Conway R, O'Connell B, Byrne D, O'Riordan D, Silke B
PhD MRCPI, MED Directorate, St James's Hospital, Dublin 8, Ireland.
MD FRC Path, FRCPI, LabMed Directorate, St. James's Hospital, Dublin 8, Ireland.
Acute Med. 2020;19(2):83-89.
Positive blood cultures predict mortality. The prognostic value of blood culture performance itself has not been fully defined.
We evaluated medical admissions from 2002-2017. We defined blood culture category as 1) no culture 2) negative culture 3) positive culture. We employed a multivariable logistic regression model to evaluate outcomes.
We evaluated 78,568 blood cultures in 106,586 admissions. 30-day in-hospital mortality for no culture was 2.8% (95%CI 2.7, 2.9), culture negative 8.9% (95%CI 8.5, 9.3) and culture positive 16.7% (95%CI 15.5, 17.9). There was significant interaction between blood culture category and illness severity, OR 1.06 (95%CI 1.05, 1.08), and comorbidity, OR 1.09 (95%CI 1.09, 1.10).
Performance and results of blood cultures are independently associated with increased mortality.
血培养阳性可预测死亡率。血培养操作本身的预后价值尚未完全明确。
我们评估了2002年至2017年的内科住院患者。我们将血培养类别定义为：1）未进行培养；2）培养阴性；3）培养阳性。我们采用多变量逻辑回归模型来评估结果。
我们评估了106,586例住院患者的78,568次血培养。未进行培养的患者30天院内死亡率为2.8%（95%置信区间2.7，2.9），培养阴性患者为8.9%（95%置信区间8.5，9.3），培养阳性患者为16.7%（95%置信区间15.5，17.9）。血培养类别与疾病严重程度之间存在显著交互作用，比值比为1.06（95%置信区间1.05，1.08），与合并症之间也存在显著交互作用，比值比为1.09（95%置信区间1.09，1.10）。
血培养的操作和结果与死亡率增加独立相关。
