EM Evidence Rundown — Issue 32

EM EVIDENCE RUNDOWN — ISSUE 32 — 2 OCTOBER 2026

EM Evidence Rundown

Emergency medicine evidence for UK clinicians — weekly — emevidence.org

Jake Turner — Senior Registrar in Emergency Medicine, ST6 — Curated with the assistance of AI (Perplexity). All content editorially reviewed.

Highlight (Trauma): Emergency thoracotomy: how to do it in 2026 (EMJ practice review, UK authors) — the first full update of the 2005 Wise et al. guide, covering indications, a stepwise open-chest sequence, human factors and governance. Resuscitation: PARAMEDIC-3 Bayesian re-analysis — 81.6% probability that IO-first is worse than IV-first for 30-day survival, but any effect is probably modest. Post-ROSC ECG: about half of angiographically confirmed occlusions had no STEMI pattern. Change This Month: Hyponatraemia meta-analysis (n=27,672): faster correction associated with lower mortality, ODS remains rare. ED-to-ICU boarding beyond 4 hours associated with doubled in-hospital mortality. Paediatric EM (5 items): Neonatal and infant airway guideline (ESAIC/BJA), eFONA in paediatric trauma (TARN), triage ondansetron RCT, corticosteroids in paediatric sepsis, suicide risk after a positive ASQ screen. Core Revision: Serotonin syndrome vs neuroleptic malignant syndrome — the hot, rigid patient (Toxicology).

BOTTOM LINE UP FRONT — 2 OCTOBER 2026

ACT ON THIS NOW

HIGHLIGHT Emergency thoracotomy 2026 (EMJ): Use the ERC threshold (≤15 min from loss of vital signs). Crucifix arm position, clamshell, open the pericardium craniocaudally anterior to the phrenic nerves. Foley catheters for cardiac wounds are now discouraged.

RESUS PARAMEDIC-3 Bayesian (n=6,082): IV-first remains the preferred adult OHCA strategy; IO is the fallback, not the default. Probability IO-first is worse for ROSC: 98.0%.

CHANGE Post-ROSC ECG (n=214): NST-OMI patterns were as common as STEMI among confirmed occlusions, yet went to angiography less often (78% vs 98%). Read the first post-ROSC ECG for OMI, not just STEMI.

CHANGE ED-to-ICU delay >4 h (n=650): adjusted in-hospital mortality 27.5% vs 13.8%. Most of the delay accrued after ICU acceptance.

UK MHRA NatPSA/2026/005: patient hoists and slings — deaths and serious harm from falls. Check your department's hoist and sling compatibility and pre-use checks.

KNOW FOR NEXT TIME

RENAL Hyponatraemia correction (MA, n=27,672): rapid correction associated with lower mortality (OR 0.40 at 8 mmol/L/day threshold); ODS 0.26% vs 0.04%. Observational — do not abandon rate limits in chronic hyponatraemia.

DIAGNOSTICS LRINEC (n=359): AUC 0.663; sensitivity 42.9% at ≥6. A low score does not exclude necrotising infection.

PAEDS EM Neonatal/infant airway (ESAIC/BJA): videolaryngoscopy first, apnoeic oxygenation, limit attempts, capnography to confirm.

PAEDS EM eFONA in paediatric trauma (TARN/VSTR): 27 cases in 22 years; 2 of 7 TARN cricothyroidotomy patients survived. Know your department's paediatric CICO plan before you need it.

CORE REVISION Serotonin syndrome: clonus is the key sign (Hunter criteria). NMS: lead-pipe rigidity, slower onset, dopamine antagonist exposure.

This week's highlight is a practice review that a generation of UK emergency physicians will recognise: twenty-one years after Wise and colleagues published "Emergency thoracotomy: how to do it", a Lincs & Notts Air Ambulance-led team has rewritten it for 2026. Alongside it, the PARAMEDIC-3 team's Bayesian re-analysis puts probabilities on the IO-versus-IV question, and two papers on cardiac arrest aftercare and ICU boarding speak directly to resus room decisions we make every shift.

WHAT'S INSIDE — ISSUE 32

• Highlight: Emergency thoracotomy — how to do it in 2026 (EMJ)
• Resus: PARAMEDIC-3 Bayesian re-analysis — IO-first vs IV-first (Resuscitation)
• Resus: The post-ROSC ECG — NST-OMI patterns and prognosis (Resuscitation)
• Renal: Sodium correction rate and mortality in hyponatraemia — meta-analysis (EMJ)
• Systems: ED-to-ICU transfer delay and mortality (J Emerg Med)
• POCUS: Focused echo for acute PE — UK-led meta-analysis (Eur J Emerg Med)
• Diagnostics: LRINEC and other NSTI scores fail validation (J Emerg Med)
• Analgesia: Inhaled methoxyflurane for trauma pain — meta-analysis (J Emerg Med)
• Quick hits: high-dose GTN in SCAPE; sepsis symptoms and antibiotic timing; NRB plus nasal cannula preoxygenation; endovascular resuscitation and CFA access reviews
• UK: MHRA hoist and sling NatPSA; MHRA September safety roundup; Scotland alcohol, drugs and self-harm pathways (EMJ); trauma teams in trauma units (EMJ)
• PEM: Neonatal and infant airway guideline; eFONA in paediatric trauma; triage ondansetron RCT; corticosteroids in paediatric sepsis; youth suicide risk after positive ASQ
• FOAMed: JournalFeed — tertiary trauma survey in moderate trauma; inhaled isopropyl alcohol BET

• Core Revision: Serotonin syndrome vs NMS (Toxicology)

Contents: 1. Key Trials & Articles — 2. Guidelines & UK Updates — 3. Paediatric Emergency Medicine — 4. FOAMed & Critical Appraisal — 5. Action Points — 6. Trials to Watch — 7. Core Revision: Serotonin Syndrome vs NMS (Toxicology)

1 — KEY TRIALS & ARTICLES

HIGHLIGHT ARTICLE — IN DEPTH

LEAD PRACTICE REVIEW UK AUTHORS FRCEM

Emergency Thoracotomy: How To Do It in 2026 — The Wise Guide, Updated After Two Decades of UK Practice

Chesters A, Hughes M, Ley E, O'Meara M, Davies G, Hormis A. Emergency thoracotomy: how to do it in 2026. Emerg Med J. 2026 (epub ahead of print; accepted 2 August 2026). doi:10.1136/emermed-2026-216110. Authors from Lincs & Notts Air Ambulance, Cambridge University Hospitals, Leeds Teaching Hospitals, University Hospitals of North Midlands and Sheffield Hallam University. No specific funding; no competing interests declared.

Why this matters: In 2005, Wise and colleagues published "Emergency thoracotomy: how to do it", describing a simplified clamshell technique that non-cardiothoracic doctors could perform in the resus room or at the roadside. It became the reference text for UK emergency medicine and prehospital practice. This 2026 update integrates current ERC, EAST and Western Trauma Association guidance with two decades of operational experience. Where evidence is limited, the authors state explicitly that recommendations reflect their collective experience.

Who should receive an emergency thoracotomy? The authors recommend the ERC 2025 time-based threshold: thoracotomy for traumatic cardiac arrest where loss of vital signs occurred up to 15 minutes before, modified where appropriate by signs of life (EAST) and CPR duration (WTA). Prehospital series consistently show that patients with cardiac tamponade are more likely to survive if thoracotomy is performed within 15 minutes of witnessed loss of signs of life, whereas those who exsanguinate rarely survive beyond 5 minutes. The authors' experience is that the greatest benefit may come when thoracotomy is started before complete loss of cardiac output — and they predict future practice will focus on earlier intervention rather than extending post-arrest thresholds.

Selected rows from table 1 of Chesters et al. (2026). The full table also includes Matsumoto 2009 (Japan, blunt, n=81, 0 survivors), Van Vledder 2017 (Netherlands, n=33, 1 survivor), EAST 2015, WTA 2024 and Ter Avest 2025.

Before the chest is opened — what has changed since 2005:

Once the chest is open — the structured sequence (table 2 of the paper):

After the chest is open — key decisions: Survival in exsanguination is extremely rare (1.9% in a large London series) and there were no survivors when tamponade and exsanguination coexisted. The authors name the emotional investment of the team as a recognised source of bias toward continuing beyond the point of benefit, and recommend senior remote advice to support objective decisions. Ventricular fibrillation in trauma usually reflects critically impaired coronary perfusion rather than a primary arrhythmia, so haemorrhage control, aortic compression, volume and effective internal massage come first. Outside hospital, external defibrillation can be attempted after partial closure of the clamshell, with pads placed so the anterior pad does not cross the incision. In the ED, ATLS 10th edition describes internal paddles at 10–20 J, increasing to a maximum of 40 J. After ROSC, expect brisk internal mammary bleeding, partially close the clamshell with the retractor loosened in situ for re-entry, and actively prevent heat loss. With ongoing aortic compression, consider intermittent occlusion (for example 45 seconds on, 15 seconds off), and avoid over-resuscitation in isolated tamponade, where patients may be euvolaemic. Intracardiac adrenaline into the right ventricle is described as a rescue option only when IV or IO access is impossible; no major guideline body recommends it.

Human factors: A minimum of two clinicians with defined roles — one operator focused exclusively on the surgical field, and a second managing ventilation timing, instruments, a continuous sharps and swab count, and the wider scene. Sharps and blood-exposure risk is specific and high (Gigli saws, shears and rib edges in a confined, poorly lit space). Brief unfamiliar team members beforehand and debrief afterwards, because the sight of an open chest can be acutely distressing. Train together with the equipment you will actually use.

Critical appraisal: This is an expert practice review, not a systematic review or trial, and the authors are transparent that many recommendations rest on operational experience. Its strengths are a clearly reported literature search (online supplement), alignment with current ERC, EAST and WTA guidance, and authorship from high-volume UK prehospital and ED practice. The supporting outcome data come from observational case series with selection and survivorship bias, small denominators (n=33 to 601) and heterogeneous definitions of timing and "signs of life"; the much-quoted survival windows are associations from these series, not thresholds validated prospectively. Several technical preferences (sutures over staples, discouraging Foley catheters and intracardiac cannulation) are reasoned opinion that some services do not share. The 15-minute ERC threshold widens the window compared with the 10 minutes used by Wise et al. and the 2011 London series, which may increase the number of procedures performed with a very low chance of survival; the paper's emphasis on mechanism (tamponade versus exsanguination) is the important counterweight. Figure 2 in the paper was generated with an AI image tool and verified by the authors, which they disclose. Overall, this is the most useful single reference currently available for UK clinicians who may perform this procedure rarely.

Governance (four foundations recommended by the authors): (1) a written checklist in three parts — before, during and after the procedure — aligned with the WHO Surgical Safety Checklist and NatSSIPs in England; (2) a service-specific SOP including authorisation criteria and training requirements; (3) immediate access to senior advice, mandatory debrief and audit of every case; (4) an equipment agreement with a local sterile services department, including rapid replacement of used kits.

UK ED practice: For a penetrating traumatic cardiac arrest arriving in your ED within 15 minutes of loss of vital signs, the decision to open the chest should be fast and pre-planned. Check now: where is your thoracotomy set, does it contain Tuff-cut shears, a Gigli saw, Finochietto retractors and non-crushing clamps, and who reassembles the retractor after sterilisation? Does your trauma unit have a thoracotomy SOP and a three-part checklist, and is every case debriefed and audited? Consider an in-situ simulation using the crucifix position and the paper's seven-step sequence. If you are a trauma unit rather than a major trauma centre, agree in advance with your network how a patient with ROSC after thoracotomy will be transferred.

CHANGE THIS MONTH BAYESIAN SECONDARY ANALYSIS OF RCT UK TRIAL FRCEM

PARAMEDIC-3 Bayesian Re-analysis: IO-First Access Is More Likely Harmful Than Beneficial in Adult OHCA, but Any Effect Is Probably Modest

Couper K, Ji C, Deakin CD, Fothergill RT, Kandiyali R, Nolan JP, Pocock H, Smyth MA, Wiles JV, Lall R, Perkins GD. Intraosseous versus intravenous access in adult out-of-hospital cardiac arrest: a Bayesian secondary analysis of the Paramedic-3 randomised controlled trial. Resuscitation. 2026 (journal pre-proof). doi:10.1016/j.resuscitation.2026.111335. ISRCTN14223494. Warwick Medical School with London, South Central and West Midlands ambulance services.

Design: Pre-planned Bayesian secondary analysis of PARAMEDIC-3, a UK multi-ambulance-service RCT (November 2021 to July 2024) that randomised adults in OHCA needing vascular access to an IO-first or IV-first strategy. The primary analysis used non-informative priors, adjusted for key baseline characteristics; sensitivity analyses used sceptical and enthusiastic priors. Outcomes: 30-day survival (primary), favourable neurological outcome at discharge, and ROSC at hospital handover.

Key results: 6,082 patients (IO 3,040; IV 3,042); 79% non-shockable rhythms, 79% at home. In the IO-first group 94.4% received IO first; in the IV-first group 64.6% received IV first; crossover 8.7%. Overall ROSC at handover 23.1%, 30-day survival 4.8%, favourable neurological outcome 2.8%. With non-informative priors, the probability that IO-first was worse than IV-first was 81.6% for 30-day survival, 61.8% for favourable neurological outcome and 98.0% for ROSC at handover. The probability of a moderate harm (OR <0.75) for 30-day survival was only 13.4%, and of a large harm (OR <0.67) 3.2%.

• 6,082 ADULTS WITH OHCA RANDOMISED (UK)
• 81.6% PROBABILITY IO-FIRST WORSE FOR 30-DAY SURVIVAL (OR <1.0)
• 98.0% PROBABILITY IO-FIRST WORSE FOR ROSC AT HANDOVER
• 13.4% PROBABILITY OF A MODERATE HARM (OR <0.75) FOR SURVIVAL

Critical appraisal: This is a large, pragmatic, UK-delivered RCT with a Bayesian analysis planned in the statistical analysis plan, although the specific priors were finalised after the frequentist results were published. The Bayesian framing is useful because the original frequentist analysis found point estimates favouring IV for all outcomes but statistical significance only for ROSC; this re-analysis expresses that as probabilities clinicians can use. Important caveats: only 64.6% of the IV-first group actually received IV first, so the comparison is of strategies, not routes, and the true difference between routes may be larger than estimated. Results were sensitive to the prior: with enthusiastic priors favouring IO, the probability that IO-first improved 30-day survival rose to 78.0%, so readers with strong prior belief in IO would not be moved much. Absolute survival was low (4.8%), so even a real effect translates into small absolute numbers. The anatomical IO site was not protocolised.

UK ED practice: In adult cardiac arrest, attempt IV access first where it can be obtained promptly, with IO as the rescue route rather than the default. This matters most for patients arriving with IO-only access from the ambulance service: consider establishing IV access early in the resus room without interrupting compressions, especially before repeat adrenaline doses. Align local ALS teaching with RCUK guidance and your ambulance trust's protocol.

CHANGE THIS MONTH RETROSPECTIVE COHORT ECG FRCEM

The Post-ROSC ECG: Half of Confirmed Coronary Occlusions Had No STEMI Pattern — and Went to the Cath Lab Less Often

Silwanis C, Groche M, Huss M, et al. The post-ROSC ECG: morphological spectrum of non-ST-elevation occlusion and prognostic markers after cardiac arrest. Resuscitation. 2026 (journal pre-proof). doi:10.1016/j.resuscitation.2026.111334.

Design: Retrospective single-centre analysis of the first post-ROSC ECG in 214 cardiac arrest survivors admitted to a tertiary ICU (2020–2025), with coronary angiography as the reference standard for occlusion myocardial infarction (OMI).

Key results: 101 of 214 had angiographically confirmed OMI. NST-OMI patterns were more prevalent than STEMI patterns overall (35% vs 24%) and among confirmed OMI (48% vs 44%), yet patients with NST-OMI patterns underwent angiography less often (78% vs 98%; p=0.002). The highest 30-day mortality was in NST-OMI patients who did not undergo angiography (75%; p=0.005). The commonest NST-OMI patterns were left main equivalent and Smith-modified Sgarbossa (8% each), Aslanger pattern (5%), and lateral OMI, shark-fin and hyperacute T waves (4% each). PR-segment prolongation was independently associated with reduced 30-day survival (aOR 0.977 per ms, 95% CI 0.956–0.998).

• 101/214 ANGIOGRAPHICALLY CONFIRMED OMI
• 48% vs 44% NST-OMI VS STEMI PATTERN AMONG CONFIRMED OMI
• 78% vs 98% ANGIOGRAPHY RATE: NST-OMI VS STEMI PATTERN
• 75% 30-DAY MORTALITY: NST-OMI WITHOUT ANGIOGRAPHY

Critical appraisal: The finding that STEMI criteria alone miss around half of post-ROSC occlusions is consistent with the wider OMI literature and is clinically important. However, this is a small, retrospective, single-centre study of ICU-admitted survivors, so it excludes patients who died early and those not admitted to ICU (selection bias). ECG interpretation was retrospective, and the paper does not establish whether the patterns were recognised in real time. The very high mortality in NST-OMI patients without angiography is confounded: those not taken to the cath lab were probably sicker, older, or judged non-survivable, so the result should not be read as angiography causing survival. The PR-segment finding is a small per-millisecond effect needing external validation before it has any role in prognostication.

UK ED practice: Read every first post-ROSC ECG specifically for OMI equivalents — Sgarbossa (modified), left main equivalent pattern, Aslanger, hyperacute T waves and shark-fin morphology — not just for STEMI criteria. If an OMI pattern is present and the arrest is plausibly cardiac, discuss with the interventional cardiology team rather than relying on the absence of ST elevation. Repeat the ECG after 15–20 minutes, as early post-ROSC ECGs can be misleading.

CHANGE WHEN GUIDELINE UPDATES SYSTEMATIC REVIEW & META-ANALYSIS FRCEM

Sodium Correction Rate in Hyponatraemia: Faster Correction Associated With Lower Mortality Across 27,672 Cases — ODS Remains Rare

Matrisch L, Rau Y, Graßhoff L, Nitschke M. On the optimal sodium correction rate in hyponatraemia and clinical outcome: a meta-analysis. Emerg Med J. 2026;43:647–654. doi:10.1136/emermed-2025-215189.

Design: Systematic review and random-effects meta-analysis of 11 retrospective studies (PubMed and Web of Science, August 2024), each scoring at least 8 on the Newcastle-Ottawa scale. Studies were grouped by the threshold separating slow from rapid correction: 8, 10 or 12 mmol/L per 24 hours. Osmotic demyelination syndrome (ODS) was pooled separately.

Key results: At the 8 mmol/L/day threshold, mortality was 10.0% with slow correction (1,836/18,310) vs 5.8% with rapid correction (283/4,843); OR 0.398 (95% CI 0.29–0.547), with substantial heterogeneity (I²=66%). At 10 mmol/L/day, OR 0.489 (95% CI 0.376–0.637; I²=0%); at 12 mmol/L/day, OR 0.570 (95% CI 0.440–0.738; I²=0%). ODS occurred in 8/18,718 (0.04%) slow vs 13/4,987 (0.26%) rapid; OR 3.96, p=0.002.

• 27,672 PATIENTS IN 11 RETROSPECTIVE STUDIES
• 5.8% vs 10.0% MORTALITY: RAPID VS SLOW (8 MMOL/L/DAY THRESHOLD); OR 0.40
• 0.26% vs 0.04% ODS: RAPID VS SLOW; OR 3.96
• ARD 4.2pp CRUDE MORTALITY DIFFERENCE (8 MMOL/L THRESHOLD, NOT ADJUSTED)

Critical appraisal: This is the largest synthesis on the question, and the direction is consistent across all three thresholds. But every included study is retrospective, so the association is heavily exposed to confounding by indication: patients whose sodium corrects quickly tend to have reversible causes (thiazides, polydipsia, hypovolaemia), whereas slow correction often reflects severe illness such as heart failure, cirrhosis or malignancy. The authors acknowledge that no causal conclusion is possible. Correction rates were defined and reported inconsistently across studies, and acute and chronic hyponatraemia were not separated. The crude absolute difference above is from pooled raw counts and should not be read as an NNT. ODS was rare in both arms, but it is a devastating, largely irreversible harm, and the six-fold relative increase is the more robust signal because the biological mechanism is established.

UK ED practice: This paper does not change ED management today. For severe symptomatic hyponatraemia (seizures, reduced consciousness), give hypertonic saline boluses as per your local and Society for Endocrinology guidance, aiming for an initial rise of around 5 mmol/L. For chronic or asymptomatic hyponatraemia, continue to avoid over-correction — especially in high-risk groups (alcohol use disorder, malnutrition, hypokalaemia, liver disease) — and hand over a clear sodium monitoring plan. The take-home is that fear of ODS should not delay treatment of symptomatic hyponatraemia.

CHANGE THIS MONTH RETROSPECTIVE COHORT SYSTEMS

ED-to-ICU Transfer Delay Over 4 Hours Associated With Doubled In-Hospital Mortality — Most of the Delay Comes After ICU Acceptance

Kabahasanoglu K, Bektas MH, Ozturk Ormeci G, Beyoglu R. Emergency department-to-intensive care unit transfer delay and mortality in critically ill adults: a 5-year retrospective cohort study. J Emerg Med. 2026;89:140–152. doi:10.1016/j.jemermed.2026.08.013.

Design: Five-year retrospective cohort (2020–2024) of adults transferred from the ED to two level-3 general ICUs at a single tertiary hospital in Turkey. Primary models used robust Poisson regression adjusted for early ED severity, with supportive analyses of decision-to-bed boarding, continuous delay, matching, and COVID-negative cohorts.

Key results: 650 adults analysed; 351 (54.0%) had total delay over 4 hours. Median total transfer time was 4.3 hours: 1.3 hours before the ICU decision and 3.0 hours after ICU acceptance. Delay over 4 hours was associated with higher in-hospital mortality (aRR 2.01, 95% CI 1.46–2.76) and ICU mortality (aRR 1.70, 95% CI 1.20–2.42). Adjusted in-hospital mortality was 27.5% vs 13.8%. Post-decision boarding over 4 hours: aOR 1.82 (95% CI 1.21–2.73); each boarding hour aOR 1.14 (95% CI 1.06–1.24).

• 54% OF ADULTS REACHING A STUDY ICU BED: TOTAL ED-TO-ICU TIME >4 H
• 27.5% vs 13.8% ADJUSTED IN-HOSPITAL MORTALITY, DELAY >4 H VS ≤4 H
• ARD +13.7pp ADJUSTED RISK DIFFERENCE (OBSERVATIONAL, NOT AN NNH)
• 3.0 h MEDIAN WAIT AFTER ICU ACCEPTANCE

Critical appraisal: The decomposition of delay into pre- and post-decision time is the paper's main contribution, and the per-hour dose-response for boarding supports a real relationship. However, this is single-centre, retrospective and observational: residual confounding is likely (patients who wait may differ in ways severity scores do not capture), and the cohort includes only patients who reached an ICU bed, excluding those who died in the ED or improved. The setting differs from the NHS, where boarding is often longer and outreach models vary. The adjusted risk difference is large and should be treated as an upper estimate. It is nonetheless consistent with UK data on long ED stays and mortality reported by RCEM.

UK ED practice: Treat an accepted ICU patient waiting in resus as an ICU patient: agree a named ICU clinician, a documented management plan, and escalation triggers at the point of acceptance. Record decision-to-bed times in your department and report long waits through your incident system — these are the data that support escalation to site and system level.

INFORMING SYSTEMATIC REVIEW & META-ANALYSIS UK AUTHORS POCUS

Focused Echo for Suspected PE: Specific Signs Rule In, but Nothing Rules Out — UK-Led Meta-analysis of 33 Studies

Mani N, Morris E, Tahvili A. Focused transthoracic echocardiography in acute pulmonary embolism: a systematic review and meta-analysis of diagnostic accuracy. Eur J Emerg Med. 2026 (epub 24 September). doi:10.1097/MEJ.0000000000001378. PMID: 42779550. POCUSUK Group, Chesterfield Royal Hospital, UCL and University of Sheffield.

Design and results: Bivariate random-effects meta-analysis of 33 studies (3,981 adults) with suspected acute PE in acute care settings. Individual focused echo signs had poor sensitivity (0.03–0.65) but high specificity: McConnell's sign 0.98, 60/60 sign 0.93, RV outflow tract early systolic notching 0.98. Combining focused echo with lower-limb DVT ultrasound gave pooled sensitivity 0.62 and specificity 0.98.

• 33 / 3,981 STUDIES / PATIENTS
• 0.98 SPECIFICITY: MCCONNELL'S SIGN AND RVOT NOTCHING
• 0.62 SENSITIVITY: ECHO PLUS DVT ULTRASOUND COMBINED

Critical appraisal: A useful UK synthesis, but the included studies vary in operator experience, population (many enriched for large or haemodynamically significant PE) and reference standard, which can inflate specificity. The abstract does not report heterogeneity statistics or likelihood ratios. Spectrum bias is likely: in stable ED patients with small or subsegmental PE, sensitivity will be even lower than reported.

UK ED practice: Use focused echo in the shocked or peri-arrest patient where a positive finding (RV dilatation with McConnell's sign, 60/60, RVOT notching, clot in transit) can support thrombolysis decisions while CTPA is unsafe. Never use a normal focused echo to rule out PE in a stable patient — follow NICE NG158 with Wells, D-dimer and CTPA.

CHANGE THIS MONTH RETROSPECTIVE COHORT FRCEM

LRINEC and Related Scores Fail as Stand-Alone Tests for Necrotising Soft Tissue Infection — Sensitivity 42.9% at the Classic Threshold

Shiau HM, Wang HS, Hung SK, Chang CP, Lin LC, Wu KH. Validation of diagnostic scoring systems in patients with clinical suspicion of necrotizing soft tissue infection: a cohort study. J Emerg Med. 2026. doi:10.1016/j.jemermed.2026.08.031.

Design and results: Retrospective cohort at a Taiwanese tertiary centre (2018–2022) of adults with clinically suspected NSTI who had surgical consultation: 359 patients, 214 with NSTI and 145 with non-necrotising infection. AUCs: LRINEC 0.663, modified LRINEC 0.693, Japanese LRINEC 0.720, SIARI 0.486. LRINEC ≥6 had sensitivity 42.9% and specificity 85.0%.

• 359 SUSPECTED NSTI WITH SURGICAL REVIEW (214 CONFIRMED)
• 0.663 LRINEC AUC
• 42.9% SENSITIVITY OF LRINEC ≥6
• 85.0% SPECIFICITY OF LRINEC ≥6

Critical appraisal: The cohort is the right one to test — patients in whom clinicians were already worried — which makes it more clinically relevant than derivation cohorts that compared NSTI with simple cellulitis. NSTI prevalence was high (60%), retrospective data collection risks incomplete laboratory values, and the single Asian centre may limit generalisability. Still, these results match previous validation studies showing poor sensitivity.

UK ED practice: Do not use a low LRINEC score to reassure yourself or a surgical colleague. Pain out of proportion, rapid progression, systemic toxicity or crepitus warrant urgent senior surgical review, broad-spectrum antibiotics per local guidance (including clindamycin for toxin suppression where advised), and no delay for imaging if the patient is deteriorating.

INFORMING SYSTEMATIC REVIEW & META-ANALYSIS ANALGESIA

Inhaled Methoxyflurane for Trauma Pain: Faster Relief and Less Rescue Analgesia, at the Cost of More Dizziness

Husain AZ, Alsahli ASM, Alghamdi AMA, et al. Efficacy of inhaled methoxyflurane for acute trauma pain management: a systematic review and meta-analysis. J Emerg Med. 2026 (journal pre-proof). doi:10.1016/j.jemermed.2026.09.008.

Design and results: Seven RCTs (1,585 participants), comparators including placebo, standard analgesic care, tramadol and morphine. Methoxyflurane shortened time to first pain relief (MD −5.93 minutes, 95% CI −6.86 to −5.00; I²=0%), improved pain scores at 5, 10, 15 and 20 minutes, and reduced rescue analgesia (RR 0.40, 95% CI 0.18–0.90). Dizziness and drowsiness were more frequent (RR 6.64 across 5 studies).

• 7 RCTs / 1,585 TRIALS / PARTICIPANTS
• −5.9 min TIME TO FIRST PAIN RELIEF VS CONTROL
• RR 0.40 RESCUE ANALGESIA (95% CI 0.18–0.90)
• RR 6.64 DIZZINESS AND DROWSINESS

Critical appraisal: Four of seven trials were at high risk of bias, and several compared methoxyflurane with placebo, where superiority is expected; the comparative-effectiveness question against established analgesia is less certain. Only one trial predominantly enrolled children, so paediatric conclusions are limited. The rescue-analgesia confidence interval is wide. The time-to-relief finding is the most robust result.

UK ED practice: Methoxyflurane (Penthrox) remains a reasonable bridge analgesic at triage and for short procedures in adults with moderate to severe trauma pain while IV access or a regional block is arranged. Follow the SmPC limits (maximum two 3 mL inhalers in 24 hours), avoid in renal impairment and in patients with malignant hyperthermia susceptibility, and plan the next analgesic step because the effect is short-lived.

QUICK HITS

Also This Week

High-dose GTN in sympathetic crashing acute pulmonary oedema (SCAPE) — meta-analysis. Six studies (n=534 in comparative analysis): no difference in intubation (RR 0.91, 95% CI 0.11–7.73) or ICU admission (RR 0.92, 95% CI 0.81–1.05) with high-dose (≥100 microgram/min or bolus) vs conventional GTN, but more hypotension (RR 1.41, 95% CI 1.02–1.94). Mixed RCT and observational data, variable comparators. High-dose GTN with early CPAP remains reasonable in hypertensive SCAPE; watch the blood pressure. Am J Emerg Med. 2026;110:453–465. PMID 42784941.

Presenting symptoms and sepsis antibiotic timing (JAMA Netw Open, n=29,647, 67 US hospitals). Gastrointestinal symptoms were associated with lower odds of timely antibiotics (with hypotension aOR 0.84; without aOR 0.80), prehospital fever with higher odds. After adjusting for symptoms, timely antibiotics were still associated with lower 30-day mortality (adjusted absolute difference −2.10 percentage points, 95% CI −3.01 to −1.10). Think sepsis in the vomiting, diarrhoeal older patient. PMID 42804206.

NRB plus nasal cannula is inferior to a flush-rate NRB for preoxygenation. Crossover study in 30 healthy volunteers: mean FeO2 85% with flush-rate NRB vs 68% (NRB + nasal cannula each at 15 L/min) and 75% (each at 25 L/min). Volunteer physiology, not patients, but a clear signal: if you can flush the NRB (≥50 L/min from a wall flowmeter), do so before RSI. Prehosp Emerg Care. 2026. PMID 42784740.

Endovascular resuscitation and common femoral artery access — two EMJ practice reviews. The EVR review (Adams et al., EMJ 2026;43:626–633, doi:10.1136/emermed-2025-215376) summarises REBOA, SAAP and ECPR, including UK-REBOA (90-day mortality 54% REBOA vs 42% standard care; posterior probability of harm 86.9%; stopped early) and the recruiting UK ERICA-ARREST pilot of prehospital REBOA in OHCA. The companion CFA review (EMJ 2026;43:572–576, doi:10.1136/emermed-2025-215295) argues that ultrasound-guided CFA access should be a deliberately practised EM skill, as it is the gateway to all endovascular techniques. Neither changes ED practice today; both are good orientation reading for registrars.

2 — GUIDELINES & UK UPDATES

MHRA NATIONAL PATIENT SAFETY ALERT

NatPSA/2026/005/MHRA: Patient Hoists and Slings — Risk of Death and Serious Harm From Falls

MHRA Safety Roundup: September 2026 (29 September 2026). gov.uk/drug-device-alerts/mhra-safety-roundup-september-2026.

Falls from hoists and slings continue to cause deaths (an average of two per year since 2015) across all manufacturers and settings. Causes include incompatible hoist and sling combinations, damaged or incorrectly seated sling loops, ineffective pre-use checks, overdue maintenance, incorrect sling size and insufficient training. The same roundup flags: Fluenz product information updated for increased risk of fainting immediately after nasal-spray administration; parenteral nutrition must be given through a filter of no larger than 1.2 µm after a fatal embolism; and RF interference with wireless CTG monitors (DSI/2026/012).

UK ED practice: Ask your nurse in charge or matron to confirm your ED has acted on the NatPSA: hoist and sling compatibility, pre-use checks and staff competence for bariatric and immobile patients, including those held long term in the department.

MHRA CLASS 4 MEDICINES DEFECT

Levetiracetam Neuraxpharm Granules (Sachets) — Patient Leaflet Missing Safety Information, EL(26)A/40

MHRA, 1 October 2026. gov.uk drug and device alert EL(26)A/40.

The product is unaffected and supply continues, but the leaflet omits warnings. Patients on these sachets presenting with fever, rigidity and altered consciousness (possible neuroleptic malignant syndrome), or with fever, rash, facial swelling and lymphadenopathy within 2–8 weeks of starting (possible DRESS), should be assessed urgently. Relevant to children and adults with swallowing difficulties.

INFORMING UK DATA NATIONAL LINKED DATA

Alcohol, Drugs and Self-Harm in Scotland's Unscheduled Care System: 2.7% of 2.7 Million Care Pathways, Top 30 Pathways Cost About £217 Million

Schölin L, McStravick M, Lloyd A, Brett J, Eddleston M, Humphries C. Alcohol, drugs and self-harm emergencies in a national unscheduled care system: a retrospective observational study. Emerg Med J. 2026 (epub 23 September). doi:10.1136/emermed-2025-215639. PMID: 42778343.

Linked data from every unscheduled care service in Scotland (NHS 24, out-of-hours primary care, ambulance, ED, acute and mental health admissions) for April 2022 to March 2023: 2.7% of 2,721,120 pathways involved alcohol, drugs and/or self-harm. The commonest journey was ambulance to ED to acute admission, with directly attributable cost of about £216.9 million for the top 30 pathways. Case identification relied on service-level flags and clinical codes, which the authors say need refining.

UK ED practice: Ambulance services and EDs feature in most of these pathways. Make sure your department has accessible alcohol care team, liaison psychiatry and drug-service referral routes, and code these presentations accurately — the data drive commissioning.

INFORMING UK SERVICE EVALUATION TRAUMA

Trauma Teams in Trauma Units: Are They Fit for Purpose? Only 2 of 10 Site Trauma Leads Said Yes

Carroll R, Lee JC, Raitt J. Trauma teams in trauma units: are they fit for purpose? Emerg Med J. 2026 (epub). doi:10.1136/emermed-2026-216024. South West London and Surrey Trauma Network.

Survey across all seven trauma units in the network (February–March 2025): 151 responses (10 site trauma leads, 65 trauma team leaders, 76 team members). Only 20% of site trauma leads felt the current system worked for their hospital, although 67.6% of team leaders were confident in their team. 36% of team members had no formal trauma training and 42% had attended only 1–2 trauma calls in the preceding 3 months; 32% of leaders reported teams were often missing members. Respondents favoured senior-led, multidisciplinary teams and several supported a tiered response.

Appraisal: Convenience and snowball sampling, a respondent pool dominated by EM staff, and distribution via site leads mean the results may not be representative. It is nevertheless a candid picture of skill decay in low-volume trauma units that will be familiar outside London.

UK ED practice: If you lead trauma in a trauma unit, consider a tiered activation model, regular in-situ multidisciplinary simulation, and a requirement that specialty representatives at trauma calls can make decisions or escalate immediately.

3 — PAEDIATRIC EMERGENCY MEDICINE

CHANGE THIS MONTH PAEDS EM GUIDELINE SUMMARY FRCEM

Neonatal and Infant Airway Management: The ESAIC/BJA Guideline for Children Under 1 Year, Translated for the ED

Long B, Gottlieb M. Guideline update: neonatal and infant airway management. Am J Emerg Med. 2026;109:259–262. doi:10.1016/j.ajem.2026.07.056. Summarises the 2024 European Society of Anaesthesiology and Intensive Care / British Journal of Anaesthesia joint guideline.

Key recommendations: predict difficulty from history and examination (micro- or retrognathia, facial asymmetry, limited mouth opening, cleft, syndromes such as Pierre Robin and Treacher Collins); use adequate anaesthesia (for example ketamine or etomidate) with neuromuscular blockade where maintaining spontaneous breathing is not necessary; videolaryngoscopy as first choice; apnoeic oxygenation during tracheal intubation of neonates; supraglottic airway for rescue oxygenation; limit attempts, changing technique, provider or both after each failure; stylet for hyperangulated blades or anterior airways; confirm with end-tidal capnography; and explicit attention to human factors. Neonatal and infant first-pass success rates are reported at 30–57%, with peri-intubation adverse events in 9–50%.

Appraisal: Many recommendations are graded on low-quality evidence, mostly from operating-theatre studies; ED evidence in this age group is sparse. The summary is US-authored, but the underlying guideline is a joint ESAIC/BJA document, so it aligns well with UK paediatric anaesthesia practice.

UK ED practice: Infant intubation in the ED should be a team event with the most experienced available airway operator, often paediatric anaesthesia or PICU/retrieval. Check that a videolaryngoscope with neonatal and infant blades, size 1 supraglottic airways and paediatric capnography are in your resus area, and rehearse the "change something after each failed attempt" rule in paediatric simulation.

INFORMING PAEDS EM UK REGISTRY (TARN)

Emergency Front of Neck Access in Paediatric Trauma: 27 Cases in Two Trauma Registries Over 22 Years

McDonald R, Coffey M, Lecky F, Mclellan S, Cameron P, Mullen S. Emergency front of neck access in the setting of paediatric trauma (letter). Emerg Med J. 2026;43:662–664. doi:10.1136/emermed-2025-215688. Royal Belfast Hospital for Sick Children and TARN.

TARN (UK and Ireland, 2000–2022) and the Victorian State Trauma Registry (2001–2022) recorded 27 eFONA procedures in children under 16 within 24 hours of injury: 20 tracheostomies (15 TARN, 5 VSTR) and 7 cricothyroidotomies (all TARN, 2 of them prehospital). Only 2 of the 7 TARN cricothyroidotomy patients survived. Guidance differs by age and organisation: DAS recommends surgical cricothyroidotomy over 8 years; APLS UK suggests emergency surgical tracheostomy under 1 year, tracheostomy or surgical cricothyroidotomy at 1–5 years, and notes surgical cricothyroidotomy has a higher success rate over 5 years.

Appraisal: A research letter with very small numbers, missing TARN tracheostomy outcome data after a cyber attack, and registry inclusion criteria that may miss cases. It cannot tell us which technique is best — only that the event is extraordinarily rare and that guidance is inconsistent.

UK ED practice: Because no ED clinician will gain experience of paediatric eFONA, your department needs a written, age-banded paediatric CICO plan (aligned with APLS UK and DAS), the right kit, and an ENT escalation route agreed in advance.

INFORMING PAEDS EM DOUBLE-BLIND RCT

Triage Nurse-Initiated Ondansetron for Gastroenteritis: No Difference in Early Discharge, but the Trial Was Underpowered

Weill O, Lucas N, Bailey B, Marquis C, Gravel J. Triage administration of ondansetron for gastroenteritis in children: a randomized controlled trial. Pediatr Emerg Care. 2026 (epub 24 September). doi:10.1097/PEC.0000000000003715. PMID: 42782131. NCT03052361.

Single-centre Canadian double-blind placebo-controlled RCT in children aged 6 months to 17 years with at least 4 vomits in 24 hours. Recruitment stopped early because of COVID-19: 91 randomised (44 ondansetron, 47 placebo). Immediate discharge after first physician assessment: 44% vs 45% (absolute difference −1%, 95% CI −20% to 19%). No differences in secondary outcomes.

• 91 CHILDREN RANDOMISED (STOPPED EARLY)
• 44% vs 45% DISCHARGED AFTER FIRST ASSESSMENT
• −1% (−20 to 19) ABSOLUTE DIFFERENCE, 95% CI

Appraisal: The confidence interval spans a 20% harm to a 19% benefit, so this trial neither supports nor refutes triage ondansetron. Ondansetron's benefit on vomiting and IV rehydration is already established; the question here was only about the timing.

UK ED practice: No change. Oral ondansetron for children over 6 months with vomiting from gastroenteritis remains reasonable to facilitate oral rehydration; nurse-initiated protocols are a local service decision.

INFORMING PAEDS EM PROSPECTIVE COHORT (PREDICT)

Corticosteroids in Children With Suspected Sepsis: Used in 1 in 10, Mostly Outside Refractory Shock

Long E, Nolan GJG, Williams A, et al.; PREDICT network. Corticosteroid use in children with community acquired sepsis: a multicentre prospective cohort study. Arch Dis Child. 2026 (epub 29 September). doi:10.1136/archdischild-2025-330072. PMID: 42810819.

5,119 children admitted with suspected sepsis through 11 EDs in Australia and New Zealand (2021–2023); in-hospital mortality 1.0%. 4.9% met Phoenix sepsis criteria (mortality 12.7%). Corticosteroids were given within 24 hours to 9.6% overall and 31.3% of Phoenix-positive children; dexamethasone 69.5%, hydrocortisone 25.4%, with wide variation between sites. Most use was outside fluid- or catecholamine-refractory shock.

Appraisal: Descriptive only — it does not assess whether steroids helped or harmed, and dexamethasone use probably reflects co-existing indications such as croup or asthma. It highlights practice variation in the absence of trial evidence.

UK ED practice: Reserve hydrocortisone in paediatric septic shock for catecholamine-refractory shock or known adrenal insufficiency, in discussion with PICU or your retrieval service, consistent with the Surviving Sepsis Campaign paediatric guidance.

INFORMING PAEDS EM PROSPECTIVE COHORT (PECARN) MENTAL HEALTH

After a Positive ED Suicide Risk Screen, 11.6% of Adolescents Attempted Suicide Within 3 Months

Foster AA, Hoffmann JA, Addo N, et al. Patient and clinical characteristics associated with subsequent suicide attempt among youth in the emergency department with a positive suicide risk screen. Ann Emerg Med. 2026 (epub 24 September). doi:10.1016/j.annemergmed.2026.07.011. PMID: 42788929.

Secondary analysis of a PECARN prospective cohort (2015–2019): 2,085 adolescents aged 12–17 with a positive Ask Suicide-Screening Questions (ASQ) screen. Within 3 months, 11.6% attempted suicide and 19.9% had a mental-health ED revisit or admission. Independent risk factors for attempt: attempt in the past month (aOR 2.52), past non-suicidal self-injury (aOR 2.06), prior mental health admission (aOR 1.39) and hopelessness (aOR 1.30).

Appraisal: A large, well-characterised US cohort; findings are associations and the US health-system context differs. The factors are clinically intuitive, but their presence or absence cannot safely stratify an individual child.

UK ED practice: A positive screen is the start of an assessment, not a disposition. Document recent attempts, self-harm history and hopelessness explicitly, involve CAMHS liaison per NICE NG225, and give safety-planning advice and follow-up to every young person and their carers.

4 — FOAMED & CRITICAL APPRAISAL

JOURNALFEED TRAUMA

Tertiary Trauma Survey Still Finds Missed Injuries in Moderate Trauma (ISS 9–15)

Knight C, Burke JR, Hind S, Smith SR. More than meets the eye: the continuing importance of tertiary trauma survey even in moderate (ISS 9–15) trauma. Injury. 2026. PMID 42285813. JournalFeed EM Speed Read, 1 October 2026.

In 293 major trauma patients (ISS ≥9), the tertiary trauma survey identified unrecognised injuries in 18.8%, most often affecting the face, head and eyes. A useful reminder that a structured tertiary survey is worth doing for every admitted trauma patient, not just the most severely injured; check the full paper for design details before citing locally.

BEST EVIDENCE TOPIC EMJ

Whiff of Doubt: Inhaled Isopropyl Alcohol for Acute ED Nausea — Supported, but Not in Pregnancy

Khamesi J, Dabis R (eds). Whiff of doubt: is inhaled isopropyl alcohol (alcohol wipes) an effective treatment for acute nausea in the emergency department? Best Evidence Topic reports. Emerg Med J. 2026 (epub). doi:10.1136/emermed-2026-216455.

Six studies (two SR/MAs, two RCTs, two implementation studies) support IPA wipe inhalation for rapid short-term relief of acute nausea; it may reduce the need for conventional antiemetics, though this was not consistent across studies, and it is low cost. This BET corroborates the RCT covered in Issue 29. No study evaluated nausea and vomiting of pregnancy or hyperemesis gravidarum, and several excluded pregnant patients, so IPA cannot yet be recommended in pregnancy.

5 — ACTION POINTS

6 — TRIALS TO WATCH

ONGOING / EXPECTED

CORE REVISION — ISSUE 32 — 2 OCTOBER 2026

Toxicology: Serotonin Syndrome vs Neuroleptic Malignant Syndrome

The hot, rigid patient — Hunter criteria, clonus, onset, drug causes, treatment — FRCEM

Exam goal: Diagnose serotonin syndrome using the Hunter criteria; distinguish it from neuroleptic malignant syndrome (NMS), anticholinergic toxicity and malignant hyperthermia; list common precipitating drugs; and describe ED management, including when to cool aggressively, sedate, and involve critical care. A NEJM review of serotonin syndrome was published this week (PMID 42804612), and the MHRA levetiracetam alert above lists NMS among its warning symptoms.

1. HUNTER SEROTONIN TOXICITY CRITERIA

In a patient who has taken a serotonergic agent, serotonin toxicity is diagnosed if ANY of the following is present:

2. SEROTONIN SYNDROME VS NMS VS ANTICHOLINERGIC TOXICITY

3. COMMON CAUSES

SSRIs and SNRIs, MAOIs (including linezolid and methylene blue, which are MAO inhibitors), tramadol, pethidine, fentanyl, dextromethorphan, MDMA and other stimulants, lithium, triptans (low risk), ondansetron and metoclopramide (rarely), St John's wort. The highest-risk combinations involve an MAOI. Remember serotonin toxicity after ED drug administration — for example tramadol, fentanyl or methylene blue given to a patient already on an SSRI.

4. ED MANAGEMENT

NMS management: stop the antipsychotic (or restart the withdrawn dopaminergic drug), supportive care, cooling, benzodiazepines, and critical care for severe cases; dantrolene and bromocriptine are used on specialist advice. Check CK, renal function and potassium for rhabdomyolysis in both conditions.

Memory aid: Serotonin syndrome is fast and twitchy (hours, clonus, hyperreflexia, diarrhoea); NMS is slow and stiff (days, lead-pipe rigidity, bradyreflexia). If it is hot and dry with urinary retention, think anticholinergic.

FRCEM exam focus: (1) Hunter criteria require exposure to a serotonergic agent plus a specific neuromuscular finding — clonus is the key sign. (2) Onset within hours vs days distinguishes serotonin syndrome from NMS. (3) Clonus and hyperreflexia vs lead-pipe rigidity and bradyreflexia. (4) Linezolid and methylene blue are MAO inhibitors. (5) Severe hyperthermia is muscular — treat with cooling, sedation and paralysis, not paracetamol. (6) Use TOXBASE and NPIS for UK dosing advice.

EM Evidence Rundown — Issue 32 — 2 October 2026 Curated by Jake Turner, Senior Registrar in Emergency Medicine (ST6). Produced with the assistance of AI (Perplexity). All content editorially reviewed. Not a substitute for clinical judgement or local guidelines. Feedback: Submit feedback | Archive: emevidence.org Unsubscribe: Reply UNSUBSCRIBE to this email. To update your preferences, visit emevidence.org