- 學位論文
基於血小板數目的心因性休克病人30天死亡率的預後預測模型 - 臺灣的單一機構研究
A prognostic prediction model based on platelet count for 30-day mortality in patients with cardiogenic shock - A single-institution study in Taiwan
摘要
背景 心因性休克是高死亡率的醫學挑戰之一，除了多重複雜原因外，其嚴重度範圍可以從迅速改善的低血壓到不可逆轉的多重器官衰竭。風險分級(Risk stratification)在定義研究族群及臨床處置決策上具有重要性。CardShock是最多被引用的心因性休克風險評分模組之一，然而在一項歐洲的外部驗證研究上(External validation)，CardShock評分對於非冠心症導致休克的預測表現並不顯著(AUC of ACS vs. Non-ACS: 0.742 vs. 0.648)。本研究關於當代心因性休克族群，研究目的為1. 臨床特徵描述 2. 風險因子識別 3. 建立預測評分模型。 設計 於台灣中部單一醫學中心的回顧性、觀察型研究，研究期間2014年至2019年，共6年。蒐集到急診呈現心因性休克的病人其人口學特徵、臨床表現、心臟超音波參數、心導管及實驗室數據，以Cox比例風險模型做存活分析，取得獨立的風險預測因子，連續變項由Youden index找出分界點(Cutoff point)，根據其風險比(Hazard ratio)建立預測模組。預後結果(Outcome)定義為30天死亡率(All-cause mortality)。 結果 共納入225位急診心因性休克病人，其中因冠心症引發休克有107位(47.6%)，冠心症病人接受再灌流的比率為91.59%，最終49位病人於30天內死亡(死亡率21.8%)，有53名病人(23.55%)呈現血小板低下(血小板數目<150x103/μL)，血小板數目與30天死亡風險呈現顯著負相關，即使在正常範圍內(血小板數目150-400x103/μL)，經多變項Cox模型分析後，得到4個風險預測因子: Age >71years (HR 2.452, 95% CI 1.327-4.531, p = 0.004)、LVEF<40% (HR 2.613, 95% CI 1.020-6.692, p = 0.045)、Lactate >24mg/dl (HR 1.967, 95% CI 1.069-3.620, p = 0.030)、PLT<200x103/µL (HR 2.574, 95% CI 1.379-4.805, p = 0.003)。以此建立預測模組，積分最高5分，預測表現以AUC評估(All: 0.774, ACS: 0.781, Non-ACS: 0.759)。 結論 由Age、LVEF、Lactate、Platelet count所組成的預測模組在整體心因性休克族群及ACS與Non-ACS次族群對30天死亡率均有良好的預測表現，不須導管室參數，可以應用於急診早期預後評估及治療決策參考。無論是否因冠心缺血引發，心因性休克於急診呈現血小板低下與30天死亡率的增加存在顯著相關性。追蹤血小板數目趨勢作為心因性休克病程中代償不良的生物標記，其可行性值得進一步評估。
並列摘要
Background Cardiogenic shock (CS) is one of the medical challenges with a high mortality rate. In addition to multiple complex etiologies, its severity spectrum can range from rapidly improving hypotension to irreversible multiple organ failure. CardShock is one of the most cited risk scoring models in CS. However, in a European external validation study, the predictive performance of CardShock scores for non-coronary shock was suboptimal (AUC of ACS vs. Non-ACS: 0.742 vs. 0.648). The objectives of this study in contemporary Asian patients with cardiogenic shock were to describe clinical characteristics, identify risk factors and develop a predictive scoring model. Design A retrospective observational 6-year study of a single medical center in central Taiwan was conducted from 2014 to 2019 to collect demographic characteristics, clinical presentation, echocardiographic parameters, angiographic and laboratory data of patients presenting with cardiac shock in the emergency department. The cutoff point was identified by the Youden index, and the prediction model was built based on the hazard ratio in multivariate Cox survival analysis. The outcome was defined as 30-day all-cause mortality. Results A total of 225 patients with cardiogenic shock were included in this study. 107 (47.6%) patients were in shock status related to acute coronary syndrome and 98 (91.59%) patients received reperfusion therapy. 49 patients eventually died within 30 days (21.8% of all-cause mortality rate). 53 patients (23.55%) presented with thrombocytopenia (Platelet counts<150x103/μL) and platelet counts were significantly negatively correlated with the risk of death within 30 days, even in the normal range (Platelet counts of 150-400x103/μL). According to the results of multivariate Cox survival analysis, four independent risk predictors were obtained: Age >71years (HR 2.452, 95% CI 1.327-4.531, p = 0.004)、LVEF <40% (HR 2.613, 95% CI 1.020-6.692, p = 0.045)、Lactate >24mg/dl (HR 1.967, 95% CI 1.069-3.620, p = 0.030)、PLT <200x103/µL (HR 2.574, 95% CI 1.379-4.805, p = 0.003). The final score of the prediction model has a maximum of 5 points, and the discrimination ability was assessed by the area under the receiver operating characteristic curve (AUC for All: 0.774; ACS: 0.781; Non-ACS: 0.759). Conclusion The prediction model, consisting of age, LVEF, lactate, and platelet counts, has a good predictive performance for 30-day all-cause mortality in the overall cohort, both ACS and Non-ACS subgroups, and it can be used for early prognostic assessment and treatment decision-making. A significantly negative correlation was observed between platelet counts and 30-day mortality regardless of causes related to ACS or Non-ACS. The trends of serial platelet count utilized as a clinical marker of decompensation in cardiogenic shock warrant further evaluation.