Context & Rationale

• Background

  • By the late 2000s, hydroxyethyl starch (HES) solutions—particularly so‑called “tetrastarches” (e.g., 130/0.4–0.42)—were widely used for intravascular volume expansion in the ICU based on physiological arguments about oncotic properties and putative fluid‑sparing effects.
  • However, earlier HES preparations had been linked to renal toxicity and bleeding, and signals of harm were emerging even with newer formulations.
  • The VISEP trial (severe sepsis; pentastarch 200/0.5 vs Ringer’s lactate) reported increased renal failure and RRT with HES, and the 6S trial (severe sepsis; HES 130/0.42 vs Ringer’s acetate) found increased 90‑day mortality and RRT with HES.
  • Robust data in a broad, heterogeneous ICU population remained limited.
• Why This Matters

  • Fluid resuscitation is ubiquitous in critical care; the choice of fluid—colloid vs crystalloid—has system‑wide implications for patient outcomes and cost.
  • CHEST was designed to provide definitive clinical effectiveness and safety data for HES in a general ICU population (beyond sepsis alone), to guide practice and policy.
• Trial Hypothesis

  Does resuscitation with 6% HES (130/0.4) in 0.9% saline, compared with 0.9% saline alone, change 90‑day all‑cause mortality in a general adult ICU population, and what are the renal and other safety effects?

Design & Methodology

Trial Design

• Design

  Investigator‑initiated, multicentre (32 ICUs in Australia/New Zealand), double‑blind, concealed, randomised controlled trial; stratified by site and trauma admission.

• Setting

  Adult ICUs in Australia & New Zealand

Population

• Inclusion Criteria

  • Adult ICU patients (≥18 y) requiring fluid resuscitation to expand/maintain intravascular volume (beyond maintenance/nutrition/replacement fluids),
  • with at least one supporting sign (e.g., HR>90 bpm; SBP<100 mmHg/MAP<75 mmHg or ≥40 mmHg drop; CVP<10 mmHg; PAWP<12 mmHg; pulse pressure variation >5 mmHg; capillary refill >1 s; urine output <0.5 mL/kg/h for 1 h);
  • clinician judged HES and saline both appropriate options;
  • consent (or deferred consent) per ethics approval
• Exclusion Criteria

  Age <18 y; prior HES allergy; serum chloride >130 mmol/L; possible pregnancy/breastfeeding; >1000 mL HES given outside ICU within 24 h pre‑randomisation; postoperative cardiac surgery/burns/liver transplant; death deemed imminent/inevitable or expected survival <90 days; treatment limitation; prior enrolment; prior resuscitation already prescribed within study ICU; transfer from another ICU after receiving resuscitation there

Intervention

• 6% HES (130/0.4)

  6% HES (130/0.4) in 0.9% sodium chloride (Voluven®) in indistinguishable 500 mL Freeflex® bags, administered for fluid resuscitation as clinically indicated.

• Dose

  Up to a maximum of 50 mL/kg per day could be used; beyond this limit, open‑label saline could be used within the 24‑h period

• End Point

  Study fluid continued until ICU discharge, death, or day 90.

Control

• 0.9% sodium chloride

  0.9% sodium chloride in identical‑appearing bags, used in the same manner and timeframe.

• Pragmatic

  Other care was per clinician discretion.

Statistical Plan

• Power Calculation

  • 7000 patients were required to detect a 3.5% absolute difference in 90‑day mortality (assuming a control group mortality of ~26% at 90 days), with 90% power and two‑sided α=0.05.
  • Using group‑sequential monitoring (two interim looks at ~2000 and ~4000 with 90‑day follow‑up).
  • The trial was also powered (90%, α=0.05) to detect a 1.5% relative risk increase in acute kidney injury, from an anticipated 5% in the control group.
• Analysis

  • Intention‑to‑treat.
  • Two‑sided tests at α=0.05.
  • Six prespecified subgroups (trauma with/without TBI; severe sepsis; pre‑existing renal impairment without oliguria/anuria; APACHE II > 25; prior HES before randomisation).
  • No multiplicity adjustment.

Other

• Blinding

  • Double‑blind (patients, clinicians, outcome assessors).
  • Allocation concealed via secure web randomisation with minimisation.
  • Packs indistinguishable.
• Follow Up

  • Daily in ICU to day 7 (renal injury metrics) and day 28 (organ failure metrics).
  • Vital status to day 90.
  • 6‑month quality‑of‑life and additional follow‑up in TBI subgroup (GOS), with health‑economic data in a subset.

Key Results

• Early Stopping

Notes

Internal Validity

• Randomisation & Allocation

  • Web‑based concealed randomisation with minimisation, stratified by site and trauma; identical packaging ensured allocation concealment.
  • Low risk of selection bias.
• Drop Outs & Exclusions

  • Of 7000 randomised, 6651 contributed to the primary analysis (3315 HES; 3336 saline), primarily because consent to use data could not be obtained in some jurisdictions.
  • Vital status ascertainment to day 90 was excellent among analysed participants.
  • This reduces precision but is unlikely to bias the treatment effect materially given blinding and balanced numbers.
• Performance/Detection Bias

  • Double‑blind design.
  • Outcomes (mortality, RRT use) are objective.
  • Some secondary organ‑failure outcomes rely on SOFA‑derived scores but were assessed under blinding.
• Protocol Adherence

  • Protocol and statistical analysis plan were prespecified and published.
  • DSMB oversight with two interim looks.
  • Alpha‑spending specified.
• Outcome Assessment

  • Primary outcome objective
  • Renal outcomes included both RRT (objective) and RIFLE categories (composite of creatinine and urine output)
• Statistical Rigor

  • ITT analysis.
  • Prespecified SAP.
  • Predefined subgroup/tertiary outcomes.
  • Multiplicity not adjusted (hierarchy declared).
  • Power target met.
• Separation of the Variable of Interest

  • Study‑fluid exposure differed:
    • 526 ± 425 mL/day (HES) vs 616 ± 488 mL/day (saline) over days 1–4
    • blood product exposure greater with HES (78 ± 250 mL vs 60 ± 190 mL)
• Key Delivery Aspects

  • Choice of rates/end‑points of resuscitation and indications for RRT were not protocolised (pragmatic design).
  • This may increase variability but also enhances real‑world relevance.
• Baseline Characteristics

  Groups were well balanced; mean age ~63 y; 60.4% male; median APACHE II 17 [IQR 12–23]; APACHE II ≥ 25 in 18.2%; severe sepsis 28.8%; trauma 7.9%; 64.5% on mechanical ventilation; 45.8% receiving vasopressors; ∼15% had received HES before randomisation; median enrolment ~11 h after ICU admission.

• Heterogeneity

  Broad ICU mix with substantial postoperative admissions; prespecified subgroup analyses showed no differential treatment effect.

• Timing

  • Enrolment typically within ~11 h of ICU admission.
  • Study fluid administered over the first days of ICU stay—appropriate window for resuscitation fluids.
• Dose

  • Maximum permitted dose 50 mL/kg/day.
  • actual mean study‑fluid volumes were modest, indicating conservative real‑world dosing.
• Crossover

  • Not applicable—fluids were protocol‑assigned.
  • Outside‑protocol use (e.g., blood products) was allowed.
• Adjunctive Therapies

  HES recipients received slightly more blood products early (mean +18 mL over 4 days).

• Outcome Assessment

  • Primary outcome objective.
  • Renal outcomes included both RRT (objective) and RIFLE categories (composite of creatinine and urine output).

• Conclusion

  • Overall, internal validity is strong: concealed randomisation, double‑blinding, objective outcomes and a prespecified SAP.
  • Pragmatic elements (unprotocolised RRT triggers) and consent‑related exclusions slightly temper—but do not undermine—the validity.

External Validity

• Population Representativeness

  • Broad adult ICU cohort across 32 Australasian ICUs
  • Includes medical, surgical (notably elective), and trauma patients; excluded burns, liver transplant, and immediate post‑cardiac surgery.
• Applicability

  • High generalisability to mixed ICUs using saline as standard crystalloid.
  • Applicability to settings with different practice patterns (e.g., early goal‑directed resuscitation pathways or balanced crystalloids as default) remains reasonable for the comparative safety signals.

• Conclusion

  • Findings are broadly generalisable to contemporary mixed ICUs.
  • Extrapolation to excluded populations (burns, liver transplant, immediate post‑cardiac surgery) should be cautious.

Strengths & Limitations

• Strengths

  • Large sample
  • Multicentre
  • Rigorous blinding and allocation concealment
  • Prespecified protocol/SAP and DSMB
  • Objective primary outcome
  • Comprehensive renal/organ‑failure assessments
  • Health‑economic and functional follow‑up planned
• Limitations

  • Consent constraints reduced numbers with primary outcome data
  • Unprotocolised triggers for RRT and transfusion; composite RIFLE metrics led to paradoxical patterns (lower “Injury” yet higher RRT)
  • Relatively low illness severity and significant elective surgical admissions may have diluted treatment effects

Interpretation / Why This Matters

• No Benefit with HES

  CHEST shows that, in an unselected ICU population, HES 130/0.4 does not improve survival and is associated with more RRT and adverse events, despite small fluid‑sparing and slightly less cardiovascular SOFA failure.

• Practice Changing

  In light of concordant evidence of harm in sepsis (VISEP, 6S) and meta‑analyses linking HES to increased AKI/mortality, CHEST helped shift practice away from HES in the ICU.

Controversies & Subsequent Evidence

• Renal Outcomes Paradox

  • CHEST observed higher RRT use with HES but a lower incidence of RIFLE‑Injury.
  • Post‑hoc analyses and commentary attribute this to discordance between the creatinine and urine‑output components (HES may preserve urine output transiently while creatinine rises), and to unprotocolised RRT decisions.
  • The net signal—more RRT—favours avoiding HES.
• Transfusion and Hepatic Effects

  HES was associated with slightly greater blood product exposure and more new hepatic failure, adding to safety concerns

• Editorial and correspondence

  • Contemporaneous commentary highlighted consistency with severe‑sepsis data and questioned any clinical justification for HES given absent mortality benefit and renal signals.
  • Letters in 2013 debated interpretation; authors’ reply and an independent reanalysis (2017) left the principal conclusions unchanged.
• Correction

  • In 2016, a journal correction updated the p‑value for the adverse‑event comparison (treatment‑related AEs) to p=0.006.
  • Other conclusions were unaffected.
• Long‑Term Outcomes

  Linked long‑term follow‑up and cost‑effectiveness analysis from a CHEST cohort found no patient‑centred benefit with HES and no economic advantage.

Summary

• Mortality

  In >6600 analysed ICU patients, 90‑day mortality was similar: 18.0% (HES) vs 17.0% (saline); RR 1.06 (95% CI 0.96–1.18)

• RRT Use

  RRT use was higher with HES: 7.0% vs 5.8%; RR 1.21 (95% CI 1.00–1.45).

• Adverse Events

  HES produced small fluid‑sparing (−90 mL/day over days 1–4) but more blood product exposure and more adverse events.

• No Benefit

  • No convincing benefit in organ failure or long‑term outcomes to offset safety signals.
  • Findings align with severe‑sepsis trials and meta‑analyses indicating renal harm.

Further Reading

Other Trials

• 6S

  Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus Ringer’s acetate in severe sepsis (6S). N Engl J Med. 2012;367:124–134

• VISEP

  Brunkhorst. Intensive insulin therapy and pentastarch resuscitation in severe sepsis (VISEP). N Eng J Med 2008;358:125-139

• CRYSTMAS

  Guidet. Assessment of hemodynamic efficacy and safety of 6% hydroxyethylstarch 130/0.4 versus 0.9% NaCl fluid replacement in patients with severe sepsis: the CRYSTMAS study. Crit Care 2012;16:R94

Systematic Review & Meta Analysis

• 2013

  Zarychanski R, Abou‑Setta AM, Turgeon AF, et al. Association of HES with mortality and AKI: systematic review and meta‑analysis. JAMA. 2013;309:678–688

• 2013

  Perner A, Haase N, Wetterslev J, et al. Hydroxyethyl starch 130/0.38–0.45 vs crystalloid/albumin in sepsis: systematic review with meta‑analysis and trial sequential analysis. BMJ. 2013;346:f839

• 2018

  Perel P, Roberts I, et al. Colloids versus crystalloids for fluid resuscitation in critically ill people. Cochrane Database Syst Rev. 2013;Issue 2:CD000567. doi:10.1002/14651858.CD000567.pub6; updated 2018:

• 2015

  Rochwerg B, Alhazzani W, et al. Fluid resuscitation in sepsis: network meta‑analysis. Intensive Care Med. 2015;41:1519–1531

Observational Studies

• 1

  Hjortrup PB, Haase N, et al. The effect of 6% HES (130/0.42) on acute kidney injury in critically ill adults. Anaesthesia. 2018;73:596–605

• 2

  Sakr. Effects of hydroxyethyl starch administration on renal function in critically ill patients. Br J Anaesth 2007;98(2):216-24

• 3

  Miyao. Renal Morbidity of 6% Hydroxyethyl Starch 130/0.4 in 9000 Propensity Score Matched Pairs of Surgical Patients. Anesth Analg 2019;130(6):1618–1627

Notes

Guidelines

• ESICM

  European Society of Intensive Care Medicine clinical practice guideline on fluid therapy in adult critically ill patients. Part 1: the choice of resuscitation fluids. Intensive Care Med 2024;epublished May 21st

• Surviving Sepsis Campaign

  Evans. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med 2021;epublished October 2nd

• French

  Joannes-Boyau. Guidelines for the choice of intravenous fluids for vascular filling in critically ill patients. Anaesth Crit Care Pain Med 2022;41(3):101058

• Perioperative

  Perioperative fluid management: evidence-based consensus recommendations from the international multidisciplinary PeriOperative Quality Initiative. Br J Anaesth 2024;133(6):1263-1275