The AHA rewrites its endocarditis statement after eleven years, a Brazilian ICU rehabilitation guideline that contradicts three null trials, and an appraisal finding every hospital sepsis guideline low quality

Three new documents, all published 8 September. Both sweeps ran clean with no query failures: the topic-wide guideline search across all journals for 8–9 September (39 hits) and the society-acronym sweep for the same window. The French-society sweep returned a confirmed zero for 1–9 September — SFAR, SPILF, SRLF and Infectious Diseases Now have published nothing this month.

1. AHA scientific statement on infective endocarditis — the first update since 2015

Society · Published: American Heart Association, Councils on Lifelong Congenital Heart Disease and Heart Health in the Young; Clinical Cardiology; Cardiovascular and Stroke Nursing; Cardiovascular Radiology and Intervention. Circulation, 8 September 2026 — published yesterday DeSimone, Marks, Dayer, Esquer Garrigos, Messika-Zeitoun, O’Gara, Pettersson, Bucholz, Andrews, Fowler, Miro, Baddour

The statement opens with its own justification, and it is a list of everything that has changed in eleven years:

Since the publication of the 2015 American Heart Association scientific statement on infective endocarditis, its incidence has increased, its epidemiology has changed, and novel diagnostic and treatment options have become available. Updated case definitions, non-culture-based pathogen identification methods, and advanced multimodality imaging techniques have emerged. A randomized controlled trial of partial oral therapy published in 2019 may change the treatment landscape for the foreseeable future. Percutaneous mechanical aspiration of right-sided vegetations is being used in select patients with uncontrolled infection despite appropriate antimicrobial therapy. The importance of an endocarditis team has become increasingly recognized in the management of this life-threatening syndrome.

Access note: this is a long-form scientific statement and only the abstract is deposited in Europe PMC. The recommendation tables are not reachable from here, so what follows is the framing the statement gives itself, not its individual recommendations. If you have institutional access to Circulation, the tables are the substance.

Interpretation. Four threads are named in that paragraph, and each is a real change.

The 2019 trial is POET, and its shadow is everywhere in this statement’s framing — “may change the treatment landscape for the foreseeable future” is a striking thing for an AHA statement to say about partial oral therapy, which American practice has been markedly slower to adopt than European practice. This archive sent POET-II from the ESC Congress on 29 August, on response-tailored antibiotic duration. Read together, the direction is unmistakable: shorter courses, earlier oral switch, duration driven by response rather than by the calendar. That the AHA is now framing its statement around it is the more significant half of the news.

“Its epidemiology has changed” is doing quiet work. Endocarditis in 2026 is increasingly a disease of injection drug use, of prosthetic material, and of cardiac implantable electronic devices — a different patient from the rheumatic-valve case the older teaching was built around, and one who reaches the cardiac surgeon and the intensivist by a different route.

Percutaneous mechanical aspiration of right-sided vegetations is the item most likely to be new to a perioperative reader. It is an alternative to surgery in patients with uncontrolled infection, and it is described as being used “in select patients” rather than recommended — appropriate caution for a technique whose evidence is registry-level.

The endocarditis team is the organisational recommendation, and it is the same shape as the heart team: a formalised multidisciplinary structure for a disease where the cardiologist, the microbiologist, the infectious diseases physician and the cardiac surgeon each hold one part of the decision and none holds it all. It also names a practical gap — most hospitals treating endocarditis do not have one.

Note where this leaves the SPILF-AEPEI position statement on antibiotic therapy and prophylaxis of infective endocarditis, still on the outstanding list unread. A side-by-side of the French and American positions on prophylaxis specifically would now be worth doing — the two traditions have differed on it for fifteen years.

Read the statement · PMID 42708190

2. ASSOBRAFIR ICU rehabilitation guidelines — a strong recommendation the trials do not support

Society · Published: ASSOBRAFIR, the Brazilian Association of Cardiorespiratory Physiotherapy and Intensive Care Physiotherapy. Brazilian Journal of Physical Therapy, 8 September 2026 — published yesterday Barbosa Neto, Martinez, Lunardi, Dias, Vieira, Ferreira, Pinto, Cordeiro, and 22 others, with Cipriano as senior author

GIN and AGREE-II methods, GRADE for evidence appraisal, four key questions. Not a tracked society, but squarely in primary scope.

The four recommendations:

1. Structured and progressive early mobilization over usual care — strong recommendation, low-to-moderate certainty
2. Inspiratory muscle training instead of usual care in mechanically ventilated adults undergoing difficult or prolonged weaning — strong recommendation, moderate certainty
3. Against passive mobilization alone versus usual care in non-cooperative ICU patients — conditional, very low certainty
4. Against inspiratory muscle training as a routine intervention in mechanically ventilated adults under simple weaning — conditional, low certainty

Interpretation. Recommendation 1 sits directly against what this archive has been recording, and the tension is worth stating plainly rather than smoothing over.

Early mobilization has now failed three randomised trials in a row. TEAM (NEJM 2022), NEXIS, and EVER — sent here on 26 August, where a structured six-step programme in 169 ventilated patients with sepsis or respiratory failure produced a null primary outcome (FSS-ICU 23.6 vs 22.2, P = 0.44) despite successfully delivering nearly three times the mobilization. A strong GRADE recommendation on low-to-moderate certainty, issued in that context, is a methodological choice that deserves scrutiny. GRADE permits strong recommendations on low certainty, but the conditions for it are narrow, and “three trials showed no functional benefit” is not obviously one of them.

Two honest readings are possible. The generous one is that the guideline is recommending structured, progressive mobilization as a process — a physiotherapy service, an escalation protocol, a default that patients are moved rather than left — and that the trials tested more mobilization against a comparator that was already mobilising, not mobilisation against nothing. That is a fair defence, and it is the same argument the “mobilise smarter, not harder” editorial position makes. The less generous one is that a physiotherapy society has issued a strong recommendation for physiotherapy.

Recommendations 2 and 4 are the more interesting pair, and they are the ones I would act on. Inspiratory muscle training is recommended only in difficult or prolonged weaning (strong, moderate certainty) and recommended against as routine in simple weaning. That is a genuinely discriminating pair of statements: it targets an intervention at the population where the physiology predicts benefit and explicitly declines to spread it across everyone. It also happens to be the only recommendation here carrying moderate certainty. Guidelines that say where not to use their own specialty’s intervention are worth more than guidelines that do not.

Recommendation 3 — against passive mobilization alone in non-cooperative patients — rests on very low certainty and is appropriately conditional. It is also the recommendation with the clearest resource implication, since passive mobilization is what actually happens to sedated patients in most units.

Read the guideline · PMID 42710394

3. SENTINEL — every hospital paediatric sepsis guideline appraised was low quality

Study · Published: SENTINEL International study group (Sepsis Epidemiology in Paediatric Acute Care International Study). Archives of Disease in Childhood, 8 September 2026 — published yesterday Solan, Williams, George, Breuer, Hearps, Yock-Corrales, Pavlicich, Krishnamurthy, and a large international collaboration. ACTRN12621000920097

Not a guideline but an appraisal of guidelines — included here because that is what this watch exists to track. 14 hospital-specific paediatric sepsis clinical practice guidelines from 13 countries across five continents, appraised with AGREE II and checked for concordance with the Surviving Sepsis Campaign 2020 paediatric standard across 12 key treatment recommendations.

• All 14 were rated low methodological quality overall
• Lowest-scoring AGREE II domains: rigour of development (range 0–50%) and editorial independence (0–58%)
• Concordance with SSC 2020 ranged from 42% to 100%
• Consistency across the 12 recommendations ranged from 43% (steroid indications) to 93% (initial laboratory investigations, time to antibiotics, fluid bolus volume)
• The authors’ conclusion: despite uniformly low methodological quality, most aligned with SSC 2020 on the time-critical recommendations; the clinically important variation was in fluid resuscitation composition and timing of steroids

Interpretation. This is the most useful document of the three, and the least likely to be read.

The headline is not the failure — it is the pattern of the failure. Fourteen hospital guidelines, none methodologically sound, and yet 93% agreement on time to antibiotics and fluid bolus volume. The things that are settled are settled everywhere, regardless of how badly the local document was written. What varies is exactly what remains genuinely uncertain in the primary literature: which fluid, and when to give steroids — 43% consistency on steroid indications is close to a coin flip.

That is a reassuring finding dressed as an alarming one. A local guideline can be poorly constructed by AGREE II standards — no systematic search, no declared conflicts, no external review — and still transmit the settled parts of the evidence correctly, because those parts are transmitted by the whole culture rather than by the document. The document is failing where the field is failing.

The editorial independence scores of 0–58% are the part worth taking personally. Zero means no conflicts of interest were declared or managed at all. For a hospital protocol that is easy to shrug at — who has a financial interest in a paediatric sepsis pathway? — but the AGREE II point is broader than money: it asks whether the funding body’s or the department’s views influenced the content. In a document deciding fluid composition, that is not a hypothetical.

Practically: if your unit has a local sepsis pathway, this study predicts it is methodologically weak, correct on antibiotics and fluid volume, and idiosyncratic on steroids and fluid choice. That is a specific and testable prediction about your own document, and checking it takes twenty minutes.

Set it beside the 2026 Surviving Sepsis Campaign guidelines already in this archive: the guideline exists and is good; the problem is the fourteen translations of it.

Read the appraisal · PMID 42711094

---

Also seen and not pursued: a correction notice to the 2026 ACC/AHA dyslipidemia guideline and a “Top 10 Concepts” summary of it (secondary-scope cardiology, and the guideline itself predates this archive), the 2026 ASNC guideline for stress testing in SPECT and PET myocardial perfusion imaging, a Spanish Society of Cardiology imaging position statement on indications for cardiac MR and CT, an IOF position statement on bone fragility in MGUS, an AACE consensus statement on diabetes technology, and the SCCM Presidential Address from the 55th Annual Congress. Two EJA items dated 9 September are correspondence replies about the 2025 ESAIC/ESPA neuromuscular block guidelines and the ESAIC preoperative assessment guidelines — letters, not guideline output, and both parent documents are already known to this archive.