Narrow pulse pressure predicts AKI even when the MAP looks fine

A quiet day for new primary literature — one genuinely new study. The other two items appeared on today’s aggregator listing but are three and six months old respectively; both are included because they continue threads already running here, and both are labelled with their real dates.

1. Narrow pulse pressure during normotension and postoperative AKI

Journal · Published: Anesthesiology, 3 September 2026 — genuinely new, today Yao, Liu, Zeng, Tang

Retrospective cohort of adult noncardiac surgery at a single academic centre in South Korea, 2011–2020. Patients stratified by median intraoperative pulse pressure: narrow (<40 mmHg), reference (40–70), wide (>70). Entropy balancing adjusted for covariates including cumulative vasopressor load and hypotension duration. A two-by-two risk matrix examined the interaction between MAP (<65 vs ≥65 mmHg) and pulse pressure, with a mechanistic substudy correlating pulse pressure against stroke volume index.

• 30,039 patients, median age 60, 54.6% male; AKI incidence 6.7%
• Narrow pulse pressure independently associated with AKI — odds ratio 1.66 (95% CI 1.42–1.94)
• The association held even when MAP was maintained above 65 mmHg

Conclusion, in the authors’ words: “A MAP-centric approach may obscure low-flow states, highlighting the potential value of incorporating pulse pressure into perioperative assessments.”

The deposited abstract truncates at the p-value for the odds ratio; the confidence interval is well clear of unity.

Interpretation. This is the third paper in a fortnight arriving at the same conclusion from a different direction, and together they make a more interesting argument than any of them alone.

• The BJA intraoperative hypotension meta-analysis found no outcome difference between higher and lower MAP targets, with achieved separation usually under 10 mmHg.
• The off-pump tissue oxygenation study showed cerebral oxygenation depends on both pressure and flow, and peripheral oxygenation on flow alone.
• This paper shows a flow surrogate predicts kidney injury at pressures we call adequate.

The common thread: MAP is the variable we can measure continuously and treat easily, which is not the same as it being the variable that matters. Pulse pressure is a crude flow surrogate, but it is free, already on every monitor, and — per the substudy — tracks stroke volume index.

Two caveats worth holding. Retrospective single-centre data cannot separate “narrow pulse pressure causes AKI” from “narrow pulse pressure marks the patient who was going to get AKI” — vascular stiffness, cardiac reserve and occult hypovolaemia all plausibly do both. And the entropy balancing adjusts for vasopressor load and hypotension duration, which is careful, but a patient with a persistently narrow pulse pressure at a normal MAP is usually a patient on vasopressors for a reason.

Still, as a prompt to look at the width of the trace rather than only its height, it is worth the reading time.

Read the paper · PMID 42479566

2. Haemoadsorption in septic shock — toward a personalised approach

Journal · Published: Critical Care — actually first published 13 June 2026, so ~3 months old, not new. It appeared on today’s aggregator listing because the issue version (Crit Care 2026;30:432) has just come out. Molnar, Toth, Ronco, Teboul, Mitzner, De Backer, Taccone et al. — a position statement.

The authors’ own account of the evidence base is unusually blunt for a group that largely supports the therapy:

Its adoption into clinical practice has largely been based on pathophysiological considerations rather than on evidence from large, well-designed randomized clinical trials. Over the past 15 years, most of the available evidence has been predominantly derived from small, single-center cohorts, reports from registries and heterogeneous prospective studies with substantial variability in patients’ selection, timing, and treatment intensity. In addition, the precise mechanisms of action of hemoadsorption remain incompletely understood… current guideline recommendations are largely based on expert opinions rather than high-certainty evidence.

Interpretation. This completes a picture the archive has been assembling. The Pittaway BJA meta-analysis found no benefit from haemoadsorption during cardiopulmonary bypass across 12 RCTs and 713 patients, only one at low risk of bias. CLEANSE reported HR 0.68 (0.44–1.07) in septic shock and is under formal critique. The Ohri Perfusion review proposed tiered patient selection. Now a position statement from the field’s own advocates — Ronco, De Backer, Teboul, Taccone — concedes that fifteen years of adoption rests on pathophysiological reasoning and expert opinion.

The honest summary across all four: this is a therapy in wide clinical use whose supporters agree it has never been properly tested, and whose response is to argue for better patient selection rather than for the definitive trial. That may well be right mechanistically. It is also the argument every unproven therapy makes.

Read the position statement · PMID 42288929

3. Ventilator weaning and extubation in acute brain injury

Journal · Published: Anaesthesia Critical Care & Pain Medicine — actually first published 20 March 2026, so ~5.5 months old, not new; the issue version (2026;45(5):101816) is what surfaced today. Al-Husinat, Araydah, Schultz, Rocco, Rose, Patroniti et al. Narrative review.

The central observation: in acute brain injury, respiratory mechanics are frequently preserved, and what actually determines extubation readiness is impaired consciousness, ineffective airway protective reflexes and excessive secretions.

• Evidence consistently shows the main determinants of extubation success are cough strength, swallowing function, ability to follow commands, and level of consciousness
• Traditional respiratory predictors — rapid shallow breathing index, maximal inspiratory pressure / negative inspiratory force — show limited discriminatory value in this population
• Emerging data favour integrated models combining neurological and respiratory variables
• The review reframes Boles’s Six Stages of Weaning through a neurocritical care lens

Interpretation. The practical content is the list of predictors that do not work. RSBI is calculated reflexively in most units and, in brain-injured patients, is close to uninformative — which matters because a normal RSBI in a patient who cannot protect their airway is a falsely reassuring number.

It also sits directly against the PUMA extubation guideline from 26 August, whose framing — that planned extubation is always elective and deferral is a legitimate strategy — is exactly the posture this population demands. And it complements the Grieco ventilation study in acute brain injury: the lung and the brain want different things throughout the admission, and weaning is where that conflict comes to a head.

Read the review · PMID 41866030