Endovascular preoperative embolization for temporomandibular joint replacement surgery

Abstract

Background and purpose

This retrospective study evaluates the safety and effectiveness of preoperative endovascular embolization in patients who underwent temporomandibular joint (TMJ) replacement surgery.

Material and methods

We included all patients treated with preoperative embolization of the internal maxillary artery (IMAX) between June 2016 and January 2019. All patients were treated by the same surgeon using standard surgical approaches and procedures. Periprocedural adverse events, blood loss during surgery and clinical follow-up are reported.

Results

Fourteen patients (12 females, median age 32.5) were treated with 21 embolizations of the IMAX (bilateral embolizations in seven patients) prior to TMJ replacement surgery with prosthetic joints (TMJ Concepts prostheses). Seven patients presented with TMJ ankylosis/degenerative joint disease/post-trauma deformity, four patients with Idiopathic Condylar Resorption and resultant mandibular displacement/hypoplasia, two patients with rheumatoid arthritis-associated condylar degeneration and resultant loss of mandibular position, and 1 patient being re-reconstructed following management of a prosthetic joint infection. Seven patients underwent bilateral prosthetic joint replacement. Four patients underwent additional facial skeletal surgery as part of their treatment. The mean blood volume loss during TMJ surgery was approximately 370 cc (range 100–800 cc). Joint space-specific blood loss was not recorded, but, as per the surgical team, was significantly decreased when compared to non-embolized patients. There were no intra-procedural complications. The median clinical follow-up was 3.5 months (range 1–24 months). The modified Rankin scale (mRS) was 0 before the procedure and at last clinical follow-up in all patients. After TMJ surgery, three patients reported paresthesia of the trigeminal nerve likely related to the residual condyle resection and two patients had mild facial nerve weakness (Temporal and/or Marginal Mandibular branch) related to the surgical exposures.

Conclusion

Endovascular preoperative embolization of the IMAX is feasible, safe and likely effective in reducing blood volume loss in complex TMJ replacement surgery.

Introduction

Temporomandibular joint (TMJ) replacement surgery is indicated for a variety of conditions that causes impairment in mastication, digestion, se.¹ This type of surgery can be associated with substantial blood loss with volumes as high as 3 L due to poor visualization and limited proximal access after iatrogenic arterial injury that might occur during surgery.^2–5 Preoperative endovascular embolization of the internal maxillary (IMAX) artery has been used as a means to reduce morbidity and mortality from large blood volume loss (BVL) that might occur during TMJ surgery.^4,6 In this paper we present our experience and technique with pre-operative embolization of the IMAX artery prior to TMJ surgery using prosthesis. We also report the BVL associated with TMJ surgery and its relationship with IMAX embolization.

Material and methods

Patient population

All patients who underwent embolization of the IMAX artery between June 2016 and January 2019 are included in the analysis.

Endovascular technique

All embolization procedures were performed with the patients under conscious sedation on a biplane angiography system, using a 5 F femoral artery access sheath. In all of our cases, an initial angiogram was performed to evaluate for ECA/ICA anastomoses from different branches of the IMAX such as the anterior tympanic artery, middle meningeal artery (MMA), accessory meningeal artery, vidian artery, pterygoid artery, artery of the foramen rotundum, infraorbital artery, anterior deep temporal artery and sphenopalatine artery. We evaluated the orbital supply since the MMA can anastomose to the OA via (1) sphenoid branches via the superior orbital fissure which anastomoses with the recurrent branch of the OA, (2) frontal branch of the MMA may anastomose with the anterior ethmoidal artery which is a branch of the OA and (3) through a variant called the meningolacrimal artery that has a potential collateral route to the OA.⁷ The presence of dominant orbital blood supply from the MMA is a contraindication for MMA occlusion.

A microcatheter was used to catheterize the distal IMAX. Subsequently, platinum detachable coils were deployed in the lumen of the IMAX until complete occlusion was achieved. We aimed to occlude the part of the IMAX deep to the mandible. The ostium of the MMA was included in the embolization in the origin of the MMA was located behind the condyle. Following embolization, the status of collateral flow of the distal segment of the IMAX (and MMA when applicable) was evaluated by common carotid artery angiogram. Bilateral embolizations were performed in the same setting (see Figure 1). The patients were admitted for observation overnight and had surgery the day after the embolization.

Figure 1.

A 17-year-old female with the medical history of mandibular hypoplasia presenting for preoperative embolization of bilateral internal maxillary arteries. (a) Pre-embolization angiogram (lateral view) of the left external carotid artery. (b) The internal maxillary artery was completely embolized with coils (asterisk) with preservation of the left middle meningeal artery (MMA) (arrow) and the occluded accessory meningeal artery (AMA) got reconstituted (double arrows). Following temporomandibular joint surgery, patient had paresthesia on the left trigeminal nerve distribution at last follow-up.

Surgical technique

All patients had preoperative CT scans. Surgical access for the total joint reconstruction was through standard retro-mandibular and pre-auricular approaches. Resection of the residual condyle was done from the inferior approach. Minimal blood loss was noted superiorly during removal of the condylar structure. Degenerated and/or residual joint-space soft-tissues and/or ankylotic bone was removed from the pre-auricular exposure. Blood loss within and/or medial to the joint space was also minimal or significantly decreased when compared to non-embolized patients from the surgeon’s perspective. In one case, a contralateral mandibular ramus sagittal osteotomy was performed to enable mandibular movement/rotation without condylar torque. In three cases, maxillary LeFort I one-piece or multiple-piece osteotomies were also performed to enable creation of stable and functional occlusions and/or achieve the desired mandibular advancement movement. In two of these cases, advancement genioplasties were also performed. The surgical blood loss in these three cases was significantly increased when compared to the TMJ prosthetic replacement-only cases. The measured perioperative BVL was obtained from the anesthesia records. Complex TMJ surgery was defined when patient had a procedure that included Le Fort osteotomy, genioplasty, lower jaw osteotomy with significant amount of soft tissue stripping. The surgical team discussed their impressions of joint replacement associated blood loss after each case. The patients were followed with the oral and maxillofacial surgery team according to standard follow-up. There was no post-discharge follow-up with the interventional neuroradiology team.

Follow-up

The modified Rankin scale (mRS) was used to evaluate clinical status of the patient before embolization and at last follow-up with the oral and maxillofacial surgery team.

Statistical analysis

Categorical variables were described as frequency (%) and continuous variables were described as mean (standard deviation) and/or median (range). BVL was normally distributed by Shapiro Wilks test for normality after stratification by variables of interest. Therefore, two-sided two-sample t-tests at the 0.05-significance level were used to make bivariate comparisons between BVL and predictors of interest. Three subjects were excluded from analyses comparing BVL and MMA occlusion due to having bilateral embolization of the IMAX with unilateral occlusion of the MMA (Subjects #6, #13 and #14). Subject #9 was observed to have a high degree of blood loss due to profuse venous bleeding from injury to the emissary veins. Therefore, the above analyses were repeated after exclusion of this subject. All hypothesis-testing was performed at the 0.05-alpha level, and all analyses were performed in R Version 3.5.3.

Results

Study population

A total of 14 patients (12 females and 2 males) were included in the analysis. Seven patients had bilateral embolization for bilateral TMJ surgery. Seven patients presented with TMJ ankylosis/degenerative joint disease/post-trauma deformity, four patients with Idiopathic Condylar Resorption and resultant mandibular displacement/hypoplasia, two patients with rheumatoid arthritis-associated condylar degeneration and resultant loss of mandibular position, and one patient being re-reconstructed following management of a prosthetic joint infection. The median age was 32.5 years (range 15–66 years). We were able to treat all patients as intended (see Table 1).

Table 1.

Participant characteristics.

	Overall (n = 14)
Age
Mean (SD)	34.9 (15.3)
Median [min, max]	32.5 [15.0, 66.0]
Sex
Female	12 (85.7%)
Male	2 (14.3%)
Type of surgery
Complex	5 (35.7%)
Non-complex	9 (64.3%)
Indication
Ankylosis of L TMJ	3 (21.4%)
Ankylosis of R TMJ	3 (21.4%)
Idiopathic bilateral condylar resorption	1 (7.1%)
Mandibular hypoplasia	3 (21.4%)
Progressing apertognathia	1 (7.1%)
Rheumatoid arthritis	2 (14.3%)
TMJ injury due to MVA	1 (7.1%)
Embolization site
Bilateral IMAX	7 (50.0%)
Unilateral IMAX	7 (50.0%)
MMA occlusion
Occluded	6 (42.9%)
Preserved	5 (35.7%)
Clinical follow-up (months)
Mean (SD)	7.09 (7.12)
Median [min, max]	3.50 [1.00, 24.0]
BVL during TMJ surgery (cubic centimeters)
Mean (SD)	370 (222)
Median [min, max]	425 [100, 800]

BVL: blood volume loss; IMAX: internal maxillary artery; TMJ: temporomandibular joint; MMA: middle meningeal artery; MVA: motor vehicle accident.

Endovascular results

We performed 21 embolizations of the IMAX artery in 14 patients. Embolization of bilateral IMAX arteries were performed in seven patients and embolization of a unilateral IMAX was performed in seven patients. Complete occlusion of the internal maxillary artery (IMAX) was achieved using coils only. We did not use particles or liquid embolic agents. Collateral flow from the contralateral IMAX and ipsilateral facial arteries was visualized in all cases of unilateral IMAX artery embolization and from bilateral facial arteries in cases of bilateral IMAX artery embolization. The embolization included the MMA ostium in 9/21 embolization procedures. Following occlusion of the ostium of the MMA, collateral circulation from the distal IMAX was observed in all cases. There were no cerebrovascular or peripheral vascular complications. There was no neurological morbidity or mortality.

Operative blood loss and clinical follow-up

All patients underwent TMJ surgery the day after the embolization. The mean BVL was 370 cc per patient (range 100–800 cc) following embolization. There was no bleeding from IMAX or MMA injury during the procedure. We analyzed our results in an attempt to find any association between blood loss and endovascular technique followed by TMJ surgery.

Patient #6 had the highest operative blood loss in the cohort (800 cc); this patient had bilateral TMJ replacement, Le Fort osteotomy and genioplasty with significant amount of soft tissue stripping including the entire genial region increasing the perioperative BVL. Patient #9 who had a unilateral TMJ replacement surgery, had the second highest operative blood loss (525 cc), due to profuse venous bleeding from injury to the emissary veins that connect the cavernous sinus to the pterygoid plexus. Patient #5 had the third highest operative blood loss with unilateral TMJ replacement surgery (500 cc). This patient had unilateral TMJ replacement and sagittal split osteotomy of the lower jaw which requires a significant amount of soft tissue stripping increasing the amount of blood loss.

Complex TMJ surgery was performed in five patients (patient #5, #6, #11, #12 and #13) and was associated with a mean of 560 cc of blood loss per TMJ replacement surgery versus 231 cc BVL per TMJ surgery without complex surgery (p = 0.005) excluding subject #9 (n = 8). This association was also significant in the whole population (p = 0.009). Bilateral IMAX embolization (n = 7) was associated with a mean of 479 cc of blood loss per TMJ replacement surgery versus 217 cc BVL in patients with unilateral IMAX embolization (p = 0.027) excluding subject #9 (n = 6). This association was not significant in the complete study sample (p = 0.063). Occlusion of the MMA (n = 5) was associated with a mean of 250 cc of blood loss per TMJ replacement surgery versus 380 cc BVL in patients where the MMA was preserved during embolization of the IMAX (p = 0.264) excluding subject #9 (n = 5). This result was consistent with the full-sample analysis. There was no MMA injury during TMJ surgery (Table 2).

Table 2.

Associations between BVL and variables of interest.

	Complete study sample			Excluding subject 9
	N	Mean (SD)	p-value	N	Mean (SD)	p-value
MMA occlusion vs. preservation	11			10
Occluded	6	296 (202)	0.451	5	250 (187)	0.264
Preserved	5	380 (152)		5	380 (152)
Embolization site	14			13
Unilateral	7	261 (185)	0.063	6	217 (157)	0.027
Bilateral	7	479 (212)		7	479 (212)
Type of surgery	14			13
Non-complex	9	264 (182)	0.009	8	231 (165)	0.005
Complex	5	560 (152)		5	560 (152)

MMA: middle meningeal artery.

Follow-up

The median clinical follow-up was 3.5 months (range 1–24 months). The preoperative and postoperative modified Rankin scale was 0 in all cases. There was no neurological mortality or morbidity related to the embolization. Mild facial nerve palsy occurred in two patients and trigeminal nerve paresthesia in three patients following the TMJ replacement surgery.

Discussion

TMJ replacement surgery is indicated in end-stage joint disease that leads to severe functional and structural impairment. This disease process result from developmental disorders (mandibular hypoplasia), neoplasia, trauma, arthritic disease, ankyloses, or failed prior joint surgery. Patients that undergo TMJ surgery are susceptible to potentially life threatening vascular injury, most commonly the IMAX artery.⁸ As shown in similar paradigms in other parts of the body, pre-operative embolizations can reduce the risk of blood loss from inadvertent vascular injury and from surgical approach/dissection.⁹ In this paper, we report the results in 14 patients that underwent pre-operative embolization of 21 internal maxillary arteries with coils only, with no periprocedural complications.

Hemostasis following injury to the IMAX artery is particularly challenging due to limited access and visibility as the IMAX, crosses medial to the condylar neck and sigmoid notch and is located 20 mm below the head of the condyle.^3,10,11 Bouloux and Perciaccante¹¹ reported failure to control bleeding in a patient with ankylosis of the TMJ despite ligation of the ECA, most likely due to collateral retrograde blood supply to the IMAX. Selective catheterization of the IMAX and occlusion of the IMAX exposed to injury during surgery has become an important method to control BVL.^2,12 We believe that occlusion of the IMAX segments and not just the ostium prevents major bleeding from retrograde/collateral blood flow.

Baseline cerebral angiograms of the internal and external carotid arteries were performed before endovascular embolization to rule out dangerous anastomoses between the external and internal carotid arteries and ophthalmic artery supply from the MMA. The major risks of the IMAX embolization procedure include inadvertent embolization of the cerebral and orbital vessels via well-known collateral pathways, and embolization of vasa nervosa supplying cranial nerves (such as the petrosal branch of the MMA). There are two ways to avoid these complications: (1) identify and avoid such collaterals before the embolization, and (2) avoid the use of particles or liquid embolic material that can easily penetrate vasa nervosa or dangerous ECA to ICA anastomosis that are often angiographically occult.

Susarla et al.⁴ reports three patients who underwent preoperative embolization using platinum coils of the IMAX artery for TMJ ankylosis with no periprocedural complications. Hossameldin et al.⁶ reports 14 patients who underwent preoperative embolization of the IMAX artery with coils for TMJ ankylosis with no periprocedural complications as well. Alderazi et al.¹ report two cases that underwent bilateral IMAX embolization using Vortex coils (Boston Scientific) distally followed by n-BCA (50% concentration) from the coil mass up to the proximal IMAX distal to the origin of the superficial temporal artery. In all of our cases, coils were used to achieve complete embolization of the target vessel. According to Alderazi et al.,¹ pushable coils were used to define the distal end of the IMAX and to protect from inadvertent embolization of ethmoidal arteries and collaterals that could potentially occur with the use of n-BCA. We believe that avoiding liquid embolics allows for a safe embolization without the risk of inadvertent embolization into small branches of the proximal IMAX such as the MMA that has numerous anastomosis with branches of the internal carotid artery, often not visualized in a conventional diagnostic angiogram. There is also risk for cranial neuropathy from inadvertent embolization of arterioles supplying nerves, such as the facial nerve from the petrosal branch of the MMA. Other materials such as Onyx carry the risk of nerve injury from DMSO infusion into very small branches.

The mean volume loss reported in our case series from TMJ surgery was 246 cc per TMJ replacement surgery, which is similar to other reports.^1,4 Susarla et al.⁴ reports one case of bilateral TMJ disorder who had an estimated blood loss of 3750 cc after preoperative embolization of only one IMAX artery since the other IMAX artery was previously ligated and unable to be embolized due to retrograde flow. Hossameldin et al.⁶ reports a mean blood loss of 136 cc during surgery of 14 patients with preoperative embolization of the IMAX artery for TMJ ankylosis. In our series, the mean BVL was 370 cc with preoperative embolization of the IMAX and the mean BVL was less when comparing complex TMJ surgery and unilateral IMAX embolization versus non-complex TMJ surgery and bilateral IMAX embolization.

In addition to IMAX, the MMA may be injured during TMJ replacement surgery involving extensive dissection.¹³ In our cohort, the MMA ostium was occluded with coils when it originated behind the condyle, to prevent blood loss in case of injury during surgery. Mean blood loss was observed to be lower among patients who underwent occlusion of the ostium. Although this result was not statistically significant, this study is likely underpowered to detect a difference. There were no neurological deficits after occluding the MMA in our series.

Mild facial nerve palsy occurred in two patients and trigeminal nerve paresthesia in three patients following TMJ surgery. Injury to the branches of the trigeminal and facial nerves is a known complication of TMJ replacement surgery. The reported incidence of facial nerve injury following TMJ replacement surgery ranges from 1 to 25% and is typically transient in nature, resolving within three to six months. Causes of neuropraxia include edema, excessive flap retraction forces, electrocautery, inadvertent suture ligation, or clamping of tissues.

One of the limitations of the study includes the relatively small sample and the lack of a control group. All procedures were performed by the same physician at a single center and the results should be generalized with caution.

Pre-operative embolization of the IMAX is feasible, safe and likely effective in reducing BVL in patients with TMJ replacement surgery.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.