Format

Send to

Choose Destination
Otol Neurotol. 2011 Aug;32(6):1031-6. doi: 10.1097/MAO.0b013e31822558de.

Prognostic factors of peripheral facial palsy: multivariate analysis followed by receiver operating characteristic and Kaplan-Meier analyses.

Author information

1
Department of Otolaryngology, Suita Municipal Hospital, 2-13-20 Katayama-cho, Suita City, Osaka 564-0082, Japan.

Abstract

OBJECTIVE:

To search for prognostic predictors and reexamine the usefulness of electroneurography (ENoG) in predicting the prognosis of peripheral facial palsy using statistical methods.

STUDY DESIGN:

Prospective study.

SETTING:

Tertiary referral center.

PATIENTS:

Consecutive 142 patients with Bell's palsy and 26 with Ramsay Hunt syndrome treated with steroid plus antiviral agents.

INTERVENTIONS:

Multivariate analysis was used to identify which factors, including Yanagihara grading score and ENoG, predict better recovery. Receiver operating characteristic (ROC) curves were constructed for ENoG and grading score. The cumulative recovery rate by ENoG was calculated using Kaplan-Meier analysis. Recovery was defined as the improvement of grading score to 36 points or more (full score, 40) without synkinesis.

RESULTS:

Multivariate analysis revealed that Ramsay Hunt syndrome, the worst grading score and ENoG were the significant prognostic predictors. The area under the ROC curve for ENoG was broader than those for grading score, indicating that ENoG was superior to grading score in terms of accuracy for prognosis prediction. The ROC curve revealed that more than 85% degeneration on ENoG had the best specificity (77.8%) and sensitivity (71.4%) to predict nonrecovery. When ENoG was subjected to the analysis of cumulative recovery rate using Kaplan-Meier plots, patients with more than 85% degeneration on ENoG had significantly poorer prognosis.

CONCLUSION:

ENoG was the most effective factor for prediction of the prognosis of peripheral facial palsy, and more than 85% degeneration had the best specificity and sensitivity to predict nonrecovery.

PMID:
21725266
DOI:
10.1097/MAO.0b013e31822558de
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Wolters Kluwer
Loading ...
Support Center