Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Review - Intervention

Combined intermittent pneumatic leg compression and pharmacological prophylaxis for prevention of venous thromboembolism

Authors' declarations of interest

Version published: 28 January 2022 Version history

https://doi.org/10.1002/14651858.CD005258.pub4
Collapse all Expand all

Background

It is generally assumed by practitioners and guideline authors that combined modalities (methods of treatment) are more effective than single modalities in preventing venous thromboembolism (VTE), defined as deep vein thrombosis (DVT) or pulmonary embolism (PE), or both. This is the second update of the review first published in 2008.

Objectives

The aim of this review was to assess the efficacy of combined intermittent pneumatic leg compression (IPC) and pharmacological prophylaxis compared to single modalities in preventing VTE.

Search methods

The Cochrane Vascular Information Specialist searched the Cochrane Vascular Specialised Register, CENTRAL, MEDLINE, Embase, CINAHL, and AMED databases, and World Health Organization International Clinical Trials Registry Platform and ClinicalTrials.gov trials registers to 18 January 2021. We searched the reference lists of relevant articles for additional studies. 

Selection criteria

We included randomised controlled trials (RCTs) or controlled clinical trials (CCTs) of combined IPC and pharmacological interventions used to prevent VTE compared to either intervention individually.

Data collection and analysis

We independently selected studies, applied Cochrane's risk of bias tool, and extracted data. We resolved disagreements by discussion. We performed fixed‐effect model meta‐analyses with odds ratios (ORs) and 95% confidence intervals (CIs). We used a random‐effects model when there was heterogeneity. We assessed the certainty of the evidence using GRADE. The outcomes of interest were PE, DVT, bleeding and major bleeding.

Main results

We included a total of 34 studies involving 14,931 participants, mainly undergoing surgery or admitted with trauma. Twenty‐five studies were RCTs (12,672 participants) and nine were CCTs (2259 participants). Overall, the risk of bias was mostly unclear or high. We used GRADE to assess the certainty of the evidence and this was downgraded due to the risk of bias, imprecision or indirectness.

The addition of pharmacological prophylaxis to IPC compared with IPC alone reduced the incidence of symptomatic PE from 1.34% (34/2530) in the IPC group to 0.65% (19/2932) in the combined group (OR 0.51, 95% CI 0.29 to 0.91; 19 studies, 5462 participants, low‐certainty evidence). The incidence of DVT was 3.81% in the IPC group and 2.03% in the combined group showing a reduced incidence of DVT in favour of the combined group (OR 0.51, 95% CI 0.36 to 0.72; 18 studies, 5394 participants, low‐certainty evidence). The addition of pharmacological prophylaxis to IPC, however, increased the risk of any bleeding compared to IPC alone: 0.95% (22/2304) in the IPC group and 5.88% (137/2330) in the combined group (OR 6.02, 95% CI 3.88 to 9.35; 13 studies, 4634 participants, very low‐certainty evidence). Major bleeding followed a similar pattern: 0.34% (7/2054) in the IPC group compared to 2.21% (46/2079) in the combined group (OR 5.77, 95% CI 2.81 to 11.83; 12 studies, 4133 participants, very low‐certainty evidence).

Tests for subgroup differences between orthopaedic and non‐orthopaedic surgery participants were not possible for PE incidence as no PE events were reported in the orthopaedic subgroup. No difference was detected between orthopaedic and non‐orthopaedic surgery participants for DVT incidence (test for subgroup difference P = 0.19). 

The use of combined IPC and pharmacological prophylaxis modalities compared with pharmacological prophylaxis alone reduced the incidence of PE from 1.84% (61/3318) in the pharmacological prophylaxis group to 0.91% (31/3419) in the combined group (OR 0.46, 95% CI 0.30 to 0.71; 15 studies, 6737 participants, low‐certainty evidence). The incidence of DVT was 9.28% (288/3105) in the pharmacological prophylaxis group and 5.48% (167/3046) in the combined group (OR 0.38, 95% CI 0.21 to 0.70; 17 studies; 6151 participants, high‐certainty evidence). Increased bleeding side effects were not observed for IPC when it was added to anticoagulation (any bleeding: OR 0.87, 95% CI 0.56 to 1.35, 6 studies, 1314 participants, very low‐certainty evidence; major bleeding: OR 1.21, 95% CI 0.35 to 4.18, 5 studies, 908 participants, very low‐certainty evidence).

No difference was detected between the orthopaedic and non‐orthopaedic surgery participants for PE incidence (test for subgroup difference P = 0.82) or for DVT incidence (test for subgroup difference P = 0.69).

Authors' conclusions

Evidence suggests that combining IPC with pharmacological prophylaxis, compared to IPC alone reduces the incidence of both PE and DVT (low‐certainty evidence). Combining IPC with pharmacological prophylaxis, compared to pharmacological prophylaxis alone, reduces the incidence of both PE (low‐certainty evidence) and DVT (high‐certainty evidence). We downgraded due to risk of bias in study methodology and imprecision. Very low‐certainty evidence suggests that the addition of pharmacological prophylaxis to IPC increased the risk of bleeding compared to IPC alone, a side effect not observed when IPC is added to pharmacological prophylaxis (very low‐certainty evidence), as expected for a physical method of thromboprophylaxis. The certainty of the evidence for bleeding was downgraded to very low due to risk of bias in study methodology, imprecision and indirectness. The results of this update agree with current guideline recommendations, which support the use of combined modalities in hospitalised people (limited to those with trauma or undergoing surgery) at risk of developing VTE. More studies on the role of combined modalities in VTE prevention are needed to provide evidence for specific patient groups and to increase our certainty in the evidence.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Are inflatable sleeves and medication effective to prevent deep vein thrombosis and pulmonary embolism after surgery?

Key message

• The use of inflatable sleeves worn on the legs (intermittent pneumatic leg compression) plus medication may reduce the rate of new cases of blood clots in the lungs and legs compared to inflatable sleeves alone.

• The use of inflatable sleeves plus medication compared to medication alone reduces the rate of new cases of blood clots in the legs and may reduce new blood clots in the lungs.

• The addition of a medication to inflatable sleeves, may increase the risk of bleeding compared to inflatable sleeves alone.  

Why is this question important?

Deep vein thrombosis (DVT) and pulmonary embolism are collectively known as venous thromboembolism, and occur when a blood clot develops inside the leg veins and travels to the lungs. They are possible complications of staying in hospital after surgery, trauma or other risk factors. These complications extend hospital stay and are associated with long‐term disability and death. Patients undergoing total hip or knee replacement (orthopaedic) surgery or surgery for colorectal cancer are at high risk of venous thromboembolism. Sluggish blood flow, increased blood clotting and blood vessel wall injury are factors that make it more likely that people will experience a blood clot. Treating more than one of these factors may improve prevention. Mechanical intermittent pneumatic leg compression involves wrapping the legs with inflatable sleeves or using foot pumps. These put gentle pressure on the legs and its veins, reducing sluggish blood flow, while medications such as aspirin and anticoagulants reduce blood clotting. These medications are known as pharmacological prophylaxis (drugs used to prevent blood clots). However, these medications can also increase the risk of bleeding.  We wanted to find out if combining compression and medication to stop blood clots was more effective than either compression or medication alone. 

What did we find?

We searched for studies that compared combined compression and medication against either compression or medication alone. We found 34 studies with a total of 14,931 participants. The mean age of participants, where reported, was 62.7 years. Most participants had either a high‐risk procedure or condition (orthopaedic surgery in 14 studies and urology, cardiothoracic, neurosurgery, trauma, general surgery, gynaecology or other types of participants in the remaining studies).

Compared to compression alone, compression plus medication was better by reducing the rate of new cases of pulmonary embolism (19 studies, 5462 participants). DVT was also reduced for compression combined with medication when compared with compression alone (18 studies, 5394 participants). The addition of a medication to compression, however, increased the risk of any bleeding compared to IPC alone, from 1% to 5.9%. Major bleeding followed a similar pattern, with an increase from 0.3% to 2.2%. Further analysis looking at different types of participants (orthopaedic and non‐orthopaedic participants) showed a similar risk for DVT. It was not possible to assess differences between subgroups for pulmonary embolism.

Compared with medication alone, combined compression and medication was better by reducing pulmonary embolism (15 studies with 6737 participants). DVT was also reduced in the combined compression and medication group (17 studies with 6151 participants). No differences were observed in rates of bleeding (six studies with 1314 participants). Further analysis looking at different subgroups of participants did not show any overall difference in incidence of pulmonary embolism or DVT between orthopaedic and non‐orthopaedic participants.

How certain are we in the evidence?

We found our confidence in the evidence ranged from high to very low. We had concerns on how the studies were carried out, because there were small numbers of clots overall and different definitions used for bleeding between the studies.

How up to date is this evidence?

This review updates our previous evidence. The evidence is current to January 2021.

Authors' conclusions

Implications for practice

Evidence suggests that combining intermittent pneumatic compression (IPC) with pharmacological prophylaxis compared with IPC, reduces the incidence of pulmonary embolism (PE) and deep vein thrombosis (DVT); both low‐certainty evidence. Combining IPC with pharmacological prophylaxis, compared with pharmacological prophylaxis alone also reduced the incidence of PE (low‐certainty evidence) and DVT (high‐certainty evidence). Very low‐certainty evidence suggests that the addition of pharmacological prophylaxis to IPC increased the risk of bleeding compared to IPC alone, a side effect not reported when adding IPC to pharmacological prophylaxis (very low‐certainty evidence), as expected for a physical method of thromboprophylaxis. We downgraded the certainty of the evidence due to risk of bias in study methodology, imprecision and indirectness, highlighting the need for further research. The results of the current review agree with guideline recommendations for hospitalised people at risk of developing venous thromboembolism (VTE). 

Implications for research

Most participants who received combined modalities in the studies reviewed were at high risk of developing VTE. Although the relative VTE reduction was large in this group, the same cannot be extrapolated for patients at moderate risk. Future studies should therefore address this question and also take into account cost‐effectiveness issues; looking at benefits in terms of reduced hospital stay, rehabilitation, mortality and long‐term complications, for example post‐thrombotic syndrome, which add to the burden of disability in the community in the long term. Future research using RCTs in other patient groups (such as patients with stroke or medical ICU patients) and confirmatory RCTs are warranted. Nevertheless, cost‐effectiveness for combined modalities has been demonstrated in certain high‐risk groups by the NICE guidelines (NICE 2009), and also in orthopaedic patients undergoing lower limb arthroplasty (Torrejon Torres 2019). Cost‐effectiveness analysis should be performed in order to define the impact of this policy on health economics in both high‐risk and moderate‐risk patients.

More studies on the role of combined modalities (as opposed to pharmacological prophylaxis alone) in orthopaedic and non‐orthopaedic patients are urgently needed.

Future research should also aim to use standardised bleeding criteria such as those defined by the International Society on Thrombosis and Haemostasis (Schulman 2010).

Further research should compare the efficacy of improved single modalities, including more effective schedule changes, with their combinations (Eriksson 2001; Kakkos 2005a; King 2007; Eriksson 2008). Only two studies in the present review used a direct oral anticoagulant (Sakai 2016; Liu 2017a). Further research on the effect of combined use of recently introduced, improved prophylactic modalities is justified.

Summary of findings

Open in table viewer
Summary of findings 1. IPC plus pharmacological prophylaxis versus IPC alone

Does combined intermittent pneumatic compression (IPC) plus pharmacological prophylaxis increase prevention of venous thromboembolism compared with IPC alone?

Patient or population: people undergoing surgery or at risk of developing VTE due to surgery, trauma or ICU stay

Settings: hospital 

Intervention: combined modalities ‐  IPC plus pharmacological prophylaxis

Comparison: IPC alone

Outcomes

Anticipated absolute effects * (95% CI)

Relative effect
(95% CI)

No of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with IPC alone

Risk with combined modalities

Incidence of PEa

(early postoperative period)

16 per 1000

7 per 1000 (4 to 12)

OR 0.51 (0.29 to 0.91)

5462 (19)

⊕⊕⊝⊝
Lowb

 

Incidence of DVTc

(early postoperative period)

38 per 1000

20 per 1000 (14 to 28)

OR 0.51 (0.36 to 0.72)

5394 (18)

⊕⊕⊝⊝
Lowb

 

Incidence of bleedingd

(early postoperative period)

10 per 1000

55 per 1000 (36 to 83)

OR 6.02 (3.88 to 9.35)

4634 (13)

⊕⊝⊝⊝
Very lowe

 

Incidence of major bleedingf

(early postoperative period)

3 per 1000

19 per 1000 (10 to 39)

OR 5.77 (2.81 to 11.83)

4133 (12)

⊕⊝⊝⊝

Very lowe

 

* 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;ICU: intensive care unit; IPC: intermittent pneumatic compression; OR: odds ratio; PE: pulmonary embolism; 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. 

aPulmonary embolism assessed by pulmonary angiography or scintigraphy, computed tomography (CT), angiography, or autopsy.
bDowngraded by two levels due to risk of bias concerns (high in one or more domains regarding all but one study) and due to imprecision as a result of a small number of overall events.
cDeep vein thrombosis assessed by ascending venography, I‐125 fibrinogen uptake test, and ultrasound scanning.
d Any type of bleeding as described by the study authors.
eDowngraded by three levels due to risk of bias concerns (high in one or more domains regarding all but one study), due to imprecision as a result of a small number of events overall, and indirectness because bleeding definitions were not uniform across the studies.
fMajor bleeding as defined by the study authors, but usually located at the surgical site or in a critical organ or site, requiring intervention or transfusion of at least units of blood, or leading to death.

Open in table viewer
Summary of findings 2. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone

Does combined intermittent pneumatic compression (IPC) plus pharmacological prophylaxis increase prevention of venous thromboembolism compared with pharmacological prophylaxis alone?

Patient or population: people undergoing surgery or at risk of developing VTE because of surgery, trauma or ICU stay

Settings: hospital 

Intervention: combined modalities ‐ IPC plus pharmacological prophylaxis

Comparison: pharmacological prophylaxis alone

Outcomes

Anticipated absolute effects * (95% CI)

Relative effect
(95% CI)

No of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with pharmacological prophylaxis alone

 

Risk with combined modalities

Incidence of PEa

(early postoperative period)

18 per 1000

9 per 1000 (6 to 13)

OR 0.46 (0.30 to 0.71)

6737 (15)

⊕⊕⊝⊝
Lowb

 

Incidence of DVTc

(early postoperative period)

93 per 1000

37 per 1000 (21 to 67)

OR 0.38 (0.21 to 0.70)

6151 (17)

⊕⊕⊕⊕
Highd

 

Incidence of bleedinge

(early postoperative period)

74 per 1000

65 per 1000 (43 to 98)

OR 0.87 (0.56 to 1.35

1314 (6)

⊕⊝⊝⊝

Very lowf

 

Incidence of major bleedingg

(early postoperative period)

11 per 1000

13 per 1000 (4 to 44)

OR 1.21 (0.35 to 4.18)

908 (5)

⊕⊝⊝⊝

Very lowf

 

* 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;ICU: intensive care unit; IPC: intermittent pneumatic compression; OR: odds ratio; PE: pulmonary embolism; 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.

aPulmonary embolism assessed by pulmonary angiography or scintigraphy, computed tomography (CT), angiography, single photon emission CT (SPECT) or autopsy.
bDowngraded by two levels due to risk of bias (which was high in one or more domains regarding all studies) and due to imprecision as a result of a small number of events overall.
cDeep vein thrombosis assessed predominantly by ascending venography, I‐125 fibrinogen uptake test, and ultrasound scanning.
dDowngraded by one level due to risk of bias (which was high in one or more domains regarding all but one study) and upgraded by one level because of a large magnitude of the effect.
eAny type of bleeding as described by the study authors.
fDowngraded by three levels due to risk of bias (which was high in one or more domains regarding all studies), due to imprecision as a result of a small number of events overall and a wide confidence interval, and indirectness because bleeding definitions were not uniform across the studies.
gMajor bleeding as defined by the study authors, but usually located at the surgical site or in a critical organ or site, requiring intervention or transfusion of at least units of blood, or leading to death.

Background

It has been proposed that combined modalities are more effective than single modalities in preventing venous thromboembolism (VTE), defined as deep vein thrombosis (DVT) or pulmonary embolism (PE), or both. This is the second update of the review first published in 2008.

Description of the condition

Deep vein thrombosis (DVT), that is, the development of thrombi (blood clots) inside the deep veins of the legs (in most instances), is a potentially fatal disease as it can be complicated by pulmonary embolism (PE), resulting from the movement of thrombi from the leg veins to the pulmonary artery or its branches. The incidence of venous thromboembolism (VTE), DVT, PE or both, is still high despite the use of contemporary prophylactic measures. VTE risk is increased by the presence of certain risk factors, including, age, malignancy, immobilisation, and the type of surgery. High‐risk patients include those undergoing total hip or knee replacement, or surgery for colorectal cancer (McLeod 2001). Experts in the field have indicated that this and similar observations are the result of failed and also omitted prophylaxis (Goldhaber 2001; Piazza 2007). The most recent guidelines recommend combined pharmacological and mechanical prophylaxis in high‐risk groups, in an effort to maximise thromboprophylaxis (ASH 2019; Gould 2012; Nicolaides 2013). It is likely that mechanical methods increase the efficacy of thromboprophylaxis and reduce death and morbidity rates without increasing bleeding risk.

Description of the intervention

Intermittent pneumatic leg compression (IPC) involves wrapping the legs with inflatable sleeves, using commercially available devices. As a result of sleeve inflation, external pressure is exerted on the legs and its veins, resulting in an increase in blood flow and this reduction of blood stasis decreases the incidence of VTE. Pharmacological prophylaxis on the other hand is achieved by mostly small doses of anticoagulants given orally or subcutaneously; these also significantly reduce the incidence of VTE. Combined IPC and pharmacological prophylaxis in the form of dual modalities concurrently used for prevention of VTE may improve the efficacy of each method used alone.

How the intervention might work

Mechanical methods reduce VTE mainly by reducing venous stasis, while anticoagulants inhibit elements of the thrombosis cascade. Because single prophylactic modalities reduce but do not completely eliminate VTE, combined modalities are expected to reduce further the frequency of VTE because of their different mechanisms of action.

Why it is important to do this review

This is the second update of a Cochrane Review first published in 2008 (Kakkos 2008). VTE is the single most common, preventable cause of postoperative death. Better use of preventive resources is expected to reduce VTE events and mortality. Use of combined modalities is suggested by current guidelines in high‐risk patients, however the evidence supporting these recommendations requires better attention (Gould 2012; Nicolaides 2013). We performed this update to assess the breadth and strength of the best available evidence by pooling data from multiple studies to overcome the limitations of small and underpowered studies. 

Objectives

The aim of this review was to assess the efficacy of combined intermittent pneumatic leg compression (IPC) and pharmacological prophylaxis compared to single modalities in preventing VTE.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) and controlled clinical trials (CCTs).  We excluded studies with non‐standard designs, such as cross‐over trials and cluster‐randomised trials because they were deemed inappropriate in this context.

Types of participants

We included any type of hospitalised patient requiring prevention of venous thromboembolism (VTE) or at risk of developing VTE. We included participants undergoing surgery and trauma and intensive care unit (ICU) patients.

Types of interventions

We included studies that assessed the combined use of IPC (including foot pumps and devices inflating calf sleeves) and pharmacological prophylaxis (including unfractionated heparin and low molecular weight heparin) compared with IPC or pharmacological prophylaxis alone. We excluded studies that used IPC for a short period of time (that is, intraoperatively).

Types of outcome measures

Primary outcomes

  • Incidence of PE, assessed by pulmonary angiography or scintigraphy, computed tomography (CT), angiography and autopsy for PE

  • Incidence of DVT (symptomatic or asymptomatic), assessed by ascending venography, I‐125 fibrinogen uptake test and ultrasound scanning

Secondary outcomes

  • Bleeding: considered as a safety outcome and including all types reported, that is, any type

  • Major bleeding (as defined by the study authors, but usually located at the surgical site or in a critical organ or site, requiring intervention or transfusion of at least two units of blood, or leading to death), and fatal bleeding reported separately

  • Fatal PE, assessed by autopsy

  • Symptomatic DVT, assessed by ascending venography, I‐125 fibrinogen uptake test and ultrasound scanning

Search methods for identification of studies

Electronic searches

The Cochrane Vascular Information Specialist conducted systematic searches of the following databases for RCTs and CCTs without language, publication year or publication status restrictions.

  • Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web searched on 18 January 2021)

  • Cochrane Central Register of Controlled Trials (CENTRAL) Cochrane Register of Studies Online (CRSO 2020, Issue 12)

  • MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE) (searched from 3 May 2016 to 13 January 2021)

  • Embase Ovid (searched from 3 May 2016 to 18 January 2021)

  • CINAHL Ebsco (searched from 3 May 2016 to 18 January 2021)

  • AMED Ovid (searched from 3 May 2016 to 18 January 2021)

The Information Specialist modelled search strategies for other databases on the search strategy designed for CENTRAL. Where appropriate, they were combined with adaptations of the highly sensitive search strategy designed by Cochrane for identifying RCTs and CCTs (as described in the Cochrane Handbook for Systematic Reviews of Interventions chapter 4, Lefebvre 2021). Search strategies for major databases are provided in Appendix 1.

The Information Specialist searched these trials registries on 18 January 2021:

Searching other resources

The review authors searched the reference lists of relevant articles and also similar systematic reviews and meta‐analyses to identify additional studies.  We also carried out clinical trial database searches up to 12 July 2021. See Appendix 2.

Data collection and analysis

Selection of studies

For this update, two review authors (SK and GK) independently selected studies for inclusion on the basis of the use of combined mechanical IPC and pharmacological modalities. We resolved any disagreements by discussion.

Data extraction and management

For this update, two review authors (SK and GK) independently extracted the data. We used a data extraction form to record the type of patient or surgical procedure, total number of participants in the study (including those randomised, excluded and also withdrawn), the interventions used, the number of participants who reached an endpoint (DVT or PE) and the methodology used to establish this. A third review author (JC) arbitrated any disagreements.

Assessment of risk of bias in included studies

For this update, we assessed the methodological quality of included studies using Cochrane's risk of bias tool (RoB 1). SK and GK independently performed the assessment according to Higgins 2011. We assessed the following domains: selection bias (random sequence generation, allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessment), attrition bias (incomplete outcome data), reporting bias (selective reporting) and other bias. We classified the domains as low risk, high risk, or unclear risk according to Higgins 2011. We resolved any disagreements after discussion.

Measures of treatment effect

We performed separate analyses for the interventions of IPC versus combined modalities, and pharmacological prophylaxis versus combined modalities for the outcomes of PE and DVT. We used odds ratios (ORs) with 95% confidence intervals (CIs) for the assessment of dichotomous outcomes.  None of our outcomes of interest were reported as continuous data.

Unit of analysis issues

We excluded studies with non‐standard designs, such as cross‐over trials and cluster‐randomised trials. The individual participant was the unit of analysis.

Dealing with missing data

In case of missing participants due to dropout, we used intention‐to‐treat analysis. Where necessary, we contacted study authors to request that they provided any missing information.

Assessment of heterogeneity

We assessed statistical heterogeneity with the I2 test (Higgins 2003). We considered I2 test levels exceeding 50% as substantial heterogeneity to justify the use of random‐effects model analysis (Deeks 2021). We also considered the magnitude and direction of effects and the strength of evidence for heterogeneity (e.g. P value from the Chi2 test, or a CI for I2 test).

Assessment of reporting biases

We assessed publication bias with funnel plots when 10 or more studies were included in a comparison and contributed to the effect estimate; as described in the Cochrane Handbook for Systematic Reviews of Interventions (Page 2021). Where the number of studies in each comparison was not greater than 10 the plots lack the power to distinguish chance from real asymmetry (Egger 1997).

Data synthesis

We used fixed‐effect models for each meta‐analysis to pool data, unless there was evidence of heterogeneity, in which case we used a random‐effects model to calculate the ORs and 95% CIs (see Assessment of heterogeneity). We only undertook meta‐analyses when it was meaningful to do so. That is, if the treatments, participants, and the underlying clinical question were similar enough for pooling to make sense. If meta‐analysis was not possible, we planned to report the results using a narrative synthesis.

Subgroup analysis and investigation of heterogeneity

To investigate heterogeneity, we performed subgroup analysis of the primary outcomes by:

  • surgery type (orthopaedic surgery compared to non‐orthopaedic surgery or other conditions);

  • type of IPC (foot IPC and other than foot IPC).

Sensitivity analysis

We planned to perform sensitivity analysis of the primary outcomes by excluding studies with a high risk for bias in any one or more domains, based on the RoB 1, and by excluding CCTs, in order to test the robustness of the evidence. We also planned to perform sensitivity analysis of the primary outcomes by excluding studies with medical patients.

Summary of findings and assessment of the certainty of the evidence

We created summary of findings tables for the comparisons of 'IPC plus pharmacological prophylaxis versus IPC alone' (summary of findings Table 1) and 'IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone' (summary of findings Table 2). We used GRADEpro GDT software to present the main findings of the review. We included the outcomes PE, DVT, incidence of bleeding, and incidence of major bleeding in the summary of findings tables. We calculated assumed control intervention risks from the mean number of events in the control groups of the selected studies for each outcome. We used the system developed by the GRADE working group for grading the certainty of the evidence as high, moderate, low and very low, based on within‐study risk of bias, directness of evidence, heterogeneity, precision of effects estimates, and risk of publication bias (Atkins 2004).

Results

Description of studies

Results of the search

See Figure 1.  We identified 12 additional studies for this 2021 update (Arabi 2019; Dong 2018; Hata 2019; Kamachi 2020; Liu 2017a; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Zhou 2020). We also identified four new ongoing studies (ChiCTR1800014257; EUCTR2007‐006206‐24; NCT02271399; NCT03559114 (PROTEST). We did not identify any new excluded studies.

Figure 1
Study flow diagram

Study flow diagram

Included studies

For this update we included 12 additional studies (Arabi 2019; Dong 2018; Hata 2019; Kamachi 2020; Liu 2017a; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Zhou 2020), making a total of 34 studies that met the inclusion criteria (Arabi 2019; Bigg 1992; Borow 1983; Bradley 1993; Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Liu 2017a; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Ramos 1996; Sakai 2016; Sang 2018; Sieber 1997; Silbersack 2004; Siragusa 1994; Stannard 1996; Tsutsumi 2012; Turpie 2007; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011; Zhou 2020; see Figure 1). We identified one additional publication of the full report to a study that had presented interim results included in the last update (Jung 2018). The included studies investigated 14,931 participants. Five publications had three arms (Borow 1983; Cahan 2000; Dickinson 1998; Sang 2018; Stannard 1996), using IPC, pharmacological prophylaxis and both, respectively.

Twenty‐five studies involving a total of 12,672 participants were RCTs (Arabi 2019; Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Hata 2019; Jung 2018; Kamachi 2020; Liu 2017a; Liu 2017b; Nakagawa 2020; Obitsu 2020; Patel 2020; Ramos 1996; Sakai 2016; Silbersack 2004; Siragusa 1994; Stannard 1996; Turpie 2007; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011; Zhou 2020). The remaining nine studies were CCTs, which were classified according to the draft guidelines of the Cochrane Non‐Randomised Studies Methods Group (NRSMG). These included five quasi‐randomised CCTs that involved a total of 1106 participants (Bigg 1992; Bradley 1993; Kurtoglu 2003; Lobastov 2021; Sang 2018), and four CCTs with concurrent controls that involved a total of 1153 participants (Borow 1983; Sieber 1997; Tsutsumi 2012; Westrich 2005). 

Fourteen included studies evaluated orthopaedic patients (Bradley 1993; Edwards 2008; Eisele 2007; Liu 2017a; Liu 2017b; Sakai 2016; Silbersack 2004; Siragusa 1994; Stannard 1996; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011); three evaluated urology patients (Bigg 1992; Patel 2020; Sieber 1997); two evaluated cardiothoracic patients (Dong 2018; Ramos 1996); one study evaluated neurosurgery patients (Dickinson 1998); one, trauma patients (Kurtoglu 2003); one, intensive care, mostly medical, patients (Arabi 2019); and 12 studies evaluated general surgery, gynaecology and other types of patients (Borow 1983; Cahan 2000; Hata 2019; Jung 2018; Kamachi 2020; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Sang 2018; Tsutsumi 2012; Turpie 2007; Zhou 2020). Some 1566 participants (10.5%) had a medical condition requiring ICU stay (Arabi 2019), and the remaining 13,365 participants (89.5%) had trauma or underwent surgery. Participant weighted mean age (in 29 studies that reported age; 11,379 participants) was 62.7 years (Arabi 2019; Bigg 1992; Bradley 1993; Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Hata 2019; Jung 2018; Kamachi 2020; Liu 2017a; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Ramos 1996; Sakai 2016; Sang 2018; Sieber 1997; Silbersack 2004; Stannard 1996; Tsutsumi 2012; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011; Zhou 2020).

Pharmacological prophylaxis included unfractionated heparin (UFH) (Bigg 1992; Bradley 1993; Cahan 2000; Ramos 1996; Patel 2020 Sieber 1997; Siragusa 1994; Stannard 1996), low molecular weight heparin (LMWH) (Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Jung 2018; Kamachi 2020; Kurtoglu 2003; Liu 2017a; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Sang 2018; Silbersack 2004; Westrich 2006; Windisch 2011; Zhou 2020), any heparin type (Arabi 2019), fondaparinux (Tsutsumi 2012; Turpie 2007), LMWH or fondaparinux (Hata 2019; Yokote 2011), UFH or warfarin (Borow 1983), warfarin or aspirin (Westrich 2005; Woolson 1991) and a direct oral anticoagulant/factor Xa inhibitor (Liu 2017b; Sakai 2016).

IPC types included foot pumps (Bradley 1993; Sakai 2016; Stannard 1996; Windisch 2011), and inflating calf sleeve devices (Edwards 2008; Eisele 2007; Kamachi 2020; Silbersack 2004; Westrich 2005; Westrich 2006), or thigh‐high sleeves (Bigg 1992; Borow 1983; Cahan 2000; Dickinson 1998; Jung 2018; Liu 2017b; Lobastov 2021; Ramos 1996; Sang 2018; Sieber 1997; Woolson 1991). One additional study reported the use of sleeves that inflated the foot, calf and thigh (Zhou 2020). Ten studies did not report the exact IPC type (Dong 2018; Hata 2019; Kurtoglu 2003; Liu 2017a; Nakagawa 2020; Obitsu 2020; Patel 2020; Siragusa 1994; Tsutsumi 2012; Yokote 2011), while in two multi‐centre studies the investigators were allowed to use the device type of their choice (Arabi 2019; Turpie 2007).

Five publications had three arms (Borow 1983; Cahan 2000; Dickinson 1998; Sang 2018; Stannard 1996), using IPC, pharmacological prophylaxis and both, respectively. Of the remaining 29 publications, prophylactic methods in the control group included: IPC in 16 studies, either without aspirin (Bigg 1992; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Nakagawa 2020; Obitsu 2020; Patel 2020; Sieber 1997; Tsutsumi 2012; Turpie 2007; Woolson 1991; Yokote 2011) or with aspirin (Westrich 2005; Westrich 2006); and pharmacological prophylaxis in 13 studies (Arabi 2019; Bradley 1993; Edwards 2008; Eisele 2007; Liu 2017a; Liu 2017b; Lobastov 2021; Ramos 1996; Sakai 2016; Silbersack 2004; Siragusa 1994; Windisch 2011; Zhou 2020). The intervention group in all studies used combined modalities and only two studies used aspirin (Stannard 1996; Woolson 1991).

Ultrasound was the main diagnostic modality to diagnose DVT and was used by most studies (Arabi 2019; Borow 1983; Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Hata 2019; Jung 2018; Kurtoglu 2003; Liu 2017a; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Sakai 2016; Sang 2018; Silbersack 2004; Siragusa 1994; Stannard 1996; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011; Zhou 2020). Where reported, PE was diagnosed mainly with scintigraphy scanning (Bigg 1992; Hata 2019; Ramos 1996; Turpie 2007; Woolson 1991) or single photon emission CT (SPECT) (Lobastov 2021), a pulmonary angiogram (Hata 2019; Ramos 1996; Turpie 2007); or a CT pulmonary angiogram (Arabi 2019; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Lobastov 2021; Liu 2017b; Nakagawa 2020; Obitsu 2020; Sakai 2016; Sang 2018; Silbersack 2004; Tsutsumi 2012; Turpie 2007; Westrich 2006; Windisch 2011; Yokote 2011).

Two studies did not report on DVT rates (Bigg 1992; Ramos 1996), and four studies did not report on PE rates (Bradley 1993; Eisele 2007; Siragusa 1994; Zhou 2020).

Nineteen studies reported on bleeding outcomes (Bigg 1992; Dickinson 1998; Hata 2019; Jung 2018; Kamachi 2020; Liu 2017a; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Sakai 2016; Sang 2018; Tsutsumi 2012; Turpie 2007; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011). Many studies provided no specific bleeding definitions (Bigg 1992; Dickinson 1998; Liu 2017a; Patel 2020; Westrich 2005; Westrich 2006; Windisch 2011; Woolson 1991). The remaining studies, which did provide bleeding definitions, did not use uniform criteria (Hata 2019; Jung 2018; Kamachi 2020; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Sakai 2016; Sang 2018; Tsutsumi 2012; Turpie 2007; Yokote 2011).

Excluded studies

For this update, we did not assess any new studies as excluded. We excluded a total of 21 studies in previous versions (Ailawadi 2001; Eskander 1997; Frim 1992; Gagner 2012; Gelfer 2006; Kamran 1998; Kiudelis 2010; Kumaran 2008; Lieberman 1994; Macdonald 2003; Mehta 2010; Nathan 2006; Patel 2010; Roberts 1975; Spinal cord injury investigators; Stannard 2006; Tsutsumi 2000; Wan 2015; Westrich 1996; Whitworth 2011; Winemiller 1999). Exclusions were because five studies' use of combined modalities was not concurrent or a different type of pharmacological prophylaxis was given in the two study groups (Eskander 1997; Gelfer 2006; Macdonald 2003; Spinal cord injury investigators; Stannard 2006); in two studies IPC use was limited to intraoperative use (Kiudelis 2010; Roberts 1975); three studies were controlled before and after studies (Frim 1992; Kamran 1998; Tsutsumi 2000); six studies were retrospective case‐control studies (Ailawadi 2001; Nathan 2006; Patel 2010; Wan 2015; Whitworth 2011; Winemiller 1999); one was a registry study (Gagner 2012); in another the single modalities group used either heparin or IPC (Kumaran 2008); two studies used aspirin for thromboprophylaxis (Lieberman 1994; Westrich 1996), and one study provided only aggregated VTE rates and not separate DVT and PE rates (Mehta 2010).

Ongoing studies

For this update, we identified four additional studies as ongoing (ChiCTR1800014257; EUCTR2007‐006206‐24; NCT02271399; NCT03559114 (PROTEST)), in addition to NCT00740987 (CIREA 2). See Characteristics of ongoing studies.

Risk of bias in included studies

See Figure 2 and Figure 3.

Figure 2
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

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

Figure 3
Risk of bias summary: review authors' judgements about each risk of bias item for each included study

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

We judged the overall risk of bias as mostly unclear or high due to concerns with selection bias, performance bias and detection bias.

Allocation

The randomisation methods were unclear in 12 of the 25 RCTs (Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Silbersack 2004; Siragusa 1994; Stannard 1996; Westrich 2006; Windisch 2011; Yokote 2011; Zhou 2020). Thirteen studies were at low risk of bias as they provided sufficient information using random tables (Liu 2017a; Ramos 1996), a centralised computer‐generated schedule (Arabi 2019; Jung 2018; Kamachi 2020; Nakagawa 2020; Obitsu 2020; Patel 2020; Sakai 2016; Turpie 2007), central randomisations without further details (Hata 2019), and sealed envelopes (Liu 2017b; Woolson 1991). As a result, the quality of RCTs was often poor regarding selection bias, which was generally at high risk. By definition, all quasi‐randomised trials and CCTs had a high risk for random sequence generation and therefore selection bias (Bigg 1992; Borow 1983; Bradley 1993; Kurtoglu 2003; Lobastov 2021; Sang 2018; Sieber 1997; Tsutsumi 2012; Westrich 2005).

A high risk for allocation concealment was evident in 11 studies (Bigg 1992; Borow 1983; Bradley 1993; Jung 2018; Kurtoglu 2003; Lobastov 2021; Sakai 2016; Sang 2018; Sieber 1997; Tsutsumi 2012; Westrich 2005). We deemed two studies with adequate randomisation methods to be at high risk for allocation concealment (Jung 2018; Sakai 2016). Only eight studies had a low risk for allocation bias (Arabi 2019; Hata 2019; Kamachi 2020; Liu 2017b; Nakagawa 2020; Obitsu 2020; Patel 2020; Turpie 2007). In the remaining studies, the risk of selection bias due to allocation concealment was unclear (Cahan 2000; Dickinson 1998; Dong 2018; Edwards 2008; Eisele 2007; Liu 2017a; Ramos 1996; Silbersack 2004; Siragusa 1994; Stannard 1996; Westrich 2006; Windisch 2011; Woolson 1991; Yokote 2011; Zhou 2020).

Blinding

A high risk of performance bias was evident in all studies except two RCTs, which were double‐blinded and so at low risk (Turpie 2007; Yokote 2011). We judged the remaining studies as being at high risk because of the lack of use of a placebo medication or device. Six studies reported blinding of outcome assessment of all outcomes so these were at low risk of detection bias (Bradley 1993; Kurtoglu 2003; Stannard 1996; Turpie 2007; Windisch 2011; Yokote 2011). This was not the case in nine studies judged to be at high risk (Dong 2018; Hata 2019; Kamachi 2020; Liu 2017b; Lobastov 2021; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018), while in the remaining studies there was unclear evidence of detection bias. This lack of blinding may have affected the detection of DVT or PE and potentially increase the heterogeneity of the results.

Incomplete outcome data

A total of 428 participants (2.9%) were excluded and a total of 88 participants (0.59%) were lost to follow‐up. Four studies reported some attrition but were judged to be at low risk of bias as the proportion of participants lost to follow‐up were unlikely to impact the results (Arabi 2019; Nakagawa 2020; Sakai 2016; Yokote 2011). One study excluded eight participants due to non‐compliance, confinement to bed for more than one week, premature transfer to a different institution, or re‐operation or discharge from hospital without ultrasonography (Silbersack 2004). We judged this unlikely to impact the results so deemed it at low risk of attrition bias. Another study excluded 11 participants because of a protocol violation (discharged before the ultrasound (6 participants)), or because they did not receive the correct study medication (5 participants); Westrich 2006). This study also reported a 26.5% loss to follow‐up, which was 0.59% of the total number of participants in this systematic review; short‐term data were provided but we judged this to be at high risk of bias (Westrich 2006). A third study excluded 24 participants because inclusion or exclusion criteria were not met, informed consent was withdrawn, adverse events occurred, or for other reasons not stated so we judged this to be at high risk (Turpie 2007). We deemed a high risk of bias present in a total of seven studies as we judged missing data or exclusions sufficient to impact the results (Dong 2018; Jung 2018; Kamachi 2020; Obitsu 2020; Ramos 1996; Turpie 2007; Westrich 2006). The remaining studies were at low risk of bias (Bigg 1992; Borow 1983; Bradley 1993; Cahan 2000; Dickinson 1998; Eisele 2007; Hata 2019; Kurtoglu 2003; Liu 2017a; Liu 2017b; Lobastov 2021; Patel 2020; Sang 2018; Sieber 1997; Siragusa 1994; Stannard 1996; Tsutsumi 2012; Westrich 2005; Windisch 2011; Woolson 1991; Zhou 2020).

Selective reporting

We identified selective reporting in only one case where symptomatic DVT was not reported, so we judged this study to be at high risk of reporting bias (Liu 2017b). All the other studies were at low risk as expected results were provided.

Other potential sources of bias

We considered two studies as being at high risk for other sources of bias, because they were prematurely stopped (Dickinson 1998; Sakai 2016). Six studies were at unclear risk because of a lack of baseline characteristic details (Bigg 1992; Borow 1983; Kurtoglu 2003; Sieber 1997; Siragusa 1994; Westrich 2005). All the remaining studies were at low risk of other bias.

Effects of interventions

See: Summary of findings 1 IPC plus pharmacological prophylaxis versus IPC alone; Summary of findings 2 IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone

Intermittent pneumatic leg compression (IPC) plus pharmacological prophylaxis versus IPC alone

See summary of findings Table 1. 

Nineteen of the included studies evaluated the role of combined modalities on the incidence of symptomatic PE (Bigg 1992; Borow 1983; Cahan 2000; Dickinson 1998; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Sieber 1997; Stannard 1996; Tsutsumi 2012; Turpie 2007; Woolson 1991; Yokote 2011). The incidence of PE was 0.65% (19/2932) in the combined group compared to 1.34% (34/2530) in the IPC‐alone group, showing a benefit for combined modalities (OR 0.51, 95% CI 0.29 to 0.91; 19 studies, 5462 participants; P = 0.02; low‐certainty evidence; Analysis 1.1). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to low for risk of bias, and imprecision as a result of a small number of events. We carried out subgroup analysis by orthopaedic and non‐orthopaedic patient groups for PE incidence.  As no PE events were reported in the orthopaedic subgroup we were unable to test for differences (Stannard 1996; Woolson 1991; Yokote 2011; Analysis 1.1).

Fatal PE was not reported.

We carried out subgroup analysis by foot IPC versus other IPC on incidence of PE. As no PE events were reported in the foot IPC subgroup we were unable to test for differences (Analysis 1.2).

Eighteen studies investigated the role of combined modalities on the incidence of DVT (Borow 1983; Cahan 2000; Dickinson 1998; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Sieber 1997; Stannard 1996; Tsutsumi 2012; Turpie 2007; Woolson 1991; Yokote 2011). The incidence of DVT was 2.03% (59/2900) in the combined group compared to 3.81% (95/2494) in the IPC group, showing a reduced incidence of DVT in favour of the combined modalities group (OR 0.51, 95% CI 0.36 to 0.72; 18 studies, 5394 participants; P = 0.0001; low‐certainty evidence; Analysis 1.3). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to low for risk of bias, and imprecision as a result of a small number of events. 

No difference was detected between the orthopaedic and non‐orthopaedic subgroups for DVT incidence (test for subgroup differences P = 0.19;  Analysis 1.3).

Ten studies reported on the occurrence of symptomatic DVT (Cahan 2000; Hata 2019; Jung 2018; Kamachi 2020; Nakagawa 2020; Sang 2018; Sieber 1997; Tsutsumi 2012; Turpie 2007; Yokote 2011). The incidence of symptomatic DVT was 0.53% (12/2245) in the combined modalities group compared to 0.70% (13/1844) in the IPC control group, showing no clear evidence of a difference between the groups (OR 0.48, 95% CI 0.21 to 1.10; 10 studies, 4089 participants; P = 0.08; Analysis 1.4). Results did not demonstrate heterogeneity (I2 = 0%).

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of symptomatic DVT (test for subgroup differences P = 0.47; Analysis 1.4).

One study investigated the role of combined modalities on the incidence of DVT using a foot IPC (Stannard 1996), but because of a lack of events we could not calculate a risk estimate. Seventeen studies investigated the role of combined modalities on the incidence of DVT using IPC other than a foot IPC (Borow 1983; Cahan 2000; Dickinson 1998; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Kurtoglu 2003; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Sieber 1997; Tsutsumi 2012; Turpie 2007; Woolson 1991; Yokote 2011). The incidence of DVT was 2.03% (59/2901) in the combined group compared to 3.81% (95/2494) in the IPC group, showing a reduced incidence of DVT in favour of the combined modalities group (OR 0.51, 95% CI 0.36 to 0.72; 17 studies, 5345 participants; P = 0.0001; Analysis 1.5). Results did not demonstrate heterogeneity (I2 = 0%).

Thirteen studies reported on the incidence of bleeding in the combined modalities and IPC groups (Bigg 1992; Dickinson 1998; Hata 2019; Jung 2018; Kamachi 2020; Nakagawa 2020; Obitsu 2020; Patel 2020; Sang 2018; Tsutsumi 2012; Turpie 2007; Woolson 1991; Yokote 2011). The incidence of bleeding was 5.88% (137/2330) in the combined group compared to 0.95% (22/2304) in the IPC group, showing an increase in bleeding in the combined group (OR 6.02, 95% CI 3.88 to 9.35; 13 studies, 4634 participants; P < 0.00001; very low‐certainty evidence; Analysis 1.6). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to very low due to risk of bias, imprecision as a result of a small number of events, and indirectness.

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of bleeding (test for subgroup difference P = 0.22; Analysis 1.6).

Major bleeding followed a similar pattern, with an incidence of 2.21% (46/2079) in the combined group compared to 0.34% (7/2054) in the IPC group (OR 5.77, 95% CI 2.81 to 11.83; 12 studies, 4133 participants; P < 0.00001; very low‐certainty evidence; Analysis 1.7). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to very low for risk of bias, imprecision as a result of a small number of events, and indirectness. 

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of major bleeding (test for subgroup difference P = 0.74;  Analysis 1.7).

Fatal bleeding was not reported during the intervention period.

No publication bias was indicated by investigating comparisons with funnel plots.

Intermittent pneumatic leg compression (IPC) plus pharmacological prophylaxis versus pharmacological prophylaxis alone

See summary of findings Table 2. 

Fifteen studies evaluated the role of combined modalities compared to pharmacological prophylaxis alone on the incidence of mainly symptomatic PE (Arabi 2019; Borow 1983; Bradley 1993; Cahan 2000; Dickinson 1998; Edwards 2008; Liu 2017a; Liu 2017b; Lobastov 2021; Ramos 1996; Sakai 2016; Sang 2018; Silbersack 2004; Stannard 1996; Windisch 2011). The incidence of PE was 0.91% (31/3419) in the combined group compared to 1.84% (61/3318) in the pharmacological prophylaxis control group showing a reduction in PE in favour of the combined modalities group (OR 0.46, 95% CI 0.30 to 0.71; 15 studies, 6737 participants; P = 0.0005; low‐certainty evidence; Analysis 2.1). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to low for risk of bias, and due to imprecision as a result of a small number of events.

No difference was detected between the orthopaedic and non‐orthopaedic subgroups for PE incidence (test for subgroup differences P = 0.82; Analysis 2.1).

Two published studies reported fatal PE (Lobastov 2021; Ramos 1996). Lobastov 2021 reported three cases of fatal PE, exclusively in the pharmacological prophylaxis‐alone group. Ramos 1996 did not provide the exact number of deaths or the treatment group they occurred in. The authors of a third study provided unpublished results; a single fatal PE occurred in the pharmacological prophylaxis‐alone group (Arabi 2019).

We carried out subgroup analysis by foot IPC versus other IPC on incidence of PE. This showed no difference between the two subgroups (test for subgroup differences P = 0.82; Analysis 2.2).

Seventeen studies investigated the role of combined modalities on the incidence of DVT (Arabi 2019; Borow 1983; Bradley 1993; Cahan 2000; Dickinson 1998; Edwards 2008; Eisele 2007; Liu 2017a; Liu 2017b; Lobastov 2021; Sakai 2016; Sang 2018; Silbersack 2004; Siragusa 1994; Stannard 1996; Windisch 2011; Zhou 2020). The incidence of DVT was 5.48% (167/3046) in the combined group compared to 9.28% (288/3105) in the pharmacological prophylaxis control group showing a reduction in DVT in favour of the combined modalities group (OR 0.38, 95% CI 0.21 to 0.70; 17 studies, 6151 participants; P = 0.002; high‐certainty evidence; Analysis 2.3). Results demonstrated substantial heterogeneity so we used a random‐effects model (I2 = 78%). The certainty of the evidence was high (downgraded by one level due to risk of bias and upgraded by one level because of a large magnitude of effect (< 0.50 and > 0.2)).

No difference was detected between the orthopaedic and non‐orthopaedic subgroups for DVT incidence (test for subgroup differences P = 0.69; Analysis 2.3).

Seven studies reported on the occurrence of symptomatic DVT (Cahan 2000; Edwards 2008; Eisele 2007; Lobastov 2021; Sakai 2016; Sang 2018; Windisch 2011). The incidence of symptomatic DVT was 0.59% (9/1515) in the combined group compared to 0.73% (11/1517) in the pharmacological prophylaxis control group, showing no clear difference between the combined and single groups (OR 0.83, 95% CI 0.34 to 2.01; 7 studies, 3032 participants; P = 0.67; Analysis 2.4). Results did not demonstrate heterogeneity (I2 = 0%). 

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of symptomatic DVT (test for subgroup differences P = 0.20; Analysis 2.4).

Four studies investigated the role of combined modalities on the incidence of DVT using a foot IPC (Bradley 1993; Sakai 2016; Stannard 1996; Windisch 2011). The incidence of DVT was 13.07% (20/153) in the combined group compared to 16.37% (28/171) in the pharmacological prophylaxis‐alone control group showing no clear difference between the combined and single groups (OR 0.40, 95% CI 0.05 to 3.47; 4 studies, 324 participants; P = 0.41; Analysis 2.5). Results demonstrated substantial heterogeneity so we used a random‐effects model (I2 = 81%).

Twelve studies investigated the role of combined modalities on the incidence of DVT using IPC other than a foot IPC (Borow 1983; Cahan 2000; Dickinson 1998; Edwards 2008; Eisele 2007; Liu 2017a; Liu 2017b; Lobastov 2021; Sang 2018; Silbersack 2004; Siragusa 1994; Zhou 2020). The incidence of DVT was 2.73% (52/1902) in the combined group compared to 9.11% (175/1922) in the pharmacological prophylaxis‐alone group, showing a reduced incidence of DVT in favour of the combined modalities group (OR 0.31, 95% CI 0.17 to 0.54; 12 studies, 3824 participants; P < 0.0001; Analysis 2.5). The results demonstrated substantial heterogeneity and we used a random‐effects model (I2 = 50%). The test for subgroup differences did not detect any clear difference between the foot IPC or IPC other than foot subgroups (P = 0.81).

Six studies reported on the incidence of bleeding in the combined and pharmacological prophylaxis groups (Dickinson 1998; Liu 2017a; Lobastov 2021; Sakai 2016; Sang 2018; Windisch 2011). These studies showed no clear difference in bleeding rates between the combined group (43/656, 6.55%) and the pharmacological prophylaxis group (49/658, 7.45%; OR 0.87, 95% CI 0.56 to 1.35; 6 studies, 1314 participants; P = 0.53; very low‐certainty evidence; Analysis 2.6). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to very low for risk of bias, indirectness and imprecision.

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of bleeding (test for subgroup difference P = 0.58; Analysis 2.6).

There was also no clear difference in major bleeding rates between the combined group (6/452, 1.33%) and the pharmacological prophylaxis group (5/456, 1.10%; OR 1.21, 95% CI 0.35 to 4.18; 5 studies, 908 participants; P = 0.76; very low‐certainty evidence; Analysis 2.7). Results did not demonstrate heterogeneity (I2 = 0%). We downgraded the certainty of the evidence to very low for risk of bias, indirectness and imprecision. 

We did not detect a difference between the orthopaedic and non‐orthopaedic subgroups in incidence of major bleeding (test for subgroup difference P = 0.82; Analysis 2.7).

Fatal bleeding during the intervention period was not reported.

No publication bias was indicated by investigating comparisons with funnel plots.

IPC plus pharmacological prophylaxis versus IPC plus aspirin

Three studies evaluated the role of combined IPC plus pharmacological prophylaxis versus IPC plus aspirin on the incidence of symptomatic PE (Westrich 2005; Westrich 2006; Woolson 1991). The studies showed a similar frequency of PE in the IPC plus pharmacological prophylaxis treatment groups (0/337, 0%) compared to the IPC plus aspirin control (2/268, 0.75%; OR 0.33, 95% CI 0.03 to 3.19; 3 studies, 605 participants; P = 0.34; Analysis 3.1). Results did not demonstrate heterogeneity (I2 = 0%).

Fatal PE was not reported.

The same studies investigated the role of combined modalities compared to IPC plus aspirin on the incidence of DVT. The studies showed a similar frequency in DVT in the IPC plus pharmacological prophylaxis treatment groups (30/337, 8.9%) compared to the IPC plus aspirin control (32/268, 11.9%; OR 0.83, 95% CI 0.48 to 1.42; 3 studies, 605 participants; P = 0.49; Analysis 3.2). Results did not demonstrate heterogeneity (I2 = 0%).

One study reported on the occurrence of symptomatic DVT (Westrich 2005), but because of the lack of events, we could not calculate a risk estimate (Analysis 3.3).

The three studies in this comparison all included orthopaedic participants only and therefore subgroup analyses between orthopaedic and non‐orthopaedic groups were not possible (Westrich 2005; Westrich 2006; Woolson 1991). No foot IPC was used in this comparison and therefore subgroup analysis was not possible. 

Three studies evaluated the role of combined IPC plus pharmacological prophylaxis versus IPC plus aspirin on the incidence of bleeding (Westrich 2005; Westrich 2006; Woolson 1991). The studies showed a similar frequency in bleeding in the IPC plus pharmacological prophylaxis treatment groups (4/341, 1.2%) compared to the IPC plus aspirin control (2/275, 0.7%) (OR 1.23, 95% CI 0.27 to 5.53; 3 studies, 616 participants; P = 0.79; Analysis 3.4). Results did not demonstrate heterogeneity (I2 = 0%).

These studies also showed a similar frequency in major bleeding in the IPC plus pharmacological prophylaxis treatment groups (2/341, 0.6%) compared to the IPC plus aspirin control (2/275, 0.7%; OR 0.80, 95% CI 0.15 to 4.17; 3 studies, 616 participants; Analysis 3.5). Results did not demonstrate heterogeneity (I2 = 0%).

Fatal bleeding during the intervention period was not reported.

Sensitivity analysis

We planned to perform sensitivity analysis of the primary outcomes by excluding studies at a high risk of bias, by excluding CCTs, and by excluding studies when substantial heterogeneity was present, in order to test the robustness of the evidence.

Exclusion of studies with high risk of bias

Assessing the included studies for risk of bias showed a high number of studies at high risk for performance bias and at unclear risk for selection and detection bias. See Figure 2 and Figure 3. We were unable to exclude studies at overall high risk because this included almost all studies, and did not leave enough studies within comparisons for meta‐analyses. Therefore, we did not perform sensitivity analysis for studies at high risk of bias.

Exclusion of controlled clinical studies
IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only

Exclusion of CCTs from the analysis did not change the overall effect on either PE or DVT. Thirteen RCTs evaluated the role of combined modalities on the incidence of mostly symptomatic PE (Cahan 2000; Dickinson 1998; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Nakagawa 2020; Obitsu 2020; Patel 2020; Stannard 1996; Turpie 2007; Woolson 1991; Yokote 2011). The incidence of PE in the combined group was 0.58% (12/2083), compared to 1.2% (25/2076) in the IPC control group, favouring combined modalities (OR 0.54, 95% CI 0.28 to 1.07; 13 RCTs, 4159 participants; P = 0.08; Analysis 4.1). Results demonstrated no substantial heterogeneity (I2 = 20%).

The same RCTs evaluated the role of combined modalities on the incidence of DVT (Cahan 2000; Dickinson 1998; Dong 2018; Hata 2019; Jung 2018; Kamachi 2020; Nakagawa 2020; Obitsu 2020; Patel 2020; Stannard 1996; Turpie 2007; Woolson 1991; Yokote 2011). These RCTs showed a reduction in DVT in favour of the combined treatment group; 2.21% (46/2083) in the combined treatment group compared to 3.81% (79/2076) in the IPC control group (OR 0.53, 95% CI 0.36 to 0.77; 13 RCTs, 4159 participants; P = 0.0009; Analysis 4.2). No heterogeneity was present (I2 = 0%).

Results on the secondary outcome of symptomatic DVT is shown in Analysis 4.3, with no clear effect of combined modalities. We could not estimate a difference between the effects of foot IPC and other IPC (Analysis 4.4). The risk of bleeding (Analysis 4.5), and risk of major bleeding (Analysis 4.6) was increased with combined modalities.

IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only

Exclusion of CCTs from the analysis did not change the overall effect on either PE or DVT. Eleven RCTs evaluated the effect of combined modalities on the incidence of mostly symptomatic PE (Arabi 2019; Cahan 2000; Dickinson 1998; Edwards 2008; Liu 2017a; Liu 2017b; Ramos 1996; Sakai 2016; Silbersack 2004; Stannard 1996; Windisch 2011). These RCTs showed a reduction in the incidence of PE in favour of the combined treatment group; 1.02% (30/2951) in the combined treatment group compared to 2.14% (60/2807) in the pharmacological prophylaxis control group (OR 0.45, 95% CI 0.29 to 0.70; 11 RCTs, 5758 participants; P = 0.0004; Analysis 5.1). Results did not demonstrate heterogeneity (I2 = 0%).

Thirteen RCTs investigated the role of combined modalities on the incidence of DVT (Arabi 2019; Cahan 2000; Dickinson 1998; Edwards 2008; Eisele 2007; Liu 2017a; Liu 2017b; Sakai 2016; Silbersack 2004; Siragusa 1994; Stannard 1996; Windisch 2011; Zhou 2020). These RCTs showed a reduction in incidence of DVT in favour of the combined treatment group; 6.17% (159/2578) in the combined treatment group compared to 9.21% (239/2594) in the pharmacological prophylaxis control group (OR 0.44, 95% CI 0.22 to 0.87; 13 RCTs, 5172 participants; P = 0.02; Analysis 5.2). Substantial heterogeneity was present (I2 = 80%) so we used a random‐effects model. 

Results on the secondary outcome of symptomatic DVT are shown in Analysis 5.3, with no clear effect of combined modalities. There was no clear difference between the effects of foot IPC and other IPC (Analysis 5.4). The risk of bleeding (Analysis 5.5) and risk of major bleeding (Analysis 5.6) was similar between the two groups.

IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only

Exclusion of CCTs from the analysis did not change the overall effect on either PE or DVT. Two RCTs evaluated the role of these combined modalities on the incidence of symptomatic PE (Westrich 2006; Woolson 1991). Pooling data from these RCTs showed no clear difference in PE; 0% (0/204) in the IPC plus pharmacological prophylaxis treatment group compared to 1.00% (2/201) in the IPC plus aspirin control group (OR 0.33, 95% CI 0.03 to 3.19; 2 RCTs, 405 participants; P = 0.34; Analysis 6.1). Results did not demonstrate heterogeneity (I2 = 0%).

The same RCTs investigated the role of combined modalities on the incidence of DVT. These showed no clear difference in DVT; 12.25% (25/204) in the IPC plus pharmacological prophylaxis treatment group compared to 14.92% (30/201) in the IPC plus aspirin control group (OR 0.79, 95% CI 0.44 to 1.39; 2 RCTs, 405 participants; P = 0.41; Analysis 6.2). Results did not demonstrate heterogeneity (I2 = 0%).

Exclusion of studies with medical patients

Exclusion of one study (Arabi 2019), with a majority of medical patients, did not alter the significance of findings of Analysis 2.1 and Analysis 2.2.

Discussion

Summary of main results

Our review showed that combined modalities may be more effective than single modalities (anticoagulation or compression alone) in reducing the incidence of PE and incidence of DVT. The use of IPC plus pharmacological prophylaxis reduced the incidence of PE and DVT compared to IPC alone (both low‐certainty evidence). The use of IPC plus pharmacological prophylaxis also reduced incidence of PE (low‐certainty evidence) and DVT (high‐certainty evidence) when compared to pharmacological prophylaxis alone. Regarding those studies that investigated the combination of compression plus anticoagulant with compression plus aspirin, our review showed no clear difference between combined and single modalities in the incidence of PE and DVT; this was likely caused by the low number of events and can be attributed to a type II error, that is, an incorrect retention of the null hypothesis, since the results are in favour of IPC combined with anticoagulation. The addition of an anticoagulant to IPC, however, increased the risk of bleeding compared to IPC alone (very low‐certainty evidence), a side effect not observed for IPC when added to anticoagulation (very low‐certainty evidence), as indeed expected for a physical method of thromboprophylaxis. These findings highlight the need to tailor the use of additional pharmacological thromboprophylaxis in patients at low risk for bleeding or those for whom bleeding does not have catastrophic consequences. This issue deserves further study since the criteria for major bleeding were not uniform across the studies, with some of them reporting on blood loss during the procedures, or through drains, or providing rates for postoperative bleeding. We judged the certainty of the evidence for the research that supports these conclusions to be mostly of low or very low certainty as a result of bias, indirectness, and imprecision.

The mechanism responsible for the improved effectiveness of combined modalities may be attributed to the fact that VTE is a multifactorial process. Virchow in 1856 suggested that venous stasis, coagulopathy, and endothelial injury are all causes of VTE (Virchow 1856). By treating the different causes of VTE it is expected that efficacy of DVT prevention would be improved. Rosendaal more recently extended Virchow's theory by proposing a model of risk factors, which considered the importance of the additive role and interaction of multiple risk factors (multiple‐hit model; Rosendaal 1999). Based on the additive role of mechanical and pharmacological modalities, the results of this review suggest that venous stasis and hypercoagulopathy may be independent risk factors. IPC reduces venous stasis by producing active flow enhancement (Kakkos 2005a), and also increases tissue factor pathway inhibitor (TFPI) plasma levels (Chouhan 1999). Unfractionated and LMWH inhibit factor X. These totally different mechanisms of action are most likely responsible for the synergy between these two modality types.

Subgroup analysis did not detect a difference in effectiveness between orthopaedic and non‐orthopaedic participants regarding DVT and PE prevention. No differences were detected in subgroup analysis for foot pumps compared to non‐foot pumps, likely a result of the small number of participants.

Sensitivity analyses restricting analysis to RCTs only did not significantly alter the overall results.

Pulmonary embolism risk‐reduction rates were mostly consistent across the studies with no heterogeneity, perhaps because symptomatic PE that was mostly reported is a clinically significant complication. In contrast, some heterogeneity was noted in the results on DVT reduction with the adjuvant use of IPC. This might have been related to the fact that the methodological quality of the assessed studies was low, and risk of bias was usually high. An alternative explanation is that the heterogeneity of the included participants who underwent various surgical procedures resulted in a variable risk of DVT. Finally, the variety of IPC devices may also account for the difference observed.

The results of our review are in agreement with the recommendations of the venous thromboembolism prevention guidelines that certain high‐risk surgical and trauma patients should receive multimodal prophylaxis (ASH 2019; Gould 2012; NICE 2009; NICE 2018; Nicolaides 2013).

Overall completeness and applicability of evidence

The studies included in this review were carried out in a wide range of patient groups undergoing orthopaedic but also urological, cardiothoracic, general surgical, neurosurgical and gynaecological procedures, and trauma patients. Most of the participants had a high‐risk procedure or condition and, therefore, the results of this review are not necessarily applicable to different patient groups, where a much lower risk may reduce the absolute risk reduction with combined modalities.

In an effort to investigate the applicability of combined modalities in orthopaedic and non‐orthopaedic participants, we performed subgroup analysis. This confirmed the efficacy of adding compression to anticoagulation in orthopaedic patients (and also non‐orthopaedic patients regarding PE), and the efficacy of adding anticoagulation to compression in non‐orthopaedic patients; these results indicate a need for further research in particular patient populations. Since studies on combined modalities are mostly performed in participants at high risk for VTE, the absolute benefit that would be observed is expected to be much lower in moderate‐risk patients, calling for cost‐effectiveness calculations and studies.

Additionally, it should be noted that the various IPC types may not have the same effectiveness and should not be used interchangeably, for example, foot pumps versus calf or calf and thigh leggings.

A potential confounding factor in the present review is the concurrent use of elastic stockings, very often used together with IPC.

Also, it should be mentioned that medical (non‐surgical) participants were only a small fraction of patients included in the present review (Arabi 2019), where IPC was found ineffective in reducing further PE and DVT.

Reporting of bleeding outcomes (major and minor bleeding) was not uniform across the studies, with some studies reporting on blood loss during the procedures or through the drains or providing rates for postoperative bleeding. The definitions used were also not uniform. This issue deserves further study.

Quality of the evidence

This review included some 14,931 participants who were studied in 34 studies (26 RCTs). This provided a body of evidence to investigate our hypothesis that combined modalities are more effective than their single counterparts. However, risk for performance bias was high in most studies, and risk for selection and detection bias was mostly unclear or high. Nevertheless, the results of the present meta‐analysis update are generally consistent with a low amount of heterogeneity in almost all comparisons.

Using GRADE assessment, the certainty of evidence for DVT and PE prevention with combined modalities varied from high to low. See summary of findings Table 1 and summary of findings Table 2. Risk of bias and imprecision accounted for downgrading, with the exception of the outcome of DVT, where the risk of bias was counteracted by a large treatment effect, in combination with a lack of imprecision.

The certainty of the evidence for bleeding and major bleeding for the comparison combined modalities versus IPC is very low. In addition to risk of bias and imprecision, we downgraded the certainty of the evidence for indirectness as definitions of bleeding and the reporting of bleeding outcomes was not uniform across studies. The certainty of the evidence for bleeding and major bleeding for the comparison combined modalities versus pharmacological prophylaxis was also very low. We downgraded the certainty of the evidence for risk of bias, indirectness (because the definition of bleeding and reporting of bleeding outcomes was not uniform across studies) and imprecision (due to the small number of participants and few events, and also wide confidence intervals).

Potential biases in the review process

The review authors have made an enormous effort to identify potential studies for inclusion in the present review. Publication bias still could have limited the validity of our results.

This review set out to assess RCTs and CCTs. Many of the CCTs are old and the reporting of the study methodology is often poor. In addition, patient care and standard practice has changed considerably over time. When assessing the incidence of DVT and PE in RCTs only, the overall direction and size of the effects were not affected. This was likely caused by the fact that many CCTs did not contribute to the analysis due to small number of reported events. However, when sufficient RCTs become available to perform meaningful analyses of the planned subgroups we will consider including RCTs only.

The review assessed symptomatic or asymptomatic DVT and symptomatic PE as outcomes. In future updates, data permitting, we will add proximal DVT and clinically important VTE (proximal DVT and symptomatic PE) as additional important outcomes.

In order to be as inclusive as possible and because not all studies reported on the type of IPC devices used, we included all IPC devices. This resulted in including some devices that may no longer be used in some parts of the world.

We performed no formal assessment of side effects of IPC. However, from the included studies we note these were rarely encountered and recorded by the studies. We will look into this in more detail in future updates.

Agreements and disagreements with other studies or reviews

The results presented here agree with previous systematic meta‐analyses on this topic, which showed that combined modalities are better than single prophylactic modalities (Ho 2013; Kakkos 2012; Sobieraj 2013; Zareba 2014). The studies included in these reviews were mostly restricted to a particular patient category or were limited by the fact that they used IPC interchangeably with elastic stockings, which is a limitation when interpretation of the results is attempted.

Study flow diagram

Figures and Tables -
Figure 1

Study flow diagram

Navigate to figure in ReviewOpen in new tab

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

Figures and Tables -
Figure 2

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

Navigate to figure in ReviewOpen in new tab

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

Figures and Tables -
Figure 3

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

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 1: Incidence of PE ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 1.1

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 1: Incidence of PE ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 2: Incidence of PE ‐ foot IPC or other IPC

Figures and Tables -
Analysis 1.2

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 2: Incidence of PE ‐ foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 3: Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 1.3

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 3: Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 4: Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 1.4

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 4: Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 5: Incidence of DVT ‐ by foot IPC or other IPC

Figures and Tables -
Analysis 1.5

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 5: Incidence of DVT ‐ by foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 6: Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 1.6

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 6: Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 7: Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 1.7

Comparison 1: IPC plus pharmacological prophylaxis versus IPC alone, Outcome 7: Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 1: Incidence of PE ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 2.1

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 1: Incidence of PE ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 2: Incidence of PE ‐ foot IPC or other IPC

Figures and Tables -
Analysis 2.2

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 2: Incidence of PE ‐ foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 3: Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 2.3

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 3: Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 4: Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 2.4

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 4: Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 5: Incidence of DVT ‐ foot IPC or other IPC

Figures and Tables -
Analysis 2.5

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 5: Incidence of DVT ‐ foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 6: Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 2.6

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 6: Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 7: Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients

Figures and Tables -
Analysis 2.7

Comparison 2: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone, Outcome 7: Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients

Navigate to figure in ReviewOpen in new tab

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 1: Incidence of PE

Figures and Tables -
Analysis 3.1

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 1: Incidence of PE

Navigate to figure in ReviewOpen in new tab

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 2: Incidence of DVT

Figures and Tables -
Analysis 3.2

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 2: Incidence of DVT

Navigate to figure in ReviewOpen in new tab

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 3: Incidence of symptomatic DVT

Figures and Tables -
Analysis 3.3

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 3: Incidence of symptomatic DVT

Navigate to figure in ReviewOpen in new tab

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 4: Incidence of bleeding

Figures and Tables -
Analysis 3.4

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 4: Incidence of bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 5: Incidence of major bleeding

Figures and Tables -
Analysis 3.5

Comparison 3: IPC plus pharmacological prophylaxis versus IPC plus aspirin, Outcome 5: Incidence of major bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 1: Incidence of PE

Figures and Tables -
Analysis 4.1

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 1: Incidence of PE

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 2: Incidence of DVT

Figures and Tables -
Analysis 4.2

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 2: Incidence of DVT

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 3: Incidence of symptomatic DVT

Figures and Tables -
Analysis 4.3

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 3: Incidence of symptomatic DVT

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 4: Incidence of DVT by foot IPC or other IPC

Figures and Tables -
Analysis 4.4

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 4: Incidence of DVT by foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 5: Incidence of bleeding

Figures and Tables -
Analysis 4.5

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 5: Incidence of bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 6: Incidence of major bleeding

Figures and Tables -
Analysis 4.6

Comparison 4: IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only, Outcome 6: Incidence of major bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 1: Incidence of PE

Figures and Tables -
Analysis 5.1

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 1: Incidence of PE

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 2: Incidence of DVT

Figures and Tables -
Analysis 5.2

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 2: Incidence of DVT

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 3: Incidence of symptomatic DVT

Figures and Tables -
Analysis 5.3

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 3: Incidence of symptomatic DVT

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 4: Incidence of DVT by foot IPC or other IPC

Figures and Tables -
Analysis 5.4

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 4: Incidence of DVT by foot IPC or other IPC

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 5: Incidence of bleeding

Figures and Tables -
Analysis 5.5

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 5: Incidence of bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 6: Incidence of major bleeding

Figures and Tables -
Analysis 5.6

Comparison 5: IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only, Outcome 6: Incidence of major bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 6: IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only, Outcome 1: Incidence of PE

Figures and Tables -
Analysis 6.1

Comparison 6: IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only, Outcome 1: Incidence of PE

Navigate to figure in ReviewOpen in new tab

Comparison 6: IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only, Outcome 2: Incidence of DVT

Figures and Tables -
Analysis 6.2

Comparison 6: IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only, Outcome 2: Incidence of DVT

Navigate to figure in ReviewOpen in new tab
Summary of findings 1. IPC plus pharmacological prophylaxis versus IPC alone

Does combined intermittent pneumatic compression (IPC) plus pharmacological prophylaxis increase prevention of venous thromboembolism compared with IPC alone?

Patient or population: people undergoing surgery or at risk of developing VTE due to surgery, trauma or ICU stay

Settings: hospital 

Intervention: combined modalities ‐  IPC plus pharmacological prophylaxis

Comparison: IPC alone

Outcomes

Anticipated absolute effects * (95% CI)

Relative effect
(95% CI)

No of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with IPC alone

Risk with combined modalities

Incidence of PEa

(early postoperative period)

16 per 1000

7 per 1000 (4 to 12)

OR 0.51 (0.29 to 0.91)

5462 (19)

⊕⊕⊝⊝
Lowb

 

Incidence of DVTc

(early postoperative period)

38 per 1000

20 per 1000 (14 to 28)

OR 0.51 (0.36 to 0.72)

5394 (18)

⊕⊕⊝⊝
Lowb

 

Incidence of bleedingd

(early postoperative period)

10 per 1000

55 per 1000 (36 to 83)

OR 6.02 (3.88 to 9.35)

4634 (13)

⊕⊝⊝⊝
Very lowe

 

Incidence of major bleedingf

(early postoperative period)

3 per 1000

19 per 1000 (10 to 39)

OR 5.77 (2.81 to 11.83)

4133 (12)

⊕⊝⊝⊝

Very lowe

 

* 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;ICU: intensive care unit; IPC: intermittent pneumatic compression; OR: odds ratio; PE: pulmonary embolism; 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. 

aPulmonary embolism assessed by pulmonary angiography or scintigraphy, computed tomography (CT), angiography, or autopsy.
bDowngraded by two levels due to risk of bias concerns (high in one or more domains regarding all but one study) and due to imprecision as a result of a small number of overall events.
cDeep vein thrombosis assessed by ascending venography, I‐125 fibrinogen uptake test, and ultrasound scanning.
d Any type of bleeding as described by the study authors.
eDowngraded by three levels due to risk of bias concerns (high in one or more domains regarding all but one study), due to imprecision as a result of a small number of events overall, and indirectness because bleeding definitions were not uniform across the studies.
fMajor bleeding as defined by the study authors, but usually located at the surgical site or in a critical organ or site, requiring intervention or transfusion of at least units of blood, or leading to death.

Figures and Tables -
Summary of findings 1. IPC plus pharmacological prophylaxis versus IPC alone
Navigate to table in Review
Summary of findings 2. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone

Does combined intermittent pneumatic compression (IPC) plus pharmacological prophylaxis increase prevention of venous thromboembolism compared with pharmacological prophylaxis alone?

Patient or population: people undergoing surgery or at risk of developing VTE because of surgery, trauma or ICU stay

Settings: hospital 

Intervention: combined modalities ‐ IPC plus pharmacological prophylaxis

Comparison: pharmacological prophylaxis alone

Outcomes

Anticipated absolute effects * (95% CI)

Relative effect
(95% CI)

No of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with pharmacological prophylaxis alone

 

Risk with combined modalities

Incidence of PEa

(early postoperative period)

18 per 1000

9 per 1000 (6 to 13)

OR 0.46 (0.30 to 0.71)

6737 (15)

⊕⊕⊝⊝
Lowb

 

Incidence of DVTc

(early postoperative period)

93 per 1000

37 per 1000 (21 to 67)

OR 0.38 (0.21 to 0.70)

6151 (17)

⊕⊕⊕⊕
Highd

 

Incidence of bleedinge

(early postoperative period)

74 per 1000

65 per 1000 (43 to 98)

OR 0.87 (0.56 to 1.35

1314 (6)

⊕⊝⊝⊝

Very lowf

 

Incidence of major bleedingg

(early postoperative period)

11 per 1000

13 per 1000 (4 to 44)

OR 1.21 (0.35 to 4.18)

908 (5)

⊕⊝⊝⊝

Very lowf

 

* 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;ICU: intensive care unit; IPC: intermittent pneumatic compression; OR: odds ratio; PE: pulmonary embolism; 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.

aPulmonary embolism assessed by pulmonary angiography or scintigraphy, computed tomography (CT), angiography, single photon emission CT (SPECT) or autopsy.
bDowngraded by two levels due to risk of bias (which was high in one or more domains regarding all studies) and due to imprecision as a result of a small number of events overall.
cDeep vein thrombosis assessed predominantly by ascending venography, I‐125 fibrinogen uptake test, and ultrasound scanning.
dDowngraded by one level due to risk of bias (which was high in one or more domains regarding all but one study) and upgraded by one level because of a large magnitude of the effect.
eAny type of bleeding as described by the study authors.
fDowngraded by three levels due to risk of bias (which was high in one or more domains regarding all studies), due to imprecision as a result of a small number of events overall and a wide confidence interval, and indirectness because bleeding definitions were not uniform across the studies.
gMajor bleeding as defined by the study authors, but usually located at the surgical site or in a critical organ or site, requiring intervention or transfusion of at least units of blood, or leading to death.

Figures and Tables -
Summary of findings 2. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone
Navigate to table in Review
Comparison 1. IPC plus pharmacological prophylaxis versus IPC alone

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1.1 Incidence of PE ‐ orthopaedic and non‐orthopaedic patients Show forest plot

19

5462

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.29, 0.91]

1.1.1 Orthopaedic patients

3

445

Odds Ratio (M‐H, Fixed, 95% CI)

Not estimable

1.1.2 Non‐orthopaedic patients

16

5017

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.29, 0.91]

1.2 Incidence of PE ‐ foot IPC or other IPC Show forest plot

19

5462

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.29, 0.91]

1.2.1 Foot IPC

1

50

Odds Ratio (M‐H, Fixed, 95% CI)

Not estimable

1.2.2 Other IPC

18

5412

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.29, 0.91]

1.3 Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients Show forest plot

18

5394

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.36, 0.72]

1.3.1 Orthopaedic patients

3

445

Odds Ratio (M‐H, Fixed, 95% CI)

0.80 [0.38, 1.69]

1.3.2 Non‐orthopaedic patients

15

4949

Odds Ratio (M‐H, Fixed, 95% CI)

0.46 [0.31, 0.68]

1.4 Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients Show forest plot

10

4089

Odds Ratio (M‐H, Fixed, 95% CI)

0.48 [0.21, 1.10]

1.4.1 Orthopaedic patients

1

250

Odds Ratio (M‐H, Fixed, 95% CI)

1.50 [0.06, 37.33]

1.4.2 Non‐orthopaedic patients

9

3839

Odds Ratio (M‐H, Fixed, 95% CI)

0.44 [0.19, 1.04]

1.5 Incidence of DVT ‐ by foot IPC or other IPC Show forest plot

18

5395

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.36, 0.72]

1.5.1 Foot IPC

1

50

Odds Ratio (M‐H, Fixed, 95% CI)

Not estimable

1.5.2 Other IPC

17

5345

Odds Ratio (M‐H, Fixed, 95% CI)

0.51 [0.36, 0.72]

1.6 Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients Show forest plot

13

4634

Odds Ratio (M‐H, Fixed, 95% CI)

6.02 [3.88, 9.35]

1.6.1 Orthopaedic patients

2

400

Odds Ratio (M‐H, Fixed, 95% CI)

2.79 [0.77, 10.18]

1.6.2 Non‐orthopaedic patients

11

4234

Odds Ratio (M‐H, Fixed, 95% CI)

6.61 [4.14, 10.56]

1.7 Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients Show forest plot

12

4133

Odds Ratio (M‐H, Fixed, 95% CI)

5.77 [2.81, 11.83]

1.7.1 Orthopaedic patients

2

400

Odds Ratio (M‐H, Fixed, 95% CI)

3.35 [0.13, 83.62]

1.7.2 Non‐orthopaedic patients

10

3733

Odds Ratio (M‐H, Fixed, 95% CI)

5.91 [2.83, 12.36]
Figures and Tables -
Comparison 1. IPC plus pharmacological prophylaxis versus IPC alone
Navigate to table in Review
Comparison 2. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

2.1 Incidence of PE ‐ orthopaedic and non‐orthopaedic patients Show forest plot

15

6737

Odds Ratio (M‐H, Fixed, 95% CI)

0.46 [0.30, 0.71]

2.1.1 Orthopaedic patients

8

1202

Odds Ratio (M‐H, Fixed, 95% CI)

0.58 [0.08, 4.49]

2.1.2 Non‐orthopaedic patients

7

5535

Odds Ratio (M‐H, Fixed, 95% CI)

0.46 [0.29, 0.71]

2.2 Incidence of PE ‐ foot IPC or other IPC Show forest plot

15

6737

Odds Ratio (M‐H, Fixed, 95% CI)

0.46 [0.30, 0.71]

2.2.1 Foot IPC

4

324

Odds Ratio (M‐H, Fixed, 95% CI)

0.32 [0.01, 8.25]

2.2.2 Other IPC

11

6413

Odds Ratio (M‐H, Fixed, 95% CI)

0.47 [0.30, 0.72]

2.3 Incidence of DVT ‐ orthopaedic and non‐orthopaedic patients Show forest plot

17

6151

Odds Ratio (M‐H, Random, 95% CI)

0.38 [0.21, 0.70]

2.3.1 Orthopaedic patients

10

3075

Odds Ratio (M‐H, Random, 95% CI)

0.34 [0.18, 0.68]

2.3.2 Non‐orthopaedic patients

7

3076

Odds Ratio (M‐H, Random, 95% CI)

0.46 [0.13, 1.61]

2.4 Incidence of symptomatic DVT ‐ orthopaedic and non‐orthopaedic patients Show forest plot

7

3032

Odds Ratio (M‐H, Fixed, 95% CI)

0.83 [0.34, 2.01]

2.4.1 Orthopaedic patients

3

477

Odds Ratio (M‐H, Fixed, 95% CI)

2.09 [0.38, 11.50]

2.4.2 Non‐orthopaedic patients

4

2555

Odds Ratio (M‐H, Fixed, 95% CI)

0.55 [0.18, 1.66]

2.5 Incidence of DVT ‐ foot IPC or other IPC Show forest plot

16

4148

Odds Ratio (M‐H, Random, 95% CI)

0.34 [0.18, 0.63]

2.5.1 Foot IPC

4

324

Odds Ratio (M‐H, Random, 95% CI)

0.40 [0.05, 3.47]

2.5.2 Other IPC

12

3824

Odds Ratio (M‐H, Random, 95% CI)

0.31 [0.17, 0.54]

2.6 Incidence of bleeding ‐ orthopaedic and non‐orthopaedic patients Show forest plot

6

1314

Odds Ratio (M‐H, Fixed, 95% CI)

0.87 [0.56, 1.35]

2.6.1 Orthopaedic patients

3

550

Odds Ratio (M‐H, Fixed, 95% CI)

0.64 [0.20, 2.07]

2.6.2 Non‐orthopaedic patients

3

764

Odds Ratio (M‐H, Fixed, 95% CI)

0.91 [0.57, 1.47]

2.7 Incidence of major bleeding ‐ orthopaedic and non‐orthopaedic patients Show forest plot

5

908

Odds Ratio (M‐H, Fixed, 95% CI)

1.21 [0.35, 4.18]

2.7.1 Orthopaedic patients

3

551

Odds Ratio (M‐H, Fixed, 95% CI)

1.07 [0.21, 5.54]

2.7.2 Non‐orthopaedic patients

2

357

Odds Ratio (M‐H, Fixed, 95% CI)

1.42 [0.21, 9.49]
Figures and Tables -
Comparison 2. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone
Navigate to table in Review
Comparison 3. IPC plus pharmacological prophylaxis versus IPC plus aspirin

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

3.1 Incidence of PE Show forest plot

3

605

Odds Ratio (M‐H, Fixed, 95% CI)

0.33 [0.03, 3.19]

3.2 Incidence of DVT Show forest plot

3

605

Odds Ratio (M‐H, Fixed, 95% CI)

0.83 [0.48, 1.42]

3.3 Incidence of symptomatic DVT Show forest plot

1

Odds Ratio (M‐H, Fixed, 95% CI)

Totals not selected

3.4 Incidence of bleeding Show forest plot

3

616

Odds Ratio (M‐H, Fixed, 95% CI)

1.23 [0.27, 5.53]

3.5 Incidence of major bleeding Show forest plot

3

616

Odds Ratio (M‐H, Fixed, 95% CI)

0.80 [0.15, 4.17]
Figures and Tables -
Comparison 3. IPC plus pharmacological prophylaxis versus IPC plus aspirin
Navigate to table in Review
Comparison 4. IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

4.1 Incidence of PE Show forest plot

13

4159

Odds Ratio (M‐H, Fixed, 95% CI)

0.54 [0.28, 1.07]

4.2 Incidence of DVT Show forest plot

13

4159

Odds Ratio (M‐H, Fixed, 95% CI)

0.53 [0.36, 0.77]

4.3 Incidence of symptomatic DVT Show forest plot

7

3064

Odds Ratio (M‐H, Fixed, 95% CI)

1.21 [0.15, 9.63]

4.4 Incidence of DVT by foot IPC or other IPC Show forest plot

17

5085

Odds Ratio (M‐H, Fixed, 95% CI)

0.52 [0.36, 0.74]

4.4.1 Foot IPC

1

50

Odds Ratio (M‐H, Fixed, 95% CI)

Not estimable

4.4.2 Other IPC

16

5035

Odds Ratio (M‐H, Fixed, 95% CI)

0.52 [0.36, 0.74]

4.5 Incidence of bleeding Show forest plot

10

4120

Odds Ratio (M‐H, Fixed, 95% CI)

5.45 [3.38, 8.80]

4.6 Incidence of major bleeding Show forest plot

9

3619

Odds Ratio (M‐H, Fixed, 95% CI)

5.90 [2.82, 12.33]
Figures and Tables -
Comparison 4. IPC plus pharmacological prophylaxis versus IPC alone ‐ RCTs only
Navigate to table in Review
Comparison 5. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

5.1 Incidence of PE Show forest plot

11

5758

Odds Ratio (M‐H, Fixed, 95% CI)

0.45 [0.29, 0.70]

5.2 Incidence of DVT Show forest plot

13

5172

Odds Ratio (M‐H, Random, 95% CI)

0.44 [0.22, 0.87]

5.3 Incidence of symptomatic DVT Show forest plot

5

2312

Odds Ratio (M‐H, Fixed, 95% CI)

1.02 [0.29, 3.54]

5.4 Incidence of DVT by foot IPC or other IPC Show forest plot

15

5431

Odds Ratio (M‐H, Random, 95% CI)

0.42 [0.23, 0.80]

5.4.1 Foot IPC

4

324

Odds Ratio (M‐H, Random, 95% CI)

0.40 [0.05, 3.47]

5.4.2 Other IPC

11

5107

Odds Ratio (M‐H, Random, 95% CI)

0.40 [0.20, 0.81]

5.5 Incidence of bleeding Show forest plot

4

594

Odds Ratio (M‐H, Fixed, 95% CI)

0.80 [0.30, 2.14]

5.6 Incidence of major bleeding Show forest plot

4

595

Odds Ratio (M‐H, Fixed, 95% CI)

1.21 [0.35, 4.18]
Figures and Tables -
Comparison 5. IPC plus pharmacological prophylaxis versus pharmacological prophylaxis alone ‐ RCTs only
Navigate to table in Review
Comparison 6. IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

6.1 Incidence of PE Show forest plot

2

405

Odds Ratio (M‐H, Fixed, 95% CI)

0.33 [0.03, 3.19]

6.2 Incidence of DVT Show forest plot

2

405

Odds Ratio (M‐H, Fixed, 95% CI)

0.79 [0.44, 1.39]
Figures and Tables -
Comparison 6. IPC plus pharmacological prophylaxis versus IPC plus aspirin ‐ RCTs only
Navigate to table in Review