SAFELY RULING OUT DEEP VEIN THROMBOSIS IN PREGNANCY WITH THE LEFt SCORE AND TRIMESTER-ADJUSTED D-DIMER CUTOFFS (the PREG-DVT Study): A PROSPECTIVE COHORT STUDY.

1. DETAILED RESEARCH PROPOSAL 3.1.1. Diagnostic tests and rationale for their use in the PREG-DVT study Assessment of clinical probability Among the multiple tools that may be useful for DVT diagnosis, the assessment of clinical probability is a crucial step because according to the Bayesian approach, the interpretation of subsequent tests depends on pretest probability. For example, a D-dimer ELISA test below a certain cut-off in a patient with a low or moderate clinical probability of DVT lowers the likelihood of the disease to the point (1 to 2%) that clinicians would decide not to treat the patient.19 However, in the presence of a high clinical probability, the post-test probability of DVT remains too high to rule out DVT despite a normal D-dimer level.20 The LEFt clinical decision score-a simple three-item tool including Left leg symptoms, Edema (≥2 cm calf circumference difference compared to the other leg), and First trimester onset of symptoms-has shown promise in the LEFt derivation study6 and retrospective validation studies.7,8 More recently, the score was tested in a prospective study (LEAD study), which confirmed its safety in association with D-dimers at the usual cutoff of 500 ng/ml to safely rule out DVT in pregnant patients. The LEFt score represents a nice milestone in this setting as it provides a standardized, objective framework for a population that has historically been managed based on clinical "gut feeling." By utilizing the LEFt score, clinicians can move away from subjective judgment and use a now well-validated, reproducible tool to categorize clinical probability. Its primary importance lies in its role as a "gatekeeper": it allows for the safe identification of women with an unlikely probability of DVT who can undergo D-dimer testing. The LEFt score is shown in Table 1.

D-dimer measurement D-dimer measurement has been shown to be very safe to rule out PE and DVT. The test is highly sensitive and has widely been used to rule out DVT and PE in many pragmatic studies.16,21 In these studies, the three-month thromboembolic rate in patients left untreated after a normal D-dimer measurement was less than 1% and is therefore fully acceptable.

The LEAD study revealed that using a fixed 500 µg/L D-dimer cutoff resulted in a diagnostic yield of 7.4% for ruling out DVT.15 This low efficiency forced more than 90% of patients to undergo further imaging, and demonstrated that the standard cutoff set at 500 ng/mL is inefficient. When exploring D-dimer at trimester adjusted thresholds of <500 ng/ml FEUs (1st trimester), <750 ng/ml FEUs (2nd trimester) and <1000 ng/ml FEUs (3rd trimester), D-dimer was negative in 142/352 (40.3%) and yielded negative predictive values of 142/142, 100% (95% CI 97.4%-100%). Therefore, using trimester adjusted D-dimer cutoffs is a nice option to increase the proportion of patients in whom imaging can be avoided.3

Lower limb compression ultrasonography A prospective multicenter study conducted across tertiary centers and private practices in evaluated 210 pregnant and postpartum women presenting with suspected DVT using a single complete lower-limb high-definition B-mode CUS combined with Doppler flow imaging of the iliac veins. Anticoagulation was safely withheld in patients with a negative scan and the 3-month thromboembolic failure rate was only 1.1% (95% CI, 0.3% to 4.0%), demonstrated that the safety margins of a negative complete ultrasound match the historical benchmarks of contrast venography in the non-pregnant population.2 In a systematic review and metanalysis evaluating serial CUS limited to the proximal veins, a total of 635 pregnant patients with symptoms of DVT had an initial negative ultrasound examination. Of those, six were diagnosed with DVT during repeat serial testing (0.94%) and three developed DVT or PE during 3-month follow-up after serial ultrasonography (0.47%). The pooled false-negative rate of a single ultrasound was 1.27% (95% confidence interval, 0.42-2.56).22 Overall, this data suggest that both a single complete CUS or serial CUS limited to the proximal veins are safe for ruling out DVT in pregnant patients.

3.1.2 Study design We plan a multicenter prospective management outcome study. The study will be carried out in Canada, France, the Netherlands and Switzerland, with the support of the INVENT and INNO-VTE networks.

Intervention Eligible pregnant women with a clinical suspicion of DVT will be included in the study. The assessment of clinical probability will be done in each patient by the LEFt score (Table 1) and will separate patients in two categories (unlikely-likely). After the assessment of clinical probability, D-dimer will be performed in patients with an unlikely clinical probability (LEFt score 0-1 points). The D-dimer cutoffs to rule out DVT will be trimester-adapted. D-dimer will be considered as negative if lower than 500 ng/mL during the first trimester, if lower than 750 ng/mL during the second trimester and if lower than 1000 ng/mL during the third trimester.

Patients with an unlikely clinical probability and a negative trimester adjusted D-dimer test will not undergo further testing and will be followed-up for a 3-month period. Patients with a positive trimester-adapted D-dimer cutoff and/ patients with a likely clinical probability (LEFt score ≥2) will undergo compression ultrasonography. Ultrasonography will be either a single whole-leg ultrasound or a proximal repeated CUS. Patients with a negative CUS will not be treated with anticoagulants. In presence of a proximal DVT or distal DVT, patients will receive therapeutic anticoagulation. Magnetic Resonance Direct Thrombus imaging (MRDTI) will be conducted in patients with negative CUS in whom pelvic deep vein thrombosis is suspected.23 The study flow chart for patients is depicted in Figure 1.

3.1.3 Follow-up of patients included in the PREG-DVT study In prospective management outcome studies for VTE, the safety of a diagnostic algorithm is defined by the "failure rate", i.e. the incidence of symptomatic venous thromboembolic events (VTE) during a structured three-month follow-up period among patients in whom the diagnosis was initially excluded. A three-month window is an international recognized standard because it is considered long enough for any missed, clinically significant clot to manifest symptomatically, yet short enough to distinguish a missed initial event from a completely new, unrelated VTE event.

The follow-up will include all patients in whom the diagnostic strategy ruled out DVT at baseline without the initiation of anticoagulant therapy. To ensure high-quality data and minimize "loss to follow-up" the following protocol will be implemented:

At the time of discharge, patients will be clearly instructed to return immediately if any symptoms suggestive of VTE (dyspnoea, chest pain, hemoptysis, or leg swelling) recur or worsen.

A telephone interview will be conducted at the end of the 90-day period. Standardized questionnaires will be used to screen for any hospitalizations, new treatments, or clinical symptoms occurred during the interval.

If a patient presents with a suspected VTE during these three months, an objective workup (CUS, CTPA, V/Q scan or Single Photon Emission Computed Tomography (SPECT) will be performed.

All suspected events, including deaths, will be reviewed by an Independent Adjudication Committee blinded to the initial diagnostic results to determine if the event qualifies as a "failure" of the algorithm.

3.1.4 Ethical concerns The ethical justification for this study is rooted in the goal of making a standard medical procedure safer and more precise, rather than introducing a completely new diagnostic tool. Since CUS is already the widely accepted clinical standard for diagnosing DVT in pregnancy, the primary ethical mandate is to optimize its necessity and execution. This approach fulfils the ethical duty of "justification," ensuring that the test is only used when essential. By validating clinical decision algorithms that safely exempt low-risk patients from diagnostic imaging, this research optimizes maternal care and resource allocation. Furthermore, establishing a definitive diagnostic confirmation or exclusion satisfies a critical ethical imperative, ensuring that high-risk anticoagulation regimens-which carry maternal-fetal hemorrhagic liabilities-are administered exclusively under objective, rigorous indication.

The protocol will be submitted to the appropriate ethics committees in all centers and written informed consent will be obtained from all study participants