Scholarlink

<p>Hypernatremia is a common postoperative complication after pituitary adenoma surgery. Despite its prevalence, there is a significant gap in research regarding the construction of predictive models for assessing risk. Therefore, it is essential to create a robust model to accurately predict the risk of postoperative hypernatremia. To explore the risk factors for developing hypernatremia after pituitary adenoma surgery and establish a prediction model. From April 2022 to June 2024, 269 patients were admitted to the Department of Neurosurgery at Qilu Hospital as part of this cohort study. Initially, the sample was split into a training set and a validation set using a ratio of 3:1 randomly. Then, a univariable analysis was conducted for each variable in the training set. Subsequently, the Least Absolute Shrinkage and Selection Operator (LASSO) regression and stepwise regression were employed to identify the candidate variables for further examination. A multivariable logistic regression analysis was then performed to develop a risk prediction model. A nomogram was created to enhance the model’s utility, and the Hosmer-Lemeshow (HL) test was implemented to evaluate the model’s goodness of fit. The predictive efficiency of the model was assessed by calculating the area under the receiver operating characteristic curve (AUC), while model discrimination was determined using the calibration plot and decision curve analysis (DCA). This study indicated that the overall incidence of hypernatremia after pituitary surgery is 8.55%. We collected data on 46 potential risk factors. After LASSO regression, we constructed three models: LR model (LASSO regression), FR model (forward stepwise regression), and BR model (backward stepwise regression). The multivariable analysis identified that LR model, including pituitary stalk involvement, surgery time, and output 1st day, is the most effective at predicting postoperative hypernatremia. The HL test of LR model indicated a <i>p</i>-value of 0.448, and the AUC value was 0.813. In the test set, the AUC was 0.894. The calibration plots and DCA plots demonstrate substantial clinical relevance. Additionally, we developed a nomogram for predicting hypernatremia after pituitary surgery based on this regression model. The prediction model serves as a clinically valuable tool for identifying patients at heightened risk of postoperative hypernatremia following pituitary surgery, facilitating timely implementation of prophylactic management protocols to optimize clinical outcomes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Construction and validation of a LASSO penalized logistic regression model predicting hypernatremia after pituitary adenoma surgery

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月4日

查看全文
  • Xiao Liu,
  • Zhe Wang,
  • Yanli Zhang,
  • Xiaorong Luan

摘要

Hypernatremia is a common postoperative complication after pituitary adenoma surgery. Despite its prevalence, there is a significant gap in research regarding the construction of predictive models for assessing risk. Therefore, it is essential to create a robust model to accurately predict the risk of postoperative hypernatremia. To explore the risk factors for developing hypernatremia after pituitary adenoma surgery and establish a prediction model. From April 2022 to June 2024, 269 patients were admitted to the Department of Neurosurgery at Qilu Hospital as part of this cohort study. Initially, the sample was split into a training set and a validation set using a ratio of 3:1 randomly. Then, a univariable analysis was conducted for each variable in the training set. Subsequently, the Least Absolute Shrinkage and Selection Operator (LASSO) regression and stepwise regression were employed to identify the candidate variables for further examination. A multivariable logistic regression analysis was then performed to develop a risk prediction model. A nomogram was created to enhance the model’s utility, and the Hosmer-Lemeshow (HL) test was implemented to evaluate the model’s goodness of fit. The predictive efficiency of the model was assessed by calculating the area under the receiver operating characteristic curve (AUC), while model discrimination was determined using the calibration plot and decision curve analysis (DCA). This study indicated that the overall incidence of hypernatremia after pituitary surgery is 8.55%. We collected data on 46 potential risk factors. After LASSO regression, we constructed three models: LR model (LASSO regression), FR model (forward stepwise regression), and BR model (backward stepwise regression). The multivariable analysis identified that LR model, including pituitary stalk involvement, surgery time, and output 1st day, is the most effective at predicting postoperative hypernatremia. The HL test of LR model indicated a p-value of 0.448, and the AUC value was 0.813. In the test set, the AUC was 0.894. The calibration plots and DCA plots demonstrate substantial clinical relevance. Additionally, we developed a nomogram for predicting hypernatremia after pituitary surgery based on this regression model. The prediction model serves as a clinically valuable tool for identifying patients at heightened risk of postoperative hypernatremia following pituitary surgery, facilitating timely implementation of prophylactic management protocols to optimize clinical outcomes.