Development and Validation of a Clinical Risk Model to Predict the Hospital Mortality in Ventilated Patients With Acute Respiratory Distress Syndrome

A Population-Based Study

Weiyan Ye; Rujian Li; Hanwen Liang; Yongbo Huang; Yonghao Xu; Yuchong Li; Limin Ou; Pu Mao; Xiaoqing Liu; Yimin Li


BMC Pulm Med. 2022;22(268) 

In This Article


Acute respiratory distress syndrome (ARDS) is an acute diffuse inflammatory lung injury, featuring increased pulmonary vascular permeability and lung weight mechanistically, and clinically hypoxemia and bilateral radiographic opacities.[1] Although numerous interventions, for example, low tidal volume ventilation[2] and conservative fluid strategies,[3] have been employed in the management of ARDS, whose hospital mortality remained significantly high in critical illness patients in intensive care unit (ICU), especially that of invasive ventilation.[4] In addition, the large variability in mortality exists in ARDS, which was thought to contribute to indeterminate or conflicting study results in most randomized clinical trials in patients with ARDS.[5]

As risk stratification for ARDS would aid in medical decision making and clinical trial design, lots of effort have been made to develop a model of predicting ARDS-induced/related mortality.[6–8] Sequential Organ Failure Assessment (SOFA)[9] was initially designed for assessing organ dysfunction/failure over time, yet it's also widely adopted for clinical outcome prognostication of critical ill patients and those with ARDS.[10–12] Other scoring systems, for instance, Simplified Acute Physiology Score II (SAPS II),[13] Acute Physiology and Chronic Health Evaluation IV (APACHE IV) score[14] and Oxford Acute Severity of Illness Score (OASIS),[15] not intentionally focused on ARDS patients though, have been associated with patient outcomes. However, these scoring systems failed to provide consistent and accurate predictive estimates of the risk of death in patient populations with a specific disease process. In addition, some of the models above require laborious data collection and not easily to be calculated at the bedside. A few previous studies attempted to establish a prognostic model by integrating some predictors of death in ARDS,[6–8,16] but whose predictive power remain controversial. Most of those studies developed their models based on the patients participating in clinical trials and receiving specific treatments, resulting in failure of external validation in real-world patients.[17]

Currently, there's no reliable predictive model available for ARDS patients' hospital outcomes based on data straight from the bedside and the patients' actual condition. The primary aim of this study is to develop a clinician-friendly prognostic model incorporating variables that may be relevant to ARDS prognosis and that clinicians could routinely collect and easily calculate to predict risk of in-hospital death in ARDS patients with invasive mechanical ventilation (IMV).