Navigation auf zora.uzh.ch

Search ZORA

ZORA (Zurich Open Repository and Archive)

Distinguishing Infected From Noninfected Abdominal Fluid Collections After Surgery: An Imaging, Clinical, and Laboratory-Based Scoring System

Gnannt, Ralph; Fischer, Michael A; Baechler, Thomas; Clavien, Pierre-Alain; Karlo, Christoph; Seifert, Burkhardt; Lesurtel, Mickael; Alkadhi, Hatem (2015). Distinguishing Infected From Noninfected Abdominal Fluid Collections After Surgery: An Imaging, Clinical, and Laboratory-Based Scoring System. Investigative Radiology, 50(1):17-23.

Abstract

OBJECTIVES: Mortality from abdominal abscesses ranges from 30% in treated cases up to 80% to 100% in patients with undrained or nonoperated abscesses. Various computed tomographic (CT) imaging features have been suggested to indicate infection of postoperative abdominal fluid collections; however, features are nonspecific and substantial overlap between infected and noninfected collections exists. The purpose of this study was to develop and validate a scoring system on the basis of CT imaging findings as well as laboratory and clinical parameters for distinguishing infected from noninfected abdominal fluid collections after surgery.
MATERIALS AND METHODS: The score developmental cohort included 100 consecutive patients (69 men, 31 women; mean age, 58 ± 17 years) who underwent portal-venous phase CT within 24 hours before CT-guided intervention of postoperative abdominal fluid collections. Imaging features included attenuation (Hounsfield unit [HU]), volume, wall enhancement and thickness, fat stranding, as well as entrapped gas of fluid collections. Laboratory and clinical parameters included diabetes, intake of immunosuppressive drugs, body temperature, C-reactive protein, and leukocyte blood cell count. The score was validated in a separate cohort of 30 consecutive patients (17 men, 13 women; mean age, 51 ± 15 years) with postoperative abdominal fluid collections. Microbiologic analysis from fluid samples served as the standard of reference.
RESULTS: Diabetes, body temperature, C-reactive protein, attenuation of the fluid collection (in HUs), wall enhancement and thickness of the wall, adjacent fat stranding, as well as entrapped gas within the fluid collection were significantly different between infected and noninfected collections (P < 0.001). Multiple logistic regression analysis revealed diabetes, C-reactive protein, attenuation of the fluid collection (in HUs), as well as entrapped gas as significant independent predictors of infection (P < 0.001) and thus was selected for constructing a scoring system from 0 to 10 (diabetes: 2 points; C-reactive protein, ≥100 mg/L: 1 point; attenuation of fluid collection, ≥20 HU: 4 points; entrapped gas: 3 points). The model was well calibrated (Hosmer-Lemeshow test, P = 0.36). In the validation cohort, scores of 2 or lower had a 90% (95% confidence interval [CI], 56%-100%) negative predictive value, scores of 3 or higher had an 80% (95% CI, 56%-94%) positive predictive value, and scores of 6 or higher a 100% (95% CI, 74%-100%) positive predictive value for diagnosing infected fluid collections. Receiver operating characteristic analysis revealed an area under the curve of 0.96 (95% CI, 0.88-1.00) for the score.
CONCLUSIONS: We introduce an accurate scoring system including quantitative radiologic, laboratory, and clinical parameters for distinguishing infected from noninfected fluid collections after abdominal surgery.

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:04 Faculty of Medicine > University Hospital Zurich > Clinic for Diagnostic and Interventional Radiology
04 Faculty of Medicine > Epidemiology, Biostatistics and Prevention Institute (EBPI)
Dewey Decimal Classification:610 Medicine & health
Scopus Subject Areas:Health Sciences > Radiology, Nuclear Medicine and Imaging
Language:English
Date:2015
Deposited On:23 Sep 2014 12:35
Last Modified:11 Nov 2024 02:40
Publisher:Lippincott, Williams & Wilkins
ISSN:0020-9996
OA Status:Green
Publisher DOI:https://doi.org/10.1097/RLI.0000000000000090
PubMed ID:25198832
Download PDF  'Distinguishing Infected From Noninfected Abdominal Fluid Collections After Surgery: An Imaging, Clinical, and Laboratory-Based Scoring System'.
Preview
  • Content: Published Version

Metadata Export

Statistics

Citations

Dimensions.ai Metrics
41 citations in Web of Science®
42 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

3019 downloads since deposited on 23 Sep 2014
328 downloads since 12 months
Detailed statistics

Authors, Affiliations, Collaborations

Similar Publications