Chapter 14 — Initial VA ECMO Management

BEST-EVIDENCED CHAPTER IN PART III — built around BLENDER (Burrell, Intensive Care Medicine 2024;50(9):1470-1483), retrieved IN FULL from the publisher rather than a secondary summary, and the first randomised trial of how to SET VA ECMO rather than whether to use it. Conservative (SaO2 92-96%) vs liberal (97-100%): no difference in ICU-free days to day 28 (median treatment effect 0, 95% CI -3.1 to 3.1), day 28 mortality 39.6% vs 39.1%, day 60 43% vs 41.1%, ECMO/ventilation duration, LOS or 6-month function; subgroups consistent. The most useful bedside number is the 29.5% vs 1.3% major protocol deviation rate — the conservative target is genuinely hard to deliver, which both weakens the null and argues against enforcing it. A DATED ADDENDUM WAS APPENDED TO CHAPTER 13 recording that its 'no randomised evidence for any VA target' statement is now partly wrong. Also owns: the initiation sequence (pump to 500 RPM first, then drainage clamp, then return clamp — a stopped centrifugal pump on an unclamped VA circuit shunts backwards); ACT above 250 s; sweep 5-6 L/min at 1:1; cardiac index 2.0-2.4 target and why hypotension during first titration is preload not pump failure (5-10 mL/kg aliquots as a DIAGNOSTIC); avoid hypocarbia; the immediate post-cannulation table adapted from Red Book Table 32-3 (ECPR-specific content left to Ch 19-20); 'rapidly down-titrate pressors'; 60-80 mmHg MAP with no demonstrated optimum; the ISCCM initiation safety checklist and the pre-initiation neurological baseline that cannot be obtained later; THE 4-HOUR DISTAL PERFUSION DEADLINE; six things to settle before the team disperses. Controversies: oxygen target after BLENDER; cardiac index vs pulse contour as the flow target (resolved as a target and a constraint, not competitors). GAP IDENTIFIED: the SAVE score appears as a BLENDER subgroup variable and is not covered anywhere in this book — close in Ch 11 or Ch 30.