Effect of Permissive Mild Hypercapnia on Cerebral Vasoreactivity in Infants: A Randomized Controlled Crossover Trial.

Background: Mechanical ventilation interferes with cerebral perfusion via changes in intrathoracic pressure and/or as a consequence of alterations in CO2. Cerebral vascular vasoreactivity is dependent on CO2, and hypocapnia can potentially lead to vasoconstriction and subsequent decrease in cerebral blood flow. Thus, we aimed at characterizing whether protective ventilation with mild permissive hypercapnia improves cerebral perfusion in infants.

Methods: Following ethical approval and parental consent, 19 infants were included in this crossover study and randomly assigned to 2 groups for which the initial ventilation parameters were set to achieve an end-tidal carbon dioxide (Etco2) of 6.5 kPa (group H: mild hypercapnia, n = 8) or 5.5 kPa (group N: normocapnia, n = 11). The threshold was then reversed before going back to the initial set value of normo- or hypercapnia. At each step, hemodynamic, respiratory, and near-infrared spectroscopy (NIRS)-derived parameters, including tissue oxygenation index (TOI) and tissue hemoglobin index (THI), concentration of deoxygenated hemoglobin (HHb) and oxygenated hemoglobin (O2Hb), were collected. Concomitantly, sevoflurane maintenance concentration, ventilatory (driving pressure) and hemodynamic parameters, as mean arterial pressure (MAP), were recorded.

Results: Targeting an Etco2 of 5.5 kPa resulted in significantly higher mean driving pressure than an Etco2 of 6.5 kPa (P < .01) with no difference between the groups in end-tidal sevoflurane, MAP, and heart rate. A large scatter was observed in NIRS-derived parameters, with no evidence for difference in Etco2 changes between or within groups. A mild decrease with time was observed in THI and MAP in infants randomly assigned to group N (P < .036 and P < .017, respectively). When pooling all groups together, a significant correlation was found between the changes in MAP and TOI (r = 0.481, P < .001).

Conclusions: Allowing permissive mild hypercapnia during mechanical ventilation of infants led to lower driving pressure and comparable hemodynamic, respiratory, and cerebral oxygenation parameters than during normocapnia. Whereas a large scatter in NIRS-derived parameters was observed at all levels of Etco2, the correlation between TOI and MAP suggests that arterial pressure is an important component of cerebral oxygenation at mild hypercapnia.