Comparison of the outcomes of three surgical treatments for end-stage temporomandibular joint disease


The aim of this study was to determine whether there are any differences between condylectomy, rib grafts, and prosthetic joints (Biomet TMJ stock prosthesis) with regard to outcomes for patients with end-stage temporomandibular joint (TMJ) disease. Fifty-six of a total 127 patients who presented with category 5 end-stage TMJ disease over 3 years (2010–2013) agreed to participate in this retrospective, comparative, cohort study. Patients were divided into four groups: preoperative ( n = 16), condylectomy ( n = 8), rib graft ( n = 16), and prosthetic joint ( n = 16). They were assessed for major postoperative complications (i.e., return to theatre) and maximum range of mandibular motion, and all completed a specific quality of life (QOL) questionnaire. Whilst the condylectomy group demonstrated the best mandibular range of motion ( P < 0.01), rib graft patients were more likely to experience complications (43.8%) necessitating a return to theatre. The prosthesis group recorded the best mean aggregate QOL score, but the difference compared to the rib graft and condylectomy groups was not statistically significant. The results of this study suggest that for dentate patients, prosthetic joints are highly dependable with no returns to theatre and favourable QOL outcomes. For edentulous patients, condylectomies alone also appear to work well. Future TMJ prosthetic designs should focus on improving mandibular range of motion, as the current stock prosthesis allows only a restricted range, no better than that achieved with rib graft ( P > 0.05) and far less than that achieved with condylectomy ( P < 0.01).

The surgical management of end-stage temporomandibular joint (TMJ) disease has improved progressively with the introduction of sophisticated imaging techniques and the evolution of prosthetic total joint replacement (TJR) systems that are reliable, durable, and effective in restoring mandibular function. The recent introduction of a new surgical classification ( Table 1 ) provides researchers and clinicians with a new description of end-stage joint disease termed category 5 TMJ, which refers to catastrophic changes to the joint as a result of which none of the joint components can be salvaged. It is the category 5 TMJs that the field of TJR surgery aims to address, so it is incumbent on the manufacturers of TJR systems that in addition to engineering, laboratory, animal, and cadaveric studies, clear clinical evidence is also sought to establish the real advantages of prosthetic joints over other alternatives.

Table 1
TMJ surgical classification as proposed by Dimitroulis (2013). All 56 patients involved in this study had category 5 joints.
Category 1 TMJ normal
Category 2 TMJ minor changes (all components salvageable)
Category 3 TMJ moderate changes (mostly salvageable)
Category 4 TMJ severe changes (partly salvageable)
Category 5 TMJ catastrophic changes (nothing is salvageable)

The steady growth of TMJ prosthetic TJRs has prompted many centres around the world to publish their experiences. More often than not, the experiences published are positive. Unfortunately, in our keen determination to show the world the benefits of prosthetic TJR systems, there is little hard evidence to show that prosthetic joints are in fact better than the age-old condylectomy (no reconstruction) and costochondral rib graft, which are now largely confined to developing nations. As our experiences with TJR systems gather momentum, we need to pause for a moment and take stock of which direction we need to take with future design improvements in existing TJR systems. The only way to achieve this is to begin by looking at the real (as opposed to imagined) advantages of existing TJR systems over the older techniques of condylectomies and rib grafts. The aim of this study was to determine if there are any differences between condylectomy, rib grafts, and prosthetic joints with regard to surgical outcomes for patients with category 5 TMJ end-stage joint disease.

Patients and methods

From a total of 127 patients who presented with category 5 TMJ end-stage joint disease over a 3-year period (May 2010 to May 2013), 85 met the inclusion criteria and 56 agreed to participate in this retrospective, comparative, cohort study ( Fig. 1 ). The 56 patients were recruited on a first-come basis until each of the four study groups contained a maximum of 16 patients. The exception was the condylectomy group, as this is a rarely performed operation; only eight patients could be recruited during the time-frame of the study. The four study groups were a preoperative group ( n = 16), condylectomy group ( n = 8), rib graft group ( n = 16), and prosthetic joint group ( n = 16) ( Table 2 ).

Fig. 1
Flow diagram showing the selection process for patient participation in the study.

Table 2
Summary of the 56 patients who participated in this study.
Number of patients Sex Age, years, mean (SD) Age range, years Side Dentition * Diagnosis Follow-up, mean (range) months
Preoperative 16 2 males 53.7 (12.4) 32–84 4 bilateral
5 left
7 right
13 full dentition
2 partial dentition
1 edentulous
12 idiopathic OA
4 iatrogenic OA
Condylectomy 8 1 male 59.8 (16.6) 30–76 1 bilateral
5 left
2 right
2 full dentition
2 partial dentition
4 edentulous
5 idiopathic OA
1 post-traumatic OA
1 iatrogenic OA
1 tumour OC
31.8 (12–60)
Rib graft 16 2 males 38.6 (11.5) 27–65 1 bilateral
12 left
3 right
16 full dentition 9 idiopathic OA
2 post-traumatic OA
1 iatrogenic OA
4 tumours OC
31.9 (12–62)
Prosthetic joint 16 2 males 55.3 (7.7) 39–67 1 bilateral
7 left
8 right
8 full dentition
8 partial dentition
11 idiopathic OA
5 iatrogenic OA
23.3 (12–48)

OA, osteoarthritis; OC, osteochondroma; SD, standard deviation.

* Partial dentition: five or more missing teeth excluding third molars.

The study was conducted according to the STROBE recommendations ( ), which are endorsed by a growing number of biomedical journals. Data for the study were collected in a retrospective, cross-sectional manner so that each patient was interviewed and assessed only once during the course of the study. A maximum of 16 patients with full histories, clinical assessments, and completed questionnaires were obtained for each group (except the condylectomy group, which had only eight patients). Each group was matched for age and sex so that any additional males above two in each group (but one for the smaller condylectomy group) were not included. Further selection criteria for inclusion in this study were as follows: (1) adult patient >25 years old; (2) history of intolerable clinical symptoms of TMJ pain and dysfunction not relieved by lesser measures, such as medications, splint therapy, physiotherapy, or previous TMJ surgery, arthrocentesis, or arthroscopy; (3) radiological evidence (magnetic resonance and/or cone beam computed tomography scan) of category 5 TMJ – i.e., catastrophic changes to the joint resulting from osteoarthritis or a benign tumour (e.g., osteochondroma); (4) minimum postoperative follow-up of 12 months for all three treatment groups ( Table 2 ).

All patients involved in this study were referred for surgical assessment and management from all around Australia. Furthermore, all surgical treatments were performed by the author. Patients in the treatment groups had already undergone their surgery and were being followed up when they were invited to participate in the study. Since the study was a retrospective look at surgical outcomes that did not interfere with the normal provision of surgical care, the project was found to comply with the principles of the National Statement on the Ethical Conduct of Human Research, as assessed by the human research ethics committee of the study hospital. Each patient provided signed consent to allow the de-identified data collected to be used in this study. The following patients were excluded from the study: (1) patients <25 years old; (2) surgical patients not treated by the author, i.e., cases done by registrars in training; (3) those with a known psychiatric history; (4) those with a systemic arthropathy, e.g., rheumatoid arthritis or other autoimmune or collagen disorders that affect joints; (5) those who could not understand the questionnaire, e.g., recent immigrants and the intellectually disabled; (6) patients with less than 12 months post-surgical follow-up.

The preoperative group were patients diagnosed with category 5 TMJ awaiting TJR surgery. Most of these patients (12 of 16) presented with idiopathic osteoarthritis and had no previous history of TMJ surgery, whilst four had previous histories of open TMJ surgery and were therefore diagnosed as iatrogenic osteoarthritis ( Table 2 ). The other three treatment groups were all operated on by the author under naso-endotracheal general anaesthesia. A standard pre-auricular incision and dissection to expose the TMJ was used on all patients. All three treatment groups underwent a condylectomy and had abdominal fat grafts; the variables that separated the three groups were: no reconstruction (i.e., condylectomy), rib graft reconstruction, and prosthetic joint replacement. The operative details for each group are described below.

The condylectomy group were patients who, for reasons of edentulousness, cost, and/or compromised medical conditions, had undergone a TMJ discectomy and high-level condylectomy across the neck of the condyle at least 5 mm above the level of the sigmoid notch. Abdominal fat grafts were placed in all cases between the condylar neck stump and base of the skull. Dentate individuals were placed in 3–6 weeks of maxillomandibular elastic fixation following surgery to control the occlusion. Edentulous patients were given physiotherapy and then advised to have a new set of dentures constructed at ≥6 weeks following surgery.

The rib graft group were patients, who for reasons of cost and younger age, decided to have autogenous reconstruction of the resected mandibular condyle using a costochondral rib graft harvested from the 7th, 8th, or 9th rib of the right chest wall. After discectomy and low-level condylectomy (at the level of the sigmoid notch) via a standard pre-auricular approach, the harvested costochondral rib graft was trimmed and inserted via a retrosubmandibular incision and approach ( Fig. 2 ). The rib was fixated to the lateral aspect of the ramus with an 8-hole 2-mm miniplate and four monocortical screws to the rib superiorly and four bicortical screws to the mandible inferiorly. An abdominal dermis fat graft was placed as interpositional material between the cartilaginous head of the rib graft and the glenoid fossa. The maxillomandibular fixation used to position the rib at the correct occlusion was released at the end of the procedure, and light training elastics were applied for 3 weeks postoperatively with physiotherapy thereafter.

Fig. 2
Clinical photograph showing a patient under naso-endotracheal general anaesthesia with pen marks depicting the pre-auricular (i.e., endaural with temporal extension) and retrosubmandibular incisions used for surgical access in the rib graft and prosthetic joint study groups. The condylectomy group only required the pre-auricular incision. The angle and lower border of the mandible is also marked for reference.

The prosthetic joint group were all treated with the Biomet TMJ stock prosthesis (Biomet TMJ Replacement System; Biomet Microfixation, Jacksonville, FL, USA). Following discectomy and very low condylectomy at least 5 mm below the level of the sigmoid notch, the articular eminence and lateral aspect of the root of the zygomatic arch above the glenoid fossa were flattened using a diamond-studded reciprocating file. An ultra-high molecular weight polyethylene fossa component was secured to the root of the zygomatic arch with at least four 2-mm self-tapping titanium–aluminium–vanadium alloy screws, after trial of three size components to establish the best fit. The mandible was then placed into maxillomandibular wire fixation (MMF) to establish the correct occlusion before the lateral aspect of the ascending ramus was surgically exposed via a retrosubmandibular incision and dissection ( Fig. 2 ). Minor flattening of the lateral aspect of the ramus was performed with the diamond-studded reciprocating file to establish best fit of the sandblasted ramus surface of the condylar prosthesis. A series of trial dummies were inserted to determine the best size and fit before the standard titanium condylar stem component (small, 45 mm; medium, 50 mm; large, 55 mm) was fitted to the lateral ramus and secured with at least four 2.7-mm self-tapping titanium screws, making sure the smooth condylar head was appropriately engaged in the prosthetic fossa. Abdominal fat was applied around the stem of the condylar prosthesis and the MMF released to check function and occlusion before the wounds were closed. There was no requirement for maxillomandibular fixation postoperatively and physiotherapy was commenced 5–7 days following surgery to allow pain and swelling to subside.

All patients who participated in the study were assessed in three stages. First, a thorough history was obtained through the existing medical records with further questioning to ensure the accuracy of the medical records. Transient side effects such as facial nerve weakness, paraesthesia, infections, minor occlusal discrepancies, and deafness were noted but not included in the study. Instead, complications were assessed in terms of return to the operating theatre for the correction of problems that would not resolve spontaneously with time. In this study, complications were measured by the number of subsequent returns to theatre for further corrective surgery related to the treatment groups. The second part of the study involved measurements of inter-incisal mouth opening in the vertical, lateral, and protrusive directions; these were recorded in millimetres using a metric ruler. Lateral excursions were divided into ipsilateral, referring to movement of the mandible towards the side of the affected or operated joint, and contralateral, referring to the side opposite the affected joint. Finally, all patients were asked to complete a TMJ surgery-specific quality of life questionnaire (TMJ-S-QOL), modified and adapted by the author from the University of Washington QOL questionnaire used for head and neck cancer patients ( Fig. 3 ).

Fig. 3
The TMJ surgery-specific quality of life questionnaire (TMJ-S-QOL) used in this study was modified and adapted from the University of Washington Quality of Life Questionnaire (Weymuller et al., 2001 ) for head and neck cancer patients and was first published by Dimitroulis et al. in 2010. The responses are graded 1–5, with 1 rated as excellent and 5 as poor for all questions.

For the mandibular range of motion data, the Student’s t -test was used to determine if each of the treatment groups was statistically different from the baseline ‘preoperative’ group, with a P -value <0.05 considered as significant. For the TMJ-S-QOL , the Wilcoxon–Mann–Whitney test (i.e., Wilcoxon rank sum test) was used to evaluate differences in the non-parametric ordinal data derived from a scale of 1 (best outcome) to 5 (worst outcome) between the groups, with a P -value <0.05 considered significant. A comparative visual analogue (0–10-mm) scale was also used for some of the questions and overlaid on the graphs.

The null hypothesis (H 0 ) tested in the statistical analysis (Microsoft Excel) was as follows: there is no difference between condylectomy, rib grafts, and prosthetic joints as far as surgical outcomes are concerned for patients with category 5 TMJ end-stage joint disease.


In the first part of the study, which looked at major complications, one case from the condylectomy group (1/8, 12.5%) and seven cases from the rib graft group (7/16, 43.8%) returned to the operating theatre for further corrective surgery. There were no returns to theatre for the prosthetic joint group within the time-frame of the study. The single case in the condylectomy group was a young fully dentate patient (30 years old) who decided on the option of bilateral condylectomies with subsequent mandibular advancement osteotomy 6 months later to correct a pre-existing anterior open bite, which was further compounded by her condylectomies. The seven return-to-theatre cases from the rib graft group included one case of fibrous ankylosis, two cases of prominent and loose fixation, and four cases of cartilaginous overgrowth, which were all successfully treated with no loss of the graft.

In the second part of the study, which looked at mandibular range of motion ( Figs. 4–7 ), the condylectomy group demonstrated the greatest range of mandibular motion in vertical (mean 41.2 mm, P < 0.01) ( Fig. 4 ), ipsilateral (mean 9.4 mm, P < 0.01) ( Fig. 5 ), and contralateral (mean 6.8 mm, P < 0.05) ( Fig. 6 ) excursions compared to all of the other groups. The preoperative group had the highest mean value (3.8 mm) for protrusive excursions, but this was not statistically significant ( P > 0.05) compared to the other groups ( Fig. 7 ). Both the rib graft and prosthetic joint groups had similar vertical (mean 37.8 mm and 35.6 mm, respectively; P > 0.05) ( Fig. 4 ) and ipsilateral (mean 6.2 mm and 6.0 mm, respectively; P > 0.05) ( Fig. 5 ) excursions, but the prosthesis group had significantly reduced contralateral (mean 1.6 mm) ( Fig. 6 ) excursions compared to the other groups, including the rib graft group (mean 3.9 mm), which was statistically significant ( P < 0.01).

Jan 19, 2018 | Posted by in Oral and Maxillofacial Surgery | Comments Off on Comparison of the outcomes of three surgical treatments for end-stage temporomandibular joint disease
Premium Wordpress Themes by UFO Themes