Determinants of Baseline Membrane Lung Resistance in Extracorporeal Membrane Oxygenation.
BACKGROUND Membrane lung (ML) resistance is commonly monitored during extracorporeal membrane oxygenation (ECMO) to forewarn of circuit obstruction, but differences in device design and patient hemorheological factors may complicate interpretation. We sought to determine whether a stable baseline resistance exists shortly after ECMO initiation and to identify factors associated with baseline variability. METHODS We conducted a single-center retrospective cohort study of adult patients receiving veno-arterial or veno-venous ECMO between 2011 and 2023. Baseline resistance was calculated as the mean inlet-outlet pressure difference normalized to flow during the first 2 and 3 h after initiation. Descriptive statistics and linear regression evaluated associations between patient and device characteristics and baseline resistance. RESULTS Among 170 patients, 33 (19.4%) used the Sorin EOS ML and 135 (79.4%) used the Quadrox-i Small Adult ML, and 2 (1.2%) used other ML models that were excluded from model-specific analysis. Resistance stabilized within 2-3 h post-initiation. Baseline resistance varied by ML model: median 31.7 mmHg·min/L for Sorin versus 5.5 mmHg·min/L for Quadrox (p < 0.0001). Each 1 unit increase in hematocrit increased baseline resistance by 0.54 (95% CI: 0.13 to 0.95) for Sorin, and by 0.05 (95% CI: 0.02 to 0.09) for Quadrox. Female sex was associated with lower resistance compared to male sex (Sorin: -6.92 mmHg·min/L, 95% CI -11.28 to -2.56; Quadrox: -0.63 mmHg·min/L, 95% CI -1.06 to -0.20). Additional associations varied by device model. CONCLUSIONS Baseline ML resistance reflects device design and patient hemorheology. Early resistance provides a stable baseline reference, though patient hemorheology may confound resistance deviations attributed to true ML obstruction. Device-specific and physiologically adjusted interpretation may personalize detection of ML obstruction.