The Impact of Delayed Transition From Noninvasive to Invasive Mechanical Ventilation on Hospital Mortality in Immunocompromised Patients With Sepsis
Yang Xu,Yi-Fan Wang,Yi-Wei Liu,Run Dong,Yan Chen,Yi Wang,Li Weng,Bin Du,China Critical Care Clinical Trials Group (CCCCTG) and China National Critical Care Quality Control Center Group,Mingyan Zhao,Songlin Yang,Xiaochun Ma,Yini Sun,Yongjie Yin,Dexin Liu,Zhenjie Hu,Yan Huo,Li Jiang,Qi Zhang,Youzhong An,Huiying Zhao,Yuan Xu,Wei He,Dawei Wu,Chen Li,Enqiang Mao,Cheng Zhu,Tiehe Qin,Shouhong Wang,Rongqing Sun,Yuexia Li,Yuhang Ai,Yin Huang,Jiandong Lin,Xiongjian Xiao,Zhenyang He,Rui Li,Yan Kang,Xuelian Liao,Shusheng Li,Xiao Ran,Xiangyou Yu,Lihua Zhou,Lipeng Zhang,Xijing Zhang,Binxiao Su,Fachun Zhou,Fang Xu,Renhua Su,Qian Li,Chuanyun Qian,Wei Zhang,Xiaoyun Hu,Xiangyuan Cao,Xigang Ma,Qingyuan Zhan,Xu Huang,Ming Hou,Xiang Zhou,Yi Yang,Linjun Wan,Zhen Wang,Wei Fang,Yongjun Liu,Jian Liu,Jing Yan,Xiaojun Yang,Feng Shen,Lina Zhang,Erzhen Chen,Weidong Wu,Meili Duan,Bingyu Qin,Kejian Qian,Lei Xu,Bin Xiong,Longxiang Su,Congshan Yang
DOI: https://doi.org/10.1097/CCM.0000000000006400
2024-08-21
Abstract:Objective: To determine whether mortality differed between initial invasive mechanical ventilation (IMV) or noninvasive ventilation (NIV) followed by delayed IMV in immunocompromised patients with sepsis. Design: Retrospective analysis using the National Data Center for Medical Service claims data in China from 2017 to 2019. Setting: A total of 3530 hospitals across China. Patients: A total of 36,187 adult immunocompromised patients with sepsis requiring ventilation. Interventions: None. Measurements and main results: The primary outcome was hospital mortality. Patients were categorized into NIV initiation or IMV initiation groups based on first ventilation. NIV patients were further divided by time to IMV transition: no transition, immediate (≤ 1 d), early (2-3 d), delayed (4-7 d), or late (≥ 8 d). Mortality was compared between groups using weighted Cox models. Over the median 9-day follow-up, mortality was similar for initial NIV versus IMV (adjusted hazard ratio [HR] 1.006; 95% CI, 0.959-1.055). However, among NIV patients, a longer time to IMV transition is associated with stepwise increases in mortality, from immediate transition (HR 1.65) to late transition (HR 2.51), compared with initial IMV. This dose-response relationship persisted across subgroups and sensitivity analyses. Conclusions: Prolonged NIV trial before delayed IMV transition is associated with higher mortality in immunocompromised sepsis patients ultimately intubated.