Accuracy of Computed Tomography in Diagnosis of Intra‐abdominal Injuries in Stable Patients With Anterior Abdominal Stab Wounds: A Systematic Review and Meta‐analysis. (2nd April 2018)

• Record Type: Journal Article
• Authors: Baron, Bonny J.; Benabbas, Roshanak; Kohler, Casey; Biggs, Carina; Roudnitsky, Valery; Paladino, Lorenzo; Sinert, Richard
• Editors: Carpenter, Christopher R.
• Is Part Of: Academic emergency medicine. Volume 25:Number 7(2018)
• Page Start: 744
• Page End: 757
• Publication Date: 2018-04-02
• Subjects: Emergency medicine -- Periodicals; 616.02505
• DOI: 10.1111/acem.13380
• Languages: English
• ISSNs: 1069-6563

Abstract

Abstract: Background: Workup for patients presenting to the emergency department (ED) following an anterior abdominal stab wound (AASW) has been debated since the 1960s. Experts agree that patients with peritonitis, evisceration, or hemodynamic instability should undergo immediate laparotomy (LAP); however, workup of stable, asymptomatic or nonperitoneal patients is not clearly defined. Objectives: The objective was to evaluate the accuracy of computed tomography of abdomen and pelvis (CTAP) for diagnosis of intraabdominal injuries requiring therapeutic laparotomy (THER‐LAP) in ED patients with AASW. Is a negative CT scan without a period of observation sufficient to safely discharge a hemodynamically stable, asymptomatic AASW patient? Methods: We searched PubMed, Embase, and Scopus from their inception until May 2017 for studies on ED patients with AASW. We defined the reference standard test as LAP for patients who were managed surgically and inpatient observation in those who were managed nonoperatively. In those who underwent LAP, THER‐LAP was considered as disease positive. We used the Quality Assessment Tool for Diagnostic Accuracy Studies (QUADAS‐2) to evaluate the risk of bias and assess the applicability of the included studies. We attempted to compute the pooled sensitivity, specificity, positive likelihood ratio (LR+), and negative likelihood ratio (LR–) using a random‐effects model with MetaDiSc software and calculate testing and treatment thresholds for CT scan applying the Pauker and Kassirer model. Results: Seven studies were included encompassing 575 patients. The weighted prevalence of THER‐LAP was 34.3% (95% confidence interval [CI] = 30.5%–38.2%). Studies had variable quality and the inclusion criteria were not uniform. The operating characteristics of CT scan were as follows: sensitivity = 50% to 100%, specificity = 39% to 97%, LR+ = 1.0 to 15.7, and LR– = 0.07 to 1.0. The high heterogeneity (I 2 > 75%) of the operating characteristics of CT scan prevented pooling of the data and therefore the testing and treatment thresholds could not be estimated. Discussion: The articles revealed a high prevalence (8.7%, 95% CI = 6.1%–12.2%) of injuries requiring THER‐LAP in patients with a negative CT scan and almost half (47%, 95% CI = 30%–64%) of those injuries involved the small bowel. Conclusions: In stable AASW patients, a negative CT scan alone without an observation period is inadequate to exclude significant intraabdominal injuries.