clinical and genetic characteristics of a large monocentric series of patients affected by thyroid...

O R I G I N A L A R T I C L E

Clinical and genetic characteristics of a large monocentric seriesof patients affected by thyroid hormone (Th) resistance andsuggestions for differential diagnosis in patients withoutmutation of Th receptor b

Enrico Macchia, Martina Lombardi, Valentina Raffaelli, Paolo Piaggi, Lorenzo Macchia, Ilaria Scattina and

Enio Martino

Department of Clinical and Experimental Medicine, Section of Endocrinology, University of Pisa, Pisa, Italy

Summary

Objective The syndrome of resistance to thyroid hormone

(RTH) is caused by a mutation of TH receptor b (TRb) in 80%

of cases. Patients without mutation (non-TR-RTH) may have a

biochemical pattern that is difficult to differentiate from that of

pituitary TSH-secreting adenoma (TSHoma). Herein, we report

a large monocentric series of RTH focusing on patients with

non-TR-RTH, to evaluate possible clinical or biochemical

parameters able to distinguish them from TSHoma.

Design and patients We retrospectively reviewed the data of

99 consecutive patients with inappropriate TSH secretion

(IST) syndrome referred to our Department between 1983 and

2011, identifying 68 patients with RTH and 31 patients with

TSHomas.

Measurements Patient records were reviewed for the main

clinical, biochemical and imaging characteristics.

Results Of our 68 patients with RTH, 16 (23�5%) did not show

a TRb mutation and did not have affected family members. Of

these 16 patients, three developed a TSHoma, during follow-up.

To distinguish non-TR-RTH from TSHoma, we identified

appropriate cut-off values for the main biochemical parameters

that demonstrated the greatest sensitivity and specificity (T3

suppression test, a-subunit/TSH molar ratio, a-subunit assay

and TRH test) and we calculated the probability for each patient

to develop a TSHoma.

Conclusions The application of the identified cut-offs could

become a very useful tool in the challenging differential diagno-

sis between sporadic non-TR-RTH and TSHoma. It would then

be possible to select the patients at higher risk of developing a

TSHoma and therefore needing a closer follow-up.

(Received 10 June 2014; returned for revision 2 July 2014; accepted

11 July 2014)

Introduction

The syndrome of resistance to thyroid hormone (RTH), nowa-

days included in the syndromes of reduced sensitivity to TH,1,2

is the non-neoplastic form of inappropriate TSH secretion (IST)

syndrome that is characterized by increased serum TH levels

associated with inappropriately detectable TSH.

Recently, mutations of the cell membrane transporter of TH,

MCT8 and of the SECISBP2 gene that is involved in TH metab-

olism have been reported as responsible for a reduced sensitivity

to TH, but with clinical, biochemical and hereditary features dif-

ferent from the classical RTH.2

In RTH, peripheral target tissues and the pituitary gland have

a reduced responsiveness to TH, as in fact demonstrated by a

reduced inhibition of serum TSH after exogenous T3 adminis-

tration, leading to a higher set point of circulating TH. The clin-

ical presentation depends on the degree of sensitivity of each

organ to TH and on the effectiveness of compensatory mecha-

nisms, thus ranging from hypothyroid to hyperthyroid manifes-

tations. A typical clinical feature is the enlargement of the

thyroid gland (goitre diffuse or nodular).3

In 80% of cases, RTH is caused by mutations of TH receptor

b (TRb) gene on chromosome 3.4 The majority of reported fam-

ilies have a single nucleotide substitution, while in some cases

deletion, insertion or duplication has been detected.4 These

mutations are typically inherited as autosomal dominant traits

although a recessive inheritance has been described in one family

in which a complete deletion of all coding sequences of the TRbgene was found.5

In the remaining 20% of patients, a mutation of TRb gene

could not be found (non-TR-RTH). In the majority of non-

TR-RTH, no specific molecular cause has been identified yet.

Among the 39 different families with non-TR-RTH so far

reported, 33 presented just sporadic cases and only six (15%)

Correspondence: Enrico Macchia, Department of Clinical and Experi-mental Medicine, Section of Endocrinology, University of Pisa, OspedaleCisanello, Via Paradisa 2, 56124, Pisa, Italy. Tel./Fax: +39 050 578772;E-mail: [email protected]

© 2014 John Wiley & Sons Ltd 1

Clinical Endocrinology (2014) doi: 10.1111/cen.12556

had at least two affected members with a dominant inheritance

pattern.6 In these six families, a defect in one of the cofactors

involved in the mediation of TH action has been postulated but

not proved.7,8

In this retrospective study, we report the clinical, biochemical

and genetic features of a large monocentric series of RTH subjects

referred to our Department. We evaluated the differences between

RTH and the various forms of hyperthyroidism; in particular, we

looked for the possible differences between non-TR-RTH, TR-

RTH and TSHoma patients, to distinguish the patients with

unrecognized TSHoma from those with non-TR-RTH.

Materials and methods

Study design and protocol

We retrospectively analysed 99 consecutive patients with an IST

referred to the Endocrinology Department of Pisa University,

between January 1983 and December 2011. We identified 68

patients with RTH and 31 patients with TSHomas and collected

data of interest. As controls groups, we also analysed 288 healthy

subjects, 421 untreated Graves’ disease and 62 untreated toxic

adenoma patients.

Patients with IST

In the 99 patients with IST, differential diagnosis between RTH

and TSH pituitary adenoma3 was based on:

• Familiarity for inappropriate TSH-secreting syndrome,

• TSH response to TSH-releasing hormone (TRH),

• TH-dependent peripheral serum parameters [sex-hormone-

binding globulin (SHBG), cholesterol, ferritin, osteocalcin, crea-

tine kinase (CK)],

• Sensitivity of central and peripheral tissues to TH, deter-

mined in selected cases by a short-term administration of incre-

mental doses of levo-thyronine (L-T3),

• a-subunit (a-SU) levels and TSH/a-SU molar ratio and

• Mutations of TRb.

In the two groups of patients, we evaluated the main clinical,

biochemical and radiological features, excluding those patients

(7 out of 99, 7�1%) who had already been submitted to total or

subtotal thyroidectomy.

To assess the familiarity of the syndrome, we collected data

on TH and TSH of all available first-degree relatives. Whenever

these tests were abnormal, we defined the IST syndrome as

familiar.

Among the RTH group, we separated patients with TRbmutation (of whom 4 were previously reported9–11) from those

without and then we performed comparative analyses between 3

series of subjects (TR-RTH, non-TR-RTH and TSHoma). In

particular, we selected the most relevant biochemical parameters

(TSH percentage reduction after administration of incremental

doses of L-T3, molar ratio, TSH response to TRH) and we per-

formed a ROC curve analysis in order to find the appropriate

cut-offs to distinguish non-TR-RTH from TSHoma.

Assays

Hormone measurements were performed in our laboratory,

using commercially available immunometric assays.

Serum TSH ultrasensitive assays have been available since 1987.

In the four patients diagnosed before 1987, serum TSH concen-

tration was above 1 mU/l. From 1987 to 1992, serum TSH was

assayed by chemiluminescent immunoassay (AutoDELFIA hTSH

Ultra Kit, Wallac, Gaithersburg, MD, USA: sensitivity 0�01 mU/l;

normal range of 0�4–3�9 mU/l). From 1992 up to now, serum

TSH was determined by another immunoassay (Immulite 2000

Third Generation, Diagnostic Products Corporation, Los Angeles,

CA, USA: sensitivity 0�004 mU/l; normal range of 0�4–3�4 mU/l).

Serum FT4 and FT3 were evaluated by different assays over

the years (immunoradiometric methods until 2004 and immu-

noenzymatic since then), setting normal range as 7–17 and 2�7–5�7 pg/ml, respectively.

Serum a-SU was measured by an immunoradiometric assay

(Immunotech SAS, Marseille, France: sensitivity 0�02 IU/l; nor-

mal value <0�9 IU/l in normogonadotropic and <1�6 IU/l in

hypergonadotropic patients). a-subunit/TSH molar ratio was

calculated dividing a-SU (lg/l) by TSH and multiplying by 10;

results were considered normal if the value was <2�4, as previ-

ously reported.12

Mutations

Genomic DNA was automatically extracted from peripheral blood

leucocytes of all patients with RTH by Maxwell 16 DNA purifier

kit (Promega Corp, Madison, WI, USA). Exons 7 through 10 of

the TRb1 gene were amplified by PCR, using specific primers

complementary to intronic sequences flanking the various TRb1exons. Reactions were carried out with Master Mix (Promega

Corp, Madison, WI, USA), using 1 mg genomic DNA and 50

pmol of each primer in a 25-ll reaction volume. After an initial

3 min denaturation at 94 °C, we submitted the samples to 40

cycles obtaining denaturation at 94 °C for 30 s, annealing at

59 °C for 30 s and extension at 72 °C for 45 s. PCR products

were checked on a 1% agarose gel and visualized with UV light

after staining with ethidium bromide. If the appropriate size

bands were present, the amplification products were purified

using Quiaquick PCR purifier kit (Quiagen, Milan, Italy).

The purified PCR materials were then directly sequenced

using BigDye Terminator cycle V1.1 Sequencing kit in Perkin

Elmer 3100 XL sequencer.

Results were compared with the normal sequence of the TRbgene (GenBank NT_022517.18). The mutations were verified by

at least two independent PCR and sequencing reactions. Recep-

tor mutations were generated by site-directed mutagenesis of

wild-type (WT) human TRb1 complementary DNA and con-

firmed by direct sequencing. For in vitro studies, both WT and

mutant TRb1 were subcloned into pGEM7z cloning vector.

Receptor proteins were synthesized by coupled transcription and

translation using the TNT T7-coupled reticulocyte lysate system

(Promega Corp., Southampton, UK). T3 binding affinities were

determined using a modification of a filter assay, and binding

© 2014 John Wiley & Sons Ltd

Clinical Endocrinology (2014), 0, 1–8

2 E. Macchia et al.

affinity constants (Ka) were calculated using Scatchard analysis

from three separate experiments, as previously reported.9

Imaging studies

Contrast-enhanced computerized tomography (CT) of the pitui-

tary region was performed in the patients diagnosed before 1991

(n = 10), while from 1992, all the patients were studied with a

pituitary magnetic resonance imaging (MRI), associated with CT

in selected cases.

A thyroid ultrasound was performed in all the patients seen

after 1986, as previously reported.13 The estimated thyroid vol-

ume was calculated according to the ellipsoid formula and

expressed in ml.

Statistical analysis

Statistical tests used to compare groups of subjects included Stu-

dent’s t-test and ANOVA for difference in mean values, Mann–

Whitney U-test and Kruskal–Wallis test for skewed variables and

Pearson’s chi-square test for difference in counts and frequency.

The Kolmogorov–Smirnov test was used to assess normality of

data. Logarithmic transformations were applied to skewed vari-

ables to approximate a normal distribution. Post hoc test were

conducted using the Bonferroni correction of significance. Pear-

son’s (r) and Spearman’s (q) correlation coefficients were

employed for Gaussian and skewed variables, respectively. Recei-

ver operating characteristic (ROC) curves were calculated to

identify the cut-off values of selected biochemical and clinical

parameters, which best discriminated between TR-RTH from

TSHoma patients in terms of sensitivity and specificity according

to the highest Youden index.

In these analyses, sensitivity was calculated as the proportion

of subjects with TSHoma correctly classified by the cut-off value,

while specificity was calculated as the proportion of subjects

with TR-RTH identified as such according to the cut-off value.

We then applied the cut-off values obtained by the ROC curve

analysis of TR-RTH and TSHoma patients, to predict the out-

come of patients with non-TR-RTH.

A P-value <0�05 was considered statistically significant. Data

are presented as mean � standard deviation (SD) or median

with interquartile range (IQR) as indicated.

Results

Clinical, biochemical and radiological features of

patients with RTH:

Among the 68 patients with RTH (35 females, 33 males) belong-

ing to 35 families, 63 were untreated and 5 were previously sur-

gically treated. The median age at diagnosis was 28 years. Forty-

four patients had at least another affected family member, 16

were sporadic, and 8 did not have available information about

first-degree relatives.

A TR b mutation was identified in 52 patients, whereas no

mutation was found in 16 (23�5%). All the 16 patients without

mutation were sporadic cases, although familiarity was not

assessable in 3 of them (first-degree relatives were deceased at

the moment of diagnosis).

The main clinical and biochemical parameters of the 63

untreated patients with RTH are reported in Table 1.

The median TSH value in patients with untreated RTH was sig-

nificantly higher than that observed in 173 healthy subjects

matched for sex and age (1�7 vs 1�4 mU/l, P = 0�004). The esti-

mated thyroid volume of the same patients with RTH was 20 ml

against the 9 ml of 78 normal subjects (P = 0�000). As expected,median percentage reduction of serum TSH both basal and after

TRH stimulation under 100 lg of L-T3 was significantly lower in

patients with RTH than in normal subjects. In particular, median

percentage reduction of TSH response to TRH under 100 lg of

L-T3 was 96�7 in normal subjects, 92�6 in TR-RTH (P = 0�001)and 82�1 in non-TR-RTH (P < 0�001). The FT3/FT4 ratio in

patients with RTH (median = 0�32, IQR: 0�28–0�37) was not dif-ferent from that found in normal subjects (median = 0�33,IQR = 0�28–0�42, adj. P = 0�54), while it was significantly lower

than the ratio observed in primary hyperthyroidism (Graves’ dis-

ease and toxic thyroid adenoma; median = 0�37, IQR: 0�32–0�44,adj. P < 0�001; toxic thyroid adenoma: median = 0�41, IQR:

0�35–0�49, adj. P < 0�001) and TSHoma (median = 0�39, IQR:

0�34–0�45, adj. P = 0�006) (Fig. 1).The overall data of identified TR b mutations are reported in

Table 2, together with the affinity constant of the mutated

receptor. Mutations R383E, L450V and G460Y were never

described before.

A significant correlation was found between a) age at diagno-

sis and thyroid volume (q = 0�44, P = 0�002, Fig 2A); b) TRbaffinity constant to TH and FT4 serum levels (q = �0�50,P = 0�001, Fig. 2B); and c) TSH serum concentrations and thy-

roid volume (q = �0�35, P = 0�02, Fig. 2C).At the time of diagnosis, all the patients with RTH had a neg-

ative pituitary CT or MRI.

Comparison between RTH and TSHoma

As expected, in the TSHoma group, serum a-SU and molar

ratio were significantly higher than in the RTH group, while

TSH response to TRH was lower (all P < 0�001). Moreover,

TSH serum levels, SHBG and age at diagnosis were significantly

higher (all P < 0�001), while CK was significantly lower in the

TSHoma group (P = 0�02) (Table 1).

After a short-term administration of incremental doses of

L-T3, the reduction of basal and TRH-stimulated serum TSH

(all P < 0�001) and of 24 h RAIU (P = 0�04) was significantly

greater in RTH than in patients with TSHoma (Table 1).

Comparison between TR-RTH and non-TR-RTH

In the TR-RTH group, serum concentrations of FT4, FT3, TSH

and a-SU and the FT3/FT4 ratio were not statistically different

from the non-TR-RTH one (P > 0�10). Molar ratio was higher in

non-TR-RTH although the statistical significance was not reached

(P = 0�08). On the contrary, age at diagnosis was significantly



Management of thyroid hormone resistance syndrome 3

higher (median values: 41 years vs 28 years, P = 0�03) in non-TR-

RTH compared with TR-RTH, while TSH response to TRH

(P = 0�01) and CK (P = 0�03) were significantly lower. After a

short-term administration of incremental doses of L-T3, median

percentage values of serum TSH both basal (P < 0�001) and after

TRH stimulation (P < 0�001) under 100 lg of L-T3 were signifi-

cantly lower in patients with non-TR-RTH (Table 1).

Comparison between non-TR-RTH and TSHoma:

Patients with non-TR-RTH had significantly lower levels of

a-SU, molar ratio, basal TSH, FT3, FT3/FT4 ratio and higher

TSH response to TRH than those with TSHoma (all P < 0�05).The percentage of TSH reduction under 100 lg of L-T3 was not

significantly different in the two groups (P = 0�84) (Table 1).

In-depth evaluation of non-TR-RTH

The evaluation of the biochemical data of non-TR-RTH showed

intermediate features between TSHoma and TR-RTH (Table 1;

Fig. 3). Therefore, we decided to re-evaluate the pituitary mor-

phology with MRI of all the available patients with non-TR-RTH

(14/16). We found out that 3 of them had developed a pituitary

microadenoma after a mean follow-up of 108 � 132 months.

After two injections of long-acting somatostatin analogues

(SSA) (Sandostatin LAR 20 mg at weeks 0 and 4), as proposed

by a Mannavola et al.,14 these three patients had a complete

normalization of TH and TSH, suggesting the existence of a real

micro-TSHoma or an autonomous hyperplasia of thyrotrophs.

Among all the biochemical and clinical parameters of our

patients with IST, we identified the most important ones to

Table 1. Clinical comparison of the different study groups

Parameter

(normal values)

All RTH

(IQR*)

TR-RTH

(IQR)

Non-TR-RTH

(IQR)

TSHoma

(IQR)

P-value

(TR-RTH vs

non-TR-RTH)

P-value

(all RTH

vs TSHoma)

P-value

(non-TR-RTH

vs TSHoma)

Age 28 (17–48) 28 (13–39) 41 (28–60) 47 (37–56) 0�026 <0�001 0�406TSH (lUI/ml) (0�4–3�4) 1�7 (1�2–2�4) 1�7 (1�2–2�6) 1�7 (1�1–2�2) 3�2 (2�1–3�8) 0�928 <0�001 0�002FT3 (pg/ml) (2�7–5�7) 7�0 (5�9–8�5) 7�0 (6�3–8�7) 6�8 (5�9–8�5) 7�9 (6�5–10�8) 0�651 0�013 0�035FT4 (pg/ml) (7–17) 25�3 (22–29�5) 25�6 (21�7–29�7) 24�9 (23�4–29�0) 24�6 (21�7–32�9) 0�974 0�966 0�892FT3/FT4 ratio

(0�28–0�42)0�32 (0�28–0�37) 0�33 (0�28–0�37) 0�29 (0�26–0�34) 0�39 (0�34–0�45) 0�337 <0�001 0�002

Tg (ng/ml) (1–50) 54�0 (22�5–120) 53�5 (26�9–101) 59�0 (19–182) 48�0 (22�5–100) 0�787 0�842 0�812Thyroid volume

(ml) (5–18)

20 (15–31) 19 (15–31) 21 (18–30) 21 (13–35) 0�576 0�657 0�925

Α-SU (mUI/l)

(<0�9;<1�6)0�4 (0�2–0�5) 0�3 (0�2–0�5) 0�41 (0�3–0�6) 1�0 (0�6–2�1) 0�148 <0�001 <0�001

Molar ratio (<2�4) 2�0 (1�1–2�7) 1�33 (0�77–0�78) 2�44 (1�56–3�24) 4�1 (2�6–7�1) 0�078 <0�001 0�022SHBG (nmol/l) (27–92) 28�0 (18�3–51�7) 28�0 (19�3–48�8) 41�5 (17�8–60�9) 73�2 (51�1–88�0) 0�616 <0�001 0�011CK (UI/ml) (10–190) 67�0 (52�5–98�5) 70�5 (57�5–102) 49�5 (39–64�5) 52�5 (37–63) 0�031 0�016 0�889Cholesterol

(mg/dl) (<200)170�5 (152–200) 176�0 (156–200) 158�5 (121–200) 158�5 (135–191) 0�101 0�185 0�832

Ferritin (lg/l) (10–150) 48�0 (31–121) 48�0 (28–121) 54�5 (35–152) 80 (43�9–119) 0�378 0�306 0�718Osteocalcin

(ng/ml) (6�8–34)17�6 (9�9–53�1) 14�4 (10�6–57�1) 18�6 (9–34�5) 23�3 (16�2–39�1) 0�944 0�285 0�551

TSH response

to TRH† (5–20)

12�1 (8�6–15�7) 13�2 (9�1–17�7) 8�6 (2�98–14�8) 1�0 (0�4–2�1) 0�014 <0�001 0�002

D-RAIU24 h (%)‡ (>50) 62�3 (48�2–87�3) 74�3 (50�4–90�8) 54�63 (29�8–73�4) 41�7 (30�4–46�5) 0�159 0�043 0�464% reduction of

TRH-stimulated

TSH after

100 lg L-T3

89�3 (64�9–93�3) 92�6 (85�7–94�2) 82�1 (49�9–96�6) 45�09 (20�3–62�5) <0�001 <0�001 0�845

% reduction of

TRH-stimulated

TSH after

200 lg L-T3

86�4 (66�4–94�1) 93�6 (88–96�1) 87�9 (68�8–97�1) 54�4 (40�0–70�1) <0�001 0�007 0�001

Data are reported as median values. Bold numbers indicate statistical different data.

Serum SHBG was measured by an immunoradiometric assay (Orion Diagnostica, Espoo, Finland); serum thyroglobulin (Tg) by immunoradiometric

assay kit (HTGK, Sorin Biomedica, Saluggia, Italy); and serum osteocalcin by a human radioimmunoassay (Nichols Institute) until 1998 and then by

enzyme-linked immunosorbent assay (Nordic Bioscience Diagnostic A/S, Herlew, Denmark).Other common biochemical parameters were measured in

the main hospital laboratory using standard techniques.

*Interquartile Range.

†TSH increment after thyrotropin releasing hormone iv.

‡Percentage variation of 24 h radioiodine uptake before and after L-T3 suppression test.



4 E. Macchia et al.

differentiate TR-RTH from TSHoma (Fig. 3), as previously

reported12:

• Basal TSH reduction after administration of 100 lg of L-T3,• Basal TSH reduction after administration of 200 lg of L-T3,• Fold reduction of TRH-stimulated TSH after 200 lg of L-T3,

• Molar ratio and

• a-SU (UI/l)

Based on these parameters, we calculated several cut-off values

able to differentiate these two groups.

The sensitivity and specificity of the cut-off values confirm

that the most important aspect for the differential diagnosis is

the degree of resistance to L-T3 (Fig. 3).

Considering every parameter singularly, we estimated the

probabilities of each patient with non-TR-RTH of developing a

TSHoma, thus selecting the ones at high risk for pituitary ade-

nomas. These results are shown in Table 3.

The assessment of a global probability score was not feasible,

due to the scarceness of data and lack of correlation among the

various parameters.

Two of the three patients in whom a pituitary microadenoma

became visible at the MRI follow-up had at least two parameters

suggesting a probability to develop a TSHoma over 90%. The

third one had just two parameters because she could not tolerate

the administration of supraphysiological doses of L-T3 due to a

tachyarrhythmia. Therefore, the cut-off values previously identi-

fied by the ROC curve analysis correctly predicted the subse-

quent diagnosis of TSHoma in these three patients.

Discussion

The resistance to thyroid hormone (RTH) is a syndrome charac-

terized by reduced tissue responsiveness to TH.

To date, over 370 families affected by RTH have been

described,6 and more than 3000 cases have been indicated in the

up-to-date commentary on Thyroid 2014.2 All the patients have

increased serum FT4 and FT3 and nonsuppressed TSH. This

syndrome is usually due to mutation in the ligand-binding

domain of the TRb gene. Inheritance is usually autosomal domi-

nant and very rarely autosomal recessive. However, a subgroup

of patients with RTH did not show any mutation of the TRbgene. At present, around 20% of patients with RTH belongs to

this category and usually show a phenotype similar to that of

subjects with TRb mutation.2,3

In this study, we present a large series of patients with an

inappropriate TSH-secreting syndrome (68 with RTH and 31

with TSHoma) from a single institution between 1983 and 2011.

This monocentric retrospective cohort of 99 patients with IST

admitted to our Department of Endocrinology was studied for a

period up to 22 years.

The differential diagnosis among patients with inappropriate

TSH secretion syndrome was done using several parameters. The

diagnostic tests having the greatest sensitivity and specificity to

differentiate the two conditions were in order of relevance: T3

suppression test, a-SU/TSH molar ratio, a-SU assay and TRH

test. All patients underwent a CT scan and/or MRI and TRbgene sequencing.

Of our 68 patients with RTH, 52 (76�5%) were mutated for

TRb. Three new mutations, never described before, were found,

and the TRb affinity constant to T3 was inversely correlated

with serum FT4.

Fig. 1 FT3/FT4 ratio in the different studied groups of subjects. FT3/FT4

ratio in the different groups of patients: thyroid toxic adenoma patients

(Adenoma); untreated Graves’ disease patients (Graves); normal subjects

(Normal); all the 99 RTH patients (RTH) and patients with a TSHoma.

Table 2. List of TRb mutations and their affinity constant

Kindred*

TRb mutation:

codon change Ka MUT/WT

A.G. A234T 0�34A.M. A317T 0�22A.G.† G251E 0�28F.M. G332R N.A.

O.D.I. G460Y N.D.B.

L.S.† L346V 0�02G.C. L450P N.A.

G.A. L450V N.D.B.

E.G. P453A 0�17C.C; V.S. P453T 0�41M.P. R243Q 0�84R.A. R316H 0�02G.A; L.A.; I.C.R. R320H 0�38G.C.; P.I.; A.G. R338W 0�20N.A.M.M. R383E N.D.B.

E.B. R429Q 0�21S.G. R438C 0�30S.L.† S314C 0�70G.S.† S314F 0�01A.G. V349M 0�23

Ka = TR b affinity constant for T3 binding; Wild type (WT)

Ka = 2�2 9 1010 M�1, MUT = mutated type; N.A. = not available;

N.D.B. = not described before.

*Index case of each family.

†Previously reported cases.9–11




Both median serum TSH values and thyroid volume were

significantly higher in RTH than in healthy subjects. However,

it is worth noting that serum TSH values are mostly in

the normal range and sometimes even in the lower part of

the normal range in spite of increased serum TH. This

phenomenon suggests either an increased TSH biological

activity 15 or a higher sensitivity to TSH by the mutated

thyrocytes, in which TRa isoform prevails, as suggested by

several studies.16–20

We observed an inverse correlation between TSH and thyroid

volume that could be explained by the presence of autonomous

areas in large goitres of some patients with RTH.

Confirming previous data,3,7 the FT3/FT4 ratio did not differ

from that of normal subjects being significantly lower than that

observed in primary and central hyperthyroidism. A possible

explanation is that type 2 thyroid deiodinase could be less sensi-

tive to TH excess in RTH but not in the other forms of hyper-

thyroidism, as supposed by others.21,22

(a) (b) (c)

Fig. 2 Clinical and biochemical correlations in RTH patients. Significant correlations between the main clinical and biochemical parameters in the 99

RTH patients (thyroid volume vs age and TSH, FT4 vs affinity constant (Ka)).

Fig. 3 Comparison of the main biochemical

parameters able to distinguish non TR-RTH from

TSHoma patients. Comparison of several

biochemical parameters in different groups of

patients (normal subjects, RTH patients with

(TR-RTH) and without (non TR-RTH) TRbmutations, and patients with TSHoma), and results

of the ROC curve analysis with the cut-offs values

able to distinguish non TR-RTH from TSHoma

patients.



6 E. Macchia et al.

In the present series, 16 of 68 (23�5%) patients with RTH did

not show a TRb mutation and very interestingly, no one had an

affected family member, at variance with other studies.4

The definitely older age at diagnosis in non-TR-RTH com-

pared with TR-RTH subjects may strengthen the hypothesis of a

nongenetic or a low penetrance genetic origin of the former type

of RTH.

When the RTH patient without TRb mutation does not pres-

ent other relatives with the biochemical pattern of IST, the dif-

ferential diagnosis between non-TR-RTH and TSHoma still

remains difficult.

Our patients with non-TR-RTH as a whole show intermediate

biochemical characteristics between RTH and TSHoma. This

observation could suggest that the non-TR-RTH population may

be a heterogeneous group of subjects either with RTH in which

we are not yet able to identify the genetic cause 8 or with TSHo-

ma too small to be detected on imaging. In fact, 3 of our 16

patients initially diagnosed as non-TR-RTH, subsequently turned

out to have developed a visible pituitary TSH-secreting microad-

enoma. Their biochemical pattern actually showed an IST

compatible with a resistance to exogenous TH, thus excluding

an initial diagnostic misinterpretation. Their pituitary imaging

noticeably changed from negative to positive for microadenoma.

The other main criteria for differential diagnosis, mentioned

above, shifted towards values compatible with TSHoma. More-

over, a persistent normalization of their IST pattern was

obtained after administration of long-acting SSA for 2 months.14

Based on the above observations, in the attempt to discriminate

between RTH and TSHoma, we used a ROC curve analysis to

obtain some appropriate cut-off values for the main biochemical

parameters (basal TSH reduction after administration of 100 lgand 200 lg of L-T3, reduction of TSH response to TRH after

200 lg of L-T3, molar ratio and a-SU). Applying these cut-offs,

we were able to assess the probability for each patient to develop a

TSHoma according to the above clinical parameters.

Among these parameters, the most useful is the basal TSH

reduction after 200 lg of L-T3: a value <75 is associated with a

probability of TSHoma near 100%.

In the three aforementioned patients who subsequently devel-

oped TSHoma, at least two parameters out of five indicated that

these subjects were at a higher risk of developing a TSHoma. In

particular, one or both of these two parameters showed a pre-

dicted probability of 100% of developing a TSHoma.

Therefore, by applying the proposed cut-off values of these five

clinical parameters, we were able to foresee, with a high confi-

dence, the outcome of the three subjects who developed a TSHo-

ma. On the other hand, the risk probability predicted by the ROC

curve in the subjects who showed at least one parameter over the

cut-off value but did not develop a TSHoma during the observa-

tion period was lower compared with those who instead devel-

oped a TSHoma. Even so, these subjects should be considered at

high risk of developing a TSHoma and require a longer follow-

up. We can suppose that some of these subjects with RTH

without TR mutations may instead harbour precursor lesions of

pituitary adenoma with possible sequential changes of thyro-

trophs from autonomous hyperplasia to neoplasia, as suggested.23

Due to the epidemiological rareness of the RTH syndrome,

the number of patients involved in this study was too low to be

Table 3. Probability of developing a TSHoma in each non-TR-RTH patient, according to different biochemical parameters

Patient

Probability of developing a TSHoma according to:

Basal TSH reduction after

100 lg of L-T3 (CVT < 68�75;APT ≥ 42%)

Basal TSH reduction

after 200 lg of L-T3

(CVT < 75; APT ≥ 99%)

Post-TRH TSH reduction

after 200 lg of L-T3

(CVT < 4�94; APT ≥ 24%)

Molar ratio

(CVT < 2�58;APT ≥ 47%)

a-SU (CVT < 0�56;APT ≥ 43%)

B.A. N.A. N.A. 5% 36% 14%

G.B. 17% 0% 0% 78% 24%

F.C. 6% 0% 1% 31% 5%

L.C. 2% 0% 1% 47% 12%

M.D.E. 2% 0% 0% 54% 49%

S.G.* 99% 100% 100% 54% 21%

E.L.R. 12% 0% 0% 46% 40%

B.L. 2% 0% 0% 45% 21%

D.L. N.A. N.A. 0% 46% 34%

P.P.* N.A. N.A. 100% 47% 22%

P.L.P. 88% 100% 21% 90% 71%

M.P. N.A. N.A. N.A. 47% 65%

F.S. 84% 98% 97% 44% 7%

V.S. N.A. N.A. N.A. 39% 49%

S.S. 3% 0% 0% 29% 4%

A.T.* 96% 100% 100% 54% 38%

CVT, Cut-off value for TSHoma obtained through a ROC curve analysis; APT, Associated probability of TSHoma calculated by logistic regression

analysis; N.A., not available.

Bold values represent those above associated probability of TSHoma.

*Patients who developed a TSH-secreting pituitary adenoma during the clinical follow-up.




able to define a general clinical score. Nonetheless, the applica-

tion of the main diagnostic biochemical parameters’ cut-offs

(Fig. 3) could become a very useful tool in the challenging dif-

ferential diagnosis between sporadic non-TR-RTH and TSHoma.

These cut-offs could facilitate the selection of those patients for

whom a trial with long-acting somatostatin analogues and a sub-

sequent close MRI follow-up would be more appropriate.

Multicentric studies would be necessary to collect a larger

number of patients with non-TR-RTH, thus permitting a better

understanding of the actual origin, genetic or acquired, of the

nonmutated form of RTH and the creation of a comprehensive

clinical score able to distinguish between non-TR-RTH and

TSHoma.

Acknowledgements

This work was partially supported by Grants from the University

of Pisa [Fondi d’Ateneo] to E.M.

References

1 Refetoff, S. (2008) Resistance to thyroid hormone: one of several

defects causing reduced sensitivity to thyroid hormone. Nature

Clinical Practice Endocrinology & Metabolism, 4, 1.

2 Refetoff, S., Bassett, J.H.D., Beck-Peccoz, P., et al. (2014) Classi-

fication and proposed nomenclature for Inherited Defects of

Thyroid Hormone Action, Cell Transport, and Metabolism. Thy-

roid, 24, 1–3.3 Refetoff, S., Weiss, R.E. & Usala, S.J. (1993) The syndrome of

resistance to thyroid hormones. Endocrine Reviews, 14, 348–399.4 Weiss, R.E., Hayashi, Y., Nagaya, T., et al. (1996) Dominant

inheritance of resistance to thyroid hormone not linked to

defects in the thyroid hormone receptors alpha or beta genes

may be due to a defective co-factor. Journal of Clinical Endocri-

nology and Metabolism, 81, 4196–4203.5 Takeda, K., Sakurai, A., DeGroot, L.J., et al. (1992) Recessive

inheritance of thyroid hormone resistance caused by complete

deletion of the protein-coding region of the thyroid hormone

receptor-b gene. The Journal of Clinical Endocrinology and

Metabolism, 74, 49–55.6 Refetoff, S. & Dumitrescu, A. (2007) Syndromes of reduced sen-

sitivity to thyroid hormone: genetic defects in hormone recep-

tors, cell transporters and deiodination. Best Practice Research

Clinical Endocrinol Metabolism, 21, 277–305.7 Dumitrescu, A.M. & Refetoff, S. (2012) Reduced sensitivity to

thyroid hormone: defects of transport, metabolism, and action.

Werner & Ingbar’s The Thyroid: A Fundamental and Clinical

Text, 10th edn. Lippincott Williams and Wilkins, Philadelphia,

845–875.8 Reutrakul, S., Sadow, P.M., Pannain, S., et al. (2000) Search for

abnormalities of nuclear corepressors, coactivators and a coregu-

lator in families with resistance to thyroid hormone without thy-

roid hormone receptor b or a genes mutations. Journal of

Clinical Endocrinology and Metabolism, 85, 3609–3617.9 Gurnell, M., Rajanayagam, O., Agostini, M., et al. (1999) Three

novel mutations at serine 314 in the thyroid hormone beta-

receptor differentially impair ligand binding in the syndrome of

resistance to thyroid hormone. Endocrinology, 140, 5901–5906.10 Macchia, E., Agostini, M., Sarkissian, G., et al. (1998) Detection

of a new de novo mutation at codon 251 of exon 8 of thyroid

hormone receptor beta gene in an Italian kindred with resistance

to thyroid hormone. Journal of Endocrinological Investigation, 21,

226–233.11 Macchia, E., Gurnell, M., Agostini, M., et al. (1997) Identifica-

tion and characterization of a novel de novo mutation (L346V)

in the thyroid hormone receptor beta gene in a family with gen-

eralized thyroid hormone resistance. European Journal of Endocri-

nology, 137, 370–376.12 Macchia, E., Gasperi, M., Lombardi, M., et al. (2009) Clinical

aspects and therapeutic outcome in thyrotropin-secreting pitui-

tary adenomas: a single center experience. Journal of Endocrino-

logical Investigation, 32, 773–779.13 Vitti, P. & Rago, T. (2003) Thyroid ultrasound as a predictor of

thyroid disease. Journal of Endocrinological Investigation, 26, 686–689.

14 Mannavola, D., Persani, L., Vannucchi, G., et al. (2005) Different

responses to chronic somatostatin analogues in patients with

central hyperthyroidism. Clinical Endocrinology (Oxford), 62,

176–181.15 Persani, L., Asteria, C., Tonacchera, M., et al. (1994) Evidence

for the secretion of thyrotropin with enhanced bioactivity in syn-

dromes of thyroid hormone resistance. Journal of Clinical Endo-

crinology and Metabolism, 78, 1034–1039.16 Abel, E.D., Ahima, R.S., Boers, M.E., et al. (2001) Critical role

for thyroid hormone receptor _2 in the regulation of paraven-

tricular thyrotropin-releasing hormone neurons. The Journal of

Clinical Investigation, 107, 1017–1023.17 Forrest, D., Hanebuth, E., Smeyne, R.J., et al. (1996) Recessive

resistance to thyroid hormone in mice lacking thyroid hormone

receptor: evidence for tissue-specific modulation of receptor

function. EMBO Journal, 15, 3006–3015.18 Gauthier, K., Chassande, O., Plateroti, M., et al. (1999) Different

functions for the thyroid hormone receptors TRalpha and TRbe-

ta in the control of thyroid hormone production and post-natal

development. EMBO Journal, 18, 623–631.19 Kaneshige, M., Kaneshige, K., Zhu, X., et al. (2000) Mice with a

targeted mutation in the thyroid hormone _ receptor gene exhi-

bit impaired growth and resistance to thyroid hormone. Proceed-

ings of the National Academy of Sciences of the United States of

America, 97, 13209–13214.20 Selmi-Ruby, S., Bouazza, L., Obregon, M.S., et al. (2014) The

targeted inactivation of TRb gene in thyroid follicular cells sug-

gests a new mechanism of regulation of thyroid hormone pro-

duction. Endocrinology, 155, 635–646.21 Schneider, M.J., Fiering, S.N., Pallud, S.E., et al. (2001) Targeted

disruption of the type 2 selenodeiodinase gene (DIO2) results in

a phenotype of pituitary resistance to T4. Molecular Endocrinol-

ogy, 15, 2137–2148.22 Galton, V.A., Schneider, M.J., Clark, A.S., et al. (2009) Life with-

out thyroxine to 3, 5, 30-triiodothyronine conversion: studies in

mice devoid of the 50-deiodinases. Endocrinology, 150, 2957–2963.

23 Mete, O. & Asa, S.L. (2013) Precursor lesions of endocrine sys-

tem neoplasms. Pathology, 45, 316–330.



8 E. Macchia et al.

clinical and genetic characteristics of a large monocentric series of patients affected by thyroid...

Documents