Primary prophylaxis for venous thromboembolism in ambulatory cancer patients receiving chemotherapy

Marcello Di Nisio; Ettore Porreca; Matteo Candeloro; Michele De Tursi; Ilaria Russi; Anne WS Rutjes

doi:10.1002/14651858.CD008500.pub4

. 2016 Dec 1;2016(12):CD008500. doi: 10.1002/14651858.CD008500.pub4

Primary prophylaxis for venous thromboembolism in ambulatory cancer patients receiving chemotherapy

Marcello Di Nisio ^1,^2,^✉, Ettore Porreca ¹, Matteo Candeloro ³, Michele De Tursi ³, Ilaria Russi ¹, Anne WS Rutjes ^4,⁵

Editor: Cochrane Vascular Group

PMCID: PMC6463937 PMID: 27906452

Abstract

Background

Venous thromboembolism (VTE) often complicates the clinical course of cancer. The risk is further increased by chemotherapy, but the trade‐off between safety and efficacy of primary thromboprophylaxis in cancer patients treated with chemotherapy is uncertain. This is the second update of a review first published in February 2012.

Objectives

To assess the efficacy and safety of primary thromboprophylaxis for VTE in ambulatory cancer patients receiving chemotherapy compared with placebo or no thromboprophylaxis.

Search methods

For this update the Cochrane Vascular Information Specialist searched the Cochrane Vascular Group Specialised Register (June 2016). In addition, the Information Specialist searched the Cochrane Central Register of Controlled Trials (CENTRAL) (2016, Issue 5). Clinical trials registries were searched up to June 2016.

Selection criteria

Randomised controlled trials comparing any oral or parenteral anticoagulant or mechanical intervention to no thromboprophylaxis or placebo, or comparing two different anticoagulants.

Data collection and analysis

We extracted data on methodological quality, participant characteristics, interventions, and outcomes including symptomatic VTE and major bleeding as the primary effectiveness and safety outcomes, respectively.

Main results

We identified five additional randomised controlled trials (2491 participants) in the updated search, considering in this update 26 trials with a total of 12,352 participants, all evaluating pharmacological interventions and performed mainly in people with locally advanced or metastatic cancer. The quality of the evidence ranged from high to very low across the different outcomes and comparisons. The main limiting factors were imprecision and risk of bias. One large trial of 3212 participants found a 64% (risk ratio (RR) 0.36, 95% confidence interval (CI) 0.22 to 0.60) reduction of symptomatic VTE with the ultra‐low molecular weight heparin (uLMWH) semuloparin relative to placebo, with no apparent difference in major bleeding (RR 1.05, 95% CI 0.55 to 2.00). When compared with no thromboprophylaxis, LMWH significantly reduced the incidence of symptomatic VTE (RR 0.54, 95% CI 0.38 to 0.75; no heterogeneity, Tau² = 0.00%) with a non‐statistically significant 44% higher risk of major bleeding events (RR 1.44, 95% CI 0.98 to 2.11). In participants with multiple myeloma, LMWH was associated with a significant reduction in symptomatic VTE compared with the vitamin K antagonist warfarin (RR 0.33, 95% CI 0.14 to 0.83), while the difference between LMWH and aspirin was not statistically significant (RR 0.51, 95% CI 0.22 to 1.17). Major bleeding was observed in none of the participants treated with LMWH or warfarin and in less than 1% of those treated with aspirin. Only one study evaluated unfractionated heparin against no thromboprophylaxis but did not report on VTE or major bleeding. When compared with placebo, warfarin was associated with a non‐statistically significant reduction of symptomatic VTE (RR 0.15, 95% CI 0.02 to 1.20). Antithrombin, evaluated in one study involving paediatric patients, had no significant effect on VTE or on major bleeding when compared with no antithrombin. The direct oral factor Xa inhibitor apixaban was evaluated in a phase II dose‐finding study that suggested a low rate of major bleeding (2.1% versus 3.4%) and symptomatic VTE (1.1% versus 13.8%) in comparison with placebo.

Authors' conclusions

In this second update, we confirmed that primary thromboprophylaxis with LMWH significantly reduced the incidence of symptomatic VTE in ambulatory cancer patients treated with chemotherapy. In addition, the uLMWH semuloparin, which is not commercially available, significantly reduced the incidence of symptomatic VTE. The risk of major bleeding associated with LMWH, while not reaching statistical significance, suggest caution and mandate additional studies to determine the risk‐to‐benefit ratio of LMWH in this setting. Despite the encouraging results of this review, routine prophylaxis in ambulatory cancer patients cannot be recommended before safety issues are adequately addressed. We need additional studies investigating targeted primary prophylaxis in people with specific types or stages of cancer associated with a higher risk of VTE.

Plain language summary

Prevention of blood clots in non‐hospitalised cancer patients receiving chemotherapy

Background

Cancer patients are more likely than people without cancer to develop blood clots in their veins (known as venous thromboembolism). Chemotherapy further increases this risk. Yet a number of factors specific to the cancer, such as the bleeding tendency at the site of the cancer, or a relative decrease in the number of platelets in the blood (thrombocytopenia) caused by chemotherapy can increase the likelihood that cancer patients will have bleeding complications with medicines used to prevent and treat blood clots (anticoagulants). This systematic review looked at the effectiveness and safety of anticoagulants when used to prevent blood clots in cancer patients receiving chemotherapy.

Key results

We included 26 randomised controlled trials involving a total of 12,352 participants (current until June 2016). Low molecular weight heparin and the ultra‐low molecular weight heparin semuloparin were associated with a significant reduction in symptomatic blood clots. We found no evidence that the risk of major bleeding is increased with semuloparin or low molecular weight heparin, but based on the uncertainty around the estimates, we cannot exclude that the risk is doubled. There was no clear survival benefit for semuloparin or low molecular weight heparin. In people with multiple myeloma, low molecular weight heparin significantly reduced the incidence of blood clots when compared with the vitamin K antagonist warfarin, while the difference with aspirin was not significant. There were no major bleeds with low molecular weight heparin or warfarin, and in participants treated with aspirin the rate was below 1%. One study evaluated unfractionated heparin and did not report on venous thromboembolism or major bleeding. There was no mention of blood clots in the two study groups. Data for warfarin in comparison with placebo were too limited to support the use of warfarin in the prevention of blood clots in cancer patients. One study in children evaluated antithrombin, which had no significant effect on blood clots or major bleeding when compared with no antithrombin. A small pilot study evaluated the oral anticoagulant apixaban and found a low rate of bleeding and blood clots compared to placebo.

Quality of the evidence

The quality of the included studies ranged from low to high, such that future studies may change our confidence in the estimates and the size of the estimates, in particular with regard to the safety of anticoagulants. The quality of findings ranged from high to very low across the different outcomes and comparisons. The main limiting factors, which were the reason for a decrease in quality in some outcomes, were imprecision and risk of bias. The relatively low number of studies, participants, and clinical events prevented us from determining the potential influence of age and type or stage of cancer on treatment effects and providing more definitive conclusions about the risk of bleeding in association with anticoagulants. None of the studies tested intermittent pneumatic compression or graduated elastic stockings for the prevention of venous thromboembolism.

Summary of findings

Background

Cancer is often complicated by venous thromboembolism (VTE), which can present as deep vein thrombosis (DVT) or pulmonary embolism (PE), or both (Khorana 2009). Cancer patients with VTE have a two‐fold or greater increased mortality compared with cancer patients without thrombosis, which could be explained by the development of fatal PEs or by a worse prognosis for patients with those cancers complicated by VTE (Sorensen 2000). VTE in cancer patients may be hard to recognise due to aspecific symptoms, which may overlap and be confused with symptoms caused by the underlying cancer disease process or cancer treatments. VTE carries significant morbidity due to the need for hospitalisation and an increased risk of recurrent VTE or bleeding complications while on anticoagulation (Hutten 2000; Prandoni 2002). The occurrence of symptomatic or unsuspected VTE may delay the delivery of cancer treatments such as chemotherapy, with a further negative impact on morbidity and potentially mortality. In addition, the occurrence of venous thromboembolic events brings further emotional strain for patients and their families, which negatively impacts their quality of life. Finally, the costs related to the management of VTE may be considerable, resulting from the expenses related to the drugs and hospitalisation (Heit 2015).

Description of the condition

The incidence of VTE is higher in people with cancer compared with people without cancer (Heit 2015). Compared with an incidence of about 0.1% in the general population, the rate of VTE in people with cancer has been reported to vary between 0.6% and about 8% (Khorana 2009). Chemotherapy has been recognised as an independent predictor for symptomatic VTE, with reported rates of from 11%, in Otten 2004, up to 75%, in Heit 2015 and Khorana 2009, depending on the type of chemotherapeutic agent used. The risk of thrombosis in cancer patients receiving chemotherapy seems to vary based on the stage of the disease, ranging from 3% to 5% in patients with early‐stage cancer to 30% in those with metastatic or advanced malignancy (Khorana 2009). The benefit‐risk ratio of primary prophylaxis in ambulatory patients with cancer who are receiving chemotherapy is not well established, and current guidelines do not recommend routine thromboprophylaxis in such patients (Lyman 2013).

Description of the intervention

Currently available drugs for the prevention of VTE are the parenteral anticoagulants, which include unfractionated heparin (UFH), low molecular weight heparins (LMWH), and fondaparinux, vitamin K antagonists (VKAs), and the oral direct thrombin inhibitor dabigatran and the factor Xa inhibitors rivaroxaban, apixaban, and edoxaban. In fact, each one of these agents may present disadvantages for long‐term prophylaxis in the ambulatory patient with cancer. Heparins and fondaparinux, as well as the new ultra‐low molecular weight heparin (uLMWH) semuloparin, require daily subcutaneous injections, which represent a considerable burden for the patient. Of note is that marketing applications for the uLMWH semuloparin have been withdrawn worldwide, and it is therefore unlikely to ever be commercially available (EMEA 2012). Both the direct thrombin inhibitor dabigatran and the factor Xa inhibitors offer the potential advantages of an oral route of administration, and in comparison with VKAs do not require routine laboratory monitoring and have fewer pharmacological interactions. VKAs and direct thrombin or factor Xa inhibitors can be difficult to administer in cancer patients with nausea or vomiting.

The use of pharmacological prophylaxis may be more challenging in people with cancer. The efficacy of pharmacological thromboprophylaxis could be reduced by the intrinsic procoagulant state induced by the cancer itself, prothrombotic treatments for cancer (e.g. chemotherapy), as well as the decline in the patient's general condition leading to immobilisation. On the other hand, the risk of bleeding events could be high even with prophylactic doses because of a number of predisposing factors such as the bleeding tendency at the site of the cancer, the relative decrease in the number of platelets in the blood (thrombocytopenia) secondary to chemotherapy, and the concomitant use of drugs (e.g. bevacizumab) that affect the vessel wall integrity (Kamphuisen 2014).

Currently available mechanical interventions for the prevention of VTE include intermittent pneumatic compression and graduated elastic stockings. These non‐pharmacological interventions may be a valid option in cancer patients who are at risk of bleeding, however evidence supporting their benefit and assuring no harm is limited.

Why it is important to do this review

The overall burden of VTE in people with cancer is steadily increasing as a result of an aging population, greater awareness, frequent staging assessments using sensitive imaging techniques, prothrombotic anticancer treatments, as well as the growing cancer population that is due to the aforementioned aging (Heit 2015). Provision of widespread primary thromboprophylaxis for ambulatory cancer patients who receive chemotherapy may help in preventing this complication. However, the efficacy of thromboprophylaxis needs to be balanced against the risks, such as (major) bleeding events.

Objectives

Our main objective was to assess the efficacy and safety of primary thromboprophylaxis for VTE in ambulatory patients with cancer receiving chemotherapy compared with placebo or no thromboprophylaxis. The secondary objective was to compare the efficacy and safety of different types of primary thromboprophylaxis by stratifying the main results per type of drug or mechanical intervention, and by aggregating results from head‐to‐head comparisons.

Methods

Criteria for considering studies for this review

Types of studies

All randomised controlled trials (RCTs) and quasi‐randomised trials were eligible.

Types of participants

Ambulatory outpatients at the time of randomisation or study entry. Participants of any age (including paediatric patients) with either a solid or haematological cancer, at any stage, and receiving chemotherapy were eligible. We excluded studies of participants receiving anticoagulation for a previous VTE or an indication other than VTE if data could not be extracted separately for participants not on anticoagulants. We excluded studies evaluating prophylaxis for catheter‐related thrombosis, since this is already the subject of another Cochrane review (Akl 2011).

Types of interventions

Interventions included any oral or parenteral anticoagulant (for example UFH, LMWH, uLMWH, fondaparinux, direct thrombin or factor Xa inhibitors, VKAs) or mechanical intervention (intermittent pneumatic compression or graduated elastic stockings), or both, used to prevent VTE in ambulatory patients with cancer that were receiving chemotherapy. Comparison interventions included either no thromboprophylaxis in the form of an inactive control intervention (placebo, no treatment, standard care) or an active control intervention (a different scheme or regimen of the same intervention, a different pharmacological type of prophylaxis, a different type of non‐pharmacological prophylaxis). We considered any frequency or duration of administration, dosage or intensity, and timing of delivery of pharmacological prophylaxis.

Types of outcome measures

Primary outcomes

The main effectiveness outcome was symptomatic VTE, objectively verified by means of Doppler (compression) ultrasonography or venography for DVT, and spiral computed tomography, ventilation/perfusion lung scan, or pulmonary angiography for PE.

The main safety outcome was major bleeding, typically defined as overt bleeding associated with a fall in haemoglobin of 2 g/dL or more, or leading to a transfusion of two or more units of packed red blood cells or whole blood; bleeding that occurred at a critical site (intracranial, intraspinal, intraocular, pericardial, intra‐articular, intramuscular with compartment syndrome, retroperitoneal); or bleeding contributing to death (Schulman 2005).

Secondary outcomes

Secondary outcomes included symptomatic PE; symptomatic DVT; unsuspected (incidental) VTE; overall (symptomatic and unsuspected) VTE; clinically relevant bleeding (major and clinically relevant non‐major bleeding); minor bleeding; one‐year overall mortality; arterial thromboembolic events; superficial venous thrombosis; quality of life; and number of participants experiencing any serious adverse event. Clinically relevant non‐major bleeding is typically defined as overt bleeding that does not meet the criteria for major bleeding, but is associated with the need for medical intervention, contact with a physician, or interruption of the study drug or with discomfort or impairment of activities of daily life. Minor bleeding was defined as a bleeding event not matching the criteria for major bleeding or clinically relevant non‐major bleeding. Serious adverse events were defined as events resulting in patient hospitalisation, prolongation of hospitalisation, persistent or significant disability, congenital abnormality or birth defect of offspring, life‐threatening events or death. For trials using LMWH as the intervention or control, heparin‐induced thrombocytopenia (HIT) and the incidence of osteoporosis, as defined by the trial authors, were recorded. We considered all outcomes as binary outcomes except for quality of life, which we considered a continuous outcome.

For the 'Summary of findings' tables, we selected the following outcomes as the most patient‐relevant ones:

symptomatic VTE;
major bleeding;
symptomatic PE;
symptomatic DVT;
one‐year mortality;
clinically relevant bleeding.

Search methods for identification of studies

Electronic searches

For this update, the Cochrane Vascular Information Specialist (CIS) searched the following databases for relevant trials:

Cochrane Vascular Specialised Register (23 June 2016);
Cochrane Central Register of Controlled Trials (CENTRAL) (2016, Issue 5) via the Cochrane Register of Studies Online.

See Appendix 1 for details of the search strategy used to search CENTRAL.

The Cochrane Vascular Specialised Register is maintained by the CIS and is constructed from weekly electronic searches of MEDLINE Ovid, Embase Ovid, CINAHL, AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings which have been searched, as well as the search strategies used are described in the Specialised Register section of the Cochrane Vascular module in the Cochrane Library (www.cochranelibrary.com).

The CIS searched the following clinical trials registries (last searched 23 June 2016) by combining the search terms 'cancer' and 'thrombosis':

ClinicalTrials.gov (www.clinicaltrials.gov);
World Health Organization International Clinical Trials Registry Platform (www.who.int/trialsearch);
ISRCTN Register (www.isrctn.com/).

Searching other resources

The review authors searched the reference lists of identified studies and contacted content experts and trialists for relevant references. One review author screened the conference proceedings of the American Society of Clinical Oncology (from 2009 to 2015) and the International Society of Thrombosis and Haemostasis (from 2003 to 2016), combining the search terms of 'venous thrombosis', 'vein thrombosis', or 'pulmonary embolism' with 'cancer' or 'tumour'. We included studies if we could obtain adequate information from either the abstract or personal communication.

Data collection and analysis

Selection of studies

Two review authors (MDN, MC) independently reviewed the titles and abstracts identified from the database searches to determine whether they met the inclusion criteria. Any disagreements were resolved through discussion between the review authors. The review authors were not blinded to the journal, institution, or results of the study. We applied no language restrictions. We reassessed studies with insufficient information if we were able to obtain additional information from the trial authors. We documented reasons for excluding studies. In the event of multiple reports relating to the same trial, we considered all of the reports.

Data extraction and management

Two review authors (MDN, MC) independently extracted the data from the included studies onto standardised forms, resolving any disagreements by consensus or by involving a third review author (AR). We collected information on methodological quality, participant characteristics, characteristics of the intervention and control groups, and outcome characteristics of every group of participants. Whenever possible, we extracted the results from an intention‐to‐treat analysis. If we could not calculate effect sizes, we contacted the trial authors for additional data.

Assessment of risk of bias in included studies

Two review authors independently assessed randomisation, blinding, and adequacy of analyses (Higgins 2011; Rutjes 2009). Disagreements were resolved by consensus.

We assessed two components of randomisation: generation of allocation sequence and concealment of allocation. We considered generation of the allocation sequence to be adequate if it resulted in an unpredictable allocation schedule. Mechanisms considered to be adequate included random number tables, computer‐generated random numbers, minimisation, coin tossing, shuffling cards, and drawing lots. We considered trials using an unpredictable allocation sequence to be randomised. We considered trials using potentially predictable allocation mechanisms, such as alternation or allocation of participants according to date of birth, date of presentation, or case record number, to be quasi‐randomised.

We considered concealment of allocation to be adequate if participants and the investigators responsible for participant selection were unable to predict before allocation which treatment was next. Methods considered adequate included central randomisation; pharmacy‐controlled randomisation using identical, pre‐numbered containers; and sequentially numbered, sealed, opaque envelopes. We considered blinding of participants and therapists to be adequate if experimental and control preparations were explicitly described as indistinguishable, or if a double‐dummy technique was used. We considered assessors to be blinded if this was explicitly mentioned by the investigators.

We considered analyses to be at low risk of bias if all randomised participants were included in the analyses according to the intention‐to‐treat principle. We classified the item 'free of selective reporting' as at low risk of bias if we had both the protocol and the full report of a given study, where the full report presented results for all outcomes listed in the protocol. We classified a study as at high risk of bias if a report did not present data on all outcomes reported in either the protocol or the methods section. We did not consider the item 'free of other bias' in this review. We assessed the reporting of primary outcomes and sample size calculations. Finally, we used GRADE to describe the quality of the overall body of evidence (Guyatt 2008; Higgins 2011), defined as the extent of our confidence in the estimates of treatment benefits and harms.

Measures of treatment effect

We have presented results as summary risk ratios (RRs) for dichotomous variables, determining a 95% confidence interval (CI) for each estimate. We used inverse‐variance random‐effects model meta‐analysis to combine the trials (DerSimonian 1986). In the case of statistically significant overall estimates, we also calculated clinical effect summary statistics such as the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH) to express the final results of the review.

Assessment of heterogeneity

We measured heterogeneity of the treatment effect between trials using the variance estimate Tau², as currently recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). A Tau² of 0.04 is typically interpreted to indicate low heterogeneity, 0.09 moderate heterogeneity, and 0.16 high heterogeneity across trials (Rutjes 2012; Spiegelhalter 2004).

Assessment of reporting biases

We evaluated publication bias and other biases related to small study size using funnel plots, plotting the RRs on the vertical axis against their standard errors on the horizontal axis (Sterne 2001). Funnel plot symmetry would be expected in the absence of any bias related to small study size. We used the Harbord–Egger's test to assess symmetry (Harbord 2006). We further explored any anomaly in stratified analyses, in which we investigated the effects of differences in types of LMWH, age, type of cancer, and suboptimal design choices on the magnitude of the effects.

Data synthesis

In the main analyses, we analysed and presented data by stratifying for the type of thromboprophylaxis used.

We planned to explore the between‐trial heterogeneity by stratifying the main outcomes for the following trial characteristics: age (below or equal to 65 years versus above 65 years); type of cancer, stage of cancer (metastatic versus non‐metastatic); type of major bleeding (according to the definition provided by Schulman 2005 versus unclear or different definition); concealment of allocation (adequate versus inadequate or unclear); blinding (adequate versus inadequate or unclear); analysis in accordance with the intention‐to‐treat principle (yes versus no or unclear); trial size (large versus small); and differences in the use of co‐interventions in the trial groups. We planned to use univariate random‐effects model meta‐regression to determine whether treatment effects were affected by these factors and by three continuous variables at trial level: dosage of intervention, treatment duration, and length of follow‐up (Thompson 1999).

We defined large trials as those randomising 1360 participants or more. This threshold was informed by the following sample size calculation: according to the Khorana 2008 score, cancer patients at high risk of symptomatic VTE who could potentially receive thromboprophylaxis have an incidence of 7.1% over a period of about 3 months. Assuming that the risk of VTE can be halved by thromboprophylaxis, a total number of about 1360 participants would be needed to detect a statistically significant difference between groups.

We performed the data analysis in Review Manager 5 (RevMan 2014). We performed stratified analyses and funnel plot exploration in STATA release 14 (Stata 2015).

'Summary of findings' table

We presented the main findings of the review concerning the quality of the evidence, magnitude of effect of the interventions examined, and sum of available data in a 'Summary of findings' table, according to the GRADE principles described by Higgins 2011 and Guyatt 2008. For the critical outcome symptomatic VTE, we applied cutoffs to define high and intermediate risk groups. We used a cutoff of 7% to define high risk, which is in line with the cutoff proposed by Khorana 2008, which is between 6.7% and 7.1% over a period of about 3 months. We used event rates between 2% and 7% to define intermediate risk groups to experience symptomatic VTEs (Khorana 2008).

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies.

Results of the search

See Figure 1.

Following title and abstract screening, we considered 27 reports (21 trials) to be potentially eligible for this update. We included eight reports related to five new trials (Elit 2012; Khorana 2015; Lecumberri 2013; Macbeth 2016; Vadhan‐Raj 2013), and 10 reports were related to previously included trials. We added four reports to the Characteristics of ongoing studies table (NCT00717938; NCT02048865; NCT02285738; NCT02555878), and two to the Characteristics of studies awaiting classification table (Ciftci 2012; NCT00771563). We added two reports to the excluded studies (Bocharov 2011; Gercor 2013). We assessed one report that was a review as not relevant.

Included studies

For this update we included five new studies (Elit 2012; Khorana 2015; Lecumberri 2013; Macbeth 2016; Vadhan‐Raj 2013). Four of these were reported as ongoing studies in the previous version of the review (Elit 2012; Khorana 2015; Macbeth 2016; Vadhan‐Raj 2013), and one was an excluded study (Lecumberri 2013). We added four additional ongoing studies (NCT00717938; NCT02048865; NCT02285738; NCT02555878), and added two additional studies to Studies awaiting classification (Ciftci 2012; NCT00771563).

In total, 26 RCTs randomising a total of 12,352 participants were eligible for inclusion in the review. The treatments evaluated consisted of the uLMWH semuloparin (Agnelli 2012), LMWH (Agnelli 2009; Altinbas 2004; Elit 2012; Haas 2012; Kakkar 2004; Khorana 2015; Klerk 2005; Larocca 2012; Lecumberri 2013; Macbeth 2016; Maraveyas 2012; Palumbo 2011; Pelzer 2015; Perry 2010; Sideras 2006; Vadhan‐Raj 2013; van Doormaal 2011; Zwicker 2013), UFH (Lebeau 1994), the VKA warfarin (Chahinian 1989; Levine 1994; Maurer 1997; Palumbo 2011; Zacharski 1981), antithrombin (Mitchell 2003), and the oral direct factor Xa inhibitor apixaban (Levine 2012). None of the included RCTs used non‐pharmacological prophylaxis, or pharmacological thromboprophylaxis with fondaparinux, the direct thrombin inhibitor dabigatran, or the direct factor Xa inhibitors rivaroxaban and edoxaban. In 15 (8056 participants, 65%) of the 26 studies, inclusion was restricted to people with locally advanced or metastatic cancer, in two studies limited cancer was included, in four studies both early and advanced disease were included, while in the remaining studies the stage was not clear (see Characteristics of included studies).

One study assessed the uLMWH semuloparin versus placebo.

Agnelli 2012 recruited participants (n = 3212) with metastatic or locally advanced solid cancer of the lung, pancreas, stomach, colon or rectum, bladder, or ovary and randomised them to the uLMWH semuloparin (20 mg once daily) versus placebo starting on the first day of a first or new regimen of chemotherapy. The intervention was continued for three months unless chemotherapy was stopped earlier.

Eigtheen studies assessed LMWH.

Sixteen studies assessed LMWH either versus placebo or no thromboprophylaxis (Agnelli 2009; Altinbas 2004; Haas 2012; Kakkar 2004; Khorana 2015; Klerk 2005; Lecumberri 2013; Macbeth 2016; Maraveyas 2012; Pelzer 2015; Perry 2010; Sideras 2006; Vadhan‐Raj 2013; van Doormaal 2011; Zwicker 2013), or compared different doses from prophylactic to full therapeutic of LMWH with each other (Elit 2012). These 16 trials varied in the duration and type of LMWH, including eight weeks to 48 months of subcutaneous (sc) dalteparin, enoxaparin, certoparin, nadroparin, bemiparin. The dose of LMWH was prophylactic in the majority of the studies and intermediate, in Pelzer 2015, or therapeutic, in Maraveyas 2012, in one study each. In two studies initial therapeutic LMWH was followed by intermediate doses (Klerk 2005; van Doormaal 2011). Fifteen of these 16 studies reported a mean age at study entry of 65 years or younger, whereas Zwicker 2013 included participants with a mean age above 65.

Agnelli 2009 recruited participants (n = 1150) with metastatic or locally advanced lung, gastrointestinal, pancreatic, breast, ovarian, or head and neck cancer and randomised them to nadroparin (3800 IU anti‐factor Xa sc, once daily) versus placebo. Study treatment started on the same day as chemotherapy and was given for the duration of the chemotherapy or up to a maximum of 120 days (± 10 days).

Altinbas 2004 recruited participants (n = 84) with histologically confirmed small cell lung carcinoma and randomised them to standard anticancer treatment with or without dalteparin (5000 IU sc, once daily). Dalteparin was stopped with disease progression or at the end of the 18 weeks of chemotherapy.

Elit 2012 recruited women (n = 77) with newly diagnosed epithelial ovarian cancer and randomised them to receive standard chemotherapy and one of three subcutaneous doses of dalteparin (50 IU/kg, 100 IU/kg, or 150 IU/kg), once daily during the first three of six cycles of three‐weekly chemotherapy.

Haas 2012 recruited participants with metastatic breast cancer (n = 353) or non‐small cell lung carcinoma (n = 547) and receiving first‐ or second‐line chemotherapy. Participants were randomised to six months of certoparin (3000 IU sc, once daily) versus placebo.

Kakkar 2004 recruited participants (n = 385) with histologically confirmed locally advanced or metastatic malignant disease of the breast, lung, gastrointestinal tract, pancreas, liver, genitourinary tract, ovary, or uterus and randomised them to dalteparin (5000 IU sc, once daily) versus placebo. Study treatment was given for one year or until the participant died, whichever occurred sooner.

Khorana 2015 recruited cancer patients (n = 98) at high risk for VTE (Khorana score ≥ 3) who initiated a new systemic chemotherapy regimen and randomised them to dalteparin (5000 IU sc, once daily) versus no thromboprophylaxis for 12 weeks.

Klerk 2005 recruited participants (n = 302) with metastasised or locally advanced solid tumours and randomised them to nadroparin versus placebo. Study treatment was given using pre‐ﬁlled syringes containing a fixed volume of nadroparin (anti‐factor Xa 9500 U/mL) or placebo according to the participant’s weight: 0.4 mL for those weighing less than 50 kg, 0.6 mL for those weighing between 50 kg and 70 kg, and 0.8 mL for those weighing more than 70 kg. Study treatment was to be administered sc twice daily during the initial 14 days of treatment and once daily thereafter for another four weeks.

Lecumberri 2013 recruited participants (n = 39) with newly diagnosed, limited‐stage small cell lung cancer and randomised them to standard chemoradiotherapy alone or combined with bemiparin 3500 IU daily for a maximum of 26 weeks.

Macbeth 2016 recruited participants (n = 2202) with histopathologic or cytologic diagnosis of primary bronchial carcinoma of any stage and histology (small cell or non‐small cell) and randomised them to standard anticancer treatment (including active supportive or palliative care) with or without dalteparin (5000 IU sc, once daily) for a maximum of 24 weeks.

Maraveyas 2012 recruited participants (n = 123) with advanced pancreatic cancer and randomised them to dalteparin (200 IU/kg sc, once daily for four weeks followed by 150 IU/kg for a further eight weeks) in combination with gemcitabine versus gemcitabine alone. Continuing dalteparin prophylaxis after 12 weeks was not recommended, but was left to the discretion of the investigator.

Pelzer 2015 recruited participants (n = 312) with histologically or cytologically confirmed advanced pancreatic cancer. Participants were randomised to standard anticancer treatment with or without enoxaparin (1 mg/kg once daily) for three months, started simultaneously with palliative systemic chemotherapy; after 12 weeks of initial chemotherapy, all participants who had not progressed received the standard therapy with or without enoxaparin (40 mg once daily) for an additional three months.

Perry 2010 recruited participants (n = 186) with newly diagnosed, pathologically confirmed World Health Organization (WHO) grade 3 or grade 4 glioma and randomised them to six months of dalteparin (5000 IU sc, once daily) versus placebo starting within the first month after surgery. Participants were allowed to continue the study medication for 12 months.

Sideras 2006 recruited participants (n = 138) with advanced breast cancer who did not respond to first‐line chemotherapy, advanced prostate cancer resistant to primary hormonal therapy, advanced lung cancer, or advanced colorectal cancer. In the first part of the study, participants were randomised to dalteparin (5000 IU sc, once daily) versus placebo, while in the second part participants were randomised to dalteparin (5000 IU sc, once daily) plus standard clinical care versus standard clinical care alone. Dalteparin (or placebo) was given for 18 weeks or until disease progression.

Vadhan‐Raj 2013 recruited participants (n = 75) with advanced stage (unresectable or metastatic) adenocarcinoma of the pancreas planning to initiate systemic chemotherapy and randomised them to chemotherapy with or without dalteparin (5000 IU sc, once daily) for 16 weeks.

van Doormaal 2011 recruited participants (n = 503) with non–small cell lung cancer (stage IIIB), hormone‐refractory prostate cancer, or locally advanced pancreatic cancer and randomised them to standard anticancer treatment with or without nadroparin. Subcutaneous nadroparin was administered for six weeks (two weeks at therapeutic dose and four weeks at half therapeutic dose). The participants were eligible to receive additional cycles of nadroparin (two weeks at therapeutic dose and four weeks washout period) for a maximum of six cycles.

Zwicker 2013 recruited participants (n = 34) with histologically confirmed advanced stage malignancy, which included adenocarcinoma of the pancreas (locally advanced or metastatic), colorectal (stage IV), non‐small cell lung cancer (stage III or IV), relapsed or stage IV ovarian, or surgically unresectable or metastatic gastric adenocarcinoma. Participants were randomised to enoxaparin (40 mg sc, once daily) for two months or observation.

In two additional studies, LMWH was compared against an active control:

Larocca 2012 recruited participants (n = 342) with newly diagnosed multiple myeloma treated with lenalidomide and low‐dose dexamethasone induction and melphalan‐prednisone‐lenalidomide consolidation. Participants were randomised to aspirin (100 mg per day) or LMWH (enoxaparin 40 mg per day). Prophylaxis was provided during the four (28‐day) cycles of induction and the six (28‐day) cycles of consolidation therapy.

Palumbo 2011 recruited participants (n = 667) with previously untreated myeloma who received thalidomide‐containing regimens and randomised them to aspirin (100 mg per day), low‐dose warfarin (1.25 mg per day) or LMWH (enoxaparin 40 mg per day). The prophylaxis was administered during the three cycles of induction therapy in participants ≤ 65 years of age and during the first six cycles of induction therapy in participants > 65 years.

Four studies compared the VKA warfarin against no thromboprophylaxis or placebo.

Chahinian 1989 recruited participants (n = 328) with extensive carcinoma of the lung and randomised them to warfarin (dose to maintain a prothrombin time 1.5 to twice the control values) versus no warfarin. Warfarin was continued throughout the course of chemotherapy.

Levine 1994 recruited participants (n = 311) with metastatic stage IV breast carcinoma who had been receiving first‐ or second‐line chemotherapy for four weeks or less and randomised them to warfarin (target of international normalised ratio (INR) 1.3 to 1.9) versus placebo. Study treatment began either at the start of chemotherapy or within the following four weeks and continued until one week after termination of chemotherapy.

Maurer 1997 recruited participants (n = 347) with limited‐stage small cell lung cancer who were to receive chemotherapy and radiotherapy and randomised them to warfarin or no warfarin. Warfarin (dose of 10 mg per day for the first three days and then at a dose to maintain the prothrombin time between 1.4 and 1.6 times the local institutional control standards) was continued through the complete course of chemotherapy and radiation therapy and was stopped three weeks after the last cycle of chemotherapy.

Zacharski 1981 recruited participants (n = 50) with small cell lung cancer and randomised them to warfarin (dose to prolong the prothrombin time to approximately two times the control value) versus no warfarin.

UFH, antithrombin, and the factor Xa inhibitor apixaban were evaluated against placebo or no thromboprophylaxis in one study each.

Lebeau 1994 recruited participants (n = 277) with limited and extensive small cell lung cancer who had not been previously treated with chemotherapy or radiotherapy. The dose of UFH was initially adapted to weight (500 IU/kg/day), then adjusted by clotting times (different techniques used, and results had to be between two and three times the control value). UFH was administered in two or three daily injections for five weeks and stopped one week after the second course of chemotherapy.

Levine 2012 recruited participants (n = 125) receiving either first‐ or second‐line chemotherapy for advanced or metastatic lung, breast, gastrointestinal, bladder, ovarian, or prostate cancer; cancer of unknown origin; myeloma; or selected lymphomas. Participants were randomised to apixaban 5 mg (n = 32), 10 mg (n = 30), 20 mg (n = 33), and placebo (n = 30). The study treatment was given for 12 weeks, beginning within four weeks of starting chemotherapy.

Mitchell 2003 recruited paediatric patients (n = 85) newly diagnosed with acute lymphoblastic leukaemia and randomised them to receive, or not, weekly infusions of antithrombin.

Excluded studies

We excluded two additional studies in this update (Bocharov 2011; Gercor 2013) making for a total of 24 excluded studies. The reasons for exclusion were: design other than an RCT (Baz 2005; Bocharov 2011; Kessler 2011; Meister 2008; Minnema 2004; Paydas 2008; Zangari 2003); studies on perioperative thromboprophylaxis (Bergqvist 1983; Heilmann 1995; Hills 1972; Macintyre 1974; Maxwell 2000; Sideras 2007; Welti 1981); inclusion of hospitalised cancer patients (Eichinger 2008; Haas 2011; Poniewierski 1987; Weber 2008); no relevant outcomes reported (Rajan 1995); no eligible intervention (Niesvizky 2007); prophylaxis was for catheter‐related thrombosis (Kwaan 2007). Three studies were terminated early: Levin 2008 because of a drug supply issue; Gercor 2013 due to the lack of eligible patients; Pandya 2002 with no reason for study termination reported.

We added four studies to ongoing studies (NCT00717938; NCT02048865; NCT02285738; NCT02555878).

Risk of bias in included studies

The 'Risk of bias' summary is shown in Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

The random sequence was adequately generated in 19 studies (Agnelli 2009; Agnelli 2012; Chahinian 1989; Elit 2012; Haas 2012; Kakkar 2004; Klerk 2005; Larocca 2012; Lecumberri 2013; Levine 1994; Levine 2012; Macbeth 2016; Maraveyas 2012; Mitchell 2003; Palumbo 2011; Pelzer 2015; Perry 2010; van Doormaal 2011; Zacharski 1981), but was unclear in the remaining studies due to poor reporting.

Allocation was adequately concealed in 18 studies (Agnelli 2009; Agnelli 2012; Elit 2012; Kakkar 2004; Klerk 2005; Larocca 2012; Lebeau 1994; Lecumberri 2013; Levine 2012; Macbeth 2016; Maraveyas 2012; Mitchell 2003; Palumbo 2011; Pelzer 2015; Perry 2010; Sideras 2006; van Doormaal 2011; Zwicker 2013), and was unclear in the remaining studies due to poor reporting.

Blinding

Eight studies had a double‐blinded design (Agnelli 2009; Agnelli 2012; Haas 2012; Kakkar 2004; Klerk 2005; Levine 1994; Levine 2012; Perry 2010), and 14 were open studies (Altinbas 2004; Elit 2012; Larocca 2012; Lebeau 1994; Lecumberri 2013; Macbeth 2016; Maraveyas 2012; Mitchell 2003; Palumbo 2011; Pelzer 2015; Sideras 2006; Vadhan‐Raj 2013; van Doormaal 2011; Zwicker 2013). In four studies blinding was unclear due to poor reporting (Chahinian 1989; Khorana 2015; Maurer 1997; Zacharski 1981).

Incomplete outcome data

Fourteen studies performed the analysis according to the intention‐to‐treat principle (Agnelli 2012; Elit 2012; Khorana 2015; Klerk 2005; Larocca 2012; Lebeau 1994; Lecumberri 2013; Macbeth 2016; Maraveyas 2012; Pelzer 2015; Perry 2010; Vadhan‐Raj 2013; Zacharski 1981; Zwicker 2013), while in nine studies the percentages of participants randomised and subsequently excluded from the analyses ranged from 1.3% to 10% (Agnelli 2009; Chahinian 1989; Haas 2012; Kakkar 2004; Levine 1994; Levine 2012; Palumbo 2011; Sideras 2006; van Doormaal 2011); we considered these to be at high risk of bias. The study involving paediatric patients used a per‐protocol analysis and excluded 22% of the participants that were initially enrolled (Mitchell 2003); we considered this study to be at high risk of bias. Attrition bias was unclear in Altinbas 2004 and Maurer 1997.

Selective reporting

We judged 18 studies to be free of selective reporting (Agnelli 2009; Agnelli 2012; Altinbas 2004; Elit 2012; Haas 2012; Kakkar 2004; Khorana 2015; Klerk 2005; Larocca 2012; Lebeau 1994; Lecumberri 2013; Levine 1994; Levine 2012; Macbeth 2016; Maraveyas 2012; Mitchell 2003; Sideras 2006; van Doormaal 2011). In four studies one or more outcomes that were reported in the results were not anticipated in the methods sections of the publications (Chahinian 1989; Maurer 1997; Vadhan‐Raj 2013; Zacharski 1981). In four studies not all outcomes were reported in the results (Palumbo 2011; Pelzer 2015; Perry 2010; Zwicker 2013).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9; Table 10

Summary of findings for the main comparison. Semuloparin versus placebo.

Semuloparin compared with placebo for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: semuloparin Comparison: placebo
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		Without semuloparin	With semuloparin
Symptomatic VTE: Follow‐up: median 3.5 months	RR 0.36 (0.22 to 0.60)	Intermediate‐risk population³		22 per 1000 fewer events (14 to 27 fewer)	3212 (1)	⊕⊕⊕⊕ high	Semuloparin decreases the incidence of symptomatic VTE across different cancer types.
		34 per 1000	12 per 1000  (8 to 21)
Major bleeding Follow‐up: median 3.5 months	RR 1.05 (0.55 to 2.0)	Intermediate‐risk population³		1 per 1000 more events (5 fewer to 11 more)	3172 (1)	⊕⊕⊝⊝ low⁴	Semuloparin may increase major bleedings across different cancer types.
		11 per 1000	12 per 1000  (6 to 23)
Symptomatic PE Follow‐up: median 3.5 months	RR 0.48 (0.22 to 1.01)	Intermediate‐risk population³		7 per 1000 fewer events (0 to 10 fewer)	3212 (1)	⊕⊕⊕⊝ moderate⁵	Semuloparin probably decreases the incidence of symptomatic PE across different cancer types.
		13 per 1000	6 per 1000  (3 to 13)
Symptomatic DVT Follow‐up: median 3.5 months	RR 0.32 (0.16 to 0.63)	Intermediate‐risk population³		14 per 1000 fewer events (8 to 18 fewer)	3212 (1)	⊕⊕⊕⊕ high	Semuloparin decreases the incidence of symptomatic DVT across different cancer types.
		21 per 1000	7 per 1000  (3 to 13)
1‐year mortality Follow‐up: 1 year	RR 1.02 (0.96 to 1.08)	Intermediate‐risk population³		11 per 1000 more events (22 fewer to 44 more)	3212 (1)	⊕⊕⊕⊝ moderate⁵	Semuloparin probably has no effect on 1‐year mortality across different cancer types.
		555 per 1000	566 per 1000  (533 to 599)
Clinically relevant bleeding Follow‐up: median 3.5 months	RR 1.40 (0.90 to 2.19)	Intermediate‐risk population³		8 per 1000 more events (2 fewer to 24 more)	3172 (1)	⊕⊕⊕⊝ moderate⁵	Semuloparin probably increases the incidence of clinically relevant bleeding across different cancer types.
		20 per 1000	28 per 1000  (18 to 44)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Low molecular weight heparin compared with no thromboprophylaxis for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: LMWH Comparison: no thromboprophylaxis (placebo or no LMWH)
Outcomes	Relative effect (95% CI)	Illustrative comparative risk (95% CI)*		Difference (95% CI)²	No of participants (studies)	Quality of the evidence (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		No thromboprophylaxis	With LMWH
Symptomatic VTE Follow‐up: median 10 months	RR 0.54 (0.38 to 0.75)	High‐risk population³		33 per 1000 fewer events (18 to 44 fewer)	3284 (9)	⊕⊕⊕⊕    high⁴	LMWH decreases the incidence of symptomatic VTE across different cancer types.
		71 per 1000	39 per 1000 (27 to 54)
Major bleeding Follow‐up: median 12 months	RR 1.44 (0.98 to 2.11)	High‐risk population³		8 per 1000 more major bleeds (0 to 19 more)	6356 (13)	⊕⊕⊝⊝    low⁵	LMWH may increase major bleedings across different cancer types.
		17 per 1000	25 per 1000 (17 to 36)
Symptomatic PE Follow‐up: median 10 months	RR 0.59 (0.40 to 0.86)	High‐risk population³		6 per 1000 fewer events (2 to 8 fewer)	5226 (7)	⊕⊕⊝⊝    low⁶	LMWH may decrease the incidence of symptomatic PE across different cancer types.
		14 per 1000	8 per 1000 (6 to 12)
Symptomatic DVT Follow‐up: median 10 months	RR 0.49 (0.35 to 0.67)	High‐risk population³		15 per 1000 fewer events (10 to 19 fewer)	5310 (8)	⊕⊕⊕⊕ high⁷	LMWH decreases the incidence of symptomatic DVT across different cancer types.
		29 per 1000	14 per 1000
1‐year mortality Follow‐up: median 12 months	RR 0.93 (0.80 to 1.09)	High‐risk population³		41 per 1000 fewer deaths (117 fewer to 53 more)	2304 (8)	⊕⊕⊝⊝    low⁸	LMWH may decrease the incidence of death across different cancer types.
		587 per 1000	546 per 1000 (470 to 640)
Clinically relevant bleeding Follow‐up: median 12 months	RR 3.40 (1.20 to 9.63)	High‐risk population³		40 per 1000 more clinically relevant bleeds (3 to 145 more)	3105 (4)	⊕⊕⊕⊝ moderate⁹	LMWH probably increases the incidence of clinically relevant bleeding across different cancer types.
		17 per 1000	57 per 1000 (20 to 162)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; LMWH: low molecular weight heparin; PE: pulmonary embolism; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Low molecular weight heparin: prophylactic dose compared with intermediate or therapeutic dose for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: prophylactic dose LMWH Comparison: intermediate or therapeutic dose LMWH
Outcomes	Control type	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
			Assumed risk¹	Corresponding risk
			Intermediate/ therapeutic dose	Prophylactic dose
			Intermediate‐risk population³
Symptomatic VTE: Follow‐up: median 3.5 months	Intermediate	RR 2.89 (0.12 to 66.75)	31 per 1000	90 per 1000 (4 to 2086)	59 per 1000 more events (28 fewer to 2055 more)	51 (1)	⊕⊕⊝⊝ low⁴	Prophylactic‐dose LMWH may be associated with a higher risk of symptomatic VTE when compared to intermediate‐dose LMWH in ovarian cancer.
Symptomatic VTE: Follow‐up: median 3.5 months	Therapeutic	RR 1.00 (0.07 to 15.15)	53 per 1000	53 per 1000  (4 to 805)	0 per 1000 fewer events (49 fewer events to 752 more)	52 (1)	⊕⊕⊝⊝ low⁴	We do not know if prophylactic‐dose LMWH is associated with a higher risk of symptomatic VTE when compared to therapeutic‐dose LMWH in ovarian cancer.
			Intermediate‐risk population³
Major bleeding Follow‐up: median 3.5 months	Intermediate	Not estimable⁵	NA	NA	NA	NA	NA	As we have insufficient data to estimate the relative risk, we do not know how prophylactic‐dose LMWH affects major bleeding in ovarian cancer.
Major bleeding Follow‐up: median 3.5 months	Therapeutic	Not estimable⁵	NA	NA	NA	NA	NA
			Intermediate‐risk population³
Symptomatic PE Follow‐up: median 3.5 months	Intermediate	RR 2.89 (0.12 to 66.75)	NA⁶	NA	NA	NA	NA	As we have insufficient data to estimate the assumed risk, we do not know how prophylactic‐dose LMWH affects symptomatic PE.
Symptomatic PE Follow‐up: median 3.5 months	Therapeutic	RR 3.00 (0.13 to 70.42)	NA⁶	NA	NA	NA	NA
			Intermediate‐risk population³
Symptomatic DVT Follow‐up: median 3.5 months	Intermediate	Not estimable⁵	NA	NA	NA	NA	NA	We do not know how prophylactic‐dose LMWH affects symptomatic DVT.
Symptomatic DVT Follow‐up: median 3.5 months	Therapeutic	RR 0.33 (0.01 to 7.82)	53 per 1000	18 per 1000  (1 to 415)	36 per 1000 fewer DVT (53 fewer to 362 more)	52 (1)	⊕⊕⊝⊝ low⁴	Prophylactic‐dose LMWH may reduce the risk of symptomatic DVT when compared to therapeutic‐dose LMWH in ovarian cancer, although this seems an implausible finding.
			Intermediate‐risk population³
1‐year mortality Follow‐up: NA	Intermediate	NA⁷	NA	NA	NA	NA	NA	We do not know how prophylactic‐dose LMWH affects all‐cause mortality when compared to intermediate or therapeutic‐dose LMWH.
1‐year mortality Follow‐up: NA	Therapeutic	NA⁷	NA	NA	NA	NA	NA
			Intermediate‐risk population³
Clinically relevant bleeding Follow‐up: median 3.5 months	Intermediate	NA⁵	NA	NA	NA	NA	NA	We do not know how prophylactic‐dose LMWH affects clinically relevant bleeding.
	Therapeutic	RR 0.33 (0.01 to 7.82)	38 per 1000⁸	13 per 1000  (0 to 301)	26 per 1000 fewer clinically relevant bleeding (38 fewer to 262 more)	52 (1)	⊕⊕⊝⊝ low⁴	Prophylactic‐dose LMWH may reduce clinically relevant bleeding when compared to therapeutic‐dose LMWH in ovarian cancer.
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; LMWH: low molecular weight heparin; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Unfractionated heparin compared with no thromboprophylaxis for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: unfractionated heparin Comparison: no thromboprophylaxis
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		Without UFH	With UFH
Symptomatic VTE: Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how UFH affects symptomatic VTE.
		NA	NA
Major bleeding Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how UFH affects major bleeding.
		NA	NA
Symptomatic PE Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how UFH affects symptomatic PE.
		NA	NA
Symptomatic DVT Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how UFH affects symptomatic DVT.
		NA	NA
1‐year mortality Follow‐up: 1 year	RR 0.86 (0.72 to 1.03)	Population at unclear risk⁴		98 per 1000 fewer events (195 fewer to 21 more)	277 (1)	⊕⊕⊝⊝ moderate⁵	UFH probably decreases the incidence of 1‐year mortality in small cell lung cancer.
		698 per 1000	600 per 1000  (502 to 719)
Clinically relevant bleeding Follow‐up: median not reported,  maximum of 4.9 years of follow‐up	RR 2.01 (0.18 to 21.96)	Population at unclear risk⁴		7 per 1000 more events (6 fewer to 151 more)	277 (1)	⊕⊝⊝⊝ low⁶	UFH may increase the risk of clinically relevant bleeding in small cell lung cancer.
		7 per 1000	14 per 1000  (1 to 158)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; UFH: unfractionated heparin; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Vitamin K antagonists compared with placebo or no thromboprophylaxis for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: VKA Comparison: placebo or no thromboprophylaxis
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		Without VKA	With VKA
Symptomatic VTE: Follow‐up: mean 6 months	RR 0.15 (0.02 to 1.2)	Intermediate‐risk population³		37 per 1000 fewer events (43 fewer to 9 more)	311 (1)	⊕⊕⊝⊝  low⁴	VKA may reduce the incidence of symptomatic VTE in breast cancer.
		44 per 1000	7 per 1000  (1 to 53)
Major bleeding Follow‐up: mean 6 months	RR 3.82 (0.97 to 15.04)	Intermediate‐risk population³		18 per 1000 more events (0 fewer to 88 more)	994 (4)	⊕⊕⊝⊝  low⁵	VKA may increase the incidence of major bleeding in breast cancer and small cell lung cancer.
		6 per 1000	24 per 1000  (6 to 95)
Symptomatic PE Follow‐up: mean 6 months	RR 1.05 (0.07 to 16.58)	Intermediate‐risk population³		0 per 1000 fewer events (6 fewer to 101 more)	311 (1)	⊕⊝⊝⊝  very low⁶	We have very little confidence in the estimated effect of VKA on symptomatic PE in breast cancer.
		6 per 1000	7 per 1000  (0 to 108)
Symptomatic DVT Follow‐up: mean 6 months	RR 0.08 (0 to 1.42)	Intermediate‐risk population³		35 per 1000 fewer events (38 fewer to 16 more)	311 (1)	⊕⊕⊝⊝  low⁴	VKA may reduce the incidence of symptomatic DVT in breast cancer.
		38 per 1000	3 per 1000  (0 to 54)
1‐year mortality Follow‐up: NA	NA	Intermediate‐risk population³		NA	NA	NA	We do not know how VKA affects 1‐year mortality.
		NA	NA
Clinically relevant bleeding Follow‐up: NA	NA	Intermediate‐risk population³		NA	NA	NA	We do not know how VKA affects clinically relevant bleeding.
		NA	NA
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VKA: vitamin K antagonists; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Vitamin K antagonists compared with aspirin for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: VKA Comparison: aspirin
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		With aspirin	With VKA
Symptomatic VTE: Follow‐up: median 2.1 years	RR 1.50 (0.74 to 3.04)	Intermediate‐risk population³		27 per 1000 more events (14 fewer to 211 more)	440 (1)	⊕⊕⊕⊝  moderate⁴	VKA probably increases the incidence of symptomatic VTE when compared to aspirin in multiple myeloma.
		55 per 1000	82 per 1000  (40 to 166)
Major bleeding Follow‐up: median 2.1 years	RR 0.14 (0.01 to 2.75)	Intermediate‐risk population³		12 per 1000 fewer events (14 fewer to 24 more)	440 (1)	⊕⊕⊝⊝  low⁵	VKA may reduce the incidence of major bleeding when compared to aspirin in multiple myeloma.
		14 per 1000	2 per 1000  (0 to 38)
Symptomatic PE Follow‐up: median 2.1 years	RR 1.00 (0.25 to 3.95)	Intermediate‐risk population³		0 per 1000 fewer events (14 fewer to 54 more)	440 (1)	⊕⊕⊕⊝  moderate⁴	VKA is probably as effective as aspirin in the prevention of symptomatic PE in multiple myeloma.
		18 per 1000	18 per 1000  (5 to 72)
Symptomatic DVT Follow‐up: median 2.1 years	RR 1.75 (0.75 to 4.09)	Intermediate‐risk population³		27 per 1000 more events (9 fewer to 112 more)	440 (1)	⊕⊕⊕⊝  moderate⁴	VKA probably increases the incidence of symptomatic DVT when compared to aspirin in multiple myeloma.
		36 per 1000	64 per 1000  (27 to 149)
1‐year mortality Follow‐up: NA	NA⁶	Intermediate‐risk population³		NA	NA	NA	We do not know how VKA affects 1‐year mortality when compared to aspirin in multiple myeloma.
		NA	NA
Clinically relevant bleeding Follow‐up: NA	NA⁶	Intermediate‐risk population³		NA	NA	NA	We do not know how VKA affects clinically relevant bleeding when compared to aspirin in multiple myeloma.
		NA	NA
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VKA: vitamin K antagonists; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Antithrombin compared with placebo for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: antithrombin Comparison: placebo
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		Without antithrombin	With antithrombin  (any dosage)
Symptomatic VTE: Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how antithrombin affects symptomatic VTE.
		NA	NA
Major bleeding Follow‐up: median 4 months	RR 0.78 (0.03 to 18.57)	Population at unclear risk⁴		4 per 1000 fewer events (16 fewer to 293 more)	85 (1)	⊕⊝⊝⊝  very low⁵	We have very little confidence in the estimated effect of antithrombin on the incidence of major bleeding.
		17 per 1000	13 per 1000  (1 to 310)
Symptomatic PE Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how antithrombin affects symptomatic PE.
		NA	NA
Symptomatic DVT Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how antithrombin affects symptomatic DVT.
		NA	NA
1‐year mortality Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how antithrombin affects 1‐year mortality.
		NA	NA
Clinically relevant bleeding Follow‐up: NA	NA³	Population at unclear risk⁴		NA	NA	NA	We do not know how antithrombin affects clinically relevant bleeding.
		NA	NA
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Apixaban compared with placebo for primary thromboprophylaxis in ambulatory cancer patients receiving chemotherapy
Patient or population: ambulatory cancer patients receiving chemotherapy Settings: outpatient clinics Intervention: apixaban Comparison: placebo
Outcomes	Relative effect (95% CI)	*Illustrative comparative risks (95% CI)**		Difference²   (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	What it means
		Assumed risk¹	Corresponding risk
		Without apixaban	With apixaban (any dosage)
Symptomatic VTE: Follow‐up: median 3 months	RR 0.08 (0.01 to 0.67)	High‐risk population³		66 per 1000 fewer events (24 to 71 fewer)	122 (1)	⊕⊕⊕⊝  moderate⁴	Apixaban probably decreases the incidence of symptomatic VTE across different cancer types.
		71 per 1000	6 per 1000  (1 to 48)
Major bleeding Follow‐up: median 3 months	RR 0.62 (0.06 to 6.63)	High‐risk population³		6 per 1000 fewer events (16 fewer to 96 more)	122 (1)	⊕⊕⊕⊝  low⁵	We are unsure whether apixaban increases the incidence of major bleeding across different cancer types.
		17 per 1000	11 (1 to 113)
Symptomatic PE Follow‐up: median 3 months	RR 0.11 (0.00 to 2.54)	High‐risk population³		12 per 1000 fewer events (14 fewer to 22 more)	122 (1)	⊕⊕⊝⊝  low⁵	We are unsure whether apixaban decreases the incidence of symptomatic PE across different cancer types.
		14 per 1000	2 per 1000 (0 to 36)
Symptomatic DVT Follow‐up: median 3 months	RR 0.08 (0.01 to 0.67)	High‐risk population³		27 per 1000 fewer events (10 to 29 fewer)	122 (1)	⊕⊕⊕⊝  moderate⁴	Apixaban probably decreases the incidence of symptomatic DVT across different cancer types.
		29 per 1000	2 per 1000  (0 to 19)
1‐year mortality Follow‐up: NA	NA⁶	High‐risk population³		NA	0 (0)	NA	We do not know how apixaban affects all‐cause mortality.
		NA	NA
Clinically relevant bleeding Follow‐up: median 3 months	RR 1.87 (0.23 to 14.91)	High‐risk population³		30 per 1000 more events (27 fewer to 480 more)	122 (1)	⊕⊕⊝⊝  low⁷	We are unsure how much apixaban increases the incidence of clinically relevant bleeding across different cancer types.
		34 per 1000	64 per 1000 (8 to 514)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; DVT: deep vein thrombosis; NA: not applicable; PE: pulmonary embolism; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Variable	N of trials	N of participants   (LMWH)	N of participants   (control)	RR (95% CI)	Heterogeneity I²/Tau²	P for interaction
All trials	9	1849	1435	0.54 (0.38 to 0.75)	0.0/0.00
Type of LMWH						0.710
Dalteparin	5	458	443	0.64 (0.39 to 1.06)	0.0/0.00
Certoparin	1	442	441	0.57 (0.24 to 1.35)	NA
Nadroparin	1	769	381	0.50 (0.22 to 1.13)	NA
Enoxaparin	1	160	152	0.43 (0.21 to 0.88)	NA
Bemiparin	1	20	18	0.10 (0.01 to 1.75)	NA
Type of dosage						0.322
Prophylactic	7	1630	1223	0.61 (0.41 to 0.92)	0.0/0.00
Higher than prophylactic	2	219	212	0.41 (0.23 to 0.75)	0.0/0.00
Type of cancer						0.610
Mixed	3	828	555	0.70 (0.33 to 1.51)	0.0/0.00
Lung	4	529	404	0.40 (0.20 to 0.80)	0.0/0.00
Pancreatic	2	219	212	0.41 (0.23 to 0.75)	0.0/0.00
Glioma	1	99	87	0.74 (0.35 to 1.57)	NA
Breast cancer	1	174	177	0.76 (0.17 to 3.36)	NA
Allocation concealment						0.960
Adequate	7	1365	952	0.54 (0.37 to 0.78)	0.0/0.00
Inadequate or unclear	2	484	483	0.55 (0.24 to 1.26)	0.0/0.00
Blinding of participants						0.353
Double‐blind	4	1500	1093	0.62 (0.40 to 0.96)	0.0/0.00
Inadequate or unclear blinding	5	349	342	0.44 (0.26 to 0.74)	0.0/0.00
Intention‐to‐treat analysis						0.627
Yes	4	338	317	0.50 (0.31 to 0.79)	0.0/0.00
No or unclear	5	1511	1118	0.59 (0.36 to 0.97)	0.0/0.00
Selective outcome reporting						0.857
Adequate	7	1590	1196	0.52 (0.34 to 0.82)	0.0/0.00
Incomplete or unclear	2	259	239	0.56 (0.33 to 0.95)	5.6/0.01

Date	Event	Description
9 July 2016	New citation required but conclusions have not changed	Searches rerun. Five additional studies were added to the included studies. Two additional studies excluded on full‐text basis. New authors joined the review team. 'Summary of findings' tables added. Conclusions not changed.
9 July 2016	New search has been performed	Searches rerun. Five additional studies were added to the included studies. Two additional studies excluded on full‐text basis.

Date	Event	Description
24 July 2013	New citation required but conclusions have not changed	Searches rerun. Twelve additional studies were added to the included studies and nine additional studies to the excluded studies. Risk of bias was reassessed in all included trials. Conclusions not changed. Change in author team.
24 July 2013	New search has been performed	Searches rerun. Twelve additional studies were added to the included studies and nine additional studies to the excluded studies.

Variable	N of trials	N of participants   (LMWH)	N of participants   (control)	RR (95% CI)	Heterogeneity I²/Tau²	P for interaction
All trials	12*	3378	2978	1.44 (0.98 to 2.11)	0.0/0.00
Type of LMWH						0.736
Dalteparin	6	1567	1552	1.29 (0.67 to 2.47)	0.0/0.00
Certoparin	1	447	451	2.19 (0.84 to 5.70)	NA
Nadroparin	3	1161	794	1.83 (0.69 to 4.85)	13.8/0.15
Enoxaparin	2	183	163	1.24 (0.56 to 2.73)	NA
Bemiparin	1	20	18	0.30 (0.01 to 6.97)	NA
Type of dosage						0.669
Prophylactic	8	2744	2340	1.56 (0.91 to 2.69)	0.0/0.00
Higher than prophylactic	4	611	627	1.32 (0.76 to 2.27)	0.0/0.00
Type of cancer						0.641
Mixed	6	1293	1027	1.24 (0.63 to 2.42)	0.0/0.00
Lung	4	1593	1472	1.49 (0.79 to 2.80)	0.0/0.00
Pancreatic	2	219	214	1.21 (0.58 to 2.51)	0.0/0.00
Glioma	1	99	87	4.39 (0.52 to 36.89)	NA
Breast cancer	1	174	178	7.16 (0.37 to 137.60)	NA
Definition of major bleeding						0.235
Standard**	9	3077	2697	1.70 (1.07 to 2.70)	0.0/0.00
Alternative or unclear	3	278	270	0.99 (0.50 to 1.98)	0.0/0.00
Trial size						0.917
Large	1	1101	1101	1.50 (0.62 to 3.66)	NA
Small	11	2254	1866	1.42 (0.93 to 2.18)	0.0/0.00
Allocation concealment						0.447
Adequate	10	2858	2468	1.33 (0.87 to 2.04)	0.0/0.00
Inadequate or unclear	2	497	499	2.00 (0.81 to 4.94)	0.0/0.00
Blinding of participants						0.175
Double‐blind	6	1897	1516	1.97 (1.11 to 3.51)	0.0/0.00
Inadequate or unclear blinding	6	1458	1451	1.11 (0.66 to 1.86)	0.0/0.00
Intention‐to‐treat analysis						0.941
Yes	7	1637	1622	1.45 (0.87 to 2.43)	0.0/0.00
No or unclear	5	1718	1345	1.41 (0.76 to 2.61)	6.0/0.03
Selective outcome reporting						0.989
Adequate	10	3096	2728	1.43 (0.91 to 2.25)	0.0/0.00
Incomplete or unclear	2	259	239	1.56 (0.59 to 4.11)	16.7/0.13

Search run on Thu Jun 23 2016

#1	MESH DESCRIPTOR Thrombosis	1211
#2	MESH DESCRIPTOR Thromboembolism	880
#3	MESH DESCRIPTOR Venous Thromboembolism	221
#4	MESH DESCRIPTOR Venous Thrombosis EXPLODE ALL TREES	1979
#5	(thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol* or microembol*):TI,AB,KY	16473
#6	MESH DESCRIPTOR Pulmonary Embolism EXPLODE ALL TREES	719
#7	(PE or DVT or VTE):TI,AB,KY	4294
#8	((vein* or ven) near thromb):TI,AB,KY	5939
#9	(blood near3 clot*):TI,AB,KY	2358
#10	(pulmonary near3 clot*):TI,AB,KY	5
#11	(lung near3 clot*):TI,AB,KY	4
#12	MESH DESCRIPTOR Antineoplastic Protocols EXPLODE ALL TREES	11008
#13	MESH DESCRIPTOR Survival EXPLODE ALL TREES	123
#14	surviv*:TI,AB,KY	51947
#15	chemotherap*:TI,AB,KY	36001
#16	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15	89289
#17	MESH DESCRIPTOR Anticoagulants EXPLODE ALL TREES	7943
#18	(anticoagul* or anti‐coagu*):TI,AB,KY	7083
#19	MESH DESCRIPTOR Heparin EXPLODE ALL TREES	3794
#20	parin:TI,AB,KY	51054
#21	UFH:TI,AB,KY	431
#22	LMWH:TI,AB,KY	775
#23	LMH:TI,AB,KY	6
#24	(Ariven or Arteven or Calcilean or Hepalean or Hepathrom or Leparan or Lipo‐Hepin or Liquaemin or Liquemin or Pabyrin or Pularin or Thromboliquine or Vetren ):TI,AB,KY	9
#25	(Clexane or klexane or lovenox ):TI,AB,KY	41
#26	(Fragmin):TI,AB,KY	178
#27	(Innohep):TI,AB,KY	11
#28	clivarin* :TI,AB,KY	20
#29	(danaproid or danaparoid):TI,AB,KY	35
#30	(antixarin):TI,AB,KY	2
#31	(Zibor or cy 222 or embolex or monoembolex):TI,AB,KY	38
#32	(rd 11885 or RD1185):TI,AB,KY	0
#33	(Kabi‐2165 or Kabi 2165):TI,AB,KY	39
#34	(emt‐966 or emt966 or emt‐967 or emt977 or pk‐10169 or pk10169):TI,AB,KY	8
#35	(fr‐860 or fr860 or cy‐216 or cy216):TI,AB,KY	51
#36	(kb101 or lomoparan or orgaran ):TI,AB,KY	28
#37	(fluxum or lohepa or lowhepa):TI,AB,KY	11
#38	(op 2123 or op2123):TI,AB,KY	1

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Symptomatic VTE: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Symptomatic VTE: LMWH vs no thromboprophylaxis	9	3284	Risk Ratio (IV, Random, 95% CI)	0.54 [0.38, 0.75]
2.1 Dalteparin	5	901	Risk Ratio (IV, Random, 95% CI)	0.64 [0.39, 1.06]
2.2 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.57 [0.24, 1.35]
2.3 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.50 [0.22, 1.13]
2.4 Enoxaparin	1	312	Risk Ratio (IV, Random, 95% CI)	0.43 [0.21, 0.88]
2.5 Bemiparin	1	38	Risk Ratio (IV, Random, 95% CI)	0.10 [0.01, 1.75]
3 Symptomatic VTE: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	0.51 [0.22, 1.17]
4 Symptomatic VTE: LMWH vs warfarin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
5 Symptomatic VTE: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
5.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	2.89 [0.12, 67.75]
5.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	1.0 [0.07, 15.15]
6 Symptomatic VTE: vitamin K antagonists vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
7 Symptomatic VTE: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
8 Symptomatic VTE: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Major bleeding: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Major bleeding: LMWH vs no thromboprophylaxis	13	6356	Risk Ratio (IV, Random, 95% CI)	1.44 [0.98, 2.11]
2.1 Dalteparin	6	3119	Risk Ratio (IV, Random, 95% CI)	1.29 [0.67, 2.47]
2.2 Certoparin	1	898	Risk Ratio (IV, Random, 95% CI)	2.19 [0.84, 5.70]
2.3 Nadroparin	3	1955	Risk Ratio (IV, Random, 95% CI)	1.83 [0.69, 4.85]
2.4 Enoxaparin	2	346	Risk Ratio (IV, Random, 95% CI)	1.24 [0.56, 2.73]
2.5 Bemiparin	1	38	Risk Ratio (IV, Random, 95% CI)	0.30 [0.01, 6.97]
3 Major bleeding: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	0.14 [0.01, 2.76]
4 Major bleeding: LMWH vs warfarin	1	440	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Major bleeding: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
5.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Major bleeding: vitamin K antagonists vs no thromboprophylaxis	4	994	Risk Ratio (IV, Random, 95% CI)	3.82 [0.97, 15.04]
7 Major bleeding: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
8 Major bleeding: antithrombin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
9 Major bleeding: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Abbreviation	Scientific description	Lay description
	Anticoagulation therapy	Blood‐thinning therapy
GES	Graduated elastic stockings	Special socks that improve blood flow in the leg veins and prevent blood from pooling in the legs
	Incidence	Number of newly diagnosed diseases, in this review cases of VTE
IPC	Intermittent pneumatic compression	A mechanical intervention using an air pump and inflatable leggings to provide pulsing pressure that pushes blood through the veins
	Primary prophylaxis	Primary protective treatment aiming at the prevention of disease development
	Thromboprophylaxis	Treatment to prevent the development of blood clots
VTE	Venous thromboembolism	Blood clots

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Symptomatic PE: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Symptomatic PE: LMWH vs no thromboprophylaxis	7	5226	Risk Ratio (IV, Random, 95% CI)	0.59 [0.40, 0.86]
2.1 Dalteparin	4	2881	Risk Ratio (IV, Random, 95% CI)	0.61 [0.40, 0.92]
2.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.50 [0.10, 2.44]
2.3 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.60 [0.14, 2.49]
2.4 Enoxaparin	1	312	Risk Ratio (IV, Random, 95% CI)	0.14 [0.01, 2.61]
3 Symptomatic PE: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	0.13 [0.02, 1.03]
4 Symptomatic PE: LMWH vs warfarin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
5 Symptomatic PE: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
5.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	2.89 [0.12, 67.75]
5.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	3.0 [0.13, 70.42]
6 Symptomatic PE: vitamin K antagonists vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
7 Symptomatic PE: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
8 Symptomatic PE: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Symptomatic DVT: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Symptomatic DVT: LMWH vs no thromboprophylaxis	8	5310	Risk Ratio (IV, Random, 95% CI)	0.49 [0.35, 0.67]
2.1 Dalteparin	5	2965	Risk Ratio (IV, Random, 95% CI)	0.50 [0.32, 0.78]
2.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.50 [0.19, 1.31]
2.3 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.46 [0.18, 1.20]
2.4 Enoxaparin	1	312	Risk Ratio (IV, Random, 95% CI)	0.45 [0.22, 0.93]
3 Symptomatic DVT: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	0.81 [0.32, 2.04]
4 Symptomatic DVT: LMWH vs warfarin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
5 Symptomatic DVT: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
5.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	0.33 [0.01, 7.82]
6 Symptomatic DVT: vitamin K antagonists vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
7 Symptomatic DVT: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
8 Symptomatic DVT: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidental VTE: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Incidental VTE: LMWH vs no thromboprophylaxis	4	4354	Risk Ratio (IV, Random, 95% CI)	0.66 [0.41, 1.08]
2.1 Dalteparin	2	2321	Risk Ratio (IV, Random, 95% CI)	0.62 [0.33, 1.14]
2.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.74 [0.21, 2.62]
2.3 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.75 [0.26, 2.14]
3 Incidental VTE: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
3.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	2.89 [0.12, 67.75]
3.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	3.0 [0.13, 70.42]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Overall VTE: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Overall VTE: LMWH vs no thromboprophylaxis	9	5366	Risk Ratio (IV, Random, 95% CI)	0.59 [0.48, 0.73]
2.1 Dalteparin	4	2494	Risk Ratio (IV, Random, 95% CI)	0.54 [0.41, 0.70]
2.2 Nadroparin	3	1955	Risk Ratio (IV, Random, 95% CI)	0.78 [0.48, 1.27]
2.3 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.65 [0.37, 1.15]
2.4 Enoxaparin	1	34	Risk Ratio (IV, Random, 95% CI)	0.16 [0.02, 1.36]
3 Overall VTE: prophylactic vs intermediate vs therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
3.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	4.81 [0.24, 95.58]
3.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	5.0 [0.25, 99.34]
4 Overall VTE: antithrombin vs placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Clinically relevant bleeding: semuloparin vs placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2 Clinically relevant bleeding: LMWH vs no thromboprophylaxis	4	3105	Risk Ratio (IV, Random, 95% CI)	3.40 [1.20, 9.63]
3 Clinically relevant bleeding: prophylactic vs intermediate vs therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
3.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	0.33 [0.01, 7.82]
4 Clinically relevant bleeding: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
5 Clinically relevant bleeding: UFH vs no thromboprophylaxis	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Minor bleeding: LMWH vs no thromboprophylaxis	7	2803	Risk Ratio (IV, Random, 95% CI)	1.23 [0.89, 1.70]
1.1 Dalteparin	4	717	Risk Ratio (IV, Random, 95% CI)	1.25 [0.72, 2.17]
1.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	1.00 [0.69, 1.45]
1.3 Certoparin	1	898	Risk Ratio (IV, Random, 95% CI)	1.96 [1.11, 3.46]
1.4 Bemiparin	1	38	Risk Ratio (IV, Random, 95% CI)	0.45 [0.09, 2.17]
2 Minor bleeding: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	0.70 [0.17, 2.84]
3 Minor bleeding: LMWH vs warfarin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
4 Minor bleeding: prophylactic vs intermediate or therapeutic LMWH	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
4.1 Prophylactic vs intermediate	1	51	Risk Ratio (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Prophylactic vs therapeutic	1	52	Risk Ratio (IV, Fixed, 95% CI)	0.2 [0.01, 3.97]
5 Minor bleeding: UFH vs no thromboprophylaxis	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
6 Minor bleeding: vitamin K antagonists vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
7 Minor bleeding: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
8 Minor bleeding: antithrombin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 1‐year mortality: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 1‐year mortality: LMWH vs no thromboprophylaxis	8	2304	Risk Ratio (IV, Random, 95% CI)	0.93 [0.80, 1.09]
2.1 Dalteparin	4	782	Risk Ratio (IV, Random, 95% CI)	0.97 [0.77, 1.21]
2.2 Nadroparin	2	1452	Risk Ratio (IV, Random, 95% CI)	0.95 [0.77, 1.18]
2.3 Enoxaparin	1	34	Risk Ratio (IV, Random, 95% CI)	0.72 [0.34, 1.51]
2.4 Bemiparin	1	36	Risk Ratio (IV, Random, 95% CI)	0.18 [0.05, 0.70]
3 1‐year mortality: UFH vs no thromboprophylaxis	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Symptomatic arterial thromboembolism: LMWH vs no thromboprophylaxis	4	3974	Risk Ratio (IV, Random, 95% CI)	0.74 [0.46, 1.18]
1.1 Dalteparin	2	2321	Risk Ratio (IV, Random, 95% CI)	0.90 [0.52, 1.53]
1.2 Nadroparin	2	1653	Risk Ratio (IV, Random, 95% CI)	0.38 [0.14, 1.03]
2 Symptomatic arterial thromboembolism: LMWH vs aspirin	2	781	Risk Ratio (IV, Random, 95% CI)	2.01 [0.37, 10.86]
3 Symptomatic arterial thromboembolism: LMWH vs warfarin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
4 Symptomatic arterial thromboembolism: vitamin K antagonists vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
5 Symptomatic arterial thromboembolism: warfarin vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Superficial venous thrombosis: LMWH vs no thromboprophylaxis	2	2033	Risk Ratio (IV, Random, 95% CI)	0.83 [0.30, 2.26]
1.1 Certoparin	1	883	Risk Ratio (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.83 [0.30, 2.26]
2 Superficial venous thrombosis: LMWH vs aspirin	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

PERMALINK

Primary prophylaxis for venous thromboembolism in ambulatory cancer patients receiving chemotherapy

Marcello Di Nisio

Ettore Porreca

Matteo Candeloro

Michele De Tursi

Ilaria Russi

Anne WS Rutjes

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

'Summary of findings' table

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Risk of bias in included studies

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Effects of interventions

Summary of findings for the main comparison. Semuloparin versus placebo.

Summary of findings 2. Low molecular weight heparin versus no thromboprophylaxis.

Summary of findings 3. Prophylactic versus intermediate or therapeutic dose low molecular weight heparin.

Summary of findings 4. Low molecular weight heparin versus aspirin.

Summary of findings 5. Low molecular weight heparin versus vitamin K antagonists.

Summary of findings 6. Unfractionated heparin versus no thromboprophylaxis.

Summary of findings 7. Vitamin K antagonists versus placebo or no thromboprophylaxis.

Summary of findings 8. Vitamin K antagonists versus aspirin.

Summary of findings 9. Antithrombin versus placebo.

Summary of findings 10. Apixaban versus placebo.

Anticoagulants versus control

uLMWH versus placebo

Primary outcomes

Secondary outcomes

LMWH versus placebo or no thromboprophylaxis

Primary outcomes

3.

4.

1. Results of stratified analyses on symptomatic venous thromboembolism.

5.

6.

2. Results of stratified analyses on major bleeding.

Secondary outcomes

LMWH versus active control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Serious adverse events: semuloparin vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected
2 Serious adverse events: LMWH vs no thromboprophylaxis	5	1531	Risk Ratio (IV, Random, 95% CI)	0.86 [0.70, 1.07]
2.1 Dalteparin	3	343	Risk Ratio (IV, Random, 95% CI)	1.22 [0.45, 3.34]
2.2 Nadroparin	1	1150	Risk Ratio (IV, Random, 95% CI)	0.89 [0.68, 1.17]
2.3 Bemiparin	1	38	Risk Ratio (IV, Random, 95% CI)	0.75 [0.43, 1.30]
3 Serious adverse events: apixaban vs placebo	1		Risk Ratio (IV, Fixed, 95% CI)	Totals not selected

Methods	PROTECHT study: multicentre, double‐blind, placebo‐controlled trial; modified intention‐to‐treat analysis, including participants who received at least 1 dose of study treatment
Participants	Ambulatory patients older than 18 years of age who were receiving chemotherapy for metastatic or locally advanced lung, gastrointestinal, pancreatic, breast, ovarian, or head and neck cancer. Mean age (SD): 62.1 (10.3) years in the nadroparin group; 63.7 (9.2) years in the placebo group; male sex, n (%): 372 (48.4) in nadroparin and 183 (48) in placebo; metastatic disease, n (%): not reported; previous VTE: 12 (1.6%) in nadroparin and 6 (1.6) in placebo. The median duration of follow‐up was 111 and 113 days in the nadroparin and placebo groups, respectively.
Interventions	Intervention: LMWH, nadroparin (3800 IU sc, once daily) Control: placebo Study treatment started on the same day as chemotherapy (the first cycle or a new course), and was given for the duration of chemotherapy or up to a maximum of 120 days (± 10 days).
Outcomes	Primary outcomes: composite of symptomatic venous or arterial thromboembolic events occurring during the study treatment plus 10 days; major bleeding that occurred between randomisation and 48 hours after the last injection of the study drug Secondary efficacy outcomes: incidental thromboembolic events incidentally diagnosed; survival at the end of study treatment and at 12 months; superficial venous thrombosis of the lower limbs; response to chemotherapy; central venous catheter‐related complications of possible thrombotic origin Secondary safety outcome: minor bleeding
Notes	Antiplatelet agents, oral anticoagulants, fibrinolytic agents, unfractionated heparin, or low molecular weight heparin other than nadroparin not allowed during the study period. Funding: Italfarmaco SpA, Milan, Italy Disclosure of potential conflicts of interest: The scientific director of Italfarmaco was involved as an author.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The randomisation list was generated by an independent statistician who used a standard permuted block of six without stratification. The list was generated with SAS version 8.2." Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "The allocation sequence was available online to the investigators using the Hypernet web‐based system. At the time the investigator accessed the web‐based system with personal codes (user ID and password) and requested the treatment allocation for a new patient who fulfilled the eligibility criteria, the system assigned the next free number in accordance with the randomisation sequence" Comment: adequate method of allocation concealment
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "Patients and investigators did not know whether study drug or placebo was being given, since pre‐filled syringes were used which were identical in appearance. Treatment assignments were masked from all study personnel and participants for the duration of the study." "All study outcomes were assessed by a central independent adjudication committee whose members were unaware of patients’ study‐group allocation" Comment: double‐blinded RCT and adequate methods of blinding
Incomplete outcome data (attrition bias)   All outcomes	High risk	Quote: "All randomised patients who received at least one dose  of the study treatment were included in the efficacy and  safety analyses" Comment: 769 out of 779 (98.7%) participants randomised were analysed in the LMWH group, 381 out of 387 (98.4%) randomised were analysed in the placebo group.
Selective reporting (reporting bias)	Low risk	All outcomes reported in the methods section were addressed in the results or discussion section.

Methods	SAVE‐ONCO study: multicentre, double‐blinded RCT, intention‐to‐treat for effectiveness and modified intention‐to‐treat analysis for safety outcomes, including participants who received at least 1 study dose
Participants	Patients with metastatic or locally advanced solid cancer of the lung, pancreas, stomach, colon or rectum, bladder, or ovary who were beginning a course of chemotherapy. Mean age: 59.8 years in the semuloparin group and 59.4 years in the placebo group. Male, n (%): 974 (60.6) in semuloparin and 956 (59.6) in placebo; metastatic disease: not reported; previous VTE: 2% in semuloparin and 2.3% in placebo. Mean duration of follow‐up: not reported
Interventions	Intervention: uLMWH semuloparin (20 mg sc, once daily)  Control: placebo The first dose of the study drug was administered on the first day of a course of chemotherapy (first regimen or a new regimen), continuing for the duration of chemotherapy (intended to be a minimum of 3 months). Median treatment duration was 3.5 months.
Outcomes	Primary efficacy outcome: composite of any symptomatic DVT, any non‐fatal PE, and death related to VTE Primary safety outcome: clinically relevant bleeding (major and non‐major) Secondary efficacy outcome: 1‐year overall survival or at the study end date
Notes	Funding, quote: "Supported by Sanofi". "The study was designed by the steering committee members and sponsored by Sanofi. Data were collected through a clinical research organization and analyzed by Sanofi. No Sanofi employees were members of the steering committee or the data and safety monitoring board" Disclosure of potential conflicts of interest: In the section "The Work Under Consideration for Publication", some of the authors declared they were employed by Sanofi or had received consulting fee or honorarium and support for travel to meetings by Sanofi‐Aventis.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was performed centrally by means of an interactive voice‐response system." Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "Randomization was performed centrally by means of an interactive voice‐response system." Comment: adequate method of allocation concealment
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "Efficacy and bleeding outcomes were assessed by a central independent adjudication committee, whose members were unaware of the study treatment" Comment: double‐blinded RCT and blinding of outcome assessors
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "All patients who underwent randomization were included in the primary efficacy population (intention‐to‐treat population), and those who underwent randomization and received at least one dose of the study treatment were included in the safety population" Comment: For safety, 1589 out of 1608 participants (98.8%) randomised are analysed in the uLMWH group, 1583 out of 1604 participants (98.7%) randomised are analysed in the placebo group.
Selective reporting (reporting bias)	Low risk	All outcomes reported in the protocol and in the methods section of the full report were addressed in the results or discussion section, except for 1 outcome mentioned in the protocol only: "Secondary efficacy variables include the initiation of curative treatment by the investigator after VTE". We did not consider this outcome to be relevant for the current review.

Methods	RCT; intention‐to‐treat analysis for survival outcomes
Participants	Patients between ages 18 and 75 years with histologically confirmed small cell lung carcinoma with an ECOG performance status of less than 3 and normal haematological, renal, and hepatic function tests. Median age: 58 years (range 34 to 75); gender, n: 33 males and 9 females in dalteparin, 35 males and 7 females in control; metastatic disease, n: 19 in dalteparin and 17 in control; previous VTE: 0/84. Median duration of follow‐up: 10 months (range 2 to 33 months)
Interventions	Intervention: LMWH, dalteparin (5000 IU sc, once daily) Control: no dalteparin Dalteparin was stopped with disease progression or at the end of the 18 weeks of chemotherapy. Median duration of treatment was 18 weeks.
Outcomes	Primary outcome: overall survival Secondary outcomes: progression‐free survival, side effects
Notes	Funding: not reported Disclosure of potential conflicts of interest: not disclosed, no COI forms available
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were randomized to receive either CT or CT plus LMWH" Comment: method of random sequence generation not reported
Allocation concealment (selection bias)	Unclear risk	Quote: "Patients were randomized to receive either CT or CT plus LMWH" Comment: method of allocation concealment not reported
Blinding (performance bias and detection bias)   All outcomes	High risk	Comment: The trial is reported as a "Chemotherapy‐only" versus Chemotherapy + LMWH" trial, without mentioning the use of placebo LMWH, or any attempt to blind.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Comment: For effectiveness is not reported. For safety, survival is analysed according to the intention‐to‐treat principle.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes reported in the methods section were addressed in the results or discussion section.

Methods	Multicentre, 3‐arm RCT, type of analyses not reported
Participants	Patients with extensive carcinoma of the lung. Patients aged 60 years or older: 55% warfarin and 60% control group; males: 68% and 67%, respectively. Metastatic or extensive disease: 100%; previous VTE: not reported
Interventions	Intervention: warfarin to maintain a prothrombin 1.5 to twice the control values Control: no warfarin Warfarin was continued throughout the course of chemotherapy, and it was withheld in participants with brain metastases during cranial irradiation and whenever platelet counts fell below 75,000/μL.
Outcomes	Main outcomes: overall survival, failure‐free survival, and cancer response (complete response, partial response, and objective response rate) to therapy Secondary outcomes: toxicity
Notes	Funding: grants from the National Cancer Institute, Department of Health and Human Services, and the T.J. Martell Foundation for Leukemia and Cancer Research Disclosure of potential conflicts of interest: not disclosed, no COI forms available 2 out of 3 available trial arms were considered in this review, as the chemotherapy provided was the same in both arms. The excluded trial arm provided a different chemotherapy regimen.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "allocation was determined by a Latin square arrangement balancing the sequence within and across institutions" Comment: adequate method of sequence generation; stratified randomisation, use of Latin square design
Allocation concealment (selection bias)	Unclear risk	Comment: method of allocation concealment not reported
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Comment: blinding not reported, use of placebo warfarin not reported
Incomplete outcome data (attrition bias)   All outcomes	High risk	Comment: 34 out of 328 participants (10%) enrolled in the study were not considered for the analysis. Exclusions per trial arm were not reported.
Selective reporting (reporting bias)	Unclear risk	Comment: All outcomes reported in the methods section were addressed in the results or discussion section. Toxicity was addressed in the results, but not explicitly reported as an outcome in the methods section.

Methods	Multicentre, open‐label, phase II randomised trial with 3 active treatment arms
Participants	Women with newly diagnosed epithelial ovarian cancer stage IIB to IV. Age, median (min‐max): 61 years (34, 74); female: 100%; metastatic disease: not reported; previous VTE, n (%): 4 (5)
Interventions	Intervention 1: standard adjuvant chemotherapy (taxane and platinum‐based) and dalteparin 50 IU/kg subcutaneously once daily during the first 3 of 6 cycles of 3‐weekly CT Intervention 2: standard adjuvant chemotherapy (taxane and platinum‐based) and dalteparin 100 IU/kg subcutaneously once daily during the first 3 of 6 cycles of 3‐weekly CT Intervention 3: standard adjuvant chemotherapy (taxane and platinum‐based) and dalteparin 150 IU/kg subcutaneously once daily during the first 3 of 6 cycles of 3‐weekly CT Study medication was started within 7 days prior to the first 21‐day cycle of chemotherapy and continued until day 21 of cycle 3.
Outcomes	Primary outcome: tumour response defined by ≥ 50% reduction in serum CA125 from baseline sustained for at least 28 days Secondary outcomes: major bleeding up to 24 hours after the last dose of dalteparin; any bleeding up to 24 hours after the last dose of dalteparin; symptomatic VTE up to 7 days after the last dose of dalteparin; death up to the last day of follow‐up; and compliance with dalteparin administration Participants were followed until the end of CT.
Notes	Funding, quote: "The Steering Committee wishes to acknowledge the financial support from both the Juravinski Cancer Centre Foundation and Pfizer Canada Inc" Disclosure of potential conflicts of interest, quote: "There are no financial disclosures from any of the authors related to this work except for Dr. Lee who has provided educational lectures and received financial reimbursement from Pfizer Canada Inc." Quote: "The study was terminated early due to poor recruitment."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Concealed randomization was performed centrally ... using a computer‐generated, permuted‐block randomization schedule." Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "Concealed randomization was performed centrally ... using a computer‐generated, permuted‐block randomization schedule." Comment: adequate method of allocation concealment
Blinding (performance bias and detection bias)   All outcomes	High risk	Quote: "Study outcomes were adjudicated by members of a Central Adjudication Committee masked to treatment assignment." Comment: open‐label study with blinded adjudication of outcomes
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "The primary analysis included all patients as randomized" Comment: All participants who were randomised were included in the analysis.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes indicated in the methods are presented in the results.

Methods	TOPIC‐1 and TOPIC‐2 studies: multicentre RCTs, intention‐to‐treat analysis for effectiveness and modified intention‐to‐treat analysis for safety outcomes
Participants	Patients with metastatic breast cancer (n = 353) or non‐small cell lung carcinoma (n = 547) receiving first‐ or second‐line chemotherapy. In the TOPIC‐1 study (breast cancer patients), the mean age (SD) was 54.6 (10.3) years in the certoparin group and 56.6 (11.0) years in the placebo group. In the TOPIC‐2 study (lung cancer patients), the mean age (SD) was 60.8 (9.5) years in the certoparin group and 60.3 (10.0) years in the placebo group. Males, n (%): TOPIC‐1: none, TOPIC‐2: 227 (83.2) in both certoparin and placebo; metastatic disease: not reported; previous VTE: 0/900
Interventions	Intervention: LMWH, certoparin (3000 IU sc, once daily) Control: placebo Study treatment was given for 6 months.
Outcomes	Primary outcomes: symptomatic or incidental VTE, major bleeding Secondary outcomes: symptomatic VTE, overall thrombosis rate (to include arterial thrombotic events, superficial venous thrombosis, and central‐line thrombosis), minor bleeding, thrombocytopenia, heparin‐induced thrombocytopenia, osteoporotic fractures, survival Post hoc: mortality, symptomatic or incidental VTE according to tumour stage
Notes	Funding: grant from Novartis Pharma, Nuremberg, Germany. Quote: "The TOPIC studies were supported by an unrestricted grant from Novartis Pharma GmbH, Germany." Disclosure of potential conflicts of interest, quote: "The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article." The study on breast cancer was prematurely halted after an interim analysis.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Using a computer‐generated randomisation list" and "Randomization was block‐stratified according to treatment with hormone‐based chemotherapy" Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomization numbers were allocated sequentially as patients were enrolled at each center." Comment: Concealment of allocation was poorly reported. It was not reported if sealed, opaque, and consecutively numbered envelopes, coded syringes, or other methods were used. In addition, it remains unclear what is meant by randomisation number in "Patients were allocated to the lowest available randomisation number available for each study center."
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "Efficacy outcomes were validated by a blinded, independent Central Thrombosis Evaluation Team; safety end points were validated by a Data Safety Monitoring Committee consisting of 2 clinicians (blinded to treatment) and an independent statistician with access to the treatment assignments." and "Only the external statistician from the Safety Committee had access to the randomization codes." Comment: double‐blinded, placebo‐controlled RCT with blinding of participants, physicians, and outcome assessors
Incomplete outcome data (attrition bias)   All outcomes	High risk	Comment: For effectiveness, 442 out of 447 (98.9%) in the LMWH group and 441 out of 453 (97.4%) in the placebo group were analysed. For safety, 447 out of 447 (100%) in the LMWH group and 451 out of 453 (99.6%) in the placebo group were analysed.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes reported in the methods section were addressed in the results or discussion section. However, the outcome osteoporotic fracture was incompletely reported; it remained unclear in which of the TOPIC‐2 trial arms the single event occurred. Post hoc analyses were reported transparently.

Methods	FAMOUS study: double‐blind, placebo‐controlled, multicentre RCT; modified intention‐to‐treat analysis for both effectiveness and safety analyses, including participants with at least 1 study dose and 1 follow‐up visit
Participants	Patients of 18 and 80 years with histologically confirmed advanced stage III or IV (locally advanced or metastatic) malignant disease of the breast, lung, gastrointestinal tract, pancreas, liver, genitourinary tract, ovary, or uterus. Age: 62 years in the dalteparin group and 60.9 years in the placebo group; female, n (%): 113 (59.5) in dalteparin and 100 (54.3) in placebo; metastatic disease, n: 161 in both dalteparin and placebo; previous VTE: 0/385. Median duration of follow‐up: 10 months in the dalteparin group and 9 months in the placebo group
Interventions	Intervention: LMWH, dalteparin (5000 IU sc, once daily) Control: placebo (0.9% normal saline) Study treatment given for 1 year or until the participant died, whichever occurred sooner
Outcomes	Primary outcomes: mortality after 1 year of therapy Secondary outcomes: symptomatic, objectively confirmed VTE disease and bleeding complications
Notes	Funding: Pharmacia Corp, New York, NY Disclosure of potential conflicts of interest: The lead author declared having acted as a consultant for Pfizer. Quote: "The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Acted as a consultant within the last 2 years: Ajay K. Kakkar, Pfizer. Received more than $2,000 a year from a company for either of the last 2 years: Ajay K. Kakkar, Pfizer"
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was performed centrally by computer‐generated code" Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "Randomization was performed centrally by computer‐generated code" Comment: adequate method of allocation concealment
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "placebo (0.9% normal saline), each supplied in 0.2‐mL prefilled  syringes" Comment: trial reported as double‐blind, with active substance or placebo provided in prefilled syringes. It is not reported whether syringes were identical in appearance.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Comment: Both for effectiveness and safety, 190 out of 196 (96.9%) were analysed in the LMWH group and 184 out of 189 (97.4%) were analysed in the placebo group.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes reported in the methods section were addressed in the results or discussion section.

Methods	PHACS trial: a multicentre RCT
Participants	Cancer patients at high risk for VTE (Khorana score ≥ 3) and initiating a new systemic chemotherapy regimen. Mean age (SD): dalteparin 60 (10) and observation 58 (12); gender, female/male: 21/29 dalteparin and 24/24 observation; metastatic disease: not reported; previous history of VTE, n (%): dalteparin 4 (8) and observation 2 (4)
Interventions	Intervention: LMWH, dalteparin 5000 IU daily subcutaneously for 12 weeks Control: no dalteparin
Outcomes	Primary outcome: any VTE over 12 weeks Participants in both arms were screened with lower extremity ultrasounds every 4 weeks of study. Primary safety endpoint was clinically relevant bleeding events over 13 weeks.
Notes	NCT00876915 Funding: not reported Disclosure of potential conflicts of interest: All authors report conflicts of interest. The study was terminated due to poor accrual.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "randomized to either dalteparin 5000 units daily subcutaneously or no prophylactic anticoagulation" Comment: method of sequence generation not reported
Allocation concealment (selection bias)	Unclear risk	Quote: "randomized to either dalteparin 5000 units daily subcutaneously or no prophylactic anticoagulation" Comment: method of allocation concealment not reported
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	Comment: It is not reported whether participants, personnel, and outcome assessors were blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Of 117 enrolled patients, 19 were not randomized due to the presence of VTE on initial screening (N =10, 8.5%) or for other reasons (N = 9)." Comment: All randomised participants were included in the analysis.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes indicated in the methods of the abstract are reported in the results.

Methods	MALT study: multicentre, double‐blinded, randomised, placebo‐controlled study with intention‐to‐treat analyses for both effectiveness and safety, including participants who received at least 1 study dose
Participants	Patients with metastasised or locally advanced solid tumours. Median age (range): 63 years (36 to 86) in the nadroparin group and 64 years (28 to 83) in the placebo group; male, n (%): 77 (52) in nadroparin and 81 (53) in placebo; metastatic disease, n (%): 137 (93) in nadroparin and 139 (90) in placebo; previous VTE: 0/302. Mean duration of follow‐up: 12 months
Interventions	Intervention: LMWH, nadroparin Control: placebo Preﬁlled syringes containing a fixed volume of nadroparin (9500 anti‐factor Xa U/mL) or placebo were provided according to participant’s weight: 0.4 mL for those weighing less than 50 kg, 0.6 mL for those weighing between 50 kg and 70 kg, and 0.8 mL for those weighing more than 70 kg. Study treatment was to be administered sc twice daily during the initial 14 days of treatment and once daily thereafter for another 4 weeks.
Outcomes	Primary efficacy outcome: death from any cause  Primary safety outcome: major bleeding Secondary safety outcome: clinically relevant non‐major bleeding
Notes	Funding: The study treatment was provided by Sanoﬁ‐Synthelabo (Paris, France). The authors state that "protocol design, data collection, and analysis were solely the responsibility of the authors". Disclosure of potential conflicts of interest: The senior author and statistician declared consultancy activities for various pharmaceutical companies, including Sanofi‐Synthelabo.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Sequentially numbered boxes of syringes with nadroparin or placebo were prepared using a central computer‐generated randomization schedule, stratified for body weight with blocks of four" Comment: adequate method of sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "Sequentially numbered boxes of syringes with nadroparin or placebo were prepared using a central computer‐generated randomization schedule, stratified for body weight with blocks of four" Comment: adequate method of allocation concealment
Blinding (performance bias and detection bias)   All outcomes	Low risk	Quote: "Prefilled syringes containing a fixed volume of nadroparin (9,500 antifactor Xa U/mL) or placebo were provided according to patient’s weight" Comment: trial reported as double‐blind, with active substance or placebo provided in prefilled syringes. It is not reported whether syringes were identical in appearance.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: All enrolled participants were included in the analysis.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes reported in the methods section were addressed in the results or discussion section. The authors reported reasons for the discontinuation of the study drug in the results section only, but this was for descriptive purposes, so unlikely to introduce bias.