Diagnostic test accuracy of D‐dimer with or without a clinical decision rule in peripartum patients with suspected venous thromboembolism: A systematic review and meta‐analysis

Abstract

AbstractBackgroundPregnancy and the peripartum period is a hypercoagulable state increasing the risk of venous thromboembolism (VTE). There may be a role in utilising D‐dimer in the peripartum setting.AimsThe purpose of this review was to summarise the latest evidence regarding the diagnostic accuracy of D‐dimer in the peripartum setting with or without the addition of clinical decision rules.MethodsWe searched PubMed and CENTRAL databases to identify articles that included studies of women who had suspected VTE, underwent a D‐dimer index test to rule out VTE and where radiological imaging or clinical follow‐up, to a minimum of 30 days, was used as the reference standard.ResultsWe included 11 studies in the systematic review and meta‐analysis. The log diagnostic odds ratio (DOR) for identifying VTE using D‐dimer was 1.56 (95% confidence interval (CI) 0.59–2.52). The pooled sensitivity was 87% (95% CI 76.8–93%), specificity was 63.2% (95% CI 47.1–76.7%), and the area under receiver operator characteristic (ROC) curves was 0.76. We included four studies evaluating D‐dimer combined with YEARS to detect VTE. The log DOR for identifying VTE using D‐dimer combined with YEARS was 1.13 (95% CI 0.005–2.25). The pooled sensitivity was 89.8% (95% CI 60.2–98.1%), specificity was 65.7% (95% CI 54.7–75.2%) and the area under ROC for studies included with the YEARS clinical decision rule was 0.49.ConclusionThis review highlighted that D‐dimer use in the peripartum period for detection of VTE had a high sensitivity and high DOR but a poor area under ROC, which may limit its use in clinical practice.