急性皮肤衰竭联合预测模型对脓毒症患者28天生存预后的判断价值

Clinical value of a predictive model based on acute skin failure in determining the 28-day prognosis of patients with sepsis

目的构建基于急性皮肤衰竭的联合预测模型,评价其对脓毒症患者28 d生存预后的预测价值,为医护人员制订有效干预措施提供依据。方法采用前瞻性调查方法,便利选取2020年5月至2023年4月上海中医药大学附属曙光医院收治的231例脓毒症患者,其中2020年5月至2022年3月的162例患者作为训练集构建预测模型,2022年4月至2023年4月的69例患者作为验证集进行外部验证。单因素和多因素Logistic回归分析脓毒症患者28 d死亡的危险因素,构建基于急性皮肤衰竭的联合预测模型并绘制列线图,验证其准确性。结果训练集162例患者中有53例死亡,病死率为32.7%;验证集69例患者中有19例死亡,病死率为27.5%,2组病死率比较差异无统计学意义(χ^(2)=0.61,P=0.437)。训练集中多因素分析显示,APACHEⅡ评分(OR=0.674,95%CI 0.509~0.631)、序贯性器官衰竭评分(OR=0.391,95%CI 0.242~0.631)、血乳酸值(OR=2.291,95%CI 1.306~4.019)、皮肤花斑评分(OR=2.950,95%CI 1.586~5.488)、皮肤湿冷(OR=3.678,95%CI 0.910~1.865)、毛细血管充盈时间>2 s(OR=6.070,95%CI 0.774~1.579)、指尖经皮血氧饱和度降低(OR=2.046,95%CI 1.312~2.076)、皮肤感觉减弱(OR=3.354,95%CI 0.796~1.124)是影响脓毒症患者28 d死亡的独立危险因素。列线图模型验证结果显示,急性皮肤衰竭联合预测模型在训练集与验证集的C-index分别为0.834、0.811,ROC曲线下面积分别为0.834、0.807。结论基于急性皮肤衰竭构建的列线图可预测脓毒症患者28 d死亡发生风险,能帮助医护人员及时采取有效的干预措施。

Objective To construct a predictive model based on acute skin failure,and to evaluate its predictive value on the 28-day prognosis of patients with sepsis,to provide a basis for medical staff to develop effective intervention measures.Methods A prospective survey method was adopted,231 patients with sepsis hospitalized in Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine from May 2020 to April 2023 were enrolled as the research subjects,of which 162 patients from May 2020 to March 2022 were allocated into the test group for construct a prediction model,and 69 patients from April 2022 to April 2023 in the validation group for external validation.Univariate and multivariate Logistic regression were implemented to analyze the risk factors of 28-day mortality in sepsis patients,construction of a joint prediction model based on acute skin failure,and drawing of a column chart to verify its accuracy.Results The 53 of 162 cases in the test group died,with mortality rate of 32.7%.The 19 of 69 cases in the validation group died,with mortality rate of 27.5%,there was no statistically significant difference in mortality rates between the two groups(χ^(2)=0.61,P=0.437).The results of multivariate analysis in the test group showed that APACHE II score(OR=0.674,95%CI 0.509-0.631),Sequential Organ Failure Assessment(OR=0.391,95%CI 0.242-0.631),lactate(OR=2.291,95%CI 1.306-4.019),skin mottling score(OR=2.950,95%CI 1.586-5.488),skin wet cold(OR=3.678,95%CI 0.910-1.865),capillary filling time>2 s(OR=6.070,95%CI 0.774-1.579),decreased fingertip transcutaneous oxygen saturation(OR=2.046,95%CI 1.312-2.076),and weakened skin sensation(OR=3.354,95%CI 0.796-1.124)were independent risk factors that affecting the 28-day mortality of patients with sepsis.The verification results of combined predictive model for acute skin failure showed that the C-index of test group and validation group were 0.834 and 0.811 respectively;the areas under ROC curve were 0.834 and 0.807,respectively.Conclusions Acute skin failure-based nomogram model can predict the 28-day mortality of patients with sepsis,and help medical staff to implement personalized intervention measures.