A Retrospective Review of the Appropriateness of D-Dimer Ordering and Interpretation Using Wells' Clinical Probability Criteria

Introduction

The D-dimer has been validated in diagnostic venous thromboembolism (VTE) algorithms. The high sensitivity of the assay allows for safe exclusion of VTE in patients with low clinical pre-test probability and a negative D-dimer. The Wells score for Deep Vein Thrombosis (DVT) and Pulmonary Embolism (PE) are validated pre-test probability tools which help guide physicians on when to order a D-dimer in patients with suspected VTE.

However, we suspect these scoring tools are often under-utilized by physicians leading to inappropriate D-dimer ordering and subsequent interventions. We sought to explore the landscape of D-dimer ordering at our institution.

Methods

We conducted a retrospective chart review of 482 patients in whom a D-dimer had been ordered over a 3-month period at the University of Alberta Hospital, a tertiary care teaching hospital in Edmonton, Canada. Charts were reviewed for patient demographics, specialty of ordering physician, apparent indication for ordering, patient risk factors for VTE and evidence of a pre-test probability (PTP) calculation.

If no PTP score was documented, we retrospectively calculated Wells DVT or PE scores. VTE was deemed likely with a calculated Wells score for DVT ≧2 or Wells score for PE >4. In the case of high PTP for PE, patients should go directly to imaging and a D-dimer should not be performed. A cut off of ≥ 0.50 mg/L was deemed a positive D-dimer (STA-LIATEST). We also reviewed subsequent investigations thought to be influenced by interpretation of the D-dimer including: ventilation/perfusion (V/Q) and pulmonary angiography (CTPA) scans, and upper and lower extremity doppler ultrasound studies. We then used multivariable logistic regression analysis to evaluate the proportion of patients who received imaging despite a low PTP and negative D-dimer.

Results

Seventy eight percent of D-dimers were ordered by Emergency physicians while 15.3% were drawn on admitted patients, and 5.8% in the outpatient setting. The indication for ordering was unknown in 87 (17.5%) of cases. Pre-test probability scores were documented in only 8 (1.6%) of cases. All of those documented were the Wells PE score. When Wells DVT and PE scores were calculated retrospectively, 30.0% and 17.1% (87 cases) were deemed ‘likely’ for VTE, respectively. However, imaging was performed in 172 cases (34.6%), including in 36 cases despite a negative D-dimer result and low PTP.

In contrast, 68 cases (17.2%) had a D-dimer performed with a high Wells PTP for PE despite the recommendation to proceed directly to imaging. VTE (either DVT or PE) was confirmed by imaging in 32 (18.6%) of cases, the majority (53.1%) had a high retrospective PTP.

Conclusions

Inappropriate ordering and interpretation of D-dimers remains a significant problem despite the implementation of clinical guidelines and pre-test probability algorithms, namely the Wells score for DVT and PE meant to guide physicians. This leads to unnecessary cost, radiation exposure, and prolonged contact with the health care system for patients. This suggests the need for quality improvement initiatives which draw physician’s attention to pre-test probability tools which can curbing subsequent inappropriate investigations and improve patient care.