Direct oral anticoagulants compared to low molecular weight heparin for the treatment of venous thromboembolism among individuals with gastrointestinal cancer: an updated meta-analysis

Abstract

Background

Cancer, particularly gastrointestinal cancer, is associated with a higher risk of venous thromboembolism. Recent studies have increasingly compared direct oral anticoagulants (DOACs) with low molecular weight heparin (LMWH) for treating venous thromboembolism (VTE) in patients with gastrointestinal cancer. This meta-analysis aimed to investigate the efficacy of DOACs compared to LMWH for VTE in patients with gastrointestinal cancer.

Methods

PubMed, the Cochrane Library, Scopus, and Web of Science were systematically searched from the inception to February 20, 2025, to identify randomized controlled trials (RCTs) or cohort studies comparing the effect of DOACs with LMWH on VTE recurrence, clinically relevant non-major bleeding, and major bleeding among patients with active gastrointestinal cancer suffering from VTE.

Results

Finally, 13 studies, including 8 cohort studies and 5 RCTs, were included. The random-effects model revealed that compared to using LMWH, the risk of VTE recurrence (risk ratio (RR) 0.75, 95% confidence interval (CI) (0.59, 0.97), I² = 0.00%) was significantly decreased and the risk of clinically relevant non-major bleeding (RR 1.64, 95% CI (1.10, 2.45), I² = 59.79%) was significantly increased when using DOACs; however, the risk of major bleeding (RR 1.16, 95% CI (0.86, 1.56), I² = 28.51%) did not significantly change. RCTs suggested no significant change (RR 0.80, 95% CI (0.50, 1.27), I² = 0.00%) in the risk of VTE recurrence, but cohort studies indicated a decreased risk (RR 0.73, 95% CI (0.54, 0.99), I² = 0.00%) of VTE recurrence when using DOACs instead of LMWH. Both cohort studies (RR 1.02, 95% CI (0.75, 1.39), I² = 20.43%) and RCTs (RR 1.65, 95% CI (0.89, 3.07), I² = 26.59%) showed no significant difference in the risk of major bleeding when using DOACs instead of LMWH. RCTs showed an elevated risk (RR 2.32, 95% CI (1.48, 3.64), I² = 0.00%) of clinically relevant non-major bleeding when using DOACs, but cohort studies reported no significant change (RR 1.40, 95% CI (0.86, 2.29), I² = 60.84%) in the risk of clinically relevant non-major bleeding.

Conclusion

Among patients with gastrointestinal cancer, compared to LMWH, DOACs may not increase the risk of VTE recurrence and major bleeding, but may increase the risk of clinically relevant non-major bleeding.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12959-025-00789-7.

Introduction

Cancer is associated with a substantially increased odds of venous thromboembolism (VTE), particularly deep vein thrombosis (DVT) and pulmonary thromboembolism (PTE), which can also increase the mortality rate of patients and shorten their survival [1]. The inflammatory response, the prothrombotic mediators released by cancer cells, tumor cell metastasis, and abnormal activation of the endothelial layer, are known to induce a hypercoagulable state in cancer, making patients with cancer prone to VTE [2]. A retrospective study reported that of 942,019 patients with cancer, 62,003 (6.6%) developed cancer-associated thrombosis [1].

Patients with gastrointestinal diseases suffer from a high risk of VTE [3]. Herein, a meta-analysis of 28 studies indicated that the 30-day and 90-day incidence rates of VTE among patients undergoing colorectal cancer resection were 195 and 91 per 1000 person-years [3]. Similarly, another meta-analysis reported that of 9768 patients who underwent esophagectomy for esophageal cancer, 4% developed VTE [4]. Notably, VTE was found to occur in approximately 16.5% of those with pancreatic cancer [5]. Because of the high risk of VTE in patients with gastrointestinal cancer [6, 7], highly effective and safe antithrombotic drugs are needed to resolve the existing thrombosis and lower the risk of future thrombosis, without increasing the risk of bleeding events [8]. The available guidelines mainly support the use of low molecular weight heparin (LMWH) as the antithrombotic drug of choice for managing patients with cancer [8]. Recently, several cohort studies and randomized controlled trials (RCTs) have investigated the efficacy and safety of direct oral anticoagulants (DOACs), including direct thrombin inhibitors and factor Xa inhibitors, compared to LMWH for patients with gastrointestinal cancer and VTE [9–12]; however, their results remained controversial to some extent.

This updated meta-analysis included a markedly higher number of studies to compare the efficacy of DOACs with LMWH regarding VTE recurrence, clinically relevant non-major bleeding, and major bleeding in the treatment of patients suffering from active gastrointestinal cancer and VTE. Our results can address the discrepancies between the results of previous studies and provide robust evidence for future clinical practice. Although there are previous meta-analyses on the same topic, our updated meta-analysis included considerably more studies and patients [13]. Besides, there are differences between the results of this updated meta-analysis and previous meta-analyses, which included fewer studies. For instance, our meta-analysis unveiled that DOACs can decrease the risk of VTE recurrence compared to LMWH, but the most recent meta-analysis on the same topic reported non-significant results [13].

Method

Search strategy

We registered the protocol of this systematic review in PROSPERO (CRD420250655802). Furthermore, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline was followed for this study [14]. PubMed, the Cochrane Library, Scopus, and Web of Science were searched from the inception to February, 20, 2025, using the following search keywords: (“DOAC” OR “rivaroxaban” OR “apixaban” OR “dabigatran” OR “edoxaban”) [title, abstract, keywords] AND (“low-molecular weight heparin” OR “low molecular weight heparin” OR “low-molecular-weight heparin” OR “LMWH” OR “enoxaparin” OR “dalteparin” OR “tinzaparin” OR “certoparin” OR “bemiparin” OR “nadroparin” OR “parnaparin” OR “reviparin”) [title, abstract, keywords] AND (“gastrointestinal cancer” OR “gastric cancer” OR “esophageal cancer” OR “intestinal cancer” OR “stomach cancer” OR “colon cancer” OR “colorectal cancer” OR “pancreatic cancer” OR “pancreatic ductal adenocarcinoma” OR “biliary cancer” OR “hepatic cancer” OR “hepatobiliary cancer” OR “gastrointestinal malignancy” OR “gastric malignancy” OR “esophageal malignancy” OR “intestinal malignancy” OR “stomach malignancy” OR “colon malignancy” OR “colorectal malignancy” OR “pancreatic malignancy” OR “pancreatic ductal adenocarcinoma” OR “biliary malignancy” OR “hepatic malignancy” OR “hepatobiliary malignancy”) [all fields].

The search strategy is reported in Supplementary Material 1. Besides, we assessed the reference lists of the retrieved trials and recent review articles to include more records. The identified studies were transferred to EndNote 9.0, and duplicate reports were removed.

Inclusion and exclusion criteria

The inclusion criteria for this study were as follows: (I) being an RCT or a cohort study; (II) comparing the efficacy of DOACs with LMWH for managing VTE in patients suffering from gastrointestinal cancer; (III) at least one of the following outcomes was reported: recurrence of VTE, major bleeding, and clinically relevant non-major bleeding.

The exclusion criteria for this study were as follows: (I) being a case-control study, case report, case series, cross-sectional study, or review article; (II) absence of DOAC or LMWH group; (III) none of the target outcomes were reported; (IV) the outcomes were not specifically reported in those with gastrointestinal cancer.

Outcome definition

The definitions of outcomes were as follows:

• Recurrent VTE: Defined as the formation of a new thrombus or involvement of a new blood vessel based on duplex ultrasonography, chest CT scan, or lung perfusion.

• Major bleeding: Defined as intracranial hemorrhage, major gastrointestinal bleeding, bleeding at other critical sites, and a decrease in hemoglobin of at least 2 g/dL over 24 h.

• Clinically relevant non-major bleeding: Acute clinically overt bleeding that did not meet the criteria for major bleeding, but necessitated nonsurgical, medical intervention by a health care professional, leading to hospitalization or a higher level of care, or prompt evaluation.

Although the overall definition of outcomes was consistent across all studies, there were some minor differences between individual studies regarding the definition of outcomes.

Data extraction and outcome measures

Two authors independently searched the literature and assessed the eligibility of the retrieved studies. All disagreements were resolved via consultation. The authors screened the titles and abstracts to assess the relevance of the articles and measure their compliance with the inclusion and exclusion criteria. Thereafter, the full text of the remaining articles was evaluated to determine their eligibility for inclusion.

We collected the following data: study characteristics (the first author’s name, publication year, and study design and settings), population (sex, age, trial registration number, and sample size), interventions (type of DOAC, type of LMWH, doses of each drug, and length of treatment), and the outcome measures (recurrence of VTE, clinically relevant non-major bleeding, and major bleeding).

Risk of bias assessment

The bias risk was measured by two researchers independently, and all discrepancies were resolved through discussions. The revised Cochrane risk-of-bias tool for randomized trials (RoB2) was employed to determine the risk of bias for RCTs included in this meta-analysis. The tool measures the risk of bias based on the randomization process, deviations from intended interventions, missing outcome data, selection of the reported results, and outcome measurement, thereby providing an overall risk of bias. One of the following ranks was assigned to each domain: some concerns, low risk of bias, and high risk of bias [15].

For cohort studies, the Newcastle-Ottawa Scale (NOS) was used to assess the risk of bias, and each study was evaluated in three domains, including selection of cohorts, comparability, and assessment of outcomes. In particular, each study could receive a total of nine scores based on answers to 8 items.

Statistical analysis

For all outcome measures, the number of events and non-events per group was collected to calculate the risk ratio (RR). For each outcome measure, we also separately repeated the meta-analysis for RCTs and cohort studies. This meta-analysis was performed employing Stata version 17.0 (StataCorp, TX, USA).

The chi-square test and I² statistics were utilized to measure between-trial heterogeneity. Specifically, I² of more than 50% was regarded as a high statistical heterogeneity. A random-effects model (DerSimonian-Laird) was employed to pool data.

Subgroup analysis

Subgroup analysis was performed based on the drugs used as LMWH and DOACs to determine the potential sources of heterogeneity.

Publication bias

We drew a funnel plot to visually investigate publication bias risk. In addition, Begg’s and Egger’s tests were employed to determine publication bias risk.

Results

Systematic search results

Via systematic search, we found 39 records in PubMed, 570 records in Scopus, 264 records in Web of Science, and 39 records in the Cochrane Library. Of 912 articles, 95 duplicate records were deleted, and 817 records remained. By screening the titles and abstracts, 748 articles were excluded due to incompliance with the inclusion criteria or irrelevance to the topic. Subsequently, the full-text version of the 69 remaining articles was evaluated to assess their eligibility for inclusion. Finally, 14 articles from 13 studies, including 8 cohort studies [6, 12, 16–21] and 5 RCTs [7, 9–11, 22, 23], were included in this meta-analysis (Fig. 1 and Table 1). Two articles reported different outcomes of the same RCT (the Caravaggio study) [7, 22]. Nearly all types of gastrointestinal tract, hepatobiliary, and pancreatic cancer were enrolled in the included studies, but most patients had colorectal cancer. Except for the cohort study conducted by Houghton et al. [19], all other cohort studies were conducted retrospectively. There were 995 participants from RCTs and 4847 participants from cohort studies, totaling 5842 participants. The mean age of participants ranged from 62.4 to 67.2 years across studies. Except for two cohort studies that conducted 12-month follow-up [12, 18], other studies assessed the outcomes until 6 months of follow-up. Rivaroxaban and apixaban were the most commonly used DOACs, while dalteparin and enoxaparin were the most commonly used forms of LMWH (Table 1).

Fig. 1.

Systematic search flowchart

Table 1.

Study characteristics

First author name	Publication year	Study design	Cancer type	Age (year, mean)	Male gender percentage	DOAC type	DOAC dose	LMHW type	LMWH dose	Length of follow-up (months)	Sample size	References
Ageno et al.	2020	RCT	Colorectal cancer (62.4%), pancreatic or hepatobiliary cancer (23.2%), and upper gastrointestinal cancer (14.4%)	67.2	62.9%	Apixaban	10 mg twice a day (bid) for 7 days, followed by 5 mg bid	Dalteparin		6	375	[7]
Agnelli et al.	2022	RCT	Colorectal cancer (61.8%), pancreatic cancer (17.8%), gastric cancer (6.7%), esophageal cancer (6.1%), biliary tract cancer (4%), hepatocellular carcinoma (1.3%), and unknown gastrointestinal cancer (2.1%)	67.2	62.9%	Apixaban	10 mg twice daily for the first 7 days, followed by 5 mg twice daily	Dalteparin		6	375	[22]
Kim et al.	2022	RCT	Esophageal cancer (14.5%), gastric cancer (41.1%), ampulla of Vater cancer (2.2%), duodenal cancer (1.1%), hepatocellular carcinoma (2.3%), biliary cancer (18.9%), and pancreatic cancer (19.9%)	63.5	53.3%	Rivaroxaban or apixaban	Rivaroxaban (15 mg twice daily for 3 weeks, then 20 mg once daily)/apixaban (10 mg twice daily for the first 7 days, then 5 mg twice daily)	Dalteparin		6	90	[11]
Mokadem et al.	2021	RCT	Colon cancer (87.5%) and liver cancer (12.5%)	60.6	42%	Apixaban	10 mg twice daily for seven days followed by 5 mg twice daily	Enoxaparin		6	48	[9]
Mulder et al.	2020	RCT	Colorectal cancer (53.1%), pancreatic cancer (20.3%), esophageal or gastric cancer (17.71%), and hepatobiliary cancer (8.5%)	64	66.2%	Edoxaban	60 mg once daily (or 30 mg in those meeting criteria for dose reduction) after a low-molecular-weight heparin lead-in of at least 5 days, or to dalteparin	Dalteparin		6	305	[10]
Young et al.	2018	RCT	Colorectal (57.6%), gastric (6.2%), gallbladder (2.2%), esophageal/gastroesophageal (16.9%), pancreatic (16.9) cancer	67	52.5%	Rivaroxaban	15 mg twice daily for 3 weeks, then 20 mg once daily	Dalteparin		6	177	[23]
Caroti et al.	2023	Retrospective cohort	Colorectal, hepatobiliary, pancreas, and upper gastrointestinal cancer	Not reported	47.8%	Rivaroxaban	15 mg twice daily for 3 weeks, then 20 mg once daily	Mostly enoxaparin		6	1871	[16]
Cohen et al.	2018	Retrospective cohort	Not reported	64.2	48.5%	Apixaban	Not reported	Only mentioned LMWH		6	1520	[17]
Guan et al.	2023	Retrospective cohort	Colon cancer, esophageal cancer, hepatic cancer, pancreatic cancer, rectal cancer, gastric cancer, anal cancer, bile duct cancer, and gallbladder cancer	62.8	52.15%	Rivaroxaban or apixaban	Not reported	Enoxaparin and dalteparin		12	141	[17]
Kim et al.	2020	Retrospective cohort	Esophageal cancer, gastric cancer, gastrointestinal stromal tumor of the stomach, hepatocellular carcinoma, biliary cancer, pancreatic cancer, and ampulla of Vater cancer	62.5	49.85%	Rivaroxaban	15 mg for the first 21 days of treatment, and then once daily at 20 mg	Enoxaparin or dalteparin		6	174	[21]
Lee et al.	2019	Retrospective cohort	Pancreatobiliary, stomach, and colorectal cancer	64.5	57%	Rivaroxaban	15 mg twice daily with food for 21 days followed by 20 mg once daily	Dalteparin, enoxaparin, or nadroparin		6	281	[20]
Recio-Boiles et al.	2019	Retrospective cohort	Pancreas, colon, rectal, NET, gastric, esophageal, appendix, biliary, and gastroesophageal junction cancer, and hepatocellular carcinoma	66.6	62.1%	Apixaban or rivaroxaban	Rivaroxaban at 15 mg BID for 3 weeks, then 20 mg daily and Apixaban at 10 mg BID for 7 d, then 5 mg twice daily	Enoxaparin		6	106	[6]
Chen et al.	2021	Retrospective cohort	Colorectal, esophageal, intestinal, stomach, pancreatic, and gallbladder cancer	66	47.9%	Apixaban, rivaroxaban, edoxaban, or dabigatran	5 mg twice per day for apixaban, 60 mg once daily for edoxaban, 150 mg twice daily for dabigatran, and 15 mg twice for the first 21 days and then 20 mg per day for rivaroxaban	Enoxaparin		12	302	[12]
Houghton et al.	2021	Prospective cohort	Upper and lower gastrointestinal tract cancer and hepatobiliary and pancreatic cancer	62.4	55%	Rivaroxaban or apixaban	Not reported	Enoxaparin	Not reported	6	452	[19]

Risk of bias assessment

Due to the open-label design of RCTs included in this study, they were found to have a high risk of bias based on the revised Cochrane risk-of-bias tool for randomized trials (RoB2). Regarding cohort studies, most of them had a small sample size, which could possibly undermine their representativeness (Supplementary material 2).

Meta-analysis

Using data from the included studies, we compared the effects of DOACs with LMWH on the risk of VTE recurrence, clinically relevant non-major bleeding, and major bleeding among those with active gastrointestinal cancer who were receiving these anti-thrombotic agents. In addition to pooling data from both RCTs and cohort studies, all analyses were separately repeated for cohort studies and RCTs. Using a random-effects model, we found that compared to using LMWH, VTE recurrence risk was significantly decreased when using DOACs (RR 0.75, 95% CI (0.59, 0.97), I² = 0.00%) (Fig. 2). Considering the study design, the difference in VTE recurrence risk was significant in cohort studies (RR 0.73, 95% CI (0.54, 0.99), I² = 0.00%), but non-significant in RCTs (RR 0.80, 95% CI (0.50, 1.27), I² = 0.00%).

Fig. 2.

The effect of DOACs compared to LMWH on the risk of VTE recurrence among patients with gastrointestinal cancer and VTE

The random-effects model indicated that compared to treatment with LMWH, major bleeding risk was not significantly changed when using DOACs (RR 1.16, 95% CI (0.86, 1.56), I² = 28.51%) (Fig. 3). Considering the study design, the difference in the risk of major bleeding remained non-significant in both cohort studies (RR 1.02, 95% CI (0.75, 1.39), I² = 20.43%) and RCTs (RR 1.65, 95% CI (0.89, 3.07), I² = 26.59%).

Fig. 3.

The effect of DOACs compared to LMWH on the risk of major bleeding recurrence among patients with gastrointestinal cancer and VTE

The random-effects model indicated that compared to treatment with LMWH, the risk of clinically relevant non-major bleeding was markedly elevated when using DOACs (RR 1.64, 95% CI (1.10, 2.45), I² = 59.79%) (Fig. 4). Considering the study design, the difference in the risk of clinically relevant non-major bleeding was significantly increased in RCTs (RR 2.32, 95% CI (1.48, 3.64), I² = 0.00%), but not in cohort studies (RR 1.40, 95% CI (0.86, 2.29), I² = 60.84%). Taken together, compared to RCTs, cohort studies supported the use of DOACs over LMWH.

Fig. 4.

The effect of DOACs compared to LMWH on the risk of clinically relevant non-major bleeding among patients with gastrointestinal cancer and VTE

Subgroup analysis

Subgroup analysis was performed based on the drugs used as LMWH (dalteparin vs. enoxaparin) and DOACs (apixaban vs. rivaroxaban). Subgroup analysis indicated no significant difference between apixaban and rivaroxaban in terms of VTE recurrence and major bleeding. However, unlike apixaban, rivaroxaban significantly increased the risk of clinically relevant non-major bleeding compared to LMWH. In addition, DOACs offered a reduced risk of VTE recurrence, clinically relevant non-major bleeding, and major bleeding when enoxaparin was administered as the comparator, compared to when dalteparin was administered as the comparator. However, these differences did not meet the threshold of statistical significance (Supplementary material 3).

Funnel plot

Visually, the funnel plot revealed some degrees of asymmetry for VTE recurrence and major bleeding (Supplementary material 4); therefore, we performed Begg’s and Egger’s test to assess the risk of publication bias both of which indicated no significant risk of publication bias for VTE recurrence (p-value = 0.876 and p-value = 0.878, respectively) and major bleeding (p-value = 0.360 and p-value = 0.264, respectively).

Discussion

This meta-analysis with 14 articles from 13 studies, including 5 RCTs and 8 cohort studies, compared the efficacy of DOACs with LMWH for the management of patients with active gastrointestinal cancer and VTE. The results indicated that compared to LMWH, the risk of VTE recurrence was significantly decreased and the risk of clinically relevant non-major bleeding was significantly increased when using DOACs; however, the risk of major bleeding did not change significantly. RCTs suggested no change in the risk of VTE recurrence and major bleeding, but indicated an increased risk of clinically relevant non-major bleeding when using DOACs instead of LMWH. In contrast, cohort studies reported no change in the risk of major bleeding and clinically relevant non-major bleeding, but mentioned a decreased risk of VTE recurrence when using DOACs instead of LMWH. In addition, subgroup analysis indicated that, unlike apixaban, rivaroxaban significantly increased the risk of clinically relevant non-major bleeding compared to LMWH.

Although the overall result suggested the non-inferiority of DOACs, it should be considered that a great proportion of data came from cohort studies, which generally supported the use of DOACs over LMWH. Compared to RCTs, cohort studies, particularly retrospective cohort studies, are more susceptible to bias and missing and incomplete data. Furthermore, randomization, patient selection, and outcome definition are more rigorously pursued in RCTs. Therefore, compared to RCTs, observational studies may overestimate the effects of DOACs, and the results of this meta-analysis should be cautiously interpreted [24].

Similar to this meta-analysis, Rungjirajittranon et al. compared the efficacy of DOACs with LMWH for the treatment of patients with gastrointestinal cancer and VTE [13]. They included 2226 patients from 11 studies (RCTs and cohort studies) and reported that compared to LMWH, DOACs did not significantly change the risk of major bleeding (RR 1.31, 95% CI (0.84–2.04) I² = 41%) and VTE recurrence (RR 0.72, 95% CI (0.49–1.04) I² = 0.00%), but increased the risk of clinically relevant non-major bleeding (RR 1.76, 95% CI (1.24–2.52) I² = 0.00%) [13].

Another recent meta-analysis with three RCTs and seven observational studies, consisting of 3054 patients who postoperatively received DOACs or LMWH as thromboprophylaxis, found no significant differences between the two arms regarding the risk of major bleeding, VTE, and clinically relevant non-major bleeding in the 30-day postoperative period [25]. Compared to our meta-analysis, this study enrolled patients with all types of cancer and had a shorter follow-up length [25]. A recent meta-analysis of 10 RCTs comprising patients with different types of cancer suffering from VTE who were treated with DOACs or LMWH reported that DOACs led to a significant reduction in the incidence of VTE compared to LMWH [26]. In addition, the meta-analysis reported no significant difference between DOACs and LMWH in terms of mortality and total bleeding [26]. The meta-analysis conducted by Zhou et al. also revealed that among patients with abdominal/pelvic cancer suffering from postoperative VTE, DOACs can be considered safe and effective alternatives for LMWH in terms of VTE, major bleeding, and clinically relevant non-major bleeding [27].

Lee et al. compared the efficacy of DOACs with LMWH or vitamin K antagonists (VKAs) in the management of VTE in patients suffering from cancer and reported that DOACs did not significantly change the risk of major bleeding compared to LMWH/VKAs. Furthermore, they found that DOACs were significantly more effective than LMWH in preventing VTE recurrence [28].

Song et al. included 14 retrospective cohort studies and 4 RCTs and indicated that DOACs are superior to LMWH in the prevention of VTE recurrence among those with cancer [29]. Despite our meta-analysis, this study was conducted in those with cancer, regardless of the type of cancer; however, its results were consistent with our results.

DOACs are easier to administer compared to LMWH, can be administered with fixed doses, and incur a lower cost of treatment for patients and healthcare systems [11, 20, 30]. For instance, a study indicated that the 6-month cost of treatment with DOACs was $ 654.65 with 0.40 quality-adjusted life-years (QALYs), but the 6-month cost of treatment with LMWH was $ 1719.31 with 0.37 QALYs [30]. Furthermore, among patients with gastrointestinal malignancy, treatment with DOACs incurred a lower cost ($USD 657.85 vs. $USD 1716.56) and provided greater health benefits (0.40 QALYs vs. 0.37 QALYs) compared to treatment with LMWH [30]. Considering these benefits and their greater efficacy in the prevention of VTE recurrence among those with active gastrointestinal cancer, future studies and clinicians can consider the use of DOACs over LMWH when necessary; however, patient selection is still of great importance because it was shown that DOACs can elevate the likelihood of clinically relevant non-major bleeding among those with gastrointestinal cancer.

Limitations

There are some limitations to our study. First, a few RCTs were available regarding the topic. Second, the open-label nature of RCTs could increase the risk of bias and potentially affect the accuracy of our results. Third, there existed some methodological discrepancies among studies, which have been mentioned in previous sections; however, degrees of methodological difference are common among the studies included in a meta-analysis. Fourth, there were minor differences between the results of cohort studies and RCTs, which necessitate future research on this topic. Fifth, although the overall definition of outcomes was consistent across all studies, there were some minor differences between individual studies regarding the definition of outcomes. Sixth, the included studies, both RCTs and cohort studies, had a small sample size; thus, future large-scale multi-center RCTs are needed to bring more data and confirm our findings.

Conclusion

This meta-analysis revealed that DOACs may not elevate the risk of major bleeding, but may decrease the risk of VTE recurrence and increase the risk of clinically relevant non-major bleeding among those with active gastrointestinal cancer and VTE. These findings suggest that DOACs are non-inferior to LMWH in the management of patients with active gastrointestinal cancer and VTE; however, there were a few clinical trials in this regard, and a major proportion of data came from cohort studies. Therefore, these findings should be interpreted with caution, and future large-scale randomized controlled trials are needed to validate these findings.

Authors’ contributions

Pingli Li (conceptualization, searching the literature, data extraction, revising the draft, and data analysis), Jie Ren (searching the literature, data extraction, and writing the draft). All authors reviewed and edited the manuscript and approved the final version of the manuscript.

Funding

This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data will be available by the corresponding author on a reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.