An internally validated prognostic model for success in revision stapes surgery for otosclerosis

Objectives/Hypothesis

To develop a prediction model that can accurately predict the chance of success following revision stapes surgery in patients with recurrent or persistent otosclerosis at 2- to 6-months follow-up and to validate this model internally.

Study Design

A retrospective cohort study of prospectively gathered data in a tertiary referral center.

Methods

The associations of 11 prognostic factors with treatment success were tested in 705 cases using multivariable logistic regression analysis with backward selection. Success was defined as a mean air-bone gap closure to 10 dB or less. The most relevant predictors were used to derive a clinical prediction rule to determine the probability of success. Internal validation by means of bootstrapping was performed. Model performance indices, including the Hosmer-Lemeshow test, the area under the receiver operating characteristics curve (AUC), and the explained variance were calculated.

Results

Success was achieved in 57.7% of cases at 2- to 6-months follow-up. Certain previous surgical techniques, primary causes of failure leading up to revision stapes surgery, and positions of the prosthesis placed during revision surgery were associated with higher success percentages. The clinical prediction rule performed moderately well in the original dataset (Hosmer-Lemeshow P = .78; AUC = 0.73; explained variance = 22%), which slightly decreased following internal validation by means of bootstrapping (AUC = 0.69; explained variance = 13%).

Conclusions

Our study established the importance of previous surgical technique, primary cause of failure, and type of the prosthesis placed during the revision surgery in predicting the probability of success following stapes surgery at 2- to 6-months follow-up.

Level of Evidence

2b. Laryngoscope, 2018

http://ift.tt/2G8znU5