Venous thromboembolism (VTE) is a common complication and the second most common preventable cause of death in cancer patients. Individuals with pancreatic cancer (PC) have been shown to carry the highest risk of VTE of any cancer type [1
], with VTE rates ranging from 5% to 41% in retrospective cohorts [3
], and up to 67% in postmortem series [21
]. In a recent large prospective cohort study of 731 PC patients, VTE occurred in 21% of patients and was associated with significant decreases in both progression-free survival (PFS) and overall survival (OS) [22
Most clinical practice guidelines (CPG) now recommend thromboprophylaxis for ambulatory cancer patients at high risk of VTE, in the absence of contra-indication [23
]. In ambulatory advanced PC patients receiving chemotherapy, the benefit of low molecular weight heparin (LMWH) for preventing VTE has been initially established in two dedicated randomized controlled trials (RCT), demonstrating that LMWH reduces the rate of VTE by 66–85%, without increasing the risk of major bleeding [25
]. Based on these findings, the International Initiative on Cancer and Thrombosis (ITAC) CPG have recommended the use of LMWH in these patients since 2013, provided they have a low risk of bleeding (Grade 1B) [23
]. However, both the burden of VTE and the benefit of primary thromboprophylaxis in PC patients continue to be under-recognized worldwide, and implementation of this CPG remains insufficient.
Recently, the CASSINI trial assessed the efficacy and safety of prophylactic doses of rivaroxaban for preventing VTE in ambulatory cancer patients receiving systemic anticancer therapy who were at intermediate-to-high-risk of VTE (Khorana score ≥ 2 prior to starting chemotherapy) [29
]. A subgroup analysis of PC patients included in the CASSINI phase III trial was subsequently released [29
]. This subgroup analysis increases the size and quality of the dataset available for a meta-analysis assessing the efficacy and safety of anticoagulants for primary thromboprophylaxis in PC patients. Careful comparison of these results with those from previous studies on this topic is warranted since the use of direct oral anticoagulants (DOAC) might improve the net clinical benefit of thromboprophylaxis in PC patients.
Herein, we report a systematic literature review and meta-analysis of all RCTs assessing the benefit of anticoagulants versus placebo or non-placebo control for the prevention of VTE in ambulatory PC patients receiving systemic chemotherapy. We performed sensitivity analyses regarding parenteral versus oral anticoagulants and prophylactic versus supra-prophylactic doses, in order to determine the optimal anticoagulant agent and dosing to be used for primary thromboprophylaxis.
The present meta-analysis pooled data from 1003 PC patients enrolled in 5 RCTs that compared anticoagulants (parenteral or oral) with placebo or no placebo control for primary VTE prevention in cancer patients receiving chemotherapy. Overall, primary thromboprophylaxis with anticoagulants was found to be associated with a 69% relative risk reduction in the rates of VTE without heterogeneity between studies, resulting in a NNT of 11.9 to prevent one VTE event. Primary thromboprophylaxis exhibited a significant net clinical benefit by drastically decreasing the risk of VTE without increasing the risk for major bleeding.
A previous Cochrane meta-analysis which assessed the benefit of primary LMWH thromboprophylaxis in unselected ambulatory cancer patients receiving chemotherapy reported that LMWH significantly reduced the rate of VTE (RR 0.54, 95% CI 0.38–0.75) with a non-statistically significant increase in the risk of major bleeding events (RR 1.44, 95% CI 0.98–2.11) [34
]. In this meta-analysis, the NNT appeared too high to support the use of thromboprophylaxis in all ambulatory cancer patients, due to an overall low rate of VTE events in the study population [34
]. Therefore, it was suggested to rather use targeted thromboprophylaxis in high-risk patients. A risk-stratified approach based on the Khorana score [35
] has been proposed to select patients at high risk of VTE. However, it is questionable whether its use is relevant for PC patients given that the Khorana score assigns +2 points for PC, thereby classifying all PC patients at intermediate risk of VTE, at least. It is therefore not surprising that the Khorana score has low predictive power in determining differences in VTE risk within the PC patient population, both in a small retrospective study of PC patients undergoing chemotherapy [36
], and more recently in the large prospective Base Clinico-Biologique de l’Adénocarcinome Pancréatique [BACAP]-VTE study [5
], the Khorana score failed to accurately predict VTE risk across PC patients. Nevertheless, the overall high incidence of VTE in PC patients has prompted recent CPGs to recommend that primary thromboprophylaxis should be considered in all PC patients undergoing chemotherapy who are at low risk of bleeding [23
Indeed, compared to other cancer patients, PC patients carry a higher risk of VTE [1
], with reported incidence rates of VTE ranging from 5% to 41% in retrospective cohorts, depending on the study population [3
]. In the largest prospective multicenter cohort of patients with newly diagnosed PC to date [22
], VTE occurred in more than 20% of patients, with a median time from PC diagnosis to VTE of 4.49 months, highlighting the need for an adequate thromboprophylaxis scheme in these patients. Moreover, PC patients who developed VTE had significantly shorter PFS and OS compared to those without VTE, even after adjustment for age and cancer stage [22
Our study further confirms the results of a previous meta-analysis [37
] demonstrating a clear benefit of LMWH primary thromboprophylaxis in PC patients with the addition of a subgroup of 273 PC patients enrolled in the CASSINI trial, which recently demonstrated the benefit of rivaroxaban in preventing VTE in intermediate-to-high-risk cancer patients (Khorana score ≥ 2) [29
To our knowledge, this is the first systematic review and meta-analysis assessing the efficacy and safety of thromboprophylaxis in the PC patient population that has included data from a subgroup of PC patients enrolled in a DOAC placebo-controlled trial for primary VTE prevention. The double-blind placebo-controlled CASSINI trial randomized cancer patients initiating chemotherapy and classified as intermediate-to-high-risk of VTE (as defined by a Khorana score ≥ 2) to receive either primary prophylaxis with 10 mg rivaroxaban or placebo, once daily, for up to 6 months [29
]. A pre-specified subgroup analysis of PC patients reported that the primary composite endpoint (deep vein thrombosis [DVT], asymptomatic proximal DVT, pulmonary embolism [PE] and VTE-related death within the first 180 days after randomization) occurred in 5 out of 135 PC patients in the rivaroxaban arm compared to 14 out of 138 (10.1%) patients in the placebo arm, without difference in major bleeding between the two arms [30
]. The cumulative incidence rates of VTE in the placebo and thrombopropylaxis groups at 6 months were consistent with those observed in the PROSPECT-CONKO 004 trial [26
] and in the subgroup analysis of PC patients included in the SAVE-ONCO trial [32
]. In sensitivity analyses, similar reductions in VTE rates were observed when using parenteral (LMWH) or oral anticoagulants (DOAC), suggesting that DOAC might be as efficient as LMWH for VTE prevention in this specific population. However, while parenteral anticoagulants were used in a total of 740 patients included in 4 studies, only a single subgroup of 273 PC patients from a non-PC DOAC placebo-controlled trial was included in our meta-analysis, which may have introduced a bias in the sensitivity analysis. Therefore, these results should be interpreted with caution and more data from dedicated trials are needed to assess the benefit of DOAC in this specific setting and confirm this finding.
Primary thromboprophylaxis with LMWH in advanced or metastatic PC patients receiving systemic anticancer therapy who are at low risk of bleeding (Grade 1B) has been recommended by the ITAC CPGs since 2013 [23
]. However, fear of bleeding in otherwise frail patients, the inconvenience of prolonged parenteral therapy and the inherent costs for such therapy remain major concerns for both patients and physicians. DOAC, if shown to have similar efficacy and safety profiles, can present an alternative to LMWH that would address the potential treatment burden associated with administering daily injections of a parenteral anticoagulant for extended periods, and improve the feasibility of thromboprophylaxis in PC patients. Nonetheless, the use of DOAC in these patients has potential limitations that should be taken into consideration to guide treatment choice. First, data on their efficacy and safety in patients with extreme body weights (i.e., cachexia or obesity) are lacking and the use of DOAC in these patients may result in over- or under-coagulation. Second, DOAC should be used with caution in the elderly who have been shown to be at high risk of bleeding. Third, careful consideration of competing risks, which include presence of comorbidities (e.g., renal or hepatic impairment), drug–drug interactions that may affect DOAC pharmacokinetics and chemotherapy common side effects (vomiting and diarrhea), that can limit DOACs absorption, is warranted. Therefore, an individualized approach is necessary, and for each single PC patient, full consideration of the appropriate balance of benefits and harms is important.
It has been suggested that PC patients receiving chemotherapy may require higher doses of anticoagulant for VTE prevention since dalteparin was administered at therapeutic doses in the FRAGEM trial [25
]. On the other hand, enoxaparin was administered at supra-prophylactic doses in the PROSPECT-CONKO 004 trial [26
]. In the present meta-analysis, similar VTE risk reductions were observed using either therapeutic, prophylactic or supra-prophylactic doses of anticoagulant, suggesting that the use of higher doses of anticoagulant does not improve the net clinical benefit of primary thromboprophylaxis. Once again, these results need to be interpreted with caution since sensitivity analyses of prophylactic doses were performed in subgroups of PC patients from non-PC anticoagulant placebo-controlled trials [29
The optimal duration of primary thromboprophylaxis in PC patients receiving chemotherapy remains an unanswered question. Thromboprophylaxis duration ranged from 3 to 6 months in RCTs included in this meta-analysis and we did not observe any difference in VTE risk reduction between studies using a 3- or 6-month duration of thromboprophylaxis. Whether longer periods of prophylaxis can be of benefit with a similar safety profile remains unknown.
The prevalence of PC is projected to increase by approximately 40% over the next decade in North America and Europe [38
]. Despite recent advancements in the management of PC patients, the prognosis remains poor, with few patients surviving to 10 years [39
]. Therefore, there is an urgent need for optimizing and integrating supportive care, especially VTE prevention, which will improve patient quality of life and, potentially, OS [36
Despite a significant association between VTE and mortality in PC patients, the FRAGEM [25
] and PROSPECT-CONKO 004 [26
] failed to demonstrate a benefit of LMWH on overall survival, which might be related to the short life expectancy of patients included in these studies [41
]. In the FRAGEM trial, 3 out of 4 deaths observed in the placebo arm in the first 3 months of treatment were secondary to VTE, compared to only one death in the dalteparin arm, which was due to sepsis [25
]. A non-randomized trial reported that the use of nadroparin improved survival in 69 consecutive patients with advanced PC treated with Gemcitabine plus cisplatin every 21 days until disease progression. The overall response total response rate was 58.8% with nadroparin compared to 12.1% without nadroparin (p
= 0.0001). Patients receiving nadroparin had longer PFS (7.3 versus 4.0 months, p
= 0.0001) and OS (13.0 versus 5.5 months, p
= 0.0001) compared to those not receiving nadroparin [42
]. Future studies to determine the role of LMWH or DOAC thromboprophylaxis in improving overall cancer-related treatment outcomes are required.
Our study has several limitations inherent to meta-analyses. First, selected studies were heterogeneous in terms of study design, study population, primary thromboprophylaxis modalities (e.g., anticoagulant agent, dosing, schedule, duration of thromboprophylaxis), VTE outcome definitions and length of follow-up. Second, data were derived from subgroup analyses of PC patients in 3 out of the 5 studies included in our meta-analysis. Finally, the analysis was not based on individual data.
In the absence of head-to-head comparison between LMWH and DOACs, it is not yet possible to conclude on the superiority of one agent over the other. Only an adequately powered randomized, double-blind, DOAC-LMWH-controlled, non-inferiority multicenter trial with a 6-month follow-up duration would allow to draw definitive conclusions.