Meta-analysis of randomized controlled trials on primary ambulatory thromboprophylaxis in patients with ovarian cancer receiving chemotherapy

Abstract

Ovarian cancer (OC) is highly associated with venous thromboembolism (VTE). The OC cells stimulate thrombin generation, and chemotherapy potentiates the prothrombotic effect of cancer cells by damaging endothelium and enhancing hypercoagulability. Recently, primary ambulatory thromboprophylaxis (PATP) has been studied as a potential treatment in cancer patients undergoing chemotherapy with an aim of reducing the incidence of VTE and potentially prolonging survival. A meta-analysis was performed of randomized controlled trials of PATP vs control in patients with OC receiving chemotherapy. The primary outcome measure was the incidence of VTE. The secondary outcome measure was the incidence of major bleeding complications. Two articles published between 2012 and 2020 fulfilled selection criteria. The incidence of VTE was 0.9% in the PATP group and 1.8% in the control group. However, the pooled risk ratio was not statistically significant at 0.69 (95% CI: 0.08 to 5.67; P = 0.73). The absolute risk difference was −0.03 (95% CI, −0.17 to 0.11; P = 0.66). There was no statistically significant reduction in VTE by providing PATP to patients with OC receiving chemotherapy. Routine PATP should not be recommended in ambulatory OC patients. Future randomized trials are necessary to define the high-risk subset of OC patients who may benefit from PATP.

Ovarian cancer (OC) is the seventh most common cancer diagnosed in women in the world.¹ Up to 20% of women with OC develop venous thromboembolism (VTE) in their lifetime.² The risk of thrombosis is determined by several factors such as specific cancer type, cancer stage, types of cancer treatments, central venous catheters, comorbidities, and patient characteristics.³ Thrombogenesis in cancer patients is multifactorial. Cancer cells stimulate thrombin generation and other prothrombotic factors. Furthermore, chemotherapy augments the procoagulant effect of cancer cells as well as endothelial damage.⁴ In the last decade, primary ambulatory thromboprophylaxis (PATP) has become an area of interest. Thus, we performed a systematic review and meta-analysis of randomized controlled trials (RCTs) to determine the outcome of PATP in OC patients receiving chemotherapy.

METHODS

The systematic review was conducted based on the Cochrane Handbook for Systematic Reviews and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.^5,6 The terms ‘thromboprophylaxis’ OR ‘anticoagulation’ OR ‘low-molecular-weight heparin (LMWH)’ OR ‘direct oral anticoagulants (DOAC)’ AND ‘ovarian cancer’ were included in our search strategy through the systematic review of MEDLINE and EMBASE databases and international conference meeting abstracts until September 30, 2021. Our database search was limited to humans and controlled clinical trials, and RCTs published in English were retrieved. The titles and abstracts of the relevant citations were evaluated to include all potential studies pertinent to our topic through the references of research articles.

To be included for the meta-analysis, the studies had to conform with the following characteristics: RCTs comparing PATP with LMWHs or DOACs vs a control group; RCTs of ambulatory patients who received chemotherapy for OC; and RCTs with reduction in VTE as a primary or secondary endpoint and major hemorrhage as a safety outcome.

Two authors (T.W.H. and K.Z.T.) independently performed data extraction from the eligible studies. The following data were collected: first author’s last name, publication year, study title and type, its primary and secondary outcomes, number of VTE events, bleeding events, study drug and dosage and duration, and number of patients included in each arm. Biases in each study were identified by using the tool recommended by the Cochrane Collaboration. The two authors independently assessed the methodological quality of each study using a standard quality assessment form. Discrepancies were resolved by consensus or by consultation with the senior author (T.H.O.). Potential biases were categorized into five main types: selection bias, performance bias, detection bias, attrition bias, and reporting bias and others and were rated as low, high, or unclear risk.^5,7 Publication bias was evaluated by Cochrane RevMan 5.3 software and assessed by funnel plots.

We used the Mantel-Haenszel method and Review Manager version 5.3 (Nordic Cochrane Centre, Copenhagen, Denmark) to calculate the pooled risk ratio and risk difference with 95% confidence interval (CI) for VTE events and bleeding. Heterogeneity was assessed with I² and Cochran’s Q statistic. A P value <0.05 was considered significant and I² > 50% was considered substantially heterogeneous. A risk ratio <1.0 favored PATP.

RESULTS

A total of 6112 potential references were identified, and 3874 duplicates were removed. After exclusions, 22 records identified from the database were assessed for eligibility for inclusion in our study. Two RCTs—SAVE-ONCO (Evaluation of AVE5026 in the Prevention of Venous Thromboembolism in Patients Undergoing Chemotherapy) and CASSINI (Study to Evaluate the Efficacy and Safety of Rivaroxaban Venous Thromboembolism Prophylaxis in Ambulatory Cancer Participants)—with 433 patients were eligible for inclusion in the final analysis^8,9 (Figure 1).

Figure 1.

Study flow diagram in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA).

The characteristics of the included studies are summarized in Table 1. A total of 433 patients with OC from subsets of two RCTs (SAVE-ONCO and CASSINI) were included in our meta-analysis.^8,9 The SAVE-ONCO trial included unselected cancer patients, whereas the CASSINI trial included cancer patients with a Khorana score (KS) ≥ 2. Prophylactic doses of anticoagulants were utilized—semuloparin in SAVE-ONCO and rivaroxaban in CASSINI. The duration of PATP ranged from 3.5 to 6 months. The randomization ratio was 1:1 in both studies.

Table 1.

Characteristics of the studies included in the meta-analysis^a

Study	Study type	Patients (n)		Anticoagulant (dose and duration)	Primary efficacy outcome measure	VTE events (n)
		Control	AC			Control	AC
SAVE-ONCO, 20128	Randomized, double-blind, multicenter trial	191	188	Semuloparin 20 mg/day for a median of 3.5 months	Reduction in VTE	1	0
CASSINI, 20199	Randomized, double-blind, randomized trial	24	30	Rivaroxaban 10 mg/day for 180 days	Reduction in VTE	1	4

AC indicates anticoagulation; VTE, venous thromboembolism.

^a Both studies were for ovarian cancer; data on type of chemotherapy was unavailable.

Both studies showed a low risk of bias (Figure 2). We found no publication bias. The I² statistic showed moderate heterogeneity among RCTs, and a random effects model was applied. In OC patients, VTE events occurred in 2 patients (0.9%) who had received PATP compared to 4 patients (1.8%) in the control group. The pooled risk ratio for VTE was not statistically significant at 0.69 (95% CI, 0.08 to 5.67; P = 0.73) (Figure 3). The absolute risk difference for VTE occurrence was −0.03 (95% CI, −0.17 to 0.11; P =  0.66) (Figure 4). Neither RCT included specific data for safety outcomes (bleeding events) in OC.

Figure 2.

Risk of bias summary.

Figure 3.

Meta-analysis forest plot: Pooled risk ratio for venous thromboembolism in patients with ovarian cancer receiving primary ambulatory thromboprophylaxis vs control.

Figure 4.

Meta-analysis forest plot: Pooled risk difference for venous thromboembolism in patients with ovarian cancer receiving primary ambulatory thromboprophylaxis vs control.

DISCUSSION

In the California Cancer Registry, which reviewed over 13,000 patients with OC, 5% were found to have VTE within 2 years of diagnosis. Development of VTE within 2 years had a significant adverse impact on survival.^10–12 Indeed, the cumulative incidence depends upon the stage of OC, ranging from 1.4% in patients with local disease up to 6.7% in those with advanced disease. In a recent prospective study, 110 patients with OC obtained a venous ultrasound prior to cancer treatment and, interestingly, the investigators noted a high incidence of deep vein thrombosis (22.7%), where 1.8% had concomitant pulmonary embolism (PE).² OC was hence denoted a highly thrombogenic neoplasm with reference to the KS and assigned 1 point in the outpatient chemotherapy setting.

VTE is the second leading cause of death in patients.^10–12 The 1-year survival of cancer patients diagnosed with VTE was 12%, whereas patients without VTE had a 36% survival.¹¹ Six-month mortality rates after VTE are three times higher for patients with cancer than for those without.¹² Furthermore, anticoagulation is frequently challenging and may potentially complicate anticancer regimens. Primary thromboprophylaxis is routinely recommended in hospitalized cancer patients and those undergoing major surgery. Nonetheless, previous trials showed a modest benefit of PATP in ambulatory cancer patients who were receiving chemotherapy.¹³

Guidelines of a recent international initiative on thrombosis and cancer recommend PATP with DOACs or LMWHs for those who are on systemic anticancer therapy at an intermediate to high risk of VTE, identified by cancer type or by a validated risk assessment model (i.e., KS ≥ 2). On the other hand, guidelines of the American Society of Clinical Oncology take a cautious approach to PATP, indicating that physicians “may offer thromboprophylaxis.”^14,15 Although OC carries a high thrombogenic potential, our meta-analysis did not demonstrate a statistically significant reduction in VTE by PATP in patients with OC who are receiving chemotherapy. Moreover, the clinical utility of KS was recently challenged in some solid tumors, including gynecologic cancers. Fleming et al showed that a high-risk KS failed to accurately predict VTE occurrence in over 700 patients with advanced OC who underwent either neoadjuvant chemotherapy or primary debulking surgery.¹⁶ Furthermore, another large retrospective analysis indicated that a KS of 2 or 3 was not statistically significantly associated with a higher risk of VTE compared to patients with a score of 1, raising concern about recent guidelines recommending PATP for patients with an intermediate to high risk of VTE (i.e., KS ≥ 2) undergoing systemic anticancer therapy.^14,15,17

The major strength of our study was that we used strict, predefined inclusion criteria and included only RCTs in our analysis, leaving out retrospective and prospective cohort studies. One of the limitations of our study was that we incorporated studies with two different anticoagulants—LMWH in SAVE-ONCO and DOAC in CASSINI—which might contribute confounding variables. Another limitation is that the SAVE-ONCO trial included unselected cancer patients, while the CASSINI trial enrolled cancer patients with a KS ≥ 2. Notably, the VTE incidence in patients with OC was low in both PATP and control arms in those two studies. In the SAVE-ONCO trial, symptomatic deep vein thrombosis and PE as well as incidental PE detected during tumor evaluation were included in the analysis of primary efficacy outcome.⁸ On the other hand, in the CASSINI trial, mandatory bilateral lower extremity duplex compression ultrasound studies were performed at week 8, week 16, and day 180. Both symptomatic and asymptomatic deep vein thrombosis as well as incidental PE detected on tumor restaging were included in the analysis of the primary efficacy outcome.^8,9 In both trials, the exact numbers for metastatic disease vs early disease for OC patients were not mentioned.^8,9 The differences in patient inclusion criteria or background patient characteristics may explain the differences in the outcomes of the studies. The last limitation is the lack of data availability on safety (bleeding events) and survival of the patients in both trials.

Future larger RCTs are necessary to classify the subset at high risk of developing VTE and determine potential strategies to significantly reduce VTE events, ultimately leading to better survival. At present, PATP should not be recommended routinely in ambulatory OC patients.