Skip to main content

Foundational Trials

SEPSISPAM

Image: Shutterstock / ChaNaWiT

Publication

  • Title: High versus low blood-pressure target in patients with septic shock
  • Acronym: SEPSISPAM
  • Year: 2014
  • Journal published in: New England Journal of Medicine
  • Citation: Asfar P, Meziani F, Hamel JF, et al; SEPSISPAM Investigators. High versus low blood-pressure target in patients with septic shock. N Engl J Med. 2014;370(17):1583-1593.

Context & Rationale

  • Background
    • Mean arterial pressure (MAP) targets in septic shock were largely consensus-driven, commonly set at ≥65 mmHg, with wide practice variation.
    • Physiological rationale supported higher MAP targets in patients with chronic hypertension (right-shifted autoregulation) and in those with renal hypoperfusion, but higher vasopressor exposure was plausibly harmful.
    • Prior evidence consisted mainly of observational associations and small studies; a definitive multicentre trial was needed to balance potential organ-protection against catecholamine-related harms.
  • Research Question/Hypothesis
    • Does targeting a higher MAP (80–85 mmHg) versus a lower MAP (65–70 mmHg) during early septic shock reduce 28-day mortality (and other clinically relevant outcomes)?
    • Is any effect modified by pre-existing chronic arterial hypertension (prespecified stratification)?
  • Why This Matters
    • MAP is a continuously titrated, ubiquitous ICU “dose” decision with immediate downstream consequences for vasopressor exposure, organ perfusion, and arrhythmia/ischaemia risk.
    • Even small absolute effects on mortality or renal failure would translate into large population-level impact given the global burden of septic shock.
    • The trial directly tested whether “more pressure” is benefit or harm in contemporary ICU resuscitation.

Design & Methods

  • Research Question: In adult septic shock, does targeting MAP 80–85 mmHg compared with 65–70 mmHg improve 28-day mortality (and secondary patient-centred outcomes), and is any effect modified by chronic arterial hypertension?
  • Study Type: Randomised, multicentre, parallel-group, open-label trial in French ICUs; central computer-based randomisation; prespecified subgroup stratification by chronic arterial hypertension; follow-up to 90 days.
  • Population:
    • Setting: 29 ICUs (France).
    • Adults with septic shock requiring vasopressors after fluid resuscitation.
    • Operational enrolment features (from screening log/CONSORT): shock >6 hours, inadequate fluid challenge, and catecholamine dose <0.1 µg/kg/min were common reasons for ineligibility.
    • Pre-randomisation resuscitation: protocol documentation indicates inclusion criteria were amended to include a minimum fluid load of 30 mL/kg and enrolment within 6 hours of shock onset.
    • Baseline severity: SAPS II 57.2 ± 16.2 (low-target) vs 56.1 ± 15.5 (high-target); SOFA 10.8 ± 3.1 vs 10.7 ± 3.1.
  • Intervention:
    • High-target MAP strategy: vasopressors titrated to maintain MAP 80–85 mmHg for the prespecified protocol period (first 5 days), primarily using norepinephrine; other vasoactive agents permitted as clinically indicated.
    • Weaning/titration guidance specified stepwise catecholamine reductions once MAP exceeded the target range (protocolised decrement schedules).
  • Comparison:
    • Low-target MAP strategy: vasopressors titrated to maintain MAP 65–70 mmHg for the prespecified protocol period (first 5 days), otherwise similar ICU care and co-interventions.
    • Other vasoactive agents permitted as clinically indicated (e.g., epinephrine, dobutamine) alongside norepinephrine.
  • Blinding: Unblinded haemodynamic management (MAP target requires bedside titration); outcomes were largely objective (mortality, RRT, prespecified adverse events), mitigating but not eliminating performance/detection bias.
  • Statistics: Power calculation: 800 patients required to detect a 10% absolute reduction in 28-day mortality (from 45% to 35%) with 80% power at the 5% significance level; analyses were intention-to-treat; interim analyses planned after 200, 400, and 600 patients using a stringent stopping threshold (Peto–Haybittle, P<0.001).
  • Follow-Up Period: 28 days (primary) and 90 days (secondary) after randomisation.

Key Results

This trial was not stopped early. Prespecified interim monitoring was planned; the final analysed cohort comprised 388 patients per group (776 total) versus a planned sample size of 800.

Outcome High-target MAP (80–85) Low-target MAP (65–70) Effect p value / 95% CI Notes
Death at day 28 (primary) 142/388 (36.6%) 132/388 (34.0%) HR 1.07 95% CI 0.84 to 1.38; P=0.57 Primary endpoint; time-to-event analysis reported.
Death at day 90 170/388 (43.8%) 164/388 (42.3%) Not reported P=0.74 Effect estimate not reported for this endpoint.
Survival at day 28 without organ support 235/388 (60.6%) 241/388 (62.1%) Not reported P=0.66 Organ support defined as vasopressors, mechanical ventilation, or renal-replacement therapy.
Renal-replacement therapy (day 1–7) 130/388 (33.5%) 139/388 (35.8%) Not reported P=0.50 Overall effect estimate not reported.
Renal-replacement therapy (day 1–7) — chronic arterial hypertension subgroup 53/167 (31.7%) 73/173 (42.2%) Not reported P=0.046 Prespecified subgroup; interaction P=0.02 (reported by trialists).
Doubling of plasma creatinine (overall) 150/388 (38.7%) 161/388 (41.5%) Not reported P=0.42 Patients treated with RRT were not included in this calculation (per supplementary definitions).
Doubling of plasma creatinine — chronic arterial hypertension subgroup 65/167 (38.9%) 90/173 (52.0%) Not reported P=0.02 Prespecified subgroup; direction suggests less renal dysfunction with higher MAP among chronic hypertension.
Newly diagnosed atrial fibrillation (serious adverse event) 26/388 (6.7%) 11/388 (2.8%) Not reported P=0.02 Higher MAP strategy associated with more AF episodes.
Acute myocardial infarction (serious adverse event) 7/388 (1.8%) 2/388 (0.5%) Not reported P=0.18 Definition included ECG changes plus troponin rise and echocardiographic regional wall motion abnormality, with angiographic confirmation when possible.
Any serious adverse event 74/388 (19.1%) 69/388 (17.8%) Not reported P=0.64 Composite of prespecified serious adverse events.
  • Despite substantially greater vasopressor exposure in the high-target strategy, 28-day mortality was similar (36.6% vs 34.0%; HR 1.07; 95% CI 0.84 to 1.38; P=0.57).
  • Renal outcomes showed effect modification: among patients with chronic arterial hypertension, high-target MAP was associated with less RRT (31.7% vs 42.2%; P=0.046) and less doubling of creatinine (38.9% vs 52.0%; P=0.02).
  • High-target MAP increased newly diagnosed atrial fibrillation (6.7% vs 2.8%; P=0.02), reinforcing a harm trade-off with “more pressure”.

Internal Validity

  • Randomisation and allocation: Central computer-based randomisation supported allocation concealment; prespecified stratification by chronic arterial hypertension; baseline characteristics were closely balanced (e.g., chronic hypertension 43.0% vs 44.6%; SAPS II 56.1 ± 15.5 vs 57.2 ± 16.2).
  • Dropout/exclusions: 798 randomised; 22 excluded post-randomisation (14 low-target; 8 high-target) due to withdrawal of consent or guardianship/prisoner status; primary analysis included 388 per group with 90-day follow-up reported for the analysed cohort.
  • Performance/detection bias: Open-label MAP titration creates scope for co-intervention differences; however, primary endpoint (mortality) is objective; key secondary endpoints (RRT, creatinine doubling) were protocol-defined.
  • Protocol adherence (target attainment): MAP targets were not achieved due to clinician-limited vasopressor dosing in 64/388 (16.5%) high-target vs 40/388 (10.3%) low-target (P=0.01); in 14/388 (3.6%) high-target patients, vasopressors were down-titrated to maintain MAP 65–70 due to adverse effects.
  • Timing: Screening data and protocol documentation indicate enrolment early in shock (delay from vasopressor start to inclusion 3.5 ± 2.2 hours high-target vs 3.6 ± 2.1 low-target), with exclusion of shock lasting >6 hours, and pre-enrolment fluid resuscitation (protocol amendment included ≥30 mL/kg).
  • Dose and separation of the variable of interest: Achieved MAPs clustered slightly above targets: low-target observed values were “for the most part” 70–75 mmHg and high-target 85–90 mmHg; vasopressor exposure was higher with high-target (median norepinephrine day 1: 0.58 [IQR 0.26–1.80] vs 0.45 [0.17–1.21] µg/kg/min; P<0.001) and duration longer (4.7 ± 3.7 vs 3.7 ± 3.2 days; P<0.001).
  • Heterogeneity: Broad septic shock population (predominantly pulmonary and abdominal sources) with prespecified hypertension stratification; treatment effects plausibly heterogeneous by vascular/renal autoregulation, but the trial was powered for an overall mortality effect rather than multiple subgroup interactions.
  • Outcome assessment/statistical rigor: Prespecified primary endpoint, intention-to-treat analysis, and stringent interim stopping thresholds; however, the achieved sample size (776 analysed vs 800 planned) and observed mortality lower than the design assumption could reduce power for modest effects.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong: randomisation and balance were robust and outcomes largely objective, but open-label titration and non-trivial target non-adherence (especially in the high-target arm) plausibly diluted treatment separation and attenuated detectable effects.

External Validity

  • Population representativeness: Typical adult ICU septic shock cohort (high baseline severity; ~74–79% mechanically ventilated; ~43–45% chronic hypertension), but conducted exclusively in French ICUs.
  • Applicability: Findings generalise best to settings using norepinephrine-first strategies with early enrolment; extrapolation to resource-limited environments, different vasopressor formularies, or markedly different resuscitation bundles should be cautious.
  • Clinical decision translatability: MAP targets are highly implementable and resemble real-world titration; however, observed MAP in the low-target group frequently exceeded the nominal target range (70–75 mmHg), so “usual care” in many ICUs may already approximate the achieved low-target physiology.

Conclusion on External Validity: External validity is good for contemporary high-income ICUs managing septic shock with norepinephrine, but may be limited where baseline practice, co-interventions, or achievable MAP control differ substantially.

Strengths & Limitations

  • Strengths:
    • Large, pragmatic, multicentre randomised trial addressing a ubiquitous ICU titration decision.
    • Prespecified subgroup stratification for chronic arterial hypertension (clinically plausible effect modifier).
    • Objective primary outcome with prespecified safety endpoints and high follow-up completeness for the analysed cohort.
    • Documented separation in vasopressor exposure and achieved MAP between strategies.
  • Limitations:
    • Open-label MAP management with potential for performance bias and clinician-driven protocol deviations.
    • Target non-adherence was common (16.5% high-target; 10.3% low-target), likely biasing towards the null.
    • Achieved MAP in the low-target arm was frequently 70–75 mmHg (above the 65–70 target), narrowing physiological contrast.
    • Final analysed sample size slightly below planned (776 vs 800), and design assumptions for mortality may have overestimated event rate.

Interpretation & Why It Matters

  • Mortality
    Routinely targeting MAP 80–85 mmHg (vs 65–70 mmHg) in septic shock did not improve 28- or 90-day survival, despite higher catecholamine exposure.
  • Renal trade-off signal
    Among patients with chronic arterial hypertension, the higher MAP strategy was associated with less RRT and less creatinine doubling, supporting a “selective higher MAP” hypothesis rather than a universal target.
  • Harms
    Higher MAP targets increased new-onset atrial fibrillation (6.7% vs 2.8%), underscoring catecholamine-mediated cardiovascular toxicity as a counterweight to organ-perfusion goals.

Controversies & Subsequent Evidence

  • Editorial framing highlighted that any benefit of higher MAP would need to outweigh increased vasopressor exposure and arrhythmia/ischaemia risk, and interpreted the chronic hypertension renal signal as plausible but not definitive given interaction testing and open-label titration. 1
  • Published correspondence raised methodological and interpretive issues (including achievable MAP separation and implications for subgroup inference), with author reply clarifying trial conduct and interpretation. 2
  • Subsequent MAP-target RCTs broadened the evidence base beyond septic shock alone: the OVATION pilot trial (vasodilatory shock, lower 60–65 vs higher 75–80 mmHg) established feasibility and suggested clinically meaningful heterogeneity hypotheses for larger trials. 3
  • The 65 Trial (older patients with vasodilatory hypotension) tested permissive hypotension (MAP 60–65) versus usual care and informed the contemporary “less vasopressor exposure” paradigm alongside SEPSISPAM. 4
  • Individual patient data meta-analysis of MAP-target RCTs (including SEPSISPAM, OVATION, and 65 Trial) suggested that lower MAP targets may reduce 90-day mortality with low certainty and did not identify robust subgroups that clearly benefit from higher targets. 5
  • Updated systematic review/meta-analysis of higher versus lower MAP targets in vasodilatory shock found no convincing mortality advantage for higher targets and reinforced arrhythmia/cardiac complication concerns, supporting a default lower MAP approach with individualisation. 6
  • Major guidelines incorporate SEPSISPAM by recommending an initial MAP target around 65 mmHg with individualisation (e.g., chronic hypertension) rather than routine higher targets. 78

Summary

  • SEPSISPAM randomised 776 ICU patients with septic shock to a high MAP target (80–85) versus a low MAP target (65–70) strategy.
  • High-target MAP did not improve 28-day mortality (36.6% vs 34.0%; HR 1.07; 95% CI 0.84 to 1.38) or 90-day mortality (43.8% vs 42.3%; P=0.74).
  • High-target MAP increased vasopressor exposure and was associated with more new-onset atrial fibrillation (6.7% vs 2.8%; P=0.02).
  • In the prespecified chronic hypertension subgroup, high-target MAP was associated with less RRT (31.7% vs 42.2%; P=0.046; interaction P=0.02) and less creatinine doubling (38.9% vs 52.0%; P=0.02).
  • Practice impact: supports MAP ~65 mmHg as a default target, with cautious individualisation (especially for chronic hypertension/renal endpoints) and attention to arrhythmia risk.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Trial registration: ClinicalTrials.gov NCT01149278.
  • In the low-target arm, achieved MAP commonly exceeded the nominal target (often 70–75 mmHg), an important consideration for implementation and effect dilution.
  • Renal subgroup findings in chronic hypertension are biologically plausible but should be interpreted alongside increased arrhythmia risk and subsequent pooled evidence.

Overall Takeaway

SEPSISPAM established that routinely targeting a higher MAP (80–85 mmHg) in septic shock does not improve survival compared with a lower target (65–70 mmHg), while increasing catecholamine exposure and atrial fibrillation. Its enduring contribution is the nuanced signal that patients with chronic arterial hypertension may experience improved renal outcomes with higher MAP targets, supporting guideline-endorsed individualisation rather than a universal “higher is better” haemodynamic doctrine.

Overall Summary

  • In septic shock, higher MAP targets did not improve mortality but increased atrial fibrillation; default MAP ~65 mmHg is supported, with selective escalation for chronic hypertension/renal endpoints when benefits outweigh arrhythmia risk.

Bibliography