Venous thromboembolism in women with ovarian cancer undergoing neoadjuvant chemotherapy prior to cytoreductive surgery: A retrospective study

Ovarian cancer is associated with a venous thromboembolism risk of at least 7.2% by 2 years from diagnosis, and although patients undergoing surgery benefit from routine thromboprophylaxis, those undergoing neoadjuvant chemotherapy do not. This study aims to determine the venous thromboembolism incidence in patients with ovarian cancer undergoing neoadjuvant chemotherapy, and explore whether any subset is at higher risk, in order to evaluate whether thromboprophylaxis is justified in some or all of these patients.


| INTRODUC TI ON
Patients with cancer have a high risk of developing venous thromboembolism (VTE), 1 and this incidence has increased over recent years, possibly as the result of increased detection, thrombogenicity of current treatments, and the use of indwelling venous catheters. 2 Chemotherapy is a significant risk factor for VTE, 3 which commonly causes substantial morbidity, including delays to cancer care, 4 and is a leading cause of mortality in ambulatory chemotherapy patients. 5,6 Ovarian cancer has one of the highest VTE incidences among solid tumors, 7 with a 2-year cumulative incidence of 7.2% in a Danish national cohort study. 8 This is likely due to a combination of widespread disease at diagnosis, venous stasis from pelvic masses, and chemotherapeutic agents, all of which are known to promote coagulation. 9,10 Epithelial ovarian cancer typically presents with advanced disease, and management involves either upfront debulking surgery followed by adjuvant chemotherapy, or neoadjuvant chemotherapy (NACT) followed by interval debulking surgery (IDS), which is non-inferior for selected patients. 11 Current practice in many countries is for women undergoing ovarian cancer surgery to have extended thromboprophylaxis for up to 30 days with low-molecular-weight heparin (LMWH). 12,13 This was found to have a number-needed-to-treat to prevent one VTE of 13-39 by a recent systematic review of major abdomino-pelvic cancer surgery. 14 However, there are no such recommendations for thromboprophylaxis for women with ovarian cancer undergoing NACT before IDS.
For women having upfront surgery, this heightened postoperative VTE risk was found to extend into the period of adjuvant chemotherapy, 15 demonstrating the thrombogenicity of this disease across the treatment timeline. Silent or subclinical VTE frequently occurs before surgery, and this may account for the high incidence of VTE detected immediately following ovarian cancer surgery. 16 There is therefore good reason to expect a heightened risk of VTE during NACT for ovarian cancer, which may manifest before or immediately after IDS.
Several studies have looked at thromboprophylaxis in ambulatory cancer patients with a variety of solid tumors. A recent Cochrane review 17 found high-quality evidence that LMWH reduced the risk of any VTE compared with placebo by 43% (relative risk 0.57, 95% CI 0.46-0.71). Factor Xa inhibitors, a class of direct oral anti-coagulants, have been studied in two major relevant placebo-controlled trials: Apixaban in the AVERT trial, which contained 25.8% of patients with ovarian cancer, 18 and Rivaroxaban in the CASSINI trial, which contained a lower proportion of patients with ovarian cancer (6.4%). 19 Each of these trials demonstrated a significant reduction in the rates of VTE in the intention-to-treat population, and pooled analysis found that direct oral anti-coagulants significantly reduced the risk of VTE (5.2% on treatment vs. 9.3% on placebo (relative risk 0.55, 95% CI 0.34-0.90; p = 0.02). 17 However, in order to establish whether one or more of these agents should be offered to ovarian cancer patients during NACT, either routinely or as an ongoing trial, it is necessary to reliably estimate the incidence of VTE in these patients.
The aims of this study were therefore to identify the risk of VTE in patients with Stage III or IV epithelial ovarian cancer undergoing NACT before IDS, and to conduct an exploratory analysis as to whether this entire cohort, or a particular subset based on demographic or tumor characteristics, had an elevated VTE risk. This informs an evaluation of whether these patients may benefit from thromboprophylaxis during NACT.

Key message
This study provides evidence that patients undergoing neoadjuvant chemotherapy for advanced epithelial ovarian cancer face an unacceptably high risk of venous thromboembolism, including pulmonary embolism. Consideration should now be given to strategies to reduce this risk, including evaluation of thromboprophylaxis.

| Statistical analyses
Data were analyzed using SPSS version 23 (IBM). To identify the association between VTE risk and patient demographics or tumor subtype, the odds ratios and 95% CI were calculated. Mann-Whitney U test was used and all analyses were two-sided with a significance level of p = 0.05.

| Demographics
Three hundred and two women were identified; 24 of these were subsequently excluded because of pre-existing anti-coagulation or a lack of sufficient data about VTE outcomes, leaving 278 women suitable for analysis.
Demographic information and tumor characteristics are shown in   to identify a higher-risk subgroup from these results, and we therefore conclude that the entire cohort is at risk.

| DISCUSS ION
To our knowledge, this study constitutes the largest review of VTE risk in patients with ovarian cancer undergoing NACT. Although baseline patient data were collected prospectively our analysis was retrospective, so some data were missing, primarily for those patients before 2010. However, we do not consider that this affected the overall analysis, with a generally high degree of detail available in the individual patient medical records. We were not able to analyze whether laboratory results such as anemia, elevated platelet or leukocyte counts were associated with VTE in our groups. These are known to be associated with VTE in ovarian and other cancers, 20 with reactive thrombocytosis being a common occurrence in ovarian cancer, associated with ascitic interleukin-6 levels, 21 so the omission of these data is a potential limitation.
We did not identify patient characteristics associated with VTE diagnosis in our exploratory analysis, which may be a result of the numbers in our series. Of note, BMI was not associated with VTE in our study, although an association is well known. 22 We speculate that this is due to the low number of patients with recorded raised BMI in our study, along with the possibility that disease and treatment-related prothrombotic mechanisms may outweigh the effect of patient-related factors such as BMI. Patients undergoing NACT have a high (usually unresectable) tumor burden with ascites, both significantly associated with VTE risk, likely due to a combination of reduced blood flow from pelvic venous compression, and the ectopic secretion of procoagulant material including tissue factor, membrane-derived microparticles, and cytokines such as interleukin-6 and tumor necrosis factorα. 9,23 In addition, NACT can lead to VTE via pathways including reduced protein C and S expression (from reduced vitamin K absorption), increased factor V expression, and direct injury to vascular endothelium. 24,25 In our study, clear cell and carcinosarcomas were associated with VTE, although the numbers are small and do not reach statistical significance in the case of clear cell disease. This reflects their at least 2.5-fold greater VTE risk, which is known from larger studies, 23,26 most likely because of increased ectopic production of tissue factor and factor VII. 9 Our data add to results from other smaller retrospective studies in the field. Chavan et al 27  preferably via a randomized controlled trial, to assess the impact on VTE rate and complications, along with a formal cost-benefit analysis.

ACK N OWLED G M ENTS
The authors would like to thank Dr Jane Thomas for her assistance and advice on data analysis.

AUTH O R CO NTR I B UTI O N S
SO contributed to data collection and analysis, manuscript drafting, and critical revision. YA contributed to design of the work, data collection, and manuscript drafting. NS contributed to design of the work and data collection. DH contributed to data collection. MT and ML contributed to critical revision of the manuscript.
AO contributed to conception of the work and to critical revision of the manuscript. All authors approved the final manuscript for publication.