The Visual Function of Emergent Transnasal Endoscopic Orbital Decompression Surgery for Patients With Steroid-Resistant Dysthyroid Optic Neuropathy

Abstract

Background:

Dysthyroid optic neuropathy (DON) is occasionally resistant to corticosteroid treatment. In these cases, emergent transnasal endoscopic orbital decompression may be the optimal therapy method. We evaluated 13 patients with steroid-resistant DON to determine the potential benefit of this surgical intervention.

Methods:

Thirteen patients with steroid-resistant DON seen at Shenzhen Eye Hospital (China) between January 2023 and April 2025 were included in our study. All patients underwent emergent transnasal endoscopic orbital decompression and received postoperative corticosteroids and neurotrophic medication. The rate of improvement in postoperative best-corrected visual acuity (BCVA) was an important criterion for successful postoperative therapy.

Results:

In all, 13 patients (range, 34–72 y) with 13 eyes (8 right eyes and 5 left eyes) were included in our study. The duration of DON ranged from 9 to 36 months (median time 17.4 mo). Eleven of 13 patients (84.61%) had a statistically significant improvement in BCVA from 0.76±0.28 LogMAR to 0.23±0.18 postoperatively (P<0.01). These 11 patients underwent decompression within <26 months of diagnosis. Three of these 11 patients demonstrated pronounced BCVA improvement: 2 of them improved from counting fingers to 0.6 and 0.02, respectively, and the other improved from hand motion to 0.3. However, the remaining 2 showed no improvement.

Conclusion:

Emergent transnasal endoscopic orbital decompression surgery is a very useful and effective treatment for dysthyroid optic neuropathy (DON) that is refractory to corticosteroid therapy. Its efficacy seems inversely correlated with disease duration rather than preoperative BCVA. As such, longer disease duration may be associated with poorer postoperative outcomes.

Dysthyroid optic neuropathy (DON) is a sight-threatening emergency in patients with thyroid-associated ophthalmopathy (TAO). It is characterized by optic nerve compression at the crowded orbital apex.¹ Other clinical manifestations of DON include deterioration in best-corrected visual acuity (BCVA), diplopia, presence of a relative afferent pupillary defect, visual field defects, visual evoked potential (VEP)-P100 anomalies, contrast sensitivity loss, optic disc changes, folded choroid, and orbital apex crowding.^2,3 Vision loss occurs rapidly in patients who do not receive urgent treatment. DON may be initially treated with corticosteroids. For cases of steroid-resistant DON, emergent orbital decompression is necessary. Orbital decompression surgery requires access to the medial and inferomedial quadrants of the orbital apex, a technically challenging task. These quadrants correspond to the areas where the optic nerve is most vulnerable to compression from an enlarged medial rectus and/or inferior rectus muscle, especially the medial rectus.^4,5

The endoscopic transnasal approach to medial orbital wall and orbital apex decompression represents a major, well-justified advance in the surgical treatment of DON.⁶ Its principal advantages include: (1) superb visualization of the decompression site under high-definition endoscopic magnification; (2) the capacity to perform precise bone removal medial to the optic nerve; (3) a lower risk of postoperative strabismus compared to techniques that involve the orbital floor; (4) direct access to the tightest area of the orbital apex, the medial quadrant, without needing to compromise orbital structural support.

In cases of dysthyroid optic neuropathy (DON) where pulse steroid therapy fails to rescue vision, emergency endoscopic transnasal orbital decompression is required.⁷ This retrospective study evaluated the outcomes of emergency endoscopic transnasal orbital decompression in 13 patients with steroid-resistant dysthyroid optic neuropathy at the hospital.

METHODS

We collected the clinical data of patients with dysthyroid optic neuropathy who were admitted to the Department of Orbitopathy and Ophthalmic Oncology of Shenzhen Eye Hospital (China) between January 2023 and April 2025. According to the 2021 European Group on Graves’ orbitopathy (EUGOGO) clinical practice guidelines,⁸ patients with dysthyroid optic neuropathy were treated with several high single doses of i.v. methylprednisolone per week. Written informed consent was waived by the Ethics Committees of Shenzhen Eye Hospital due to the retrospective nature of this study, the research abided by the declaration of Helsinki, and the research was conducted in compliance with the Ethics Committees of Shenzhen Eye Hospital (No. 2026KYPJ029).

We must do an ophthalmological examination to assess the visual acuity of patients, which helps to exclude other diseases of optic neuropathy, including traumatic optic neuropathy, retinopathy, and preexisting optic neuropathy. BCVA was measured using a Snellen chart, and BCVAs <0.01 were documented as either counting fingers (CF), hand motion (HM), light perception (LP), or no light perception (NLP). Patients also underwent preoperative and postoperative visual field (VF) (Humphrey HFA; Carl Zeiss, Germany) testing, visual evoked potential (VEP) (ROLAND, Germany) testing, optical coherence tomography (OCT) (Carl Zeiss Meditec, United States, or Heidelberg Engineering, Germany), anterior segment photography (CSO, Italy), and fundus photography. Patients with relative afferent pupillary defects underwent computed tomography (CT) and magnetic resonance imaging (MRI) to determine the underlying pathology. These imaging modalities were also used to specifically measure any changes in the orbital structures, the degree of optic nerve compression, and the degree of extraocular muscle enlargement. Postoperatively, all patients were given medications, such as hormones, nutritional nerves, and vasodilators. Postoperative BCVA measurement, VF testing, and VEP testing were also done.

Surgery

All surgeries were performed by a single surgeon with 10 years of experience. The surgery proceeded as follows: the procedure was performed under general anesthesia with the patient in a supine position and the head elevated at 30 degrees. The nasal cavity was disinfected with 0.05% povidone–iodine and then packed with gauze soaked in 1% lidocaine (with 1:1000 epinephrine) for vasoconstriction. An endoscopic total sphenoethmoidectomy was performed by using a 0-degree endoscope (Karl Storz, Tuttlingen, Germany),^9,10 which provided access to the sphenoid sinus ~1 to 1.3 cm above the nasal septum. The posterior ethmoidal wall was removed with a tricut blade (4.0 mm; Ref. 1,884,004; Medtronic, USA), which provided visualization of the optic nerve canal and internal carotid artery. The maxillary sinus was accessed via the superior edge of the inferior turbinate or the maxillary ostium. This provided better exposure of the ethmoid and sphenoid sinuses (Fig. 1A), allowing for adequate exposure of the bony medial orbital wall (Fig. 1B). Blood was cleared from the surgical site to provide adequate exposure of the optic canal (Fig. 1C). An anterior maxillotomy was additionally performed to improve visualization of the posterior wall. An incision was made on the lateral nasal wall to resect the lamina papyracea, the anterior portion of the optic nerve canal, and the medial wall of the pterygopalatine fossa to obtain adequate exposure. A microdrill (XPS3000; Medtronic, USA) was used at the rotational speed of 5000 rpm to thin the medial and anterior walls of the optic nerve canal and pterygopalatine fossa, which were then carefully removed with a dissector (Fig. 1D, E). Meticulous care was taken throughout to avoid injury to the optic nerve, ethmoidal arteries, and sphenopalatine artery. At the posterior maxillary wall, a 3 to 4 mm segment of bone and the lamina papyracea were resected inferiorly along the inferomedial strut. Preserving this inferomedial strut (Fig. 1F) helps prevent inferomedial displacement of the globe. A 20-gauge MVR blade (MVR-Lance 20G; MANI, Japan) was used to dissect the soft tissues comprising the superior and inferior quadrants of the orbital apex. These tissues were reflected toward the maxillary line, which created an incision that was parallel to the medial rectus muscle and the annulus of Zinn (Fig. 1G). Extraconal and intraconal orbital fat (Fig. 1H) was removed using a low-suction cutting instrument (New Direction Medical Optical Instrument Co., Ltd., Dezhou, China). This resection was performed gently, under direct visualization to avoid vascular injury. The amount of fat removed was determined by the degree of proptosis reduction required. The hypertrophic medial rectus muscle was visualized, and any herniation into the nasal cavity was noted (Fig. 1I). Given the technically demanding nature of intraconal fat decompression, this procedure was not routinely done in patients with mild proptosis. The site of orbital soft-tissue herniation was covered with a bioabsorbable nasal dressing (Merogel; Medtronic Xomed, USA) soaked in triamcinolone.

FIGURE 1.

Main surgery procedures. Surgical image: (A) Exposure of the ethmoid sinus (ES) and sphenoid sinus (MS). (B) The area between the upper and lower yellow dashed lines indicates the portion of the medial orbital wall bone requiring removal. (C) The location marked by the yellow dotted circle is the optic canal orifice. (D) The yellow arrow indicates the orbital plate to be removed. (E) Removal of bone at the optic canal orbital opening. (F) Preserving the inferomedial strut. (G) Incise the orbital fascia longitudinally along the long axis of the orbit, superiorly and inferiorly, respectively. Incision of the annulus of Zinn area. (H) Removal of partial intraorbital fat (OF). (I) The hypertrophic medial rectus muscle (MR) was visible. Herniation of the medial rectus muscle into the nasal cavity.

In summary, the performed decompression extended posteriorly from the posterior lacrimal crest and the lesser wing of the sphenoid bone to the anterior portion of the optic canal. The inferior boundary was defined by the line connecting the posterior wall of the maxillary sinus and the medial wall of the pterygopalatine fossa. This adequately reduced pressure at the orbital apex and alleviated optic nerve compression. All patients underwent decompression of both extraconal and intraconal orbital fat.

Statistical Analysis

A paired-sample t test was used to compare the best-corrected visual acuity (BCVA) before and after surgery. Statistical significance was considered to be P<0.05.

RESULTS

Patient Characteristics

Thirteen patients (range, 34–72 y) with steroid-resistant DON were included in our study. The median follow-up time was 3.7 months (range, 2–5 mo). Patient demographics and baseline thyroid function are clearly shown in Table 1, Supplemental Digital Content 1, http://links.lww.com/SCS/J718.

Clinical Characteristics

CT imaging provided measurements for preoperative and postoperative medial rectus diameter, degree of optic nerve decompression, and extent of medial wall decompression (Fig. 2A–F). Preoperative visual fields showed a defect (Fig. 3A), and postoperative visual fields improved significantly (Fig. 3B, C).

FIGURE 2.

Preoperative and postoperative CT images. (A–C) Preoperative orbital computed tomography (CT) thin-slice scan. The diameter of the medial rectus muscle was 7.1 mm. Preoperative exophthalmos measured 22.4 mm. (D–F) Postoperative orbital computed tomography (CT) thin-slice scan. The yellow dashed line and the red dashed line reflect the degree of protrusion of the medial rectus muscle into the nasal cavity. The diameter of the medial rectus muscle was 8.5 mm. Postoperative exophthalmos measured 18.4 mm.

FIGURE 3.

Preoperative and postoperative visual fields. Visual field: (A) Before surgery, almost complete visual field defect; (B) 1 month after surgery, the nasal visual field showed minimal improvement; (C) at the 3-month postoperative follow-up, the visual field defect had completely resolved.

The amplitude of N1, P1, N2, and N3 had decreased apparently (Fig. 4) in the affected eye compared with the fellow eye before surgery. Comparatively, the postoperative VEP test showed significant improvements in these defects (Fig. 5).

FIGURE 4.

Before surgery, the amplitude of N1, P1, N2, and N3 appeared to have decreased in the right eye, compared with the left eye.

FIGURE 5.

At the 3-month postoperative follow-up, the amplitudes of N1, P1, N2, and N3 waveforms in the operated eye had improved significantly, demonstrating no significant difference from the fellow eye.

Best-Corrected Visual Acuity (BCVA)

The time from DON to surgery ranged from 9 to 36 months (median time 17.4 mo). Eleven (84.61%) patients noted improved BCVA postoperatively. Three of these 11 patients pronounced significant BCVA improvement, 2 improved from CF to 0.6 and 0.02, respectively, and 1 improved from HM to 0.3. There was no BCVA improvement in 2 patients. These 2 patients had the longest interval between diagnosis and surgery, with both intervals lasting 36 months. Table 2, Supplemental Digital Content 2, http://links.lww.com/SCS/J719, details the aforementioned information. Preoperative assessment revealed a reduced mean in BCVA of 0.76±0.28 LogMAR (range 1.32–0.28) in all 13 eyes. At the final postoperative BCVA, the mean BCVA improved to 0.23±0.18 (P<0.01).

Complication

The surgery proceeded uneventfully. Four cases developed diplopia after surgery. In 2 of these cases, diplopia resolved after 2 months of visual training; however, in the other 2, it resolved after receiving strabismus surgery.

DISCUSSION

The incidence of TAO is 15 per 100,000, and DON develops in 3% to 5% of these cases.¹¹ DON is the most serious and disabling complication of TAO. Without timely treatment, it can result in visual loss slowly over time. However, the underlying mechanism of DON remains unclear. Increased retrobulbar pressure, compression of optic nerve, rarely, stretching of the optic nerve.¹² In DON, auto-immune mediated inflammation of the retrobulbar fat and extraocular muscles (EOMs) typically results in orbit apex syndrome.¹³ Anatomically, the medial rectus muscle is positioned closest to the optic nerve. An increase in medial rectus diameter is most closely associated with an increased risk for optic nerve compression.¹⁴

CT imaging allows for direct evaluation of the spatial relationship between the optic nerve and the hypertrophied medial rectus muscle. In our study, the preoperative diameter of the medial rectus muscle was 7.1 mm (Fig. 2A–C), which is significantly greater than its expected diameter. This supports the theory that increased pressure within the orbital apex causes compression of the optic nerve, leading to vision loss. Tansnasal endoscopic orbital decompression allows the medial rectus muscle to herniate into the nasal cavity (Fig. 2D–F), which alleviates pressure on the optic nerve. As with the results of Figures 3A and 4, it demonstrated visual field defects and evoked potential (VEP) abnormality. The patients of TAO with the thickening of the medial rectus had not suffered from vision loss or vision deterioration, which indirectly tested the effectiveness of transnasal endoscopic orbital decompression.

Any interventions for DON aim to relieve pressure at the orbital apex by expanding the bony orbit's orbital volume and/or decreasing orbital soft-tissue volume. Endoscopic transethmosphenoid surgery is widely used for traumatic optic neuropathy because it provides clear access to the orbital apex without requiring neurosurgery or risking neurovascular structures.^15–18 Our surgery, which includes adequate removal of the medial wall, incision of the orbital fascia, widening of the optic canal, and incision of the annulus of Zinn, was safe and effective for patients with steroid-resistant DON. The majority of patients noted improved BCVA postoperatively. Our surgical technique was based on the study published by Wu et al.¹⁶ Adequate orbital apex decompression is crucial for the treatment of compressive optic neuropathy. Figure 1D, E illustrates the removal of the bony anterior wall of the optic canal orifice as well as the pterygopalatine fossa. Figure 1G depicts the incision performed at the annulus of Zinn and surrounding periorbital tissue. Figure 1G, H demonstrates the incision of the orbital fascia parallel to the medial rectus muscle and the aspiration of orbital fat, respectively. These steps, combined with the preservation of the inferomedial strut (Fig. 1F), maximized the possible improvements in BCVA after surgery while minimizing the risk of diplopia. In our study, 11 patients out of 13 patients (84.61%) had a statistically significant improvement in BCVA from 0.76±0.28 LogMAR to 0.23±0.18 postoperatively (P<0.01). Three patients had significantly improved BCVA from HM or CF to at least >0.2. VF and VEP test results also showed improvements between 2 and 4.5 months postoperatively (Figs. 3B, C and 5). In their retrospective study, Wang et al¹⁹ reported a BCVA improvement from 0.79±0.77 LogMAR preoperatively to 0.21±0.27 LogMAR postoperatively (P<0.001) among patients treated with combined direct orbital and endoscopic endonasal decompression. Rotenberg et al²⁰ also demonstrated BCVA improvement in their patients postoperatively. Our results suggest that transnasal endoscopic orbital decompression adequately addresses optic nerve compression by creating a new anatomic space for the hypertrophied medial rectus muscle (Fig. 2A–F), postoperative BCVA is directly correlated with the surgeon skills, as well as the patient’s age, preoperative BCVA, and duration of DON.¹² Among the patients in this study, DON duration was ≤26 months in 11 cases. These patients all noted improvement in BCVA after surgery. Compared with the other patients, 2 had durations of 36 months; BCVA improvement was not noted in these patients after surgery. Previous research has demonstrated that disease duration is a critical determinant of BCVA among patients with TAO and DON.²¹ Kemchoknatee et al²² reported that prolonged DON is 5 times less likely to respond to intravenous methylprednisolone. These results were reflected in our study. Specifically, the patients with the longest disease durations did not show improvements in BCVA following surgery. The specific mechanisms underlying the poor postoperative outcomes in patients with a prolonged DON remain unclear. Nevertheless, decompression should be recommended in these patients, along with more modern therapy, such as teprotumumab and tocilizumab. These were offered to the 2 patients in our study. However, they declined due to financial constraints.

Diplopia is the most common and major complication of orbital decompression surgery for DON. In a previous study, almost 12% of patients who underwent decompression surgery presented diplopia postoperatively.²³ In our research, 4 patients developed new diplopia, 2 received visual training for 2 months, and finally, diplopia disappeared. However, in 2 other patients who did not receive adequate treatment, diplopia persisted and ultimately resolved following strabismus surgery.

Our study has some limitations. First, our study examined a small sample size of patients. Expanding this research to other centers may increase the significance of our study. A larger sample size may also be able to identify rare complications, such as cerebrospinal fluid leak, which were not observed in our study cohort.

CONCLUSION

Our study examined the benefits of modified endoscopic endonasal orbital decompression surgery in 13 patients with steroid-resistant DON. Our data suggest that emergent transnasal endoscopic orbital decompression surgery is a useful and effective treatment for DON, especially with regard to BCVA improvement postoperatively. Our data also suggests an inverse correlation between disease duration and surgical success.

Supplementary Material

scs-37-e1084-s001.docx ^{(17.2KB, docx)}

scs-37-e1084-s002.docx ^{(19.4KB, docx)}