We retrospectively investigated the prognostic factors for recovery by multivariate analyses in patients with peripheral facial nerve palsy who had undergone transmastoid facial nerve decompression at our department.
Forty-four patients with Bell’s palsy or Ramsay Hunt syndrome were enrolled in the study. They were 26 men and 18 women, aged 17 to 80 years with an average of 55.4 years. Twenty-seven patients had Bell’s palsy and 17 patients had Hunt syndrome. We performed multiple regression analysis and binary logistic regression analysis with the final palsy score (Scorefinal) or increase in the palsy score (ΔScore) as the outcome variables. Six explanatory variables were set; preoperative palsy score (Scorepre), days from onset to surgery, patient age, gender (male = 1, female = 0), causative disease (Bell = 1, Hunt = 0), and percent electroneurographic response.
Eleven patients (25.0%) completely recovered (Scorefinal ≥36), the Scorefinal was restored to ≥20 in 40 patients (90.9%), and there was no patient with Scorefinal <10. In multivariate analyses, the partial regression coefficient was negative for days from onset, age and gender, and positive for causative disease and electroneurographic response. On the other hand, the sign of the partial regression coefficient for Scorepre changed with the outcome variable; positive with Scorefinal and negative with ΔScore. The probability of ΔScore ≥30, calculated by the binary logistic regression equation, was larger in patients with ΔScore ≥30 than in those with ΔScore <30 (0.442 ± 0.230 vs. 0.260 ± 0.151, P = 0.003). The positive and negative predictive values for ΔScore ≥30 by this analysis were 0.750 and 0.778, respectively.
These results suggest that early surgery, young age, female gender, Bell’s palsy, and high electroneurographic response may be favorable for recovery from palsy after facial nerve decompression, and that binary logistic regression analysis is useful for practical prediction of recovery.
