Skip to main content

Foundational Trials

PREVENT (pcd)

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Adjunctive Intermittent Pneumatic Compression for Venous Thromboprophylaxis
  • Acronym: PREVENT (Pneumatic CompREssion for PreVENting Venous Thromboembolism)
  • Year: 2019
  • Journal published in: New England Journal of Medicine
  • Citation: Arabi YM, Al-Hameed F, Burns KEA, et al. Adjunctive intermittent pneumatic compression for venous thromboprophylaxis. N Engl J Med. 2019;380(14):1305-1315.

Context & Rationale

  • Background
    • Critically ill patients are at substantial risk of venous thromboembolism (VTE) due to immobility, systemic inflammation, invasive devices, and frequent comorbidity.
    • Pharmacologic thromboprophylaxis (unfractionated heparin or low-molecular-weight heparin) reduces VTE, but does not eliminate risk, and practice varies regarding adding mechanical prophylaxis.
    • Intermittent pneumatic compression (IPC) is widely used when pharmacologic prophylaxis is contraindicated; however, the incremental benefit of routinely adding IPC to heparin-based prophylaxis in ICU patients remained uncertain.
    • IPC has non-trivial opportunity costs (staff time, interruptions to mobilisation/skin care), potential harms (skin injury, ischaemia), and device heterogeneity; therefore, a definitive pragmatic randomised trial was methodologically and clinically important.
  • Research Question/Hypothesis
    In critically ill adults receiving pharmacologic thromboprophylaxis with unfractionated or low-molecular-weight heparin, does adjunctive IPC reduce incident proximal lower-limb deep-vein thrombosis compared with pharmacologic prophylaxis alone?
  • Why This Matters
    • Determines whether routine “dual-modality” prophylaxis (IPC + heparin) is clinically justified in ICU patients who can receive anticoagulant prophylaxis.
    • Clarifies resource allocation and “low-value” care risks (large ICU device burden vs uncertain benefit).
    • Informs guideline development by providing adequately powered, international RCT evidence with systematic ultrasound surveillance and explicit safety ascertainment.

Design & Methods

  • Research Question: Among critically ill adults receiving pharmacologic thromboprophylaxis, does adjunctive IPC (vs no routine IPC) reduce incident proximal lower-limb DVT detected on twice-weekly ultrasound?
  • Study Type: Pragmatic, investigator-initiated, international, multicentre, parallel-group randomised controlled trial; open-label delivery; allocation concealed; stratified by site and by pharmacologic prophylaxis type (unfractionated vs low-molecular-weight heparin); ultrasound images interpreted by radiologists unaware of allocation.
  • Population:
    • Setting: Medical-surgical-trauma intensive care units across Saudi Arabia, Canada, Australia and India.
    • Inclusion: Adult per local definition (≥14/16/18 years); weight ≥45 kg; expected ICU stay ≥72 hours; eligible for pharmacologic thromboprophylaxis with unfractionated or low-molecular-weight heparin; enrolment within 48 hours of ICU admission.
    • Key exclusions: >24 hours of IPC in the current ICU admission; ICU stay >48 hours; pharmacologic prophylaxis other than unfractionated/low-molecular-weight heparin; inability/contraindication to apply IPC or obtain adequate lower-limb ultrasound (e.g., extensive dressings, burns, active infection); acute lower-limb ischaemia/severe peripheral arterial disease/compartment syndrome; major lower-limb fracture or limb amputation preventing bilateral IPC; therapeutic-dose heparin; pregnancy; limitation of life support or life expectancy ≤7 days; inferior vena cava filter; allergy to sleeve material.
  • Intervention:
    • IPC applied to both lower limbs for at least 18 hours/day, in addition to pharmacologic prophylaxis with unfractionated or low-molecular-weight heparin.
    • IPC devices were centre-selected (multiple commercial devices were used across sites); sleeves were routinely removed for skin inspection/care during the intervention period.
    • IPC was continued until ICU discharge, death, attainment of full mobility, trial day 28, or discontinuation for clinical reasons (e.g., suspected/confirmed thrombosis or limb complication).
  • Comparison:
    • Pharmacologic thromboprophylaxis with unfractionated or low-molecular-weight heparin alone (no routine IPC).
    • IPC could be used in the control group during interruptions of pharmacologic prophylaxis (reflecting pragmatic usual care during transient contraindications).
  • Blinding: Open-label for clinicians/patients (device-based intervention); radiologists interpreting ultrasound images were unaware of trial-group assignment.
  • Statistics: A total sample size of ~2000 was planned to detect a 3% absolute risk reduction in incident proximal DVT (from 7% to 4%) with 80% power at a two-sided 5% significance level; primary analysis was modified intention-to-treat (excluding patients withdrawing consent for intervention/data collection and those identified as ineligible after randomisation), with prespecified sensitivity analyses and a published statistical analysis plan. 12
  • Follow-Up Period: Twice-weekly lower-limb ultrasound surveillance from after trial day 3 until ICU discharge, death, full mobility, or day 28; clinical outcomes (including mortality) assessed to day 28 and day 90.

Key Results

This trial was not stopped early. Recruitment continued until the prespecified sample size target was achieved.

Outcome Pneumatic compression group Control group Effect p value / 95% CI Notes
Primary: incident proximal lower-limb DVT (ultrasound-detected; after trial day 3) 37/957 (3.9%) 41/985 (4.2%) RR 0.93 95% CI 0.60 to 1.44; P=0.74 Denominators reflect patients evaluable for the primary outcome (excluding prevalent DVT and missing ultrasound).
Incident proximal or distal lower-limb DVT 49/957 (5.1%) 55/985 (5.6%) RR 0.92 95% CI 0.63 to 1.33; P not reported Ultrasound-detected (incident definition aligned with the primary outcome timeframe).
Pulmonary embolism 8/991 (0.8%) 10/1012 (1.0%) RR 0.82 95% CI 0.32 to 2.06; P not reported Clinically detected and investigated (not routine screening).
Composite: lower-limb DVT, pulmonary embolism, or death from any cause at 28 days 231/990 (23.3%) 243/1011 (24.0%) RR 0.97 95% CI 0.83 to 1.14; P not reported Composite including mortality; denominators reflect available 28-day follow-up.
Death from any cause at 28 days 145/990 (14.6%) 167/1011 (16.5%) RR 0.89 95% CI 0.72 to 1.09; P not reported Two patients discharged before day 28 were lost to follow-up for 28- and 90-day mortality.
Death from any cause at 90 days 258/990 (26.1%) 270/1011 (26.7%) RR 0.98 95% CI 0.84 to 1.13; P not reported Longer-term mortality signal absent.
Safety: stage I lower-limb skin injury (non-blanchable erythema) 25/991 (2.5%) 22/1012 (2.2%) RR 1.16 95% CI 0.66 to 2.04; P not reported Staged using the National Pressure Ulcer Advisory Panel classification (highest stage during trial period).
Safety: limb ischaemia affecting toes 5/991 (0.5%) 10/1012 (1.0%) RR 0.51 95% CI 0.18 to 1.49; P not reported Clinically ascertained safety endpoint.
Safety: serious adverse events attributed to IPC (e.g., stage III/IV ulcer or IPC-related ischaemia) 0/991 (0.0%) 0/1012 (0.0%) Not estimable Not reported No serious IPC-related adverse events reported.
  • The primary endpoint was neutral: incident proximal DVT 3.9% with adjunct IPC vs 4.2% with pharmacologic prophylaxis alone (RR 0.93; 95% CI 0.60 to 1.44; P=0.74).
  • Clinically important endpoints (pulmonary embolism, 28-day and 90-day mortality) were also similar between groups, with wide confidence intervals for low-frequency events.
  • Protocol adherence and exposure separation were substantial (median IPC use 22 hours/day in the IPC arm vs 0 hours/day in controls), supporting interpretation that large benefits of routine adjunct IPC are unlikely in heparin-eligible ICU patients.

Internal Validity

  • Randomisation and allocation: Central randomisation with stratification by site and planned heparin type; allocation concealment maintained until assignment.
  • Drop out/exclusions: Modified intention-to-treat excluded patients who withdrew consent for intervention/data collection and those identified as ineligible after randomisation; primary-outcome evaluable denominators were 957/991 vs 985/1012 because prevalent DVTs and missing ultrasound assessments were excluded from the incident DVT analysis.
  • Performance/detection bias: Trial delivery was unblinded; however, ultrasound images were interpreted by radiologists unaware of allocation, reducing detection bias for the primary outcome.
  • Protocol adherence: IPC was delivered with high intensity: median 22.0 (IQR 18.0–23.2) hours/day in the IPC group vs 0 (IQR 0–0) hours/day in controls during the intervention period.
  • Separation of the variable of interest: Total IPC duration was 144 (IQR 72–260) hours in the IPC group vs 0 (IQR 0–0) hours in controls; daily compliance ≥18 hours occurred on 6422/7717 (83.2%) intervention days in the IPC arm.
  • Crossover/contamination: 103/1012 (10.2%) control patients received IPC at any time during the intervention period (permitted during interruptions of pharmacologic prophylaxis), but typical daily exposure remained 0 hours/day at the group level.
  • Baseline characteristics: Groups were broadly balanced (e.g., age ~57 years; APACHE II ~20; ~54–55% mechanically ventilated at enrolment), supporting comparability of prognosis at baseline.
  • Outcome assessment: The primary endpoint was objectively defined (ultrasound-detected proximal DVT) with a prespecified “incident” window (after calendar day 3) to reduce misclassification of prevalent thrombosis as incident events.
  • Statistical rigour: A prespecified protocol and statistical analysis plan were published before trial completion; multiple sensitivity analyses (including alternate incident-DVT cut-offs and handling of missing baseline ultrasound) were consistent with the primary neutral result. 12

Conclusion on Internal Validity: Strong overall: concealed randomisation, high adherence with substantial exposure separation, minimal missingness for key endpoints, and blinded radiology interpretation support a robust primary inference; the main limitations are intrinsic to the intervention (open-label delivery) and permitted IPC use in controls during pharmacologic interruptions.

External Validity

  • Population representativeness: The cohort reflects typical mixed medical-surgical-trauma ICU populations who are eligible for heparin-based prophylaxis and anticipated to remain in ICU ≥72 hours.
  • Important exclusions: Findings do not directly apply to patients in whom pharmacologic prophylaxis is contraindicated (the common scenario where IPC is used as the primary prophylaxis modality), nor to patients with limb conditions precluding IPC application/ultrasound assessment.
  • Intervention feasibility: Achieved IPC intensity (median 22 hours/day) may exceed adherence in many real-world ICUs; thus, if anything, the trial tests a “best-case” implementation scenario for adjunct IPC.
  • Practice setting: International conduct supports generalisability, but local thromboprophylaxis practices, IPC device availability, and ultrasound surveillance capability vary; settings without systematic ultrasound screening may prioritise symptomatic VTE and patient-centred endpoints differently.

Conclusion on External Validity: Moderate-to-strong for heparin-eligible ICU populations in resourced settings; limited for patients requiring mechanical prophylaxis alone or in environments where high-intensity IPC adherence is not feasible.

Strengths & Limitations

  • Strengths:
    • Large, pragmatic, multicentre international RCT addressing a common ICU practice question with high resource implications.
    • Systematic DVT surveillance (twice-weekly ultrasound) and blinded radiology interpretation for the primary endpoint.
    • Substantial exposure separation (high IPC dose in intervention arm; minimal group-level exposure in controls).
    • Protocol and statistical analysis plan published a priori, supporting analytic transparency.
  • Limitations:
    • Open-label delivery; ultrasound operators and bedside teams necessarily aware of allocation.
    • Primary endpoint was largely ultrasound-detected thrombosis (a surrogate for symptomatic VTE), and event rates were lower than anticipated, reducing statistical power for modest effects.
    • Permitted IPC use in controls during heparin interruptions introduces pragmatic contamination (potential dilution), although typical exposure remained low.
    • Does not answer whether IPC alone is effective when anticoagulant prophylaxis is contraindicated.

Interpretation & Why It Matters

  • Implication for routine ICU practice
    In ICU patients eligible for unfractionated/low-molecular-weight heparin prophylaxis, routine addition of adjunct IPC did not reduce incident proximal DVT, pulmonary embolism, or mortality, and therefore does not support default “dual-modality” prophylaxis as a universal strategy.
  • Resource allocation and implementation
    Given the device burden, nursing workload, and potential skin/ischaemic complications, PREVENT supports prioritising consistent pharmacologic prophylaxis and reserving IPC for periods when anticoagulant prophylaxis is interrupted or contraindicated.
  • Methodological lesson
    Modern ICU thromboprophylaxis trials may face lower-than-expected DVT event rates under contemporary care, increasing reliance on large pragmatic designs and careful endpoint selection (surrogate vs patient-centred outcomes) to maintain inferential precision.

Controversies & Subsequent Evidence

  • Neutral primary endpoint with lower-than-anticipated event rate: The main interpretive tension is whether the neutral result reflects true lack of incremental benefit or limited power for modest effects in a lower-risk-than-expected population; the confidence interval spans both clinically meaningful benefit and harm.
  • Surrogate vs patient-centred outcomes: The trial’s primary endpoint was ultrasound-detected proximal DVT (important, but not equivalent to symptomatic VTE or mortality), raising debate about what magnitude of reduction would justify routine adjunct IPC.
  • Open-label pragmatism and control-group IPC during heparin interruptions: The pragmatic allowance of IPC in controls during pharmacologic interruptions may dilute between-group differences, although achieved median exposure remained 0 hours/day in controls.
  • Editorial perspective: The accompanying editorial framed PREVENT as evidence against routine adjunct IPC in heparin-eligible ICU patients, while emphasising that mechanical prophylaxis remains relevant when pharmacologic prophylaxis is unsafe or withheld. 3
  • Subsequent subgroup-focused analysis: A post hoc analysis restricted to patients with heart failure reported incident proximal DVT 11/125 (8.8%) with IPC vs 6/132 (4.5%) in controls (RR 1.94; 95% CI 0.74 to 5.08), with no improvement in ventilation-free days (median 21 vs 25 days). 4
  • Guidelines: Contemporary critical care guidance subsequent to PREVENT (e.g., Surviving Sepsis Campaign 2021) reflects a trend away from routine adjunct mechanical prophylaxis when pharmacologic prophylaxis is feasible, while maintaining mechanical prophylaxis as an option when bleeding risk precludes anticoagulants. 5

Summary

  • In heparin-eligible ICU patients, adjunct IPC did not reduce incident proximal DVT versus pharmacologic prophylaxis alone (3.9% vs 4.2%; RR 0.93; 95% CI 0.60 to 1.44; P=0.74).
  • No signal of benefit was seen for pulmonary embolism or mortality (28-day mortality 14.6% vs 16.5%; 90-day mortality 26.1% vs 26.7%).
  • Adherence and exposure separation were high (median IPC 22 hours/day in the IPC arm vs 0 hours/day in controls), strengthening inference against a large incremental effect.
  • IPC-related harms (skin injury/ischaemia) were uncommon and similar between groups; no serious IPC-related adverse events were reported.
  • Findings apply to adjunct IPC in addition to heparin prophylaxis, not to IPC as a substitute when anticoagulants are contraindicated.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • This PREVENT trial evaluates adjunctive intermittent pneumatic compression for VTE prophylaxis; it is distinct from the similarly named “PReVENT” trial evaluating ventilator tidal volume strategies.
  • Confidence intervals for secondary outcomes were not adjusted for multiplicity in the primary publication; interpret secondary endpoints accordingly.

Overall Takeaway

PREVENT provided definitive, pragmatic RCT evidence that routine adjunct IPC added to heparin-based thromboprophylaxis does not meaningfully reduce incident proximal DVT in heparin-eligible ICU patients. With substantial protocol adherence and clear exposure separation, the trial shifts the default approach away from universal “dual-modality” prophylaxis towards selective IPC use when pharmacologic prophylaxis is interrupted or contraindicated.

Overall Summary

  • Adjunct IPC + heparin did not reduce incident proximal DVT versus heparin alone (3.9% vs 4.2%).
  • No improvement in pulmonary embolism or mortality was demonstrated.
  • High-dose IPC delivery in the intervention arm strengthens inference against a large incremental benefit in heparin-eligible ICU patients.

Bibliography