Study objectives: Validated clinical pretest probability tools are lacking in the diagnosis of pulmonary embolism (PE) in pregnancy, and the negative predictive value of D-dimers in this population has not been appropriately tested. The goal of this study was to evaluate outcomes of negative multidetector computed tomography with pulmonary angiography (MDCT-PA) for the diagnosis of PE in pregnancy in the absence of such measures.
Methods: Imaging and medical record data, including clinical presentation, risk factors, and all imaging studies performed for the diagnosis of venous thromboembolism (VTE), of 343 pregnant women with PE were reviewed retrospectively. The imaging was performed at a large tertiary-care women's hospital between 2004 and 2008 using the same-generation MDCT scanner. Primary outcome measure was the occurrence of VTE events 3 months following MDCT-PA or 6 weeks postpartum (whichever came later).
Results: Dyspnea (75.6%) and chest pain (45.6%) were the most common complaints. Fifty-seven percent of patients had at least one additional risk factor other than pregnancy. Body mass index >30 was the most common risk factor (49%). Eight scans were positive for PE and one patient with a negative MDCT-PA had a positive upper-extremity ultrasound, yielding a diagnosis of VTE in 2.9%. Negative scans were technically adequate in 79.1%, technically limited in 20%, and inconclusive in 0.9%. Follow-up showed no symptomatic VTE events after index imaging.
Conclusion: This study showed that MDCT-PA may safely exclude clinically significant PE in pregnancy.