Monk Muhammad, Torres Jordan, Vickery Kimberly, Jayaraman Gnananandh, Sarva Siva T, Kesavan Ramesh
Internal Medicine, HCA Houston Kingwood/University of Houston College of Medicine, Kingwood, USA.
Internal Medicine, Univeristy of Houston/HCA Healthcare Kingwood, Houston, USA.
Cureus. 2023 Feb 24;15(2):e35423. doi: 10.7759/cureus.35423. eCollection 2023 Feb.
Background Over the past three years, COVID-19 has been a major source of mortality in intensive care units around the world. Many scoring systems have been developed to estimate mortality in critically ill patients. Our intent with this study was to compare the efficacy of these systems when applied to COVID-19. Methods The was a multicenter, retrospective cohort study of critically ill patients with COVID-19 admitted to 16 hospitals in Texas from February 2020 to March 2022. The Simplified Acute Physiology Score (SAPS) II, Acute Physiology and Chronic Health Evaluation (APACHE) II, Sequential Organ Failure Assessment (SOFA) score, and 4C Mortality scores were calculated on the initial day of ICU admission. Primary endpoints were all-cause mortality, ICU length of stay, and hospital length of stay. Results Initially, 62,881 patient encounters were assessed, and the cohort of 292 was selected based on the inclusion of the requisite values for each of the scoring systems. The median age was 56 +/- 14.93 years and 61% of patients were male. Mortality was defined as patients who expired or were discharged to hospice and was 78%. The different scoring systems were compared using logistic regression, receiver operating characteristic (ROC) curve, and area under the ROC curve (AUC) analysis to compare the accuracy of prediction of the mortality and length of stay. The multivariate analysis showed that SOFA, APACHE II, SAPS II, and 4C scores were all significant predictors of mortality. The SOFA score had the highest AUC, though the confidence intervals for all of the models overlap therefore one model could not be considered superior to any of the others. Linear regression was performed to evaluate the models' ability to predict ICU and hospital length of stay, and none of the tested systems were found to be significant predictors of length of stay. Conclusion The SOFA, APACHE II, ISARIC 4-C, and SAPS II scores all accurately predicted mortality in critically ill patients with COVID-19. The SOFA score trended to perform the best.
背景 在过去三年中，新冠病毒病一直是全球重症监护病房死亡的主要原因。已经开发了许多评分系统来估计危重症患者的死亡率。我们进行这项研究的目的是比较这些系统应用于新冠病毒病时的效果。方法 这是一项对2020年2月至2022年3月在德克萨斯州16家医院收治的新冠病毒病危重症患者进行的多中心回顾性队列研究。在重症监护病房入院首日计算简化急性生理学评分（SAPS）II、急性生理学与慢性健康状况评价（APACHE）II、序贯器官衰竭评估（SOFA）评分和4C死亡评分。主要终点为全因死亡率、重症监护病房住院时长和住院时长。结果 最初，评估了62881例患者就诊情况，并根据每个评分系统所需值的纳入情况选择了292例患者作为队列。中位年龄为56±14.93岁，61%的患者为男性。死亡率定义为死亡或出院至临终关怀机构的患者，为78%。使用逻辑回归、受试者工作特征（ROC）曲线和ROC曲线下面积（AUC）分析比较不同评分系统，以比较死亡率和住院时长预测的准确性。多变量分析显示，SOFA、APACHE II、SAPS II和4C评分均为死亡率的显著预测因素。SOFA评分的AUC最高，不过所有模型的置信区间重叠，因此不能认为一个模型优于其他任何模型。进行线性回归以评估模型预测重症监护病房和住院时长的能力，未发现任何测试系统是住院时长的显著预测因素。结论 SOFA、APACHE II、国际严重急性呼吸道感染协作组（ISARIC）4-C和SAPS II评分均准确预测了新冠病毒病危重症患者的死亡率。SOFA评分表现最佳。
