Abstract
IMPORTANCE: Advanced imaging to diagnose pulmonary embolism (PE) during pregnancy increases costs, ionizing radiation exposure, and the risk of overdiagnosis. D-dimer testing can exclude PE and compression ultrasonography can diagnose deep vein thrombosis (DVT), both reducing advanced imaging needs. How these 2 strategies are used in community PE diagnostics is unknown. OBJECTIVE: To analyze the use and efficiencies of D-dimer and ultrasonography for the diagnosis of antenatal PE. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study was performed in 21 US community hospitals from October 1, 2021, to March 31, 2023. Participants included pregnant patients who underwent D-dimer testing, compression ultrasonography, or advanced imaging for suspected PE. EXPOSURES: D-dimer testing or ultrasonography. MAIN OUTCOMES AND MEASURES: The main outcome was avoidance of advanced imaging, measured by the number needed to test (NNT) to avoid 1 advanced imaging study using low to intermediate D-dimer values or a positive ultrasonography finding. Variables associated with D-dimer and ultrasonography were evaluated using quasi-Poisson regression reported as adjusted relative risks (ARRs). RESULTS: A total of 720 outpatient encounters in 699 patients were included in the analysis (median age, 30.2 [IQR, 24.3-36.1] years). D-dimer testing was obtained in 524 patients (72.8%). Advanced imaging was not pursued in 149 of 155 patients (96.1%) with low (<0.5 mg/L) and 82 of 179 (45.8%) with intermediate (0.5 to <1.0 mg/L) D-dimer values, yielding an NNT to avoid 1 advanced imaging study of 2.3 (95% CI, 2.0-2.6). No 90-day venous thromboembolism or death occurred following rule-outs. D-dimer testing was less commonly used in third (vs first) trimester (ARR, 0.85; 95% CI, 0.75-0.97); with unilateral signs or symptoms of DVT (vs none) (ARR, 0.67; 95% CI, 0.54-0.82), and during labor and delivery unit (vs emergency department) evaluation (ARR, 0.27; 95% CI, 0.20-0.35). Compression ultrasonography was performed in 229 patients (31.8%), with DVT diagnosed in 3 (1.3%). Advanced imaging was not pursued in 2 of these patients, yielding an NNT to avoid 1 advanced imaging study of 115 (95% CI, 32-417). Ultrasonography use and results varied by DVT symptoms. Among 55 patients with DVT symptoms, 46 (83.6%) underwent ultrasonography, with positive findings in 3 (yield of 6.5%). Among 665 patients without symptoms, 183 (27.5%) underwent ultrasonography, with negative findings in all (yield of 0). CONCLUSIONS AND RELEVANCE: In this cohort study, use of D-dimer testing was efficient and safe with opportunities for expansion across settings and patients. These findings suggest that ultrasonography efficiency was low but could be improved with symptom-driven use in antenatal PE diagnostics.