Skip to main content

Foundational Trials

Hemodiafe

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial
  • Acronym: Hemodiafe
  • Year: 2006
  • Journal published in: The Lancet
  • Citation: Vinsonneau C, Camus C, Combes A, Costa de Beauregard MA, Klouche K, Boulain T, et al; Hemodiafe Study Group. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet. 2006;368:379-85.

Context & Rationale

  • Background
    Acute kidney injury requiring renal replacement therapy (RRT) in ICU patients with multiple-organ dysfunction syndrome (MODS) carries high mortality.
    Intermittent haemodialysis (IHD) and continuous renal replacement therapy (CRRT, here continuous venovenous haemodiafiltration [CVVHDF]) were both widely used, but with strong physiological arguments and major practice variation.
    Pre-trial evidence consisted largely of single-centre experience, small randomised studies, and confounded observational comparisons; it was unclear whether modality choice affected survival, renal recovery, or treatment tolerance in the sickest patients.
  • Research Question/Hypothesis
    In ICU patients with acute renal failure and MODS severe enough to warrant RRT, does CVVHDF improve 60-day survival compared with IHD?
  • Why This Matters
    Modality selection has major consequences for haemodynamic management, nursing workload, anticoagulation strategy, and resource allocation (machines, staffing, consumables).
    A definitive multicentre trial was needed to inform evidence-based modality choice and to determine whether any survival advantage was large enough to justify routine use of one strategy over the other.

Design & Methods

  • Research Question: Among ICU patients with acute renal failure and MODS, does CVVHDF (vs IHD) increase 60-day survival?
  • Study Type: Prospective, multicentre, parallel-group, open-label (non-blinded), randomised trial in 21 medical or multidisciplinary ICUs in France (Oct 1999–Mar 2003); allocation stratified by centre; outcome assessors stated as blinded.
  • Population:
    • Core setting: Critically ill adults with acute renal failure and MODS requiring RRT.
    • Key inclusion (protocol-defined): Need for RRT plus severe MODS (logistic organ dysfunction [LOD] score ≥6) and biochemical criteria (serum urea ≥36 mmol/L and/or serum creatinine ≥310 μmol/L); after ~8 months, oliguria (<320 mL over 16 h despite fluid loading) was added due to low enrolment.
    • Key exclusions (selected): Age <18 years; pregnancy; pre-existing chronic renal failure (pre-ARF serum creatinine >180 μmol/L); obstructive or vascular aetiology of acute renal failure; ongoing angiotensin-converting enzyme inhibitor exposure at inclusion; major bleeding risk (prothrombin time <20% and/or platelets <30×109/L) or uncontrolled haemorrhage; less severe illness (SAPS II ≤37); moribund/expected survival <8 days.
  • Intervention:
    • CVVHDF: Delivered with a dedicated CRRT system (Aquarius; Hospal) using a polymer membrane; mean blood flow 145.7 mL/min (95% CI 138.7–152.8).
    • Delivered solute clearance prescription (trial-reported): Dialysate flow 1099 mL/h (1053–1145) and ultrafiltration/replacement flow 1278 mL/h (1222–1334); average delivered “dose” reported as 29 mL/kg per h.
    • Net fluid removal (trial-reported): 2107 mL/day (1738–2476).
    • Anticoagulation: Unfractionated heparin when feasible; non-heparin strategies (e.g., saline flushes) if contraindicated.
  • Comparison:
    • IHD: Recommended daily sessions, with pragmatic haemodynamic-tolerance strategies (high sodium dialysate ~145 mmol/L and cool dialysate ~35°C during the first 24 h after randomisation).
    • Delivered treatment (trial-reported): Mean session duration 5.2 h (95% CI 4.9–5.6), mean blood flow 278 mL/min (269–288), dialysate flow 500 mL/min (fixed); net ultrafiltration 2213 mL/day (1944–2482).
    • Switching/rescue: Switching to the other modality permitted if clinically necessary (e.g., haemodynamic instability, access problems, circuit issues).
  • Blinding: Non-blinded treatment allocation; outcome assessors reported as blinded (important because some secondary endpoints include potentially clinician-influenced decisions such as RRT discontinuation and switching).
  • Statistics: Power calculation: initially 480 patients (240/group) to detect a 10% absolute increase in survival (50% to 60%) with 80% power at α=0.05; after early observed survival around 22%, sample size was recalculated to 360 patients (180/group) to test the same absolute difference; primary analysis was intention-to-treat.
  • Follow-Up Period: Survival to day 60 (primary), with additional day 28 and day 90 survival; renal recovery/dialysis dependence assessed in survivors after ICU and hospital discharge (as reported).

Key Results

This trial was not stopped early. Recruitment concluded after the prespecified (recalculated) sample size of 360 participants was reached.

Outcome CVVHDF IHD Effect p value / 95% CI Notes
Survival at day 60 (primary endpoint) 32.6% (95% CI 25.6–39.5) 31.5% (24.8–38.2) Absolute difference 1.1% 95% CI -8.8 to 11.1; P=0.98 Intention-to-treat; no evidence of survival benefit with CVVHDF.
Survival at day 28 38.9% (31.7–46.1) 41.8% (34.6–48.9) Not reported P=0.65 Secondary endpoint.
Survival at day 90 28.5% (21.8–35.2) 27.2% (20.8–33.6) Not reported P=0.95 Secondary endpoint.
Duration of renal support 11 days (95% CI 8–14) 11 days (8–13) Not reported P=0.84 Reported as mean (95% CI).
Length of ICU stay 19 days (95% CI 15–22) 20 days (16–23) Not reported P=0.73 Reported as mean (95% CI).
Length of hospital stay 32 days (95% CI 22–42) 30 days (24–35) Not reported P=0.66 Reported as mean (95% CI).
Dialysis dependence after ICU discharge (among assessed survivors) 4/61 (6.6%) 6/61 (9.8%) Not reported P=0.5 Assessment denominators reported as 61 in each arm.
Hypothermia (adverse event) 31 (17%) 10 (5%) Not reported P=0.0005 Clinically important modality-associated harm signal (temperature control).
Unplanned switch to the other modality 17 (10%) 6 (3%) Not reported P=0.01 Switching permitted; higher switching from CVVHDF to IHD.
  • Neutral primary endpoint: Day 60 survival was essentially identical (CVVHDF 32.6% vs IHD 31.5%; absolute difference 1.1%; 95% CI -8.8 to 11.1; P=0.98).
  • System-level outcomes similar: Renal support duration (11 vs 11 days), ICU stay (19 vs 20 days), and hospital stay (32 vs 30 days) did not differ (all P>0.6).
  • Safety/tolerability signal: Hypothermia occurred more often with CVVHDF (17% vs 5%; P=0.0005), and unplanned switching was more frequent from CVVHDF to IHD (10% vs 3%; P=0.01).

Internal Validity

  • Randomisation and allocation: Central computer-generated randomisation, balanced blocks of four, stratified by centre; allocation reported as concealed until enrolment was complete.
  • Attrition: 360 randomised (184 IHD; 176 CVVHDF); 1 patient in the CVVHDF group withdrew consent after randomisation and was excluded from analysis; no other loss to follow-up reported for survival endpoints.
  • Performance/detection bias: Open-label delivery creates risk for clinician-mediated outcomes (switching, RRT discontinuation), although the primary outcome (mortality) is objective; outcome assessors were reported as blinded.
  • Protocol adherence and delivered separation:
    • IHD delivered: mean session duration 5.2 h; blood flow 278 mL/min; dialysate flow 500 mL/min.
    • CVVHDF delivered: blood flow 145.7 mL/min; dialysate flow 1099 mL/h; ultrafiltration/replacement flow 1278 mL/h; average dose 29 mL/kg per h.
    • Biochemical “separation” in azotaemia was small (mean serum urea 14.8 mmol/L CVVHDF vs 15.7 mmol/L IHD), suggesting broadly comparable solute control at the level measured.
  • Crossover/switching: Unplanned switching was more frequent in the CVVHDF arm (10%) than the IHD arm (3%), potentially diluting any true modality effect and reflecting pragmatic feasibility constraints.
  • Baseline comparability: Groups were well matched on key severity indicators (e.g., mean SAPS II ~64–65; mean LOD ~10; high prevalence of mechanical ventilation and vasopressor support), supporting exchangeability.
  • Timing: Enrolment occurred early-to-intermediate in ICU course (median ~2–3 days after ICU admission), but not “ultra-early”; the pathobiology and reversibility of AKI may already have been established in many patients.
  • Heterogeneity: Multicentre conduct increases generalisability, but centre-level practice differences (dialysis prescription, anticoagulation, catheter care) and permitted switching introduce treatment heterogeneity; no major centre interaction was highlighted in the published report.
  • Statistical rigour: Intention-to-treat analysis aligned with trial design; however, the sample size revision and lower-than-anticipated survival reduce precision, and the confidence interval around the primary effect still permits modest benefit or harm.

Conclusion on Internal Validity: Overall, internal validity appears moderate to strong for the mortality endpoint (robust randomisation; objective outcome; minimal attrition), but is limited for modality-specific mechanistic inference due to open-label care, substantial switching (especially from CVVHDF to IHD), and imperfect comparability of delivered “dose” metrics across modalities.

External Validity

  • Population representativeness: Highly representative of severe ICU AKI with MODS (very high rates of mechanical ventilation and vasopressor use; majority septic), but not representative of milder AKI (SAPS II ≤37 excluded) or patients with significant pre-existing chronic kidney disease.
  • Intervention feasibility: Both modalities were delivered with relatively optimised prescriptions (e.g., haemodynamic-tolerance strategies for IHD; quantified CVVHDF dose), which may exceed the consistency achievable in less-resourced settings.
  • Applicability across systems: Findings are most applicable to centres that can deliver both IHD and CVVHDF at high quality with experienced staff; the comparative balance may differ where one modality is delivered suboptimally or where staffing constraints limit continuous therapies.

Conclusion on External Validity: Generalisability is good to high-acuity ICUs managing severe AKI with MODS in well-resourced settings, but more limited for less severe AKI, pre-existing advanced CKD, and settings where consistent delivery of either modality (particularly CVVHDF) is constrained.

Strengths & Limitations

  • Strengths: Multicentre randomised design; clinically important primary endpoint (60-day survival); severe, high-risk population (high event rate); pragmatic allowance of switching reflecting real-world practice; detailed reporting of delivered treatment parameters and adverse events.
  • Limitations: Non-blinded treatment assignment; substantial switching/crossover (especially from CVVHDF to IHD) potentially diluting modality effects; “dose” comparability between IHD and CVVHDF remains methodologically challenging (CVVHDF dose quantified; intermittent delivered clearance not directly expressed in equivalent terms); inclusion criteria amended early in the trial (oliguria added) with potential for population drift.

Interpretation & Why It Matters

  • Clinical practice implication
    When both modalities are delivered with careful prescriptions, CVVHDF did not improve survival over IHD in severe ICU AKI with MODS; modality choice should therefore prioritise haemodynamic stability, fluid balance needs, anticoagulation constraints, staffing, and local expertise.
  • Operational implication
    A “modality-first” approach is less defensible than a “delivery-first” approach: ensuring timely initiation, minimising interruptions, preventing complications (e.g., hypothermia during CVVHDF), and enabling rational switching may matter more than rigid adherence to one modality.
  • Mechanistic inference
    The trial challenges the assumption that continuous modalities inherently improve outcomes via superior haemodynamic tolerance; the observed hypothermia excess with CVVHDF underscores that continuous therapies can introduce modality-specific harms if temperature management is not optimised.

Controversies & Subsequent Evidence

  • Interpretation of a “neutral” result: The accompanying editorial argued that, in a setting where both treatments are delivered competently, the absence of a mortality difference supports prioritising high-quality delivery (dose, continuity, complication prevention) over claims of intrinsic modality superiority.1
  • Switching as both strength and threat: Permitting switching improves pragmatic relevance, but higher switching from CVVHDF to IHD (10% vs 3%) can dilute effect estimates and makes “pure modality” comparisons harder to interpret as causal.
  • Broader evidence base: Systematic reviews aggregating randomised evidence have not demonstrated a consistent mortality advantage for continuous over intermittent modalities; differences, where present, tend to relate to haemodynamic tolerance and selected renal-recovery outcomes rather than survival.2
  • Guideline position: KDIGO recommends individualising modality choice, commonly favouring continuous modalities in haemodynamically unstable patients, while acknowledging the low-to-moderate certainty of evidence for hard outcomes.3
  • Sepsis-focused guidance: The Surviving Sepsis Campaign guidelines similarly do not endorse a universal survival benefit of one RRT modality and emphasise choosing a strategy consistent with haemodynamics, fluid management goals, and available resources.4

Summary

  • In 360 ICU patients with acute renal failure and MODS, CVVHDF did not improve 60-day survival compared with IHD (32.6% vs 31.5%; absolute difference 1.1%; 95% CI -8.8 to 11.1; P=0.98).
  • Secondary outcomes (28- and 90-day survival, ICU/hospital length of stay, duration of renal support) were similarly neutral.
  • CVVHDF was associated with more hypothermia (17% vs 5%; P=0.0005), highlighting a preventable modality-associated harm.
  • Unplanned switching was more frequent from CVVHDF to IHD (10% vs 3%; P=0.01), which both reflects real-world constraints and dilutes modality separation.
  • The trial supports a pragmatic approach: choose modality based on haemodynamics, fluid goals, anticoagulation constraints, and local capability rather than expecting a survival benefit from “continuous” therapy alone.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • In this summary, “Not reported” indicates that the specific bibliographic detail (e.g., volume/pages) was not available from the provided trial documents and accessible sources used to construct this page.
  • Delivered clearance metrics are inherently more comparable within, than across, intermittent and continuous modalities; the trial reported serum urea and treatment parameters but did not express a unified delivered-dose metric across arms.

Overall Takeaway

Hemodiafe is a landmark pragmatic ICU RRT modality trial because it tested, at multicentre scale and in a very high-risk MODS population, whether “continuous” therapy improves survival over optimised intermittent haemodialysis—and found no survival advantage. Its enduring contribution is to shift emphasis from modality dogma to the quality, continuity, and safety of RRT delivery (including recognising and preventing modality-specific harms such as hypothermia).

Overall Summary

  • CVVHDF did not improve 60-day survival compared with IHD in severe ICU AKI with MODS.
  • Key system outcomes (RRT duration, ICU/hospital stay) were similar; temperature-related complications were more frequent with CVVHDF.
  • Modern practice should individualise modality based on haemodynamics, fluid goals, anticoagulation feasibility, and delivery capability rather than expecting a mortality benefit from one modality alone.

Bibliography

  • 1Kellum JA, Palevsky PM. Renal support in acute kidney injury. Lancet. 2006;368:344-5.
  • 2Rabindranath KS, Adams J, Macleod AM, Muirhead N, et al. Intermittent versus continuous renal replacement therapy for acute renal failure in adults. Cochrane Database Syst Rev. 2007;(3):CD003773.
  • 3Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl. 2012;2:1-138.
  • 4Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock 2021. Intensive Care Med. 2021;47:1181-1247.