Direct Oral Anticoagulants versus Low-molecular-weight Heparin in Prevention of Venous Thromboembolism Recurrence in Cancer Patients: A Systematic Review and Meta-analysis

Venous thromboembolism (VTE) is a common and potentially fatal complication in cancer patients. Patients with cancer are at a higher risk of developing VTE due to several factors, including malignancy-related factors such as surgery, chemotherapy and immobility^1–3 The incidence of VTE in cancer patients is estimated to be four to seven times higher compared to patients without cancer⁴ Despite improvements in treatment modalities for cancer, the risk of VTE has increased, with patients with tumours of the pancreas, stomach and brain having the highest risk. The prevalence of cancer-associated VTE has increased due to multiple factors, including longer patient survival, anti-cancer therapies, increased detection of incidental VTE during surveillance imaging and wider use of central venous catheters³

Direct oral anticoagulants (DOACs) are a newer class of anticoagulants that have been shown to be as effective as low-molecular-weight heparin (LMWH) for the prevention and treatment of VTE in cancer patients. Several randomised controlled trials (RCTs) have demonstrated that DOACs are non-inferior to LMWH in terms of efficacy and safety outcomes, including recurrence of VTE and major bleeding^3,5–7 Additionally, DOACs have several practical advantages over LMWH, including oral administration, no need for routine laboratory monitoring and a lower risk of heparin-induced thrombocytopenia. However, there are also limitations to the use of DOACs, such as the lack of a specific antidote, the potential for drug interactions, and the need for dose adjustments in patients with renal impairment. Overall, the availability of DOACs has expanded the options for VTE prophylaxis and treatment in cancer patients, but careful consideration of patient factors and preferences is necessary to determine the optimal therapy^7,8

The researchers performed a systematic review and meta-analysis to compare the effectiveness of DOACs and LMWH in the prevention of VTE recurrence among cancer patients.

Aim

The aim was to conduct a systematic review and meta-analysis on RCTs of DOACs on patients with cancer-associated VTE.

The primary outcome investigated was VTE recurrence after treatment with DOACs. Secondary outcomes included major bleeding (MB), clinically relevant non-major bleeding (CRNMB) and all-cause mortality after treatment with DOACs. VTE recurrence is defined as the reappearance of deep vein thrombosis (DVT), pulmonary embolism (PE) or arterial thromboembolism, confirmed through autopsy or imaging modalities such as duplex ultrasonography, venography, CT, MRI, conventional pulmonary angiography or ventilation-perfusion (VQ) scanning. MB was defined as overt bleeding with a haemoglobin drop of 2 g/dl or more, the need for transfusion of at least two units of packed red blood cells, or bleeding occurring in critical locations such as intracranial, intraspinal/epidural, intraocular, retroperitoneal, pericardial, intraarticular or intramuscular with compartment syndrome, or any bleeding resulting in death⁹ Clinically relevant non-major bleeding (CRNMB) was defined as overt bleeding that does not qualify as MB but requires medical intervention, leads to unscheduled contact with healthcare providers, or necessitates temporary cessation of anticoagulant therapy¹⁰

Materials and Methods

Search Strategy

The systematic review and meta-analysis were conducted in accordance with the updated Preferred Reporting for Systematic Review and Meta-Analysis (PRISMA) guidelines.

A comprehensive systematic search of PubMed, Cochrane and Google Scholar was performed from inception to April 2023. The search strategies and keywords were (“new oral anticoagulants” or “DOAC” or “direct oral anticoagulants” or “DOAC” or “dabigatran” or “edoxaban” or “apixaban” or “Rivaroxaban”) AND (“low molecular weight heparin” or “LMWH” or “enoxaparin” or “dalteparin”) AND (“venous thromboembolism” or “VTE” or “venous thrombosis” or “deep venous thrombosis” or “DVT” or “pulmonary embolism” or “PE”) AND (“cancer” OR “adenocarcinoma” OR “carcinoma” or “leukemia” or “lymphoma”).

Study Eligibility

Two reviewers (WBO and PBN) independently searched the databases to identify all potentially eligible studies and reviewed the full articles for inclusion. Only RCTs were included in this review. Studies on patients with cancer-associated VTE were included. Studies were considered potentially eligible for the meta-analysis if the following inclusion criteria were met: participants were adult patients aged 18 years and older with cancer-associated VTE; intervention was anticoagulant therapy with DOACs (dabigatran, apixaban, edoxaban or rivaroxaban); control was anticoagulant therapy with LMWH (enoxaparin, dalteparin). Clinical trials with participants with the following criteria were excluded from the review: platelet count <100,000/μl; creatinine clearance <30 ml/min/1.73 m²; active bleeding or with high risk of bleeding; pregnancy or breastfeeding; severe liver disease. Studies that compared DOACs with medications other than LMWH or placebo were excluded.

Study Selection

Selected articles were then compared and the decision to include the article was reached through a consensus. In cases where a consensus could not be achieved, the third author (FMR) was consulted. Only manuscripts with extractable primary data were included in the final analysis.

Data Extraction and Risk of Bias Assessment

A data collection form was created and used to extract information from each article. This included the author, demographics of the study population, inclusion and exclusion criteria, intervention and comparison interventions, primary outcome of recurrence of VTE and secondary outcomes of MB, CRNMB and all-cause mortality. Quality assessment was done using the Cochrane risk of bias tool. Each article was critically appraised and graded as high, low or unclear. Discrepancies were settled through discussion and consensus.

Data Synthesis and Analysis

Data synthesis and analysis were performed using RevMan 5.4 (Cochrane Collaboration) for computational analysis of outcomes in the study. Mantel-Haenszel statistical method was used for pooled RR, which were calculated for dichotomous outcomes. RR with 95% CI and p-value <0.05 were considered significant. The intention-to-treat principle was followed whenever possible for all analyses. Forest plots were used to present these results.

Visual inspection of the forest plots and I² statistics was used to assess heterogeneity of the study. Statistical calculation was performed using RevMan 5.4. I² values of 25%, 26–50% and >50% were classified as low, moderate and high statistical heterogeneity, respectively. A random effects model was used for moderate-to-high statistical heterogeneity, while a fixed effect model was used for low statistical heterogeneity.

A subgroup analysis was conducted separately for the site of cancer, whether gastrointestinal (GI) or non-GI, and MB. Another subgroup analysis was performed based on types of DOAC – apixaban, rivaroxaban and edoxaban – on VTE recurrence, MB and CRNMB.

Sensitivity analyses were performed, excluding studies one by one to identify possible sources of statistical heterogeneity.

Results

Study Selection

The initial search yielded 2,728 papers. After duplicates were removed, 97 studies remained. Of these, 61 were excluded after reviewing titles and abstracts. A total of 36 full-text articles were assessed for eligibility, with eight clinical trials meeting the inclusion criteria. Notably, the study selection highlighted a limited number of studies evaluating dabigatran for VTE recurrence prevention (Supplementary Figure 1).

Study Characteristics

Supplementary Table 1 summarises the key characteristics of the studies. Collectively, these studies, published between 2013 and 2021, included a total of 4,745 patients, with 2,408 receiving DOACs and 2,337 receiving LMWH across multiple countries spanning two continents.

Risk of Bias

A summary of the risk of bias is shown in Supplementary Figure 2. The risk of bias is generally low for all studies. However, performance bias and detection bias were generally difficult to avoid because of the difference in route of administration of the interventions used. DOACs and LMWH are generally administered differently, one being given orally and the other subcutaneously; hence, it is difficult to blind the participants and assessors to which treatment was given.

Quantitative Analysis

VTE Recurrence

VTE recurrence was observed in 5.15% (124 of 2,408) of cancer patients treated with DOACs compared to 8.04% (188/2,337) of those treated with LMWH (Figure 1). Patients receiving DOACs had a significantly lower risk of VTE recurrence than those treated with LMWH, with RR 0.64; 95% CI [0.52–0.80]; p<0.0001; I²=0%.

Major Bleeding

MB occurred in 3.90% (94/2,408) of cancer patients treated with DOACs, compared to 3.63% (85/2,337) in those treated with LMWH (Figure 2, top panel). There was no statistically significant difference in the risk of MB between the two groups, with an RR of 1.02 (95% CI [0.68–1.54]; p=0.91; I²=33%). Subgroup analysis by cancer site revealed no significant difference in the risk of MB between DOAC and LMWH in non-GI sites (Figure 2, bottom panel). However, a slightly higher incidence of MB was observed in GI sites among patients receiving DOACs compared to those in the LMWH group.

Clinically Relevant Nonmajor Bleeding

CRNMB was observed in 11.07% (222/2,004) of cancer patients treated with DOACs, compared to 9.0% (175/1,945) in those receiving LMWH (Figure 3). Although the incidence of CRNMB was higher in the DOAC group, the difference between the two treatment groups was not statistically significant (RR 1.23; 95% CI: 0.85–1.79; p=0.28; I²=68%).

All-cause Mortality

All-cause mortality was reported in 25.58% (512/2,001) of patients in the DOAC group, compared to 25.28% (490/1,938) in the LMWH group (Figure 4). There was no significant difference in the risk of all-cause mortality between the two groups (RR 1.02; 95% CI [0.92–1.14]; p=0.65; I²=0%).

Type of DOAC

Subgroup analyses by DOAC type were conducted for VTE recurrence, MB and CRNMB (Figure 5, Supplementary Figures 3–4). Rivaroxaban was associated with a statistically significant lower risk of VTE recurrence compared to LMWH (RR 0.60; 95% CI [0.38–0.96]; p=0.03; I²=0%). In contrast, edoxaban was linked to a statistically significant higher incidence of major bleeding compared to LMWH (RR 1.72; 95% CI [1.02–2.91]; p=0.03) with I² not applicable as only the Hokusai-VTE trial was included¹¹ No statistically significant differences were observed between LMWH and any type of DOAC for CRNMB.

Sensitivity Analysis

The summary effect estimates remained consistent with the primary analyses when repeated using a random effects model. Sensitivity analysis, conducted by sequentially removing individual studies, yielded similar results for VTE recurrence, MB, CRNMB and all-cause mortality, aligning with the primary findings (Supplementary Tables 2–5). These results highlight the robustness of the analysis, as the sensitivity testing confirmed the stability and reliability of the primary outcomes.

Discussion

The meta-analysis demonstrated that anticoagulant therapy with DOACs in cancer patients with VTE is associated with a reduced risk of recurrent VTE compared to LMWH. Notably, in terms of safety, there were no significant differences between DOACs and LMWH in the risks of MB and CRNMB. However, our study did not include dabigatran due to limited literature.

These findings align with previous studies on the topic. For instance, a 2021 meta-analysis by Giustozzi et al. found that DOACs reduced the risk of cancer-associated VTE recurrence without a significant increase in MB when compared to LMWH¹² Similarly, a study by Dong et al. reported a substantial reduction in VTE recurrence with DOAC use, noting that while there was a trend toward increased MB with DOACs, the difference was not statistically significant¹³ Another meta-analysis by Frere et al. of six RCTs demonstrated a significantly lower risk of VTE recurrence with DOACs compared to LMWH (RR 0.67; 95% CI [0.52–0.85]; p=0.001; I²=0%)⁴ Although MB and CRNMB were more frequent in patients on DOACs, the difference was not statistically significant and there was no difference in all-cause mortality between the two treatment groups. The consistent results across these studies underscore the favourable pharmacokinetic and pharmacodynamic profiles of DOACs, which include a rapid onset and offset of action, a short half-life, a broad therapeutic window and minimal drug–drug and drug–food interactions¹⁴

Rivaroxaban

Among the DOACs, our study highlighted rivaroxaban as particularly effective at preventing cancer-associated VTE recurrence. Several studies have evaluated the safety and efficacy of rivaroxaban in cancer patients, showing a favourable balance between its benefits and potential bleeding risks. In the EINSTEIN CHOICE trial, rivaroxaban was shown to significantly reduce the risk of VTE recurrence compared to LMWH (1.3% versus 7.1%, p<0.001)¹⁵ Similarly, the SELECT-D trial reported that rivaroxaban was superior to LMWH for preventing VTE recurrence in cancer patients (4.4% versus 10.2%, HR 0.43; 95% CI [0.31–0.60]; p<0.001)¹⁶ Furthermore, the CASTA-DIVA trial demonstrated Rivaroxaban’s non-inferiority to LMWH in preventing VTE recurrence (0.7 versus 3.0%, p=0.006) and indicated a similar safety profile between the two treatments¹⁷ These studies suggest that Rivaroxaban serves as an effective alternative to LMWH for cancer-associated VTE prevention.

The pharmacokinetic properties of Rivaroxaban, such as its rapid onset of action and shorter half-life, contribute to its efficacy. Rivaroxaban reaches peak plasma concentration within 2–4 hours of oral administration and has a half-life of approximately 4–6 hours, promoting consistent anticoagulation and potentially reducing treatment failures seen with LMWH¹⁸ This stability in therapeutic effect has likely contributed to the lower VTE recurrence rates observed in clinical trials.

However, the bleeding risk associated with rivaroxaban remains a critical consideration. MB is a particularly concerning complication in cancer patients as it can increase morbidity and may necessitate interruptions in anticoagulation therapy. Studies indicate that while rivaroxaban is effective, it may be associated with higher rates of bleeding compared to LMWH, especially in patients with GI cancers or elevated baseline bleeding risks¹⁹ For example, the SELECT-D trial reported a higher incidence of CRNMB events with Rivaroxaban, with rates of 13% compared to 4% in the LMWH group²⁰ Although CRNMB events are not typically life-threatening, they can disrupt cancer treatment and necessitate additional interventions, such as transfusions, thereby affecting patient quality of life and overall treatment outcomes.

Apixaban

Our study found that compared to LMWH, apixaban had a statistically non-significant association with VTE recurrence as well as with MB and CRNMB. Interestingly, these findings contrast with the findings of a meta-analysis conducted by Baloch et al., which demonstrated a lower risk of VTE recurrence with apixaban compared to LMWH and other DOACs. When examining safety outcomes, both our study and that of Baloch et al. found no significant difference in the risk of MB events between apixaban and LMWH²¹ The differences between our studies could be attributed to sample size, study populations, or methodological variations. The pooled data may also capture a broader patient population with varying cancer types and risk profiles, providing a more comprehensive view of apixaban’s efficacy and safety across diverse clinical settings.

Edoxaban

The use of edoxaban for the treatment of cancer-associated VTE recurrence has seen increased adoption in recent years¹¹ While our study found a significant association between edoxaban and MB when compared to LMWH, only one clinical trial on edoxaban was included in the analysis. In broader studies outside of cancer-associated VTE recurrence, edoxaban has shown a lower bleeding risk compared to other anticoagulants. For example, a meta-analysis assessing the use of DOACs for stroke prevention in patients with AF reported that edoxaban was associated with a reduced risk of major bleeding compared to rivaroxaban²² These findings suggest that further research is warranted to confirm its bleeding risk profile specifically in cancer-associated VTE recurrence.

Safety Considerations for DOACs in Patients with Different Malignancy Types

Although not statistically significant in the meta-analysis, patients with GI malignancies have an increased risk of bleeding compared to those with non-Gl malignancies when treated with DOACs. This is likely due to the location of the malignancy and the presence of pathologies such as ulceration or inflammation that can increase vascular fragility and bleeding⁶ GI malignancy has been identified as an independent predictor of major bleeding in patients undergoing anticoagulant therapy for cancer-associated VTE⁶ This underscores the importance of closely monitoring symptoms and signs of bleeding in patients with gastrointestinal malignancy who are on anticoagulant therapy.

Rivaroxaban in Relation to GI Malignancy Bleeding

Although most subgroups included a limited number of patients, bleeding events were primarily observed among those with cancer of the lung, upper GI, genitourinary tract and hematologic malignancies, with the majority of incidents occurring during the active phase of malignancy. Regarding composite outcomes for MB, patients treated with rivaroxaban experienced fewer adverse events than those receiving enoxaparin and a vitamin K antagonist¹⁵ This subgroup analysis, as reported by Prins et al. in 2014, highlights the need for further detailed examination¹⁵

Apixaban in Relation to GI Malignancy Bleeding

In the Caravaggio trial, major bleeding occurred in 22/585 patients (3.8%) in the apixaban group. Although this was not statistically significant, it suggests that oral apixaban is non-inferior to subcutaneous dalteparin for treating cancer-associated VTE without an increased risk of MB²³ Nevertheless, it is essential to recognise gastrointestinal malignancy as a known risk factor for significant bleeding. The rates of major bleeding, including GI bleeding cases, between the apixaban and dalteparin groups were comparable. These findings differ from those of previous studies, which reported higher bleeding incidences with direct oral anticoagulants. Key risk factors for GI bleeding include surgery, invasive procedures and declining renal function.

Edoxaban in Relation to Gastrointestinal Malignancy Bleeding

The HOKUSAI-VTE cancer trial found that patients with GI malignancies treated with edoxaban faced an elevated bleeding risk¹¹ The incidence of MB was significantly higher in those treated with edoxaban compared to dalteparin (6.9% versus 4.0%; HR 1.77; 95% CI [1.03–3.04]; p=0.04). This finding aligns with previous studies on DOACs. The increase in major upper GI bleeding predominantly affected patients who entered the trial with pre-existing GI cancer. Further research is needed to investigate the factors contributing to this elevated bleeding risk, taking into account variables such as treatment duration, disease extent and patient comorbidities¹¹ Additionally, Marshall et al. observed that prolonged treatment with edoxaban increases the risk of GI bleeding¹⁶

Patient Adherence to DOAC Therapy Compared to LMWH Therapy

Patient adherence, or the extent to which patients follow prescribed anticoagulant regimens, is crucial for the effective management of cancer-associated VTE²⁴ Although the American Society of Clinical Oncology recommends LMWH for the prophylaxis and treatment of VTE in cancer patients, a study by Guo et al. demonstrated lower adherence rates for LMWH compared to DOACs²⁵ In a study conducted by Mohamed et al., the therapeutic duration of LMWH was consistently shorter than that of rivaroxaban, which was statistically significant²⁶ Lower long-term adherence to LMWH in some patients is attributed to practical limitations, including the subcutaneous administration route, which can cause pain and bruising, as well as higher costs compared to DOACs^19,27,28 This finding underscores the need for more practical and convenient treatment options for cancer patients with VTE.

Strengths and Limitations

Eight RCTs were included in the meta-analysis, demonstrating consistent effectiveness in reducing the risk of cancer-associated VTE recurrence and in the secondary outcome of all-cause mortality (I²=0%), indicating no heterogeneity across studies. Sensitivity analyses were conducted and yielded similar results for VTE recurrence, MB, CRNMB and all-cause mortality, consistent with the primary analysis. Additionally, subgroup analyses were performed to assess the primary and secondary outcomes based on different types of DOACs.

However, several limitations should be considered. First, variations in follow-up durations among the included clinical trials (ranging from 3 to 12 months) may have influenced the study outcomes. Different follow-up periods can affect the interpretation of long-term outcomes such as VTE recurrence and mortality. Shorter follow-up periods may underestimate the incidence of recurrent events or mortality as some late recurrences or complications may not be captured. Conversely, longer follow-up periods can provide more comprehensive data but introduce variability, especially if patients transition to different treatments or care settings over time.

The meta-analysis revealed moderate to substantial heterogeneity in MB (I²=33%) and CRNMB (I²=68%). Subgroup analyses involving rivaroxaban (for MB and CRNMB) and apixaban also exhibited moderate to substantial heterogeneity for CRNMB. The observed heterogeneity in MB and CRNMB outcomes likely arises from several factors. Variability in study designs among the included trials, such as ADAM-VTE, Caravaggio, Hokusai-VTE, SELECT-D, Guntupalli et al. 2020, and Prins et al. 2024, led to differences in patient populations, treatment protocols and outcome definitions.^{7,11,15,16,23,29} Patient characteristics, including age, sex, comorbidities and the presence of GI malignancies influenced bleeding risks. Additionally, differences in the specific DOACs used and their dosing regimens contributed to varying bleeding rates. Discrepancies in follow-up durations and methods for assessing bleeding events further contributed to this variability, affecting the strength of conclusions that could be drawn from the safety outcomes in the analysis.

In most study groups, particularly in the Hokusai VTE trial, DOACs were primarily compared to dalteparin, focusing on MB and CRNMB¹¹ This variation in the control arm may have influenced the safety outcomes associated with DOACs.

Recommendations

To enhance clinical decision-making and personalise anticoagulation therapy for cancer-associated VTE, the authors propose specific recommendations tailored to cancer type and patient risk factors.

First, patient selection based on bleeding risk is crucial. For patients with low-to-moderate bleeding risk, DOACs such as apixaban or rivaroxaban are generally appropriate choices due to their ease of administration and favourable performance in preventing recurrence. However, for patients with high bleeding risk – especially those with GI malignancies – LMWH remains preferable, as increased bleeding risks have been observed with DOACs in these patients. When prescribing DOACs for GI cancer patients, clinicians should proceed with caution and monitor closely for signs of GI bleeding.

Cancer type and treatment setting also play important roles in selecting anticoagulation therapy. For patients with GI malignancies, due to their heightened bleeding risk, DOACs should be cautiously considered only when their benefits clearly outweigh the risks. Regular monitoring for GI bleeding signs is essential if DOACs are prescribed. In contrast, for patients with non-GI malignancies, DOACs may be preferable due to a safety profile comparable to LMWH and the convenience of oral administration, which can improve patient adherence and quality of life. Furthermore, individualised assessment of comorbidities and renal function is recommended. Since DOACs require dosing adjustments based on renal function, clinicians should carefully evaluate renal status and other comorbidities before initiating therapy. In patients with impaired renal function or other bleeding risks, LMWH may provide a safer profile.

Given the importance of patient adherence in influencing VTE outcomes, DOACs are generally favoured over LMWH for patients concerned about the frequency and discomfort of subcutaneous injections. Improved adherence with DOACs can significantly enhance therapeutic success, especially for patients who require long-term VTE management. Considering the variability in follow-up durations in existing studies, this meta-analysis is most applicable for guiding treatment over a 3 to 12-month period. As long-term safety and efficacy of DOACs require further research, decisions regarding extended prophylaxis should be individualised, taking into account factors such as cancer progression and ongoing treatments.

Finally, further clinical trials with standardised and extended follow-up periods are needed to refine our understanding of DOAC safety and efficacy across different malignancy types. Research focusing on the primary prevention of cancer-associated VTE with DOACs, particularly in high-risk populations, would address a critical gap in the current literature. These recommendations provide actionable guidelines for clinicians, emphasising the importance of tailoring anticoagulation therapy to cancer type, individual risk profile, and patient preferences to optimise treatment outcomes and enhance patient safety in managing cancer-associated VTE.

Conclusion

This meta-analysis demonstrates that DOACs offer an effective and safe alternative to LMWH in preventing VTE recurrence among selected cancer patients. While DOACs reduce the recurrence risk of VTE without significantly increasing risks of MB or CRNMB, clinicians should exercise caution in patients with high bleeding risk, particularly those with GI malignancies.

Click here to view Supplementary Material.