Patients with lower limb injuries are commonly discharged from the ED with the affected area immobilised. Rigid casting of the lower limb is known to be a risk factor for the development of venous thromboembolism (VTE), making thromboprophylaxis in this population an important consideration for clinicians in the ED. The use of structured risk assessment methods (RAMs) to evaluate VTE risk and recommend thromboprophylaxis to those at higher risk is widespread in the UK. However, the evidence informing this practice is nearly exclusively based on studies of patients with rigid lower limb casts but many patients with knee injuries, including some with significant thrombotic risk factors, are managed in semi-rigid (‘cricket’) knee splints. These are both removable and allow free movement of the ankle, but the baseline risk of VTE and the performance of different RAMs in this population are not known.
Consecutive patients (≥14 years) discharged from the ED at Aberdeen Royal Infirmary, between 1 January 2010 and 31 December 2021, in a semi-rigid knee splint were identified retrospectively and followed up to 3 months after splint removal for the development of symptomatic VTE. Secondarily, data permitting the assessment of five different RAMs (NICE, GEMNet, an Aberdeen tool, the Plymouth score (V.2) and the L-TRiP(cast) score) were extracted systematically and compared.
In 510 patients (mean age 32 (SD 16) years, 62% male) none received thromboprophylaxis and all completed follow-up. Two patients developed symptomatic VTE (0.4%, 95% CI 0.1% to 1.4%). The different RAMs varied considerably in the proportions identified for thromboprophylaxis from GEMNet (47%) to the L-TRiP(cast) score (2%), but no RAM was able to identify the two patients who progressed to VTE.
In our cohort of patients managed in semi-rigid removable knee splints, the risk of symptomatic VTE was low, about 1 in 250, and current methods of VTE risk assessment did not prove clinically useful.
Diagnosis of venous thromboembolism (VTE) requires chest CT angiography for pulmonary embolism and venous ultrasound for deep vein thrombosis. To reduce imaging, guidelines recommend D-dimer levels to rule-out VTE in patients with a low pre-test probability. The most widely used D-dimer cut-off is 500 ng/mL. This cut-off has low specificity, meaning many patients without disease require imaging.
In this retrospective chart review, we evaluated the diagnostic performance of the D-dimer/fibrinogen ratio (DFR) for identifying thromboembolism and compared it to the performance of two different D-dimer cut-offs (500 ng/mL and 1000 ng/mL) in patients who underwent a chest CT angiography or a venous ultrasound in the ED of San Raffaele Hospital, Italy, in 2017. Patients had a retrospective Wells score calculated after chart review, identifying both high-risk and low-risk pre-test probability patients for this study and low probability patients were further stratified into low-risk of deep vein thrombosis or pulmonary embolism.
Enrolled patients included 92 with suspected pulmonary embolism and 154 with suspected deep vein thrombosis; of whom 67 (27%) were diagnosed with VTE. The most accurate cut-off for DFR in terms of discriminative power was 2.65. In the whole sample and in low-risk patients, this cut-off had the same sensitivity values of the 500 ng/mL D-dimer cut-off (97% (95% CI: 89.8% to 99.2%)), while slightly lower sensitivity values were found for the 1000 ng/mL D-dimer cut-off (95.5% (95% CI: 87.6% to 98.5%)). Specificity was higher for the 2.65 DFR cut-off (55.3% (95% CI: 48.0% to 62.4%)) in the whole sample compared with both 500 ng/mL D-dimer cut-off (22.9% (95% CI: 17.4% to 29.6%)) and 1000 ng/mL D-dimer cut-off (45.8% (95% CI: 38.7% to 53.1%)). Similar results were found in all subgroups.
A DFR, with a cut-off of 2.65, may improve the specificity for VTE patients when compared with D-dimer alone in high-risk VTE emergency medicine populations. This is exploratory information only, needing evaluation in prospective, multicentre studies, prior to consideration for use in routine clinical work.