Bony realignment in surgically treated orbital blowout fractures based on computed tomography

Abstract

Background

This study investigated postoperative bony realignment using follow-up computed tomography (CT) scans of patients with blowout fracture repair using an alloplastic implant.

Methods

This is a retrospective study of patients who underwent surgical treatment for orbital blowout fractures and had follow-up CT at least 4 months after the surgery. Preoperative, early, and late postoperative CT scans were obtained to assess the outcome measures including reduction completeness, reconstructed wall composition, and presence of implant-related complications. CT scans of patients with complications were compared with those of patients without complications for radiologic changes.

Results

Our study comprised 48 orbits from 48 patients. The mean age was 28.1 years, the mean time from injury to surgery was 18.3 days, and the mean time from surgery to the latest CT scan was 47.2 months (4–212 months). Most orbits (n = 41, 91.7%) showed complete reduction postoperatively. Among the 36 patients without complications, 89% showed total and 11% showed partial bony coverage. Among the 12 patients with complications, 42% showed total, 42% showed partial, and 16% showed no bony coverage. Partial or no bony coverage was noted in 11.1% (4/36) of patients without complications and in 58.3% (7/12) with complications. Reconstructed wall composition was significantly related to the presence of complications (p = 0.002). The group with total bony coverage was younger than the group with partial or no bony coverage (p = 0.048).

Conclusions

A displaced bone in the sinus is repositioned in the early postoperative stage along with the implant substantially. Partial or no bony coverage was related to the complicated cases. Young age was found to be a favoring factor for total bony coverage over the implant.

Background

Blunt orbital trauma may result in a blowout fracture of the orbital floor or medial wall by increased orbital pressure or buckling force [1]. The indications and timing of surgical repair are areas of continued interest, and of some controversy with considerable variations among specialists [2,3,4].

There have been many reports on the surgical effects of clinical factors of diplopia or enophthalmos after blowout fracture repair [5,6,7,8]. However, there are few studies solely dedicated to bony alignment after the surgery using an alloplastic implant [9]. In studies using rabbit models, Aral et al. [10] demonstrated that dense connective tissue and bone formation were observed in the group with an alloplastic implant in histologic examination. Kontio et al. [11] reported that the alloplastic implants were surrounded by elastic capsules, which gradually ossified in a sheep model. There was a study in a clinical setting demonstrating bony remodeling in orbital medial wall fractures with computed tomography (CT) evaluation [9].

In general, long-term follow-up CT scans are not usually taken to check bony realignment after blowout fracture repair unless there is a specific reason such as a patient’s request or symptoms. We aimed to evaluate postoperative bony realignment after orbital blowout fracture surgery based on these long-term follow-up CT scans using the outcome measures of completeness of reduction and composition of the reconstructed wall.

Methods

We performed a retrospective, non-comparative study on patients with surgically treated orbital blowout fractures within 2 months of trauma and having follow-up CT scans at least 4 months after the surgery in a single tertiary institution from 2002 to 2019. Patients whose preoperative CT scans were unavailable or who had undergone primary surgery at another hospital or had other fractures than orbital fractures were excluded. Follow-up CT scans are not usually taken unless there is a patient’s request or a symptom indicating complication. The study was approved by Samsung Medical Center’s institutional review board and adhered to the tenets of the Declaration of Helsinki.

These included patients with either large (≥ 50% of orbital wall area) or small fractures (< 50% of orbital wall area), and with various symptoms of clinically significant diplopia, ocular motility restriction, enophthalmos (2 mm or more), or persistent oculocardiac reflex. The types of fracture were categorized as follows: (a) hinged fractures with an edge of the fractured bone attached on one side; (b) comminuted fractures with a fractured bone freely floating in the paranasal sinus without a hinge; (c) trapdoor fractures that presented as a minimally displaced orbital fracture with incarceration of the orbital contents at the fracture site [12].

Surgeries were performed under general anesthesia by 2 surgeons (Y.-D. Kim or K. I. Woo) using the conventional reduction method. Exposure of the inferior and medial orbital walls was achieved by an inferior transconjunctival and a retrocaruncular approach, respectively. When dissecting the prolapsed tissue from the fractured site, freely floating bony fragments not attached to the sinus mucosa were removed, but those attached to the mucosa were not manipulated further and were left as such. After the orbital tissue was retrieved completely from the fractured site, a pre-fashioned alloplastic implant was placed at the fractured site.

Alloplastic orbital implants used were porous polyethylene (Medpor^®, Porex Surgical Inc., U.S.A.), nylon foil (Supramid^®, S. Jackson Inc., U.S.A.), and various bioresorbable copolymer implants. The choice of the implant largely depended on patient factors, fracture type, or surgeons’ preference. Small fractures in younger patients were more likely to be repaired with bioresorbable implants. Large fractures were likely to be repaired with either bioresorbable or permanent implants.

Preoperative CT scans, and early and late postoperative scans (later than 4 months after the surgery) were retrieved for the study. If there were multiple follow-up CT scans after the surgery, the latest CT scan was chosen for analysis. In our practice, postoperative CT scans are taken at an early postoperative period of less than 3 days from the surgery.

Outcome measures included completeness of reduction and composition of the reconstructed wall. For completeness of reduction, the follow-up CT scan was compared with the preoperative CT scan to determine if the herniated orbital tissue at the fractured site was completely repositioned into the orbit after the surgery. In the composition of the reconstructed wall, bony coverage was considered as total when the bone was realigned along with the implant and the joining of bony edges was well-formed. Partial bony coverage referred to a condition in which a portion of the reformed orbital wall consisted of no bony lining on the CT scan. No bony coverage referred to a condition in which a reformed orbital wall did not contain any bony structure.

Two ophthalmologists with extensive experience in oculoplastic surgery (J. W. Park and C. Lee) first evaluated and classified the completeness of the reduction and the composition of the reconstructed wall. Then, two more highly experienced oculoplastic surgeons (Y.-D. Kim or K. I. Woo) convened to perform a follow-up evaluation and reach a consensus on the classification. To minimize subjective evaluations, the evaluations were conducted anonymously.

Cases with complications were collected and grouped separately. To find out the factors associated with the occurrence of complications, the rate of completeness of reduction and the rate of composition of the reconstructed wall were compared between the groups of presence and absence of complications. Further analysis was performed to determine if there was any association between the presence of complications and features of the fractures or the implant type (permanent or bioresorbable). The time-related factors were also analyzed with the rate of complete reduction and the rate of composition of the realignment.

Results

This study comprised 48 orbits (all unilateral cases) of 48 patients (35 males; 73%) who presented at a mean age at surgery of 28.1 ± 15.5 years (range, 7–79 years). The mean time from injury to the initial CT scan was 4.3 days (range, 0–25 days). The types of injuries were assault (n = 13; 27.1%), sports-related trauma (n = 13; 27.1%) followed by fall on level ground (n = 9; 18.8%), other blunt trauma (n = 8; 16.7%), traffic accidents (n = 3; 6.3%), and fall from a height (n = 1; 2.1%). There was one patient (2.1%) with an unknown cause.

Isolated orbital floor fracture was the most common type of fracture in our series (n = 21; 43.8%), followed by the medial orbital wall (n = 14; 29.2%), and the combined orbital floor and medial wall fracture (n = 13; 27.1%). The most common type of fracture was comminuted (n = 26; 54.2%), followed by hinged (n = 17; 35.4%), and trapdoor type (n = 5; 10.4%). The common implant type was permanent (n = 41; 85.4%), and a bioresorbable implant was used in a small group (n = 7; 14.6%).

The mean time from injury to surgery was 18.3 days (range, 2–58 days). Twenty-seven patients (56.3%) had a CT scan within 3 days after surgery. The mean time from the surgery to the latest CT scan was 47.2 months (range, 4-212 months). The reasons for follow-up CT in our study included: (1) evaluation for an additional periocular trauma; (2) patients’ request for postoperative evaluation; (3) newly developed symptoms of a possible complication. Half of the patients (n = 24) underwent follow-up CT for evaluation of a new trauma or on their request without specific symptoms, and the other half (n = 24) underwent CT scanning to evaluate newly developed orbital symptoms or events. The most common symptoms warranting follow-up CT were diplopia and pain (n = 11), followed by swelling (n = 7), ocular motility restriction (n = 6), proptosis (n = 5), enophthalmos (n = 1), and epistaxis (n = 1).

In terms of completeness of reduction, of the 48 orbits, 91.7% (n = 44) showed complete reduction after surgery. Of the 4 patients with incomplete reduction, two patients showed a late-onset abscess pocket with fulminant sinusitis. The other two asymptomatic patients showed incomplete reduction of the herniated orbital tissue on follow-up CT.

For the composition of the reconstructed wall, total bony coverage was observed in 37 patients (77.1%). The rough and irregular bony line in the preoperative CT scan changed, with time, to a smooth bony line along the implant. A displaced bony fragment in the sinus cavity away from the implant right after the surgery seemed to migrate towards the implant and become realigned with the bony wall (Fig. 1). Among the patients who showed total bony realignment, the earliest time observed for total bony coverage was 4 months, seen in two patients. The completion time of bony realignment was not the scope of this study and warrants further study.

Fig. 1

Computed tomography (CT) scans showing total bony realignment after surgically treated orbital blowout fractures. First row: a patient in his late 20s underwent surgery with a permanent implant (Medpor^® Sheet) for an isolated orbital floor fracture. The displaced bone in the maxillary sinus at preoperative and immediate postoperative CT is realigned along the implant. (A) Preoperative, (B) immediate postoperative, and (C) 2.5-year follow-up scan. Second row: a patient in his mid 30s who underwent surgery with a permanent implant (Medpor^® Titan implant) for an isolated orbital floor fracture showed total bony realignment along the implant. (D) Preoperative, (E) immediate postoperative, and (F) 8-month follow-up scan. Third row: a teenage patient who underwent surgery with a permanent implant (Medpor^® Titan implant) for combined orbital floor and medial wall fractures showed total bony realignment along the implant. (G) Preoperative, (H) immediate postoperative, and (I) 11-month follow-up scan. The arrows represent the locations of the displaced bone from initial to follow-up imaging

The other 11 patients (22.9%) did not show total bony coverage; 9 had partial and 2 had no bony coverage (Fig. 2). Associated findings in these patients were large fracture (n = 2), and complications such as sinusitis (n = 4), peri-implant abscess (n = 1), chronic inflammation (n = 1), hematic cyst (n = 1), and implant malposition (n = 4). In 2 patients with no bony coverage, one presented with maxillary sinusitis, implant malposition, and incomplete reduction. The other patient also presented with acute orbital inflammatory symptoms of maxillary sinusitis about 10 years after surgery.

Fig. 2

Computed tomography (CT) scans showing complications associated with the implant. First row: a case with complete reduction but the partial bony coverage after surgically treated left large orbital floor fracture with a permanent implant (Medpor^® Barrier sheet) accompanied by chronic inflammation. (A) Preoperative, (B) immediate postoperative, and (C) 19-month follow-up scan. Second row: a case with complete reduction with a large hematic cyst located on the medial wall after combined orbital floor and medial wall fractures repair with a permanent implant (Medpor^® Barrier sheet). Inferior wall showed complete reduction and total bony coverage. (D) Preoperative, (E) 14-month, and (F) 26-month follow-up scan. As the hematic cyst grew larger, the bony lining was pushed into the ethmoid sinus with total bony coverage. Third row: a case with complete reduction but no bony coverage under the permanent implant (Medpor^® Barrier sheet) with late-onset fulminant sinusitis. (G) Preoperative, and (H, I) 10-year follow-up scans. The patient underwent implant removal with endoscopic sinus surgery

There were improvements in the preoperative and postoperative ocular motility restriction and enophthalmos outcome according to the pattern of bone coverage on CT scan. In patients with total bony coverage, ocular motility restriction was observed in 23 (62.2%) before surgery, and in 3 (8.1%) patients after surgery, the improvement was statistically significant (p < 0.001). Enophthalmos was present in 10 (27%) of patients before surgery, but all patients improved after surgery (n = 0) (p = 0.002). Among patients with partial or no bony coverage, the composition of patients with EOM restriction before and after surgery does not match, but the number remained the same (n = 4; 57%) (p = 1). Enophthalmos was present in 4 (57%) patients before surgery, but all patients improved after surgery (n = 0) (p = 0.125).

Complication associated with the orbital implant was observed in 12 patients, including hematic cyst in 5 (41.7%), sinusitis with orbital inflammation in 4 (33.3%), peri-implant chronic inflammation in 2 (16.7%), and peri-implant abscess in 1 patient (8.3%) (Fig. 2).

A total of 48 patients were divided into two groups with or without complications. When the demographics of the groups with and without complications were compared, there were no significant differences in age at surgery, mean time from injury to initial CT, or mean time from injury to surgery. The mean time from surgery to the latest CT scan was significantly longer in the group with complications, at 72.3 months in the complicated group and 38.8 months in the uncomplicated group (p = 0.028) (Table 1).

Table 1 Comparison of demographic features between patients with and without complications

In the group without complications, 94.4% (34/36) of patients showed complete reduction after surgery; in the group with complications, 83.3% (10/12) of patients showed complete reduction (p > 0.05). The composition of the reconstructed wall was significantly associated with complications (p = 0.002). Partial or no bony coverage was seen in 11.1% (4/36) of patients without complications, compared with 58.3% (7/12) of patients with complications. (Table 2)

Table 2 Comparison of radiologic findings between patients with and without complications

Among factors of completeness of reduction, composition of reconstructed wall, and presence of complications, none of them showed a significant correlation with fracture location (medial or floor or combined), size (small or large), type (trapdoor, hinged or comminuted), or implant type (permanent or bioresorbable) (p > 0.05).

A comparison of radiologic findings and age and time-related factors is shown in Table 3. The distribution of age at surgery was significantly different between complete and incomplete reduction and total and partial bony coverage. The group with complete reduction tended to be younger (27.2 ± 15.7 years) than the group with incomplete reduction (38.0 ± 9.3 years) (p = 0.043). The group with total bony coverage tended to be younger (25.2 ± 11.9 years) than the group with partial or no bony coverage (37.4 ± 19.8 years) (p = 0.048). Neither duration from injury to surgery nor duration from surgery to the latest CT scan demonstrated a significant relationship on completeness of reduction or composition of reconstructed wall.

Table 3 Correlation between radiologic changes and age and time-related factors

Discussion

The fate and configuration of the fractured wall repaired with an alloplastic implant have not been described substantially. In this study, we focused on the configuration of the reconstructed wall and underlined the importance of completeness of bony wall reconstruction on the occurrence of complications.

For blowout fractures, the benefits of corrective surgery have been validated by several studies [5,6,7,8]. The purposes of reconstructive surgery for orbital wall fractures are to minimize cosmetic complications associated with enophthalmos by restoring the collapsed orbital tissue to the fractured site and to minimize functional complications such as ocular motility restriction and diplopia by releasing the incarcerated orbital tissue to the fractured site [13, 14].

There have been few reports of radiological feature changes after orbital blowout fracture surgery [5, 9]. To the best of the authors’ knowledge, there is one report of postoperative bony wall change after conventional open reduction of isolated medial wall fractures, in which an absorbable mesh implant was used [9]. In their report, most bony depressions were improved significantly after surgery, the rough and irregular bony line changed to a smooth bony line as time passed. Fibrosis appeared to fill the defective gap between the reduced medial orbital bone and periosteum [9, 15,16,17]. The inserted absorbable mesh simply functioned as a template for bony realignment.

Radiologic changes in untreated blowout fractures of the orbit have been recently described [18]. On comparison of CT scans taken immediately after trauma and after an average of 4.6 months with conservative treatment, a large proportion of patients showed improvement in radiologic findings, with smoothening of bony contour, joining of bony edges, reduction in orbital content herniation, features of neobone formation, and reduction in both orbital and fracture volumes [18]. The connective tissue septa surrounding the eyeball and the extraocular muscles to the orbital walls may predispose conservative healing process [17].

The healing process after fracture reduction might be different from that via conservative treatment without surgery. In conservatively treated cases, the torn portion of the periosteum, separated bony edges, and sinus mucosa heal with the tissue healing processes at the fractured sites. In surgically managed cases, the prolapsed orbital tissue and the periosteum are restored using an implant, and bony repositioning or regrowth, and sinus mucosal healing are accomplished at the surface of the implant. As shown in Fig. 1, a displaced bony fragment as a hinged type in the sinus at the immediate postoperative period is appropriately repositioned, as seen on the late postoperative CT scan, with the sinus mucosal healing process.

The bony realignment after the orbital blowout fracture surgery appears to progress with time along the inserted implant [9]. This is somewhat consistent with results from an experimental study in a rabbit model using microscopic analysis. After reconstruction with a porous alloplastic implant in an orbital floor fracture, both vascular and soft-tissue ingrowth into the implant were seen at 1 week, bone ingrowth was seen at 3 weeks, and mucosal defects were closed by 4 weeks [19]. Another experimental study in a goat model with a bioresorbable alloplastic implant reported that encapsulation and new bone formation were progressively being formed, and at 79 weeks, the new bone had fully covered the plate on the antral and orbital sides [20].

In this study, 12 patients had complications, with 5 showing total bony coverage and 7 showing partial or no bony coverage. Of the 5 patients with total bony coverage, 4 had a hematic cyst and 1 had chronic inflammation around the implant. A hematic cyst is formed by an accumulation of hematoma within an established fibrous capsule around the implant; therefore, the hematoma formation may occur after total bony coverage in these patients [12]. Of the 7 patients showing partial or no bony coverage, 4 patients had sinusitis, and one patient each had a peri-implant abscess, chronic inflammation, and hematic cyst. Moreover, the rate of partial or no bony coverage was 58% in complicated cases and 11% in non-complicated cases. In complicated cases with orbital inflammation associated with sinusitis, there was no case of total bony coverage. Therefore, lack of total bony coverage seemed to be associated with such complications.

Bony realignment along with an implant can be affected by many factors, but there have been no studies clearly demonstrating these effects. Our study showed that the location, size, and type of blowout fractures, and used implant type (permanent or bioresorbable) did not have any significant association with radiologic changes. The group with complete reduction and the group with total bony coverage showed significantly younger age distribution. This suggests that the younger the age, the more active the bony alignment may occur. In the study on the conservative treatment of orbital blowout fractures [18], the reduction in orbital volume was statistically significant from the initial to the follow-up scans, and this reduction in orbital volume was greater in younger patients, indicating a greater potential for spontaneous radiologic healing in the young [18].

Duration from injury to surgery did not show a significant correlation with radiologic changes. Although 14 days after trauma is commonly cited as a timeline target for orbital blowout fracture repair, in a study, ocular motility and diplopia recovered similarly when orbital fracture repair was done within 14 days as opposed to 15 to 29 days following trauma [21]. Additionally, it has been reported that surgical repair of orbital blowout fractures occurring more than 6 weeks or more from injury can achieve marked improvement in both the functional and cosmetic aspects [22]. Since the time of surgery was not largely delayed in our cases (less than 2 months, mean = 18.3 days), significant differences could not be observed. The duration from surgery to the latest CT scan was significantly longer in patients with complications. Delayed onset of implant-related infection or inflammation ranging from 1.5 to 20 years can happen after implantation [23,24,25].

There were some limitations in this study. Firstly, the retrospective nature of the study subjected it to an inherent selection bias. However, it is worthwhile since we collected and analyzed follow-up CT scans performed for a variety of causes, even though they are not usually taken. Secondly, the bony wall configuration and extent of bony coverage were evaluated with a CT scan with the bone setting, and not with an endoscopic examination or histopathologic study.

Conclusions

In conclusion, our study is worthwhile to analyze the pattern of bony realignment after orbital blowout fracture surgery based on computed tomography. Most patients showed good bony realignment after surgery using an alloplastic implant. Complicated cases were associated with partial or no bony coverage along the implant in the CT scan. Good bony realignment over the implant was associated with younger age. Enhancing the early and total bony healing over the surgical site seems to be important to prevent late-onset complications.