High Altitude Medical Research Institute, Tibet Autonomous Region People's Hospital, Lhasa, China.
High Altitude Medical Research Institute, Tibet Autonomous Region People's Hospital, Lhasa, China
BMJ Open. 2023 Nov 3;13(11):e074161. doi: 10.1136/bmjopen-2023-074161.
To develop the first prediction model based on the common clinical symptoms of high-altitude pulmonary edema (HAPE), enabling early identification and an easy-to-execute self-risk prediction tool.
A total of 614 patients who consulted People's Hospital of Tibet Autonomous Region between January 2014 and April 2022 were enrolled. Out of those, 508 patients (416 males and 92 females) were diagnosed with HAPE and 106 were patients without HAPE (33 females and 72 males). They were randomly distributed into training (n=431) and validation (n=182) groups. Univariate and multivariate analysis were used to screen predictors of HAPE selected from the 36 predictors; nomograms were established based on the results of multivariate analysis. The receiver operating characteristic curve (ROC) was developed to obtain the area under the ROC curve (AUC) of the predictive model, and its predictive power was further evaluated by calibrating the curve, while the Decision Curve Analysis (DCA) was developed to evaluate the clinical applicability of the model, which was visualised by nomogram.
All six predictors were significantly associated with the incidence of HAPE, and two models were classified according to whether the value of SpO (percentage of oxygen in the blood) was available in the target population. Both could accurately predict the risk of HAPE. In the validation cohort, the AUC of model 1 was 0.934 with 95% CI (0.848 to 1.000), and model 2 had an AUC of 0.889, 95% CI (0.779 to 0.999). Calibration plots showed that the predicted and actual HAPE probabilities fitted well with internal validation, and the clinical decision curve shows intervention in the risk range of 0.01-0.98, resulting in a net benefit of nearly 99%.
The recommended prediction model (nomogram) could estimate the risk of HAPE with good precision, high discrimination and possible clinical applications for patients with HAPE. More importantly, it is an easy-to-execute scoring tool for individuals without medical professionals' support.
开发首个基于高原肺水肿（HAPE）常见临床症状的预测模型，以便能够进行早期识别和易于执行的自我风险预测工具。
共纳入 2014 年 1 月至 2022 年 4 月期间在西藏自治区人民医院就诊的 614 名患者。其中，508 名患者（男 416 例，女 92 例）被诊断为 HAPE，106 名患者无 HAPE（女 33 例，男 72 例）。他们被随机分配到训练（n=431）和验证（n=182）组。使用单变量和多变量分析从 36 个预测因素中筛选出 HAPE 的预测因素；基于多变量分析的结果建立了列线图。开发了受试者工作特征曲线（ROC）以获得预测模型的 ROC 曲线下面积（AUC），并通过校准曲线进一步评估其预测能力，同时开发决策曲线分析（DCA）以评估模型的临床适用性，通过列线图进行可视化。
所有六个预测因素均与 HAPE 的发生率显著相关，并根据目标人群中 SpO（血氧百分比）值的可用性对两种模型进行分类。两种模型都能准确预测 HAPE 的风险。在验证队列中，模型 1 的 AUC 为 0.934，95%CI（0.848 至 1.000），模型 2 的 AUC 为 0.889，95%CI（0.779 至 0.999）。校准图显示，内部验证中预测和实际 HAPE 概率拟合良好，临床决策曲线显示在 0.01-0.98 的风险范围内进行干预，可获得近 99%的净收益。
推荐的预测模型（列线图）可以准确、高区分度地估计 HAPE 的风险，并可能具有临床应用价值，适用于 HAPE 患者。更重要的是，它是一种易于执行的评分工具，适用于没有医疗专业人员支持的个体。
