The safety and efficacy of using age-adjusted D-dimers in hospitalised patients in a diverse urban centre: a real-world data study

Abstract

ObjectiveDespite recommendations, age adjusted thresholds (AAT) for D Dimers are not routinely used as part of venous thromboembolism (VTE) screening in many healthcare settings due to concerns about missing cases, especially in older and co-morbid adults. The National Institute for Health and Care Excellence in the UK has highlighted that evidence to support AAT is not plentiful. This study assessed the real-world use of AAT D-dimers for VTE in a large cohort of acutely hospitalised patients.MethodsThis retrospective data study included all adult patients attending a large hospital with a suspected VTE between January 2017 to December 2021. The predictive accuracy of D-dimer was assessed against gold standard imaging. Outcomes of false negative (with AAT) and false positives (with standard thresholds) cases were assessed.Results27,526 suspected VTE attendances were included, with a 4.3% confirmed VTE diagnosis rate. The ST D-dimer exhibited high sensitivity (91.1%) but modest specificity (65.2%). The AAT demonstrated slightly lower sensitivity (87.0%) but higher specificity (71.7%, p<0.001). The performance of ST thresholds declined with age, with false positive rates increasing from 17.4% to 80.0% in people aged < 50 years and > 90 years respectively. The AAT accurately identified 1,700 true negatives misclassified as false positives by the ST. 14 patients in this group were admitted with a bleed within 30 days. AAT misdiagnosed 24 cases as false negatives, with most being small sub-segmental pulmonary emboli or non-occlusive DVTs. Using AAT thresholds could have avoided 64 scans per 1,000 attendances, saving approximately £235,310 of imaging costs in this cohort.ConclusionThe age-adjusted D-dimer threshold enhances diagnostic precision and could decrease unnecessary imaging and anticoagulation, reducing investigations with time and cost savings with no significant safety signal.