Skip to main content

Trial Summaries & Critiques

RATE

Image: Shutterstock / ME_Image

ECMO machine beside a patient in the ICU

Publication

  • Title: Standard-dose unfractionated heparin versus low-dose unfractionated heparin and low-molecular-weight heparin in extracorporeal life support (RATE): an open-label, randomised, non-inferiority trial
  • Acronym: RATE
  • Year: 2026
  • Journal published in: Lancet
  • Citation: van Minnen O, Oude Lansink-Hartgring A, van Amstel RBE, van den Bogaard B, Bunge JJH, Delnoij TSR, et al. Standard-dose unfractionated heparin versus low-dose unfractionated heparin and low-molecular-weight heparin in extracorporeal life support (RATE): an open-label, randomised, non-inferiority trial. Lancet. 2026;published online July 7.

Context & Rationale

  • Background:
    • ECMO exposes blood to artificial circuit surfaces, activates haemostasis, and creates simultaneous thrombotic and bleeding risk.12
    • Traditional practice has used systemic UFH with aPTT targets around 2.0-2.5 times baseline, but this target was extrapolated from non-ECMO anticoagulation and was not validated in adequately powered ECMO trials.34
    • Before RATE, systematic reviews of VV and VA ECMO anticoagulation practice described substantial heterogeneity in anticoagulation targets, monitoring, bleeding definitions, thrombotic events, and outcome reporting.56
    • Modern centrifugal pumps, coated circuits, and heparin-coated cannulae have reduced clinically manifest thrombotic events, whereas bleeding and transfusion exposure remain common and prognostically important.7
  • Research Question/Hypothesis:
    • RATE asked whether low-dose UFH or therapeutic LMWH would be non-inferior to standard-dose UFH for the net clinical outcome of severe bleeding during ECMO, severe thromboembolic complications during ECMO, or all-cause mortality at 6 months.
  • Why This Matters:
    • Anticoagulation intensity is a daily ECMO decision with major consequences: more heparin might prevent rare catastrophic thrombosis, but might also amplify common and lethal bleeding.
    • A lower-intensity strategy that preserved thrombotic safety could reduce bleeding-related harm, transfusion exposure, monitoring burden, and cost.
    • The question is relevant across both respiratory and circulatory ECMO because the trial included both VV and VA ECMO.

Design & Methods

  • Research Question:
    • In adults receiving ECMO without a separate vital indication for full-dose anticoagulation, are low-dose UFH and therapeutic LMWH non-inferior to standard-dose UFH for severe bleeding, severe thromboembolism, or death at 6 months?
  • Study Type:
    • Multicentre, open-label, three-arm, randomised, non-inferiority trial.
    • Conducted in seven Dutch closed-format mixed ICUs.
    • Registered at ClinicalTrials.gov and the Dutch trial register.
  • Population:
    • Adults aged 18 years or older receiving VV or VA ECMO in ICU for severe respiratory or circulatory failure considered reversible by the treating team, or as a bridge to transplantation.
    • Key exclusions were procedural ECMO, absence of deferred consent, a compelling indication for full-dose anticoagulation such as mechanical mitral valve, proven pulmonary embolism or intracardiac thrombus, and heparin-induced thrombocytopenia.
    • Between Oct 22, 2020, and Sept 12, 2024, 330 patients were enrolled; 320 had deferred consent and 6-month outcome data and were analysed.
  • Intervention:
    • Low-dose UFH: continuous intravenous UFH targeting aPTT 1.5-2.0 times centre-specific baseline, approximately 45-60 s.
    • Therapeutic LMWH: therapeutic subcutaneous LMWH twice daily, adjusted for bodyweight and renal function.
    • Anticoagulation was started after cannulation and could be deferred for up to 24 h if active bleeding was present.
  • Comparison:
    • Standard-dose UFH: continuous intravenous UFH targeting aPTT 2.0-2.5 times centre-specific baseline, approximately 60-75 s.
    • A UFH loading dose of 5000 IU was commonly given at percutaneous cannulation according to local practice.
    • UFH dose adjustment followed local protocols; aPTT was measured four to six times daily in UFH groups.
  • Blinding:
    • The trial was open-label because dosing route, monitoring, and anticoagulation targets differed by group.
    • Follow-up personnel were masked to treatment allocation.
  • Statistics:
    • The trial assumed a 70% primary-outcome event rate with standard-dose UFH and 60% with each intervention; with a 7.5 percentage point non-inferiority margin, 91 patients per group gave approximately 80% power at one-sided alpha 2.5%, and the target sample size was increased to 330 to allow for loss to follow-up.
    • The primary analysis followed the intention-to-treat principle in patients with deferred consent and 6-month outcome data.
    • Absolute risk differences with two-sided 95% CIs were the primary comparison.
    • RRs were estimated using robust Poisson regression adjusted for centre and ECMO mode.
    • Missing data were not imputed.
  • Follow-Up Period:
    • Primary outcome follow-up was to 6 months.
    • Support-free days were assessed at day 60.
    • Health-related quality of life and costs were assessed at 6 months.
  • Protocol:
    • The published protocol planned a phase 3, three-arm, open-label, non-inferiority trial with deferred consent and the same core anticoagulation strategies, but described more than ten Dutch and Belgian referral centres, whereas the final trial enrolled from seven Dutch centres.8

Key Results

This trial continued to completion. It enrolled the target sample of 330 patients; the final report describes faster-than-anticipated enrolment and lower-than-expected loss to follow-up, rather than stopping for benefit, harm, or futility.

Outcome Standard-dose UFH Low-dose UFH Therapeutic LMWH Effect p value / 95% CI Notes
Primary composite: severe bleeding during ECMO, severe thromboembolism during ECMO, or all-cause mortality at 6 months 87/107 (81%) 78/108 (72%) 79/105 (75%) Low-dose vs standard: ARD -9.1%; RR 0.89
LMWH vs standard: ARD -6.1%; RR 0.93		 Low-dose: 95% CI -20.3 to 2.1; RR 95% CI 0.77 to 1.03
LMWH: 95% CI -17.2 to 5.0; RR 95% CI 0.80 to 1.07		 Both met non-inferiority because the upper 95% CI was below the 7.5 percentage point margin. Adjusted RR was 0.88 (95% CI 0.81 to 1.02) for low-dose UFH and 0.91 (95% CI 0.79 to 1.05) for LMWH.
Severe bleeding during ECMO 70/107 (65%) 63/108 (58%) 62/105 (59%) Low-dose: ARD -7.1%
LMWH: ARD -6.4%		 Low-dose: 95% CI -20.0 to 5.9
LMWH: 95% CI -19.4 to 6.7		 Among 195 patients with severe bleeding, 100 (51%) subsequently died.
Severe thromboembolic complications during ECMO 12/107 (11%) 11/108 (10%) 9/105 (9%) Low-dose: ARD -1.0%
LMWH: ARD -2.6%		 Low-dose: 95% CI -9.3 to 7.2
LMWH: 95% CI -10.7 to 5.4		 Among 32 patients with severe thromboembolic complications, 17 (53%) died.
All-cause mortality at 6 months 54/107 (50%) 45/108 (42%) 46/105 (44%) Low-dose: ARD -8.8%
LMWH: ARD -6.7%		 Low-dose: 95% CI -21.0 to 5.0
LMWH: 95% CI -20.1 to 6.8		 Mortality numerically favoured lower-intensity strategies, but superiority was not established.
aPTT separation: lowest / highest average daily aPTT 48 s (IQR 43-54) / 68 s (60-77) 42 s (38-47) / 56 s (49-65) 34 s (30-40) / 38 s (33-50) Clear separation between allocated strategies Low-dose vs standard: P<0.001 for lowest and highest aPTT
LMWH vs standard: P<0.001 for lowest and highest aPTT		 Anti-Xa was measured in a subset of 71 patients after it was added midway through the trial and was not used for dose adjustment.
Red blood cell transfusion, total units 4 (IQR 1-11) 3 (0-9) 3 (1-9) No statistically significant difference Low-dose: P=0.17
LMWH: P=0.36		 Per ECMO day: 0.50 (0.20-1.19), 0.44 (0.00-1.00), and 0.50 (0.14-1.00), respectively.
Oxygenator exchanges per ECMO day 0.02 (SD 0.11) 0.02 (0.04) 0.01 (0.02) LMWH lower than standard-dose UFH Low-dose: P=0.86
LMWH: P=0.037		 Total oxygenator exchanges: 0.22 (SD 0.54), 0.24 (0.59), and 0.10 (0.46), respectively.
Days alive and free from ECMO at day 60 0 (IQR 0-53) 39 (0-55) 44 (0-54) Numerically more ECMO-free days in both intervention groups Low-dose: P=0.45
LMWH: P=0.24		 Median value of 0 in the standard-dose group reflects high mortality before day 60.
EQ-5D-5L utility among survivors at 6 months 0.80 (IQR 0.59-0.89) 0.83 (0.65-0.89) 0.79 (0.66-0.89) Preserved quality of life among survivors Low-dose: P=0.50
LMWH: P=0.70		 Baseline EQ-5D-5L utility was not collected.
Mean total costs €127,601 €118,890 €119,866 Lower mean costs in both intervention groups Standard: 95% CI €107,565 to €150,463
Low-dose: €102,685 to €136,904
LMWH: €103,561 to €137,679		 Cost-effectiveness probabilities across €0-€50,000 willingness-to-pay thresholds were 73-81% for low-dose UFH and 72-74% for LMWH.
Exploratory subgroup: VV-ECMO 6-month mortality, low-dose UFH vs standard-dose UFH 21/46 (46%) 9/40 (23%) 16/42 (38%) Low-dose vs standard: RR 0.49; ARD -23.15% RR 95% CI 0.26 to 0.95; ARD 95% CI -42.51 to -3.80 Exploratory only; no formal interaction testing or multiplicity control.
Crossovers to another anticoagulation regimen 23/107 (22%) 14/108 (13%) 29/105 (28%) Most crossovers occurred after the composite endpoint had already been reached Not reported Crossovers before endpoint: 1, 5, and 13 patients, respectively; per-protocol analysis was consistent with intention-to-treat.
  • Both reduced-intensity strategies satisfied non-inferiority and numerically reduced the primary composite, severe bleeding, and 6-month mortality compared with standard-dose UFH.
  • The trial did not show excess severe thrombosis with either low-dose UFH or therapeutic LMWH: 11%, 10%, and 9% in the standard-dose UFH, low-dose UFH, and LMWH groups, respectively.
  • The most practice-relevant signal is safety of lower-intensity anticoagulation; superiority for mortality, bleeding reduction, or cost-effectiveness remains suggestive rather than definitive.

Internal Validity

  • Randomisation and Allocation:
    • Randomisation was central, web-based, concealed until assignment, used variable block sizes, and was stratified by site and ECMO mode.
  • Drop-out or Exclusions:
    • Of 330 enrolled patients, 320 were included in the 6-month analysis.
    • The analysis groups were 107, 108, and 105 patients, with no missing data for the primary outcome or covariates in the main model.
  • Performance/Detection Bias:
    • Bedside treatment was open-label, which could influence transfusion, procedure, monitoring, and crossover decisions.
    • The primary components were clinically important and largely objective, and follow-up personnel were masked.
  • Protocol Adherence and Crossover:
    • Crossovers occurred in 23/107 (22%) with standard-dose UFH, 14/108 (13%) with low-dose UFH, and 29/105 (28%) with LMWH.
    • Only 1, 5, and 13 patients respectively crossed over before the endpoint.
    • Per-protocol findings were consistent with intention-to-treat findings.
  • Baseline Characteristics:
    • Baseline characteristics were broadly comparable.
    • APACHE IV scores were numerically higher in the low-dose UFH and LMWH groups: 61 (IQR 48-89), 70 (52-90), and 73 (55-96), respectively, which would tend to bias against the intervention groups.
  • Heterogeneity:
    • The trial intentionally combined VV and VA ECMO.
    • This improves breadth but introduces biological heterogeneity because thrombotic risk, bleeding risk, cannulation, shock physiology, and circuit flow differ substantially between modes.
  • Timing:
    • Anticoagulation was initiated after cannulation and could be deferred up to 24 h for active bleeding.
    • This was clinically appropriate but introduced early post-cannulation variability in exposure.
  • Dose and Separation of the Variable of Interest:
    • Dose separation was achieved: highest average daily aPTT was 68 s (IQR 60-77) with standard-dose UFH, 56 s (49-65) with low-dose UFH, and 38 s (33-50) with LMWH.
    • Lowest average daily aPTT also separated: 48 s (IQR 43-54), 42 s (38-47), and 34 s (30-40), respectively.
    • Anti-Xa was 0.37 IU/mL (0.25-0.46), 0.25 IU/mL (0.19-0.49), and 0.41 IU/mL (0.27-0.65) in the subset measured.
  • Key Delivery Aspects:
    • The trial tested pragmatic anticoagulation management rather than a tightly controlled pharmacometric protocol.
    • Local aPTT assays, UFH adjustment rules, transfusion practice, and LMWH products were not fully standardised.
  • Outcome Assessment:
    • The composite endpoint was clinically relevant but broad, combining ECMO-period bleeding and thrombosis with 6-month mortality.
    • This captures net clinical harm but makes mechanistic interpretation harder.
  • Statistical Rigor:
    • The primary non-inferiority analysis was prespecified, absolute-risk based, and concordant with adjusted and per-protocol analyses.
    • Two primary treatment comparisons, an interim analysis, numerous secondary outcomes, and subgroup analyses create multiplicity; secondary and subgroup findings should be considered exploratory.
    • Routine screening for venous thromboembolism during or after ECMO was not customary in participating centres, so clinically silent thrombosis may have been underestimated.
    • Detailed screening data distinguishing missed eligible patients from ineligible patients were not systematically collected, limiting assessment of selection bias.

Conclusion on Internal Validity: Internal validity is moderate-to-strong. The randomisation, allocation concealment, objective primary components, complete primary follow-up, and achieved anticoagulation separation support causal inference, but open-label treatment, local management variability, crossover, composite-outcome complexity, and non-inferiority-margin judgement temper certainty.

External Validity

  • Population Representativeness:
    • The trial included both VA ECMO (192/320; 60%) and VV ECMO (128/320; 40%), making it more representative than studies limited to VV ECMO.
    • Patients were very ill, with median SOFA scores of 11, 11, and 12, and 6-month mortality of 50%, 42%, and 44%, respectively.
  • Important Exclusions:
    • The findings do not apply to patients with mechanical mitral valves, proven pulmonary embolism, intracardiac thrombus, heparin-induced thrombocytopenia, or procedural ECMO, because these groups were excluded.
  • Healthcare Setting:
    • All final enrolling centres were Dutch ICUs.
    • Three centres contributed most participants: Amsterdam University Medical Center enrolled 119, University Medical Center Groningen enrolled 92, and Leiden University Medical Center enrolled 60.
    • This may limit generalisability to low-volume ECMO programmes, different transfusion cultures, or resource-limited settings.
  • Technology Context:
    • Most patients had heparin-coated cannulae: 85/107 (79%), 90/108 (83%), and 86/105 (82%).
    • Extrapolation to older, non-coated, or highly thrombogenic circuits should be cautious.
  • Monitoring Context:
    • The UFH strategy was aPTT-guided; centres using anti-Xa-guided UFH can reasonably view RATE as evidence for lower-intensity anticoagulation, but not as a direct comparison of monitoring strategies.
    • Contemporary guidance increasingly emphasises anti-Xa because aPTT is affected by inflammation, factor deficiency, lupus anticoagulant, and reagent variability.4
  • LMWH Applicability:
    • Therapeutic LMWH may reduce monitoring burden and was not associated with excess severe thrombosis.
    • Clinical uptake will depend on renal function, procedural frequency, reversibility concerns, local anti-Xa capacity, and confidence in subcutaneous absorption in shock.

Conclusion on External Validity: External validity is moderate. RATE is highly applicable to experienced adult ECMO centres using modern circuits and caring for patients without a separate full-anticoagulation indication, but less directly applicable to paediatric ECMO, HIT, mandatory full-dose anticoagulation, procedural ECMO, low-resource settings, or centres using substantially different monitoring systems.

Strengths & Limitations

  • Strengths:
    • First adequately powered randomised trial of ECMO anticoagulation intensity.
    • Three-arm design testing both lower-intensity UFH and therapeutic LMWH against standard-dose UFH.
    • Inclusion of both VV and VA ECMO increases relevance across contemporary ECMO practice.
    • Concealed central randomisation, stratification by site and ECMO mode, objective primary components, complete primary-outcome follow-up, and clear aPTT separation strengthen causal inference.
    • Patient-centred follow-up included 6-month mortality, health-related quality of life, and cost-effectiveness.
  • Limitations:
    • Open-label bedside management could affect transfusion, bleeding intervention, monitoring, and crossover decisions.
    • The composite primary endpoint mixes bleeding, thromboembolism, and mortality across different time windows and mechanisms.
    • The non-inferiority margin of 7.5 percentage points was clinically prespecified but remains judgement-based.
    • Local aPTT assays and UFH adjustment protocols were not standardised; time to therapeutic range and time in range were not analysed.
    • Subgroup and secondary-outcome findings were not multiplicity-adjusted and should not be interpreted as definitive evidence of superiority.
    • Detailed screening logs, routine VTE screening, and baseline EQ-5D-5L utility were not available.

Interpretation & Why It Matters

  • Clinical meaning:
    • For many adult ECMO patients without a separate full-anticoagulation indication, standard-dose UFH targeting aPTT 2.0-2.5 times baseline is probably more intense than necessary.
  • Practice implication:
    • Low-dose UFH targeting approximately 45-60 s is the most immediately translatable strategy because it preserves familiar UFH titration, reversibility, and procedural flexibility while reducing anticoagulation exposure.
  • LMWH implication:
    • Therapeutic LMWH is a plausible alternative in selected patients and had fewer oxygenator exchanges per ECMO day.
    • Renal function, bleeding risk, procedural plans, and reversibility make it less universally deployable than low-dose UFH.
  • What it refutes:
    • RATE weakens the assumption that higher heparin intensity is required to prevent clinically important thrombosis during modern ECMO.
  • What it does not prove:
    • The trial establishes non-inferiority; it does not prove that low-dose UFH or LMWH is superior for mortality, bleeding, or quality of life.

Controversies & Subsequent Evidence

  • Comparator intensity:
    • The standard-dose UFH arm used aPTT 2.0-2.5 times baseline, but many contemporary ECMO programmes have already drifted towards lower targets.
    • The accompanying Lancet Comment highlighted that the comparator may poorly mirror current practice in many settings.9
  • Non-inferiority interpretation:
    • The upper 95% CI was below the 7.5 percentage point non-inferiority margin for both comparisons.
    • The margin itself remains a clinical judgement; clinicians should interpret the absolute effects and component outcomes, not only the non-inferiority declaration.
  • Composite endpoint:
    • Severe bleeding, severe thromboembolism, and 6-month mortality are all important, but they differ in timing, mechanism, preventability, and susceptibility to anticoagulation intensity.
    • A favourable composite could mask divergent component effects, although this was not seen numerically in RATE.
  • Subgroup signals:
    • VV-ECMO mortality with low-dose UFH was lower than standard-dose UFH: 9/40 (23%) versus 21/46 (46%); RR 0.49; 95% CI 0.26 to 0.95.
    • This is clinically provocative but exploratory because no formal interaction testing or multiplicity control was performed.
  • Pre-RATE evidence synthesis:
    • A recent systematic review and meta-analysis reported that low-dose UFH was associated with lower mortality and intracranial haemorrhage than standard-dose UFH, without an apparent increase in systemic thrombosis.
    • RATE therefore converted a plausible observational signal into randomised evidence.10
  • Monitoring debate:
    • RATE tested lower aPTT targets rather than anti-Xa-guided anticoagulation.
    • It supports lower anticoagulation intensity more directly than it resolves the optimal monitoring assay.
  • LMWH debate:
    • LMWH had fewer oxygenator exchanges per ECMO day and non-inferior clinical outcomes.
    • The LMWH group had the highest crossover rate at 29/105 (28%), raising implementation questions in unstable ECMO patients.
  • Guidelines:
    • ELSO anticoagulation guidance and ISTH ECMO anticoagulation guidance predate RATE and therefore do not incorporate its randomised evidence.
    • Future updates will need to decide whether lower-intensity anticoagulation becomes the default rather than an exception.34

Summary

  • RATE randomised 330 adult ECMO patients to standard-dose UFH, low-dose UFH, or therapeutic LMWH; 320 were analysed at 6 months.
  • Low-dose UFH and LMWH were non-inferior to standard-dose UFH for severe bleeding, severe thromboembolism, or 6-month mortality.
  • Primary composite event rates were 81%, 72%, and 75%, respectively; severe bleeding was 65%, 58%, and 59%; severe thromboembolism was 11%, 10%, and 9%.
  • Anticoagulation separation was achieved: highest average daily aPTT was 68 s, 56 s, and 38 s in the standard-dose UFH, low-dose UFH, and LMWH groups.
  • The trial supports lower-intensity anticoagulation in selected adult ECMO patients, but superiority claims, subgroup signals, and LMWH implementation should be interpreted cautiously.

Overall Takeaway

RATE is an important practice-shaping ECMO anticoagulation trial because it directly challenges the historical assumption that full-dose UFH targets are required during modern ECMO. In adults without a separate indication for full-dose anticoagulation, lower-intensity anticoagulation appears clinically safe and may reduce bleeding-related harm, but implementation should be protocolised, monitored, and adapted to patient-specific thrombotic and bleeding risk.

Overall Summary

  • Low-dose UFH and therapeutic LMWH were non-inferior to standard-dose UFH for net clinical harm during and after ECMO.
  • The most actionable change is reconsideration of routine aPTT targets of 2.0-2.5 times baseline in adult ECMO patients without another full-anticoagulation indication.
  • RATE should shift the default question from “how much heparin is needed to protect the circuit?” to “how little anticoagulation safely balances bleeding, thrombosis, and survival?”

Bibliography

Added July 8th, 2026