Publication

• Title: Effect of Reduced Exposure to Vasopressors on 90-Day Mortality in Older Critically Ill Patients With Vasodilatory Hypotension: A Randomized Clinical Trial
• Acronym: 65 Trial
• Year: 2020
• Journal published in: JAMA
• Citation: Lamontagne F, Richards-Belle A, Thomas K, Harrison DA, Sadique MZ, Grieve RD, et al; 65 Trial Investigators. Effect of reduced exposure to vasopressors on 90-day mortality in older critically ill patients with vasodilatory hypotension: a randomized clinical trial. JAMA. 2020;323(10):938-949.

Context & Rationale

• Background

  Vasopressors are widely used in ICU vasodilatory hypotension to avoid hypotension associated with organ injury and death, yet carry plausible harms (excess vasoconstriction, cardiac/metabolic effects) and are a potentially modifiable “dose” of physiological support.
  Guidelines commonly recommended an initial MAP target of 65 mm Hg in septic shock, while practice patterns and observational cohorts reported systematically higher MAPs, suggesting routine “overshoot” above a nominal target.
  Pre-trial RCT evidence on MAP targets in septic shock largely contrasted higher targets (~80–85 mm Hg) vs lower targets (~65–70 mm Hg), showing no consistent overall mortality benefit, while raising the possibility of effect modification by chronic hypertension and age.¹²³
• Research Question/Hypothesis

  In critically ill adults aged ≥65 years receiving vasopressors for vasodilatory hypotension, does a permissive hypotension strategy (target MAP 60–65 mm Hg while receiving vasopressors) reduce 90-day all-cause mortality compared with usual care?
• Why This Matters

  Older ICU patients have high baseline mortality and high exposure to vasopressors; even modest reductions in vasopressor dose/duration could translate to meaningful patient-centred benefit if vasopressor-related harms are clinically important.
  The trial also directly tests a “less is more” physiological support paradigm (restraint rather than escalation), with implications for standard haemodynamic targets and how ICU teams operationalise “usual care”.

Design & Methods

• Research Question: Whether targeting a lower MAP (60–65 mm Hg) while receiving vasopressors reduces 90-day mortality in older (≥65 years) ICU patients with vasodilatory hypotension, compared with usual care.
• Study Type: Pragmatic, multicentre, parallel-group, randomised clinical trial; open-label haemodynamic strategy; conducted in UK National Health Service general adult ICUs; centrally randomised with site as a random effect in adjusted models.
• Population:
  • Setting: 65 UK adult general ICUs participating in national audit infrastructure (ICNARC Case Mix Programme).
  • Key inclusion: age ≥65 years; receiving a vasopressor infusion for vasodilatory hypotension; randomised within 6 hours of starting vasopressors; clinician intent that vasopressors would be required for at least a further 6 hours; “adequate” fluid resuscitation completed or ongoing; not previously enrolled.
  • Key exclusions (protocol-defined): vasopressors used solely for haemorrhage, acute ventricular failure, or post–cardiopulmonary bypass vasoplegia; ongoing treatment for brain injury or spinal cord injury; death deemed imminent; prior enrolment in the trial.
• Intervention:
  • Permissive hypotension strategy: while receiving vasopressors, target MAP 60–65 mm Hg; vasopressor infusions reduced when MAP exceeded the upper range, and increased if MAP fell below the lower range; strategy applied during vasopressor administration in ICU.
  • Operational focus: reduce exposure to vasopressors (dose and/or duration) by avoiding unnecessary escalation/maintenance above a lower MAP range.
• Comparison:
  • Usual care: MAP management and vasopressor titration at clinician discretion (no mandated MAP target or titration algorithm from the trial).
  • Co-interventions: general ICU care at the treating team’s discretion in both groups.
• Blinding: Unblinded (haemodynamic strategy not feasibly blinded); mortality outcomes were objective and derived from routine data linkage; patient-reported outcomes were collected by questionnaire.
• Statistics: Final planned total sample size 2600 participants; powered to detect a 6% absolute reduction in 90-day mortality (from 35% to 29%) with 90% power at a 5% 2-sided significance level; allowance for 2.5% loss to follow-up; 1 interim analysis after primary outcome ascertainment for 500 participants using a Peto-Haybittle stopping boundary (P<.001).
  Primary analysis: intention-to-treat principle with a modified analysis set reflecting participants who permitted use of routine data; unadjusted comparison for primary outcome (Fisher exact test) with prespecified adjusted modelling (multilevel logistic regression with covariate adjustment and random effect of site).
• Follow-Up Period: Primary endpoint at 90 days; additional survival follow-up to the longest available time point via routine data; selected patient-centred outcomes assessed at 90 days and (in a subset/early recruits) at 1 year.

Key Results

This trial was not stopped early._attach_it was completed to the revised planned sample size (n=2600), with one planned interim analysis and no early termination for efficacy or harm.

• Primary outcome: 90-day mortality was numerically lower with permissive hypotension (41.0% vs 43.8%), but did not meet prespecified superiority (absolute difference −2.85%; 95% CI −6.75 to 1.05; P=0.15); adjusted modelling yielded a statistically significant OR (0.82; 95% CI 0.68 to 0.98).
• Physiological separation was clear for MAP while on vasopressors (median 66.7 vs 72.6 mm Hg; difference −5.9) and for vasopressor exposure (median duration 33 vs 38 hours; dose 17.7 vs 26.4 mg norepinephrine equivalents).
• Prespecified subgroup: chronic hypertension showed a larger apparent benefit (214/560 [38.2%] vs 253/571 [44.3%]; adjusted OR 0.67 [0.51–0.88]; interaction P=0.047), requiring cautious interpretation given multiplicity and the trial’s primary result.

Internal Validity

• Randomisation and Allocation: Central randomisation with allocation concealment at enrolment; large sample (n=2600) and multicentre design reduce selection bias.
• Drop out or exclusions: Of 2600 randomised, 15 refused permission to use data; 1 lost to follow-up by 90 days; 5 declined retrospective consent for follow-up after 90 days but were included in the primary endpoint set; primary outcome analysed in 2463 patients (1221 vs 1242), representing a modified intention-to-treat set driven by consent for data use.
• Performance/Detection Bias: Open-label haemodynamic strategy; potential for clinician behaviour changes (fluids, adjunct vasopressors, thresholds for organ support), though mortality is objective and secondary outcomes include routine-data endpoints.
• Protocol Adherence: Nonadherence reported in 153/1358 (11.3%) of permissive hypotension participants and 63/1430 (4.4%) of usual care participants; separation was achieved but not absolute.
• Baseline Characteristics: Groups were broadly well matched (median age ~75; similar prevalence of chronic hypertension ~46%); imbalance noted in pre-admission dependency requiring assistance with daily activities (34.4% permissive vs 30.9% usual care), which is prognostically relevant and partly motivates attention to adjusted analyses.
• Heterogeneity: Conducted across 65 ICUs; adjusted models included random effect of site; no strong evidence provided that site-level variation undermined the primary inference, but pragmatic delivery implies some inevitable between-site heterogeneity in usual care MAP targets and titration practices.
• Timing: Randomisation within 6 hours of vasopressor initiation targets an early treatment window; median vasopressor exposure before randomisation was 186 minutes in both groups, limiting immortal-time concerns within the enrolled cohort.
• Dose: MAP target difference was modest at the whole-population level, reflecting that “usual care” already often targets near 65 mm Hg; the intervention’s effect was therefore predominantly a reduction in “overshoot” above 65 and reduced vasopressor exposure rather than profound hypotension.
• Separation of the Variable of Interest: While receiving vasopressors, median MAP was 66.7 (64.5–69.8) vs 72.6 (69.4–76.5) mm Hg (difference −5.9); vasopressor duration 33.0 (19.0–55.0) vs 38.0 (22.0–63.0) hours (difference −5.0); norepinephrine-equivalent dose 17.7 (6.0–47.6) vs 26.4 (8.6–71.8) mg (difference −8.7).
• Key Delivery Aspects: Pragmatic implementation and standard ICU staffing improve realism; however, the control arm’s “usual care” may have included similar MAP targets, which plausibly attenuated effect size and increases reliance on exposure (dose/duration) differences rather than absolute MAP separation.
• Crossover: No formal crossover design; protocol nonadherence (higher in permissive arm) is the main mechanism of contamination and would bias towards the null for an exposure-reduction strategy.
• Adjunctive therapy use: Organ support use was broadly similar (advanced respiratory and renal support), suggesting limited compensatory escalation in the control arm that could mask benefits; nonetheless, the trial was not designed to standardise co-interventions.
• Outcome Assessment: Primary endpoint objective (mortality) with near-complete ascertainment; patient-reported outcomes (EQ-5D-5L, cognitive decline via IQCODE) are susceptible to missingness and survivor bias, and 1-year outcomes were available for only a subset.
• Statistical Rigor: Prespecified unadjusted primary analysis did not reach significance; adjusted models favoured permissive hypotension and became statistically significant, raising the methodological question of how strongly conclusions should weight adjusted vs unadjusted estimates in a pragmatic RCT with minor baseline imbalance and missing-by-consent exclusions.

Conclusion on Internal Validity: Moderate: randomisation and objective mortality follow-up are strong, but open-label delivery, consent-related exclusions from the primary analysis set, modest separation from an active “usual care” comparator, and a clinically relevant baseline imbalance increase uncertainty around the magnitude (and statistical robustness) of any mortality effect.

External Validity

• Population Representativeness: Older (≥65) critically ill adults in general UK ICUs; high prevalence of sepsis/septic shock and perioperative/medical admissions; vasopressor agents reflected UK practice (norepinephrine and metaraminol common).
• Applicability: Most applicable to high-resource ICUs where invasive BP monitoring and continuous vasopressor titration are standard; generalisability may be limited in settings with fewer staff or less frequent BP measurement, and in populations explicitly excluded (acute brain/spinal cord injury, specific cardiogenic/haemorrhagic indications for vasopressors).
• Implementation Considerations: The intervention is behaviourally and operationally simple (a target range plus active down-titration), but requires cultural acceptance of MAP 60–65 in older patients and consistent nursing/medical practice to avoid “target overshoot”.

Conclusion on External Validity: Good for older vasodilatory shock patients in comparable ICU systems; caution is warranted when extrapolating to patients with explicit higher-perfusion-pressure requirements, markedly different vasopressor protocols, or resource-limited environments.

Strengths & Limitations

• Strengths: Large pragmatic multicentre RCT in a high-risk group; early randomisation after vasopressor initiation; clear, actionable intervention; near-complete mortality follow-up; clinically meaningful separation in vasopressor exposure metrics; prespecified subgroup framework including chronic hypertension.
• Limitations: Open-label design; modified intention-to-treat analysis set due to post-randomisation data permission; modest absolute MAP separation vs usual care; higher nonadherence in the intervention arm; risk of overinterpretation of adjusted and subgroup findings given a non-significant unadjusted primary analysis.

Interpretation & Why It Matters

• Clinical practice

  A permissive hypotension strategy (MAP 60–65 while on vasopressors) reduced vasopressor exposure (median duration 33 vs 38 hours; median norepinephrine-equivalent dose 17.7 vs 26.4 mg) without clear increases in serious adverse events, but did not demonstrate a statistically significant reduction in 90-day mortality in the prespecified primary (unadjusted) analysis.
• Mechanistic implication

  The main modifiable element in real-world practice may be avoiding MAP overshoot above a nominal target, rather than selecting a substantially different target; the intervention’s benefit (if any) may reflect reduced unnecessary vasopressor dose and time rather than tolerance of profound hypotension.
• Subgroup signal

  The apparent benefit in chronic hypertension (adjusted OR 0.67) suggests that “one MAP fits all” is biologically and clinically unlikely; however, this should be treated as hypothesis-generating until replicated or confirmed in pooled analyses designed to test effect modification.

Controversies & Subsequent Evidence

• Primary (unadjusted) vs adjusted inference: The prespecified unadjusted primary analysis was non-significant (P=0.15), while adjusted modelling crossed conventional statistical thresholds (adjusted OR 0.82; 95% CI 0.68–0.98), prompting debate about how much weight should be placed on adjusted estimates in a pragmatic RCT with minor but prognostically relevant baseline imbalance.⁴
• “Usual care” as an active comparator: Editorial commentary highlighted that a MAP target of 65 mm Hg in routine practice can translate into achieved MAPs well above 65 because clinicians increase vasopressors when pressure is below target but may not actively reduce infusions when MAP is above target; the 65 trial’s strategy can be interpreted as an explicit countermeasure to this behavioural asymmetry rather than simply “allowing low MAP”.⁴
• Correspondence on interpretation: Published letters raised methodological concerns about interpretation of non-significant primary results and highlighted reliance on adjusted/subgroup findings; the investigators’ reply emphasised prespecified analysis structure and consistency of the point estimate favouring permissive hypotension across analyses.⁵⁶
• Subgroup multiplicity: The chronic hypertension interaction (P=0.047) risks false-positive inference given multiple subgroup tests and the primary outcome’s statistical non-significance; it remains clinically provocative but insufficient alone to mandate different MAP targets for all older hypertensive patients.
• Subsequent evidence synthesis: A more recent individual patient-level meta-analysis of MAP target trials in vasodilatory shock reported an overall mortality estimate close to unity with uncertainty intervals that include modest benefit; it supports continued individualisation rather than universal adoption of lower targets across all patients.⁷
• Guideline uptake: Post-trial international guidance continues to recommend an initial MAP target around 65 mm Hg in septic shock with clinical individualisation (including consideration of chronic hypertension and perfusion), rather than a blanket change to lower targets for all older patients.⁸

Summary

• Pragmatic UK multicentre RCT in ICU adults aged ≥65 with vasodilatory hypotension requiring vasopressors, randomised within 6 hours of vasopressor initiation.
• Permissive hypotension (MAP 60–65 on vasopressors) reduced vasopressor exposure (median duration −5 hours; norepinephrine-equivalent dose −8.7 mg) and lowered achieved MAP while on vasopressors (median −5.9 mm Hg).
• Primary outcome: 90-day mortality 41.0% vs 43.8% (absolute difference −2.85%; 95% CI −6.75 to 1.05; P=0.15); adjusted modelling suggested benefit (OR 0.82; 95% CI 0.68–0.98).
• No clear differences in advanced respiratory support, renal support, or overall serious adverse events; patient-reported HRQoL and cognitive outcomes were broadly similar among respondents.
• Chronic hypertension subgroup showed a larger apparent benefit (adjusted OR 0.67; interaction P=0.047), but remains hypothesis-generating.

Overall Takeaway

The 65 trial is a landmark pragmatic test of haemodynamic restraint in older ICU patients: targeting MAP 60–65 mm Hg while on vasopressors reduced vasopressor exposure but did not produce a statistically significant mortality reduction at 90 days in the prespecified primary analysis. Its enduring contribution is methodological and clinical—highlighting the importance of avoiding “overshoot” above targets and strengthening the case for individualised perfusion goals rather than reflexive escalation for all.

Bibliography

• 1Asfar 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.
• 2Lamontagne F, Meade MO, Hébert PC, et al; Canadian Critical Care Trials Group. Higher versus lower blood pressure targets for vasopressor therapy in shock: a multicentre pilot randomised controlled trial. Intensive Care Med. 2016;42(4):542-550.
• 3Lamontagne F, Day AG, Meade MO, et al. Pooled analysis of higher versus lower blood pressure targets for vasopressor therapy in septic and vasodilatory shock. Intensive Care Med. 2018;44(1):12-21.
• 4Marshall JC. Choosing the best blood pressure target for vasopressor therapy. JAMA. 2020.
• 5Das A, Bersten A, Bellomo R, et al. Reducing vasopressor exposure in patients with vasodilatory hypotension. JAMA. 2020.
• 6Lamontagne F, Richards-Belle A, Thomas K, et al. Reducing vasopressor exposure in patients with vasodilatory hypotension—Reply. JAMA. 2020.
• 7Angriman F, et al. Blood pressure targets for adults with vasodilatory shock: a patient-level meta-analysis. NEJM Evid. 2025.
• 8Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181-1247.