This Week in Emergency Medicine — Aug 26, 2026

Welcome to This Week in Emergency Medicine. This week we're covering 10 notable papers spanning procedural care and sedation, high-stakes airway and resuscitation practice, and the growing pressure on emergency departments to move faster with better technology. Let's dive in.

We'll start with procedures, because two big consensus documents landed in Annals of Emergency Medicine this week. Green, Roback, Bhatt and colleagues have published a two-part multidisciplinary Delphi consensus on unscheduled procedural sedation, approved by the American College of Emergency Physicians board. Part one covers principles, oversight and quality monitoring [2], and part two addresses clinical practice at the bedside [1]. These are policy documents rather than peer-reviewed original research, and the abstracts don't summarise the recommendations, so the practical move is to read the full statements — particularly if you sit on a sedation credentialing or quality committee, because the oversight and monitoring half is where most departments have the least standardisation.

Staying with procedures, CJEM published a nice piece of reality-testing from Atkinson and colleagues on distal radius fracture reduction — one of the commonest reasons we do that procedural sedation in the first place. Seventy patients who underwent sedation and closed reduction had their films independently reviewed, blinded, by orthopaedic surgeons and by radiologists. Orthopaedic surgeons judged just over half of the post-reduction films adequate; radiologists judged about three quarters of them adequate. Agreement between the two specialties was moderate at best. Slippage by follow-up also depended on who was reading — roughly three in ten reductions by the surgeons' criteria, about one in six by the radiologists'. And here's the part that matters most for us: only seven patients, one in ten, went to surgery, and no demographic or radiographic pattern reliably separated them from those managed non-operatively. In other words, radiographic adequacy alone did not explain who ended up in the operating room. That should temper how hard we chase the perfect post-reduction film, and it argues for shared, explicit definitions of adequacy with our orthopaedic colleagues rather than repeated attempts under deepening sedation.

For a procedure at the opposite end of the frequency spectrum, the Emergency Medicine Journal has published an updated how-to guide for emergency thoracotomy from Chesters and colleagues [3], the first substantial refresh since the influential 2005 guide. It's a narrative practical guide rather than new data, integrating current international guideline recommendations with two decades of operational experience — refined indications, clearer survival windows, stepwise procedural detail, troubleshooting, and importantly, human factors and team safety. It is explicitly written for clinicians who may perform this once in a career or never. Worth printing for the resuscitation room wall.

And on rescue airways, Resuscitation brings a nationwide Danish cohort from Jensen and colleagues covering every prehospital emergency cricothyroidotomy in the country from 2016 to 2023 [7]. In eight years they found 71 validated cases. Among nearly forty thousand out-of-hospital cardiac arrests, the procedure was performed in about one in a thousand. Procedural success was high overall, close to nine in ten, and even higher — over ninety percent — in patients who were not in arrest. Aspiration and oropharyngeal bleeding were the leading indications, each about a third of cases. Survival, though, was sobering: among the arrest patients, a quarter achieved return of spontaneous circulation, but 30-day survival was under five percent and one-year survival around two percent. The message is not that the procedure fails — technically, in a physician-staffed system, it works. The message is that it marks a catastrophically ill population, and that because you will almost certainly never do this in routine practice, your competence has to come from deliberate simulation rather than clinical volume.

Turning to flow and efficiency, two papers push in opposite directions on the question of whether new technology actually speeds us up. In the American Journal of Emergency Medicine, Mahler and colleagues report a prospective three-site United States study of point-of-care high-sensitivity troponin, running samples simultaneously on a handheld analyser and in the central lab across 602 chest pain patients [5]. Turnaround time fell from just under an hour to about sixteen minutes — a difference of roughly 36 minutes. Based on surveyed physician disposition decisions, potential time to disposition was about fifty minutes shorter and potential length of stay about an hour shorter. The important caveat is in the word potential: the downstream time savings are modelled from what clinicians said they would do, not from observed dispositions, so treat the hour of length of stay as a ceiling rather than a promise.

Contrast that with Nature Medicine, which published a prospective evaluation of a large language model clinical decision support system deployed in a tertiary emergency department [9]. Leibovitch and colleagues ran a DECIDE-AI stage one evaluation across 1,138 patients in two parallel units over four weeks. The system's outputs were clinically sound — expert reviewers rated 99 of 100 sampled outputs as appropriate, and no adverse events were detected. But adoption collapsed over the study, falling from around two thirds of encounters to under a third, and disengagement tracked with workload — the longer into a shift, the less it was used. Length of stay was identical between the two wings at 4.9 hours, and the trend toward shorter consultation cycle time was not statistically significant. Physicians did preferentially reach for it for radiology consultations, roughly tripling the odds of use in that context. The authors are unusually blunt: these data inform randomised trial design but do not justify clinical deployment. The barrier isn't accuracy, it's sustained human engagement when the department is busy — which should reframe how your institution evaluates any artificial intelligence pitch.

Finally, three papers on core medical management and prognostication. Internal and Emergency Medicine published a systematic review and meta-analysis by Tesch and colleagues of nine randomised trials and 1,415 patients comparing the modified Valsalva manoeuvre with the standard one for stable supraventricular tachycardia [4]. The modified technique — supine repositioning with passive leg raise after the strain — converted about half of patients, versus under a fifth with the standard manoeuvre, roughly two and a half times the success rate, with no heterogeneity across trials and no major adverse events in either arm. That is about as clean an evidence base as vagal manoeuvres will ever get. If you are still doing a seated strain into a syringe and stopping there, change your practice.

The BMJ published a state-of-the-art review on acute hyperkalaemia from Rech and colleagues [6], synthesising epidemiology, pathophysiology, risk stratification and both acute and chronic management, including the newer potassium binding agents and high-risk populations. It's a narrative review, but it directly addresses the substantial practice variability that exists in how we define, risk-stratify and treat this condition. And in Resuscitation, Yver and colleagues report a multimodal prognostic model after paediatric cardiac arrest [10], testing whether day-one blood brain injury biomarkers and an MRI injury score add to clinical variables for predicting death or severe disability at one year. In the 118-child outcomes cohort, biomarkers did improve model fit and MRI added incremental value, with the best model achieving good discrimination — but the biomarker benefit did not hold in the full imputed cohort, and the authors are clear that external validation is required before any of this touches a family conversation.

If you only have time for one paper this week, make it the modified Valsalva meta-analysis in Internal and Emergency Medicine [4]. It is free, immediately actionable on your next shift, and the effect size on a common presentation is large enough to keep a meaningful number of patients away from adenosine.

Here are the key takeaways from this week in Emergency Medicine. First, the modified Valsalva converts roughly half of stable supraventricular tachycardia patients and outperforms the standard manoeuvre by about two and a half fold — it should be your default. Second, orthopaedic surgeons and radiologists agree only moderately on whether a distal radius reduction is adequate, and adequacy did not predict who went to surgery, so integrate clinical context rather than chasing the film. Third, point-of-care high-sensitivity troponin genuinely slashes turnaround time, but the downstream length of stay savings in this study were modelled from surveyed intentions, not observed. Fourth, a large language model decision support system in a real emergency department produced clinically appropriate output yet failed to change length of stay, with adoption falling away as shifts got busy — engagement, not accuracy, is the bottleneck. And fifth, prehospital cricothyroidotomy is technically successful nearly nine times in ten but is vanishingly rare, so your skill has to come from simulation. Alongside that, review the new ACEP procedural sedation consensus documents and the updated emergency thoracotomy guide.

That's your roundup for This Week in Emergency Medicine. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And think of one colleague — in any specialty — who never has time to keep up with the literature. Tell them about AudioScholar: a free ten-minute weekly for every specialty, to listen to in any podcast app, or to read at audioscholar dot C C.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.