Nomogram and scoring models based on BMI z-score, serum chlorine and urine specific gravity levels for predicting oral rehydration solution efficacy in pediatric postural orthostatic tachycardia syndrome.

Background: Oral rehydration solution (ORS) is predominantly utilized in the management of hypovolemic postural orthostatic tachycardia syndrome (POTS). This study aimed to identify effective indicators and develop models to assess the impact of ORS on pediatric patients diagnosed with POTS.

Methods: We utilized a retrospective analysis of totally 158 pediatric patients with POTS receiving a 3-month course of ORS treatment. All patients were classified into training set (n = 98) and validation set (n = 60). Within the training set, univariate analysis and binary logistic regression were employed to select candidate predictors. To predict the efficacy of ORS in pediatric patients with POTS, a nomogram model and a scoring model were constructed and demonstrated. Additionally, the predictive ability and calibration performance were evaluated using receiver operating characteristic (ROC) curves, Hosmer-Lemeshow (H-L) goodness of fit test, and calibration plots. Decision curve analysis (DCA) was employed to assess the clinical applicability of the predictive models.

Results: Body mass index (BMI) z-score, serum chlorine, and urine specific gravity (USG) before treatment were identified as significant and independent predictors of efficacy of ORS in pediatric patients with POTS. Consequently, these indicators were included in the predictive models. A nomogram model was constructed in the training set (AUC = 0.87, which yields a sensitivity of 84.5% and a specificity of 85.0%) and validated in the validation set (the sensitivity, specificity, and accuracy were 87.5%, 85%, and 86.7%, respectively). A scoring model was advanced in the training set (AUC = 0.88, which yields a sensitivity of 79.3% and a specificity of 82.5%) and validated in the validation set (the sensitivity, specificity, and accuracy were 77.5%, 80.0%, and 78.3%, respectively). The H-L test results indicated a good model fit. The calibration plots and DCA for both models exhibited excellent calibration and satisfactory net benefit.

Conclusions: Based on pre-treatment BMI z-score, serum chlorine, and USG, a nomogram model and a scoring model were developed and validated. The models can effectively assess the efficacy of ORS in pediatric patients with POTS, offering an accurate and individualized therapeutic strategy.