Skip to main content

Trial Summaries & Critiques

PRETREAT

Image: Shutterstock / Mehmet Cetin

Publication

  • Title: Proactive vs Reactive Treatment of Hypotension During Surgery: The PRETREAT Randomized Clinical Trial
  • Acronym: PRETREAT
  • Year: 2025
  • Journal published in: JAMA
  • Citation: Kant M, van Klei WA, Hollmann MW, de Klerk ES, Otterspoor LC, Besselink MG, Kappen TH, Veelo DP; PRETREAT Study Group. Proactive vs Reactive Treatment of Hypotension During Surgery: The PRETREAT Randomized Clinical Trial. JAMA. 2025;334(21):1905-1914.

Context & Rationale

  • Background
    • Intraoperative hypotension is frequent in noncardiac surgery and is consistently associated with postoperative organ injury and mortality in observational datasets; whether hypotension is causal, or a marker of patient/surgical risk, remains uncertain.
    • Randomised trials of blood pressure–targeted strategies have reported inconsistent clinical benefits (e.g., improved postoperative organ dysfunction in INPRESS vs neutral effects on major cardiovascular outcomes in BBB and POISE-3). 123
    • More recent perioperative trials have evaluated individualised blood pressure targets using different personalisation principles (e.g., IMPROVE-multi), reflecting ongoing uncertainty about the optimal target and the relevant high-risk phenotypes. 4
    • Consensus statements emphasise avoidance of sustained hypotension while acknowledging uncertainty about universal MAP thresholds/targets and which phenotypes benefit most. 5
    • Meta-analyses of randomised trials have not shown consistent improvements in patient-centred outcomes despite reductions in hypotension exposure. 67
  • Research Question/Hypothesis
    • In adults undergoing elective noncardiac surgery, does a proactive, risk-stratified intraoperative MAP strategy (targets ≥70/80/90 mm Hg depending on hypotension risk) reduce postoperative disability compared with usual (reactive) care?
  • Why This Matters
    • Functional disability at 6 months is a patient-important outcome integrating complications and recovery trajectories.
    • Proactive strategies typically require earlier and higher-dose vasoactive therapy; benefit must be demonstrated to justify exposure, workload, and cost.
    • Findings inform whether perioperative haemodynamic optimisation should remain “blood pressure–centred” or pivot towards broader perfusion/oxygen delivery strategies.

Design & Methods

  • Research Question: Whether proactive, risk-stratified targeting of higher intraoperative MAP reduces patient-reported disability at 6 months after elective noncardiac surgery, compared with usual care.
  • Study Type:
    • Investigator-initiated, pragmatic, adaptive two-phase, parallel-group randomised clinical trial. 8
    • Adaptive phase allowed iterative refinements to the intervention (with 3-weekly monitoring of hypotension) to ensure a ≥30% reduction in hypotension exposure (AUC MAP <65 mm Hg) before entering the fixed trial phase.
    • Setting: operating theatres at 2 tertiary hospitals in the Netherlands.
    • Allocation: 1:1 with risk-stratified randomisation; Bayesian primary analysis.
  • Population:
    • Setting: Operating theatres (elective noncardiac surgery); enrolment June 2022 to October 2024.
    • Inclusion: Adults ≥18 years; elective noncardiac surgery; general or central neuraxial anaesthesia; expected hospital stay ≥1 night.
    • Risk strata at randomisation: Low risk (333 vs 344), intermediate risk (908 vs 906), high risk (377 vs 379).
    • Blood pressure measurement and hypotension definition: Invasive arterial measurements were recorded once per minute when used; otherwise, noninvasive measurements were taken every 3–5 minutes; hypotension was defined as MAP <65 mm Hg for ≥2 consecutive invasive measurements or ≥1 noninvasive measurement.
    • Key exclusions: Low-risk procedures; obstetric surgery; transplant surgery; ASA physical status 5; concurrent participation in another interventional trial; inability to meet study requirements/legal incapacity.
  • Intervention:
    • Risk stratification: Preoperative intraoperative-hypotension risk score categorised patients as low/intermediate/high risk using a centre-specific prediction model. 9
    • Blood pressure targets (from induction through surgery): Minimum MAP ≥65 mm Hg for all participants; risk-stratified minimum MAP targets ≥70 mm Hg (low risk), ≥80 mm Hg (intermediate risk), ≥90 mm Hg (high risk).
    • Delivery: Anaesthetists received patient-specific targets and a treatment guideline; vasoactive drug selection and dosing remained clinician-directed.
    • Preoperative co-interventions standardised: ACE inhibitors/angiotensin receptor blockers and diuretics were withheld on the day of surgery (except for heart failure with reduced ejection fraction); β-blockers and calcium-channel blockers were continued.
  • Comparison:
    • Usual care: Intraoperative management at the treating anaesthetist’s discretion, typically aiming to avoid sustained MAP <65 mm Hg without risk-stratified higher targets.
    • Contamination minimisation: Control anaesthetists were not informed that their patients were enrolled in the trial.
  • Blinding: Treating anaesthetists could not be blinded in the intervention group; control clinicians were unaware of enrolment; patients were not informed of group assignment until after completion of the 6‑month follow-up questionnaires; disability and quality-of-life outcomes were collected electronically.
  • Statistics:
    • Sample size / effect size sought: 5000 planned (2500 per group) to detect a 5-point absolute reduction in mean WHODAS score at 6 months (from 17 to 12) in the intermediate-risk stratum, allowing 20% loss to follow-up.
    • Alpha and beta/power: Not reported (Bayesian design).
    • Analysis type: Bayesian regression models; patients analysed according to randomisation among those receiving the allocated strategy with a complete perioperative dataset (modified intention-to-treat); sensitivity analyses for missing follow-up included multiple imputation and delta-based approaches.
  • Follow-Up Period: 30 days and 6 months after surgery (patient-reported disability and quality of life); last follow-up October 24, 2024.

Key Results

This trial was stopped early. Recruitment stopped after 2 interim analyses for futility; 3522/5000 planned patients were randomised (3247 received the allocated strategy and were analysed), because the posterior probability of a clinically meaningful (≥5-point) improvement in 6‑month disability was very low.

Outcome Proactive strategy Usual care Effect p value / 95% CI Notes
Primary: WHODAS 2.0 score at 6 months (0–100), mean (SD) [n] 17.7 (20.1) [1155] 18.2 (20.5) [1151] Mean difference −0.5 95% CrI −1.9 to 0.9 Posterior probability of ≥5-point improvement <1% (overall).
WHODAS 2.0 score at 30 days (0–100), mean (SD) [n] 7.8 (12.7) [1618] 7.1 (12.1) [1629] Mean difference 0.7 95% CrI −0.3 to 1.6 No signal for clinically important benefit.
Vasopressor initiation according to assigned guideline, % (n/N) 60.1% (973/1618) 29.5% (481/1629) OR 3.73 95% CrI 3.21 to 4.31 Protocol adherence / behavioural separation metric.
Area under threshold for MAP <65 mm Hg (mm Hg·min), mean (SD) 28.5 (80.9) 70.6 (133.8) Mean difference −42.3 95% CrI −49.8 to −34.8 Reduced hypotension exposure (efficacy requirement).
Average intraoperative MAP overall (mm Hg), mean (SD) 86.9 (6.8) 81.1 (7.1) Not reported Not reported Process outcome from Supplement. 9
Acute kidney injury (KDIGO), % (n/N) 2.0% (33/1618) 2.2% (36/1629) Risk difference −0.2 95% CrI −1.1 to 0.7 Low event rates.
ICU admission, % (n/N) 3.3% (53/1618) 3.8% (62/1629) Risk difference −0.5 95% CrI −1.9 to 1.0 No clinically meaningful difference.
All-cause death by 6 months, % (n/N) 0.2% (4/1618) 0.3% (6/1629) Risk difference −0.1 95% CrI −0.4 to 0.2 Very rare events; imprecise.
Hospital length of stay, days, median (IQR) 2.0 (2.0–4.0) 2.0 (2.0–4.0) Difference −0.02 95% CrI −0.2 to 0.1 No separation in utilisation outcomes.
  • Despite clear haemodynamic separation (higher MAP and lower hypotension exposure), there was no improvement in 6‑month disability; the posterior probability of a ≥5‑point WHODAS improvement was <1%.
  • High-risk patients had numerically lower 6‑month disability (mean difference −1.8; 95% CrI −5.7 to 2.0), but the credible interval crossed 0 and the posterior probability of a ≥5‑point benefit was 2%.
  • Early stopping for futility and 29% missing 6‑month disability data reduce precision for small effects and limit inference for rare harms.

Internal Validity

  • Randomisation and allocation concealment:
    • Computer-generated permuted-block randomisation (block sizes 4/6/8), stratified by predicted intraoperative-hypotension risk; allocation performed in an electronic data system immediately preoperatively.
    • Two-centre design; centre handled as a fixed effect in models.
  • Baseline characteristics and illness severity:
    • Age: 59 years (IQR 44–69) vs 59 years (IQR 45–69); baseline MAP: 104 mm Hg (IQR 93–114) vs 104 mm Hg (IQR 93–115).
    • ASA 3–4: 25.3% vs 25.4%; invasive arterial blood pressure monitoring: 18.6% vs 17.1%.
    • Preoperative disability was modest (WHODAS median 12.5 vs 14.6), and event rates for major morbidity and mortality were low, limiting power for rare outcomes.
  • Post-randomisation exclusions and missing outcome data:
    • 3522 randomised; 3247 received the allocated strategy and had a complete perioperative dataset (intervention: 1618; control: 1629).
    • Reasons for non-receipt included withdrawal of consent before anaesthesia (n=139), cancellation of surgery (n=122), withdrawal after anaesthesia started (n=5), and technical monitoring failure (n=9).
    • Primary outcome (WHODAS at 6 months) available for 2306/3247 (71.0%); 29.0% missing.
    • Sensitivity analyses for missing follow-up (including multiple imputation and delta-based approaches) were reported, but missingness remains a plausible source of bias for a patient-reported endpoint.
  • Performance and detection bias:
    • Treating anaesthetists were unblinded in the intervention group (unavoidable given patient-specific MAP targets).
    • Control clinicians were not informed of trial participation, reducing behavioural contamination but limiting the ability to blind all providers.
    • Patients were blinded until after completion of 6‑month follow-up; electronic questionnaires and analysis were consistent with blinded outcome assessment.
  • Protocol adherence and separation of the variable of interest:
    • Guideline-adherent vasopressor initiation: 60.1% (973/1618) vs 29.5% (481/1629); OR 3.73; 95% CrI 3.21 to 4.31.
    • Average intraoperative MAP overall: 86.9 ± 6.8 mm Hg vs 81.1 ± 7.1 mm Hg. 9
    • Average MAP during surgery: 88.3 ± 6.7 mm Hg vs 82.7 ± 7.0 mm Hg. 9
    • Hypotension exposure (AUC MAP <65): 28.5 ± 80.9 mm Hg·min vs 70.6 ± 133.8 mm Hg·min; mean difference −42.3; 95% CrI −49.8 to −34.8.
    • Residual hypotension remained common: MAP <65 mm Hg at any time in 47.7% vs 71.0%. 9
    • Noradrenaline use: 58.3% (943/1618) vs 46.4% (756/1629); total dose 97.4 μg (IQR 0–597.3) vs 0.0 μg (IQR 0–315.0). 9
  • Timing and “dose” of intervention:
    • Targets were applied from induction through surgery; mean overall MAP was 86.9 vs 81.1 mm Hg, which may be insufficient to change downstream disability if hypotension is not a dominant causal pathway.
  • Statistical rigour and early stopping:
    • Bayesian modelling with prespecified interim assessment of haemodynamic efficacy and a DSMB-recommended futility framework; stopping for futility reduces exposure to an ineffective strategy but can reduce precision for small effects.
    • Primary analyses excluded patients who did not undergo surgery or withdrew consent (modified intention-to-treat), which may introduce bias if exclusions were differential by group (not suggested by available data).

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomisation and allocation concealment were robust and the intervention achieved substantial haemodynamic separation, but early stopping for futility, post-randomisation exclusions, and 29% missing 6‑month disability outcomes reduce precision and leave some risk of bias for a patient-reported primary endpoint.

External Validity

  • Population representativeness:
    • Elective noncardiac surgery patients with expected postoperative admission (≥1 night) at two tertiary centres; results may not apply to ambulatory surgery, emergency surgery, obstetrics, transplant surgery, or the sickest ASA 5 patients.
    • The cohort included a broad surgical mix (including major abdominal surgery), but event rates for major morbidity and mortality were low, limiting transferability to very high-risk populations.
  • Intervention feasibility and transportability:
    • Implementation depends on access to a validated hypotension risk model and on clinician adherence to risk-stratified MAP targets; the prediction model was centre-specific. 9
    • Vasoactive drug choice and dosing were clinician-directed, which mirrors routine practice but may lead to different effects in systems with different vasopressor/fluid strategies.

Conclusion on External Validity: Generalisability is moderate for elective, admitted noncardiac surgery in well-resourced perioperative systems, but is uncertain for ambulatory or emergency settings, and for centres without a validated risk prediction infrastructure.

Strengths & Limitations

  • Strengths:
    • Large, pragmatic, multicentre randomised trial with risk-stratified targets and prespecified haemodynamic efficacy criteria.
    • Patient-important primary outcome (6‑month disability) with patient blinding until completion of follow-up.
    • Substantial haemodynamic separation achieved (higher MAP and lower hypotension exposure) enabling a fair test of the biological hypothesis.
    • Bayesian analytic framework with DSMB oversight and prespecified interim monitoring.
  • Limitations:
    • Stopped early for futility (3522/5000 randomised), reducing precision for small effects and limiting evaluation of rare harms.
    • Modified intention-to-treat analysis excluded post-randomisation withdrawals/cancellations; 29% missing primary outcome at 6 months.
    • Two-centre setting and centre-specific hypotension risk model may limit transportability.
    • Intervention relied on clinician-directed vasoactive selection and titration, potentially introducing variability in the achieved “dose” of vasopressor exposure.

Interpretation & Why It Matters

  • Clinical implication
    In broadly selected elective noncardiac surgery patients, raising intraoperative MAP targets using a proactive, risk-stratified approach reduced hypotension exposure but did not improve longer-term functional recovery; routine adoption of higher MAP targets for disability prevention is not supported.
  • Mechanistic implication
    The dissociation between haemodynamic separation and patient-centred outcomes supports the interpretation that intraoperative hypotension may be an imperfect surrogate for downstream disability, or that benefits are confined to narrower phenotypes not captured by the trial’s risk stratification.
  • Practice focus
    Perioperative haemodynamic management may need to prioritise individualised perfusion targets (e.g., baseline blood pressure, cardiac output/oxygen delivery, and context-specific autoregulation) rather than a uniform “higher MAP” strategy.

Controversies & Other Evidence

  • What PRETREAT adds to the blood pressure–targeting literature:
    • PRETREAT and the contemporaneous IMPROVE-multi trial both achieved meaningful reductions in hypotension exposure and/or higher MAP, yet neither demonstrated improvement in patient-centred outcomes, reinforcing limits of a purely blood pressure–centred strategy. 410
    • These findings contrast with INPRESS (high-risk surgery with invasive monitoring and baseline-relative targets), and align with BBB and POISE-3 in demonstrating that higher blood pressure targets do not reliably translate into better clinical outcomes. 123
  • Methodological debates highlighted by the accompanying editorial:
    • A strategy that reduces hypotension exposure but fails to improve functional outcomes challenges the causal interpretation of hypotension and raises the possibility that hypotension is a marker of risk rather than a modifiable driver of disability.
    • Increasing MAP targets generally necessitates greater vasoactive exposure; the net clinical value depends on the balance between avoiding low perfusion pressures and introducing drug-related harms—an equilibrium that may vary by phenotype. 10
    • Patient-centred outcomes are essential but vulnerable to missing data; trials using long-term questionnaires need robust follow-up strategies and transparent handling of missingness. 10
  • Subsequent and adjacent evidence:
    • BP-CARES (major abdominal surgery) evaluated intensive vs conventional intraoperative blood pressure management; interpretation should be integrated with PRETREAT and IMPROVE-multi when considering higher MAP targets in high-risk abdominal surgery. 11
    • Systematic reviews of RCTs report no consistent improvement in postoperative outcomes with higher blood pressure targets, despite reduced hypotension exposure, supporting a move towards more selective targeting. 67
  • Guidelines and expert syntheses after PRETREAT:
    • POQI XI (2024) emphasised measurement quality and avoidance of sustained hypotension while acknowledging uncertainty about universal thresholds and advocating context- and patient-specific targets. 5
    • An updated guideline from the Association of Anaesthetists and the British and Irish Hypertension Society (2026) provides contemporary recommendations on peri-operative blood pressure measurement and management incorporating recent trial evidence. 12
    • A 2026 narrative review synthesises contemporary evidence and frames intraoperative blood pressure management as a precision problem (risk, autoregulation, and co-interventions), rather than a single-threshold target. 13

Summary

  • PRETREAT randomised elective noncardiac surgery patients to proactive, risk-stratified higher MAP targets (≥70/80/90 mm Hg; minimum ≥65) versus usual care.
  • The intervention increased guideline-adherent vasopressor initiation (60.1% vs 29.5%) and raised mean intraoperative MAP (86.9 ± 6.8 vs 81.1 ± 7.1 mm Hg), with lower hypotension exposure (AUC MAP <65: 28.5 ± 80.9 vs 70.6 ± 133.8 mm Hg·min).
  • There was no improvement in 6‑month disability (WHODAS 17.7 vs 18.2; mean difference −0.5; 95% CrI −1.9 to 0.9), and the posterior probability of a clinically meaningful benefit (≥5 points) was <1%.
  • The trial was stopped early for futility after 3522/5000 planned randomisations, and 29% of analysed patients lacked 6‑month disability follow-up.
  • In the broader evidence base, PRETREAT supports the view that simply targeting higher intraoperative MAP, even with clear haemodynamic separation, does not reliably improve patient-centred outcomes across unselected elective surgery populations.

Overall Takeaway

PRETREAT provides a rigorous test of a pragmatic, risk-stratified “higher MAP” strategy: it changed intraoperative management and reduced hypotension exposure, yet did not improve longer-term functional recovery. The trial therefore challenges blood pressure as a sufficient surrogate target for postoperative disability and supports a more selective, phenotype-driven approach to haemodynamic optimisation.

Overall Summary

  • Proactive, risk-stratified higher MAP targets reduced hypotension exposure but did not improve 6‑month disability, and the trial was stopped early for futility.

Bibliography