SESSION TITLE: Quality & Clinical Improvement II
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Tuesday, October 28, 2014 at 11:00 AM - 12:15 PM
PURPOSE: The objective of this study is to validate the currently used cut-off value of D-dimer (DD) assay in the workup of Pulmonary Embolism (PE) in a multicenter acute care institute in Detroit. We hypothesized that our current cut-off value is leading to a high number of negative CT-pulmonary angiography (CT-PA).
METHODS: We conducted a retrospective analysis on 768 patients who had DD and CT-PA as part of PE workup at three tertiary centers in Detroit, during a six-month period. Receiver-Operator Characteristics (ROC) curve was performed. In addition, we searched for the best cut-off point that maintains high sensitivity and carries less false positive rate simultaneously.
RESULTS: Our sample had female (66%) and African American (90%) predominance with a median age of 53 years. Out of 768 cases, PE was confirmed in 57 (7.4%). The DD value in patients with confirmed PE had an average of 9.17 mg/L compared to 2.54 mg/L in the patients with negative CT-PA (P< 0.023). The area under the curve (AUC) shown in ROC was 0.85 (P< 0.0001). ROC analysis suggested a cut-off point of 3.09 mg/L with a sensitivity of 74% and a specificity of 84%, yet such sensitivity has a negative impact on using D-dimer as a tool for ruling out PE. Our institution’s current DD cut-off value (<0.59 mg/L) has a sensitivity and a specificity of 98.2% and 11.4%, respectively. However, the specificity was significantly improved to 40% by increasing this cut-off to be <0.92 mg/L while maintaining the same sensitivity. Applying this cut-off might have resulted in 204 less CT-PA scans (27%) out of the sample.
CONCLUSIONS: CT-PA in the workup of PE carries multiple risks such as contrast-induced nephropathy, extravasation of contrast material and radiation exposure. Our data shows that the number of CT-PA in our institute can be substantially reduced using the adjusted D-dimer cut-off value (<0.92 mg/L) without compromising its diagnostic accuracy. Such practice would potentially improve patient safety and utilization of hospital resources.
CLINICAL IMPLICATIONS: Based on these results, we would apply the adjusted cut-off value of D-dimer assay to our institution in a prospective initiative. We recommend similar studies in multi-center database to validate the accuracy of this vital test.
DISCLOSURE: The following authors have nothing to disclose: Sulaiman Alhassan, Alaa Abu Sayf, Hicham Krayem, Rajika Munasinghe, Camelia Arsene, John Flack
No Product/Research Disclosure Information