Skip to main content

Foundational Trials

REDOXS

Image: Shutterstock / ChaNaWiT

Publication

  • Title: A randomized trial of glutamine and antioxidants in critically ill patients
  • Acronym: REDOXS
  • Year: 2013
  • Journal published in: New England Journal of Medicine
  • Citation: Heyland DK, Muscedere J, Wischmeyer PE, Cook D, Jones G, Albert M, et al. A randomized trial of glutamine and antioxidants in critically ill patients. N Engl J Med. 2013;368(16):1489-1497.

Context & Rationale

  • Background
    • Glutamine and antioxidant “pharmaconutrition” were widely used in ICU practice on the basis of mechanistic plausibility (gut integrity, immune function, oxidative stress) and signal from smaller trials.
    • Existing evidence was heterogeneous (population severity, routes of delivery, dosing, co-interventions) and at risk of small-study effects, leaving uncertainty about net benefit versus harm in contemporary ICU cohorts.
    • There was particular uncertainty in early, severe critical illness with multi-organ failure, where metabolic handling and organ-specific toxicity might differ from elective surgical or stable ICU populations.
  • Research Question/Hypothesis
    • In mechanically ventilated adults with at least two acute organ failures, does early supplementation with glutamine and/or antioxidants (enteral plus parenteral delivery) reduce 28-day mortality compared with placebo?
    • Factorial hypothesis: each intervention (glutamine; antioxidants) would independently reduce mortality, with no clinically important interaction.
  • Why This Matters
    • These supplements were already embedded in practice and guidelines in some settings; clarifying benefit/harm had immediate implications for patient safety and resource use.
    • A pragmatic, adequately powered, blinded ICU trial was needed to resolve uncertainty created by mechanistic enthusiasm and inconsistent prior clinical evidence.

Design & Methods

  • Research Question: Among mechanically ventilated adults with ≥2 acute organ failures, does early high-dose glutamine and/or an antioxidant cocktail (enteral + parenteral) reduce 28-day mortality versus placebo?
  • Study Type: Randomised, multicentre, international, double-blind, placebo-controlled 2×2 factorial trial in adult ICUs (40 ICUs across Canada, the United States, and Europe); study treatments initiated within 24 hours of ICU admission and continued until ICU discharge, death, or day 28.
  • Population:
    • Key inclusion: Adults admitted to ICU; mechanically ventilated; ICU length of stay <24 hours; ≥2 acute organ failures related to current illness, defined as: respiratory failure (PaO2/FiO2 ≤300), hypoperfusion (vasopressor support or mean arterial pressure <60 mmHg, or lactate ≥2.5 mmol/L after adequate fluid resuscitation), renal dysfunction (serum creatinine ≥171 μmol/L or urine output <500 mL/24 h), and/or thrombocytopenia (platelets ≤50×109/L).
    • Key exclusions: ICU stay ≥24 hours; burns >40% TBSA; severe traumatic brain injury; grand mal seizures or status epilepticus requiring anticonvulsants; transplantation; Child–Pugh class C liver disease; metastatic cancer; DNR/comfort measures only; current haemodialysis; creatinine clearance <10 mL/min without haemodialysis; pregnancy; body weight <50 kg or >200 kg.
  • Intervention:
    • Factorial allocation (each patient received two blinded assignments): glutamine versus glutamine-placebo, and antioxidants versus antioxidant-placebo.
    • Glutamine strategy: parenteral glutamine 0.35 g/kg/day (maximum 40 g/day) infused continuously over 20–24 hours, plus enteral glutamine 30 g/day.
    • Antioxidant strategy: parenteral selenium 500 μg/day (maximum 700 μg/day) infused over 20–24 hours, plus enteral antioxidant cocktail containing selenium 300 μg/day (maximum 400 μg/day) and additional antioxidant micronutrients.
  • Comparison:
    • Placebo strategy: matching blinded placebos for both parenteral and enteral components (parenteral saline placebo; enteral placebo).
    • All groups received usual ICU care, including usual nutrition support (enteral nutrition and/or parenteral nutrition) at clinician discretion; study supplements were delivered in addition to standard feeding, not as a replacement.
  • Blinding: Double-blind for both interventions (patients, treating clinicians, and outcome assessors); use of matched placebo preparations for enteral and parenteral study agents.
  • Statistics: 1,200 patients required to detect a 25% relative reduction in 28-day mortality (from 30.0% to 22.5%) with 80% power at a two-sided 5% significance level for each intervention main effect; primary analysis by intention-to-treat with interim monitoring (final primary-outcome threshold P<0.044); regression models used to estimate adjusted odds ratios (28-day, ICU, and hospital mortality) and an adjusted hazard ratio (6-month mortality).
  • Follow-Up Period: Primary endpoint at 28 days; additional follow-up to hospital discharge and 6 months.

Key Results

This trial was not stopped early. It completed enrolment (1,223 randomised; 1,218 included in the primary analysis because 28-day vital status could not be ascertained for 5 patients).

Outcome Active Control Effect p value / 95% CI Notes
28-day mortality (glutamine main effect) 198/611 (32.4%) 166/607 (27.3%) Adjusted OR 1.28 95% CI 1.00 to 1.64; P=0.05 Primary outcome; prespecified final threshold P<0.044 (interim monitoring), therefore not statistically significant at the prespecified alpha.
28-day mortality (antioxidants main effect) 197/617 (31.9%) 167/601 (27.8%) Adjusted OR 1.09 95% CI 0.86 to 1.40; P=0.48 No mortality benefit; factorial main effect estimate.
In-hospital mortality (glutamine main effect) 227/611 (37.2%) 188/607 (31.0%) Adjusted OR 1.46 95% CI 1.07 to 2.00; P=0.02 Directionally consistent harm beyond day 28; objective endpoint.
In-hospital mortality (antioxidants main effect) 217/617 (35.2%) 198/601 (32.9%) Adjusted OR 1.15 95% CI 0.92 to 1.44; P=0.24 No mortality benefit.
6-month mortality (glutamine main effect) 267/611 (43.7%) 226/607 (37.2%) Adjusted HR 1.18 95% CI 1.02 to 1.36; P=0.02 Sustained excess mortality; analysed as time-to-event.
6-month mortality (antioxidants main effect) 250/617 (40.5%) 243/601 (40.4%) Adjusted HR 1.01 95% CI 0.87 to 1.16; P=0.81 No mortality benefit.
High urea (>50 mmol/L) 13.4% 4.0% Not reported P<0.001 Safety signal associated with glutamine assignment.
Time to discontinuation of mechanical ventilation and alive (days) 11.0 8.7 Not reported P=0.03 Competing risk of death makes this endpoint clinically interpretable alongside mortality.
  • Early high-dose glutamine (enteral + parenteral) was associated with higher in-hospital mortality (37.2% vs 31.0%) and higher 6-month mortality (43.7% vs 37.2%), despite the 28-day mortality P value (0.05) not meeting the prespecified final alpha (P<0.044).
  • The antioxidant strategy (enteral cocktail + parenteral selenium) did not reduce mortality at 28 days, in hospital, or at 6 months.
  • Biochemical harm signal was present: high urea (>50 mmol/L) occurred in 13.4% with glutamine versus 4.0% without glutamine (P<0.001).

Internal Validity

  • Randomisation and allocation concealment: Central randomisation with concealed allocation; stratified by centre; double-blind matched placebo design limits selection and performance bias.
  • Drop-out / exclusions: 1,223 randomised; 5 excluded from the primary analysis because 28-day vital status was not ascertainable (primary analysis n=1,218); magnitude of missingness is small but non-zero for the primary endpoint.
  • Performance/detection bias: Mortality outcomes are objective; blinding of two study agents and placebos reduces differential co-intervention and outcome assessment bias.
  • Protocol adherence: Enteral study nutrients delivered at ~69.1–72.4% of prescribed across groups; parenteral study nutrients at ~88.3–89.8% of prescribed across groups; adherence appears adequate to test the intervention as delivered.
  • Timing (delivered within intended window): Median ICU stay before randomisation ~18 hours (e.g., 18.1 hours placebo vs 18.5 hours glutamine); median time from first organ dysfunction to initiation ~21–23 hours for enteral study supplements and ~22–23 hours for parenteral study supplements, supporting “early” delivery in acute organ failure.
  • Dose (separation of exposure): Glutamine assignment markedly increased nitrogen delivery (e.g., 20.4 ± 7.0 g with glutamine vs 6.9 ± 4.4 g without glutamine; P<0.001), confirming strong separation of the biological exposure.
  • Baseline comparability / severity: Groups were broadly similar at baseline (mean APACHE II ~26.5; high prevalence of hypoperfusion and respiratory failure by inclusion definition), supporting exchangeability; enrolled cohort was sufficiently severe to plausibly benefit (or be harmed) by pharmaconutrition.
  • Heterogeneity: Multinational, multicentre ICU enrolment increases clinical heterogeneity, but stratified randomisation by centre and objective endpoints mitigate internal validity threats; any treatment effect heterogeneity is primarily an interpretation/generalisation issue rather than a bias issue.
  • Separation of the variable of interest (clinical endpoints): Glutamine was associated with worse time-to-event endpoints incorporating survival (e.g., time to discontinuation of ventilation and alive 11.0 vs 8.7 days; P=0.03), and delayed discharge alive from ICU (17.1 vs 13.1 days; P=0.03) and hospital (51.0 vs 40.1 days; P=0.04), consistent with biological and clinical separation rather than contamination.
  • Outcome assessment and definitions: Primary (28-day mortality) and major secondary outcomes (in-hospital and 6-month mortality) were prespecified and clinically meaningful; statistical models were prespecified with interim monitoring (final P<0.044 for the primary outcome) and intention-to-treat analysis.

Conclusion on Internal Validity: Internal validity is strong for the question actually tested (early, high-dose, combined-route glutamine and an antioxidant strategy in ventilated adults with acute multi-organ failure), with robust concealment/blinding and clear exposure separation; interpretive complexity mainly relates to the factorial structure and the clinical meaning of “early high-dose” supplementation.

External Validity

  • Population representativeness: Enrolled a high-acuity ICU phenotype (mechanical ventilation plus ≥2 acute organ failures, randomised within ~18 hours of ICU admission), representative of severely ill ICU patients rather than broader, lower-risk ICU cohorts.
  • Important exclusions: Findings should not be extrapolated to patients with major burns (>40% TBSA), severe traumatic brain injury, status epilepticus requiring anticonvulsants, Child–Pugh C liver disease, or those receiving chronic haemodialysis/very low creatinine clearance without dialysis.
  • Applicability by intervention “formulation”: The tested strategy was combined enteral + parenteral supplementation at relatively high doses; applicability to lower-dose parenteral glutamine used only as a component of parenteral nutrition (without concurrent enteral glutamine) is uncertain.
  • Healthcare-system transferability: Multinational ICU participation supports transferability to similar high-resource ICU systems; external validity is weaker for resource-limited settings where nutrition delivery, monitoring, and renal replacement availability differ.

Conclusion on External Validity: External validity is moderate: findings generalise best to early, severe, ventilated multi-organ failure in modern ICUs, but are less directly applicable to stable ICU patients, different dosing/route strategies, or settings with substantially different nutrition support infrastructure.

Strengths & Limitations

  • Strengths: Large, multicentre, international ICU trial; rigorous double-blind placebo-controlled design; factorial efficiency addressing two common pharmaconutrition strategies; early enrolment and meaningful exposure separation; patient-centred endpoints including 6-month mortality.
  • Limitations: Tested a specific high-dose, combined enteral+parenteral strategy (not all glutamine/antioxidant use patterns); small proportion missing for 28-day vital status (5/1,223); factorial interpretation depends on the assumption of no clinically meaningful interaction; limited routine biomarker targeting (baseline deficiency/“need” not used to select patients or titrate dose).

Interpretation & Why It Matters

  • Clinical meaning
    • REDOXS supports avoiding early high-dose glutamine (particularly combined enteral + parenteral dosing) in mechanically ventilated adults with acute multi-organ failure, given consistent signals of harm (higher in-hospital and 6-month mortality) and biochemical toxicity (high urea).
    • The antioxidant strategy tested (enteral cocktail + parenteral selenium) did not improve mortality outcomes, arguing against routine use of this approach as a mortality-modifying therapy in this population.
    • The trial materially shifted the evidentiary standard for pharmaconutrition in critical illness: plausible mechanisms and small-trial signals are insufficient when large blinded trials show harm in the target phenotype.

Controversies & Subsequent Evidence

  • Patient selection and renal dysfunction: A post hoc analysis comparing the individual factorial groups versus placebo found excess mortality in glutamine-containing arms and reported the greatest harm among patients with baseline renal dysfunction, with no subgroup showing a mortality reduction.1
  • Signal replication in related “pharmaconutrition” strategies: The MetaPlus randomised clinical trial (immune-modulating enteral formula including glutamine and other nutrients) reported increased 6-month mortality, reinforcing concerns that complex nutrient “cocktails” can be harmful in unselected critically ill populations.3
  • How to reconcile with earlier trials: Commentary focusing on REDOXS and MetaPlus highlighted that dose, route (combined enteral+parenteral), and early administration in severe organ failure may be key determinants of harm, challenging extrapolation from earlier, lower-dose or parenteral-nutrition–anchored glutamine trials.2
  • Context from other factorial ICU supplementation trials: The SIGNET trial (glutamine and selenium) is frequently cited as showing no convincing mortality benefit from these strategies in critically ill adults, in contrast to early small studies.4
  • Antioxidant/selenium strategy uncertainty: A large randomised trial of sodium selenite with procalcitonin-guided therapy in severe sepsis/septic shock did not demonstrate a mortality benefit attributable to selenium, adding to uncertainty about selenium as a mortality-modifying intervention in contemporary sepsis care.5
  • Meta-analytic synthesis after REDOXS: Meta-analyses incorporating REDOXS reported attenuation or loss of prior apparent benefit of glutamine supplementation and highlighted heterogeneity by population and intervention strategy, consistent with the need to avoid routine glutamine in severe multi-organ failure.6
  • Guideline evolution: Subsequent ICU nutrition guidelines recommend against routine glutamine supplementation in critically ill adults (particularly in multi-organ failure and/or renal dysfunction) and do not endorse routine antioxidant “cocktail” strategies as mortality-reducing therapy.7891011

Summary

  • REDOXS was a large, blinded, 2×2 factorial ICU trial testing early combined enteral+parenteral glutamine and an antioxidant strategy in ventilated adults with acute multi-organ failure.
  • Glutamine was associated with higher in-hospital mortality (37.2% vs 31.0%; adjusted OR 1.46; 95% CI 1.07 to 2.00; P=0.02) and higher 6-month mortality (43.7% vs 37.2%; adjusted HR 1.18; 95% CI 1.02 to 1.36; P=0.02).
  • The 28-day mortality result for glutamine (32.4% vs 27.3%; adjusted OR 1.28; 95% CI 1.00 to 1.64; P=0.05) did not meet the prespecified final alpha (P<0.044), but the overall pattern favoured harm rather than benefit.
  • The antioxidant strategy did not reduce mortality at 28 days, in hospital, or at 6 months.
  • A clear biochemical safety signal occurred with glutamine: high urea (>50 mmol/L) 13.4% versus 4.0% (P<0.001).

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Several additional glutamine/selenium supplementation trials and systematic reviews exist; only references with DOI-verified links available from the sources used in this summary are listed here.

Overall Takeaway

REDOXS is a landmark ICU nutrition trial because it overturned the prevailing assumption that glutamine and antioxidant supplementation are broadly beneficial in severe critical illness. In ventilated adults with early multi-organ failure, early high-dose combined-route glutamine was associated with harm (higher in-hospital and 6-month mortality) and a biochemical toxicity signal, driving a sustained shift away from routine pharmaconutrition and towards phenotype- and strategy-specific evidence.

Overall Summary

  • In early, severe, ventilated multi-organ failure, high-dose glutamine (enteral + parenteral) increased longer-term mortality and produced a marked high-urea signal.
  • The antioxidant strategy tested did not reduce mortality.
  • Practice and guidelines subsequently moved away from routine glutamine supplementation in critical illness.

Bibliography

  • 1Heyland DK, Elke G, Cook D, et al. Glutamine and antioxidants in the critically ill patient: a post hoc analysis of a large-scale randomized trial. JPEN J Parenter Enteral Nutr. 2015;39(4):401-409. Link
  • 2van Zanten ARH. Consequences of the REDOXS and METAPLUS trials. JPEN J Parenter Enteral Nutr. 2016. Link
  • 3van Zanten ARH, et al. Effect of high-protein enteral nutrition enriched with immune-modulating nutrients on outcomes in critically ill patients: the MetaPlus randomized clinical trial. JAMA. 2014;312(5):514-524. Link
  • 4Andrews PJ, et al. Glutamine and selenium for critically ill patients: a randomised controlled trial. BMJ. 2011;342:d1542. Link
  • 5Bloos F, et al. Effect of sodium selenite administration and procalcitonin-guided therapy on mortality in patients with severe sepsis or septic shock: a randomized clinical trial. JAMA Intern Med. 2016;176(9):1266-1276. Link
  • 6Chen QH, Yang Y, He HL, et al. The effect of glutamine therapy on outcomes in critically ill patients: a meta-analysis of randomized controlled trials. Crit Care. 2014;18:R8. Link
  • 7McClave SA, Taylor BE, Martindale RG, et al. Guidelines for the provision and assessment of nutrition support therapy in the adult critically ill patient. JPEN J Parenter Enteral Nutr. 2016;40(2):159-211. Link
  • 8Rice TW, et al. Guidelines for the provision of nutrition support therapy in the adult critically ill patient: 2022 update. JPEN J Parenter Enteral Nutr. 2022. Link
  • 9Reignier J, et al. Clinical practice guidelines for early enteral nutrition management in critically ill adults. Intensive Care Med. 2017;43(3):380-398. Link
  • 10Singer P, Blaser AR, Berger MM, et al. ESPEN guideline on clinical nutrition in the intensive care unit. Clin Nutr. 2019;38(1):48-79. Link
  • 11Reignier J, et al. Expert consensus-based clinical practice guidelines for nutritional support in critically ill adults and children (excluding neonates and burn patients). Ann Intensive Care. 2025. Link