Clinical Cancer Research Vol. 10, 7238-7243, November 1, 2004
© 2004 American Association for Cancer Research
Plasma Coagulation Markers in Patients with Solid Tumors and Venous Thromboembolic Disease Receiving Oral Anticoagulation Therapy
Sabah Sallah1,
Aisha Husain1,
Vaia Sigounas2,
Jim Wan3,
Francesco Turturro1,
George Sigounas2 and
Nam P. Nguyen4
1 Thrombosis and Hemostasis Program and Feist-Weiller Cancer Center, Louisiana State University Health Sciences Center, Shreveport, Louisiana; 2
Division of Hematology, East Carolina University, Greenville, North Carolina; 3
Department of Biostatistics, University of Tennessee, Memphis, Tennessee; and 4
Radiation Oncology, Southwestern University, Dallas, Texas
 |
ABSTRACT
|
|---|
Purpose: To correlate the concentration of plasma coagulation markers at baseline and during follow-up in patients with solid tumors and venous thromboembolic disease with the risk of recurrence and death.
Experimental Design: Patients (N = 223) with first episode of venous thromboembolic disease received oral anticoagulation with warfarin for a target international normalized ratio of 2 to 3. Plasma coagulation markers were measured before instituting warfarin and at 3 monthly intervals, thereafter.
Results: The median duration of oral anticoagulation was 6.7 months (range 2 weeks to 11 months). Major bleeding episodes occurred in 18 patients (8%), and minor hemorrhagic events occurred in 15 (6.7%) patients. Patients with advanced malignancy (P = 0.032), history of surgery (P = 0.057), and those with poor performance status (P = 0.001) were more likely to encounter major bleeding episodes. Recurrence of venous thromboembolic disease was diagnosed in 31 patients (14%). At univariate analysis, advanced stage of cancer (P = 0.03), performance status > 1 (P = 0.001), treatment with chemotherapy (P = 0.01), the presence of metastatic liver disease (P = 0.03), higher D-dimer (P = 0.001), and thrombin antithrombin complex levels (P = 0.01) were features predictive of recurrent venous thromboembolic disease. At multivariate analysis, poor performance status (P = 0.01) and D-dimer levels (P = 0.001) were predictors of recurrent venous thromboembolic disease. Persistent activation of coagulation as indicated by an upward trend in D-dimer (P = 0.001) and antithrombin (P = 0.001) was observed in patients who developed recurrent thrombosis. Similar upward trends in D-dimer (P = 0.001), antithrombin (P = 0.001), and prothrombin fragment F1 + 2 (P = 0.001) was observed in the 76 patients who died during the study period and in the patients who received chemotherapy.
Conclusions: Successful oral anticoagulation with warfarin in patients with cancer and venous thromboembolic disease is more likely to be achieved in patients with early stage tumors and good performance status. The persistence of activation of hemostasis as shown by plasma coagulation markers is a strong predictor of recurrence and poor outcome.
|
INTRODUCTION
|
|---|
Thrombotic complications are among the most common causes of death in patients with cancer. It is estimated that venous thromboembolic disease occurs in 
6.8% of patients with an underlying malignancy (1)
. A constellation of factors rather than a single event contributes to the hypercoagulable state of cancer. Chemotherapy, surgery, immobilization, and the presence of comorbid conditions are associated with increased risk of clotting in these patients (2, 3, 4)
. However, the leading factor that predisposes cancer patients to thrombosis seems to be related to the interaction of tumor cells with the hemostatic system (5
, 6)
. Cancer cells possess a host of procoagulant properties mediated through excessive release of angiogenic factors, cytokine release as well as direct effect on the vessel wall. Coagulation activation in cancer patients is evident by increased generation of thrombin shown by measurable hemostatic markers such as prothrombin fragment 1 + 2 (F1 + 2), thrombin antithrombin complex, and D-dimer (5, 6, 7, 8)
.
Despite the well-described occurrence of venous thromboembolic disease in malignancy, the subject of anticoagulation in cancer patients continues to be an interesting one (9, 10, 11, 12)
. In this study, we prospectively assessed the use of oral anticoagulation in patients with malignancy. The objective was to evaluate the utility of certain hemostatic markers in patients undergoing oral anticoagulation.
|
PATIENTS AND METHODS
|
|---|
Inclusion and Exclusion Criteria.
Between 1997 and 2003, patients with solid tumors and a diagnosis of first episode of venous thromboembolic disease were enrolled in this prospective study. Eligible patients had to be 18 years of age or older. Performance status was graded as per Eastern Cooperative Oncology Group. Patients with a performance status >2 were not included in the study. The diagnosis of extremity deep venous thrombosis was established with Doppler ultrasonography and/or venography. Ventilation-perfusion scan, spiral tomogram, or pulmonary angiogram was used to confirm the diagnosis of pulmonary embolism. Patients were excluded from the study if they had a diagnosis of primary tumor of the central nervous system, history of thromboembolic disorder, platelet count <80.000/mm3 at the time of venous thromboembolic disease, history of a hemorrhagic disorder, or if there was evidence of disseminated intravascular coagulation. Patients with metastatic lesions to the brain were not excluded. The study was approved by the Institutional Review Board, and patients signed an informed consent.
Anticoagulation.
All of the patients received initially unfractionated or low molecular weight heparin followed by warfarin. In general, warfarin was started 24 to 48 hours after heparin at a dose of 5 to 7.5 mg adjusted for international normalized ratio between 2 to 3. The average overlap between heparin and warfarin was 9 days. The duration of oral anticoagulation was predetermined at 6 months or as long as risk factors for thrombosis persisted. Immobilization and/or occurrence of venous thromboembolic disease while receiving chemotherapy were considered indications for oral anticoagulation beyond 6 months.
Coagulation Studies.
Blood samples for prothrombin time, international normalized ratio, and activated partial thromboplastin time were measured on the automated MLA Electra 1600 C instrument with photo-optometric clot detection system. Standard therapeutic range for international normalized ratio is 2.0 to 3.0, and normal activated partial thromboplastin time is 22 to 32.5 seconds. The D-dimers were assayed by semiquantitative agglutination of mouse anti-D antibody-coated latex particles (American Bioproducts, Parsippany, NJ); normal value <0.5 µg/mL. Prothrombin fragment 1 + 2 was determined by Enzygnost F1 + 2 (Dade Behrings, Marburg, Germany). Thrombin-antithrombin complex was done with antithrombin assay (Enzyme Research, Inc., South Bend, IN). The D-dimer, prothrombin F1 + 2, and thrombin-antithrombin were obtained before anticoagulation (baseline) at the 3-month interval during anticoagulation and at 2 weeks and 3 months after discontinuation of warfarin. Each measurement consisted of three sequential blood samples obtained at 0, 1, and 24 hours with the average of the three values reported as one measurement.
Definition of Variables.
Outcomes assessed in this study included the utility of plasma levels of coagulation markers in relation to the risk of recurrence of venous thromboembolic disease and death. Measurements of D-dimer, prothrombin F1 + 2, and antithrombin over sequential time points would circumvent the bias produced by abnormalities in coagulation markers that may be caused by the acute thrombotic event. Therefore, evaluation of the rate of change in these values is more meaningful than reporting absolute numbers.
Major bleeding was defined as a drop in hemoglobin concentration by 
2 g/dl, the need for transfusion support, or any serious hemorrhage such as intracranial or intra-abdominal bleeding. Bleeding episodes requiring medical attention, but not fulfilling the above definition, were considered as minor bleeding. A recurrent thrombotic event was defined as absence of compressible vein in a previously confirmed compressible venous segment on ultrasound or a new venographically demonstrable intraluminal filling defect.
Statistical Analysis.
Two-sample t test and 
2 test were used to compare patients with and without bleeding or with and without recurrence. For categorical variables, odds ratios and the corresponding 95% confidence intervals were also produced. Logistic regression was used for univariate and multivariate analysis. In the multivariate analysis, we used stepwise regression to select significant variables. The rate of decline of plasma coagulation markers over time was estimated by regression. These rates, together with their respective standard errors, were then compared between groups, such as recurrence versus no recurrence, chemotherapy versus surgery or radiation, and alive versus dead.
|
RESULTS
|
|---|
Characteristics of Patients.
A total of 223 patients with solid tumors and venous thromboembolic disease were enrolled into this study. The median age of the 111 men and 112 women was 57 years (range 29 to 81 years). There were 45 patients (20%) with breast cancer, 39 (18%) with lung cancer, 38 (17%) with genitourinary cancer, 36 (16%) with colorectal cancer, 32 (14%) with head and neck cancer, 19 (9%) with sarcoma, and 13 (6%) with liver cancer. The baseline clinical and laboratory data of all of the patients are shown in Table 1
. Conditions that contributed to the poor performance status of patients with stage I/II tumors included congestive heart failure, severe pulmonary disease, diabetes with complications, severe coronary artery disease, progressive arthritis, renal failure, and chronic liver disease. The median follow-up of the patients was 23 months (range 2 weeks to 44 months).
View this table:
[in this window]
[in a new window]
|
Table 1 Baseline clinical and laboratory features of all 223 patients with VTD and of the subset of patients that was diagnosed with recurrent thrombosis
|
|
Anticoagulation.
The median duration of oral anticoagulation was 6.7 months (range 2 weeks to 11 months). The average number of international normalized ratio measurements per patient was 11. The percentage of time spent in international normalized ratio 2 to 3 was 76%, international normalized ratio >3 was 16%, and international normalized ratio <2 was 8%. Major bleeding episodes occurred in 18 patients (8%; 3.6/100 patient years), whereas minor hemorrhagic events were observed in 15 patients (6.7%; 3/100 patient years). The median time from initiation of warfarin until the presentation with bleeding was 8 weeks (range 4 to 19 weeks). The median international normalized ratio in patients with major hemorrhage compared with those without bleeding was 3.9 versus 2.95 (P = 0.001). A significantly higher proportion of patients with advanced cancer had major bleeding episodes when compared with patients with early or localized malignancy (P = 0.032; odds ratio 2.4; 95% confidence interval 1.1, 5.4). Patients undergoing surgery (P = 0.05; odds ratio 1.43; 95% confidence interval 0.18, 1.04) and patients with poor performance status (P = 0.001; odds ratio 3.3; 95% confidence interval 1.5, 7.0) were more likely to have major bleeding complications. The presence of metastatic liver disease and administration of chemotherapy, P = 0.39 and P = 0.58, respectively, did not correlate with an increased risk for bleeding.
Recurrence of Venous Thromboembolic Disease and Coagulation Markers.
The characteristics of patients with recurrence are shown in Table 1
. Recurrence of venous thromboembolic disease was diagnosed in 31 patients (14%) with an overall incidence of 6.42/100 patient years. The median time of recurrence was 7.3 weeks (range 3 to 12 weeks). The difference in time to achieve therapeutic international normalized ratio between patients with and without recurrent venous thromboembolic disease was not statistically significant (P = 0.73). Advanced stage of cancer (P = 0.03), performance status >1 (P = 0.001), treatment with chemotherapy (P = 0.01), and the presence of metastatic liver disease (P = 0.03) were clinical variables predictive of recurrence of venous thromboembolic disease in univariate analysis.
An average of 4 measurements (12 samples) of D-dimer, prothrombin F1 + 2, and antithrombin per patient was done. The median baseline values of these plasma coagulation markers was as follows: (a) D-dimer 4.2 µg/mL (range 0.3 to 14.8); (b) prothrombin F1 + 2, 3.7 nmol/l (range 0.5 to 18.3); and (c) antithrombin 5.3 ng/mL (range 0.4 to 20.3). The patients who developed recurrent venous thromboembolic disease were more likely to have higher median baseline levels of D-dimer (6.8 versus 4.1 µg/mL; P = 0.03) and of thrombin-antithrombin (8.3 versus 5.3; P = 0.02) compared with the group of patients without recurrence. In a univariate analysis, upward trend in serial measurements of D-dimer (P = 0.01), and thrombin-antithrombin (P = 0.01) was predictive of recurrence of venous thromboembolic disease (Table 2
; Fig. 1
). In multivariate analysis, performance status and D-dimer levels were predictive of recurrent thrombosis; P = 0.01 and P = 0.001, respectively.
View this table:
[in this window]
[in a new window]
|
Table 2 Logistic regression of the clinical and laboratory variables in relation to recurrence of VTD in 223 patients with solid tumors
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1. A–C. The rate of change in the plasma levels of D-dimer (A), prothormbin F1 + 2 (B), and antithrombin (C) over the sequential time points as estimated by regression. The median change in the rate is represented by the midpoint, whereas the extremes represent 1 SD. Changes in prothrombin F1 + 2 were not predictive of recurrence (P = 0.07), however, changes in all of the other markers were statistically significant with respect to compared variables (P = 0.001). (TAT, thrombin antithrombin)
|
|
During the study period, 76 (34%) patients died. Of these patients, 67 (88%) had advanced or metastatic tumors and 23 (30%) had recurrent venous thrombosis of the lower extremity (17 patients) or pulmonary embolism (4 patients). Death was attributed to tumor progression in 65 (85.5%) of the patients and to pulmonary embolism in 2 of the patients. Intracranial and gastrointestinal bleeding was the cause of death in 2 patients, and acute renal failure and sepsis was the cause of death in 3 patients. The cause of death was not clear in 3 patients. Higher median baseline levels as well as upward trend in the sequential determination of the coagulation markers (Figs. 1)
were observed in the 76 patients who died: D-dimer (7.8 versus 4.2 µg/mL; P = 0.001), prothrombin F1 + 2, 7.1 versus 3.7 nmol/l; P = 0.001), and thrombin-antithrombin (8.8 versus 5.3 ng/mL; P = 0.001). Persistent activation of coagulation, as indicated by higher serial levels of these markers, was also detected in the 105 patients who received chemotherapy (Figs. 1)
.
|
DISCUSSION
|
|---|
In the current study, we evaluated the time course of markers of in vivo hypercoagulability in patients with cancer and a diagnosis of venous thromboembolic disease while undergoing anticoagulation with warfarin. We have shown in this trial that the persistent elevation or slower rate of decline in plasma coagulation markers correlated strongly with recurrence of venous thromboembolic disease. In particular, our finding in regards to D-dimer levels as a marker of persistent hemostatic activation and a predictor of recurrence of thrombosis is confirmatory to a recent report by ten Wolde et al. (13)
. The D-dimer represents the degradation products of fibrin that have been crosslinked by activated factor XIII. Activation of factor XIII requires the presence of thrombin, and increase in the levels of D-dimer indicates thrombin generation. It is important to note that D-dimer levels can be elevated in cancer patients even in the absence of venous thromboembolic disease, but the magnitude of this elevation is much less than what is observed in patients diagnosed with venous thromboembolic disease. Also, the negative predictive value for D-dimer in patients with cancer and venous thromboembolic disease may vary, depending on the assay used (14)
. Therefore, serial measurements of D-dimer are expected to be more informative in regards to outcome than a single isolated level. The time course of D-dimer levels in the current study was in parallel with thrombin-antithrombin and prothrombin F1 + 2. However, both thrombin-antithrombin and prothrombin F1 + 2 are not routine tests in clinical practice.
Prothrombin fragment F1 + 2 is generated during the conversion of prothrombin to thrombin by activated factor X, whereas the thrombin-antithrombin complex is formed during inhibition of thrombin by antithrombin. Therefore, these two tests reflect activation of the coagulation system and subsequent thrombin release. The upward trend in plasma levels of D-dimer, prothrombin F1 + 2, and thrombin-antithrombin over the sequential time points of anticoagulation correlated strongly with an increase risk of death in this investigation. It is interesting that similar observations on the correlation between activation of blood coagulation, as reflected by increased generation of fibrinopeptide A, and poor outcome of patients with cancer have been reported by Rickles et al. (15)
20 years ago. The persistent elevation of these markers in cancer patients may represent an increasing tumor burden and possibly failure to respond to the therapeutic maneuvers targeted against the underlying malignancy. This is an important observation because the majority of our patients died because of tumor progression rather than bleeding or recurrence. This argument is counterbalanced by the fact that these patients had less opportunity to develop complications of anticoagulation because of their short survival. This is in accordance with the findings from a recent study where death of 90% of patients with venous thromboembolic disease was attributed to cancer progression (16)
.
To address the incidence of recurrence and bleeding in patients with cancer, few studies have been selected for comparison with the current trial (Table 3)
. It is possible that differences in the design and the population of patients in these studies have contributed to some disagreement in the final results. For example, two of the cited studies (10
, 17)
evaluated patients with and without malignancy, and in both studies, patients with genitourinary and gastrointestinal tumors comprised the majority of patients with cancer, whereas in the current study, the predominant tumors were breast and lung. The types of tumors reported in the current series correspond well with the distribution of these cancers in the United States. It is important to note, however, that in accordance with our results, the extent of the underlying malignancy was observed to be a risk factor for both recurrence and bleeding in one study (17)
.
While the preparation of this manuscript was in progress, Lee et al. (16)
reported their interesting observations from a study comparing low molecular weight heparin versus oral anticoagulation for the prevention of recurrence of venous thrombotic disease (CLOT study). The findings from the CLOT study show convincingly that low molecular weight heparin is superior to oral anticoagulation in terms of preventing recurrence of venous thromboembolic disease and with comparable risk of bleeding. The risk of recurrence in patients receiving oral anticoagulation was double the rate observed in patients on low molecular weight heparin (15.7 versus 8%). Because our trial was a single-arm study, it is difficult to compare these 2 studies directly. However, it is important to address the difference in the cohorts in both trials. Specifically, 68% of the patients in the study by Lee et al. (16)
had metastatic disease versus 56% in the current study. This may account for the slightly lower rates of recurrence and mortality in our patients compared with those receiving oral anticoagulation in the CLOT investigation. These differences may have some clinical impact. Although it is clear from the CLOT data that low molecular weight heparin is superior to oral anticoagulation, it is possible that a subset of patients with cancer and venous thromboembolic disease may derive similar benefit from warfarin. These are mainly the patients with early stage tumors and with good performance status. Without clear evidence, it is intriguing to hypothesize that the favorable results obtained with the use of low molecular weight heparin may be, in part, related to a greater influence on coagulation activation.
Because of the obvious difference in the cost of low molecular weight heparin versus warfarin, this issue deserves additional investigation. Finally, in agreement with other studies, our results showed that the majority of clotting and bleeding events in cancer patients occur during the early period of anticoagulation (9, 10, 11)
. Therefore, careful monitoring, especially during this period, is warranted for those patients who are at increased risk for these complications.
In summary, our study investigated both clinical and biochemical factors that predicted complications of oral anticoagulation with warfarin in cancer patients. Because of the exclusion criteria, the current trial did not guarantee enrolling the patients in a consecutive manner. This might have impacted on the estimation of the risk of bleeding and recurrence of venous thromboembolic disease, especially because patients with a performance status >2 were not enrolled. However, the exclusion of some patients should not influence the factors predicting recurrence and outcome because all of the variables were entered in regression analysis. Our study showed that persistent activation of hemostasis, as indicated by in vivo markers, seems to correlate strongly with an increased risk for recurrence and poor outcome. It is not clear whether this is a function of resistant malignancy and failure of anticoagulation or other factors related to cancer development and thrombogenesis. This qualification aside, successful anticoagulation with warfarin cannot be viewed as a binary outcome; for example, resolution of thrombi does not indicate remission of cancer. On the other hand, and in view of our results, strategies designed to maximize effective anticoagulation in patients at risk and the sue of surrogate markers for coagulation activation such as D-dimer, prothrombin F1 + 2, and thrombin-antithrombin should be evaluated in future trials.
|
FOOTNOTES
|
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Sabah Sallah, Louisiana State University Health Sciences Center, 1501 Kings Highway, Shreveport, LA 71103. Phone: (318) 675-4451; Fax: (318) 675-4338; E-mail: asll{at}novonordisk.com
Received 3/ 5/04;
revised 6/ 9/04;
accepted 7/30/04.
|
REFERENCES
|
|---|
- Sallah S, Wan JY, Nguyen NP. Venous thrombosis in patients with solid tumors: determination of frequency and characteristics. Thromb Haemost 2002;87:575-9.[Medline]
- Saphner T, Tormey DC, Gray R. Venous and arterial thrombosis in patients who received adjuvant therapy for breast cancer. J Clin Oncol 1991;9:286-94.[Abstract]
- Rickles FR, Edwards RL. Activation of blood coagulation in cancer: Trousseau’s syndrome revisited. Blood 1983;62:14-31.[Free Full Text]
- Sallah S, Wan JY, Nguyen NP, et al Disseminated intravascular coagulation in solid tumors: clinical and pathologic study. Thromb Haemost 2001;86:828-33.[Medline]
- Falanga A, Levine MN, Consonni R, et al The effect of very-low-dose warfarin on markers of hypercoagulation in metastatic breast cancer: Results from a randomized trial. Thromb Haemost 1998;79:23-7.[Medline]
- Iversen LH, Thorlacius-Ussing O. Relationship of coagulation test abnormalities to tumor burden and post operative DVT in resected colorectal cancer. Thromb Haemost 2002;87:402-8.[Medline]
- Wojtukiewicz MZ, Rucinska M, Zimnoch L, et al Expression of prothrombin fragment 1+2 in cancer tissue as an indicator of local activation of blood coagulation. Thromb Res 2002;97:335-42.
- Weitz LC, Israel VK, Waisman JR, et al Chemotherapy-induced activation of hemostasis: effect of a low molecular weight heparin (dalteparin sodium) on plasma markers of hemostatic activation. Thromb Haemost 2002;88:213-30.[Medline]
- Penning-van Beest FJ, van Meegen E, Rosendaal FR, et al Characteristics of anticoagulant therapy and comorbidity related to over anticoagulation. Thromb Haemost 2001;86:569-74.[Medline]
- Hutten BA, Prins MH, Gent M, et al Incidence of recurrent thromboembolic and bleeding complications among patients with venous thromboembolism in relation to both malignancy and achieved International Normalized Ratio: a retrospective analysis. J Clin Oncol 2000;18:3078-83.[Abstract/Free Full Text]
- Bona RD, Sivjee KY, Hickey AD, et al The efficacy and safety of oral anticoagulation in patients with cancer. Thromb Haemost 1995;74:1055-8.[Medline]
- Van der Meer FJ, Rosendaal FR, Vandenbroucke JP, et al Assessment of a bleeding risk index in two cohorts of patients treated with oral anticoagulants. Thromb Haemost 1996;76:12-6.[Medline]
- ten Wolde M, Kraaijenhagen RA, Prins MH, et al The clinical usefulness of D-dimer testing in cancer patients with suspected deep venous thrombosis. Arch Intern Med 2002;162:1880-4.[Abstract/Free Full Text]
- Lee AY, Julian JA, Levine MN, et al Clinical utility of a rapid whole-blood D-dimer assay in patients with cancer who present with acute deep venous thrombosis. Ann Intern Med 1999;131:417-23.[Abstract/Free Full Text]
- Rickles FR, Edwards RL, Barb C, et al Abnormalities of blood coagulation in patients with cancer: fibrinopeptide A generation and tumor growth. Cancer (Phila) 1983;51:301-7.
- Lee AY, Levine MN, Baker RI, et al Low-molecular weight heparin versus coumadin for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 2003;49:146-53.
- Prandoni P, Lensing AW, Piccioli A, et al Recurrent venous thromboembolism and bleeding complications during anticoagulant treatment in patients with cancer and venous thrombosis. Blood 2002;100:3484-8.[Abstract/Free Full Text]
Top
ABSTRACT
INTRODUCTION
