Repair of facial nerve after removal of cerebellopontine angle tumors: a comparative study.

The results of repair of 18 facial nerves were examined by means of a modified House-Brackmann grading system. Six were repaired by end-to-end anastomosis and 12 by nerve graft. The reliability of the simplified House-Brackmann grading system was also assessed, using the kappa statistic to analyze the agreement between pairs of observers who examined the function of 40 nerves in 37 patients. Facial nerves studied had been either preserved, repaired or grafted, or divided and treated by faciohypoglossal nerve anastomosis. One nerve was not treated. The grading system proved to be somewhat unreliable, with complete agreement between observers in only 25% of cases. Facial nerve repair produced a fair return of function in just under two-thirds of the cases. The ability of an examiner ignorant of the patient's history to assess from the end result how the nerve had been managed was also estimated. Observers showed little ability to decide correctly on the previous treatment of the nerve when the patient showed moderate dysfunction postoperatively. The implications of these findings for grading systems and for management of the facial nerve in acoustic nerve tumor surgery are discussed.