Mean flow index (Mxa) algorithm option can influence cerebral autoregulation classification in critically ill septic...

The Mean flow index (Mxa) is widely used to assess dynamic cerebral autoregulation in different clinical populations. This calculation is based on defined characteristics, including blocks, overlap periods, and epochs of the whole recordings. This study aimed to investigate the reproducibility of different Mxa calculations, using variable blocks, overlap periods, and epochs. We retrospectively analyzed 50 transcranial Doppler recordings from septic shock patients, acquired within 48 h of ICU admission. Mxa was computed using eight signal-processing strategies that varied by block duration (5-10 s), overlap percentage (20%, 50%, 80%), and epoch length (3-5 min), as well as a continuous approach without epochs. Each configuration was labeled using the format epoch-block-overlap. Mxa values were compared using repeated measures analyses, intraclass correlation coefficients (ICC), Bland-Altman plots, and polychoric correlation heatmaps. Median Mxa values ranged from 0.36 to 0.45 across configurations, with no statistically significant differences in within-patient comparisons (p > 0.05). ICCs demonstrated excellent agreement (ICC > 0.90) between approaches using the same epoch duration. Agreement declined modestly when comparing configurations with different epoch lengths (e.g., ICC = 0.782 between 3-10-50 and 5-10-50). Fixed-effects analysis did not identify any individual segmentation parameter as a significant source of variability. Mxa values calculated using different combinations of block, overlap, and epoch duration were consistent within patients, particularly when epoch length was maintained. These findings support the reproducibility of Mxa and suggest flexibility in processing strategies, provided methodological consistency is maintained. Further validation is warranted.