A curated cross-year highlights reel — the trials from 2024–2026 most likely to change how you practice, organized by theme rather than year. For full 11-part summaries, follow the links through to Year Book → [Year] → [Category].
🟢 Genuinely Practice-Changing Positives
Trial | Year | Why It Matters |
2024 | First mechanical circulatory support device ever to show a mortality benefit in cardiogenic shock (45.8% vs 58.5%), after decades of neutral IABP/device trials. | |
2024 | First randomized trial to meet its primary endpoint for surgical ICH evacuation, after STITCH, STITCH II, and MISTIE III all failed. Benefit concentrated in lobar hemorrhages. | |
2025 | Definitively validates capillary-refill-time-guided resuscitation — a simple, free, bedside target — as superior to usual care in septic shock. | |
2025 | First trial to show biomarker-guided, phenotype-matched immunotherapy improves sepsis organ dysfunction — overturning the assumption that all patients follow one hyperinflammation-to-immunoparalysis trajectory. | |
2024 | Simple qSOFA-based electronic sepsis screening cut 90-day mortality at scale (NNS~206) — suggests automated early recognition beats downstream bundle compliance. | |
2026 | Meaningfully expand thrombectomy/thrombolysis eligibility — basilar occlusion out to 24h, and medium-vessel-occlusion strokes previously outside standard criteria. |
🔴 Important Harm Signals — Reversed or Refined Practice
Trial | Year | Why It Matters |
2024 | High-protein enteral feeding (2.0 vs 1.3 g/kg/day) caused WORSE quality of life — directly reversing decades of nutrition-guideline dogma. Joined by TARGET Protein (2025) and a mechanistic trial (2026) showing higher protein doses don't even boost muscle synthesis. | |
2025 | Stopped early for harm: targeting high MAP (80-85) in older septic shock significantly increased mortality — refuting the SEPSISPAM hypothesis that hypertensive patients need higher pressure targets. | |
2025 | Inhaled sevoflurane sedation in ARDS caused worse mortality, ventilator-free days, and AKI vs propofol — a genuine caution against a growing sedation trend. | |
2026 | Elective LV unloading (Impella) in high-risk, non-shock PCI showed excess cardiovascular mortality (HR 1.91) despite widespread and increasing real-world adoption. | |
2024 | HFNO was noninferior to NIV in 4 of 5 respiratory-failure groups — but WORSE specifically in immunocompromised patients with hypoxemia. A critical subgroup-specific caution. | |
2026 | rFVIIa for ICH reduced hematoma growth but gave no functional benefit and tripled thromboembolic events — even in the population most likely to benefit. |
🔄 Practice-Challenging Equipoise
Trial | Year | Why It Matters |
2025 | Deferring routine arterial catheterization in favor of noninvasive BP monitoring was noninferior — challenging one of ICU's most entrenched invasive-monitoring habits. | |
2024 | Together mark a real shift toward shorter, biomarker- or protocol-guided antibiotic courses — PCT-guided stopping and 7-day (vs 14-day) treatment both proved safe. | |
2024 | A fascinating contrast: liberal transfusion significantly helped in broad acute brain injury (TRAIN) but showed no benefit specifically in aneurysmal SAH (SAHaRA) — transfusion strategy may need to be brain-injury-subtype-specific. | |
2026 | Two simple, deployable interventions (plasma exchange in liver failure; EDTA catheter locks) with real, meaningful effect sizes — unglamorous but genuinely useful additions to practice. |
How to use this page
This is a starting point, not a replacement for the full handbook — each trial above deserves its complete 11-part read (Internal/External Validity, Controversies, etc.) before changing practice. Use the Master Trial Index database to filter by category, year, or verdict for anything not listed here.