Skip to main content

Foundational Trials

6S

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Hydroxyethyl Starch 130/0.42 versus Ringer’s Acetate in Severe Sepsis
  • Acronym: 6S
  • Year: 2012
  • Journal published in: The New England Journal of Medicine
  • Citation: Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus Ringer’s acetate in severe sepsis. N Engl J Med. 2012;367:124-34.

Context & Rationale

  • Background
    • Hydroxyethyl starch (HES) solutions were widely used for intravascular volume expansion on the premise of greater volume effect vs crystalloids.
    • Earlier generation starches were linked to coagulopathy and kidney injury; “modern” tetrastarches (HES 130/0.4–0.42) were marketed as safer, but high-quality sepsis-specific outcomes data were limited.
    • Sepsis carries high AKI risk; any nephrotoxic signal from resuscitation fluids would have major patient-centred and health-system consequences.
  • Research Question/Hypothesis
    • In ICU patients with severe sepsis, does resuscitation with 6% HES 130/0.42 (vs Ringer’s acetate) increase the risk of death or dialysis dependence at 90 days?
  • Why This Matters
    • Addressed a ubiquitous ICU exposure with plausible mechanistic toxicity (renal tubular injury, tissue storage, coagulopathy) using hard outcomes rather than surrogates.
    • Directly informed fluid choice policies and subsequent regulatory and guideline restrictions on starches in critical illness.

Design & Methods

  • Research Question: Does HES 130/0.42, compared with Ringer’s acetate, increase the composite of death or end-stage kidney failure (dialysis dependence) at 90 days in severe sepsis?
  • Study Type: Multicentre, randomised, blinded, parallel-group trial in Scandinavian ICUs; pragmatic management except for trial-fluid assignment.
  • Population:
    • Setting: Adult ICUs in Denmark, Norway, Sweden, and Finland.
    • Key inclusion: Severe sepsis requiring fluid resuscitation in ICU.
    • Key exclusions: Not fully enumerated in the main paper text excerpt available here; protocolised enrolment and fluid caps applied.
  • Intervention:
    • Study fluid: 6% HES 130/0.42 in Ringer’s acetate (Tetraspan).
    • Administration rule: Trial fluid used whenever intravascular volume expansion was needed in ICU for a maximum of 90 days.
    • Dose cap: Maximum daily dose 33 mL/kg ideal body weight; if exceeded, unmasked Ringer’s acetate was used for further resuscitation.
    • Avoidance: Other synthetic colloids prohibited; albumin permitted only if trial fluid and crystalloids were insufficient and limited to 200 mL/day.
  • Comparison:
    • Study fluid: Ringer’s acetate (crystalloid).
    • Same administration rule and caps: Used for intravascular volume expansion as needed; same restriction on synthetic colloids and albumin.
  • Blinding: Blinded group assignments with concealed allocation; trial fluids provided in masked fashion; open-label Ringer’s acetate permitted if dose cap exceeded.
  • Statistics: A total of 800 patients were required to show an absolute between-group difference of 10 percentage points in the primary outcome with 80% power at the 5% significance level; primary analyses performed in the intention-to-treat population.
  • Follow-Up Period: 90 days for primary endpoint; additional 28-day outcomes and ICU-period secondary outcomes reported.

Key Results

This trial was not stopped early. Completed recruitment (n=798) with 90-day follow-up.

Outcome HES 130/0.42 Ringer’s acetate Effect p value / 95% CI Notes
Primary composite: dead or dependent on dialysis at day 90 202/398 (51%) 173/400 (43%) RR 1.17 95% CI 1.01 to 1.36; P=0.03 Composite driven by mortality; dialysis dependence at day 90 was 1 patient in each group.
Death at day 90 201/398 (51%) 172/400 (43%) RR 1.17 95% CI 1.01 to 1.36; P=0.03 Mortality difference mirrors primary composite.
Use of renal-replacement therapy 87/398 (22%) 65/400 (16%) RR 1.35 95% CI 1.01 to 1.80; P=0.04 Renal toxicity signal consistent with mechanistic concerns for starches.
Severe bleeding 38/398 (10%) 25/400 (6%) RR 1.52 95% CI 0.94 to 2.48; P=0.09 Directionally higher bleeding with HES; CI includes no difference.
Alive and out of hospital (mean % of days in 90-day follow-up) 29 34 — P=0.048 Functional/resource-relevant endpoint favoured Ringer’s acetate.
  • HES 130/0.42 increased the primary composite at 90 days (51% vs 43%; RR 1.17; 95% CI 1.01 to 1.36; P=0.03), with the difference almost entirely explained by mortality.
  • Renal-replacement therapy was more frequent with HES (22% vs 16%; RR 1.35; 95% CI 1.01 to 1.80; P=0.04), supporting a clinically meaningful nephrotoxicity signal.
  • Trial-fluid volumes were similar between groups (day 1: 1500±1208 vs 1484±1223 mL), strengthening causal attribution to fluid type rather than gross fluid dose.

Internal Validity

  • Randomisation and Allocation:
    • Concealed allocation with blinded trial fluids; multicentre design reduces centre-specific bias.
    • Randomisation process described with adequate safeguards; baseline characteristics were well balanced.
  • Dropout / Exclusions:
    • Randomised n=798 (398 HES; 400 Ringer’s); primary endpoint available at day 90 in ITT analyses.
    • Missing data noted for selected secondary outcomes (e.g., doubling creatinine missing for 38 vs 34 patients), but primary mortality/dialysis dependence robust.
  • Performance/Detection Bias:
    • Blinding reduces differential co-intervention; use of open-label Ringer’s when cap exceeded could dilute separation (bias towards the null).
  • Protocol Adherence:
    • Study-fluid administered when volume expansion needed; daily cap 33 mL/kg ideal body weight enforced with switch to open-label crystalloid.
    • Other synthetic colloids prohibited; albumin tightly restricted (≤200 mL/day) to minimise contamination.
  • Baseline Characteristics:
    • Patients were severely ill (SOFA at randomisation median 6 (IQR 4–9) vs 6 (4–9)); high vasopressor and ventilation use typical of ICU severe sepsis.
  • Heterogeneity:
    • Severe sepsis population heterogeneous by infection source and organ dysfunction; however, this improves pragmatic inference for ICU sepsis.
  • Timing:
    • Intervention delivered during ICU course when volume expansion indicated, for up to 90 days; early exposure likely dominated as most trial fluid was given in the first days.
  • Dose:
    • Daily cap limited extreme dosing while still reflecting common clinical use; adverse signals persisted within this “controlled” exposure window.
  • Separation of the Variable of Interest:
    • Trial fluid day 1: 1500±1208 mL (HES) vs 1484±1223 mL (Ringer’s).
    • Trial fluid day 2: 999±963 vs 1035±1007 mL; day 3: 595±855 vs 554±730 mL.
    • Open-label fluid day 1: 1658±1764 vs 1563±1670 mL; similar thereafter.
    • Blood product exposure differed: any blood products 66% vs 59% (RR 1.20; 95% CI 1.02 to 1.42; P=0.03); packed red cells 43% vs 36% (RR 1.28; 95% CI 1.04 to 1.58; P=0.02).
  • Outcome Assessment:
    • Primary endpoint (death/dialysis dependence) objective and clinically meaningful; renal-replacement therapy use also clinically salient though practice-influenced.
  • Statistical Rigor:
    • Pre-specified clinically important effect size (10% absolute) and adequate sample achieved; primary outcome statistically significant with narrow CI excluding unity by a small margin.

Conclusion on Internal Validity: Strong. Concealed, blinded allocation, robust primary endpoint, balanced baseline features, and similar trial-fluid volumes support a credible causal inference that HES 130/0.42 increased harm in severe sepsis.

External Validity

  • Population Representativeness:
    • Severe sepsis ICU population across multiple Scandinavian centres; illness severity and supportive therapy needs resemble ICU sepsis cohorts internationally.
    • Exclusions and albumin restrictions may differ from some practices, but the comparison remains clinically relevant (colloid vs balanced crystalloid).
  • Applicability:
    • Findings apply most directly to ICU patients with severe sepsis (and, by extension, septic shock populations), where AKI risk is high.
    • Results discourage HES use in critically ill/inflammatory states even when “modern” tetrastarch formulations and dose caps are used.

Conclusion on External Validity: High for ICU severe sepsis practice. While fluid strategies and regulatory environments vary, the harm signal is clinically plausible and consistent enough to generalise across many high-income ICU contexts.

Strengths & Limitations

  • Strengths:
    • Robust design: multicentre, randomised, blinded, concealed allocation with pragmatic management otherwise preserved.
    • Patient-centred primary endpoint (death or dialysis dependence at 90 days).
    • Clinically meaningful renal and functional secondary outcomes; strong separation by fluid type with similar volumes.
  • Limitations:
    • Open-label crystalloid permitted after dose cap, potentially diluting differences (bias towards null rather than towards harm).
    • Renal-replacement therapy use is partly practice-dependent; nevertheless, mortality and composite endpoint were objective.
    • Protocol restrictions on albumin (≤200 mL/day) may not mirror all ICUs, but minimised confounding.

Interpretation & Why It Matters

  • Clinical practice
    In severe sepsis, HES 130/0.42 should be avoided for resuscitation because it increased 90-day mortality and renal-replacement therapy compared with a balanced crystalloid.
  • Mechanistic coherence
    Renal injury signal (RRT and creatinine outcomes) aligns with biological plausibility of starch tissue storage and tubular toxicity; bleeding signal directionally consistent with starch-associated coagulopathy.
  • Policy impact
    6S became central evidence for guideline recommendations against HES in sepsis and for national procurement/ICU formulary changes.

Controversies & Subsequent Evidence

  • Correspondence raised questions about whether protocolised haemodynamic goals (e.g., cardiac output monitoring) might have altered fluid requirements, and whether differences in co-interventions (including transfusion) could influence outcomes12.
  • Authors’ replies emphasised that fluid volumes were similar and that the intervention contrast was primarily fluid type; they also noted that selected haemodynamic variables (e.g., CVP) were not routinely captured, limiting post hoc mechanistic adjudication12.
  • Subsequent trials and meta-analyses in critical illness broadly corroborated kidney injury risk with HES and informed strong guideline recommendations against HES in sepsis and critically ill adults (see Further Reading; current sepsis guidance discourages starch use)3.

Summary

  • Multicentre, randomised, blinded sepsis trial comparing HES 130/0.42 with Ringer’s acetate for ICU volume expansion.
  • Primary endpoint (death or dialysis dependence at 90 days) occurred more often with HES: 51% vs 43% (RR 1.17; 95% CI 1.01 to 1.36; P=0.03).
  • Renal-replacement therapy was higher with HES: 22% vs 16% (RR 1.35; 95% CI 1.01 to 1.80; P=0.04).
  • Trial-fluid volumes were similar between groups, supporting toxicity of starch rather than dose differences as a contributor.
  • 6S is foundational evidence underpinning modern avoidance of HES in sepsis and critical illness.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • For several key follow-up trials and guidance documents, PubMed navigation links are provided where DOI landing pages were not available from the provided source material.

Overall Takeaway

6S is a landmark because it definitively linked a widely used “modern” tetrastarch (HES 130/0.42) to worse patient-centred outcomes in severe sepsis: higher 90-day mortality and greater need for renal-replacement therapy versus a balanced crystalloid. By combining strong internal validity with clinically meaningful endpoints and similar administered volumes, it shifted international practice away from starches and continues to underpin guideline recommendations against HES in critical illness.

Overall Summary

  • In severe sepsis, HES 130/0.42 increased death or dialysis dependence at 90 days and increased renal-replacement therapy—despite similar fluid volumes—supporting a true toxicity signal.

Bibliography