End-tidal CO₂ changes during short breath-hold maneuvers: Variability, paradigm effects, and spirometric correlates

Breath-hold functional MRI (bh-fMRI) is a widely used method for mapping cerebrovascular reactivity (CVR). Normalization of BOLD signal changes to end-tidal CO₂ changes (Δp_etCO₂) is suggested to improve between-subject comparability. The aim was to characterize magnitude and variability of Δp_etCO₂ during short breath-hold maneuvers. Twenty healthy volunteers participated in this study. Ten of these participants performed five breathing paradigms: baseline breathing, and end-expiratory breath-holds of 6, 9, and 12s, as well as a deep breathing maneuver, each repeated seven times. p_etCO₂ was recorded continuously via nasal cannula. Within- and between-subject variability were quantified using the coefficient of variation (CV). Differences in Δp_etCO₂ between paradigms were assessed using repeated-measures ANOVA. Ten additional participants underwent the 9s breath-hold paradigm and spirometry; associations between Δp_etCO₂ and pulmonary volumes, age, and baseline p_etCO₂ were assessed using Pearson correlation. Within-subject within-paradigm CVs were low (3.4-4.7%), indicating high reproducibility. Between-subject CVs were substantially higher (19.7-35.3%). Δp_etCO₂ differed significantly between the 6 and 9s paradigms (p = 0.014, p_FDR = 0.017) but not between the 9 and 12s paradigms (p = 0.209, p_FDR = 0.209). The deep breathing induced a reproducible p_etCO₂ decrease. A very strong inverse correlation was observed between Δp_etCO₂ and inspiratory vital capacity (IVC) (r = -0.797 (95%CI: -0.950--0.336), p = 0.006, p_FDR = 0.030). Short breath-hold maneuvers induce reproducible, yet variable CO₂ responses within individuals. Pulmonary anatomy is a key determinant of Δp_etCO₂. Inadvertent deep breathing prior to breath-holding may induce p_etCO₂ decreases of non-negligible magnitude, with potential implications for bh-fMRI interpretation.