Schurink Carolina A M, Nieuwenhoven Christianne A Van, Jacobs Jan A, Rozenberg-Arska Maja, Joore Hans C A, Buskens Erik, Hoepelman Andy I M, Bonten Marc J M
Department of Medicine, Division of Acute Medicine and Infectious Diseases, Eijkman-Winkler Laboratory for Microbiology , Julius Centre for Health Sciences and Primary Health Care, University Hospital Maastricht, University Medical Centre Utrecht and Department of Medical Microbiology, PO Box 85500, HP F02.1263508 GA , Utrecht, The Netherlands.
Intensive Care Med. 2004 Feb;30(2):217-224. doi: 10.1007/s00134-003-2018-2. Epub 2003 Oct 18.
Although quantitative microbiological cultures of samples obtained by bronchoscopy are considered the most specific tool for diagnosing ventilator-associated pneumonia, this labor-intensive invasive technique is not widely used. The Clinical Pulmonary Infection Score (CPIS), a diagnostic algorithm that relies on easily available clinical, radiographic, and microbiological criteria, could be an attractive alternative for diagnosing ventilator-associated pneumonia. Initially, the CPIS scoring system was validated upon 40 quantitative cultures of bronchoalveolar lavage fluid from 28 patients, and only few other studies have evaluated this scoring system since then. Therefore, little is known about the accuracy of this score.
We compared the scores of a slightly adjusted CPIS with results from quantitative cultures of bronchoalveolar lavage fluid in 99 consecutive patients with suspicion of ventilator-associated pneumonia, using growth of > or =10(4) cfu/ml in bronchoalveolar lavage fluid as a cut-off for diagnosing ventilator-associated pneumonia. In addition, the CPIS were calculated for 52 patients by two different intensivists to determine the inter-observer variability.
Ventilator-associated pneumonia was diagnosed in 69 (69.6%) patients. When using a CPIS >5 as diagnostic cutoff, the sensitivity of the score was 83% and its specificity was 17%. The area under the Receiver Operating Characteristic curve was 0.55. The level of agreement for prospectively measured Clinical Pulmonary Infection Score (< or =6 and >6) was poor (kappa =0.16).
When compared to quantitative cultures of bronchoalveolar lavage fluid, the CPIS has a low sensitivity and specificity for diagnosing ventilator-associated pneumonia with considerable inter-observer variability.
虽然通过支气管镜检查获取的样本进行定量微生物培养被认为是诊断呼吸机相关性肺炎最具特异性的工具，但这种劳动强度大的侵入性技术并未得到广泛应用。临床肺部感染评分（CPIS）是一种依赖于易于获得的临床、影像学和微生物学标准的诊断算法，可能是诊断呼吸机相关性肺炎的一种有吸引力的替代方法。最初，CPIS评分系统是基于对28例患者的40份支气管肺泡灌洗 fluid 的定量培养进行验证的，从那时起只有少数其他研究评估了该评分系统。因此，对该评分的准确性了解甚少。
我们将经过轻微调整的CPIS评分与99例连续怀疑患有呼吸机相关性肺炎患者的支气管肺泡灌洗 fluid 定量培养结果进行比较，以支气管肺泡灌洗 fluid 中生长≥10⁴ cfu/ml作为诊断呼吸机相关性肺炎的临界值。此外，由两名不同的重症监护医生对52例患者计算CPIS，以确定观察者间的变异性。
69例（69.6%）患者被诊断为呼吸机相关性肺炎。当使用CPIS>5作为诊断临界值时，该评分的敏感性为83%，特异性为17%。受试者工作特征曲线下面积为0.55。前瞻性测量的临床肺部感染评分（≤6和>6）的一致性水平较差（kappa=0.16）。
与支气管肺泡灌洗 fluid 的定量培养相比，CPIS在诊断呼吸机相关性肺炎时敏感性和特异性较低，观察者间变异性较大。
