Orbital Fractures Treated in Auckland from 2016 to 2020: Review of Patient Outcomes

Abstract

Orbital reconstruction is a complex procedure demanding accurate placement of implants to restore volume and anatomic shape to the orbits. Intra-operative computed-tomography (CT) and rapid-prototype (RP) biomodels have been recently introduced as surgical aids to improve outcomes. Investigation is required to determine if there is a reduction in post-operative ophthalmic complications. Study Design: Retrospective cohort study. Objective: To evaluate the impact of intra-operative CT and RP biomodels on the incidence of post-operative diplopia, paraesthesia, cosmetic issues and ability to return to work following orbital reconstruction surgery. Methods: Adult (>18 years) patients treated at Counties Manukau District Health Board, Auckland, by the Department of Oral and Maxillofacial Surgery for isolated orbital fractures were retrospectively enrolled into this study. An audit of clinical records was undertaken to determine incidences of diplopia, altered sensation, cosmetic concerns and ability to return to work. These findings were compared against our previous audit which documented these findings in patients treated between 2010 and 2015, prior to the introduction of intra-operative CT and RP biomodels. Results: Routine use of intra-operative CT and RP biomodels was associated with a reduced incidence of post-operative diplopia. No significant difference was observed with regards to paraesthesia and cosmetic deficits. Conclusions: The relatively low radiation exposure and cost associated with intra-operative CT and RP biomodels is justified with improved outcomes in subjective diplopia. Titanium as a material for orbital reconstruction was confirmed to be associated with low complication rates.

Introduction

Technological advancements in the management of orbital fractures have been pivotal in the improvement of surgical outcomes. Surgical repair of these complex maxillofacial injuries requires accurate placement of implants to restore volume and three-dimensional (3D) anatomic shape of the orbital cavity.[1] Changes to the anatomic contour and volume of the orbits can result in progressive and prolonged diplopia, restriction in ocular motility, enophthalmos, hypoglobus, entropion and orbital pain.[2,3] Incorrect placement or adaptation of the reconstructive implant may subject patients to difficult revision surgeries with associated cost, risk and medicolegal considerations.[1] Careful diagnosis, planning, material selection and surgical precision are demanded in the repair of orbital fractures.

Pre-operative computed tomography (CT) scans are standard-of-care in the evaluation of facial fractures.[1,4,5] Post-operative CT scans are often routinely undertaken to confirm adequate retrieval of herniated orbital contents, implant positioning and restoration of orbital volume and form.[1,4,6,7] The obvious disadvantage of post-operative scanning is delayed appreciation of malpositioned implants and need for revision surgery.[6]

Intra-operative imaging has been routinely used in orthopaedic surgery for over 3 decades to assess adequacy of reduction and fixation of long bone and pelvic injuries.[4] Previous studies have demonstrated the impact of intraoperative imaging to guide decision-making in spinal surgery, neurosurgery and orthopaedics.[8,9,10]

Intra-operative imaging in maxillofacial surgery has gained significant popularity since it was first described for midface injuries in 1999.[11] Orbital reconstruction can be challenging due to limited access, swelling, distorted anatomy and bleeding. It can be difficult to directly visualize the bony fracture margins, especially posteriorly and supero-medially.[1,4] Intra-operative CT imaging assists in assessing implant positioning when direct visualization is compromised. The newer intra-operative CT scanners have improved resolution, are portable and more affordable.[4,12] Newer intra-operative CT scanners have been shown to have comparable radiation dosages with conventional CT systems. Radiation exposure is predicted to range between 600 and 800 μSv for both intra-operative and conventional CT scanners whilst cone-beam CT systems range between 40 and 80 μSv.[13] Extra operating time from intra-operative CT is reportedly around 15 minutes.[13] Research by Borad et al strongly supported the use of intra-operative CT, which resulted in a change in management for 44% of orbital fracture repairs.[14] This has been supported by other recent investigations which have found similar revision rates for repair of orbital defects following intra-operative imaging.[15,16] Revisions ranged from implant repositioning, reshaping or replacement of the implant.[1,14] The use of intra-operative scanning should reduce the need for postoperative CT imaging and return to theatre. There should be an overall reduction in cost to the service.

The importance of correct implant positioning has also given rise to rapid-prototype (RP) biomodels. This relatively recent innovation uses additive manufacturing technology to 3D print accurate reproductions of the osseous anatomy.[17,18] Biomodels enable 3-dimensional physical appreciation of the orbital anatomy and traumatic defect and allows pre-surgical contouring of the implant to facilitate close adaptation to the osseous walls.[17,18] The use of pre-contoured implants reduces theatre time and surgical trauma and improves surgical outcomes.[18,19] The affordability of RP biomodels has been demonstrated,[19] often costing less than $5 NZD for unilateral orbital wall defects.[18]

A study conducted at our institution between 2010 and 2015 investigated isolated orbital fracture repairs prior to the introduction of RP biomodels and intra-operative scanning.[3] The findings indicated generally favourable long-term results with 86% of patients reporting no diplopia at 12 months or more post-surgery. This is consistent with other investigations which have reported similar outcomes without the aid of intra-operative imaging and RP biomodels.[20,21,22,23]

Recent studies incorporating the use of intra-operative CT and biomodels have found even higher degrees of success with authors often reporting complete resolution of diplopia, enophthalmos and other post-operative complications.[24,25,26]

The aim of the present study is therefore to investigate the benefit of introducing intra-operative CT scanning and RP biomodels to our service for orbital fracture repair.

Methods

Ethics approval was obtained by the Auckland Health Research Ethics Committee (reference code: AH22851). Patients diagnosed with isolated orbital fractures and who underwent orbital reconstructive surgery from January 2016 to December 2020 at the Oral and Maxillofacial Surgery Department, Counties Manukau District Health Board, Auckland, New Zealand, were retrospectively recruited to the study. All patients verbally consented to participate in this study. The inclusion criteria were: adult patients (>18 years), unilateral orbital wall fractures involving the floor and/or medial walls, no prior history of orbital trauma or surgery, surgery performed within 1 month of the injury and a follow-up of at least 6 months. Patients with orbital fractures involving the orbital rim or associated with other midface structures, bilateral orbital trauma, endocrine orbitopathy, anophthalmia, unior bilateral loss of vision, defects caused by pathological lesions, craniofacial malformations, past radiation or chemotherapy to the orbits or cognitive impairment.

All patients had a pre-operative CT scan (1.0 mm thickness slices). All patients were assessed for ocular injury by the ophthalmology service. Surgery was offered if there was suspected muscle entrapment (early treatment), if the fractures exceeded 50% of the orbital floor, or if there was significant enophthalmos or diplopia 1 week post-injury (delayed treatment). All surgeries were performed within 1 month of the injury. Using the pre-operative imaging, ipsilateral or mirror-imaged RP biomodels of the orbits were generated using the 3D Slicer 4.10.1 (Brigham and Women’s Hospital, Harvard, MA) software and Tiertime UP300 (Tiertime, Beijing, CN) 3D printer. Synthes Matrix MIDFACE titanium preformed plates (Synthes, West Chester, PA) were precontoured on the RP biomodels to achieve close adaptation to the intact orbital fracture margins with special attention to the shelf of palatine bone infero-medially, if intact. The contoured implant was then sent for autoclave sterilization to be available for surgery.

Electronic medical records were obtained from Counties Manukau District Health Board Clinical Records and reviewed for patient demographics (age, sex and occupation), trauma aetiology, fracture location, implant material, preoperative diplopia, soft tissue/nerve injuries and enophthalmos. Additionally, information regarding number of intra-operative O-arm spins taken, need for further surgical intervention, restoration of orbital form, surgeon experience and peri-operative complications were documented. All patients were contacted by phone and invited to undertake a verbal questionnaire. Information was gathered regarding subjective diplopia, paraesthesia, cosmetic outcomes and ability to return to work. Subjective diplopia was assessed using a standardized questionnaire adapted from Holmes et al.[27]

Surgical Technique

Orbital defects were approached by a combination of transconjunctival or subcilary incisions to gain subperiosteal access to the orbital floor and/or medial wall and the fracture margins. The pre-contoured implantsa was carefully inserted to cover the defect. Once the implant was positioned, an intra-operative CT scan was completed with an O-arm (Meditronics, Minneapolis, MN) for verification. If the positioning was suboptimal, the implant would be repositioned or removed/ recontoured/replaced and a new scan taken. When the surgeon was satisfied, the implant would be fixed to the infra-orbital rim with 1–2 midface screws. A forced duction test would be undertaken to ensure free ocular motility prior to closure. Patients were followed up by the surgical team for at least 12 months post-operatively. Clinical evaluation included presence/absence of diplopia, paraesthesia and enophthalmos.

Statistical Analysis

Findings from this study were tested against the rates of post-operative complications reported in our previous investigation, using an independent Chi-squared test. P value <.05 is considered statistically significant.

Results

142 orbital reconstruction surgeries were completed by the Department of Oral and Maxillofacial Surgery, Counties Manukau District Health Board, New Zealand, between January 2016 and December 2020.

Table 1. Demographic and Aetiological Data.

Demographics	N = 142
Gender
Male	119
Female	23
Age group
<18 (excluded)	4
18–30	59
31–49	52
50–65	15
>65	12
Ethnicity
NZ European	38
NZ M¯aori	29
Pasifika	30
Asian	24
European	21
Mechanism
Interpersonal violence	71
Sporting injury	33
Falls	19
Motor vehicle accident	10
Work-related injury	8
Attempted suicide	1

demonstrates patient demographics in this cohort. The male to female ratio was approximately 5:1 (M = 119, F = 23) with the 18– 30 age bracket accounting for the majority of presentations (mean age = 37.0). Interpersonal violence was the leading mechanism for orbital fractures (50%), followed by sporting injuries, falls and motor vehicle accidents, respectively. This demographic representation closely resembles the data reported by Anand and Sealey (2017).[3] M¯aori and Pasifika populations were over-represented, receiving 20.4% and 21.1%, respectively, of all orbital fracture surgeries, despite only making up 16.5% and 8.1% of the New Zealand population (2018 Census ethnic groups dataset, Stats NZ). This finding is consistent with poorer health outcomes for Māori and Pasifika generally.[28,29]

Table 2. Fracture Location and Diplopia Outcome.

Fracture Location	N (Total)	N (Contactable)	Pre-operative Diplopia	Post-operative Diplopia
Orbital floor	96	56	44	0
Orbital floor + medial wall	33	27	25	3
Medial wall	13	6	5	1

demonstrates the prevalence of each fracture site location, with isolated orbital floor fractures being the most common presentation. Orbital floor + medial wall and isolated medial wall fractures were proportionally more common, representing 23.2% and 9.2% of all orbital fractures, respectively, compared with 5.8% and 2.9% found by Anand and Sealey.[3] Unlike the previous study, all orbital reconstructions in the current study were completed using titanium.

Of the 142 patients treated, 89 patients were contactable, with 46 being loss to follow-up and 7 patients excluded (due to pre-existing ophthalmological issues). The findings are illustrated in

Table 3. Questionnaire Outcomes.

	2010–2015 [3]	2016–2020
	N = 64	N = 89	P-Value
Pre-operative diplopia
	Not reported	75	N/A
Persistent post-operative diplopia
	9	4	P = .036
Paraesthesia
None	41	51	P = .40
Mild	11	28
Moderate	8	6
Profound	4	4
Cosmetic concerns
None	54	66	P = .13
Enophthalmos	6	11
Eyelid position changes	3	11
Proptosis	0	1
Lagopthalmos	1	0
Ability to return to work
	60	89	P = .15

.

A statistically significant reduction in the prevalence of post-operative subjective diplopia was observed in this cohort when compared with the previous cohort. Only 4 patients in the current cohort reported persistent subjective diplopia (P = .036) following orbital reconstruction. Of these 4 patients, 3 had been treated by ophthalmology in the intervening years, receiving various attempts at further treatment, without any subjective improvements. No statistical difference was observed between the 2 cohorts with respect to post-operative sensory changes (P = .40) or cosmetic outcome (P = .13). Age, gender, mechanism of injury or surgical delay was not significantly correlated with any of the measured outcomes.

Subjective reduction in post-operative diplopia was seen in the current cohort, regardless of fracture location (P < .0001). Isolated orbital floor fractures were least likely to be associated with pre-operative diplopia. Combined orbital floor + medial wall fractures were most likely to be associated with preoperative diplopia. Diplopia associated with isolated orbital floor fractures was more likely to resolve post-operatively than diplopia associated with either combined orbital floor + medial wall fractures (P = .016) or isolated medial wall defects (P = .0044). No iatrogenic diplopia occurred in the 14 patients who did not have diplopia, pre-operatively.

Table 4. Operator Experience.

N = 89	Desired Outcome	Adverse Outcome	N = 142
Consultant			P = .61
	32	1
Registrar
	53	3

compares surgical outcomes to operator experience, with no significant difference demonstrated (P = .61). All patients returned to work or had the capacity to return to work (3 had retired since). The difference was not statistically significant (P = .15) between the current cohort and that reported in our other study.[3]

Table 5. O-arm Acquisitions Required for Satisfactory Implant Placement.

O-arm Spins	N = 142
1	131
2	11

illustrates that most cases required only 1 intraoperative O-arm acquisition (mean = 1.08). Most implants were seen to be well-positioned and did not require any modification.

Two patients required a return to theatre: 1 due to persistent diplopia and another due to pain with ocular movement, underaction of the inferior rectus muscle and diplopia. In both cases, there was a complex fracture pattern involving both the orbital floor + medial wall. Both patients had the titanium implants removed. Implant removal was successful in resolving the symptoms for 1 patient. The other patient reported ongoing issues with pain and global diplopia but did consent to further treatment. No comparison in rates of return to theatre could be drawn to the prior study[3] as this was not reported.

Of the 46 patients who could not be contacted, 36 had been discharged from the service after clinical review, documenting resolution of any concerns. The last documentation (1 week post-op) for 3 patients noted subjective diplopia, and for 2 patients, ongoing swelling and pain. A total of 5 patients failed to attend for arranged post-operative follow-up and repeated attempts to contact them were unsuccessful. Three of these patients sought further treatment from ophthalmology based on available documentation.

Discussion

Previously, orbital reconstruction was undertaken without the aid of biomodels or intra-operative CT scanning. Surgical success was often judged clinically. In cases of concern, a post-operative CT would be taken. Identifying a malpositioned plate might necessitate a return to theatre.[12,30] The introduction of RP biomodels and intraoperative CT represent significant technological advances in the management of orbital defects. The utility of RP biomodels and intra-operative imaging is well-established in the maxillofacial literature.[15,17,18] RP biomodels enable the surgeon to better appreciate the 3-dimensional anatomy and adapt an implant to the defect. This reduces operating time, surgical trauma and need for revision surgery.[12,15,18] The present study supports these observations. Comparing our 2 cohorts, it appears that the introduction of these technologies is associated with a significant reduction in postoperative diplopia. These outcomes are independent of operator experience.

Important anatomic landmarks of the orbit (such as the orbital process of the palatine bone and the anterior and posterior ethmoid canals) can be challenging to identify, particularly in cases where there is extensive comminution.[31,32] Access to a RP biomodel enables intimate appreciation of the fracture geometry and its correlation with anatomic landmarks.

Given that no significant difference was found between surgeon experience and post-operative diplopia, it is likely that access to RP biomodels and immediate feedback of plate positioning by intra-operative CT aids less experienced surgeons and is protective for patients.

Although subjective reports of cosmetic deficit (particularly enophthalmos and eyelid position changes) had increased in the current investigation, no statistical difference could be established when compared with the previous cohort. The rates of long-term enophthalmos found in this study closely resemble figures reported in a prior investigation also using RP biomodels.[33] The lack of an observable difference between the cohorts however, may be a consequence of the greater proportion of more complex defects within our current population. Whilst orbital floor + medial wall fractures only made up 5.8% of the cohort between 2010 and 2015, the current investigation found this fracture pattern in 23.2% of cases. This is supported by the fact that nine of 11 patients complaining of long-term enophthalmos, demonstrated orbital floor + medial wall fracture patterns. Other studies have found the magnitude of enophthalmos to proportionally worsen as the extent of the fracture increases.[34,35] A study investigating only isolated orbital floor defects found no cases of post-operative enophthalmos to occur when using RP biomodels and intra-operative CT imaging.[36] A systematic review found the risk of postoperative enophthalmos with conventional reconstruction techniques to be around 19.2%, whereas patient-specific implants (PSI) had a lower risk of 11.2%.[37] This supports the use of PSI in larger or more complex defects when the additional cost can be met.

Intra-operative CT scanning exposes patient to additional radiation (compared with post-operative cone-beam CT) and prolongs theatre time (approx. 15 mins).[13] In this study, 11 patients required a second scan. Nguyen et al[17] reported 25/110 patients required 2 or more scans and Lim et al[36] documented 3/21 requiring multiple image acquisitions to correct plate malposition. We can summise that in the absence of intra-operative imaging, many of the patients who required additional scans, would have likely required a return to theatre and further exposure to other anaesthetic and surgical risks. Radiation from intra-operative scanning seems justified.

In this study, titanium was the only material used for orbital reconstruction. The success of titanium for orbital surgery is attributed to its biocompatibility, mechanical properties, shape stability and ability to osseointegrate.[7,38,39] The malleability and strength of titanium enables it to be used where there is loss of important supporting bony landmarks or extensive comminution. Titanium remains one of the most commonly used alloplastic materials for repair of orbital defects.[39] The results from our investigation support the use of titanium as a first choice material for orbital repair.

Three patients in our study were referred to the Department of Opthalmology for assessment of post-operative, non-resolving diplopia. All patients underwent surgery to have the titanium plates removed. Fibrous tissue ingrowth into the titanium mesh was noted in all cases. One patient had the titanium replaced with a polytetrafluoroethylene (Teflon) sheet. Another received a steroid injection into the rectus sheath, and the remaining patient had the titanium implant removed without further intervention. Diplopia persisted in all 3 patients. It is postulated that trauma to the inferior rectus muscle itself or the delicate complex network of fibrous septa that functionally unite the muscle, fibrofatty tissue and periorbita, may, in some patients with orbital fractures, account for diplopia that is unresponsive to restoration of orbital volume and form, irrespective of the material used. A further 3 patients who were lost to follow-up in the current study also received intervening treatment from ophthalmology. However, it is unclear whether these patients demonstrated any marked improvement, having received care from a different domicile.

A limitation of our study was the use of a telephone questionnaire, which is subjective by nature. This was necessary because of COVID lockdowns, although had the likely benefit of achieving a higher level of participation that might otherwise have been achieved. Another important limitation is the lack of access to the questionnaire or raw data from our prior investigation. In particular, variations in the formulated questionnaire between the 2 studies limit the accuracy of the comparison. For similar reasons, rates of pre-operative and post-operative diplopia are difficult to compare as preoperative diplopia was assessed and reported by the treating surgeons in the acute setting, as opposed to the telephone questionnaire used for the post-operative phase, which could lead to under-reporting in the latter.

We have shown the use of RP biomodels and intraoperative CT scanning in the management of isolated orbital fractures; it is associated with a significant reduction in postoperative diplopia. Titanium is confirmed as an excellent first choice material for orbital reconstruction.