Development and validation of a nomogram to predict prolonged residual dizziness lasting at least 7 days after benign...

This study developed and validated a nomogram to predict the risk of residual dizziness lasting ≥ 7 days in patients with benign paroxysmal positional vertigo (BPPV).

We retrospectively analyzed 907 BPPV patients, randomly divided 7:3 into training (n = 635) and validation (n = 272) sets. Using residual dizziness (≥ 7 days) as the outcome, predictors were selected via LASSO and multivariable logistic regression to construct and validate a nomogram.

According to findings from LASSO regression and logistic regression screening, older age [odds ratio (OR) = 1.054], diabetes mellitus (DM; OR = 5.564), and depression (OR = 9.070) were identified as independent risk factors, while vestibular rehabilitation training (OR = 0.197) was a protective factor. The model demonstrated strong predictive performance. Receiver operating characteristic (ROC) analysis showed that the training set had an area under the curve (AUC) of 0.836 [95% confidence interval (CI): 0.805-0.867], with an optimal threshold of 0.589, sensitivity of 0.682, and specificity of 0.873; the validation set had an AUC of 0.802 (95% CI: 0.750-0.853), with an optimal threshold of 0.588, sensitivity of 0.643, and specificity of 0.907. Calibration curves showed good agreement between predicted and actual risks, and decision curve analysis confirmed its clinical utility across a wide risk threshold range.

The nomogram model, constructed using age, DM, depression, and vestibular rehabilitation training, is a practical tool for identifying BPPV patients at high risk for prolonged residual dizziness.