Zworth Reading
EMMax’s EM Weekly Update
Highlight of the Week
YEARS Algorithm for Diagnosis of Suspected Pulmonary Embolism in Patients With Cancer (Hydra Trial)
JAMA · Multicentre open-label RCT, blinded central adjudication
The Hydra trial included 698 adults with active cancer and clinical suspicion of acute PE across 21 hospitals in 6 European countries (76% enrolled from the ED). Key exclusion: life expectancy <3 months. Median age 65, 60% female, 48% metastatic disease. Patients were randomized to the YEARS algorithm (n=352) or CTPA-only (n=346). YEARS uses 3 clinical items (signs of DVT, hemoptysis, PE most likely diagnosis) with a risk-dependent D-dimer threshold: 0 items → cutoff <1000 ng/mL rules out; ≥1 items → cutoff <500 ng/mL rules out. In the per-protocol analysis, VTE or PE-related death within 90 days occurred in 5/282 (1.8%) YEARS patients vs 15/273 (5.5%) CTPA-only patients — absolute risk difference -3.7% (99.9% CI -8.8% to 1.4%), meeting noninferiority (margin 2.6%). Intention-to-diagnose analysis gave a difference of -2.6% (99.9% CI -7.5% to 2.4%), also meeting noninferiority. CTPA was avoided in 77/352 (22%) of YEARS patients.
This is a well-conducted noninferiority trial that shows the YEARS algorithm is safe in this population. Some things I liked: (1) the inclusion criteria are broad and pragmatic — any active cancer other than basal/squamous skin, with clinical suspicion of PE — which is close to how we actually think in the ED; (2) the exclusion of patients with estimated life expectancy <3 months is important context and appropriate; (3) follow-up was thorough (90-day phone call + GP/oncologist contact + chart review, only 1 patient lost); (4) randomization was appropriate and outcome adjudication was centrally blinded; (5) noninferiority was met in BOTH per-protocol and intention-to-diagnose analyses. A few honest observations to sit with: 'safe' doesn't mean missing zero PEs. In the YEARS group 5 patients (1.8%) had a diagnosed VTE during the 3-month follow-up (2 of whom actually did get a CT in the ED). But 1.8% is well within the conventional <2% threshold we consider acceptable for a PE-exclusion strategy, and the noninferiority margin still held. The open-label design is a limitation but unavoidable. The trial was stopped at interim after noninferiority was declared, which is appropriate but leaves the trial smaller than originally planned but also potentially a limitation. Population is European; generalizability to Canadian academic EDs seems reasonable given similar cancer demographics and imaging access.
Bottom line: Important study to know about. This one also fits squarely in the Practice-Changing EM category. It addresses a common ED question (workup of suspected PE in the cancer patient) with a directly actionable answer. You can safely use the YEARS algorithm in patients with active cancer to rule out PE without CTPA, sparing about 1 in 5 patients from a scan in a population that is already exposed to a lot of imaging. This has practical implications for both patient experience (contrast, radiation, IV access) and department flow. Curious to hear criticisms and looking forward to further analyses, but on face value this is a genuine step forward for a common ED question.
FOAM Radar
EM@3AM: Cortical Hand-Knob Stroke Mimicking Peripheral Nerve InjuryCase-based education
emDocs · LOE 5 — case report / teaching point
Case of isolated hand weakness from a cortical hand-knob infarct presenting without classic stroke features. The motor homunculus representation of the hand is a discrete cortical region; small infarcts here can mimic peripheral nerve or radicular pathology.
Bottom line: Good reminder that isolated hand weakness without sensory loss can be central. Low threshold for stroke workup in sudden-onset isolated motor deficits, especially in older patients with vascular risk factors.
Guidelines Update
Society of Obstetricians and Gynaecologists of Canada (SOGC)
- Screen for iron deficiency (not just anemia) using ferritin. Hemoglobin alone under-detects the problem, particularly in menstruating and pregnant patients
- Oral iron remains first-line for most patients; consider alternate-day dosing to improve absorption and tolerability
- IV iron is appropriate for oral intolerance, inadequate response, malabsorption, severe anemia in late pregnancy, or need for rapid replacement
- Recognize that iron deficiency without anemia is clinically significant. Fatigue, exercise intolerance, cognitive symptoms, and restless legs may all be present at normal Hb
- Uses GRADE methodology with multidisciplinary input (obstetrics, gynaecology, anaesthesiology, hematology, patient partners)
GRADE methodology, systematic literature review 2015-2025. Recommendation strength varies by topic; the general framework of ferritin-based screening and early oral iron is well-supported by RCT and observational evidence.
Conflicts with existing guidance: Aligns broadly with recent international guidance (BSH, ACOG, Iron Consortium 2025) — Canadian framework may differ modestly in thresholds and specific IV iron formulation preferences.
Bottom line: Not immediately practice-changing for the ED, but useful reference for the many pregnant and postpartum patients we see with fatigue, dizziness, or dyspnea whose anemia workup starts in the ED. Two practical takeaways: (1) send ferritin, not just Hb, as iron deficiency without anemia is a real entity and worth flagging in your discharge letter; (2) it is reasonable to start oral iron from the ED with PCP or obstetrics follow-up, and IV iron is an appropriate consideration in late pregnancy with symptomatic anemia.
Adjacent Specialties
Cardiac Findings in Swimming-Induced Pulmonary Edema — Implications for Acute AssessmentLOE 3 (Prospective observational cohort)
Chest · Pulmonary / Sports Medicine
In patients with swimming-induced pulmonary edema (SIPE), what cardiac abnormalities are present acutely and do they resolve?
45 SIPE patients assessed within 2 hours of the event, matched to 45 controls from the same swimming events. 43% of SIPE patients had mildly impaired systolic function on echo (vs 10% controls). 30% had elevated PA pressures (vs 0% controls). Troponin was elevated in 67% of SIPE patients (median 47 pg/mL). All echo abnormalities resolved at follow-up except one patient. Two patients had concurrent MI with significant chest pain and markedly high troponin.
Clever study design. Prospective assessment within 2 hours with matched controls from the same swimming event over a 3 year period, which is unusually clean for this kind of question. Small sample but adequate to characterize the phenomenon. The key clinical insight is establishing a reference range: mild troponin elevation and transient systolic dysfunction are expected in SIPE and don't necessarily indicate ACS. The two concurrent MI cases had clinically obvious presentations.
Bottom line: SIPE is uncommon, but triathlon season is in full swing and every now and then someone will show up with respiratory symptoms that started during a swim. Worth being aware of the entity. When you see it: expect mild troponin elevation and transient echo abnormalities — that's the disease, not necessarily ACS.
Major Journals Scan
Rivaroxaban Then Aspirin vs. Aspirin Alone after Total Hip or Knee Arthroplasty (EPCAT III)LOE 2 (Multicentre double-blind RCT)
New England Journal of Medicine
Why it matters to EPs: We're not initiating anticoagulation in these patients, but this literature matters for us in two ways. First, we regularly see post-op arthroplasty patients presenting with symptoms or complications, and it is worth knowing what prophylaxis they may be on. Second, there's potential applicability to other populations where we do have prevention decisions to make (or advise on) — the flying-with-a-fracture patient, the patient with multiple VTE risk factors sustaining a leg injury with immobilization, and so on.
5,429 patients randomized after hip or knee arthroplasty. Rivaroxaban 10 mg daily for 5 days followed by aspirin 81 mg vs aspirin 81 mg alone from post-op day 1. Primary outcome (symptomatic VTE within 90 days): 0.45% rivaroxaban+aspirin vs 0.48% aspirin alone — noninferior, essentially identical. Bleeding events similar (1.66% vs 2.04%). Canadian trial (EPCAT III), fully blinded, adequate power.
VTE rates in both arms are reassuringly low, which is probably a testament to modern surgical technique and early mobilization more than either drug. Aspirin-only prophylaxis is now robustly evidence-based for this population. Applicability beyond arthroplasty is uncertain (the trial doesn't tell us anything about DOAC vs aspirin in other prophylaxis settings), but the study fits into a broader trend of aspirin performing as well as heavier anticoagulation for VTE prevention in ambulatory-recovery populations.
Bottom line: Not practice-changing for the ED, but informative. Post-arthroplasty patients on aspirin-only prophylaxis are appropriately managed, not undertreated. Broader implication (not proven by this trial): in populations where we counsel or advise on VTE risk — cast immobilization + long flight, multiple risk factors + minor lower-limb injury — aspirin may be a reasonable option in the risk-benefit conversation, though this hasn't been directly tested.
Methodology Flag
Annals of Emergency Medicine
Very large dataset (6.7 million patients) with the headline finding that longer ED stays are associated with higher 30-day mortality. Guaranteed to be cited in policy conversations about ED crowding and boarding.
- The primary comparison is patients staying >4 hours vs <4 hours in the ED. It's not clear how the 4-hour cutoff was chosen — it feels arbitrary and doesn't capture the fact that ED length of stay is a continuous variable, not something with a magic threshold. To their credit, the authors also compare 3 vs 2 hours, 4 vs 2, and so on, which somewhat mitigates this concern.
- There is enormous confounding potential. This isn't necessarily classical confounding-by-indication (sicker patients staying longer, though that also exists). >4-hour patients tended to arrive at particular hours of the day, be located in specific parts of the country, and were less likely to arrive by ambulance. All of these variables plausibly relate to both LOS AND mortality through pathways the model can't fully capture.
- Retrospective observational design cannot establish causation despite the headline framing.
- Health-systems studies conducted in other countries are inherently difficult to extrapolate from. The structure of EDs, prehospital systems, admission thresholds, and downstream care in England is quite different from Canadian practice. Findings may not translate.
- Excluded 'immediate acuity' patients, which may paradoxically enrich for patients whose severity wasn't fully recognized in triage. another confounding pathway.
What it does contribute: Confirms at population scale what most of us already believe: prolonged ED stays are a marker of system strain and probably do worsen outcomes for at least some patients. The data are useful for health services research and policy conversations about resource allocation, staffing, and access-block. The signal is broadly consistent across the various time-threshold sensitivity analyses the authors performed, which is a strength.
Bottom line: Correlation, not causation. BUT - it's completely plausible that ED length of stay is associated with increased mortality for a whole host of reasons (missed monitoring, delayed diagnostics, medication timing, delirium, staff cognitive load), and measures to reduce ED length of stay should absolutely be pursued regardless of what this specific study proves. Just be careful when this gets cited to support administrative targets, because the causal claim is stronger than the evidence.