Navid Farahmand, Claude B Sirlin, Michèle A Brown, Gordon P Shragg, Dale Fortlage, David B Hoyt, and Giovanna Casola.
Departments of Radiology and Surgery and General Clinical Research Center, University of California at San Diego, 200 W Arbor Drive, San Diego, CA 92103-8756, USA.
Radiology. 2005 May 1;235(2):436-43.
PurposeTo determine retrospectively the accuracy of screening ultrasonography (US) in patients with hypotension (systolic blood pressure Materials And MethodsThe investigational review board approved the study and waived informed consent. The study group consisted of 128 hypotensive patients with blunt abdominal trauma who underwent screening US over a 9-year period. Abdomens were scanned for free fluid and for parenchymal heterogeneity in visceral organs; scans that depicted these were considered positive. Prospective reports were used to calculate diagnostic performance. Patients were retrospectively given a fluid score according to the number of fluid pockets visualized (0, 1, or > or =2) (consensus by three readers) and were assigned to a low- or high-risk group according to the presence of hematuria and/or axial fracture on radiographs. Screening US results were compared with findings with the best available reference standard (computed tomography [CT]), repeat US, other diagnostic test, laparotomy, autopsy, clinical course). Data were compared by using chi(2) or Fisher exact test, depending on expected frequencies, with Bonferroni correction for multiple comparisons. Continuous variables were compared by using unpaired Student t test or Mann-Whitney U test, depending on data distribution.ResultsThe study included 77 male and 51 female patients (mean age, 42 years). Sensitivity was 85% (44 of 52) for detection of any injuries, 97% (30 of 31) for surgical injuries (ie, injuries requiring surgery), and 100% (10 of 10) for fatal injuries. Specificity was 96% (73 of 76), 82% (80 of 97), and 69% (81 of 118), and accuracy was 91% (117 of 128), 86% (110 of 128), and 71% (91 of 128), for respective injury categories. One nonfatal surgical injury was missed in a high-risk patient. For each injury category, frequency of injury in patients with a fluid score of 2 or more was nine times that in patients with a score of 0 (P < .001 for all comparisons). Frequency of false-negative US findings in high-risk patients was eight times that in low-risk patients (P < .01).ConclusionIn patients who are hypotensive after blunt abdominal trauma and not hemodynamically stable enough to undergo diagnostic CT, negative US findings virtually exclude surgical injury, while positive US findings indicate surgical injury in 64% of cases.(c) RSNA, 2005.