The purpose of this study was to assess the prevalence of a ‘bad’ split after sagittal ramus osteotomies (SRO) and report the results of initial mandibular healing. A retrospective cohort study derived from patients treated by a single surgeon at one institution between 2004 and 2013 was performed. An index group consisting of a series of subjects with a spectrum of bimaxillary dentofacial deformities also involving the chin and symptomatic chronic obstructive nasal breathing was identified. The SRO design, bicortical screw fixation technique, and perioperative management were consistent. Outcome variables included the occurrence of a ‘bad’ split and the success of initial SRO healing. Two hundred sixty-two subjects undergoing 524 SROs met the inclusion criteria. Their average age was 25 years (range 13–63 years) and 134 were female (51%). Simultaneous removal of a third molar was performed during 209 of the SROs (40%). There were no ‘bad’ splits. All subjects achieved successful bone union, the planned occlusion, and return to a chewing diet and physical activities by 5 weeks after surgery. The presence of a third molar removed during SRO was not associated with an increased frequency of a ‘bad’ split or delayed mandibular healing.
The sagittal ramus osteotomy (SRO) is the most common procedure performed for the correction of dentofacial deformities involving the mandible. The technique was first reported by Trauner and Obwegeser in 1955, and has been modified over the years by a number of clinicians. Individual surgeons continue to prefer minor variations in SRO technique, primarily in attempts to avoid unfavourable splitting. The literature shows the incidence of ‘bad’ splits to vary widely (1–23%). A buccal plate fracture of the proximal segment, a lingual plate fracture of the distal segment, or a coronoid process fracture of the proximal segment are the most commonly recorded unfavourable fractures. These ‘less than ideal’ ramus osteotomies, often categorized as a ‘bad’ split, may explain the high incidence reported in some series.
When splitting an SRO of the mandible, the majority of ‘less than ideal’ fractures may result in fixation challenges; however, they are rarely responsible for either short-term healing difficulties or long-term negative sequelae. When a fragment of the buccal plate separates from the proximal segment it is removed if small, or if the fragment is large, it is maintained with additional fixation. When the lingual plate separates from the distal segment, the fragment is generally removed, but it may be left in place without additional fixation. When the coronoid process separates from the proximal segment, it can be removed to limit the chance of ankylosis during the healing process. The present authors believe that these ‘less than ideal’ fractures should only be classified as a ‘bad’ split if more extensive procedures are required that alter initial favourable healing. The most concerning form of a ‘bad’ split is when it occurs in the ascending ramus region. When performing an SRO, if the medial ramus osteotomy is either incomplete through the cortex or too ‘high’, there is a tendency for the split to creep up and include the condyle within the distal segment of the mandible. The present authors define a ‘bad’ split as those SROs in which (1) the condyle remains within the distal segment, requiring an additional osteotomy to separate it as a third piece, or (2) the proximal segment ‘shatters’ during the splitting process, resulting in a separate condylar fragment.
Surgeons continue to refine their operative techniques to limit ‘bad’ splitting of an SRO. Concerns persist that risk factors for a ‘bad’ split include simultaneous removal of a third molar in the line of osteotomy, the subject’s age, and the pattern of the dentofacial deformity. The purpose of this retrospective cohort study was to report the prevalence of ‘bad’ splitting of an SRO, as defined above, in a series of patients. It was hypothesized that when using an osteotomy technique designed to limit ‘bad’ splitting, adequate bone overlap for rigid bicortical screw fixation would still remain for favourable initial mandibular healing. It was also speculated that the simultaneous extraction of a third molar in the line of an SRO, the subject’s age and sex, and the pattern of the presenting developmental dentofacial deformity would not increase the risk of a ‘bad’ split or unfavourable initial mandibular healing. The specific aims of this study were to document (1) the prevalence of a ‘bad’ split during SRO when using a consistent surgical technique to treat a spectrum of developmental dentofacial deformities, (2) the variation in the occurrence of a ‘bad’ split in the presence of a third molar (impacted, erupted, or partially erupted) that is simultaneously removed, and (3) the efficacy of the described SRO technique and bicortical screw fixation method to achieve initial favourable mandibular healing.
Materials and methods
Study design and sample
To address the research objectives, a retrospective cohort study was designed and implemented. The study sample was derived from an index group of patients treated by one surgeon (JCP) in a private practice setting, with surgery carried out at a single hospital between 2004 and 2013. The study patients all had a bimaxillary developmental dentofacial deformity also involving the chin and symptomatic chronic obstructive nasal breathing. These subjects underwent surgery that at a minimum included a Le Fort I osteotomy and bilateral SROs, osseous genioplasty, septoplasty (submucosal resection), and reduction of the inferior turbinates.
Patients were excluded if their jaw deformity had been operated on previously, or if their condition was syndromic, cleft-related, post-traumatic, or tumour-related. Patients not residing in North America were also excluded, as long-term follow-up was geographically inconsistent. Other exclusions included subjects using nicotine products for at least 3 weeks prior to surgery, those currently taking bisphosphonate medication, immunosuppressed individuals, and those with poorly controlled diabetes. All subjects in the study were confirmed to be cardiovascularly stable and without pulmonary disease, renal disease, or a known coagulation disorder.
The use of antibiotics, incision site preparation, wound care after surgery, extent of allowed early postoperative jaw function, general physical activity level, and diet restrictions early after surgery were consistent for the study. Intravenous prophylactic antibiotics were given 15–30 min prior to the surgical incision. Adults weighing less than 80 kg were given 1 g of cefazolin. Adults weighing more than 80 kg were given 2 g of cefazolin. Postoperative doses were given intravenously every 8 h until hospital discharge. This was followed by an additional course of cephalexin (500 mg every 6 h, elixir form) extending for a 5-day total antibiotic regimen. Subjects with a history of penicillin allergy were given clindamycin as an alternative.
Incision site preparation and wound management after surgery were consistent for all subjects throughout the study. In the operating room, after induction of anaesthesia, the head and neck were prepped with povidone–iodine from the scalp to the clavicles. The teeth, tongue, and gingiva were scrubbed with a toothbrush using chlorhexidine 0.12% solution. At the end of the operation, the mouth and teeth were again scrubbed with a toothbrush using chlorhexidine. The subjects were instructed to brush their teeth with a toothbrush five times per day using chlorhexidine, commencing the morning after surgery and for the following 2 weeks. After 2 weeks, the elastics were removed for 6–10 h per day, and limited mandibular opening was encouraged. Light elastics were reapplied each evening. The subjects remained on a full liquid diet for 5 weeks. Physical activities were sedentary during the same timeframe. Subjects were followed on an outpatient basis by the operating surgeon at weekly intervals through the initial 5 weeks after surgery. After documentation of initial favourable healing by physical examination and radiographic analysis, ongoing care was transferred back to the treating orthodontist. Four weeks later (9 weeks after surgery), communication between the surgeon and patient/family confirmed successful achievement of a full chewing diet and return to vigorous physical activities. If not, appropriate physical therapy was arranged for. Further routine surgeon assessment was conducted after the completion of orthodontic treatment and at a minimum of 12 months after surgery. This study was approved by the necessary institutional review board.
The predictor variables studied for a ‘bad’ split and unfavourable initial osteotomy site healing were grouped into the following categories: demographic, anatomic, and operative.
The demographic variables collected included age at the time of operation and sex.
The anatomic variables studied included the presence and position of the mandibular third molar and the pattern of the developmental dentofacial deformity.
The first anatomic variable was the presence or absence of a third molar at the osteotomy site. If present, the third molar was further classified into one of three categories: fully impacted (no exposure through the oral mucosa), fully erupted (full crown exposure through the oral mucosa), or partially erupted (partial exposure of the crown through the oral mucosa). The second anatomic variable was the pattern of the developmental dentofacial deformity. At presentation, each subject was classified into one of six abnormal jaw growth patterns: primary mandibular deficiency, maxillary deficiency with relative mandibular excess, asymmetric mandibular excess, short face, long face, or bimaxillary dental protrusion. In addition there was an atypical category, which included dentofacial deformities strongly influenced by parafunctional habits such as thumb sucking.
Operative variables documented included the design of the SRO, method of osteotomy fixation, and management of a third molar in the line of osteotomy.
The SRO design was consistent for all study subjects. A unique aspect of the SRO technique used in the study is the medial osteotomy, which is ‘low and short’ ( Fig. 1 A) . This is found to be a practical way to limit an unwanted connection between the condyle and the distal segment of the mandible during the splitting process (‘bad’ split).
The osteotomy is initiated with a reciprocating saw (short straight blade) placed parallel to and just above the mandibular plane of occlusion. The osteotomy is typically located below the mandibular foramen and will by design result in limited lingual plate attached to the distal segment. The lateral osteotomy continues anteriorly for a varied length depending on the degree of the planned mandibular advancement, but always with extension to at least between the first and second molars. Greater planned mandibular advancements require further anterior extension of the lateral osteotomy prior to executing the vertical cut ( Fig. 1 A). This is done to provide adequate bone contact at the osteotomy site after the distal mandible is repositioned into its new location. The connection between the lateral and vertical osteotomies is rounded to limit buccal shelf fragmentation during the splitting process. The vertical osteotomy is bevelled through the buccal cortex at 45°, and then carried just through the inferior border for the same reason ( Fig. 1 B). The inferior border osteotomy is extended through the buccal cortex but not all the way through the lingual cortex as this could propagate a ‘bad’ split ( Fig. 1 B). Once the three cortical osteotomies are fully connected, the ramus is ‘split.’ Using this technique, the splitting process typically involves the use of four instruments and seven steps.
A 7-mm osteotome is placed into the lateral cortical cut adjacent to the second molar and tapped in (approximately 5 mm). If an impacted third molar is present, the osteotome is first inserted just posterior to, then just anterior to the tooth.
A 5-mm osteotome is placed in the lateral cortical cut just anterior to the first osteotome. It is tapped in approximately 5 mm confirming movement and then removed.
The 5-mm osteotome is next placed into the lower aspect of the vertical cut, just above the inferior border, and tapped in through the cortex. Due to proximity of the inferior alveolar nerve (IAN), which may be located just deep to the buccal cortex, penetration of the osteotome is limited once through the cortex.
Twisting both osteotomes simultaneously yields movement and confirms that all cortical cuts are complete.
The 5-mm osteotome is removed and the Smith spreader forceps placed deeper into the anterior aspect of the lateral cut.
The 7-mm osteotome is removed and a broader-based instrument (15 mm) is inset.
The spreader forceps and the broad-based osteotome are spread/twisted simultaneously. No significant pressure is required to complete the split. If resistance is encountered during the spreading/twisting process, this indicates a persistent bicortical bone connection. The persistent bone connection is typically below the IAN and adjacent to the gonial notch. If so, while protecting the IAN, the reciprocating saw (short straight blade) is typically used to complete the separation.
It was planned to fixate the SRO with three bicortical screws of 2.3 mm in diameter (Stryker Corporation, Kalamazoo, MI, USA) ( Fig. 2 ). The length of the screws varied from 14 mm to 18 mm according to the width of the overlapping medial and lateral cortices of the mandible. The intention was to accomplish effective rigid fixation by widely separating the placement of the screws from each other while achieving bicortical contact and preventing injury to the dental roots and the IAN. The drill holes and screws were placed through a transbuccal trocar (Stryker Corporation, Kalamazoo, MI, USA). A graduated drill bit was used to determine the proper length of the screws.
If extraction of a mandibular third molar was required, it was the senior surgeon’s (JCP) preference to remove it at the time of orthognathic surgery. If the mandibular third molar was fully impacted, it was removed from the distal segment through the osteotomy site just after splitting the ramus. If the mandibular third molar was partially or fully erupted, it was typically removed after the ramus vestibular incision and soft tissue dissection, and prior to splitting of the mandible. This was to limit unwanted fracture of the lingual plate. The soft tissue incision design and wound closure for the SRO varied according to the eruption pattern of the mandibular third molar.
The outcome variables studied included the occurrence of a ‘bad’ split and the initial ramus osteotomy healing. The outcome variable of a ‘bad’ split was documented by review of each subject’s operative report and analysis of the postoperative lateral cephalogram and panoramic radiographs. The outcome variable of initial ramus osteotomy site healing was recorded as either favourable or unfavourable. The absence of non-union, malunion, and fibrous union at 2 months after surgery was required as evidence of favourable initial SRO healing. Recording of favourable initial healing specifically required timely (1) confirmation of bony union on physical examination (lack of osteotomy site mobility), (2) documentation of healing by radiographic analysis (i.e., close approximation and correct positioning of the segments to each other and condylar position in the glenoid fossa), (3) documentation of the subject’s occlusion when in function (i.e., comparison to model planning), (4) confirmation of the subject’s return to the treating orthodontist for routine care (without delay and without special needs), and (5) confirmation of the subject’s achievement of expected oro-motor function without compromise (i.e., full chewing diet). Any subject not meeting each of these five milestones early in the healing process (by 2 months after surgery) was considered a red flag for unfavourable SRO healing. Any need for reoperation of the mandible within 1 year after surgery was also reported as a potential indication of unfavourable healing.
Collection, management, and analysis of data
The data were abstracted from the subjects’ hospital and outpatient medical records and recorded on a standardized data collection form by one researcher/observer (EC). This included review of each subject’s standard interval facial and occlusal photographs and radiographs, as well as each subject’s analytic model planning. Each subject underwent lateral cephalometric and panoramic radiography within 2 months before and approximately 5 weeks after surgery. The preoperative radiographs were used to confirm the anatomy and diagnosis. The postoperative radiographs were reviewed to confirm the following: approximation and position of the osteotomy segments, location of the condyle in the glenoid fossa, placement of fixation, removal of a third molar(s), and the occurrence/absence of a ‘bad’ split. The data were entered into a database created using Microsoft Access (Microsoft Inc., Redmond, WA, USA). The data were transferred to a software package for statistical analysis (Stata 13.1; StataCorp LP, College Station, TX, USA). Descriptive statistics were computed for all study variables. The goal was also to analyze positive findings in the study through bivariate and logistic regression analysis. No subjects were lost to follow-up and no records or data points were missing for any of the subjects for any parameters assessed in the study.