Abstract
BACKGROUND: To analyze the risk factors for benign paroxysmal positional vertigo (BPPV) and to construct a predictive nomogram model. METHODS: In this retrospective study, 312 participants were enrolled, including 164 BPPV patients and 148 healthy subjects without BPPV. Risk predictors for BPPV were identified using univariate and multivariate analyses, and a clinical nomogram was constructed. The predictive accuracy was assessed by unadjusted concordance index (C-index) and calibration plot. RESULTS: Univariate and multivariate regression analysis identified stroke (95% CI, 0.575-5.954; P=0.022), hyperlipidemia (95% CI, 0.471-4.647; P=0.003), chronic suppurative otitis media (95% CI, 1.222-45.528; P=0.005), cervical spondylosis (95% CI, 1.232-3.017; P=0.005), and osteoporosis (95% CI, 1.130-3.071; P=0.001) were the independent risk factors for BPPV. These risk factors were used to construct a clinical predictive nomogram. The regression equation was: logit (P) = -6.820 + 0.450 * stroke + hyperlipidemia * 0.312 + chronic suppurative otitis media * 0.499 + cervical spondylosis * 0.916 + osteoporosis * 0.628. The calibration curves demonstrated excellent accuracy of the predictive nomogram. Decision curve analysis showed that the predictive model is clinically applicable when the threshold probability was between 20% and 60%. CONCLUSIONS: Stroke, hyperlipidemia, chronic suppurative otitis media, cervical spondylosis and osteoporosis are independent risk predictors for BPPV. The developed nomogram is useful in predicting the risk of BPPV.