Evaluation of Risk Factors for Venous Thromboembolismin Chinese Women With Epithelial Ovarian Cancer
Xiangxiang Wu, MD,* Xiang Xue, PhD,Þ Jie Tang, MD,* Xi Cheng, MD,* Wenjuan Tian, MD,*Rong Jiang, MD,* and Rongyu Zang, MD*
Objective: Venous thromboembolism (VTE) is a life-threatening complication that oftenoccurs in ovarian tumors. However, the risk factors for VTE are still undetermined.Methods: We retrospectively analyzed VTE occurrence and its potential risk factors in 254Chinese patients with ovarian tumor at Fudan University Cancer Hospital from July 2007 toJune 2011.Results: The VTE incidence was 7.1% (13/183) in epithelial ovarian cancer (EOC), and noVTE was found in ovarian borderline or benign tumor. D-dimer levels were significantlyhigher in EOC than in ovarian benign and borderline tumors. Furthermore, D-dimer levelsincreased with the advancement of EOC stages. Correlation analysis suggested that D-dimerlevels werewell correlated with platelet counting (PLT), prothrombin time (PT), white bloodcell counting (WBC), cancer antigen (CA) 125, and CA153. Univariate logistic regressionanalysis found that D-dimer levels greater than 788 Kg/L, PLT levels greater than 261� 109/L, PT greater than 11.7 seconds, CA125 greater than 760 U/mL, and ascites greater than1500 mL are risk factors for VTE in EOC. Moreover, multivariate analysis grouped primaryEOC, low differentiated grade, D-dimer greater than 788Kg/L, PT greater than 11.7 seconds,and CA125 greater than 760 U/mL as prediction factors for VTE.Conclusions: In addition to D-dimer and ascites, high levels of PLT, PT, and CA125, whichare highly correlated with D-dimer, are independent risk factors for VTE
Received August 27, 2012, and in revised form September 28, 2012.Accepted for publication Sept 30, 2012.
(Int J Gynecol Cancer 2013;23: 65Y72)
V enous thromboembolism (VTE), including deep veinthrombosis (DVT) and pulmonary thromboembolism (PE),
is the most fatal risk factor for ovarian tumors.1 On the other
hand, patients with primary VTE diagnosis are at an increasedrisk for ovarian tumors.2 In a large population study, Rodriguezet al3 reported that theVTE incidencewas 5.2% (672/13,031) inovarian cancer within 24 months after diagnosis. Currently, theinvestigation of predictive risk factors for VTE in ovariantumors is a hot topic.3 High D-dimer levels,4 tissue factor,5
advanced age,6,7 larger body mass index (BMI),7,8 advancedcancer stages, chronic comorbid conditions,3 tumor diameter ofmore than 20 cm,9 and transfusion10 were reported as high riskfactors for the development of VTE in ovarian cancer in dif-ferent studies. However, the incidence and risk factors for VTEin ovarian tumors in Chinese women are still undefined.
D-dimer is a degradation product of fibrin, whichincreases with fibrin formation and degradation; thus, it couldbe used as a sensitive molecular marker for plasmin activity.11
ORIGINAL STUDY
International Journal of Gynecological Cancer & Volume 23, Number 1, January 2013 65
*Ovarian Cancer Program, Department of Gynecologic Oncology,Fudan University Cancer Hospital, Shanghai, China; and †Departmentof Molecular and Integrative Physiology, University of Michigan,Ann Arbor, MI.Address correspondence and reprint requests to Rongyu Zang, MD,
Department of Gynecologic Oncology, Cancer Hospitalof Fudan University, 399 Ling-Ling Rd, Shanghai 200032,China. E-mail: [email protected].
The authors declare no conflicts of interest.Copyright * 2013 by IGCS and ESGOISSN: 1048-891XDOI: 10.1097/IGC.0b013e318276dd87
Furthermore, it is important for maintaining the normal per-meability of blood vessels, flow status of blood, and tissuerepair. The increase of D-dimer indicates the enhancement ofsecond fibrinolysis, and it could be used as a marker for hy-percoagulable state and thromboembolism.12 Plasma D-dimerlevels have been used for VTE screening and evaluating the riskof VTE occurrence with other clinical indices.13,14 Kodamaet al14 analyzed the D-dimer levels in 75 patients with gyne-cologic tumor at preoperation, intraoperation, and post-operation. They found that 21 patients developed VTE, and theD-dimer levels at day 3 postoperation, blood transfusion, andnonYtypeObloodare independent risk factors forVTE.14Satohet al4 found that the levels of D-dimerwere higher than500Kg/Lin 90.2% of patients among 60 cases of ovarian cancer and12 cases of ovarian borderline tumor, 18 cases (25%) devel-oped DVT, and 8 cases (11.1%) developed PE.4 Although theincidences of VTE are quite high in these 2 small-sampleprospective investigations, they are relatively lower in larger-sample studies. Mereu et al6 found that 16 cases (7.8%) ofthromboembolism occurred in a retrospective analysis of 203cases of patients with ovarian cancer who received chemo-therapy for the first time.
Here we retrospectively analyzed and reported the inci-dence of VTE in Chinese patients with ovarian tumor in ourhospital. Furthermore, we evaluated D-dimer levels and furtheranalyzed its relationship with hematologic indices. Moreover,we investigated other potential predictive risk factors forVTE inepithelial ovarian cancer (EOC).We found that, inEOCpatientswith D-dimer greater than 788 Kg/L, platelet counting (PLT)greater than 261 � 109/L, prothrombin time (PT) greater than11.7 seconds, cancer antigen (CA) 125 greater than 760 U/mL,and ascites greater than 1500mL are at high risk for VTE. Thisstudy is the first report to our knowledge that investigated theincidence and risk factors of VTE in Chinese patients withEOC. This should provide useful and relevant information forthe prevention, diagnosis, and treatment ofVTE in patientswithovarian cancer.
MATERIALS AND METHODS
Study Design and Data CollectionWe reviewed general information, clinical pathology,
and follow-up for patients with ovarian tumor who underwentsurgery in our hospital from July 2007 to January 2011.Pathologic changes for patients who transferred from otherhospitals were further confirmed by the department of pa-thology in our hospital. Patients with EOC were categorizedinto different stages according to the pathologic stage stan-dard from the international Federation of Gynecology andObstetrics (FIGO). All cases were grouped into different typesand graded according to a revised standard of the world healthorganization (WHO).
General information (age, height, weight, and medicalcomplications), operation-related information (operation date,operation methods, intraoperative ascites, hemorrhage, andblood transfusion), pathology information (pathologic types,differentiation grades, and FIGO stages), thromboembolismprophylaxisYrelated information (postoperative usage time oflow-molecular-weight heparin sodium injection), detection-
related information (typeB ultrasonic examination forDVTandchest radiograph for PE), and occurrence-related information(confirmation time and clinic symptoms) were recorded. Wealso collected clinical hematologic data such as D-dimer levels,PT,white blood cell counting (WBC), andPLT.Patient follow-upwas carried out at the outpatient department or by telephone.
Statistical AnalysisD-dimer levels among ovarian tumors were compared
by 1-way analysis of variance followed by Dunn multiplecomparison test. Differences of D-dimer levels between pri-mary and recurrent EOC were examined by Student t test.D-dimer levels among different stages of EOC were analyzedby Mann-Whitney U test and linear regression. The asso-ciations of D-dimer levels in ovarian tumors with PLT, WBC,PT, BMI, or age were analyzed by bivariate Pearson corre-lation analysis. For logistic regression analysis, the clinicalcharacteristics of EOC were categorized into continuous orcategorical variables, input into a database, and analyzed withunivariate and multivariate logistic regression models.According to regression coefficients A and odds ratio (OR),variables were attributed to be risk factors for VTE in EOC, ornot. The sensitivity and specificity of risk factors for the VTEoccurrence were analyzed by receiver operator characteristic(ROC) curve. The positive predictive value (PPV) and neg-ative predictive value (NPV) for D-dimer levels were calcu-lated. All these data were analyzed with GraphPad Prism 5.0(GraphPad Software, Inc, La Jolla, Calif ) or SPSS 15.0(International Business Machines Corp, Armonk, NY), andP G 0.05 was considered as statistically significant.
RESULTS
VTE Incidence and Clinic Characteristics inOvarian Tumors
We reviewed clinical information for patients withovarian tumor who had surgery in our hospital during July2007 to January 2011. The follow-ups were carried at June2011, and by that time, at least 5 months from surgery hadpassed for all patients. We successfully followed up 254patients, among whom 72.0% (183/254) had EOC, 31.7%(58/254) had ovarian benign tumor, and 5.1% (13/254) hadovarian borderline tumor. The incidence of VTE in EOCpatients was 7.1% (13/183), among which the incidence ofDVTwas 5.5% (10/183), and the incidence of PE was 1.6%(3/183; Table 1). No VTE was observed in ovarian benign orborderline tumors. We found 35 cases of death within thefollow-up, among which 5 cases (5/13) had VTE and 30 cases(30/170) did not have VTE.
For EOC, most cases (42.1%) were 51 to 60 years old,and most cases (63.9%) had a BMI at 18.6 to 24.9 kg/m2
(Table 1). According to the FIGO postoperative pathologicstage standard, there were 8 (4.4%), 13 (7.1%), 129 (70.5%),and 11 (6%) cases of EOC from stage I to IV, respectively. Theincidences of serous carcinoma, mucinous carcinoma, endo-metrioid carcinoma, clear cell carcinoma, undifferentiated,and other types were 66.1% (121/183), 3.3% (6/183), 7.1%(13/183), 3.8% (7/183), and 18.6% (34/183), respectively.The incidences for medical complication with hypertension,
Wu et al International Journal of Gynecological Cancer & Volume 23, Number 1, January 2013
diabetes, both hypertension and diabetes, and other diseaseswere 16.4% (30/183), 4.4% (8/183), 4.4% (8/183), and 14.7%(27/183), respectively. The percentages for primary and re-current cases were 62.3% (114/183) and 37.7% (69/183),respectively.
D-Dimer Levels in Ovarian TumorsThe mean D-dimer levels were 261.1 Kg/L [95% con-
fidence interval (CI), 194.9Y328.3 Kg/L], 173.8 Kg/L (95%CI, 110.0Y236.5 Kg/L), and 539.8 Kg/L (95% CI, 469.0Y610.7 Kg/L) for ovarian benign tumor, ovarian borderlinetumor, and EOC, respectively. One-way analysis followed byDunn multiple comparison test showed that the D-dimer levelsin EOC were significantly higher than ovarian benign andovarian borderline tumors (Fig. 1A). Furthermore, D-dimerlevels in primary EOC (mean, 696.6 Kg/L; 95% CI,602.4Y790.7 Kg/L) were significantly higher than recurrentEOC (mean, 280.3 Kg/L; 95% CI, 209.7Y350.9 Kg/L;Fig. 1B). The mean D-dimer levels were 129.1 Kg/L (95% CI,74.24Y184.0 Kg/L), 433.2 Kg/L (95% CI, 272.0Y594.3 Kg/L),615.4 Kg/L (95% CI, 514.8Y715.9 Kg/L), and 973.9 Kg/L(95% CI, 574.0Y1374 Kg/L) from stage I to IV EOC, re-spectively. One-way analysis of variance analysis followed byDunn multiple comparison test showed that D-dimer levels forpatients with stage I EOC were significantly lower than thosewith stage II and III (Fig. 1C). Furthermore, the D-dimer levelsfor stage IV EOC were significantly higher than that for stageIII EOC. Moreover, linear regression showed that the D-dimerlevels were highly correlated with EOC stages (P = 0.0070;Fig. 1D).
For EOC,we further analyzed the correlation of the D-dimerlevels with some clinical characteristics. We found that D-dimerlevels were tightly correlated with PLT (Pearson r= 0.5245, P G0.0001), WBC (Pearson r = 0.4522, P G 0.0001), and CA125(Pearson r = 0.4117, P G 0.0001) by bivariate Pearson correlationanalysis (Fig. 2AYC). They were also correlated with CA153(Pearson r = 0.3664,P = 0.0237) and PT (Pearson r = 0.2280,P =0.0043; Fig. 2D, E), but they had no correlation with CA724(Pearson r= 0.2909,P= 0.0650), CA199 (Pearson r=j0.01308,P = 0.8858), age (Pearson r = 0.04378, P = 0.5719), BMI(Pearson r = j0.1060, P = 0.1740), and hemoglobin(Pearson r = j0.0691, P = 0.3722; data not shown).
Logistic Regression Analysis of Risk Factorsfor VTE in EOC
We categorized clinical characteristics including he-matologic indices and operative factors from 183 cases ofEOC into continuous or categorical variables, generated theROC curve with VTE occurrence as state variable and otherclinic characteristics as test variables, and assessed the pre-diction value of different variables for VTE occurrence by thearea under curve (AUC) value. We found that variables withAUC of greater than 0.5 indicating high risks for VTE werePT [mean (SD), 0.693 (0.098)], CA125 [mean (SD), 0.653(0.088)], D-dimer [mean (SD), 0.609 (0.105)], PLT [mean(SD), 0.577 (0.078)], and ascites [mean (SD), 0.555 (0.104);Table 2). According to ROC curve, when PT threshold was11.7 seconds, the sensitivity was 72.7%, the specificity was67.2%, PPV was 14.5%, and NPV was 97.0%; when CA125
TABLE 1. Clinic characteristics for patients with ovarian cancer (total cases n = 183)
threshold was 760 U/mL, the sensitivity was 72.7%, thespecificity was 61.2%, PPVwas 12.5%, and NPVwas 96.7%;when D-dimer threshold was 788 Kg/L, the sensitivity was54.5%, the specificity was 76.9%, PPV was 15.3%, and NPVwas 95.7%; when PLT threshold was 261 �109/L, the sen-sitivity was 72.7%, the specificity was 53.0%, PPVwas 10.6%, and NPV was 96.2%; when ascites thresholdwas 1500 mL, the sensitivity was 45.4%, the specificity was79.8%, PPV was 14.7%, and NPV was 95.0% (Table 2).
After finding out that the variable corresponded to thesmallest value of the sum for specificity square and sensitivitysquare that is closest to the upper left corner on the ROCcurve, we grouped different consecutive variables. Throughsubstituting these variables into logistic regression analysiswith enter method, we found that when D-dimer was greaterthan 788 Kg/L, regression index A was 1.249, the OR was3.49 (95% CI, 1.06Y11.49; P = 0.040), and the incidence ofVTE was 6 (14.6%) of 41; when PLTwas greater than 261 �109/L, regression index A was 1.440, the OR was 4.22 (95%CI, 1.12Y15.89; P = 0.033), and the incidence of VTE was 10(11.8%) of 85; when PT was greater than 11.7 seconds, re-gression index A was 1.736, the OR was 5.67 (95% CI,1.48Y21.81; P = 0.012), and the incidence of VTE was 9(14.1%) of 64; when CA125 was greater than 760 U/mL,regression index A was 1.539, the OR was 4.67 (95% CI,1.21Y17.96; P = 0.025), and the incidence of VTE was 9(13.8%) of 65; when ascites was greater than 1500 mL, re-gression index A was 1.510, the OR was 4.53 (95% CI,1.43Y14.4; P = 0.010), and the incidence of VTE was 7(17.5%) of 40. These data show that D-dimer greater than
788Kg/L, PLT greater than 261� 109/L, PT greater than 11.7seconds, CA125 greater than 760 U/mL, and intraoperativeascites greater than 1500 mL were high risk factors for VTEoccurrence (Table 3). There was no significant differenceobserved in other clinical characteristics such as pathologictypes and operative factors.
Multivariate regression analysis revealed that the ORfor primary EOC, low-differentiated cancer, D-dimer levelsgreater than 788 Kg/L, PT greater than 11.7 seconds, andCA125 greater than 760 U/mL were 17.67 (95% CI,1.15Y206.39; P = 0.026), 8.44 (95% CI, 1.10Y64.39; P =0.040), 9.83 (95% CI, 1.30Y74.50; P = 0.027), 7.63 (95% CI,1.23Y47.12; P = 0.029), and 21.88 (95%CI, 1.41Y330.12; P =0.026), respectively (Table 4).
DISCUSSIONVenous thromboembolism is a common complication
after ovarian tumor surgery. The relationship between D-dimerand the VTE occurrence is a hot topic. However, it is in-consistent about the threshold value of D-dimer for VTEprediction and the best time for D-dimer detection. It is alsoinconsistent about the risk factors for VTE in EOC, which hasa high incidence of VTE. In addition, there are great dis-crepancies among different areas and populations for theoccurrence, prevention, diagnosis, and treatment of VTE. Inthe present study, we collected 183 cases of EOC, 58 cases ofovarian benign tumor, and 13 cases of ovarian borderlinetumor from Fudan University Cancer Hospital, analyzed theirD-dimer levels and related clinical characteristics, and eval-uated the risk factors for VTE in Chinese women with EOC.
FIGURE 1. D-dimer levels in Chinese women with ovarian tumors. A, D-dimer levels in ovarian benign tumor, ovarianborderline tumor, and EOC. B, D-dimer levels in primary and recurrent EOC. C, D-dimer levels in different stages ofEOC. D, Linear regression analysis of D-dimer levels with different stages in EOC. *P G 0.05, **P G 0.01, ***P G 0.001.
Wu et al International Journal of Gynecological Cancer & Volume 23, Number 1, January 2013
In our study, the D-dimer levels in EOC were higherthan ovarian benign and borderline tumors. This is consistentwith a previous report that the plasma D-dimer levels arehigher in 56 cases of ovarian epithelial cancer than those in 65cases with ovarian benign tumor.13 We also found that theD-dimer levels increase with the advancement of EOC stages,which is consistent with previous reports.4,13 Interestingly, wefound that D-dimer levels in EOC were highly correlated with
PLT, WBC, CA125, CA153, and PT. However, there is nosuch correlation in ovarian benign tumor and ovarian bor-derline tumor (data not shown). This could be related to thehigh coagulation status in EOC and indicates that D-dimerlevel is highly correlated with the anticoagulation system.15
It is reported that PLT levels were increased in patientswith cancer and the interaction was increased between PLT,tumor monocytes, and the vasculature.16,17 In a multicenter
FIGURE 2. Correlation analysis of D-dimer level with hematologic index in Chinese women with EOC. A, PLT,(B) WBC, (C) CA125, (D) CA153 and (E) PT.
TABLE 2. The AUC values, sensitivity, specificity, PPV, and NPV of risk factors for thrombosis in ovarian cancer
Risk factors Cutoff value AUC, mean (SD) Sensitivity, % Specificity, % PPV, % NPV, %
retrospective study, Khorana et al17 showed that the VTEincidence within 2.5 months of follow-up was 4% for patientswith cancer with PLT of 350,000/KL or greater before che-motherapy, whereas the VTE incidence was only 1.25% forpatients with PLT of less than 200,000/KL (P for trend =0.0003). In addition, patients with VTE had a higher PLT thanthose patients without VTE before every chemotherapycourse (P G 0.001). When expanding their analysis, WBCgreater than 11,000/KL before chemotherapy was also inde-pendently correlated with increased VTE.18 By multivariateanalysis, increased basal PLT (OR, 1.8; 95% CI, 1.1Y3.2) andWBC (OR, 2.2; 95% CI, 1.2Y4.0) were independently cor-related with VTE. This was confirmed by some other
reports.19,20 Thus, the D-dimer level and PLT are correlatedwith VTE occurrence.
By logistic regression analysis for factors including generalcharacteristics, pathology types and stages, and operation-relatedfactors, we found that D-dimer levels greater than 788 Kg/L,PLT levels greater than 261 � 109/L, PT greater than11.7 seconds, CA125 greater than 760 U/mL, and ascitesgreater than 1500mL are independent risk factors for VTE inEOC. After analysis of D-dimer levels in 60 cases EOC and12 cases ovarian borderline tumor, Satoh et al21 reportedthat, when the D-dimer threshold was 1500 Kg/L, the sen-sitivity and NPV were 100%, but the specificity was only47.2%. By contrast, with our D-dimer threshold (788 Kg/L),
not only was our NPV high (95.7%), but the specificity(76.9%) was also relatively higher. Previously, PT was in-creased in patients with breast cancer compared with patientswithout cancer before chemotherapy and was altered bychemotherapy as early as within 24 hours in patients sub-sequently developing VTE.22 Here we clearly determinedthat the cutoff value for VTE in EOC was 11.7 seconds.Gadducci et al23 reported that plasma levels of D-dimer andCA125 are higher in 56 patients with ovarian cancer than in65 ovarian benign tumors. The combination of D-dimer andCA125 was shown to be useful for differentiating benignfrommalignant ovarian tumors.23 Here we found that CA125could also be useful for VTE prediction. Massive ascites, as asign of peritoneal carcinomatosis and advanced tumor dis-ease, was identified as an independent predictor for VTEbefore.24,25 This was further confirmed in our study.
CONCLUSIONSIn the present study, we found that the D-dimer levels in
EOC were higher than ovarian benign and borderline tumors.The D-dimer levels were higher in advanced stages of EOC.D-dimer levels in EOC were highly correlated with PLT, WBC,PT, CA125, and CA153. D-dimer greater than 788 Kg/L, PLTgreater than 261� 109/L, PT greater than 11.7 seconds, CA125greater than 760 U/mL, and intraoperative ascites greater than1500 mL were high risk factors for VTE occurrence.
ACKNOWLEDGMENTWe thank Erik Anderson for proofreading the
manuscript.
REFERENCES1. Einstein MH, Pritts EA, Hartenbach EM. Venous
thromboembolism prevention in gynecologic cancer surgery: asystematic review. Gynecol Oncol. 2007;105:813Y819.
2. Murchison JT, Wylie L, Stockton DL. Excess risk of cancer inpatients with primary venous thromboembolism: a national,population-based cohort study. Br J Cancer. 2004;91:92Y95.
3. Rodriguez AO, Wun T, Chew H, et al. Venousthromboembolism in ovarian cancer. Gynecol Oncol.2007;105:784Y790.
4. Satoh T, Oki A, Uno K, et al. High incidence of silent venousthromboembolism before treatment in ovarian cancer. Br JCancer. 2007;97:1053Y1057.
5. Uno K, Homma S, Satoh T, et al. Tissue factor expression as apossible determinant of thromboembolism in ovarian cancer.Br J Cancer. 2007;96:290Y295.
6. Mereu L, Tateo S, Klersy C, et al. Stratification of venousthromboembolism risk in ovarian cancer patients duringchemotherapy. Int J Gynecol Cancer. 2009;19:79Y83.
7. Fotopoulou C, duBois A, Karavas AN, et al. Incidence of venousthromboembolism in patients with ovarian cancer undergoingplatinum-/paclitaxel-containing first-line chemotherapy: anexploratory analysis by the ArbeitsgemeinschaftGynaekologische Onkologie Ovarian Cancer Study Group.J Clin Oncol. 2008;26:2683Y2689.
8. Tateo S, Mereu L, Salamano S, et al. Ovarian cancer and venousthromboembolic risk. Gynecol Oncol. 2005;99:119Y125.
9. Suzuki N, Yoshioka N, Ohara T, et al. Risk factors forperioperative venous thromboembolism: a retrospective study inJapanese women with gynecologic diseases. Thromb J.2010;8:17.
10. Abu-Rustum NR, Richard S, Wilton A, et al. Transfusionutilization during adnexal or peritoneal cancer surgery: effectson symptomatic venous thromboembolism and survival.Gynecol Oncol. 2005;99:320Y326.
11. Adam SS, Key NS, Greenberg CS. D-dimer antigen: currentconcepts and future prospects. Blood. 2009;113:2878Y2887.
12. Blackwell K, Haroon Z, Broadwater G, et al. Plasma D-dimerlevels in operable breast cancer patients correlate with clinicalstage and axillary lymph node status. J Clin Oncol.2000;18:600Y608.
13. Gadducci A, Baicchi U, Marrai R, et al. Preoperative evaluationof D-dimer and CA 125 levels in differentiating benign frommalignant ovarian masses. Gynecol Oncol. 1996;60:197Y202.
14. Kodama J, Seki N, Masahiro S, et al. D-dimer level as a riskfactor for postoperative venous thromboembolism in Japanesewomen with gynecologic cancer. Ann Oncol. 2010;21:1651Y1656.
15. Acosta S, Alhadad A, Svensson P, et al. Epidemiology, risk andprognostic factors in mesenteric venous thrombosis. Br J Surg.2008;95:1245Y1251.
16. Zwicker JI, Furie BC, Furie B. Cancer-associated thrombosis.Crit Rev Oncol Hematol. 2007;62:126Y136.
17. Khorana AA, Francis CW, Culakova E, et al. Risk factors forchemotherapy-associated venous thromboembolism in aprospective observational study. Cancer. 2005;104:2822Y2829.
18. Khorana AA, Kuderer NM, Culakova E, et al. Development andvalidation of a predictive model for chemotherapy-associatedthrombosis. Blood. 2008;111:4902Y4907.
19. Trujillo-Santos J, Di Micco P, Iannuzzo M, et al. Elevated whiteblood cell count and outcome in cancer patients with venousthromboembolism. Findings from the RIETE Registry. ThrombHaemost. 2008;100:905Y911.
20. Tefferi A, Gangat N, Wolanskyj A. The interaction betweenleukocytosis and other risk factors for thrombosis in essentialthrombocythemia. Blood. 2007;109:4105.
21. Satoh T, Matsumoto K, Uno K, et al. Silent venousthromboembolism before treatment in endometrial cancer andthe risk factors. Br J Cancer. 2008;99:1034Y1039.
22. Kirwan CC, McDowell G, McCollum CN, et al. Early changesin the haemostatic and procoagulant systems after chemotherapyfor breast cancer. Br J Cancer. 2008;99:1000Y1006.
23. Gadducci A, Marrai R, Baicchi U, et al. The measurement ofplasma D-dimer (DD) levels in the follow-up of patients withovarian cancer. Anticancer Res. 1995;15:2683Y2686.
24. Maksimovic M, Gojnic M, Maksimovic Z, et al. Surgicaltreatment of ovarian cancer and early detection of venousthromboembolism. Eur J Gynaecol Oncol. 2011;32:415Y418.
25. Fotopoulou C, Karavas A, Trappe R, et al. Venousthromboembolism in recurrent ovarian cancer-patients: asystematic evaluation of the North-Eastern German Society ofGynaecologic Oncology Ovarian Cancer Study Group(NOGGO). Thromb Res. 2009;124:531Y535.
Wu et al International Journal of Gynecological Cancer & Volume 23, Number 1, January 2013
