Skip to main content

Foundational Trials

REVIVAL

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Phase-3 trial of recombinant human alkaline phosphatase for patients with sepsis-associated acute kidney injury (REVIVAL)
  • Acronym: REVIVAL
  • Year: 2024
  • Journal published in: Intensive Care Medicine
  • Citation: Pickkers P, Angus DC, Bass K, Bellomo R, van den Berg E, Bernholz J, et al. Phase-3 trial of recombinant human alkaline phosphatase for patients with sepsis-associated acute kidney injury (REVIVAL). Intensive Care Med. 2024;50:68-78.

Context & Rationale

  • Background
    • Sepsis-associated acute kidney injury (SA-AKI) is common in critical care, with high short-term mortality and frequent progression to acute kidney disease and chronic kidney disease.
    • Most SA-AKI care is supportive (source control, haemodynamic optimisation, avoidance of nephrotoxins, and renal replacement therapy when indicated), with no established disease-modifying pharmacotherapy.
    • Alkaline phosphatase (ALP) is mechanistically plausible in SA-AKI: dephosphorylation of endotoxin (LPS) and extracellular ATP may reduce inflammatory signalling, endothelial injury, and tubular stress.
    • A small randomised trial of bovine intestinal ALP in sepsis-induced AKI showed renal biomarker and creatinine clearance signals consistent with renoprotection.1
    • A subsequent phase 2 RCT of recombinant human ALP (ilofotase alfa; STOP-AKI) did not meet its primary endpoint (day-7 creatinine clearance) but suggested favourable kidney-centred outcomes (including MAKE90) and a mortality signal, supporting a phase 3 trial powered for mortality.2
  • Research Question/Hypothesis
    • In adults with sepsis requiring vasopressor support and recent-onset AKI, does early treatment with ilofotase alfa (vs placebo) reduce 28-day all-cause mortality?
    • Does ilofotase alfa improve kidney-centred and resource outcomes (MAKE90, RRT use, and days alive free of organ support)?
  • Why This Matters
    • SA-AKI remains a major driver of ICU mortality, RRT utilisation, and long-term kidney disability.
    • Phase 3 confirmation was needed because prior signals were generated in smaller studies with different endpoints and potentially different SA-AKI phenotypes.
    • The trial also informs how SA-AKI trials should balance mortality endpoints against kidney-centred composite outcomes in a heterogeneous syndrome.

Design & Methods

  • Research Question: Whether ilofotase alfa reduces 28-day all-cause mortality compared with placebo in adults with sepsis requiring vasopressors and recent-onset AKI.
  • Study Type: Randomised, double-blind, placebo-controlled, multicentre, international, phase 3, two-arm parallel-group trial with a preplanned group-sequential interim analysis for futility; ICU and intermediate care settings (107 sites).
  • Population:
    • Core clinical phenotype: adults with sepsis requiring continuous vasopressors and acute kidney injury (KDIGO creatinine/urine output criteria) of recent onset.
    • Haemodynamic inclusion threshold: vasopressor support at ≥0.1 μg/kg/min norepinephrine (or equivalent) for ≥1 hour (after the initial hour, any dose while on continuous infusion was acceptable).
    • Timing constraints: <72 hours of vasopressor treatment before starting trial drug; AKI diagnosed <24 hours before starting trial drug.
    • AKI inclusion criteria: increase in creatinine ≥0.3 mg/dL in 48 h, or ≥1.5× pre-AKI reference creatinine, or urine output <0.5 mL/kg/h for 6 h, or creatinine >2.0 mg/dL if no baseline creatinine was available.
    • Kidney function strata/cohorts: main population (pre-AKI eGFR ≥45 mL/min/1.73 m², and no suspected/confirmed COVID-19); a moderate CKD cohort (pre-AKI eGFR 25 to <45 mL/min/1.73 m² at selected sites, and no suspected/confirmed COVID-19); a COVID-19 cohort (confirmed COVID-19 infection; introduced during the pandemic).
    • Key exclusions: RRT initiated >24 hours before first dose; no longer requiring continuous vasopressors at randomisation; continuous vasopressors for >72 hours before first dose; AKI not attributable to sepsis; imminent death; urinary tract obstruction; severe pancreatitis; severe liver disease (Child–Pugh class C); chronic kidney disease below cohort thresholds (main/COVID excluded pre-AKI eGFR <45; moderate CKD cohort excluded pre-AKI eGFR <25).
    • Consent: written informed consent per local regulation (including emergency/deferred processes where applicable).
  • Intervention:
    • Ilofotase alfa (recombinant human alkaline phosphatase) 1.6 mg (1000 U)/kg, infused intravenously over 1 hour, once daily for 3 consecutive days.
    • Weight handling: dosing based on actual body weight up to 120 kg; patients >120 kg received a fixed 192 mg dose.
    • Administered early in the SA-AKI course (subject to timing constraints above).
  • Comparison:
    • Matched placebo infusion (intravenous), given once daily for 3 consecutive days, plus usual care for sepsis and AKI.
    • RRT initiation/termination was not protocolised; sites used conventional clinical indications.
  • Blinding: Double-blind (participants, clinicians, investigators, and outcome assessment); clinicians were asked not to measure serum ALP until day 14 to reduce the risk of functional unblinding from enzyme exposure.
  • Statistics: Maximum sample size of 1400 patients (main population) planned to provide ~85% power to detect an 8% absolute reduction in 28-day mortality (from 35% to 27%) at a one-sided α=0.025; first interim analysis planned after 400 participants with a non-binding futility threshold based on estimated power <15%; primary endpoint analysed in a modified intention-to-treat (mITT) population (all randomised patients in whom trial drug administration started, analysed by assigned group); secondary endpoints were prioritised for analysis in a combined population per amended statistical analysis plan (protocol published).3
  • Follow-Up Period: Primary endpoint at day 28; kidney/organ-support endpoints through day 90; deaths recorded through day 180; adverse events collected through day 28.

Key Results

This trial was stopped early. Recruitment was stopped after the first planned interim analysis (main trial mITT population N=411), because the predictive probability of success for the mortality endpoint was 1.2% (pre-specified futility threshold <15%).

Outcome Ilofotase alfa Placebo Effect p value / 95% CI Notes
28-day all-cause mortality (primary endpoint) 92/330 (27.9%) 88/319 (27.6%) Difference 0.46% 95% CI -6.38 to 7.29; P=0.448 Combined population; one-sided test; objective endpoint; trial stopped early for futility.
90-day all-cause mortality 112/330 (33.9%) 111/319 (34.8%) Difference -0.86% 95% CI -8.17 to 6.45; P=0.409 Deaths recorded through day 180 (overall mortality curve separation not demonstrated).
MAKE90A (death, RRT at day 90 or RRT through day 28, eGFR decline >25% at day 90, or rehospitalisation) 184/330 (55.8%) 204/319 (63.9%) Difference -7.91% 95% CI -15.40 to -0.42; P=0.019 Secondary endpoint; driven mainly by the RRT component (see below).
RRT at day 90 or RRT through day 28 (MAKE90A component) 90/330 (27.3%) 118/319 (37.0%) Difference -9.66% 95% CI -16.86 to -2.47; P=0.008 RRT initiation was based on conventional clinical indications (not protocolised).
eGFR decline >25% at day 90 versus pre-AKI reference (MAKE90A component) 79/330 (23.9%) 78/319 (24.5%) Difference -0.56% 95% CI -7.31 to 6.18; P=0.434 No signal for preservation of eGFR at day 90 in the overall combined population.
Rehospitalisation by day 90 (MAKE90A component) 112/330 (33.9%) 126/319 (39.5%) Difference -5.60% 95% CI -12.85 to 1.66; P=0.066 Borderline signal; ascertainment depends on health-system capture and follow-up completeness.
MAKE90B (death, RRT at day 90, eGFR decline >25% at both day 28 and day 90, or rehospitalisation) 223/330 (67.6%) 227/319 (71.3%) Difference -3.76% 95% CI -11.24 to 3.71; P=0.162 Alternative composite definition; did not reach nominal statistical significance.
Days alive and free of organ support to day 28 Median 17 (IQR 0–24) Median 14 (IQR 0–24) Not reported P=0.217 One-sided p value; patients dying before day 28 counted as 0 days.
Any adverse event (safety population) 224/330 (67.9%) 240/320 (75.0%) Not reported P=0.0446 Adverse events collected to day 28; denominators reflect treated safety population.
Serious adverse event (safety population) 128/330 (38.8%) 125/320 (39.1%) Not reported P=0.9800 No excess of serious adverse events with ilofotase alfa.
  • There was no evidence of a mortality benefit: 28-day mortality 27.9% with ilofotase alfa vs 27.6% with placebo (difference 0.46%; 95% CI -6.38 to 7.29; P=0.448).
  • A kidney-centred composite signal was observed for MAKE90A (55.8% vs 63.9%; difference -7.91%; 95% CI -15.40 to -0.42; P=0.019), largely attributable to less RRT through day 90 (27.3% vs 37.0%; difference -9.66%; 95% CI -16.86 to -2.47; P=0.008).
  • A post hoc interaction analysis suggested greater kidney benefit at lower pre-AKI eGFR (interaction P=0.024), with predicted probability of MAKE90A at eGFR 59.8 of 71% (placebo) vs 57% (ilofotase) and at eGFR 89 of 58% (placebo) vs 56% (ilofotase).

Internal Validity

  • Randomisation and allocation: central randomisation (1:1) with stratification by site and baseline modified SOFA (≤9 vs >9); allocation concealment supported by identical infusion processes.
  • Dropout/exclusions: 676 enrolled and 655 randomised; 650 received trial drug (mITT; 329 assigned ilofotase and 321 assigned placebo); 5 randomised but not treated (2 ilofotase, 3 placebo); 1 treated patient lacked efficacy data in the combined population analysis.
  • Performance/detection bias: double-blinding (including avoidance of serum ALP measurement until day 14) makes differential co-interventions less likely, particularly for objective outcomes (mortality).
  • Protocol adherence: all treated patients received at least one infusion; completion of all 3 doses was not reported in the main manuscript.
  • Baseline characteristics: groups were well matched in severity (median mSOFA 9 in both; mechanical ventilation 74.8% vs 75.6%; AKI stage 3: 21.8% vs 25.2%; median lactate 2.2 vs 2.4 mmol/L; median pre-AKI eGFR 74.0 vs 73.9 mL/min/1.73m²).
  • Heterogeneity: multiple cohorts (main, moderate CKD, COVID-19) increase clinical heterogeneity; outcomes were presented for combined and cohort-specific populations.
  • Timing: eligibility required early administration relative to vasopressor initiation and AKI diagnosis (within prespecified windows); exact time-to-first-dose distributions were not reported.
  • Dose: fixed regimen (1.6 mg/kg daily ×3) without dose-ranging; weight capping (120 kg) may affect exposure in the heaviest patients.
  • Separation of the variable of interest: crossover was rare (3/655 randomised); safety denominators show near-complete separation (330 ilofotase-treated vs 320 placebo-treated).
  • Outcome assessment: mortality is objective; MAKE90 components include clinician-driven RRT initiation and rehospitalisation, which are less standardised across sites and systems.
  • Statistical rigour: preplanned interim futility stopping occurred; early stopping reduces power and precision for mortality; one-sided hypothesis testing was used; secondary endpoints were analysed under an amended plan that prioritised a combined population analysis.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong: randomisation, blinding, and low crossover support robust estimation for mortality, but early stopping and the complexity of secondary endpoint definitions/populations introduce uncertainty around the kidney composite signal.

External Validity

  • Population representativeness: patients were critically ill (high ventilator use, shock requiring vasopressors) across 107 international sites, resembling many high-acuity ICU sepsis populations.
  • Key exclusions affecting generalisability: severe CKD below cohort thresholds, chronic dialysis, prolonged vasopressor exposure (>72 h), and delayed AKI presentation beyond the recruitment windows.
  • Applicability: the intervention (three 1-hour IV infusions) is feasible in well-resourced ICUs with pharmacy support; applicability may be reduced in settings where early AKI recognition, vasopressor timing documentation, or follow-up/rehospitalisation capture is limited.
  • Subpopulation uncertainty: signals suggested potential heterogeneity by baseline kidney function (greater kidney benefit at lower pre-AKI eGFR), but this remains hypothesis-generating.

Conclusion on External Validity: Generalisability is moderate, best aligned to adults with septic shock and early AKI in high-resource ICUs; translation to advanced CKD, delayed presenters, and resource-limited systems is uncertain.

Strengths & Limitations

  • Strengths:
    • International, multicentre, double-blind, placebo-controlled design across 107 sites.
    • Objective primary endpoint (28-day all-cause mortality) with mortality tracking to 180 days.
    • Preplanned group-sequential design with independent interim monitoring for futility.
    • Biologically grounded intervention with prior phase 2 signal supporting equipoise.
    • Large treated population for a sepsis pharmacotherapy trial (650 treated; 649 in efficacy analysis of combined population).
  • Limitations:
    • Stopped early for futility, reducing power and precision to detect modest mortality effects.
    • Clinical heterogeneity introduced by multiple cohorts (main, moderate CKD, COVID-19) and evolving trial context.
    • Kidney composite outcomes include clinician- and system-dependent elements (RRT initiation, rehospitalisation) that are variably standardised across sites.
    • Secondary endpoint analysis prioritised a combined population under an amended plan, complicating inference about kidney benefit while preserving primary endpoint analysis in mITT.
    • Completion of all three doses and granular timing-to-treatment metrics were not reported in the main manuscript.

Interpretation & Why It Matters

  • Clinical practice
    • Ilofotase alfa did not reduce 28-day mortality in SA-AKI and should not be adopted as routine therapy on the basis of REVIVAL.
    • Supportive care remains the standard of care for SA-AKI (haemodynamics, sepsis management, and RRT based on clinical indications).
  • Kidney signal
    • The MAKE90A reduction (absolute difference -7.91%) was largely attributable to lower RRT utilisation (absolute difference -9.66%), without an overall day-90 eGFR preservation signal.
    • Effect modification by baseline pre-AKI eGFR (interaction P=0.024) supports a hypothesis that benefit—if real—may concentrate in specific phenotypes (e.g., lower baseline eGFR).
  • Trials methodology
    • REVIVAL illustrates the difficulty of translating phase 2 biomarker/renal signals into a phase 3 mortality benefit in a biologically heterogeneous syndrome.
    • It reinforces the need for carefully selected endpoints (including kidney-centred and hierarchical outcomes) and potential phenotype enrichment in SA-AKI trials.

Controversies & Subsequent Evidence

  • Early stopping and power: the futility stop at the first interim analysis (predictive probability of success 1.2%) materially reduced the ability to detect smaller mortality effects and widened uncertainty around mortality estimates.
  • Endpoint choice in SA-AKI: the null mortality result despite kidney composite signals aligns with the broader SA-AKI endpoint debate (mortality may be insensitive to kidney-targeted biological effects in heterogeneous ICU populations).4
  • Composite kidney endpoints: MAKE90A’s signal was largely driven by RRT use; RRT initiation is clinician-mediated and practice-variable, which can influence composite behaviour even in blinded trials (particularly across international sites with different thresholds and resources).
  • Protocol and analysis adaptations: the protocol and statistical approach (including cohort structure and planned analyses) were published, but the combination of pandemic-era cohort additions and secondary endpoint prioritisation increased interpretive complexity for kidney outcomes.3
  • Phenotype heterogeneity: post hoc modelling suggested greater kidney benefit at lower pre-AKI eGFR, consistent with the hypothesis that baseline kidney vulnerability modifies therapeutic responsiveness; SA-AKI consensus work emphasises the biological heterogeneity underpinning such effect modification and the need for improved phenotyping.5
  • Guideline incorporation: contemporary critical care guidance for SA-AKI continues to emphasise sepsis treatment, haemodynamic optimisation, and RRT based on clinical indications; alkaline phosphatase is not embedded as standard therapy.6
  • Positioning within the alkaline phosphatase evidence base: REVIVAL’s mortality-neutral result contrasts with earlier smaller studies and emphasises the need for confirmatory, phenotype-informed trials if further development continues.12

Summary

  • REVIVAL tested ilofotase alfa (1.6 mg/kg IV daily ×3) versus placebo in adults with sepsis requiring vasopressors and recent-onset AKI across 107 international sites.
  • The trial stopped early for futility after the first interim analysis (predictive probability of success 1.2%).
  • There was no mortality benefit: 28-day mortality 27.9% (ilofotase) vs 27.6% (placebo); 90-day mortality 33.9% vs 34.8%.
  • MAKE90A was lower with ilofotase (55.8% vs 63.9%), driven primarily by lower RRT utilisation (27.3% vs 37.0), without an overall day-90 eGFR preservation signal.
  • Post hoc analysis suggested baseline kidney function may modify treatment response (greater kidney composite benefit at lower pre-AKI eGFR), supporting phenotype-enriched future trials rather than broad adoption.

Overall Takeaway

REVIVAL provides the most rigorous phase 3 test to date of recombinant human alkaline phosphatase in SA-AKI and demonstrates no reduction in 28-day mortality, with the trial stopped early for futility. A kidney composite signal (MAKE90A) driven mainly by reduced RRT use is hypothesis-generating, particularly given evidence of effect modification by baseline kidney function. Clinically, the trial consolidates supportive care as the standard, while methodologically it reinforces the need for phenotype enrichment and carefully selected kidney-centred endpoints in SA-AKI trials.

Overall Summary

  • No mortality benefit: 28-day mortality 27.9% (ilofotase) vs 27.6% (placebo); trial stopped early for futility.
  • MAKE90A improved (55.8% vs 63.9%), largely through lower RRT exposure (27.3% vs 37.0).
  • Post hoc analyses suggest baseline eGFR may modify kidney benefit; confirmatory phenotype-enriched trials would be needed before practice adoption.

Bibliography