Skip to main content

Foundational Trials

VASST

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Vasopressin versus norepinephrine infusion in patients with septic shock
  • Acronym: VASST (Vasopressin and Septic Shock Trial)
  • Year: 2008
  • Journal published in: New England Journal of Medicine
  • Citation: Russell JA, Walley KR, Singer J, Gordon AC, Hebert PC, Cooper DJ, et al.; VASST Investigators. Vasopressin versus norepinephrine infusion in patients with septic shock. N Engl J Med. 2008;358(9):877-887.

Context & Rationale

  • Background
    Septic shock is characterised by vasoplegia and impaired vascular responsiveness, commonly treated with catecholamine vasopressors (particularly norepinephrine).
  • Research Question/Hypothesis
    Whether low-dose vasopressin (a non-catecholamine vasopressor) would reduce all-cause mortality compared with norepinephrine when used as a blinded study vasopressor in established septic shock.
  • Why This Matters
    If vasopressin improved survival (or reduced organ dysfunction) without excess ischaemic toxicity, it would support a major shift in first-line or early vasopressor strategy in septic shock.

Design & Methods

  • Research Question: In adults with septic shock receiving vasopressor support, does low-dose vasopressin (0.01–0.03 U/min) compared with norepinephrine (5–15 μg/min) reduce all-cause mortality?
  • Study Type: Multicentre, randomised, double-blind, parallel-group controlled trial in ICUs (27 centres across Canada, Australia, and the USA), stratified by centre and severity-of-shock stratum at randomisation.
  • Population:
    • Adults with septic shock, enrolled within 24 hours of meeting eligibility criteria.
    • Required ongoing vasopressor therapy with norepinephrine ≥5 μg/min (or equivalent) for at least 6 hours (with a “fast-track” option for higher-dose vasopressor requirements).
    • Randomisation stratified by severity-of-shock at randomisation: “less severe” (5–14 μg norepinephrine/min) vs “more severe” (≥15 μg norepinephrine/min).
    • Key exclusions included acute coronary syndromes (including ST-elevation MI), New York Heart Association class III/IV heart failure, mesenteric ischaemia, Raynaud’s phenomenon, pregnancy, previous vasopressin infusion during current admission, and >24 hours elapsed since meeting eligibility criteria.
  • Intervention:
    • Blinded vasopressin infusion 0.01–0.03 U/min, delivered via a constant study-drug infusion rate of 15 mL/hour.
    • Open-label vasopressors (excluding vasopressin) titrated to achieve a target mean arterial pressure of 65–75 mmHg.
    • Weaning strategy: open-label vasopressors reduced first to low doses before tapering the blinded study drug.
  • Comparison:
    • Blinded norepinephrine infusion 5–15 μg/min, delivered via an identical constant study-drug infusion rate of 15 mL/hour.
    • Open-label vasopressors (excluding vasopressin) titrated similarly to achieve mean arterial pressure 65–75 mmHg.
    • Open-label vasopressin was prohibited (crossover not permitted except in exceptional circumstances).
  • Blinding: Double-blind; pharmacists prepared indistinguishable study infusions, and clinicians/patients/outcome assessors were intended to remain blinded to group allocation.
  • Statistics: A total of 776 patients were required to detect an absolute mortality difference of 10% (from an expected 60%) with 80% power at the 5% significance level; the planned sample size was increased to account for the primary analysis being restricted to patients who underwent randomisation and infusion; analyses included (i) a primary analysis based on randomisation and infusion and (ii) an analysis based on randomisation; prespecified interim analyses used an O’Brien–Fleming approach.
  • Follow-Up Period: Mortality assessed to 28 days (primary end point) and 90 days (key secondary end point).

Key Results

This trial was not stopped early. Prespecified interim analyses were performed; the trial continued to completion.

Outcome Vasopressin Norepinephrine Effect p value / 95% CI Notes
All-cause mortality at 28 days (primary; analysis based on randomisation and infusion) 140/396 (35.4%) 150/382 (39.3%) RR 0.90 95% CI 0.75 to 1.08; P=0.26 (ARR 3.9%; 95% CI −2.9 to 10.7) Adjusted OR 0.88; 95% CI 0.62 to 1.26; P=0.48
All-cause mortality at 90 days (analysis based on randomisation and infusion) 172/392 (43.9%) 188/379 (49.6%) RR 0.88 95% CI 0.76 to 1.03; P=0.11 (ARR 5.7%; 95% CI −1.3 to 12.8) Adjusted OR 0.81; 95% CI 0.57 to 1.16; P=0.26
All-cause mortality at 28 days (analysis based on randomisation) 144/405 (35.6%) 154/395 (39.0%) RR 0.91 95% CI 0.76 to 1.09; P=0.33 (ARR 3.4%; 95% CI −3.4 to 10.2) Adjusted analysis not reported for this cohort
All-cause mortality at 90 days (analysis based on randomisation) 179/405 (44.2%) 195/394 (49.5%) RR 0.89 95% CI 0.77 to 1.04; P=0.15 (ARR 5.4%; 95% CI −2.0 to 12.7) Adjusted analysis not reported for this cohort
28-day mortality (less severe shock stratum: 5–14 μg norepinephrine/min at randomisation) 52/196 (26.5%) 65/182 (35.7%) RR 0.74 95% CI 0.55 to 1.01; P=0.05 (ARR 9.2%; 95% CI −0.1 to 18.5) Prespecified stratum analysis
90-day mortality (less severe shock stratum: 5–14 μg norepinephrine/min at randomisation) 69/193 (35.8%) 83/180 (46.1%) RR 0.78 95% CI 0.61 to 0.99; P=0.04 (ARR 10.4%; 95% CI 0.4 to 20.4) Prespecified stratum analysis
Serious adverse events (≥1 event) 41/396 (10.3%) 40/382 (10.5%) Not reported P=1.00 Composite of prespecified serious adverse events
Digital ischaemia (serious adverse event) 8/396 (2.0%) 2/382 (0.5%) Not reported P=0.11 Low absolute event rates
Hyponatraemia (serious adverse event) 5/396 (1.3%) 0/382 (0.0%) Not reported P=0.06 Event definition prespecified
  • Overall 28-day mortality was 35.4% with vasopressin vs 39.3% with norepinephrine (RR 0.90; 95% CI 0.75 to 1.08; P=0.26; primary analysis).
  • In the prespecified “less severe shock” stratum, 90-day mortality was 35.8% with vasopressin vs 46.1% with norepinephrine (RR 0.78; 95% CI 0.61 to 0.99; P=0.04).
  • Serious adverse events occurred in 10.3% vs 10.5% (P=1.00), with low absolute rates of digital ischaemia (2.0% vs 0.5%; P=0.11).

Internal Validity

  • Randomisation and Allocation: Central telephone randomisation using a computer-generated sequence with variable block sizes; stratified by centre and severity-of-shock stratum (supports allocation concealment).
  • Dropout or exclusions: 802 patients were randomised; 21 did not receive an infusion; 2 withdrew consent after infusion; 1 was lost to follow-up before day 28; primary efficacy analysis included 778 patients (vasopressin 396; norepinephrine 382).
  • Performance/Detection Bias: Double-blind drug preparation and delivery; the primary outcome (mortality) is objective and resistant to ascertainment bias.
  • Protocol Adherence: Mean infusion rates of study drug differed by <2 mL/hour between groups; norepinephrine infusion rates were significantly lower in the vasopressin group during the first four study days (P<0.001).
  • Baseline Characteristics: Broadly similar illness severity (APACHE II: 27.1 ± 7.6 vs 27.0 ± 7.1); small imbalances included age (61.8 ± 16.0 vs 59.3 ± 16.4) and new coagulation dysfunction (22.0% vs 29.7%).
  • Heterogeneity: Multinational, multicentre enrolment increases clinical heterogeneity; stratified randomisation by centre and severity-of-shock was used to mitigate imbalance.
  • Timing: Patients were enrolled within 24 hours of meeting eligibility criteria; mean time from meeting criteria to study drug was 11.5 ± 9.4 hours vs 11.9 ± 8.9 hours.
  • Dose: Vasopressin dose was limited to 0.01–0.03 U/min (low-dose strategy), consistent with a catecholamine-sparing rather than “rescue high-dose” approach.
  • Separation of the Variable of Interest: Active separation was achieved by comparing different blinded study vasopressors while standardising the infusion rate (15 mL/hour), and by demonstrating lower norepinephrine infusion requirements in the vasopressin group (P<0.001 during days 1–4).
  • Adjunctive therapy use: Concomitant therapies (e.g., corticosteroids) were common and reported; days alive and free of corticosteroid use did not differ significantly (median 9.4 vs 10.8 days; P=0.11).
  • Outcome Assessment: Mortality to 28 and 90 days; organ dysfunction-free days were prespecified and reported as medians with interquartile ranges.
  • Statistical Rigor: Prespecified interim analyses; primary analysis distinguished between “randomised and infused” and “randomised” cohorts; adjusted odds ratios were reported for the primary cohort.

Conclusion on Internal Validity: Overall, internal validity appears strong to moderate given robust allocation concealment, blinding, and objective primary outcome assessment, with the main caveat being the analytic separation between “randomised” and “randomised + infused” cohorts (and associated exclusions).

External Validity

  • Population Representativeness: Adult ICU patients with established septic shock requiring ongoing norepinephrine; exclusions (e.g., acute coronary syndromes, severe heart failure, mesenteric ischaemia, pregnancy) limit applicability to some high-risk cardiovascular/ischaemic phenotypes.
  • Applicability: Highly applicable to ICU practice in well-resourced systems where norepinephrine is standard first-line therapy and vasopressin is considered as an alternative or adjunct vasopressor.
  • Temporal generalisability: Conducted in 2001–2006; background sepsis care and adjuncts have evolved since then, but the core haemodynamic problem and vasopressor choices remain relevant.

Conclusion on External Validity: External validity is good for adult ICU septic shock requiring vasopressors, but is more limited for patients with major contraindications to vasopressin, for non-ICU contexts, and for paediatric populations.

Strengths & Limitations

  • Strengths: Large multicentre ICU trial; double-blind design; prespecified severity stratification; clinically definitive primary outcome; reporting of organ dysfunction-free days and serious adverse events.
  • Limitations: Primary analysis depended on the “randomised and infused” cohort; substantial screening-to-enrolment attrition; vasopressin dose capped at 0.03 U/min (does not test higher-dose rescue strategies); conducted in an era with different background sepsis therapeutics than current practice.

Interpretation & Why It Matters

  • Clinical signal
    VASST did not demonstrate a statistically significant overall reduction in 28-day mortality with low-dose vasopressin compared with norepinephrine, but it provided a quantified estimate of effect compatible with modest benefit or no benefit.
  • Phenotype matters
    Prespecified severity-of-shock strata suggested differential effects, with the clearest signal in the “less severe shock” group at 90 days (RR 0.78; 95% CI 0.61 to 0.99; P=0.04), reinforcing the concept that vasopressor responsiveness and risk–benefit may vary by shock severity.
  • Practice impact
    The trial helped reframe vasopressin as an evidence-based non-catecholamine option (typically as an adjunct rather than a replacement), with careful attention to ischaemic complications and patient selection.

Controversies & Subsequent Evidence

  • Interpretation of subgroup findings: The prespecified severity-of-shock strata showed a mortality signal in “less severe shock” (particularly at 90 days), while the treatment-by-stratum interaction was not statistically significant (reported interaction P=0.10), raising concerns about overinterpretation of subgroup effects.
  • Trial framing and urgency: The accompanying editorial emphasised cautious interpretation of the overall neutral primary outcome and the practical urgency of timely haemodynamic support in septic shock.1
  • Correspondence debate: Published correspondence highlighted the potential clinical importance of the “less severe shock” subgroup signal and discussed how clinicians should weigh subgroup findings in practice; an author reply was published in the same correspondence thread.2
  • Vasopressin–corticosteroid interaction hypothesis: A subsequent analysis reported a statistically significant interaction between vasopressin assignment and corticosteroid use (interaction P=0.008), with lower 28-day mortality among patients receiving corticosteroids who were assigned vasopressin (35.9% vs 44.7%; P=0.03), but higher (non-significant) mortality among those not receiving corticosteroids (33.7% vs 21.3%; P=0.06).3
  • Later randomised evidence (early vasopressin strategy): The factorial VANISH trial (early vasopressin vs norepinephrine; hydrocortisone vs placebo) did not demonstrate a clear reduction in kidney failure-free days with early vasopressin, contributing to the overall view that vasopressin is not a universal mortality-modifying first-line vasopressor in septic shock.4
  • Evidence synthesis: An individual patient data meta-analysis of vasopressin/terlipressin trials has been used to contextualise VASST, generally supporting catecholamine-sparing effects without consistent overall survival benefit, and encouraging phenotype- and timing-aware interpretation of vasopressin strategies.5
  • Guideline translation: Modern international guidelines continue to recommend norepinephrine as first-line vasopressor in septic shock and consider vasopressin as an adjunct (often to reduce norepinephrine dose), reflecting the largely neutral mortality signal from VASST alongside broader evidence.6

Summary

  • VASST (NEJM 2008) compared low-dose vasopressin (0.01–0.03 U/min) with norepinephrine (5–15 μg/min) as a blinded study vasopressor strategy in adult ICU septic shock.
  • The trial was not stopped early; primary analysis (randomised + infused) found no statistically significant reduction in 28-day mortality (35.4% vs 39.3%; RR 0.90; 95% CI 0.75 to 1.08; P=0.26).
  • Prespecified severity strata suggested a signal in “less severe shock,” including lower 90-day mortality with vasopressin (35.8% vs 46.1%; RR 0.78; 95% CI 0.61 to 0.99; P=0.04).
  • Serious adverse events were similar overall (10.3% vs 10.5%; P=1.00), with low absolute rates of digital ischaemia (2.0% vs 0.5%; P=0.11).
  • VASST informed subsequent research and guidelines that generally position vasopressin as an adjunct to norepinephrine rather than a universal first-line replacement.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • VASST tested a low-dose vasopressin strategy (max 0.03 U/min) against a norepinephrine strategy under blinded conditions, with additional open-label vasopressors titrated to a common mean arterial pressure target.
  • Where an article uses an electronic article identifier rather than page range (e.g., “R154”), this is reproduced as published.

Overall Takeaway

VASST was a landmark, rigorously blinded multicentre ICU trial that demonstrated low-dose vasopressin is not a universal mortality-reducing replacement for norepinephrine in septic shock, while providing important prespecified stratum signals and safety quantification. Its enduring influence is in shaping modern practice towards norepinephrine first-line with vasopressin considered as an adjunct in selected patients, and in framing subsequent trials and meta-analyses around phenotype, timing, and co-therapies.

Overall Summary

  • In adult ICU septic shock, vasopressin (0.01–0.03 U/min) did not significantly reduce 28-day mortality vs norepinephrine in the overall cohort, but suggested a possible benefit in less severe shock strata.
  • Serious adverse events were similar overall, with low absolute rates of digital ischaemia and hyponatraemia signals requiring clinical vigilance.
  • The trial strongly influenced guideline positioning of vasopressin as an adjunct rather than a universal first-line replacement vasopressor.

Bibliography