Keywords

FormalPara Learning Objectives

  • Surgical rehabilitation of the patient with GO follows four steps: orbital decompression, strabismus surgery, eyelid lengthening procedures and blepharoplasty [1].

  • Each previous step may influence a subsequent step.

  • Surgery is performed during the quiescent stage of GO. There is one exception: dysthyroid optic neuropathy that is not responding to medical treatment.

  • Not a single orbital decompression technique is best. Probably the best is to tailor the orbital decompression technique to the patient and to the surgeon [2].

Introduction and History

In the first half of the nineteenth century, Caleb Hillier Parry, Robert James Graves and Carl Adolph von Basedow, independently described signs and symptoms of a new disease, which we now call Graves’ disease [3,4,5]. From the very beginning, a wide variety of treatment modalities have been considered for this disease, of which surgical intervention played an important role. An interesting paper about the pioneers of Graves’ disease was published by Alper in 1995 [6].

Management of patients with Graves’ orbitopathy (GO), nowadays, consists of medical treatment (Chap. 17), if patients have active and moderately to severe disease, or surgical treatment, if patients have burnt-out (Chap. 17) disease. The most important aim of medical treatment is inactivation of the disease process, whereas surgical treatment aims at functional and cosmetical rehabilitation. Usually, a period of 3–6 months of unchanging symptoms and signs is awaited before surgery is undertaken. The only exception is dysthyroid optic neuropathy (DON), not responding to medical treatment. Immediate orbital decompression is indicated in such cases. The surgical rehabilitation of a patient with GO often requires four steps: orbital decompression, strabismus surgery, eyelid lengthening and blepharoplasty procedures. As each previous step may influence a subsequent step, this sequence of surgeries has to be respected. To get an impression how many treatment modalities an ‘average’ GO-patient, referred to a tertiary referral institute, undergoes, we evaluated a 1-year’s cohort of 95 consecutive new patients. After a follow-up of 5 years, 24 patients had been treated with intravenously administered prednisolone pulse treatment, 50 had undergone orbital decompression, 39 one or more strabismus corrections, and 51 eyelid corrections (at the last visit to the clinic, 78% of patients had no complaints anymore, 16% had diplopia in the extremes of gaze, 5% had diplopia in all directions and one had complaints of facial pain). In conclusion, 50% of them underwent orbital decompression.

In GO, there is a discrepancy between the bony orbital volume and its contents (Figs. 18.1 and 18.2). Nature itself tries to compensate by ‘spontaneous’ orbital decompression; the walls, especially the medial wall, become concave [7]. This phenomenon, unfortunately, is rare and its effects limited. The solution, therefore, is either making the orbital bony volume bigger by creating an orbital fracture (Fig. 18.3), or decreasing the orbital fat volume; in other words, removing orbital fat.

Fig. 18.1
A coronal view C T scan of the eyes reveals the enlargement of the extraocular muscles.

Increase of orbital contents due to enlargement of the extraocular muscles

Full size image
Fig. 18.2
An axial view C T scan of the eyes reveals the enlargement of the orbital tissues.

Increase of orbital contents due to fat tissue increase

Full size image
Fig. 18.3
A coronal view C T scan of the eyes reveals the enlarged left orbit as compared to the right orbit.

Enlarged left orbit after extensive 3-wall decompression

Full size image

In 1888, Krönlein excised an orbital tumor via a lateral orbitotomy, thus after having removed an orbital bony boundary [8]. Dollinger, in 1890, was the first to perform an orbital decompression in a patient with GO [9]. Hirsch, in 1930, described a route via the canine fossa to approach the orbital floor to correct ‘eines excessiven Exophthalmus’ and, in 1950, to treat ‘Malignant Exophthalmos’ [10, 11] Naffziger, in 1931, adapted a transcranial technique and removed the orbital roof for progressive exophthalmos following thyroidectomy [12]. Sewall and Kistner, in 1936, approached the orbit via the medial wall [13]. Walsh and Ogura described a transantral approach for orbital decompression [14]. Hence, over time, approaches via all four orbital walls have been explored. The history of more than hundred years of orbital decompression has been recorded by Daniel Rootman [15].

Sole fat removal to correct exophthalmos in Graves’ disease has been described by Moore in 1920 [16], but had initially a rather questionable reputation in Western-Europe due to a lack of pertinent data. This changed when better studies were published [17, 18]. It was reported that 1 cc of fat removal corresponded with 1 mm of proptosis reduction [19]. Several studies on fat removal for GO-patients came from Asian countries [20, 21]. The outcomes of these studies cannot be simply applied on patients in Western countries as their orbits differ from Asian orbits and perhaps the presentation of GO in Asians also differs from that in Westerners [22]. Moreover, it is difficult to understand how fat removal alone could reduce the exophthalmos when muscle enlargement is the principle cause of exophthalmos [23]. Here, we enter a major problem in the literature and our knowledge of orbital decompression in GO-patients. Although this literature is abundant (> than 1000 Pubmed publications on orbital decompressive surgery in GO or TED), almost all studies are retrospective and not comparative. Outcome criteria are not well-defined and evaluation of complications differ significantly. These shortcomings have made that there has been little progress in our understanding of orbital decompression despite the huge numbers of publications. One of the very view comparative studies on orbital decompression is a EUGOGO-study [2] with 18 different approaches in 139 patients, that shows that the more orbital walls are removed, the more proptosis reduction can be expected, no matter what technique has been used. In addition, fat removal, adds to the proptosis reduction.

Next to fat removal and bony orbital decompression, there is a third way to camouflage disfiguring proptosis: by enlargement of the orbital rim [24]. This technique has never found many followers as far as we know. In conclusion, the overwhelming majority of orbital decompressions is a bony decompression, with or without fat-removal.

Bony Orbital Decompression: Indications and Approaches

The strongest indication for GO is DON not responding to other treatment modalities [25]. Orbital decompression appears to be very effective in restauration of visual functions [25]. After Leo Koornneef, in Amsterdam, in 1985, had introduced the 3-wall coronal approach (Fig. 18.4), we started to do orbital decompressions for disfiguring proptosis as well [26, 27]. Other indications were persistent pain or a corneal ulcer as a complication of severe proptosis and lagophthalmos or the very exceptional globe (sub-) luxation.

Fig. 18.4
An intraoperative photograph of the coronal approach to decompress the orbital wall.

The coronal approach

Full size image

The most common techniques and approaches at that time were the transantral/nasal (Ogura-Walsh) [14], the Lynch (medial), the lateral (with or without the orbital rim), the translid and the transconjunctival approaches. Or combinations of these, such as the ‘balanced’ decompression (medial and lateral wall, in order to decrease iatrogenic diplopia) [28]. Other techniques that quickly gained popularity were the deep lateral wall approach via a superior eyelid crease incision [29, 30] and the endoscopic approach [31]. Basically, with these techniques, one to three orbital walls are (partly) removed, the periosteum is incised and the orbital fat is allowed to herniate into the surrounding cavities (sinuses) or in later years is removed. The techniques differ in the way the walls are approached. Some complications are specifically related to the chosen technique. For instance, postoperative numbness of the cheek is caused by dissection of the bony canal of the infraorbital nerve and thus only seen after orbital floor decompression. A complication that is seen after any kind of orbital decompression is induced or worsened diplopia. Minor modifications have been proposed in order to improve the outcome or to reduce complications. It is assumed that in DON especially the medial wall has to be removed in order to get the best results, because the medial wall is closest to the medial rectus muscle, which is thought to contribute the most to compression of the optic nerve. However, a reduction alone of intra-orbital pressure, obtained by any orbital decompression, appears already to be effective [26, 32].

In line with these modifications are the ‘balanced’ decompression [28] and the decompression with leaving the periosteum intact [33], which are assumed to cause less diplopia. These assumptions sound plausible, but we do not know if they are true, as comparative studies with well-described inclusion and outcome criteria and fixed protocols for i.e., the assessment of diplopia have never been carried out. To provide such well-described criteria, we have suggested to use the Goldman perimeter and Sullivan’s score (see: Chap. 6), with which diplopia can be scored in an easy and reproducible fashion [34].

Clearly, the goals of orbital decompressive surgery are normalization of the visual functions (if they were impaired) and restauration of the premorbid face. It is often thought that the more proptosis reduction the better, but this is not true. Far more important is a symmetrical position of the globes. Symmetrical Hertel values are not always obtained after a single operation and redo-decompressions can result in a satisfactory outcome in the end.

Orbital decompression is one step in the functional and cosmetic rehabilitation of the patient with GO, but not the only one. Strabismus and eyelid correction contribute markedly to the final result. Thus, the final evaluation cannot be made before all steps have been taken [35].

For the reasons aforementioned, we cannot say which decompression is best. We, ourselves, started with the coronal approach but shifted to the ‘swinging eyelid’ technique, which was described by McCord already in 1981 [36]. Before our change, we compared both techniques in a prospective way. We found that both techniques, in our hands, were equally effective, but the swinging eyelid is much faster, less imposing to the patient than the coronal, with a shorter hospitalization time and less morbidity [37]. For the medial wall we chose an additional retrocaruncular approach.

Swinging Eyelid Technique and Retrocaruncular Approach

To obtain decompression of the globe, an increase of volume of the bony orbit usually combined with a decrease of periorbital fat is the recommended treatment. Hence, an easy access to the orbital walls, but also to the periorbital fat is mandatory. Especially, the transconjunctival approach combined with a swinging eyelid incision gives in our opinion an optimal approach to the orbital floor, the medial and the lateral wall. If the more posterior and superior regions of the medial orbit have to be approached, an additional transcaruncular approach can be done (Fig. 18.5).

Fig. 18.5
A diagram of decompressing the orbital walls using the transconjunctival approach. The lower lid is kept open using a wire speculum.

Transconjunctival approach for orbital decompression (with permission from AO Foundation)

Full size image

During any orbital decompression, the pupil has to be checked repeatedly throughout the entire procedure. We therefore do not recommend a non-transparent cornea-eye shield. At regular intervals, the retractors should be released and the pupil should be checked. If the pupil dilates and anisocoria develops, the procedure must be interrupted. This unwanted effect on the pupil is most likely caused by too much and or too long pressure on the ciliary ganglion located in the orbital apex. Continued elevated intraorbital pressure may damage the parasympathetic pathways in the ganglion resulting in deficient innervation to the sphincter pupillae muscle and to the ciliary muscle. Pupillary dilatation will occur and should be considered a warning sign to relief apical pressure (Chap. 5). Limited alertness to this sign may result in permanent pupillary dilatation and loss of accommodation. During the entire procedure, the cornea is kept wet, either by eyedrops or by gels or by ointment.

We prefer diathermal cutting with a protective sleeve exposing just about 5 mm of the needle, thus reducing the risk of collateral thermal damage to the eye. The lateral eyelid technique is relatively straight forward. The procedure begins with a cantholysis: an incision is made in the lateral canthal fold of about 8–10 mm, dissecting the dermis, the inferior lateral canthal tendon and the periosteum (Fig. 18.6). After dissection, the lower eyelid gives way, exposing the conjunctiva. Subsequently, a flexible metal spatula puts tension on the conjunctiva on the posterior side of the infra orbital rim. A Desmarres retractor is placed on the anterior part of the orbital rim, thus creating a stretched path for dissection of the conjunctiva and periosteum (Fig. 18.7). The incision can be either pre- or post-septal. Subsequently, the incision is extended on the medial side up to the lacrimal punctum (Fig. 18.8). Leaving the retractors in place, the orbit itself can be accessed. Starting on top or just posteriorly of the orbital rim, the periorbit can be elevated from the lateral wall, the orbital floor and the lower part of the medial wall. The flexible spatula can be replaced to gently lift or shift the bulbus to reveal and give access to the different orbital walls (Fig. 18.9). Care is taken not to detach the inferior oblique muscle located medially of the lacrimal punctum. Now that the different orbital walls are accessible, osteotomies can be made perpendicular to the orbital rim and parallel to the infraorbital nerve using a piezotome (Fig. 18.10), the instrument which allows us to precisely make the necessary bone cuts and helps control to leave the bone-boundary of the infraorbital canal intact. Fragments of the thin orbital floor can be removed easily. In our opinion, it is paramount to preserve a small bony bridge covering the infraorbital nerve both to protect the nerve itself (Fig. 18.11), as well as to prevent post-operative hypoglobus.

Fig. 18.6
An intraoperative photograph of the eye with the patient lying in the supine position. An incision is made in the canthal fold.

The cantholysis

Full size image
Fig. 18.7
An intraoperative photograph reveals the infraorbital rim with the Desmarres retractor.

Exposure of the infraorbital rim and incision of the conjunctiva

Full size image
Fig. 18.8
An intraoperative photograph of the exposure of the orbital contents.

Access to the orbital contents

Full size image
Fig. 18.9
An intraoperative photograph of the eye. The bulbus is shifted with the help of a spatula to access the different orbital walls.

Access to the orbital walls

Full size image
Fig. 18.10
An intraoperative photograph of the eye depicts the use of piezotome to adjust the bone to be perpendicular to the orbital rim and parallel to the infraorbital nerve.

Use of piezotome

Full size image
Fig. 18.11
An intraoperative photograph of the eye reveals the bony bridge under the eye.

Bony bridge covering the infraorbital nerve is kept intact

Full size image

The lateral wall can be approached in a similar way using the piezotome to set out the boundaries for the osteotomy. The bone of the lateral wall exists of two bony plates (Chap. 2) and is much more rigid and sturdy compared to the orbital floor. Therefore, an osteotome has to be used as well to mobilize the different parts of the wall (Fig. 18.12). Care should be taken not to damage the temporal muscle; this will lead to profuse bleeding. Tamponade with Surgicel or Spongostan will normally resolve this bleeding. The bone removal of the orbital floor can be continuous with the lateral wall, no strut has to be preserved. On the medial side, however, a key area of bone has to be preserved for the proper support and position of the globe and to prevent iatrogenic obliteration of the infundibulum.

Fig. 18.12
An intraoperative photograph of the eye reveals the removal of the lateral wall using the instrument named osteotome.

Use of osteotome to remove lateral wall from inside out

Full size image

The swinging eyelid technique will provide ample approach of the anterior and inferior part of the medial wall. If further decompression is needed, an additional retrocaruncular approach can be used. To facilitate this approach, a suture is placed in both the lower and upper eyelid that are subsequently used as retractor (Fig. 18.13). An additional retractor is placed medially to expose the caruncle. A vertical incision is made just lateral/posterior of the caruncle using diathermy. Subsequently, blunt dissection is performed posteriorly till the medial wall is exposed. The exposure will be posterior to the lacrimal sac. Using this technique, a good exposure of the more posterior and superior parts of the medial wall is accomplished. Care should be taken to locate and preserve the anterior and posterior ethmoidal arteries or obtain good cauterization. As a rule of thumb, 2–3 mm of the most medial part of the orbital floor and the 2–3 mm of the most inferior part of the medial wall should be kept in place. After the bony decompression, further decompression is obtained by periorbital fat removal. The first step is careful dissection of the periorbita in a perpendicular plane to the orbital rim. Subsequently using Metzenbaum scissors, the periorbita is further opened allowing the periorbital fat to bulge into the newly created space of the bony orbit. The amount of fat to be harvested is individually tapered (Fig. 18.14). Special care is taken not to damage the orbital muscles or the peri-muscular fat as this will enhance the risk of contusion and adhesions of these delicate orbital soft tissues potentially resulting in an increase of postoperative diplopia.

Fig. 18.13
An intraoperative photograph represents the opening of the inner corner of the eye to access the medial wall. A wire speculum is fixed on the lids to keep the eye open.

Access to the medial wall

Full size image
Fig. 18.14
A photograph of the syringe filled with the removed fat from the eyes.

Assessment of quantity of removed fat

Full size image

After hemostasis and conformation of the proper symmetrical position of the globes, the wounds can be closed. Using a resorbable 5 × 0 suture, the closure starts with aligning the lateral canthus. The suture runs through the periosteum, the inferior tarsal tendon and the superior tarsal tendon. Care is taken to position the upper eyelid just anterior of the lower eyelid (Fig. 18.15). After correct alignment of the lateral canthus, the dermis is closed with a non-resorbable 6 × 0 suture. There is a lot of controversy regarding the closure of the conjunctiva, we prefer not to close the conjunctiva to lessen the chance of creating an entropion and no adverse reaction have been noticed. Figure 18.16 shows the results of a 3-wall orbital decompression (plus eyelid corrections) as described above.

Fig. 18.15
An intraoperative photograph of the suture procedure of the upper and lower eyelids of the left eye.

Repositioning of the lateral canthus

Full size image
Fig. 18.16
2 sets of 3 photographs compare the before and after results of the wall orbital decompression and eyelid corrections surgery. Set 1 reveals swelling in both the eye bags under the eyes of the patient. Set 2 reveals normal eyes without any swelling.

(a–f) 52 year-old lady, who underwent bilateral 3-wall orbital decompression and eyelid corrections. Hertel values dropped from 22 to 18 mm and from 23 to 19 mm. Diplopia disappeared completely

Full size image

Complications

The most feared complication of orbital decompression, no doubt, is a retrobulbar hemorrhage causing blindness. We have seen retrobulbar bleedings at the time the patients awoke from their general anesthesia, started to cough and raised their intraorbital pressure. Proptosis quickly recurred and the orbit felt very stiff. Immediate evacuation of the blood reversed the threatening situation. Therefore, operated patients have to be watched closely at the moment of extubation, but also during the next 24 h postoperative. We have never seen hemorrhages after that period of time. The most frequent complication is induction or worsening of diplopia [38]. Here, a lot of misunderstanding exists, because everyone defines these conditions differently. We, therefore, suggested fixed models to assess diplopia and changes of diplopia [34], but till present, these have reached few followers. Numbness of the region innervated by the infraorbital nerve is rather common after floor decompressions, but mostly resolves with time. Obstruction of the ostiomeatal complex after medial wall and/or floor decompression can cause drainage problems (Fig. 18.17) and recurrence of exophthalmos and retrobulbar pain. Preservation of the medial orbital strut prevents this complication. Re-opening of the passage of the maxillary sinus to the nose and aeration of the sinus cures the problem. Hypoglossus is another complication of orbital floor decompression, however since the sagittal bone strut i.e., the roof of the infraorbital canal is left intact, no post-operative hypoglobus has been observed.

Fig. 18.17
A C T scan of the face reveals a dark area under the left eye.

Obstruction of the left antrum after medial wall and floor orbital decompression

Full size image

CSF-leakage and anosmia are rare complications after decompressions including the medial wall. Remarkably, the incidence of postoperative infections of the orbits are extremely low. There is no need to prescribe peri-operative antibiotics, prophylactically.

Conclusion

Since the fifties of the last century, all over the world, numerous bony orbital decompressions have been performed. What can be said is that all bony decompressions are more or less effective in the restauration of the visual functions and in reduction of proptosis and -at the same time- are safe [2]. The chance of persistent blindness is around 1% [39]. There are many complications described, but most of them are infrequent and -moreover-can be taken care off.