Thrombosis Research 134 (2014) 763–768
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Review Article
Clinical Decision Rules and D-dimer in Venous Thromboembolism: Current controversies and future research priorities Marc A. Rodger a,b,c,d,⁎, Gregoire Le Gal a,c,d, Philip Wells a,c,d, Trevor Baglin e, Drahomir Aujesky f, Marc Righini g, Gualtiero Palareti h, Menno Huisman i, Guy Meyer j a
Hematology, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada Medicine, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada Obstetrics and Gynaecology, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada d Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON Canada e Cambridge Haemophilia and Thrombophilia Centre, Addenbrookes Hospital, Cambridge University Hospitals NHS Trust, Cambridge, UK f Division of General Internal Medicine, Bern University Hospital, Bern, Switzerland g Division of Angiology and Hemostasis, Department of Medical Specialties, Geneva University Hospital and Faculty of Medicine, Geneva, Switzerland h Unit of Angiology and Blood Coagulation, University Hospital of Bologna, Italy i Department of Thrombosis and Hemostasis, LUMC, Leiden, the Netherlands j Université Paris Descartes Sorbonne Paris Cité and Hopital europeen Georges Pompidou APHP, Paris, France b c
a r t i c l e
i n f o
Article history: Received 14 May 2014 Received in revised form 25 July 2014 Accepted 27 July 2014 Available online 4 August 2014 Keywords: Clinical Decision Rules D-dimer Diagnosis Treatment Prognosis Venous Thromboembolism
a b s t r a c t Venous thromboembolism (VTE) is a potentially lethal clinical condition that is suspected in patients with common clinical complaints, in many and varied, clinical care settings. Once VTE is diagnosed, optimal therapeutic management (thrombolysis, IVC filters, type and duration of anticoagulants) and ideal therapeutic management settings (outpatient, critical care) are also controversial. Clinical prediction tools, including clinical decision rules and D-Dimer, have been developed, and some validated, to assist clinical decision making along the diagnostic and therapeutic management paths for VTE. Despite these developments, practice variation is high and there remain many controversies in the use of the clinical prediction tools. In this narrative review, we highlight challenges and controversies in VTE diagnostic and therapeutic management with a focus on clinical decision rules and D-Dimer. © 2014 Elsevier Ltd. All rights reserved.
Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodologic Standards for Clinical Predictors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: Pretest Probability Assessment for Diagnostic Management of DVT or PE: Simplified Rules or The Original Rules? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: Traditional or Age-Adjusted D-Dimer Cut Off for DVT or Pulmonary Embolism Diagnostic Management? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: VTE Diagnosis and Pregnancy: Use of D-Dimer and Clinical Decision Rules . . . . . . . . . Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE: Duration of Anticoagulation . . . . . . . . . . . . Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE- PE Risk Stratification for Selection of PE Treatment Setting Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of Interest Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
764 764
. . . . .
764
. . . . . . . . .
765 766 766 767 767 767 767 767 767
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
⁎ Corresponding author at: The Ottawa Hospital, Ottawa Blood Disease Center, 501 Smyth Road, Box 201; Ottawa, ON, Canada, K1H 8L6. Tel.: +613 737 8899x74641; fax: +613 739 6102. E-mail address: [email protected] (M.A. Rodger).
http://dx.doi.org/10.1016/j.thromres.2014.07.031 0049-3848/© 2014 Elsevier Ltd. All rights reserved.
764
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
Introduction Venous thromboembolism (VTE), comprising both deep vein thrombosis (DVT) and pulmonary embolism (PE), are common, potentially lethal yet treatable clinical conditions [1]. Clinical Decision Rules (CDRs) are decision making tools, using combinations of simple available clinical predictors to define an outcome in the present, in other words a diagnosis (or a probability of disease), or an outcome in the future, in other words a prognosis (or probability of an outcome), either of which leads to a diagnostic course of action or a therapeutic course of action [2]. CDRs and/or D-dimer have a crucial place in the diagnostic and therapeutic management of VTE. These tools permit us to judiciously and safely use diagnostic imaging for VTE diagnosis, select the right treatment setting for initial therapeutic management of VTE (intensive care unit, ward or home) and select which patients will derive net benefit from anticoagulant therapies. The focus of this narrative review will be to review current controversies in the use of clinical decision rules and/or D-dimer in the diagnostic and therapeutic management of VTE. We will highlight wide international practice variations, even among experts in thrombosis, and also highlight challenges in knowledge translation for most clinicians that impact putting research findings into routine clinical practice. These challenges are largely reflective of the absence of level IA evidence to guide practice in these areas and the need to develop and validate simple and widely available tools that are easy to adopt in routine clinical practice. Methodologic Standards for Clinical Predictors High quality evidence to support routine use of CDRs, and other clinical prediction tools, requires that they are developed and validated strictly following methodologic guidelines (see Table 1). Standards for their development and validation were first published more than 20 years ago [3], were updated and which have formed the basis for a quality assessment framework we should consider before adopting CDRs in our daily practice [2]. These standards require adherence to methodologic guidelines at the development, validation and impact analysis stage (see Table 1). If this is the case, in the long run, they usually perform up to expectations. However, on the other end of the spectrum, many CDRs are developed with minimal adherence to methodologic guidelines, are never validated and ultimately, generally, these rules fail to meet expectations. Similarly, single predictors (e.g. Ddimer), can be viewed as the simplest of CDRs and their use in routine clinical practice should be limited to those clinical applications where validation and impact analysis have also been conducted. CDRs select and combine the best independent predictors (risk factors, symptoms, clinical signs and results of simple diagnostic tests) for a diagnosis or prognosis. The most useful CDRs are accurate, reproducible, simple and easy to apply. CDRs should be sensible i.e. have a clear purpose (e.g. exclude DVT), be relevant (e.g. exclude clinically important DVT), demonstrate content validity (e.g. be composed of well recognised independent predictors), be concise (e.g. simple rules containing limited items will be remembered), and be easy to use in the intended clinical application (e.g. don’t require a computer to calculate at the bedside). The use of the rule should provide a probability of disease (e.g. N10% likelihood of DVT) or prognosis (e.g. 15% risk of recurrent VTE) and should imply a course of diagnostic (e.g. needs an ultrasound to rule out DVT) or therapeutic action (e.g. continue anticoagulants indefinitely). Construction of valid, accurate and reproducible CDRs follows a strict methodology. Levels of evidence can be attributed to CDRs depending on whether or not they have gone through all the methodological steps (Levels 1 to 4). Level 4 corresponds to a rule that is derived but not prospectively validated: it needs to be further evaluated before clinical application. Validation of the CDR in an independent cohort of patients is a crucial next step. In fact, the rule is built as a "tailored suit"
Table 1 Methodologic characteristics and criteria required to develop and validate clinical decision rules and why they are important to follow. Stage
Methodologic Criteria
Why?
Development Describe patient characteristics
To ensure generalizability and applicability of the results to reader’s clinical practice Broad spectrum of disease To ensure that CDR can segregate patients with subtle or not obvious disease To ensure that the CDR predicts Outcome clearly defined, important, assessed blindly and the right outcome in an unbiased manner is the gold standard To ensure all potential predictors Complete set of potential can enter CDR. To ensure that predictors that are clearly predictors collected in a defined and blindly and reproducible, accurate and prospectively collected unbiased manner. To maximise generalizability Reproducibility of potential predictors assessed only reliable predictors (Kappa N0.6) should enter rule. To ensure adequate power so Statistical techniques used to important predictors have derive the rule are identified narrow estimates of effect. To and valid ensure rule is not “overfit” and only works in the derivation cohort To ensure the CDR is easy to use Final rule chosen based on simplicity and ability to provide and remember so it is widely adoptable. To ensure thresholds a clear course of action. for dividing outcomes are chosen for optimal repartition of patients across groups and optimal proportion of patients with outcomes in each group. Validation Apply the rule in an To check the proportions of independent cohort of patients patients classified by the rule in each clinical probability group. Impact Determine impact of use of the To determine if the rule is used. assessment rule in the real world To determine the rule’s safety, accuracy and clinical utility in real world use.
for the derivation cohort. In other words, it is important to ensure that the “suit” will fit all the populations in which it is intended to be applied. Level 3 rules have been prospectively validated but in only one narrow sample: physicians may consider their use with caution and only if patients in their clinical setting are similar to those included in validation study. Level 2 rules have demonstrated their accuracy in either a large prospective study including a broad spectrum of the disease, or in several different smaller settings. They can be used in various settings with confidence in their accuracy. Finally, the impact of use of the rule, its clinical utility and safety of managing patients on the basis of the rule “in the real world” should be demonstrated in a prospective management outcome study. In these studies, use of the rule in usual practice is measured along with the performance of the rule in usual practice. Level 1 rules have been prospectively validated in a different population and the impact of use of the rule has been measured and demonstrates a change in clinician behaviour with beneficial consequences. Level 1 rules can be used with the confidence that they can change clinician behaviour and improve patients' outcomes.
Challenges and Controversies in VTE Diagnosis: Pretest Probability Assessment for Diagnostic Management of DVT or PE: Simplified Rules or The Original Rules? The original Wells DVT CDR and the Wells PE CDR in addition to the Geneva CDR for PE, have been well evaluated in clinical research over the last 2 decades [4]. Nonetheless it is evident from published practice patterns that clinicians often do not use these rules [5]. The risk of this underuse is either over or under diagnosis of VTE [5]. Some of the
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
knowledge translation barriers that have been highlighted include 1) the need to recall more than 7 items with each of these rules, 2) the need to recall the points assigned for each predictor in each of these rules, 3) the need to calculate the total number of points and 4) with the Wells DVT CDR and Wells PE CDR the “alternative diagnosis” item which requires subjective assessment that may lead to discomfort for less experienced physicians. These challenges/complexities may dissuade practicing clinicians from using these rules. However, new technology including bedside applications (“apps”) offer one solution to these recall, point assignment and calculation complexities and as younger “tech friendly” generations become involved in clinical practice perhaps these considerations will become less important. Simplified rules have also been developed in order to reduce the complexities associated with these rules. In a prospective study the simplified decision rules – either simplified Wells rule or simplified Geneva rule – worked equally well; total failure rates of the simplified CDR’s and D-dimer combinations were similar (1 failure, 0.5% to 0.6% [upper-limit 95% CI, 2.9% to 3.1%]) [6]. The use of these simplified rules, after additional validation, will hopefully lead to better and more prevalent use of the diagnostic CDRs. However, there is some skepticism about whether they are simple enough (e.g. still involve recall of N 7 predictors) and that they will be more widely used than the original rules. Challenges and Controversies in VTE Diagnosis: Traditional or AgeAdjusted D-Dimer Cut Off for DVT or Pulmonary Embolism Diagnostic Management? After pre-test probability assessment, clinicians can use D-dimer as a next step to determine if imaging is required or VTE is excluded. Epidemiologic research has demonstrated that in the CT scan era, the diagnosis of PE is increasing in frequency with a reduction in case fatality rate without a change in pulmonary embolism mortality pointing to overdiagnosis by CT scan [7]. There are also concerns with CT scan radiation leading to cancer and other CT complications. It is felt that improved use of clinical decision rules and improved use of D-dimer may be one avenue to tackle these risks of over diagnosis and imaging complications. It is important to emphasise that not all commercially available DDimer assays have been independently and prospectively validated in DVT and/or PE diagnostic cohort studies [8,9]. Furthermore, the suggested cut-offs or thresholds for many D-Dimer assays have not been prospectively evaluated. As highlighted, above the absence of this crucial validation step reduces confidence in these D-Dimer assays and their suggested cut-offs. Clinicians and laboratory physicians must understand the level of evidence to support their local choice in D-Dimer assay and the diagnostic cut-offs of their local assays. Conventionally, a universal D-dimer cut-off, that is reagent specific, is used for all patient populations. This approach has also been extensively validated for many but not all D-Dimer assays. However, in certain populations, D-dimer levels are commonly well above these universal cut-offs lessening the diagnostic yield. In other words, there is concern that the numbers needed to test in certain patient populations including elderly patients, patients with malignancy and pregnant women leads to numbers needed test that are not cost-effective and are perceived by clinicians to be ineffective (i.e. rarely lead to diagnostic imaging averted). To tackle this issue, ongoing validation studies are exploring the diagnostic test characteristics of age-adjusted D-dimer cut-offs in elderly patients. Two different approaches using age-adjusted D-dimer cut-offs have been proposed. The first one is a fixed age-adjusted D-dimer cut-off where a cut-off of 750 μg/L or 1000 μg/L is proposed in patients of more than 60 years. In 2007, Harper et al. evaluated an age-dependent cut-off value of 500 μg/L for patients younger than 60 years and 1,000 μg/L for those aged N 60 years, in a cohort of 1897 patients in suspected VTE testing using Vidas® ELISA assay [10].The results showed that all the patients
765
with a confirmed DVT or PE in the age group N 80 years had a Ddimer concentration greater than 1000 ng/ml and only one out of 58 VTE patients in the 60–80 years age group had a D-Dimer result between 500 and 1,000 μg/L (his VTE was a below-knee DVT). At this threshold, assay sensitivity was 98% (but with a lower limit of the 95% confidence interval of 90%) in patients aged more than 60 years with a specificity of 55% in the 60 to 80 years age group (versus 25.3% for a cut-off of 500 μg/L) and a specificity of 27% in patients aged more than 80 years (versus 5% for a cut-off of 500 μg/L). The second approach is a variable age adjusted cut-off. This age-adjusted D-Dimer cut-off was proposed after gathering four prospective diagnostic management studies that included a total of 5,132 patients with suspected PE in order to derive and internally validate an age adjusted D-Dimer cutoff value [11]. The Vidas® assay was used in three of these studies, while either the Vidas® or Tinaquant® assays were used in the fourth. Patients N 50 with a non-high clinical probability were divided into 10 year age groups and ROC curves of the D-Dimer test for each age group were performed to find the best cut-off value. The D-Dimer cutoff level against age group was plotted and a linear regression analysis was performed to obtain the regression coefficient representing the increase in D-Dimer cut-off value per decade. This data suggested that the increase in the optimal cutoff value was 11.2 μg/L for each additional year of age. To facilitate clinical practicality and to be conservative, a 10 μg/L increase per year was proposed. Otherwise said, the optimal cut-off for a patient would be his age (in years) × 10 μg/L in patients older than 50 years (for example, a patient aged 82 years would have a cut-off value of 82 × 10 = 820 μg/L) for exclusion of pulmonary embolism. Overall, the age-adjusted cut-off allowed increasing the proportion of patients in whom the diagnosis of PE could be ruled out on the basis of a negative D-Dimer and a non-high clinical probability from 36 to 42%, without increasing the false negative rate. In patients aged N 80, the proportion of negative D-Dimer increased from 6% using the conventional 500 μg/L cut-off, to 21% using the age adjusted cut-off. This turned in a reduction from 18 to 5 in the number of patients needed to test to obtain one negative D-Dimer test. From a safety point of view, only one out of 153 patients from this age group with a D-Dimer level below their age-adjusted cut-off had PE. The age-adjusted cut-off was also assessed in a retrospective analysis of five management cohort studies that included 2,818 outpatients with suspected DVT [12]. Different D-Dimer assays were used: Tinaquant®, Vidas®, HemosIL-DD®, Liatest® and MDA®. Similarly to what was observed in patients with suspected PE, using the ageadjusted cut-off was associated with an important increase in the proportion of patients with a negative D-Dimer test: 44% using the ageadjusted cut-off, versus 20% at the conventional cut-off, without compromising test safety: the false negative rate was not different using either cut-off value. Finally, the variable age-adjusted cut-off and a fixed age-adjusted cut-off were retrospectively assessed among 1,374 consecutive patients aged N50 years with suspected DVT [13]. Patients were classified according to the dichotomized Wells’ score as “likely” (≥2) or “unlikely” (≤ 1). The performances of three different D-Dimer cut-off values were analyzed: the conventional cut-off value of 500 μg/L for patients of all ages, the age-adjusted cut-off value in patients older than 50 years (cut-off value = age in years × 10 μg/L), as well as the fixed cut-off value of 750 μg/L for patients aged 60 years and older. The exclusion and false negative rates were computed for each D-Dimer cut-off value. The two methods with adaptation of the D-Dimer cut-off value combined with an unlikely clinical probability of DVT, resulted again in a considerable and similar increase in the proportion of suspected elderly patients in whom DVT could be safely and correctly excluded: 21% with a cut-off value of 500 μg/L, 36% with the age-adjusted cutoff value, and 34% with a cut-off value of 750 μg/L in patients aged N 80 years. The main limitation of all these studies lies in their retrospective design. Admittedly, all were diagnostic studies using a standardized
766
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
strategy and imaging tests were performed in all patients with a DDimer level above the conventional cut-off. However, the missing step is to prospectively validate the adjusted cut-offs in a management outcome studies. The ADJUST study, a multicenter multinational prospective diagnostic management outcome study (NCT01134068) that uses the age-adjusted cut-off in outpatients with suspected PE is currently ongoing. The results of such a study are needed before any recommendation can be made for clinical practice. Overall, and based on preliminary information it appears that the safety issue with age-adjusted D-dimer cut-offs is not a concern and that there will be improved specificity but the extent of the specificity gain is uncertain and may be insufficient to change clinical practice or have an impact on routine clinical practice. A major concern with the age-adjusted D-Dimer cut-offs in elderly patients (and further consideration for malignancy adjusted D-Dimer, pregnancy adjusted D-Dimer or pretest probability adjusted D-Dimer) is that these approaches, in toto, will only increase the complexity of D-Dimer use in clinical practice and may lead to the average clinician to abandoning or misunderstanding DDimer testing.
women in the first, second, and third trimesters [19]. This finding also requires validation prior to adoption of a D-dimer only diagnostic strategy for suspected DVT in pregnancy. Exploration of a D-dimer alone strategy for suspected PE has never been examined. Chan and colleagues also demonstrated in a post-hoc analysis that altering the D-Dimer cut-off levels in pregnancy to explore if improvements in specificity could be achieved without compromising safety [20]. While this approach looks promising it requires prospective validation prior to adoption in clinical practice for suspected DVT and exploration for suspected PE in pregnancy. Furthermore, altering the D-Dimer threshold may be difficult for busy clinicians, including obstetricians who often provide primary care for these patients to recall and adopt in their practices. Overall, while ongoing studies will help clarify the validity of nonimaging based diagnostic approaches (i.e. clinical decision rules and D-Dimer) for suspected DVT in pregnancy we suggest all women with suspected VTE should be assigned a pre-test probability by clinical gestalt, have D-dimer testing but should undergo a diagnostic imaging based algorithm [21].
Challenges and Controversies in VTE Diagnosis: VTE Diagnosis and Pregnancy: Use of D-Dimer and Clinical Decision Rules
Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE: Duration of Anticoagulation
While clinical assessment using clinical decision rules has been demonstrated to be very useful in assigning pre-test probability for DVT [14] and PE [15,16] outside of pregnancy, the studies deriving and validating these CDRs model did not include pregnant patients. The proportion of confirmed VTE is usually lower among pregnant women with suspected VTE. Moreover, it is very likely that the distribution of physical findings (e.g. left leg swelling) and risk factors (e.g. trauma, immobilization, surgery, malignancy) would be different in pregnant women with and without VTE who are suspected of having VTE and thus the performance of clinical assessment may differ in this sub-population and as such the existing CDRs should not be used in pregnant women. Similarly, DDimer use for suspected VTE in pregnancy has not been adequately evaluated. D-dimer levels increase throughout normal pregnancy [17]. Near term and in the post-partum period most pregnant women will have abnormal D-Dimer levels. Therefore, D-dimer assays are, in general and more so in pregnancy, sensitive but non-specific markers for venous thromboembolism. Thankfully, pregnancy specific CDR and DDimer research is beginning to appear in the literature that may one day simplify management of suspected DVT and PE in pregnancy. In a prospective cohort study, Chan et al. reported on 194 pregnant women with suspected DVT [18]. Expert clinicians collected clinical information and synthesized this information in a subjective pre-test probability assessment. 17/194 women had DVT on an initial ultrasound, 182 women had a normal initial ultrasound and 152 then underwent serial leg compression ultrasounds (day 3 and/or day 7)). If ultrasounds were negative, participants were followed for 3 months for further clinical events and 1 additional patient developed DVT (total number of DVT = 18). The subjective opinion of the expert clinicians managing these patients demonstrated high negative predictive value was 98.5% (95% CI 94.6% to 99.6%). A clinical decision rule was derived. The “LEFT” clinical decision rule considers 3 variables in pregnant women with suspected DVT 1) left leg presentation, 2) ≥ 2 cm calf circumference difference and 3) first trimester presentation; if none of the LEFT variables are present the negative predictive value is 100% (95% CI 95.8% to 100%). CDRs often perform worse in validation studies hence this rule should not be used in clinical practice until it is validated but it appears very promising. A recent small study demonstrated promising diagnostic performance for a whole blood agglutination D-Dimer in pregnant women. This D-dimer had a sensitivity of 100% (CI, 77% to 100% [13 of 13 patients]), specificity of 60% (CI, 52% to 68%), and a negative predictive value of 100% (CI, 95% to 100% [81 of 81]). The D-dimer was positive in 0% (CI, 0% to 60%), 24% (CI, 14% to 37%), and 51% (CI, 40% to 61%) of
In unprovoked VTE patients the duration of anticoagulation is one of the most important unanswered questions clinicians face on a daily basis. Clinicians and patients must balance the long-term risks of recurrent VTE off anticoagulants with the long-term risk of major bleeding to come to a treatment decision regarding the duration of anticoagulation. In unselected unprovoked VTE patients, these long term risks are closely balanced [22]. Attention has turned towards identifying tools to risk stratify patients into sub-groups at low risk of VTE recurrence where anticoagulants can be safely discontinued and sub-groups at high risk of VTE recurrence where anticoagulants should clearly be continued. The tools explored to date include single predictors and clinical decision rules. The single predictors that have been examined including residual venous obstruction [23] and D-dimer [24]. The authors note significant practice variation where residual venous obstruction and D-Dimer is being used quite commonly in Italian centers but not in North American centers. The absence of a clear ability to importantly risk stratify patients with these predictors is likely the cause of this practice variation. Attention has turned to the development of CDRs to guide duration of anticoagulation. The DASH [25], Vienna [26] and “Men continue and HERDOO2” [27] CDRs have been developed but not yet validated and hence are not ready to be adopted in clinical practice. Some concerns have been raised in the derivation of these rules. For example, the Vienna and DASH scores did not include all important predictors at the derivation stage. Specifically, these rules did not collect and therefore could not include the most potent predictor of recurrent VTE identified in the REVERSE study [27], namely the post thrombotic findings of hyperpigmentation, redness or edema. Both the Vienna and DASH rules also require D-Dimer measurement 1 month after discontinuation of anticoagulation which some may consider impractical in busy clinical practices and may be risky as many patients must come off anticoagulants for a month to define if they are at high risk. During this month they may be exposed to recurrent VTE and indeed fatal recurrent VTE as has been reported in this off treatment testing interval [28]. The Vienna prediction rule also requires use of a complicated nomogram that may be challenging to adoption in routine clinical practice although the availability of an online version facilitates its use in clinical practice. The “Men continue and HERDOO2” rule is undergoing prospective validation in a large multi-national cohort study. The rule is simple to apply as it has 5 predictors (male gender, PTS findings of hyperpigmentation, edema or redness in either leg (HER), Vidas® D-Dimer N250 ug/ ml Older age (N 65), and Obesity (BMI over 30) on anticoagulants), an
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
easy to recall mnemonic is applied while patients are on anticoagulant treatment and thereby doesn’t require a repeat visit to apply nor expose patients to an off anticoagulant treatment testing interval. The rule suggests that all men and women with 2 or more of the HER DOO features continue anticoagulants longer term. There is also important controversy regarding the recurrent VTE threshold below which it is safe to discontinue anticoagulants or above which it is clearly of benefit to continue anticoagulants. Some experts have suggested that the lower threshold should be less than 5% [29] others have suggested that it should be less than 3% [22]. This discussion hinges on the absolute bleeding risk associated with ongoing anticoagulant therapy and the case fatality rates of major bleeding and recurrent VTE. Recent literature suggests that the major bleeding risk is low (1-2%) in vitamin K antagonist experienced patients and may be even lower with the novel oral anticoagulants (b1%) [30]. The latter argues in favor of using a lower recurrent VTE threshold to ensure patients wouldn’t benefit from ongoing anticoagulant therapy. Finally, there is a clear need to develop bleeding risk stratification tools to be married with recurrent VTE risk stratification tools to tailor patient's therapy choices. Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE- PE Risk Stratification for Selection of PE Treatment Setting Patients experiencing PE face their highest mortality risks from PE in the minutes and days that follow the event. Treatment setting for the initial management of PE is an area of wide international practice variation. In some centers in Canada PE is treated in an outpatient setting in up to 55% of patients [31,32]. In these centers patients are selected for outpatient therapy based on the absence of 1) need for oxygen, 2) hemodynamic instability, 3) history of pre-syncope or syncope as presenting symptom and 4) contraindications to anticoagulant therapy. In many U.S. and European centers PE patients are treated exclusively as inpatients. Clinical decision rules have been derived and validated to objectively identify patients with a low mortality risk who may be treated as outpatient. The most extensively validated clinical decision rule is the Pulmonary Embolism Severity Index (PESI) [33]. The PESI comprises 11 easily available clinical variables and accurately stratifies patients into five risk classes (I-V) with increasing risk of all-cause 30day mortality, ranging from 1.1% in class I to 24.5% in class V. Patients in risk classes I and II are considered low-risk. In an international randomized non-inferiority trial, outpatient treatment with lowmolecular-weight heparin of selected low-risk patients based on the PESI was non-inferior to inpatient treatment in terms of efficacy and safety, was well accepted by patients, and reduced time spent in the hospital [34]. This clinical trial demonstrated that up to 30% of patients with PE could be safely treated in an outpatient setting based on the PESI and that doctors do not have to rely on markers of myocardial dysfunction or injury (e.g., brain natriuretic peptide, troponin, or echocardiographic measurements of right ventricular function) to safely identify candidates for outpatient care. Recently, a simplified 6-item version of the PESI (sPESI) has become available but requires prospective validation before its use can be recommended [35]. In a multicenter prospective management study from the Netherlands, 51% of patients with PE who were selected on the basis of 11 simple clinical criteria (HESTIA criteria) were safely managed as outpatients [36]. The safety of outpatient care based on the HESTIA criteria alone versus HESTIA criteria plus NTpro-BNP is currently being evaluated in the randomized VESTA trial (Netherlands Trial Registry no. 2603).At the other end of the spectrum, there is an identified need for prognostic PE rules that identify hemodynamically stable, high-risk patients that may require ICU admission or IVC filter insertion. Several biomarkers and right ventricular dilatation (RVD) on echocardiography or CT-scan have been associated with an increased risk of death or major complication related to PE. But most studies did not adjust for
767
clinical variables [37]. More recently, multicenter cohort studies using multivariable analysis suggest that Brain Natriuretic Peptide and RVD have prognostic value independently from the clinical variables and the PESI but these results have not been confirmed externally and the clinical significance of such findings is questionable [38,39]. In addition, the clinical benefit of the two main therapeutic options in these patients with so-called intermediate-risk PE, i.e. thrombolytic therapy and vena cava interruption remains to be confirmed. Conclusion In many areas of diagnostic and therapeutic management of the PE and DVT clinical decision rules and D-dimer are promising tools to improve cost-effective clinical care. However, knowledge translation barriers and wide international practice variation points to the need for future research. This future research will need to simplify these rules so they are ultimately adopted in routine clinical practice after adequate validation so that practice variation becomes limited. Technology may improve adoption of what we now perceive to be complex rules by simple automation of solicitation of clinical decision rule predictors, calculation of categorization and suggested subsequent management. With these improvements, optimal diagnostic and therapeutic management of VTE is likely to be adopted worldwide. Conflict of Interest Statement Marc Rodger has received grant funding from Biomerieux for the conduct of studies exploring the role of D-Dimer in predicting the risk of recurrent VTE in unprovoked VTE patients. Funding No funding was obtained to write this review Acknowledgements Marc Rodger was supported through the Heart and Stroke Foundation with a Career Investigator Award and is the recipient of a University of Ottawa Faculty of Medicine Research Chair Awards and a University of Ottawa Department of Medicine Research Salary Award. Dr. Phil Wells held a Canada Research Chair in Thromboembolic Diseases. References [1] Anderson Jr FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester study. Arch Intern Med 1991;151:933–8. [2] McGinn TG, Guyatt GH, Wyer PC, Naylor CD, Stiell IG, Richardson WS. Users' guides to the medical literature: XXII: how to use articles about clinical decision rules. Evidence-Based Medicine Working Group. JAMA 2000;284:79–84. [3] Wasson JH, Sox HC, Neff RK, Goldman L. Clinical prediction rules. Applications and methodological standards. N Engl J Med 1985;313:793–9. [4] Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ 2005;331:259–68. [5] Roy PM, Meyer G, Vielle B, Le Gall C, Verschuren F, Carpentier F, et al. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med 2006;144:157–64. [6] Douma RA, Mos IC, Erkens PM, Nizet TA, Durian MF, Hovens MM, et al. Performance of 4 Clinical Decision Rules in the Diagnostic Management of Acute Pulmonary Embolism: A Prospective Cohort Study. Ann Intern Med 2011;154:709–18. [7] Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med 2011;171:831–7. [8] Adam SS, Key NS, Greenberg CS. D-dimer antigen: current concepts and future prospects. Blood 2009;113:2878–87. [9] Stein PD, Hull RD, Patel KC, Olson RE, Ghali WA, Brant RF, et al. D-Dimer for the exclusion of acute venous thrombosis and pulmonary embolism. A systematic review. Ann Intern Med 2004;140:589–602. [10] Harper PL, Theakston E, Ahmed J, Ockelford P. D-dimer concentration increases with age reducing the clinical value of the D-dimer assay in the elderly. Intern Med J 2007;37:607–13.
768
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
[11] Douma RA, Le GG, Sohne M, Righini M, Kamphuisen PW, Perrier A, et al. Potential of an age adjusted D-dimer cut-off value to improve the exclusion of pulmonary embolism in older patients: a retrospective analysis of three large cohorts. BMJ 2010;340: c1475. [12] Douma RA, Tan M, Schutgens RE, Bates SM, Perrier A, Legnani C, et al. Using an agedependent D-dimer cut-off value increases the number of older patients in whom deep vein thrombosis can be safely excluded. Haematologica 2012;97:1507–13. [13] Schouten HJ, Koek HL, Oudega R, Geersing GJ, Janssen KJ, van Delden JJ, et al. Validation of two age dependent D-dimer cut-off values for exclusion of deep vein thrombosis in suspected elderly patients in primary care: retrospective, cross sectional, diagnostic analysis. BMJ 2012;344:e2985. [14] Wells PS, Owen C, Doucette S, Fergusson D, Tran H. Does this patient have deep vein thrombosis? JAMA 2006;295:199–207. [15] Wells PS, Ginsberg JS, Anderson DR, Kearon C, Gent M, Turpie AGG, et al. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med 1998;129:997–1005. [16] Wicki J, Perneger TV, Junod AF, Bounameaux H, Perrier A. Assessing clinical probability of pulmonary embolism in the emergency ward: a simple score. Arch Intern Med 2001;161:92–7. [17] Kjellberg U, Andersson NE, Rosen S, Tengborn L, Hellgren M. APC resistance and other haemostatic variables during pregnancy and puerperium. Thromb Haemost 1999;81:527–31. [18] Chan WS, Lee A, Spencer FA, Crowther M, Rodger M, Ramsay T, et al. Predicting deep venous thrombosis in pregnancy: out in "LEFt" field? Ann Intern Med 2009;151: 85–92. [19] Chan WS, Chunilal S, Lee A, Crowther M, Rodger M, Ginsberg JS. A red blood cell agglutination D-dimer test to exclude deep venous thrombosis in pregnancy. Ann Intern Med 2007;147:165–70. [20] Chan WS, Lee A, Spencer FA, Chunilal S, Crowther M, Wu W, et al. D-dimer testing in pregnant patients: towards determining the next "level" in the diagnosis of DVT. J Thromb Haemost 2010;8:1004–11. [21] Rodger M. Evidence base for the management of venous thromboembolism in pregnancy. Hematology Am Soc Hematol Educ Program 2010;2010:173–80. [22] Rodger M, Carrier M, Gandara E, Le GG. Unprovoked Venous Thromboembolism: Short term or Indefinite Anticoagulation? Balancing Long-Term Risk and Benefit. Blood Rev 2010;24:171–8. [23] Donadini MP, Ageno W, Antonucci E, Cosmi B, Kovacs MJ, Le GG, et al. Prognostic significance of residual venous obstruction in patients with treated unprovoked deep vein thrombosis. A patient-level meta-analysis. Thromb Haemost 2014;111:172–9. [24] Verhovsek M, Douketis JD, Yi Q, Shrivastava S, Tait RC, Baglin T, et al. Systematic review: D-dimer to predict recurrent disease after stopping anticoagulant therapy for unprovoked venous thromboembolism. Ann Intern Med 2008;149: 481–90 [W-94]. [25] Tosetto A, Iorio A, Marcucci M, Baglin T, Cushman M, Eichinger S, et al. Predicting disease recurrence in patients with previous unprovoked venous thromboembolism: a proposed prediction score (DASH). J Thromb Haemost 2012;10:1019–25.
[26] Eichinger S, Heinze G, Jandeck LM, Kyrle PA. Risk assessment of recurrence in patients with unprovoked deep vein thrombosis or pulmonary embolism: the Vienna prediction model. Circulation 2010;121:1630–6. [27] Rodger MA, Kahn SR, Wells PS, Anderson DA, Chagnon I, Le Gal G, et al. Identifying unprovoked thromboembolism patients at low risk for recurrence who can discontinue anticoagulant therapy. CMAJ 2008;179:417–26. [28] Cosmi B, Legnani C, Tosetto A, Pengo V, Ghirarduzzi A, Testa S, et al. Usefulness of repeated D-dimer testing after stopping anticoagulation for a first episode of unprovoked venous thromboembolism: the PROLONG II prospective study. Blood 2010; 115:481–8. [29] Baglin T, Bauer K, Douketis J, Buller H, Srivastava A, Johnson G. Duration of anticoagulant therapy after a first episode of an unprovoked pulmonary embolus or deep vein thrombosis: guidance from the SSC of the ISTH. J Thromb Haemost 2012;10: 698–702. [30] Castellucci LA, Cameron C, Le Gal G, Rodger MA, Coyle D, Wells PS, et al. Safety and efficacy outcomes of oral anticoagulants and antiplatelet drugs in the secondary prevention of venous thromboembolism: systematic review and network metaanalysis. BMJ 2013;347:f5133. [31] Erkens PM, Gandara E, Wells PS, Shen AY, Bose G, Le Gal G, et al. Safety of outpatient treatment in acute pulmonary embolism. J Thromb Haemost 2010;8:2412–7. [32] Kovacs MJ, Hawel JD, Rekman JF, Lazo-Langner A. Ambulatory management of pulmonary embolism: a pragmatic evaluation. J Thromb Haemost 2010;8:2406–11. [33] Aujesky D, Obrosky DS, Stone RA, Auble TE, Perrier A, Cornuz J, et al. Derivation and validation of a prognostic model for pulmonary embolism. Am J Respir Crit Care Med 2005;172:1041–6. [34] Aujesky D, Roy PM, Verschuren F, Righini M, Osterwalder J, Egloff M, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet 2011;378:41–8. [35] Jimenez D, Aujesky D, Moores L, Gomez V, Lobo JL, Uresandi F, et al. Simplification of the pulmonary embolism severity index for prognostication in patients with acute symptomatic pulmonary embolism. Arch Intern Med 2010;170:1383–9. [36] Zondag W, Mos IC, Creemers-Schild D, Hoogerbrugge AD, Dekkers OM, Dolsma J, et al. Outpatient treatment in patients with acute pulmonary embolism: the Hestia Study. J Thromb Haemost 2011;9:1500–7. [37] Sanchez O, Trinquart L, Colombet I, Durieux P, Huisman MV, Chatellier G, et al. Prognostic value of right ventricular dysfunction in patients with haemodynamically stable pulmonary embolism: a systematic review. Eur Heart J 2008;29:1569–77. [38] Sanchez O, Trinquart L, Caille V, Couturaud F, Pacouret G, Meneveau N, et al. Prognostic factors for pulmonary embolism: the prep study, a prospective multicenter cohort study. Am J Respir Crit Care Med 2010;181:168–73. [39] Sanchez O, Trinquart L, Planquette B, Couturaud F, Verschuren F, Caille V, et al. Echocardiography and pulmonary embolism severity index have independent prognostic roles in pulmonary embolism. Eur Respir J 2013;42:681–8.
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Review Article
Clinical Decision Rules and D-dimer in Venous Thromboembolism: Current controversies and future research priorities Marc A. Rodger a,b,c,d,⁎, Gregoire Le Gal a,c,d, Philip Wells a,c,d, Trevor Baglin e, Drahomir Aujesky f, Marc Righini g, Gualtiero Palareti h, Menno Huisman i, Guy Meyer j a
Hematology, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada Medicine, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada Obstetrics and Gynaecology, University of Ottawa and The Ottawa Hospital, Ottawa, ON Canada d Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON Canada e Cambridge Haemophilia and Thrombophilia Centre, Addenbrookes Hospital, Cambridge University Hospitals NHS Trust, Cambridge, UK f Division of General Internal Medicine, Bern University Hospital, Bern, Switzerland g Division of Angiology and Hemostasis, Department of Medical Specialties, Geneva University Hospital and Faculty of Medicine, Geneva, Switzerland h Unit of Angiology and Blood Coagulation, University Hospital of Bologna, Italy i Department of Thrombosis and Hemostasis, LUMC, Leiden, the Netherlands j Université Paris Descartes Sorbonne Paris Cité and Hopital europeen Georges Pompidou APHP, Paris, France b c
a r t i c l e
i n f o
Article history: Received 14 May 2014 Received in revised form 25 July 2014 Accepted 27 July 2014 Available online 4 August 2014 Keywords: Clinical Decision Rules D-dimer Diagnosis Treatment Prognosis Venous Thromboembolism
a b s t r a c t Venous thromboembolism (VTE) is a potentially lethal clinical condition that is suspected in patients with common clinical complaints, in many and varied, clinical care settings. Once VTE is diagnosed, optimal therapeutic management (thrombolysis, IVC filters, type and duration of anticoagulants) and ideal therapeutic management settings (outpatient, critical care) are also controversial. Clinical prediction tools, including clinical decision rules and D-Dimer, have been developed, and some validated, to assist clinical decision making along the diagnostic and therapeutic management paths for VTE. Despite these developments, practice variation is high and there remain many controversies in the use of the clinical prediction tools. In this narrative review, we highlight challenges and controversies in VTE diagnostic and therapeutic management with a focus on clinical decision rules and D-Dimer. © 2014 Elsevier Ltd. All rights reserved.
Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodologic Standards for Clinical Predictors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: Pretest Probability Assessment for Diagnostic Management of DVT or PE: Simplified Rules or The Original Rules? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: Traditional or Age-Adjusted D-Dimer Cut Off for DVT or Pulmonary Embolism Diagnostic Management? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Controversies in VTE Diagnosis: VTE Diagnosis and Pregnancy: Use of D-Dimer and Clinical Decision Rules . . . . . . . . . Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE: Duration of Anticoagulation . . . . . . . . . . . . Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE- PE Risk Stratification for Selection of PE Treatment Setting Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of Interest Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
764 764
. . . . .
764
. . . . . . . . .
765 766 766 767 767 767 767 767 767
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
⁎ Corresponding author at: The Ottawa Hospital, Ottawa Blood Disease Center, 501 Smyth Road, Box 201; Ottawa, ON, Canada, K1H 8L6. Tel.: +613 737 8899x74641; fax: +613 739 6102. E-mail address: [email protected] (M.A. Rodger).
http://dx.doi.org/10.1016/j.thromres.2014.07.031 0049-3848/© 2014 Elsevier Ltd. All rights reserved.
764
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
Introduction Venous thromboembolism (VTE), comprising both deep vein thrombosis (DVT) and pulmonary embolism (PE), are common, potentially lethal yet treatable clinical conditions [1]. Clinical Decision Rules (CDRs) are decision making tools, using combinations of simple available clinical predictors to define an outcome in the present, in other words a diagnosis (or a probability of disease), or an outcome in the future, in other words a prognosis (or probability of an outcome), either of which leads to a diagnostic course of action or a therapeutic course of action [2]. CDRs and/or D-dimer have a crucial place in the diagnostic and therapeutic management of VTE. These tools permit us to judiciously and safely use diagnostic imaging for VTE diagnosis, select the right treatment setting for initial therapeutic management of VTE (intensive care unit, ward or home) and select which patients will derive net benefit from anticoagulant therapies. The focus of this narrative review will be to review current controversies in the use of clinical decision rules and/or D-dimer in the diagnostic and therapeutic management of VTE. We will highlight wide international practice variations, even among experts in thrombosis, and also highlight challenges in knowledge translation for most clinicians that impact putting research findings into routine clinical practice. These challenges are largely reflective of the absence of level IA evidence to guide practice in these areas and the need to develop and validate simple and widely available tools that are easy to adopt in routine clinical practice. Methodologic Standards for Clinical Predictors High quality evidence to support routine use of CDRs, and other clinical prediction tools, requires that they are developed and validated strictly following methodologic guidelines (see Table 1). Standards for their development and validation were first published more than 20 years ago [3], were updated and which have formed the basis for a quality assessment framework we should consider before adopting CDRs in our daily practice [2]. These standards require adherence to methodologic guidelines at the development, validation and impact analysis stage (see Table 1). If this is the case, in the long run, they usually perform up to expectations. However, on the other end of the spectrum, many CDRs are developed with minimal adherence to methodologic guidelines, are never validated and ultimately, generally, these rules fail to meet expectations. Similarly, single predictors (e.g. Ddimer), can be viewed as the simplest of CDRs and their use in routine clinical practice should be limited to those clinical applications where validation and impact analysis have also been conducted. CDRs select and combine the best independent predictors (risk factors, symptoms, clinical signs and results of simple diagnostic tests) for a diagnosis or prognosis. The most useful CDRs are accurate, reproducible, simple and easy to apply. CDRs should be sensible i.e. have a clear purpose (e.g. exclude DVT), be relevant (e.g. exclude clinically important DVT), demonstrate content validity (e.g. be composed of well recognised independent predictors), be concise (e.g. simple rules containing limited items will be remembered), and be easy to use in the intended clinical application (e.g. don’t require a computer to calculate at the bedside). The use of the rule should provide a probability of disease (e.g. N10% likelihood of DVT) or prognosis (e.g. 15% risk of recurrent VTE) and should imply a course of diagnostic (e.g. needs an ultrasound to rule out DVT) or therapeutic action (e.g. continue anticoagulants indefinitely). Construction of valid, accurate and reproducible CDRs follows a strict methodology. Levels of evidence can be attributed to CDRs depending on whether or not they have gone through all the methodological steps (Levels 1 to 4). Level 4 corresponds to a rule that is derived but not prospectively validated: it needs to be further evaluated before clinical application. Validation of the CDR in an independent cohort of patients is a crucial next step. In fact, the rule is built as a "tailored suit"
Table 1 Methodologic characteristics and criteria required to develop and validate clinical decision rules and why they are important to follow. Stage
Methodologic Criteria
Why?
Development Describe patient characteristics
To ensure generalizability and applicability of the results to reader’s clinical practice Broad spectrum of disease To ensure that CDR can segregate patients with subtle or not obvious disease To ensure that the CDR predicts Outcome clearly defined, important, assessed blindly and the right outcome in an unbiased manner is the gold standard To ensure all potential predictors Complete set of potential can enter CDR. To ensure that predictors that are clearly predictors collected in a defined and blindly and reproducible, accurate and prospectively collected unbiased manner. To maximise generalizability Reproducibility of potential predictors assessed only reliable predictors (Kappa N0.6) should enter rule. To ensure adequate power so Statistical techniques used to important predictors have derive the rule are identified narrow estimates of effect. To and valid ensure rule is not “overfit” and only works in the derivation cohort To ensure the CDR is easy to use Final rule chosen based on simplicity and ability to provide and remember so it is widely adoptable. To ensure thresholds a clear course of action. for dividing outcomes are chosen for optimal repartition of patients across groups and optimal proportion of patients with outcomes in each group. Validation Apply the rule in an To check the proportions of independent cohort of patients patients classified by the rule in each clinical probability group. Impact Determine impact of use of the To determine if the rule is used. assessment rule in the real world To determine the rule’s safety, accuracy and clinical utility in real world use.
for the derivation cohort. In other words, it is important to ensure that the “suit” will fit all the populations in which it is intended to be applied. Level 3 rules have been prospectively validated but in only one narrow sample: physicians may consider their use with caution and only if patients in their clinical setting are similar to those included in validation study. Level 2 rules have demonstrated their accuracy in either a large prospective study including a broad spectrum of the disease, or in several different smaller settings. They can be used in various settings with confidence in their accuracy. Finally, the impact of use of the rule, its clinical utility and safety of managing patients on the basis of the rule “in the real world” should be demonstrated in a prospective management outcome study. In these studies, use of the rule in usual practice is measured along with the performance of the rule in usual practice. Level 1 rules have been prospectively validated in a different population and the impact of use of the rule has been measured and demonstrates a change in clinician behaviour with beneficial consequences. Level 1 rules can be used with the confidence that they can change clinician behaviour and improve patients' outcomes.
Challenges and Controversies in VTE Diagnosis: Pretest Probability Assessment for Diagnostic Management of DVT or PE: Simplified Rules or The Original Rules? The original Wells DVT CDR and the Wells PE CDR in addition to the Geneva CDR for PE, have been well evaluated in clinical research over the last 2 decades [4]. Nonetheless it is evident from published practice patterns that clinicians often do not use these rules [5]. The risk of this underuse is either over or under diagnosis of VTE [5]. Some of the
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
knowledge translation barriers that have been highlighted include 1) the need to recall more than 7 items with each of these rules, 2) the need to recall the points assigned for each predictor in each of these rules, 3) the need to calculate the total number of points and 4) with the Wells DVT CDR and Wells PE CDR the “alternative diagnosis” item which requires subjective assessment that may lead to discomfort for less experienced physicians. These challenges/complexities may dissuade practicing clinicians from using these rules. However, new technology including bedside applications (“apps”) offer one solution to these recall, point assignment and calculation complexities and as younger “tech friendly” generations become involved in clinical practice perhaps these considerations will become less important. Simplified rules have also been developed in order to reduce the complexities associated with these rules. In a prospective study the simplified decision rules – either simplified Wells rule or simplified Geneva rule – worked equally well; total failure rates of the simplified CDR’s and D-dimer combinations were similar (1 failure, 0.5% to 0.6% [upper-limit 95% CI, 2.9% to 3.1%]) [6]. The use of these simplified rules, after additional validation, will hopefully lead to better and more prevalent use of the diagnostic CDRs. However, there is some skepticism about whether they are simple enough (e.g. still involve recall of N 7 predictors) and that they will be more widely used than the original rules. Challenges and Controversies in VTE Diagnosis: Traditional or AgeAdjusted D-Dimer Cut Off for DVT or Pulmonary Embolism Diagnostic Management? After pre-test probability assessment, clinicians can use D-dimer as a next step to determine if imaging is required or VTE is excluded. Epidemiologic research has demonstrated that in the CT scan era, the diagnosis of PE is increasing in frequency with a reduction in case fatality rate without a change in pulmonary embolism mortality pointing to overdiagnosis by CT scan [7]. There are also concerns with CT scan radiation leading to cancer and other CT complications. It is felt that improved use of clinical decision rules and improved use of D-dimer may be one avenue to tackle these risks of over diagnosis and imaging complications. It is important to emphasise that not all commercially available DDimer assays have been independently and prospectively validated in DVT and/or PE diagnostic cohort studies [8,9]. Furthermore, the suggested cut-offs or thresholds for many D-Dimer assays have not been prospectively evaluated. As highlighted, above the absence of this crucial validation step reduces confidence in these D-Dimer assays and their suggested cut-offs. Clinicians and laboratory physicians must understand the level of evidence to support their local choice in D-Dimer assay and the diagnostic cut-offs of their local assays. Conventionally, a universal D-dimer cut-off, that is reagent specific, is used for all patient populations. This approach has also been extensively validated for many but not all D-Dimer assays. However, in certain populations, D-dimer levels are commonly well above these universal cut-offs lessening the diagnostic yield. In other words, there is concern that the numbers needed to test in certain patient populations including elderly patients, patients with malignancy and pregnant women leads to numbers needed test that are not cost-effective and are perceived by clinicians to be ineffective (i.e. rarely lead to diagnostic imaging averted). To tackle this issue, ongoing validation studies are exploring the diagnostic test characteristics of age-adjusted D-dimer cut-offs in elderly patients. Two different approaches using age-adjusted D-dimer cut-offs have been proposed. The first one is a fixed age-adjusted D-dimer cut-off where a cut-off of 750 μg/L or 1000 μg/L is proposed in patients of more than 60 years. In 2007, Harper et al. evaluated an age-dependent cut-off value of 500 μg/L for patients younger than 60 years and 1,000 μg/L for those aged N 60 years, in a cohort of 1897 patients in suspected VTE testing using Vidas® ELISA assay [10].The results showed that all the patients
765
with a confirmed DVT or PE in the age group N 80 years had a Ddimer concentration greater than 1000 ng/ml and only one out of 58 VTE patients in the 60–80 years age group had a D-Dimer result between 500 and 1,000 μg/L (his VTE was a below-knee DVT). At this threshold, assay sensitivity was 98% (but with a lower limit of the 95% confidence interval of 90%) in patients aged more than 60 years with a specificity of 55% in the 60 to 80 years age group (versus 25.3% for a cut-off of 500 μg/L) and a specificity of 27% in patients aged more than 80 years (versus 5% for a cut-off of 500 μg/L). The second approach is a variable age adjusted cut-off. This age-adjusted D-Dimer cut-off was proposed after gathering four prospective diagnostic management studies that included a total of 5,132 patients with suspected PE in order to derive and internally validate an age adjusted D-Dimer cutoff value [11]. The Vidas® assay was used in three of these studies, while either the Vidas® or Tinaquant® assays were used in the fourth. Patients N 50 with a non-high clinical probability were divided into 10 year age groups and ROC curves of the D-Dimer test for each age group were performed to find the best cut-off value. The D-Dimer cutoff level against age group was plotted and a linear regression analysis was performed to obtain the regression coefficient representing the increase in D-Dimer cut-off value per decade. This data suggested that the increase in the optimal cutoff value was 11.2 μg/L for each additional year of age. To facilitate clinical practicality and to be conservative, a 10 μg/L increase per year was proposed. Otherwise said, the optimal cut-off for a patient would be his age (in years) × 10 μg/L in patients older than 50 years (for example, a patient aged 82 years would have a cut-off value of 82 × 10 = 820 μg/L) for exclusion of pulmonary embolism. Overall, the age-adjusted cut-off allowed increasing the proportion of patients in whom the diagnosis of PE could be ruled out on the basis of a negative D-Dimer and a non-high clinical probability from 36 to 42%, without increasing the false negative rate. In patients aged N 80, the proportion of negative D-Dimer increased from 6% using the conventional 500 μg/L cut-off, to 21% using the age adjusted cut-off. This turned in a reduction from 18 to 5 in the number of patients needed to test to obtain one negative D-Dimer test. From a safety point of view, only one out of 153 patients from this age group with a D-Dimer level below their age-adjusted cut-off had PE. The age-adjusted cut-off was also assessed in a retrospective analysis of five management cohort studies that included 2,818 outpatients with suspected DVT [12]. Different D-Dimer assays were used: Tinaquant®, Vidas®, HemosIL-DD®, Liatest® and MDA®. Similarly to what was observed in patients with suspected PE, using the ageadjusted cut-off was associated with an important increase in the proportion of patients with a negative D-Dimer test: 44% using the ageadjusted cut-off, versus 20% at the conventional cut-off, without compromising test safety: the false negative rate was not different using either cut-off value. Finally, the variable age-adjusted cut-off and a fixed age-adjusted cut-off were retrospectively assessed among 1,374 consecutive patients aged N50 years with suspected DVT [13]. Patients were classified according to the dichotomized Wells’ score as “likely” (≥2) or “unlikely” (≤ 1). The performances of three different D-Dimer cut-off values were analyzed: the conventional cut-off value of 500 μg/L for patients of all ages, the age-adjusted cut-off value in patients older than 50 years (cut-off value = age in years × 10 μg/L), as well as the fixed cut-off value of 750 μg/L for patients aged 60 years and older. The exclusion and false negative rates were computed for each D-Dimer cut-off value. The two methods with adaptation of the D-Dimer cut-off value combined with an unlikely clinical probability of DVT, resulted again in a considerable and similar increase in the proportion of suspected elderly patients in whom DVT could be safely and correctly excluded: 21% with a cut-off value of 500 μg/L, 36% with the age-adjusted cutoff value, and 34% with a cut-off value of 750 μg/L in patients aged N 80 years. The main limitation of all these studies lies in their retrospective design. Admittedly, all were diagnostic studies using a standardized
766
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
strategy and imaging tests were performed in all patients with a DDimer level above the conventional cut-off. However, the missing step is to prospectively validate the adjusted cut-offs in a management outcome studies. The ADJUST study, a multicenter multinational prospective diagnostic management outcome study (NCT01134068) that uses the age-adjusted cut-off in outpatients with suspected PE is currently ongoing. The results of such a study are needed before any recommendation can be made for clinical practice. Overall, and based on preliminary information it appears that the safety issue with age-adjusted D-dimer cut-offs is not a concern and that there will be improved specificity but the extent of the specificity gain is uncertain and may be insufficient to change clinical practice or have an impact on routine clinical practice. A major concern with the age-adjusted D-Dimer cut-offs in elderly patients (and further consideration for malignancy adjusted D-Dimer, pregnancy adjusted D-Dimer or pretest probability adjusted D-Dimer) is that these approaches, in toto, will only increase the complexity of D-Dimer use in clinical practice and may lead to the average clinician to abandoning or misunderstanding DDimer testing.
women in the first, second, and third trimesters [19]. This finding also requires validation prior to adoption of a D-dimer only diagnostic strategy for suspected DVT in pregnancy. Exploration of a D-dimer alone strategy for suspected PE has never been examined. Chan and colleagues also demonstrated in a post-hoc analysis that altering the D-Dimer cut-off levels in pregnancy to explore if improvements in specificity could be achieved without compromising safety [20]. While this approach looks promising it requires prospective validation prior to adoption in clinical practice for suspected DVT and exploration for suspected PE in pregnancy. Furthermore, altering the D-Dimer threshold may be difficult for busy clinicians, including obstetricians who often provide primary care for these patients to recall and adopt in their practices. Overall, while ongoing studies will help clarify the validity of nonimaging based diagnostic approaches (i.e. clinical decision rules and D-Dimer) for suspected DVT in pregnancy we suggest all women with suspected VTE should be assigned a pre-test probability by clinical gestalt, have D-dimer testing but should undergo a diagnostic imaging based algorithm [21].
Challenges and Controversies in VTE Diagnosis: VTE Diagnosis and Pregnancy: Use of D-Dimer and Clinical Decision Rules
Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE: Duration of Anticoagulation
While clinical assessment using clinical decision rules has been demonstrated to be very useful in assigning pre-test probability for DVT [14] and PE [15,16] outside of pregnancy, the studies deriving and validating these CDRs model did not include pregnant patients. The proportion of confirmed VTE is usually lower among pregnant women with suspected VTE. Moreover, it is very likely that the distribution of physical findings (e.g. left leg swelling) and risk factors (e.g. trauma, immobilization, surgery, malignancy) would be different in pregnant women with and without VTE who are suspected of having VTE and thus the performance of clinical assessment may differ in this sub-population and as such the existing CDRs should not be used in pregnant women. Similarly, DDimer use for suspected VTE in pregnancy has not been adequately evaluated. D-dimer levels increase throughout normal pregnancy [17]. Near term and in the post-partum period most pregnant women will have abnormal D-Dimer levels. Therefore, D-dimer assays are, in general and more so in pregnancy, sensitive but non-specific markers for venous thromboembolism. Thankfully, pregnancy specific CDR and DDimer research is beginning to appear in the literature that may one day simplify management of suspected DVT and PE in pregnancy. In a prospective cohort study, Chan et al. reported on 194 pregnant women with suspected DVT [18]. Expert clinicians collected clinical information and synthesized this information in a subjective pre-test probability assessment. 17/194 women had DVT on an initial ultrasound, 182 women had a normal initial ultrasound and 152 then underwent serial leg compression ultrasounds (day 3 and/or day 7)). If ultrasounds were negative, participants were followed for 3 months for further clinical events and 1 additional patient developed DVT (total number of DVT = 18). The subjective opinion of the expert clinicians managing these patients demonstrated high negative predictive value was 98.5% (95% CI 94.6% to 99.6%). A clinical decision rule was derived. The “LEFT” clinical decision rule considers 3 variables in pregnant women with suspected DVT 1) left leg presentation, 2) ≥ 2 cm calf circumference difference and 3) first trimester presentation; if none of the LEFT variables are present the negative predictive value is 100% (95% CI 95.8% to 100%). CDRs often perform worse in validation studies hence this rule should not be used in clinical practice until it is validated but it appears very promising. A recent small study demonstrated promising diagnostic performance for a whole blood agglutination D-Dimer in pregnant women. This D-dimer had a sensitivity of 100% (CI, 77% to 100% [13 of 13 patients]), specificity of 60% (CI, 52% to 68%), and a negative predictive value of 100% (CI, 95% to 100% [81 of 81]). The D-dimer was positive in 0% (CI, 0% to 60%), 24% (CI, 14% to 37%), and 51% (CI, 40% to 61%) of
In unprovoked VTE patients the duration of anticoagulation is one of the most important unanswered questions clinicians face on a daily basis. Clinicians and patients must balance the long-term risks of recurrent VTE off anticoagulants with the long-term risk of major bleeding to come to a treatment decision regarding the duration of anticoagulation. In unselected unprovoked VTE patients, these long term risks are closely balanced [22]. Attention has turned towards identifying tools to risk stratify patients into sub-groups at low risk of VTE recurrence where anticoagulants can be safely discontinued and sub-groups at high risk of VTE recurrence where anticoagulants should clearly be continued. The tools explored to date include single predictors and clinical decision rules. The single predictors that have been examined including residual venous obstruction [23] and D-dimer [24]. The authors note significant practice variation where residual venous obstruction and D-Dimer is being used quite commonly in Italian centers but not in North American centers. The absence of a clear ability to importantly risk stratify patients with these predictors is likely the cause of this practice variation. Attention has turned to the development of CDRs to guide duration of anticoagulation. The DASH [25], Vienna [26] and “Men continue and HERDOO2” [27] CDRs have been developed but not yet validated and hence are not ready to be adopted in clinical practice. Some concerns have been raised in the derivation of these rules. For example, the Vienna and DASH scores did not include all important predictors at the derivation stage. Specifically, these rules did not collect and therefore could not include the most potent predictor of recurrent VTE identified in the REVERSE study [27], namely the post thrombotic findings of hyperpigmentation, redness or edema. Both the Vienna and DASH rules also require D-Dimer measurement 1 month after discontinuation of anticoagulation which some may consider impractical in busy clinical practices and may be risky as many patients must come off anticoagulants for a month to define if they are at high risk. During this month they may be exposed to recurrent VTE and indeed fatal recurrent VTE as has been reported in this off treatment testing interval [28]. The Vienna prediction rule also requires use of a complicated nomogram that may be challenging to adoption in routine clinical practice although the availability of an online version facilitates its use in clinical practice. The “Men continue and HERDOO2” rule is undergoing prospective validation in a large multi-national cohort study. The rule is simple to apply as it has 5 predictors (male gender, PTS findings of hyperpigmentation, edema or redness in either leg (HER), Vidas® D-Dimer N250 ug/ ml Older age (N 65), and Obesity (BMI over 30) on anticoagulants), an
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
easy to recall mnemonic is applied while patients are on anticoagulant treatment and thereby doesn’t require a repeat visit to apply nor expose patients to an off anticoagulant treatment testing interval. The rule suggests that all men and women with 2 or more of the HER DOO features continue anticoagulants longer term. There is also important controversy regarding the recurrent VTE threshold below which it is safe to discontinue anticoagulants or above which it is clearly of benefit to continue anticoagulants. Some experts have suggested that the lower threshold should be less than 5% [29] others have suggested that it should be less than 3% [22]. This discussion hinges on the absolute bleeding risk associated with ongoing anticoagulant therapy and the case fatality rates of major bleeding and recurrent VTE. Recent literature suggests that the major bleeding risk is low (1-2%) in vitamin K antagonist experienced patients and may be even lower with the novel oral anticoagulants (b1%) [30]. The latter argues in favor of using a lower recurrent VTE threshold to ensure patients wouldn’t benefit from ongoing anticoagulant therapy. Finally, there is a clear need to develop bleeding risk stratification tools to be married with recurrent VTE risk stratification tools to tailor patient's therapy choices. Challenges and Controversies in VTE Prediction: Therapeutic Management of VTE- PE Risk Stratification for Selection of PE Treatment Setting Patients experiencing PE face their highest mortality risks from PE in the minutes and days that follow the event. Treatment setting for the initial management of PE is an area of wide international practice variation. In some centers in Canada PE is treated in an outpatient setting in up to 55% of patients [31,32]. In these centers patients are selected for outpatient therapy based on the absence of 1) need for oxygen, 2) hemodynamic instability, 3) history of pre-syncope or syncope as presenting symptom and 4) contraindications to anticoagulant therapy. In many U.S. and European centers PE patients are treated exclusively as inpatients. Clinical decision rules have been derived and validated to objectively identify patients with a low mortality risk who may be treated as outpatient. The most extensively validated clinical decision rule is the Pulmonary Embolism Severity Index (PESI) [33]. The PESI comprises 11 easily available clinical variables and accurately stratifies patients into five risk classes (I-V) with increasing risk of all-cause 30day mortality, ranging from 1.1% in class I to 24.5% in class V. Patients in risk classes I and II are considered low-risk. In an international randomized non-inferiority trial, outpatient treatment with lowmolecular-weight heparin of selected low-risk patients based on the PESI was non-inferior to inpatient treatment in terms of efficacy and safety, was well accepted by patients, and reduced time spent in the hospital [34]. This clinical trial demonstrated that up to 30% of patients with PE could be safely treated in an outpatient setting based on the PESI and that doctors do not have to rely on markers of myocardial dysfunction or injury (e.g., brain natriuretic peptide, troponin, or echocardiographic measurements of right ventricular function) to safely identify candidates for outpatient care. Recently, a simplified 6-item version of the PESI (sPESI) has become available but requires prospective validation before its use can be recommended [35]. In a multicenter prospective management study from the Netherlands, 51% of patients with PE who were selected on the basis of 11 simple clinical criteria (HESTIA criteria) were safely managed as outpatients [36]. The safety of outpatient care based on the HESTIA criteria alone versus HESTIA criteria plus NTpro-BNP is currently being evaluated in the randomized VESTA trial (Netherlands Trial Registry no. 2603).At the other end of the spectrum, there is an identified need for prognostic PE rules that identify hemodynamically stable, high-risk patients that may require ICU admission or IVC filter insertion. Several biomarkers and right ventricular dilatation (RVD) on echocardiography or CT-scan have been associated with an increased risk of death or major complication related to PE. But most studies did not adjust for
767
clinical variables [37]. More recently, multicenter cohort studies using multivariable analysis suggest that Brain Natriuretic Peptide and RVD have prognostic value independently from the clinical variables and the PESI but these results have not been confirmed externally and the clinical significance of such findings is questionable [38,39]. In addition, the clinical benefit of the two main therapeutic options in these patients with so-called intermediate-risk PE, i.e. thrombolytic therapy and vena cava interruption remains to be confirmed. Conclusion In many areas of diagnostic and therapeutic management of the PE and DVT clinical decision rules and D-dimer are promising tools to improve cost-effective clinical care. However, knowledge translation barriers and wide international practice variation points to the need for future research. This future research will need to simplify these rules so they are ultimately adopted in routine clinical practice after adequate validation so that practice variation becomes limited. Technology may improve adoption of what we now perceive to be complex rules by simple automation of solicitation of clinical decision rule predictors, calculation of categorization and suggested subsequent management. With these improvements, optimal diagnostic and therapeutic management of VTE is likely to be adopted worldwide. Conflict of Interest Statement Marc Rodger has received grant funding from Biomerieux for the conduct of studies exploring the role of D-Dimer in predicting the risk of recurrent VTE in unprovoked VTE patients. Funding No funding was obtained to write this review Acknowledgements Marc Rodger was supported through the Heart and Stroke Foundation with a Career Investigator Award and is the recipient of a University of Ottawa Faculty of Medicine Research Chair Awards and a University of Ottawa Department of Medicine Research Salary Award. Dr. Phil Wells held a Canada Research Chair in Thromboembolic Diseases. References [1] Anderson Jr FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester study. Arch Intern Med 1991;151:933–8. [2] McGinn TG, Guyatt GH, Wyer PC, Naylor CD, Stiell IG, Richardson WS. Users' guides to the medical literature: XXII: how to use articles about clinical decision rules. Evidence-Based Medicine Working Group. JAMA 2000;284:79–84. [3] Wasson JH, Sox HC, Neff RK, Goldman L. Clinical prediction rules. Applications and methodological standards. N Engl J Med 1985;313:793–9. [4] Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ 2005;331:259–68. [5] Roy PM, Meyer G, Vielle B, Le Gall C, Verschuren F, Carpentier F, et al. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med 2006;144:157–64. [6] Douma RA, Mos IC, Erkens PM, Nizet TA, Durian MF, Hovens MM, et al. Performance of 4 Clinical Decision Rules in the Diagnostic Management of Acute Pulmonary Embolism: A Prospective Cohort Study. Ann Intern Med 2011;154:709–18. [7] Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med 2011;171:831–7. [8] Adam SS, Key NS, Greenberg CS. D-dimer antigen: current concepts and future prospects. Blood 2009;113:2878–87. [9] Stein PD, Hull RD, Patel KC, Olson RE, Ghali WA, Brant RF, et al. D-Dimer for the exclusion of acute venous thrombosis and pulmonary embolism. A systematic review. Ann Intern Med 2004;140:589–602. [10] Harper PL, Theakston E, Ahmed J, Ockelford P. D-dimer concentration increases with age reducing the clinical value of the D-dimer assay in the elderly. Intern Med J 2007;37:607–13.
768
M.A. Rodger et al. / Thrombosis Research 134 (2014) 763–768
[11] Douma RA, Le GG, Sohne M, Righini M, Kamphuisen PW, Perrier A, et al. Potential of an age adjusted D-dimer cut-off value to improve the exclusion of pulmonary embolism in older patients: a retrospective analysis of three large cohorts. BMJ 2010;340: c1475. [12] Douma RA, Tan M, Schutgens RE, Bates SM, Perrier A, Legnani C, et al. Using an agedependent D-dimer cut-off value increases the number of older patients in whom deep vein thrombosis can be safely excluded. Haematologica 2012;97:1507–13. [13] Schouten HJ, Koek HL, Oudega R, Geersing GJ, Janssen KJ, van Delden JJ, et al. Validation of two age dependent D-dimer cut-off values for exclusion of deep vein thrombosis in suspected elderly patients in primary care: retrospective, cross sectional, diagnostic analysis. BMJ 2012;344:e2985. [14] Wells PS, Owen C, Doucette S, Fergusson D, Tran H. Does this patient have deep vein thrombosis? JAMA 2006;295:199–207. [15] Wells PS, Ginsberg JS, Anderson DR, Kearon C, Gent M, Turpie AGG, et al. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med 1998;129:997–1005. [16] Wicki J, Perneger TV, Junod AF, Bounameaux H, Perrier A. Assessing clinical probability of pulmonary embolism in the emergency ward: a simple score. Arch Intern Med 2001;161:92–7. [17] Kjellberg U, Andersson NE, Rosen S, Tengborn L, Hellgren M. APC resistance and other haemostatic variables during pregnancy and puerperium. Thromb Haemost 1999;81:527–31. [18] Chan WS, Lee A, Spencer FA, Crowther M, Rodger M, Ramsay T, et al. Predicting deep venous thrombosis in pregnancy: out in "LEFt" field? Ann Intern Med 2009;151: 85–92. [19] Chan WS, Chunilal S, Lee A, Crowther M, Rodger M, Ginsberg JS. A red blood cell agglutination D-dimer test to exclude deep venous thrombosis in pregnancy. Ann Intern Med 2007;147:165–70. [20] Chan WS, Lee A, Spencer FA, Chunilal S, Crowther M, Wu W, et al. D-dimer testing in pregnant patients: towards determining the next "level" in the diagnosis of DVT. J Thromb Haemost 2010;8:1004–11. [21] Rodger M. Evidence base for the management of venous thromboembolism in pregnancy. Hematology Am Soc Hematol Educ Program 2010;2010:173–80. [22] Rodger M, Carrier M, Gandara E, Le GG. Unprovoked Venous Thromboembolism: Short term or Indefinite Anticoagulation? Balancing Long-Term Risk and Benefit. Blood Rev 2010;24:171–8. [23] Donadini MP, Ageno W, Antonucci E, Cosmi B, Kovacs MJ, Le GG, et al. Prognostic significance of residual venous obstruction in patients with treated unprovoked deep vein thrombosis. A patient-level meta-analysis. Thromb Haemost 2014;111:172–9. [24] Verhovsek M, Douketis JD, Yi Q, Shrivastava S, Tait RC, Baglin T, et al. Systematic review: D-dimer to predict recurrent disease after stopping anticoagulant therapy for unprovoked venous thromboembolism. Ann Intern Med 2008;149: 481–90 [W-94]. [25] Tosetto A, Iorio A, Marcucci M, Baglin T, Cushman M, Eichinger S, et al. Predicting disease recurrence in patients with previous unprovoked venous thromboembolism: a proposed prediction score (DASH). J Thromb Haemost 2012;10:1019–25.
[26] Eichinger S, Heinze G, Jandeck LM, Kyrle PA. Risk assessment of recurrence in patients with unprovoked deep vein thrombosis or pulmonary embolism: the Vienna prediction model. Circulation 2010;121:1630–6. [27] Rodger MA, Kahn SR, Wells PS, Anderson DA, Chagnon I, Le Gal G, et al. Identifying unprovoked thromboembolism patients at low risk for recurrence who can discontinue anticoagulant therapy. CMAJ 2008;179:417–26. [28] Cosmi B, Legnani C, Tosetto A, Pengo V, Ghirarduzzi A, Testa S, et al. Usefulness of repeated D-dimer testing after stopping anticoagulation for a first episode of unprovoked venous thromboembolism: the PROLONG II prospective study. Blood 2010; 115:481–8. [29] Baglin T, Bauer K, Douketis J, Buller H, Srivastava A, Johnson G. Duration of anticoagulant therapy after a first episode of an unprovoked pulmonary embolus or deep vein thrombosis: guidance from the SSC of the ISTH. J Thromb Haemost 2012;10: 698–702. [30] Castellucci LA, Cameron C, Le Gal G, Rodger MA, Coyle D, Wells PS, et al. Safety and efficacy outcomes of oral anticoagulants and antiplatelet drugs in the secondary prevention of venous thromboembolism: systematic review and network metaanalysis. BMJ 2013;347:f5133. [31] Erkens PM, Gandara E, Wells PS, Shen AY, Bose G, Le Gal G, et al. Safety of outpatient treatment in acute pulmonary embolism. J Thromb Haemost 2010;8:2412–7. [32] Kovacs MJ, Hawel JD, Rekman JF, Lazo-Langner A. Ambulatory management of pulmonary embolism: a pragmatic evaluation. J Thromb Haemost 2010;8:2406–11. [33] Aujesky D, Obrosky DS, Stone RA, Auble TE, Perrier A, Cornuz J, et al. Derivation and validation of a prognostic model for pulmonary embolism. Am J Respir Crit Care Med 2005;172:1041–6. [34] Aujesky D, Roy PM, Verschuren F, Righini M, Osterwalder J, Egloff M, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet 2011;378:41–8. [35] Jimenez D, Aujesky D, Moores L, Gomez V, Lobo JL, Uresandi F, et al. Simplification of the pulmonary embolism severity index for prognostication in patients with acute symptomatic pulmonary embolism. Arch Intern Med 2010;170:1383–9. [36] Zondag W, Mos IC, Creemers-Schild D, Hoogerbrugge AD, Dekkers OM, Dolsma J, et al. Outpatient treatment in patients with acute pulmonary embolism: the Hestia Study. J Thromb Haemost 2011;9:1500–7. [37] Sanchez O, Trinquart L, Colombet I, Durieux P, Huisman MV, Chatellier G, et al. Prognostic value of right ventricular dysfunction in patients with haemodynamically stable pulmonary embolism: a systematic review. Eur Heart J 2008;29:1569–77. [38] Sanchez O, Trinquart L, Caille V, Couturaud F, Pacouret G, Meneveau N, et al. Prognostic factors for pulmonary embolism: the prep study, a prospective multicenter cohort study. Am J Respir Crit Care Med 2010;181:168–73. [39] Sanchez O, Trinquart L, Planquette B, Couturaud F, Verschuren F, Caille V, et al. Echocardiography and pulmonary embolism severity index have independent prognostic roles in pulmonary embolism. Eur Respir J 2013;42:681–8.
