Tafelski S, Nachtigall I, Stengel S, Wernecke K, Spies C
Department of Anesthesiology and Intensive Care, Charité-Universitaetsmedizin Berlin, Campus Charité Mitte and Campus Virchow-Klinikum, Berlin, Germany -
Minerva Anestesiol. 2015 Mar;81(3):264-71. Epub 2014 Sep 15.
Sepsis remains one of the most challenging burdens of critically ill patients. But for interventional studies significant heterogeneity remains in classifying patients. PIRO (Predisposition, Response, Infection and Organ dysfunction) has been introduced as innovative option for improved patient characterization. Aim of this study was to evaluate precision to predict hospital mortality of three different proposed PIRO classification systems.
Data were prospectively obtained data in five ICUs in a university hospital in Berlin, Germany including adult patients with sepsis. Three different scoring systems were compared using patient characteristics to classify the population according to all cause hospital mortality risk (Moreno- PIRO, Rubulotta- PIRO and Howell-PIRO).
Two-hundred-seventy-eight sepsis patients were included and reclassified using different PIRO models. All cause hospital mortality was 16.2%. Hospital mortality increased with higher PIRO scores with odds ratios of 1.070 (95% CI 1.041-1.100) for Moreno-PIRO, 1.282 (95% CI 1.079-1.524) for Rubulotta-PIRO and 1.256 (95% CI 1.146-1.367) for Howell-PIRO. Area under the curves for Moreno-PIRO was 0.743 (95% CI: 0.687-0.793), for Rubulotta-PIRO 0.646 (95% CI: 0.587-0.702) and for Howell-PIRO 0.751 (95% CI: 0.696-0.801). Moreno-PIRO and Howell-PIRO were statistically different compared with Rubulotta-PIRO (P=0.046 and P=0.035).
Proposed PIRO classifications demonstrated slight differences between models without prioritization of one approach and all seemed feasible for patient classification. Future PIRO-development is needed to straighten predisposition, infection, and especially the response category.
脓毒症仍然是重症患者面临的最具挑战性的负担之一。但在干预性研究中，患者分类存在显著的异质性。PIRO（易感性、反应性、感染和器官功能障碍）已被引入作为改善患者特征描述的创新选择。本研究的目的是评估三种不同的PIRO分类系统预测医院死亡率的准确性。
前瞻性收集德国柏林一家大学医院五个重症监护病房的成人脓毒症患者的数据。使用患者特征比较三种不同的评分系统，根据全因医院死亡风险对人群进行分类（莫雷诺 - PIRO、鲁布洛塔 - PIRO和豪厄尔 - PIRO）。
纳入278例脓毒症患者，并使用不同的PIRO模型进行重新分类。全因医院死亡率为16.2%。医院死亡率随PIRO评分升高而增加，莫雷诺 - PIRO的比值比为1.070（95%可信区间1.041 - 1.100），鲁布洛塔 - PIRO为1.282（95%可信区间1.079 - 1.524），豪厄尔 - PIRO为1.256（95%可信区间1.146 - 1.367）。莫雷诺 - PIRO的曲线下面积为0.743（95%可信区间：0.687 - 0.793），鲁布洛塔 - PIRO为0.646（95%可信区间：0.587 - 0.702），豪厄尔 - PIRO为0.751（95%可信区间：0.696 - 0.801）。与鲁布洛塔 - PIRO相比，莫雷诺 - PIRO和豪厄尔 - PIRO在统计学上有差异（P = 0.046和P = 0.035）。
所提出的PIRO分类在模型之间显示出细微差异，没有一种方法具有优先性，并且所有方法似乎都适用于患者分类。未来需要进一步发展PIRO，以明确易感性、感染，特别是反应类别。
