This Week in Emergency Medicine — Aug 11, 2026

Welcome to This Week in Emergency Medicine. This week we're covering seven notable papers spanning post-cardiac-arrest critical care, risk stratification and decision rules in trauma, and equity and quality in acute pain and toxicology care. Let's dive in.

We start with two large trials that both ask whether more intensive post-arrest intervention actually helps, and both answer no. In the New England Journal of Medicine, the LOGICAL trial randomised 1,840 unresponsive, mechanically ventilated adults across 53 intensive care units in Australia, New Zealand and Ireland to conservative or liberal oxygen therapy after resuscitation [1]. Both arms shared a lower saturation limit of 90 percent; the conservative arm capped the upper alarm at 95 percent and weaned inspired oxygen down to room air where tolerated, while the liberal arm had no upper limit and a minimum inspired oxygen fraction of 0.3. At 180 days, survival with a favourable functional outcome on the Extended Glasgow Outcome Scale was 38.2 percent with conservative oxygen versus 39.7 percent with liberal oxygen, a relative risk of 0.97 with confidence intervals crossing one and a p-value of 0.65. This was a flatly neutral result, not a signal of benefit, and no adverse events were reported in either arm. The practical reading is that within these fairly generous boundaries, obsessive oxygen titration in the first hours after arrest is not what determines outcome, though avoiding frank hyperoxia remains sensible physiology. Alongside it, JAMA published ICECAP, an adaptive-allocation trial across 71 United States hospitals that randomised 1,158 comatose out-of-hospital arrest survivors, already cooled below 34 degrees within four hours, to hypothermia durations ranging from 6 to 72 hours [2]. Using a Bayesian duration-response model with a weighted modified Rankin Scale at 90 days, the trial stopped early at interim analysis. In the nonshockable cohort, which made up the large majority at 883 patients, the posterior probability that six hours — the shortest duration tested — was optimal was 0.51, and the shockable cohort looked similar. There were no differences in secondary outcomes or mortality across durations. Read together with LOGICAL, the message for the emergency clinician is consistent: prolonging or intensifying these particular interventions does not buy neurological recovery, and our attention is better spent on high-quality resuscitation, early coronary evaluation and avoiding fever and secondary insults.

Turning to risk stratification, PLOS Medicine reports a prospective multicentre cohort of 1,620 European patients aged 65 and older who presented to five emergency departments with mild traumatic brain injury after a ground-level fall and underwent computed tomography [3]. Mean age was nearly 85. Significant acute traumatic intracranial haemorrhage occurred in 72 patients, about 4.4 percent, and only five patients — three in a thousand — needed urgent neurosurgery. Eight predictors emerged as both strongly associated and reliably assessed: a visible forehead or scalp impact, a Glasgow Coma Scale below the patient's baseline, focal neurological deficit, signs of basal skull fracture, acute confusion, vomiting, loss of consciousness and headache. Two derived rules, PIWI 1 and PIWI 2, each achieved 100 percent sensitivity with specificities between roughly 25 and 44 percent, which would have avoided between a quarter and 42 percent of head CTs. The caveats matter: only scanned patients were enrolled, raising selection bias, there was no centralised imaging review, and this is internal validation only. So do not stop scanning anticoagulated octogenarians on the strength of this paper — but the visible impact location is a variable most of us do not systematically record, and it deserves attention. Prehospital Emergency Care takes a different approach to prediction, using more than two million 9-1-1 emergency medical services activations for motor vehicle crashes in the National Emergency Medical Services Information System to model who receives prehospital blood transfusion [5]. Random forest and ridge-penalised regression models discriminated well, with areas under the curve of about 0.90, driven by composite physiologic measures and shock designations, and projected that roughly 2 to 4 percent of crash patients nationally would be high-probability transfusion candidates. This is a systems-planning tool for deciding where blood products should live, not a bedside triage rule.

Our final theme is the quality and equity of acute symptom management. In the American Journal of Emergency Medicine, a systematic review and meta-analysis of six studies and 3,367 children with sickle cell vaso-occlusive episodes found no significant difference between early and delayed opioid administration for hospital admission, emergency department discharge, length of stay or pain reassessment, including in a subgroup analysis split at 60 minutes [4]. The authors note individual studies suggested benefit from timely repeat dosing and intranasal fentanyl, so the fair interpretation is not that speed is irrelevant, but that a single early dose without a structured re-dosing pathway may not change disposition. From the same journal, a ten-year retrospective analysis of 12,650 children undergoing facial laceration repair in two paediatric emergency departments found that non-Hispanic White children had modestly but significantly higher adjusted odds of receiving anxiolysis or sedation, about 17 percent higher than the grand mean, while preferred language showed no significant association [6]. That is a discretionary, comfort-related decision showing a racial gradient, and it argues for protocolised default anxiolysis rather than case-by-case judgement. Rounding out the journal, a scoping review of carbon monoxide-induced myocardial injury reports that 20 to 40 percent of patients with moderate to severe poisoning show biomarker or electrocardiographic evidence of myocardial injury, associated with a roughly twofold increase in long-term mortality and adverse cardiovascular events, often in young patients without known coronary disease [7]. The authors argue for routine electrocardiography, troponin and bedside echocardiography in significant exposures.

If you only have time for one paper this week, make it the JAMA ICECAP trial [2]. It resolves a question that has driven post-arrest protocols for two decades, and the answer — that longer cooling adds nothing — will reshape targeted temperature management pathways that emergency physicians initiate.

Here are the key takeaways from this week in emergency medicine. First, neither conservative oxygen targeting nor prolonged hypothermia improved functional outcomes after cardiac arrest; both trials were neutral, so focus on the fundamentals rather than protocol intensification. Second, in older adults after a ground-level fall, history and examination findings, including visible forehead impact, may safely reduce head CT use — but these rules need external validation before you change your scanning threshold. Third, in paediatric sickle cell crisis, simply giving the first opioid faster did not change admission or length of stay; build pathways around repeat dosing and intranasal routes. Fourth, sedation and anxiolysis for laceration repair varied by race and ethnicity, which is a strong argument for standardised protocols. And fifth, consider an electrocardiogram and troponin in anyone with significant carbon monoxide exposure, because cardiac injury is common and frequently silent.

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 if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

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