Key words
thyroid treatment - thyroidectomy - ATD - course of GO
Introduction
Graves’ orbitopathy (GO) is an inflammatory orbital disease that is mostly associated
with autoimmune Graves’ disease. During the course of GO, patients may develop lid
retraction, marked periorbital swelling, impaired ocular motility due to fibrotic
changes of the extraocular muscles, disfiguring proptosis, and optic nerve compression.
Patients with severe disease suffer from major functional deficits and facial disfigurement.
About half of the patients with Graves’ hyperthyroidism develop GO, which is closely
related to onset and course of the thyroiditis: it usually occurs simultaneously,
occasionally after onset of thyroid disease due to an unfavorable course, and only
rarely before clinically evident hyperthyroidism. Autoantigens responsible for GO
include molecules expressed by thyroid epithelial cells as well as by orbital tissues.
Especially, interaction between the TSH-receptor (TSHR) and TSH-receptor autoantibodies
(TRAb) is pivotal to the pathogenesis of hyperthyroidism and orbital disease [1]
[2]. Anti-inflammatory therapy may inactivate orbital inflammation in 70–80% of cases;
however, complete remission of orbital changes is rare [3]. Commonly, surgical intervention (orbital surgery, squint- and lid corrections)
is necessary to restore function and appearance in moderate to severe disease stages.
As prerequisites for ophthalmosurgery are stable euthyroidism and inactive GO for
at least 6 months [4], it takes 1.5–2 years until surgical rehabilitation is initiated.
In Europe, treatment of hyperthyroidism usually includes thyreostatic pharmacotherapy
for 12–18 months [5]. After termination of thyreostatics, the overall relapse rate of hyperthyroidism
is roughly 50%. Large thyroid volume, nicotine consumption, advanced age and, notably,
high serum TRAb activities increase the risk of relapse [6]. Studies using a second generation human TRAb assay show that relapse or remission
of hyperthyroidism correlate with TRAb serum levels and the severity of eye disease
6, 12, or 18 months after initiation of thyreostatic therapy (reviewed in [7]). As early as 6 months after the beginning of antithyroid drug treatment in patients
with TRAb serum levels ≥ 10 IU/l remission rates of hyperthyroidism are rather low
with 3% [8]. Comparably low remission rates of about 8% occur in patients with severe GO [9]. These low remission rates raise the question of the optimal time point of definite
thyroid treatment for these patients. In those cases either thyroidectomy and/or radioiodine
therapy, is indicated, particularly as removal of thyroid antigens by attenuating
autoimmunity, may be beneficial for GO [10].
Antithyroid drug (ATD) therapy is most favorable in the early phase of the disease,
and, if chosen as initial therapy, it is accompanied by a 10% risk of further deterioration
of GO in comparison to 16% after thyroidectomy, and 33% after radioiodine therapy
as published by Tallstedt et al. [11]. However, in this randomized clinical trial only patients without and with mild
GO stages were included. Other observational studies with widely variable inclusion
criteria led to the assumption that patients with active orbital inflammation may
benefit from early thyroidectomy [12]
[13]. Several studies concerning the influence of thyroidectomy with or without accompanying
radioiodine therapy (total ablation) are available in the later phase of GO after
a longer course of antithyroid drug therapy (Marcocci 11±4 months or De Bellis 21.23±11.2).
Marcocci and colleagues reported that the course of Graves’ orbitopathy is not influenced
by near total thyroidectomy in comparison to methimazole therapy [14]. De Bellis et al. reported a much later and less marked improvement of GO with persistence
of TRAb in the methimazole therapy group in comparison to thyroidectomy with and without
postoperative radioiodine therapy [15]. However, more recent studies substantiate the hypothesis that patients with moderate
to severe active GO might benefit from an early thyroidectomy in the sense of faster
and more profound improvement of GO [16]
[17]
[18]
[19].
We now report findings from a retrospective case control study on the effect of early
thyroidectomy on moderate to severe, active GO when compared to standard antithyroid
drug treatment.
Patients and Methods
We performed a retrospective cohort study, including a total of 92 GO patients of
Caucasian origin. All patients had a CAS>4 at baseline, having active GO. Patients
were included into early thyroidectomy group (Tx-group), if they underwent total thyroidectomy
6±2 months after initiation of thyreostatic therapy, and GO was moderate to severe,
and if they had developed orbitopathy within 6 months before or after the onset of
hyperthyroidism, at least short before baseline examination, and if follow-up after
thyroidectomy was 6±2 months (n=46). These patients were compared to a control group
(n=46), which were followed after 6 months of ATD for another 6±2 with further antithyroid
treatment (ATD-group). Controls were consecutively picked from a patient database
that had been initiated in 11/2 000, containing data on more than 1 000 patients suffering
from Graves’ orbitopathy at the time of this study. Control patients were eligible,
if they were first examined 6±2 months after initiation of antithyroid drug therapy,
if antithyroid drugs had been administered for at least 12±2 months (until the end
of follow-up period), if they had developed orbitopathy within 6 months before or
after the onset of hyperthyroidism at least short before baseline examination, and
if they matched for a patient in the Tx-group regarding previous anti-inflammatory
therapy and severity of GO. If there were more matched patients, patients with the
closest referrals to the thyroidectomy patients were chosen for the more common situations.
The effect of both therapeutical approaches on the course of orbital inflammation
was evaluated at baseline and after 6±2 months. Clinical activity score (CAS), NOSPECS
severity score, serum TRAb activity, inactivation rate, and the necessity of anti-inflammatory
therapy served as outcome parameters. Visual acuity, lid fissure width, downward motility
of the upper lid, proptosis, signs of inflammation (rubor and swelling of lids and
conjunctiva), and extraocular motility (monocular excursions measured with Kestenbaum
glasses) were assessed in each patient. Consecutive examinations were performed by
the same investigator. In case of reduced visual acuity, the cornea was stained with
fluorescein, and if slit lamp and fundus examination revealed normal anatomical conditions,
visual evoked potentials were carried out to examine optic nerve function. The clinical
activity score (CAS) was assessed as described by Mourits [20]. Disease severity was estimated using a previously described modified NOSPECS score
[21].
Anti-inflammatory treatment during follow-up
During follow-up, anti-inflammatory therapy was offered to all patients with active
orbitopathy (CAS≥4) who had no therapy yet, who had i. v. steroids but not the full
cumulative dose of 4.5 g and who had motility impairment but no irradiation yet. Prednisone
(Prednisolut) was given intravenously in a cumulative dose of 2.68 g for 6 weeks [once
every week 500 mg (3 ×) or 250 mg (3 ×)], and on all other days 10 mg oral fluocortisone
(Fluocortulone), which was tapered out over 10 days and this was repeated when activity
persisted. In cases with impaired ocular motility, additional orbital irradiation
(total absorbed doses between 12–20 Gray) was performed. However, despite CAS≥4 10
patients of the Tx-group and 13 of the ATD-group did not receive any treatment during
observation period. These patients had a good quality of life despite CAS≥4 and did
not wish any further treatment.
TRAb assay
TRAb serum levels were measured with a second generation TSH-binding inhibitory (TBII)
assay based on the human recombinant TSH-receptor (TRAK human LIA®, B.R.A.H.M.S AG, Hennigsdorf/Berlin, Germany) [22]. This assay is calibrated in international units (IU), based on the WHO-reference
standard MRC 90/672. 50% inhibition of tracer binding corresponds to 7–8 IU/l. Values≥1.5 IU/l
(approx. 10% tracer binding) were regarded as positive, values between 1 and 1.5 IU/l
as borderline, and values<1.0 IU/l as negative.
Thyroidectomy
Surgery was performed under general anesthesia. Total thyroidectomy was conducted
through a central neck incision. Routinely, both recurrent laryngeal nerves were visualized
and dissected free. The parathyroid glands were meticulously preserved and in case
of insufficient perfusion autotransplanted prior the closure of the skin. A video-assisted
approach was used in selected cases.
Statistical analysis
Group comparisons were tested for statistical difference with the non-parametric Mann-Whitney
test (Graph pad prism 5.0) analyzing parameters that did not follow Gaussian distribution,
such as TRAb activities, CAS, and NOSPECS. Fisher’s exact T-Test, Wilcoxon rank and
Chi-Square test were used for cross tables of patient-groups vs. binary variable as
stated (e. g., gender, smoking status, inactivation, and prior anti-inflammatory therapy).
All p-values equal to or lower than 5% (≤0.05) were regarded as significant.
Ethical approvement
This study was approved by the Medical Ethics Committee of the University of Duisburg-Essen,
Germany. Written consent to be included in our database and to have blood exams performed
was obtained from all participants.
Results
Both cohorts were very well matched. At baseline ([Table 1]) there were no statistically significant inter-group differences concerning age
(median age: Tx-group 48.8 years, ATD-group 50.8 years, p=0.6) or gender (Tx-group
87% female/13% male patients, ATD-group 76% female/24% male patients, p=0.28). Seventy
percent of patients of the Tx-group and 59% of the ATD-group were smokers (p=0.38).
In the Tx-group, mean CAS was 5.9 vs. 5.7 in the ATD-group (p=0.39), and in both groups
all patients had active GO (CAS≥4). Severity of GO at baseline was comparable in both
groups (mean NOSPECS score: Tx-group 7, ATD-group 7.35, p=0.46). TRAb activities in
serum were higher in the Tx-group (mean 23.8, median 18.6 IU/l) in comparison to ATD-group
(mean 16.3, median 12.8 IU/l), the difference was not significant (p=0.07). In both
groups, 61% of patients had received steroids, and 20% of these patients had had orbital
irradiation, 39% had not received any anti-inflammatory therapy prior to observation
period (p=1.0). Protrusio of the eye measured with a Hertel Exophthalmometer was pathological
(≥17 mm) in 70% of patients in Tx-group and in 67% of patients in ATD-group (p=0.87).
All patients in both group presented with a pathological Edema score of 2 or higher,
mean Edema score in Tx-group was 7.4 and in ATD-group 6.9 (p=0.2) ([Table 1], [Fig. 1]).
Fig. 1 a: Reduction of Clinical Activity of GO: Mean CAS score declined significantly from
baseline (0) to 6 months follow-up (1) as well in Early Thyroidectomy group (Early
Tx-group; from 5.9 to 2.1, *** p<0.0001) as in ATD-group (from 5.7 to 2.8, *** p<0.0001).
However, patients of the Early Tx group presented with a significantly lower CAS of
GO when compared to ATD-group at the end of follow-up (* p=0.03). b: Reduction of severity of GO. Mean NOSPECS score declined significantly from baseline
(0) to follow-up (1) as well in Early Tx-group (from 7 to 5.7, ** p<0.0042) as in
ATD group (ATD; from 7.35 to 5.9, **p<0.0033), but with no significant differences
between both cohorts (p=0.56). Data in a and b are shown as mean±SEM. c Reduction of TRAb serum level (IU/l log). Serum TRAb activities declined significantly
as well in Early Tx-group (from median 18.6 to 5.2, *** p=0.0002) as in ATD-group
(from median 12.8 to 3.2, *** p=0.0001), but with no significant differences between
both cohorts as calculated using Wilcoxon rank test (p0=0.07, p6months=0.32]. Data are shown as median with range.
Table 1 Patient characteristics at baseline.
|
Patients characteristics at baseline
|
Early thyroidectomy group
|
p
|
ATD-Group
|
|
Number of patients
|
46
|
46
|
|
Age (years)a
|
48.8 (20–68)
|
0.6 ns
|
50.8 (27–78)
|
|
Gender Female/Male (%)
|
87/13
|
0.28 ns
|
76/24
|
|
Smoker/Non-smoker (%)
|
70/30
|
0.38 ns
|
59/41
|
|
Classification mild*/severe
|
9/37
|
0.57 ns
|
6/40
|
|
Duration of GD since initiation ATD (months) a
|
6 (4–8)
|
0.93 ns
|
6 (4–8)
|
|
Duration of GO since initiation ATD (months) a
|
0 (−6 to 6)
|
0.22 ns
|
0 (−6 to 6)
|
|
Duration of GD until Thyroidectomy (months) a
|
6 (4–8)
|
|
Patients with active GO (CAS≥4) (%) b
|
100
|
1.0 ns
|
100
|
|
Activity of GO (CAS score) c
|
5.9±1.5
|
0.39 ns
|
5.7±1.5
|
|
Severity of GO (NOSPECS score) c
|
7±2.1
|
0.46 ns
|
7.35±2.4
|
|
TRAb level a
|
18.6 (0.9–123.7)
|
0.07 ns
|
12.8 (0.4–97.5)
|
|
Anti-inflammatory therapy prior observation period (%) (No therapy/Steroids/Steroids
and orbital irradiation)
|
39/41/20
|
1.0 ns
|
39/41/20
|
|
Protrusio: Hertel R/L pathological/normal (%) d
|
70/30
|
0.87 ns
|
67/33
|
|
Mean Edema Score (Injection/Chemosis/Swelling/Irritation)
|
7.4
|
0.2 ns
|
6.9
|
|
Mean Edema score pathological (≥2) (%)
|
100
|
1.0 ns
|
100
|
*Inclusive 1 mild-moderate in Thyroidectomy- and 3 in ATD-group
a Data expressed as median (range)
b Active GO: CAS≥4, Inactive GO: CAS<4
c Data expressed as mean±SD
d Hertel pathological (≥17 mm)/normal (<17 mm)
At follow-up 6±2 months after baseline disease severity (mean NOSPECS Tx-group 5.7,
ATD-group 5.9, p=0.56) as well as serum TRAb activities (median TRAb level Tx-group
5.2, ATD-group 3.2, p=0.36) had declined significantly in both cohorts, but with no
significant inter-group differences between thyroidectomized and control patients
([Table 2], [Fig. 1]). Protrusio of the eye was still pathological (≥17 mm) in 66.3% of patients in Early
Tx-group and in 60.9% of patients in ATD-group (p=0.54). All patients in both groups
presented with a pathological Edema score of 2 or higher, mean Edema score in Tx-group
was 5.1 and in ATD-group 4.6 (p=0.38) ([Table 2]). Anti-inflammatory therapy during follow-up was not different in both groups. Additional
doses of i. v. steroids pulses were administered in 12 patients of the Tx-group and
10 patients of the ATD-group. Twenty-four (Tx-group) as well as 23 (ATD-group) patients
received steroids in combination with orbital irradiation due to persistent motility
deficits (cumulative steroid doses remained < 4.5 g). Ten thyroidectomized patients
vs. 13 in ATD-group had not received any anti-inflammatory therapy according to patients
wishes despite CAS≥4 due to good quality of life (p=0.8) ([Table 2], [Fig. 2]).
Fig. 2 Anti-inflammatory therapy (imt) of GO-patients before and during follow-up was comparable:
At baseline, as well as at 12 m follow-up, patients in both groups had received comparable
imt. At baseline, in each group, 9 patients had received steroids in combination with
orbital irradiation, 19 patients had received steroids only, and 18 patients had had
no imt at all (matched pairs) (p=1). During follow-up, additional doses of steroids
were given to 12 patients of the Tx-group and 10 patients of the ATD-group due to
still active GO. Those who also had motility deficits, received steroids in combination
with orbital irradiation: 24 (Tx-group) as well as 23 (ATD-group) (cumulative steroid
doses remained<4.5 g). Ten thyroidectomized patients vs. 13 in ATD-group had not required
any further immunosuppressive treatment at the end of the observation period (p=0.8).
Table 2 Patient characteristics at follow-up.
|
Patients characteristics at follow-up
|
Early thyroidectomy group
|
p
|
ATD-Group
|
|
Smoker/Non-smoker (%)
|
60.9/39.1
|
0.5 ns
|
52.2/47.8
|
|
Duration of GD since initiation ATD (months) a
|
12 (10–14)
|
0.7 ns
|
12 (10–14)
|
|
Activity of GO (CAS score) c
|
2.1±1.3
|
0.02*
|
2.8±1.5
|
|
Patients with active/inactive GO (%) b
|
10.9/89.1
|
32.60/67.4
|
|
Inactivation rate (%)
|
89.1
|
0.02*
|
67.4
|
|
Severity of GO (NOSPECS score) c
|
5.7±2.2
|
0.56 ns
|
5.9±2.2
|
|
TRAb level a
|
5.2 (0.1–65.2)
|
0.32 ns
|
3.2 (0.1–53.1)
|
|
Anti-inflammatory therapy during observation period (%):
Baseline until follow-up
(No therapy/Steroids/Steroids and orbital irradiation)
|
22/26/52
|
0.8 ns
|
28/22/50
|
|
Protrusio: Hertel R/L pathological/normal (%) d
|
66.3/33.7
|
0.54 ns
|
60.9/39.1
|
|
Hertel difference between RE and LE (%) e
|
23.9/76.1
|
0.86 ns
|
21.7/78.3
|
|
Mean Edema Score (Injection/Chemosis/Swelling/Irritation)
|
5.1
|
0.38 ns
|
4.6
|
|
Mean Edema score pathological (≥ 2) (%)
|
45
|
0.03*
|
38
|
a Data expressed as median (range)
b Active GO: CAS≥4, Inactive GO: CAS<4
c Data expressed as mean±SD
d Hertel pathological (≥ 17 mm)/normal (< 17 mm)
e Hertel difference between Right Eye (RE) and Left Eye (LE) pathological (≥ 1.5 mm)/normal
(< 1.5 mm)
*significant; ns: not significant
Clinical activity of GO declined in both groups, but more in the Tx-group (mean CAS
at 12 months Tx-group 2.1, ATD-group 2.8, * p=0.03) ([Table 2], [Fig. 1]). The inactivation rate (Tx-group 89.1%, ATD-group 67.4%, * p=0.02) was significantly
higher in the early thyroidectomized cohort ([Table 2], [Fig. 3]).
Fig. 3 Rate of inactivation: In the Early Tx-group, the inactivation rate was significantly
higher than in ATD-group (* p=0.02; calculated using Fisher’s exact test). GO was
considered as being active if CAS≥4 and inactive if CAS<4. In the Early Tx-group,
from 46 cases of active GO at baseline, 41 were inactive after 6±2 month follow-up
(89.1%). In the ATD-group, from 46 cases of active GO at baseline, 31 were inactive
after 6±2 month follow-up (67.4%), while15 patients still presented with active GO.
Discussion
The management of active Graves’ orbitopathy (GO) can be a challenging therapeutic
dilemma due to its pathogenic complexity, disease heterogeneity, clinical unpredictability,
and ocular morbidity. The relation between orbital inflammation and treatment of hyperthyroidism
has been subject of many studies (reviewed in [23]
[24]), and it has been shown, that consistent euthyroidism has a beneficial influence
on the course of orbitopathy [25]. However, the effect of early thyroid ablation is still a matter of debate.
Ours is the first study to retrospectively assess the effect of early thyroidectomy
on active, recently onset moderate to severe GO. We showed that early thyroidectomy
after about 6 months of antithyroid treatment in GO-patients with low chance of remission
did not influence final outcomes concerning remaining defects of GO in comparison
to patients being further treated with thyreostatics. However, those patients benefit
from early thyroidectomy, as they present with significantly less active GO and they
show a significantly higher rate of inactivation of GO at 6 months follow-up. Our
result is in line with reports from De Bellis et al. who reported on the effect of
thyroidectomy at a later time point after 21–22±9–11 months of antithyroid drug treatment
[15]. Also, in this phase of the disease they observed a later and less marked improvement
of GO in patients on methimazole treatment in comparison to patients that underwent
thyroidectomy.
Concerning the treatment of hyperthyroidism, usually patients with GO and very low
probability of remission (high TRAb, smokers, large goiter) have 2 options: long lasting
antithyroid drug therapy or thyroidectomy. Radioiodine therapy is not the treatment
of choice in smokers with high TRAb and moderate to severe active orbitopathy, as
it carries a high risk of deteriorating orbitopathy [26]. Even after longtime antithyroid treatment, patients with persisting high serum
activities of TRAbs have a higher risk of relapsing hyperthyroidism during any attempt
to reduce the dose of thyreostatics, although this risk decreases slightly over time
[27]
[28]
[29].
Total thyroidectomy, on the other hand, is associated with a low risk of relapsing
thyroid disease and consecutive reactivation of orbitopathy (reviewed in [30]). Once thyroid hormone substitution is properly attuned and serum levels are within
normal parameters, clinical follow-ups have to be performed only few times per year.
Complications of thyroidectomy are relatively rare – these include transient recurrent
laryngeal nerve palsies (1.3% calculated on Nerves at Risk, NaR) or transient hypocalcaemia
(approx. 7.4%) [31]. Inactivation rates of GO can be increased if thyroidectomy is followed by a small
dose radioiodine therapy called ‘total ablation’ [32]
[33], which nearly prevents recurrence of hyperthyroidism [34].
Therefore, lifelong hormone substitution vs. long term ATD have to be carefully evaluated
in patients with low chance of remission. Early thyroidectomy should be considered
preferably in patients who need to reach stable, inactive GO as quickly as possible,
for example, patients with diplopia who need to drive, or women with the desire to
have children.
The higher inactivation rate after early thyroidectomy shortens average active disease
duration. Our examined cohort comprised mainly patients with moderate to severe GO
– these are usually patients who do not go into remission [9]. Since these are largely patients who will undergo ophthalmosurgical repair of remaining
defects, restorative functional and aesthetic oculoplastic surgery can be performed
earlier in these patients. This consequently influences quality of life and especially
the duration of sick leave of these patients with all associated socio-economic aspects.
Patients of the herein studied cohorts had already received extensive anti-inflammatory
treatment prior to and during observation. We do not know, whether early thyroidectomy
has an even greater impact on patients with untreated orbitopathy, in comparison to
continuous thyreostatic treatment. This will be subject of a prospective randomized
trial.
The immunological consequences of the removal of the thyroid have not been completely
elucidated yet. Several lines of evidence suggest that the thyroid is a major site
of autoantibody synthesis and antigen (Ag) presentation in the autoimmune thyroid
diseases. Thyrocytes express HLA-DR [35] and CD40 [36]
[37] and may be involved in the induction of Graves’ autoimmunity. The colocalization
of thyrocytes, mature professional antigen-presenting cells and CD4+T-helper cells
led to the extended concept that these cells might cooperatively stimulate thyroid
autoimmunity [38]
[39]. IgG from patients with Graves` disease maintain the process since they induce Interleukin-16
and RANTES expression in cultured human thyrocytes: chemokines which modulate T cell
activation [40]. And consequently, removal of the thyroid is followed by a continuous decrease of
TRAb in most of the patients [11]
[41]. Takamura et al. showed even 100% TRAb elimination after total thyroidectomy in
the long term follow-up of their cohort [42].
However, antithyroid drug treatment is also accompanied be a continuous TRAb decrease
[41]. Methimazole treatment led to downregulation of major histocompatibility class II
gene expression in FRTL-5 thyrocytes [43]. In experimental models of autoimmune uveitis methimazole inhibits uveitis at least
in part by preventing the recruitment and/or maturation of antigen presenting cells,
resulting in reduced generation of Ag-specific T cells [44]. However, since most of the patients are treated by titration regime the effect
of total thyroidectomy on the autoimmune disease may outweigh the effect of low doses
of thyreostatics and so be accompanied by a faster and more effective inactivation
of GO.
Another aspect is epitope spreading. Although in general there is a close correlation
between TRAb levels and the course of GO [45], there are patients with very high TRAb levels and no overt eye symptoms. Or there
are patients who develop orbitopathy only later in the course with a relapse of thyroid
hyperthyroidism. It has been shown that TSHR-Ab binding results in different signaling
cascades. Stimulating antibodies use signaling pathways similar to the TSH activation.
Both, TSH-blocking and neutral TSHR-antibodies, use other signaling networks which
results in variable signal responses. For instance, via Gαq, the c-Raf-ERK-p90RSK
signaling cascade is activated, while it is not activated by TSH [46]
[47]. For orbital fibroblasts it has been shown that signaling pathways of TSHR and the
growth factor IGF1 influence each other (PKA and PI3K) with consequences for hyaluronan
production [48]. These observations help to explain how TSHR-Abs may contribute to different clinical
phenotypes.
Since there is clear evidence for Ag presentation in the thyroid, an early removal
of the thyroid might prevent epitope spreading towards more antibodies specific for
orbitopathy. Therefore, prevention of epitope spreading could be included in the list
of arguments pro a definite therapy of the thyroid during the course of Graves’ orbitopathy.
In conclusion, early thyroidectomy might be a good therapeutic option for patients
with moderate to severe Graves’ orbitopathy who have only low chances of remission.
In this first, retrospective study, early thyroidectomy does not reduce remaining
defects in comparison to antithyroid drug therapy. However, it speeds up inactivation
of GO which may impact patients’ sick leave and quality of life since rehabilitative
surgery can be started earlier. To further study and define the effect of early thyroidectomy
on the course and outcome of GO, future prospective randomized trials with large cohorts
and without concomitant immunosuppressive therapy are necessary.
