EM Evidence Rundown — Issue 37

EM EVIDENCE RUNDOWN — ISSUE 37 — 5 NOVEMBER 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 (Claude). All content editorially reviewed.

Full archive and PDF downloads at emevidence.org

Lead: 0/1 h vs 0/3 h troponin pathways: an individual-patient meta-analysis of 5 RCTs (110,933 presentations) found similar 30-day death or type 1 MI (0.8% vs 1.0%; OR 0.83, 95% CI 0.45 to 1.53) and no clear gain in ED length of stay (314 vs 308 minutes). Change Tonight: MHRA, 5 November: Aspire tropicamide 1% single-dose batches E941A, E941B and E941C are recalled because the packs carry the wrong expiry date (30/04/2027) and their true expiry, 31 October 2026, has passed, and the Sandoz co-amoxiclav 250/62.5 mg/5 mL SmPC dosing table gave about twice the intended dose. Paediatric EM (4 items): video vs direct laryngoscopy, trainee first-attempt intubation success, continuous oximetry in young infants with bronchiolitis, and lactate after drowning; pyuria and the national PEWS roll-out are in Quick Hits.

BOTTOM LINE UP FRONT — 5 NOVEMBER 2026

ACT ON THIS NOW

MHRA Tropicamide and co-amoxiclav: check ED eye-room stock for Aspire tropicamide 1% batches E941A to E941C (true expiry 31 October 2026) and quarantine them; dose children from the notice’s Annex 1 or the local paediatric formulary, not the Sandoz co-amoxiclav 250/62.5 mg/5 mL SmPC table.

WORTH KNOWING ON SHIFT

RCT Ventilation during CPR (crossover RCT, 99 intubated adults in ongoing CPR after the decision to stop had been made): the ventilator gave a 146.2 mL lower measured tidal volume than bagging (95% CI −187.0 to −105.5), and 31.3% of bagged medians exceeded 10 mL/kg.

UK Thoracotomy (TETRiS, 116 UK cases): ROSC 43.1%, survival to discharge 6.9%, and no survivor beyond 10 minutes of arrest.

RCT Renal colic (n=251, unregistered): 30-minute success 76.7% with IV ibuprofen, 74.4% with dexketoprofen and 66.3% with IV paracetamol (p=0.280); NSAIDs stay first line (NICE NG118).

COHORT FAST: 24.8% of FAST-negative trauma patients still had a severe internal injury (including pelvic fractures and haemothorax, which FAST is not designed to find); a negative scan does not exclude it.

KNOW FOR NEXT TIME

UK Ramping (UK EDs): at sites that held patients in ambulances, each 5-point rise in the share of escalation care held there was associated with higher 28-day mortality (adjusted HR 1.021, 95% CI 1.002 to 1.041); any use vs none was not (HR 0.957); observational.

DIAGNOSTIC PE rules after the CTPA decision: YEARS kept 96.5% sensitivity but would have spared imaging in only 13.8% of the 398 patients with a D-dimer.

SR/MA Bicarbonate (4 ICU RCTs, n=1,544): no clear mortality reduction (RR 0.94, 95% CrI 0.73 to 1.22) but probably less renal replacement therapy (RR 0.74, 95% CrI 0.59 to 0.98).

DIAGNOSTIC Head injury blood test: GFAP/UCH-L1 sensitivity 91% (95% CI 59 to 100%) but specificity 22%; not a substitute for the NICE NG232 CT rules.

COHORT Cannabinoid hyperemesis: 72-hour return 8.2% after haloperidol vs 15.1% after ondansetron (OR 0.50, 95% CI 0.41 to 0.62); off-label in the UK, ECG first.

PAEDS Video laryngoscopy: crude first-pass success 72.2% vs 69.0% with direct laryngoscopy (pooled RR 1.11, 95% CI 1.06 to 1.16), very low certainty.

EM Evidence Rundown, Issue 37. Page 1 of 25

This week leads on the chest pain pathway. A large individual-patient meta-analysis finds no clear difference in 30-day death or MI between the 0/1 hour and 0/3 hour troponin pathways, though the estimate is imprecise, and no clear time saving from the faster one. Resuscitation has three papers: a crossover trial of ventilator vs bag ventilation in real patients during CPR, the UK TETRiS evaluation of resuscitative thoracotomy, and signs of life in traumatic arrest. A small trial tests intravenous analgesia for renal colic, and a Bayesian meta-analysis revisits bicarbonate for metabolic acidosis. In UK news, a TERN analysis links ambulance ramping with mortality, and MHRA has issued a tropicamide recall and flagged a dosing error in a co-amoxiclav SmPC. Paediatrics has four items. Core Revision covers paediatric resuscitation. There is also a correction to our Issue 28 summary of TARGET-CTCA, in Quick Hits.

WHAT’S INSIDE — ISSUE 37

• Lead: 0/1 h vs 0/3 h troponin pathways (IPD meta-analysis, Lancet); PE decision rules after the CTPA request
• Resuscitation: ventilator vs bag during CPR; TETRiS resuscitative thoracotomy; signs of life in traumatic arrest
• Toxicology and pain: haloperidol for cannabinoid hyperemesis; IV ibuprofen, dexketoprofen or paracetamol for renal colic
• Critical care and trauma: bicarbonate for metabolic acidosis; what a positive or negative FAST means; GFAP/UCH-L1 in mild head injury
• UK: MHRA tropicamide recall and co-amoxiclav dosing error; ambulance ramping vs corridor care (UNCORKED)
• PEM (4): video vs direct laryngoscopy, trainee intubation, bronchiolitis oximetry, lactate after drowning
• FOAMed: EMCrit on massive lithium overdose, set against the Priadel SmPC and EXTRIP
• Quick Hits (13): including the TARGET-CTCA correction, PE severity classes, POCUS for PE, xylazine strips, low-level pyuria and national PEWS

• Core Revision: Clinical — paediatric resuscitation essentials

Contents: 1. Key Trials & Articles — 2. Guidelines & UK Updates — 3. Paediatric Emergency Medicine — 4. FOAMed & Critical Appraisal — 5. Quick Hits — Core Revision — 6. Action Points — Trials to Watch

EM Evidence Rundown, Issue 37. Page 2 of 25

1 — KEY TRIALS & ARTICLES

LEAD

INFORMING PRACTICE

SR/MA

0/1 h vs 0/3 h Troponin Pathways (IPD Meta-Analysis, 5 RCTs, 110,933 Presentations): 30-Day Death or Type 1 MI 0.8% vs 1.0%, With No Clear Gain in ED Length of Stay

Boeddinghaus J, Kunz M, Lopez-Ayala P, et al. Performance of 0/1 h versus 0/3 h pathways in patients with suspected myocardial infarction: a systematic review and individual participant data meta-analysis of randomised controlled trials. Lancet. 2026 (online 29 October). doi:10.1016/S0140-6736(26)01767-8. PMID: 42956796. PROSPERO CRD420261383452. Source route: PubMed (found while resolving EMA Daily 3 Nov); abstract only, no free full text, and no JournalFeed, SGEM or other commentary found by 5 November. Background: the 0/1 h vs 0/3 h RCT reported by JournalFeed on 23 July was covered in Issue 22 (24 July 2026). Abstract only

Two-stage random-effects IPD meta-analysis of all five RCTs comparing a 0/1 h with a 0/3 h hs-troponin pathway (55,489 vs 55,444 presentations; every trial supplied IPD). The primary safety outcome, all-cause death or new type 1 MI within 30 days, occurred in 446 (0.8%) vs 562 (1.0%) presentations (pooled OR 0.83, 95% CI 0.45 to 1.53; I² 62.9%). The primary efficacy outcomes did not clearly differ: median ED length of stay 314 vs 308 minutes (geometric mean ratio 1.03, 95% CI 0.85 to 1.25; I² 99.2%) and direct ED discharge 67.0% vs 65.8% (OR 1.31, 95% CI 0.79 to 2.16; I² 97.8%). The authors conclude that both pathways had low short-term event rates but that the effect on flow may not transfer uniformly across health systems.

• 0.8% vs 1.0% 30-DAY DEATH OR TYPE 1 MI, 0/1 H VS 0/3 H
• OR 0.83 95% CI 0.45 TO 1.53; NO NNT, CI CROSSES NO EFFECT
• 314 vs 308 min ED LENGTH OF STAY; GMR 1.03 (0.85 TO 1.25)
• 67.0% vs 65.8% DIRECT ED DISCHARGE; OR 1.31 (0.79 TO 2.16)

CRITICAL APPRAISAL Clear question; protocol registered

Partly PROSPERO CRD420261383452; safety outcome narrowed from any MI to type 1 MI after IPD receipt but before analysis (direction of bias unclear)

Design: SR/MA | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: either pathway is acceptable; do not expect a switch from 0/3 h to 0/1 h to shorten ED stays on its own | Confidence in that: moderate

Main concerns: heterogeneity over 97% for flow outcomes means the pooled averages may not describe any real ED, in either direction; post-registration narrowing (after IPD receipt, before analysis) of the safety outcome to type 1 MI could understate harm from either pathway if type 2 MI or injury were missed Strengths: individual participant data from every eligible trial, with harmonised outcome definitions

UK practice: NICE HTG552 (formerly DG40, 2020) rec 1.2 recommends hs-troponin early rule-out with a single sample at or near the limit of detection, or a second sample at 30 minutes to 3 hours, alongside clinical judgement; it does not choose between 0/1 h and 0/3 h. We found no RCEM best practice guideline on troponin pathways (rcem.ac.uk search, 5 November 2026).

EM Evidence Rundown, Issue 37. Page 3 of 25

INFORMING PRACTICE DIAGNOSTIC

PE Rules After the CTPA Has Been Ordered (Indian ED Cohort, n=408): YEARS Sensitivity 96.5%, but Only 13.8% Could Have Avoided Imaging

Pannu AK, Saroch A, Bhatia V, et al. Performance of pulmonary embolism diagnostic strategies in patients undergoing computed tomography pulmonary angiography. Intern Emerg Med. 2026 (online 28 October). doi:10.1007/s11739-026-04573-x. PMID: 42955693. Source route: PubMed sweep; abstract only, no free full text, and no secondary commentary found by 5 November. Abstract only

Prospective study of consecutive ED patients aged 13 or over in whom a clinician had already decided to request CTPA for suspected PE (authors from PGIMER, Chandigarh, India; number of centres not stated in the abstract). After that decision, investigators applied the modified Wells and simplified revised Geneva scores and the YEARS and PEGeD strategies, with segmental or more proximal PE on CTPA as the primary outcome. Significant PE was found in 144 of 408 (35.3%); Wells discriminated better than Geneva (AUC 0.725, 95% CI 0.671 to 0.779, vs 0.656, 95% CI 0.598 to 0.714), but 21.6% of Wells “PE unlikely” patients had significant PE. In 398 patients with a D-dimer, YEARS and PEGeD each had 96.5% sensitivity but would have spared imaging in only 13.8% and 13.6%.

• 35.3% SEGMENTAL OR LARGER PE (144/408)
• 96.5% YEARS SENSITIVITY (CI NOT REPORTED)
• 13.8% ELIGIBLE TO AVOID IMAGING UNDER YEARS
• 21.6% SIGNIFICANT PE IN WELLS PE-UNLIKELY (SCORE ALONE)

CRITICAL APPRAISAL Representative spectrum

Partly Consecutive, but only patients already sent for CTPA (prevalence 35.3%); likely overstates the miss rate of a low-probability category compared with an unselected UK ED

Design: Diagnostic | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: keep the NG158 two-level Wells plus D-dimer pathway; YEARS adds little once a clinician has decided on CTPA | Confidence in that: moderate

Main concerns: CTPA-selected, high-prevalence cohort: the missed-PE figures are likely worse than in UK practice; abstract does not give specificity or the counts behind 96.5%, so the safety of YEARS here cannot be checked Strengths: every patient had the reference standard (CTPA)

UK practice: NICE NG158 (2020, updated 2023) recs 1.1.17 and 1.1.21: use the two-level PE Wells score; if PE is unlikely (4 points or less), offer a D-dimer with the result within 4 hours if possible. Rec 1.1.14: consider an age-adjusted D-dimer threshold over 50. The recommendations page does not mention YEARS or PEGeD (checked 5 November 2026).

EM Evidence Rundown, Issue 37. Page 4 of 25

INFORMING PRACTICE RCT

Ventilator vs Bag in Intubated Adults During CPR (Crossover RCT, n=99): The Ventilator Gave a 146 mL Lower Measured Tidal Volume and Missed Non-Inferiority, While 31.3% of Bagged Medians Exceeded 10 mL/kg

Vinay S, Manu Ayyan S, Amrithanand VT, et al. Mechanical ventilation vs bag-valve ventilation during cardiopulmonary resuscitation: a randomized crossover noninferiority trial. Resusc Plus. 2026;32:101530. doi:10.1016/j.resplu.2026.101530. PMID: 42985663. CTRI/2024/08/072738. Source route: PubMed sweep; PMC full text (PMC13733466). Appraised from full text

Single-centre randomised two-period crossover trial in an Indian tertiary ED in real patients, not manikins: 104 intubated adults in ongoing CPR were enrolled only after the treating physician had decided termination criteria were met (99 completed both periods; 46.5% out-of-hospital arrests). Each received 1 minute of ventilator breaths (volume control, 8 mL/kg predicted body weight, rate 10, PEEP 0, flow trigger off) and 1 minute of bag-valve ventilation through the tube (1,600 mL bag, visible chest rise, no feedback) in random order with compressions continuing. For the primary outcome, the difference in each patient’s median device-calculated inspired tidal volume (margin −150 mL), the adjusted mean difference was −146.2 mL (95% CI −187.0 to −105.5), so non-inferiority was not shown, and no clinical outcomes were measured.

Source: Vinay et al, Table 2 and Results (n=99 paired).

• −146.2 mL VENTILATOR MINUS BAG (95% CI −187.0 TO −105.5)
• −150 mL NON-INFERIORITY MARGIN: NOT MET
• 31.3% vs 0% MEDIANS ABOVE 10 ML/KG, BAG VS VENTILATOR (CI NOT REPORTED)
• 26.3% vs 45.5% MEDIANS BELOW 6 ML/KG, BAG VS VENTILATOR (CI NOT REPORTED)

CRITICAL APPRAISAL Focused question; registered primary outcome

Partly Prospectively registered; the two registered secondary outcomes (end-tidal CO2, expired volume) were not collected

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: if you put an arrested patient on a ventilator, use the RCUK settings and confirm chest rise and capnography; if you bag, give small timed breaths | Confidence in that: moderate

Main concerns: unblinded bagging staff may have bagged more carefully than usual (our judgement; not assessed), so the excess with bagging may be larger in practice; chest compressions can distort flow sensing, so true delivered volumes are uncertain in either arm (direction unknown); a late peri-termination window in a single centre limits generalisation to early resuscitation Strengths: within-patient crossover in real arrests with compressions running, prospectively registered, no funding or conflicts declared

UK practice: The RCUK Adult advanced life support guidelines (2025) say that after intubation you ventilate at 10 per minute without pausing compressions. If you use a ventilator, set volume control at 6–8 mL/kg predicted body weight (or visible chest rise), maximum oxygen, rate 10, inspiratory time 1–2 s, PEEP 0–5 cm H2O, peak pressure alarm 60–70 cm H2O and flow trigger off, and switch to manual ventilation if it is not effective. The trial protocol was close to this, so its low ventilator volumes are a warning to watch the delivered breath, not only the set one. Issue 35 covered a meta-analysis (mostly observational) that linked mechanical ventilation in CPR with higher ROSC; this trial measured volumes only.

EM Evidence Rundown, Issue 37. Page 5 of 25

INFORMING PRACTICE

UK

OBSERVATIONAL

FRCEM

TETRiS (UK National Evaluation, n=116 Resuscitative Thoracotomies): ROSC in 43.1%, Survival to Discharge in 6.9%, and No Survivor Beyond 10 Minutes of Arrest

Ramage L, Tucker H, Marsden ME, et al; TETRiS Collaborative. Traumatic Emergency Thoracotomy for Resuscitation in Shock (TETRiS): a national service evaluation of resuscitative thoracotomy practice outside the operating theatre. Resuscitation. 2026 (online 28 October 2026): 111371. doi:10.1016/j.resuscitation.2026.111371. PMID: 42950975. Source route: JournalFeed EM SpeedRead 5 Nov; PubMed abstract (closed access, not available to us in full); no published commentary found. Abstract only

Prospective 12-month service evaluation of resuscitative thoracotomy across 20 UK HEMS services and 26 major trauma centre EDs (PHOTON, TERN and NaTRIC networks), with no comparator group. Of 116 thoracotomies, 82 (70.7%) were done on scene; the median age was 30 years and 81 (69.8%) had penetrating injury. Suspected tamponade was the commonest indication (72, 62.1%), but the operative finding was exsanguinating haemorrhage in 82 (70.7%). ROSC occurred in 50 (43.1%) and 8 (6.9%) survived to discharge; survival was better with tamponade than exsanguination and with ED than prehospital thoracotomy (shorter arrests, better physiology), and nobody survived beyond 10 minutes of arrest.

• 43.1% ROSC (50/116; CI NOT REPORTED)
• 6.9% SURVIVAL TO DISCHARGE (8/116; CI NOT REPORTED)
• 70.7% THORACOTOMY ON SCENE (82/116)
• 0 SURVIVORS AFTER MORE THAN 10 MIN OF ARREST

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Prospective national capture; definitions of arrest duration and physiology not reported in the abstract

Design: Observational | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: keep the RCUK 15-minute elapsed-time limit, and treat a penetrating chest arrest with suspected tamponade arriving within minutes as the case most likely to benefit | Confidence in that: moderate

Main concerns: the 10-minute observation rests on 8 survivors, so using it as a new cut-off would overstate futility; the ED vs prehospital difference reflects shorter arrests and better physiology, which overstates any benefit of location; abstract only: case definitions, missing data and funding not reported Strengths: prospective, national and includes thoracotomies done on scene, which retrospective hospital series miss

UK practice: The RCUK adult traumatic cardiac arrest/peri-arrest algorithm (2025) puts reversible causes ahead of compressions, including relieving tamponade in penetrating chest injury, and gives resuscitative thoracotomy when expertise, equipment and environment allow and elapsed time is under 15 minutes. NICE NG39 (2016) makes no recommendation on resuscitative thoracotomy (we searched the recommendations). The UK how-to guide for the procedure (Chesters et al, EMJ 2026) was covered in EM Issue 32.

EM Evidence Rundown, Issue 37. Page 6 of 25

INFORMING PRACTICE OBSERVATIONAL

Signs of Life in Traumatic Cardiac Arrest (Kanto Registry, n=462): 10.2% With Any Electrical Activity vs 2.5% Without Survived Beyond the ED, and 1 Patient Had a Good Neurological Outcome

Yamashita I, Yamamoto R, Suzuki M, et al. Signs of life and clinical outcomes in patients with traumatic out-of-hospital cardiac arrest: a nationwide multicenter observational study. J Emerg Med. 2026;90:161–169. doi:10.1016/j.jemermed.2026.09.006. PMID: 42874196. Source route: JournalFeed EM SpeedRead 4 Nov; open-access PDF (CC BY 4.0) via BMA Library. Appraised from full text

Retrospective descriptive analysis of the prospective SOS-KANTO 2017 cardiac arrest registry (42 hospitals in Tokyo and its suburbs, September 2019 to March 2021): 462 adults already in traumatic arrest when the ambulance arrived, grouped by three signs of life at scene (pupillary light reflex, agonal gasping, and any rhythm other than asystole on the first monitor). Survival beyond the ED (primary outcome) was 22 of 462 (4.8%): 4 of 16 (25.0%) with electrical activity plus gasping, 10 of 117 (8.5%) with electrical activity alone and 8 of 318 (2.5%) with no signs of life, and none of the 7 with a pupil reflex or gasping alone. Only 1 patient (0.2%) had a Cerebral Performance Category of 1–2 at 30 days, and there was no adjusted analysis.

• 10.2% vs 2.5% SURVIVAL BEYOND ED, ANY ELECTRICAL ACTIVITY VS NONE (14/137 VS 8/325, CALCULATED; CI NOT REPORTED)
• 25.0% ELECTRICAL ACTIVITY PLUS GASPING (4/16; 95% CI 10.2 TO 49.5)
• 2.5% NO SIGNS OF LIFE (8/318; 95% CI 1.3 TO 4.9)
• 1 of 462 GOOD NEUROLOGICAL OUTCOME AT 30 DAYS (0.2%)

CRITICAL APPRAISAL Exposure measured reliably

Partly Signs of life taken from ambulance records; limb movement not recorded; “electrical activity” is any non-asystolic rhythm

Design: Observational | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: asystole with no signs of life in traumatic arrest (mechanism not recorded, but likely mostly blunt) carried very poor odds here, but the registry cannot support a termination rule | Confidence in that: moderate

Main concerns: no trauma detail (mechanism, thoracotomy, blood), so it cannot show which patients were salvageable (direction unknown); survival beyond the ED overstates meaningful recovery: 1 good neurological outcome in 462; a population drawn from a country where over 95% of trauma is blunt (mechanism not recorded) likely understates salvageability in penetrating tamponade seen in UK major trauma centres Strengths: prospectively collected Utstein-style registry data across 42 hospitals (retrospective analysis), with sensitivity analyses for missing data that gave similar results

UK practice: The RCUK adult traumatic cardiac arrest/peri-arrest algorithm (2025) asks you to treat hypovolaemia, oxygenation, tamponade and tension pneumothorax before compressions, and to consider terminating resuscitation if there is no ROSC; it does not list signs-of-life criteria. Read this with TETRiS above: TETRiS ties survival to arrest duration and residual physiology; here residual cardiac electrical activity was the sign most often present in survivors, and no single sign was enough to select or exclude patients.

EM Evidence Rundown, Issue 37. Page 7 of 25

INFORMING PRACTICE RCT

Renal Colic, IV Ibuprofen vs Dexketoprofen vs Paracetamol (Unregistered RCT, n=251): 30-Minute Success 76.7%, 74.4% and 66.3%, Not Significantly Different

Yılmaz MM, Türkseven ÇH, Alkaya M, et al. Opioid-free analgesia for acute renal colic: a randomized, double-blind, head-to-head comparison of intravenous ibuprofen, dexketoprofen trometamol and paracetamol in the emergency department. Intern Emerg Med. 2026 (published online 30 October 2026). doi:10.1007/s11739-026-04579-5. PMID: 42961621. Trial not registered (stated in the abstract). Source route: PubMed sweep (abstract and publisher page); no blog or journal commentary found on 5 November 2026. Abstract only

Single-centre, double-blind, three-arm RCT in a university ED in Türkiye. 251 adults with acute renal colic received IV ibuprofen 400 mg (n=86), dexketoprofen trometamol 50 mg (n=82) or paracetamol 1 g (n=83) over 15 minutes; with three active arms, each drug is the comparator for the others. The primary outcome, at least a 50% fall in 100-mm VAS at 30 minutes, occurred in 76.7%, 74.4% and 66.3% (p=0.280; no CI reported). Rescue fentanyl was given in 32.6%, 31.7% and 43.4% (p=0.217). After adjustment for baseline VAS (higher in the dexketoprofen arm), 30-minute VAS was lower with dexketoprofen than with paracetamol (−9.5 mm, 95% CI −16.5 to −2.5), and the adjusted odds of rescue were lower (OR 0.40, 95% CI 0.20 to 0.84).

• +10.5 pp SUCCESS, IBUPROFEN VS PARACETAMOL (CALCULATED; 95% CI −3.1 TO 24.0)
• −11.7 pp RESCUE FENTANYL, DEXKETOPROFEN VS PARACETAMOL (CALCULATED; 95% CI −26.3 TO 3.0)
• OR 0.40 ADJUSTED RESCUE, DEXKETOPROFEN VS PARACETAMOL (95% CI 0.20 TO 0.84)
• 251 RANDOMISED, 1 CENTRE

Source: Yılmaz et al, abstract. Counts back-calculated from the reported percentages and arm sizes (calculated). No NNT is given because the primary result is null.

CRITICAL APPRAISAL Focused question; registered primary outcome

No Not registered, so the primary outcome and the adjusted analyses cannot be checked against a plan; likely overstates the secondary dexketoprofen findings

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; keep NSAIDs first line as NICE advises and use IV paracetamol when an NSAID cannot be given | Confidence in that: moderate

Main concerns: unregistered trial with a null primary and a positive adjusted secondary: likely overstates the dexketoprofen advantage; baseline pain imbalance in the dexketoprofen arm makes the adjusted comparison depend on the model; the trial may be underpowered, so a clinically important difference cannot be excluded (risk of a false-negative result) Strengths: double-blind head-to-head comparison of the IV non-opioid options at standard doses

UK practice: NICE NG118 (2019, renal and ureteric stones) recommends an NSAID by any route first line for suspected renal colic (1.2.1), IV paracetamol if NSAIDs are contraindicated or not working (1.2.2), and opioids only if both fail or are contraindicated (1.2.3). The RCEM Management of Acute Pain in Adults in the ED (revised September 2024) names NSAIDs for ureteric colic and rectal diclofenac 100 mg as particularly useful. The ibuprofen 400 mg solution for infusion SmPC (eMC, revised 2025) is licensed for acute moderate pain only when other routes are not possible; eMC lists dexketoprofen only as oral tablets.

EM Evidence Rundown, Issue 37. Page 8 of 25

INFORMING PRACTICE OBSERVATIONAL

Cannabinoid Hyperemesis (TriNetX Cohort): 72-Hour ED Return 8.2% After Haloperidol vs 15.1% After Ondansetron in Propensity-Matched Adults

O’Brien J, Patel M, Brown J, et al. Comparative Effectiveness of Haloperidol, Droperidol, and Ondansetron for Cannabinoid Hyperemesis Syndrome in the Emergency Department. Acad Emerg Med. 2026;33(11):e70436. doi:10.1111/acem.70436. PMID: 42983225. Source route: PubMed sweep (abstract); no blog or journal commentary found on 5 November 2026. Background RCT: Ruberto AJ, et al. HaVOC. Ann Emerg Med. 2021;77(6):613–619. PMID: 33160719. Abstract only

Multicentre retrospective cohort in the TriNetX research network (2014 to 2024) of adults with vomiting-related and cannabis-related ICD-10 diagnoses who received haloperidol, droperidol or ondansetron in the ED; three pairwise propensity-matched comparisons were made, and the matched group sizes are not given in the abstract. The primary outcome was ED return within 72 hours: 8.2% with haloperidol vs 15.1% with ondansetron (OR 0.50, 95% CI 0.41 to 0.62) and 7.4% with droperidol vs 14.0% with ondansetron (OR 0.49, 95% CI 0.32 to 0.76). Droperidol and haloperidol did not differ (7.5% vs 9.8%; OR 0.75, 95% CI 0.47 to 1.20). Coded adverse events were described only as rare.

• 8.2% 72-H RETURN, HALOPERIDOL (MATCHED)
• 15.1% 72-H RETURN, ONDANSETRON (MATCHED)
• 7.4% 72-H RETURN, DROPERIDOL VS 14.0% ONDANSETRON
• 7.5% vs 9.8% DROPERIDOL VS HALOPERIDOL (OR 0.75, 0.47 TO 1.20)

Source: O’Brien et al, abstract. Multivariable Cox model for 72-h return: haloperidol HR 0.62 (0.49 to 0.79), droperidol HR 0.68 (0.56 to 0.83); ondansetron not associated. Cohort sizes, doses and routes not reported in the abstract.

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly CHS defined by ICD-10 codes, not by clinical criteria; returns to EDs outside the network are missed

Design: Observational | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; it supports choosing haloperidol over ondansetron for suspected CHS, off-label and with an ECG | Confidence in that: moderate

Main concerns: confounding by indication (who is given haloperidol) likely overstates the benefit; code-based CHS definition dilutes the cohort with other vomiting, direction unknown; coded adverse events likely understate dystonia and QT harms Strengths: direction matches HaVOC, an earlier RCT of haloperidol vs ondansetron (30 patients analysed; haloperidol better than ondansetron 8 mg by 2.3 cm on a 10-cm scale, 95% CI 0.6 to 4.0, with 2 returns for acute dystonia, both on the higher dose)

UK practice: Neither drug is licensed for CHS in the UK. The haloperidol 5 mg/mL injection SmPC (eMC, revised 2025) licenses it for postoperative nausea and vomiting at 1 to 2 mg IM, recommends a baseline ECG before IM dosing and requires continuous ECG monitoring if given IV. The droperidol 2.5 mg/mL SmPC (eMC, revised 2024) lists only postoperative indications (postoperative nausea and vomiting, and opioid-induced nausea and vomiting during postoperative PCA) and is contraindicated with QTc over 440 ms (men) or 450 ms (women), hypokalaemia, hypomagnesaemia or a heart rate under 55. The droperidol SmPC also advises checking electrolytes and creatinine in patients at risk of arrhythmia.

EM Evidence Rundown, Issue 37. Page 9 of 25

INFORMING PRACTICE

SR/MA

FRCEM

Sodium Bicarbonate for Metabolic Acidosis in Critical Illness (Bayesian SR/MA, 4 RCTs, n=1,544): No Clear Mortality Benefit (RR 0.94, 95% CrI 0.73 to 1.22) but a 97% Probability of Less Renal Replacement Therapy (RR 0.74, 0.59 to 0.98)

Kitisin N, Walker HGM, Jung B, et al. Sodium bicarbonate therapy for metabolic acidosis in critically ill patients: a Bayesian systematic review and meta-analysis of randomised controlled trials. Intensive Care Med. 2026 (published 2 November 2026). doi:10.1007/s00134-026-08602-3. PMID: 42970124. PROSPERO CRD420251246519. Source route: open-access full text on the publisher’s site; Tables 1–2 and the supplement were not in the text available. Appraised from full text

Four RCTs (BICAR-ICU, BICARICU-2, the SODa-BIC pilot and SODa-BIC; 768 patients assigned to IV sodium bicarbonate vs 776 assigned to placebo or usual care) in adult ICU patients with metabolic acidosis. For the primary outcome, all-cause mortality at day 28 or 30, the pooled RR was 0.94 (95% credible interval 0.73 to 1.22), a 72.1% posterior probability of any benefit. RRT initiation was less frequent with bicarbonate (RR 0.74, 95% CrI 0.59 to 0.98; probability of benefit 97.2%), and every adverse-event CrI included no effect. Absolute event rates (Table S6) were not in the text available, so no ARD or NNT is given; SODa-BIC itself was appraised in EM Issue 18.

• Not available ABSOLUTE RISKS (TABLE S6 NOT IN TEXT AVAILABLE)
• RR 0.94 DAY 28 OR 30 MORTALITY (95% CRI 0.73 TO 1.22)
• RR 0.74 RRT INITIATION (95% CRI 0.59 TO 0.98)
• 72.1% / 97.2% PROBABILITY OF BENEFIT: MORTALITY / RRT

Source: Kitisin et al, Study characteristics (text). CRITICAL APPRAISAL Clear question; protocol registered

Partly PROSPERO registered; protocol primary (longest follow-up) changed to day 28 or 30 mortality as 90-day data were incomplete

Design: SR/MA | Certainty: moderate (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; no case for bicarbonate to improve survival; for severe acidaemia with AKI, decide with ICU | Confidence in that: moderate

Main concerns: RRT is clinician-decided, partly open label, and correcting pH removes one of its indications, so the kidney benefit is likely overstated; pooling milder with severe acidaemia likely biases mortality towards the null for the sickest Strengths: RCTs only, two double blind; stable across priors and a frequentist re-analysis.

UK practice: The UK MOSAICC (ISRCTN14027629) trial is randomising critical care adults with pH under 7.30 (PaCO2 under 6.5 kPa) and KDIGO stage 2–3 AKI to IV 8.4% sodium bicarbonate targeting pH 7.30 or more (maximum 500 ml in 24 hours) or none; primary outcome 90-day mortality, target 1,672, recruiting to 30 September 2027 (registry). No UK guideline we checked makes a recommendation on bicarbonate for metabolic acidosis: NICE NG148 (acute kidney injury, 2019) mentions it only as volume expansion before iodine-based contrast (1.2.8).

EM Evidence Rundown, Issue 37. Page 10 of 25

INFORMING PRACTICE OBSERVATIONAL

FAST in Adult Trauma (4 US Level I Centres, n=2,151): ED Transfusion in 53.4% if Positive vs 23.3% if Negative, and Severe Internal Injury in a Quarter of FAST-Negative Patients

Driver L, Castro A, LaColla J, et al. Association Between Focused Assessment with Sonography in Trauma (FAST) Results, Transfusion, and Injury in Trauma: A Multicenter Retrospective Cohort Study. J Am Coll Emerg Physicians Open. 2026;7(5):100487. doi:10.1016/j.acepjo. 2026.100487. PMID: 42614755. Source route: JournalFeed Spoon Feed 2 November; PMC full text (PMC13482603, CC BY). Appraised from full text

Retrospective cohort of adult trauma activations with a documented FAST at 4 US level I centres (2021 to 2023), using the clinician’s recorded interpretation without image review. The primary outcome, red cell or whole blood transfusion before leaving the ED, occurred in 179/335 (53.4%) FAST-positive vs 418/1,792 (23.3%) FAST-negative patients (adjusted OR 3.38, 95% CI 2.53 to 4.50, adjusted for age, sex, weight, physiology, lactate, bicarbonate, mechanism and prehospital hypotension, with site as a random effect). Severe internal injury (solid organ injury, haemoperitoneum, pelvic fracture, retroperitoneal haematoma, major vascular injury or haemothorax) was more common after a positive scan (adjusted OR 4.82, 95% CI 3.66 to 6.35). The 24 indeterminate scans behaved like positives (transfusion adjusted OR 4.11, 95% CI 1.58 to 10.72).

• 53.4% ED TRANSFUSION, FAST POSITIVE
• 23.3% ED TRANSFUSION, FAST NEGATIVE
• 24.8% SEVERE INTERNAL INJURY DESPITE NEGATIVE FAST
• aOR 3.38 TRANSFUSION, POSITIVE VS NEGATIVE (2.53 TO 4.50)

Source: Driver et al, Table 2. CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Clinician-documented FAST, no image review; composite includes injuries FAST is not designed to see (pelvic fracture)

Design: Observational | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing new: a negative FAST is not a reason to stop looking for haemorrhage | Confidence in that: high Disagreement: The authors support FAST as an early risk stratification tool alongside physiology; NICE NG39 limits FAST in suspected haemorrhage to the minimum needed to direct intervention in unstable patients and says it should not decide who needs CT.

UK practice: NICE NG39, Major trauma: assessment and initial management (2016) limits FAST to the minimum needed to direct intervention in unstable patients not responding to volume resuscitation (1.5.29) and says a negative FAST does not exclude intraperitoneal or retroperitoneal haemorrhage (1.5.30). Do not use FAST before immediate CT in major trauma, or to decide who needs CT (1.5.32 and 1.5.33).

EM Evidence Rundown, Issue 37. Page 11 of 25

INFORMING PRACTICE DIAGNOSTIC

GFAP/UCH-L1 Blood Test Before Head CT in Mild TBI (Single-Centre US, n=104): Sensitivity 91% (95% CI 59 to 100%), Specificity 22%; 4 in 5 Patients Would Still Be Scanned

Pandey S, Chen Y, Yue JK, et al. Associations Between Diagnostic Blood-Based Biomarkers and Neuroimaging Findings in Mild Traumatic Brain Injury. Ann Emerg Med. 2026 (published 27 October 2026). doi:10.1016/j.annemergmed.2026.08.731. PMID: 42949696. Source route: PubMed sweep; PubMed abstract only (not in BMA/EMJ/Cochrane or a free copy); no secondary commentary found. Abstract only

Single-centre validation of a US FDA-approved plasma GFAP/UCH-L1 test in 104 adults with suspected mild TBI who were having a head CT at a US level I trauma centre (October 2023 to June 2024), with blood taken within 12 hours of injury. The reference standard was intracranial injury on CT (11 patients, 10.6%, calculated). The test was positive (GFAP over 30 pg/mL or UCH-L1 over 360 pg/mL) in 83 patients: sensitivity 0.91 (95% CI 0.59 to 1.00), specificity 0.22 (95% CI 0.14 to 0.31), NPV 0.95 (95% CI 0.76 to 1.00). Back-calculating from the abstract, 1 of 11 injuries was missed and 21 of 104 patients (20.2%) tested negative; the median result was back in 1.09 hours vs 3.28 hours for CT.

• 91% SENSITIVITY (95% CI 59 TO 100%)
• 22% SPECIFICITY (95% CI 14 TO 31%)
• 1.16 LR+ (CALCULATED; CI NOT REPORTED)
• 0.42 LR− (CALCULATED; CI NOT REPORTED)

All values calculated from the back-calculated 2×2 table (LR+ 1.16, LR− 0.42); the 10.6% row is the observed cohort. CRITICAL APPRAISAL Representative spectrum; consecutive enrolment

Unclear Only patients already having CT; enrolment method not reported

Design: Diagnostic | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; keep scanning by the NICE NG232 criteria | Confidence in that: moderate

UK practice: NICE NG232, Head injury: assessment and early management (2023) sets the adult CT head criteria (recommendations 1.5.8 and 1.5.9) and makes no recommendation for blood biomarkers. The committee could not recommend biomarkers for predicting acute complications after mild TBI because specificity was not high enough, and made research recommendation 2 on their diagnostic accuracy. It also noted that manufacturers recommend testing within 3 hours of injury; this study allowed up to 12.

EM Evidence Rundown, Issue 37. Page 12 of 25

2 — GUIDELINES & UK UPDATES

CHANGE TONIGHT

SAFETY

SAFETY ALERT

UK

PAEDS

MHRA, 5 November: Sandoz Co-Amoxiclav 250/62.5 mg/5 mL SmPC Dosing Table Gave About Twice the Intended Dose; Aspire Tropicamide 1% Single-Dose Eye Drop Batches E941A–C Recalled (True Expiry 31 October 2026)

Medicines and Healthcare products Regulatory Agency. Class 4 Medicines Defect Notification: Sandoz Limited, Co-Amoxiclav powder for oral suspension 250/62.5 mg/5mL, EL(26)A/46. 5 November 2026. gov.uk (notice with Annex 1: PDF). Class 3 Medicines Recall: Aspire Pharma Limited, Tropicamide 1% Eye Drops, solution in single dose containers 20 x 0.5ml, EL(26)A/47. 5 November 2026. gov.uk. MHRA Safety Roundup October 2026, published 27 October, updated 30 October 2026: gov.uk. Source route: MHRA alerts pages. Appraised from full text

Sandoz has told the MHRA that the weight-based dosing table in section 4.2 of its co-amoxiclav 250/62.5 mg/5 mL suspension SmPC instructed a volume giving approximately twice the intended dose. All batches are affected, the powder itself is not defective, the SmPC has been removed from the MHRA and eMC websites, and corrected dosing is in Annex 1 of the notice. Aspire Pharma is recalling three batches of tropicamide 1% single-dose eye drops printed with an expiry of 30/04/2027 when the correct expiry is 31/10/2026, a date that has now passed. Separately, the 30 October update to the October Safety Roundup adds the Direct Healthcare Professional Communication (dated 23 October) for the antidepressant boxed suicide-risk warning and patient card that Issue 35 covered.

Alert

Product and batches

Problem

What MHRA asks

EL(26)A/46, Class 4 (caution in use)

Sandoz co-amoxiclav 250/62.5 mg/5 mL powder for oral suspension, PL 04416/0515; all batches

SmPC table gave about twice the dose; possible dose-related GI upset, dehydration, hypersensitivity; crystalluria very rare

Do not dose from the Sandoz SmPC; use Annex 1. Patients continue as prescribed and seek immediate medical advice if they have GI symptoms after a higher dose

EL(26)A/47, Class 3

Aspire tropicamide 1% eye drops, 20 x 0.5 mL single-dose, PL 35533/0198; batches E941A, E941B, E941C

Wrong expiry printed (30/04/2027; correct 31/10/2026)

Stop supplying, quarantine and return to supplier; no adverse incidents reported

Safety Roundup update, 30 October

Antidepressants (28 medicines in the review)

DHPC sent for the boxed suicide-risk warning, patient card and PSSD wording

Discuss suicidal thoughts when starting or changing dose; advise urgent help for new or worse suicidal thoughts

Source: MHRA EL(26)A/46 and EL(26)A/47 (5 November 2026) and Safety Roundup October 2026 (updated 30 October); gov.uk, accessed 5 November 2026.

CRITICAL APPRAISAL Who issued it and for whom

Yes

MHRA Defective Medicines Report Centre; the co-amoxiclav notice is distributed to pharmacy/wholesaler level and tagged for 39 specialties including emergency medicine and paediatrics; the tropicamide recall is for pharmacy and dispensing GPs, and patients are told to carry on using the drops

What it rests on

Yes

Manufacturer notifications (SmPC error; labelling error); the co-amoxiclav notice does not mention harm reports, and none have been reported for tropicamide

Scope

Yes

All Sandoz 250/62.5 mg/5 mL batches; three named tropicamide batches

Evidence vs opinion

Yes

Regulatory action; the corrected Annex 1 text is awaiting approval of the SmPC variation

Disagreement with UK guidance

No

None found

Design: Safety alert | Certainty: high (GRADE-style, our judgement) | What it changes for a UK ED: Tonight, check ED eye-room stock (or ask pharmacy) for Aspire tropicamide 1% batches E941A to E941C, whose true expiry was 31 October 2026, and quarantine them; dose paediatric co-amoxiclav from Annex 1 of the notice or the local paediatric formulary, not the Sandoz SmPC table | Confidence in that: high

UK practice: MHRA EL(26)A/46 Annex 1 (2026): for children under 40 kg the lower dose is 20 mg/5 mg/kg/day to 40 mg/10 mg/kg/day in three divided doses; with 50 mg/mL amoxicillin, a 6 kg child at 20 mg/kg/day gets 0.8 mL every 8 hours. The table marks 60 mg/15 mg/kg/day as not recommended up to 11 kg because there are no clinical data above 40 mg/10 mg/kg/day in children under 2, and children of 40 kg or more should have adult formulations. Report suspected reactions through the Yellow Card scheme.

EM Evidence Rundown, Issue 37. Page 13 of 25

INFORMING PRACTICE

OBSERVATIONAL

UK

UNCORKED (131 UK EDs, 18,744 Admissions): Sites That Held More of Their Escalation Care in Ambulances Had Higher 28-Day Mortality (Unadjusted 8.9% at the Highest Use vs 6.4% With None; Adjusted HR 1.021 per 5 Points)

McHenry RD, Birse F, Clarke B, Roberts T, on behalf of the Trainee Emergency Research Network. Ambulance or corridor? The association between site-level use of ambulance ‘ramping’ as Emergency Department escalation areas and 28-day mortality in admitted patients: a secondary analysis of the UNCORKED study. Scand J Trauma Resusc Emerg Med. 2026 (article in press). doi:10.1186/s13049-026-01710-5. PMID not yet assigned. ISRCTN16396025. Source route: full text (PDF; open access, CC BY-NC-ND). Appraised from full text

Post hoc secondary analysis of UNCORKED, the TERN prospective snapshot study, restricted to adults (16 and over) in 134 type 1 EDs in England, Wales and Northern Ireland at five time points in March 2025; 131 EDs used escalation areas and 18,744 admitted patients were followed to 28 days (1,240 deaths, 6.6%). The exposure was a site-level Ambulance: Escalation Index, the share of all escalation-area time spent in ambulances waiting more than 15 minutes to hand over, and the comparator was sites holding a smaller share (or none) in ambulances. Any use of ambulance escalation was not associated with mortality (HR 0.957, 95% CI 0.808 to 1.134). Among the 82 sites (62.6%) that used it, each 5-percentage-point rise in the index was associated with higher 28-day mortality (adjusted HR 1.021, 95% CI 1.002 to 1.041). It was adjusted for the site’s total escalation burden, admission rate, trauma designation, age, sex, snapshot and diagnosis group. The authors call the finding hypothesis-generating and say it is not an endorsement of corridor care.

Source: McHenry et al, Table 1 (unadjusted). Mortality is not monotonic across bands; the adjusted estimate is the dose term among sites using ambulances.

• 8.9% vs 6.4% 28-DAY MORTALITY, INDEX OVER 15% VS NO AMBULANCE USE (UNADJUSTED)
• HR 1.021 PER 5-POINT RISE IN INDEX, ADJUSTED (95% CI 1.002 TO 1.041)
• HR 0.957 ANY VS NO AMBULANCE ESCALATION (95% CI 0.808 TO 1.134)

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Mortality from hospital and GP records at 28 days; the exposure is a site ratio from five snapshots, not each patient’s own time in an ambulance, and ambulance waits under 15 minutes were not counted (misclassification, likely towards the null)

Design: Observational | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet for clinical care; supports escalation plans that do not simply shift waits from corridors into ambulances | Confidence in that: moderate

UK practice: The NHS England letter on additional actions to virtually eliminate corridor care (March 2026) defines corridor care as at least 45 minutes in a clinically inappropriate area of an ED or ward and aligns that threshold with the W45 ambulance protocol. It says ambulance handover delays should continue to be reported separately and monitored alongside corridor care so that reducing corridor care does not lengthen waits outside the hospital. The RCEM position statement on care in temporary escalation spaces (December 2024) says safe, good-quality care in corridors is not possible, and RCEM (4 March 2026) called the 45-minute threshold hard to justify. Issue 35 covered the paediatric UNCORKED analysis.

EM Evidence Rundown, Issue 37. Page 14 of 25

3 — PAEDIATRIC EMERGENCY MEDICINE

INFORMING PRACTICE

SR/MA

PAEDS

Video vs Direct Laryngoscopy for Children Intubated in an Emergency (ILCOR SR/MA, 18 Studies): Crude First-Pass Success 72.2% vs 69.0% (Pooled RR 1.11), Crude Severe Events 7.7% vs 11.8% (Pooled RR 0.76); Very Low Certainty, No Arrest or Survival Data

Couto TB, Acworth J, Rech L, et al. Video laryngoscopy versus direct laryngoscopy for pediatric emergency tracheal intubation: a systematic review and meta-analysis. Resusc Plus. 2026;32:101524. doi:10.1016/j.resplu.2026.101524. PMID: 42960018. PROSPERO CRD420251081694. Source route: PMC full text (PMC13717893); PubMed sweep. Appraised from full text

ILCOR Paediatric Life Support Task Force review of children intubated in the ED, PICU or prehospital setting (neonates at birth, NICU and theatre excluded), comparing video laryngoscopy (VL) with direct laryngoscopy (DL) as the first device. Eighteen studies were included, 17 observational (mainly retrospective cohorts and registries, seven from NEAR4KIDS) and one RCT of 120 children in a PICU. Crude pooled first-pass success was 7,737/10,720 (72.2%) with VL vs 9,651/13,994 (69.0%) with DL (pooled RR 1.11, 95% CI 1.06 to 1.16). Crude severe intubation-associated event rates were 7.7% vs 11.8% (pooled RR 0.76, 95% CI 0.62 to 0.93); overall success (RR 1.00, 0.99 to 1.01) and desaturation (RR 0.88, 0.77 to 1.01) did not differ. No study reported survival or cardiac arrest data, and GRADE certainty was very low for every outcome.

• 72.2% vs 69.0% FIRST PASS, VL VS DL (CRUDE)
• RR 1.11 FIRST PASS, POOLED (95% CI 1.06 TO 1.16)
• 7.7% vs 11.8% SEVERE EVENTS, VL VS DL (CRUDE)
• RR 0.76 SEVERE EVENTS, POOLED (95% CI 0.62 TO 0.93)

Source: Table 4 and subgroup analysis by design. Percentages are crude pooled rates; RRs are pooled across studies. All outcomes very low certainty (GRADE). Pooled estimates are unadjusted.

CRITICAL APPRAISAL Clear question; protocol registered

Yes ILCOR PICOST; PROSPERO CRD420251081694

Design: SR/MA | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet beyond current RCUK advice: use the device your team is trained on, and where VL is the local default for children, the data do not argue against it | Confidence in that: moderate

Main concerns: confounding by indication (VL kept for expected difficult airways) likely biases towards DL, while the rise in VL use over time alongside other airway changes likely overstates its benefit; net direction unknown; pooled estimates are unadjusted and seven studies come from overlapping NEAR4KIDS registry periods, which overstates precision Strengths: ILCOR-commissioned, registered, librarian-led search with GRADE and a prespecified registry-overlap sensitivity analysis

Disagreement: The authors suggest VL “may be considered as a first-line device” where equipment and training allow; RCUK 2025 leaves it to local protocols and rescuer experience; the one RCT pointed the other way for severe events (RR 1.24, 95% CI 0.51 to 3.01).

UK practice: Resuscitation Council UK Paediatric Life Support guidelines (2025) say to use video or direct laryngoscopy “depending on local protocols and rescuer experience” and not to use unfamiliar equipment, to intubate only if competent with a checklist, and to limit attempts to four. This review does not change that; the authors call for RCTs and arrest data.

EM Evidence Rundown, Issue 37. Page 15 of 25

INFORMING PRACTICE

OBSERVATIONAL

PAEDS

Who Intubates Children in the ED (NEAR4PEM Registry, 642 Trainee Intubations): First-Attempt Success 75% for EM Residents, 74% for PEM Fellows, 58% for Paediatric Residents

Jackson J, Greenwald E, Kerrey BT, et al. Who’s on Airway? Trainee Tracheal Intubation Frequency and Outcomes Within the National Emergency Airway Registry for Pediatric Emergency Medicine (NEAR4PEM). Ann Emerg Med. 2026 (published online 27 October 2026). doi: 10.1016/j.annemergmed.2026.08.729. PMID: 42949693. Source route: PubMed sweep; abstract only (publisher page not accessible, not in PMC); no JournalFeed, SGEM, REBEL EM, St Emlyn’s or First10EM commentary found. Abstract only

Retrospective analysis of the NEAR4PEM registry of intubations in academic paediatric EDs (January 2021 to December 2023). Of 1,404 encounters, 642 first attempts were by trainees: emergency medicine residents 343 (53%), paediatric emergency medicine (PEM) fellows 254 (40%) and paediatric residents 45 (7%). The primary outcome was first-attempt success, 471/642 (73%) overall: 75% for EM residents, 74% for PEM fellows and 58% for paediatric residents. After adjustment, EM residents and PEM fellows did not differ in first-attempt success (aOR 0.84, 95% CI 0.56 to 1.28) or adverse airway outcomes (aOR 1.4, 95% CI 0.96 to 2.18), while paediatric residents had lower success (aOR 0.41, 95% CI 0.20 to 0.82) and more adverse airway outcomes (aOR 2.14, 95% CI 1.08 to 4.27) than EM residents.

• 75% FIRST-ATTEMPT SUCCESS, EM RESIDENTS (N=343)
• 74% PEM FELLOWS (N=254)
• 58% PAEDIATRIC RESIDENTS (N=45)
• aOR 2.14 ADVERSE AIRWAY OUTCOME, PAEDS VS EM RESIDENTS (1.08 TO 4.27)

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Registry-recorded first-attempt success and adverse events (collection method not described in the abstract); under-reporting would understate harm in every group

Design: Observational | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; a prompt to check who in your department is expected to intubate a child and how they are supervised | Confidence in that: moderate

Main concerns: selection of easier airways for less experienced trainees likely understates the difference between groups; individual intubation experience was not measured, so job title stands in for skill (direction of that bias unknown) Strengths: multicentre registry data with a clear, patient-relevant primary outcome

UK practice: Resuscitation Council UK Paediatric Life Support guidelines (2025) say to intubate a child only if competent, with the drugs and equipment immediately available, to use an intubation checklist and to have a plan for unexpected difficulty, including calling for additional expertise. The registry does not measure individual experience, so it cannot say which UK grade should hold the laryngoscope.

EM Evidence Rundown, Issue 37. Page 16 of 25

INFORMING PRACTICE

DIAGNOSTIC

PAEDS

Continuous Pulse-Oximetry Model for Infants 90 Days or Under With Bronchiolitis (Single Centre, n=117): Sensitivity 78% in a 36-Infant Validation Set; a Low-Risk Result Still Left About 1 in 4 Deteriorating

McLaren SH, Megjhani M, Camargo CA, et al. Continuous Pulse Oximetry Analysis to Predict Outcomes in Young Infants With Bronchiolitis Presenting to the Emergency Department. Pediatr Emerg Care. 2026 (published ahead of print 19 October 2026). doi:10.1097/PEC. 0000000000003735. PMID: 42894352. Source route: JournalFeed EM SpeedRead 3 Nov 2026; open-access full text on the publisher’s site. Appraised from full text

Retrospective derivation and internal validation study in one US paediatric ED (November 2021 to November 2025) of infants aged 90 days or under with bronchiolitis. Heart rate and SpO2 recorded every second during the first 2 hours were reduced to three features (median heart rate, heart-rate IQR and SpO2 IQR) in a penalised logistic regression; the outcome was high-flow, CPAP, BiPAP or invasive ventilation within 24 hours, which occurred in 57 of 117 (49%). In the validation set (n=36, 18 events) the primary model had an AUROC of 0.78 (95% CI 0.62 to 0.93), sensitivity 0.78 (0.58 to 0.95) and specificity 0.67 (0.43 to 0.87), short of the prespecified 90% sensitivity target. Four of 18 infants who deteriorated were called low risk; three needed CPAP and one BiPAP, and none was intubated.

• 78% SN, VALIDATION (58 TO 95)
• 67% SP, VALIDATION (43 TO 87)
• LR+ 2.3 CALCULATED, CI NOT REPORTED
• LR− 0.33 CALCULATED, CI NOT REPORTED

Calculated from the validation counts: 14 true positives, 4 false negatives, 12 true negatives, 6 false positives (Tables 2 and 3). CRITICAL APPRAISAL Level of evidence

Partly Derivation with a 70/30 internal split; no external validation; authors call it proof of concept

a short monitored period to label a young infant low risk | Confidence in that: high

Main concerns: exclusion of infants without usable monitor data (60% of those eligible, calculated) means the model was not tested in the low-risk infants it is meant for; direction unknown; a 36-infant validation set gives very wide CIs, and the 70/30 split from one centre likely overstates performance elsewhere Strengths: prespecified operating point and outcome, honest reporting of each missed infant

UK practice: NICE NG9 (bronchiolitis, 2015, updated 2021) asks for pulse oximetry in every child with bronchiolitis in secondary care (1.3.1) and admission for apnoea, persistent SpO2 below 90% (below 92% under 6 weeks or with underlying conditions, when breathing air), inadequate oral fluid intake (50% to 75% of usual) or severe respiratory distress (1.3.2). It lists age under 3 months as a risk factor for severe disease (1.3.3), which covers every infant in this study. NG9 says nothing about analysing continuous monitor data, and this study gives no reason to discharge a young infant who meets any NG9 admission criterion.

EM Evidence Rundown, Issue 37. Page 17 of 25

INFORMING PRACTICE

OBSERVATIONAL

PAEDS

Lactate and Bicarbonate After Paediatric Drowning (Single US Centre, n=208): Mean Lactate 4.8 vs 12.9 mmol/L in Survivors vs Non-Survivors; AUC 0.97 for Death but 0.77 for Admission

Jamasbi A, Nguyen L, Papa L, Weatherford T. Initial Emergency Department Concentrations of Serum Bicarbonate and Lactate Are Associated With Survival to Hospital Discharge in Children Following Submersion Injuries. Pediatr Emerg Care. 2026 (published online 4 November 2026). doi:10.1097/PEC.0000000000003729. PMID: 42978141. Source route: PubMed sweep; abstract only (no free full text); no secondary commentary found. Abstract only

Cross-sectional cohort of consecutive children seen after submersion injury at one US paediatric level 1 trauma centre (February 2018 to November 2019), with Utstein-style data collection. Of 208 children (mean age 47 months, 66% given bystander CPR), 11 (5%) died, 135 (65%) were admitted and 52 (25%) went to ICU. Mean venous lactate was 4.8 mmol/L (95% CI 4.3 to 5.4) in survivors vs 12.9 mmol/L (10.9 to 14.9) in non-survivors, and mean venous bicarbonate 19 vs 11 mmol/L. Lactate discriminated death well (AUC 0.97, 95% CI 0.95 to 1.00) but admission less well (AUC 0.77, 0.67 to 0.88), and bicarbonate less well again (AUC 0.93 for death, 0.69 for admission).

• 4.8 mmol/L MEAN LACTATE, SURVIVORS (4.3 TO 5.4)
• 12.9 mmol/L MEAN LACTATE, NON-SURVIVORS (10.9 TO 14.9)
• AUC 0.97 LACTATE FOR DEATH (0.95 TO 1.00)
• AUC 0.77 LACTATE FOR ADMISSION (0.67 TO 0.88)

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Venous lactate and bicarbonate with hard outcomes; how many children had bloods taken, and when, is not reported

Design: Observational | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; a very high lactate after drowning marks a sick child, but no threshold is available to support discharge | Confidence in that: moderate

Main concerns: no adjustment for arrest on arrival or CPR duration likely overstates what the blood gas adds to the clinical picture; with 11 deaths and no cut-off reported, the AUC of 0.97 cannot be turned into a safe discharge rule; the abstract conclusion describes “Elevations in venous bicarbonate and lactate levels”, but bicarbonate was lower in non-survivors (11 vs 19 mmol/L) Strengths: consecutive cases with Utstein-style data and hard outcomes

UK practice: Resuscitation Council UK Paediatric Life Support guidelines (2025) (drowning section) prioritise early reversal of hypoxia, 100% oxygen, treatment of hypothermia, a search for an underlying cause and a check of glucose and electrolytes; it gives no lactate or bicarbonate threshold and no discharge criteria. We found no UK guidance that uses blood gas values to decide disposition after drowning.

EM Evidence Rundown, Issue 37. Page 18 of 25

4 — FOAMED & CRITICAL APPRAISAL

INFORMING PRACTICE FOAMED

EMCrit 437 (Weingart With Josh King): Massive Lithium Overdose — No Charcoal, Consider a Binder Then Whole Bowel Irrigation, Repeat Levels Often and Keep a Low Threshold for Dialysis When a Raised Level Comes With Confusion

Weingart S, with King J. EMCrit 437 – Massive Lithium Overdose. EMCrit Podcast, 29 October 2026. emcrit.org (show notes read; audio not heard). Background: Decker BS, Goldfarb DS, Dargan PI, et al. Extracorporeal treatment for lithium poisoning: systematic review and recommendations from the EXTRIP Workgroup. Clin J Am Soc Nephrol. 2015;10(5):875–887. doi:10.2215/CJN.10021014. PMID: 25583292. Source route: EMCrit website. Appraised from full text

Weingart talks to a US medical toxicologist and nephrologist about the large acute lithium ingestion. Charcoal does not bind lithium; King suggests one or two doses of sodium polystyrene sulfonate (off-label, with gut necrosis risk if repeated) followed by whole bowel irrigation, and says gastric lavage is now rarely done in time. He stresses that patients can look well early and deteriorate later, that sinus bradycardia is the usual cardiac finding and a wide QRS in a mixed overdose points to a co-ingestant, and that the aim of dialysis is to prevent lasting neurotoxicity rather than death. Practical advice: repeat levels as often as the lab can turn them round, restore normovolaemia and then give double maintenance fluids, without diuretics or forced diuresis, and involve the poisons service and nephrology early.

Source: Decker et al, CJASN 2015, abstract (PubMed); EXTRIP reports mEq/L, which equals mmol/L for lithium. 166 articles, mostly case reports.

CRITICAL APPRAISAL Who wrote it and for whom

Yes US ED intensivist podcast with a toxicologist-nephrologist who is medical director of the Maryland Poison Center; for emergency and critical care clinicians

Design: FOAMed | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing new for protocols; for a large lithium ingestion, involve NPIS and the renal team early and repeat levels | Confidence in that: moderate Disagreement: King now reserves gastric lavage for very high-lethality drugs such as colchicine and flecainide, whereas the Priadel SmPC says consider it within 1 hour for more than 4 g of a non-sustained-release preparation in an adult, and the RCEM/NPIS guideline says gastric lavage should not be routine and only after discussion with NPIS. EXTRIP stops dialysis once lithium is under 1.0 mmol/L or the patient improves, while the SmPC says continue until no lithium is left in serum or dialysis fluid; the binder is off-label and not mentioned in either UK document.

UK practice: The Priadel SmPC, section 4.9 (eMC, updated April 2025) says there is no antidote, activated charcoal does not adsorb lithium, diuretics should not be used, all patients should be observed for at least 24 hours, and haemodialysis is the treatment of choice for severe intoxication or overdose with renal impairment. The RCEM/NPIS unidentified poisoning guideline (April 2025) lists lithium among poisons poorly adsorbed by charcoal and says whole bowel irrigation and renal replacement therapy are agent specific and should follow TOXBASE or NPIS advice.

EM Evidence Rundown, Issue 37. Page 19 of 25

5 — QUICK HITS

Correction to Issue 28, TARGET-CTCA (UK RCT, n=3,170; re-read from the accepted manuscript; 7.1% as in the published abstract, where the manuscript prints 7.0%): Routine outpatient CTCA after MI rule-out with intermediate hs-troponin did not reduce MI or cardiac death over a median 3.0 years (112 of 1,587, 7.1%, vs 7.3%; adjusted HR 0.95, 95% CI 0.73 to 1.23). Issue 28 said this interval “definitively rules out a clinically meaningful benefit”; it does not, as it is compatible with anything from a 27% reduction to a 23% increase. Issue 28 also described these patients as having a negative hs-troponin, when all had an intermediate result, and advised reserving CTCA for intermediate-risk patients (HEART score 4 or more, unexplained ECG changes or strong clinical suspicion); the trial selected patients by troponin alone, and its HEART-stratified subgroups did not suggest greater benefit above a score of 3. Practice is unchanged: follow NICE CG95 (recs 1.2.6.5, 1.2.6.6 and 1.3.4.3). Lee KK, et al. N Engl J Med 2026 (TARGET-CTCA). doi:10.1056/NEJMoa2608903. PMID: 42670980

AHA/ACC 2026 vs ESC 2019 PE severity classes (retrospective cohort, n=2,251, two-campus Chinese hospital; abstract only): The AHA/ACC categories discriminated 30-day mortality better than the ESC strata (AUC 0.885 vs 0.763; difference 0.122, 95% CI 0.084 to 0.160), and the 37 patients in AHA/ACC categories D or E but ESC intermediate risk had 75.7% mortality by 30 days. Inpatients only and retrospective; the gap narrowed slightly (0.112, 95% CI 0.038 to 0.186) when categories A and E were excluded. Fan J, et al. Chest 2026. PMID: 42810427

Multimarker calculator vs AHA/ACC vs ESC in stable PE (prospective registry, n=11,568, China; abstract only): 30-day mortality was 2.9%; AUC was 0.704 for the multimarker calculator and 0.680 for the 7-category AHA/ACC scheme vs 0.606 for the ESC model (CIs not reported in the abstract). The original calculator overestimated intermediate-high-risk mortality (predicted 16.2%, observed 9.4%) until recalibrated, so it needs validation outside China before UK use. Tao Y, et al. Eur Heart J Acute Cardiovasc Care 2026. PMID: 42948904

POCUS for PE under the 2026 AHA/ACC guideline (narrative review, US authors, J Emerg Med pre-proof): The authors give POCUS three roles: an adjunct in narrow high-yield scenarios (the unstable patient, or when definitive imaging is delayed or unsafe), risk stratification after PE is confirmed, and no role in ruling PE out in low-risk or unselected patients. Single findings are specific but insensitive (McConnell’s sign sensitivity 29%, 95% CI 20% to 40%, specificity 99%; two-point leg compression sensitivity 44%, 95% CI 36% to 51%, from a 2022 meta-analysis of 70 studies), so a normal scan does not exclude PE. O’Hare C, et al. J Emerg Med 2026 (pre-proof). doi:10.1016/j.jemermed.2026.10.022. No PMID yet

AI-ECG decision tree to rule out ACS after ROSC (abstract only; single-centre retrospective, Korea, n=329 development and 71 temporal validation): the tree classed 52.0% and 45.1% of post-arrest patients as “ACS not likely” with no adjudicated ACS case misclassified, mostly in non-shockable rhythms (AUROC 0.824 in that subgroup). Zero false negatives was a design constraint and the number of ACS cases is not in the abstract, so the lower confidence bound of sensitivity cannot be judged. Choi A, et al. Resuscitation 2026. PMID: 42956925

Eight equipment-free bedside signs vs NEWS (abstract only; retrospective, one US hospital, n=4,581 patients at rapid response calls): a model of jaundice, supplemental oxygen, capillary refill, a cold body, distressed breathing, altered mentation, dry skin and cyanosis discriminated death within 24 hours better than NEWS (AUC 0.790 vs 0.707; standard errors 0.014 and 0.020, CI not reported). Ward patients at rapid response arrival, not ED triage, and the abstract reports no external validation. Sebat C, et al. Resusc Plus 2026. PMID: 42985299

Xylazine test strips on urine (diagnostic, 250 fentanyl-positive specimens, single US centre; abstract only): against LC-MS, strips had sensitivity 26.8% (95% CI 16.9 to 38.6%) and specificity 77.7% (95% CI 70.8 to 83.5%), LR+ 1.2 and LR− 0.94; 91.5% of LC-MS-positive specimens were below the strip’s 1,000 ng/mL threshold. Strips made for drug samples do not work as a urine test; fentanyl positivity was not confirmed by LC-MS. Chun R, et al. Clin Toxicol 2026. PMID: 42960873

Cardiac toxicity in poisoning (cohort, 9,743 presentations, Australian toxicology service; abstract only): cardiac complications occurred in 473 (4.9%), mostly hypotension (368, 3.8%) or isolated bradycardia (139 of 164 dysrhythmias, 85%); ventricular dysrhythmias occurred in 6 (0.06%), with sotalol, methadone, escitalopram, dothiepin (dosulepin) and nortriptyline, and no death was attributed to a primary dysrhythmia. One specialist-service population; sinus tachycardia not counted as a complication; no CIs given. Berling I, et al. Clin Toxicol 2026. PMID: 42961000

Venous pO2 on the VBG and deterioration (abstract only; ambispective single-centre US ED cohort, n=973): median venous pO2 was 4.4 kPa (33 mmHg; kPa calculated) and was not associated with ICU admission, death, vasopressors or positive-pressure ventilation within 48 hours after adjustment for age and lactate (aOR 1.054 per 10

EM Evidence Rundown, Issue 37. Page 20 of 25

mmHg, 95% CI 0.957 to 1.161); adding it moved the AUROC from 0.649 to 0.654. The authors call for consecutive-enrolment multicentre confirmation and cannot exclude smaller associations. Corbo J, et al. Acad Emerg Med 2026;33(10). PMID: 42947281

Procalcitonin for pneumonia in acute breathlessness (abstract only; diagnostic SR/MA, 6 studies, n=2,341, 576 with pneumonia): pooled sensitivity 0.67 (95% CI 0.58 to 0.75) and specificity 0.76 (95% CI 0.67 to 0.83), LR+ 2.8 (95% CI 2.0 to 3.8) and LR− 0.44 (95% CI 0.34 to 0.56). Only two databases were searched, and the abstract does not report heterogeneity; procalcitonin alone cannot confirm or exclude pneumonia. Cui JY, et al. Am J Emerg Med 2026;111:103–109. PMID: 42971102

Low-level pyuria accounts for a disproportionate share of antibiotic overtreatment in children (cross-sectional, n=1,150, 4 US ED and urgent care settings; abstract only): trace leukocyte esterase, 10–69 WBC/mm³ or 5–10 WBC/high-power field was associated with starting antibiotics (aOR 6.62, 95% CI 3.72 to 11.79 vs no pyuria) and accounted for 17.0% (10.3% to 23.2%) of treatment in children whose culture excluded UTI, against 1.8% (1.1% to 2.4%) of treatment for confirmed UTI. US prescribing in one region; read low-grade pyuria alongside age, symptoms and the rest of the dipstick. Wattier RL, et al. Pediatrics 2026. PMID: 42982088

National PEWS to be rolled out in English EDs (RCEM press release, 2 Nov 2026; BMJ news): RCEM welcomed the rollout of the National Paediatric Early Warning System to A&Es in England, said it had been involved in its development from the outset, and welcomed the commitment to supporting resources and eventual digital implementation. It said requiring software developers to build it into their systems will be key to its success. The press release gives no start date or chart details. RCEM, 2 November 2026; Furneaux R. BMJ 2026;395:e101323. PMID: 42977314

New RCEM Learning modules (30 October to 5 November): “A Malevolent Mosquito” (returned traveller with fever), “Can’t Bear To Weightbear” (a 79-year-old with non-traumatic hip pain and fever), “Take my breath away” (a 27-year-old with acute breathlessness and right-sided chest pain), the SBA “A child with fever and epistaxis” (tagged acute leukaemia) and Curriculum Cup: Gastroenterology and Hepatology Part 2 (25 questions in 25 minutes). rcemlearning.co.uk · module · module · SBA · quiz

EM Evidence Rundown, Issue 37. Page 21 of 25

CORE REVISION — ISSUE 37 — 5 NOVEMBER 2026

Core Revision: Clinical — Paediatric Resuscitation Essentials

Anchored on RCUK 2025 Paediatric Life Support — FRCEM

Exam goal: Pick the best weight estimate, give RCUK 2025 energies and doses from memory, explain the child’s airway and why most arrests are non-shockable, and summarise THAPCA.

1. WEIGHT ESTIMATION

2. SHOCKS AND DRUGS: RCUK 2025 PALS ALGORITHM

EM Evidence Rundown, Issue 37. Page 22 of 25

Memory aid — “4, 10, 5, 10, 2”: 4 J/kg shock; 10 mcg/kg adrenaline; 5 mg/kg amiodarone; 10 mL/kg crystalloid bolus; 2 mL/kg 10% glucose. Shockable-rhythm drugs follow the 3rd and 5th shocks.

3. WHY THE PAEDIATRIC AIRWAY DIFFERS

4. WHY MOST PAEDIATRIC ARRESTS ARE NON-SHOCKABLE

5. AFTER ROSC: TEMPERATURE AND TARGETS

FRCEM focus: rank the weight methods; recite 4 J/kg, 10 mcg/kg (max 1 mg) and 5 mg/kg (max 300 mg, then 150 mg); explain the UK–AHA first-shock difference; quote THAPCA as neutral. Sources are in each table.

EM Evidence Rundown, Issue 37. Page 23 of 25

6 — ACTION POINTS

EM Evidence Rundown, Issue 37. Page 24 of 25

TRIALS TO WATCH

Q4 2026

NICE sepsis in pregnancy Draft partial update of NICE NG255 (GID-NG10468) out for consultation until 11 November 2026; final publication still expected 21 January 2027.

2026–2027

LONGER HORIZON

PENG vs fascia iliaca block RCEM’s fascia iliaca block guideline (updated 2025) lists PENG vs fascia iliaca block as a research recommendation; the PENG trial in Issue 36 compared PENG only with IV morphine.

Jake Turner — Curated with the assistance of AI (Claude). All content editorially reviewed. Feedback form · emevidence.org · emevidence999@gmail.com

EM Evidence Rundown, Issue 37. Page 25 of 25