Interventions for preventing venous thromboembolism in adults undergoing knee arthroscopy

Abstract

Background

Knee arthroscopy (KA) is a routine orthopedic procedure recommended to repair cruciate ligaments and meniscus injuries and in eligible patients, to assist the diagnosis of persistent knee pain. KA is associated with a small risk of thromboembolic events. This systematic review aims to assess if pharmacological or non‐pharmacological interventions may reduce this risk. This review is the second update of the review first published in 2007.

Objectives

To assess the efficacy and safety of interventions, whether mechanical, pharmacological, or in combination, for thromboprophylaxis in adult patients undergoing KA.

Search methods

For this update, the Cochrane Vascular Information Specialist searched the Cochrane Vascular Specialised Register, the CENTRAL, MEDLINE, Embase and CINAHL databases, and the World Health Organization International Clinical Trials Registry Platform and ClinicalTrials.gov trials registries, on 14 August 2019.

Selection criteria

We included randomized controlled trials (RCTs) and controlled clinical trials (CCTs), whether blinded or not, of all types of interventions used to prevent deep vein thrombosis (DVT) in males and females aged 18 years and older undergoing KA. There were no restrictions on language or publication status.

Data collection and analysis

Two authors independently selected studies for inclusion, assessed trial quality with the Cochrane 'Risk of bias' tool, and extracted data. A third author addressed discrepancies. We contacted study authors for additional information when required. We used GRADE to assess the certainty of the evidence.

Main results

This update adds four new studies, bringing the total of included studies to eight and involving 3818 adult participants with no history of thromboembolic disease undergoing KA. Studies compared daily subcutaneous (sc) low‐molecular‐weight heparin (LMWH) versus control (five studies); oral rivaroxaban 10 mg versus placebo (one study); daily sc LMWH versus graduated compression stockings (GCS) (one study); and aspirin versus control (one study).

The incidence of pulmonary embolism (PE) in all trials combined was low, with seven cases in 3818 participants.There were no deaths in any of the intervention or control groups.

LMWH versus control

When compared with control, LMWH probably results in little to no difference in the incidence of PE in patients undergoing KA (risk ratio (RR) 1.81, 95% confidence interval (CI) 0.49 to 6.65; 1820 participants; 3 studies; moderate‐certainty evidence). LMWH showed no reduction of the incidence of symptomatic DVT (RR 0.61, 95% CI 0.18 to 2.03; 1848 participants; 4 studies; moderate‐certainty evidence). LMWH may reduce the risk of asymptomatic DVT but the evidence is very uncertain (RR 0.14, 95% CI 0.03 to 0.61; 369 participants; 2 studies; very low‐certainty evidence). There was no evidence of an increased risk of all adverse events combined (RR 1.85, 95% CI 0.95 to 3.59; 1978 participants; 5 studies; moderate‐certainty evidence). No evidence of a clear effect on major bleeding (RR 0.98, 95% CI 0.06 to 15.72; 1451 participants; 1 study; moderate‐certainty evidence), or minor bleeding was observed (RR 1.79, 95% CI 0.84 to 3.84; 1978 participants; 5 studies; moderate‐certainty evidence).

Rivaroxaban versus placebo

One study with 234 participants compared oral rivaroxaban 10 mg versus placebo. No evidence of a clear impact on the risk of PE (no events in either group), symptomatic DVT (RR 0.16, 95% CI 0.02 to 1.29; moderate‐certainty evidence); or asymptomatic DVT (RR 0.95, 95% CI 0.06 to 15.01; very low‐certainty evidence) was detected. Only bleeding adverse events were reported. No major bleeds occurred in either group and there was no evidence of differences in minor bleeding between the groups (RR 0.63, 95% CI 0.18 to 2.19; moderate‐certainty evidence).

Aspirin versus control

One study compared aspirin with control. No PE, DVT or asymptomatic events were detected in either group. Adverse events including pain and swelling were reported but it was not clear what groups these were in. No bleeds were reported.

LMWH versus GCS

One study with 1317 participants compared the use of LMWH versus GCS. There was no clear difference in the risk of PE (RR 1.00, 95% CI 0.14 to 7.05; low‐certainty evidence). LMWH use did reduce the risk of DVT compared to people using GCS (RR 0.17, 95% CI 0.04 to 0.75; low‐certainty evidence). No clear difference in effects was seen between the groups for asymptomatic DVT (RR 0.47, 95% CI 0.21 to 1.09; very low‐certainty evidence); major bleeding (RR 3.01, 95% CI 0.61 to 14.88; moderate‐certainty evidence) or minor bleeding (RR 1.16, 95% CI 0.64 to 2.08; moderate‐certainty evidence). Levels of thromboembolic events were higher in the GCS group than in any other group.

We downgraded the certainty of the evidence for imprecision resulting from overall small event numbers; risk of bias due to concerns about lack of blinding, and indirectness as we were uncertain about the direct clinical relevance of asymptomatic DVT detection.

Authors' conclusions

There is a small risk that healthy adult patients undergoing KA will develop venous thromboembolism (PE or DVT). There is moderate‐ to low‐certainty evidence of no benefit from the use of LMWH, aspirin or rivaroxaban in reducing this small risk of PE or symptomatic DVT. There is very low‐certainty evidence that LMWH use may reduce the risk of asymptomatic DVT when compared to no treatment but it is uncertain how this directly relates to incidence of DVT or PE in healthy patients. No evidence of differences in adverse events (including major and minor bleeding) was seen, but data relating to this were limited due to low numbers of events in the studies reporting within the comparisons.

Plain language summary

Interventions for preventing venous thromboembolism in adults undergoing knee arthroscopy

Background

Knee arthroscopy (KA) is used for the treatment of many kinds of knee injuries. Patients may benefit from the intervention but there is also a small risk that they develop venous thromboembolism. Venous thromboembolism occurs when a blood clot forms in the deep veins of their legs. Patients may experience symptoms like calf pain, swelling (called symptomatic deep vein thrombosis, or DVT), and there is a risk that these clots may move to their lungs (called pulmonary embolism, or PE).

This systematic review included clinical trials that evaluated the use of medications that patients need to take after KA. These medications help to prevent blood clots by thinning the blood.

This review is an update of a review first published in 2007. Searches for this update were done on 14 August 2019.

Study characteristics and main results

Pooled data from five clinical trials show that the use of low‐molecular‐weight heparin (LMWH), a medicine that patients inject into their abdomens every day, does not clearly reduce the risk of PE or symptomatic DVT. The use of LWMH may reduce the risk of asymptomatic DVT. The use of LMWH had no clear effect on swelling or bleeding.

In one study, the use of oral rivaroxaban did not reduce the risk PE or DVT, and also did not increase any bleeding when compared to placebo. One study comparing aspirin with a control reported no PE or DVT. The use of LMWH compared to compression stockings did not reduce PE but did reduce DVT. There was higher‐than‐expected DVT in the group wearing compression stockings.

There is a small risk that healthy adult patients undergoing KA will develop venous thromboembolism (PE or DVT). There is uncertain evidence that the use of LMWH, aspirin or rivaroxaban is of benefit in reducing this small risk of PE or symptomatic DVT.

Certainty of the evidence

The certainty of evidence differed across interventions and outcomes. We downgraded the certainty of the evidence for PE and symptomatic DVT from high to moderate because the results were imprecise and because the number of cases was small. We downgraded the certainty of the evidence for asymptomatic clots to very low because of bias risk in the clinical trials, as participants were aware of what treatment they had. This may have influenced the medical staff who were looking for clots without symptoms. We also downgraded the certainty of evidence for indirectness, as there is uncertainty about the relevance of asymptomatic DVT in the context of KA.

Summary of findings

Summary of findings 1. Low‐molecular‐weight heparin compared to control (no prophylactic treatment) for preventing venous thromboembolism in adults undergoing knee arthroscopy.

LMWH compared to control (no prophylactic treatment) for preventing VTE in adults undergoing KA
Patient or population: preventing VTE in adults undergoing KA Setting: hospital Intervention: LMWH Comparison: control (no prophylactic treatment)
Outcomes	№ of participants (RCTs)	Relative effect (95% CI)	Anticipated absolute effects* (95% CI)		Certainty of the evidence (GRADE)	What happens
			Without LMWH	With LMWH
PE: assessed with CT arteriography (follow up: 30 days to 3 months)	1820 (3 RCTs)	RR 1.81 (0.49 to 6.65)	Study population		⊕⊕⊕⊝ MODERATE a	LMWH probably leads to no difference in rates of PE
			3 per 1000	6 per 1000 (2 to 22)
Symptomatic DVT: assessed with presence of symptoms and confirmed with a positive compression ultrasound (follow up: 30 days to 3 months)	1848 (4 RCTs)	RR 0.61 (0.18 to 2.03)	Study population		⊕⊕⊝⊝ LOW b	LMWH may lead to little or no reduction in symptomatic DVT
			11 per 1000	7 per 1000 (2 to 22)
Asymptomatic DVT: assessed with compression ultrasound (follow up: 30 days to 8 weeks)	369 (2 RCTs)	RR 0.14 (0.03 to 0.61)	Study population		⊕⊝⊝⊝ VERY LOW c	It is very uncertain if LMWH prevents asymptomatic DVT
			81 per 1000	11 per 1000 (2 to 49)
Mortality (follow up: 30 days to 3 months)	1978 (5 RCTs)	See comment			Not estimable	No deaths occurred in any of the included studies
All adverse events: including injections allergies, minor and major bleeding, others (follow up: 30 days to 3 months)	1978 (5 RCTs)	RR 1.85 (0.95 to 3.59)	Study population		⊕⊕⊕⊝ MODERATE d	No clear increased risk of adverse events
			13 per 1000	24 per 1000 (13 to 47)
Major bleeding (follow up: 30 days to 3 months)	1451 (1 RCT)	RR 0.98 (0.06 to 15.72).	Study population		⊕⊕⊕⊝ MODERATE e	LMWH probably leads to no difference in rates of major bleeding
			1 per 1000	1 per 1000 (0 to 22)
Minor bleeding (follow up: 30 days to 3 months)	1978 (5 RCTs)	RR 1.79 (0.84 to 3.84)	Study population		⊕⊕⊕⊝ MODERATE f	LMWH probably leads to no difference in rates of minor bleeding
			11 per 1000	20 per 1000 (9 to 43)
*The risk in the intervention group (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;KA: knee arthroscopy; LMWH: low‐molecular‐weight heparin; PE: pulmonary embolism; RCTs: randomized controlled trials; RR: risk ratio; VTE: venous thromboembolism
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aWe downgraded by one step due to imprecision.
^bWe downgraded by two steps due to imprecision and risk of bias.
^cWe downgraded by three steps to very low‐certainty due to imprecision, risk of bias and indirectness.
^dWe downgraded by one step due to imprecision.
^e.We downgraded by one step due to imprecision.
^fWe downgraded by one step due to imprecision.

Background

Description of the condition

Knee arthroscopy (KA) is indicated to repair the anterior cruciate ligament (ACL), meniscal tears, and in eligible patients, to assist the diagnosis of persistent knee pain (AAOS 2014). Complications of KA are uncommon, with rates of less than 1%. The most frequent complications are local hematoma, pyogenic arthritis and thromboembolic events. In a large cohort study of 20,770 KA procedures (Maletis 2012), the risk of deep venous thrombosis (DVT) was 0.25% (95% confidence interval (CI), 0.18% to 0.31%). For pulmonary embolism (PE) the risk was 0.017% (95% CI, 0.11% to 0.22%). Two prospective studies described incidence rates of asymptomatic DVT of 7.84% to 18% (Delis 2001; Demers 1998). The risk of thromboembolic events after KA was higher in smokers, women taking oral contraceptives, people with a previous history of thromboembolic diseases or a body mass index > 30.

Proximal DVT (above the popliteal vein) increases the risk of PE. Half of the patients presenting with proximal symptomatic DVT will have evidence of PE in lung scans (Kearon 2003). Immobilization and limb surgery also increase the risk for thromboembolic disease because of decreased blood flow and an increase in inflammatory markers that activate the coagulation cascade (Delis 2001; Maletis 2012). The clinical relevance of asymptomatic distal DVT is controversial. The consensus is that isolated asymptomatic calf DVT rarely progresses to symptomatic DVT (Gaskill 2015; Kearon 2003).

The diagnosis of asymptomatic DVT by sonography (ultrasound) has a variation in accuracy dependent on the technique used. Higher sensitivity is achieved using duplex ultrasound (71%), and higher specificity (98%) is seen using compression ultrasound alone (Goodacre 2005). The use of thromboprophylaxis is recommended in hip replacement and knee replacement surgeries as thromboembolic events are relatively frequent in those clinical circumstances, probably due to the longer duration of the surgery needed. In addition to this, it is not always possible to regain early mobility.

Description of the intervention

Clinical guideline recommendations for pharmacological prophylaxis in orthopedic surgery include use of apixaban, aspirin, dabigatran etexilate, fondaparinux sodium, low‐molecular‐weight heparin (LMWH) and rivaroxaban. The use of graduated elastic compression stockings (GCS) and intermittent pneumatic compression is indicated in the guidelines when there is a contra indication to use any other pharmacological strategy (Falck‐Ytter 2012; NICE 2018). The aim of all the above interventions is to prevent venous thromboembolism.

How the intervention might work

Pharmacological interventions include the parenteral anticoagulants (LMWH, fondaparinux); the oral direct thrombin inhibitors (dabigatran); the oral factor Xa inhibitors (rivaroxaban and apixaban), and oral antiplatelets (aspirin). They all work by inhibiting stages of the coagulation process (coagulation cascade). LMWH selectively binds to antithrombin (AT), a serine protease inhibitor, and creates a conformational change of the molecule. This change promotes inhibition of activated Factor X. Fondaparinux is a synthetic pentasaccharide Factor Xa inhibitor, again by binding to AT but is also selective for Factor Xa (Ritter 2019). Dabigatran reversibly binds to the active site on the thrombin molecule, preventing thrombin‐mediated activation of coagulation factors (Ritter 2019). Rivaroxaban inhibits both free Factor Xa and Factor Xa bound in the prothrombinase complex. It is a highly selective direct Factor Xa inhibitor with a rapid onset of action. Aspirin inhibits platelet generation of thromboxane A2, resulting in an antithrombotic effect (Ritter 2019). Mechanical interventions facilitate the circulation of the venous flow in lower limbs preventing thrombus formation (Benko 2001).

Why it is important to do this review

Arthroscopy procedures of the knee are frequent and it is relevant to clinical practice to define if the use of thromboprophylaxis reduces the already low risk of thromboembolic disease. There is uncertainty about whether prophylaxis used routinely in orthopedic surgery should be used in KA, due to the very low risk of severe thromboembolic events (proximal DVT and PE), the uncertain significance of asymptomatic distal DVT (which is also of low prevalence), and the early mobility regained by patients after the procedure. As new randomized studies have been completed since the previous version of this review was published (Ramos 2008), it is important to include new evidence of thromboprophylaxis to assist decision making in people with no previous history of thromboembolic events undergoing KA.

Objectives

To assess the effectiveness and safety of interventions, mechanical or pharmacological, or in combination, for thromboprophylaxis in adult patients undergoing KA.

Methods

Criteria for considering studies for this review

Types of studies

We included randomized controlled trials (RCTs) and controlled clinical trials (CCTs), whether blinded or not (i.e. double‐blinded, single‐blinded or unblinded trials were eligible for inclusion).

Types of participants

We included studies where participants were aged 18 years and older and undergoing KA.

Types of interventions

We included any type of intervention, whether mechanical or pharmacological, and whether used singly or in combination, to prevent DVT in patients undergoing KA. We excluded interventions in which investigators used arthroscopic‐assisted osteosynthesis of tibial plateau fractures.

Types of outcome measures

Primary outcomes

• Pulmonary embolism confirmed by ventilation/perfusion (V/Q) lung scan, spiral computed tomography (CT), or pulmonary angiography

• Symptomatic DVT (distal and proximal combined). Clinically, venographically or sonographically diagnosed. Defined as 'symptomatic' when patients spontaneously experienced DVT symptoms (calf pain, swelling, edema); or when patients referred to symptoms of DVT while undergoing echo‐Doppler or compression ultrasound in the context of the research trial

• Asymptomatic DVT. Defined as 'asymptomatic' when patients were actively screened for DVT using echo‐Doppler or compression ultrasound and patients did not indicate any symptoms during ultrasound

• All‐cause mortality

Secondary outcomes

• Overall number of reported side effects and adverse effects (pain, allergies, minor and major bleeding)

• Major bleeding (overt and requiring transfusion of red blood cells or surgical interventions, and which may lead to permanent disability)

• Minor bleeding (did not meet the above major criteria for intervention for example, hematoma, hemarthrosis, thrombocytopenia (reduction of circulating platelets) of < 80,000/mm³ or decrease > 50% of the initial count)

Search methods for identification of studies

We did not put any restrictions on publication status or language.

Electronic searches

For this update, the Cochrane Vascular Information Specialist first searched the following databases for relevant trials (19 March 2017)

• the Cochrane Vascular Specialised Register (19 March 2017)

• the Cochrane Central Register of Controlled Trials (CENTRAL; 2017, Issue 2) via the Cochrane Register of Studies Online.

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

The Cochrane Vascular Information Specialist subsequently conducted systematic top‐up searches of the following databases for randomized controlled trials and controlled clinical trials without language, publication year or publication status restrictions:

• the Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web) (searched to 14 August 2019)

• the Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Register of Studies Online (CRSO 2019, Issue 7)

• MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE) (searched from 1 January 2017 to 14 August 2019)

• Embase via Ovid (searched from 1 January 2017 to 14 August 2019)

• CINAHL via EBSCO (searched from 1 January 2017 to 14 August 2019)

• AMED via Ovid (searched from 1 January 2017 to 14 August 2019).

The Information Specialist modeled search strategies for the listed databases on the search strategy designed for CENTRAL. Where appropriate, they were combined with adaptations of the Cochrane highly sensitive search strategy for identifying reports of randomized controlled trials and controlled clinical trials (Lefebvre 2011). Search strategies for major databases are provided in Appendix 2.

The Information Specialist also performed searches of the following trials registries on 14 August 2019:

• ClinicalTrials.gov (clinicaltrials.gov);

• World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch).

Searching other resources

In addition, we searched the reference lists of identified studies, and articles and abstracts of international meetings of orthopedics, hematology and thrombosis journals from the year 1998 to 2017. We contacted specialists known to be involved in phlebology and interested in thromboembolic disease for details of unpublished and ongoing trials. There were no restrictions on language.

Data collection and analysis

Selection of studies

Two authors (JR, GB) independently screened the initial reference list, identified relevant citations and addressed if they met inclusion criteria through use of a standardized form. Disagreements were resolved by a third author (CP).

Data extraction and management

Three authors (JR, JC, GB) extracted the data into a custom spreadsheet and entered data into Review Manager 5 (Review Manager 2014). When necessary, data requests were sent to trial authors for additional information or data.

Assessment of risk of bias in included studies

We used the Cochrane 'Risk of bias' tool (Higgins 2011). The risk of bias for each of the included studies was independently assessed by three authors (JC, JR, GB).The following domains were evaluated for each study.

• Sequence generation (selection bias)

• Allocation concealment (selection bias)

• Blinding of participants and personnel (performance bias)

• Blinding of outcome assessors (ascertainment bias)

• Incomplete outcome data (attrition bias)

• Selective outcome reporting bias

• Other sources of bias

Each domain was explicitly judged as being at either low, high or unclear risk of bias (either from lack of information or uncertainty over the potential for bias). Disagreements among review authors were resolved by consensus.

Measures of treatment effect

We investigated treatment effect on our outcomes of interest by calculating the risk ratios (RR) with 95% confidence intervals (CIs). We planned to include clinical trials that had zero events. We did not have any outcomes in which continuous data would be reported.

Unit of analysis issues

We took each individual participant to be the unit of randomization in the included trials. In the original protocol, we stated that the unit of analysis would be by 'knee.' None of the trials included people undergoing bilateral KA.

Dealing with missing data

We did not identify any studies with missing data from the relevant outcomes (thromboembolic events and adverse events). Had that been the case, we would have contacted the authors for clarification, in relation to missing data.

Assessment of heterogeneity

We assessed the heterogeneity of estimated effects in the included studies by visual inspection of the forest plot generated from meta‐analysis of studies. The degree of statistical heterogeneity was assessed based on the test for heterogeneity and the I² statistic. If the results appeared to be very different or the I² statistic was greater than 50%, we considered that this likely represented substantial heterogeneity. This information was also considered with other indicators of statistical heterogeneity, such as the Chi² statistic and degrees of freedom (df). We assumed statistically significant heterogeneity when Chi² exceeded df and the P value was less than 0.05. If heterogeneity was present, we aimed to explore sources of heterogeneity potentially relevant to the clinical question.

Assessment of reporting biases

We planned to assess the presence of publication bias and other reporting bias using funnel plots if sufficient studies (10 or more) had been included in the meta‐analysis (Higgins 2011).

Data synthesis

When studies compared similar interventions and reported comparable outcomes, we pooled effect estimates using the inverse variance method through a fixed‐effect model and calculated 95% CI.

If heterogeneity was present we used a random‐effects model, and, when possible, carried out subgroup analysis to explain the source of heterogeneity. We conducted narrative synthesis when studies were too heterogeneous or a given intervention was tested by a single study.

We used Review Manager 5 to perform the analyses (Review Manager 2014).

Subgroup analysis and investigation of heterogeneity

We did not have any pre‐planned subgroup analyses. In the case of heterogeneity we did attempt to explain differences as per study design by looking at subgroups of studies, specifically in how the outcomes were defined and assessed.

Sensitivity analysis

We carried out sensitivity analysis to determine any impact of using either intention‐to‐treat (ITT) data or per‐protocol (PP) data in the van Adrichem 2017 study.

Summary of findings and assessment of the certainty of the evidence

We prepared a 'Summary of findings' (SoF) table to present the main findings from the meta‐analysis for the outcomes of PE, symptomatic DVT, asymptomatic DVT, all‐cause mortality, adverse effects (all), major bleeding and minor bleeding.

We created a SoF table for the comparison of LMWH versus control. As the remaining comparisons included data from one study only, we did not create separate SoF tables for them.

We used GRADEprofiler software to create the SoF table (GRADEpro GDT). The GRADE criteria were used to rank the certainty of the evidence for each outcome based on risk of bias, inconsistency, indirectness, imprecision and publication bias (Guyatt 2008).

We provided a rationale for each instance where it was appropriate to downgrade the certainty of the evidence.

Results

Description of studies

Results of the search

See Figure 1.

1.

Study flow diagram.

Included studies

Please see the Characteristics of included studies table for further details.

We identified four new studies for this update (Camporese 2008; Camporese 2016; Kaye 2015; van Adrichem 2017), bringing the total to eight included studies with 3818 participants (Camporese 2008; Camporese 2016; Canata 2003; Kaye 2015; Michot 2002; Roth 1995; van Adrichem 2017; Wirth 2001).

Overall, four different treatment comparisons were used to prevent thromboembolic disease in patients undergoing knee arthroscopy.

All the studies excluded pregnant participants, or those with previous thromboembolic disease or cancer. Kaye 2015 excluded smokers and Wirth 2001 screened for additional risk factors (obesity, nicotine abuse, oral contraceptives and family history of thrombosis). If three or more of these factors were present, participants were excluded.

LMWH versus control (no intervention)

Five studies with a total of 1942 participants compared LMWH to a control condition with no treatment (Canata 2003; Michot 2002; Roth 1995; van Adrichem 2017; Wirth 2001).

Rivaroxaban versus placebo

One study with 241 participants compared rivaroxaban to placebo (Camporese 2016).

Aspirin versus control (no intervention)

One study with 170 participants compared aspirin to no treatment (Kaye 2015).

LMWH versus compression stockings

One study with 1317 participants compared LMWH for seven or 14 days to use of compression stockings (Camporese 2008).

All studies evaluated thromboembolic disease (PE, or symptomatic DVT or asymptomatic DVT or all), mortality, minor and major bleeding or other relevant adverse events.

The types of KA performed were meniscectomy, ACL reconstruction and removing of loose fragments and, in a small proportion of participants, diagnostic procedures. Most of the studies did not tabulate events in relation to the type of surgery performed or to participant characteristics such as body mass index or smoking status.

Excluded studies

No new studies were excluded for this update. Marlovits 2004 was excluded previously as it compared LMWH treatment for 3 to 8 days with LMWH for 20 days without a control arm. Please see the Characteristics of excluded studies table for further details.

Risk of bias in included studies

Please see Figure 2 and Figure 3 for a summary of the overall risk of bias.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

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

Overall, the included studies were of acceptable methodological quality, given the type of intervention. No domains in any study were judged to be at high risk of bias in relation to randomization, allocation, selection bias or reporting bias. We flagged a few issues, such as a lack of participant or personnel blinding, and the lack of a placebo for the studies comparing LMWH to control. Our judgment is that the lack of placebo would not have impacted more objective outcomes such as death, PE or symptomatic DVT (participants seeking medical advice due to DVT symptoms with confirmation by a positive echo‐Doppler or compression ultrasound). There is a risk of bias in outcomes which could be influenced by study personnel or participants being aware of, or disclosing, participant status in treatment arms.

All studies except van Adrichem 2017 used asymptomatic venous thrombosis in their combined outcome. Camporese 2008 and Michot 2002 defined symptomatic DVT after actively asking participants if they had any symptoms of DVT (e.g. calf pain, edema, redness) prior to confirmation with color‐coded Doppler and compression ultrasonography, respectively. This was done during study visits or through telephone interviews. We judged that there is an unclear risk of bias either through indirectness of the outcome (a positive Doppler sonograph is not a usual procedure after KA) and in the way symptomatic DVT was defined. Actively asking participants about their symptoms before giving them Doppler ultrasound could introduce recall bias. Moreover, the use of compression ultrasound has an estimate specificity for distal veins of 97.8% (Goodacre 2005). We addressed the impact of any potential bias on our confidence in the effect estimates of individual outcomes while assessing the certainty of evidence with GRADE,

Allocation

Allocation was adequately carried out by five studies, as part of the randomization process used methods including sealed opaque envelopes, sequentially numbered, unlabelled boxes and research staff being unaware of block sizes in randomization strategies; these studies were at low risk of bias (Camporese 2008; Camporese 2016; Kaye 2015; van Adrichem 2017; Wirth 2001). The remaining three studies did not provide sufficient detail to show adequate concealment of allocation; these studies were judged to be at unclear risk of selection bias (Canata 2003; Michot 2002; Roth 1995).

Blinding

Camporese 2016 used oral medications as the intervention and a placebo in the control group. This study was at low risk of both performance and detection bias as participants, personnel and outcome assessors were blinded to the treatment group. In van Adrichem 2017, participants were aware of treatment arm, but outcomes were assessed by checking hospital registries. There was low risk of this influencing the outcome measurement. In the remaining studies, where the participants and/or personnel were aware of the treatment group, but outcome assessors were not aware, we judged that the outcomes of PE, symptomatic DVT and adverse events would not have been affected. However, for asymptomatic DVT we judged that there was a risk of detection bias. These studies were judged to be at unclear risk of bias, as it is possible that the participants' knowledge of the groups they were in could have influenced outcome assessment. (Camporese 2008; Canata 2003; Kaye 2015; Michot 2002; Wirth 2001). Roth 1995 was at an unclear risk of performance and detection bias as sufficient details were not provided.

Incomplete outcome data

Five studies were at low risk of attrition bias because they reported complete outcome data, or because any losses to follow up were accounted for and balanced between groups (Camporese 2008; Camporese 2016; Kaye 2015; Michot 2002; Wirth 2001). van Adrichem 2017 described losses and reported results both PP and by ITT. There were no differences in the estimates so we judged the risk of attrition bias to be low. Study authors reported combined outcomes and individual outcomes (all thromboembolic events, including DVT, PE and asymptomatic DVT) so we were able to report the outcomes of interest individually. Roth 1995 was at unclear risk of attrition bias because investigators presented incomplete information on drop outs or losses to follow‐up. Canata 2003 was assessed as being at unclear risk of attrition bias because there were no details on missing data or losses to follow‐up.

Selective reporting

Seven studies were at low risk of reporting bias as there are no concerns in relation to selective reporting (Camporese 2008; Camporese 2016; Canata 2003; Kaye 2015; Michot 2002; van Adrichem 2017; Wirth 2001). Roth 1995 did not have complete reporting for all the outcomes of interest and so is at unclear risk of reporting bias.

Other potential sources of bias

In Camporese 2008, the data and safety monitoring board stopped the 14‐day treatment group during a planned interim analysis. The study continued as a trial with two groups (7‐day treatment versus control). We were unsure if this could have biased the results so we judged there to be an unclear risk of other bias. In Michot 2002, recruitment was stopped prior to reaching the desired sample size as it was decided by the trialists that withholding LMWH was unethical. We judged this study to be at unclear risk of other bias. All remaining studies were at low risk of other bias.

Effects of interventions

See: Table 1

LMWH versus control (no intervention)

Pulmonary embolism

Of five studies comparing LMWH versus control (Canata 2003; Michot 2002; Roth 1995; van Adrichem 2017; Wirth 2001), three reported on PE (Michot 2002; van Adrichem 2017; Wirth 2001). Overall, events were few and pooling the data from these showed little to no difference in the rates of PE in the LMWH group (6/914) compared to the control group (3/906); (RR 1.81, 95% CI 0.49 to 6.65; 1820 participants; 3 studies); moderate‐certainty evidence; Analysis 1.1). The certainty of the evidence was downgraded one step to moderate‐certainty due to imprecision around the estimate of effect. Canata 2003 and Roth 1995 did not report PE.

1.1. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 1: Pulmonary embolism

Symptomatic DVT

Four studies reported symptomatic DVT as a separate outcome (Canata 2003; Roth 1995; van Adrichem 2017; Wirth 2001). There was no reduction in symptomatic DVT rate when comparing LMWH versus control (RR 0.61, 95% CI 0.18 to 2.03; 1848 participants; 4 studies; low‐certainty evidence; Analysis 1.2). We downgraded the certainty of the evidence due to imprecision (small absolute number of events and small sample size) and risk of bias due to lack of a valid placebo, uncertainty about blinding, and reporting of symptoms with potential recall bias.

1.2. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 2: Symptomatic DVT

Asymptomatic DVT

Two studies reported on asymptomatic DVT (Michot 2002; Wirth 2001). LMWH may reduce the prevalence of DVT diagnosed by compression ultrasonography, compared to control but the evidence is very uncertain (RR 0.14, 95% CI 0.03 to 0.61; 369 participants; 2 studies; very low‐certainty evidence; Analysis 1.3).

1.3. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 3: Asymptomatic DVT

We downgraded the evidence to very low‐certainty due to imprecision, risk of bias and indirectness. The use of compression ultrasonography to actively seek for asymptomatic DVT is not routinely recommended in KA. As described in the risk of bias section, above, participants were not blinded to the intervention. It is not certain that participants did not discuss with assessors their allocated intervention. Additionally, this is not an outcome of interest for participants that have low risk of thromboembolic disease or that will return to full mobilization within the week after the procedure.

Mortality

No deaths occurred in either group in any study.

All reported adverse events (pain, allergies, minor and major bleeding)

Five studies reported a total of 39 adverse events in 1978 participants (Canata 2003; Michot 2002; Roth 1995; van Adrichem 2017; Wirth 2001). Adverse events included minor bleeding at the injection site, pain at the injection site, or allergies presented as inflammation at the injection site. Thrombocytopenia was not assessed by any study. In the LMWH intervention there was uncertainty about an increased risk of all adverse events combined (RR 1.85, 95% CI 0.95 to 3.59; 1978 participants; 5 studies; moderate‐certainty evidence; Analysis 1.4). We downgraded one step due to imprecision, probably related to the low numbers of events.

1.4. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 4: Adverse events

Major bleeding

None of the trials in this intervention group reported any cases of major bleeding except for van Adrichem 2017, with one case in each treatment arm (RR 0.98, 95% CI 0.06 to 15.72; 1451 participants; 1 study, moderate‐certainty evidence; Analysis 1.5). We downgraded one step due to imprecision.

1.5. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 5: Major bleeding

Minor bleeding

All five studies assessed minor bleeding (Canata 2003; Michot 2002; Roth 1995; van Adrichem 2017; Wirth 2001). No evidence of clear differences in minor bleeding events was seen in LMWH versus control (RR 1.79, 95% CI 0.84 to 3.84; 1978 participants; 5 studies; moderate‐certainty evidence; Analysis 1.6). We downgraded one step due to imprecision.

1.6. Analysis.

Comparison 1: Low molecular weight heparin versus control (no prophylactic treatment), Outcome 6: Minor bleeding

As van Adrichem 2017 presented both PP and ITT data, we carried out sensitivity analysis investigating both data. There were no differences in the results and we included the ITT data in our analysis.

Rivaroxaban versus placebo

Pulmonary embolism

Only Camporese 2016 investigated this comparison and there were no PE events during the study period (234 participants; 1 study).

Symptomatic DVT

There was no evidence that use of oral rivaroxaban 10 mg reduced the risk of symptomatic DVT (RR 0.16, 95% CI 0.02 to 1.29; 234 participants; 1 study; moderate‐certainty evidence; Analysis 2.1). The certainty of the evidence was downgraded one step due to the imprecision of the estimate.

2.1. Analysis.

Comparison 2: Rivaroxaban versus placebo, Outcome 1: Symptomatic DVT

Asymptomatic DVT

There was no evidence that oral rivaroxaban 10 mg reduced the risk of DVT as diagnosed by compression ultrasonography compared against placebo (RR 0.95, 95% CI 0.06 to 15.01; 234 participants; 1 study; very low‐certainty evidence; Analysis 2.2). We downgraded due to imprecision, risk of bias and indirectness of the outcome.

2.2. Analysis.

Comparison 2: Rivaroxaban versus placebo, Outcome 2: Asymptomatic DVT

Mortality

No deaths occurred in either group.

All reported adverse events (pain, allergies, minor and major bleeding)

Camporese 2016 did not report on any adverse events in addition to major and minor bleeding, which are reported below.

Major bleeding

No major bleeding events were reported.

Minor bleeding

There was no evidence that the use of rivaroxaban increased the risk of minor bleeding (RR 0.63, 95% CI 0.18 to 2.19; 234 participants; 1 study; moderate‐certainty evidence; Analysis 2.3). The quality of the evidence was downgraded one step due to the imprecision of the estimate.

2.3. Analysis.

Comparison 2: Rivaroxaban versus placebo, Outcome 3: Minor bleeding

Aspirin versus control (intervention)

Pulmonary embolism

Only Kaye 2015 reported on this comparison. There were no PE events during the study period. All Kaye 2015 outcomes provided moderate‐certainty evidence due to no events in either of the groups.

Symptomatic DVT

There were no DVT events in Kaye 2015,

Asymptomatic DVT

There were no asymptomatic DVT events in Kaye 2015. Asymptomatic events were investigated using compression ultrasonography.

Mortality

No deaths occurred in either group.

All reported adverse events (pain, allergies, minor and major bleeding)

Kaye 2015 reported complications including pain and swelling (15 cases, 9%) and residual joint line tenderness (5 cases, 3%) but not what groups they occurred in. Three participants developed knee swelling. Two of these participants were randomized to the aspirin group.

Major bleeding

No major bleeding events were reported by Kaye 2015.

Minor bleeding

No minor bleeding events were reported by Kaye 2015.

LMWH versus compression stockings

Pulmonary embolism

Only Camporese 2008 investigated this comparison. There was no clear difference in the risk of PE between the LWMH group or the compression stockings group (RR 1.00, 95% CI 0.14 to 7.05; 1317 participants; 1 study, low‐certainty evidence; Analysis 3.1). We downgraded by two steps due to imprecision with low number of events and risk of bias concerns.

3.1. Analysis.

Comparison 3: Low molecular weight heparin versus compression stockings, Outcome 1: Pulmonary embolism

Symptomatic DVT

The use of LMWH reduced the risk of symptomatic DVT compared to compression stockings (RR 0.17, 95% CI 0.04 to 0.75; 1317 participants; 1 study; low‐certainty evidence; Analysis 3.2). We downgraded this outcome by two steps due to imprecision of the absolute effects, and risk of bias.

3.2. Analysis.

Comparison 3: Low molecular weight heparin versus compression stockings, Outcome 2: Symptomatic DVT

Asymptomatic DVT

It was not clear if the use of LMWH may reduce the risk of asymptomatic DVT as diagnosed by compression ultrasonography compared to compression stockings (RR 0.47, 95% CI 0.21 to 1.09; 1317 participants; 1 study, very low‐certainty evidence; Analysis 3.3). We have downgraded by three steps due to imprecision of the estimates, low number of events, indirectness and risk of bias concerns.

3.3. Analysis.

Comparison 3: Low molecular weight heparin versus compression stockings, Outcome 3: Asymptomatic DVT

Mortality

No deaths occurred in either group.

All reported adverse events (pain, allergies, minor bleeding and major bleeding)

In addition to major bleeding and minor bleeding, Camporese 2008 reported on pain, tenderness and edema. No evidence of differences was detected (RR 1.17, 95% CI 0.95 to 1.43; 1317 participants; 1 study; moderate‐certainty evidence; Analysis 3.4). We downgraded the certainty of the evidence by one step due to imprecision due to the small number of events and small sample size. In the compression stockings group, 152 participants reported symptoms (pain, 92; tenderness, 23; edema, 24; discoloration, 10; and nonvaricose collateral symptoms, 3), compared with 131 participants in the LMWH group (pain, 85; tenderness, 18; edema, 18; discoloration, 8; and nonvaricose collateral symptoms, 2). Thrombocytopenia was not assessed.

3.4. Analysis.

Comparison 3: Low molecular weight heparin versus compression stockings, Outcome 4: Adverse events

Major bleeding

There was no clear difference between the groups for major bleeding (RR 3.01, 95% CI 0.61 to 14.88; 1317 participants; 1 study; moderate‐certainty evidence; Analysis 3.4). We downgraded by one step for imprecision due to the small number of events and small sample size.

Minor bleeding

There was no clear difference between the groups for minor bleeding (RR 1.16, 95% CI 0.64 to 2.08; 1317 participants; 1 study; moderate‐certainty evidence; Analysis 3.4). We downgraded by one step for imprecision due to small sample size and small number of events.

Discussion

Summary of main results

There was moderate‐certainty evidence that the risk of PE was not reduced by any of the interventions identified in this systematic review (LMWH versus control, aspirin versus placebo, rivaroxaban versus placebo, and LMWH versus compression stocking). The overall risk of PE was low, with 7 cases in 3499 participants (less than 0.2%) evenly distributed between control and interventions, with no mortality due to PE or any other cause.

The use of LMWH showed no clear reduction in the risk of symptomatic DVT (patients presenting with symptoms of DVT at the emergency department, or symptoms with a positive screening) compared to control (low‐certainty evidence). Similarly, no evidence of clear impact on DVT risk was seen with the use of aspirin or rivaroxaban. A reduced rate of DVT was observed with LMWH use compared with compression stockings (low‐certainty evidence). Interestingly, the incidence of symptomatic DVT was higher in the compression stocking group than in any other group arm across all studies, This group had a 2% incidence of symptomatic DVT while the literature reports rates below 0.5% (Maletis 2012).

A reduced risk of asymptomatic DVT was observed in the comparison LMWH versus control but the evidence was very uncertain (very low‐certainty evidence). No evidence of risk reduction was seen in rivaroxaban versus placebo (very low‐certainty evidence), no asymptomatic DVT events were detected in aspirin versus placebo, and it is not clear if rates were reduced in LMWH versus compression stocking (very low‐certainty evidence).

There was no evidence of clear differences in adverse events for any comparison (moderate‐certainty evidence), and results were limited by low numbers of studies reporting clearly on this. When reported there was no evidence of differences in the rates of major or minor bleeding (both moderate‐certainty evidence).

In conclusion, for the outcome of PE and symptomatic DVT in low risk patients undergoing KA, there is moderate to very low‐certainty evidence of no clear benefit in the use of LMWH, aspirin or rivaroxaban compared to placebo or no intervention for preventing PE or DVT.

Overall completeness and applicability of evidence

This evidence is applicable for otherwise healthy patients undergoing KA for minor conditions and at low risk of DVT. Study participants had no previous history of any thromboembolic events, were not pregnant, with no history of any coagulation disorder that put the patient at risk of thromboembolic events, and had accessible health care systems. Patients taking oral contraceptive pills were excluded in only one study.

Patients who smoked or were obese, were included in the population and appeared to be well balanced between intervention and control groups. The low rate of events did not allow subgrouping according to those variables, but study authors did not report more events in participants with those characteristics.

Only one trial (van Adrichem 2017) reflected what a 'real life' scenario would be ‐ identifying study patients that attended an emergency department or were admitted for PE or DVT without actively looking for thromboembolic events. PE and symptomatic DVT were obtained from hospital registries up to three months after the procedure. It is unlikely that the study authors would have missed significant events but it is still a possibility. In any case, this reflects what occurs in daily practice within a healthcare system that has good patient accessibility.

It may be the case, that in healthcare systems with poor accessibility, or healthcare systems that are highly fragmented with a poor network, thromboembolic events maybe missed. As there is still a very small risk for thromboembolic events following knee arthroscopy, the healthcare team looking after these patients (general practitioners, orthopedics, physiotherapist, Emergency Department team) need to advise patients to seek medical care if any thromboembolic symptoms develop.

The evidence presented in this review may not be applicable in situations where patients are at higher risk of PE or DVT, if they can't reach their surgeons or do not have access to healthcare, or if they are unable to achieve early mobility after the procedure.

Quality of the evidence

The certainty of the evidence is moderate to low for the outcomes PE, symptomatic DVT and adverse effects (including major and minor bleeding). Our certainty of the evidence for asymptomatic DVT is very low. Outcomes are more prone to bias in cases where interventions are not blinded and we downgraded for risk of bias concerns when we felt this could have impacted outcome assessment.

There was significant imprecision in all the clinical outcomes. This is imprecision is mostly due to the low prevalence of symptomatic thromboembolic events after knee arthroscopy in low risk patients. It is possible that larger sample sizes may lead to different results, but this may be unlikely due to the low event rate expected after this particular procedure.

For the outcome of asymptomatic DVT, there is very low‐certainty evidence that the use of LMWH compared to control reduces its occurrence. The clinical significance of DVT diagnosed through a echo‐Doppler or compression ultrasonography scan is uncertain in the context of a knee arthroscopy in a healthy patient and so we also downgraded our certainty in the evidence for indirectness.

Potential biases in the review process

A thorough search was carried out and we are confident all relevant studies have been assessed, however the available evidence is limited by the small number of studies eligible for inclusion. A major limitation in all the studies is the low event rate for severe thromboembolic events (symptomatic DVT or PE). The event rate was similar to that reported in observational studies. This confirms that KA in healthy individuals has a small risk of symptomatic thrombosis, but it is also makes it difficult to detect any benefit from the proposed interventions. Our certainty of the evidence remains low to moderate. As discussed above, the main risk of bias concern is the lack of an adequate placebo, which then prevented appropriate blinding. Researchers asked participants to cover the injection site. Those who assessed outcomes were blinded to the allocated treatment arm, providing the participants did not break this blinding.

Agreements and disagreements with other studies or reviews

A systematic review looking only at LMWH reached similar conclusions to our review (Huang 2018). Our review aimed to include all types of interventions to prevent DVT including pharmacological and mechanical interventions. Compared with the review by Huang 2018, our review also included studies using aspirin and rivaroxaban as pharmacological interventions (Camporese 2016; Kaye 2015). Of note, Huang 2018 included Marlovits 2004, which we excluded as it compared short duration LMWH with extended duration LMWH. Huang 2018 also excluded distal DVT from their analysis.

Authors' conclusions

Implications for practice.

There is a small risk that healthy adult patients undergoing KA will develop PE or DVT. There is moderate‐ to low‐certainty evidence of no benefit from the use of LMWH, aspirin or rivaroxaban in reducing this small risk of PE or symptomatic DVT. There is very low‐certainty evidence that LMWH use may reduce the risk of asymptomatic DVT when compared to no treatment but we are not sure how this directly relates to incidence of DVT or PE in healthy patients. No evidence of differences in adverse events (including major and minor bleeding) was seen, but data relating to this were limited due to the small numbers of studies reporting these outcomes.

Implications for research.

New studies should focus on the benefit of the prophylaxis in high‐risk patients; such as obese patients, patients not able to gain early mobility or patients with limited access to acute emergency care. Conducting new clinical trials with low risk patients will require a very large sample size which would also increase the likelihood of adverse events. For new trials, researchers should minimize the risk of bias. Study designs should observe patients' trajectories and measure only symptomatic thromboembolic outcomes without actively seeking venous thromboembolism with diagnostic tests such as compression ultrasound. The clinical trial conducted by van Adrichem 2017 is an example of a design which can answer the research question without introducing additional bias.

Feedback

Anticoagulant feedback, February 2011

Summary

Feedback received on this review, and other reviews and protocols on anticoagulants, is available on the Cochrane Editorial Unit website at http://www.editorial-unit.cochrane.org/anticoagulants-feedback.

What's new

Date	Event	Description
14 August 2019	New search has been performed	Searches rerun. Four new studies included. No new studies excluded.
14 August 2019	New citation required but conclusions have not changed	Searches rerun. Four new studies included. No new studies excluded. New authors joined team. Review text updated to reflect current Cochrane standards. Risk of bias assessed using Cochrane's 'Risk of bias' tool and 'Sumary of findings' table included. Conclusions not changed.

History

Protocol first published: Issue 2, 2005
Review first published: Issue 2, 2007

Date	Event	Description
14 February 2011	Amended	Link to anticoagulant feedback added
5 August 2008	New citation required but conclusions have not changed	Amendment to correct spelling error in title and labelling of one of the graphs. Additional reference for ongoing study (KANT Study) added.
31 July 2008	Amended	Converted to new review format.

Appendices

Appendix 1. CENTRAL search March 2017

#1	MESH DESCRIPTOR Thrombosis	1261
#2	MESH DESCRIPTOR Thromboembolism	919
#3	MESH DESCRIPTOR Venous Thromboembolism	257
#4	MESH DESCRIPTOR Venous Thrombosis EXPLODE ALL TREES	2036
#5	(thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol*):TI,AB,KY	18960
#6	MESH DESCRIPTOR Pulmonary Embolism EXPLODE ALL TREES	746
#7	(PE or DVT or VTE):TI,AB,KY	4979
#8	((vein* or ven*) near thromb*):TI,AB,KY	6702
#9	(blood near3 clot*):TI,AB,KY	2963
#10	(pulmonary near3 clot*):TI,AB,KY	5
#11	(lung near3 clot*):TI,AB,KY	4
#12	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11	24595
#13	MESH DESCRIPTOR Arthroscopy EXPLODE ALL TREES	1189
#14	MESH DESCRIPTOR Knee EXPLODE ALL TREES	656
#15	MESH DESCRIPTOR Knee Joint EXPLODE ALL TREES	2624
#16	MESH DESCRIPTOR Anterior Cruciate Ligament Reconstruction EXPLODE ALL TREES	242
#17	MESH DESCRIPTOR Anterior Cruciate Ligament EXPLODE ALL TREES	713
#18	MESH DESCRIPTOR Debridement	502
#19	(cruciate and ligament and re*):TI,AB,KY	1560
#20	meniscectomy:TI,AB,KY	259
#21	chondroplasty:TI,AB,KY	26
#22	(articular debridement):TI,AB,KY	1
#23	(synovial and resection):TI,AB,KY	24
#24	(tibial plateau fracture):TI,AB,KY	16
#25	(articular and debridement):TI,AB,KY	68
#26	arthroscop*:TI,AB,KY	2550
#27	(knee near3 surg*):TI,AB,KY	2847
#28	(knee near3 debridement):TI,AB,KY	12
#29	(knee near3 lavage):TI,AB,KY	17
#30	(Meniscal repair):TI,AB,KY	35
#31	(meniscal surgery):TI,AB,KY	15
#32	(abrasion arthroplasty):TI,AB,KY	2
#33	(loose body removal):TI,AB,KY	1
#34	(ACL reconstruction):TI,AB,KY	648
#35	(Lateral release):TI,AB,KY	39
#36	synovectomy:TI,AB,KY	79
#37	debridement:TI,AB,KY	1531
#38	#13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37	8728
#39	#12 AND #38	292

Appendix 2. Database search strategies 2018 and 2019

CENTRAL	#1 MESH DESCRIPTOR Thrombosis 1625 #2 MESH DESCRIPTOR Thromboembolism 1130 #3 MESH DESCRIPTOR Venous Thromboembolism 463 #4 MESH DESCRIPTOR Venous Thrombosis EXPLODE ALL TREES 2392 #5 (thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol*):TI,AB,KY 23978 #6 MESH DESCRIPTOR Pulmonary Embolism 879 #7 (PE or DVT or VTE):TI,AB,KY 6619 #8 ((vein* or ven*) near thromb*):TI,AB,KY 8286 #9 (blood near3 clot*):TI,AB,KY 4096 #10 (pulmonary near3 clot*):TI,AB,KY 9 #11 (lung near3 clot*):TI,AB,KY 7 #12 #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 31413 #13 MESH DESCRIPTOR Arthroscopy EXPLODE ALL TREES 1316 #14 MESH DESCRIPTOR Knee EXPLODE ALL TREES 716 #15 MESH DESCRIPTOR Knee Joint EXPLODE ALL TREES 2937 #16 MESH DESCRIPTOR Anterior Cruciate Ligament Reconstruction EXPLODE ALL TREES 315 #17 MESH DESCRIPTOR Anterior Cruciate Ligament EXPLODE ALL TREES 637 #18 MESH DESCRIPTOR Debridement 559 #19 (cruciate and ligament and re*):TI,AB,KY 2003 #20 meniscectomy:TI,AB,KY 313 #21 chondroplasty:TI,AB,KY 35 #22 (articular debridement):TI,AB,KY 1 #23 (synovial and resection):TI,AB,KY 28 #24 (tibial plateau fracture):TI,AB,KY 27 #25 (articular and debridement):TI,AB,KY 80 #26 arthroscop*:TI,AB,KY 3303 #27 (knee near3 surg*):TI,AB,KY 3496 #28 (knee near3 debridement):TI,AB,KY 14 #29 (knee near3 lavage):TI,AB,KY 23 #30 (Meniscal repair):TI,AB,KY 51 #31 (meniscal surgery):TI,AB,KY 24 #32 (abrasion arthroplasty):TI,AB,KY 3 #33 (loose body removal):TI,AB,KY 2 #34 (ACL reconstruction):TI,AB,KY 888 #35 (Lateral release):TI,AB,KY 45 #36 synovectomy:TI,AB,KY 91 #37 debridement:TI,AB,KY 1985 #38 #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 11062 #39 #12 AND #38 387 #40 01/01/2016 TO 08/08/2018:CD 393004 #41 #39 AND #40 152	8.8.18‐ 152 14.8.19 ‐ 99
Clinicaltrials.gov	Venous Thrombosis OR Thrombosis OR Thromboembolism OR Pulmonary Embolism | Arthroscopy OR knee OR Anterior Cruciate Ligament OR Debridement | Start date on or after 01/01/2017 | Start date on or after 01/01/2016 | Last update posted on or before 08/08/2018	8.8.18 ‐ 15 14.8.19 ‐1
ICTRP Search Portal	Venous Thrombosis OR Thrombosis OR Thromboembolism OR Pulmonary Embolism | Arthroscopy OR knee OR Anterior Cruciate Ligament OR Debridement | Start date on or after 01/01/2017 | Start date on or after 01/01/2016 | Last update posted on or before 08/08/2018	8.8.18 ‐ 6 14.8.19 ‐ 3
Medline (Ovid MEDLINE® Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE® Daily and Ovid MEDLINE®) 1946 to present (2017, 2018 AND 2019 ONLY)	1 THROMBOSIS/ 65932 2 THROMBOEMBOLISM/ 22659 3 Venous Thromboembolism/ 8455 4 exp Venous Thrombosis/ 51651 5 (thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol*).ti,ab. 302393 6 Pulmonary Embolism/ 36228 7 (PE or DVT or VTE).ti,ab. 47721 8 ((vein* or ven*) adj thromb*).ti,ab. 61557 9 (blood adj3 clot*).ti,ab. 10109 10 (pulmonary adj3 clot*).ti,ab. 191 11 (lung adj3 clot*).ti,ab. 47 12 or/1‐11 391017 13 exp ARTHROSCOPY/ 21054 14 exp KNEE/ 13141 15 exp Knee Joint/ 54459 16 exp Anterior Cruciate Ligament Reconstruction/ 3594 17 exp Anterior Cruciate Ligament/ 10041 18 exp DEBRIDEMENT/ 14696 19 (cruciate and ligament and re*).ti,ab. 18543 20 meniscectomy.ti,ab. 2615 21 chondroplasty.ti,ab. 219 22 articular debridement.ti,ab. 11 23 (synovial and resection).ti,ab. 736 24 tibial plateau fracture.ti,ab. 428 25 (articular and debridement).ti,ab. 679 26 arthroscop*.ti,ab. 26943 27 (knee adj3 surg*).ti,ab. 6509 28 (knee adj3 debridement).ti,ab. 121 29 (knee adj3 lavage).ti,ab. 44 30 Meniscal repair.ti,ab. 821 31 meniscal surgery.ti,ab. 158 32 abrasion arthroplasty.ti,ab. 72 33 loose body removal.ti,ab. 92 34 ACL reconstruction.ti,ab. 5820 35 Lateral release.ti,ab. 541 36 synovectomy.ti,ab. 2733 37 debridement.ti,ab. 21810 38 or/13‐37 130627 39 12 and 38 2401 40 randomized controlled trial.pt. 465865 41 controlled clinical trial.pt. 92545 42 randomized.ab. 418209 43 placebo.ab. 190756 44 drug therapy.fs. 2036694 45 randomly.ab. 294968 46 trial.ab. 435296 47 groups.ab. 1820659 48 or/40‐47 4254364 49 exp animals/ not humans.sh. 4482204 50 48 not 49 3677822 51 39 and 50 572 52 (2017* or 2018*).ed. 1550949 53 51 and 52 55	8.8.18 ‐ 55 14.8.19 ‐ 59
Embase 1974 to present (2017, 2018 AND 2019 ONLY)	1 thrombosis/ 113668 2 thromboembolism/ 56775 3 exp vein thrombosis/ 105035 4 (thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol*).ti,ab. 405573 5 lung embolism/ 74000 6 exp venous thromboembolism/ 120720 7 (PE or DVT or VTE).ti,ab. 72223 8 ((vein* or ven*) adj thromb*).ti,ab. 86363 9 (blood adj3 clot*).ti,ab. 12337 10 (pulmonary adj3 clot*).ti,ab. 266 11 (lung adj3 clot*).ti,ab. 72 12 or/1‐11 563623 13 exp arthroscopy/ 25760 14 exp knee/ 55425 15 exp anterior cruciate ligament reconstruction/ 6875 16 exp anterior cruciate ligament/ 10368 17 exp debridement/ 31642 18 (cruciate and ligament and re*).ti,ab. 20132 19 meniscectomy.ti,ab. 2966 20 chondroplasty.ti,ab. 268 21 articular debridement.ti,ab. 16 22 (synovial and resection).ti,ab. 983 23 tibial plateau fracture.ti,ab. 445 24 (articular and debridement).ti,ab. 772 25 arthroscop*.ti,ab. 31084 26 (knee adj3 surg*).ti,ab. 8508 27 (knee adj3 debridement).ti,ab. 131 28 (knee adj3 lavage).ti,ab. 60 29 Meniscal repair.ti,ab. 902 30 meniscal surgery.ti,ab. 167 31 abrasion arthroplasty.ti,ab. 85 32 loose body removal.ti,ab. 102 33 ACL reconstruction.ti,ab. 6522 34 Lateral release.ti,ab. 584 35 synovectomy.ti,ab. 3139 36 debridement.ti,ab. 24937 37 or/13‐36 149266 38 12 and 37 4857 39 randomized controlled trial/ 487282 40 controlled clinical trial/ 459277 41 random$.ti,ab. 1251895 42 randomization/ 79353 43 intermethod comparison/ 218664 44 placebo.ti,ab. 262465 45 (compare or compared or comparison).ti. 439187 46 ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab. 1678409 47 (open adj label).ti,ab. 61556 48 ((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab. 200591 49 double blind procedure/ 144058 50 parallel group$1.ti,ab. 20847 51 (crossover or cross over).ti,ab. 89652 52 ((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)).ti,ab. 271172 53 (assigned or allocated).ti,ab. 319349 54 (controlled adj7 (study or design or trial)).ti,ab. 280289 55 (volunteer or volunteers).ti,ab. 217206 56 trial.ti. 233978 57 or/39‐56 3856424 58 38 and 57 981 59 (2017* or 2018*).em. 2737847 60 58 and 59 127 61 from 60 keep 1‐127 127	8.8.18 ‐ 127 14.8.19 ‐ 181
CINAHL (2017, 2018 AND 2019 ONLY)	S55 S53 AND S54 9 S54 EM 2017 OR EM 2018 407,857 S53 S12 AND S39 AND S52 118 S52 S40 OR S41 OR S42 OR S43 OR S44 OR S45 OR S46 OR S47 OR S48 OR S49 OR S50 OR S51 344,262 S51 MH "Random Assignment" 39,143 S50 MH "Single‐Blind Studies" or MH "Double‐Blind Studies" or MH "Triple‐Blind Studies" 32,850 S49 MH "Crossover Design" 11,250 S48 MH "Factorial Design" 921 S47 MH "Placebos" 8,370 S46 MH "Clinical Trials" 92,989 S45 TX "multi‐centre study" OR "multi‐center study" OR "multicentre study" OR "multicenter study" OR "multi‐site study" 4,528 S44 TX crossover OR "cross‐over" 0 S43 AB placebo* 28,502 S42 TX random* 220,497 S41 TX trial* 252,026 S40 TX "latin square" 143 S39 S13 OR S14 OR S15 OR S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31 OR S32 OR S33 OR S34 OR S35 OR S36 OR S37 OR S38 30,719 S38 TX debridement 6,175 S37 TX synovectomy 160 S36 TX Lateral release 201 S35 TX ACL reconstruction 1,351 S34 TX loose body removal 53 S33 TX abrasion arthroplasty 17 S32 TX meniscal surgery 331 S31 TX Meniscal repair 272 S30 TX knee n3 lavage 9 S29 TX knee n3 debridement 27 S28 TX knee n3 surg* 5,665 S27 TX arthroscop* 6,153 S26 TX arthroscop* 6,153 S25 TX articular and debridement 228 S24 TX tibial plateau fracture 269 S23 TX synovial and resection 69 S22 TX articular debridement 15 S21 TX chondroplasty 54 S20 TX meniscectomy 498 S19 TX cruciate and ligament and re* 6,305 S18 (MH "Debridement+") 4,302 S17 (MH "Anterior Cruciate Ligament") 1,294 S16 (MH "Anterior Cruciate Ligament Reconstruction") 1,623 S15 (MH "Knee Joint+") 6,997 S14 (MH "Knee") 6,388 S13 (MH "Arthroscopy") 4,526 S12 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 45,219 S11 TX lung n3 clot* 22 S10 TX pulmonary n3 clot* 29 S9 TX blood n3 clot* 914 S8 TX (vein* or ven*) n thromb* 123 S7 TX PE or DVT or VTE 11,061 S6 (MH "Pulmonary Embolism") 4,782 S5 TX thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol* 36,515 S4 (MH "Venous Thrombosis+") 6,383 S3 (MH "Venous Thromboembolism") 3,103 S2 (MH "Thromboembolism") 3,242 S1 (MH "Thrombosis") 4,657	8.8.18 ‐ 9 14.8.19 ‐ 35
AMED (Allied and Complementary Medicine) 1985 to July 2018 (2017, 2018 AND 2019 ONLY)	1 thrombosis/ 201 2 thromboembolism/ 72 3 (thrombus* or thrombopro* or thrombotic* or thrombolic* or thromboemboli* or thrombos* or embol*).ti,ab. 648 4 (PE or DVT or VTE).ti,ab. 246 5 ((vein* or ven*) adj thromb*).ti,ab. 311 6 (blood adj3 clot*).ti,ab. 34 7 (pulmonary adj3 clot*).ti,ab. 0 8 (lung adj3 clot*).ti,ab. 0 9 or/1‐8 873 10 exp Arthroscopy/ 444 11 exp Knee/ 1527 12 exp Knee joint/ 3706 13 exp Anterior Cruciate Ligament Reconstruction/ 268 14 exp Anterior cruciate ligament/ 1456 15 exp debridement/ 99 16 (cruciate and ligament and re*).ti,ab. 1559 17 meniscectomy.ti,ab. 100 18 chondroplasty.ti,ab. 12 19 articular debridement.ti,ab. 0 20 (synovial and resection).ti,ab. 3 21 tibial plateau fracture.ti,ab. 9 22 (articular and debridement).ti,ab. 24 23 arthroscop*.ti,ab. 801 24 (knee adj3 surg*).ti,ab. 325 25 (knee adj3 debridement).ti,ab. 1 26 (knee adj3 lavage).ti,ab. 2 27 Meniscal repair.ti,ab. 20 28 meniscal surgery.ti,ab. 7 29 abrasion arthroplasty.ti,ab. 3 30 loose body removal.ti,ab. 6 31 ACL reconstruction.ti,ab. 352 32 Lateral release.ti,ab. 30 33 synovectomy.ti,ab. 39 34 debridement.ti,ab. 335 35 or/10‐34 7585 36 9 and 35 27 37 exp CLINICAL TRIALS/ 3766 38 RANDOM ALLOCATION/ 314 39 DOUBLE BLIND METHOD/ 661 40 Clinical trial.pt. 1212 41 (clinic* adj trial*).tw. 5410 42 ((singl* or doubl* or trebl* or tripl*) adj (blind* or mask*)).tw. 2849 43 PLACEBOS/ 590 44 placebo*.tw. 3118 45 random*.tw. 17631 46 PROSPECTIVE STUDIES/ 1109 47 or/37‐46 22657 48 36 and 47 5 49 ("2017" or "2018").yr. 2693 50 48 and 49 0	8.8.18 ‐ 0 14.8.19 ‐ 0

Data and analyses

Comparison 1. Low molecular weight heparin versus control (no prophylactic treatment).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Pulmonary embolism	3	1820	Risk Ratio (IV, Fixed, 95% CI)	1.81 [0.49, 6.65]
1.2 Symptomatic DVT	4	1848	Risk Ratio (IV, Fixed, 95% CI)	0.61 [0.18, 2.03]
1.3 Asymptomatic DVT	2	369	Risk Ratio (IV, Fixed, 95% CI)	0.14 [0.03, 0.61]
1.4 Adverse events	5	1978	Risk Ratio (IV, Fixed, 95% CI)	1.85 [0.95, 3.59]
1.5 Major bleeding	1	1451	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.06, 15.72]
1.6 Minor bleeding	5	1978	Risk Ratio (IV, Fixed, 95% CI)	1.79 [0.84, 3.84]

Comparison 2. Rivaroxaban versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Symptomatic DVT	1	234	Risk Ratio (IV, Fixed, 95% CI)	0.16 [0.02, 1.29]
2.2 Asymptomatic DVT	1	234	Risk Ratio (IV, Fixed, 95% CI)	0.95 [0.06, 15.01]
2.3 Minor bleeding	1	234	Risk Ratio (IV, Fixed, 95% CI)	0.63 [0.18, 2.19]

Comparison 3. Low molecular weight heparin versus compression stockings.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Pulmonary embolism	1	1317	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.14, 7.05]
3.2 Symptomatic DVT	1	1317	Risk Ratio (IV, Fixed, 95% CI)	0.17 [0.04, 0.75]
3.3 Asymptomatic DVT	1	1317	Risk Ratio (IV, Fixed, 95% CI)	0.47 [0.21, 1.09]
3.4 Adverse events	1		Risk Ratio (IV, Fixed, 95% CI)	Subtotals only
3.4.1 Pain, tenderness and edema	1	1317	Risk Ratio (IV, Fixed, 95% CI)	1.17 [0.95, 1.43]
3.4.2 Major bleeding	1	1317	Risk Ratio (IV, Fixed, 95% CI)	3.01 [0.61, 14.88]
3.4.3 Minor bleeding	1	1317	Risk Ratio (IV, Fixed, 95% CI)	1.16 [0.64, 2.08]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Camporese 2008.

Study characteristics
Methods	Study design: assessor‐blinded, randomized, controlled trial Method of randomization: a list of random assignments was generated by using nQuery Advisor with a block size of 10 participants and no stratification Concealment of allocation: sealed opaque envelope Losses to follow up: none Intention to treat: yes
Participants	Country: Italy Setting: hospital outpatient department Numbers: 1761; group A: compression stockings (660 patients) 7 days; group B LMWH 7 days (657 patients); group C LMWH 14 days (444 patients). LMWH 14 days arm was stopped early and is not included in this analysis Age: (mean) control group: 42.3 (14.4), treatment 1: 41.9 (15.1), treatment 2: 42.5 (16.7) Sex distribution male/female: control group: 1.66:1, treatment 1: 1.62:1, treatment 2: 1.60:1 Inclusion criteria: patients undergoing diagnostic arthroscopy or arthroscopy‐assisted knee surgery for partial meniscectomy, cartilage shaving, cruciate ligament reconstruction, synovial resection, or combined surgical procedures were eligible for the study Exclusion criteria: patients younger than 18 years of age, pregnant, previous VTE, active cancer, known thrombophilia, receiving mandatory anticoagulation, hypersensitive to LMWH, recent major bleeding event, severe renal or hepatic failure, anticipated poor adherence, geographic inaccessibility, or tourniquet thigh time greater than 1 hour All patients had bilateral, whole‐leg, color‐coded compression ultrasonography at the end of prophylaxis (8 or 15 days) or earlier if clinically indicated. Before ultrasonography, a nurse interviewed patients by using a standardized questionnaire to determine whether they had symptoms of DVT or PE and whether they were adhering to the prophylactic regimen. Patients reporting 1 or more symptoms were considered symptomatic.
Interventions	Full length graduated compression stockings for 7 days (group A); or Nadroparin 0.4 mL sc once daily for 7 days (group B); or Nadroparin 0.4 mL sc once daily for 14 days (group C)
Outcomes	The primary efficacy end point was the 3‐month cumulative incidence of asymptomatic proximal DVT, symptomatic VTE, and all‐cause mortality The secondary efficacy end point was the primary efficacy end point plus the 3‐month cumulative incidence of asymptomatic distal DVT The secondary safety end point was the 7‐ or 14‐day cumulative incidence of bleeding events
Notes	One of the selected outcomes 'asymptomatic DVT' is considered in this particular trial a surrogate of DVT. Patients outside a clinical trial setting will not routinely have a compression ultrasound scan. We will discuss this particularity in the summary of findings tables using GRADE for individual outcomes as less objective outcomes at higher risk of bias.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The investigators describe a random component in the sequence generation process
Allocation concealment (selection bias)	Low risk	Participants and investigators enrolling participants could not foresee assignment. Sequentially numbered, opaque, sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants were aware of their allocated treatment
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Ultrasound operators and nurses gathering DVT symptoms were blinded to the intervention. Hard outcomes such as PE and symptomatic DVT unlikely to be affected by lack of blinding but the outcome asymptomatic DVT has a risk of bias due to lack of placebo and also due to imprecision of the diagnostic test used
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for missing outcome data unlikely to be related to intervention
Selective reporting (reporting bias)	Low risk	The study protocol is available and all of the study’s pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way. However, the authors present and based their conclusions from the combined outcome (death, PE, symptomatic and asymptomatic DVT), individual outcome data can be extracted from the study tables
Other bias	Unclear risk	The data and safety monitoring board stopped the 14 day treatment group during a planned interim analysis and the study continued as a 2‐group trial ‐ control vs 7 day treatment. The 14 day treatment group is not included in this analysis

Camporese 2016.

Study characteristics
Methods	Study design: Phase II, Multicentre, double‐blind, placebo‐controlled randomized study Method of randomization: patients were centrally randomized. Investigators called a randomization centre, reporting the stratification variables, and a telephone operator, using a computer‐generated sequence communicated the drug‐box number assigned to the patient Concealment of allocation: each drug‐box was unlabelled, except for a sequential code only known to the statistical consultant, filled with seven white tablets of identical appearance, either containing rivaroxaban or placebo Losses to follow up: none Intention to treat: yes
Participants	Country: Italy Setting: hospital, multicenter Number: 241 patients (122 treatment group, 119 placebo group) Age: (mean) treatment group 44.9 years (12.8); control group 45.9 years (13.9) Sex: treatment group 64% males 36% females; control group 71% males, 29% females Inclusion criteria: consecutive patients aged at least 18 years, scheduled for non‐diagnostic arthroscopy‐assisted knee surgery, not combined with open surgery were eligible for inclusion Exclusion criteria: concomitant strong concurrent CYP3A4‐inhibitors and/or P‐gp‐inhibitors; proven hypersensitivity to the study drug; pregnancy or lactation;advanced hepatic disease; known thrombophilia; mandatory anticoagulation; previous objectively documented VTE; known severe bleeding tendency; clinically significant active bleeding; severe renal failure ; recent (6 to 12 weeks) major surgery; current involvement in another clinical trial
Interventions	Treatment: 10 mg oral rivaroxaban once daily, starting between 8 ‐ 10 hours at the end of the procedure Control: placebo once daily, starting between 8 ‐ 10 hours at the end of the procedure Duration: 6 days
Outcomes	The primary efficacy end point was the 3‐month cumulative incidence of combined outcomes: asymptomatic proximal DVT, symptomatic VTE, and all‐cause mortality The primary safety end point was the 7‐ or 14‐day cumulative incidence of major and clinically relevant non major bleeding events The secondary efficacy end point was the primary efficacy end point plus the 3‐month cumulative incidence of asymptomatic distal DVT The secondary safety end point was the 7‐ or 14‐day cumulative incidence of bleeding events
Notes	One of the selected outcomes 'asymptomatic DVT' is considered a surrogate of DVT. Patients outside a clinical trial setting will not normally have a compression ultrasound study to look for asymptomatic DVT. We will discuss this particularity in the summary of findings tables using the GRADE classification for individual outcomes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated sequence
Allocation concealment (selection bias)	Low risk	Each drug‐box was unlabelled, except for a sequential code only known to the statistical consultant, filled with seven white tablets of identical appearance, either containing rivaroxaban or placebo
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of participants and key study personnel ensured, and unlikely that the blinding could have been violated
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding of outcome assessment ensured, and unlikely that the blinding could have been broken
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups
Selective reporting (reporting bias)	Low risk	The study protocol is available and all of the study’s pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way
Other bias	Low risk	No other source of bias identified

Canata 2003.

Study characteristics
Methods	Study design: prospective, randomized controlled trial Method of randomization: not reported Concealment of allocation: not reported Losses to follow‐up: not reported Intention to treat: not reported
Participants	Country: Italy Setting: hospital Number: 36; 18 in each group Age: (mean) treatment group 29.6 years (range 22 to 59); control group 32.5 years (range 16 to 59) Sex: treatment group 12 males, 6 females; control group 13 males, 5 females Inclusion criteria: symptomatic ACL‐deficient knees Exclusion criteria: not reported
Interventions	Treatment: LMWH (enoxaparin sodium) sc daily (no dose specified) Control: no intervention Duration: 6 days
Outcomes	Compression color‐coded sonography in case of clinically‐suspected VTE; major and minor bleeding; range of motion
Notes	Surgery: ACL reconstruction performed by the same surgeon using a single technique and the same rehabilitation protocol. This is a small sized study presented in a conference proceeding with some information missing from the abstract. Both groups were balanced in terms of demographics and DVT risk factors
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The conference proceeding does not provide information on the randomization method
Allocation concealment (selection bias)	Unclear risk	The conference proceeding does not provide information on the randomization method
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Patients were asked not to reveal treatment arm. Personnel were blinded to the allocated treatment
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcome assessors blinded to allocated treatment
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Outcome data was not complete
Selective reporting (reporting bias)	Low risk	The protocol was not available but the publication (conference proceeding) reported relevant outcomes
Other bias	Low risk	No other source of bias identified

Kaye 2015.

Study characteristics
Methods	Study design: single‐center, randomized, single‐blind prospective trial Method of randomization: random Concealment of allocation: sealed envelope Losses to follow up: none Intention to treat: yes
Participants	Country: USA Setting: hospital outpatient department Number: 170; treatment group 66; control group 104 Age: (mean) treatment group 46 years control group 43.3, overall 44.4 (± 14.4) Sex: 104 men (61%) and 66 women (39%) Inclusion criteria: patients from 18 to 70 years who were scheduled for arthroscopic knee surgery were included Exclusion criteria: pregnancy; had been diagnosed with thrombocytopenia, anemia, coagulopathy; were undergoing active treatment for DVT; if they had a history of bleeding, of DVT, or of PE; had active infections or undergoing active treatment for cancer; had hypersensitivity or allergy to aspirin; were smokers; or were using oral contraceptive or hormonal medications All patients had a bilateral, whole leg, compression venous duplex ultrasonography 10 to 14 days postoperatively
Interventions	Treatment: 325 mg aspirin for 14 days starting on postoperative day 1 Control: control group (nonpharmacological) advised not to take any NSAIDs for the first 14 days postoperatively Duration: 4 weeks: (1) primary outcome was the development of postoperative DVT or PE within 10 to 14 days postoperatively. (2) secondary outcome of any complication development was assessed at 1 and 4‐week postoperative visits
Outcomes	The primary outcome measure was the development of postoperative DVT or PE. A secondary outcome measure was the development of any complications. Patients were evaluated at two postoperative visits, at 1 week and 4 weeks postoperatively
Notes	One of the selected outcomes 'asymptomatic DVT' is considered a surrogate of DVT. Patients outside a clinical trial setting will not normally have a compression ultrasonography study to look for DVT. We will discuss this particularity in the summary of findings tables using GRADE for individual outcomes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequencial randomization used but no further details provided
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Patients and personnel were not blinded to the intervention, but radiologists were blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	A single radiologist, who was unaware of the allocation group of the patients, performed the ultrasound evaluations'. Operators were blinded to the allocated treatment. We consider that the risk of bias for symptomatic thromboembolic events was low but for asymptomatic thromboembolic events there is a higher risk of bias
Incomplete outcome data (attrition bias) All outcomes	Low risk	Losses explained. Because the loss of follow up was balanced between groups the impact on the outcomes is probably low
Selective reporting (reporting bias)	Low risk	No evidence of reporting bias
Other bias	Low risk	No other source of bias identified

Michot 2002.

Study characteristics
Methods	Study design: prospective, single‐blind, randomized controlled trial Method of randomization: not stated Concealment of allocation: sealed envelope Losses to follow up: treatment group 5; control group 1 Intention to treat: yes
Participants	Country: Switzerland Setting: hospital outpatient department Number: 130; treatment group 66; control group 64 Age: (mean) treatment group 42.0 years (SD 14.7); control group 46.5 years (SD 13.2) Sex: treatment group 40 males, 26 females; control group 46 males, 18 females Inclusion criteria: patients requiring diagnostic or therapeutic arthroscopic knee surgery as outpatients; aged 18 to 80 years Exclusion criteria: inability or unwillingness to give written informed consent; past medical history of DVT or PE, known deficiency of AT III, Protein C or Protein S; ongoing anti‐thrombotic therapy, history of GI bleeding in the previous 2 weeks; hypersensitivity to heparin; history of CVA in the previous 6 months; or severe renal or hepatic failure All patients underwent bilateral compression ultrasonography of the legs the day before the operation Preoperative ultrasonography was inconclusive in 5 patients and revealed DVT in 2 patients
Interventions	Treatment: first dose of LMWH sc (2,500 IU anti‐FXa deltaparin; Low Liquemin, Roche, Basel, Switzerland) 60 to 120 minutes before starting the procedure Six hours after the end of the operation, a second, weight‐adapted dose (2500 IU if weight < 70 kg, 5000 if > 70 kg) was administered, and daily up to 30 days postoperatively Control: no treatment Duration: 30 days
Outcomes	Patients were seen on days 12 and 31 post operation for systematic questioning for symptoms of DVT and PE, or bleeding complications and bilateral compression ultrasonography (US). If US was not conclusive, venography was performed
Notes	A wide variety of procedures were carried out. One of the selected outcomes 'asymptomatic DVT' is considered a surrogate of DVT. Patients outside a clinical trial setting will not normally have a compression ultrasonography study to look for DVT. We will discuss this particularity in the summary of findings tables using GRADE for individual outcomes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients were assigned to a control group or an active treatment group according to sealed envelopes, groups were balanced, no details on how randomization was achieved
Allocation concealment (selection bias)	Low risk	Sealed envelopes used and groups were balanced
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding not possible and personnel and patients aware of treatment arm
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Ultrasonography was performed by 1 of 2 angiologists experienced in the technique. Pathologic findings were reviewed by the second of these 2 physicians. Described the ultrasound operator as blinded as patients instructed to cover thighs and not to reveal the treatment arm so there is possible risk of bias
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 4% lost to follow up
Selective reporting (reporting bias)	Low risk	No evidence of reporting bias
Other bias	Unclear risk	Sample size was calculated at 400 patients but the trial was stopped at 130 as it was decided that withholding LMWH was unethical

Roth 1995.

Study characteristics
Methods	Study design: prospective, randomized controlled trial Method of randomizations: not stated Concealment of allocation: not stated Exclusions post randomizations: 22 due to non‐compliance Intention to treat: no Losses of follow up: not reported
Participants	Country: Germany Setting: hospital outpatient department Number: 122, (n = 61 in each group) (144 randomized, 22 excluded) Age: included patients more than 60 years old Sex: not stated Inclusion criteria: patients undergoing ambulatory arthroscopic meniscus intervention, sinovectomy, chondroplasty, loose‐bodies resection Included patients with independent risk factors for thrombosis
Interventions	Treatment: 0.3 mL sc Fraxiparine 2 hours before the operation and self administered daily (except the first two doses) for 4 days after surgery Control: no treatment Duration: 4 days; no full weight bearing was allowed up to the 5th day
Outcomes	Lower limb DVT and PE; bleeding; hematomas; allergic reactions were clinically evaluated 6 to 8 weeks after surgery
Notes	22 patients were excluded due to non‐compliance
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not addressed. Patients were advised to contact the clinic if they experienced symptoms of DVT
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not described
Selective reporting (reporting bias)	Unclear risk	Did not fully report on all planned outcomes
Other bias	Low risk	No other potential sources of bias identified

van Adrichem 2017.

Study characteristics
Methods	Study design: two parallel, pragmatic, multicenter, randomized, controlled, open‐label trials with blinded outcome evaluation: the POT‐KAST trial, which included patients undergoing knee arthroscopy, and the POT‐CAST trial, which included patients treated with casting of the lower leg Method of randomization: block randomization with variable block sizes Concealment of allocation: to ensure concealment of treatment assignment, the data management unit, physicians, and researchers were unaware of the randomization scheme and block sizes Losses to follow up: treatment group 5; control group 1 Intention to treat: yes
Participants	Country: Netherlands Setting: hospital outpatient department Number: 1543, 773 treatment group and 770 controls Age: (mean) 48.5 ± 12.5 years Sex: 55.8% were men Inclusion criteria: patients 18 years of age or older who were scheduled to undergo knee arthroscopy for meniscectomy, diagnostic arthroscopy, removal of loose bodies, or other indications were eligible for inclusion in the POT‐KAST trial. Patients 18 years of age or older who presented to the emergency department and were treated for at least 1 week with casting of the lower leg (with or without surgery before or after casting but without multiple traumatic injuries) were eligible for inclusion in the POT‐CAST trial
Interventions	Treatment: In the POT‐KAST trial, LMWH was administered once daily for the 8 days after arthroscopy; the first dose was administered postoperatively but before discharge on the day of surgery. In the POT‐CAST trial, LMWH was administered for the full period of immobilization; the first dose was administered in the emergency department. In both trials, patients in the treatment group received nadroparin or dalteparin (according to the preference at the hospital), administered sc; a dose of 2850 IU of nadroparin or 2500 IU of dalteparin was used for patients who weighed 100 kg or less, and a double dose (in one daily injection) was used for patients who weighed more than 100 kg. Only the POT‐KAST data was extracted for this meta‐analysis Control: no treatment Duration: 30 days
Outcomes	The primary safety outcome was the cumulative incidence of major bleeding. The cumulative incidence of clinically relevant non‐major bleeding was a secondary outcome, and all other cases of hemorrhage were recorded as minor bleeding. All possible primary and secondary outcome events were evaluated and assessed by an independent outcome adjudication committee whose members were unaware of the treatment assignments
Notes	Authors presented both ITT and PPA analysis due to drop out rates
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Eligible patients were randomly assigned to the treatment group or the control group in a 1:1 ratio. Block randomization with variable block sizes was used. Randomization was performed centrally with the use of ProMISe software (Leiden University Medical Center) by a data‐management unit
Allocation concealment (selection bias)	Low risk	Due to central randomization, the data management unit, physicians, and researchers were unaware of the randomization scheme and block sizes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Patients were aware of the treatment assignment. Patients with PE and DVT were identified through hospital registries, therefore the lack of blinding did not impact the study outcomes
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All possible primary and secondary outcome events were evaluated and assessed by an independent outcome adjudication committee whose members were unaware of the treatment assignments
Incomplete outcome data (attrition bias) All outcomes	Low risk	After randomization, 30 patients (10 in the treatment group and 20 in the control group) were excluded because they had not met the inclusion criteria or had met exclusion criteria. Of the remaining participants, 37 withdrew consent and 25 were lost to follow‐up. A total of 731 patients in the treatment group and 720 in the control group were included in the intention‐to‐treat population
Selective reporting (reporting bias)	Low risk	Digital (online) or postal questionnaires on the occurrence of trial outcome events and adherence to the trial regimen were sent 2 weeks and 6 weeks after the start of follow‐up. When an outcome event was suspected to have occurred in a patient, detailed information was collected from the patient’s electronic hospital files and radiology reports. Data were collected centrally in an online database management system. The authors reported all outcomes and presented the results per protocol and ITT.
Other bias	Low risk	No other potential sources of bias detected

Wirth 2001.

Study characteristics
Methods	Study design: prospective, randomized, controlled, blinded‐outcome assessment trial Method of randomization: not stated Concealment of allocation: not stated Loss of follow‐up: no Intention to treat: yes
Participants	Country: Germany Setting: hospital Number: 239, treatment group 117; control group 122; 222 completed the trial, 111 in each group Age: year ± SD, treatment group 37.6 ± 13.0; control group 38.5 ±11.6 Sex: treatment group 81 males, 36 females; control group 98 males, 24 females Inclusion criteria: elective knee arthroscopy Exclusion criteria: pregnant; < 18 years; history of DVT; or contraindication to contrast venography or trial medication Patients also screened for additional risk factors (obesity, nicotine abuse, oral contraceptives and family history of thrombosis). If 3 or more present, patients were excluded
Interventions	Treatment: once daily injection of reviparin (1,750 anti Xa IU equivalent to 0.25 mL, sc) (Clivarin; Knoll Deutchland GmbH, Ludwingshafen, Germany) Control: no treatment Duration: 7 to 10 days
Outcomes	Primary: DVT diagnosed by compression color‐coded ultrasonography and clinically symptomatic PE at days 7 to 10, or earlier in case of clinical symptoms
Notes	4 patients (not having had tourniquet and/or general anesthesia) were excluded to achieve homogeneity. Population size was calculated to be 196 patients in each group to detect a significant difference with an alpha of 0.05 and a power of 0.80, 262 patients were eligible, 23 met exclusion criteria. An interim analysis was planned after completion of 50% or 15 thromboembolic events, whichever occurred first. One of the selected outcomes ''asymptomatic DVT'' is considered a surrogate of DVT. Patients outside a clinical trial setting will not normally have a compression ultrasonography study to look for DVT. We will discuss this particularity in the summary of findings tables using GRADE for individual outcomes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly assigned to 1 of 2 groups, no further details on the randomization sequence were given but groups were balanced so we judged the study to have a low risk of bias
Allocation concealment (selection bias)	Unclear risk	Limitied detail provided in relation to allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Patients were not blinded to the treatment but this was not likely to affect objective outcomes like PE or symptomatic DVT
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Ultrasonography was performed by a panel of 3 experts blinded to the randomization and the clinical status of the patients
Incomplete outcome data (attrition bias) All outcomes	Low risk	8% lost to follow up
Selective reporting (reporting bias)	Low risk	Complete outcome reporting. No protocol available but all expected outcomes reported.
Other bias	Low risk	No other potential source of bias identified

ACL: anterior cruciate ligament
AT III: antithrombin III
CVA: cerebrovascular accident
DVT: deep vein thrombosis
GI: gastrointestinal
ITT: intention‐to‐treat
LMWH: low‐molecular‐weight heparin
NSAIDs: non steroidal anti‐inflammatory drugs
PE: pulmonary embolism
sc: subcutaneous
VTE: venous thromboembolism

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Marlovits 2004	Aimed to evaluate the efficacy of an extended treatment with LMWH. Patients were randomly assigned to receive LMWH until days 3 to 8 or until 20 days after surgery. As the goal of this trial was to assess the efficacy of an extended prophylaxis treatment, there is no comparison with no treatment or appropriate control group

LMWH: low‐molecular‐weight heparin

Differences between protocol and review

For this 2019 update, we have separated the outcome of 'Proximal and distal DVT events' to symptomatic (symptomatic DVT, and asymptomatic thrombotic events (diagnosed by echo‐Doppler and/or compression ultrasound scan). We have added a 'Summary of findings' table and updated the review text to reflect current Cochrane standards, including the use of GRADE criteria to assess the certainty of the evidence.

Contributions of authors

CP: extracted data from selected trials, analyzed the results, reviewed risk of bias, completed the SOF, revised the review text including conclusion
JC: extracted data and entered data into the tables, identified and selected potential trials
GB: identified and selected potential trials, evaluated trial quality, and extracted data
JR: identified and selected potential trials, evaluated trial quality, and extracted data

Sources of support

Internal sources

• No sources of support supplied

External sources

• Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK

  The Cochrane Vascular editorial base is supported by the Chief Scientist Office.

• National Academy of Medicine, Centre for Epidemiologic Research, Cochrane Center, Argentina

Declarations of interest

CP: none known
JC: has declared that he has received payment for consultancy from Arthrex, Smith and Nephew and CONMED. None of those are relevant to this work. His institution receives money from AOSSM, OREF, and AANA
GB: none known
JR: none known

New search for studies and content updated (no change to conclusions)