Department of Emergency Medicine, Seoul National University Hospital, Seoul, Korea.
Am J Emerg Med. 2011 May;29(4):412-7. doi: 10.1016/j.ajem.2009.11.015. Epub 2010 Apr 2.
Abdominal contrast-enhanced computed tomography (A-CECT) is widely used in emergency departments despite the risk of contrast-induced nephropathy. We attempted to develop a risk stratification nomogram for nephropathy in patients receiving emergency A-CECT.
Seven hundred fifty patients who received emergency A-CECT between August 2003 and January 2007, with available serum creatinine (SCr) measurements before and after A-CECT were included. Nephropathy was defined as either an absolute increase of 0.5 mg/dL or greater (44 μmol/L) or a relative increase of 25% or more in the SCr from baseline. A nomogram was developed based on multivariate logistic regression analysis using clinical variables available before A-CECT. The model was internally validated with a bootstrapping method, and performance was assessed by area under the receiver operating characteristics curve (AUC) and calibration curve.
Nephropathy was observed in 34 of 750 patients. A nomogram was developed using age (odds ratio, 1.04 per 1-year increment) and baseline SCr (odds ratio, 2.51 per 1-mg/dL increment) as risk factors. Diagnostic accuracy of the model was fair by bias-corrected calibration plot. The AUC of the model was 0.794 (95% confidence interval, 0.734-0.854), and the AUC with bootstrapping samples of 200 repetitions was 0.794 (95% confidence interval, 0.737-0.851).
The risk of nephropathy after emergency A-CECT can be individually predicted by internally validated nomogram using clinical variables available before the procedure.
尽管存在对比剂肾病的风险,腹部增强 CT(A-CECT)仍广泛用于急诊科。我们试图为接受急诊 A-CECT 的患者开发一种肾病风险分层列线图。
纳入 2003 年 8 月至 2007 年 1 月期间接受急诊 A-CECT 的 750 例患者,这些患者均有 A-CECT 前后的血清肌酐(SCr)测量值。肾病定义为 SCr 绝对值增加 0.5mg/dL(44μmol/L)或更多,或与基线相比相对增加 25%或更多。基于 A-CECT 前可用的临床变量,使用多变量逻辑回归分析开发了一个列线图。该模型通过自举法进行内部验证,并通过接收者操作特征曲线(AUC)和校准曲线评估性能。
750 例患者中有 34 例出现肾病。使用年龄(每 1 岁增加的优势比为 1.04)和基线 SCr(每 1mg/dL 增加的优势比为 2.51)作为风险因素开发了一个列线图。通过偏倚校正校准图,该模型的诊断准确性为中等。该模型的 AUC 为 0.794(95%置信区间,0.734-0.854),自举样本重复 200 次的 AUC 为 0.794(95%置信区间,0.737-0.851)。
使用该方法可根据术前可用的临床变量,通过内部验证的列线图来预测急诊 A-CECT 后发生肾病的风险。