Age-Adjusted D-Dimer Cutoff Levels to Rule Out Deep Vein Thrombosis

Key Points

Question

Does an age-adjusted D-dimer cutoff safely increase the proportion of patients in whom deep vein thrombosis (DVT) can be excluded?

Findings

In this multinational prospective outcome study including 3205 outpatients presenting to the emergency department with suspected DVT, the age-adjusted D-dimer cutoff safely excluded DVT. Among patients with D-dimer level between 500 µg/L and their age-adjusted cutoff, none developed venous thromboembolism at 3 months. Use of the age-adjusted cutoff resulted in a 7.4% absolute increase in the proportion of patients in whom the diagnosis could be excluded.

Meaning

An age-adjusted D-dimer cutoff may safely rule out DVT and increase diagnostic efficiency, reducing the need for unnecessary imaging.

Abstract

Importance

The age-adjusted D-dimer cutoff (age × 10 µg/L in patients 50 years or older), safely increases the diagnostic yield of D-dimer in patients with suspected pulmonary embolism but has not been validated in patients with suspected leg deep vein thrombosis (DVT).

Objective

To prospectively validate whether using an age-adjusted D-dimer cutoff allows clinicians to safely rule out DVT.

Design, Setting, and Patients

Multicenter, multinational prospective management outcome study conducted in 27 centers in Belgium, Canada, France, and Switzerland between January 2015 and October 2022 (last follow-up visit, January 30, 2023) and including outpatients presenting to the emergency department with suspected DVT.

Interventions

Patients were assessed by a sequential diagnostic strategy based on the assessment of clinical pretest probability by the Wells score, a highly sensitive D-dimer test, and leg compression ultrasonography. Patients in whom DVT was ruled out were followed up for a 3-month period.

Main Outcome and Measure

The primary outcome was the rate of adjudicated symptomatic venous thromboembolic events during follow-up in patients in whom DVT was ruled out based on a D-dimer value between the conventional cutoff of 500 µg/L and their age-adjusted cutoff.

Results

A total of 3205 patients were included. Median age was 59 years, and 1737 (54%) were female. DVT prevalence was 14%. Among the 2169 patients with a non-high or unlikely clinical probability, 531 (24.5% [95% CI, 22.7%-26.4%]) had a D-dimer level less than 500 µg/L, and 161 additional patients (7.4% [95% CI, 6.4%-8.6%]) had a D-dimer level between 500 µg/L and their age-adjusted cutoff. No failures were identified in patients with a D-dimer level 500 µg/L or greater but below the age-adjusted cutoff (0% [95% CI, 0%-2.3%]). Among patients 75 years or older, using the age-adjusted cutoff instead of the 500-µg/L cutoff increased the proportion of negative D-dimer from 33 of 379 (8.7% [95% CI, 6.3%-12.0%]) to 99 of 379 (26.1% [95% CI, 22.0%-30.8%]), without any false-negative test results.

Conclusions and Relevance

The age-adjusted D-dimer cutoff may safely rule out DVT and was associated with a larger number of patients in whom DVT could be effectively ruled out.

Trial Registration

ClinicalTrials.gov Identifier: NCT02384135.

This prospective management outcome study conducted at multiple centers in Belgium, Canada, France, and Switzerland assesses whether using an age-adjusted D-dimer cutoff allows clinicians to safely rule out deep vein thrombosis (DVT) in adult patients presenting with suspected DVT.

Introduction

Deep vein thrombosis (DVT) of the lower limbs is a common medical condition that represents a significant diagnostic challenge for clinicians. Rapid and accurate diagnosis is critical, because untreated DVT can lead to life-threatening complications such as pulmonary embolism. Traditional diagnostic strategies for DVT involve the use of clinical decision rules, such as the Wells score, D-dimer measurement, and imaging with lower limb venous compression ultrasonography.¹

The D-dimer test, which measures fibrin degradation products, has become a widely used tool for excluding DVT in patients with low to intermediate or unlikely clinical probability. However, the specificity of the test decreases with age, leading to a reduced clinical usefulness in older adult patients.² Over the last decade, the safety of adjusting the D-dimer cutoff value based on a patient’s age has been extensively studied in patients with suspected pulmonary embolism.^3,4 The use of the age-adjusted D-dimer cutoff for pulmonary embolism (defined as D-dimer level < patient’s age × 10 in µg/L in patients aged ≥50 years) is now endorsed by major clinical practice guidelines.^5,6 However, whether this can be extrapolated to patients with suspected lower limb DVT remains to be established.

Retrospective data showed promising results: in studies of patients with suspected DVT in which all patients with a D-dimer level greater than 500 µg/L underwent compression ultrasonography, a low venous thromboembolism (VTE) rate at initial presentation or during follow-up was reported among those with a negative age-adjusted D-dimer level.^7,8,9 Therefore, the study hypothesis was that using the age-adjusted D-dimer cutoff would reduce the need for ultrasound imaging without compromising diagnostic safety.

To confirm this hypothesis, a large prospective international diagnostic management study was designed, using the age-adjusted D-dimer cutoff in combination with pretest probability assessment and compression ultrasonography in patients with suspected leg DVT.

Methods

Study Setting

The Age-Adjusted D-Dimer Cutoff Levels to Rule Out Deep Vein Thrombosis: a Prospective Outcome Study (ADJUST-DVT) was designed as a multicenter, multinational prospective diagnostic management outcome study, involving 27 hospitals in 4 countries (Belgium, Canada, France, and Switzerland). The ethics committees of all participating institutions approved the study, and all patients provided written informed consent before enrollment. A data and safety monitoring board oversaw study progress and participants’ safety. The trial protocol is available in Supplement 1.

Patients

Patients who presented to the emergency department of the participating hospitals were eligible if they had a clinical suspicion of lower limb DVT. Patients were excluded if they were younger than 18 years, were pregnant, were already receiving anticoagulant therapy (eg, previous VTE, atrial fibrillation), had concomitant symptoms of pulmonary embolism, a life expectancy of less than 3 months, were unwilling or unable to provide informed consent, or were not available for follow-up.

Diagnostic Strategy

Clinical pretest probability was assessed using the 2-level or the 3-level Wells score for DVT (eTable in Supplement 2).¹⁰ While Canadian centers used the 2-level Wells score, stratifying patients into unlikely or likely clinical probability categories, centers in Europe used the 3-level Wells version, stratifying patients into low, intermediate, or high clinical probability. The assessment of the Wells score was performed by emergency department physicians. Patients with a high or likely clinical probability directly proceeded to compression ultrasonography. In patients with a low/intermediate or unlikely pretest probability, a D-dimer test was performed. D-dimer result was interpreted according to the age-adjusted cutoff: in patients younger than 50 years, DVT was excluded in those with a D-dimer value below 500 µg/L. In patients 50 years or older, the D-dimer test was considered negative if the result was below their age multiplied by 10, in µg/L. Several quantitative high-sensitivity D-dimer assays were used: the VIDAS D-dimer exclusion test of first or second generation (bioMérieux), second-generation Tinaquant and Cobas h 232 and Cardiac Reader (Roche Diagnostics), Liatest D-dimer (Stago), D-dimer HS500 (IL Diagnostics), Innovance D-dimer and Immulite 2000 D-dimer (Siemens), AxSym D-dimer assay (Abbott), MDA auto D-dimer (Trinity Biotech), and Nycocard D-dimer (Nycomed Pharma).

No further testing was required in patients with a negative D-dimer test, and they were not prescribed anticoagulant therapy. Patients with a positive D-dimer test result underwent compression ultrasonography. Patients with a positive ultrasound result started anticoagulant therapy, while patients with a negative result were not prescribed anticoagulant treatment.

Follow-Up

All patients underwent follow-up for 3 months. Patients were instructed to seek urgent care in case of recurrent leg symptoms or new respiratory symptoms. At the end of follow-up, all patients included in the study were interviewed by telephone by a study coordinator using a structured script. Patients were asked to disclose all health-related events since their index visit: consultation with any physician, admission to hospital, change in medications, diagnostic testing, or hemorrhagic complication. The family physician was contacted whenever a possible VTE event was disclosed, and medical records were reviewed if a patient was readmitted to the hospital for any cause.

Outcomes

The primary outcome was the rate of adjudicated symptomatic VTE, comprising leg DVT and/or pulmonary embolism, during the 3-month follow-up period among patients with a D-dimer value between the conventional cutoff of 500 µg/L and their age-adjusted cutoff (failure rate).

Secondary outcomes included the proportion of patients with a low-intermediate or unlikely pretest probability who had a D-dimer result between 500 µg/L and their age-adjusted cutoff. This group reflects the additional diagnostic yield of applying the age-adjusted cutoff. We estimated the 3-month thromboembolic risk specifically in this subgroup of patients.

We also analyzed the additional diagnostic yield of the age-adjusted D-dimer cutoff according to age groups (18-49, 50-64, 65-74, and ≥75 years).

Symptomatic VTE during follow-up was defined according to usual criteria: leg DVT was confirmed by an abnormal proximal compression ultrasonography result and pulmonary embolism by either a high-probability ventilation-perfusion scan or computed tomography pulmonary angiography showing segmental or more proximal intraluminal defects. Deaths were classified using the International Society on Thrombosis and Haemostasis (ISTH) definition for VTE-related death.¹¹ Only pulmonary embolism–related deaths, corresponding to the categories (A1-A2-A3) of the ISTH document, were included in the primary outcome. Other venous thromboses, such as cerebral, splanchnic, gonadal, or upper extremity vein thromboses, were not included in our primary outcome. All suspected VTE events and deaths were independently adjudicated by 2 blinded experts using the CanVECTOR Verdict web-based platform. Disagreements were resolved by a third adjudicator and, if needed, by a consensus meeting.

Statistical Analysis

Descriptive statistics included mean and standard deviations, or median with interquartile ranges for continuous variables, depending on their distribution, and proportions for categorical variables. We used the Wilson score method without continuity correction to compute 2-sided 95% CIs around estimated proportions. Sample size was estimated based on our previous retrospective validation dataset and pulmonary embolism management outcome study.^12,13 We aimed to include a sufficient number of patients to provide accurate estimates of our primary and secondary outcomes. To validate the safety of ruling out DVT on the basis of a D-dimer value between 500 µg/L and the age-adjusted cutoff, the upper limit of the 95% CI for the 3-month thromboembolic event rate needed to be below 3%. This failure rate corresponds to that observed after a negative phlebography¹⁴ and is widely accepted as an acceptable failure rate to validate diagnostic strategies for VTE.^15,16 Achieving this would require no more than 2 events in 240 patients in this subgroup. In our prior pulmonary embolism validation cohort, approximately 10% of patients older than 50 years with an unlikely clinical probability had a D-dimer value between 500 µg/L and their age-adjusted cutoff. Therefore, we estimated that 2400 patients older than 50 years with an unlikely or non-high clinical probability of DVT were required. Considering that two-thirds of patients were expected to be older than 50, a total of 3200 patients with suspected DVT needed to be enrolled. Given the lack of a specific guideline for reporting prospective diagnostic management outcome studies, we adhered to a STARD guideline. Statistical analyses were performed using SPSS version 30.0 (IBM Statistics).

Results

Between January 15, 2015, and October 6, 2022, 3205 patients were enrolled. General characteristics are depicted in Table 1.

Table 1. Characteristics of Included Patients.

Characteristic	No. (%) (N = 3205)
Sex
Female	1737 (54.2)
Male	1468 (45.8)
Age, median (IQR), y	59 (45-71)
BMI, median (IQR)a	28 (24-32)
Risk factors
Personal history of VTE	543 (16.9)
Surgery within 1 mo	313 (9.8)
Long travel	308 (9.6)
Active malignancy	277 (8.6)
Bed rest	256 (8.0)
Oral contraceptives	168 (5.2)
Signs and symptoms
Calf pain	1974 (61.6)
Pitting edema	1278 (39.9)
Increase in calf circumference >3 cm	845 (26.4)
Collateral veins (nonvaricose)	328 (10.2)
Alternative diagnosis at least as likely as DVT	1108 (34.6)
Pretest clinical probability
Non-high/unlikely clinical probability	2169 (67.7)
High/likely clinical probability	1036 (32.3)

Abbreviations: BMI, body mass index; DVT, deep vein thrombosis; VTE, venous thromboembolism.

^a Calculated as weight in kilograms divided by height in meters squared.

Diagnostic Workup

The study flow is summarized in Figure 1. A non-high or unlikely pretest probability was observed in 2169 patients (67.7%). Among these, 692 patients (31.9% [95% CI, 30.0%-33.9%]) had a negative D-dimer result according to the age-adjusted cutoff: 531 (24.5% [95% CI, 22.7%-26.4%]) had a D-dimer level less than 500 µg/L, and 161 additional patients (7.4% [95% CI, 6.4%-8.6%]) had a D-dimer level between 500 µg/L and their age-adjusted cutoff. Therefore, use of the age-adjusted cutoff resulted in a 7.4% (95% CI, 6.4%-8.6%) absolute increase, or a 23.3% (95% CI, 20.3%-26.6%) relative increase, in the proportion of negative D-dimer results. Figure 2 depicts the correlation between patients’ age and D-dimer levels, in those with and without DVT, and illustrates the absence of DVT or pulmonary embolism in patients with D-dimer level between 500 µg/L and the age-adjusted cutoff.

Figure 1. Flow of Study of D-Dimer Cutoff Levels to Rule Out Deep Vein Thrombosis.

Compression ultrasonography was repeated in case of an indeterminate result. Complete follow-up, no anticoagulants refers to the number of patients who were not prescribed anticoagulants, did not withdraw consent, were not lost to follow-up, and did not die during follow-up. DVT indicates deep vein thrombosis; VTE, venous thromboembolism.

Figure 2. Scatterplot of D-Dimer Results vs Age in Non-High/Unlikely Clinical Probability Patients.

Horizontal line indicates the conventional 500-µg/L cutoff value; diagonal line, the age-adjusted cutoff (age x 10, in µg/L, in patients aged ≥50 years). Patients with D-dimer values greater than 2000 µg/L were omitted for readability. DVT indicates deep vein thrombosis.

The most commonly used assays were the VIDAS Exclusion D-dimer test (n = 749), the Innovance (n = 719), the D-dimer HS 500 (n = 342), and the Liatest (n = 205). Other assays were used in fewer than 50 patients each.

Further testing with compression ultrasonography was required in the 1418 patients with a D-dimer level above the age-adjusted cutoff, in the 59 patients in whom D-dimer testing was not performed (protocol deviations), and in the 1036 patients with a high or likely clinical probability of DVT. Moreover, despite a negative D-dimer test result, 185 patients with D-dimer level less than 500 µg/L and 94 patients with D-dimer level between 500 µg/L and the age-adjusted cutoff underwent compression ultrasonography. One patient with a D-dimer level less than 500 µg/L was diagnosed with a proximal DVT.

Overall, a proximal DVT was diagnosed in 432 of 3205 patients (13.5%): 148 of 2169 patients (6.8%) with non-high/unlikely clinical probability and 284 of 1036 patients (27.4%) with high/likely clinical probability.

In addition, a DVT limited to distal veins was diagnosed in 11 of the 185 patients who had a D-dimer level less than 500 µg/L, in 0 of 94 patients with D-dimer level between 500 µg/L and the age-adjusted cutoff, in 170 patients of 1477 with a non-high/unlikely clinical probability and positive D-dimer results, and in 87 patients of 1036 patients with a high/likely clinical probability. The overall prevalence of distal and proximal DVT at index visit was 21.8% (700/3205).

Three-Month Follow-Up

Patients With a D-Dimer Level Less Than 500 µg/L

During the 3-month follow-up period, of the 531 patients with a D-dimer level less than 500 µg/L, 12 patients received anticoagulants during follow-up: 11 were treated for a distal DVT and 1 for another indication. Fourteen patients were lost to follow-up, and 2 withdrew consent. Among the 503 remaining patients, 1 died from a cause unrelated to venous thromboembolism, and 14 had suspected VTE during follow-up. Of these 14 events, 1 was adjudicated as having a confirmed proximal DVT, contralateral to the initial suspicion. Thus, the 3-month risk of venous thromboembolism among patients with a D-dimer level less than 500 µg/L who completed the 3-month follow-up without anticoagulants was 1 in 502 (0.2% [95% CI, 0.0%-1.1%]) (per-protocol analysis). In the intention-to-diagnose analysis, the failure rate was 2 in 531 (0.4% (95% CI, 0.1%-1.4%]).

Patients With a D-Dimer Level Between 500 µg/L and Their Age-Adjusted Cutoff

Of the 161 patients with a D-dimer level between 500 µg/L and their age-adjusted cutoff, no patient was lost to follow-up and 2 patients received anticoagulant treatment for another indication than VTE. Of the remaining 159 patients, none died, and 5 underwent testing for suspected VTE during follow-up. None of these patients had an adjudicated confirmed VTE. Therefore, in the intention-to-diagnose analysis, the failure rate was 0 in 161 (0.0% [95% CI, 0.0%-2.3%]). The per-protocol analysis showed similar results (0/159; 0.0% [95% CI, 0.0%-2.4%]).

Patients With a D-Dimer Level Above the Age-Adjusted Cutoff

Of the 1477 patients with a D-dimer level above the age-adjusted cutoff or in whom D-dimer testing was not performed, a proximal DVT was diagnosed at initial testing in 147 patients. Of the remaining 1330 patients, 15 were lost to follow-up, 1 withdrew consent, and 211 were given anticoagulants (170 for distal DVT, 33 for superficial vein thrombosis, 8 for another indication). Of the remaining 1103 patients, 14 died from a cause other than VTE, and 47 presented a suspected VTE during follow-up. Five of these 47 suspected events were adjudicated as confirmed or possible events (pulmonary embolism, n = 4; DVT, n = 1). The 3-month risk of VTE among patients with a D-dimer level greater than their age-adjusted cutoff who completed the 3-month follow-up without anticoagulants was 5 in 1089 (0.5% [95% CI, 0.2%-1.1%]) (per-protocol analysis). In the intention-to-diagnose analysis, the failure rate was 5 in 1330 (0.4% [95% CI, 0.2%-0.9%]).

Patients With a High or Likely Clinical Probability

Of the 1036 patients with likely or high clinical probability, a proximal DVT was diagnosed at initial testing in 284. Among the 752 patients with no proximal DVT, 9 were lost to follow-up, 2 withdrew consent, and 109 were given anticoagulants (87 for distal DVT, 16 for superficial vein thrombosis, and 6 for another reason). Of the remaining 632 patients, 15 died of another cause than VTE, and 31 presented with a suspicion of a thromboembolic event. Six of these 31 suspected events were adjudicated as confirmed or possible VTE events (pulmonary embolism, n = 3; DVT, n = 3). The 3-month risk of VTE among patients with a high/likely pretest probability who completed the 3-month follow-up with no anticoagulants was 6 in 617 (1.0% [95% CI, 0.5%-2.1%]) (per-protocol analysis). In the intention-to-diagnose analysis, the failure rate was 6 in 752 (0.8% [95% CI, 0.4%-1.7%]). The overall failure rate was 13 in 2774 (0.5% [95% CI, 0.3%-0.8%]) in the intention-to-diagnose analysis and 12 in 2367 (0.5% [95% CI, 0.3%-0.9%]) in the per-protocol analysis.

Subgroup Analysis According to Patient Age Groups

Table 2 shows main study results stratified by age groups. The proportion of non-high/unlikely clinical probability patients with a D-dimer level below 500 µg/L was 59.6% in patients younger than 50 years but only 8.7% in patients older than 75 years. Using the age-adjusted cutoff led to an absolute increase in the proportion of patients with negative D-dimer result of 6.1% among those aged 50 to 64 years, 14.8% among those aged 65 to 74 years, and 17.4% among those 75 years or older. Overall, 597 (18.6%) patients were 75 years or older; of these, 379 (63.5%) had a non-high or unlikely pretest clinical probability. The proportion of patients with D-dimer level less than 500 µg/L was 33 of 379 (8.7%). Another 66 patients (17.4%) had a D-dimer level between the conventional cutoff and their age-adjusted cutoff. Therefore, the proportion of patients older than 75 years with a non-high or unlikely clinical probability and a negative D-dimer result using the age-adjusted cutoff was 99 of 379 (26.1%). Of those patients, 1 received anticoagulant therapy for an indication other than VTE. There were no patients lost-to follow up, no deaths, and none withdrew consent. One patient of 99 with a D-dimer less than 500 µg/L had a confirmed VTE (proximal DVT contralateral to the index initial suspicion) during follow-up (1.0% [95% CI, 0.2%-5.5%]).

Table 2. Subgroup Analysis of the Primary Outcome According to Age.

Age, y	No.	No. (%)			Absolute increase in the proportion of negative D-dimer (age-adjusted vs 500 µg/L), %	3-mo VTE rate, non-high/unlikely clinical probability and D-dimer < age-adjusted cutoff, No. (%) [95% CI]
		Non-high/unlikely clinical probability	D-dimer <500 µg/L among non-high/unlikely	D-dimer < age-adjusted among non-high/unlikely
18-49	1010	770 (76.2)	290 (59.6)	290 (59.6)	0	1 (0.3) [0.1-0.9]
50-64	1001	641 (64.0)	153 (23.9)	192 (30.0)	6.1	0 [0.0-2.0]
65-74	597	379 (63.5)	55 (14.5)	111 (29.3)	14.8	0 [0.0-3.4]
≥75	597	379 (63.5)	33 (8.7)	99 (26.1)	17.4	1 (1.0) [0.2-5.5]

Abbreviation: VTE, venous thromboembolism.

Two- vs 3-Level Wells Score

Table 3 shows the breakdown of patients managed according to the 2-level vs the 3-level Wells score. More patients were eligible for D-dimer testing when using the 3-level Wells score, but the chance of a negative result was lower. The failure rates were 1 in 209 (0.5% [95% CI, 0.1%-2.7%]) with the 2-level score and 1 in 483 (0.2% [95% CI, 0.0%-1.2%]) with the 3-level score.

Table 3. Two- vs 3-Level Wells Score.

Wells score	No. (%)	No. (%)		No./total (%)			3-mo VTE rate in patients with a negative age-adjusted D-dimer No./No. (%) [95% CI]
		Confirmed DVT	Non-high/unlikely clinical probability	Negative D-dimer in patients with a non-high/unlikely probability	Non-high/unlikely and negative D-dimer	Patients requiring compression ultrasonography
In 2 categories (North America)	1181 (36.8)	180 (15.2)	561 (47.5)	209/561 (37.3)	209 (17.7)	972/1181 (82.3)	1/209 (0.5%) [0.1%-2.7%]
In 3 categories (Europe)	2024 (63.2)	252 (12.5)	1608 (79.4)	483/1608 (30.0)	483 (23.9)	1541/2024 (76.1)	1/483 (0.2%) [0.0%-1.2%]

Abbreviations: DVT, deep vein thrombosis; VTE, venous thromboembolism.

Discussion

Among outpatients with suspected leg DVT, using the age-adjusted D-dimer cutoff (defined as D-dimer level < patient’s age × 10 in µg/L in patients ≥50 years) allows to safely rule out DVT. The failure rate, as defined by the 3-month rate of symptomatic VTE, was low in patients with a negative age-adjusted D-dimer level (2/692; 0.3% [95% CI, 0.1%-1.1%]). There were no events among the 161 patients with a D-dimer level greater than 500 µg/L but below their age-adjusted D-dimer cutoff (0/161; 0.0% [95% CI, 0.0%-2.3%]).

These results are consistent with those from prospective trials evaluating the age-adjusted D-dimer testing in patients with suspected pulmonary embolism.^3,4 While several retrospective studies in patients with suspected DVT have supported the potential safety and clinical usefulness of the age-adjusted cutoff across various assays and settings,^17,18 a prospective outcome study, in which treatment of patients with suspected DVT would be managed without anticoagulants on the basis of a negative age-adjusted D-dimer test result has been lacking.

Using the age-adjusted D-dimer threshold significantly increases the number of patients 75 years or older who can safely avoid compression ultrasonography—from 5.5% using the conventional 500-µg/L cutoff to 16.6% using the age-adjusted threshold. Put differently, while only 1 in 18 older adults could be ruled out for DVT using the standard cutoff, this increased to 1 in 6 patients older than 75 years with the age-adjusted approach. This effectively restores the diagnostic yield of the D-dimer test in older adults to levels more comparable to those of the general population. In clinical practice, this has meaningful implications: enabling more patients to be discharged without imaging, reducing emergency department length of stay, and avoiding unnecessary empiric anticoagulation. Importantly, these advantages were not offset by any compromise in diagnostic safety.

Other efforts have been made to spare imaging in patients with suspected VTE, using clinical probability–adjusted D-dimer cutoffs. For instance, in patients with suspected pulmonary embolism, higher D-dimer thresholds (eg, 1000 µg/L) have been prospectively validated in patients with a low pretest probability.^19,20 However, similar efforts have been limited in the DVT setting. A prospective study, the 4D study,²¹ applied a 1000-µg/L cutoff in patients with a low clinical probability of DVT and showed a 3-month VTE rate of 0.6% (95% CI, 0.3%-1.2%).

This study has several strengths. This was a large international collaboration. All suspected VTE events and deaths during follow-up were adjudicated by an independent committee blinded to patient’s initial D-dimer result. Another strength is the choice of focusing on the subgroup of patients with a D-dimer level 500 µg/L or greater but below their age-adjusted cutoff as the primary outcome, which allowed assessing the safety in this specific subgroup of interest without diluting the result by considering the whole population of patients including those with a D-dimer level less than 500 µg/L. Beyond reducing unnecessary imaging and conserving health care resources, this study provides important prospective validation of the age-adjusted cutoff to safely rule out leg DVT and pulmonary embolism. Many centers have been hesitant to implement the age-adjusted cutoff in their emergency departments due to concerns about inconsistent cutoffs for DVT and pulmonary embolism. Acknowledging that other challenges to implementation may exist, including the lack of standardization across institutions or medicolegal hesitations in deviating from the conventional 500 µg/L cutoff, this study supports the consistent use of the age-adjusted cutoff in patients with suspected pulmonary embolism and those with suspected leg DVT.

Limitations

Among study limitations, 11 different commercial D-dimer assays were used. Therefore, not all included patients had treatment managed using the exact same diagnostic tests. Acknowledging that standardization of D-dimer assays remains a challenge,²² the study only used highly sensitive D-dimer assays, which have been reported to have comparable diagnostic performances.^3,23 Moreover, this could increase the generalizability of the findings to a wide number of settings. This may also be said about the assessment of clinical probability, which was performed by either the 2- or 3-level Wells score, without impacting the failure rate.

Second, this study was not designed as a randomized controlled study. Therefore, the 3-month VTE risk could not be compared with that of a control group whose treatment would have been managed using the conventional 500-µg/L cutoff. However, given the low rate of VTE events, a significant difference between the 2 strategies would have been unlikely. Moreover, the use of the 3-month VTE risk is widely used as the standard reference for the validation of VTE diagnostic strategies.

Third, compared with our previous pulmonary embolism study, a lower-than-expected number of patients had a D-dimer level 500 µg/L or greater but below their age-adjusted cutoff. The current study population was younger than the one enrolled in the ADJUST-PE study (median age, 57 years vs 63 years), resulting in a lower number of patients in whom the cutoff was adjusted based on age, which could account for this difference.

Fourth, study recruitment was longer than anticipated, due to challenges in securing peer-reviewed funding in each participating country and obtaining regulatory and ethical approvals at each site. Recruitment in Canada and Switzerland started in 2015, but sites in Belgium and France joined the study in 2019 and 2020, respectively. The COVID-19 pandemic later halted the study at most participating sites.

Fifth, several patients with negative D-dimer results underwent compression ultrasonography at initial assessment (185/531 in patients with a D-dimer level less than 500 µg/L and 94/161 in patients with D-dimer between 500 µg/L and their age-adjusted cutoff). The reason why a physician would decide to override the study diagnostic algorithm was not collected. One can speculate that physicians remained concerned about a possible DVT despite a negative D-dimer result or wanted to rule out an alternative diagnosis. These additional tests led to the diagnosis of 11 distal DVTs and 1 proximal DVT in patients with a D-dimer level less than 500 µg/L. No proximal or distal DVT was diagnosed in patients with a D-dimer level between 500 µg/L and their age-adjusted cutoff. The clinical relevance, management, and prognosis of distal DVT, a fortiori with a negative D-dimer result, are widely debated.²⁴ Many studies previously demonstrated a good outcome in patients with suspected DVT and D-dimer level below 500 µg/L,^25,26,27 despite others reporting a limited sensitivity of D-dimer tests for distal DVT.²⁸ Clinical practice guidelines strongly endorse the safety of not treating with anticoagulants patients with suspected DVT, a non-high pretest probability, and negative highly sensitive D-dimer result.^1,29

Last, it is important to mention that these results cannot be extrapolated to other thrombosis locations/types, eg, upper extremity, cerebral, splanchnic, or gonadal vein thromboses.

Conclusions

A negative D-dimer result using the age-adjusted D-dimer cut-off may safely rule out leg DVT in outpatients with a non-high or unlikely pretest probability, with a very low risk of subsequent symptomatic venous thromboembolism. The age-adjusted cutoff significantly increases the proportion of patients in whom DVT can be excluded without the need for leg compression ultrasonography.

Supplement 1.

Trial Protocol

Supplement 2.

eTable. The Simplified Wells Score

Supplement 3.

ADJUST-DVT Investigators

Supplement 4.

Data Sharing Statement