- Email: [email protected]
Operative Management of Temporomandibular Joint Ankylosis: A Systematic Review and Meta-Analysis
ANESTHESIA/FACIAL PAIN J Oral Maxillofac Surg 70:531-536, 2012
Operative Management of Temporomandibular Joint Ankylosis: A Systematic Review and Meta-Analysis Alexander Katsnelson, DMD, MS,* Michael R. Markiewicz, DDS, MPH, MD,† David A. Keith, BDS, FDSRCS, DMD,‡ and Thomas B. Dodson, DMD, MPH§ Purpose: Two common treatments of temporomandibular joint ankylosis are gap arthroplasty and
ankylosis resection and reconstruction of the ramus– condyle unit with a costochondral graft. The purpose of the present study was to answer the following clinical question: “Among patients with temporomandibular joint ankylosis, do those patients who undergo gap arthroplasty, compared with those who undergo ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft have better postoperative mandibular range of motion?” Methods: A systematic search of the published data was performed to identify eligible studies. The
primary predictor variable was treatment type (ie, gap arthroplasty or ankylosis resection and ramus– condyle unit reconstruction). The main outcome was the change in maximal incisal opening postoperatively. A random effects model was used to compute the pooled weighted mean difference between the pre- and postoperative maximal incisal opening in both treatment groups. Results: Four studies met the inclusion criteria. Those undergoing gap arthroplasty had a significantly greater maximal incisal opening than the group undergoing ankylosis resection and ramus– condyle unit reconstruction. The weighted mean difference between the 2 groups was 2.4 mm (95% confidence interval 0.9 to 4.0; P ⫽ .002). Conclusions: Subjects with temporomandibular joint ankylosis who underwent gap arthroplasty had significantly better postoperative maximal incisal opening than those undergoing ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft. © 2012 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 70:531-536, 2012 The operative management of temporomandibular joint (TMJ) ankylosis is challenging. Common treatments of TMJ ankylosis include gap arthroplasty (GA) and resection of the ankylosis and reconstruction of the
ramus– condyle unit with a costochondral graft or prosthetic joint.1–3 The primary treatment goals are restoration of function and prevention of reankylosis and, in the growing patient, symmetric mandibular growth.
*Formerly, Resident in Training, Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, Boston, MA; currently in private practice, Chicago, IL. †Resident in Training, Department of Oral and Maxillofacial Surgery, Oregon Health and Science University, Portland, OR. ‡Professor of Oral and Maxillofacial Surgery, Harvard School of Dental Medicine. Visiting Oral and Maxillofacial Surgeon, Massachusetts general hospital, Boston, MA. §Visiting (Attending) Oral and Maxillofacial Surgeon, Director, Center for Applied Clinical Investigation, Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, and Professor, Department of Oral and Maxillofacial Surgery, Harvard School of Dental Medicine, Boston, MA.
This project was supported in part by the Massachusetts General Hospital Department of Oral and Maxillofacial Surgery Education and Research Fund (to A. Katsnelson), Center for Applied Clinical Investigation (to T. B. Dodson), and Massachusetts General Physicians Organization (to T. B. Dodson). Address correspondence and reprint requests to Dr Dodson: Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, 55 Fruit Street, Warren Building, Suite 1201, Boston, MA 02114; e-mail: [email protected] © 2012 American Association of Oral and Maxillofacial Surgeons
0278-2391/12/7003-0$36.00/0 doi:10.1016/j.joms.2011.10.003
531
532 Although numerous reports have addressed the management of TMJ ankylosis, a search of the published data demonstrated no evidence of any randomized trials or systematic reviews addressing the surgery outcomes of different treatments. The purpose of the present report was to complete a systematic review and meta-analysis to answer the following clinical question: “Among patients with TMJ ankylosis, do those patients who undergo GA compared with those who undergo ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft have better mandibular range of motion (ROM)?” For the remainder of our report, ankylosis resection and ramus– condyle unit reconstruction using a costochrondral graft has been abbreviated “AR.” We hypothesized that the postoperative ROM would be the same between the 2 treatment groups. The specific aims of the present project were to perform a systematic review of the published data and identify studies for analysis and to execute a meta-analysis to determine whether GA or AR produces better outcomes in terms of mandibular ROM.
Materials and Methods STUDY DESIGN
To address the research question, we designed a systematic literature review and meta-analysis. SAMPLE
Our review of the published data was designed and conducted in concordance with the QUOROM (Quality of Reporting of Meta-analyses) guidelines.4 Two of us (A.K. and M.R.M.) searched the published studies from January 1966 through May 2010, using the National Library of Medicine (PubMed, available from http://www.pubmed.gov), the National Institutes of Health clinical trials registry (available from http:// www.clinicaltrials.gov), and the Cochrane Central Register of Controlled Trials (available from http://www. mrw.interscience.wiley.com/cochrane) using specific medical subject headings and key words, including “temporomandibular joint.” “ankylosis/gap arthroplasty/costochondral graft.” In addition, the on-line databases of the “British Journal of Oral and Maxillofacial Surgery,” “Journal of the American Dental Association,” “Journal of Dental Research,” “Journal of Oral and Maxillofacial Surgery,” “International Journal of Oral and Maxillofacial Surgery,” and “Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology,” as well as the on-line abstract indexes of the conference proceedings of the American Association of Oral and Maxillofacial Surgeons and the International Association for Dental Research annual meetings were searched for
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS
studies. Electronic copies of the reports identified by the search were obtained. Bibliographies from the relevant reports were also searched for studies that might have included data relevant to the study topic. STUDY CRITERIA
Studies were included only if all the following eligibility criteria were met: the 2 treatment types (GA and AR) were compared, descriptive statistics and the primary outcome variable (ie, maximal incisal opening [MIO]) were reported or the mean and standard deviation (SD) could be computed from the report, and the pre- and postoperative data were available. DATA EXTRACTION
Using a standardized data extraction form, one of us (A.K.) extracted and tabulated all data. STUDY VARIABLES
The primary predictor variable was the treatment type (ie, GA or AR). The primary outcome variable was the MIO. The primary comparison of interest was to estimate the differences between the preoperative and postoperative MIOs between the 2 study groups (GA and AR). The reported mean difference between the preoperative measurements, SD of the difference, and sample size were used to compute the weighted mean difference (WMD) for the outcome variable (MIO) for the GA and AR groups. All MIO measurements were converted to millimeters. The other study variables included the total sample size, number of subjects in both treatment groups, and mean age. The anatomic location of ankylosis was recorded as unilateral or bilateral. STATISTICAL ANALYSIS
We used a conservative approach to estimate the SD of the change from baseline: SD (change SD from SD baseline) ⫽
冑
(Preoperative SD)2 ⫹ (Postoperative SD)2 2
This method has been previously published by Markiewicz et al.5 This formula results in a larger, and most likely overestimated, SD than the formula recommended in the Cochrane Handbook.6 The primary summary measure of association was the overall DerSimonian and Laird random effects pooled mean difference in postoperative MIO between the GA and AR groups.7 A similar method of mandibular ROM measurement was used across all studies (MIO); therefore, a WMD meta-analysis was performed. Unlike the standardized mean difference,
533
KATSNELSON ET AL
the WMD allows for direct interpretation by readers in common units used across all studies in the metaanalysis. Meta-regression analyses were not performed owing to the small number of trials included in the pooled analysis. To test for the presence of publication bias, the relative symmetry of the individual study estimates was assessed around the overall estimates using Begg’s funnel plots, in which the mean differences were plotted against their corresponding standard error. Egger’s test for publication bias was also performed.8 All analyses were conducted using Stata, version 9.2 (StataCorp, College Station, TX). P ⬍ .05 was considered statistically significant.
Results After eliminating duplicate reports or those reporting previously published results, 1,518 eligible reports were identified for possible study inclusion. The vast majority (n ⫽ 1,464) were excluded for 1 or more of the following reasons: case report, retrospective study, not of human subjects, review, editorial or commentary, or did not test the association of interest (Fig 1). After retrieving the full texts of the pertinent studies (n ⫽ 54), 50 were excluded: 3 were single case reports, 16 used a different treatment technique (ie, total prosthetic joint), 20 had no comparison group, 4 did not assess the association of interest, 3 did not have the MIO or other important information,
FIGURE 1. Flow diagram. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
1 was a discussion for a report without any original data, and 3 were review studies. The final sample included 4 studies used for data extraction. The study characteristics of the included reports9 –12 are presented in Table 1. After treatment, the subjects undergoing GA had a significantly larger MIO than those who had undergone AR (WMD 2.4 mm, 95% confidence interval 0.9 to 4.0, P ⫽ .002; Fig 2). Egger’s test did not reveal any significant publication bias (P ⫽ .7; Fig 3).
Discussion The purpose of the present study was to assess whether GA or AR are equivalent treatments of TMJ ankylosis. We hypothesized that the mean improvement in MIO would be the same between the 2 treatment groups. After a systematic review of the published data, we found 4 reports comparing these 2 treatments. The analysis of these studies showed that patients who underwent GA had a significantly larger postoperative MIO than did the patients who had undergone AR. The goals of operative management are to improve mandibular ROM, prevent reankylosis, and, in the growing patient, allow symmetric mandibular growth. The 3 principal surgical methods include GA, interpositional arthroplasty with placement of interpositional material, and ankylosis resection and reconstruction of the ramus– condyle unit with autogenous or alloplastic grafts.13 Owing to the limited number of studies available, we elected to compare GA and AR. GA is based on leaving the TMJ joint region empty by resecting the ankylotic bone.14 Esmarch was said to be the first surgeon to perform an osteotomy for treating mandibular ankylosis in 1851. The first condylectomy was performed by Humphrey in 1854. Abbe introduced the GA technique in 1880, and interpositional material was used by Risdon in 1934.15 Because of the distance between the resected bone surfaces, interpositional material is essential to prevent reankylosis and promote functional joint activity. The placement of interposition material is an attempt to form a partition between the condyle and the base of the skull.16 However, the recurrence rate was high (53%) in the study by Topazian16 on TMJ ankylosis. The use of an autogenous costochondral graft for TMJ reconstruction was first described by Gillies17 in 1920. Studies by Kaban et al18 and Ware and Brown19 showed that costochondral grafts have proved best among the autogenous tissues used for reconstruction of the mandibular condyle. The main reason for its success is the similarity of the costochondral graft to the condyle and its capacity to regenerate and grow. Saeed and Kent20 reported a high incidence of reankylosis in AR patients. The most prevalent hy-
534
9 ⫾ 5.45 32 5 Brazil ManganelloSouza et al12
Jordan India
Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
32.7 ⫾ 11.9 6.6 ⫾ 4.21 12.7 9
29.9 ⫾ 2.3 34 ⫾ 2 6.5 ⫾ 0.91 1.3 ⫾ 1.73 Unknown 9.3 14 9 6.6 ⫾ 1.9 0.95 ⫾ 1.49 8 22
Turkey
Tanrikulu et al9 Qudah et al10 Balaji11
Unknown 27.2
17
Location Investigator
Abbreviations: AR, ankylosis resection and ramus– condyle unit reconstruction using a costochrondral graft; GA, gap arthroplasty; MIO, maximal incisal opening.
Unilateral 5 Bilateral 2 23.4 ⫾ 2.74 Unknown 32.7 ⫾ 2.644 Unilateral 5 Bilateral 4 26.1 ⫾ 12.62 Unknown
24.3 ⫾ 6.8 26.4 ⫾ 5.56 2.1 ⫾ 3.9 Unknown 7
31.89 ⫾ 4.28 29.28 ⫾ 6.42 Unilateral 10 Bilateral 7 31 ⫾ 2.88 24.4 ⫾ 3.45 Unknown 36.7 ⫾ 2.13 35.75 ⫾ 2.6 Unilateral 15 Bilateral 7 33 ⫾ 4.24 24 ⫾ 6.9 Unknown 2.61 ⫾ 4.78
Laterality Preoperative
MIO Improvement Postoperative (mm) MIO (mm)
Mean Age (yr) Patients (n)
GA group Table 1. STUDY CHARACTERISTICS
Unknown
Preoperative
MIO Improvement Postoperative (mm) MIO (mm)
Patients (n)
Mean Age (yr)
AR
Laterality
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS
pothesis for reankylosis is that the overgrowth of the mandible after costochondral grafting occurs because of the transplantation of excessive amounts of cartilage.21,22 In accordance with that theory, Perrott et al23 reported that when only 2 to 4 mm of cartilage were transplanted with the rib, no overgrowth of the grafts occurred. In the study by Qudah et al,10 although only 2 to 3 mm of cartilage was transplanted, 2 of 16 patients in the costochondral graft group showed overgrowth of the mandible on the operated side during the second year of follow-up. Another disadvantage of the costochondral graft is donor side morbidity, including fracture at the costochondral junction, pleural tear, persistent chest pain, and neuritis.24 The third treatment option for TMJ ankylosis is ankylosis release and reconstruction of the ramus– condyle unit with a total joint prosthesis. Wolford et al25 demonstrated that the total joint prosthesis works well in the long term (5 to 8 years of follow-up) and is a viable technique for TMJ reconstruction. It can be useful, especially for patients with previous multiple TMJ surgeries and abnormal TMJ anatomy.25 However, no study has compared total TMJ prosthesis and costochondral graft placement or GA; therefore, we could not include that treatment option in our metaanalysis. A major drawback of the present study was the limited number of studies included in the meta-analysis. With such a small sample of studies, correctly estimating the publication bias and controlling for confounding using meta-regression analysis was difficult. In addition, with such a small number of studies, each study had a large effect on the final pooled result. It is, however, common for meta-analyses of surgical studies to have limited data pools. Instead of not analyzing the included studies, we chose to perform the meta-analysis with the caveat of a precaution to the reader and allow them to interpret the results. We did not find any publication bias in our funnel plots and Egger’s test. Although funnel plots are straightforward to interpret and have been shown to be a useful tool for the assessment of publication bias, their validity has been questioned.8 Several studies have shown that the shape of a funnel plot is largely determined by the arbitrary choice of the method used to construct the funnel plot.26 Also, misinterpretation of funnel plots can lead investigators to discredit and abandon valid evidence simply because of the asymmetry of a funnel plot.27 A drawback of a meta-analysis with a limited number of studies is that a single study with a large sample size can exert a large effect on the pooled results. For example, in the present study, the results of the metaanalysis were heavily weighted by the study by Balaji.11 Disregarding the study by Balaji,11 the treat-
535
KATSNELSON ET AL
MIO higher in AR group
Source
MIO higher in GA group
Tanrikulu (2005) Qudah (2005) Balaji (2003) Manganello-Souza (2003)
Pooled weighted mean difference (95% CI)
2.4 mm (0.9 mm, 4.0 mm) p = 0.002
0
- 12.0
12.0
GA and AR mean MIO difference (mm)
FIGURE 2. Pooled WMD of MIO between AR and GA groups. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
ment effect favored GA; however, the difference in the MIO between the GA and AR groups was not significant (WMD 1.51 mm, 95% confidence interval ⫺0.94 to 3.96 mm, P ⫽ .2). The results of the present study suggest that GA treatment is associated with better mandibular ROM compared with GA; however, the absolute difference in MIO was small (2.4 mm). Thus, the surgeon needs to consider other factors when considering the treatment options, such as age (ie, growing children) or the patient’s ability to tolerate harvesting a costochondral graft. From the findings of our literature review and metaanalysis, patients with TMJ ankylosis who underwent GA had significantly better MIO than subjects undergoing AR.
Weighted Mean Difference
20
10
0
-10 0
2
4
6
Standard Error of Weighted Mean Difference
FIGURE 3. Begg’s funnel plot with pseudo-95% confidence intervals demonstrating symmetric distribution without systematic heterogeneity of individual study treatment effects compared with standard error of each study, indicating lack of publication bias in meta-analysis. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
References 1. Loveless TP, Bjornland T, Dodson TB, et al: Efficacy of temporomandibular joint ankylosis surgical treatment. J Oral Maxillofac Surg 68:1276, 2010 2. Elgazzar RF, Abdelhady AI, Saad KA, et al: Treatment modalities of TMJ ankylosis: Experience in Delta Nile, Egypt. Int J Oral Maxillofac Surg 39:333, 2010 3. Mercuri LG, Ali FA, Woolson R: Outcomes of total alloplastic replacement with periarticular autogenous fat grafting for management of reankylosis of the temporomandibular joint. J Oral Maxillofac Surg 66:1794, 2008 4. Clarke M: The QUORUM statement. Lancet 355:756, 2000 5. Markiewicz MR, Brady MF, Ding EL, et al: Corticosteroids reduce postoperative morbidity after third molar surgery: A systematic review and meta-analysis. J Oral Maxillofac Surg 66: 1881, 2008 6. Higgins JPT, Green S: Cochrane Handbook for Systematic Reviews of Interventions, 4.2, vol 6, 2006. Available at: http:// www.cochrane.org/resources/handbook/. Accessed July 27, 2007 7. DerSimonian R, Laird N: Meta-analysis in clinical trials. Control Clin Trials 7:1986 8. Egger M, Davey Smith G, Schneider M, et al: Bias in metaanalysis detected by a simple, graphical test. BMJ 315:629, 1997 9. Tanrikulu R, Erol B, Görgün B, et al: The contribution to success of various methods of treatment of temporomandibular joint ankylosis (a statistical study containing 24 cases). Turk J Pediatr 47:261, 2005 10. Qudah MA, Qudeimat MA, Al-Maaita J: Treatment of TMJ ankylosis in Jordanian children—A comparison of two surgical techniques. J Craniomaxillofac Surg 33:30, 2005 11. Balaji SM: Modified temporalis anchorage in craniomandibular reankylosis. Int J Oral Maxillofac Surg 32:480, 2003 12. Manganello-Souza LC, Mariani PB: Temporomandibular joint ankylosis: Report of 14 cases. Int J Oral Maxillofac Surg 32:24, 2003 13. Huang IY, Lai ST, Shen YH, et al: Interpositional arthroplasty using autogenous costal cartilage graft for temporomandibular joint ankylosis in adults. Int J Oral Maxillofac Surg 36:909, 2007 14. Karaca C, Barutcu A, Baytekin C, et al: Modifications of the inverted T-shaped silicone implant for treatment of temporomandibular joint ankylosis. J Craniomaxillofac Surg 32:243, 2004
536 15. Salins PC: New perspectives in the management of craniomandibular ankylosis. Int J Oral Maxillofac Surg 29:337, 2000 16. Topazian RG: Comparison of gap and interposition arthroplasty in the treatment of temporomandibular joint ankylosis. J Oral Surg 24:405, 1966 17. Gillies HD: Plastic Surgery of the Face. London, Oxford University Press, 1920 18. Kaban LB, Perrott DH, Fisher K: A protocol for management of temporomandibular joint ankylosis. J Oral Maxillofac Surg 48:1145, 1990 19. Ware WH, Brown SL: Growth centre transplantation to replace mandibular condyles. J Maxillofac Surg 9:50, 1981 20. Saeed NR, Kent JN: A retrospective study of the costochondral graft in TMJ reconstruction. Int J Oral Maxillofac Surg 32:606, 2003 21. Peltomäki T, Rönning O: Interrelationship between size and tissue-separating potential of costochondral transplants. Eur J Orthod 13:459, 1991
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS 22. Peltomäki T: Growth of a costochondral graft in the rat temporomandibular joint. J Oral Maxillofac Surg 50:851, 1992 23. Perrott DH, Umeda H, Kaban LB: Costochondral graft construction/reconstruction of the ramus/condyle unit: Long-term follow-up. Int J Oral Maxillofac Surg 23:321, 1994 24. Ko EW, Huang CS, Chen YR: Temporomandibular joint reconstruction in children using costochondral grafts. J Oral Maxillofac Surg 57:789, 1999 25. Wolford LM, Pitta MC, Reiche-Fischel O, et al: TMJ Concepts/Techmedica custom-made TMJ total joint prosthesis: 5-Year follow-up study. Int J Oral Maxillofac Surg 32:268, 2003 26. Tang JL, Liu JL: Misleading funnel plot for detection of bias in meta-analysis. J Clin Epidemiol 53:477, 2000 27. Lau J, Ioannidis JP, Terrin N, et al: The case of the misleading funnel plot. BMJ 333:597, 2006
Operative Management of Temporomandibular Joint Ankylosis: A Systematic Review and Meta-Analysis Alexander Katsnelson, DMD, MS,* Michael R. Markiewicz, DDS, MPH, MD,† David A. Keith, BDS, FDSRCS, DMD,‡ and Thomas B. Dodson, DMD, MPH§ Purpose: Two common treatments of temporomandibular joint ankylosis are gap arthroplasty and
ankylosis resection and reconstruction of the ramus– condyle unit with a costochondral graft. The purpose of the present study was to answer the following clinical question: “Among patients with temporomandibular joint ankylosis, do those patients who undergo gap arthroplasty, compared with those who undergo ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft have better postoperative mandibular range of motion?” Methods: A systematic search of the published data was performed to identify eligible studies. The
primary predictor variable was treatment type (ie, gap arthroplasty or ankylosis resection and ramus– condyle unit reconstruction). The main outcome was the change in maximal incisal opening postoperatively. A random effects model was used to compute the pooled weighted mean difference between the pre- and postoperative maximal incisal opening in both treatment groups. Results: Four studies met the inclusion criteria. Those undergoing gap arthroplasty had a significantly greater maximal incisal opening than the group undergoing ankylosis resection and ramus– condyle unit reconstruction. The weighted mean difference between the 2 groups was 2.4 mm (95% confidence interval 0.9 to 4.0; P ⫽ .002). Conclusions: Subjects with temporomandibular joint ankylosis who underwent gap arthroplasty had significantly better postoperative maximal incisal opening than those undergoing ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft. © 2012 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 70:531-536, 2012 The operative management of temporomandibular joint (TMJ) ankylosis is challenging. Common treatments of TMJ ankylosis include gap arthroplasty (GA) and resection of the ankylosis and reconstruction of the
ramus– condyle unit with a costochondral graft or prosthetic joint.1–3 The primary treatment goals are restoration of function and prevention of reankylosis and, in the growing patient, symmetric mandibular growth.
*Formerly, Resident in Training, Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, Boston, MA; currently in private practice, Chicago, IL. †Resident in Training, Department of Oral and Maxillofacial Surgery, Oregon Health and Science University, Portland, OR. ‡Professor of Oral and Maxillofacial Surgery, Harvard School of Dental Medicine. Visiting Oral and Maxillofacial Surgeon, Massachusetts general hospital, Boston, MA. §Visiting (Attending) Oral and Maxillofacial Surgeon, Director, Center for Applied Clinical Investigation, Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, and Professor, Department of Oral and Maxillofacial Surgery, Harvard School of Dental Medicine, Boston, MA.
This project was supported in part by the Massachusetts General Hospital Department of Oral and Maxillofacial Surgery Education and Research Fund (to A. Katsnelson), Center for Applied Clinical Investigation (to T. B. Dodson), and Massachusetts General Physicians Organization (to T. B. Dodson). Address correspondence and reprint requests to Dr Dodson: Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital, 55 Fruit Street, Warren Building, Suite 1201, Boston, MA 02114; e-mail: [email protected] © 2012 American Association of Oral and Maxillofacial Surgeons
0278-2391/12/7003-0$36.00/0 doi:10.1016/j.joms.2011.10.003
531
532 Although numerous reports have addressed the management of TMJ ankylosis, a search of the published data demonstrated no evidence of any randomized trials or systematic reviews addressing the surgery outcomes of different treatments. The purpose of the present report was to complete a systematic review and meta-analysis to answer the following clinical question: “Among patients with TMJ ankylosis, do those patients who undergo GA compared with those who undergo ankylosis resection and ramus– condyle unit reconstruction with a costochondral graft have better mandibular range of motion (ROM)?” For the remainder of our report, ankylosis resection and ramus– condyle unit reconstruction using a costochrondral graft has been abbreviated “AR.” We hypothesized that the postoperative ROM would be the same between the 2 treatment groups. The specific aims of the present project were to perform a systematic review of the published data and identify studies for analysis and to execute a meta-analysis to determine whether GA or AR produces better outcomes in terms of mandibular ROM.
Materials and Methods STUDY DESIGN
To address the research question, we designed a systematic literature review and meta-analysis. SAMPLE
Our review of the published data was designed and conducted in concordance with the QUOROM (Quality of Reporting of Meta-analyses) guidelines.4 Two of us (A.K. and M.R.M.) searched the published studies from January 1966 through May 2010, using the National Library of Medicine (PubMed, available from http://www.pubmed.gov), the National Institutes of Health clinical trials registry (available from http:// www.clinicaltrials.gov), and the Cochrane Central Register of Controlled Trials (available from http://www. mrw.interscience.wiley.com/cochrane) using specific medical subject headings and key words, including “temporomandibular joint.” “ankylosis/gap arthroplasty/costochondral graft.” In addition, the on-line databases of the “British Journal of Oral and Maxillofacial Surgery,” “Journal of the American Dental Association,” “Journal of Dental Research,” “Journal of Oral and Maxillofacial Surgery,” “International Journal of Oral and Maxillofacial Surgery,” and “Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology,” as well as the on-line abstract indexes of the conference proceedings of the American Association of Oral and Maxillofacial Surgeons and the International Association for Dental Research annual meetings were searched for
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS
studies. Electronic copies of the reports identified by the search were obtained. Bibliographies from the relevant reports were also searched for studies that might have included data relevant to the study topic. STUDY CRITERIA
Studies were included only if all the following eligibility criteria were met: the 2 treatment types (GA and AR) were compared, descriptive statistics and the primary outcome variable (ie, maximal incisal opening [MIO]) were reported or the mean and standard deviation (SD) could be computed from the report, and the pre- and postoperative data were available. DATA EXTRACTION
Using a standardized data extraction form, one of us (A.K.) extracted and tabulated all data. STUDY VARIABLES
The primary predictor variable was the treatment type (ie, GA or AR). The primary outcome variable was the MIO. The primary comparison of interest was to estimate the differences between the preoperative and postoperative MIOs between the 2 study groups (GA and AR). The reported mean difference between the preoperative measurements, SD of the difference, and sample size were used to compute the weighted mean difference (WMD) for the outcome variable (MIO) for the GA and AR groups. All MIO measurements were converted to millimeters. The other study variables included the total sample size, number of subjects in both treatment groups, and mean age. The anatomic location of ankylosis was recorded as unilateral or bilateral. STATISTICAL ANALYSIS
We used a conservative approach to estimate the SD of the change from baseline: SD (change SD from SD baseline) ⫽
冑
(Preoperative SD)2 ⫹ (Postoperative SD)2 2
This method has been previously published by Markiewicz et al.5 This formula results in a larger, and most likely overestimated, SD than the formula recommended in the Cochrane Handbook.6 The primary summary measure of association was the overall DerSimonian and Laird random effects pooled mean difference in postoperative MIO between the GA and AR groups.7 A similar method of mandibular ROM measurement was used across all studies (MIO); therefore, a WMD meta-analysis was performed. Unlike the standardized mean difference,
533
KATSNELSON ET AL
the WMD allows for direct interpretation by readers in common units used across all studies in the metaanalysis. Meta-regression analyses were not performed owing to the small number of trials included in the pooled analysis. To test for the presence of publication bias, the relative symmetry of the individual study estimates was assessed around the overall estimates using Begg’s funnel plots, in which the mean differences were plotted against their corresponding standard error. Egger’s test for publication bias was also performed.8 All analyses were conducted using Stata, version 9.2 (StataCorp, College Station, TX). P ⬍ .05 was considered statistically significant.
Results After eliminating duplicate reports or those reporting previously published results, 1,518 eligible reports were identified for possible study inclusion. The vast majority (n ⫽ 1,464) were excluded for 1 or more of the following reasons: case report, retrospective study, not of human subjects, review, editorial or commentary, or did not test the association of interest (Fig 1). After retrieving the full texts of the pertinent studies (n ⫽ 54), 50 were excluded: 3 were single case reports, 16 used a different treatment technique (ie, total prosthetic joint), 20 had no comparison group, 4 did not assess the association of interest, 3 did not have the MIO or other important information,
FIGURE 1. Flow diagram. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
1 was a discussion for a report without any original data, and 3 were review studies. The final sample included 4 studies used for data extraction. The study characteristics of the included reports9 –12 are presented in Table 1. After treatment, the subjects undergoing GA had a significantly larger MIO than those who had undergone AR (WMD 2.4 mm, 95% confidence interval 0.9 to 4.0, P ⫽ .002; Fig 2). Egger’s test did not reveal any significant publication bias (P ⫽ .7; Fig 3).
Discussion The purpose of the present study was to assess whether GA or AR are equivalent treatments of TMJ ankylosis. We hypothesized that the mean improvement in MIO would be the same between the 2 treatment groups. After a systematic review of the published data, we found 4 reports comparing these 2 treatments. The analysis of these studies showed that patients who underwent GA had a significantly larger postoperative MIO than did the patients who had undergone AR. The goals of operative management are to improve mandibular ROM, prevent reankylosis, and, in the growing patient, allow symmetric mandibular growth. The 3 principal surgical methods include GA, interpositional arthroplasty with placement of interpositional material, and ankylosis resection and reconstruction of the ramus– condyle unit with autogenous or alloplastic grafts.13 Owing to the limited number of studies available, we elected to compare GA and AR. GA is based on leaving the TMJ joint region empty by resecting the ankylotic bone.14 Esmarch was said to be the first surgeon to perform an osteotomy for treating mandibular ankylosis in 1851. The first condylectomy was performed by Humphrey in 1854. Abbe introduced the GA technique in 1880, and interpositional material was used by Risdon in 1934.15 Because of the distance between the resected bone surfaces, interpositional material is essential to prevent reankylosis and promote functional joint activity. The placement of interposition material is an attempt to form a partition between the condyle and the base of the skull.16 However, the recurrence rate was high (53%) in the study by Topazian16 on TMJ ankylosis. The use of an autogenous costochondral graft for TMJ reconstruction was first described by Gillies17 in 1920. Studies by Kaban et al18 and Ware and Brown19 showed that costochondral grafts have proved best among the autogenous tissues used for reconstruction of the mandibular condyle. The main reason for its success is the similarity of the costochondral graft to the condyle and its capacity to regenerate and grow. Saeed and Kent20 reported a high incidence of reankylosis in AR patients. The most prevalent hy-
534
9 ⫾ 5.45 32 5 Brazil ManganelloSouza et al12
Jordan India
Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
32.7 ⫾ 11.9 6.6 ⫾ 4.21 12.7 9
29.9 ⫾ 2.3 34 ⫾ 2 6.5 ⫾ 0.91 1.3 ⫾ 1.73 Unknown 9.3 14 9 6.6 ⫾ 1.9 0.95 ⫾ 1.49 8 22
Turkey
Tanrikulu et al9 Qudah et al10 Balaji11
Unknown 27.2
17
Location Investigator
Abbreviations: AR, ankylosis resection and ramus– condyle unit reconstruction using a costochrondral graft; GA, gap arthroplasty; MIO, maximal incisal opening.
Unilateral 5 Bilateral 2 23.4 ⫾ 2.74 Unknown 32.7 ⫾ 2.644 Unilateral 5 Bilateral 4 26.1 ⫾ 12.62 Unknown
24.3 ⫾ 6.8 26.4 ⫾ 5.56 2.1 ⫾ 3.9 Unknown 7
31.89 ⫾ 4.28 29.28 ⫾ 6.42 Unilateral 10 Bilateral 7 31 ⫾ 2.88 24.4 ⫾ 3.45 Unknown 36.7 ⫾ 2.13 35.75 ⫾ 2.6 Unilateral 15 Bilateral 7 33 ⫾ 4.24 24 ⫾ 6.9 Unknown 2.61 ⫾ 4.78
Laterality Preoperative
MIO Improvement Postoperative (mm) MIO (mm)
Mean Age (yr) Patients (n)
GA group Table 1. STUDY CHARACTERISTICS
Unknown
Preoperative
MIO Improvement Postoperative (mm) MIO (mm)
Patients (n)
Mean Age (yr)
AR
Laterality
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS
pothesis for reankylosis is that the overgrowth of the mandible after costochondral grafting occurs because of the transplantation of excessive amounts of cartilage.21,22 In accordance with that theory, Perrott et al23 reported that when only 2 to 4 mm of cartilage were transplanted with the rib, no overgrowth of the grafts occurred. In the study by Qudah et al,10 although only 2 to 3 mm of cartilage was transplanted, 2 of 16 patients in the costochondral graft group showed overgrowth of the mandible on the operated side during the second year of follow-up. Another disadvantage of the costochondral graft is donor side morbidity, including fracture at the costochondral junction, pleural tear, persistent chest pain, and neuritis.24 The third treatment option for TMJ ankylosis is ankylosis release and reconstruction of the ramus– condyle unit with a total joint prosthesis. Wolford et al25 demonstrated that the total joint prosthesis works well in the long term (5 to 8 years of follow-up) and is a viable technique for TMJ reconstruction. It can be useful, especially for patients with previous multiple TMJ surgeries and abnormal TMJ anatomy.25 However, no study has compared total TMJ prosthesis and costochondral graft placement or GA; therefore, we could not include that treatment option in our metaanalysis. A major drawback of the present study was the limited number of studies included in the meta-analysis. With such a small sample of studies, correctly estimating the publication bias and controlling for confounding using meta-regression analysis was difficult. In addition, with such a small number of studies, each study had a large effect on the final pooled result. It is, however, common for meta-analyses of surgical studies to have limited data pools. Instead of not analyzing the included studies, we chose to perform the meta-analysis with the caveat of a precaution to the reader and allow them to interpret the results. We did not find any publication bias in our funnel plots and Egger’s test. Although funnel plots are straightforward to interpret and have been shown to be a useful tool for the assessment of publication bias, their validity has been questioned.8 Several studies have shown that the shape of a funnel plot is largely determined by the arbitrary choice of the method used to construct the funnel plot.26 Also, misinterpretation of funnel plots can lead investigators to discredit and abandon valid evidence simply because of the asymmetry of a funnel plot.27 A drawback of a meta-analysis with a limited number of studies is that a single study with a large sample size can exert a large effect on the pooled results. For example, in the present study, the results of the metaanalysis were heavily weighted by the study by Balaji.11 Disregarding the study by Balaji,11 the treat-
535
KATSNELSON ET AL
MIO higher in AR group
Source
MIO higher in GA group
Tanrikulu (2005) Qudah (2005) Balaji (2003) Manganello-Souza (2003)
Pooled weighted mean difference (95% CI)
2.4 mm (0.9 mm, 4.0 mm) p = 0.002
0
- 12.0
12.0
GA and AR mean MIO difference (mm)
FIGURE 2. Pooled WMD of MIO between AR and GA groups. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
ment effect favored GA; however, the difference in the MIO between the GA and AR groups was not significant (WMD 1.51 mm, 95% confidence interval ⫺0.94 to 3.96 mm, P ⫽ .2). The results of the present study suggest that GA treatment is associated with better mandibular ROM compared with GA; however, the absolute difference in MIO was small (2.4 mm). Thus, the surgeon needs to consider other factors when considering the treatment options, such as age (ie, growing children) or the patient’s ability to tolerate harvesting a costochondral graft. From the findings of our literature review and metaanalysis, patients with TMJ ankylosis who underwent GA had significantly better MIO than subjects undergoing AR.
Weighted Mean Difference
20
10
0
-10 0
2
4
6
Standard Error of Weighted Mean Difference
FIGURE 3. Begg’s funnel plot with pseudo-95% confidence intervals demonstrating symmetric distribution without systematic heterogeneity of individual study treatment effects compared with standard error of each study, indicating lack of publication bias in meta-analysis. Katsnelson et al. Operative Management of TMJ Ankylosis. J Oral Maxillofac Surg 2012.
References 1. Loveless TP, Bjornland T, Dodson TB, et al: Efficacy of temporomandibular joint ankylosis surgical treatment. J Oral Maxillofac Surg 68:1276, 2010 2. Elgazzar RF, Abdelhady AI, Saad KA, et al: Treatment modalities of TMJ ankylosis: Experience in Delta Nile, Egypt. Int J Oral Maxillofac Surg 39:333, 2010 3. Mercuri LG, Ali FA, Woolson R: Outcomes of total alloplastic replacement with periarticular autogenous fat grafting for management of reankylosis of the temporomandibular joint. J Oral Maxillofac Surg 66:1794, 2008 4. Clarke M: The QUORUM statement. Lancet 355:756, 2000 5. Markiewicz MR, Brady MF, Ding EL, et al: Corticosteroids reduce postoperative morbidity after third molar surgery: A systematic review and meta-analysis. J Oral Maxillofac Surg 66: 1881, 2008 6. Higgins JPT, Green S: Cochrane Handbook for Systematic Reviews of Interventions, 4.2, vol 6, 2006. Available at: http:// www.cochrane.org/resources/handbook/. Accessed July 27, 2007 7. DerSimonian R, Laird N: Meta-analysis in clinical trials. Control Clin Trials 7:1986 8. Egger M, Davey Smith G, Schneider M, et al: Bias in metaanalysis detected by a simple, graphical test. BMJ 315:629, 1997 9. Tanrikulu R, Erol B, Görgün B, et al: The contribution to success of various methods of treatment of temporomandibular joint ankylosis (a statistical study containing 24 cases). Turk J Pediatr 47:261, 2005 10. Qudah MA, Qudeimat MA, Al-Maaita J: Treatment of TMJ ankylosis in Jordanian children—A comparison of two surgical techniques. J Craniomaxillofac Surg 33:30, 2005 11. Balaji SM: Modified temporalis anchorage in craniomandibular reankylosis. Int J Oral Maxillofac Surg 32:480, 2003 12. Manganello-Souza LC, Mariani PB: Temporomandibular joint ankylosis: Report of 14 cases. Int J Oral Maxillofac Surg 32:24, 2003 13. Huang IY, Lai ST, Shen YH, et al: Interpositional arthroplasty using autogenous costal cartilage graft for temporomandibular joint ankylosis in adults. Int J Oral Maxillofac Surg 36:909, 2007 14. Karaca C, Barutcu A, Baytekin C, et al: Modifications of the inverted T-shaped silicone implant for treatment of temporomandibular joint ankylosis. J Craniomaxillofac Surg 32:243, 2004
536 15. Salins PC: New perspectives in the management of craniomandibular ankylosis. Int J Oral Maxillofac Surg 29:337, 2000 16. Topazian RG: Comparison of gap and interposition arthroplasty in the treatment of temporomandibular joint ankylosis. J Oral Surg 24:405, 1966 17. Gillies HD: Plastic Surgery of the Face. London, Oxford University Press, 1920 18. Kaban LB, Perrott DH, Fisher K: A protocol for management of temporomandibular joint ankylosis. J Oral Maxillofac Surg 48:1145, 1990 19. Ware WH, Brown SL: Growth centre transplantation to replace mandibular condyles. J Maxillofac Surg 9:50, 1981 20. Saeed NR, Kent JN: A retrospective study of the costochondral graft in TMJ reconstruction. Int J Oral Maxillofac Surg 32:606, 2003 21. Peltomäki T, Rönning O: Interrelationship between size and tissue-separating potential of costochondral transplants. Eur J Orthod 13:459, 1991
OPERATIVE MANAGEMENT OF TMJ ANKYLOSIS 22. Peltomäki T: Growth of a costochondral graft in the rat temporomandibular joint. J Oral Maxillofac Surg 50:851, 1992 23. Perrott DH, Umeda H, Kaban LB: Costochondral graft construction/reconstruction of the ramus/condyle unit: Long-term follow-up. Int J Oral Maxillofac Surg 23:321, 1994 24. Ko EW, Huang CS, Chen YR: Temporomandibular joint reconstruction in children using costochondral grafts. J Oral Maxillofac Surg 57:789, 1999 25. Wolford LM, Pitta MC, Reiche-Fischel O, et al: TMJ Concepts/Techmedica custom-made TMJ total joint prosthesis: 5-Year follow-up study. Int J Oral Maxillofac Surg 32:268, 2003 26. Tang JL, Liu JL: Misleading funnel plot for detection of bias in meta-analysis. J Clin Epidemiol 53:477, 2000 27. Lau J, Ioannidis JP, Terrin N, et al: The case of the misleading funnel plot. BMJ 333:597, 2006