Elevated D-dimer levels predict recurrence in patients with idiopathic venous thromboembolism: a meta-analysis

Transcription

1 Journal of Thrombosis and Haemostasis, 7: DOI: /j x ORIGINAL ARTICLE Elevated D-dimer levels predict recurrence in patients with idiopathic venous thromboembolism: a meta-analysis E. BRUINSTROOP,* F. A. KLOK,* M. A. VAN DE REE, F. L. OOSTERWIJKà and M. V. HUISMAN* *Leiden University Medical Center, Section of Vascular Medicine, Department of General Internal Medicine-Endocrinology, Leiden; Diakonessenhuis, Department of Internal Medicine, Bosboomstraat 1, Utrecht; and àuniversity Medical Center Utrecht, Utrecht, the Netherlands To cite this article: Bruinstroop E, Klok FA, van de Ree MA, Oosterwijk FL, Huisman MV. Elevated D-dimer levels predict recurrence in patients with idiopathic venous thromboembolism: a meta-analysis. J Thromb Haemost 2009; 7: Summary. Background: The evidence on the optimal duration of treatment in patients with an idiopathic venous thromboembolic event (VTE) is inconclusive. D-dimer testing to predict recurrent VTE has been evaluated in several studies. Objectives: We performed a meta-analysis of studies of patients with idiopathic VTE treated with oral anticoagulation therapy (OAT) to assess the prognostic value of elevated D-dimer levels 1 month after discontinuation of OAT for VTE recurrence. Patients/Methods: The MEDLINE, EMBASE and Cochrane databases were searched to identify relevant studies. Studies were eligible for inclusion if they included patients with idiopathic VTE and in addition reported results for this group separately, had measured D-dimer approximately 1 month after discontinuation of OAT and had reported on recurrence of VTE. A random-effects model was used to pool study results. Results: Data from four studies (1539 patients) were included in the current analysis. All studies reported on the number of recurrent events in the normal and elevated D-dimer groups. Overall, 125 of 751 patients (16.6%) with elevated D-dimer levels experienced recurrent VTE during the period of follow-up compared with 57 of 788 patients (7.2%) with normal D-dimer levels. Elevated D-dimer levels were significantly associated with recurrent VTE (odds ratio, 2.36; 95% CI, 1.65 to 3.36). Conclusions: Elevated D-dimer levels measured 1 month after discontinuation of OAT identify patients with idiopathic VTE at higher risk of recurrence. Correspondence: Menno V. Huisman, Leiden University Medical Center, Section of Vascular Medicine, Department of General Internal Medicine-Endocrinology, Room c4-68, PO Box 9600, 2300 RC Leiden, the Netherlands. Tel.: ; fax: m.v.huisman@lumc.nl Current address of E. Bruinstroop: Academic Medical Center, University of Amsterdam, Department of Endocrinology and Metabolism, F5-162, 1105 AZ Amsterdam, the Netherlands. Received 6 September 2008, accepted 2 January 2009 Keywords: fibrin fragment D, prognosis, pulmonary embolism, recurrence, venous thrombosis. Introduction The treatment of patients with a venous thromboembolic event (VTE) is challenging, as available evidence on the optimal length of treatment is inconclusive [1]. A primary goal in the treatment of VTE is reducing the risk of recurrence. It is well established that different risk categories can be identified in patients with VTE. After a first idiopathic event, patients have a risk of recurrence in the first year of approximately 5%, with the risk decreasing in the following years [2]. Patients experiencing VTE due to transient risk factors (i.e. trauma, immobilization) are at lesser risk of recurrence. In contrast, chronic risk factors (i.e. antiphospholipid antibody syndrome) are associated with an increased risk of recurrence. Different treatment regimens can be applied to these three groups [3]. Both discontinuation and prolongation of oral anticoagulant therapy (OAT) are associated with risks, considering the morbidity and possible mortality of a recurrent event compared with the risk of serious hemorrhage of 1 3% during 1 year with treatment, with a case fatality rate of 20% [4]. The decision regarding continuation of treatment is most challenging in the group of idiopathic venous thrombosis. Several factors have been investigated as possible prognostic markers for recurrence, including age, gender, the duration of OAT, factor (F) VIII coagulant activity and APTT levels [5 7]. D-dimer testing is widely accepted as a useful test in the diagnosis of VTE [8], but is not currently used as a prognostic indicator. Nevertheless, D-dimer levels after discontinuation of oral anti-coagulation therapy have been shown to be a promising new prognostic marker for VTE recurrence [9]. Evidence from the literature is, however, inconclusive on this issue. We performed a meta-analysis aimed at assessing the prognostic value of D-dimer 1 month after discontinuation of OAT for recurrent events in patients with an idiopathic deep venous thrombosis, pulmonary embolism or both.

2 612 E. Bruinstroop et al Methods The methods for this meta-analysis are in accordance with Meta-Analysis of Observational Studies in Epidemiology: A Proposal for Reporting [10]. Data sources and searches We searched the MEDLINE database until March Furthermore, reference lists of retrieved articles and review articles were reviewed manually to implement our search. Search criteria included synonyms of the terms Ôvenous thromboembolic eventõ, Ôd dimerõ and ÔrecurrenceÕ (Fig. 1). The synonyms were searched in Title/Abstract. The search was not limited to the English language. Finally, the EMBASE and Cochrane databases were reviewed until March 2008 to ensure the completeness of the search. Study selection Studies were included in this analysis if they reported prospectively on consecutive patients with an objective diagnosis of first idiopathic VTE treated with OAT, with D-dimer levels measured 1 month after discontinuation of OAT, and objectively determined VTE recurrence. One author (E.B.) performed the electronic search and listed the trials that were eligible for inclusion in the study. Study selection was initially performed by review of title. Candidate full-text articles were then reviewed and selected for data retrieval. Two authors (E.B. and F.O.) independently reviewed each study for quality assessment and Deep venous thrombosis OR venous thrombosis OR pulmonary embolism OR venous thromboembolism (30952 hits) AND D dimer (3665 hits) AND Recurrence OR risk OR follow up OR prognosis ( hits) 391 papers 368 Deleted by review of title 93 Review papers 1 Meta-analysis 127 Diagnosis first event 4 Diagnosis recurrence 55 Risk factors VTE 36 D dimer/vte and other disease 21 D dimer assay 7 Therapy VTE 2 Complications VTE 2 Outcome independent Of D-dimer 12 Biochemics Hemostatic factors 4 Case Report 1 Author reply 1 Comment 2 Clinical Guideline 23 papers undergoing full text review 19 Deleted by full text review 5 D-dimer measured at diagnosis 2 D-dimer measured before diagnosis 1 D-dimer measured during OAT 3 D-dimer measured > 1 month after discontinuation 2 Different outcome 3 Different predictor 2 Different study population 1 Report on same patients 4 studies included in the analysis Fig. 1. Flow diagram for study selection. VTE, venous thromboembolic event; OAT, oral anticoagulant therapy.

3 D-dimer and recurrent venous thromboembolism 613 Table 1 Quality assessment studies Eichinger et al Palareti et al Palareti et al Baglin et al Description of patient sample + +/) + +/) Description of inclusion and exclusion criteria Potential selection bias ) ) ) ) Completeness of follow-up (% loss-to-follow-up of patients eligible for study) 6% 1% 0% 1% A priori definition of study outcomes Objectivity of outcomes Definition and measurement of prognostic variables and treatment /) extracted data on studies and patient characteristics, as well as outcomes. Disagreements were resolved through revision by an additional reviewer (F.K.) and by discussion. Data extraction and quality assessment Studies were assessed for the presence of eight features: description of patient sample characteristics, description of inclusion and exclusion criteria, potential selection bias, completeness of follow-up, aprioridefinition of study outcomes, objectivity of outcomes, and definition and measurement of prognostic variables and treatment. All included studies reported on the number of recurrent events in the normal and elevated D-dimer groups. This allowed the creation of a 2 2tablebasedonD-dimer levels (normal or elevated) and recurrence (yes or no). Data synthesis and analysis Meta-analyses of all outcomes are reported using fixed-effects models. CochranÕs chi-square test and the I2 test for heterogeneity were used to assess between-study heterogeneity. Statistically significant heterogeneity was considered present if chi-square P <0.10andI 2 > 50%. Pooled odds ratios (ORs) were reported with 95% confidence intervals (CIs). Analyses were performed with Review Manager 4.2 (The Cochrane Collaboration, Oxford, England). The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agreed to the manuscript as written. Results A total of 391 articles were established in our search (Fig. 1). Articles were excluded by review of title if it concerned review papers, meta-analyses, case reports, author replies, comment or clinical guidelines, or papers on diagnosis of first or recurrent events, on risk factors of VTE, on D-dimer related to other disease, on technical aspects of D-dimer assays, on therapy or complications of VTE, or on biochemics of hemostatic factors, and if it concerned other variables then D-dimer for our outcome. Twenty-three studies underwent full text review. Articles were excluded if they reported on measurement of D-dimer before diagnosis [11,12], at diagnosis [13 17], during treatment [18] and for longer then 1 month after discontinuation of treatment [6,19,20]. Other studies did not match study population [21,22], predictor [7,23,24] or outcome [25,26]. One study was excluded due to overlapping study population [27]. A search of the EMBASE and Cochrane databases did not yield any more relevant results. Quality assessment Overall, four studies [28 31] were selected for quality assessment (Table 1). All studies described the patient sample, although two studies have only described the total study population and not idiopathic patients separately [29,31]. The inclusion and exclusion criteria were clearly described in all studies, and there was no selection bias in the selected studies. All studies had a reasonable amount of loss to follow-up. The outcomes were aprioridefined and objectively determined. The definitions of the prognostic variable and initial therapy were defined, although in the study by Baglin et al. [31] the minimal initial treatment period was not stated. Baglin et al. [31] reported that OAT was continued for a period of 6 months; the other three studies defined a minimum treatment period of 3 months. Study population The study populations of the included studies are depicted in Table 2. Eichinger et al. [28] included 610 consecutive patients with idiopathic VTE. Of the patients with idiopathic VTE, 175 patients did not complete the total duration of follow-up due to several reasons, but were followed-up until time of exclusion or death. Therefore, they were included in the original study and the present meta-analysis. Palareti et al. (2003) [29] categorized 599 patients into different subgroups (unprovoked VTE, VTE due to permanent risk factors, VTE due to temporary risk factors and unknown). In this meta-analysis the data of the 282 patients with unprovoked VTE were taken from the original study. Palareti et al. (2006) [30] included 608 patients in an intention to treat analysis. Patients with elevated levels of D-dimer were randomized to resume or stop OAT. For this meta-analysis we analyzed the data from 505 patients with normal D-dimer levels (n = 385) and patients with elevated D-dimer levels assigned to stop OAT (n = 120). Three patients were lost to follow-up, but were followed until time of exclusion and they were included in the analysis of the original study. Twentytwo patients resumed OAT because of clinical conditions

4 614 E. Bruinstroop et al Table 2 Study populations of the included studies Study Inclusion and exclusion criteria of original studies Eligible original study (n) Excluded from analysis in original study (n) Study population in original study (n) Eichinger et al Palareti et al Palareti et al Baglin et al Inclusion: consecutive patients with unprovoked VTE older then 18 years, after a 3 month treatment period Exclusion: more than 1 previous VTE (n = 332), surgery, pregnancy, trauma within previous 3 months (n = 393), deficiency of natural coagulation inhibitor (n = 59), lupus anticoagulant (n = 32), cancer (n = 282), long-term anticoagulant treatment (n = 336) Inclusion: consecutive patients with first episode of DVT of the lower limbs and/or PE, who discontinued OAT after a 3 month treatment period Inclusion: first episode of symptomatic, unprovoked VTE including proximal DVT of the lower limbs, PE or both, years, after a 3 month treatment period Exclusion: secondary to removable or non-removable risk factors, indications or contra-indications for OAT, limited life expectancy, living far from study center Inclusion: consecutive patients with objectively confirmed first episode of proximal lower limb DVT, with or without symptomatic pulmonary embolism Exclusion: post-operative VTE (within 6 weeks), calf vein thrombosis, pregnancy-associated thrombosis, antiphospholipid antibody syndrome, cancer, indication for prolonged anticoagulation, elevated anticardiolipin titres, lupus anticoagulant total* (0 resumed OAT) patients followed up until time of exclusion: 11 diagnosed with cancer 105 resumed OAT for other reasons than VTE 17 became pregnant 37 lost to follow up 5 died of other cause total 15 no thrombophilia screening performed 5 lupus anticoagulant detected 8 lost to follow-up 8 D-dimer testing not possible total 3 declined participation 5 recurrent VTE during first month after stop OAT 11 lupus anticoagulant or antithrombin deficiency diagnosed and OAT resumed total 1 recurrent VTE during first month after stop OAT 8 died during initial OAT 27 diagnosis of cancer 19 diagnosis of antiphosholipid antibody syndrome 11 continued OAT 599 total ( ) 317 provoked VTE 282 unprovoked VTE* 608 total 385 normal D-dimer levels* (14 resumed OAT) 120 elevated D-dimer levels assigned to stop OAT* (8 resumed OAT) 103 elevated D-dimer levels assigned to resume OAT patients followed up until time of exclusion: 3 lost to follow-up 272 total (0 resumed OAT) 130 provoked VTE 142 unprovoked VTE* patients followed up until time of exclusion: 5 lost to follow-up 72 died of other cause (10 postmortem examination) OAT, oral anticoagulant therapy; VTE, venous thromboembolic event.*patients included in meta-analysis. Number of patients that resumed oral anticoagulation therapy (OAT) not reported.

5 D-dimer and recurrent venous thromboembolism 615 Table 3 Study characteristics of included studies Follow-up in years (range) Outcome Index event %PE %DVT %both D-dimer assay % Male Mean age in years (range) Study 3.17 Recurrent symptomatic DVT or recurrent symptomatic pulmonary embolism, confirmed by venography, color duplex sonography, VP scannng or spiral CT Eichinger et al (43% < 45 y) Asserachrom D-dimer (Boehringer 57.2 à Mannheim) 3 weeks after stop OAT 1.45* Objectively documented DVT recurrence and/or fatal or non-fatal PE (first event or recurrence) VIDAS D-dimer ELISA (biomerieux) days after stopping OAT Palareti et al (12 91)* 50* 3.7* 73.6* 22.7* 1.37 Objectively confirmed recurrent VTE Clearview Simplify D-dimer assay (Inverness Medical Professional Diagnostics) days after stopping OAT Palareti et al Objectively determined cases of unprovoked recurrence with new or extended clot, resulting in reintroduction of OAT 3.22* ( ) MDA D-dimer assay (Trinity Biotech Ltd. Co.) 1 2 months after stopping OAT Baglin et al (18 99)* 52* 42* 58* DVT, deep venous thrombosis; PE, pulmonary embolism; VTE, venous thromboembolic event; OAT, oral anticoagulant therapy; CT, computed tomography; VP, ventilation perfusion lung scan. *Characteristics of total study population and not selected subgroup used for analysis as this was not reported. Patients with both PE and DVT were categorizsed as symptomatic PE, à 4.4% axillary veins. (eight patients in the elevated D-dimer group), but were analysed in the group to which they were randomized. Baglin et al. [31] studied 272 patients with unprovoked and provoked VTE. Of these 272 patients, 77 did not complete the total duration of follow-up due to loss to follow-up or death but were included in their analysis until time exclusion or death. For this analysis, we extracted the data from the 142 patients with unprovoked VTE. D-dimer assay The study characteristics are shown in Table 3. Different D-dimer assays were applied in the selected studies. Palareti et al. (2003) [29] used a quantitative measurement of D-dimer (VIDAS D-dimer ELISA, biomerieux, Lyon, France) days after discontinuation of OAT with a cut-off value of 500 ng ml )1. Eichinger et al. [28] used the quantitative Asserachrom D-dimer (Boehringer Mannheim, Mannheim, Germany) 3 weeks after discontinuation of OAT and divided the values into four subgroups. In their final analysis, the cutoff with the best prognostic value was 250 ng ml )1.Palareti et al. (2006) [30] used the qualitative Clearview Simplify D-dimer assay (Inverness Medical Professional Diagnostics, Bedford, UK and Luisville, CO, USA) days after discontinuation of OAT. Baglin et al. [31] used a quantitative MDA D-dimer assay (Trinity Biotech Ltd. Co., Wicklow, Ireland) 1 2 months after stopping OAT. For their analysis they applied a cut-off value of 500 ng ml )1. Recurrent VTE The results on recurrent VTE are shown in Fig. 2. Eichinger et al. [28] measured recurrent symptomatic DVT or PE, confirmed by radiological investigations for a mean follow-up period of 3.17 years. In the elevated D-dimer group 63 patients (15.7%) had recurrent VTE, compared with 16 patients (7.7%) in the normal D-dimer group. Elevated D-dimer was significantly associated with recurrence (OR, 2.25; 95% CI, ). Palareti et al. (2003) [29] considered objectively documented DVT and/or fatal or non-fatal PE as outcomes. During a mean of 1.45 follow-up years, they found 23 patients (16.5%) with recurrent VTE in the elevated D-dimer group and 10 patients (7.0%) with recurrence in the normal D-dimer group. Elevated D-dimer was significantly associated with recurrence (OR, 2.64; 95% CI, ). Palareti et al. (2006) [30] objectively evaluated all outcome events and deaths during a mean follow-up period of 1.37 years. Eighteen patients (15%) with recurrent VTE were observed in the elevated D-dimer group, compared with 24 patients (6.2%) in the normal D-dimer group. Elevated D-dimer was significantly associated with recurrence (OR, 2.65; 95% CI, ). Baglin et al. [31] followed patients yearly until closure of the study (mean follow-up of 3.22 years). Clinical events were objectively confirmed by review of clinical records and

6 616 E. Bruinstroop et al Study or sub-category Eichinger 2003 Palareti 2003 Palareti 2006 Baglin 2006 Elevated D-dimer n/n 63/401 23/139 18/120 21/91 Normal D-dimer n/n 16/209 10/143 24/385 7/51 Total (95% CI) Total events: 125 (Elevated D-dimer), 57 (Normal D-dimer) Test for heterogeneity: χ 2 = 0.45, df = 3 (P = 0.93), I 2 = 0% Test for overall effect: Z = 4.75 (P < ) OR (fixed) 95% CI Weight % OR (fixed) 95% CI [1.26, 4.00] [1.21, 5.77] [1.39, 5.08] [0.74, 4.80] 2.36 [1.65, 3.36] Fig. 2. D-dimer level and recurrent venous thromboembolism at follow-up. n = patients with recurrent VTE, N = all patients in the elevated or normal D-dimer group. radiological evidence. All cases with confirmed new or extended clots, resulting in reintroduction of OAT, were recorded. Only patients with unprovoked recurrence were included in the analysis of the original study. In the elevated D- dimer group, 21 patients (23%) experienced recurrence, compared with seven patients (14%) in the normal D-dimer group. Elevated D-dimer was non-significantly associated with recurrence (OR, 1.89; 95% CI, ). There were seven additional cases of provoked recurrence, four in patients with elevated D-dimer levels and three in patients with normal D- dimer levels. As it is not clear if these patients had a first idiopathic or provoked event, they were not included in the present analysis. Overall, 125 of 751 patients with elevated D-dimer levels experienced recurrence during the period of follow-up (16.6%) compared with 57 of 788 with normal D-dimer levels (7.2%). Elevated D-dimer levels 1 month after discontinuation of OAT for an idiopathic VTE were significantly associated with recurrent VTE (OR, 2.36; 95% CI, 1.65 to 3.36) (Fig. 2). Discussion In the present study we observed a more than twofold odds of recurrence in patients with idiopathic VTE and an elevated D-dimer level measured 1 month after discontinuation of OAT. These findings suggest that this widely available test could help in the selection of a subpopulation of patients with idiopathic venous thromboembolism suitable for continuation of OAT. Palareti et al. [30] showed in a randomized controlled trial that patients with elevated D-dimer values that continued with treatment had a significantly lower incidence of recurrence than patients with normal D-dimer values, who had discontinued treatment. Treatment decisions based on this prognostic marker could therefore lower the incidence of VTE recurrence. This meta-analysis focused on D-dimer as a predictor of recurrence. Of the 1539 patients in our analysis, 751 patients (49%) had an elevated D-dimer level. In this group, 125 patients (17%; 95% CI, 14 20%) eventually had a recurrent event and could have therefore benefited from continuation of treatment. Furthermore, 57 patients (7%; 95% CI, 6 9%) with normal levels of D-dimer had a recurrent VTE during the follow-up period. These results show that the use of the D-dimer test in isolation does not appear to be adequate to tailor the duration of anticoagulation. Baglin et al. [31] concluded that of the clinical risk factors male sex was most strongly associated with recurrence. A large prospective cohort study on the risk factors of recurrence, which has not measured levels of D-dimer, concluded that thrombophilia (adjusted HR, 1.91; 95% CI, ), presentation with DVT (adjusted HR, 1.50; 95% CI, ), shorter duration of anticoagulation (adjusted HR, 1.46; 95% CI, ) and aging (adjusted HR, 1.09; 95% CI, ) were independent risk factors for recurrence in patients with unprovoked VTE, while male sex was not (adjusted HR, 1.21; 95% CI, ) [5]. Strategies that incorporate a number of high-performing baseline and postbaseline predictors could be more effective in predicting recurrent VTE and should be tested. If OAT is continued, there is no evidence regarding the duration of this extended therapy. A strength of this meta-analysis is that we looked exclusively at idiopathic thrombosis, a cohort with relatively high recurrence rates [5]. Of note, in the studies by Palareti et al. (2003) [29] and Baglin et al. [31], patients with provoked VTE were included as well. In the first study [29] the OR of elevated D- dimer and recurrence increased to 3.12 (95% CI, ) when all subpopulations were taken into account, which was mainly caused by the subgroup of patients with thrombophilia, (OR, 8.43; 95% CI, ). In the overall study population of the second study [31] with idiopathic and provoked VTE, the OR of elevated D-dimer and recurrence was 1.24 (95% CI, ). This study showed that in the group of provoked VTE an elevated level of D-dimer is associated with less recurrence (OR, 0.61; 95% CI, ). In conclusion, it is in our view necessary to look at idiopathic and provoked VTE separately in relation to risk of recurrent VTE. A limitation of our study is that different D-dimer assays with different cut-off points were used in the included studies. There is no concluding evidence concerning the optimal cut-off value for D-dimer assays as a prognostic tool. In the second study by Palareti et al. [30] a qualitative method was used that showed good concordance with the quantitative VIDAS D-dimer test (cut-off point 500 ng ml )1 ) used in the first study by the same author [29] (j = 0.78; 95% CI, ) [32]. Eichinger et al. [28] attempted to retrospectively determine the ideal cut-off value, which might influence the results. They proposed a cut-off of 250 ng ml )1, which was obviously lower

7 D-dimer and recurrent venous thromboembolism 617 than the cut-off value of this assay used in the diagnostic workup (500 ng ml )1 ). If a cut-off of 500 ng ml )1 wasusedinthis study, the predicting quality of the test would decrease (OR, 1.45; 95% CI, ). Baglin et al. [31] prospectively set the cut-off at 500 ng ml )1, and retrospectively performed a ROC curve analysis on the quantitative data. ROC curve analysis showed an area under the curve of 0.55 (95% CI, ) and did not indicate a more preferable cut-off. In a recent study [33] an extended analysis was performed on 321 plasma samples of patients from one of the included studies [30]. Next to the qualitative D-dimer assay, four quantitative assays were assessed. In contrast to the results from Eichinger et al. [28] it was shown in retrospect that all qualitative assays have higher preferable cut-off values for prognosis compared with the cutoff values for diagnosis of VTE. An accurate cut-off value for diagnosis for the different D-dimer assays should be established in future research and prospectively tested. All studies in this meta-analysis have measured D-dimer 1 month after discontinuation of treatment. Importantly, in a clinical setting, stopping anticoagulation and resuming after 1 month is not practical and exposes high-risk patients to a long period without anticoagulants. A final limitation is the different duration of follow-up among the included studies. However, in all studies, patients with elevated D-dimer levels had a consistently higher risk of recurrence compared with patients with normalized D-dimer concentrations, independent of follow-up duration. Our findings are consistent with a recently published systematic review [34]. There are, however, some important differences between both reviews. First, small subpopulations of three additional studies were included in the other review. These studies were excluded from our analysis because in the complete study population D-dimer levels were measured in a wide range of 12 days up to 2 years after stopping OAT [6], the results of patients with idiopathic VTE were not given separately [35] and a third study was only available in abstract form [36]. Second, we report a pooled odds ratio compared with an annualized risk in the study by Verhovsek et al. While the latter represents an absolute risk, it seems to be confounded by follow-up duration, as in their review there is a significant negative correlation between total follow-up (patient-years of follow-up/patients) and annualized risk in the elevated D-dimer group (linear regression P = 0.04). This shows that annualized risk is probably not static over the years, but decreases over time, as also shown by Prandoni et al. [5]. The low number of patients in the three additional studies and differences in follow-up duration have in our view largely contributed to the significant heterogeneity (chi-square 23,8, I 2 75%, P <0.001) in their review [34] compared with our study, in which no significant heterogeneity was observed (chi-square 0.45, I 2 0%, P =0.93). In summary, elevated D-dimer levels measured 1 month after discontinuation of OAT identify patients with idiopathic VTE at a significantly higher risk of recurrent VTE. Future studies should focus on optimizing a multi-risk factor model, determining optimal D-dimer cut-off levels, and optimal duration of extended therapy in patients with elevated D-dimer levels. If these factors can be optimized, D-dimer is a very promising new prognostic marker, and very useful for the clinician in selecting patients at high risk of recurrent VTE. Disclosure of Conflict of Interests The authors state that they have no conflict of interest. References 1 Agnelli G, Prandoni P, Santamaria MG, Bagatella P, Iorio A, Bazzan M, Moia M, Guazzaloca G, Bertoldi A, Tomasi C, Scannapieco G, Ageno W; Warfarin Optimal Duration Italian Trial Investigators. Three months versus one year of oral anticoagulant therapy for idiopathic deep venous thrombosis. Warfarin Optimal Duration Italian Trial Investigators. NEnglJMed2001; 345: Heit JA, Mohr DN, Silverstein MD, Petterson TM, OÕFallon WM, Melton LJ III. Predictors of recurrence after deep vein thrombosis and pulmonary embolism: a population-based cohort study. Arch Intern Med 2000; 160: Kearon C. Duration of anticoagulation for venous thromboembolism. J Thromb Thrombolysis 2001; 12: Palareti G, Leali N, Coccheri S, Poggi M, Manotti C, DÕAngelo A, PengoV,ErbaN,MoiaM,CiavarellaN,DevotoG,BerrettiniM, Musolesi S. Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaborative study (ISCOAT). Italian Study on Complications of Oral Anticoagulant Therapy. Lancet 1996; 348: PrandoniP,NoventaF,GhirarduzziA,PengoV,BernardiE,PesaventoR,LottiM,TormeneD,SimioniP,PagnanA.Theriskof recurrent venous thromboembolism after discontinuing anticoagulation in patients with acute proximal deep vein thrombosis or pulmonary embolism. A prospective cohort study in 1,626 patients. Haematologica 2007; 92: Shrivastava S, Ridker PM, Glynn RJ, Goldhaber SZ, Moll S, Bounameaux H, Bauer KA, Kessler CM, Cushman M. D-dimer, factor VIII coagulant activity, low-intensity warfarin and the risk of recurrent venous thromboembolism. JThrombHaemost2006; 4: Legnani C, Mattarozzi S, Cini M, Cosmi B, Favaretto E, Palareti G. Abnormally short activated partial thromboplastin time values are associated with increased risk of recurrence of venous thromboembolism after oral anticoagulation withdrawal. Br J Haematol 2006; 134: Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, Kovacs G, Mitchell M, Lewandowski B, Kovacs MJ. Evaluation of D- dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003; 349: Cosmi B, Palareti G. The new role of D-dimer as a predictor of recurrent venous thromboembolism. JLabMed2005; 29: Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283: Folsom AR, Cushman M, Heckbert SR, Rosamond WD, Aleksic N. Prospective study of fibrinolytic markers and venous thromboembolism. J Clin Epidemiol 2003; 56: Cushman M, Folsom AR, Wang L, Aleksic N, Rosamond WD, Tracy RP, Heckbert SR. Fibrin fragment D-dimer and the risk of future venous thrombosis. Blood 2003; 101: KlokFA,DjurabiRK,NijkeuterM,EikenboomHC,LeebeekFW, Kramer MH, Kaasjager K, Kamphuisen PW, Büller HR, Huisman

8 618 E. Bruinstroop et al MV. High D-dimer level is associated with increased 15-d and 3 months mortality through a more central localization of pulmonary emboli and serious comorbidity. Br J Haematol 2008; 140: Grau E, Tenias JM, Soto MJ, Gutierrez MR, Lecumberri R, Perez JL, Tiberio G, RIETE Investigators. D-dimer levels correlate with mortality in patients with acute pulmonary embolism: Findings from the RIETE registry. Crit Care Med 2007; 35: Aujesky D, Roy PM, Guy M, Cornuz J, Sanchez O, Perrier A. Prognostic value of D-dimer in patients with pulmonary embolism. Thromb Haemost 2006; 96: Janata K, Holzer M, Domanovits H, Mullner M, Bankier A, Kurtaran A, Bankl HC, Laggner AN. Mortality of patients with pulmonary embolism. Wien Klin Wochenschr 2002; 114: Goldenberg NA, Knapp-Clevenger R, Manco-Johnson MJ. Elevated plasma factor VIII and D-dimer levels as predictors of poor outcomes of thrombosis in children. N Engl J Med 2004; 351: Kuruvilla J, Wells PS, Morrow B, MacKinnon K, Keeney M, Kovacs MJ. Prospective assessment of the natural history of positive D-dimer results in persons with acute venous thromboembolism (DVT or PE). Thromb Haemost 2003; 89: Ombandza-Moussa E, Samama MM, Horellou MH, Elalamy I, Conard J. Potential use of D-dimer measurement in patients treated with oral anticoagulant for a venous thromboembolic episode. Int Angiol 2003; 22: Palareti G, Guazzaloca G, Legnani C, Fortunato G, Grauso F, De Rosa V, Rodorigo G, Cosmi B, Coccheri S. High negative predictive value for recurrence of venous thromboembolism with d-dimer carried out three months after the suspension of treatment with oral anticoagulants in patients over 60. Minerva Cardioangiol 1998; 46: Cosmi B, Legnani C, Cini M, Guazzaloca G, Palareti G. The role of D-dimer and residual venous obstruction in recurrence of venous thromboembolism after anticoagulation withdrawal in cancer patients. Haematologica 2005; 90: PalaretiG,LegnaniC,CosmiB,GuazzalocaG,PancaniC,Coccheri S. Risk of venous thromboembolism recurrence: high negative predictive value of D-dimer performed after oral anticoagulation is stopped. Thromb Haemost 2002; 87: Cosmi B, Legnani C, Cini M, Guazzaloca G, Palareti G. D-dimer levels in combination with residual venous obstruction and the risk of recurrence after anticoagulation withdrawal for a first idiopathic deep vein thrombosis. J Thromb Haemost 2005; 94: Palareti G, Legnani C, Cosmi B, Guazzaloca G, Cini M, Mattarozzi S. Poor anticoagulation quality in the first 3 months after unprovoked venous thromboembolism is a risk factor for long-term recurrence. JThrombHaemost2005; 3: Kaczynska A, Kostrubiec M, Pacho R, Kunikowska J, Pruszczyk P. Elevated D-dimer concentration identifies patients with incomplete recanalization of pulmonary artery thromboemboli despite 6 months anticoagulation after the first episode of acute pulmonary embolism. Thromb Res 2008; 122: Janssen MC, Verbruggen H, Wollersheim H, Hoogkamer B, van Langen H, Novakova IR. D-dimer determination to assess regression of deep venous thrombosis. Thromb Haemost 1997; 78: Cosmi B, Legnani C, Cini M, Favaretto E, Palareti G. D-dimer and factor VIII are independent risk factors for recurrence after anticoagulation withdrawal for a first idiopathic deep vein thrombosis. Thromb Res 2008; 122: Eichinger S, Minar E, Bialonczyk C, Hirschl M, Quehenberger P, Schneider B, Weltermann A, Wagner O, Kyrle PA. D-dimer levels and risk of recurrent venous thromboembolism. JAMA 2003; 290: Palareti G, Legnani C, Cosmi B, Valdre L, Lunghi B, Bernardi F, Coccheri S. Predictive value of D-dimer test for recurrent venous thromboembolism after anticoagulation withdrawal in subjects with a previous idiopathic event and in carriers of congenital thrombophilia. Circulation 2003; 108: Palareti G, Cosmi B, Legnani C, Tosetto A, Brusi C, Iorio A, Pengo V, Ghirarduzzi A, Pattacini C, Testa S, Lensing AW, Tripodi A; PROLONG Investigators. D-dimer testing to determine the duration of anticoagulation therapy. N Engl J Med 2006; 355: Baglin T, Palmer CR, Luddington R, Baglin C. Unprovoked recurrent venous thrombosis: prediction by D-dimer and clinical risk factors. J Thromb Haemost 2008; 6: Cini M, Legnani C, Cavallaroni K, Bettini F, Palareti G. A new rapid bedside assay for D-dimer measurement (Simplify D-dimer) in the diagnostic work-up for deep vein thrombosis. J Thromb Haemost 2003; 1: Legnani C, Palareti G, Cosmi B, Cini M, Tosetto A, Tripodi A. Different cut-off values of quantitative D-dimer methods to predict the risk of venous thromboembolism recurrence: a post-hoc analysis of the PROLONG study. Haematologica 2008; 93: Verhovsek M, Douketis JD, Yi Q, Shrivastava S, Tait RC, Baglin T, Poli D, Lim W. Systematic review: D-dimer to predict recurrent disease after stopping anticoagulant therapy for unprovoked venous thromboembolism. Ann Intern Med 2008; 149: W PoliD,AntonucciE,CiutiG,Abbate R, Prisco D. Combination of D- dimer, F1 + 2 and residual vein obstruction as predictors of VTE recurrence in patients with first VTE episode after OAT withdrawal. J Thromb Haemost 2008; 6: TaitRC,LoweG,McCollM,McMahonA,RobertsonL,KingL, Clarke S, Conkie J, Roditi G, Walker I. Predicting risk of recurrent venous thrombosis using a 5-point scoring system including fibrin D- dimer. J Thromb Haemost 2007; 5(Supplement 2): O-M-060.