRβ1 gain-of-function syndrome [61, 62]. A Chinese case described the p.(Met313Thr) variant as unreported; the variant had already been published in 2002 [73, 124].

7.11. Pregnancy, perioperative care and transition

In the pregnancy of a woman with RTHβ, the guideline suggests multidisciplinary management by endocrinologist and obstetrician, given the increased risk of miscarriage, small-for-gestational-age newborn and low birth weight (weak; low), recommends monitoring fetal growth and HR and considering antithyroid drugs if there is fetal tachycardia or growth restriction (strong; low), and suggests considering antithyroid drugs when maternal FT4 exceeds 150% of the upper limit of normal, possibly after prenatal diagnosis to identify unaffected fetuses and ideally within a clinical trial (weak; very low) [2]. In the series of 18 pregnancies, the stated aim was to keep maternal FT4 no higher than 50% above the upper limit of normal, with an internal inconsistency in the abstract regarding the threshold (which is also described as below 20% above the upper limit), so the guideline threshold, 150% of the upper limit, is the one that should guide practice [2, 95]. In RTHα, data are insufficient and the guideline suggests continuing levothyroxine (weak; very low) and recommends not using liothyronine in pregnancy (strong; very low) [2]. Tiratricol crosses the placenta and is contraindicated in pregnancy in the European summary of product characteristics [93].

The French PNDS diverges from the European guideline on this point: when maternal manifestations worsen, especially cardiovascular ones, it proposes a beta-blocker first, with the obstetrician's agreement, and considers it undesirable to use antithyroid drugs to lower maternal hormones, because of the risk of transplacental passage to a fetus of unknown thyroid status; intervention in the face of maternal-fetal repercussions is described as under discussion, without solid data [41]. The proposed surveillance includes monthly maternal TSH, FT4 and FT3, fetal ultrasound scans at an adapted frequency to assess the fetal hormonal status indirectly, decision at a multidisciplinary meeting with the reference centre, paediatric assessment of the newborn and thyroid function (TSH and FT4) at 3 days, independently of filter-paper screening, and at 3 months, with the same strategy when the affected parent is the father [41].

The sources consulted do not describe specific perioperative recommendations for these disorders; the most relevant practical implication is that the tachycardia and tachyarrhythmias of RTHβ are mediated by TRα and respond to cardioselective beta-blockade, whereas lowering hormones with antithyroid drugs provokes an exaggerated rise in TSH [2, 89]. In the transition to adulthood, the guideline recommends lifelong follow-up by an endocrinologist in RTHα and continued cardiovascular and metabolic surveillance in RTHβ (Section 11) [2].

7.12. Areas of therapeutic uncertainty

The European guideline lists as uncertainties, in RTHβ, prenatal diagnosis and the use of antithyroid drugs in pregnancy, the benefit of neonatal diagnosis and early intervention, the effect of lipid-lowering drugs or TRIAC on cardiovascular events, fracture risk and the role of antiresorptives, and further evaluation of TRIAC and liothyronine in ADHD; in RTHα, biomarkers and treatment targets, the possibility that supraphysiological doses cause tissue thyrotoxicosis, preconception and pregnancy management and whether diagnosis and levothyroxine from birth prevent adverse neurodevelopmental outcomes; and, in selenoprotein deficiency, whether chronic oxidative damage predisposes to neurodegeneration, premature ageing or neoplasia, whether selenium corrects the tRNA-Sec defect and whether long-term antioxidants prevent complications [2].

7.13. French PNDS management and divergences from the European guideline

The French PNDS starts from the premise that, in most patients, the hormone concentrations resulting from the THRB variant ensure a clinically acceptable balance, so the aim is not to alter the thyroid balance: it recommends withholding treatment in paucisymptomatic forms acceptable to the patient, and discussion at a multidisciplinary meeting of the reference centre or a competence centre whenever tiratricol, liothyronine, levothyroxine, beta-blockers, methylphenidate or radical treatment is considered [41]. The protocol acknowledges that the level of evidence is low and that management rests mainly on expert opinion [41].

For goitre, surveillance suffices in most cases, because the volume is small; in large goitres, possibly linked to prolonged mildly or frankly raised TSH, medical treatment seeks to lower TSH [41]. Supraphysiological liothyronine reduced the goitre in cases and small series, in parallel with suppression of TSH below 0.1 mIU/L, sometimes with a beta-blocker for tachycardia; tiratricol, with an affinity for TRβ greater than that of T3 and equal to that of T3 for TRα, reduced TSH, TH and goitre volume in more than 30 published cases, but failed in others, which suggests dependence on the variant, and the doses used ranged from 0.7 to 4.2 mg/day (mean 1.4 mg/day); the combination of high-dose levothyroxine with a beta-blocker and calcium and vitamin D supplements has been published, but is considered experimental [41]. Goitre surgery is not recommended, because partial thyroidectomy is followed by recurrence and total thyroidectomy requires very high doses of levothyroxine, with a risk of cardiac thyrotoxicosis; it is discussed for large goitre, lesions suspicious for malignancy and severe uncontrolled cardiothyrotoxicosis, and replacement then aims at the patient's preoperative hormone concentrations; iodine-131 is not recommended either [41]. For nodules, investigation is the usual one (history, examination, ultrasound and cytology), surgery is reserved for macronodules over 10 mm highly suspicious for malignancy, preoperative molecular analysis of the nodule is not routinely available, the choice between lobectomy and total thyroidectomy in a carcinoma follows the 2021 recommendations for thyroid cancer, and for benign single nodules requiring volume reduction ethanol ablation or thermal ablation may be discussed with the reference centre [41].

For neuropsychological manifestations, assessment is systematic and conventional ADHD treatment, such as methylphenidate, may be discussed, with improvement in many cases in the authors' experience [41]. In a randomised, placebo-controlled, double-blind crossover trial with 17 children with ADHD (8 with RTHβ and 9 without), supraphysiological liothyronine (1.5 to 2 µg/kg/day according to the PNDS, with suppressed TSH) improved the Conners hyperactivity index in 5 of the 8 children with RTHβ, with no change in the other 3; in the 9 children without RTHβ it had no effect in 7, improved 1 and worsened 1, and increased commission errors on the continuous performance test in 5 of 9 [41, 125]. The PNDS summarises the result as efficacy in children with RTHβ and lack of efficacy in the others; the study figures show an inconsistent individual benefit in a very small sample [125]. Tiratricol was associated with less hyperactivity in some cases, beta-blockers prescribed for tachycardia appeared to improve hyperactivity, and antithyroid drugs alone promote goitre and are not indicated [41].

For tachycardia, the protocol proposes a non-cardioselective beta-blocker, propranolol 1 to 2 mg/kg/day in children and a dose adapted to the clinical picture in adults, after an electrocardiogram and exclusion of contraindications, allowing for an additional benefit on neurobehavioural manifestations; arrhythmias are treated according to cardiology recommendations, with an experienced cardiologist, avoiding amiodarone as first line, because it hinders the interpretation of hormonal changes, with no contraindication to other antiarrhythmics or ablation [41]. When there is doubt about the relationship between TH and the cardiac rhythm or functional status, a time-limited trial of antithyroid drugs may be discussed or, as a definitive measure, total thyroidectomy to "choose" the hormone level with levothyroxine, at the cost of hormonal exposure of the other tissues and with regular cardiological surveillance [41]. For hearing, hearing loss and recurrent otitis are managed with an otorhinolaryngologist and a hearing aid specialist, with unilateral or bilateral hearing aids, cochlear implantation according to HAS criteria in sensorineural deafness and speech therapy; for abnormalities of height and bone mineral density, the recommendations for the general population apply [41]. In France, the propranolol oral solution Hemangiol®, authorised only for infantile haemangioma, is used off-label in children who cannot swallow tablets (Table 7.1) [41, 114].

Table 7.4. Divergences between the 2024 French PNDS and the 2024 ETA guideline in RTHβ.

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by inheritance

In heterozygous RTHβ, inheritance is autosomal dominant and each child of an affected parent has a 50% probability of inheriting the variant; about 23% of variants arise de novo, so the risk for siblings of a sporadic case whose parents have normal thyroid function tests is low, but not nil, because of possible germline mosaicism [2, 18]. In biallelic THRB deletions, inheritance is autosomal recessive with clinically and biochemically normal heterozygous carriers, and the risk for siblings of a case is 25% [19–21]. Homozygosity for missense variants results from the union of two affected parents, each with heterozygous RTHβ, in which case the probability of homozygous offspring is 25% and the expected phenotype is severe, with a risk of thyrotoxic cardiomyopathy [19, 57]. In RTHα, inheritance is autosomal dominant, often through a de novo variant: in the two adolescents of Moran 2017 neither parent was affected, and 9 of the 14 variants then described had arisen sporadically [22]. In selenoprotein deficiency due to SECISBP2 or TRU-TCA1-1, inheritance is autosomal recessive, with a 25% risk for siblings; the heterozygous parent of the second patient with the tRNA-Sec variant had normal thyroid function tests, except for moderately increased rT3 [2, 26]. In that patient, homozygosity resulted from paternal uniparental isodisomy, without the variant in the mother: in this situation, the recurrence risk for siblings is low and not 25%, and only study of both parents distinguishes the two mechanisms [15, 26]. In D1 deficiency, inheritance is dominant through haploinsufficiency, and the guideline recommends testing relatives, especially the parents [2, 6].

8.2. Mosaicism and its consequences for counselling

Somatic THRB mosaicism has two implications. The first is diagnostic: negative sequencing in blood does not exclude the disease, especially when abnormal thyroid function tests are transmitted to offspring [2, 58]. The second is reproductive: in the first case described, the mosaic father transmitted the variant to both sons, who had full RTH, which indicates germline involvement and a transmission risk that may approach that of a heterozygote [18]. The allele fraction measured in blood does not predict the fraction in the germline, as shown by the variation between tissues in the case with 12% in blood, 17.9% in oral mucosa and 0% in conjunctiva [58].

8.3. Cascade testing

After diagnosis in the RTHβ index case, the guideline recommends testing first-degree relatives, first biochemically and then genetically (strong; high); abnormal tests in first-degree relatives increase the likelihood of RTHβ and constitute a diagnostic element (strong; high) [2]. Family study is, in practice, an important route to diagnosis: it accounted for 46.4% of cases in the Spanish series and 23.1% in the Turkish paediatric series [35, 37]. The Oxford Desk Reference recommends measuring TSH, FT4 and FT3 in all first-degree relatives of the proband when investigating congenital hypothyroidism, precisely because the resistance forms may go unnoticed [86]. In RTHα, diagnosis of an index case by an intellectual disability panel led to the identification of four never-treated affected adults in the same family, with a more marked phenotype than in patients treated since childhood, which shows the value of family study in adults too [38].

The French PNDS proposes study of first-degree relatives, at a minimum, with TSH and FT4, and targeted testing for the known variant when the profile is compatible; outside the neonatal period, normal tests exclude RTHβ without the need for repetition, and in the children of an affected parent, father or mother, testing is done at 3 days and 3 months of life, with targeted genetic testing after parental consent [41]. The stated aim of this neonatal screening is to institute cardiovascular, auditory and neurodevelopmental surveillance, and the protocol recalls the two possible errors in the neonatal period: confusing RTHβ with congenital hypothyroidism because of raised TSH, or considering thyroid function normal because TSH is not raised [41].

8.4. Pregnancy in a woman with RTHβ: counselling data

In an Azorean family (Portugal) with the p.(Arg243Gln) variant, 167 members were studied and 36 couples with complete information analysed: the mean miscarriage rate was 22.9% when the mother was affected (n=9), 2.0% when the father was affected (n=9) and 4.4% in unaffected relatives (n=18) (χ²=8.66; P=0.01), with an excess of miscarriages in affected mothers (z=3.10; P=0.002) [72]. Affected mothers had 20 affected and 11 unaffected children (P=0.07) and affected fathers 15 and 12 (P=0.35) [72]. Unaffected newborns of affected mothers were significantly smaller, with a weight-for-gestational-age standard deviation score of −1.79 (standard deviation 0.86) against −0.06 (1.11) to −0.22 (0.70) in the other groups (P <0.001), and only these had undetectable TSH on newborn screening [72]. The authors' interpretation is that the unaffected fetus, sensitive to the excess maternal hormones, suffers fetal thyrotoxicosis, and that miscarriages possibly affect mainly unaffected fetuses [72]. In a retrospective review of 13 women with THRB variants in 18 pregnancies, prenatal diagnosis by amniocentesis made it possible to identify fetuses without the variant and to treat those mothers with antithyroid drugs to restrain maternal FT4; there was no significant difference in birth weight corrected for gestational age or in TSH at birth between infants with and without the variant born to mothers with RTHβ [95]. Neonatal thyrotoxicosis in the child of a mother with RTHβ has been described with the p.(Met313Thr) variant, published together with maternal infertility [73].

8.5. Prenatal and preimplantation diagnosis

Prenatal diagnosis of the fetal genotype in RTHβ has a mainly therapeutic purpose: to identify the unaffected fetus, whose risk of fetal thyrotoxicosis and growth restriction justifies lowering maternal hormones with antithyroid drugs [2, 95]. In a case of selective pituitary RTH with the p.(Thr337Ala) variant, TRIAC was stopped at the start of pregnancy and resumed after recurrence of thyrotoxic manifestations, and chorionic villus sampling showed that the fetus was heterozygous; the authors advocated prenatal diagnosis and treatment of maternal hyperthyroidism to avoid fetal thyrotroph hyperplasia and goitre, an approach from an isolated report that contrasts with the contraindication of tiratricol in pregnancy in the European summary of product characteristics [8, 93]. The guideline places this procedure at the level of a suggestion, ideally within a clinical trial, and lists it as an area of uncertainty [2]. Any prenatal test presupposes prior identification of the familial variant; the sources consulted do not describe the use of preimplantation genetic testing in these disorders.

The French PNDS, after recalling the 2010 proposal of fetal genotyping in amniotic fluid to guide treatment of the affected mother, considers it undesirable to offer prenatal diagnosis systematically, for lack of effective specific therapy, especially in pregnancy, and because the type of variant does not predict prognosis; it prefers neonatal screening of the children of affected parents [41]. The two sources agree that prenatal diagnosis is not routine practice; they differ in that the European guideline keeps it as an option to be investigated and the French protocol advises against it (Table 7.4) [2, 41].

8.6. Counselling of children and adolescents

Cascade testing of first-degree relatives, including minors, has immediate clinical consequences in RTHβ, through surveillance of growth, development, hearing and cardiac function, and in RTHα, because the results of levothyroxine are better when treatment starts in early childhood and untreated adults are more affected [2, 38, 108]. In the recessive forms, the relevant information for heterozygous carriers is the absence of disease: in THRB deletions, carriers are normal; in the tRNA-Sec defect, the heterozygous parent had only mildly increased rT3 [19, 26]. In D1 deficiency, study of the parents serves to establish dominant transmission, in a picture with no described clinical disease [2].

8.7. Disclosure of the diagnosis and genetic counselling in the French PNDS

The French PNDS organises disclosure in two stages, by the paediatric or adult endocrinologist: in the first, the suspicion is discussed in light of the clinical picture and the hormone profile, with the differentials (assay interference, genetic abnormalities of the transport proteins, thyrotropinoma), and molecular study is proposed according to the HAS and Agence de la biomédecine good practice rules for constitutional genetics; in the second, the result is given and the disease, transmission, surveillance, possible treatments, the importance of follow-up during transition and the organisation of the centres are explained, with an offer of psychological support and contact with a patient association [41]. The protocol asks that genetic counselling be provided by a medical team experienced in RTHβ, explaining the 50% transmission risk and the exceptional existence of biallelic and mosaic forms, and that it be revisited when there are plans for parenthood [41].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What the available figures measure

There is no validated population prevalence estimate for any of the entities in this article. The values cited are of two kinds: local newborn screening incidences, with large denominators but numerators of 2 to 4 cases, and cumulative counts of published cases, which depend on the intensity of investigation in each country. Neither should be presented as a national prevalence.

Table 9.1. Epidemiological data by entity and by region.

9.2. Dependence on the screening method

The difference between the two screening incidences is not geographical but methodological: Oregon measures total T4 and the Basque Country measures TSH and T4 simultaneously, so both programmes detect hyperthyroxinaemia, whereas programmes that measure TSH alone identify neither RTHβ nor RTHα [34, 35, 86]. In Korea, a series of 5 probands with RTH and 6 with MCT8 deficiency led the authors to conclude that TSH-based screening has a limited role in the diagnosis of these disorders [90]. An abnormal screening result can nevertheless lead to diagnosis: in a Turkish paediatric series, 15.4% of cases were detected by raised TSH on newborn screening [37]. Patients with mild forms and those living in regions with TSH-only screening remain undiagnosed, so all the frequencies cited here underestimate the true frequency; RTHα is considered underdiagnosed because thyroid function tests appear near-normal [2, 7, 10].

9.3. Populations and cohorts by region

The published series come mainly from the USA (the NIH prospective study and the Chicago laboratory), the United Kingdom (Cambridge and Wales), Italy, Spain, Turkey and Korea [35, 37, 39, 40, 55, 56, 71, 90]. The largest family described with RTHβ is Azorean (Portugal), with 167 members studied and the p.(Arg243Gln) variant, and it constitutes the main data set on pregnancy in this disease (Section 8) [72]. From Brazil come three children with THRA variants, the series that documents the response to growth hormone in RTHα [87]. Also from Brazil comes a cross-sectional study from Ribeirão Preto on mineral metabolism, with 14 patients with RTHβ and 24 controls (Section 3.4) [84]. In France, the reference centre database recorded atrial fibrillation in 20% of patients (Section 3.3) [41]. Among the families with rare cases are the first Norwegian family with RTHα, a Polish series of 5 children with THRA variants, an Argentinian and a Turkish family with SBP2 deficiency and families with homozygous RTHβ of Mexican, Australian Aboriginal and Turkish origin [19, 23, 36, 38]. The two cohorts that established the excess of cardiovascular morbidity and mortality are from Wales and Italy, and their results agree [39, 40]. The sources consulted contain no incidence or prevalence data from Portugal or Brazil for any of these entities, nor national cohorts beyond the Azorean family, the three Brazilian children and the Ribeirão Preto series; this absence reflects the sources consulted and does not demonstrate the absence of published data.

9.4. Variability between populations

The albumin variant responsible for familial dysalbuminaemic hyperthyroxinaemia is more frequent in some ethnic groups, and the entity is the most common cause of euthyroid hyperthyroxinaemia in Caucasian populations [80, 97]: the frequency cited is 1:10,000 in Caucasians and about 1:100 in Hispanics, with a described concentration in Hispanics of Puerto Rican origin, and the p.(Arg242Pro) variant has been described in a Japanese family [78, 79]. Complete TBG deficiency is more frequent in Japan (1:1200–1:1900 newborns) than in Caucasian populations (1:5000–1:15,000) [100]. For TBG variants, high frequencies are described in some African, Australian Aboriginal and Inuit populations [98, 99]. For RTHβ, RTHα and the metabolic defects, the sources consulted describe no founder variants or differences in frequency between populations, with the exception of the local concentration of one variant in the Azorean family, which is the effect of an extended family on an island and not a population frequency [72]. The Turkish paediatric cohort shows that the aetiological composition of hyperthyroxinaemia with non-suppressed TSH differs from what is expected: familial dysalbuminaemic hyperthyroxinaemia was more frequent than RTHβ among the children tested [71].

9.5. European reference network, centres and patient associations

The relevant European network is Endo-ERN, the European Reference Network on Rare Endocrine Conditions, which brings together more than 100 reference centres in 27 European Union Member States and Norway and is organised into eight main thematic groups; group 8 is the thyroid group and covers RTHβ, RTHα, selenoprotein deficiency and deiodinase defects [43, 126]. In the list of reference centres consulted, 56 of the 105 centres belong to the thyroid thematic group, concentrated in Italy (16 centres), the Netherlands (7) and Germany (7); in Portugal, the reference centre registered in this group is the Centro Hospitalar Universitário de São João, in Porto [126]. The network does not publish its own recommendations for these entities and refers to the ETA guideline, the society that issues the European guidelines published in the European Thyroid Journal [43, 127]. The American Thyroid Association has no specific guideline for resistance to thyroid hormone, and its hyperthyroidism and hypothyroidism guidelines apply to the comorbidities [2].

For patients and families there are Thyroid Federation International, an international federation of thyroid patient associations, and the British Thyroid Foundation, which provides patient information [128, 129]. No patient association specific to resistance to thyroid hormone was identified in Portugal or Brazil in the sources consulted, where general thyroid disease associations exist. For research and referral, there are two active observational registries, DEEPTYPE in Berlin and ADAM-THAD in Milan, the latter dedicated to RTHβ and RTHα [104].

In France, RTHβ is followed by the reference centre for rare diseases of the thyroid and hormone receptors (CRMR-TRH), based at Angers University Hospital, and by its 24 competence centres, integrated into the FIRENDO rare endocrine disease health network; deafness may be referred to the reference centre for genetic deafness at the Necker-Enfants Malades Hospital (Paris) [41]. The care pathway is coordinated by the paediatric endocrinologist until age 18 and then by the adult endocrinologist, with the general practitioner in support, and draws, according to the manifestations, on cardiology, psychology, psychomotor therapy, psychiatry, paediatric neurology, rheumatology, otorhinolaryngology, thyroid surgery, obstetrics, hormonal and molecular biology, genetics, social services, speech therapy, hearing aid services and school support assistants [41]. The disease is not on the list of long-term conditions, and exemption from patient co-payment is requested as an off-list condition, with a care protocol established between the attending physician, the specialist and the health insurance medical adviser; there is no therapeutic education programme validated by the HAS, and explanation of the disease, results, transmission and the need for follow-up forms part of patient and family education [41]. There is no specific French association: Vivre Sans Thyroïde, integrated into FIRENDO and the reference centre, and the Association Française des Malades de la Thyroïde deal with thyroid diseases in general [41].

9.6. Special populations

In children, RTHβ requires assessment of growth and development by a paediatric endocrinologist, neuropsychological assessment for ADHD and cognitive deficits, educational support, audiometry and prompt treatment of ear, nose and throat infections [2]. In women of childbearing potential, RTHβ entails pregnancy planning and multidisciplinary follow-up, and tiratricol is contraindicated in pregnancy in the European Union [2, 93]. In adult men with selenoprotein deficiency, semen analysis documents azoospermia and infertility [2, 5]. In adults with RTHβ over 30 years of age, cardiovascular risk assessment is recommended at diagnosis and in continued surveillance (Section 11) [2]. In never-treated adults with RTHα, late diagnosis is associated with a more marked phenotype, which justifies looking for the disease in the adult relatives of an index case [38].

For disability support, the French PNDS provides for systematic neuropsychological, auditory and cardiovascular assessments in children, which may justify a school support assistant, support from an early medico-social action centre before age 6, follow-up in a medico-psycho-pedagogical centre or a special education and home care service, and an application to the departmental centre for disabled persons (MDPH), supported by a FIRENDO self-questionnaire on daily life difficulties; in adults, the same process allows recognition of disabled worker status or a compensation allowance, which in the authors' experience are rarely needed [41]. Regular and prolonged follow-up may justify school absences and time off work [41].

10. Illustrative Clinical Case

10.1. Extreme RTHβ due to a mosaic THRB variant

A Caucasian boy was born at 35 weeks and 3 days weighing 2402 g (50th centile) [53]. At 2 months, he was assessed for restlessness, excessive crying and reflux, with FT4 above 100 pmol/L, total T3 of 8.2 nmol/L and TSH of 33 mU/L, with negative TRAb; epileptic seizures began at 1.5 years [53]. Sequencing of THRB in leucocytes at 3 months was read as normal; review of the data three months later, prompted by strong clinical suspicion, showed small peaks in exon 10, corresponding to an 8-base-pair duplication (c.1305_1312dup in the article's nomenclature) causing a frameshift and premature stop codon (p.R438Lfs445X), present in a small fraction of reads, compatible with mosaicism; the variant was confirmed in buccal mucosa, at a lower fraction, and in fibroblasts [53].

At 8 years, the picture was one of severe developmental delay, refractory epilepsy, absence of walking and speech, continuous agitation with repetitive movements, mean HR of 105 bpm (+2.18 standard deviations) with atrial extrasystoles, chronic diarrhoea and continuous tube feeding because of nocturnal hypoglycaemia; weight was between −1 and −2 standard deviations, body mass index between 11.6 and 13.8 kg/m², height above the familial target and bone age advanced by about 4 years [53]. Tests showed T3 of 10.3 nmol/L (reference 1.7–2.9), FT4 of 132.4 pmol/L (12.7–23.3) and TSH of 5.88 mIU/L (0.60–5.22), with very high bone alkaline phosphatase and SHBG, indicating that the liver, where the fraction of cells with the variant would be lower, responded to the excess hormones; thyroid volume was 41 mL [53].

Treatment began with methimazole, 5 to 10 mg (0.2 to 0.4 mg/kg), preferred to propylthiouracil for its hepatic profile; TRIAC was added after 6 months, with a fall in FT4 and T3, but a rise in TSH and the appearance of goitre; TRIAC alone raised FT4 above the initial value [53]. The final regimen, methimazole 15 mg/day with TRIAC 3150 µg/day, brought FT4 close to normal, with mildly raised TSH and no obstructive goitre, over more than 18 months, without adverse effects apart from the goitre; alertness, agitation, weight, the need for tube feeding and the bone age advancement improved [53]. In functional studies, the mutant receptor did not respond to T3 in the TRβ1 and TRβ2 isoforms, and neither released the corepressors NCOR1 and SMRT nor recruited the coactivator SRC1 [53]. Figure 1 of the article shows the response to treatment and Figure 2 the receptor activity [53].

The TRIAC dose used in this patient, 3150 µg/day, exceeds the range of 1.4 to 2.8 mg/day recommended by the guideline for RTHβ, and combining an antithyroid drug with TRIAC is not the usual approach: the guideline recommends avoiding antithyroid drugs in RTHβ without significant comorbidity and reserves the combination for severe forms with cardiac risk, such as homozygous RTHβ [2, 57]. The case also shows that sequencing read as normal does not exclude the disease when the allele fraction is low, and the authors themselves propose combining an antithyroid drug with TRIAC for severely affected patients, which is a proposal based on a single case [53].

10.2. RTHα: two adolescents with variants of contrasting severity

The two patients described by Moran and colleagues in 2017 illustrate the correlation between the functional properties of the mutant receptor and the phenotype and response to treatment [22].

The first patient (P1) had a heterozygous de novo variant NM_199334.5:c.788C>T p.(Ala263Val) [22]. Labour was induced at 39 weeks and 4 days for macrosomia, with a birth weight of 4.58 kg (98th centile) and shoulder dystocia; weight remained at the 98th centile until 2 years, whereas length, initially at the 91st centile (56.5 cm), fell to the 9th centile (87 cm) at two and a half years [22]. At 8 months, the mother's concern about an excessively placid infant with slow movements was confirmed by mild developmental delay: he sat at 8 months, walked at 24 months and only spoke after 2 years; in childhood, coordination difficulties were documented at 9 years (motor skills below the 5th centile) and justified occupational therapy; several persistent primary teeth were extracted at 15 years [22]. Referral to paediatrics for suspected pubertal delay at 16 years, when he was already in mid-puberty, documented discordant thyroid function tests: TSH 3.20 mIU/L (reference 0.27–4.20), FT4 9.0 pmol/L (12–22) and FT3 8.0 pmol/L (3.9–6.7) [22]. On assessment, at 17 years, height was proportionate (50th centile, subischial leg length +1 standard deviation), with mild macroglossia, a slightly broad face, a few skin tags, head circumference of 62.5 cm (>99th centile), thickened calvarium, delayed dentition and loaded bowel loops on abdominal imaging, with normal femoral epiphyses; off treatment, FT4 was low-normal with high FT3, subnormal rT3, normal CK and low-normal resting energy expenditure [22]. With mild educational support he completed secondary school and entered university [22].

The second patient (P2) had the heterozygous de novo variant NM_199334.5:c.821T>C p.(Leu274Pro) [22]. He was born at 41 weeks and 5 days weighing 3.26 kg (9th centile), with coronal hypospadias and umbilical hernia; poor weight gain in the first 6 months (weight below the 0.4th centile) required high-calorie formula [22]. Developmental delay, noted at 9 weeks, persisted: he sat at 12 months, stood at 22 months and walked at two and a half years, with gross motor skills at the 12-month level and a two-word vocabulary on formal assessment at two and a half years [22]. He also had dysmorphic facies, truncal hypotonia, eruption of the first tooth at 2 years, myopic astigmatism, genu valgum, ankyloglossia, skeletal dysplasia (Wormian bones, short tubular bones of the hand, delayed closure of the anterior fontanelle, coxa valga, femoral epiphyseal dysgenesis, mesomelic limb shortening, bone age of 1 year and 3 months at 3 years and 3 months), bilateral hydroceles and constipation [22]. At 11 years, tests interpreted as central hypothyroidism (TSH 2.4 mIU/L, reference 0.4–5.5; FT4 9.0 pmol/L, reference 10–18.7) led to thyroxine being started, and puberty proceeded normally; he attends a special education school and has continuing learning difficulties [22]. At 15 years, investigation was resumed because of abnormal tests and global developmental delay: disproportionate short stature (0.4th centile, with normal sitting height and subischial leg length of −2 standard deviations), marked macroglossia, hypertelorism, flat nasal bridge, infiltrated skin, skin tags, head circumference of 60.5 cm (99th centile), multiple Wormian bones at 4 years with a now thickened calvarium, femoral epiphyseal dysgenesis no longer visible, loaded bowel loops and posterior scalloping of the lumbar vertebrae; before treatment, FT4 was subnormal with high FT3, low rT3 and raised CK [22].

Table 10.1. Comparison of the two patients with RTHα (Tables 1 and 2 of Moran 2017).

In functional assays, T3-dependent transcriptional activation by the p.(Ala263Val) mutant was reduced at low T3 concentrations (0.01 to 100 nM) and comparable to that of the normal receptor at high concentrations (1 to 10 µM), whereas the p.(Leu274Pro) mutant showed activity only at the highest concentration (10 µM); consistently, the first mutant released the corepressor at 100 nM and recruited the coactivator at 1 µM, against 10 µM for both in the second [22]. Expression of the target gene KLF9 in peripheral blood mononuclear cells normalised with 1 µM T3 in cells with p.(Ala263Val) and remained reduced with 10 µM in cells with p.(Leu274Pro) [22].

The response to thyroxine followed these differences. In P1, 18 months of treatment, with an appropriate rise in FT4, FT3 and rT3, were accompanied by improved linear growth and a marked reduction in weight (total fat mass from 26,950 g, 31%, to 20,520 g, 23%), with fasting insulin falling from 156 to 75 pmol/L (reference 0–60) and adiponectin rising from 3.4 to 7.1 µg/mL; the patient reported better concentration and school performance, more motivation and social confidence, better coordination and less constipation [22]. In P2, treated since age 11 with similar hormonal changes, linear growth barely changed and weight continued to increase, reversing only with the highest dose of 62.5 µg from age 14; he reported greater bowel regularity, but alertness, energy, cold extremities and the abnormal gait were unchanged [22]. In both patients, CK and total and LDL cholesterol fell, SHBG rose slightly and the subnormal haematocrit did not change [22]. Figures 1 and 2 of the article show height and facial features and the radiological changes (thickened calvarium, loaded bowel loops, delayed dentition, femoral epiphyseal dysgenesis, Wormian bones and vertebral scalloping), and Figure 5 the height and weight curves with the start of thyroxine [22].

TSH suppression, the fall in cholesterol and the rise in SHBG with treatment indicate a thyromimetic effect in TRβ-expressing tissues, such as the pituitary and liver, whose long-term consequences are unknown; this is the basis of the guideline's uncertainty about dose and targets in RTHα [2, 22]. The doses used (87.5 and 62.5 µg/day) are those of this report and do not constitute a recommendation, and the guideline considers the data insufficient to define a levothyroxine dose in RTHα [2].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis of RTHβ

RTHβ is often asymptomatic and many patients need no treatment, but it is not a benign biochemical curiosity [2, 37]. In the Welsh cohort of 61 genetically confirmed patients, all-cause mortality was 2.84 times that of matched controls (95% CI 1.59–5.08), atrial fibrillation 10.56 times (4.72–23.63), heart failure 6.35 times (2.26–17.86) and MACE 3.49 times (2.04–5.99), with the first adverse event occurring at a median age of 56 years against 67 in controls [39]. In the Italian cohort of 284 patients, the cumulative prevalence of sinus or supraventricular tachycardia was 40% and that of atrial fibrillation 18%, MACE occurred at a median age of 59.4 years and premature mortality corresponded to a mean of 11 years of life lost; on multivariate analysis, age at diagnosis, raised FT4 and male sex were associated with MACE [40]. In both cohorts, FT4 is the modifiable determinant identified, with increasing risk above 30 pmol/L, and previous thyroid ablation did not alter outcomes [39, 40]. In homozygous forms, thyrotoxic cardiomyopathy is the main cause of early death: in one of the siblings described by Moran and colleagues, heart failure was fatal at age 13 [57].

11.2. Prognosis of the other entities

In RTHα, never-treated adults have a more marked phenotype than patients treated since childhood, and the results of levothyroxine are better when treatment starts in early childhood; constipation and growth improve in most, the neurocognitive response is variable and anaemia changes little or not at all [2, 38, 108]. There are no mortality data in RTHα in the sources consulted. In selenoprotein deficiency, the prognosis depends on the progressive muscular dystrophy with risk of hypoventilation, the progressive hearing loss and the aneurysmal dilatation of the thoracic aorta, and most known patients are children or young adults, so the course in later adulthood is unknown [2, 88]. In the tRNA-Sec defect, the spectrum is not defined with two patients, and one of them remained asymptomatic over 6 years of untreated follow-up [2, 26]. D1 deficiency is, in the three published families, a biochemical phenotype without described clinical disease [2, 6].

11.3. Surveillance

Table 11.1. Recommended surveillance, by entity.

The European guideline does not set numerical intervals for most of these examinations, and the recommendations for periodic surveillance of the comorbidities of RTHβ (thyroid autoimmunity, tachyarrhythmias, low bone density and dyslipidaemia) are formulated as a strong recommendation with moderate quality of evidence, without a defined periodicity [2]. The numerical thresholds used here are the same as in Sections 3, 7 and 9: FT4 of 30 pmol/L as the cardiovascular risk threshold and maternal FT4 above 150% of the upper limit of normal as the threshold for considering antithyroid drugs in pregnancy [2, 39, 40].

The French PNDS sets, by expert consensus, an age-based schedule, always adjusted to the mode of discovery, age at diagnosis and complications (Table 11.2); the justification is that there are no data to predict the appearance or timing of neuropsychological, cardiovascular and thyroid complications [41]. The transition assessment, preferably as a day case, repeats the full assessment and defines the priorities of adult surveillance [41]. The PNDS intervals are compatible with the European guideline, which recommends periodic surveillance without a defined periodicity, and do not alter the action thresholds in Table 11.1 [2, 41].

Table 11.2. RTHβ surveillance schedule according to the French PNDS.

Source: section 5.3 and appendix 8 of the French PNDS; "Yes" indicates a one-off assessment at that time; pregnancy and the newborn of an affected parent are in Table 11.1 [41].

11.4. Preventing iatrogenic harm during follow-up

Long-term follow-up has to reckon with two recurrent errors. The first is interpreting hyperthyroxinaemia as primary hyperthyroidism and treating it with antithyroid drugs, radioiodine or thyroidectomy: ablation did not reduce cardiovascular events or mortality, necessitates supraphysiological levothyroxine to avoid pituitary thyrotroph hyperplasia and, in the case of serum transport protein variants, treats a euthyroid patient [39, 40, 78, 89]. The second is interpreting the low FT4 with normal TSH of RTHα as central hypothyroidism and starting levothyroxine at a replacement dose, which is insufficient to overcome the resistance of TRα-dependent tissues [2, 22]. In a Turkish paediatric series, one of 13 children was receiving levothyroxine for a misdiagnosis of hypothyroidism [37].

12. Summary and Concluding Remarks

The disorders in this article share a common mechanism, a reduction in the biological effect of a chemically intact thyroid hormone, and are separated by the step of the pathway affected: nuclear action in the resistances due to THRB and THRA, and intracellular metabolism in selenoprotein deficiency and type 1 deiodinase deficiency. What makes them recognisable is the discordance of the thyroid function tests. In RTHβ, free T4 and T3 are genuinely raised with non-suppressed TSH, because pituitary feedback depends on the mutant β receptor, and SHBG remains normal because the liver is equally resistant. In RTHα, circulating hormones are near-normal and the signal is a low T4/T3 ratio with low rT3, accompanied by disproportionate short stature, macrocephaly, constipation, dyspraxia and mild anaemia. In selenoprotein deficiency, deficient conversion of T4 to T3 produces high free T4 and rT3 with low or normal free T3, in a multisystem picture with low plasma selenium. In D1 deficiency, the abnormality is isolated, with raised rT3 and rT3/T3 ratio.

Before attributing a discordant profile to one of these genes, two steps cannot be omitted: ruling out drugs and intercurrent illness, and excluding, with the laboratory, analytical interference and serum transport protein variants. Familial dysalbuminaemic hyperthyroxinaemia was more frequent than RTHβ in a paediatric cohort with hyperthyroxinaemia and non-suppressed TSH and interferes with almost all free T4 immunoassays; thyroxine-binding globulin excess was the most common cause of persistent hyperthyroxinaemia in a T4-based newborn screening programme. Once genuine hyperthyroxinaemia is confirmed, distinguishing RTHβ from TSH-secreting pituitary adenoma requires tests in first-degree relatives, dynamic tests, pituitary imaging and, sometimes, a therapeutic trial of a somatostatin ligand. THRB sequencing confirms most cases, but a negative result does not exclude the disease: about 10% to 15% of families with the phenotype have no identifiable variant, and some of these correspond to somatic mosaicism that only next-generation sequencing, in blood and other tissues, reveals.

Treatment starts from a decision about the need to treat. In RTHβ, most patients need no specific therapy, and the error with consequences is treating hyperthyroxinaemia as primary hyperthyroidism: the European guideline recommends avoiding antithyroid drugs and ablation in the absence of significant comorbidity, and the two large cohorts showed that ablation did not reduce cardiovascular events or mortality. When there are thyrotoxic symptoms, cardioselective beta-blockade is used and, in selected cases discussed with experienced centres, tiratricol, at 1.4 to 2.8 mg per day divided into two or three doses, monitored by free T4, because the drug distorts T3 immunoassays. Tiratricol has an authorised product in the European Union since February 2025 and in the USA since September 2026, but only for MCT8 deficiency, so use in RTHβ is off-label, through a managed access programme or a compounded preparation, except in France, where tiratricol has a national authorisation for RTHβ. In RTHα, management is a therapeutic trial of levothyroxine in all patients, at a dose above replacement, with the caveat that TSH-suppressive doses expose TRβ-dependent tissues to a thyromimetic effect whose long-term consequence is unknown; results are better when treatment starts in early childhood. In selenoprotein deficiency, subnormal free T3 is treated with liothyronine, an antioxidant is considered and selenium is not supplemented, because a supplementation trial showed that selenium does not correct the incorporation defect. In D1 deficiency there is no established treatment.

The prognosis of RTHβ justifies active surveillance. All-cause mortality was almost three times that of controls in a record-linkage cohort, with the first adverse event 11 years earlier, and an independent cohort in another country found a mean of 11 years of life lost; free thyroxine is the modifiable determinant identified, with increasing risk above 30 pmol/L. From this follow cardiovascular risk assessment in all patients over 30 years, periodic metabolic assessment, thyroid ultrasound, densitometry in adults and, in children, audiometry, neuropsychological assessment and monitoring of growth. In the pregnancy of a woman with RTHβ, the risk falls on the unaffected fetus, which suffers excessive hormone exposure: in a large Azorean family, the miscarriage rate in affected mothers was 22.9% and newborns without the variant were significantly smaller, with undetectable TSH on screening. In France, the 2024 national protocol sets an age-based surveillance schedule, from 6 to 12 months in the young child to 3 to 5 years in the adult, and diverges from the European guideline on the preferred beta-blocker, the use of antithyroid drugs in pregnancy and prenatal diagnosis.

Finally, what is not known defines the clinical terrain. There is no validated prevalence estimate for any of these disorders: the available figures are local screening incidences with two to four cases and cumulative counts of published cases, and detection depends on the screening programme measuring T4 and not just TSH. There is no registered interventional trial in these disorders, only two observational registries, and all therapy rests on cases, series and retrospective cohorts from reference centres, with their inherent biases. Each of these gaps is an argument for these patients to be followed in experienced centres, linked to the European reference network for rare endocrine conditions, and included in the existing registries.

13. References

1. Refetoff S, Bassett JH, Beck-Peccoz P, Bernal J, Brent G, Chatterjee K, et al. Classification and proposed nomenclature for inherited defects of thyroid hormone action, cell transport, and metabolism. Thyroid. 2014;24(3):407–9. doi:10.1089/thy.2013.3393.nomen
2. Persani L, Rodien P, Moran C, Visser WE, Groeneweg S, Peeters R, et al. 2024 European Thyroid Association Guidelines on diagnosis and management of genetic disorders of thyroid hormone transport, metabolism and action. Eur Thyroid J 2024;13(4):ETJ240125. doi:10.1530/etj-24-0125. PMID: 38963712.
3. Orphanet. Generalized resistance to thyroid hormone. ORPHA:3221. Available from: https://www.orpha.net/en/disease/detail/3221
4. Orphanet. Pituitary resistance to thyroid hormone. ORPHA:165994. Available from: https://www.orpha.net/en/disease/detail/165994
5. Schoenmakers E, Agostini M, Mitchell C, Schoenmakers N, Papp L, Rajanayagam O, et al. Mutations in the selenocysteine insertion sequence-binding protein 2 gene lead to a multisystem selenoprotein deficiency disorder in humans. J Clin Invest. 2010;120(12):4220–35. doi:10.1172/JCI43653
6. França MM, German A, Fernandes GW, Liao XH, Bianco AC, Refetoff S, et al. Human type 1 iodothyronine deiodinase (DIO1) mutations cause abnormal thyroid hormone metabolism. Thyroid. 2021;31(2):202–7. doi:10.1089/thy.2020.0253
7. GeneGuidance. THRB — gene record (thyroid hormone resistance, OMIM 188570/274300/145650; ORPHA 566243). Version of 2026-09-26.
8. OMIM® Online Mendelian Inheritance in Man. *190160 THYROID HORMONE RECEPTOR, BETA; THRB. Johns Hopkins University; last edited 03/02/2025. https://omim.org/entry/190160. Accessed 28 October 2026.
9. VariantValidator. Hybrid variant validation tool, GRCh38, MANE Select. Available from: https://variantvalidator.org/
10. GeneGuidance. THRA — gene and condition record (Hypothyroidism, congenital, nongoitrous, 6; OMIM 614450; ORPHA 566231). Version of 2026-09-26.
11. OMIM® Online Mendelian Inheritance in Man. *190120 THYROID HORMONE RECEPTOR, ALPHA-1; THRA. Johns Hopkins University; last edited 08/03/2022. https://omim.org/entry/190120. Accessed 28 October 2026.
12. GeneGuidance. SECISBP2 — gene and condition record (Thyroid hormone metabolism, abnormal; OMIM 609698; ORPHA 171706). Version of 2026-09-26.
13. OMIM® Online Mendelian Inheritance in Man. *607693 SELENOCYSTEINE INSERTION SEQUENCE-BINDING PROTEIN 2; SECISBP2. Johns Hopkins University; last edited 25/04/2022. https://omim.org/entry/607693. Accessed 28 October 2026.
14. GeneGuidance. TRU-TCA1-1 — gene record (Thyroid hormone metabolism, abnormal, 3; OMIM 620198). Version of 2026-09-26.
15. OMIM® Online Mendelian Inheritance in Man. *165060 tRNA SELENOCYSTEINE (ANTICODON TCA) 1-1; TRU-TCA1-1. Johns Hopkins University; last edited 10/01/2023. https://omim.org/entry/165060. Accessed 28 October 2026.
16. GeneGuidance. DIO1 — gene record (Abnormal thyroid hormone metabolism 2; OMIM 619855). Version of 2026-09-26.
17. OMIM® Online Mendelian Inheritance in Man. *147892 DEIODINASE, IODOTHYRONINE, TYPE I; DIO1. Johns Hopkins University; last edited 10/01/2023. https://omim.org/entry/147892. Accessed 28 October 2026.
18. Mamanasiri S, Yesil S, Dumitrescu AM, Liao XH, Demir T, Weiss RE, et al. Mosaicism of a thyroid hormone receptor-beta gene mutation in resistance to thyroid hormone. J Clin Endocrinol Metab. 2006;91(9):3471–7. doi:10.1210/jc.2006-0727
19. Ferrara AM, Onigata K, Ercan O, Woodhead H, Weiss RE, Refetoff S. Homozygous thyroid hormone receptor β-gene mutations in resistance to thyroid hormone: three new cases and review of the literature. J Clin Endocrinol Metab. 2012;97(4):1328–36. doi:10.1210/jc.2011-2642
20. Takeda K, Sakurai A, DeGroot LJ, Refetoff S. Recessive inheritance of thyroid hormone resistance caused by complete deletion of the protein-coding region of the thyroid hormone receptor-beta gene. J Clin Endocrinol Metab. 1992;74(1):49–55. doi:10.1210/jcem.74.1.1727829
21. OMIM® Online Mendelian Inheritance in Man. #274300 THYROID HORMONE RESISTANCE, GENERALIZED, AUTOSOMAL RECESSIVE; GRTHR. Johns Hopkins University; last edited 12/12/2024. https://omim.org/entry/274300. Accessed 28 October 2026.
22. Moran C, Agostini M, McGowan A, Schoenmakers E, Fairall L, Lyons G, et al. Contrasting phenotypes in resistance to thyroid hormone alpha correlate with divergent properties of thyroid hormone receptor α1 mutant proteins. Thyroid. 2017;27(7):973–82. doi:10.1089/thy.2017.0157
23. OMIM® Online Mendelian Inheritance in Man. #614450 HYPOTHYROIDISM, CONGENITAL, NONGOITROUS, 6; CHNG6. Johns Hopkins University; last edited 05/09/2019. https://omim.org/entry/614450. Accessed 28 October 2026.
24. OMIM® Online Mendelian Inheritance in Man. #609698 THYROID HORMONE METABOLISM, ABNORMAL, 1; THMA1. Johns Hopkins University; last edited 10/01/2023. https://omim.org/entry/609698. Accessed 28 October 2026.
25. OMIM® Online Mendelian Inheritance in Man. #620198 THYROID HORMONE METABOLISM, ABNORMAL, 3; THMA3. Johns Hopkins University; last edited 10/01/2023. https://omim.org/entry/620198. Accessed 28 October 2026.
26. Geslot A, Savagner F, Caron P. Inherited selenocysteine transfer RNA mutation: clinical and hormonal evaluation of 2 patients. Eur Thyroid J. 2021;10(6):542–7. doi:10.1159/000518275
27. OMIM® Online Mendelian Inheritance in Man. #619855 THYROID HORMONE METABOLISM, ABNORMAL, 2; THMA2. Johns Hopkins University; last edited 26/04/2022. https://omim.org/entry/619855. Accessed 28 October 2026.
28. OMIM® Online Mendelian Inheritance in Man. #188570 THYROID HORMONE RESISTANCE, GENERALIZED, AUTOSOMAL DOMINANT; GRTHD. Johns Hopkins University; last edited 12/12/2024. https://omim.org/entry/188570. Accessed 28 October 2026.
29. OMIM® Online Mendelian Inheritance in Man. #145650 THYROID HORMONE RESISTANCE, SELECTIVE PITUITARY; PRTH. Johns Hopkins University; last edited 08/11/2024. https://omim.org/entry/145650. Accessed 28 October 2026.
30. Orphanet. Resistance to thyroid hormone due to a mutation in thyroid hormone receptor beta. ORPHA:566243 (active; ICD-10 E07.8). Available from: https://www.orpha.net/en/disease/detail/566243
31. Orphanet. Resistance to thyroid hormone due to a mutation in thyroid hormone receptor alpha. ORPHA:566231 (active; ICD-10 E03.8). Available from: https://www.orpha.net/en/disease/detail/566231
32. Orphanet. Short stature-delayed bone age due to thyroid hormone metabolism deficiency. ORPHA:171706 (active; ICD-10 E03.1). Available from: https://www.orpha.net/en/disease/detail/171706
33. Sarret C, Oliver Petit I, Tonduti D. Allan-Herndon-Dudley Syndrome. 2010 Mar 9 [updated 2020 Jan 16]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK26373/
34. LaFranchi SH, Snyder DB, Sesser DE, Skeels MR, Singh N, Brent GA, et al. Follow-up of newborns with elevated screening T4 concentrations. J Pediatr. 2003;143(3):296–301. doi:10.1067/S0022-3476(03)00184-7
35. Vela A, Pérez-Nanclares G, Ríos I, Rica I, Portillo N, Castaño L, et al. Thyroid hormone resistance from newborns to adults: a Spanish experience. J Endocrinol Invest. 2019;42(8):941–9. doi:10.1007/s40618-019-1007-4
36. Fu J, Korwutthikulrangsri M, Gönç EN, Sillers L, Liao XH, Alikaşifoğlu A, et al. Clinical and molecular analysis in 2 families with novel compound heterozygous SBP2 (SECISBP2) mutations. J Clin Endocrinol Metab. 2020;105(3):e6–e11. doi:10.1210/clinem/dgz169
37. Kolbaşı B, Unal E, Durak H, Akalın A, Gider T, Yıldırım R, et al. Clinical and laboratory characteristics of pediatric patients diagnosed with thyroid hormone resistance: a single-center experience. Endocrine. 2026;91(1):210. doi:10.1007/s12020-026-04680-x
38. Dahll LK, Westbye AB, Vinorum K, Sejersted Y, Barøy T, Thorsby PM, et al. Clinical and biochemical characteristics of untreated adult patients with resistance to thyroid hormone alpha. J Endocr Soc. 2023;7(8):bvad089. doi:10.1210/jendso/bvad089
39. Okosieme OE, Usman D, Taylor PN, Dayan CM, Lyons G, Moran C, et al. Cardiovascular morbidity and mortality in patients in Wales, UK with resistance to thyroid hormone β (RTHβ): a linked-record cohort study. Lancet Diabetes Endocrinol. 2023;11(9):657–66. doi:10.1016/S2213-8587(23)00155-9
40. Campi I, Censi S, Prodam F, Petrone L, Brigante G, Porcelli T, et al. Increased cardiovascular morbidity and reduced life expectancy in a large Italian cohort of patients with resistance to thyroid hormone β (RTHβ). Eur J Endocrinol. 2024;191(4):407–15. doi:10.1093/ejendo/lvae117
41. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Syndrome de Résistance aux Hormones Thyroïdiennes par variant pathogène de THRB. Saint-Denis La Plaine: HAS; 09/02/2024 (text by the Centre de Référence des Maladies Rares de la Thyroïde et des Récepteurs Hormonaux, version of 10/01/2024). Available from: https://www.has-sante.fr/jcms/p_3493351
42. Schomburg L, Dumitrescu AM, Liao XH, Bin-Abbas B, Hoeflich J, Köhrle J, et al. Selenium supplementation fails to correct the selenoprotein synthesis defect in subjects with SBP2 gene mutations. Thyroid. 2009;19(3):277–81. doi:10.1089/thy.2008.0397
43. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Thyroid main thematic group. Available from: https://endo-ern.eu/specific-expertise/thyroid/
44. Refetoff S, DeWind LT, DeGroot LJ. Familial syndrome combining deaf-mutism, stippled epiphyses, goiter and abnormally high PBI: possible target organ refractoriness to thyroid hormone. J Clin Endocrinol Metab. 1967;27(2):279–94. doi:10.1210/jcem-27-2-279
45. Chatterjee VK, Nagaya T, Madison LD, Datta S, Rentoumis A, Jameson JL. Thyroid hormone resistance syndrome. Inhibition of normal receptor function by mutant thyroid hormone receptors. J Clin Invest. 1991;87(6):1977–84. doi:10.1172/JCI115225
46. Usala SJ, Menke JB, Watson TL, Wondisford FE, Weintraub BD, Berard J, et al. A homozygous deletion in the c-erbA beta thyroid hormone receptor gene in a patient with generalized thyroid hormone resistance. Mol Endocrinol. 1991;5(3):327–35.
47. Bochukova E, Schoenmakers N, Agostini M, Schoenmakers E, Rajanayagam O, Keogh JM, et al. A mutation in the thyroid hormone receptor alpha gene. N Engl J Med. 2012;366(3):243–9. doi:10.1056/NEJMoa1110296
48. Moran C, Agostini M, Visser WE, Schoenmakers E, Schoenmakers N, Offiah AC, et al. Resistance to thyroid hormone caused by a mutation in thyroid hormone receptor (TR)α1 and TRα2: clinical, biochemical, and genetic analyses of three related patients. Lancet Diabetes Endocrinol. 2014;2(8):619–26. doi:10.1016/S2213-8587(14)70111-1
49. Dumitrescu AM, Liao XH, Abdullah MSY, Lado-Abeal J, Majed FA, Moeller LC, et al. Mutations in SECISBP2 result in abnormal thyroid hormone metabolism. Nat Genet. 2005;37(11):1247–52. doi:10.1038/ng1654
50. Schoenmakers E, Carlson B, Agostini M, Moran C, Rajanayagam O, Bochukova E, et al. Mutation in human selenocysteine transfer RNA selectively disrupts selenoprotein synthesis. J Clin Invest. 2016;126(3):992–6. doi:10.1172/JCI84747
51. Orphanet. OBSOLETE: Peripheral resistance to thyroid hormones. ORPHA:97927. Available from: https://www.orpha.net/en/disease/detail/97927
52. OMIM® Online Mendelian Inheritance in Man. *300095 SOLUTE CARRIER FAMILY 16 (MONOCARBOXYLIC ACID TRANSPORTER), MEMBER 2; SLC16A2. Johns Hopkins University; last edited 23/06/2017. https://omim.org/entry/300095. Accessed 28 October 2026.
53. van Geest FS, Liao W, Voorhoeve PG, Leen WG, Zwaveling-Soonawala N, Chatterjee VK, et al. Extreme resistance to thyroid hormone caused by a novel mosaic thyroid hormone receptor beta mutation. Eur Thyroid J. 2025;14(6):e250092. doi:10.1530/ETJ-25-0092
54. Liao W, Marelli F, Moran C, Campi I, Groeneweg S, Mohamed Z, et al. Molecular mechanisms underlying the effectiveness of triiodothyroacetic acid (Triac) in patients with resistance to thyroid hormone β (RTHβ): a multicentre, international, retrospective cohort study. EBioMedicine. 2026;133:106512. doi:10.1016/j.ebiom.2026.106512
55. Brucker-Davis F, Skarulis MC, Grace MB, Benichou J, Hauser P, Wiggs E, et al. Genetic and clinical features of 42 kindreds with resistance to thyroid hormone. The National Institutes of Health prospective study. Ann Intern Med. 1995;123(8):572–83. doi:10.7326/0003-4819-123-8-199510150-00002
56. Pohlenz J, Weiss RE, Macchia PE, Pannain S, Lau IT, Ho H, et al. Five new families with resistance to thyroid hormone not caused by mutations in the thyroid hormone receptor beta gene. J Clin Endocrinol Metab. 1999;84(11):3919–28. doi:10.1210/jcem.84.11.6080
57. Moran C, Habeb AM, Kahaly GJ, Kampmann C, Hughes M, Marek J, et al. Homozygous resistance to thyroid hormone β: can combined antithyroid drug and triiodothyroacetic acid treatment prevent cardiac failure? J Endocr Soc. 2017;1(9):1203–12. doi:10.1210/js.2017-00204
58. Donnars A, Leplat A, Grosheny C, Briet C, Illouz F, Bouzamondo N, et al. Clinically symptomatic resistance to thyroid hormone β syndrome because of THRB gene mosaicism. J Clin Endocrinol Metab. 2022;107(9):e3548–52. doi:10.1210/clinem/dgac347
59. Adams M, Matthews C, Collingwood TN, Tone Y, Beck-Peccoz P, Chatterjee KK. Genetic analysis of 29 kindreds with generalized and pituitary resistance to thyroid hormone. Identification of thirteen novel mutations in the thyroid hormone receptor beta gene. J Clin Invest. 1994;94(2):506–15. doi:10.1172/JCI117362. PMID: 8040303. PMCID: PMC296123.
60. Forrest D, Ng L, Cho YW, Liu Y, Liu H. The human THRB thyroid hormone receptor gene and the puzzle of retinal disease phenotypes. Dev Biol. 2026;535:93–110. doi:10.1016/j.ydbio.2026.04.004
61. Mahler EA, Moeller LC, Wall K, Saßmannshausen M, Kron B, Bolz HJ, et al. Mutation of the thyroid hormone receptor beta gene (THRB) causes vitelliform macular dystrophy with high intrafamilial variability. Genes (Basel). 2025;16(10):1240. doi:10.3390/genes16101240
62. Hönes GS, Liao XH, Mahler EA, Herrmann P, Eckstein A, Führer D, et al. THRB splice site variants lead to exon 4 skipping and TRβ1 gain-of-function syndrome. medRxiv [preprint]. 2026 Apr 15:2026.04.15.26349265. doi:10.64898/2026.04.15.26349265
63. Sun H, Wu H, Xie R, Wang F, Chen T, Chen X, et al. New case of thyroid hormone resistance α caused by a mutation of THRA/TRα1. J Endocr Soc. 2019;3(3):665–9. doi:10.1210/js.2019-00011
64. van Hasselt PM, Clayton PT, Houwen RHJ. Disorders in the Transport of Copper, Iron, Magnesium, Manganese, Selenium and Zinc. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 34, section 34.5 Selenium.
65. National Center for Biotechnology Information. ClinVar. Bethesda (MD): NLM. Available from: https://www.ncbi.nlm.nih.gov/clinvar/
66. Vindry C, Guillin O, Wolff P, Marie P, Mortreux F, Mangeot PE, et al. A homozygous mutation in the human selenocysteine tRNA gene impairs UGA recoding activity and selenoproteome regulation by selenium. Nucleic Acids Res. 2023;51(14):7580–601. doi:10.1093/nar/gkad482
67. Furman AE, Hannoush Z, Echegoyen FB, Dumitrescu A, Refetoff S, Weiss RE. Novel DIO1 gene mutation acting as phenotype modifier for novel compound heterozygous TPO gene mutations causing congenital hypothyroidism. Thyroid. 2021;31(10):1589–91. doi:10.1089/thy.2021.0210
68. Panicker V, Cluett C, Shields B, Murray A, Parnell KS, Perry JR, et al. A common variation in deiodinase 1 gene DIO1 is associated with the relative levels of free thyroxine and triiodothyronine. J Clin Endocrinol Metab. 2008;93(8):3075–81. doi:10.1210/jc.2008-0397
69. GeneGuidance. DIO2 — gene record. Version of 2026-09-26.
70. GeneGuidance. DIO3 — gene record. Version of 2026-09-26.
71. Sekizkardes H, Eltan M, Demir S, Giray D, Yavas Abali Z, Abali S. Diagnostic utility of genetic testing in hyperthyroxinemia with non-suppressed TSH. Clin Endocrinol (Oxf). 2026;104(6):693–701. doi:10.1111/cen.70126
72. Anselmo J, Cao D, Karrison T, Weiss RE, Refetoff S. Fetal loss associated with excess thyroid hormone exposure. JAMA. 2004;292(6):691–5. doi:10.1001/jama.292.6.691
73. Blair JC, Mohan U, Larcher VF, Rajanayagam O, Burrin JM, Perry LA, et al. Neonatal thyrotoxicosis and maternal infertility in thyroid hormone resistance due to a mutation in the TRbeta gene (M313T). Clin Endocrinol (Oxf). 2002;57(3):405–9. doi:10.1046/j.1365-2265.2002.01588.x
74. Di Geronimo C, Landi M, Mugnai F, Masutti M, Baldassarri M, Fallerini C, et al. Early-onset thyroid hormone resistance beta: a neonatal case with THRB (c.949G>A; p.Ala317Thr) pathogenic variant. J Paediatr Child Health. 2026. doi:10.1111/jpc.70624
75. Pan YA, Wang L, Malabu UH, Rangaswamaiah S. A case of thyroid hormone resistance with coincidental pituitary macroadenoma. JCEM Case Rep. 2026;4(4):luag051. doi:10.1210/jcemcr/luag051
76. Özdemir Akgün S, Bolaç Özyılmaz GL, Çetinçelik Ü, Dağdeviren Çakır A, Uçar A. A rare cause of thyroid hormone abnormalities in an adolescent: a case report of SBP2 (SECISBP2) deficiency. Mol Syndromol. 2026. doi:10.1159/000551282
77. Yang J, Ahn JM, Jeong HR. Thyroid hormone metabolism defect due to compound heterozygous SECISBP2 mutations: first reported case in Korea. J Pediatr Endocrinol Metab. 2026;39(3):291–6. doi:10.1515/jpem-2025-0500
78. Khoo S, Lyons G, McGowan A, Gurnell M, Oddy S, Visser WE, et al. Familial dysalbuminaemic hyperthyroxinaemia interferes with current free thyroid hormone immunoassay methods. Eur J Endocrinol. 2020;182(6):533–8. doi:10.1530/EJE-19-1021
79. OMIM® Online Mendelian Inheritance in Man. #615999 HYPERTHYROXINEMIA, FAMILIAL DYSALBUMINEMIC; FDAH. Johns Hopkins University; last edited 04/05/2021. https://omim.org/entry/615999. Accessed 28 October 2026.
80. GeneGuidance. ALB — gene record (familial dysalbuminemic hyperthyroxinemia, OMIM 615999). Version of 2026-09-26.
81. Đukić M, Kovačević Z, Tomšić KZ, Muzurović E. Thyroid hormone resistance beta and autoimmune thyroid disease — a family case study highlighting diagnostic and therapeutic challenges. Cardiovasc Diabetol Endocrinol Rep. 2026;12(1):23. doi:10.1186/s40842-026-00291-7
82. Davis TME, Davis WA, Moran C, Lyons G, Bryden E, Chatterjee K. Characteristics of dyslipidemia in adults with resistance to thyroid hormone β. Endocr Pract. 2025;31(11):1395–8. doi:10.1016/j.eprac.2025.07.001
83. Saudubray JM, García-Cazorla Á. Clinical Approach to Inborn Errors of Metabolism in Paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, sections 1.6.3 (Table 1.35) and 1.6.12 (Table 1.43).
84. Cardoso LF, de Paula FJ, Maciel LM. Resistance to thyroid hormone due to mutations in the THRB gene impairs bone mass and affects calcium and phosphorus homeostasis. Bone. 2014;67:222–7. doi:10.1016/j.bone.2014.07.017
85. Hauser P, Zametkin AJ, Martinez P, Vitiello B, Matochik JA, Mixson AJ, et al. Attention deficit-hyperactivity disorder in people with generalized resistance to thyroid hormone. N Engl J Med. 1993;328(14):997–1001. doi:10.1056/NEJM199304083281403
86. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 116–119 (Congenital hypothyroidism).
87. Jorge AAL, Andrade NLM, Rezende RC, Crisostomo LG, Dantas NCB, Cellin LP, et al. Clinical characteristics of children with THRA mutations: variable phenotype and good response to recombinant human growth hormone therapy. Horm Res Paediatr. 2025;98(5):457–64. doi:10.1159/000539348
88. Schoenmakers E, Marelli F, Jørgensen HF, Visser WE, Moran C, Groeneweg S, et al. Selenoprotein deficiency disorder predisposes to aortic aneurysm formation. Nat Commun. 2023;14:7994. doi:10.1038/s41467-023-43851-6
89. Moran C, Martin-Grace J, Lyons G, Watson L, Taylor K, Oddy S, et al. TRIAC therapy relieves hyperthyroid symptoms, lowering T4, T3, and metabolic rate in resistance to thyroid hormone β. J Clin Endocrinol Metab. 2026;111(4):980–7. doi:10.1210/clinem/dgaf583
90. Choi JH, Cho JH, Kim JH, Yoo EG, Kim GH, Yoo HW. Variable clinical characteristics and molecular spectrum of patients with syndromes of reduced sensitivity to thyroid hormone: genetic defects in the THRB and SLC16A2 genes. Horm Res Paediatr. 2018;90(5):283–90. doi:10.1159/000493468
91. Campi I, Feldt-Rasmussen U, Gruson D, Halsall D, Raverot V, van den Berg S, et al. 2026 ETA guideline on interference in immunoassay measurements used in assessment of thyroid function. Eur Thyroid J. 2026;15(4):e260011. doi:10.1530/ETJ-26-0011
92. U.S. Food and Drug Administration. SYNTHROID (levothyroxine sodium) tablets — Prescribing Information. NDA 021402; label version effective 2024-02-20. Available via openFDA: https://api.fda.gov/drug/label.json (set_id 1e11ad30-1041-4520-10b0-8f9d30d30fcc)
93. European Medicines Agency. Emcitate (tiratricol): EPAR — medicine overview and product information (Annex I: Summary of Product Characteristics). Amsterdam: EMA; marketing authorisation 12/02/2025; EMA/1545/2025. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/emcitate
94. Beck-Peccoz P, Lania A, Beckers A, Chatterjee K, Wemeau JL. 2013 European Thyroid Association guidelines for the diagnosis and treatment of thyrotropin-secreting pituitary tumors. Eur Thyroid J. 2013;2(2):76–82. doi:10.1159/000351007
95. Pappa T, Anselmo J, Mamanasiri S, Dumitrescu AM, Weiss RE, Refetoff S. Prenatal diagnosis of resistance to thyroid hormone and its clinical implications. J Clin Endocrinol Metab. 2017;102(10):3775–82. doi:10.1210/jc.2017-01251
96. Tzoraki M, Zannettouli AT, Vildiridi A, Karampas G, Metallinou D, Duntas L. Thyrotropin-secreting pituitary adenomas in pregnancy: insights from a systematic review. Hormones (Athens). 2025;24(4):1023–34. doi:10.1007/s42000-025-00705-8
97. Caridi G, Lugani F, Angeletti A, Campagnoli M, Galliano M, Minchiotti L. Variations in the human serum albumin gene: molecular and functional aspects. Int J Mol Sci. 2022;23(3):1159. doi:10.3390/ijms23031159
98. GeneGuidance. SERPINA7 — gene record (thyroxine-binding globulin QTL, OMIM 300932). Version of 2026-09-26.
99. Pappa T, Ferrara AM, Refetoff S. Inherited defects of thyroxine-binding proteins. Best Pract Res Clin Endocrinol Metab. 2015;29(5):735–47. doi:10.1016/j.beem.2015.09.002
100. OMIM® Online Mendelian Inheritance in Man. #300932 THYROXINE-BINDING GLOBULIN QUANTITATIVE TRAIT LOCUS; TBGQTL. Johns Hopkins University; last edited 11/01/2022. https://omim.org/entry/300932. Accessed 28 October 2026.
101. Mandel S, Hanna C, Boston B, Sesser D, LaFranchi S. Thyroxine-binding globulin deficiency detected by newborn screening. J Pediatr. 1993;122(2):227–30. doi:10.1016/s0022-3476(06)80117-4
102. GeneGuidance. TTR — gene record. Version of 2026-09-26.
103. Mimoto MS, Refetoff S. Clinical recognition and evaluation of patients with inherited serum thyroid hormone-binding protein mutations. J Endocrinol Invest. 2020;43(1):31–41. doi:10.1007/s40618-019-01084-9
104. U.S. National Library of Medicine. ClinicalTrials.gov (API v2). Available from: https://clinicaltrials.gov/
105. U.S. Food and Drug Administration. TENORMIN (atenolol) tablets — Prescribing Information. NDA 018240; label effective 2025-09-18. Available via openFDA (set_id 6e850b1e-28a4-4ff8-8ed7-bb0b29faa28a)
106. U.S. Food and Drug Administration. Propranolol hydrochloride tablets — Prescribing Information. ANDA 071972; label effective 2024-08-09. Available via openFDA (set_id 013a485a-4227-46ae-b19f-c550302fcb96)
107. Sidana S, Singh R, Gupta S, Goyal G. Resistance to thyroid hormone: clinical and biochemical paradox. BMJ Case Rep. 2025;18(11):e267068. doi:10.1136/bcr-2025-267068
108. Erbaş IM, Çakır MD, Yener AS, Demir K. Long-term follow-up results and treatment outcomes of children and adults with resistance to thyroid hormone alpha. J Endocrinol Invest. 2023;46(9):1855–63. doi:10.1007/s40618-023-02043-1
109. Westbye AB, Dahll LK, Bredahl MK, Thorsby PM, Hammerstad SS. Thyroxine treatment of adult RTHα patients: safety, efficacy, and metabolomic changes. Thyroid. 2026;36(7):788–94. doi:10.1177/10507256261460201
110. U.S. Food and Drug Administration. CYTOMEL (liothyronine sodium) tablets — Prescribing Information. NDA 010379; label version effective 2026-05-18. Available via openFDA (set_id 51452b31-ff68-4e0c-b982-c15502ebf1d3)
111. U.S. Food and Drug Administration. EMCITATE (tiratricol) tablets, for oral suspension — Highlights of Prescribing Information; and NDA 220963 approval letter. Silver Spring (MD): FDA; revised 09/2026. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/label/2026/220963Orig1s000lbl.pdf and https://www.accessdata.fda.gov/drugsatfda_docs/appletter/2026/220963Orig1s000ltr.pdf
112. INFARMED. Eutirox 25–200 microgramas comprimidos (levotiroxina sódica) — Resumo das Características do Medicamento. Marketing authorisation holder: Merck, S.A. Lisboa: INFARMED, Infomed (med_id 31940). Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=31940&tipo_doc=rcm
113. INFARMED. Inderal 40 mg comprimidos revestidos (propranolol) — Resumo das Características do Medicamento (approved 8 July 2022). Marketing authorisation holder: Atnahs Pharma Netherlands B.V. Lisboa: INFARMED, Infomed (med_id 10157). Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=10157&tipo_doc=rcm
114. European Medicines Agency. Hemangiol (propranolol) 3.75 mg/mL oral solution: EPAR — product information (Annex I: Summary of Product Characteristics). Amsterdam: EMA; first authorisation 23/04/2014, latest renewal 15/01/2019. Available from: https://www.ema.europa.eu/en/documents/product-information/hemangiol-epar-product-information_en.pdf
115. INFARMED. Tenormin 100 mg e Tenormin Mite 50 mg comprimidos revestidos por película (atenolol) — Resumo das Características do Medicamento (text revised 28 January 2026). Marketing authorisation holder: Atnahs Pharma Netherlands B.V. Lisboa: INFARMED, Infomed (med_id 8361). Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=8361&tipo_doc=rcm
116. U.S. Food and Drug Administration. RITALIN (methylphenidate hydrochloride) tablets — Prescribing Information. NDA 010187; label effective 2025-02-05. Available via openFDA (set_id d6fb2750-cdab-4749-ba0d-7534840a5892)
117. U.S. Food and Drug Administration. Methimazole tablets — Prescribing Information. Label effective 2026-08-20 (AiPing Pharmaceutical). Available via openFDA (set_id 06bb4bcc-5e47-48b6-a76c-648f846704e8)
118. U.S. Food and Drug Administration. Propylthiouracil tablets — Prescribing Information (boxed warning: hepatotoxicity). ANDA 080016; label effective 2023-12-26. Available via openFDA (set_id 53c5f586-3e72-479d-89ba-379683c8f6ba)
119. INFARMED. Infomed — Emcitate 350 µg comprimido dispersível (EMEA/H/C/5220; ATC H03AA04; registration 5928668; not marketed). Lisboa: INFARMED. Available from: https://extranet.infarmed.pt/INFOMED-fo/
120. INFARMED. Infomed — Neo-Tiroimade (liothyronine) 0.005 mg and 0.025 mg, marketing authorisation revoked, no marketed presentations (med_id 5962, 5963). Lisboa: INFARMED.
121. Agência Nacional de Vigilância Sanitária. Open data — registered medicines (DADOS_ABERTOS_MEDICAMENTOS.csv): tiratricol TRIAC (reg. 105730058, Aché) and TRIMAG (reg. 104971349, União Química), inactive status; liothyronine TIROIDAN (reg. 105830188, Germed), inactive. Brasília: ANVISA. Available from: https://dados.anvisa.gov.br/dados/
122. Agence nationale de sécurité du médicament et des produits de santé (ANSM). Emcitate (tiratricol) — accès dérogatoires: autorisation d'accès compassionnel arrêtée le 29/05/2025; autorisation d'accès précoce refusée le 24/07/2025; AMM le 12/02/2025. Saint-Denis: ANSM; publié le 03/11/2020, mis à jour le 12/03/2026. Available from: https://ansm.sante.fr/tableau-acces-derogatoire/emcitate
123. Schweizer U, Fabiano M. Genetics of selenoproteins and selenoprotein metabolism — an overview of current concepts and emerging aspects. Redox Biol. 2026;91:104083. doi:10.1016/j.redox.2026.104083
124. Yang J, Wang C, Quan L, Jiang S. Identification of a novel THRB mutation causing thyroid hormone resistance syndrome. Arch Endocrinol Metab. 2026;70(2):e260006. doi:10.20945/2359-4292-2026-0006
125. Weiss RE, Stein MA, Refetoff S. Behavioral effects of liothyronine (L-T3) in children with attention deficit hyperactivity disorder in the presence and absence of resistance to thyroid hormone. Thyroid. 1997;7(3):389-93. doi:10.1089/thy.1997.7.389. PMID: 9226208.
126. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Reference centres. Available from: https://endo-ern.eu/about/reference-centres/
127. European Thyroid Association. Available from: https://www.eurothyroid.com/
128. Thyroid Federation International. Available from: https://www.thyroid-federation.org/
129. British Thyroid Foundation. Available from: https://www.btf-thyroid.org/