急診急性呼吸衰竭插管病人長期呼吸器依賴及死亡之風險因子及預測評分系統

背景於急診接受氣管內管置放後機械通氣常常為呼吸衰竭及重症病人不可或缺的急救處置之一，經插管處置後，若病人長期無法脫離呼吸器，除增加死亡機率，接受氣切手術並改以氣切管連接呼吸器通氣為常見處置。過往病人困難脫離呼吸器之預測指標及評分多於插管且入住加護病房後取得，本研究希望藉由收集於急診接受插管病人當次就診之數據，提供病人插管前各項參數，做為預測未來困難脫離呼吸器或死亡之預測因子及建立相關評分系統。設計本研究為臺灣中部單一醫學中心之回溯性觀察型研究，收集2016年至2019年四年間於急診插管且符合年滿18歲、非外傷或預防性插管病人，且排除於急診接受心肺復甦之病人，紀錄當次急診就診臨床表徵與實驗檢查數據，將所收集個案隨機分為分析組與驗證組。於分析組對照未來14天內困難脫離呼吸器或於當次住院期間死亡之資料，以logistic regression model 取得預測風險因子。經選取適當風險因子後，設計評分系統並將該評分系統運用於驗證組，以確認其預測效力。結果本研究共收集1883位急性呼吸衰竭個案，其中947位分於分析組，936位分於驗證組，兩組成功於14天內脫離呼吸器且存活出院病人數分別為522人及498人，經過於分析組中研究，選取血小板較低 (< 200x103/μL)、白蛋白較低 (<3 g/dL)、血中二氧化碳濃度 (PaCO2)較低 (<35 mmHg)、BMI較低 (<25 kg/m2)、收縮壓較低 (<130 mmHg)、CRP較高 (>6 mg/dL)、乳酸值較高 (>28 mg/dL)、呼吸速率較慢 (< 22/min)、意識較差 (GCS <9)、不具有肺部相關疾病 (包含COPD、限制型肺病及氣喘)及有急性腎損傷之11項不利14天內成功脫離呼吸器且存活出院之指標後加總計分成預測模組。該預測模組以AUC呈現之預測表現力於分析組及驗證組分別為0.642及0.632 (95% CI: 0.607-0.677及0.597-0.668)。結論及建議 由本研究所選取之11項風險因子，為目前首篇於急診收集數據以預測未來困難脫離呼吸器及死亡之研究，旨於以最接近插管處置之前統計資料設計之評分系統，未來急診病人在插管前，做為事先預測未來困難脫離呼吸器相關風險之目的。未來可考慮以其做為預測困難脫離呼吸器之依據或以本研究提出之其他相關因子進一步研究設計，發展更具有預測效力之評分系統。

BackgroundEndotracheal intubation with mechanical ventilation is often an essential emergency intervention for patients with respiratory failure and critical conditions in the emergency department. If a patient is unable to wean off the ventilator for an extended period following intubation, it not only increases the risk of mortality but also frequently leads to the need for a tracheostomy. Historically, predictive indicators and scoring systems for difficult weaning from mechanical ventilation have been obtained after intubation and ICU admission. This study aims to collect data from patients intubated in the emergency department during their current visit, providing predictive factors and a scoring system to foresee future difficulties in weaning off the ventilator or mortality before intubation.DesignThis is a retrospective observational study conducted at a single medical center in central Taiwan. It collected clinical manifestations and laboratory data from patients who were intubated in the emergency department and met the inclusion criteria between 2016 and 2019,who were at least 18 years old, and who were non-traumatic or non-prophylactic intubations, excluding those who underwent cardiopulmonary resuscitation in the emergency department. The collected cases were randomly divided into an analysis group and a validation group. In the analysis group, data on difficult weaning from the ventilator within 14 days or mortality during the current hospitalization were compared, and predictive risk factors were identified using a logistic regression model. After selecting the appropriate risk factors, a scoring system was developed and applied to the validation group to confirm its predictive efficacy.ResultsA total of 1883 cases of acute respiratory failure were collected, with 947 in the analysis group and 936 in the validation group. The number of patients who successfully weaned from mechanical ventilation within 14 days and survived to discharge was 522 in the analysis group and 498 in the validation group.The study identified 11 adverse indicators for successful weaning from the ventilator and survival to discharge within 14 days: lower platelet count (<200x103/μL), lower albumin (<3 g/dL), lower blood carbon dioxide concentration (PaCO2 <35 mmHg), lower BMI (<25 kg/m2), lower systolic blood pressure (< 130 mmHg), higher CRP (>6 mg/dL), higher lactate (>28 mg/dL), slower respiratory rate (<22/min), poorer consciousness (GCS <9), absence of lung-related diseases (including COPD, restrictive lung disease, and asthma), and the presence of acute kidney injury. These indicators were aggregated into a predictive model. The predictive performance of this system, presented as AUC, was 0.642 (95% CI: 0.607-0.677) for the analysis group and 0.632 (95% CI: 0.597-0.668) for the validation group.Conclusion and RecommendationsThis study identified 11 risk factors as the first research to collect data in the emergency department to predict future difficulties in weaning from the ventilator and mortality. The scoring system, designed using statistical data closest to the intubation procedure, aims to preemptively predict future related risks before intubation. Future research could consider using this system as a predictive basis or further refine the scoring system with additional related factors proposed in this study for better predictive efficacy.