- Email: [email protected]
Markers of plasma coagulation and fibrinolysis after acute deep venous thrombosis
Markers of plasma coagulation and fibrinolysis after acute deep venous thrombosis Mark H. Meissner, MD,a Brenda K. Zierler, RN, PhD,a Robert O. Bergelin, MS,a Wayne C. Chandler, MD,b Richard A. Manzo, BS, RVT,a and D. Eugene Strandness, Jr, MD,a Seattle, Wash Objective: Plasma markers of coagulation and fibrinolysis have proved sensitive in the initial diagnosis of acute deep venous thrombosis (DVT). The purpose of this study was to examine the evolution and utility of measuring D-dimer and prothrombin fragment 1+2 (F 1+2) levels after an acute DVT. Methods: Subjects with DVT confirmed by ultrasonography had quantitative plasma Ddimer and F 1+2 levels determined before anticoagulation. Ultrasound scan and coagulation studies were repeated at 3, 7, and 14 days; 1 month; and every 3 months for 1 year. Results: Sixty-one patients with a median initial thrombus score of 3 (interquartile range, 2-7) were followed up for 266 days (interquartile range, 91.5-364 days). Initial D-dimer levels were elevated in 92.7% of patients and were associated with thrombus extent (P = .003), whereas F 1+2 levels were increased in 94.5% of patients and were lower in patients with isolated calf vein thrombosis (P = .001). Initial D-dimer (P = .002) and F 1+2 levels (P = .009) were significantly higher in the 26 (43%) patients with recurrent thrombosis during follow-up. Initial D-dimer levels of 2000 ng/mL or greater were predictive of recurrent events after both proximal and isolated calf vein thrombosis. Although interval increases in these markers had little value in detecting recurrent thrombotic events, D-dimer levels of 1000 ng/mL or greater and 500 ng/mL or greater had respective sensitivities of 89.3% and 100% in detecting early and late recurrences. Corresponding specificities were 35.6% and 53.9%. Conclusions: Initial D-dimer levels are determined by total thrombus load and remain elevated long after an acute DVT. F 1+2 levels are less sensitive to thrombus score and return to baseline more quickly. Initial levels of these markers may have some utility in predicting the risk of ultrasound scan–documented recurrences, whereas increased D-dimer levels are a sensitive but nonspecific marker of these events. (J Vasc Surg 2000;32:870-80.)
The development of sensitive assays for stable markers of activated coagulation aznd fibrinolysis has provided new insight into the pathophysiology of acute deep venous thrombosis (DVT) and may have some role in its diagnosis. Although the activated enzymes of coagulation and fibrinolysis From the Departments of Surgerya and Laboratory Medicine,b University of Washington School of Medicine. Competition of interest: nil. Supported by National Institutes of Health grant No. 619578. Presented at the Twelfth Annual Meeting of the American Venous Forum, Phoenix, Ariz, Feb 4, 2000. Reprint requests: Mark H. Meissner, MD, Department of Surgery, Box 359796, Harborview Medical Center, 325 9th Ave, Seattle, WA 98195 (e-mail: [email protected]). Copyright © 2000 by The Society for Vascular Surgery and The American Association for Vascular Surgery, a Chapter of the International Society for Cardiovascular Surgery. 0741-5214/2000/$12.00 + 0 24/6/110359 doi:10.1067/mva.2000.110359
870
are evanescent, measurable byproducts of zymogen activation include D-dimer fibrin degradation products formed by the action of plasmin on cross-linked fibrin; prothrombin fragment 1+2 (F 1+2) generated by factor Xa in the cleavage of prothrombin to thrombin; fibrinopeptide A formed in the thrombin-mediated conversion of fibrinogen to fibrin; and thrombin-antithrombin complex formed by the combination of thrombin with its primary inhibitor. 1,2 Among these, F 1+2, fibrinopeptide A, and thrombin-antithrombin complex similarly reflect the degree of thrombin activation, whereas D-dimer reflects fibrinolytic activity and the presence of intravascular thrombus. The initial changes in these markers after thrombosis and treatment have been described,3-5 and plasma D-dimer levels have been incorporated into some diagnostic algorithms.6,7 However, the long-
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
871
Fig 1. Correlation between extent of thrombus (SVS/NA-ISCVS thrombus score) at presentation and initial plasma D-dimer level (R = 0.39, P = .003).
Table I. Thrombotic risk factors term changes occurring in these markers have not been examined in detail. The evolution of these markers over time is critical in their application to recurrent venous thrombosis, which remains a fundamental challenge in the management of DVT. Although the adequacy of initial anticoagulation appears to be critical in preventing recurrent thrombosis,8 the ability to identify subgroups at particular risk for such events is limited. This is especially important for thrombosis isolated to the calf veins because treatment is primarily based on the risk of proximal propagation. Because the development of recurrent DVT is presumably determined by the relative balance between coagulation and fibrinolysis, these markers may have some utility in identifying patients at highest risk. Changes in these markers may also be relevant to the diagnosis of recurrent DVT. Although venous duplex ultrasonography canning is now the study of choice for the initial diagnosis of DVT, its limitations in the diagnosis of recurrent DVT are well recognized. Ultrasound scan interpretation in this setting may be limited by variable degrees of residual occlusion, partial recanalization, intimal thickening, and collateral formation. Results of compression ultrasound studies may remain abnormal in 27% to 70% of patients after 1 year.9-12 Favorable reports regarding the use of D-dimer in the initial diagnosis suggest that it may also have a role in the diagnosis of recurrent DVT. The purpose of this study was therefore to evaluate the initial determinants of quantitative D-dimer and F 1+2 levels, to document changes in these lev-
Risk factor Age > 40 y Malignancy > 3 day’s immobilization Estrogen therapy Surgery Cardiac disease Prolonged (> 6 h) travel Limb trauma Family history Previous DVT Pregnant Known prothrombotic state Obese
No. (%) 44 23 18 12 11 6 6 6 5 3 2 1 0
(72.1) (37.7) (29.5) (19.7) (18.0) (9.8) (9.8) (9.8) (8.2) (4.9) (3.3) (1.6) (0)
els after initial presentation, and to evaluate their utility in the diagnosis of recurrent DVT. METHODS All patients presenting to the University of Washington Medical Center with an acute DVT confirmed by duplex ultrasonography were asked to participate in a long-term study of the natural history of DVT. Patients with a limited life expectancy receiving prophylactic or therapeutic doses of anticoagulants before enrollment or having a residence or medical condition preventing their return were excluded from enrollment. All treatment decisions were made by the patient’s primary physician. The Human Subjects Committee at the University of Washington approved the study protocol, and informed consent was obtained from all subjects. Immediately after enrollment and before administration of anticoagulants, blood was drawn for
872 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 2. Correlation between extent of thrombus (SVS/NA-ISCVS thrombus score) at presentation and initial plasma F 1+2 level (R = 0.28, P = .08).
determination of plasma D-dimer cross-linked fibrin degradation products and prothrombin fragment 1+2 levels. A complete medical history was obtained, and thrombotic risk scores were calculated according to the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery (SVS/ISCVS) reporting standards in venous disease.13 Patients were then requested to return for follow-up visits at intervals of 3, 7, and 14 days and 1, 3, 6, 9, and 12 months after initial presentation. At each visit, patients were questioned regarding limb symptoms, their lower extremities were examined, and duplex ultrasonography of the lower extremity veins was performed. Blood was also drawn for determination of plasma D-dimer and F 1+2 levels. Duplex scanning of both lower extremities was performed in the 15-degree reverse Trendelenburg position with an Ultramark 9 instrument (Advanced Technology Laboratories, Bothell, Wash). A phased array 3-2 MHz transducer was used for the inferior vena cava and iliac segments, and a linear array 7-4 MHz transducer was used for the veins below the inguinal ligament. All superficial and deep venous segments from the inferior vena cava to the calf veins were assessed for the presence of thrombus at each visit. Individual venous segments were defined as patent, partially occluded, or completely occluded on the basis of established B mode and Doppler criteria,14 and the extent of thrombosis was determined on a 24-point scale according to the reporting standards in venous disease.13 Recurrent thrombotic events were defined by ultrasound criteria to include
propagation of thrombus to initially uninvolved venous segments, progression of nonocclusive thrombus within a segment to complete occlusion, rethrombosis of a partially or completely recanalized segment, or new thrombus within an initially uninvolved contralateral limb. The events were further defined as occurring early (≤ 30 days) or late (> 30 days) after initial presentation. Blood for all laboratory studies was drawn with a two-tube technique, discarding the first 4 to 5 mL. Samples were anticoagulated by the addition of 4.5 mL of whole blood to 0.5 mL of 0.105 mol/L sodium citrate and centrifuged at 2000g for 15 minutes at room temperature. Citrate plasma was removed and frozen at –70°C before analysis. Quantitative D-dimer levels were determined with an enzyme-linked immunosorbent assay (ELISA), as previously described.15 For Ddimer, the normal range (< 400 ng/mL) was defined as identical to that used for the exclusion of DVT in the clinical laboratory with a rapid D-dimer assay (sensitivity, 94.6%; specificity, 35%).16 F 1+2 (normal, 0.4-1.1 nmol/L) was similarly assessed with an ELISA method from Behring Inc.17 Because most variables were not uniformly distributed, all data are reported as the median and interquartile range (IQR, 25th to 75th percentile). Bivariate analysis was performed with linear regression or the Mann-Whitney test for continuous variables and the χ2 test for categoric variables. Determinants significant at a P value of .10 or less were entered into multivariate stepwise regression and logistic regression models. The sensitivity and specificity of various D-dimer and F 1+2 levels in
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
873
Table II. Follow-up D-dimer and F 1+2 levels Follow-up
D-dimer
interval
No.
Day 0 Day 3 Day 7 Day 14 Day 30 3 mo 6 mo 9 mo 1y
55 50 50 46 46 43 33 27 20
F 1+2
Median (IQR) 2320 2033 1719 1034 660 440 407 534 646
(1364-5733) (1248-3252) (1008-3444) (475-2749) (272-1362) (208-999) (209-861) (356-1184) (383-1542)
No. ≥ 400*
No.
51 47 46 37 31 26 18 18 14
55 51 50 48 46 44 34 27 20
(92.7%) (94%) (92%) (80.4%) (67.4%) (60.5%) (54.5%) (66.7%) (70%)
Median (IQR) 2.2 1.5 1.2 0.5 0.7 0.6 0.6 1.0 0.9
(1.7-3.5) (1.0-2.2) (0.6-2.0) (0.3-1.7) (0.4-1.6) (0.3-1.2) (0.3-1.2) (0.5-1.3) (0.4-1.3)
No. ≥ 1.1* 52 34 27 21 20 14 9 11 7
(94.5%) (66.7%) (54%) (43.8%) (43.5%) (31.8%) (26.5%) (40.7%) (35%)
*Number of patients with values above the normal range.
Table III. D-dimer and F 1+2 levels at 6 months’ follow-up D-Dimer < 400 ng/mL (n = 15)* 6-mo thrombus score Initial thrombus score Age (y) Risk score Risk factors Irreversible factors†
2.0 2.0 48 2.0 2.0 1.0
(0.0-3.0) (2.0-4.0) (29-64) (1.0-3.0) (1.0-2.0) (1.0-1.0)
≥ 400 ng/mL (n = 18)* 1.0 2.0 60 4.5 3.0 2.0
(0.75-2.25) (1.0-7.0) (51-73) (2.0-6.0) (1.75-3.25) (1.0-2.0)
F 1+2 P value .57 .85 .05 .01 .009 .04
< 1.1 nmol/L (n = 24)* 1.0 2.5 51 2.0 2.0 1.0
(1.0-3.0) (2.0-6.75) (33-63) (2.0-4.5) (1.0-3.0) (0.25-2.0)
≥ 1.1 nmol/L (n = 9)* 1.0 2.0 60 4.0 3.0 2.0
(0-3.5) (1.0-4.5) (56-75) (3.0-5.5) (2.5-3.5) (1.0-2.0)
P value .68 .33 .03 .09 .03 .07
*Median (IQR). †Number of irreversible risk factors, including age over 40 years, malignancy, cardiac disease, previous history of DVT, family history of thrombosis, known prothrombotic state, and obesity.
predicting recurrent thrombotic events were evaluated with receiver-operator characteristic (ROC) analysis with Rockit statistical software (Department of Radiology, University of Chicago, Chicago, Ill). Sensitivity for the diagnosis of recurrent thrombosis was defined as true-positive results in patients with recurrence divided by total patients with recurrence, and specificity was defined as true-negative results in patients without recurrence divided by total patients without recurrence. Statistical significance was defined at a P value of .05 or less. RESULTS Between October 1997 and September 1999, 61 patients with an acute DVT were enrolled and followed up for a median of 266 days (IQR, 91.5-364 days). Subjects included 30 (49.2%) women and 31 (50.8%) men with a median age of 54.6 years (IQR, 39.5-64.3 years). The median SVS/NA-ISCVS risk score among these patients was 4 (IQR, 2-6), with a median of 2.0 (IQR, 1.5-3.0) risk factors per patient (Table I). Thrombosis was bilateral in nine (14.8%) patients and unilateral in the right and left limbs in 22 (36.1%) and 30 (49.2%) patients, respectively.
These included 34 (55.7%) proximal thrombi and 27 (44.3%) isolated calf vein thrombi with a median thrombus score of 3.0 (IQR, 2.0-7.0). Of the 57 patients completing early follow-up, 42 (73.7%) were initially treated with intravenous unfractionated or subcutaneous low-molecular-weight heparin, 2 (3.5%) were treated with warfarin alone, and 13 (22.8%) were not treated with anticoagulants. Among those treated with unfractionated heparin, the median duration of treatment was 5.0 days (IQR, 4.0-9.5 days), and a median of 80% (IQR, 50100) of partial thromboplastin times was within the therapeutic range. Fifty-five patients had blood samples obtained at the time of presentation. Among these patients, median D-dimer and F 1+2 levels were 2320 ng/mL (IQR, 1364-5733 ng/mL) and 2.2 nmol/L (IQR, 1.7-3.5 nmol/L), respectively. D-dimer and F 1+2 levels were elevated in 51 (92.7%) and 52 (94.5%) patients, respectively. Initial D-dimer levels were significantly related to total thrombus score (P = .003), whereas F 1+2 levels were only marginally related to the extent of thrombus (P = .08, Figs 1 and 2). In comparison with those with proximal
874 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 3. Median D-dimer levels during follow-up stratified according to the presence of ultrasound scan–defined recurrent thrombosis. Dashed line indicates diagnostic limit for acute DVT (≥ 400 ng/mL). Plasma levels at day 0 are significantly different (P = .002) between those with and without subsequent recurrent events. Number of patients evaluated at each follow-up interval and the corresponding IQR are shown.
thrombosis, patients with isolated calf vein thrombosis had significantly lower D-dimer (1967 ng/mL [IQR, 1128-3213 ng/mL] vs 3356 ng/mL [IQR, 2016-8469 ng/mL]; P = .02) and F 1+2 (2.0 nmol/L [IQR, 1.5-2.4 nmol/L] vs 2.7 nmol/L [IQR, 2.0-4.4 nmol/L]; P = .004) levels. A history of prolonged travel was the only thrombotic risk factor related to initial D-dimer values, with median levels being higher among those without a travel history (2745 ng/mL [IQR, 1562.5-6679.5 ng/mL] vs 1462.5 ng/mL [IQR, 271.8-2240.5 ng/mL]; P = .03). A history of cardiac disease (3.6 nmol/L [IQR, 2.23-4.68 nmol/L] vs 2.1 nmol/L [IQR, 1.6-3.1 nmol/L]; P = .05) was associated with higher initial F 1+2 levels, whereas recent limb trauma (1.6 nmol/L [IQR, 0.75-1.85 nmol/L] vs 2.4 nmol/L [IQR, 1.8-3.73 nmol/L]; P = .001) was associated with lower levels. However, in a multivariate analysis including extent of thrombosis, proximal versus distal thrombosis, and risk factors, thrombus score was the only significant predictor of initial D-dimer levels (P = .003), whereas the presence of proximal thrombosis was the only significant predictor of F 1+2 levels (P = .001). Both D-dimer and F 1+2 levels decreased rapid-
ly after initial presentation. D-dimer levels at day 3 were independent of initial treatment, with median levels among those treated with unfractionated or low-molecular-weight heparin (2037 ng/mL [IQR, 1041-3376 ng/mL]) being similar to levels of those who were untreated or treated with warfarin alone (2028 ng/mL [IQR, 1693-2578 ng/mL]; P = .73). However, median F 1+2 levels at day 3 were significantly lower among those treated with unfractionated or low-molecular-weight heparin (1.1 nmol/L [IQR, 0.8-2.0 nmol/L] vs 2.0 nmol/L [IQR, 1.83.0 nmol/L]; P = .001). Median F 1+2 levels normalized by 14 days, whereas median D-dimer levels approached normal levels by 3 months (Table II). However, after 1 year of follow-up, D-dimer and F 1+2 levels remained elevated in 70% and 35% of patients, respectively. In comparison with those with normal D-dimer and F 1+2 levels, subjects with persistently elevated levels at 6 months were significantly older and had higher risk scores, greater numbers of thrombotic risk factors, and more irreversible risk factors. In contrast, extent of thrombosis both at presentation and at the 6-month follow-up interval was not significantly different from those in patients with nor-
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
875
Table IV. Interval D-dimer increases Recurrent thrombosis
Interval D-dimer increase
Yes No Total
mal D-dimer and F 1+2 levels (Table III). Similar trends were present for those with elevated levels at 1 year, although because of small patient numbers, differences in risk score, number of risk factors, and number of irreversible risk factors did not reach statistical significance. Thirty-seven recurrent thrombotic events were observed in 26 (43%) patients. Although clustered early after presentation, recurrent thrombosis was observed throughout the 1-year follow-up interval. Fifteen (40.5%) events were observed within the first 3 days of follow-up, 14 (37.8%) between days 3 and 30, and 4 (10.8%) in each of the intervals between 1 and 6 months and 6 and 12 months. Neither thrombus score (P = .15) nor overall risk score (P = .35) was significantly related to recurrent thrombotic events. However, initial D-dimer (P = .002) and F 1+2 (P = .009) levels were significantly higher among patients with recurrent thrombosis (Figs 3 and 4). In a logistic regression model, initial F 1+2 levels (P = .04) were more highly related to recurrent events than initial D-dimer levels (P = .33). However, initial D-dimer levels performed slightly better in the receiver-operator characteristic analysis. In predicting recurrent thrombosis, an initial Ddimer level of 2000 ng/mL or greater provided a sensitivity of 87% and a specificity of 56.3%, whereas F 1+2 levels of 2.0 nmol/L or greater had a sensitivity of 82.6% and specificity of 43.8%. Similarly, for thrombosis confined to the calf, a D-dimer level of 2000 ng/mL or greater had a sensitivity and specificity of 88.9% and 76.5% in predicting recurrent events, whereas F 1+2 levels of 2.0 nmol/L or greater had a sensitivity of 66.7% and specificity of 52.9%. Among patients with isolated calf vein thrombosis, recurrent thrombosis developed in 67% of those with an initial D-dimer level of 2000 ng/mL or greater in comparison with 7% of those with levels of less than 2000 ng/mL. For F 1+2, recurrent thrombosis occurred in 43% of patients with isolated calf vein thrombosis and an F 1+2 level of 2.0 nmol/L or greater in comparison with 25% of those with levels of less than 2.0 nmol/L. Persistent D-dimer and F 1+2 elevations limited
Yes
No
Total
10 18 28
102 151 253
112 169 281
Table V. Interval F 1+2 increases Recurrent thrombosis
Interval F 1+2 increase Yes No Total
Yes
No
Total
8 20 28
79 181 260
87 201 288
the utility of the threshold values used for the initial diagnosis of DVT. F 1+2 levels of 1.1 nmol/L or greater had an overall sensitivity and specificity of 62.9% and 49.3%, respectively, for the diagnosis of recurrent thrombotic events. If only early (≤ 30 days) events were considered, sensitivity improved to 71.4%, and specificity decreased to 39.6%. For the diagnosis of late recurrence (> 30 days), the sensitivity was only 28.6%, with a specificity of 67.5%. For D-dimer levels, different critical values could be selected on the basis of the timing of recurrent thrombosis. For early recurrence, D-dimer levels of 1000 ng/mL or greater provided an optimal sensitivity and specificity of 89.3% and 35.6%, respectively. Positive and negative predictive values in this population were 15.1% and 96.3%, respectively. For late recurrence, D-dimer levels of 500 ng/mL or greater provided a sensitivity of 100% and specificity of 53.9%, with positive and negative predictive values of 11.7% and 88.3%, respectively. In comparison with the value at the previous visit, the sensitivity and specificity of an increase in D-dimer levels were 35.7% and 59.7%, respectively, with a positive and negative predictive value of 8.9% and 89.3% in this population (Table IV). For interval increases in prothrombin F 1+2 levels, sensitivity and specificity were 28.6% and 69.6%, respectively, with a positive and negative predictive value of 9.2% and 90.0% (Table V). DISCUSSION The development of specific assays for stable markers of activated coagulation and intravascular thrombus formation has provided new insight into the pathophysiology of acute DVT. In contrast to the measurement of zymogen levels and standard clotting tests, these assays are able to detect low-level activation
876 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 4. Median F 1+2 levels during follow-up stratified according to the presence of ultrasound scan–defined recurrent thrombosis. Dashed line indicates upper limit of normal range (1.1 nmol/L). Plasma levels at day 0 are significantly different (P = .009) between those with and without subsequent recurrent events. Number of patients evaluated at each follow-up interval and the corresponding IQR are shown.
of the coagulation system and localized thrombotic events.1,2 Several thrombotic risk factors have been associated with increased levels of thrombin activation markers, confirming the importance of activated coagulation in thrombogenesis. Increased levels of these markers have been described in association with advanced age,2,18 surgery,19 trauma,20 oral contraceptives,21,22 and malignancy.23 Elevated levels of these markers have also been shown to be useful in excluding a diagnosis of acute DVT. Although of variable sensitivity, the uniformly low specificity of these tests requires confirmatory testing of patients with positive results. D-dimer levels appear to be more sensitive than the markers of thrombin activation,1,3,4,16,24 an observation that may reflect fluctuation in thrombin activity and the longer half-life of D-dimer. In 11 published series of symptomatic acute DVT, the average weighted sensitivity and specificity of D-dimer determined with the ELISA technique have been reported to be 96.8% and 35.2%, respectively.25 Lower sensitivities of 47% to 66% have been reported for F 1+2 levels.3,24 Although not a study of diagnostic accuracy, the results of this study are consistent with these reports. D-dimer and F 1+2 levels were elevated at the time of
presentation in 92.7% and 94.5% of patients, respectively. However, this study further demonstrates that initial D-dimer levels are most significantly related to the extent of thrombosis, whereas F 1+2 levels are more highly associated with the presence of proximal thrombus. The relationship between D-dimer level and thrombus extent most likely reflects the quantity of thrombus susceptible to the action of plasmin. The magnitude of the correlation between plasma Ddimer level and thrombus extent determined by ultrasound (R = 0.39) is similar to that reported for venographically determined thrombus extent (R = 0.340.36).3,26 Although statistically significant, this relatively weak correlation suggests that fibrinolytic differences between individuals may also be important. The relationship between F 1+2 level and thrombus location is more interesting because it implies that thrombus confined to the calf veins is associated with less extensive thrombin activation. Although it is clear that most thrombi begin in the calf veins,27,28 this observation further suggests that isolated calf vein thrombi may differ pathophysiologically from proximal thrombi. That is, they are not simply early thrombi that have yet to propagate, but rather reflect a more limited prothrombotic state.
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Although this was a small study, thrombotic risk factors were not independently associated with initial D-dimer or F 1+2 levels. Patients having risk factors associated with underlying elevations of thrombin activation products might have been expected to have higher levels of these markers after the development of DVT. The finding that these risk factors are not associated with increased F 1+2 levels after thrombosis may reflect either the strong amplification of the coagulation cascade associated with clinical thrombosis29 or the existence of a thrombotic threshold at which acute DVT occurs independent of baseline prothrombin activation.30 The latter would imply that a similar level of thrombin activity is required for thrombosis to occur, regardless of the underlying risk factors. A larger number of patients with levels measured before thrombosis would be required to substantiate this hypothesis. Several studies have examined changes in plasma coagulation and fibrinolytic markers during the first 10 days of treatment,3-5 but the long-term changes after an episode of DVT have not been well characterized. Although some have found markers of thrombin activation to normalize early after effective treatment is instituted,4 most have found these markers to rapidly decrease, but not fully normalize, during the first 10 days. Consistent with these observations, median D-dimer and F 1+2 levels in this study rapidly declined after initial presentation. However, despite this overall trend, these markers remained abnormal for prolonged periods in a substantial number of patients. Although median F 1+2 values returned to normal with the first 14 days, a prothrombotic state, implied by elevated F 1+2 levels, persisted in 27% to 44% of the patients evaluated between 14 days and 1 year after thrombosis. This was associated with a 43% incidence of ultrasounddocumented recurrent events. Consistent with a persistent thrombus load, median D-dimer levels approached normal by 3 months but remained elevated throughout the first year in most patients. The initial decrease in both markers was likely caused by a rapid reduction in thrombin activity and the generation of new lysable fibrin, whereas the slower subsequent decline in D-dimer levels reflects the decreased susceptibility of older thrombus to lysis.4 In contrast to the initial levels, persistent elevations of both markers were related to age and thrombotic risk, which is consistent with ongoing hypercoagulability and perhaps fibrin generation. These observations have important implications for the diagnosis of recurrent thrombosis. Fluctuations in the levels of both markers make interval
Meissner et al
877
increases in their levels of little diagnostic significance. This contrasts with other small series suggesting an increase in D-dimer levels of greater than 3 µg/mL within the first 10 days to be a specific but insensitive indicator of recurrent thrombosis.3 Furthermore, persistent elevations will a priori reduce the specificity and positive predictive value of these markers with the criteria used for the initial diagnosis of DVT. Despite these limitations, these tests may still have some utility in the diagnosis of recurrent DVT if sensitivity and negative predictive value are sufficiently high. Although most patients with elevated levels will not have recurrent thrombosis (low positive predictive value), low levels can exclude the diagnosis. This is particularly true for D-dimer, where values of 1000 ng/mL or greater and 500 ng/mL or greater had sensitivities of 89% and 100%, respectively, in detecting early and late recurrences. As in the initial diagnosis of DVT, the low specificities and positive predictive value of these measurements limit their utility to the exclusion, rather than the confirmation, of recurrent disease. However, given the potential limitations of other diagnostic modalities, including venous duplex ultrasonography and venography, Ddimer measurements may have some clinical value in this setting. A D-dimer level of less than 500 ng/mL in a patient with pain, swelling, and a duplex scan demonstrating thrombus of indeterminate age 9 months after presentation could potentially allow the patient to be spared the risk and inconvenience of reinstituting anticoagulation. Perhaps of greater value is the relationship between initial levels of these markers and subsequent recurrent events. Although specificity was again limited, D-dimer levels of 2000 ng/mL or greater and F 1+2 levels of 2.0 nmol/L or greater had sensitivities of 87% and 83%, respectively, in predicting recurrent thrombotic events. This may have important implications, particularly for the management of isolated calf vein thrombosis. Although management remains controversial, anticoagulant treatment of these thrombi is often recommended on the basis of the observation that approximately 20% will propagate to a more proximal level.31,32 In this regard an initial D-dimer level of 2000 ng/mL or greater was highly associated with the aggregate outcome of recurrent thrombosis, which included proximal propagation. Two thirds of patients with isolated calf vein thrombosis and a D-dimer level of 2000 ng/mL or greater had recurrent thrombosis in comparison with less than 10% of those with lower levels. Although larger confirmatory studies are obviously required, patients with initial D-dimer lev-
JOURNAL OF VASCULAR SURGERY November 2000
878 Meissner et al
els of 2000 ng/mL or greater are perhaps most appropriately treated with standard anticoagulation, whereas serial noninvasive follow-up may be an option in those with levels below 2000 ng/mL. Other uses of these markers remain speculative, but they may have some role in more precisely defining the appropriate duration of anticoagulation in individual patients. The relationship between persistently elevated D-dimer and F 1+2 levels and thrombotic risk factors tends to support current recommendations that patients with irreversible thrombotic risk factors warrant a longer duration of anticoagulation. Such measurements could also conceivably be useful in identifying those patients with idiopathic DVT at highest risk for recurrence and requiring extended anticoagulation. Validation of such speculation would obviously require larger series examining the relationship between these markers at the completion of a standard course of anticoagulation and the incidence of late recurrent events. It is important to recognize the shortcomings of this study, most notably the small number of patients. The number of patients in some subgroups, such as those having specific risk factors with late recurrence, was quite small. It is certainly possible that important relationships between risk factors, D-dimer and F 1+2 levels, and recurrent thrombosis were overlooked. Furthermore, although consecutive patients meeting entry criteria were enrolled, and the ultrasound scans and laboratory tests were independently interpreted, this was an exploratory study not designed primarily to document the diagnostic accuracy of these markers. Neither was it designed as a management trial to evaluate the use of these markers in clinical decision making. It is also important to note that quantitative ELISA assays were used in this study and that the results may not be applicable to semiquantitative latex and hemagglutination tests. Finally, a standardized treatment regimen was not used in this study. Although D-dimer levels were independent of initial treatment, early F 1+2 levels were significantly related to anticoagulation with unfractionated or low-molecular-weight heparin. Although unlikely to change the D-dimer results, anticoagulation of the 26.3% of patients who were not initially treated could alter the time course of F 1+2 levels, potentially making this marker more suitable for the diagnosis of recurrent thrombotic events. Despite these limitations, this study further defines the evolution of stable markers of activated coagulation and fibrinolysis after an episode of acute DVT. Initial F 1+2 levels are most significantly associated with the presence of proximal thrombus, where-
as D-dimer levels reflect the extent of thrombus and perhaps individual differences in the fibrinolytic system. Initial levels of both markers are predictive of recurrent thrombotic events. This observation may have particular relevance to the treatment of isolated calf vein thrombosis and warrants further investigation. Although levels of both markers rapidly decline after presentation, they remain elevated in a substantial number of patients for at least 1 year. This limits the specificity and positive predictive value of these markers in recurrent DVT, although they appear to have some utility in excluding the diagnosis. REFERENCES 1. Boisclair MD, Ireland H, Lane DA. Assessment of hypercoagulable states by measurement of activation fragments and peptides. Blood Rev 1990;4:25-40. 2. Bauer KA, Rosenberg RD. The pathophysiology of the prethrombotic state in humans: insights gained from studies using markers of hemostatic system activation. Blood 1987;70:343-50. 3. The DVTENOX Study Group. Markers of hemostatic system activation in acute deep venous thrombosis—evolution during the first days of heparin treatment. Thromb Haemost 1993;70:909-14. 4. Speiser W, Mallek R, Koppensteiner R, Stumpflen A, Kapiotis S, Minar E, et al. D-dimer and TAT measurement in patients with deep venous thrombosis: utility in diagnosis and judgement of anticoagulant treatment effectiveness. Thromb Haemost 1990;64:196-201. 5. Bridey F, Phlippoteau C, Simmoneau G, Musset D, Dreyfus M. Is D-Dimer measurement a marker of recurrence of thromboembolic disease? [Abstract]. Thromb Haemost 1991;65:981. 6. Bernardi E, Prandoni P, Lensing AW, Agnelli G, Guazzaloca G, Scannapieco G, et al. D-dimer testing as an adjunct to ultrasonography in patients with clinically suspected deep vein thrombosis: prospective cohort study. The Multicentre Italian D-dimer Ultrasound Study Investigators Group. BMJ 1998;317:1037-40. 7. Michiels JJ. Rational diagnosis of deep vein thrombosis (RADIA DVT) in symptomatic outpatients with suspected DVT: simplification and improvement of decision rule analysis for the exclusion and diagnosis of DVT by the combined use of a simple clinical model, a rapid sensitive D-dimer test and compression ultrasonography (CUS). Semin Thromb Hemost 1998;24:401-7. 8. Caps MT, Meissner MH, Tullis MJ, Polissar NL, Manzo RA, Zierler BK, et al. Venous thrombus stability during the acute phase of therapy. Vasc Med 1999;4:9-14. 9. Murphy TP, Cronan JJ. Evolution of deep venous thrombosis: a prospective evaluation with US. Radiology 1990;177:543-8. 10. Mantoni M. Deep venous thrombosis: longitudinal study with duplex US. Radiology 1991;179:271-3. 11. Prandoni P, Cogo A, Bernardi E, Villalta S, Polistena P, Simioni P, et al. A simple ultrasound approach for detection of recurrent proximal vein thrombosis. Circulation 1993;88:1730-5. 12. Baxter GM, Duffy P, MacKechnie S. Colour Doppler ultrasound of the post-phlebitic limb: sounding a cautionary note. Clin Radiol 1991;43:301-4.
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
13. Porter J, Moneta G, International Consensus Committee on Chronic Venous Disease. Reporting standards in venous disease: an update. J Vasc Surg 1995;21:635-45. 14. Meissner MH. Venous duplex scanning. In: Rutherford RB, editor. Vascular surgery. 5th ed. Philadelphia: W.B. Saunders Company; 2000. p. 214-29. 15. van Beek EJR, van den Ende B, Berckmans RJ, ven der Heide YT, Brandjes DPM, Sturk A, et al. A comparative analysis of D-Dimer assays in patients with clinically suspected pulmonary embolism. Thromb Haemost 1993;70:408-13. 16. Escoffre-Barbe M, Oger M, Leroyer C, Grimaux M, Le Moigne E, Nonent M, et al. Evaluation of a new rapid DDimer assay for clinically suspected deep venous thrombosis (Liatest D-Dimer). Am J Clin Pathol 1998;109:748-53. 17. Pelzer H, Schwarz A, Stuber W. Determination of human prothrombin activation fragment 1+2 in plasma with an antibody against a synthetic peptide. Thromb Haemost 1991;65:153-9. 18. Cadroy Y, Pierrejean D, Fontan B, Sie P, Boneu B. Influence of aging on the activity of the hemostatic system: prothrombin fragment 1+2, thrombin-antithrombin III complexes and D-dimers in 80 healthy subjects with age ranging from 20 to 94 years. Nouv Rev Fr Hematol 1992;34:43-6. 19. Kambayashi J, Sakon M, Yokota M, Shiba E, Kawasaki T, Mori T. Activation of coagulation and fibrinolysis during surgery, analyzed by molecular markers. Thromb Res 1990;60:157-67. 20. Gando S, Tedo I, Kubota M. Posttrauma coagulation and fibrinolysis. Crit Care Med 1992;20:594-600. 21. von Kaulla E, Droegemueller W, Aoki N, von Kaulla KN. Antithrombin III depression and thrombin generation acceleration in women taking oral contraceptives. Am J Obstet Gynecol 1971;109:868-73. 22. Quehenberger P, Loner U, Kapiotis S, Handler S, Schneider B, Huber J, et al. Increased levels of activated factor VII and decreased plasma protein S activity and circulating thrombomodulin during use of oral contraceptives. Thromb Haemost 1996;76:729-34.
DISCUSSION Dr Thomas W. Wakefield (Ann Arbor, Mich). Dr Meissner and coworkers continue their excellent studies on venous thrombosis, and this is really a nicely done study and the manuscript is very scholarly. The current work suggests that measurements of Ddimer were significantly elevated in patients who suffered DVT, with thrombus score predicting initial D-dimer levels, while F 1+2 levels were also elevated with the presence of proximal thrombus the only significant predictor of F 1+2 levels. One year after thrombosis D-dimer and F 1+2 levels remained elevated in 70% and 35% of patients, however, which necessitated a change in the values noted important for defining recurrent DVT. Importantly, initial levels of D-dimer greater than 2000 ng/mL and F 1+2 greater than 2 nmol/L carried greater than 80% sensitivity for the diagnosis of recurrent DVT. This is key, as we know from other data in the literature that one of the most important factors in the development of chronic venous insufficiency is the presence of ipsilateral recurrent DVT. I am going to ask two questions. First, can you speculate on the cause of the persistent elevations in D-dimer and F 1+2 levels at 1 year after DVT formation? I can certainly
Meissner et al
879
23. Falanga A, Ofosu FA, Oldani E, Dewar L, Lui L, Barbui T. The hypercoagulable state in cancer patients: evidence for impaired thrombin inhibitions. Blood Coagul Fibrinolysis 1994;5 Suppl 1:S19-23. 24. Boneu B, Pelzer H, Boccalon H. D-dimers, thrombin antithrombin III complexes, and prothrombin fragments 1+2: diagnostic value in clinically suspected deep vein thrombosis. Thromb Haemost 1991;65:28-32. 25. Bounameaux H, de Moerloose P, Perrier A, Reber G. Plasma measurement of D-dimer as a diagnostic aid in suspected venous thromboembolism: an overview. Thromb Haemost 1994;71:1-6. 26. Mirshani M, Soria C, Soria J, Mirshani M, Faivre R, Kieffer Y, et al. Changes in plasma fibrin degradation products as a marker of thrombus evolution in patients with deep vein thrombosis. Thromb Res 1988;51:295-302. 27. Nicolaides AN, Kakkar VV, Field ES, Renney JT. The origin of deep vein thrombosis: a venographic study. Br J Radiol 1971;44:653-63. 28. Kakkar VV, Howe CT, Nicolaides AN, Clarke MB. Deep vein thrombosis of the leg. Is there a “high risk” group. Am J Surg 1970;120:527-30. 29. Amiral J, Fareed J. Thromboembolic diseases: biochemical mechanisms and new possibilities of biological diagnosis. Semin Thromb Hemost 1996;22:41-8. 30. Rosendaal FR. Venous thrombosis: a multicausal disease. Lancet 1999;353:1167-73. 31. Kakkar VV, Flanc C, Howe CT, Clarke MB. Natural history of post-operative deep-vein thrombosis. Lancet 1969;2:230-2. 32. Lagerstedt CI, Olsson C, Fagher BO, Oqvist BW, Albrechtsson U. Need for long-term anticoagulant treatment in symptomatic calf vein thrombosis. Lancet 1985;2:515-8.
Submitted Feb 17, 2000; accepted May 16, 2000.
understand that persistent F 1+2 levels might indicate a baseline prothrombotic state, but D-dimer levels should reflect the breakdown of formed thrombus. We know that after 10 days to 2 weeks, thrombus inside of veins turns to scar tissue. With this in mind, where is the D-dimer coming from at these later follow-up time points? In fact, you show levels of D-dimer increased at 9 months and 1 year compared with the 6-month time point in your study. Secondly, can you make some comments concerning the type of D-dimer assay used in your study compared with the rapid bedside D-dimer assays that are on the market today? Would your data change or remain the same using a simple rapid bedside test? I congratulate the authors on a fine piece of work and thank the American Venous Forum for the opportunity to discuss this paper. Dr Mark Meissner. Thank you, Dr Wakefield. With respect to your question regarding the late D-dimer elevations I think you are exactly right. As you can see from the graph, D-dimer levels declined rapidly initially, which likely reflects a decrease in generation of new lysable fibrin. With the absence of new lysable fibrin, D-dimer levels
JOURNAL OF VASCULAR SURGERY November 2000
880 Meissner et al
plateau. I think it is most important to note that the F 1+2 levels stayed elevated in a substantial number of patients and that there was likely ongoing fibrin generation in these patients accounting for the persistently elevated Ddimer levels. I would propose, although I have no data to prove it, that the D-dimer levels late during follow-up result from the generation of new thrombin that is the result of the elevated F 1+2 levels. Regarding the assay used in this study, this was a quantitative assay, which as most people are aware, is a long test
O
requiring approximately 4 hours to perform. For research purposes, the tests were run in batches. The sensitivity and specificity levels and our cutoff level of 400 were based on the assay we use clinically, which is a rapid ELISA test that returns results within 1 hour. Based on the literature, these results are probably applicable to rapid ELISA tests. I think they are probably not applicable to latex agglutination tests, which I do not think have similar sensitivity and specificity at the levels used in this study. Thank you very much.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL, 32887
OR FAX TO: 407-363-9661
OR PHONE: 800-654-2452 Outside the US, call 407-345-4000
The development of sensitive assays for stable markers of activated coagulation aznd fibrinolysis has provided new insight into the pathophysiology of acute deep venous thrombosis (DVT) and may have some role in its diagnosis. Although the activated enzymes of coagulation and fibrinolysis From the Departments of Surgerya and Laboratory Medicine,b University of Washington School of Medicine. Competition of interest: nil. Supported by National Institutes of Health grant No. 619578. Presented at the Twelfth Annual Meeting of the American Venous Forum, Phoenix, Ariz, Feb 4, 2000. Reprint requests: Mark H. Meissner, MD, Department of Surgery, Box 359796, Harborview Medical Center, 325 9th Ave, Seattle, WA 98195 (e-mail: [email protected]). Copyright © 2000 by The Society for Vascular Surgery and The American Association for Vascular Surgery, a Chapter of the International Society for Cardiovascular Surgery. 0741-5214/2000/$12.00 + 0 24/6/110359 doi:10.1067/mva.2000.110359
870
are evanescent, measurable byproducts of zymogen activation include D-dimer fibrin degradation products formed by the action of plasmin on cross-linked fibrin; prothrombin fragment 1+2 (F 1+2) generated by factor Xa in the cleavage of prothrombin to thrombin; fibrinopeptide A formed in the thrombin-mediated conversion of fibrinogen to fibrin; and thrombin-antithrombin complex formed by the combination of thrombin with its primary inhibitor. 1,2 Among these, F 1+2, fibrinopeptide A, and thrombin-antithrombin complex similarly reflect the degree of thrombin activation, whereas D-dimer reflects fibrinolytic activity and the presence of intravascular thrombus. The initial changes in these markers after thrombosis and treatment have been described,3-5 and plasma D-dimer levels have been incorporated into some diagnostic algorithms.6,7 However, the long-
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
871
Fig 1. Correlation between extent of thrombus (SVS/NA-ISCVS thrombus score) at presentation and initial plasma D-dimer level (R = 0.39, P = .003).
Table I. Thrombotic risk factors term changes occurring in these markers have not been examined in detail. The evolution of these markers over time is critical in their application to recurrent venous thrombosis, which remains a fundamental challenge in the management of DVT. Although the adequacy of initial anticoagulation appears to be critical in preventing recurrent thrombosis,8 the ability to identify subgroups at particular risk for such events is limited. This is especially important for thrombosis isolated to the calf veins because treatment is primarily based on the risk of proximal propagation. Because the development of recurrent DVT is presumably determined by the relative balance between coagulation and fibrinolysis, these markers may have some utility in identifying patients at highest risk. Changes in these markers may also be relevant to the diagnosis of recurrent DVT. Although venous duplex ultrasonography canning is now the study of choice for the initial diagnosis of DVT, its limitations in the diagnosis of recurrent DVT are well recognized. Ultrasound scan interpretation in this setting may be limited by variable degrees of residual occlusion, partial recanalization, intimal thickening, and collateral formation. Results of compression ultrasound studies may remain abnormal in 27% to 70% of patients after 1 year.9-12 Favorable reports regarding the use of D-dimer in the initial diagnosis suggest that it may also have a role in the diagnosis of recurrent DVT. The purpose of this study was therefore to evaluate the initial determinants of quantitative D-dimer and F 1+2 levels, to document changes in these lev-
Risk factor Age > 40 y Malignancy > 3 day’s immobilization Estrogen therapy Surgery Cardiac disease Prolonged (> 6 h) travel Limb trauma Family history Previous DVT Pregnant Known prothrombotic state Obese
No. (%) 44 23 18 12 11 6 6 6 5 3 2 1 0
(72.1) (37.7) (29.5) (19.7) (18.0) (9.8) (9.8) (9.8) (8.2) (4.9) (3.3) (1.6) (0)
els after initial presentation, and to evaluate their utility in the diagnosis of recurrent DVT. METHODS All patients presenting to the University of Washington Medical Center with an acute DVT confirmed by duplex ultrasonography were asked to participate in a long-term study of the natural history of DVT. Patients with a limited life expectancy receiving prophylactic or therapeutic doses of anticoagulants before enrollment or having a residence or medical condition preventing their return were excluded from enrollment. All treatment decisions were made by the patient’s primary physician. The Human Subjects Committee at the University of Washington approved the study protocol, and informed consent was obtained from all subjects. Immediately after enrollment and before administration of anticoagulants, blood was drawn for
872 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 2. Correlation between extent of thrombus (SVS/NA-ISCVS thrombus score) at presentation and initial plasma F 1+2 level (R = 0.28, P = .08).
determination of plasma D-dimer cross-linked fibrin degradation products and prothrombin fragment 1+2 levels. A complete medical history was obtained, and thrombotic risk scores were calculated according to the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery (SVS/ISCVS) reporting standards in venous disease.13 Patients were then requested to return for follow-up visits at intervals of 3, 7, and 14 days and 1, 3, 6, 9, and 12 months after initial presentation. At each visit, patients were questioned regarding limb symptoms, their lower extremities were examined, and duplex ultrasonography of the lower extremity veins was performed. Blood was also drawn for determination of plasma D-dimer and F 1+2 levels. Duplex scanning of both lower extremities was performed in the 15-degree reverse Trendelenburg position with an Ultramark 9 instrument (Advanced Technology Laboratories, Bothell, Wash). A phased array 3-2 MHz transducer was used for the inferior vena cava and iliac segments, and a linear array 7-4 MHz transducer was used for the veins below the inguinal ligament. All superficial and deep venous segments from the inferior vena cava to the calf veins were assessed for the presence of thrombus at each visit. Individual venous segments were defined as patent, partially occluded, or completely occluded on the basis of established B mode and Doppler criteria,14 and the extent of thrombosis was determined on a 24-point scale according to the reporting standards in venous disease.13 Recurrent thrombotic events were defined by ultrasound criteria to include
propagation of thrombus to initially uninvolved venous segments, progression of nonocclusive thrombus within a segment to complete occlusion, rethrombosis of a partially or completely recanalized segment, or new thrombus within an initially uninvolved contralateral limb. The events were further defined as occurring early (≤ 30 days) or late (> 30 days) after initial presentation. Blood for all laboratory studies was drawn with a two-tube technique, discarding the first 4 to 5 mL. Samples were anticoagulated by the addition of 4.5 mL of whole blood to 0.5 mL of 0.105 mol/L sodium citrate and centrifuged at 2000g for 15 minutes at room temperature. Citrate plasma was removed and frozen at –70°C before analysis. Quantitative D-dimer levels were determined with an enzyme-linked immunosorbent assay (ELISA), as previously described.15 For Ddimer, the normal range (< 400 ng/mL) was defined as identical to that used for the exclusion of DVT in the clinical laboratory with a rapid D-dimer assay (sensitivity, 94.6%; specificity, 35%).16 F 1+2 (normal, 0.4-1.1 nmol/L) was similarly assessed with an ELISA method from Behring Inc.17 Because most variables were not uniformly distributed, all data are reported as the median and interquartile range (IQR, 25th to 75th percentile). Bivariate analysis was performed with linear regression or the Mann-Whitney test for continuous variables and the χ2 test for categoric variables. Determinants significant at a P value of .10 or less were entered into multivariate stepwise regression and logistic regression models. The sensitivity and specificity of various D-dimer and F 1+2 levels in
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
873
Table II. Follow-up D-dimer and F 1+2 levels Follow-up
D-dimer
interval
No.
Day 0 Day 3 Day 7 Day 14 Day 30 3 mo 6 mo 9 mo 1y
55 50 50 46 46 43 33 27 20
F 1+2
Median (IQR) 2320 2033 1719 1034 660 440 407 534 646
(1364-5733) (1248-3252) (1008-3444) (475-2749) (272-1362) (208-999) (209-861) (356-1184) (383-1542)
No. ≥ 400*
No.
51 47 46 37 31 26 18 18 14
55 51 50 48 46 44 34 27 20
(92.7%) (94%) (92%) (80.4%) (67.4%) (60.5%) (54.5%) (66.7%) (70%)
Median (IQR) 2.2 1.5 1.2 0.5 0.7 0.6 0.6 1.0 0.9
(1.7-3.5) (1.0-2.2) (0.6-2.0) (0.3-1.7) (0.4-1.6) (0.3-1.2) (0.3-1.2) (0.5-1.3) (0.4-1.3)
No. ≥ 1.1* 52 34 27 21 20 14 9 11 7
(94.5%) (66.7%) (54%) (43.8%) (43.5%) (31.8%) (26.5%) (40.7%) (35%)
*Number of patients with values above the normal range.
Table III. D-dimer and F 1+2 levels at 6 months’ follow-up D-Dimer < 400 ng/mL (n = 15)* 6-mo thrombus score Initial thrombus score Age (y) Risk score Risk factors Irreversible factors†
2.0 2.0 48 2.0 2.0 1.0
(0.0-3.0) (2.0-4.0) (29-64) (1.0-3.0) (1.0-2.0) (1.0-1.0)
≥ 400 ng/mL (n = 18)* 1.0 2.0 60 4.5 3.0 2.0
(0.75-2.25) (1.0-7.0) (51-73) (2.0-6.0) (1.75-3.25) (1.0-2.0)
F 1+2 P value .57 .85 .05 .01 .009 .04
< 1.1 nmol/L (n = 24)* 1.0 2.5 51 2.0 2.0 1.0
(1.0-3.0) (2.0-6.75) (33-63) (2.0-4.5) (1.0-3.0) (0.25-2.0)
≥ 1.1 nmol/L (n = 9)* 1.0 2.0 60 4.0 3.0 2.0
(0-3.5) (1.0-4.5) (56-75) (3.0-5.5) (2.5-3.5) (1.0-2.0)
P value .68 .33 .03 .09 .03 .07
*Median (IQR). †Number of irreversible risk factors, including age over 40 years, malignancy, cardiac disease, previous history of DVT, family history of thrombosis, known prothrombotic state, and obesity.
predicting recurrent thrombotic events were evaluated with receiver-operator characteristic (ROC) analysis with Rockit statistical software (Department of Radiology, University of Chicago, Chicago, Ill). Sensitivity for the diagnosis of recurrent thrombosis was defined as true-positive results in patients with recurrence divided by total patients with recurrence, and specificity was defined as true-negative results in patients without recurrence divided by total patients without recurrence. Statistical significance was defined at a P value of .05 or less. RESULTS Between October 1997 and September 1999, 61 patients with an acute DVT were enrolled and followed up for a median of 266 days (IQR, 91.5-364 days). Subjects included 30 (49.2%) women and 31 (50.8%) men with a median age of 54.6 years (IQR, 39.5-64.3 years). The median SVS/NA-ISCVS risk score among these patients was 4 (IQR, 2-6), with a median of 2.0 (IQR, 1.5-3.0) risk factors per patient (Table I). Thrombosis was bilateral in nine (14.8%) patients and unilateral in the right and left limbs in 22 (36.1%) and 30 (49.2%) patients, respectively.
These included 34 (55.7%) proximal thrombi and 27 (44.3%) isolated calf vein thrombi with a median thrombus score of 3.0 (IQR, 2.0-7.0). Of the 57 patients completing early follow-up, 42 (73.7%) were initially treated with intravenous unfractionated or subcutaneous low-molecular-weight heparin, 2 (3.5%) were treated with warfarin alone, and 13 (22.8%) were not treated with anticoagulants. Among those treated with unfractionated heparin, the median duration of treatment was 5.0 days (IQR, 4.0-9.5 days), and a median of 80% (IQR, 50100) of partial thromboplastin times was within the therapeutic range. Fifty-five patients had blood samples obtained at the time of presentation. Among these patients, median D-dimer and F 1+2 levels were 2320 ng/mL (IQR, 1364-5733 ng/mL) and 2.2 nmol/L (IQR, 1.7-3.5 nmol/L), respectively. D-dimer and F 1+2 levels were elevated in 51 (92.7%) and 52 (94.5%) patients, respectively. Initial D-dimer levels were significantly related to total thrombus score (P = .003), whereas F 1+2 levels were only marginally related to the extent of thrombus (P = .08, Figs 1 and 2). In comparison with those with proximal
874 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 3. Median D-dimer levels during follow-up stratified according to the presence of ultrasound scan–defined recurrent thrombosis. Dashed line indicates diagnostic limit for acute DVT (≥ 400 ng/mL). Plasma levels at day 0 are significantly different (P = .002) between those with and without subsequent recurrent events. Number of patients evaluated at each follow-up interval and the corresponding IQR are shown.
thrombosis, patients with isolated calf vein thrombosis had significantly lower D-dimer (1967 ng/mL [IQR, 1128-3213 ng/mL] vs 3356 ng/mL [IQR, 2016-8469 ng/mL]; P = .02) and F 1+2 (2.0 nmol/L [IQR, 1.5-2.4 nmol/L] vs 2.7 nmol/L [IQR, 2.0-4.4 nmol/L]; P = .004) levels. A history of prolonged travel was the only thrombotic risk factor related to initial D-dimer values, with median levels being higher among those without a travel history (2745 ng/mL [IQR, 1562.5-6679.5 ng/mL] vs 1462.5 ng/mL [IQR, 271.8-2240.5 ng/mL]; P = .03). A history of cardiac disease (3.6 nmol/L [IQR, 2.23-4.68 nmol/L] vs 2.1 nmol/L [IQR, 1.6-3.1 nmol/L]; P = .05) was associated with higher initial F 1+2 levels, whereas recent limb trauma (1.6 nmol/L [IQR, 0.75-1.85 nmol/L] vs 2.4 nmol/L [IQR, 1.8-3.73 nmol/L]; P = .001) was associated with lower levels. However, in a multivariate analysis including extent of thrombosis, proximal versus distal thrombosis, and risk factors, thrombus score was the only significant predictor of initial D-dimer levels (P = .003), whereas the presence of proximal thrombosis was the only significant predictor of F 1+2 levels (P = .001). Both D-dimer and F 1+2 levels decreased rapid-
ly after initial presentation. D-dimer levels at day 3 were independent of initial treatment, with median levels among those treated with unfractionated or low-molecular-weight heparin (2037 ng/mL [IQR, 1041-3376 ng/mL]) being similar to levels of those who were untreated or treated with warfarin alone (2028 ng/mL [IQR, 1693-2578 ng/mL]; P = .73). However, median F 1+2 levels at day 3 were significantly lower among those treated with unfractionated or low-molecular-weight heparin (1.1 nmol/L [IQR, 0.8-2.0 nmol/L] vs 2.0 nmol/L [IQR, 1.83.0 nmol/L]; P = .001). Median F 1+2 levels normalized by 14 days, whereas median D-dimer levels approached normal levels by 3 months (Table II). However, after 1 year of follow-up, D-dimer and F 1+2 levels remained elevated in 70% and 35% of patients, respectively. In comparison with those with normal D-dimer and F 1+2 levels, subjects with persistently elevated levels at 6 months were significantly older and had higher risk scores, greater numbers of thrombotic risk factors, and more irreversible risk factors. In contrast, extent of thrombosis both at presentation and at the 6-month follow-up interval was not significantly different from those in patients with nor-
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Meissner et al
875
Table IV. Interval D-dimer increases Recurrent thrombosis
Interval D-dimer increase
Yes No Total
mal D-dimer and F 1+2 levels (Table III). Similar trends were present for those with elevated levels at 1 year, although because of small patient numbers, differences in risk score, number of risk factors, and number of irreversible risk factors did not reach statistical significance. Thirty-seven recurrent thrombotic events were observed in 26 (43%) patients. Although clustered early after presentation, recurrent thrombosis was observed throughout the 1-year follow-up interval. Fifteen (40.5%) events were observed within the first 3 days of follow-up, 14 (37.8%) between days 3 and 30, and 4 (10.8%) in each of the intervals between 1 and 6 months and 6 and 12 months. Neither thrombus score (P = .15) nor overall risk score (P = .35) was significantly related to recurrent thrombotic events. However, initial D-dimer (P = .002) and F 1+2 (P = .009) levels were significantly higher among patients with recurrent thrombosis (Figs 3 and 4). In a logistic regression model, initial F 1+2 levels (P = .04) were more highly related to recurrent events than initial D-dimer levels (P = .33). However, initial D-dimer levels performed slightly better in the receiver-operator characteristic analysis. In predicting recurrent thrombosis, an initial Ddimer level of 2000 ng/mL or greater provided a sensitivity of 87% and a specificity of 56.3%, whereas F 1+2 levels of 2.0 nmol/L or greater had a sensitivity of 82.6% and specificity of 43.8%. Similarly, for thrombosis confined to the calf, a D-dimer level of 2000 ng/mL or greater had a sensitivity and specificity of 88.9% and 76.5% in predicting recurrent events, whereas F 1+2 levels of 2.0 nmol/L or greater had a sensitivity of 66.7% and specificity of 52.9%. Among patients with isolated calf vein thrombosis, recurrent thrombosis developed in 67% of those with an initial D-dimer level of 2000 ng/mL or greater in comparison with 7% of those with levels of less than 2000 ng/mL. For F 1+2, recurrent thrombosis occurred in 43% of patients with isolated calf vein thrombosis and an F 1+2 level of 2.0 nmol/L or greater in comparison with 25% of those with levels of less than 2.0 nmol/L. Persistent D-dimer and F 1+2 elevations limited
Yes
No
Total
10 18 28
102 151 253
112 169 281
Table V. Interval F 1+2 increases Recurrent thrombosis
Interval F 1+2 increase Yes No Total
Yes
No
Total
8 20 28
79 181 260
87 201 288
the utility of the threshold values used for the initial diagnosis of DVT. F 1+2 levels of 1.1 nmol/L or greater had an overall sensitivity and specificity of 62.9% and 49.3%, respectively, for the diagnosis of recurrent thrombotic events. If only early (≤ 30 days) events were considered, sensitivity improved to 71.4%, and specificity decreased to 39.6%. For the diagnosis of late recurrence (> 30 days), the sensitivity was only 28.6%, with a specificity of 67.5%. For D-dimer levels, different critical values could be selected on the basis of the timing of recurrent thrombosis. For early recurrence, D-dimer levels of 1000 ng/mL or greater provided an optimal sensitivity and specificity of 89.3% and 35.6%, respectively. Positive and negative predictive values in this population were 15.1% and 96.3%, respectively. For late recurrence, D-dimer levels of 500 ng/mL or greater provided a sensitivity of 100% and specificity of 53.9%, with positive and negative predictive values of 11.7% and 88.3%, respectively. In comparison with the value at the previous visit, the sensitivity and specificity of an increase in D-dimer levels were 35.7% and 59.7%, respectively, with a positive and negative predictive value of 8.9% and 89.3% in this population (Table IV). For interval increases in prothrombin F 1+2 levels, sensitivity and specificity were 28.6% and 69.6%, respectively, with a positive and negative predictive value of 9.2% and 90.0% (Table V). DISCUSSION The development of specific assays for stable markers of activated coagulation and intravascular thrombus formation has provided new insight into the pathophysiology of acute DVT. In contrast to the measurement of zymogen levels and standard clotting tests, these assays are able to detect low-level activation
876 Meissner et al
JOURNAL OF VASCULAR SURGERY November 2000
Fig 4. Median F 1+2 levels during follow-up stratified according to the presence of ultrasound scan–defined recurrent thrombosis. Dashed line indicates upper limit of normal range (1.1 nmol/L). Plasma levels at day 0 are significantly different (P = .009) between those with and without subsequent recurrent events. Number of patients evaluated at each follow-up interval and the corresponding IQR are shown.
of the coagulation system and localized thrombotic events.1,2 Several thrombotic risk factors have been associated with increased levels of thrombin activation markers, confirming the importance of activated coagulation in thrombogenesis. Increased levels of these markers have been described in association with advanced age,2,18 surgery,19 trauma,20 oral contraceptives,21,22 and malignancy.23 Elevated levels of these markers have also been shown to be useful in excluding a diagnosis of acute DVT. Although of variable sensitivity, the uniformly low specificity of these tests requires confirmatory testing of patients with positive results. D-dimer levels appear to be more sensitive than the markers of thrombin activation,1,3,4,16,24 an observation that may reflect fluctuation in thrombin activity and the longer half-life of D-dimer. In 11 published series of symptomatic acute DVT, the average weighted sensitivity and specificity of D-dimer determined with the ELISA technique have been reported to be 96.8% and 35.2%, respectively.25 Lower sensitivities of 47% to 66% have been reported for F 1+2 levels.3,24 Although not a study of diagnostic accuracy, the results of this study are consistent with these reports. D-dimer and F 1+2 levels were elevated at the time of
presentation in 92.7% and 94.5% of patients, respectively. However, this study further demonstrates that initial D-dimer levels are most significantly related to the extent of thrombosis, whereas F 1+2 levels are more highly associated with the presence of proximal thrombus. The relationship between D-dimer level and thrombus extent most likely reflects the quantity of thrombus susceptible to the action of plasmin. The magnitude of the correlation between plasma Ddimer level and thrombus extent determined by ultrasound (R = 0.39) is similar to that reported for venographically determined thrombus extent (R = 0.340.36).3,26 Although statistically significant, this relatively weak correlation suggests that fibrinolytic differences between individuals may also be important. The relationship between F 1+2 level and thrombus location is more interesting because it implies that thrombus confined to the calf veins is associated with less extensive thrombin activation. Although it is clear that most thrombi begin in the calf veins,27,28 this observation further suggests that isolated calf vein thrombi may differ pathophysiologically from proximal thrombi. That is, they are not simply early thrombi that have yet to propagate, but rather reflect a more limited prothrombotic state.
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
Although this was a small study, thrombotic risk factors were not independently associated with initial D-dimer or F 1+2 levels. Patients having risk factors associated with underlying elevations of thrombin activation products might have been expected to have higher levels of these markers after the development of DVT. The finding that these risk factors are not associated with increased F 1+2 levels after thrombosis may reflect either the strong amplification of the coagulation cascade associated with clinical thrombosis29 or the existence of a thrombotic threshold at which acute DVT occurs independent of baseline prothrombin activation.30 The latter would imply that a similar level of thrombin activity is required for thrombosis to occur, regardless of the underlying risk factors. A larger number of patients with levels measured before thrombosis would be required to substantiate this hypothesis. Several studies have examined changes in plasma coagulation and fibrinolytic markers during the first 10 days of treatment,3-5 but the long-term changes after an episode of DVT have not been well characterized. Although some have found markers of thrombin activation to normalize early after effective treatment is instituted,4 most have found these markers to rapidly decrease, but not fully normalize, during the first 10 days. Consistent with these observations, median D-dimer and F 1+2 levels in this study rapidly declined after initial presentation. However, despite this overall trend, these markers remained abnormal for prolonged periods in a substantial number of patients. Although median F 1+2 values returned to normal with the first 14 days, a prothrombotic state, implied by elevated F 1+2 levels, persisted in 27% to 44% of the patients evaluated between 14 days and 1 year after thrombosis. This was associated with a 43% incidence of ultrasounddocumented recurrent events. Consistent with a persistent thrombus load, median D-dimer levels approached normal by 3 months but remained elevated throughout the first year in most patients. The initial decrease in both markers was likely caused by a rapid reduction in thrombin activity and the generation of new lysable fibrin, whereas the slower subsequent decline in D-dimer levels reflects the decreased susceptibility of older thrombus to lysis.4 In contrast to the initial levels, persistent elevations of both markers were related to age and thrombotic risk, which is consistent with ongoing hypercoagulability and perhaps fibrin generation. These observations have important implications for the diagnosis of recurrent thrombosis. Fluctuations in the levels of both markers make interval
Meissner et al
877
increases in their levels of little diagnostic significance. This contrasts with other small series suggesting an increase in D-dimer levels of greater than 3 µg/mL within the first 10 days to be a specific but insensitive indicator of recurrent thrombosis.3 Furthermore, persistent elevations will a priori reduce the specificity and positive predictive value of these markers with the criteria used for the initial diagnosis of DVT. Despite these limitations, these tests may still have some utility in the diagnosis of recurrent DVT if sensitivity and negative predictive value are sufficiently high. Although most patients with elevated levels will not have recurrent thrombosis (low positive predictive value), low levels can exclude the diagnosis. This is particularly true for D-dimer, where values of 1000 ng/mL or greater and 500 ng/mL or greater had sensitivities of 89% and 100%, respectively, in detecting early and late recurrences. As in the initial diagnosis of DVT, the low specificities and positive predictive value of these measurements limit their utility to the exclusion, rather than the confirmation, of recurrent disease. However, given the potential limitations of other diagnostic modalities, including venous duplex ultrasonography and venography, Ddimer measurements may have some clinical value in this setting. A D-dimer level of less than 500 ng/mL in a patient with pain, swelling, and a duplex scan demonstrating thrombus of indeterminate age 9 months after presentation could potentially allow the patient to be spared the risk and inconvenience of reinstituting anticoagulation. Perhaps of greater value is the relationship between initial levels of these markers and subsequent recurrent events. Although specificity was again limited, D-dimer levels of 2000 ng/mL or greater and F 1+2 levels of 2.0 nmol/L or greater had sensitivities of 87% and 83%, respectively, in predicting recurrent thrombotic events. This may have important implications, particularly for the management of isolated calf vein thrombosis. Although management remains controversial, anticoagulant treatment of these thrombi is often recommended on the basis of the observation that approximately 20% will propagate to a more proximal level.31,32 In this regard an initial D-dimer level of 2000 ng/mL or greater was highly associated with the aggregate outcome of recurrent thrombosis, which included proximal propagation. Two thirds of patients with isolated calf vein thrombosis and a D-dimer level of 2000 ng/mL or greater had recurrent thrombosis in comparison with less than 10% of those with lower levels. Although larger confirmatory studies are obviously required, patients with initial D-dimer lev-
JOURNAL OF VASCULAR SURGERY November 2000
878 Meissner et al
els of 2000 ng/mL or greater are perhaps most appropriately treated with standard anticoagulation, whereas serial noninvasive follow-up may be an option in those with levels below 2000 ng/mL. Other uses of these markers remain speculative, but they may have some role in more precisely defining the appropriate duration of anticoagulation in individual patients. The relationship between persistently elevated D-dimer and F 1+2 levels and thrombotic risk factors tends to support current recommendations that patients with irreversible thrombotic risk factors warrant a longer duration of anticoagulation. Such measurements could also conceivably be useful in identifying those patients with idiopathic DVT at highest risk for recurrence and requiring extended anticoagulation. Validation of such speculation would obviously require larger series examining the relationship between these markers at the completion of a standard course of anticoagulation and the incidence of late recurrent events. It is important to recognize the shortcomings of this study, most notably the small number of patients. The number of patients in some subgroups, such as those having specific risk factors with late recurrence, was quite small. It is certainly possible that important relationships between risk factors, D-dimer and F 1+2 levels, and recurrent thrombosis were overlooked. Furthermore, although consecutive patients meeting entry criteria were enrolled, and the ultrasound scans and laboratory tests were independently interpreted, this was an exploratory study not designed primarily to document the diagnostic accuracy of these markers. Neither was it designed as a management trial to evaluate the use of these markers in clinical decision making. It is also important to note that quantitative ELISA assays were used in this study and that the results may not be applicable to semiquantitative latex and hemagglutination tests. Finally, a standardized treatment regimen was not used in this study. Although D-dimer levels were independent of initial treatment, early F 1+2 levels were significantly related to anticoagulation with unfractionated or low-molecular-weight heparin. Although unlikely to change the D-dimer results, anticoagulation of the 26.3% of patients who were not initially treated could alter the time course of F 1+2 levels, potentially making this marker more suitable for the diagnosis of recurrent thrombotic events. Despite these limitations, this study further defines the evolution of stable markers of activated coagulation and fibrinolysis after an episode of acute DVT. Initial F 1+2 levels are most significantly associated with the presence of proximal thrombus, where-
as D-dimer levels reflect the extent of thrombus and perhaps individual differences in the fibrinolytic system. Initial levels of both markers are predictive of recurrent thrombotic events. This observation may have particular relevance to the treatment of isolated calf vein thrombosis and warrants further investigation. Although levels of both markers rapidly decline after presentation, they remain elevated in a substantial number of patients for at least 1 year. This limits the specificity and positive predictive value of these markers in recurrent DVT, although they appear to have some utility in excluding the diagnosis. REFERENCES 1. Boisclair MD, Ireland H, Lane DA. Assessment of hypercoagulable states by measurement of activation fragments and peptides. Blood Rev 1990;4:25-40. 2. Bauer KA, Rosenberg RD. The pathophysiology of the prethrombotic state in humans: insights gained from studies using markers of hemostatic system activation. Blood 1987;70:343-50. 3. The DVTENOX Study Group. Markers of hemostatic system activation in acute deep venous thrombosis—evolution during the first days of heparin treatment. Thromb Haemost 1993;70:909-14. 4. Speiser W, Mallek R, Koppensteiner R, Stumpflen A, Kapiotis S, Minar E, et al. D-dimer and TAT measurement in patients with deep venous thrombosis: utility in diagnosis and judgement of anticoagulant treatment effectiveness. Thromb Haemost 1990;64:196-201. 5. Bridey F, Phlippoteau C, Simmoneau G, Musset D, Dreyfus M. Is D-Dimer measurement a marker of recurrence of thromboembolic disease? [Abstract]. Thromb Haemost 1991;65:981. 6. Bernardi E, Prandoni P, Lensing AW, Agnelli G, Guazzaloca G, Scannapieco G, et al. D-dimer testing as an adjunct to ultrasonography in patients with clinically suspected deep vein thrombosis: prospective cohort study. The Multicentre Italian D-dimer Ultrasound Study Investigators Group. BMJ 1998;317:1037-40. 7. Michiels JJ. Rational diagnosis of deep vein thrombosis (RADIA DVT) in symptomatic outpatients with suspected DVT: simplification and improvement of decision rule analysis for the exclusion and diagnosis of DVT by the combined use of a simple clinical model, a rapid sensitive D-dimer test and compression ultrasonography (CUS). Semin Thromb Hemost 1998;24:401-7. 8. Caps MT, Meissner MH, Tullis MJ, Polissar NL, Manzo RA, Zierler BK, et al. Venous thrombus stability during the acute phase of therapy. Vasc Med 1999;4:9-14. 9. Murphy TP, Cronan JJ. Evolution of deep venous thrombosis: a prospective evaluation with US. Radiology 1990;177:543-8. 10. Mantoni M. Deep venous thrombosis: longitudinal study with duplex US. Radiology 1991;179:271-3. 11. Prandoni P, Cogo A, Bernardi E, Villalta S, Polistena P, Simioni P, et al. A simple ultrasound approach for detection of recurrent proximal vein thrombosis. Circulation 1993;88:1730-5. 12. Baxter GM, Duffy P, MacKechnie S. Colour Doppler ultrasound of the post-phlebitic limb: sounding a cautionary note. Clin Radiol 1991;43:301-4.
JOURNAL OF VASCULAR SURGERY Volume 32, Number 5
13. Porter J, Moneta G, International Consensus Committee on Chronic Venous Disease. Reporting standards in venous disease: an update. J Vasc Surg 1995;21:635-45. 14. Meissner MH. Venous duplex scanning. In: Rutherford RB, editor. Vascular surgery. 5th ed. Philadelphia: W.B. Saunders Company; 2000. p. 214-29. 15. van Beek EJR, van den Ende B, Berckmans RJ, ven der Heide YT, Brandjes DPM, Sturk A, et al. A comparative analysis of D-Dimer assays in patients with clinically suspected pulmonary embolism. Thromb Haemost 1993;70:408-13. 16. Escoffre-Barbe M, Oger M, Leroyer C, Grimaux M, Le Moigne E, Nonent M, et al. Evaluation of a new rapid DDimer assay for clinically suspected deep venous thrombosis (Liatest D-Dimer). Am J Clin Pathol 1998;109:748-53. 17. Pelzer H, Schwarz A, Stuber W. Determination of human prothrombin activation fragment 1+2 in plasma with an antibody against a synthetic peptide. Thromb Haemost 1991;65:153-9. 18. Cadroy Y, Pierrejean D, Fontan B, Sie P, Boneu B. Influence of aging on the activity of the hemostatic system: prothrombin fragment 1+2, thrombin-antithrombin III complexes and D-dimers in 80 healthy subjects with age ranging from 20 to 94 years. Nouv Rev Fr Hematol 1992;34:43-6. 19. Kambayashi J, Sakon M, Yokota M, Shiba E, Kawasaki T, Mori T. Activation of coagulation and fibrinolysis during surgery, analyzed by molecular markers. Thromb Res 1990;60:157-67. 20. Gando S, Tedo I, Kubota M. Posttrauma coagulation and fibrinolysis. Crit Care Med 1992;20:594-600. 21. von Kaulla E, Droegemueller W, Aoki N, von Kaulla KN. Antithrombin III depression and thrombin generation acceleration in women taking oral contraceptives. Am J Obstet Gynecol 1971;109:868-73. 22. Quehenberger P, Loner U, Kapiotis S, Handler S, Schneider B, Huber J, et al. Increased levels of activated factor VII and decreased plasma protein S activity and circulating thrombomodulin during use of oral contraceptives. Thromb Haemost 1996;76:729-34.
DISCUSSION Dr Thomas W. Wakefield (Ann Arbor, Mich). Dr Meissner and coworkers continue their excellent studies on venous thrombosis, and this is really a nicely done study and the manuscript is very scholarly. The current work suggests that measurements of Ddimer were significantly elevated in patients who suffered DVT, with thrombus score predicting initial D-dimer levels, while F 1+2 levels were also elevated with the presence of proximal thrombus the only significant predictor of F 1+2 levels. One year after thrombosis D-dimer and F 1+2 levels remained elevated in 70% and 35% of patients, however, which necessitated a change in the values noted important for defining recurrent DVT. Importantly, initial levels of D-dimer greater than 2000 ng/mL and F 1+2 greater than 2 nmol/L carried greater than 80% sensitivity for the diagnosis of recurrent DVT. This is key, as we know from other data in the literature that one of the most important factors in the development of chronic venous insufficiency is the presence of ipsilateral recurrent DVT. I am going to ask two questions. First, can you speculate on the cause of the persistent elevations in D-dimer and F 1+2 levels at 1 year after DVT formation? I can certainly
Meissner et al
879
23. Falanga A, Ofosu FA, Oldani E, Dewar L, Lui L, Barbui T. The hypercoagulable state in cancer patients: evidence for impaired thrombin inhibitions. Blood Coagul Fibrinolysis 1994;5 Suppl 1:S19-23. 24. Boneu B, Pelzer H, Boccalon H. D-dimers, thrombin antithrombin III complexes, and prothrombin fragments 1+2: diagnostic value in clinically suspected deep vein thrombosis. Thromb Haemost 1991;65:28-32. 25. Bounameaux H, de Moerloose P, Perrier A, Reber G. Plasma measurement of D-dimer as a diagnostic aid in suspected venous thromboembolism: an overview. Thromb Haemost 1994;71:1-6. 26. Mirshani M, Soria C, Soria J, Mirshani M, Faivre R, Kieffer Y, et al. Changes in plasma fibrin degradation products as a marker of thrombus evolution in patients with deep vein thrombosis. Thromb Res 1988;51:295-302. 27. Nicolaides AN, Kakkar VV, Field ES, Renney JT. The origin of deep vein thrombosis: a venographic study. Br J Radiol 1971;44:653-63. 28. Kakkar VV, Howe CT, Nicolaides AN, Clarke MB. Deep vein thrombosis of the leg. Is there a “high risk” group. Am J Surg 1970;120:527-30. 29. Amiral J, Fareed J. Thromboembolic diseases: biochemical mechanisms and new possibilities of biological diagnosis. Semin Thromb Hemost 1996;22:41-8. 30. Rosendaal FR. Venous thrombosis: a multicausal disease. Lancet 1999;353:1167-73. 31. Kakkar VV, Flanc C, Howe CT, Clarke MB. Natural history of post-operative deep-vein thrombosis. Lancet 1969;2:230-2. 32. Lagerstedt CI, Olsson C, Fagher BO, Oqvist BW, Albrechtsson U. Need for long-term anticoagulant treatment in symptomatic calf vein thrombosis. Lancet 1985;2:515-8.
Submitted Feb 17, 2000; accepted May 16, 2000.
understand that persistent F 1+2 levels might indicate a baseline prothrombotic state, but D-dimer levels should reflect the breakdown of formed thrombus. We know that after 10 days to 2 weeks, thrombus inside of veins turns to scar tissue. With this in mind, where is the D-dimer coming from at these later follow-up time points? In fact, you show levels of D-dimer increased at 9 months and 1 year compared with the 6-month time point in your study. Secondly, can you make some comments concerning the type of D-dimer assay used in your study compared with the rapid bedside D-dimer assays that are on the market today? Would your data change or remain the same using a simple rapid bedside test? I congratulate the authors on a fine piece of work and thank the American Venous Forum for the opportunity to discuss this paper. Dr Mark Meissner. Thank you, Dr Wakefield. With respect to your question regarding the late D-dimer elevations I think you are exactly right. As you can see from the graph, D-dimer levels declined rapidly initially, which likely reflects a decrease in generation of new lysable fibrin. With the absence of new lysable fibrin, D-dimer levels
JOURNAL OF VASCULAR SURGERY November 2000
880 Meissner et al
plateau. I think it is most important to note that the F 1+2 levels stayed elevated in a substantial number of patients and that there was likely ongoing fibrin generation in these patients accounting for the persistently elevated Ddimer levels. I would propose, although I have no data to prove it, that the D-dimer levels late during follow-up result from the generation of new thrombin that is the result of the elevated F 1+2 levels. Regarding the assay used in this study, this was a quantitative assay, which as most people are aware, is a long test
O
requiring approximately 4 hours to perform. For research purposes, the tests were run in batches. The sensitivity and specificity levels and our cutoff level of 400 were based on the assay we use clinically, which is a rapid ELISA test that returns results within 1 hour. Based on the literature, these results are probably applicable to rapid ELISA tests. I think they are probably not applicable to latex agglutination tests, which I do not think have similar sensitivity and specificity at the levels used in this study. Thank you very much.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL, 32887
OR FAX TO: 407-363-9661
OR PHONE: 800-654-2452 Outside the US, call 407-345-4000