Skip to main content

Foundational Trials

CATS

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Norepinephrine plus dobutamine versus epinephrine alone for management of septic shock: a randomised trial
  • Acronym: CATS
  • Year: 2007
  • Journal published in: The Lancet
  • Citation: Annane D, Vignon P, Renault A, Bollaert P-E, Charpentier C, Martin C, et al; CATS Study Group. Norepinephrine plus dobutamine versus epinephrine alone for management of septic shock: a randomised trial. Lancet. 2007;370(9588):676-684.

Context & Rationale

  • Background
    • Septic shock management routinely requires vasoactive drugs after fluid resuscitation, yet the optimal catecholamine strategy (single agent versus combinations; vasopressor versus inotrope “stacking”) was uncertain at the time.
    • Norepinephrine was widely used as a first-line vasopressor, while dobutamine was commonly added when clinicians suspected low cardiac output or persistent hypoperfusion despite restored arterial pressure.
    • Epinephrine was an alternative “single catecholamine” strategy (vasopressor + inotrope properties), but concerns persisted regarding tachyarrhythmias and metabolic effects (notably lactate generation and acidaemia), potentially confounding lactate-guided resuscitation targets.
    • Prior comparative evidence was limited and largely non-definitive; pragmatic, blinded randomised comparisons of real-world catecholamine strategies were rare.
  • Research Question/Hypothesis
    • In adults with early septic shock requiring catecholamines, does a norepinephrine-based strategy with dobutamine “as needed” (guided by haemodynamic targets) improve outcomes compared with epinephrine alone?
    • Implicit hypothesis: norepinephrine + dobutamine would be superior (or at least safer) by avoiding epinephrine-associated metabolic/chronotropic adverse effects while achieving comparable haemodynamic endpoints.
  • Why This Matters
    • Catecholamine choice is a core, high-frequency ICU decision with plausible effects on macrohaemodynamics, microcirculation, lactate interpretation, arrhythmias, and organ perfusion.
    • A strategy-level comparison (rather than a narrowly controlled pharmacology experiment) directly informs bedside practice and guideline recommendations.
    • Demonstrating equivalence (or clinically important differences) would shape escalation pathways when septic shock is refractory to initial therapy.

Design & Methods

  • Research Question: Among adults with early septic shock requiring catecholamines, does norepinephrine plus dobutamine (as indicated by protocolised haemodynamic targets) reduce 28-day mortality compared with epinephrine alone?
  • Study Type: Prospective, multicentre, randomised, double-blind, parallel-group trial in 19 adult ICUs in France; allocation stratified by centre and balanced in blocks of six.
  • Population:
    • Setting: adult ICUs (France).
    • Key inclusion features (early septic shock): within <24 hours of meeting shock criteria; suspected/identified infection plus systemic inflammatory response; hypotension requiring catecholamines after initial resuscitation.
    • Shock/resuscitation thresholds: mean arterial (mean blood) pressure <70 mm Hg or systolic blood pressure <90 mm Hg; received a fluid bolus of ≥1000 mL or had pulmonary capillary wedge pressure 12–18 mm Hg; and required dopamine >15 μg/kg/min or any epinephrine or norepinephrine.
    • Hypoperfusion/organ dysfunction (≥2 required): PaO2/FiO2 <280 mm Hg if mechanically ventilated; urine output <0.5 mL/kg/h or <30 mL/h (for ≥1 hour); arterial lactate >2 mmol/L; platelet count <100×109/L.
    • Key exclusions: pregnancy; obstructive cardiomyopathy; acute myocardial ischaemia; pulmonary embolism; advanced-stage cancer/malignant haematopathy/AIDS where escalation was to be withheld/withdrawn; prolonged severe neutropenia (<0.5×109/L for >1 week); participation in another clinical trial.
  • Intervention:
    • Epinephrine strategy: epinephrine started at 0.2 μg/kg/min plus placebo infusion to maintain blinding to dobutamine allocation.
    • Protocolised titration: epinephrine increased by 0.2 μg/kg/min increments to achieve target mean blood pressure ≥70 mm Hg; additional placebo “inotrope” adjustments mirrored dobutamine titration rules.
    • Resuscitation co-protocol: repeated fluid challenges of 15–20 mL/kg colloids or crystalloids based on haemodynamic assessment; escalation guided by mean blood pressure, pulmonary capillary wedge pressure/echo hypovolaemia assessment, and cardiac index thresholds.
  • Comparison:
    • Norepinephrine + dobutamine strategy: norepinephrine started at 0.2 μg/kg/min plus dobutamine 5 μg/kg/min (dobutamine could be increased by 5 μg/kg/min increments when protocol criteria met, typically low cardiac index).
    • Norepinephrine increased by 0.2 μg/kg/min increments to maintain mean blood pressure ≥70 mm Hg; dobutamine used “whenever needed” per protocol to address low cardiac index/hypoperfusion while maintaining blinding.
    • Switching rule: permitted switching between strategies if epinephrine/norepinephrine or dobutamine/placebo failed (or vice versa), to preserve rescue therapy while maintaining the trial’s strategy comparison.
  • Blinding: Double-blind; study drugs prepared as identical syringes by pharmacists, with placebo dobutamine used in the epinephrine arm; haemodynamic titration algorithm applied in both groups to minimise performance bias while maintaining masking.
  • Statistics: Power calculation: 340 patients were required to detect a 20% absolute reduction in mortality (from 60% to 40%) with 95% power at a two-sided 5% significance level; primary analysis was intention-to-treat.
  • Follow-Up Period: Primary endpoint at 28 days; mortality also assessed to day 90; haemodynamic endpoints assessed during catecholamine infusion period.

Key Results

This trial was not stopped early. A total of 330 patients were randomised (planned 340), with complete 28-day endpoint ascertainment and 90-day follow-up.

Outcome Epinephrine Norepinephrine + dobutamine Effect p value / 95% CI Notes
28-day all-cause mortality (primary) 64/161 (40%) 58/169 (34%) RR 0.86 95% CI 0.65 to 1.14; P=0.31 RR reported for norepinephrine + dobutamine versus epinephrine
90-day all-cause mortality 84/161 (52%) 85/169 (50%) Not reported P=0.73 Unadjusted comparison
Adjusted 28-day mortality (logistic regression; age-adjusted) 64/161 (40%) 58/169 (34%) OR 0.82 95% CI 0.48 to 1.42; P=0.48 Model reported in the primary paper
Time to haemodynamic success Median NR Median NR Log-rank χ² 0.18 P=0.67 Haemodynamic success defined as mean blood pressure ≥70 mm Hg for ≥12 consecutive hours
Time to vasopressor withdrawal Median NR Median NR Log-rank χ² 2.83 P=0.09 Withdrawal defined as first interruption of study drugs for ≥24 hours
Pressor-free days until day 28 20 (0–24) 22 (6–25) Not reported P=0.05 Median (IQR)
Total fluid infused until catecholamine withdrawal 7.2 (2.2–12.3) L 8.5 (2.3–15.2) L Not reported P=0.09 Median (IQR)
Ventricular arrhythmias during catecholamine infusion (serious adverse event) 12/161 (7%) 8/169 (5%) Not reported Not reported Counts reported; no between-group p value provided for individual SAEs
  • No statistically significant difference in 28-day or 90-day mortality between strategies, with point estimates compatible with modest benefit or harm.
  • Norepinephrine + dobutamine achieved more pressor-free days to day 28 (median 22 vs 20 days), but without a signal for improved survival.
  • Epinephrine produced early metabolic/chronotropic differences (higher lactate early; lower arterial pH day 1–4; higher heart rate day 1–4), but serious adverse event counts were broadly similar; a published correction clarified the arterial pH statement and corrected a figure panel1.

Internal Validity

  • Randomisation and Allocation: Central randomisation by an independent statistician; stratified by centre; balanced in blocks of six; allocation concealment maintained via pharmacy preparation of identical syringes and placebo infusion.
  • Drop out or exclusions: Very low post-randomisation attrition; 1 participant discontinued intervention and withdrew consent on day 3; primary endpoint and 90-day follow-up were otherwise complete for analysed cohorts.
  • Performance/Detection Bias: Double-blinding (including placebo dobutamine) reduces differential co-interventions; protocolised haemodynamic algorithm likely standardised escalation decisions across sites.
  • Protocol Adherence: High delivery fidelity to initial allocation (156/161 received epinephrine; 166/169 received norepinephrine); the protocol allowed switching between strategies for failure, which protects safety but may reduce exposure separation (frequency of switching not reported).
  • Baseline Characteristics: Groups broadly comparable in severity (SAPS II and SOFA similar) and haemodynamic/biochemistry profiles; epinephrine group was slightly older (median 66 vs 63 years), a potential modest prognostic imbalance.
  • Heterogeneity: Strategy-level intervention introduces heterogeneity in dobutamine exposure (only “whenever needed” per algorithm) and in monitoring intensity (pulmonary artery catheter/echo-guided); however, protocolisation partially mitigates variability.
  • Timing: Enrolment targeted early shock (<24 hours of criteria met) and required contemporaneous resuscitation steps (fluid bolus and catecholamine requirement), supporting biological plausibility for intervention impact.
  • Dose: Starting doses were fixed (0.2 μg/kg/min for norepinephrine or epinephrine; 5 μg/kg/min for dobutamine/placebo) with prespecified increments; whether these doses are “optimal” across all phenotypes (e.g., profound vasoplegia vs myocardial depression) is uncertain, but dose titration was prespecified.
  • Separation of the Variable of Interest:
    • Pressor-free days until day 28: 20 (0–24) vs 22 (6–25).
    • Total fluid infused until catecholamine withdrawal: 7.2 (2.2–12.3) L vs 8.5 (2.3–15.2) L.
    • Time-to-event separation: log-rank P=0.67 for haemodynamic success; log-rank P=0.09 for vasopressor withdrawal (medians not reported).
    • Biochemical/chronotropic separation (lactate, pH, heart rate): direction and statistical significance reported; exact daily values not reported in text.
  • Key Delivery Aspects: The comparison is between a single-agent catecholamine strategy and a norepinephrine-first strategy with selective inotrope addition; this is clinically pragmatic but makes mechanistic attribution (vasopressor vs inotrope effects) less clean.
  • Outcome Assessment: Mortality is objective; haemodynamic success and vasopressor withdrawal were prespecified with clear operational definitions, though they are more clinician/protocol dependent than mortality.
  • Statistical Rigor: Intention-to-treat analysis used; effect estimates presented with confidence intervals; adjusted models reported; planned sample size (340) not fully reached (330), reducing power for the prespecified large effect size.

Conclusion on Internal Validity: Overall, internal validity is moderate-to-strong: randomisation, allocation concealment, and blinding were robust, with minimal attrition; however, the “strategy” comparator (selective dobutamine, permitted switching) and limited precision (slightly under target sample size; wide CIs) temper causal certainty for modest treatment effects.

External Validity

  • Population Representativeness: Adults with relatively early, catecholamine-requiring septic shock in French ICUs; exclusions removed several high-risk subgroups (e.g., acute myocardial ischaemia/pulmonary embolism; advanced malignancy with limitation-of-care decisions; profound prolonged neutropenia).
  • Applicability: Results are most applicable to ICU settings with capacity for protocolised haemodynamic monitoring and titration; generalisability may be reduced where dobutamine is used differently, where lactate-guided resuscitation is central, or where epinephrine is reserved strictly for refractory shock.
  • Sepsis Era Considerations: Conducted under pre-Sepsis-3 definitions and earlier bundles; absolute event rates and background care (e.g., antimicrobial timing, ventilation practices) may differ from contemporary practice, although catecholamine pharmacology is stable.

Conclusion on External Validity: Generalisability is moderate: the trial addresses a common ICU problem with pragmatic dosing and targets, but was conducted in a specific national ICU context with exclusions that may limit inference for the sickest or those with competing cardiac pathology.

Strengths & Limitations

  • Strengths:
    • Double-blind, multicentre randomised design in a high-acuity ICU population.
    • Pragmatic comparison of clinically plausible catecholamine strategies rather than an artificial fixed-dose pharmacology experiment.
    • Protocolised titration and clear operational definitions for haemodynamic endpoints.
    • Very low attrition with complete ascertainment of mortality endpoints.
  • Limitations:
    • Power calculation targeted a very large absolute mortality reduction (20%); the achieved sample size and observed effect sizes imply limited ability to exclude smaller but clinically important differences.
    • Comparator is a mixed “strategy” (norepinephrine plus selective dobutamine) and allowed switching for failure; exposure separation and mechanistic attribution are therefore imperfect.
    • Metabolic outcomes (lactate, pH trajectories) were directionally reported but not fully tabulated with exact values in the main text; one correction was issued for the arterial pH statement/figure.
    • Conducted in one country and one care era; external validity to different systems and modern bundles is incomplete.

Interpretation & Why It Matters

  • Clinical practice
    • Epinephrine (single-agent strategy) did not worsen survival compared with norepinephrine + dobutamine strategy, supporting its plausibility as an alternative catecholamine approach in selected settings.
    • The observed metabolic signal (higher lactate; lower pH early with epinephrine) reinforces caution when using lactate trends as a sole resuscitation target during epinephrine therapy.
  • Mechanistic framing
    • The trial contrasts a “vasopressor-first with optional inotrope” paradigm against a “single catecholamine with mixed α/β effects” paradigm, highlighting that catecholamine choice affects both haemodynamics and interpretability of metabolic biomarkers.
  • Methodology
    • Wide confidence intervals and a large targeted effect size mean the trial is better interpreted as excluding very large mortality differences than establishing strict equivalence.

Controversies & Subsequent Evidence

  • Editorial critique emphasised that the trial’s design compares complex strategies and may be underpowered for modest mortality differences; it highlighted the clinical importance of epinephrine’s metabolic footprint when clinicians increasingly use lactate as a therapeutic target2.
  • Subsequent large vasopressor trials shifted practice and reinforced norepinephrine as the preferred first-line agent in shock, with dopamine associated with more arrhythmic events in broad shock populations (SOAP II) and early norepinephrine strategies tested in modern resuscitation bundles (CENSER)34.
  • Guidelines published after CATS converged on norepinephrine as first-line vasopressor in septic shock, with epinephrine positioned as an add-on or alternative when additional agent(s) are required and vasopressin added as an adjunct in selected patients5.
  • Meta-analytic syntheses of catecholamine strategies generally support norepinephrine as the default vasopressor choice compared with dopamine, while leaving residual uncertainty for head-to-head epinephrine strategies where data remain comparatively sparse and strategy heterogeneity is substantial6.

Summary

  • In 330 adults with early catecholamine-requiring septic shock, epinephrine alone did not differ from norepinephrine + dobutamine strategy for 28-day mortality (40% vs 34%).
  • Time to haemodynamic success was similar between groups; time to vasopressor withdrawal showed a non-significant trend favouring norepinephrine + dobutamine.
  • Norepinephrine + dobutamine achieved more pressor-free days to day 28 (22 vs 20 days), without a survival signal.
  • Epinephrine produced early higher lactate and lower arterial pH with higher heart rate, underscoring biomarker-interpretation issues during epinephrine therapy; a published correction clarified one pH statement/figure.
  • The trial is best read as excluding very large mortality differences between these strategies; it does not settle smaller but clinically relevant differences, and later large trials/guidelines anchored practice around norepinephrine-first strategies.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • This trial evaluated two protocolised catecholamine strategies (not merely drug A vs drug B), and should be interpreted in that “strategy” frame.
  • The lactate/pH effects of epinephrine are clinically relevant when using lactate clearance as a resuscitation endpoint.

Overall Takeaway

CATS was a landmark pragmatic, double-blind strategy trial in septic shock, showing that epinephrine-alone and a norepinephrine-plus-selective-dobutamine approach produced broadly similar survival with differing metabolic signatures. Its main contribution is not a definitive mortality verdict, but a disciplined demonstration that epinephrine’s early lactate/acidaemia effects may coexist with comparable hard outcomes, while later large trials and guidelines nonetheless anchored practice around norepinephrine-first escalation pathways.

Overall Summary

  • Mortality at 28 and 90 days did not differ between epinephrine-alone and norepinephrine + dobutamine strategies in early septic shock.
  • Norepinephrine + dobutamine yielded more pressor-free days to day 28, but without a survival signal.
  • Epinephrine produced early lactate/pH changes that can complicate lactate-targeted resuscitation decisions.

Bibliography