Validation of the age-adjusted D-dimer in an ambulatory deep vein thrombosis diagnostic service: beware of superficial thrombophlebitis.

Abstract

We are aware the National Institute for Health and Care Excellence has recently recommended the use of age-adjusted D-dimer for venous thromboembolism (VTE) investigation which would help reduce the burden of diagnostic imaging on overstretched radiology departments.1 D-dimer assays are not standardised between National Health Service Trusts and therefore we attempted to validate (audit ID 2797) age-adjusted D-dimer using data collected in our ambulatory deep vein thrombosis (DVT) service from years 2015 to 2019. Locally we use the HemosIL D-Dimer HS, Werfen, MA, USA, with a reference range of ≤230 ng/ml (D-dimer Units). Patients included were ≥50 years of age that had been referred to the ambulatory DVT service for investigation and were divided into two groups. The first was a D-dimer ≤230 ng/ml (that had nevertheless had a US Doppler scan) as this is the local reference range and second those with a D-dimer 231 ng/ml to ×10 the upper limit of age, and in both groups patients had a low pre-test probability Well's score (the age-adjusted group reflects the patients in the ADJUST-PE study where a D-dimer cutoff of 500 ng/ml (Fibrinogen Equivalent Units) to x10 upper limit of age was used which approximately corresponds to the reference range of our local D-dimer assay; 1 ng/ml D-dimer Unit ≈ 2 ng/ml Fibrinogen Equivalent Units).2 We maintain a prospective database for audit purposes of patients seen in the DVT service and this was interrogated. With 95% confidence to detect a 3% upper limit of failure rate of the strategy no more than 2/240 patients could have a DVT, which was the method used in the ADJUST-PE study, though whether this threshold is acceptable is debatable.2, 3 The results of the validation are found in Table I. Our validation highlights the limitations in a patient cohort with a high prevalence of superficial thrombophlebitis, in both those with a D-dimer below 230 ng/ml and in the age-adjusted category although performance for DVT was satisfactory. In our experience, it can be difficult to clinically distinguish STP and DVT; previously it has been shown that the D-dimer is unreliable in patients being investigated for STP.4 STP requires specific treatment as it can be associated with morbidity and risk of developing DVT or pulmonary embolism (PE).5 Previously in a large observational study of 844 patients with STP, 210 (25%) had VTE in the form of either DVT (the majority) or pulmonary embolism (PE).6 This highlights the need to make an assessment for concomitant DVT or PE in STP patients, in order to ensure adequate treatment. Discrimination between STP and VTE is of importance as the management of the two, and anticoagulant doses, differs. We have developed a specific policy for management of radiologically confirmed STP as cases over the past 10 years have risen dramatically, from around 40 cases per annum before 2010 to more than 150 per annum now.7 In summary, we would highlight our experience and the need for awareness of the limitations of age-adjusted D-dimer in suspected DVT patients. Given the issues identified in this audit, where the age-adjusted D-dimer could cause confusion in patients with STP and with high prevalence in our validation cohort, we are not going to implement age-adjusted D-dimer in the DVT service but we recognise that even with a normal D-dimer a patient may have STP. Implementation of age-adjusted D-dimer would have theoretically saved on around 10% (≈1000 patients per annum are referred) patients from having a scan that were seen in the DVT service within the audit period however given the prevalence of STP in the cohort and the need to exclude a concomitant DVT in these patients we are sceptical about reducing the number of scans. Alternatively, we will explore a recently published algorithm that safely reduced diagnostic imaging for PE patients that is simpler to use than age-adjusted D-dimer.8 In this study, a low clinical pre-test probability of PE was combined with a D-dimer threshold of 1000 ng/ml; 1285/2017 patients fell into this category and none were found to have a PE after 3 months follow-up, although a range of D-dimer assays were utilised, including only 14 patients tested with the HemosIL D-Dimer HS, highlighting the need for further validation. The issue of a lack of standardisation of D-dimer assays is also an on-going problem both within studies and for application in daily clinical practice. Our data also highlight the need for ambulatory DVT services to be aware of STP and develop pathways to manage this condition. WT, PS & ES designed the audit. MH & AD collected data. WT, WP & ES wrote the first draft. All authors approved the final draft. None relevant to this letter.