Herwanto Velma, Shetty Amith, Nalos Marek, Chakraborty Mandira, McLean Anthony, Eslick Guy D, Tang Benjamin
Department of Intensive Care Medicine, Nepean Hospital, Sydney, NSW, Australia.
Centre for Immunology and Allergy Research, the Westmead Institute for Medical Research, Sydney, NSW, Australia.
Crit Care Explor. 2019 Sep 17;1(9):e0043. doi: 10.1097/CCE.0000000000000043. eCollection 2019 Sep.
We performed a meta-analysis to assess whether the newly introduced quick Sequential Organ Failure Assessment score could predict sepsis outcomes and compared its performance to systematic inflammatory response syndrome, the previously widely used screening criteria for sepsis.
We searched multiple electronic databases including MEDLINE, the Cochrane Library, Embase, Web of Science, and Google Scholar (up to March 1, 2019) that evaluated quick Sequential Organ Failure Assessment score, systemic inflammatory response syndrome, or both (International Prospective Register of Systematic Reviews [PROSPERO]: CRD42018103327).
Studies were included if the outcome was mortality, organ dysfunction, admission to ICU, ventilatory support, or prolonged ICU stay and if prediction performance was reported as either area under the curve, odds ratio, sensitivity, or specificity.
The criterion validity of the quick Sequential Organ Failure Assessment score and systemic inflammatory response syndrome criteria were assessed by measuring its predictive validity for primary (mortality) and secondary outcomes in pooled metrics as mentioned. The data were analyzed using random effects model, and heterogeneity was explored using prespecified subgroups analyses.
We screened 1,340 studies, of which 121 studies (including data for 1,716,017 individuals) were analyzed. For mortality prediction, the pooled area under the curve was higher for quick Sequential Organ Failure Assessment score (0.702; 95% CI, 0.685-0.718; = 99.41%; < 0.001) than for systemic inflammatory response syndrome (0.607; 95% CI, 0.589-0.624; = 96.49%; < 0.001). Quick Sequential Organ Failure Assessment score consistently outperformed systemic inflammatory response syndrome across all subgroup analyses (area under the curve of quick Sequential Organ Failure Assessment vs. area under the curve of systemic inflammatory response syndrome < 0.001), including patient populations (emergency department vs ICU), study design (retrospective vs prospective), and countries (developed vs resource-limited). Quick Sequential Organ Failure Assessment score was more specific (specificity, 74.58%; 95% CI, 73.55-75.61%) than systemic inflammatory response syndrome (specificity, 35.24%; 95% CI, 22.80-47.69%) but less sensitive (56.39%; 95% CI, 50.52-62.27%) than systemic inflammatory response syndrome (78.84%; 95% CI, 74.48-83.19%).
Overall, quick Sequential Organ Failure Assessment score outperforms systemic inflammatory response syndrome in predicting sepsis outcome, but quick Sequential Organ Failure Assessment score has relative strengths/weaknesses (more specific but less sensitive) compared with systemic inflammatory response syndrome.
我们进行了一项荟萃分析，以评估新引入的快速序贯器官衰竭评估（qSOFA）评分是否能预测脓毒症的预后，并将其性能与系统性炎症反应综合征（SIRS）进行比较，SIRS是先前广泛使用的脓毒症筛查标准。
我们检索了多个电子数据库，包括MEDLINE、Cochrane图书馆、Embase、科学网和谷歌学术（截至2019年3月1日），这些数据库评估了快速序贯器官衰竭评估评分、系统性炎症反应综合征或两者（国际前瞻性系统评价注册库[PROSPERO]：CRD42018103327）。
如果研究结果是死亡率、器官功能障碍、入住重症监护病房（ICU）、通气支持或ICU住院时间延长，并且预测性能报告为曲线下面积、比值比、敏感性或特异性，则纳入该研究。
通过测量快速序贯器官衰竭评估评分和系统性炎症反应综合征标准对上述汇总指标中主要（死亡率）和次要结局的预测有效性，评估其标准效度。使用随机效应模型分析数据，并使用预先指定的亚组分析探讨异质性。
我们筛选了1340项研究，其中121项研究（包括1716017名个体的数据）进行了分析。对于死亡率预测，快速序贯器官衰竭评估评分的汇总曲线下面积（0.702；95%可信区间，0.685 - 0.718；I² = 99.41%；P < 0.001）高于系统性炎症反应综合征（0.607；95%可信区间，0.589 - 0.624；I² = 96.49%；P < 0.001）。在所有亚组分析中，快速序贯器官衰竭评估评分始终优于系统性炎症反应综合征（快速序贯器官衰竭评估评分的曲线下面积与系统性炎症反应综合征的曲线下面积比较，P < 0.001），包括患者群体（急诊科与ICU）、研究设计（回顾性与前瞻性）和国家（发达国家与资源有限国家）。快速序贯器官衰竭评估评分比系统性炎症反应综合征更具特异性（特异性为74.58%；95%可信区间，73.55 - 75.61%），但敏感性（56.39%；95%可信区间，50.52 - 62.27%）低于系统性炎症反应综合征（78.84%；95%可信区间，74.48 - 83.19%）。
总体而言，快速序贯器官衰竭评估评分在预测脓毒症预后方面优于系统性炎症反应综合征，但与系统性炎症反应综合征相比，快速序贯器官衰竭评估评分有相对的优势/劣势（更具特异性但敏感性较低）。
