EM Evidence Rundown — Issue 36

EM EVIDENCE RUNDOWN — ISSUE 36 — 29 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 (Claude). All content editorially reviewed.

Full archive and PDF downloads at emevidence.org

Lead: Crashing PE in 2026: spot normotensive shock before the blood pressure falls, and know the UK limits on lysis and catheter treatment; only 38% of UK physicians in a 2024 survey reported a PE response team. Change Tonight: Nova StatStrip ketone strips: on 16 October MHRA updated its alert and the manufacturer’s field safety notice to add a further affected batch, so re-check stock against the 12 listed lots. Paediatric EM (3 items): CT clearance of the cervical spine, IV ketamine for suicidal adolescents, and air vs oxygen at birth; croup steroid dose, intranasal ketamine sedation, limb imaging and the saphenous vein are in Quick Hits.

BOTTOM LINE UP FRONT — 29 OCTOBER 2026

ACT ON THIS NOW

MHRA Ketone strips: MHRA’s 16 October update to the Nova StatStrip recall alert adds a further affected batch; re-check adult and paediatric ED stock against the 12 listed lots.

GUIDELINE Unknown poisoning (RCEM/NPIS): stabilise first, use the toxidrome and TOXBASE, send a paracetamol level whenever self-harm is possible, and do not act on point-of-care urine drug screens; all recommendations are consensus (level C).

WORTH KNOWING ON SHIFT

FEATURE Crashing PE: lactate over 2 mmol/L, capillary refill over 3 s or mottling can flag shock before hypotension; NICE NG158 supports systemic lysis only with haemodynamic instability.

RCT Hip fracture (n=34): a PENG block cut pain by a median 6 vs 3 points at 30 minutes against one dose of IV morphine, with no desaturation after the block (0/17 vs 3/17).

PAEDS Cervical spine: in 12,693 children CT sensitivity for injuries needing surgery or a halo was 94.7% (98.7% after expert re-review).

KNOW FOR NEXT TIME

SR/MA Induction agents (21 RCTs, n=6,031): ketamine caused more peri-intubation instability than etomidate (RR 1.31, 95% CI 1.15 to 1.48), with no difference in mortality (RR 0.96).

SR/MA Steroids in septic shock: against placebo, hydrocortisone 200–299 mg/day alone did not reduce 28-day mortality (OR 0.92, 95% CI 0.81 to 1.04) and hydrocortisone plus fludrocortisone did (OR 0.78, 0.64 to 0.96), but the two regimens did not differ significantly from each other (OR 1.17, 0.92 to 1.48).

SR/MA DKA fluids: two-bag systems were associated with less hypoglycaemia (9.2% vs 23.9%) in mostly non-randomised studies; JBDS-IP already runs 10% glucose alongside saline.

COHORT Hyperkalaemia: coded arrest, VT or VF within 48 h was 6.68% at potassium 7.0 or more vs under 1% in matched controls; a risk marker, not proof of cause.

RCT Late thrombolysis (HOPE): alteplase at 4.5 to 24 h with perfusion mismatch gave mRS 0–1 in 40.3% vs 26.3%, with sICH 3.8% vs 0.5%.

PAEDS Newborn resuscitation (Cochrane): starting in air may reduce death before discharge (RR 0.72, 95% CI 0.57 to 0.91; low certainty).

EM Evidence Rundown, Issue 36. Page 1 of 27

This week’s theme is the crashing pulmonary embolism. The lead feature sets out how to recognise shock before the blood pressure falls, what the new American categories mean against NICE, and what a district general hospital can actually offer. Three papers follow: STORM-PE’s 90-day walking distances, a review of ECMO with catheter clot removal, and the argument over STRATIFY’s deaths. Elsewhere, two network meta-analyses look at induction agents and steroids in septic shock, and a PENG block trial adds to the case for regional anaesthesia in hip fracture. RCEM and the National Poisons Information Service have a new guideline on the patient with an unknown poison. Paediatrics has three items, including a large study of CT clearance of the cervical spine.

Correction to Issue 35: the Quick Hit on the point-of-care hs-troponin systematic review (van Daalen et al, Eur J Emerg Med 2026, PMID: 42894477) said the review reported its results without confidence intervals. It did not pool results, but it does give study-level 95% CIs (its Table 3). The Issue 35 web page is being corrected.

WHAT’S INSIDE — ISSUE 36

• Lead feature: Crashing PE in 2026, with STORM-PE at 90 days, ECMO plus catheter clot removal, and the STRATIFY safety debate
• Airway and shock: induction agents for intubation; hydrocortisone regimens in septic shock
• Metabolic: two-bag DKA fluids; hyperkalaemia and short-term outcomes
• Trauma and analgesia: observing a blunt traumatic pneumothorax; PENG block vs IV morphine for hip fracture
• UK: RCEM/NPIS guideline on unidentified poisoning; MHRA update on StatStrip ketone strips
• PEM (3): CT clearance of the cervical spine, IV ketamine for suicidal adolescents, air vs oxygen at birth
• FOAMed and appraisal: HOPE late thrombolysis with SGEM; what IVC collapse means; choosing vasoactives in cardiogenic shock
• Quick Hits (14): including adaptive D-dimer, ECPR timing, biphasic anaphylaxis, troponin in SVT, croup dexamethasone dose and nitazenes

• Core Revision: ECG — broad complex tachycardia, VT vs SVT with aberrancy

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 36. Page 2 of 27

1 — KEY TRIALS & ARTICLES

LEAD

INFORMING PRACTICE

REVIEW

UK

FRCEM

Crashing PE in 2026: Look for Shock Before the Blood Pressure Falls, and Know Your Escalation Route: Only 38% of UK Physicians in a 2024 Survey Reported a PE Response Team

Feature built on: Schramm DC, Morozow A, Mandal I, et al. An overview of interventional pulmonary embolism management in the United Kingdom in 2026. Clin Radiol. 2026;100:107406. doi:10.1016/j.crad.2026.107406. PMID: 42442124. Hunsicker O, Daum N, Hernandez G, et al. Acute pulmonary embolism: clinical and echocardiographic recognition of normotensive shock. Crit Care. 2026;30(1):386. doi:10.1186/s13054-026-06214-3. PMID: 42502164. Guideline: Creager MA, Barnes GD, Giri J, et al. 2026 AHA/ACC/ACCP/ACEP/CHEST/SCAI/SHM/SIR/SVM/SVN Guideline for the Evaluation and Management of Acute Pulmonary Embolism in Adults. Circulation. 2026;153(12):e977–e1051. doi: 10.1161/CIR.0000000000001415. PMID: 41712677 (quoted via the summaries below). Barnes GD, Burnett A, Giri J, Naydenov S. The 2026 AHA/ACC guideline for the management of acute pulmonary embolism. Endovascular Today, July 2026. Weingart S. EMCrit 424: 2026 Crashing PE Update, 2 May 2026. Source route: Schramm PDF via BMA Library (supplied by Jake); Hunsicker PMC full text; EMCrit page PDF supplied by Jake (cited and summarised, not reproduced); PRAGUE-26 appraised in Issue 35. Appraised from full text

Recognise it. The 2026 AHA/ACC guideline replaces “massive” and “submassive” with categories A to E. Category D1 is transient hypotension (such as syncope), D2 is normotensive shock (raised lactate or acute kidney injury despite a normal blood pressure), E1 is persistent hypotension with cardiogenic shock and E2 is refractory shock or arrest. Hunsicker et al list the guideline’s markers (lactate above 2 mmol/L, rising creatinine, urine under 0.5 mL/kg/h, altered mental state, cardiac index 2.2 L/min/m² or less) and add capillary refill over 3 seconds, mottling and echo. UK practice uses ESC 2019 terms (Schramm); NG158 uses only stable and unstable.

Sources: Schramm 2026; NICE NG158, HTG705, HTG376, Actilyse SmPC (checked 29 Oct 2026); RCUK 2025.

Opinion, not guidance (EMCrit 424 and our reading), a DGH pathway: (1) In PE with RV strain, check and repeat lactate, capillary refill, mottling, urine output and an RV view on echo. (2) Call your regional PE, IR or cardiology team at D2, not when the blood pressure falls. (3) In E1 without a contraindication, consider full-dose alteplase, as NG158 1.6.2 advises. (4) If lysis is likely, EMCrit advises femoral lines first to keep ECMO or thrombectomy possible. (5) Arrest from suspected PE: lyse and consider 60–90 minutes of CPR (RCUK 2025).

• 38% UK PHYSICIANS REPORTING A PERT (2024 SURVEY, SCHRAMM REF 56)
• 10% UNSURE IF ONE EXISTED
• 57% CDT CENTRES ALSO OFFERING THROMBECTOMY (BSIR 2023)
• 2,756 UK DEATHS CODED TO PE, 2024

CRITICAL APPRAISAL Who wrote it and for whom

Yes Schramm: UK radiologists; one author holds a Penumbra grant (may tilt it towards catheter therapy)

Design: Review | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Agree a written escalation route and add serial perfusion markers to PE work-up; no change to who gets lysis | Confidence in that: moderate Disagreement: AHA/ACC 2026 (quoted by EMCrit 424): catheter lysis or thrombectomy “may be considered” in D1–D2; NICE allows catheter lysis only with special arrangements or research (HTG376), and intermediate-risk thrombectomy only in research (HTG705).

UK practice: NICE NG158 (2020, updated 2023) recs 1.6.2–1.6.3: consider systemic lysis with haemodynamic instability; not when stable, even with RV dysfunction. NICE HTG705 (formerly IPG778, November 2023) 1.1, 1.4: thrombectomy with special arrangements only in high-risk PE if lysis is unsuitable or failed; else research.

EM Evidence Rundown, Issue 36. Page 3 of 27

INFORMING PRACTICE RCT

STORM-PE 90-Day Outcomes (RCT, n=100): After Catheter Clot Aspiration, Patients Walked Further (479 vs 368 m), but Only About Two-Thirds of Each Arm Did the Walk Test

Rosovsky RP, Konstantinides SV, Weinberg I, et al. Mechanical Thrombectomy Versus Anticoagulation Alone in Intermediate-High Risk PE: Improved Functional 90-Day Outcomes in STORM-PE RCT. JACC Cardiovasc Interv. 2026;19(20):2775–2790. doi:10.1016/j.jcin.2026.07.042. PMID: 42836810. NCT05684796. Sponsored and funded by Penumbra. Source route: PDF via BMA Library (supplied by Jake). Primary results covered in Issues 7 and 12. Appraised from full text

STORM-PE randomised 100 adults with intermediate-high-risk PE at 22 sites (USA, Canada, Poland, New Zealand) to computer-assisted vacuum thrombectomy (Lightning Flash 16F) plus anticoagulation (n=47) or anticoagulation alone (n=53); the primary endpoint, RV/LV ratio change at 48 hours, was reported earlier. This paper reports secondary outcomes to 90 days, without adjustment for multiple comparisons. Among those who did the test, 6-minute walk distance at 90 days was 478.7 ± 159.5 m vs 368.0 ± 127.8 m (P=0.003; CI for the difference not reported), and 97.4% (37/38) vs 75.6% (34/45) were NYHA class I (OR 11.96, 95% CI 1.47 to 97.61). Quality-of-life scores, symptomatic recurrence and mortality did not differ significantly; there were 4 deaths (8.5%) with thrombectomy vs 1 (1.9%) with anticoagulation, none judged device- or procedure-related.

• 479 vs 368 m 90-DAY WALK DISTANCE, COMPLETERS (P=0.003; CI NOT REPORTED)
• 66.0% vs 64.2% DID THE 90-DAY WALK TEST (31/47 VS 34/53)
• NNT 5 NYHA CLASS I AT 90 DAYS, CALCULATED: ARD 21.8 PP (95% CI 6.9 TO 36.2)
• 8.5% vs 1.9% DEATHS BY 90 DAYS (4/47 VS 1/53; P=0.133)

Outcome at 90 days Thrombectomy + AC (n=47) AC alone (n=53) Effect

Source: Rosovsky et al, Results and abstract. ARD, NNT and their CIs calculated by us (Newcombe method) among patients with NYHA data; AC = anticoagulation.

CRITICAL APPRAISAL Registered question; these outcomes prespecified

Partly NCT05684796; functional and QoL outcomes were secondary, and several analyses (imputation, % predicted walk) were post hoc

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: thrombectomy for intermediate-risk PE stays a research or specialist-team decision | Confidence in that: moderate Disagreement: The authors conclude that the trial “supports CAVT as an effective and safe reperfusion strategy”; their own limitations section notes about one-third missing walk tests and no multiplicity adjustment. NICE keeps thrombectomy for intermediate-risk PE to research.

UK practice: NICE HTG705 (formerly IPG778, November 2023) rec 1.4: for intermediate-risk PE, percutaneous thrombectomy “should only be used in research”; NICE is updating it (GID-HTG10169), with publication expected 22 July 2027. NICE NG158 (2020, updated 2023) rec 1.6.3: do not offer systemic thrombolysis when haemodynamically stable.

EM Evidence Rundown, Issue 36. Page 4 of 27

INFORMING PRACTICE SR/MA

ECMO Plus Catheter Clot Removal vs ECMO Alone in High-Risk PE (SR/MA, 4 Cohorts, n=2,824): No Proven Mortality Benefit, and Only 47 Patients Could Be Pooled

Hasan B, Aldin ST, Buda KG, et al. Extracorporeal Membrane Oxygenation Combined With Mechanical Thrombectomy Versus Extracorporeal Membrane Oxygenation Alone for High-Risk Pulmonary Embolism: A Systematic Review and Meta-analysis. J Cardiothorac Vasc Anesth. 2026 (online 31 August). doi:10.1053/j.jvca.2026.08.165. PMID: 42786027. Source route: PDF via BMA Library (supplied by Jake). Appraised from full text

Mayo Clinic reviewers searched PubMed and Embase to 6 October 2025 for adults with high-risk PE on VA-ECMO, comparing ECMO plus catheter thrombectomy with ECMO alone; the primary outcome was in-hospital mortality. Four retrospective cohorts (n=2,824) qualified, but only two single-centre studies (n=47) could be pooled: OR 0.36 (95% CI 0.01 to 13.59; I² 69%). The two larger studies were not pooled because one was adjusted and one mixed catheter with surgical embolectomy: a US national database (n=2,285; mechanical reperfusion, including surgery) gave adjusted OR 0.75 (0.47 to 1.19), and the international CATAPULTE cohort gave 90-day mortality of 32% vs 39% after propensity matching (HR 0.68, 0.45 to 1.03). Major bleeding did not differ (adjusted OR 0.93, 0.58 to 1.48), and certainty was very low for every mortality outcome.

• OR 0.36 IN-HOSPITAL DEATH, 2 POOLED COHORTS, N=47 (0.01 TO 13.59)
• aOR 0.75 IN-HOSPITAL DEATH, US DATABASE N=2,285 (0.47 TO 1.19)
• 32% vs 39% 90-DAY DEATH, EARLY REPERFUSION VS ECMO ALONE (CATAPULTE)
• HR 0.68 90-DAY DEATH, CATAPULTE (0.45 TO 1.03)

CRITICAL APPRAISAL Clear question; protocol registered

Partly PICO stated; PROSPERO registration was retrospective and the ID is not given

thrombectomy are tertiary decisions, so escalate early | Confidence in that: high

UK practice: NICE HTG705 (formerly IPG778, November 2023) rec 1.1: for high-risk PE when thrombolysis is unsuitable or has failed, thrombectomy only with special arrangements for governance, consent and audit. RCUK Special Circumstances Guidelines (2025): consider surgical embolectomy or catheter-directed treatment as an alternative to rescue fibrinolysis, and consider ECPR as rescue therapy in selected arrests.

EM Evidence Rundown, Issue 36. Page 5 of 27

INFORMING PRACTICE

SR/MA

FRCEM

Induction Agents for Intubating the Critically Ill (NMA, 21 RCTs, n=6,031): More Peri-Intubation Instability With Ketamine Than Etomidate (RR 1.31), No Difference in Mortality

McDougall G, Forestell B, Sadeghirad B, et al. Induction Agents for Tracheal Intubation of Critically Ill Patients: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials. Chest. 2026 (published online 18 August 2026). doi:10.1016/j.chest.2026.07.5239. PMID: 42612901. PROSPERO CRD42024600739. Source route: JournalFeed Article of the Day 27 Oct; PDF via BMA Library (supplied by Jake; journal proof). Appraised from full text

Network meta-analysis of 21 RCTs (6,031 critically ill patients intubated in the ED, ICU, operating room or prehospital; 6 ED trials), searched to 10 December 2025 and including the 2,365-patient RSI trial; no primary outcome was named. Most patients received etomidate (2,749) or ketamine (2,600), and no arm used propofol alone. Against etomidate, ketamine increased haemodynamic instability during intubation (mostly systolic BP under 90 mmHg; RR 1.31, 95% CI 1.15 to 1.48), as did ketamine-opioid and midazolam-opioid combinations, without a mortality difference. Etomidate increased adrenal suppression; a 1:1 ketamine-propofol mix looked best for instability but rests on 2 trials and 129 patients.

Source: McDougall et al, Results and Figure 1 (network estimates).

• Not reported ABSOLUTE EVENT RATES; NNT NOT CALCULABLE
• RR 1.31 INSTABILITY, KETAMINE VS ETOMIDATE (95% CI 1.15 TO 1.48)
• RR 0.96 MORTALITY, KETAMINE VS ETOMIDATE (95% CI 0.88 TO 1.05)
• RR 1.87 ADRENAL SUPPRESSION, ETOMIDATE VS KETAMINE (95% CI 1.07 TO 3.25)

CRITICAL APPRAISAL Clear question; protocol registered

Yes PROSPERO CRD42024600739; RCTs comparing 2 or more sedative agents

Design: SR/MA | Certainty: moderate (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet: in shocked patients expect more hypotension with ketamine than etomidate, have a vasopressor ready and keep opioid co-induction small (authors’ suggestion) | Confidence in that: moderate

UK practice: Etomidate is licensed in the UK for induction (0.3 mg/kg IV); the Hypnomidate SmPC (eMC, revised February 2024) warns that single doses can cause transient adrenal insufficiency and advises caution in critically ill patients, including sepsis. The review shows no survival difference, so agent choice should still follow local RSI policy and the patient’s physiology.

EM Evidence Rundown, Issue 36. Page 6 of 27

INFORMING PRACTICE SR/MA

Steroid Regimens in Septic Shock (NMA, 21 RCTs, n=7,908): Against Placebo, Hydrocortisone 200–299 mg/day Alone Did Not Reduce 28-Day Mortality (OR 0.92) and With Fludrocortisone 50 µg It Did (OR 0.78), but the Two Regimens Did Not Differ Significantly

Kwon S, Kwak SG, Hong HL. Comparative efficacy and safety of hydrocortisone-based regimens in septic shock: A network meta-analysis. J Crit Care. 2027;97:155750 (online 22 September 2026). doi:10.1016/j.jcrc.2026.155750. PMID: 42784878. PROSPERO CRD420251015672. Source route: JournalFeed EM Speed Read 26 Oct; PDF via BMA Library (supplied by Jake). Appraised from full text

Network meta-analysis of 21 RCTs (7,908 adults with septic shock, or severe sepsis in older trials) comparing placebo, hydrocortisone 200–299 mg/day (H200), 300 mg/day or more (H300), and H200 plus fludrocortisone 50, 100 or 200 µg/day. For the primary outcome, 28-day mortality (14 trials, 7,303 participants), H200 alone did not differ from placebo, whereas H200 plus fludrocortisone 50 µg and H300 were associated with lower mortality (table). H200 alone against H200 plus fludrocortisone 50 µg was not significant (OR 1.17, 95% CI 0.92 to 1.48), and the H300 signal comes from five small single-centre trials from 1998 to 2006, which the authors flag for a small-study effect.

Source: Kwon et al, section 3.3 and Fig. 3 (network estimates). Absolute event rates were not reported.

• Not reported ABSOLUTE EVENT RATES; NNT NOT CALCULABLE
• OR 0.92 28-DAY MORTALITY, H200 VS PLACEBO (95% CI 0.81 TO 1.04)
• OR 0.78 28-DAY MORTALITY, H200 + FLUDROCORTISONE 50 ΜG VS PLACEBO (95% CI 0.64 TO 0.96)
• OR 0.53 H300 VS PLACEBO (95% CI 0.29 TO 0.96); FIVE SMALL TRIALS, 1998–2006

CRITICAL APPRAISAL Clear question; protocol registered

Yes PROSPERO CRD420251015672; RCTs of hydrocortisone-based regimens only

Design: SR/MA | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing for the ED: do not escalate hydrocortisone above 200 mg/day on this evidence; adding fludrocortisone is for your critical care team | Confidence in that: moderate

UK practice: NICE NG253 (2025, last updated 23 September 2026) makes no recommendation on corticosteroids (we searched the full guideline) and advises discussing vasopressors with critical care (rec 1.8.11). Hydrocortisone injection is licensed for shock unresponsive to conventional therapy when adrenocortical insufficiency may be present (Solu-Cortef SmPC (eMC, revised May 2026)). UK fludrocortisone is oral only (tablets or solution) and licensed for Addison’s disease and salt-losing adrenogenital syndrome (fludrocortisone 0.05 mg tablets SmPC (eMC, revised December 2023)), so septic shock use is off-label.

EM Evidence Rundown, Issue 36. Page 7 of 27

INFORMING PRACTICE

SR/MA

PAEDS

FRCEM

Two-Bag vs One-Bag Fluids in DKA (SR/MA, 21 Studies, Adults and Children): Hypoglycaemia 9.2% vs 23.9% (RR 0.50), From Almost Entirely Non-Randomised Data

Srikrishnaraj A, Souter AR, Woods N, et al. Two-bag Versus One-bag Method for Adult and Pediatric Diabetic Ketoacidosis Management. Ann Emerg Med. 2026;87(3):346–364. doi:10.1016/j.annemergmed.2025.07.032. PMID: 40913602. PROSPERO CRD42024541648. Source route: PDF via BMA Library (supplied by Jake); open access (CC BY). Appraised from full text

Systematic review of hospital DKA treatment comparing two bags (same electrolytes, different dextrose content, run together and titrated to glucose) with one bag (swapped for a dextrose fluid as glucose falls). It included 21 studies (4 RCTs, 16 retrospective cohorts, 1 quality-improvement study): 9 adult (3,329 visits) and 12 paediatric (1,385 visits). For the primary outcome of hypoglycaemia (glucose under 3.89 mmol/L [70 mg/dL]), pooled from 14 studies (all but one non-randomised), rates were 9.2% (215/2,345) with two bags vs 23.9% (473/1,975) with one bag (RR 0.50, 95% CI 0.41 to 0.59; I² 51.8%). In adults, hypokalaemia was less frequent (RR 0.84, 95% CI 0.76 to 0.93).

• 23.9% HYPOGLYCAEMIA, ONE BAG (POOLED)
• 9.2% HYPOGLYCAEMIA, TWO BAGS
• RR 0.50 HYPOGLYCAEMIA (95% CI 0.41 TO 0.59)
• −1.76 h TIME TO RESOLUTION (−2.80 TO −0.71)

Source: Srikrishnaraj et al, Results and GRADE tables (which print 0.42 to 0.61 and 0.38 to 0.84 for the first and third rows). One paediatric RCT (n=30) is in the hypoglycaemia analysis. The adult and child counts (3,150 and 970) are as printed and sum to 4,120, not 4,320. No NNT: the data are almost all non-randomised.

CRITICAL APPRAISAL Clear question; protocol registered

Yes PROSPERO CRD42024541648; outcomes defined a priori

Design: SR/MA | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; audit how quickly 10% glucose starts once glucose is below 14 mmol/L | Confidence in that: moderate Disagreement: The authors back the two-bag method for adults and children; neither JBDS-IP nor BSPED titrates two bags from the start of treatment.

UK practice: The adult JBDS-IP 02, The Management of Diabetic Ketoacidosis in Adults (revised March 2023) already runs two fluids together: below 14 mmol/L, start 10% glucose at 125 ml/h alongside the 0.9% sodium chloride and consider reducing insulin from 0.1 to 0.05 units/kg/h. It cites a UK survey of 283 people in which glucose fell below 4.0 mmol/L in 27.6%. The BSPED DKA guideline (2021, v3) instead swaps to one premixed bag (0.9% sodium chloride, 5% glucose, 20 mmol potassium chloride in 500 ml) at 14 mmol/L.

EM Evidence Rundown, Issue 36. Page 8 of 27

INFORMING PRACTICE OBSERVATIONAL

Hyperkalaemia and 48-Hour Outcomes (US Propensity-Matched Cohort, 140,288 Patients): Arrest, VT or VF Rose From 3.05% at K 5.5–5.99 to 6.68% at K 7.0 or More, vs Under 1% in Matched Controls

O’Brien J, Patel MJ, Schrock JW. Serum hyperkalemia and short-term clinical outcomes in emergency department patients: A multicenter propensity-matched study. Am J Emerg Med. 2026;110:423–429. doi:10.1016/j.ajem.2026.09.028. PMID: 42777312. Source route: PDF via BMA Library (supplied by Jake). Appraised from full text

TriNetX cohort of adults in US EDs (2010 to 2024; dialysis and haemolysed samples excluded). Each band of first potassium was propensity-matched 1:1 with its own controls (3.5–5.0 mmol/L) on demographics, comorbidities, creatinine, bicarbonate, calcium and vital signs. The primary outcome, coded cardiac arrest, VT or VF within 48 hours, was 3.05% vs 0.77% at 5.5–5.99 mmol/L (RR 3.97, 95% CI 3.65 to 4.32) and 6.68% vs 0.83% at 7.0 mmol/L or more (RR 8.08, 95% CI 6.70 to 9.75). Coded VT and VF alone stayed under 1% in every band and did not rise steadily with potassium.

• 3.05% ARREST/VT/VF 48 H, K 5.5–5.99 (CONTROLS 0.77%)
• 6.68% ARREST/VT/VF 48 H, K 7.0 OR MORE (CONTROLS 0.83%)
• 5.79% DEATH WITHIN 48 H, K 7.0 OR MORE (CONTROLS 0.21%)
• RR 8.08 ARREST/VT/VF, K 7.0 OR MORE (6.70 TO 9.75)

First K (mmol/L) Matched per arm Arrest/VT/VF 48 h, K vs control RR (95% CI) Death 48 h, K vs control

Source: O’Brien et al, Table 1 and section 3.2. CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly First potassium only; outcomes from billing codes, including elective cardioversion

Design: Observational | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing to change in thresholds; it supports UKKA’s monitoring of potassium 6.5 mmol/L or more | Confidence in that: moderate

UK practice: The UK Kidney Association guideline, Treatment of Acute Hyperkalaemia in Adults (2020, updated October 2023 and July 2026) grades hyperkalaemia as mild (5.5–5.9 mmol/L), moderate (6.0–6.4) or severe (6.5 or more). It recommends an urgent 12-lead ECG at 6.0 or more, and continuous 3-lead monitoring at 6.5 or more, with hyperkalaemic ECG changes at any level, or at 6.0–6.4 if clinically unwell or a rapid rise is expected (14.1–14.2). After insulin–glucose it advises 10% glucose at 50 ml/h for 5 hours when pre-treatment glucose is under 7.0 mmol/L. The July 2026 update downgraded IV calcium for ECG changes to “suggest” (2C) after the 2025 ILCOR review.

EM Evidence Rundown, Issue 36. Page 9 of 27

INFORMING PRACTICE OBSERVATIONAL

Blunt Traumatic Pneumothorax of 15–50% (9 Dutch Hospitals): Observation Failed in 29.2% (14 of 48), With No Difference in Complications After Weighting

Hartman A, Vellekoop L, Van Delft CH, et al; CONSTRAIN study group. A multicenter retrospective evaluation of treatment variation in patients with a blunt traumatic pneumothorax; noninvasive treatment versus tube thoracostomy (CONSTRAIN-Retro). J Trauma Acute Care Surg. 2026;100:863–870. doi:10.1097/TA.0000000000005113. PMID: 42468511. Source route: PDF supplied by Jake (24 Oct 2026; open access, CC BY 4.0). Appraised from full text

Retrospective cohort of 275 clinically stable adults admitted with a unilateral pneumothorax after blunt trauma to nine Dutch hospitals (2021–2022), with size measured on the ED AP chest X-ray by the Collins method; occult (CT-only) pneumothoraces were excluded. The primary outcome was the index treatment: 126 (45.8%) were observed and 149 (54.2%) had a chest drain, with 77.9% of pneumothoraces up to 15% observed, every one over 50% drained, and the widest variation at 15–50% (48 observed, 83 drained). In that medium group, 14 of 48 observed patients (29.2%) needed a secondary drain, 12 of them for progression. After inverse probability weighting (IPTW), complications did not differ (32.8% vs 30.2%; OR 1.14, 95% CI 0.40 to 3.28), ICU admission was lower with observation (2.4% vs 17.5%; OR 0.13, 95% CI 0.03 to 0.64) and median hospital stay was 4 vs 5 days (p=0.313).

• 29.2% SECONDARY DRAIN AFTER OBSERVATION (14/48)
• 32.8% vs 30.2% COMPLICATIONS, IPTW (OR 1.14, 0.40 TO 3.28)
• 2.4% vs 17.5% ICU ADMISSION, IPTW (OR 0.13, 0.03 TO 0.64)
• 4 vs 5 days MEDIAN HOSPITAL STAY, IPTW (P=0.313)

Source: Table 3 of the paper. The outcome table has 82 drained patients where Table 2 has 83; the death denominators (47 and 81) are as printed.

CRITICAL APPRAISAL Exposure and outcome measured reliably

Partly Size from AP films by a single researcher with no reliability testing; AP films miss anterior air (authors); missing data not imputed, so complications may be undercounted

Design: Observational | Certainty: very low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; observing a stable 15–50% blunt pneumothorax is a reasonable trauma-team option, but plan for about 3 in 10 needing a drain later | Confidence in that: moderate

UK practice: NICE NG39 (major trauma: assessment and initial management, 2016) covers decompression of tension pneumothorax (1.4.1–1.4.2) and chest imaging (1.4.3–1.4.5) but makes no recommendation on observing versus draining a stable non-tension traumatic pneumothorax, and gives no size threshold. We found no NICE guidance on the drain decision in this group, so it follows local trauma protocols.

EM Evidence Rundown, Issue 36. Page 10 of 27

INFORMING PRACTICE RCT

PENG Block vs IV Morphine for Hip Fracture in Older Adults (Thai ED RCT, n=34): Pain Fell 6 vs 3 Points at 30 Minutes, and No Block Patient Desaturated

Sri-on J, Fusakul Y, Phisaiphun K, et al. Ultrasound-guided pericapsular nerve group block versus intravenous morphine for pain management in older adults with hip fractures: a randomised controlled trial in the emergency department. Emerg Med J. 2026;43(9):530–536. doi:10.1136/emermed-2025-215388. PMID: 41856550. TCTR20230905007. Source route: EMJ full text (free article) read 29 Oct 2026; the results tables were not captured. Appraised from full text

Open-label, single-centre RCT in an urban teaching hospital ED in Bangkok: 34 adults aged 65–99 with a hip fracture and pain of 5 or more on an 11-point verbal scale (VNRS), enrolled on weekdays 08:00–16:00, received an ultrasound-guided pericapsular nerve group (PENG) block (bupivacaine 0.5% 1 mg/kg diluted to 20 mL, all by the principal investigator) or a single dose of IV morphine 0.1 mg/kg over 10 minutes. The primary outcome, fall in VNRS at 30 minutes, was larger with PENG: median −6 (IQR −6 to −5) vs −3 (IQR −5 to −2), p=0.001, and the adjusted repeated-measures estimate over 2 hours was β −1.55 (95% CI −2.63 to −0.47). Desaturation below 94% occurred in 0 of 17 vs 3 of 17 (17.6%), and 1 morphine patient (5.9%) needed rescue fentanyl. Quadriceps power was unchanged at 2 hours in both groups and the authors report no difference in 48-hour delirium.

• −6 vs −3 MEDIAN VNRS FALL AT 30 MIN (P=0.001)
• β −1.55 ADJUSTED PAIN OVER TIME (95% CI −2.63 TO −0.47)
• 0% vs 17.6% DESATURATION BELOW 94% (0/17 VS 3/17)
• ARD 17.6 points DESATURATION, CALCULATED (95% CI −4.1 TO 41.0); NOT SIGNIFICANT, SO NO NNT

CRITICAL APPRAISAL Focused question; registered primary outcome

Yes TCTR20230905007; VNRS change at 30 min; university research fund, no competing interests declared

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet; keep fascia iliaca block as the ED standard; PENG is reasonable where trained operators already use it | Confidence in that: moderate

UK practice: NICE CG124 (hip fracture, 2011, last updated January 2023) recommends considering nerve blocks if paracetamol and opioids do not give enough preoperative pain relief, or to limit opioid dose, given by trained personnel and not as a substitute for early surgery (1.3.6). The RCEM Best Practice Guideline: Fascia Iliaca Block in the Emergency Department (first published 2020, updated October 2025) says fascia iliaca block should be available and performed in EDs as routine pain management for hip fracture, by clinicians with a competency assessment and with at least 1 hour of monitoring. It calls PENG a “promising alternative” and lists PENG vs fascia iliaca block as a research recommendation.

EM Evidence Rundown, Issue 36. Page 11 of 27

ALSO IN THE PE DEBATE

STRATIFY Safety Signal Revisited: 6 of 141 Deaths After Low-Dose Lysis vs 0 of 69 With Heparin at 3 Months (P=0.08); a Trial Too Small to Settle It

Ghotraa J, Rivera-Lebron B, Rali P, et al. Reassessing safety signals in low-dose thrombolysis for intermediate-high-risk pulmonary embolism. Cardiovasc Res. 2026 (online 16 September). doi:10.1093/cvr/cvag203. PMID: 42745413. Reply: Kjaergaard J, Bang LE, Ersbøll MK, Carlsen J. Cardiovasc Res. 2026 (online 17 October). doi:10.1093/cvr/cvag208. PMID: 42888829. Sonne-Holm E, Bang LE, Carlsen J, et al. Long-term outcomes after heparin versus low-dose thrombolysis in patients with intermediate-high-risk pulmonary embolism - a secondary analysis of the low dose thrombolysis, ultrasound assisted thrombolysis or heparin for intermediate high risk pulmonary embolism randomized controlled trial. J Thromb Haemost. 2026 (online 19 September 2026). doi:10.1016/j.jtha.2026.09.023. PMID: 42762902. Background: Kjaergaard J, et al. STRATIFY. Cardiovasc Res. 2026;122(4):539–549. doi:10.1093/cvr/cvag038. PMID: 41610160. NCT04088292. Source route: none of the three has an abstract and the BMA Library could not supply them; we used the PubMed records, the opening extract of the Ghotraa letter on the Cardiovascular Research website, the STRATIFY abstract and EMCrit 424 (2 May 2026).

STRATIFY, a Danish investigator-initiated open-label trial, randomised 210 adults with intermediate-high-risk PE to alteplase 20 mg over 6 hours by ultrasound-assisted catheter (n=71), the same dose IV (n=70), or heparin alone (n=69). Its primary outcome, CT thrombus burden, fell by 3.6 points more with lysis than with heparin (95% CI 2.2 to 5.0), with no difference between catheter and IV delivery; the trial’s own abstract concludes that “the rate of death and risk of bleeding complications were increased” with low-dose lysis. Ghotraa et al argue that STRATIFY “was not designed to evaluate rare adverse events”: its single intracranial haemorrhage in the catheter arm (1 of 71) has a 95% CI of 0.04% to 7.6%, and 3-month deaths were 6 vs 0 (P=0.08). The investigators’ reply and their long-term follow-up in JTH have no abstract and we could not read them, so their content is not reported here. For a UK ED: Nothing new; low-dose lysis in stable PE stays off-protocol (NG158 1.6.3).

EM Evidence Rundown, Issue 36. Page 12 of 27

2 — GUIDELINES & UK UPDATES

CHANGE THIS MONTH

UK

GUIDELINE

FRCEM

RCEM/NPIS Guideline on Suspected but Unidentified Poisoning: 8 Consensus Recommendations, All Level C (Consensus) — Toxidrome First, TOXBASE Always

Welby-Everard P, Pucci M, Bradberry S, et al. Management of patients with suspected but unidentified poisoning in the emergency department: a joint Royal College of Emergency Medicine and National Poisons Information Service best practice guideline. Emerg Med J. 2026;43(7):427–434. doi:10.1136/emermed-2025-215194. PMID: 42103461. First published by RCEM in April 2025: rcem.ac.uk (PDF). Source route: EMJ full text via Jake’s login (29 October) and the RCEM April 2025 PDF. Appraised from full text

RCEM and NPIS set out a toxidrome-based approach for ED patients thought to be poisoned with an unknown agent; it does not replace TOXBASE or NPIS advice. The 8 summary recommendations: consider poisoning whenever an undifferentiated patient has diagnostic uncertainty, rule out trauma, never delay stabilisation to identify the agent, use the NPIS “Poisoning with an unknown substance” tables, repeat ECGs and gases, assume street and internet drugs may contain other substances, and discharge only when physiologically and biochemically stable, with capacity and no further deterioration expected. The text adds checkable practice points: a paracetamol level in all possible self-harm and to be considered in all unknown poisoning, no point-of-care urine drug screens to guide management (low sensitivity and specificity), no gag-reflex testing, and the GCS never used for prognosis in poisoning.

Source: RCEM/NPIS guideline, April 2025 version (Table 4, Table 5 and text). Children’s doses are in the guideline. CRITICAL APPRAISAL Developer and funding; conflicts managed

Partly Welby-Everard, Pucci, Elamin and colleagues for RCEM and NPIS (Best Practice Sub-Committee and Toxicology SIG acknowledged); declarations of interest “None”; funding not reported

Design: Guideline | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: This month, check the ED poisoning proforma against it: paracetamol level in all possible self-harm and considered in any unknown poisoning, manual QT on the nomogram, no point-of-care urine drug screens, no phenytoin for toxic seizures | Confidence in that: moderate

UK practice: This is the UK guideline for this question; use it with TOXBASE and the RCEM/NPIS Guideline on Antidote Availability for Emergency Departments (Version 7, September 2023). That list sets Category A antidotes as immediately available in the ED, Category B held in the hospital and available within 1 hour, and Category C held supra-regionally and released through NPIS. The guideline asks for a mental health assessment after self-harm in line with RCEM guidance.

EM Evidence Rundown, Issue 36. Page 13 of 27

SAFETY UPDATE (FIRST COVERED IN ISSUE 34) AND RCEM LEARNING

MHRA StatStrip Ketone Alert Updated: 12 Lots Now Listed for Recall; Replacement UK Stock Was Expected About 4 Weeks After 9 October

Medicines and Healthcare products Regulatory Agency. Nova Biomedical StatStrip ketone test strips: risk of disruption to blood ketone testing due to quality-control failures and supply interruption (DSI/2026/013). Published 9 October 2026, updated 16 October 2026. gov.uk. Repeated in MHRA Safety Roundup October 2026, published 27 October 2026: gov.uk. Source route: MHRA Safety Roundup. Update to the Issue 34 item.

Issue 34 ran this alert as Change Tonight. On 16 October MHRA updated the alert and the manufacturer’s field safety notice to add a further batch number, and the 27 October Safety Roundup repeats the advice. The alert now lists 12 affected lots of part number 46951 (50-strip dual pack): 2224295249, 2225013249, 2225072249, 2225079249, 2225111249, 2225118249, 2225167249, 2226068249, 2226070249, 2226110249, 2226173249 and 2226117249. Nova said replacement UK stock was not expected for about 4 weeks from the 9 October alert; the alert still says the Department of Health and Social Care will issue clinical guidance, and we found none published by 29 October. For a UK ED: Tonight, re-check StatStrip stock in adult and paediatric ED against the 12-lot list and confirm the backup ketone method still works.

New on RCEM Learning since Issue 35

A Deceptive Case of Adolescent Limb Shortening (29 Oct); Journal Club: Are pulse oximeters accurate for every skin tone? (28 Oct); Delirious after Dialysis (28 Oct); Peripheral numbness – a quick win? (26 Oct); Red Breasted (23 Oct); Chest pain following a distressing event (22 Oct). Source: RCEM Learning modules feed, checked 29 October 2026.

EM Evidence Rundown, Issue 36. Page 14 of 27

3 — PAEDIATRIC EMERGENCY MEDICINE

INFORMING PRACTICE

DIAGNOSTIC

PAEDS

Paediatric C-Spine CT (72 US Trauma Centres, n=12,693): Sensitivity 94.7% for Injuries Needing Surgery or a Halo, 98.7% After Expert Re-Review; Two Injuries Invisible on CT

Russell KW, Kahan AM, Eldredge RS, et al; Western Pediatric Cervical Spine Study Group. Defining pediatric cervical spine clearance with computed tomography: a prospective observational study by the Western Pediatric Surgery Research Consortium and Western Trauma Association. J Trauma Acute Care Surg. 2026;101(5):844–857. doi:10.1097/TA.0000000000005120. PMID: 42611682. NCT05608460. Source route: PMC full text (PMC13690346); JournalFeed Article of the Day 28 Oct 2026. Appraised from full text

Prospective study at 72 US trauma centres (2022 to 2025) of children under 18 imaged after blunt trauma; 12,693 of 19,651 had a cervical spine CT (median age 12; 19% with GCS 13 or less). The reference standard was a clinically significant injury, defined as one treated with surgery or a halo, which occurred in 150 (1.2%). CT as read at the treating hospital had a sensitivity of 94.7% (95% CI 89.4% to 97.5%) and an NPV of 99.9% (99.9% to 100%); six of the eight misses were visible on panel re-review, giving a sensitivity of 98.7% (94.8% to 99.8%). Two injuries were invisible on CT even on re-review; the one in an obtunded child (atlanto-occipital dislocation needing a halo) would have needed 1,942 MRIs in obtunded children with a normal CT to find (yield 0.05%).

• 94.7% SN, LOCAL READ (89.4 TO 97.5)
• 98.7% SN AFTER RE-REVIEW (94.8 TO 99.8)
• 88.5% SP, LOCAL READ (87.9 TO 89.2)
• LR− 0.06 LOCAL READ (CALCULATED)

Calculated from the paper’s counts (142 true positives, 8 false negatives, 1,588 abnormal CTs, 12,543 without injury), counting every abnormal CT as positive.

CRITICAL APPRAISAL Representative spectrum; consecutive enrolment

Partly 72 centres, but only the 65% given CT by local practice were analysed

Design: Diagnostic | Certainty: moderate (GRADE-style, our judgement) | What it changes for a UK ED: In a child who meets NG232 criteria for CT, a normal multiplanar CT in an alert child with a reliable examination and no neurological signs supports removing the collar without routine MRI; if obtunded, discuss with the spinal or paediatric trauma team | Confidence in that: moderate Disagreement: The JournalFeed reviewer (28 Oct 2026) would be slow to remove the collar, leaning towards a neurosurgical opinion, after a normal CT with GCS 8 or less, focal deficits or persistent midline tenderness; the JournalFeed editor would clear persistent midline tenderness if the neurological examination is reliable.

UK practice: NICE NG232 (head injury, 2023) 1.6.4 limits CT cervical spine in under-16s with head injury to set risk factors (for example GCS 12 or less, intubation, focal neurological signs, or multisystem scanning with clinical suspicion); for those with neck pain or tenderness and a risk factor such as a dangerous mechanism, 3-view X-rays (1.6.5–1.6.6). NG232 asks for MRI in addition to CT if there are neurological signs (1.6.9) and says to consider MRI for ligamentous and disc injuries suggested by CT or clinical findings (1.6.11). NG232 covers head-injured children; for spinal injury without head injury, NICE NG41 (spinal injury, 2016) asks for MRI in under-16s with a strong suspicion of cervical cord or column injury (1.5.2) and otherwise considers plain X-rays (1.5.3).

EM Evidence Rundown, Issue 36. Page 15 of 27

INFORMING PRACTICE

RCT

PAEDS

IV Ketamine 0.2 mg/kg for Suicidal Adolescents in the ED (RCT, n=29): No Significant Difference From Placebo in Depression or Suicidality Scores; Dissociation in 60% vs 7%

Vayngortin T, Patel E, Seneviratne S, et al. Intravenous Ketamine for Depressed and Suicidal Adolescents in the Emergency Department: A Randomized Double-Blind Trial. West J Emerg Med. 2026;27(5):1374–1383. doi:10.5811/westjem.53968. PMID: 42900929. NCT05217706. Source route: PMC full text (PMC13660616). Appraised from full text

Single-centre, double-blind, placebo-controlled RCT in a US children’s hospital ED. Adolescents aged 12 to 17 with treatment-resistant depression (at least two antidepressants for 4 weeks or more each) and suicidal ideation needing psychiatric admission received a single dose of ketamine 0.2 mg/kg IV (maximum 35 mg) or saline placebo over 2 minutes; 29 were analysed (15 vs 14). The primary outcome was a reduction of 50% or more on both the Beck Youth Inventory and the Suicidal Ideation Questionnaire (SIQ) within 3 hours: no score differed significantly between groups at any time (SIQ fall of 50% or more at 1 hour, 2/14 vs 0/14, P=0.48), although all four adolescents with a 50% fall on any scale within 3 hours had received ketamine. Dissociation was reported by 9/15 (60.0%) vs 1/14 (7.1%), with no effects beyond the 1-hour observation period.

• 4/15 vs 0/14 ANY SCALE HALVED WITHIN 3 H (POST HOC, NOT TESTED)
• P=0.48 SIQ 50% FALL AT 1 H, 14.3% VS 0%
• 60.0% vs 7.1% DISSOCIATION, KETAMINE VS PLACEBO
• NNH 2 DISSOCIATION (CALCULATED, CI NOT REPORTED)

Source: Tables 2 and 4. Denominators vary because children who were asleep or could not be reached were omitted at that time point. CRITICAL APPRAISAL Focused question; registered primary outcome

Yes NCT05217706; composite of 50% falls on BYI and SIQ within 3 hours

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing; ketamine for suicidality in young people remains a research question, and ED care stays assessment, safety and referral | Confidence in that: high

UK practice: NICE NG134 (depression in children and young people, 2019) recommends fluoxetine, with psychological therapy, as the antidepressant for moderate to severe depression in young people (1.6.10 and 1.6.16) and does not mention ketamine; neither does NICE NG225 (self-harm, 2022) (we searched both). For a young person who attends after self-harm, NG225 1.7.13 asks the ED to offer referral to a children and young people’s crisis response service, or equivalent, for a psychosocial assessment as soon as possible after arrival.

EM Evidence Rundown, Issue 36. Page 16 of 27

INFORMING PRACTICE

SR/MA

PAEDS

Air vs 100% Oxygen to Start Resuscitation at Birth in Term and Late Preterm Infants (Cochrane, 12 Trials, n=2,054): Lower Mortality Before Discharge With Air, RR 0.72, Low Certainty

Rüegger CM, Dawson JA, Cracknell J, et al. Air versus supplemental oxygen for resuscitation of term or late preterm infants at birth. Cochrane Database Syst Rev. 2026;10(10):CD014781. doi:10.1002/14651858.CD014781.pub2. PMID: 42937251. Source route: PMC (PMC13712104), which holds the structured abstract and plain language summary only; the full review could not be retrieved. Abstract only

Cochrane review of randomised, quasi-randomised and cluster trials in infants of 35 weeks’ gestation or more who needed respiratory support at birth, searched to 12 January 2026: 12 studies, 2,054 infants. Every trial compared air (21%) with 100% oxygen, none titrated to saturation targets. Air may reduce all-cause mortality before discharge (RR 0.72, 95% CI 0.57 to 0.91; I² 0%; 10 studies, 1,879 infants; low certainty), with little or no difference in HIE stage II to III (RR 0.88, 95% CI 0.71 to 1.10; low certainty). The effect on moderate to severe neurodevelopmental impairment is very uncertain (RR 1.24, 95% CI 0.74 to 2.10; 3 studies, 414 infants), and no study reported it after 12 months by the protocol definition.

• RR 0.72 DEATH BEFORE DISCHARGE (0.57 TO 0.91)
• RR 0.88 HIE STAGE II–III (0.71 TO 1.10)
• RR 1.24 NDI, VERY LOW CERTAINTY (0.74 TO 2.10)
• Not reported ABSOLUTE RISKS (NOT IN THE ABSTRACT)

CRITICAL APPRAISAL Clear question; protocol registered

Yes Published Cochrane protocol (CD014781); outcomes pre-specified

Design: SR/MA | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing new: start in air for infants of 32 weeks or more, titrate to the RCUK saturation targets, and go to 100% oxygen only when compressions start | Confidence in that: moderate

UK practice: Resuscitation Council UK Newborn resuscitation and support of transition of infants at birth (2025) (published 27 October 2025): infants of 32 weeks or more who need respiratory support start in 21% oxygen and those under 32 weeks in 30% or more; titrate to acceptable right-hand saturations (70–75% at 3 minutes, 80–85% at 5 minutes, 85–95% at 10 minutes) and increase to 100% oxygen when chest compressions start. With a self-inflating bag outside hospital, RCUK says to resuscitate in air unless compressions are needed.

EM Evidence Rundown, Issue 36. Page 17 of 27

4 — FOAMED & CRITICAL APPRAISAL

INFORMING PRACTICE RCT

HOPE (RCT, n=372): Alteplase 4.5 to 24 Hours After Stroke With a CT Perfusion Mismatch Gave mRS 0–1 in 40% vs 26%, With More Symptomatic Bleeding (3.8% vs 0.5%)

Zhou Y, He Y, Campbell BCV, et al. Alteplase for acute ischemic stroke at 4.5 to 24 hours: the HOPE randomized clinical trial. JAMA. 2025;334(9):788–797. doi:10.1001/jama.2025.12063. PMID: 40773205. NCT04879615. Discussed in SGEM#524 (guest skeptic R Stanton, 24 October 2026): thesgem.com. Source route: SGEM#524; PMC full text (PMC12332759). Appraised from full text

Open-label RCT with blinded end-point assessment at 26 stroke centres in China. Adults 4.5 to 24 hours from onset (or from the midpoint of last known well and symptom recognition), NIHSS 4–26, with salvageable tissue on CT perfusion (core 70 ml or less, mismatch ratio at least 1.2, mismatch volume at least 10 ml) and no initial plan for thrombectomy received alteplase 0.9 mg/kg (n=186) or standard medical treatment including antiplatelets (n=186). The primary outcome, mRS 0–1 at 90 days, occurred in 40.3% (75/186) vs 26.3% (49/186): adjusted RR 1.52 (95% CI 1.14 to 2.02), unadjusted risk difference 13.98 percentage points (95% CI 4.50 to 23.45). Symptomatic intracranial haemorrhage within 36 hours occurred in 3.8% (7/185) vs 0.5% (1/182), and 90-day mortality was 10.8% in both groups; only 51 patients (13.7%) were treated beyond 9 hours from known onset, while 51.6% had wake-up or unwitnessed stroke.

• 40.3% vs 26.3% MRS 0–1 AT 90 DAYS
• NNT 8 CALCULATED; ARD 14.0 PERCENTAGE POINTS (95% CI 4.5 TO 23.5)
• NNH 31 SICH, CALCULATED; ARD 3.2% (95% CI 0.3 TO 6.2)
• 10.8% vs 10.8% 90-DAY MORTALITY; ARR 0.91 (95% CI 0.52 TO 1.62)

CRITICAL APPRAISAL Focused question; registered primary outcome

Yes NCT04879615; mRS 0–1 at 90 days; secondary outcomes not adjusted for multiplicity

Design: RCT | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Nothing yet for the ED; patients beyond 4.5 hours with a perfusion mismatch remain a stroke-team decision | Confidence in that: moderate Disagreement: The HOPE authors say the results support extending the window in selected patients when thrombectomy is not planned. SGEM#524 (Stanton) says HOPE is not enough to treat routinely out to 24 hours and that extended-window thrombolysis remains a shared decision, citing a Bayesian meta-analysis (Garg and Brophy, Am J Emerg Med 2026) that found a 1% probability of a clinically important benefit on utility-weighted mRS and a 72% probability of at least one extra death per 100 treated.

UK practice: NICE NG128 (2019, amended March 2025) covers thrombolysis within 4.5 hours of onset through TA990 (tenecteplase) and TA264 (alteplase), and allows trained ED staff to give it within an acute stroke service (1.4.1, 1.4.3). The alteplase SmPC (eMC, updated May 2026) licenses it within 4.5 hours of last known well. The National Clinical Guideline for Stroke for the UK and Ireland (2023) already says consider alteplase 4.5 to 9 hours from known onset, or within 9 hours of the midpoint of sleep, with a CT/MR perfusion mismatch (ratio over 1.2, mismatch over 10 ml, core under 70 ml); it makes no recommendation on thrombolysis up to 24 hours.

EM Evidence Rundown, Issue 36. Page 18 of 27

INFORMING PRACTICE FOAMED

PulmCCM (Kenny): Inspiratory IVC Collapse Means CVP Has Fallen Below Abdominal Pressure, Not That the Patient Needs Fluid; a 2026 Crossover Study Found Collapse in 86% of Assisted Breaths

Kenny JS. Inspiratory collapse of the inferior vena cava: what is it telling us? ICU Physiology in 1000 Words, first published 17 July 2014; republished by PulmCCM (Substack, paid post). Supporting: Visanumahimachai P, Wongthanee A, Trongtrakul K. Inferior vena cava dynamics during triggered breaths under different levels of pressure support and PEEP: a randomized crossover physiologic study. J Crit Care. 2026;94:155593. doi:10.1016/j.jcrc.2026.155593. PMID: 42025595. Wenwen Y, Ping X, Yue D, et al. Intensive Crit Care Nurs. 2025;89:104015. doi:10.1016/j.iccn.2025.104015. PMID: 40187099 (abstract only). Source route: PDF uploaded by Jake (24 October); J Crit Care PDF via BMA Library. Appraised from full text

Kenny argues that the IVC collapses on inspiration when the pressure inside it, which tracks CVP, falls below intra-abdominal pressure, so the sign depends on the pleural pressure swing, venous return and cardiac function, not on volume alone. It therefore tells you neither volume status nor volume responsiveness, and PEEP, auto-PEEP, intra-abdominal hypertension, ascites and breathing pattern all confound it. With an unpressurised abdomen and a large inspiratory effort, collapse means a low CVP, which may reflect hypovolaemia, high resistance to venous return, venodilatation (sepsis, venodilators) or a hyperdynamic heart. A 2026 randomised crossover study of 60 ICU patients on pressure support fits this: collapse occurred in 550 of 642 observations (86%), fell as support rose, and respiratory effort (P0.1) mediated 38.5% (95% CI 19.2 to 57.7) of the effect of pressure support on the collapsibility index; fluid responsiveness was not measured.

• 86% INSPIRATORY COLLAPSE, 550/642 OBSERVATIONS
• 38.5% EFFECT OF SUPPORT MEDIATED BY EFFORT (95% CI 19.2 TO 57.7)
• 0.82 / 0.82 SN / SP OF IVC INDICES FOR FLUID RESPONSIVENESS IN SHOCK, SR/MA (I² 83% / 79%)

CRITICAL APPRAISAL Who wrote it and for whom

Yes Jon-Emile Kenny (Heart-Lung.org), 2014 essay in an ICU physiology series; not peer reviewed

Design: FOAMed | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Do not use IVC collapse alone to decide on a fluid bolus in a spontaneously breathing patient; read it with the breathing effort and the abdomen | Confidence in that: moderate

UK practice: NICE NG253 (2025, updated September 2026) asks for 250 ml crystalloid boluses up to 1,000 ml, with reassessment after each bolus (1.8.6 to 1.8.8). It does not mention IVC ultrasound or fluid-responsiveness testing (we searched six recommendation chapters on 29 October 2026).

EM Evidence Rundown, Issue 36. Page 19 of 27

INFORMING PRACTICE

REVIEW

FRCEM

Choosing Vasoactives in Cardiogenic Shock (Circulation Primer): Phenotype First, Then the Drug, Because Only 5 Mostly Indirect RCTs Compare Agents

Bohula EA, Morrow DA. Vasoactive Agent Selection Strategies in the Management of Cardiogenic Shock. Circulation. 2026;154(8):775–784. doi:10.1161/CIRCULATIONAHA.126.079463. PMID: 42636314. Source route: PDF supplied by Jake (24 Oct 2026); JournalFeed Article of the Day, 19 Oct 2026. Appraised from full text

A narrative clinical primer from two Boston cardiologists. It defines shock by hypoperfusion, not hypotension: lactate above 2 mmol/L, creatinine at least twice the upper limit of normal or urine output below 0.5 mL/kg/h, ALT above 3 times normal, cool or mottled peripheries, or altered mental state; at least 15% of patients with cardiogenic shock are normotensive. The authors’ four steps are: protect the MAP; define the phenotype (classic LV-predominant, vasodilatory, normotensive, RV-predominant or dynamic LV outflow obstruction) and treat reversible causes; set goals (MAP 65–75 mmHg, cardiac index above 2.2 L/min/m², PCWP 14–18 and RAP 5–10 mmHg); then titrate by repeated reassessment. Their own practice is to often consider adrenaline over noradrenaline in hypotensive classic shock, use vasopressin first and avoid phenylephrine in RV-predominant shock, use a vasodilator and/or inodilator in normotensive shock, and use fluid, a pure vasopressor and a beta-blocker for outflow obstruction.

• 52.5% vs 48.5% SOAP II 28-DAY DEATH, DOPAMINE VS NORADRENALINE, ALL SHOCK TYPES, N=1,679 (OR 1.17, 95% CI 0.97 TO 1.42)
• 24.1% vs 12.4% SOAP II ARRHYTHMIAS, DOPAMINE VS NORADRENALINE
• 37% vs 7% OPTIMACC REFRACTORY SHOCK, ADRENALINE VS NORADRENALINE (POST HOC END POINT PER THE REVIEW, N=57)
• 65% / 29% / 6% AHA REGISTRY USE: NORADRENALINE / ADRENALINE / DOPAMINE

CRITICAL APPRAISAL Who wrote it and for whom

Yes Two TIMI Study Group cardiologists writing from cardiac ICU practice; their group receives institutional grants from many companies, including device makers Abbott and Abiomed

Design: Review | Certainty: low (GRADE-style, our judgement) | What it changes for a UK ED: Diagnose shock by perfusion, not blood pressure; phenotype with examination and POCUS before choosing a drug; noradrenaline first remains defensible; involve cardiology and critical care early | Confidence in that: moderate Disagreement: JournalFeed (19 Oct 2026) concludes that norepinephrine (noradrenaline) “is a reasonable place to start” for hypotensive cardiogenic shock, adding adrenaline if hypotension persists, whereas the authors often consider adrenaline first. The primer lists nitroprusside for normotensive shock, but NICE CG187 says do not offer sodium nitroprusside to people with acute heart failure (1.3.9).

UK practice: NICE CG187 (acute heart failure, 2014, last updated November 2021) says do not routinely offer inotropes or vasopressors in acute heart failure (1.3.10), but consider them in potentially reversible cardiogenic shock, given in a cardiac care unit, high dependency unit or other setting with at least level 2 care (1.3.11). It names no first-line agent. The primer’s norepinephrine and epinephrine are UK noradrenaline and adrenaline.

EM Evidence Rundown, Issue 36. Page 20 of 27

5 — QUICK HITS

Adaptive D-dimer for suspected PE (SR/MA, 6 studies, n=12,194; abstract only): In prospective management studies using age-adjusted or clinical-probability-adjusted cut-offs, 3-month failure after rule-out without CTPA was 0.12% (95% CI 0.06 to 0.26%), and 0.50% (95% CI 0.20 to 1.11%) for D-dimer above 500 ng/mL but below the adaptive cut-off. Efficiency varied widely (I² 99.3%); NICE NG158 rec 1.1.14 already says consider an age-adjusted threshold over 50 but does not mention probability-adjusted cut-offs. Graziani M, et al. Chest 2026. PMID: 42790624

Operator experience and intubation (secondary analysis of 8 US RCTs, n=2,839 ED and ICU intubations): comparing operators with 190 vs 15 previous intubations, the adjusted odds of first-pass success were higher (OR 1.75, 95% CI 1.30 to 2.36), and learning curves plateaued at 35–50 intubations; experience was self-reported and only 26.0% of intubations were by EM clinicians. Clark AT, et al. Ann Emerg Med 2026. PMID: 40848028

ECPR timing in refractory OHCA (abstract only; target trial emulation, Korean registry, n=58,668, 909 ECPR): weighted 30-day survival with ECPR vs continued CPR was 31.0% vs 16.0% at the 15-minute landmark (adjusted risk difference +15.0 percentage points, 95% CI 8.0 to 22.0) and 10.9% vs 3.2% at 60 minutes (+7.7, 95% CI 3.7 to 11.7); observational, so selection of ECPR candidates is likely to overstate the benefit. Sim JH, et al. Eur Heart J Acute Cardiovasc Care 2026. PMID: 42926135

Suxamethonium vs rocuronium for RSI with hyperkalaemia (retrospective, one US health system, n=434, first K over 5.5 mmol/L): 24-hour mortality was 10.5% (13/124) with suxamethonium vs 10.0% (31/310) with rocuronium (adjusted OR for paralytic choice 0.911, 95% CI 0.414 to 2.00), and cardiac arrest within 1 hour 0.8% vs 1.9%. Choice was not random: all 50 dialysis patients and most of those given calcium before RSI (29% vs 4%) received rocuronium. The UK SmPC still lists pre-existing hyperkalaemia as a contraindication to suxamethonium. Simmer PE, et al. Am J Emerg Med 2026;100:154–164. PMID: 41380422

Blood gas temperature entry (UK retrospective cohort, Nottingham, 6,767 COVID-19 admissions, 80,423 arterial gases; EMJ letter): of 2,031 patients with a bedside temperature of 39°C or more, 1,585 (78%) had an analyser temperature more than 1°C different; in febrile samples, pH was 7.35 or less in 31% of temperature-corrected vs 21% of uncorrected results (CI not reported). Febrile patients can look less acidotic near the NIV threshold; single trust, one analyser type, no patient outcomes. Crooks CJ, et al. Emerg Med J 2026;43(11):729–730. PMID: 42045008

Biphasic anaphylaxis (Derby, 8-year retrospective records review, n=324): 1 of 324 patients (0.3%, 95% CI 0.05% to 1.73%) had a true biphasic reaction within 7 days, and none were children; two other apparent cases had not fully settled (persisting tachycardia) before worsening. Coded single-centre data may miss re-presentations elsewhere, and 18 patients needing critical care were excluded. Tilbury N, et al. Emerg Med J 2026 (letter). PMID: 42754390

Troponin in ED paroxysmal SVT (5 branches of Taipei City Hospital, propensity-matched cohort, 360 matched encounters): a positive troponin was linked to admission (18.1% vs 3.4%; adjusted OR 7.95, 95% CI 2.81 to 22.54) but not clearly to 30-day MACE (7.4% vs 4.1%; adjusted OR 2.05, 95% CI 0.66 to 6.41). Matching on age and sex alone and only 18 events, so it does not show that withholding troponin is safe. Yu SH, et al. Emerg Med J 2026. PMID: 42901747

White cell count and cefepime vs piperacillin-tazobactam (post hoc analyses of the ACORN RCT and an instrumental-variable study; abstract only): white cell count modified the effect of antibiotic choice on 28-day mortality in both datasets (interaction OR 0.95, 95% CI 0.92 to 0.98, and 0.95, 95% CI 0.94 to 0.96); in ACORN, a baseline count of 16 or higher favoured piperacillin-tazobactam (OR 0.51, 95% CI 0.29 to 0.90). Post hoc and hypothesis-generating; no change to empirical choice. Rzewnicki DI, et al. Am J Respir Crit Care Med 2026. PMID: 42275170

Croup: dexamethasone 0.15 vs 0.6 mg/kg (retrospective cohort, n=14,205 children, 4 paediatric EDs; abstract only): ED return within 7 days was 2.9% with 0.15 mg/kg (n=172) vs 4.1% with 0.6 mg/kg (n=13,265) (difference 1.2%, 95% CI −1.31 to 3.76), but the low dose was associated with more hospitalisation (aOR 3.18, 95% CI 1.52 to 6.00). Only 172 episodes were treated with the low dose and allocation was not random, so the hospitalisation signal may reflect who was given it. Aregbesola A, et al. Pediatr Emerg Care 2026. PMID: 42899825

Intranasal ketamine for paediatric procedural sedation (BET, 8 papers from 150; abstract only): the authors conclude that intranasal ketamine has a slightly lower likelihood of success than IV ketamine but is a useful needle-

EM Evidence Rundown, Issue 36. Page 21 of 27

free alternative, to be considered with shared decision-making in severely needle-phobic children. The abstract gives no pooled success rates. Jeyabelen H, Manak I. Emerg Med J 2026. PMID: 42399089

Reducing imaging in children with limb injuries (SR/MA, 8 studies, 7,793 children; abstract only): ankle decision rules reduced radiography (OR 0.11, 95% CI 0.07 to 0.16; I² 38%); a wrist rule reduced imaging (OR 0.06, 95% CI 0.03 to 0.11) but missed eight injuries, and one trial of triage nurses using a rule increased imaging (OR 5.44, 95% CI 2.96 to 10.02). Most studies were before-after designs with incomplete reporting. Kirkland SW, et al. Emerg Med J 2026. PMID: 41443967

Distal thigh great saphenous vein as rescue access (prospective cross-sectional study, n=166, 22 children, single US centre): ultrasound visualised the vein in 95.8% to 98.2% of thigh regions (median depth 9.9 to 11.7 mm, diameter 2.8 to 3.0 mm) and in at least one region in all 22 children. Cannulation success was not studied, and only 22 participants were children. Saadi Neto E, et al. J Emerg Med 2026. PMID: 42391823

Naloxone for nitazene overdose (EMJ Best Evidence Topic, 5 papers; abstract only): Across the five included papers, 0.4 to 4.4 mg of naloxone was required to treat nitazene overdoses, which the authors say does not support using higher doses. The abstract does not give the study designs or patient numbers, and dose frequency was not explored. Blanksby I, Harrison L. Emerg Med J 2026. PMID: 42392847

IV ibuprofen vs ketorolac for acute low back pain (triple-blind RCT, n=73, three university hospital EDs; abstract only): Ibuprofen 800 mg IV gave lower VAS pain scores than ketorolac 30 mg IV at 30 minutes (32.2 vs 40.7; p=0.038), and more patients reached mild pain (70.6% vs 35.9%), with similar rescue analgesia and adverse events. The abstract reports no confidence intervals and the trial is small. Jafarkhani M, et al. Intern Emerg Med 2026. PMID: 42902081

EM Evidence Rundown, Issue 36. Page 22 of 27

CORE REVISION — ISSUE 36 — 29 OCTOBER 2026

Core Revision — ECG: Broad Complex Tachycardia — VT vs SVT with Aberrancy

Why to assume VT, the ECG features, the Brugada and Vereckei algorithms, pre-excited AF and RCUK 2025 management — FRCEM

Exam goal: Spot the ECG features that make a regular broad complex tachycardia (BCT, QRS 0.12 s or more) ventricular tachycardia (VT), explain why no algorithm can rule VT out, and manage stable and unstable BCT to the Resuscitation Council UK (RCUK) 2025 algorithm with UK doses.

1. WHY ASSUME VT

2. ECG FEATURES FAVOURING VT

EM Evidence Rundown, Issue 36. Page 23 of 27

Morphology

Not typical RBBB or LBBB. RBBB-like: taller left rabbit ear in V1 (the most specific VT finding), monophasic R or qR in V1. LBBB-like: initial R over 30–40 ms in V1–2, notched S (Josephson sign), any Q wave in V6

Typical RSR’ with a taller right ear suggests SVT with RBBB; SVT with LBBB has no Q wave in V6

3. THE BRUGADA AND VERECKEI ALGORITHMS

Memory aid — “VT is A, B, C, D, E, F”: AV dissociation; Brugada sign (RS over 100 ms); Capture and fusion beats, and Concordance in V1–V6; Duration over 160 ms; Extreme (northwest) axis; Factors in the history (age over 35, structural heart disease, previous MI). None present does not exclude VT: if in doubt, treat as VT, and never give verapamil for a BCT.

4. MANAGEMENT: RCUK 2025 TACHYARRHYTHMIA ALGORITHM (UPDATED MARCH 2026)

EM Evidence Rundown, Issue 36. Page 24 of 27

FRCEM focus: list the features of VT; recite Brugada step by step; explain why verapamil is dangerous (Stewart: deterioration in all 13 misdiagnosed VT episodes given verapamil); know that the stable BCT amiodarone dose is 300 mg over 10–60 min, and 10–20 min after failed shocks in the unstable patient; recognise pre-excited AF and avoid AV nodal blockers. Sources: RCUK Adult ALS guidelines 2025 and Adult tachyarrhythmia algorithm V3 (March 2026), resus.org.uk; LITFL ECG Library (VT versus SVT; Monomorphic VT; AF/flutter in pre-excitation); Brugada et al, Circulation 1991; Vereckei et al, Heart Rhythm 2008; Jastrzebski et al, Europace 2012; Stewart et al, Ann Intern Med 1986; amiodarone, adenosine and verapamil injection SmPCs (eMC).

EM Evidence Rundown, Issue 36. Page 25 of 27

6 — ACTION POINTS

EM Evidence Rundown, Issue 36. Page 26 of 27

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 expected 21 January 2027.

2026–2027

LONGER HORIZON

PENG vs fascia iliaca block RCEM lists a head-to-head comparison as a research recommendation; this week’s trial 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 36. Page 27 of 27