Most patients with thyroid eye disease (TED) have a mild, self-limited condition that requires neither medical intervention nor surgical rehabilitation, but a small subset of patients (4%–8%) develops compressive optic neuropathy (CON).1,2 The milder forms of TED can be managed with observation and supportive measures, but CON obligates therapeutic intervention. The gold-standard treatment for TED-CON has been orbital decompression surgery.3 Surgery increases the orbital volume and decreases compressive forces generated by pathologically enlarged extraocular muscles against the optic nerve at the orbital apex. Relief of pressure is immediate, but decompression surgery does not typically abbreviate the active phase of TED, which may persist for 12 to 18 months in a nonsmoker and 2 to 3 years in a smoker. Consequently, recurrence of CON is a well-established sequel to initially successful orbital decompression when performed in the active phase.4–6 Long-term failure in these cases results from the unabated expansion of the orbital soft-tissues—typical of the active phase—that ultimately fills the surgically enlarged orbit.4–6 Additional decompression surgery together with medical treatment is considered in these cases. Moreover, complications including orbital hemorrhage, enophthalmos, or hypoglobus due to overcorrection of proptosis and an increased rate of diplopia occur more frequently when orbital decompression surgery is performed in the active phase.7,8
As a consequence, many advocate medical treatment of CON during the acute phase and reserve decompression surgery for failed medical treatment and stable-phase rehabilitation.9 Both oral and parenteral corticosteroids have been used to treat active-phase TED for at least 50 years.10,11 Patients treated with corticosteroids during the acute phase of disease generally appreciate a favorable clinical response with rapid improvement in optic nerve function, but corticosteroid treatment during the stable, noninflammatory phase will not improve optic neuropathy and surgical decompression will be required. Corticosteroids reliably reverse optic neuropathy during the acute phase of TED, but because there is no evidence that they shorten the active phase of the disease, administration must continue throughout the acute phase to prevent recrudescence of the optic neuropathy. In the recent study by Shams et al,12 development of optic neuropathy only occurred in patients treated with steroids alone, not in patients who also received orbital radiotherapy (ORT).
ORT has been used for 60 years as an alternative to acute-phase decompression surgery and as an adjunct to corticosteroids in treating steroid responsive TED-CON.13 In a retrospective study of 29 patients treated with oral corticosteroids and ORT for CON, only 1 (3%) required urgent surgical decompression.14 Published studies evaluating the efficacy of ORT on TED-CON describe small cohorts.14,15 In this manuscript, we report the largest retrospective evaluation of the effectiveness of ORT in the treatment of individuals with TED-CON.
The Institutional Review Board of Columbia University approved this retrospective review of de-identified data from one of the author’s (M.K.) private practice between 1990 and 2012. The work is Health Insurance Portability and Accountability Act compliant and adheres to the tenets of the Declaration of Helsinki.
Patients with TED who underwent ORT with or without simultaneous corticosteroid therapy as primary treatment between 1990 and 2012 were included if they met the following criteria:
- Active (as defined by evidence of changing clinical measures, active inflammation, and reversibility of features with corticosteroids) TED with CON based on clinical and radiographic features.
- ORT to a total dose of 2000 cGy.
- Minimum follow up of 6 months following radiotherapy.
Excluded from our study were individuals with the following criteria:
- Orbital decompression surgery prior to ORT.
- Insufficient follow-up data.
Bilateral ORT was performed at one radiation–oncology center according to standardized protocols.14 All patients received bilateral ORT. This included patients with unilateral CON but evidence of bilateral active TED. A linear accelerator with the beam directed 5 degrees posteriorly delivered a total dose of 2000 cGy, fractionated in 10 treatments over the course of 2 weeks, to the retroocular tissues of each patient.
Corticosteroid treatment was initiated when CON was diagnosed on clinical examination, and the mechanism of compression (pathologically enlarged extraocular muscles) was confirmed by orbital imaging. The initial dose of oral prednisone was approximately 1 mg/kg/day, and gastrointestinal prophylaxis (proton-pump inhibitor) was used in all patients to prevent the development of peptic ulcer disease. Patients were re-examined in 2 weeks, and ORT was advised if measurable improvement in optic neuropathy was detected. If continued improvement in optic nerve function was noted, corticosteroids were tapered on a biweekly schedule. Specifically, the rate of taper was typically a 20 mg/day decrease every 2 weeks to 1/4 mg/kg/day and then a 10 mg/day decrease every 2 weeks to 10 mg/day. The completion of the steroid taper then slowed to a 2 mg/day every 1 to 2 weeks to allow for recovery of adrenal function. If during the corticosteroid taper, optic nerve function worsened, by any of the measurements listed below, the dose was increased and optic nerve function allowed to recover before the taper was restarted.
All ophthalmologic examinations included measurements on the International Thyroid Eye Disease Society–Vision loss/Inflammation/Strabismus/Appearance (ITEDS-VISA) form. The recorded clinical measures of optic nerve function included 1) best-corrected Snellen visual acuity, 2) color vision. Ishihara color plates were used for the first 5 years of the study after which Hardy Rand and Rittler color plates were used, 3) presence of a relative afferent pupillary defect, and 4) Humphrey 24-2 automated perimetry testing (mean deviation).
The patient’s age, sex, date of TED onset, initial and cumulative dose of corticosteroid (if given), date of ORT, corticosteroid side effects, smoking status (nonsmokers and previous/current smokers), and imaging results (CT or MRI, including presence of apical compression/optic nerve stretch) were recorded.
The primary efficacy measure of the study was failure of radiotherapy, which was defined as the need for urgent surgical decompression to reverse the optic neuropathy. The decision to proceed to urgent surgical decompression was based on the persistent recurrence of CON on attempted steroid taper following ORT. The threshold level of optic neuropathy, however, varied as considerations of the patients’ tolerance for the procedure and its morbidity varied. In general, the hierarchy of clinical findings placed greater urgency for surgical treatment of loss of visual acuity and visual field than for the dischromatopsia or small residual afferent pupil defect. The secondary outcome measure was the need for elective surgery, including orbital decompression, strabismus surgery, or eyelid retraction repair during the stable phase of the disease.
Snellen visual acuity measures were converted to LogMAR to determine the means and the magnitudes of changes. Demographic data were analyzed with descriptive statistics. Groupings of continuous variables were compared with paired and unpaired t tests, and discrete variables were compared with chi-square. Two-tailed comparisons were performed, and p values of <0.05 were held to be statistically significant.
The medical records of 122 patients (193 affected orbits) who received ORT for TED-CON were reviewed, and 104 patients (163 orbits) met the inclusion criteria. The mean (±standard deviation) age of the cohort was 61.5 ± 12.3 years; the majority (77 of 104; 74%) of the patients were female (p = 0.0004), and 32.7% (34/104) of the patients were current or previous smokers. Based on the patients’ historical accounts, TED had been present for a mean of 8.1 months prior to entering the study. Mean follow up following radiotherapy was 40.6 months (success group, 39.3 months; failure group, 60.8 months).
Corticosteroid therapy and ORT delivery during the acute phase of TED-CON successfully treated 98 of 104 (94%) patients or 152 of 163 (93.3%) orbits, thereby eliminating the need for acute-phase surgery. Successfully treated patients (SG) averaged 61.8 years of age and were predominantly female (76%). Those who required acute-phase surgery because of failed corticosteroids and ORT (FG) averaged 57.7 years of age (p = 0.43, compared with SG) with no sex predilection (50% female; p = 0.15, compared with SG) and tended to have a shorter duration of TED before beginning corticosteroids (3.8 vs. 7.5 months, p = 0.09). Smokers accounted for 30.6% (30/98) of the SG group and 67% (4/6) of the FG group (p = 0.07; Table 1).
Eleven patients (15 orbits) did not receive prednisone after radiotherapy, and none of these patients required urgent orbital decompression.
The baseline mean visual acuity in the SG group measured 0.35 ± 0.15 LogMAR (20/44 Snellen equivalent) and improved to 0.16 ± 0.13 LogMAR (20/29 Snellen equivalent, p = 0.0052) after ORT. Sixty of 95 patients had a relative afferent pupillary defect at baseline compared with only 5 of 89 patients after ORT (p < 0.0001). Complete color testing data from 140 eyes showed a correct color plate identification rate of 54.1 ± 36.6% at baseline compared with 90.5 ± 15.3% after ORT (p < 0.001). Complete Humphrey threshold visual field data from 121 eyes showed a mean deviation of −5.34 dB at baseline, which improved to −2.61 dB after ORT (p < 0.001). Compared with the corticosteroid treatment group, patients who did not receive corticosteroids had comparable pre-ORT visual acuities (0.36 ± 0.32 LogMAR vs. 0.30 ± 0.22 LogMAR; p = 0.56), post-ORT visual acuities (0.16 ± 0.13 LogMAR vs. 0.20 ± 0.13 LogMAR; p = 0.31), prevalences of relative afferent pupillary defect (30% vs. 55%; p = 0.11), and mean changes in visual field sensitivity (+2.72 ± 4.28 dB vs. 0.60 ± 3.31 dB; p = 0.11), but they had less robust improvement in color plate recognition (39.9 ± 36.0% vs. 8.0 ± 18.7%; p = 0.003; Table 2).
The baseline mean visual acuity in the FG measured 0.32 LogMAR (20/42 Snellen equivalent), decreased to 0.34 LogMAR (20/44 Snellen equivalent) after corticosteroids and ORT but before surgery, and ultimately improved to 0.13 LogMAR (20/27 Snellen equivalent) postoperatively. Four of the 6 patients had a relative afferent pupillary defect at baseline, compared with 3 of the 6 preoperatively and 2 of 6 postoperatively (p = 0.25). Patients correctly identified a mean of 26.5 ± 22.9% color plates at baseline, 33.3 ± 29.35% preoperatively and 76.5 ± 25.5% postoperatively (p = 0.0003, baseline vs. postoperative). Complete Humphrey threshold visual field data from 8 eyes showed a mean deviation of −4.06 ± 5.22 dB at baseline, −10.19 ± 5.71 dB after ORT but before surgery (p = 0.01), and −4.19 ± 4.98 dB postoperatively (p = 0.96 vs. baseline; p = 0.03 vs. after ORT but before surgery).
Patients in both the SG and FG (postdecompression) experienced similar improvements in mean visual acuity (Δ −0.19 LogMAR vs. Δ −0.19 LogMAR; p = 0.95), color vision (+36.4% vs. +50.0%; p = 0.22), and Humphrey visual field sensitivity (Δ +2.73 dB vs. Δ −0.13; p = 0.06).
The orbital apices appeared crowded in 94 of 95 (98.9%) SG patients and 4 of 6 (66.6%) FG patients (p < 0.0001). The optic nerves appeared stretched in 1.1% (1 of 95) of SG patients and 33.3% (2 of 6) of FG patients (p < 0.0001).
Cumulative Corticosteroid Dose
The mean cumulative prednisone dose given to all patients was 3.05 g (equivalent to 2.44 g of methylprednisolone). The cumulative corticosteroid dose given to the SG (2.85 ± 1.74 g of prednisone) was significantly lower than that received by the FG (6.21 ± 1.57 g of prednisone; p < 0.0001). Nonsmokers and smokers received similar cumulative doses of prednisone (3.30 ± 1.46 g vs. 3.48 ± 2.15 g; p = 0.64).
The use of corticosteroids deviated from the previously mentioned strategy in 11 patients, who opted not to take corticosteroids (CS).
Sixty-two of 98 (63.3%) successfully treated patients did not require surgery. Twenty of 152 (13.2%) orbits underwent elective orbital decompression during the stable phase of TED, 23 of 304 (7.6%) eyelids underwent repair of retraction, and 23 (23.5%) patients required strabismus procedures. The subgroup of the SG patients who did not receive corticosteroids experienced similar surgical rates: 63.7% did not require surgery, 13.3% underwent elective decompression surgery, 6.7% underwent eyelid retraction repair, and 30% required strabismus surgery. All orbits in the FG required surgical decompression for CON and 4 of 6 (66.7%) underwent additional stable-phase surgery: 3 (50%) required strabismus surgery, and 3 of 22 (13.6%) eyelids underwent retraction repair (Table 3).
Ninety-six of the 104 patients received corticosteroids, and 39 patients reported no corticosteroid-related side effects. Three patients did not take steroids beyond the 2-week pre-ORT trial during which CON reversibility was determined, and 8 patients took no corticosteroids due to a previous history of side effects or having multiple medical comorbidities that unacceptably increased the risk of complications. Corticosteroid side effects experienced by study patients included weight gain, insomnia, elevated blood glucose, facial swelling, dyspepsia, anxiety, and headache.
No patients developed adverse effects attributable to ORT, and none suffered radiation-induced optic neuropathy or retinopathy.
The active phase of TED is self-limited, lasting 12–18 months in nonsmokers and as much as 3 years in smokers. The clinical signs and symptoms of TED fluctuate during the active phase and typically improve in response to the administration of either oral or intravenous (IV) corticosteroids. The benefits of corticosteroids are, however, generally limited to the duration of the treatment, and the recurrence of the inflammatory signs and symptoms of TED after withdrawal of corticosteroids is a reliable barometer of ongoing active disease. In a prospective, randomized trial of patients with moderate to severe active TED, Bartalena et al17 administered 3 different cumulative doses of intravenous methylprednisolone (2.25 g, 4.98 g, and 7.47 g) over a period of 3 months. Ophthalmic measures of disease improved in 52% of cases in the high dose group, 35% in the medium dose group, and 28% in the low dose group. The high dose group experienced the greatest improvement in ocular motility, clinical activity score, and quality of life.17 Active disease frequently recurred, however, after corticosteroids were discontinued. Of the patients in the high dose, medium dose, and low dose groups that improved at 12 weeks, active TED relapsed 3 months after withdrawal of corticosteroids in 33%, 21%, and 40%, respectively. Moreover, 6% of these cases eventually developed CON. These data suggest that while corticosteroid treatment effectively suppresses the active phase of TED, it does not modify the natural history of the immune process by reducing the duration of the active phase.
Most cases of moderately severe and nearly all cases of mild TED are treated expectantly because the side effects from the lengthy course of corticosteroid treatment that would be required if employed for the duration of the active phase outweigh the benefits of disease suppression. Moreover, spontaneous improvement of TED is sufficiently common that most patients enjoy a durable remission and require neither medical nor surgical treatment.
TED-CON occurs in approximately 1% of patients with TED and can lead to devastating loss of vision if not treated. The active phase of CON usually responds well to moderate or high doses of corticosteroids, but CON often recurs when steroids are withdrawn. In Kahaly’s18 12-week study comparing IV with oral (PO) corticosteroids for CON, 1 of 3 (33.3%) patients in the PO group and 5 of 5 (100%) in the IV group responded favorably. In Marcocci’s19 12-month study, 3 of 9 (33.3%) patients in the PO group and 11 of 14 (78.6%) patients in the IV group responded favorably. Unfortunately, neither study included a sufficient follow-up period to identify the rate of recurrent CON after steroid withdrawal, but the previously referenced Bartalena trial suggests that recurrent CON could be anticipated in a significant fraction of patients treated with corticosteroids alone.
Recurrent TED-CON after initially successful corticosteroid administration requires additional therapy. Reinstitution of corticosteroid therapy may successfully treat CON, but the risk of steroid-related morbidity increases with prolonged treatment. Many of these patients undergo orbital decompression surgery to resolve the volumetric imbalance resulting from the pathologic re-expansion of the orbital soft tissues.20 As with corticosteroid treatment, orbital decompression does not shorten the duration of the active phase of TED. Early postoperative reversal of the optic neuropathy may be anticipated but without modification of the ongoing active disease phase or coincidental disease remission, re-compression of the optic nerve can occur.4 Recurrent CON has also been described when the extraocular muscles swell rapidly soon after successful orbital decompression, as has been seen following surgical decompression of a compartment syndrome.4
ORT for TED was first reported in 1936, and the modern protocol was developed by Donaldson et al16 in 1973.13 Small, retrospective studies have since reported the effects of ORT on moderate to severe TED, either alone or more commonly in combination with either PO or IV corticosteroids. The response rate for these patients treated with ORT and PO corticosteroids (70%) is comparable to that with ORT and IV corticosteroids (80%).19,21–24
Three prospective trials have compared ORT to sham treatment; however, these did not specifically address patients with CON.25–27 In patients with moderate TED, ORT most effectively improved extraocular motility and the severity of diplopia.25 In patients with moderate–severe TED (excluding TED-CON), improved ocular elevation occurred in 60% of ORT-treated patients compared with 31% in the sham-radiation group.26 In a trial of 42 patients with mild–moderate disease, one orbit of each patient was treated with ORT and the other with sham radiation, followed 6 months later with ORT or sham to the other orbit.27 No differences between the 2 orbits were measured at 6 months, and only minimal improvements in muscle volume and proptosis were noted at 12 months.27 Patients within the cohort were diagnosed with TED an average of 18 months prior to enrollment, suggesting that the cohort was predominantly composed of stable-phase patients who would not be expected to benefit from treatment. Moreover, patients with TED-CON were excluded from the study.
For the past 40 years we have employed a combination of ORT and PO corticosteroids for CON-TED as an alternative to acute-phase surgical decompression or prolonged corticosteroid treatment. Only 6 of 104 patients in our cohort required urgent-phase orbital decompression, leading us to believe that this combination significantly reduces the need for acute-phase surgery.
Comparative trials suggest that ORT is at least as effective as PO corticosteroids for TED, but only 1 of these studies included patients with TED-CON.14,21,28,29 In a small retrospective review, Kazim et al14 showed that only 1 of 29 patients with TED-CON treated with a combination of ORT and corticosteroids required urgent orbital decompression, while 6 of 16 treated with PO corticosteroids alone required orbital decompression. Based on these results, we have routinely combined oral corticosteroids with ORT for patients with steroid-responsive TED-CON who do not have an increased risk of radiation retinopathy (patients with diabetes mellitus or poorly controlled systemic arterial hypertension). The current larger study has produced similarly favorable results when ORT is combined with PO corticosteroids.
Corticosteroid administration to patients with TED-CON serves 3 purposes. Corticosteroids rapidly reduce inflammation associated with the active phase of TED and abruptly improve CON-TED. A favorable response to corticosteroids establishes the potential therapeutic value of ORT, but when a 2-week trial of prednisone (1 mg/kg) fails to reverse TED-CON, then the addition of ORT is unlikely be effective. In such cases, we proceed to surgical decompression. The effect of ORT is first appreciated by 4 weeks, but the full benefit may not be seen until 3 months. The dose of corticosteroids can be tapered during this period at a rate titrated against the response of the CON.21–23,31,32 In the current study, the combination of oral corticosteroids and radiotherapy produced a durable response rate of 94%. In the few cases that failed combination medical treatment (6%), orbital decompression surgery successfully reversed the optic neuropathy. These patients experienced similar improvements in visual acuity, color vision, and Humphrey visual field testing, suggesting that they were not harmed by failed medical therapy and the delay in surgical intervention. We also believe that the absence of recurrent CON after withdrawing corticosteroids provides evidence that TED enters the stable phase. While advancing fibrosis may progress as measured by an interval of progressive strabismus, it is unclear as to whether this is a sign of TED activity. However, steroid reversible orbital edema and its attendant CON are signs of disease activity, which reliably respond to ORT and do not recur.
We achieved a high response rate despite the use of PO rather than IV corticosteroids. The cumulative prednisone dose used in our study is equivalent to the low dose group in the article by Bartalena, yet our relapse rate as measured by recurrent CON requiring urgent orbital decompression was much lower (6% vs. 40%).17 The end point of our study was different from that of Bartalena—recurrent optic neuropathy following withdrawal of corticosteroids that required urgent decompression versus recurrent orbital inflammatory signs as measured by Clinical Activity Score. While these end points are not directly comparable, we believe that steroid-responsive TED-CON is a very reliable measure of active TED. We attribute the considerable difference in outcomes between the 2 studies to the addition of ORT to our treatment protocol. We did not deliver corticosteroids intravenously because we have seen favorable resolution of CON with oral prednisone. Nonetheless, we would encourage a trial to compare the effects of oral versus intravenous corticosteroids when administered with ORT in treatment of TED-CON.
Further evidence that ORT produces a favorable disease-modifying effect emerged from a recent large retrospective review by Shams et al.12 The efficacy of PO or IV corticosteroid monotherapy was compared with corticosteroid therapy combined with ORT in a group of patients with moderately severe active-phase TED who did not have CON.12 Strikingly, 25 of 144 patients treated with corticosteroid monotherapy developed TED-CON during the study (5 required urgent decompression surgery), whereas 0 of 105 patients treated with ORT–corticosteroid combination therapy developed TED-CON.
After ORT and corticosteroids have reversed CON and stabilizes patients with TED, orbital decompression surgery can be performed electively in the stable phase, when the functional and cosmetic results are most predictable, or it can be deferred altogether. Because the risks of surgical decompression (bleeding, infection, visual loss, increased diplopia/strabismus, and sinusitis) are greater in the acute phase, circumventing its use in patients who can tolerate orbital radiation is advantageous. When orbital decompression surgery is performed during the acute phase to recover optic nerve function, but it is impossible to predict the final position of the globes, so residual proptosis or enophthalmos is more common. Surgical decompression in the stable phase produces more durable and predictable results when the goals are limited to proptosis reduction. In this clinical setting, graded surgical decompression can be used to achieve the desired reduction in proptosis. In this study, orbital decompression was avoided in 80% of patients, and only 13% of orbits underwent elective stable-phase decompression. Thirty-nine percent of successfully treated patients underwent strabismus, eyelid retraction, and orbital decompression surgery in the stable phase, compared with 67.7% of patients in the FG. While the difference in the percentage of patients who proceeded to orbital decompression surgery is apparent, the difference in the rate of strabismus surgery is also in part related to the rate of decompression surgery because, on average, 50% of patients who undergo medial wall decompression required in cases of CON will develop new or worsened diplopia.
Failure of combined ORT and corticosteroid therapy was more common in patients with radiographic evidence of optic nerve stretch and those without crowding of the orbital apices. The number of patients in the failure group was small, and larger studies are needed to establish the significance of these findings. While optic nerve stretch might be considered as an etiology separate from the apical crowding seen in typical TED-CON, we included these 3 cases (1 successfully treated and 2 which failed treatment) as we thought that they represent a form of generalized orbital compression. However, we do think that despite the small numbers that it is important to recognize this as a relative contraindication to the use of ORT. Smoking has been associated with failure of corticosteroid therapy, and our data showed a strong tendency (p = 0.07) toward failure of combined CS and ORT in active and previous smokers.33
The safety profile of ORT is well established. Except in patients with diabetes mellitus (particularly those with diabetic retinopathy), poorly controlled systemic arterial hypertension, or those in whom the ORT is improperly delivered, the risks of modern ORT protocols for treating orbitopathy (other than the production of cataract) are low. None of the patients in our series developed radiation retinopathy or optic neuropathy.
The primary goal of acute phase medical therapy is to effectively treat the complications of TED-CON while reducing the need for surgery. Orbital decompression remains a valuable treatment option for TED-CON, but the current study shows that ORT is an effective alternative in patients with steroid-responsive TED-CON. ORT eliminates the need for urgent orbital decompression surgery in 94% and elective decompression in 87% of TED-CON cases.
The prime limitation of the current study is its retrospective nature. Moreover, the study lacks a control group, so comparison to studies that included acute-phase orbital decompression or IV corticosteroids should be cautionary. Specifically, we had no control CS-only group because the pattern of practice had evolved to the combined use of ORT + CS as primary therapy consequent to the prior smaller studies that supported ORT + CS as compared with CS alone.14,21,23 The major strength of this study is the size of the cohort (104 patients/163 orbits), the largest to ever evaluate combined ORT–corticosteroid treatment of TED-CON. This study shows favorable outcomes in patients treated with ORT for TED-CON, which may limit the need for corticosteroid treatment and modify the natural history of TED. We recommend a multicenter, prospective trial to further investigate the use of ORT for the treatment of TED-CON.
1. Bartley GBThe epidemiologic characteristics and clinical course of ophthalmopathy associated with autoimmune thyroid disease in Olmsted County, Minnesota. Trans Am Ophthalmol Soc 1994;92:477–588.
2. Neigel JM, Rootman J, Belkin RI, et alDysthyroid optic neuropathy. The crowded orbital apex syndrome. Ophthalmology 1988;95:1515–21.
3. Perumal B, Meyer DRTreatment of severe thyroid eye disease: a survey of the American Society of Ophthalmic Plastic and Reconstructive Surgery (ASOPRS). Ophthal Plast Reconstr Surg 2015;31:127–31.
4. Kauh CY, Gupta S, Douglas RS, et alCompressive optic neuropathy and repeat orbital decompression: a case series. Ophthal Plast Reconstr Surg 2015;31:385–90.
5. Wenz R, Levine MR, Putterman A, et alExtraocular muscle enlargement after orbital decompression for Graves’ ophthalmopathy. Ophthal Plast Reconstr Surg 1994;10:34–41.
6. Zhang-Nunes SX, Dang S, Garneau HC, et alCharacterization and outcomes of repeat orbital decompression for thyroid-associated orbitopathy. Orbit 2015;34:57–65.
7. McNab AAOrbital decompression for thyroid orbitopathy. Aust N Z J Ophthalmol 1997;25:55–61.
8. Dumont N, Bouletreau P, Guyot L[Reoperation after orbital decompression for Graves’ ophthalmopathy]. Rev Stomatol Chir Maxillofac 2012;113:81–6.
9. Wakelkamp IM, Baldeschi L, Saeed P, et alSurgical or medical decompression as a first-line treatment of optic neuropathy in Graves’ ophthalmopathy? A randomized controlled trial. Clin Endocrinol (Oxf) 2005;63:323–8.
10. Brown J, Coburn JW, Wigod RA, et alAdrenal steroid therapy of severe infiltrative ophthalmopathy of Graves’ disease. Am J Med 1963;34:786–95.
11. Snyder NJ, Green DE, Solomon DHGlucocorticoid-induced disappearance of long-acting thyroid stimulator in the ophthalmopathy of graves’ disease. J Clin Endocrinol Metab 1964;24:1129–35.
12. Shams PN, Ma R, Pickles T, et alReduced risk of compressive optic neuropathy using orbital radiotherapy in patients with active thyroid eye disease. Am J Ophthalmol 2014;157:1299–305.
13. Jones AOrbital X-ray therapy of progressive exophthalmos. Br J Radiol 1951;24:637–46.
14. Kazim M, Trokel S, Moore STreatment of acute Graves orbitopathy. Ophthalmology 1991;98:1443–8.
15. Rush S, Winterkorn JM, Zak RObjective evaluation of improvement in optic neuropathy following radiation therapy for thyroid eye disease. Surv Ophthalmol 2001;45(5):458–9.
16. Donaldson SS, Bagshaw MA, Kriss JPSupervoltage orbital radiotherapy for Graves’ ophthalmopathy. J Clin Endocrinol Metab 1973;37:276–85.
17. Bartalena L, Krassas GE, Wiersinga W, et alEuropean Group on Graves’ Orbitopathy. Efficacy and safety of three different cumulative doses of intravenous methylprednisolone for moderate to severe and active Graves’ orbitopathy. J Clin Endocrinol Metab 2012;97:4454–63.
18. Kahaly GJ, Pitz S, Hommel G, et alRandomized, single blind trial of intravenous versus oral steroid monotherapy in Graves’ orbitopathy. J Clin Endocrinol Metab 2005;90:5234–40.
19. Marcocci C, Bartalena L, Tanda ML, et alComparison of the effectiveness and tolerability of intravenous or oral glucocorticoids associated with orbital radiotherapy in the management of severe Graves’ ophthalmopathy: results of a prospective, single-blind, randomized study. J Clin Endocrinol Metab 2001;86:3562–7.
20. Bartalena L, Pinchera A, Marcocci CManagement of Graves’ ophthalmopathy: reality and perspectives. Endocr Rev 2000;21:168–99.
21. Bartalena L, Marcocci C, Chiovato L, et alOrbital cobalt irradiation combined with systemic corticosteroids for Graves’ ophthalmopathy: comparison with systemic corticosteroids alone. J Clin Endocrinol Metab 1983;56:1139–44.
22. Tsujino K, Hirota S, Hagiwara M, et alClinical outcomes of orbital irradiation combined with or without systemic high-dose or pulsed corticosteroids for Graves’ ophthalmopathy. Int J Radiat Oncol Biol Phys 2000;48:857–64.
23. Ng CM, Yuen HK, Choi KL, et alCombined orbital irradiation and systemic steroids compared with systemic steroids alone in the management of moderate-to-severe Graves’ ophthalmopathy: a preliminary study. Hong Kong Med J 2005;11:322–30.
24. Johnson KT, Wittig A, Loesch C, et alA retrospective study on the efficacy of total absorbed orbital doses of 12, 16 and 20 Gy combined with systemic steroid treatment in patients with Graves’ orbitopathy. Graefes Arch Clin Exp Ophthalmol 2010;248:103–9.
25. Prummel MF, Terwee CB, Gerding MN, et alA randomized controlled trial of orbital radiotherapy versus sham irradiation in patients with mild Graves’ ophthalmopathy. J Clin Endocrinol Metab 2004;89:15–20.
26. Mourits MP, van Kempen-Harteveld ML, García MB, et alRadiotherapy for Graves’ orbitopathy: randomised placebo-controlled study. Lancet 2000;355:1505–9.
27. Gorman CA, Garrity JA, Fatourechi V, et alA prospective, randomized, double-blind, placebo-controlled study of orbital radiotherapy for Graves’ ophthalmopathy. Ophthalmology 2001;108:1523–34.
28. Wiersinga WM, Smit T, Schuster-Uittenhoeve AL, et alTherapeutic outcome of prednisone medication and of orbital irradiation in patients with Graves’ ophthalmopathy. Ophthalmologica 1988;197:75–84.
29. Prummel MF, Mourits MP, Blank L, et alRandomized double-blind trial of prednisone versus radiotherapy in Graves’ ophthalmopathy. Lancet 1993;342:949–54.
30. Noth D, Gebauer M, Müller B, et alGraves’ ophthalmopathy: natural history and treatment outcomes. Swiss Med Wkly 2001;131:603–9.
31. Kahaly G, Böckmann H, Beyer J, et alLong-term observation of endocrine ophthalmopathy and retrospective appraisal of therapeutic measures. J Endocrinol Invest 1990;13:287–92.
32. Marcocci C, Bartalena L, Bogazzi F, et alOrbital radiotherapy combined with high dose systemic glucocorticoids for Graves’ ophthalmopathy is more effective than radiotherapy alone: results of a prospective randomized study. J Endocrinol Invest 1991;14:853–60.
33. Prummel MF, Wiersinga WMSmoking and risk of Graves’ disease. JAMA 1993;269:479–82.