BACKGROUND & AIMS:
BACKGROUND:Severity of illness scores are helpful in predicting mortality; however, no standardized scoring system has been validated in patients with Staphylococcus aureus bacteremia (SAB). The modified Rapid Emergency Medicine Score (REMS), the CURB-65 (confusion, urea, respiratory rate, blood pressure and age 65) and the Charlson weighted index of comorbidity (CWIC) were compared in predicting outcomes at the onset of SAB.
METHODS:All adult inpatients with SAB from July 15, 2008, to December 31, 2009, were prospectively assessed. The 3 scoring systems were applied: REMS, CURB-65 and CWIC. The end points were attributable and overall mortality.
RESULTS:A total of 241 patients with SAB were reviewed during the study period. The all-cause mortality rate was 22.8% and attributable mortality 14.1%. Patients who died had higher mean CURB-65 score and REMS than those who lived, whereas the difference in the CWIC score was not significant. Two logistic regression models based on CURB-65 score or REMS, after controlling for CWIC, revealed that both scores were independent predictors of mortality, with an odds ratio of 3.38 (P < 0.0001) and 1.45 (P < 0.0001) for CURB-65 and REMS, respectively. Receiver operating characteristic analysis revealed that a cutoff point of 3.0 (CURB-65) and 6.0 (REMS) provided the highest sensitivity and specificity. The area under the curves for all-cause mortality were 0.832 and 0.806, and for attributable mortality 0.845 and 0.819, for CURB-65 and REMS, respectively.
CONCLUSIONS:REMS and CURB-65 scores outperformed CWIC as predictors of mortality in SAB and may be effective in predicting the severity of illness at the onset of bacteremia.
背景与目标:
背景:疾病评分的高低有助于预测死亡率。但是,尚无标准的评分系统在金黄色葡萄球菌菌血症(SAB)患者中得到验证。在预测SAB发作的结果时,比较了修改后的快速急诊医学评分(REMS),CURB-65(精神错乱,尿素,呼吸频率,血压和65岁)和Charlson合并症加权指数(CWIC)。
方法:对2008年7月15日至2009年12月31日期间所有成人SAB住院患者进行前瞻性评估。应用了3种评分系统:REMS,CURB-65和CWIC。终点归因于总死亡率。
结果:在研究期间,共对241名SAB患者进行了回顾。全因死亡率为22.8%,归因死亡率为14.1%。死亡患者的CURB-65平均得分和REMS均高于活着的患者,而CWIC得分的差异并不显着。在控制了CWIC之后,基于CURB-65得分或REMS的两个逻辑回归模型显示,两个得分都是死亡率的独立预测因子,CURB-65的优势比为3.38(P <0.0001)和1.45(P <0.0001)。和REMS分别。接收器工作特性分析显示,临界点3.0(CURB-65)和6.0(REMS)提供了最高的灵敏度和特异性。对于CURB-65和REMS,全因死亡率的曲线下面积分别为0.832和0.806,归因死亡率分别为0.845和0.819。
结论:REMS和CURB-65得分优于CWIC作为SAB死亡率的预测指标,可能有效地预测菌血症发作时疾病的严重程度。