Onodera Yoko, Ishiguro Takashi, Uozumi Ryuji, Isono Taisuke, Nishida Takashi, Kobayashi Yoichi, Takaku Yotaro
Department of Respiratory Medicine, Saitama Cardiovascular and Respiratory Center, Japan.
Department of Industrial Engineering and Economics, Tokyo Institute of Technology, Japan.
Intern Med. 2025 Aug 1;64(15):2276-2283. doi: 10.2169/internalmedicine.4727-24. Epub 2025 Jan 3.
Objective Community-acquired pneumonia is an acute infectious disease with potentialy life-threatening consequences. Because invasive mechanical ventilation (IMV) requires the attention of many medical staff, early risk prediction at the time of admission is expected to lead to a predictable course of patient care and the appropriate allocation of medical resources. There are a limited number of reports on predictive factors for IMV, such as SMART-COP. Therefore, further studies are required. Methods We retrospectively reviewed cases of patients with community-acquired pneumonia other than coronavirus disease 2019 admitted to our institution from 2002 to 2019. We performed competing risks analysis with the need for IMV from the day after admission as the outcome and used multivariable analysis to identify predictive factors of IMV from admission characteristics. Results Among 2,227 patients (mean age 67.3 years, 69.0% male), 39 patients required IMV on or after the day following admission. A multivariable analysis showed that predictive factors of IMV were respiratory rate >30 breaths/min [subdistribution hazard ratio (SHR), 5.53; 95% confidence interval (CI), 2.09 to 14.67; p=0.001], PaO/FiO ratio <250 (SHR, 8.02; 95% CI, 2.78 to 23.13; p<0.001), and Legionella pneumonia (SHR, 4.87; 95% CI, 1.56 to 15.13; p=0.006). Conclusion This study revealed that among other factors including mainly vital signs, specific infection by a microorganism itself (Legionella in this study) was a predictive factor for the need of IMV.
目的 社区获得性肺炎是一种可能危及生命的急性传染病。由于有创机械通气（IMV）需要众多医护人员的关注，因此入院时的早期风险预测有望带来可预测的患者护理过程及医疗资源的合理分配。关于IMV预测因素的报道有限，如SMART-COP。因此，需要进一步研究。方法 我们回顾性分析了2002年至2019年在我院住院的非2019冠状病毒病社区获得性肺炎患者的病例。我们以入院次日起对IMV的需求为结局进行竞争风险分析，并使用多变量分析从入院特征中识别IMV的预测因素。结果 在2227例患者（平均年龄67.3岁，男性占69.0%）中，39例患者在入院次日或之后需要IMV。多变量分析显示，IMV的预测因素为呼吸频率>30次/分钟[亚分布风险比（SHR），5.53；95%置信区间（CI），2.09至14.67；p=0.001]、动脉血氧分压/吸入氧浓度比值<250（SHR，8.02；95%CI，2.78至23.13；p<0.001）和军团菌肺炎（SHR，4.87；95%CI，1.56至15.13；p=0.006）。结论 本研究表明，在主要包括生命体征等其他因素中，微生物本身的特定感染（本研究中的军团菌）是需要IMV的预测因素。
