Ch 28 (s28.8, Controversy 1); Ch 31; Part VII
A microembolic signal count on ECMO is a research measurement, not a clinical one. It has no established action threshold, did not predict brain injury in the largest cohort, and does not track circuit clot, anticoagulation intensity or flow. Do NOT alter anticoagulation on the basis of a signal count. The one replicated finding is that signals are several times more common on venoarterial than venovenous support.
Acute brain injury and a composite of arterial thromboembolic events; also cannulation mode, circuit clot or fibrin, and haemodynamic, laboratory and ECMO parameters
DOI 10.1016/j.xjtc.2022.07.026
Presence and number of microembolic signals on transcranial Doppler
Single centre; intermittent rather than continuous monitoring; signals analysed as a per-patient dichotomy rather than a time-varying exposure. Directly contradicts an earlier, smaller study from the SAME institution (Cho 2020, ASAIO J) in which all 10 patients with arterial-sided thrombotic events had positive signals.
Signals in 35.0 percent of venoarterial and 4.7 percent of venovenous patients (p under 0.001). Signal presence was NOT associated with cannulation mode (23.4 percent peripheral versus 25.8 percent central), NOT associated with clot or fibrin in the circuit, and NOT associated with any studied haemodynamic, laboratory or ECMO parameter. Acute brain injury occurred in 38 percent of venoarterial and 31.1 percent of venovenous patients; on multivariable analysis neither acute brain injury nor the arterial thromboembolic composite was associated with the presence or number of signals.
145 adult ECMO patients undergoing transcranial Doppler with microembolic signal monitoring at a single tertiary centre, November 2017 to February 2022: 100 venoarterial (187 examinations) and 45 venovenous (65 examinations)