A certain diagnosis of acute PE or DVT can only be established by means of imaging . On the other hand, ruling out acute VTE can be safely achieved by using a standardized clinical probability assessment and a D-Dimer blood test. The main advantage of such an approach is a reduction in the required number of imaging tests, which are in general time consuming, costly and associated with radiation exposure and other complications.
Fibrin D-dimer is the final product of the plasmin-mediated degradation of cross-linked fibrin. Its plasma concentration is dependent on fibrin generation and subsequent degradation by the endogenous fibrinolytic system . D-dimer levels are typically elevated in patients with acute venous thrombosis. Accordingly, the sensitivity of an elevated D-dimer concentration for VTE is very high . In contrast, as D-dimer levels can be elevated in other clinical conditions that are associated with enhanced fibrin (e.g. malignancy, trauma, increased age, disseminated intravascular coagulation, inflammation, infection, sepsis, postoperative states and pre-eclampsia), the specificity for acute thrombosis is rather poor . Thus the diagnostic strength of D-dimer tests lies in ruling out DVT or PE.
A large variety of assays are available for D-dimer measurement, and they are all based on the use of monoclonal antibodies that recognize epitopes of the D-dimer fragment that are not present on fibrinogen or non-crossed fragments of fibrin . From numerous accuracy and management studies, it has been demonstrated that the sensitivities of the D-dimer enzyme-linked immunofluorescence assay (ELFA) (DVT 96%; PE 97%), microplate enzyme-linked immunosorbent assay (ELISA) (DVT 94%; PE 95%) and latex quantitative assay (DVT 93%; PE 95%) are superior to those of the whole-blood D-dimer assay (DVT 83%; PE 87%), latex semiquantitative assay (DVT 85%; PE 88%) and latex qualitative assay (DVT 69%; PE 75%), and have virtually no interobserver variability. The latex qualitative (DVT 99%, PE 99%) and whole-blood D-dimer assays (DVT 71%; PE 69%) have the highest specificities and are more operator dependent [3, 4]. Higher sensitivity yields a higher negative predictive value and thus less concern regarding false-negative test results. As no D-dimer assay has a perfect 100% sensitivity, the use of the D-dimer test should be restricted to patients with a non-high clinical probability.
In general, the D-dimer threshold for a normal test result is 500 μg L−1. In non-thrombotic clinical conditions associated with increased fibrin formation or decreased D-dimer clearance, the specificity of D-dimer testing with this particular cut-off is unquestionably decreased. For instance, in patients with active malignancy, the specificity was 16%, compared with 41% in patients without cancer, with a comparable sensitivity of 100% . Analogous numbers were reported in elderly patients, patients with renal function impairment and those with a high inflammatory state [6-8]. Hence, it is likely that in such cases a higher cut-off results in higher specificity without a relevant fall in sensitivity. Indeed, the results of a retrospective analysis of two cohorts with suspected PE totalling 1331 patients aged over 50 years old suggests that an age dependent cut-off defined as patient's age × 10 μg L−1 is safe for use in clinical practice . Acute PE could be excluded in 42% with the new cut-off value in combination with an unlikely clinical probability, compared to 36% with the standard cut-off value (< 500 μg L−1), without radiological imaging. A second and third post-hoc analysis in different patient populations provided further validation of the use of such an age-dependent D-dimer threshold [9, 10]. No large studies evaluating other corrections for clinical circumstances associated with elevated D-dimer concentrations have been performed. A recent study, in which doubling of the threshold for a positive D-dimer to 1000 μg L−1 was evaluated in 678 consecutive patients with an unlikely clinical probability for PE, independent of other co-morbid conditions or age, indicated that the number of patients in whom PE could be ruled out without the need for a CT scan as well as the rate of false-negative D-dimer tests, roughly doubled . The failure rate of the combination of an unlikely clinical probability and a D-dimer of < 1000 μg L−1 was 5.3% (11 of 208 patients). Of note, 10 of 11 missed PEs were subsegmental. Nonetheless, the safe use of an age-dependent or other altered D-dimer threshold should be confirmed in a prospective management study before it can be implemented in daily clinical care, both for PE and DVT.
Clinical decision rules
The introduction of clinical decision rules has led to a standardized evaluation of the clinical pre-test probability in patients with suspected PE or DVT. For DVT, several decision rules are available . Nonetheless, the first published one by Wells and colleagues is the most widely validated and used [12, 13]. It utilizes information from medical history and physical examination and consists of nine items. One point is given for each item and two points are deducted when an alternative diagnosis is considered more likely than DVT (Table 1). The decision rule initially categorized patients into low (0 points; 4.0–8.0% risk), intermediate (1–2 points, 13–23% risk) and high pre-test probability (≥ 3 points, 44–61% risk), and was later dichotomized into a low risk (< 2 points, 3.8–7.6% risk) or high risk category (≥ 2 points, 24–32% risk) for practical purposes [13, 14]. Notably, because of suggested disadvantages of the Wells rule, for example the lack of complete objectiveness due to the subjective element of considering an alternative diagnosis and lack of validation in selected cohorts including pregnant patients and the elderly, several alternative rules have been proposed . Nonetheless, the reported inter-observer variability of the Wells score is good (κ = 0.85), and it has been shown that assessing the score is independent of the physicians' experience [13, 15]. Moreover, as alternative decision rules have not been validated in prospective management studies as extensively as the Wells rule, we recommend using the Wells rule for DVT. As stated before, because DVT cannot be ruled out by either a D-dimer test or clinical probability assessment alone, both tests should be used in conjunction. Approximately 40% of patients will have the combination of an unlikely probability (Wells score ≤ 2 points) and a normal D-dimer test result, yielding a negative likelihood ratio of 0.08 for the presence of DVT, in which case it is safe to withhold anticoagulant therapy without further testing .
Table 1. Wells rule for DVT
|Calf swelling ≥ 3 cm compared with asymptomatic calf||1|
|Swollen unilateral superficial veins||1|
|Unilateral pitting edema||1|
|Previous documented DVT or PE||1|
|Swelling of entire leg||1|
|Localized tenderness along the deep venous system||1|
|Recently bedridden ≥ 3 days or major surgery||1|
|Alternative diagnosis at least as likely as DVT||−2|
| High||> 2|
| PE unlikely||≤ 1|
| PE likely||> 1|
Regarding decision rules for acute PE, two are extensively validated in several outcome studies, the Wells rule for PE and the revised Geneva score. The first one consists of seven variables (Table 2), including a judgement on whether PE is the most likely diagnosis . As with the Wells rule for DVT, this latter subjective item is the one that is most criticized, not least because it carries a major weight in the score. Using this rule, patients are classified as low (< 2 points; 2.0–5.9% risk), intermediate (2–6 points, 17–24% risk) or high pre-test probability (≥ 6 points, 54–78% risk), or alternatively as PE unlikely (≤ 4 points, 2.3–9.4% risk) or PE likely (> 4 points, 28–52% risk) . The Christopher investigators showed that the combination of an unlikely clinical probability and a normal quantitative D-dimer test safely ruled out the presence of PE with a low 3-month VTE recurrence rate of 0.49% . A meta-analysis including all high-quality prospective studies investigating the safety of ruling out PE based on a normal D-dimer blood test result and an unlikely clinical probability according to the Wells rule, confirmed this very low risk in 1660 consecutive patients with a pooled negative predictive value of 99.7% (95% CI, 99.0–99.9) and a very low PE-related mortality risk of 0.06% (95% CI, 0.0017–0.46) . In contrast, patients with a likely clinical probability should undergo further testing regardless of the D-dimer test outcome as venous thromboembolism can be diagnosed in 9.3% (95% CI, 4.8–17) of patients with a negative D-dimer test result in this population . Several attempts to construct a more objective decision rule were made. Of these, the revised Geneva score is the best validated rule and contains nearly the same items (Table 2), except for a clinical judgement of the likelihood of PE [20, 21]. This rule also has both a three- and a two-level outcome: low (< 3 points; 6.6–13% risk), intermediate (3–11 points, 24–31% risk) or high pre-test probability (≥ 11 points, 58–82% risk), and PE unlikely (≤ 2 points, 13–19% risk) or PE likely (> 2 points, 28–35% risk) [20-22]. In combination with a normal D-dimer test result, acute PE can be ruled out with high certainty in patients with a non-high (0% PE; 95% CI, 0.0–2.2) or less likely probability (0.54% PE; 95% CI, 0.0–3.0) according to the revised Geneva score [21, 22].
Table 2. Clinical decision rules for PE
|Wells rule||Revised Geneva score|
|Previous PE or DVT||1.5||1||Previous DVT or PE||3||1|
|Heart rate > 100/min||1.5||1||Heart rate 75–94/min||3||1|
| || || ||Heart rate ≥ 95/min||5||2|
|Surgery or immobilization < 4 weeks||1.5||1||Surgery or fracture within 1 month||2||1|
|Active malignancy||1||1||Active malignancy||2||1|
|Clinical signs of DVT||3||1||Unilateral lower limb pain||3||1|
|Alternative diagnosis less likely than PE||3||1||Pain on lower limb deep vein palpation and unilateral edema||4||1|
| || || ||Age > 65 years||1||1|
|Clinical probability||Clinical probability|
| Low||< 2|| || Low||0–3|| |
| Intermediate||2–6|| || Intermediate||4–10|| |
| High||> 6|| || High||≥ 11|| |
| PE unlikely||≤ 4||≤ 1|| PE unlikely||≤ 5||≤ 2|
| PE likely||> 4||> 1|| PE likely||> 5||> 2|
For practical purposes, both rules have been simplified by assigning only one point to each item (Table 2), without a resulting decrease in diagnostic accuracy [23, 24]. Several recent meta-analyses regarding the clinical utility of the above-discussed clinical decision rules in the management of acute PE have been published [19, 25-27]. They all conclude that the available decision rules show similar accuracy. However, a formal prospective management study comparing the (simplified) Wells rule and (simplified) revised Geneva score was lacking until recently. In the Prometheus study, both the original and the simplified Wells rule and revised Geneva score were directly compared in 807 consecutive patients . The four decision rules showed similar performance for exclusion of acute PE in combination with D-dimer testing. The 3-month venous thromboembolic recurrence rates of all four scores ranged between 0.5% and 0.6%, while 30% of patients could be managed without the need for imaging. Therefore, the authors concluded that either the simplified or original Wells and Geneva decision scores can be used in clinical practice with equal safety and clinical utility. The specific rule that is used should depend on local preference.
In summary, the initial diagnostic work-up of acute thrombosis is dependent on several non-invasive diagnostic tests that should be used sequentially, always starting with probability assessment using a validated clinical decision rule, followed by a quantitative D-dimer test in case of an unlikely or non-high pre-test probability. D-dimer testing in patients with a likely or high probability is redundant for diagnostic purposes. Acute DVT or PE can be ruled out in patients with the combination of a non-high or unlikely clinical probability and normal D-dimer test result. All other patients should be referred to a radiologist for an imaging test. As use of such a validated diagnostic algorithm is associated with lower healthcare costs and a decreased complication risk, we recommend implementation of such a standardized approach. Regrettably and despite clear international guidelines [28, 29] in addition to overwhelming evidence, adherence to guidelines in the real world of clinical practice is poor [30-32]. For instance, in a random record review of Canadian patients in whom a D-dimer test was ordered because of a suspected acute venous thrombosis, pre-test assessment was not documented in 64% of the cases . In the case of a positive d-dimer result, further imaging was not performed in 25–42% of patients, independent of clinical probability [30, 31]. In a large nationwide German assessment of the utilization of diagnostic methods for suspected DVT, almost every patient received an imagining examination, whereas clinical pre-test estimation was not used at all. Lastly, D-dimer tests were used as an adjunct to imaging rather than as a tool for exclusion of the disease without imaging .