Retrospective review of microsurgical repair of 222 lingual nerve injuries.

PURPOSE Injury to the lingual nerve (LN) is a known complication associated with several oral and maxillofacial surgical procedures. We have reviewed the demographics, timing, and outcome of microsurgical repair of the LN. MATERIALS AND METHODS A retrospective chart review was completed of all patients who had undergone microsurgical repair of the LN by one of us (R.A.M.) from March 1986 through December 2005. A physical examination, including standardized neurosensory testing, was completed of each patient preoperatively. All patients were followed up periodically after surgery for at least 1 year, with neurosensory testing repeated at each visit. Sensory recovery was determined from the patient's final neurosensory testing results and evaluated using the guidelines established by the Medical Research Council Scale. The following data were collected and analyzed: patient age, gender, nerve injury etiology, chief sensory complaint (numbness or pain, or both), interval from injury to surgical intervention, intraoperative findings, surgical procedure, and neurosensory status at the final evaluation. The patients were classified according to whether they achieved "useful sensory recovery" or better, according to the Medical Research Council Scale, or had unsatisfactory or no improvement in sensation. Logistic regression methods and associated odds ratios (OR) were used to quantify the association between the risk factors and improvement. Receiver operating characteristic curve analysis was used to find the age threshold and duration that maximally separated the patient outcomes. RESULTS A total of 222 patients (51 males and 171 females; average age 31.1 years, range 15 to 61) underwent LN repair and returned for at least 1 year of follow-up. The most common cause of LN injury was mandibular third molar removal (n = 191, 86%), followed by sagittal split mandibular ramus osteotomy (n = 14, 6.3%). Most patients complained preoperatively of numbness (n = 122, 55%) or numbness with pain (n = 94, 42.3%). The average interval from injury to surgery was 8.5 months (range 1.5 to 96). The most commonly performed operation was excision of a proximal stump neuroma with neurorrhaphy (n = 154, 69%), followed by external decompression with internal neurolysis (n = 29, 13%). Nineteen patients (8.6%) underwent an autogenous nerve graft procedure (greater auricular or sural nerve) for reconstruction of a nerve gap. A collagen cuff was placed around the repair site in 8 patients (3.6%; external decompression with internal neurolysis in 2 and neurorrhaphy in 6). Recovery from neurosensory dysfunction (defined by the Medical Research Council Scale as ranging from "useful sensory function" to a "complete return of sensation") was observed in 201 patients (90.5%; 146 patients with complete recovery and 55 patients with recovery to "useful sensory function"), and 21 patients (9.5%) had no or inadequate improvement. Using the logistic regression model, a shorter interval between nerve injury and repair resulted in greater odds of improvement (OR 0.942, P = .0064); with each month that passed, the odds of improvement decreased by 5.8%. The receiver operating characteristic analysis revealed that patients who waited more than 9 months for repair were at a significantly greater risk of nonimprovement. Statistical significance was observed between patient age and outcome (OR 0.945, P = .0067) representing a 5.5% decrease in the chance of recovery for every year of age in patients 45 years old and older. The odds of a return of acceptable neurosensory function were better when the patient's presenting symptom was pain and not numbness (OR 0.04, P < .001). CONCLUSIONS Microsurgical repair of LN injury has the best chance of successful restoration of acceptable neurosensory function if done within 9 months of the injury. The likelihood of recovery after nerve repair decreased progressively when the repair occurred more than 9 months after injury and with increasing patient age.